north pacific research board bering sea integrated ... · marine population assessment and...

NPRB BSIERP B.70 FEAST Forage/Euphausiid Abundance in Space and Time

NORTH PACIFIC RESEARCH BOARD

BERING SEA INTEGRATED ECOSYSTEM RESEARCH PROGRAM

FINAL REPORT

Forage Euphausiid Abundance in Space and Time (FEAST)

NPRB BSIERP Project B70 Final Report

Ivonne Ortiz1, Kerim Aydin2, André Punt3

1 Joint Institute for the Study of the Oceans and Atmosphere, College of the Environment, University of Washington, 7600 Sand Point Way NE, Seattle, WA 98115, [email protected] or [email protected]. 2 National Oceanic & Atmospheric Administration, National Marine Fisheries Service, Alaska Fisheries Science Center, Resource Ecology and Fisheries Management Division, 7600 Sand Point Way NE, Seattle, WA 98115. [email protected] 3 Marine Population Assessment and Population Dynamics Group, School of Aquatic and Fishery Sciences, College of the Environment, University of Washington, 1122 NE Boat St., Box 355020, Seattle, WA 98195 [email protected]

June 2014, revised March 2015

1


Abstract

The aim of the project was to develop the Forage and Euphausiid Abundance in Space and Time model (FEAST) as a spatially explicit, length-based bioenergetics model for multiple fish species and couple it to other models being developed for oceanography (Northeast Pacific version 5 Regional Ocean Modeling System, NEP5 ROMS), lower trophic levels (Nutrient-Phytoplankton-Zooplankton-Detritus, NPZD) and fishing effort (Fishing effort Allocation Model In Nash Equilibrium, FAMINE). The NEP5 ROMS-NPZD-FEAST-FAMINE model was to be used to run a 1970-2005 hindcast, three 50-year forecasts using input from three different climate models, and serve as the operating model for management strategy evaluation. The aim of this integrated ecosystem model was to serve both as a tool to test hypothesis of how the system worked as well as a management tool. The first required a higher level of detail in the vertical component of the three dimensional structure of the model (with longer model run times), and the second required no vertical component and short run times. The dependencies across models highlighted the need to ensure minimum performance levels in a sequential manner starting with the oceanography, then the lower trophic levels, fish and finally fishing effort. This need for minimum performance levels caused a cumulative lag in model progress and left no time to simulate the forecasts with FEAST and conduct a management strategy evaluation with FEAST as the operating model. The project developed the FEAST model such that it is independent of the complexity of the underlying three dimensional structure. It can be used within an integrated ecosystem model based on either a simplified oceanography model that has a smaller geographical extent and only 10 vertical layers (Bering 10K ROMS-NPZD), or the original NEP5 ROMS-NPZD with 60 vertical layers; feedback with the NPZD can be one-way or two-way. Even though the FEAST model was not used as the operating model for the management strategy, it was designed to satisfy all the data format and requirements of both the FAMINE and the MSE. The code and scripts in place could be used in any follow-up modeling work. The hindcast was performed using the BERING 10K ROMZ-NPZD-FEAST-FAMINE model and was extended to include the BSIERP field years, it covers the period 1970-2009. Proxies of future climate conditions for the MSE were provided by selecting ecosystem/ production indices from Bering 10K ROMS-NPZD forecasts based on three different climate models to use in a multi-species model (MSMt). Results from the hindcast and multi-year simulations show the use of dynamic prey fields allow a more realistic bioenergetics model which reformulates activity costs based on swim speed and fish length, as opposed to the common use of fixed activity costs. Fish movement patterns were also more realistic when movement rules were based on prey availability (measured as individual weight gain) and predator avoidance (measured as predation mortality) as opposed to prey availability alone. Results from the hindcast simulation also highlight the potential role of the fall bloom, and underscores the importance of developing pilot studies for late fall and winter, a period for which there is no data available. The project demonstrated the feasibility of developing a high spatial resolution model that links climate, oceanography and lower trophic levels to fish and fisheries that is useful for both hypothesis testing and management strategy evaluation. Other key conclusions pertain to the lessons learned in developing/coordinating a complex integrated ecosystem model, integrating field data into the model development process and the use of an integrated ecosystem model framework to help focus, guide and prioritize current and new lines of research.

Key Words: Arrowtooth flounder, bioenergetics, climate change, eastern Bering Sea, end-to-end-models, fish movement integrated ecosystem models, management strategy evaluation, Pacific cod, simulation, walleye pollock.

Recommended Citation: Ortiz, I., K. Aydin, A. Punt. Forage Euphausiid in Space and Time (FEAST). North Pacific Research Board BSIERP Project B70 Final Report, 175 p

2


Contents Abstract ............................................................................................................................2 Study Chronology ............................................................................................................4 Chapter 1. Introduction ....................................................................................................5 Chapter 2. Objectives .......................................................................................................7 Chapter 3. Modeling fish bioenergetics in the eastern Bering Sea (placeholder) ............11 Chapter 4. Modeling fish movement, a case study of walleye pollock, Pacific cod and

arrowtooth flounder in the eastern Bering Sea (to be updated) ...........................16 Chapter 5. Climate to fisheries: Exploring processes in the eastern Bering Sea based on a

40 year hindcast ...................................................................................................22 Chapter 6. Use of FEAST as an operating model for Management Strategy Evaluation 49 Chapter 7. Blended Forecasts and Short-Term Forecasts ................................................73 Chapter 8. Setting up a fully integrated ecosystem model (Climate to fisheries) designed

for MSE and hypothesis testing ...........................................................................99 Chapter 9. End-to-End Modeling as part of an Integrated Research Program in the Bering

Sea ........................................................................................................................126 Chapter 10 Conclusions ...................................................................................................148 Collaborative processes and best practices ......................................................................150 Next steps and future work ..............................................................................................152 Guidance for field work ...................................................................................................152 Ongoing and future modeling work .................................................................................153 Chapter 11. BSIERP and Bering Sea Project connections ..............................................155 Shortcomings ...................................................................................................................156 Meetings with other projects ............................................................................................156 Projects/ activities stemming from BSIERP Program .....................................................157 Management or policy implications.................................................................................158 Publications ......................................................................................................................158 Poster and oral presentations at scientific conferences or seminars ................................160 Outreach/workshops ........................................................................................................163 Acknowledgements ..........................................................................................................164 Literature cited: ................................................................................................................165 Appendix A ......................................................................................................................170

3


Study Chronology This was a new project and was the first NPRB-funded project for PI André Punt.The project is part of the modeling effort of BSIERP. The award for the project was made to the University of Washington. It began on October, 1 2007 and ended February 28, 2014. This project was closely related to two other BSIERP modelling projects (B.71 and B.73) as well as one project from NSF, described below: NSF Award ID 0732534 Downscaling Global Climate Projections to the Ecosystems of the Bering Sea with Nested Biophysical Models, PI Nicholas Bond, JISAO, UW. This project had 4 components, all pertinent to FEAST: i) climate forcing (forcing files from three selected climate model projections with models based on their ability to capture key features of the Bering Sea shelf dynamics –seasonal ice and Pacific Decadal oscillation), ii) numerical ocean modeling (Regional Ocean System Model, ROMS for the North East Pacific), iii) lower-trophic modeling (Nutrient-Phytoplankton-Zooplankton NPZ), and iv) linking ROMS and NPZ to FEAST. B.72 Economic-Ecological Models of Pollock and Cod, PI Michael Dalton, AFSC, NOAA and André Punt, UW.This project provided the spatially explicit catch data by fisheries for walleye pollock, Pacific cod and Arrowtooth flounder used in the 1971-2009 hindcast. This project also provided the FAMINE model which implemented the fleet dynamics model that linked output from the assessments and harvesdt control rules with FEAST. B.73 Management Strategy Evaluation, PI André Punt, UW and Jim Ianelli, AFSC, NOAA. This project implemented the use of FEAST as the operating model for MSE and currently developed methods (stock assessments, MSMt and Ecosim) as “assessment models”. . The Forage and Euphausiid Abundance in Space and Time (FEAST) model is a spatially explicit multi-species fish bioenergetics model. As the upper level trophic model linking lower trophic levels and fishing effort, the coupling and coordination of the oceanography, lower trophic levels, fish, fishing effort and management strategy evaluation into one integrated ecosystem model was de facto under this project. This report uses updates from semi-annual progress reports for project B.70 submitted in October 2008 and every April and October from 2009 to 2013, covering reporting periods from October 1 to March 31, and from April 1 to September 30, respectively. This report also includes material from the final project reports and/or papers published as part of projects B72 Economics, B73 Management Strategy Evaluation, and NSF ID 0732534 Downscaling climate models.

4


Chapter 1. Introduction Both climate variability and ecosystem based approaches to management have been recognized as key aspects for sustainable fisheries management. With increased computer power the ability to develop modeling frameworks that encompass both aspects has increased, and thus models such as Ecopath with Ecosim (EwE Christensen and Walters 2004), OSMOSE (Shin and Cury 2001, 2004), SEAPODYM (Lehodey, et al. 2003) and Atlantis (Fulton et al. 2005) have been applied to a variety of ecosystems. More recent developments include the use of regional oceanographic models as a platform to include lower trophic level (or Nutrient-Phytoplankton-Zooplankton, NPZ) models and fish species (Kishi et al. 2007, Travers-Trolet et al. 2014). The high spatial resolution of these models is of particular interest due to the expected changes that global warming may bring on species distributions (Mueter and Litzow, 2008; Kotwicki and Lauth, 2013), and in subarctic and arctic regions, effects on ice (Clement et al., 2004; Grebmeier et al., 2006; Stabeno et al., 2012a) . One closely watched region is the eastern Bering Sea shelf, where the Islands of St. Lawrence, St. Matthews and the Pribilofs are home to some of the largest breeding colonies of planktivorous seabirds in Alaska (Stephensen and Irons, 2003). The eastern Bering Sea shelf is also a primary feeding ground for fin, sei and humpback whales (Moore et al., 2002), as well as northern fur seals. The fisheries catches in this region are amongst the highest in the world (Fissel et al., 2012). One of the most striking and ecologically important features of the eastern Bering Sea shelf is the formation of a cold pool. Starting in late fall, water temperature decreases to -1.7ºC, the freezing point of salt water, and sea ice starts to form in the northern Bering Sea. This ice cannot hold all the salt (or brine) in the sea water, creating colder water of higher density (higher salinity) that remains through summer, a cold pool. This cold pool is present in the Northern Bering Sea even during warm years. South of 60°N, ice cover and cold pool extent vary in warm and cold years as the ice is advected towards the southeastern shelf, depending on cross-shelf transport and wind direction (Danielson, et al. 2011, and Stabeno et al. 2012b). A second “cold pool” forms in the middle shelf of the southeastern Bering Sea that results from the sea ice remaining in the area in late spring. Thus the extent and duration of the ice cover over the central and southeastern shelf is governed by processes independent than those regulating the amount of ice in the northern Bering. With climate projections forecasting overall less ice in the southern portion of the Bering Sea shelf (Wang et al., 2012), the assessment of the potential effects of future climate on the extent, abundance, and resilience of current fisheries has become a key issue for management. Diverse measures have been implemented as part of an EAFM for the Alaskan groundfish fisheries for well over a 15 years (Witherell et al., 2000) and several tools are in use for ecosystem assessment (Hollowed et al., 2013). Within the Eastern Bering Sea, the primary model that provides an ecosystem context for the single species stock assessments, is a mass balance model developed by Aydin et al. (2007). A selected suite of physical, biological and fisheries related ecosystem indicators provide the core information for an annual ecosystem report card and ecosystem assessment chapter (Zador, 2013) and a multiple model approach is used to simulate future ecosystem status (Livingston et al., 2005; Holsman et al., in review). While these combined sources provide information on temporal and spatial trends, they do so separately, and are not necessarily linked to climate projections. This project developed FEAST (Forage and Euphausiid Abundance in Space and Time), a high resolution length-based multispecies fish bioenergetics model and coupled it with other models to implement an integrated ecosystem model for the Bering Sea. The integrated ecosystem model had the dual goal of investigating biophysical processes and climate impacts, as well as aiding fisheries

5


management, thus addressing both bottom-up and top-down forcing mechanisms. This integrated ecosystem model represents the three dimensional dynamics of the two way interactions between physical oceanography (currents, winds, temperature) using a 10km resolution Regional Ocean Modeling System (ROMS), lower trophic levels (nutrients, phytoplankton, small and large zooplankton) by means of a Nutrient-Phytoplankton-Zooplankton-detritus (NPZD), fish (importantly walleye pollock, Pacific cod and arrowtooth flounder) via the Forage and Euphausiid Abundance in Space and Time (FEAST) model and fisheries removals (by sector, gear and species). FEAST takes water temperatures from the Regional Oceanographic Model System (ROMS), prey fields from the lower trophic level model (Nutrient-Phytoplankton-Zooplankton, NPZ), returns mortality values from fish predation, and is set up to provide the fish field for an economics model, and serve as operating model for management strategy evaluation. The model can be run in hindcast or forecast mode (Figure 1). FEAST is also the model that provided the framework into which much of the fish studies and field work data were incorporated, thus serving as a synthesis tool for the most current and detailed understanding of fish dynamics (with a strong focus on pollock) and their response to bio-physical environmental processes in the Bering Sea shelf and slope.

Figure 1. Stucture of the integrated model for the Bering Sea showing one way and two way feedback between components. Results from the 1970-2009 hindcast with the integrated ecosystemn model Bering 10K ROMS-NPZD-FEAST-FAMINE were processed and regridded; weekly output for 293 fields were added to the Bering Sea Projetc Data Archive. The datasets can be viewed, and files ordered, at http://data.eol.ucar.edu/codiac/dss/id=245.B70-001. A list of the fields is included in Appendix A. The dataset containing the mode code can be viewed, and files ordered at http://data.eol.ucar.edu/ codiac/dss/id=245.B70-002.

6

http://data.eol.ucar.edu/codiac/dss/id=245.B70-00

http://data.eol.ucar.edu/%20codiac/dss/id=245.B70-00

http://data.eol.ucar.edu/%20codiac/dss/id=245.B70-00


Chapter 2. Objectives The objectives listed here pertain only to FEAST, the fish portion of Bering 10K-ROMZ-NPZD-FEAST-FAMINE model (no objectives or results from the ROMS/NPZD/FAMINE/MSE portions are included here; results from these can be found in their corresponding reports mentioned in the study chronology section). Due to the extended amount of time required to get a hindcast with suitable results, no forecasts were simulated (see Chapter 10). Accordingly, project objectives were modified in April 2013: Original objectives Final objectives (as detailed in workplan for

second extension) 1. Assess fish productivity and growth and movement in response to physical processes and two way feedback with lower trophic level processes that occur on seasonal and multi-decadal scales (e.g., bloom timing, warm/cold years) based on a hindcast simulation from mid 1970’s to 2005.

1. Assess fish productivity and growth and movement in response to physical processes and two way feedback with lower trophic level processes that occur on seasonal and multi-decadal scales (e.g., bloom timing, warm/cold years) based on a hindcast simulation from 1971 to 2009.

a) Develop bioenergetics model of fish growth for walleye pollock, arrowtooth flounder, Pacific cod, herring, salmon, capelin, sandlance, eulachon, myctophids, squids and shrimp.

a) Develop bioenergetics model of fish growth for walleye pollock, arrowtooth flounder, Pacific cod, herring, salmon, capelin, sandlance, eulachon, myctophids, squids, crabs, shrimp, and miscellaneous zooplankton.

b) Model reproductive output and cycle of fish groups

b) Model reproductive output and cycle of fish groups.

c) Develop movement model for fish (couple with NPZ and ROMS).

c) Develop movement model for fish and fully couple to Bering 10K ROMS-NPZD from NSF 0732534

2. Assess effects of fishing effort distribution on fish stocks dependent on forecasts –postponed (couple with Economic model)

2. Simulate hindcast with spatially explicit fisheries removals as estimated by project B.72 (Economics).

3. Evaluate policy effects on fish stocks and economics (linked to MSE) (dependent on forecasts – postponed)

3. Coding fully coupled to economics and MSE

4. Compare the results of 1 and 2 and 3 under three distinct future climate scenarios based on 3 climate models run with the same emission scenario for years 2000-2050 (input from NSF 0732534) (dependent on forecasts-postponed)

4. Select ecosystem/ production indices for multi-species model (MSMt) from Bering 10K-ROMS-NPZD 2003-2039) forecasts based on three different climate models (from NSF 0732534).

5. Set-up a fully coupled integrated ecosystem model (Climate to fisheries) designed for MSE.

7


Objective 1a. Develop a bioenergetics model of fish growth for walleye pollock, arrowtooth flounder, Pacific cod, herring, salmon, capelin, sandlance, eulachon, myctophids, squids, crabs, shrimp, epifauna, and miscellaneous zooplankton. Bioenergetics models provide a mechanistic framework to determine factors that affect feeding and growth as well as quantifying consumption under dynamic environmental and ecological conditions. We developed a length-based bioenergetics model with dynamic prey fields representing a simplified food web for the eastern Bering Sea. The model is primarily based on direct sampling for diet, growth and size structure of walleye pollock, Pacific cod and arrowtooth flounder. The model takes prey estimates of euphausiids and copepods as well as temperatures from the Bering 10K ROMS-NPZD model. This provides a method for quantifying trophic interactions in a temporal, spatial and ontogenetic framework. This model is described in Chapter 3. We conducted restrospective analyses of length at age, weight at length and prey consumption data (1982-2005) for pollock, cod, and arrowtooth flounder. A preference/ length-based prey selection was parameterized based on 1982-2005 geo-referenced feeding habits. Prey fields are actually dynamic fields (they change through time), thus taking advantage of the changing zooplankton and moving fish prey. The prey selection model establishes the trophic links between the zooplankton groups from the NPZD model and the fish groups from FEAST. The final bioenergetics-prey selection model is thus based on dynamic prey and temperature fields, yielding temporal and spatially explicit growth in fish, with fat and skinny fish defined by their condition factor (proportion of current weight above or below expected weight at length). Objective 1b. Model reproductive output and cycle of fish groups with nudging in recruitment at start of every year. We used published results for each fish group to parameterize a size based approach to maturity and fecundity. We separated maturity at length from fecundity at length to allow for increased fecundity in older/longer females. Both maturity and fecundity were based on logistic curves, multiplied by the assumed proportion of females in the population (assumed 50% if no published data was available). FEAST has a minimum size of 1 cm for any given fish group, so larvae are not explicitly modeled. Instead, the number of fish at length in the third week of the year are used to estimate the total number of eggs produced, and that total number of eggs is released as fish with a normal length distribution for young of the year, at a later period of variable date and duration depending on the species. This is also detailed in Chapter 3, as part of the growth and bioenergetics. Objective 1c. Develop movement model for fish groups and fully coupled to Bering 10K ROMS-NPZD. Fish movement is horizontal, that is, only in two dimensions (2D) 2D. There is no vertical movement (up and down the water column) given the sparcity fish data available on the subject. The 2D movement allows for basic distribution and overlap of predators and their prey. Movement speed and direction is modeled based on habitat quality, which is measured as the potential biomass gain in any given cell by fish at a given length. Direction (along the x and/or y axis, but not diagonally) is selected based on the neighboring cell (out of 4) with the highest estimated biomass gain. This last is calculated based on the product of individual weight gain and predation mortality. The speed is based on a “happiness” function, where the ratio of maximum potential biomass gain to current biomass gain is transformed into low or high speed (based on length of fish). Net biomass loss represents poor habitat quality where the fish is “sad” and therefore moves faster to find better

8


a better habitat; biomass gain represents a good habitat, and “happy” fish stay longer (that is, they move slower). This algorithm allowed us to replicate general distribution patterns for pollock, cod and arrowtooth when compared to data from retrospective distribution analyses. Details are in included in Chapter 4, Fish movement. The movement model is fully coupled to the Bering 10K ROMS-NPZD; Gibson and Spitz (2011) details the NPZD model while Hermal et al., (2013) describes the Bering 10K ROMS-NPZD model. Objective 2. Simulate hindcast with spatially explicit fisheries removals as estimated by project B.72 (Economics). The hindcast is based on the integrated ecosystem model Bering 10K ROMS-NPZD-FEAST-FAMINE. The oceanography (and platform for the integrated ecosystem model) is a simplified version of the ROMS-NEP5 oceanographic model. The Bering 10K ROMS retains most parameter values, has 10 vertical layers (down from 60 vertical layers), and a smaller geographical extent –reduced from Russia-California to Russia-Gulf of Alaska. From the hindcast, we were able to evaluate biophysical processes, both in space and time, such as spring blooms of phytoplankton and zooplankton. We found fall blooms could potentially have an important role on the ecosystem dynamics, as detailed in Chapter 5 which focuses on the 1970-2009 hindcast of the vertically integrated model. Chapter 8 includes a brief description of the fisheries data and model included in the hindcast, as well as specific actions related to integrating the Bering 10K ROMS-NPZD-FEAST-FAMINE model. Objective 3. FEAST coding fully coupled to economics and MSE (detailed in B.73 MSE). The FEAST model is structured so that model outputs provide all data needed to: 1) simulate surveys 2) run single species stock assessments for pollock, Pacific cod and arrowtooth flounder, 3) run a multispecies stock assessment with temperature for pollock Pacific cod and arrowtooth flounder, and 4) run a simplified Ecosim model for the eastern Bering Sea. The Ecosim model was developed by S. Gaichas (not part of the FEAST project) and I. Ortiz. We met regularly with both the Economics and the MSE group to ensure the model structure suited all the required estimates for MSE as well as divided the fishing effort into categories that suited the Economics model. We coded all data retrieval and formatting for input into the MSE routine and the corresponding coding to incorporate input back from the MSE (catch quotas from distinct management strategies) and catch allocation. Coding in FEAST was accomplished by close collaboration with the MSE team (B.73) and was lead by Aydin and Moffitt. Essentially, output data from FEAST is extracted, processed and written into csv files from which the MSE code procedes to extract the data and generate simulated survey data used as input in the stock assessment routines. The MSE also includes two models for catch allocation. ADMB, Matlab, and Stock Synthesis were implemented on the BEAST computing cluster located at the Alaska Fisheries Science Center, and the routines developed to communicate between the operating model and the single species assessment models have been tested. Details of the full methods and coupling of FEAST and MSE and the use of FEAST as the operating model for MSE are included in Chapter 3 of the final report to NPRB of project B.73 MSE, of which Aydin and Ortiz are co-authors. This chapter is included in this report as Chapter 6. Objective 4. Select ecosystem/ production indices for multi-species model (MSMt) for Bering 10K ROMS-NPZD 2003-2039 forecasts based on three different climate models.

9


Multiple general circulation models were evaluated to select the climate model projections from which forcing files were extracted to simulate forecasts with the Bering 10K ROMS-NPZD. General circulation models were selected based on their performance in the Bering Sea, mainly their ability to capture ice cover and the PDO; details of this evaluation are found in Wang et al. (2010) and the study was part of the NSF 0732534 project. The models selected are CGMM-t47, ECHO-G and MIROCM which correspond to large, small and moderate declines respectively in winter ice on the Bering Sea shelf. These forecasts and results from their analysis can be found in Hermann et al. (in review) submitted to the fourth special issue of the Bering Sea Project; other components (oceanography and lower trophic levels) were also part of project NSF Award ID 0732534. MSMt is a multispecies stock assessment model in which weight-at-age and predation mortality vary as a function of bottom temperature, allowing MSMt to capture climatic-driven changes in growth and predation effects on biomass and attendant harvest rates. From each of the ROMS/NPZ forecast output, we developed time-series of bottom temperature, summer cold pool extent, total mean summer zooplankton biomass, and other environmental indices to be used s forcing functions on recruitment and predation in the forecast simulations of MSMt. Details on MSMt and FEAST derived timeseries for forecasts can be found in Chapter 7. Blended Forecasts and Short-Term Forecasts. The first section (Blended Forecasts) is reproduced from the B.73 MSE final report to NPRB (Chapter 5) and also in Holsman and Aydin (in review). The section corresponding to short-term forecasting is original content for this report. Objective 5. Set up a fully integrated ecosystem model (Climate to fisheries) designed for MSE. Coupling and setting up all the different components of the integrated ecosystem model Bering 10K ROMS-NPZD-FEAST-FAMINE for MSE required several specific modifications/tasks;these are described in Chapter 8. A full evaluation of the process of implementing integrated models, taking Bering 10K ROMS-NPZD0FEAST-FAMINE as the case study is included in Chapter 9.

10


Chapter 3. The FEAST foraging and bioenergetics model

Introduction

The FEAST model is based on the Wisconsin bioenergetics model for individual fish, coupled to other ecosystem components (predator and prey) using the energy-based foraging maximization. It can be implemented for individual fish (Lagrangian individual-based modeling) or for populations (Eularian state variable modeling).1

The Wisconsin bioenergetics model predicts the expected daily growth of a fish, given a specified feeding regime and water temperature; conversely, given the observed growth of a fish, the model predicts the consumption required to maintain that growth. The model user typically assumes that the fish engages in an “average” energetic activity level (ACT) over the course of the predictive period; ACT is a multiplier of the fish’s basal metabolic rate used to account for “active” energy expenditure (such as foraging or migration). This value is typically determined in laboratory experiments where both consumption and growth are known, but it has been noted that lab-measured activity levels may diverge considerably from actual active energy expenditure in a wild fish.

Foraging models, on the other hand, aim to predict the amount of food consumed by a fish, or functional response between predator and prey, given a particular prey field (and other environmental conditions) in the fish’s environment. Foraging models may take into account volume searched per unit time (combining swimming speed and visual acuity), processes such as predator satiation (the Type II functional response), prey refugia at low prey densities (Type III functional response), the movement of both predators and prey into and out of “foraging arenas”, and/or selection between multiple prey types. However, while functional foraging models predict the amount of prey consumed by a predator at a given prey density and composition, they do not tend to take into account variable energy expenditure required to obtain that prey; instead generally using a fixed “growth efficiency” to translate food consumed into usable energy.

The FEAST foraging and bioenergetics model combines the foraging and bioenergetics models by relating the energy obtained in foraging (consumption) and the energy expended during foraging (active respiration) directly to “foraging velocity” 𝑉𝑉� , used as a measure of a fish’s activity in given conditions. As described below, consumption (energetic gain) increases with 𝑉𝑉� up to a point of satiation, while respiration (loss) increases indefinitely with 𝑉𝑉� ; therefore, for biologically realistic parameter values, there is a “choice” of swimming velocity that maximizes net energetic gain.

The FEAST method thus calculates the expected daily consumption and respiration for a fish as a function of 𝑉𝑉� , then determines the value of 𝑉𝑉� for which net energy gain is maximized (𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖). 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 is considered to be the swimming speed “chosen” by the predator for that day, and is used to calculate the daily consumption, respiration, and resulting growth and prey mortality for the fish in the model. Parameters such as ACT, the bioenergetics P (proportion of maximum consumption), or instantaneous growth efficiency can be back-calculated from these results as diagnostics. It’s important to note that, while the functional forms relating to 𝑉𝑉� to energy expenditure and foraging are chosen based on the physics of fish movement, this “velocity” includes foraging maneuvers and

1 For this section, state variables are shown in bold (e.g., W), input parameters calculated outside the model are shown in italics (𝑒𝑒.𝑔𝑔. , 𝑞𝑞𝑝𝑝𝑝𝑝) while intermediate or model-calculated values or functions are shown with a hat (e.g. 𝐸𝐸�𝑣𝑣).

11


is scaled to prey through selectivity components, so does not represent a true velocity in terms of fish movement.

For a single fish of a given species, the variables tracked for the fish are its length L (cm), its wet weight W (g), and its energetic density ED (joules/g wet weight). The FEAST model predicts the fish’s daily foraging activity, consumption and growth of the fish, given the additional inputs of daily water temperature (°C) and the available prey field (𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖, described below). These values can be used to predict the growth trajectory of an individual fish at a given location.

To fold this into an Eulerian model, each predator species is divided into fixed length class bins, and tracks the number (N, fish/m2) of each length class, and the condition factor (CF, ratio), and the energy density (ED, joules/g wet weight) for fish in that bin. The weight 𝑊𝑊� for the fish in that bin is calculated as 𝑪𝑪𝑪𝑪 ∙ 𝐴𝐴𝑖𝑖 𝐿𝐿�𝐵𝐵𝑙𝑙 , where 𝐿𝐿� is the midpoint of the fish length in that bin, and 𝐴𝐴𝑖𝑖 and 𝐵𝐵𝑖𝑖 are parameters for the long-term average length/weight relationship for that species. For each bin, then, length 𝐿𝐿�, weight 𝑊𝑊� , and energy density ED are used as in the individual fish model to calculate expected growth; growth in length is expressed as graduation of a proportion of N to the next length class.

To calculate both consumption and respiration, the FEAST model requires indices of prey density for parameter calibration/fitting. As the model uses “catchability” constants to scale between predator and prey, the units for prey items may differ; in this model, for plankton, the units used are gCarbon/m2, while for fish, the prey units are biomass of fish (in wet weight/m2) at length. For bioenergetics calculations, the catchability constants are assumed to translate units to wet weight.

Foraging equations

For each predator, the prey density used to calculate functional responses is the prey available to that predator, 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖, which varies based on size and species-dependent selectivity and catchability (success of capturing an encountered prey) using the following formulation:

(1)

𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 = � 𝑞𝑞𝑝𝑝𝑝𝑝 ∙ Γ�(𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑,𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑)𝑝𝑝𝑝𝑝𝑖𝑖𝑝𝑝 𝑠𝑠𝑝𝑝,𝑖𝑖𝑎𝑎𝑖𝑖,𝑖𝑖𝑖𝑖𝑙𝑙

∙ 𝐵𝐵�𝑝𝑝𝑝𝑝𝑖𝑖𝑝𝑝 𝑠𝑠𝑝𝑝,𝑖𝑖𝑎𝑎𝑖𝑖,𝑖𝑖𝑖𝑖𝑙𝑙

where 𝐵𝐵�𝑝𝑝𝑝𝑝𝑖𝑖𝑝𝑝 𝑠𝑠𝑝𝑝,𝑖𝑖𝑎𝑎𝑖𝑖,𝑖𝑖𝑖𝑖𝑙𝑙 is the density of prey in the environment, 𝑞𝑞𝑝𝑝𝑝𝑝 is a catchability constant for each predator/prey species pair. The Γ�(𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑,𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑) function bases prey selection and capture on the log of the ratio between predator size (𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑) and prey size (𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑). For fish, size is indexed by fork length; for crabs, size is indexed by carapace width. For plankton, a fixed average length is used to typify each group. The function is a gamma function with parameters 𝛼𝛼𝑝𝑝𝑝𝑝 and 𝛽𝛽𝑝𝑝𝑝𝑝 specific to each predator/prey species pair:

(2)

Γ��𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑,𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑� = ��̂̂�𝑆𝑝𝑝𝑖𝑖𝑟𝑟𝑖𝑖𝑟𝑟�̂�𝑆𝑚𝑚𝑖𝑖𝑚𝑚

�𝛼𝛼𝑝𝑝𝑝𝑝−1

𝑒𝑒−(�̂�𝑆𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟−�̂�𝑆𝑚𝑚𝑟𝑟𝑥𝑥)

𝛽𝛽𝑝𝑝𝑝𝑝

12


where �̂̂�𝑆𝑝𝑝𝑖𝑖𝑟𝑟𝑖𝑖𝑟𝑟 = 𝑙𝑙𝑙𝑙𝑔𝑔(𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑/𝑳𝑳𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑) and �̂�𝑆𝑚𝑚𝑖𝑖𝑚𝑚 = (𝛼𝛼𝑝𝑝𝑝𝑝 − 1)βpp and is the log ratio of maximum selection (for fitting purposes, �̂�𝑆𝑚𝑚𝑖𝑖𝑚𝑚 may be easier to estimate from diet length composition data than one of alpha or beta). Diet Composition (𝐷𝐷𝐷𝐷� , proportion in diet by weight) of each prey item in the predator’s diet is simply the proportion of the total 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 due to a particular prey item.

The area of water that a fish sweeps out while foraging is proportional to its swimming velocity while foraging (𝑉𝑉�) and also a function of its length: as a fish increases its length, the area it sweeps out in front of while foraging for prey increases with visual acuity (eye size) in proportion to body length. Swept area for the fish is thus 𝐸𝐸�𝑣𝑣 = 𝐴𝐴𝑖𝑖𝑙𝑙𝑒𝑒𝑳𝑳

𝐵𝐵𝑒𝑒𝑒𝑒𝑒𝑒 where 𝐴𝐴𝑖𝑖𝑙𝑙𝑒𝑒 and 𝐵𝐵𝑖𝑖𝑙𝑙𝑒𝑒 are allometric

parameters for predator body length, and 𝑳𝑳 is the length of the predator. At low prey densities, consumption is assumed to be linearly proportional to the overall encounter rate 𝐸𝐸�𝑣𝑣 ∙ 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉� . However, at high prey densities, consumption may be slowed down either by handling time (time for the predator to consume the prey) or digestion rate limitations (satiation). Using the standard Type II foraging equation to adjust foraging rates for handling time gives consumption, in grams weight weight/day, as a function of velocity 𝑉𝑉� and available prey as follows:

(3)

�̂�𝐷𝑤𝑤𝑤𝑤 = 𝐸𝐸�𝑣𝑣 ∙ 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉�

1 + 𝐸𝐸�𝑣𝑣 ∙ 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉��̂�𝐷𝑚𝑚𝑖𝑖𝑚𝑚

where �̂�𝐷𝑚𝑚𝑖𝑖𝑚𝑚 = 𝑓𝑓𝑒𝑒(𝑻𝑻) ∙ 𝐴𝐴𝑠𝑠𝑳𝑳 𝐵𝐵𝑠𝑠

is the maximum daily consumption rate of the predator. Note that other formulations of the functional response are possible here (e.g. the Type III functional response) although that would require re-derivation of the velocity maximization method described below. The wet weight of consumption is translated into joules of assimilated energy by converting 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 and �̂�𝐷𝑚𝑚𝑖𝑖𝑚𝑚 to joules, using 𝐽𝐽𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 = 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝐸𝐸𝐷𝐷𝑝𝑝𝑝𝑝𝑖𝑖𝑝𝑝 ∙ 𝑃𝑃𝑖𝑖𝑠𝑠𝑠𝑠𝑖𝑖𝑚𝑚 and 𝐽𝐽𝑚𝑚𝑖𝑖𝑚𝑚 = �̂�𝐷𝑚𝑚𝑖𝑖𝑚𝑚 ∙ 𝐸𝐸𝐷𝐷𝑝𝑝𝑝𝑝𝑖𝑖𝑝𝑝 ∙𝑃𝑃𝑖𝑖𝑠𝑠𝑠𝑠𝑖𝑖𝑚𝑚, where 𝐸𝐸𝐷𝐷𝑝𝑝𝑝𝑝 and 𝑃𝑃𝑖𝑖𝑠𝑠𝑠𝑠𝑖𝑖𝑚𝑚 are average energy density of the prey, and proportion of prey weight that is not lost to egestion or excretion, weighted by the relative contribution of each prey type to 𝐵𝐵�𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 using DC). Therefore, the total joules obtained as the result of foraging is:

(4)

�̂�𝐷𝑗𝑗 = 𝐸𝐸�𝑣𝑣 ∙ 𝐽𝐽𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉�

1 + 𝐸𝐸�𝑣𝑣 ∙ 𝐽𝐽𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉�𝐽𝐽𝑚𝑚𝑖𝑖𝑚𝑚

Respiration equations

Respiration can be broken into resting (basal metabolism) and active components. The Wisconsin model calculates a resting (basal) respiration rate (𝑅𝑅�𝑓𝑓𝑖𝑖𝑚𝑚) as a function of body weight and water

temperature, 𝑅𝑅�𝑓𝑓𝑖𝑖𝑚𝑚 = 𝑓𝑓𝑝𝑝(𝑻𝑻) ∙ 𝐴𝐴𝑝𝑝𝑳𝑳 𝐵𝐵𝑟𝑟

. To scale from this basal rate to actual daily respiration (including active movement and foraging), most bioenergetics model implementations use a fixed constant, ACT, so that total respiration 𝑅𝑅�𝑗𝑗 = 𝑅𝑅�𝑓𝑓𝑖𝑖𝑚𝑚 ∙ 𝐴𝐴𝐷𝐷𝐴𝐴. ACT is meant to account for a range of

13


activities in which a fish might engage, and might differ substantially between laboratory and the field.

Here, we take the approach of linking ACT to the velocity 𝑉𝑉� of the foraging equation, above, by setting 𝐴𝐴𝐷𝐷𝐴𝐴 = 1 + 𝑅𝑅�𝑣𝑣𝑖𝑖𝑖𝑖𝑉𝑉�𝐵𝐵𝑣𝑣𝑒𝑒𝑙𝑙 so that :

(5)

𝑅𝑅�𝑗𝑗 = 𝑅𝑅�𝑓𝑓𝑖𝑖𝑚𝑚(1 + 𝑅𝑅�𝑣𝑣𝑖𝑖𝑖𝑖 𝑉𝑉�𝐵𝐵𝑣𝑣𝑒𝑒𝑙𝑙)

Here, the energy required to maintain a specific velocity increases with velocity, due to drag. Large fish are better able to overcome drag than small fish, so 𝑅𝑅�𝑣𝑣𝑖𝑖𝑖𝑖 is set to be a decreasing function of body length as follows: 𝑅𝑅�𝑣𝑣𝑖𝑖𝑖𝑖 = 𝑅𝑅𝑚𝑚 + 𝑅𝑅0/(1 + �𝑅𝑅0

𝑅𝑅1− 1� 𝑳𝑳𝑅𝑅𝑝𝑝𝑟𝑟𝑝𝑝) where Rm, R0, R1, and Rpow are

determined by fitting the model to measured growth trends.

Maximizing net growth

For a fish of a given species, length, prey field, and temperature, the net energy, as a function of foraging velocity 𝑉𝑉� is therefore:

(6)

𝐽𝐽𝑙𝑙𝑖𝑖𝑟𝑟 = �̂�𝐷𝑗𝑗 − 𝑅𝑅�𝑗𝑗 =𝐸𝐸�𝑣𝑣 ∙ 𝐽𝐽𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉�

1 + 𝐸𝐸�𝑣𝑣 ∙ 𝐽𝐽𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉�𝐽𝐽𝑚𝑚𝑖𝑖𝑚𝑚

− 𝑅𝑅�𝑓𝑓𝑖𝑖𝑚𝑚(1 + 𝑅𝑅�𝑣𝑣𝑖𝑖𝑖𝑖 𝑉𝑉�𝐵𝐵𝑣𝑣𝑒𝑒𝑙𝑙)

Note that, for reasonable biological values of parameters, consumption (gain) is an asymptotic function of 𝑉𝑉� while respiration (loss) is an increasing function of 𝑉𝑉� . Therefore, for any given set of conditions, there will be a velocity 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 at which 𝐽𝐽𝑙𝑙𝑖𝑖𝑟𝑟 is a maximum. 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 can be found by setting the derivative of 𝐽𝐽𝑙𝑙𝑖𝑖𝑟𝑟 with respect to 𝑉𝑉� to 0 and solving for 𝑉𝑉� :

(7)

𝑑𝑑𝐽𝐽𝑙𝑙𝑖𝑖𝑟𝑟𝑑𝑑𝑉𝑉�

= 0 = 𝐸𝐸�𝑣𝑣 ∙ 𝐽𝐽𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ �𝐽𝐽𝑚𝑚𝑖𝑖𝑚𝑚�

2

� 𝐸𝐸�𝑣𝑣 ∙ 𝐽𝐽𝑖𝑖𝑣𝑣𝑖𝑖𝑖𝑖𝑖𝑖 ∙ 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 + 𝐽𝐽𝑚𝑚𝑖𝑖𝑚𝑚�2 − 𝐵𝐵𝑣𝑣𝑖𝑖𝑖𝑖𝑅𝑅�𝑓𝑓𝑖𝑖𝑚𝑚𝑅𝑅�𝑣𝑣𝑖𝑖𝑖𝑖�𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖�

𝐵𝐵𝑣𝑣𝑒𝑒𝑙𝑙−1

This equation can be solved for 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 numerically, but some values of 𝐵𝐵𝑣𝑣𝑖𝑖𝑖𝑖 lead to analytical solutions. In particular, values found experimentally for 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 are between 1.8-2.3. Setting 𝐵𝐵𝑣𝑣𝑖𝑖𝑖𝑖 to 2.0 allows 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 to be found as the solution to a cubic equation using Cardano’s Method. Once the solution for 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 is found, it can be substituted for �̂�𝐷𝑗𝑗 and 𝑅𝑅�𝑗𝑗 to determine 𝐽𝐽𝑙𝑙𝑖𝑖𝑟𝑟 at 𝑉𝑉�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (�̂�𝐷𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 , 𝑅𝑅�𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖, and 𝐽𝐽𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖) as the “actual” daily rates realized by the fish in the model. Derived values such as ACT or the bioenergetics P (consumption as proportion of maximum consumption) can also be calculated.

14


Allocation of growth, reproduction, and mortality

Once 𝐽𝐽𝑙𝑙𝑖𝑖𝑟𝑟for a fish has been determined, growth (or loss) from that energy must be allocated between length, weight, and caloric density. If the fish is experiencing loss, length is not lost; 1/3 of loss is attributed to weight loss while 2/3 of loss is attributed to caloric density loss; this is based on a set of laboratory experiments for adult gadids. If net energetic change is positive, a “target” energetic density is calculated as a linear function of fish length: 𝐸𝐸𝐷𝐷�𝑟𝑟𝑖𝑖𝑝𝑝𝑎𝑎𝑖𝑖𝑟𝑟 = 𝐸𝐸𝐷𝐷𝑏𝑏𝑖𝑖𝑠𝑠𝑖𝑖 + 𝐸𝐸𝐷𝐷𝑚𝑚 ∙ 𝑳𝑳, where EDbase and EDm are fit from collected fish data. If the fish’s energy density is less than 1.05x the target level, half of gained energy is assigned to caloric growth. Likewise, if a fish’s condition factor (CF) is less than 1.05, half of gained energy is assigned to weight growth. Any remaining energy after these calculations is assigned to growth in length. If spawning is being tracked, then during a set spawning season, weight (condition factor) is lost in proportion to fixed, length-based weight, maturity, and fecundity curves.

For population-scale simulations, mortality from predation is calculated by summing consumption by prey across all predators, once consumption has been determined as described above. Additionally, starvation is simulated by removing all fish from the population if their condition factor or caloric density drops below 25% of long-term average for that body length. An additional mortality term, M0, is used to provide closure from top predators not contained in the model.

15


Chapter 4. Fish movement in FEAST Abstract We developed an algorithm for fish movement based on maximum biomass gain which is part of larger model that uses a regional ocean system model (ROMS) for the Bering Sea as a platform, and is linked to a lower trophic level model and a multispecies bioenergetics length/preference prey selection model. The purpose of the model was to explore fish distribution as a result of prey abundance and environmental conditions, there is no directed movement in the model. We compared the model results for walleye pollock, Pacific cod and arrowtooth flounder from multi-year simulations during cold and war years and compare them to the summer distribution based on bottom trawl surveys. The model shows skill capturing the general distribution for each species as well as differences between cold and warm years. However, fall distribution seems to fit better the summer survey (particularly for pollock) and this may be related to the delayed timing of the phytoplankton and zooplankton blooms in the model compared to mooring data as well as the fact that the survey is conducted over a period of 3 months, from June to end of August. Introduction In recent years, interest in the development of end to end models using a regional ocean system model (ROMS) as a platform has increased, as these models have the ability to incorporate climate, oceanography, lower trophic levels, and recently fish as well (Travers-Trolet et al., 2014, Ortiz et al., in review). Forecast projections of water temperature and ice cover in the Bering Sea (Wang et al, 2012, Hermann et al. 2014) have highlighted the importance of spatial dynamics in biophysical processes and the response of fish to these environmental changes. Mueter and Litzow (2008) analyzed historical distribution of fish and showed latitudinal changes in distribution across a large suite of species in the Bering Sea shelf. As part of the Bering Sea Project, an integrated ecosystem modeling approach was developed using a 10km resolution Regional Ocean Modeling System (ROMS) (Hermann et a.l 2013), coupled to lower trophic levels (nutrients, phytoplankton, small and large zooplankton) by means of a Nutrient-Phytoplankton-Zooplankton-detritus (NPZ) (Gibson and Spits, 2011), and fish (importantly walleye pollock, Pacific cod and arrowtooth flounder) via the Forage and Euphausiid Abundance in Space and Time (FEAST) model (Ortiz et al., in review) which is based on bioenergetics and a length-preference based prey selection model (Aydin et al., in review).. The purpose of the movement algorithm was to explore fish distribution as a result of prey abundance and environmental conditions, with no directed movement in the model (a common approach to ensure the right directionality/timing of movement in other models, such as Atlantis) or kinetic (temperature based) movement. We compared the model results for walleye pollock, Pacific cod and arrowtooth flounder from multi-year simulations during cold and warm years and compare them to the summer distribution based on bottom trawl surveys, as well as zooplankton distribution.

16


Methods The fish in FEAST are modeled using explicit horizontal movement and no vertical movement. Each fish species is assumed to be in a single, fixed vertical layer; therefore, there is no vertical movement tracked for fish. Fish swim speed is based on length and a diffusion rate inversely proportional to local habitat quality along a gradient. Direction of movement is based on habitat quality measured as a function of net population biomass gain rate and predation mortality. So fish in a high quality habitat swim in circles and diffuse slower than fish in poor quality habitat which swim longer distances in straighter lines. Briefly, the horizontal movement function is described as movement of a single numerical density (N) in a single spatial dimension x over time t such that movement is in response to local conditions (λ) and diffusion D:

∂∂

⋅∂∂

+

∂∂

⋅∂∂

=

∂∂

xND

xxx

xmove

tN λ

(1)

We compared the model results for walleye pollock, Pacific cod and arrowtooth flounder from multi-year simulations during cold and war years and compare them to the summer distribution based on bottom trawl surveys. We also compared the fish distrubtion to that of the large crustacean zooplankton as modeled in FEAST. The model does capture seasonal changes in the general distribution for each species as well as differences between cold and warm years. However, fall distribution seems to fit better the summer survey (particularly for pollock) and this may be related to the delayed timing of the phytoplankton and zooplankton blooms in the model compared to mooring data as well as the fact that the survey is conducted over a period of 3 months, from June to end of August. Figures 4.1 to 4.3 show the modeled seasonal distribution in a warm (2004) and cold (2008) of pollock, cod and arrowtooth respectively. Figure 4.4 shows the seasonal distribution in a warm (2004) and cold year (2008) of large crustacean zooplankton as modeled by FEAST. We also compared the disfferences in zooplankton distribution between simulations with two-way feedback with the NPZD model and one-way feedback. Compared distributions for fall 2004 show higher prey densities in the one way feedback version where the mortality of the large crustacean zooplankton is always proportional to biomass (quadratic mortality). In cosntrast, the two way feedback simulation had more even distribution across the shelf, with less high density concentrations due to the fish moving towards areas where prey is more available (Figure 4.5). References Aydin, K., Ortiz, I., Hermann, A.J., In review. New approach to modeling fish bioenergetics, a case

study for walleye pollock, Pacific cod and arrowtooth flounder in the Eastern Bering Sea. Deep Sea Res. II 00, 00–00.

Gibson, G.A. and Y.H. Spitz. 2011. Impacts of biological parameterization, initial conditions, and environmental forcing on parameter sensitivity and uncertainty in a marine ecosystem model for the Bering Sea. Journal of Marine Systems 88: 214-231.

Hermann, A.J., Gibson, G.A., Bond, N.A., Curchitser, E.N., Hedstrom, K., Cheng, W., Wang, M., and Stabeno, P.J. in review. Projected futures biophysical states of the Bering Sea. Deep Sea Reasearch II 00:00-00.

Hermann, A. J., Gibson, G.A., Bond, N.A., Curchitser, E.N., Hedstrom, K., Cheng, W., Wang, M., Stabeno, P.J., Eisner, L. and Cieciel, K.D.. 2013. A multivariate analysis of observed and

17


modeled biophysical variability on the Bering Sea shelf: multidecadal hindcasts (1970-2009) and forecasts (2010-2040). Deep Sea Research II, doi:10.1016/j.dsr2.2013.04.007.

Ortiz, I., Aydin, K., Hermann, A.J., Gibson, G., In review-b. Climate to fisheries: a vertically

integrated model for the eastern Bering Sea. Deep Sea Res. II 00, 00–00. Travers-Trolet, Morgane; Shin, Yunne-Jai; Shannon, Lynne J; Moloney, Coleen L; Field, John G.

2014b. Combined Fishing and Climate Forcing in the Southern Benguela Upwelling Ecosystem: An End-to-End Modelling Approach Reveals Dampened Effects:. PLoS One 9. E94286.

Wang, M., Overland, J.E., Bond, N.A., 2010. Climate projections for selected large marine ecosystems. J. Mar. Syst. 79(3-4), doi: 10.1016/j.jmarsys.2008.11.028, 258-266

18

http://www.pmel.noaa.gov/publications/search_abstract.php?fmContributionNum=3042



Figure4.1. Seasonal pollock distribution in a warm (2004, top) and cold (2008, bottom) as modeled by FEAST and compared to data from summer bottom trawl survey. From left to right winter (week 5), spring (week 18), summer (week 32) and autumn (week 45).

Figure 4.2. Seasonal distribution of Pacific cod in a warm (2004, top) and cold (2008, bottom) as modeled by FEAST and compared to data from summer bottom trawl survey. From left to right winter (week 5), spring (week 18), summer (week 32) and autum (week 45).

19


Figure 4.3. Seasonal distribution of arrowtooth flounder in a warm (2004, top) and cold (2008, bottom) as modeled by FEAST and compared to data from summer bottom trawl survey. From left to right winter (week 5), spring (week 18), summer (week 32) and autum (week 45).

Figure 4.4 Seasonal distribution of large crustacean zooplankton in a warm (2004, top) and cold (2008, bottom) as modeled by FEAST. From left to right winter (week 5), spring (week 18), summer (week 32) and autum (week 45).

20


Figure 4.5 Large crustacean zooplankton (euphausiids and copepods). Two way coupling between FEAST (left panel) and one way coupling (right panel) for fall 2004.

21


Chapter 5. Climate to fisheries: Exploring processes in the eastern Bering Sea based on a 40 year hindcast Ivonne Ortiz1.2, Kerim Aydin2, Albert J. Hermann3, Georgina Gibson4

1School of Aquatic and Fishery Sciences, Box 355020, University of Washington, Seattle WA 98195, USA. E-mail: [email protected]

2NOAA Alaska Fisheries Science Center, 7600 Sand Point Way N.E., Building 4, Seattle, WA 98115 3Joint Institute for the Study of the Atmosphere and Ocean, University of Washington, Seattle, WA 98195, U.S.A. 4International Arctic Research Center, University of Alaska Fairbanks, Fairbanks, AK 99775, U.S.A. Corresponding author address University of Washington School of Aquatic and Fishery Sciences 1122 NE Boat St. Seattle WA 98105 E: [email protected] P: 206.526.4692; F: 206.526.6723 Keywords: Alaska, Bering Sea, end-to-end modelling, integrated ecosystem modelling, hindcast Citation: Ortiz, I, Aydin, K., Hermann, A., and Gibson, G. Climate to fisheries: Exploring processes in the eastern Bering Sea based on a 40 year hindcast. Prepared for submission to Deep-Sea Research II

22


Abstract We used Bering 10K-ROMS-NPZD-FEAST an integrated ecosystem model, to simulate a 40 year hindcast from 1971 to 2009 and explore both seasonal and interannual spatiotemporal patterns. The integrated model is comprised by 4 components: a regional oceanographic model (ROMS), a lower trophic level model (NPZD), a multispecies length-based bioenergetic fish model (FEAST, Forage Euphausiid Abundance in Space and Time) and spatially explicit historical fisheries removals. General biophysical spatio-temporal patterns persist, despite annual variability. Weekly temperature anomalies by region reflect the dominant water masses. Although cold/warm regimes are evident across the shelf and slope, their timing varies by region. The same applies to the strength of euphausiid (Order Euphausiacea) biomass anomalies. The tendency of euphausiid abundance to covary inversely with temperature is observed as in previous simulations not coupled to explicitly modeled fish predation; however we note that this tendency is stronger in off-shelf regions and weaker in years prior to 1977. The model exhibits skill in reproducing numbers of age-1+ walleye pollock (Gadus chalcogrammus) and cod (Gadus macrocephalus); however, this is largely driven by the number of age-1 fish entering the model, which is set as an initial condition and tuned rather than modeled. The mortality of pollock and cod at older ages is higher compared to that of stock assessments. Introduction Within the NE Pacific, the eastern Bering Sea is of particular interest given the volume and worth of its fisheries (over a billion tons and a billion USD in product value from 1992-2012, Fissel et al., 2012), the large and numerous populations of seabirds and marine mammals (Friday et al., 2012; Allen and Angliss, 2013; Delinger 2006), and its vulnerability to climate change given its rapid response to atmospheric forcing (Stabeno 2007). This rapid response includes changes in seasonal ice coverage extent and duration (and hence variability in bottom temperature and perhaps biological productivity) (Hunt et al. 2011; Stabeno et al., 2012) as well as circulation (Stabeno et al., 2010; Danielson et al., 2012). Climate variability, and in particular climate change through global warming (IPCC 2007), has been recognized as one of the main challenges of sustainable fisheries in the future given its implications for marine resources abundance, distribution, and commercial catch (Brander, 2013; Salinger, 2013). Management approaches to these issues include single-species stock assessments as well as spatially-explicit ecosystem models of diverse complexity (Hollowed et al., 2013). End-to-end ecosystem models have proliferated in particular, given their inclusion of both human components and climate impacts, as well as their ability to include processes at multiple scales, albeit with their own set of challenges (Travers, et al. 2007; Rose et al., 2010). In this regard, downscaled earth systems models coupled to lower trophic level model of varying complexity are starting to become more common with some even including key fish groups (Travers et al 2009; Travers et al 2014a, 2014b; Kishi et al. 2011). End-to-end models have also been recognized as effective strategic tools and are considered essential to an ecosystem approach to fisheries management (EAFM) (Fulton, 2010; Fulton et al. 2014), and remain an active area of research, in particular with respect to (1) the evaluation of trade-offs between different actions and (2) direct linkages between climate variability and marine resources as mediated by oceanography and phytoplankton/zooplankton productivity. Diverse measures have been implemented as part of an EAFM for the Alaskan groundfish fisheries for well over a 15 years (Witherell et al., 2000) and several tools are in use for ecosystem assessment (Hollowed et al., 2013). Within the Eastern Bering Sea, the primary model that provides

23


an ecosystem context for the single species stock assessments is a mass balance model developed by Aydin et al. (2007). A selected suite of physical, biological and fisheries related ecosystem indicators provide the core information for an annual ecosystem report card and ecosystem assessment chapter (Zador, 2013) and a multiple model approach is used to simulate future ecosystem status (Livingston et al., 2005; Jurado-Molina et al. 2005). As part of the Bering Sea Project, a large scale, multi-disciplinary, and multi-institutional ecosystem research program, we developed the Bering 10K ROMS-NPZD-FEAST model with the dual goals of 1) investigating biophysical processes and climate impacts, as well as 2) aiding fisheries management by addressing both bottom-up and top-down forcing mechanisms on fish stocks and ecosystems (Wiese et al., 2012). The Bering 10K ROMS-NPZD-FEAST model represents the three dimensional dynamics of the two way interactions between physical oceanography and lower trophic levels; the Regional Oceanographic Model System (ROMS) provides information to the Nutrients Phytoplankton Zooplankton and Detritus (NPZD) such as currents, temperature, ice thickness and snow cover, the NPZD provides phytoplankton density to ROMS. The mode also has a two way feedback between NPZD and FEAST: the NPZD model provides prey fields for euphauisid and small and large copepods to the FEAST model and FEAST provides fish predation of zooplankton to the NPZD. The spatially explicit Fishery removals were included by sector, gear and species (Figure 5.1). The objective of this paper are to 1) describe how the components of the integrated model are linked, 2) discuss some of the processes and emergent properties of the model, and 3) validate some aspects pertinent to fisheries management resulting from a 40 year hindcast from 1971 to 2009. Methods The hindcast covers the years 1970 to 2009. The Common Ocean Reference Experiment reanalysis (CORE; Large and Yeager, 2008) was utilized as physical forcing and physical boundary conditions for the years 1969-2004 while the Climate Forecast System Reanalysis (CFSR; Saha et al., 2010) was utilized for the years 2003-2009. These two reanalyses were combined to have a continuous hindcast from 1979 to 2009 since CORE spans from 1950-2004 and CFSR spans 1979-present. Overlapping runs for 2003 and 2004 allowed a comparison of results using the two reanalyses; these were used to adjust CFSR for compatibility with CORE (Hermann et al. 2013). Oceanography ROMS-Bering10K is a coupled ocean-sea ice circulation model whose spatial grid is a subset of the ROMS-NEP5 model described and evaluated by Danielson et al. (2011), which itself builds on a model described by Curchitser et al. (2005). The model used a regular grid that has a spatial resolution of ~10 km and 10 vertical layers. The subgrid extends from the western Gulf of Alaska to the Russian coast and slightly past the Bering Strait (Figure 5.2). Danielson et al. (2011) showed ROMS-NEP5 closely reproduces ice cover and spring ice retreat onset. The Bering10K simulation includes modifications to the heat and salinity fluxes of NEP5, which were calibrated using extensive mooring data (Hermann et al. 2013); additional model-data comparisons for temperature and salinity have been conducted using the most recent iteration of the physical model (the same version as is used here) by Hermann et al. (this issue). Model outputs for depth-integrated temperature at two moorings (M2, and M4) run slightly warmer in winter (Hermann et al. 2013), though near surface temperatures can be colder than observed in the northern Bering Sea (Hermann et al, this issue). Model coupling includes feedback from the NPZ to ROMS Bering10K through phytoplankton density, which affects shortwave penetration (heat absorption) in the upper water

24


column. This and related modifications on the model described by Hermann et al. (2013) are detailed in Hermann et al. (this issue), where downscaled projections for circulation, temperature, salinity and NPZ elements are described. Lower Trophic levels The NPZ model was specifically designed to incorporate the ice dynamics of the Bering Sea, and includes nutrients (nitrate, ammonium, iron), ice algae, small and large phytoplankton, small copepods, oceanic/shelf large copepods (Neocalanus sp), oceanic/shelf krill (euphausiids, Order Euphausiacea), jellyfish (Class Scyphozoa), fast and slow sinking (pelagic) detritus, benthic detritus and benthic infauna (how about some examples of what this comprises) (Figure 5.3). Spatio-temporal dynamics are affected by ice thickness, temperature, salinity, solar radiation and circulation patterns provided by ROMS-Bering10K. The living compartments used as food supply for fish are the euphausiids, copepods and benthic infauna. The NPZ model has been described and tested by Gibson and Spitz (2011) as well as reviewed by a team of field biologists as part of a synthesis project funded by the National Science Foundation (Mordy and Lomas 2012). In particular, the euphausiids do not have vertical migration, nor feed on detritus and their biomass (as for all the zooplankton groups) goes very close to zero every winter. Fish The FEAST model aims at representing primarily the pelagic portion of the Eastern Bering Sea (EBS) food web as quantified by the mass balance food web model for the EBS. A thorough analysis of the full mass balance model, as well as sensitivity analysis is discussed in Aydin et al. (2007). FEAST is a multispecies bioenergetics model for forage and predatory fish species that simulates fish interaction via length-based and preferential prey predation. FEAST uses temperature from Bering 10K for the bioenergetics. FEAST is coupled to the NPZ model from which it gets daily estimated dry weight of euphausiids, small copepods, large copepods, and benthic infauna; in return, FEAST provides biomass in dry weight consumed by fish for each zooplankton species. This predation mortality is in addition to the quadratic mortality in the NPZ; in one-way feedback mode this term accounts for both predation and other sources. For each fish group the model keeps track of numbers of fish and condition factor (relative amount of non-length growth), and a fish group can be age- and length- structured (walleye pollock, Pacific cod and arrowtooth flounder), length structured only (Pacific herring, capelin, eulachon, Pacific sandlance, myctophids, squids, and salmon) or biomass pools with neither explicit age nor explicit length structure (crab, shrimp and epifauna). Age/length structured groups have 10 ages in addition to age zeroes, 14 length bins of 4 cm each for ages 1 to 10, and 2 cm length bins for age zeroes. The length-based groups have 20 length bins of 2 cm; biomass pools are assumed to have an average length. All symbols for formulas can be found in Table 5.1. The weight of all individuals of species sp of age a within a length bin l is calculated as:

( ) aspBlaspspL

lasp CFLAWspL ,,,,, ⋅= (1)

where spLA ,

spLB are length-weight conversions over all available data (long-term baseline) and

weight is wet weight (g). The biomass of an age and length class in a cell is thus:

25


lasplasplasp NWB ,,,,,, ⋅= (2)

The bionenergetics, growth and prey selection have been documented in detail by Aydin et al. (in prep). Temporal resolution and spatial distribution The "biological and forcing" timestep of FEAST is daily, although it is integrated on a finer time scale. Thus, growth increments, etc. are based on "daily" inputs which average sub-daily behavior and external forcing functions (e.g. fishing) which are updated daily, or coarser. Although the fish can move throughout the total grid, the portion of interest and seeded with fish for the initial conditions is restricted to shelf/slope processes with a depth cutoff of 200m for the shelf, 3500m for the slope. The northern shelf boundary corresponds to the U.S. Exclusive Economic Zone, and the furthest southwest (Aleutian) point corresponds to 172°W along the Aleutian Archipelago. Movement The fish in FEAST are modeled using explicit horizontal movement and no vertical movement. Each fish species is assumed to be in a single, fixed vertical layer; therefore, there is no vertical movement tracked for fish. Fish swim speed is based on length and a diffusion rate inversely proportional to local habitat quality along a gradient. Direction of movement is based on habitat quality measured as a function of net population biomass gain rate and predation mortality. So fish in a high quality habitat swim in circles and diffuse slower than fish in poor quality habitat which swim longer distances in straighter lines. Briefly, the horizontal movement function is described as movement of a single numerical density (N) in a single spatial dimension x over time t such that movement is in response to local conditions (λ) and diffusion D:

∂∂

⋅∂∂

+

∂∂

⋅∂∂

=

∂∂

xND

xxx

xmove

tN λ

(3)

Fishing effort allocation The fishing effort allocation for the hindcast is based on historical sector/gear/species catch data downscaled to weekly removals by Alaska Department of Fish and Game STAT6 statistical areas. Removals in each STAT6 area are further downscaled to the FEAST grid by allocating removals proportional to biomass in each grid cell so that:

𝐹𝐹 𝑟𝑟𝑒𝑒𝑞𝑞𝑟𝑟𝑟𝑟𝑟𝑟𝑒𝑒𝑑𝑑𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 = ℎ𝑟𝑟𝑖𝑖𝑖𝑖𝑙𝑙𝑟𝑟𝑟𝑟𝑖𝑖𝑖𝑖𝑙𝑙 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖ℎ𝑖𝑖,𝑗𝑗,𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 𝐷𝐷𝑃𝑃𝐶𝐶𝐹𝐹𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖⁄ (4)

where:

𝐷𝐷𝑃𝑃𝐶𝐶𝐹𝐹𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 = 𝑞𝑞𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 ∗ 𝑁𝑁𝑠𝑠𝑝𝑝,𝑖𝑖 ∗ 𝑊𝑊𝑠𝑠𝑝𝑝,𝑖𝑖 (5)

To avoid removals exceed the amount of fish available at any given location:

𝐹𝐹𝑖𝑖𝑖𝑖𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 = min (0.99,𝐹𝐹 𝑟𝑟𝑒𝑒𝑞𝑞𝑟𝑟𝑟𝑟𝑟𝑟𝑒𝑒𝑑𝑑𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖) (6)

The fisheries include: Catcher-processor for pollock trawl, Pacific trawl, pots and longline, other species trawl, pots and longline; catcher vessels pollock trawl, Pacific cod trawl, pots and hook and

26


line, other species trawl, pots, hook and line; herring gillnets and seine. Initial conditions, and field data FEAST needs starting conditions for state variables Nsp,l (Numbers at length for each species); the condition factor was assumed to be one (CFsp,l). Initial conditions for the fish were derived from the historical database of the RACE Bottom Trawl summer survey (BTS) conducted by the Alaska Fisheries Science Center and stock assessment estimates in the case of pollock, cod and arrowtooth. Three types of spatial distribution at length were derived from the BTS: one for warm years (1998, 2001 – 2005), one for cold years (1995, 1997, 1999, 2007 – 2010) and one for neutral years (1996, 2000, 2006) (year classification follows Stabeno et al. 2012); all years prior to 1995 are considered neutral years. For the initial conditions, we allocated the number of fish estimated for 1971 by the stock assessment using the spatial distribution of a neutral year. Data used to estimate initial conditions for fish are described in Table 5.2.

(8) 𝑁𝑁𝑖𝑖,𝑗𝑗,𝑝𝑝,𝑖𝑖,𝑖𝑖,𝑠𝑠𝑝𝑝,𝑛𝑛𝑒𝑒𝑟𝑟𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙 = �∑𝑁𝑁𝑛𝑛𝑒𝑒𝑟𝑟𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙𝑖𝑖𝑟𝑟𝑟𝑟𝑁𝑁𝑒𝑒𝑁𝑁𝑖𝑖,𝑗𝑗,𝑝𝑝,𝑖𝑖,𝑠𝑠𝑝𝑝 ∑𝑁𝑁𝑛𝑛𝑒𝑒𝑟𝑟𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙𝑖𝑖𝑟𝑟𝑟𝑟𝑁𝑁𝑒𝑒𝑁𝑁𝑝𝑝,𝑖𝑖,𝑠𝑠𝑝𝑝⁄ � ∗ 𝑁𝑁𝑖𝑖𝑖𝑖𝑙𝑙𝑖𝑖𝑁𝑁𝑖𝑖,𝑖𝑖,𝑝𝑝,𝑠𝑠𝑝𝑝 Model run The ROMS-Bering10K and NPZ were initialized using time specific conditions from the hindcast by Hermann et al. (2013) which uses the same model parametrizations. We started the model in July 1, 1970 and ran a simulation with fish movement but no mortality through December 31, 1970 (spin-up). Starting January 1, 1971 the fish mortality was turned on for the remainder of the simulation (Jan 1, 1971 - December 30, 2009). A forcing file containing daily catches by sector, gear, species and length for each grid cell supplies the catch data for the calculation of the fishing effort. A second forcing file supplies the estimated age 1 recruits from the (EBS area-integrated) stock assessment for pollock, cod and arrowtooth flounder. At the beginning of the year, the total number of age zero fish are corrected to that of the stock assessment estimate, while preserving the spatial distribution of the model. We computed spatial averages of weekly model output for phytoplankton and zooplankton biomass per meter squared integrated over the top 300m (or the total depth of the water column, whichever is shallower); the averages were based on the standard BSIERP regions, which were chosen to minimize within-region variance and to maximize variance across regions (Ortiz et al, 2012) (Fig. 5). Weekly \anomalies for these regions were also computed for the euphausiid biomass in the top 300m as well as for the depth-averaged temperature in the top 300 m. Validation and model-data comparisons of fish bioenergetics and movement are presented by Aydin et al. and Ortiz et al. (in review) respectively. Here we show model-data comparisons for the timeseries of numbers at age available from the stock assessment for pollock and Pacific cod (NPFMC 2014). The timeseries of the modeled numbers at age were created by extracting the model output corresponding to regions 1 through 16, the assumed distribution of the fish stock for walleye pollock and primary distribution for the Pacific cod. Results We present results from the hindcast in two sections: the first explores spatio-temporal patterns for phytoplankton and zooplankton, across regions; the second refers to FEAST validation on aspects relevant to management.

27


Mean weekly patterns by region We focus here on the timing as opposed to the magnitude of spring blooms. The spring blooms start off in the eastern Bering Sea with large phytoplankton in the southern and off-shelf regions, with timing of the bloom in the inner northern shelf regions lagging by up to six weeks. The blooms typically last for ~2 months. Small phytoplankton begin to bloom several weeks after large phytoplankton, however small and large phytoplankton peaks occur closer together in northern regions than in southern regions. Both large and small phytoplankton appear to fuel large microzooplankton (Figure 5.6). Since the large and small phytoplankton bloom are typically consecutive events, regionally coherent patterns emerge despite the interannual variability in the magnitude of the peak biomasses. Figure 5.7 shows the weekly time series for years 1971-2009 for the BSIERP regions. Figure 5.8 shows the sum of small and large phytoplankton in the water column from the model output for regions 3, 6, 9 and 10 compared to chlorophyll a measurements at surface and medum depths at moorings 2, 4, 5 and 8 respectively. The timing of the spring and fall bloom appears to be somewhat delayed in the model, although the actual lag varies across years and location. However, these measurements do not include subsurface blooms except for M2. The spatial coherence and timing sequence is likewise evident in the 1971-2009 mean weekly biomass for the small copepods and euphausiids. The biomass of small copepods bloom occurs when?, shortly followed by an increase in the biomass of euphausiids (Figures 5.9 and 5.10). As with phytoplankton, these increases occur later in the northernmost and inner shelf regions. The earlier euphausiid increase from the model output in the outer shelf/ off-shelf area resembles that of Thysanoessa inermis, likewise, the later bloom in the middle shelf would be akin to spawning of T. raschii in this area. The fall increase in the biomass of small copepods and euphausiids in the model output seems to be fueled, at least partially, by the fall phytoplankton bloom; however we did not quantify the relative contribution of microzooplankton vs phytoplankton to the biomass increase of either small copepods or euphausiids. The increase in both phytoplankton and zooplankton biomass in the fall is in particular evidence in this model, is a subject of speculation and very little direct observational data (FOCI pers. comm.). Euphausiid data compiled by Hunt et al. (in review) for the eastern Bering Sea indicates late summer spawning has indeed been observed for T. raschii (middle shelf) and while interannual variability may account for occasional late spawning, it seems unlikely the spawning season could be extended into mid-late fall. However, 30 years of walleye pollock diet data indicates that while copepod consumption by pollock is limited to the spring and summer, euphausiid consumption is important year-round with a possible peak in late fall as well as in the spring (Figure 5.11, see Buckley et al (this issue) for a description of the data source). Hence, while fall blooms of euphausiids are possible, the relative magnitude of the ones observed in the model output may overestimate true values. Given the importance of this time period for determining pollock survival in the model, direct measurements of fall zooplankton would be quite desirable for improving this model. Results from a hindcast simulated with a previous version of the ROMS-Bering10K-NPZ, (Hermann et al. 2013) found model output of large crustaceans (Neocalanus and euphausiids) tend to covary inversely to temperature on the outer shelf, which Hunt et al. (2011) and Coyle et al. (2011) have also observed in the field. This was explored further by comparing the euphausiid biomass and temperature anomalies by region from 1971 to 2009 (Figure 5.12). The tendency appears to be stronger in more recent years during the string of warm and cold years, but is particularly evident in off-shelf/slope areas (regions 15 and 16). In contrast, years prior to 1977 show a weaker tendency and Hermann et al (2013) reported a positive correlation between

28


temperature and large crustacean zooplankton on the inner and northern shelf. The strength of this tendency may be linked to the dominant water mass, and with flow reversals due to wind currents in the northern shelf (Danielson, et al, 2012), the dominant water masses shift between Anadyr and Bering shelf/Strait waters, disrupting what would be a multi-year monotonic trend in other areas. Validation of number of fish age 1+ in model output for pollock and cod Figure 5.13 shows the compared timeseries from model output as well as the stock assessment estimates (NPFMC 2014). The stock assessment age structure includes ages over 10 years old; in the model output, age 10 is considered a plus group (that is, it aggregated ages 10 and older, albeit the length number of length bins remained at 14, the same as other single year ages). Model output closely follows the stock assessment estimates for pollock an dcod; the lower model estimates for number of fish age 1+ is due to higher mortality in the older ages than that in the stock assessment. The majority of the pollock age 1are distributed in northern areas throughout the years, particularly during warm years (Figure 5.14). This coincides with what been observed in the field during BASIS and the annual bottom trawl summer survey (Alex Andrews, pers. Comm. NMFS, NOAA). Hindcast results show individuals enter into Russian waters (not shown), and it is well known that Alaskan and Russian pollock mix. The expansion into inner waters, towards the Norton Sound and St Lawrence was also observed in BASIS surveys. In the hindcast, the northern shelf area also saw higher proportions of age 1 pollock between 1977 and 1987, decreasing towards 1991 and staying low throughout 2003-2004, unfortunately survey data with the geographic coverage needed to corroborate whether this was indeed the case is not available. Model output for the cold years 2006-2008 in the northern regions 12 and 13 actually show positive anomalies for euphausiids biomass and temperature at a time when both of these were actually negative for most other regions. This might partially have driven the movement of age 1 pollock towards these areas, the higher proportion of age 1 pollock in 2004 in the northern regions is likely related to the late warming of those regions compared to the rest. Discussion Understanding the spatiotemporal succession of zooplankton blooms and their interannual variability as flows of available prey and energy is important to understanding how and when these resources become available to upper trophic levels. Timing in particular is not only key to young fish, but also to migratory birds and whales. The succession of resources –or lack thereof- contributes to the success or failure of particular age classes or reproductive seasons. While overall trends in primary and secondary production are indeed relevant as an indicator of maximum energy available for transfer in the system, it may be that pockets of high prey abundances or more suitable temperatures provide a spatial refuge for forage and upper trophic levels. Buckley et al., (in review) suggests pollock might feed more than previously thought on copepods, and that pollock may feed on euphausiids primarily when copepods are not available. If so, the succession of zooplankton blooms and length of copepod vs euphausiid blooms might be more influential in driving pollock survival and/or distribution, than considering their abundance only. Because of the implications climate variability may have on primary and secondary production, it becomes then even more important to isolate driving factors prone to change –such as temperature, ice extent and duration, winds and currents- from those that will remain stable such as length of day and tidal currents.

29


The spatiotemporal patterns of phytoplankton, zooplankton and age 1 pollock are an emergent property of the model. The difference in the expected magnitude of euphausiid fall blooms can be ascribed to the variety of life history strategies that had to be simplified in this iteration of the model. One of the most beneficial aspects of the hindcast has been the synthesis it has prompted and how it has guided research to bridge current gaps, both in design and temporal focus. Already improvements to the models used here have a tighter link to benthic energy, and forecasts have shown potential changes in habitat (cold pool) as perceived by groundfish (Hermann et al., this issue). The ability to run simulations of integrated ecosystem models have proven a necessary tool to elucidate potential costs and benefits, as well as future challenges (Fulton, 2011; 2014) In the face of all the uncertainty, simulations such as these, tightly coupled to field programs, will become more common as testbeds for process exploration and management evaluation, increasing their relevance for future management and strategic planning. Acknowledgements Ivonne Ortiz and Al Hermann were partially supported by the North Pacific Research Board. This is NPRB publication number XXXX. This is BEST-BSIERP Bering Sea Project publication number XXXX

30


References Allen, B.M. and Angliss, R.P. Alaska Marine Mammal Stock Assessmen , 2012. U.S. Dept.

Commer. NOAA Technical Memorandum NMFS-AFSC-245. 282 p. downloaded 05/2013 http://www.nmfs.noaa.gov/pr/sars/pdf/ak2012.pdf.

Andrews, A.G., Heintz, R.A., Farley, E.V. in review. Age-1 walleye pollock in the eastern Bering Sea: distribution, abundance, diet, and energy density.

Aydin, K., Ortiz, I., Hermann, A.J., In review. New approach to modeling fish bioenergetics, a case study for walleye pollock, Pacific cod and arrowtooth flounder in the Eastern Bering Sea. Deep Sea Res. II 00, 00–00.

Aydin, K; Gaichas, S; Ortiz, I; Kinzey, D; Friday, N. 2007. Comparison of the Bering Sea, Gulf of Alaska, and Aleutian Islands Large Marine Ecosystems through Food Web Modeling. NOAA Technical Memorandum NOAATMNMFSAFSC178.

Brander, K. 2013. Climate and current anthropogenic impacts on fisheries. Climatic Change. 119: 9-21

Coyle, K.O., Eisner,L.B., Mueter,F.J., Pinchuk,A.I., Janout,M.A., Cieciel,K.D., Farley,E. V., Andrews, A.G. 2011. Climate change in the southeastern Bering Sea: impacts on pollock stocks and implications for the oscillating control hypothesis. Fish. Ocean 20(2), 139–

156. Curchitser, E. N., D. B. Haidvogel, A. J. Hermann, E. L. Dobbins, T. M. Powell, and A.

Kaplan.2005. Multi‐scale modeling of the North Pacific Ocean: Assessment and analysis of simulated basin‐scale variability (1996–2003), Journal of Geophysical Research. , 110, C11021, doi:10.1029/2005JC002902.

Danielson, S., E. Curchitser, K. Hedstrom, T. Weingartner, and P. Stabeno. 2011. On ocean and sea ice modes of variability in the Bering Sea. Journal of Geophysical Research, 116, C12034, doi:10.1029/2011JC007389.

Danielson, S., Hedstrom, K., Aagaard, K., Weingartner, T., Curchitser, E. 2012. Wind-induced reorganization of the Bering shelf circulation. Geophysical Research Letters 39 L08601 doi: 10.1029/2012GL051231.

Denlinger, L.M. 2006. Alaska Seabird Information Series. Unpubl. Rept., U.S. Fish and Wildl. Serv., Migr. Bird Manage. Nongame Program, Anchorage, AK. Downloaded 05/2013 http://www.fws.gov/alaska/mbsp/mbm/seabirds/pdf/asis_complete.pdf. Friday, N.A., Waite, J.M., Zerbini, A.N., and Moore, S.E. 2012. Cetacean distribution and abundance in relation to oceanographic domains on the eastern Bering Sea shelf: 1999-2004. Deep Sea Research II 65-70:260-272.

Fulton E.A., Smith, A,D.M, Smith, D.C., Johnson, P. (2014) An Integrated Approach Is Needed for Ecosystem Based Fisheries Management: Insights from Ecosystem-Level Management Strategy Evaluation. PloS ONE 9(1): e84242. Doi:10.1371/journal.pone.0084242

Fulton, E A. 2011. Interesting times: winners, losers, and system shifts under climate change around Australia ICES Journal of Marine Science 68: 1329-1342.

Fulton, E.A. 2010. Approaches to end-to-end ecosystem models. Journal of Marine Systems, 81:171-183.

Gibson, G.A. and Y.H. Spitz. 2011. Impacts of biological parameterization, initial conditions, and environmental forcing on parameter sensitivity and uncertainty in a marine ecosystem model for the Bering Sea. Journal of Marine Systems 88: 214-231.

Hermann, A.J., Gibson, G.A., Bond, N.A., Curchitser, E.N., Hedstrom, K., Cheng, W., Wang, M., and Stabeno, P.J. in review. Projected futures biophysical states of the Bering Sea. Deep Sea Reasearch II 00:00-00.

Hermann, A. J., Gibson, G.A., Bond, N.A., Curchitser, E.N., Hedstrom, K., Cheng, W., Wang, M., Stabeno, P.J., Eisner, L. and Cieciel, K.D. 2013. A multivariate analysis of observed and modeled biophysical variability on the Bering Sea shelf: multidecadal hindcasts (1970-

31

http://www.nmfs.noaa.gov/pr/sars/pdf/ak2012.pdf


2009) and forecasts (2010-2040). Deep Sea Research II, doi:10.1016/j.dsr2.2013.04.007. Herring stock assessment ADFW Hollowed, Anne Babcock; Curchitser, Enrique N; Stock, Charles A; Zhang, Chang Ik. 2013. Trade-

offs associated with different modeling approaches for assessment of fish and shellfish responses to climate change. Climatic Change 119: 111-129.

Hollowed, Anne B; Aydin, Kerim Y; Essington, Timothy E; Ianelli, James N; Megrey, Bernard A; André E. Punt and Anthony D.M. Smith. 2011. Experience with quantitative ecosystem assessment tools in the northeast Pacific Fish and Fisheries 12: 189-208

Hunt, G.L., Ressler, P., De Robertis, A., Aydin, K., Gibson, G., Sigler, M., Ortiz, I., Lessard, E., Williams, B., Pinchuk, A., and Buckley, T. In review. Deep Sea Res. II 00:00–00

Hunt, G.L., Coyle, K.O., Eisner, L.B., Edward V., Heintz, R.A., Mueter, F., Napp, J.M., Overland, J.E., Ressler, P.H., Salo, S., Stabeno, P.J. 2011. Climate impacts on eastern Bering Sea foodwebs: a synthesis of new data and an assessment of the Oscillating Control Hypothesis. ICES Journal of Marine Science 68: 1230-1243.

IPCC (2007) Climate change 2007: working group I: the physical science basis. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL (eds.) Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change, 2007. Cambridge University Press, Cambridge and New York.

Jurado-Molina, J. Livingston, P.A., and Ianelli, J.N. 2005. Incorporating predation interactions in a statistical catch-at-age model for predator-prey system in the Bering Sea. Canadian Journal of Aqautic and Fisheries Science 62:1865-1873.

Kishi, M.J, Ito, S., Megrey, B.A., Rose, K.A., Werner, F.E. 2011. A review of the NEMURO and NEMURO.FISH models and their application to marine ecosystem investigations. Journal of Oceanography 67: 3-16.

Large and Yeager, 2008. The global climatology of an interannually varying air-sea 1009 flux data set, Clim. Dyn. 33, 341-364

Livingston, P.A., Aydin, K., Boldt, J., Ianelli, J., Jurado-Molina, J. 2005. A framework for ecosystem impacts assessment using an indicator approach. ICES Journal of Marine Science, 62:592-597.

North Pacific Fishery Management Council. 2009. Stock assessment and fishery evaluation report for the groundfish fisheries of the Gulf of Alaska and Bering Sea/Aleutian Islands area: Economic status of the Groundfish fisheries off Alaska, 2012. North Pacific Fishery Management Council 605 West 4th Ave., Suite 306 Anchorage, AK 99501

North Pacific Fishery Management Council. 2014. Stock assessment and fishery evaluation report for the groundfish fisheries of the Gulf of Alaska and Bering Sea/Aleutian Islands area: Economic status of the Groundfish fisheries off Alaska, 2012. North Pacific Fishery Management Council 605 West 4th Ave., Suite 306 Anchorage, AK 99501.

Ortiz, I., Aydin, K., Hermann, A.J., In review-a. Modeling fish movement, a case study of walleye pollock, Pacific cod and arroetooth flounder in the eastern Bering Sea. Deep Sea Res. II 00, 00–00.

Ortiz, I., F. K. Wiese and A. Grieg. 2012. Marine Regions Boundary Data for the Bering Sea Shelf and Slope. UCAR/NCAR-Earth Observing Laboratory/Computing, Data, and Software Facility. Dataset. http://dx.doi.org/10.5065/D6DF6P6C

Page LM, Espinosa-Pérez H, Findley LT, Gilbert CR and others (2013) Common and scientific names of fishes from the United States, Canada, and Mexico, 7th edn. Spec Publ Am Fish Soc No. 34, Bethesda, MD

Rose, K.A., Allen, J.I., Artioli, Y., Barange, M., Blackford, J., Carlotti, F., Cropp, R., Daewell, U., Edwards, K., Flynn, K., Hill, S,L., HilleRisLambers, R., Huse, G., Mackinson, S., Megrey, B., Moll, A,. Rivkin, R., Salihoglu, B., Schrum, C., Shannon, L., Shin, Y., Smith, S.L., Smith, C., Solidoro, C., St. John, M., Zhou, M.. 2010. End-to-End modeling for the analysis of marine ecosystems: challenges, issues and next steps. Mar. Coast. Fish. 2, 115–130.

32

http://dx.doi.org/10.5065/D6DF6P6C


Saha S and coauthors, 2010. The NCEP Climate Forecast System Reanalysis. Bull. Amer. 873 Meteor. Soc. 91, 1015.1057. doi: 10.1175/2010BAMS3001.1

Salinger, M.J. 2013. A brief introduction to the issue of climate and marine fisheries. Climatic Change. 119: 23-35.

Sigler, M.F., Stabeno, P.J., Eisner, L.B., Napp, J.M., Mueter, F.J. in press. Spring and fall phytoplankton blooms in a productive subarctic ecosystem, the eastern Bering Sea, during 1995-2011. Deep Sea Research II.

Stabeno, P J; Bond, NA; Salo, SA. 2007. On the recent warming of the southeastern Bering Sea shelf. Deep Sea Research II 54: 2599-2618.

Stabeno, P.J., Kachel, N., Moore, S.E., Napp, J., Sigler,M., Zerbini, A. 2012 Comparison of warm and cold years on the southeastern Bering Sea shelf and some implications for the ecosystem. Deep Sea Research II. 65-70: 31-45.

Stabeno, P., Napp, J., Mordy, C., Whitledge, T. 2010. Factors influencing physical structure and lower trophic levels of the eastern Bering Sea shelf in 2005: Sea ice, tides and winds. Progress in Oceanography 85: 180-196.

Travers-Trolet, M; Shin, Y-J; Field, JG. 2014a. An end-to-end coupled model ROMS-N2P2Z2D2-OSMOSE of the southern Benguela foodweb: 33umerical33zation, calibration and pattern-oriente validation. African Journal of Marine Science 36: 11-29.

Travers-Trolet, Morgane; Shin, Yunne-Jai; Shannon, Lynne J; Moloney, Coleen L; Field, John G. 2014b. Combined Fishing and Climate Forcing in the Southern Benguela Upwelling Ecosystem: An End-to-End Modelling Approach Reveals Dampened Effects:. PloS One 9. E94286.

Travers, M; Shin, Y J; Jennings, S; Machu, E; Huggett, JA; Field, J.G. and P.M. Cury. 2009. Two-way coupling versus one-way forcing of plankton and fish models to predict ecosystem changes in the Benguela. Ecological Modelling 220: 3089-3099.

Travers, M., Shin, Y.J.; Jennings, S., Cury, P. 2007. Towards end-to-end models for investigating the effects of climate and fishing in marine ecosystems Progress in Oceanography 75: 751-770.

Wiese, F. K., W. J. Wiseman, Jr., and T. I. Van Pelt. 2012. Bering Sea Linkages. Deep Sea Research II 65-70: 2-5.

Witherell, D., C. P. Pautzke, and D. Fluharty. 2000. An ecosystem-based approach for Alaska groundfish fisheries. ICES Journal of Marine Science 57:771-777

Zador, S. (Ed). 2013. Ecosystem consideration 2013. North Pacific Fishery Management Council 605 West 4th Ave., Suite 306 Anchorage, AK 99501. 235p.

33


Table 5.1. List of symbols from equations.

Eq 1, 2 laspW ,, wet weight (in grams) of an individual of species sp in age class a

within length bin l spLA

parameter A (factor) of length weight conversion for species sp (longterm baseline using all available data)

laspL ,, length (in centimeters) of species sp in age class a, length bin l

spLB

parameter B (exponent) of length weight conversion for species sp (longterm baseline using all available data)

aspFC , condition factor (proportion of total weight growth allocated to non-

length growth) laspB ,, biomass of species sp in age class a for length bin l

(weightsp,a,,l*numberssp,a,,l) laspN ,, numbers at age a of length l for species sp (gridcell specific)

Eq 3

movet

N lasp

∂

∂ ,,

change through movement of a single 34umerical density through time

∂∂

⋅∂∂

xx

xλ change in local conditions λ through spatial dimension x

∂∂

⋅∂∂

xND

x change in number of individuals N through spatial dimension x due

to diffusion D Eq 4, 5, 6

𝐹𝐹 𝑟𝑟𝑒𝑒𝑞𝑞𝑟𝑟𝑟𝑟𝑟𝑟𝑒𝑒𝑑𝑑𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 Fishing rate required by gear gr species sp and length l ℎ𝑟𝑟𝑖𝑖𝑖𝑖𝑙𝑙𝑟𝑟𝑟𝑟𝑖𝑖𝑖𝑖𝑙𝑙 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖ℎ𝑖𝑖,𝑗𝑗,𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 Fisheries removal in cell I,j by gear gr, species sp, length l

𝐷𝐷𝑃𝑃𝐶𝐶𝐹𝐹𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 Catch per unit fished by gear gr, species sp, and length l 𝑞𝑞𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 Catchability by gear gr of species sp at length l 𝑁𝑁𝑠𝑠𝑝𝑝,𝑖𝑖 Number of fish of species sp at length l 𝑊𝑊𝑠𝑠𝑝𝑝,𝑖𝑖 Individual weight of fish species sp at length l

𝐹𝐹𝑖𝑖𝑖𝑖𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙𝑎𝑎𝑝𝑝,𝑠𝑠𝑝𝑝,𝑖𝑖 Fishing rate actually used to avoid negative removals by gear gr, species sp and length l

Eq 8 𝑁𝑁𝑖𝑖,𝑗𝑗,𝑝𝑝,𝑖𝑖,𝑖𝑖,𝑠𝑠𝑝𝑝,𝑛𝑛𝑒𝑒𝑟𝑟𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙 Number of fish in a neutral year at cell i,j, in year y, of age a, length

l and species sp 𝑁𝑁𝑛𝑛𝑒𝑒𝑟𝑟𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙𝑖𝑖𝑟𝑟𝑟𝑟𝑁𝑁𝑒𝑒𝑁𝑁𝑖𝑖,𝑗𝑗,𝑝𝑝,𝑖𝑖,𝑠𝑠𝑝𝑝, Number of fish in a neutral survey years (1996, 2000, 2006) at cell

i,j, of length l and species sp 𝑁𝑁𝑛𝑛𝑒𝑒𝑟𝑟𝑖𝑖𝑟𝑟𝑖𝑖𝑙𝑙𝑖𝑖𝑟𝑟𝑟𝑟𝑁𝑁𝑒𝑒𝑁𝑁,𝑝𝑝,𝑖𝑖,𝑖𝑖,𝑠𝑠𝑝𝑝, Number of fish in a neutral survey years (1996, 2000, 2006) of

length l and species sp 𝑁𝑁𝑖𝑖𝑖𝑖𝑙𝑙𝑖𝑖𝑁𝑁𝑖𝑖,𝑖𝑖,𝑝𝑝,𝑠𝑠𝑝𝑝 Number of fish estimated by stock assessment at age a length l, in

year y, for species sp.

34


Table 5.2. Data used for fish initial conditions Group Numbers/ Biomass Spatial

distribution Length

Pollock SAFE estimate for 1971 (NPFMC 2009) plus 2% assumed to inhabit the Northern Bering Sea (based on survey)

RACE neutral year

Length at age data from RACE and BASIS surveys

Cod SAFE estimate for 1971 (NPFMC 2009)

RACE neutral year

Length at age data from RACE survey

Arrowtooth flounder

SAFE estimate for 1982 (NPFMC 2009) minus 17% assumed to inhabit outside the Bering sea shelf and slope.

RACE neutral year

Length at age data from RACE survey

Herring SAFE 1971 RACE 1982 Length at age data from RACE survey

Capelin Survey for 1982*q from Ecopath RACE 1982 Length data from RACE survey

Eulachon Survey for 1982*q from Ecopath RACE 1982 Length data from RACE survey

Sandlance Survey for 1982*q from Ecopath RACE 1982 Length data from RACE survey

Myctophids Ecopath estimate converted to no. of fish

RACE 1982 and along slope

Weight at length data from RACE survey

Squids Ecopath biomass estimate RACE 1982 and along slope

RACE 1982 and along slope

Shrimp Survey for 1982*q from Ecopath RACE 1982 Length data from RACE survey

Crab Survey for 1982*q from Ecopath RACE 1982 Length data from RACE survey

Epifauna Survey for 1982*q from Ecopath RACE 1982 Length data from RACE survey

35


Figure 5.1. Data flow and feedbacks across the components of the Bering 10K ROMS-NPZD-FEAST model for the Bering Sea.

Figure 5.2. Geographical extent of ROMS-Bering10K is delimited by the yellow rectangle

36


extending from Russia to the Gulf of Alaska and from south of the Aleutian Chain to slightly past the Bering Strait.

Figure 5.3. Nutrient-Phytoplankton-Zooplankton-Detritus model for the Bering Sea, (from Gibson and Spitz, 2011).

37


Figure 5.4. Trophic structure and coupled processes represented by the integrated ecosystem model for the eastern Bering Sea.

38


Figure 5.5. BSIERP regions used for spatial averaging of results.

39


Figure 5.6. Mean weekly biomass of phytoplankton and microzoplankton by region.

40


outer/off-shelf middle shelf middle shelf inner shelf

S

outh

N

orth

Figure 5.7. Mean weekly biomass time series from 1971-2009 for large phytoplankton (green), small phytoplankton (blue) and large microzooplankton (brown) by region.

41


Figure 5.8. Surface and medium depth Chlorophyll measurements (Chlas, Chlam) at moorings M2, M4, M5 and M8 and FEAST model output for the sum of small and large phytoplankton (PhLS) in the water column for the corresponding regions 3, 6, 9 and 10. Cholorophyll data courtesy of Michael Sigler (see Sigler et al., in press).

42


Figure 5.9. Small copepod and euphausiid weekly mean biomass from 1971-2009 hindcast, shown by BSIERP region 1-16.

43


outer/off-shelf middle shelf middle shelf inner shelf

S

outh

N

orth

Figure 5.10. Mean weekly biomass time series from 1971-2009 for small copepods (brown), euphausiids (green), large shelf copepods (blue) and large oceanic copepods (purple) by region.

44


Figure 5.11. 95% confidence limits for frequency of occurrence of euphausiids (gray shaded area) and copepods (dotted lines) in pollock stomachs, by week of year for BSIERP domains 1, 3, 4, and 8 from 30 years of pollock diet data collected on surveys and by observers.

45


1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Figure 5.12. Euphausiid biomass anomaly (bars) and depth-averaged temperature anomaly (red line) by BSIERP region (1-16 top to bottom) from 1971-2009 hindcast.

46


Figure 5.13. Compared timeseries of number of fish age 1+ from 1971-2009 model output (“FEAST”) and stock assessment (“SAFE”) estimates for pollock (top panel) and cod (bottom panel) (NPFMC 2014).

47


Figure 5.14. Propotion of age 1 pollock by BSIERP region (1-16). Regions in the northern shelf are shown in color (9-14), all others are shown in grey (1-8, 15-16).

48


Chapter 6. Use of FEAST as an operating model for Management Strategy Evaluation

This chapter is taken from the final report for B.73 Management Strategy Evaluation where it appeared as chapter 3. The key steps associated with using FEAST as operating model are to: (a) identify the stock assessment methods and associated harvest control rules, (b) specify how the data used by the stock assessment methods are generated by the operating model, (c) specify the specific scenarios (combinations of stock assessment methods, harvest control rules and specifications for the operating model) to be tested, and (d) specify the performance measures used to summarize estimation and management performance. Stock assessment methods Three classes of stock assessment method were identified as potential estimation methods and as the basis for providing the input for control rules. These include the single species methods which are used in practice by NMFS and the North Pacific Fishery Management Council along with alternatives devised which consider trophic interactions (MSMt and Ecosim), Single-species assessments Single species assessments are currently used by the AFSC to provide management advice for Eastern Bering Sea (EBS) walleye pollock (e.g., Ianelli et al., 2012), Pacific cod (e.g. Thompson and Lauth, 2012), and arrowtooth flounder (e.g., Spies et al., 2012). The assessments for these stocks are based on software developed specifically for those stocks coded using AD Model Builder (Fournier et al., 2012). To streamline the assessment process for Pacific cod, data inputs based on the Stock Synthesis framework (Methot and Wetzel, 2013) were adopted to retain the essential model aspects (i.e., fit to length frequency data from fisheries and surveys) but avoid some of the complexity in generating model input and converting that to ABC values. The assessments were recompiled and transferred to the high performance computer cluster located at AFSC. All single species assessments have the following features in common:

1) They are fundamentally age-structured and use an annual time step 2) Estimates of annual fishing mortality rates are conditioned on the total catch (retained and

discards) estimates (i.e., an annual term fits the observed catch biomass precisely) 3) Fishery data (catch biomass and catch proportions at age) are aggregated over seasons and

areas within each year 4) Proportions at age from surveys and fisheries are fitted using estimated (or assumed)

multinomial sample sizes 5) Survey indices (abundance or biomass) are modeled using lognormal assumptions and

annually-specified observation errors (variances)

Other specific characteristics for each species follows. Eastern Bering Sea pollock The assessment of EBS Pollock considers the period 1964-present. This assessment is formulated as a Bayesian assessment, with priors on all key parameters. The population dynamics model on which the assessment is based on sex-aggregated, and the fisheries for EBS pollock are combined into a single fleet with allowance made for changes over time in fishing practices by modelling

49


fishery selectivity as time-varying. The model is fitted to fishery catch-at-age data (1979-present), index and the age-composition data from the shelf survey (1982-present), index and age-composition data from the acoustic-trawl survey (1979-present), and an index derived from acoustic data collected opportunistically on bottom-trawl survey vessels. The selectivity pattern for the shelf survey is asymptotic with time-varying slope and inflection points. The numbers of age-1 animals in the acoustic trawl survey are treated separately from the data from the remaining ages. The likelihood function for the age-composition data is taken to be the robust normal for proportions distribution of Fournier et al. (1990) while the likelihood functions for the index and catch data are assumed to be lognormal (Ianelli et al., 2012). Pacific cod The Pacific cod model commences in 1977 and is a simplified version of the stock synthesis configuration used in annual assessment process. This was required because of complexities associated with a) the growth model specification, b) converting Stock Synthesis output into a form used for model projections and harvest control rule specifications, and run/estimation time. A fixed growth model was assumed (to fit observed length frequency via a conversion matrix) and was specified to be the same as used in the 2012 assessment model (Thompson et al., 2012). Arrowtooth flounder The arrowtooth flounder model tracks sex-specific data on length frequencies from survey and fishery observations. Where available, age compositions replace observed length frequency data. Survey data from the eastern Bering Sea shelf are used from 1982 onwards, whereas intermittent data from the deeper slope region (NMFS “Bering Sea slope” trawl survey) and the Aleutian Islands trawl survey are also used (Spies et al,. 2012). Multi-species stock assessment model with temperature Temperature specific Multi-species Stock-assessment Model (MSMt) (Holsman et al., in review)is a modification of a previous multi-species age-structured statistical model that combines a catch-at-age stock assessment model with multispecies virtual population analysis (MSVPA) in a statistical framework (Jurado-Molina et al., 2005). In MSMt, weight-at-age and predation mortality vary as a function of bottom temperature, allowing MSMt to capture climatic-driven changes in growth and predation effects on biomass and attendant harvest rates. Weight-at-age is determined from temperature-specific von Bertalanffy growth functions fit to otolith-based size-at-age data (Holsman et al., in review). MSMt dynamically estimates time-varying natural mortality (i.e., 𝑀𝑀𝑝𝑝𝑖𝑖,𝑝𝑝 = 𝑀𝑀1𝑝𝑝𝑖𝑖 + 𝑀𝑀2𝑝𝑝𝑖𝑖,𝑝𝑝) based on the numerical abundance and biomass of predators and prey where 𝑀𝑀1𝑝𝑝𝑖𝑖 is the age (a) specific residual mortality for each species p, and 𝑀𝑀2𝑝𝑝𝑖𝑖,𝑝𝑝 is the annual age-specific predation mortality for each species. Predation morality in MSMt is the combined outcome of temperature-dependent predator rations estimated from bioenergetics models of consumption, and a foraging sub-model that allocates predator consumption to various species in the model. The foraging model is based on patterns of size- and species-specific prey preference that reflect the relative availability of prey species in the system and is based on trophic patterns in diet data from 1980-2012 averaged over the entire EBS. MSMt is statistically fit to fishery and survey data (1979+) for catch biomass, survey biomass, fishery and survey size- at age-composition, length to weight relationships, predator size and species preference, bioenergetics-based temperature-specific predator rations, and maturity (Holsman et al., in prep). Published size-and temperature-specific algorithms for predator rations (Holsman et al., in review) are also used. Emergent quantities estimated by the model include biomass consumed (by predators in the model), recruitment, fishery, survey, and predator selectivity, annually varying natural mortality, age-specific abundance, population biomass, and

50


harvest rate. Ecosim Ecosim (Walters 2000; Christensen and Walters, 2004) is a dynamic whole-of-ecosystem model. It simulates predator-prey relationships between functional groups, implicit refuges from predation, and time-varying diets. Unlike FEAST, Ecosim is spatially-aggregated. The mass balance model of the EBS continental shelf system is defined by the North Pacific Fishery Management Council (NPFMC) management areas between 500 and 531 (but does not include area 530), which coincide roughly with International Pacific Halibut Commission (IPHC) management areas 4C-4E in the EBS. The continental shelf and slope to approximately 1,000 m are included in the model following AFSC bottom trawl surveys. Unlike in the AI or GOA, nearshore areas of less than 50 m depth are included in the shallowest depth stratum for the EBS. Within the NPFMC management areas listed above, the area of the EBS shelf/slope covered by NMFS trawl surveys is 495,218 km2. This total shelf area was used to calculate biomass and production per unit area as model inputs. There are ten spatial strata in the EBS model (Aydin et al., 2007): six on the EBS shelf, three on the northern Alaska Peninsula (“Horseshoe”), and one along the EBS slope. The shelf habitat types are defined as “shallowest” habitats from 0-50 m depth, “shallow” habitats from 50-100 m depth, and “middle” habitats from 100-200 m depth. The entire EBS slope habitat ranges from 2001,000 m depth. Habitats north of the Alaska Peninsula in the Horseshoe area are classified similarly to GOA and AI, with shallow, middle and deep regions referring to the 0-100 m, 100200 m, and 200-500 m depth layers, respectively. Table 6.2 lists the species included in the model. Note that these are the model group names, which do not always correspond to single taxonomic species. Full descriptions of the species included in each of these groups are found in Appendix A of Aydin et al. (2007). Species were categorized as one of either migratory (moving specifically across model boundaries), stock (primarily contained within each model’s boundaries), complexes (stocks consisting of multiple species) or local (subpopulation/different species may occur in different subdomains of each of the three models). Further, species were modeled as either biomass pools or aged (initially split into juvenile and adult biomass accounting; this would be elaborated into a fully age-structured model during future dynamic simulations). Juvenile groups were included to account for ontogenetic diet shifts and to represent age structure for protected pinnipeds and commercially important fish species. See Appendix A of Aydin et al. (2007) for detailed pinniped juvenile definitions. In general, “juveniles” of each major groundfish species are defined to be those individuals less than 20 cm long. This size threshold was based on observations of groundfish predator diets, where fish smaller than 20 cm were much more common in diets than those above 20 cm in length. Using a size threshold to define all juvenile groups means that the age of juveniles may vary by species. The approximate ages corresponding to juvenile groups for each species in these models are discussed in each species group description in Appendix A of Aydin et al. (2007). Pacific salmon (Oncorhynchus spp.) represent a unique model group, as a large proportion of the critical stages in their life cycle occur outside of modeled areas, and their presence occurs in compressed bursts of migration throughout the year. These bursts represent a large component of both food supply and predation, and yet their temporal compression prevents scaling their brief in-system growth rates to the remainder of their life cycle. Therefore, outmigrating and immigrating salmon are considered to be separate (unlinked) species and treated as an input parameter rather than a state variable for dynamic simulations. The substantial catch of incoming adult salmon is included in the EBS models, although this fishery operates differently than other modeled fisheries

51


(terminal fishery). Data Generation The data used by the single species assessment models are shown in Table 6.1, which also shows which data types are used by MSMt and Ecosim. The multi-species models have additional data requirements. Both MSMt and Ecosim require diet data from the shelf and slope surveys. Diet data for pollock, Pacific cod, and arrowtooth flounder by length is required for MSMt. Ecosim requires diet data for the entire modeled food web (Table 6.2). Species groups were chosen for use with the FEAST operating model. Additional Ecosim data requirements tailored for use with the FEAST operating model are shown in Tables 6.3 and 6.4. Transforming FEAST age and length bins to single species assessment bins Binning of fish length and age data was optimized in FEAST to reduce the runtime of the simulations while keeping the bins small enough to capture the dynamics of the fish in the system. The FEAST age bin [0:1:10]2 and lower length bins for walleye pollock [0:2:76], Pacific cod [0:2:102], and arrowtooth flounder [0:2:74] differ from the age and length bins that have been used in the single-species assessments for these species: walleye pollock ([1:1:15] and [25:2:35 36:1:46 48:2:62]), Pacific cod ([1:1:12] and [9:3:45 50:5:105]), and arrowtooth flounder ([1:1:21] and [10 16:2:40 43:3:70 75]). Therefore the fish data need to be transformed to the appropriate bins for the assessments, and the length bins used in the Pacific cod and arrowtooth flounder assessments were reduced to [9:3:45 50:5:100] and [10 16:2:40 43:3:70] respectively. It was necessary to transform the age and length bins for each grid cell and day from FEAST needed for the MSE. Length data The length data from FEAST bins were first transformed to 1 cm bins. Linear interpolation was used to calculate the density of fish in 1 cm bins from the 2 or 4 cm bins in FEAST. These 1 cm bins were then used in all intermediate calculations. The final step was to aggregate the length data by 1 cm length bins into the length bins used in the assessments. Age data Fish density at ages 1 through 9 were unchanged from the FEAST age bins, but age-10 fish needed to be extrapolated into older age bins for the single-species assessments under the assumption of an exponential decay in abundance with age. If no data were available for the years prior to the year for which age data were needed (i.e., the start year of the simulation), age-10 fish were extrapolated into older age bins using only the current run year natural and fishing mortality rates (mortality rates used were particular to each FEAST grid cell defined by the location of the haul and averaged over the year). For years in which previous years’ data were available, the historical density of age 9 fish on July 1 (the middle of the survey season) averaged over the EBS, along with the historical mean natural and fishing mortality rates, were used to determine the proportion of age 10 fish that would be expected in age bins 10 through the maximum age bin in the assessment, i.e.:

2 This vector notation refers to [minimum value: step: maximum value]. Specific values are separated by a space.

52


, , , ,1,9 1,9

, , , , , , , ,9,9 9,9 ,10 ,10

( ), ,1,9

1( ) ( ), , , ,

, 9,910

(, ,9,9

s survey h s survey hy y

s survey h s survey h s survey h s survey hy a y a i i

y

M Fs survey hy

yM F M Fs survey h s survey h

y a y ai y a

Ms survey hy j

D e

P D e e

D e

− −

− + − +

−

− +−

−− + − +

− += − +

−− +

= ⋅

⋅

∏, , , , , , , ,

9,9 9,9 ,10 ,10

15) ( )

10

s survey h s survey h s survey h s survey hj y j i i

yaF M F

j a i y j

e+ − +

−++ − +

= = − +

∑ ∏

if 10

if 11

if

s

s

a

a x

a x

=

≤ <

=

(6.1)

where , /,s survey h

y aP is the relative density at the location of the hth haul during survey survey for

animals of species s and age a during year y, , /,

s survey hy aD is the density at the location of the hth haul

during survey survey of animals of species s and age a during year y, , /,

s survey hy aM is the rate of

natural mortality for animals of species s and age a during year y in the location where the hth haul during survey survey took place, , /

,s survey h

y aF is the fishing mortality rate for animals of species s and

age a during year y at the location where the hth haul during survey survey took place, and sx is the maximum age-class for species s. The densities by age-class for survey survey for ages a > 9 were computed using the equation:

, , , , , , , ,, ,10 , ,

10/

sxs survey h s survey h s survey h s survey hy a y y a y a

aD D P P

=

= ∑ (6.2)

Condition factors Condition factors, which in the FEAST model are used to calculate weight, are defined for each FEAST age and length combination. They need to be transformed into the length and age bins used for single-species assessment. For length, the condition factor for the 1 cm bins were set equal to the condition factor of the nearest smaller FEAST length bin. The condition factors for the age bins greater than age-10 were set equal to the age-10 condition factor for each particular grid cell and day (haul). 3.2.1.4 Sex-ratios for arrowtooth flounder The single-species stock assessment for arrowtooth flounder is sex-structured owing to sex-specific differences in mortality, and hence length compositions. However, the FEAST model does not split species by sex. It was therefore necessary to split the length compositions for simulated shelf and slope surveys and the catch for arrowtooth flounder to sex. Assuming that the sex-ratio is independent of length (e.g. that 60% of all arrowtooth are female) would lose information. Consequently, the length-compositions by sex for the fishery catch, the shelf survey and the slope survey for 1981-2009 (Figure 6.1) were used to split the simulated fishery catch and survey length-compositions to sex. Generating simulated survey data Stochastic verses deterministic survey data Two types of simulated data sets are generated using FEAST: a deterministic data set and several stochastic data sets. The deterministic data set represents the expectations of the results of the shelf, slope, and Echo Integration Trawl (EIT) surveys. Given that the survey stations are pre-specified,

53


there is no uncertainty associated with the choice of survey grid and survey stations within the grid. Stochasticity is introduced into the data generation process because of subsampling (taking weight data from only a few fish from the survey catch for instance) and additional observation error so that the generated data mimic the noise associated with the observed data. Diet data MSMt and ECOSIM require diet data collected from stomach samples. Diet data were only simulated for collections during the summer bottom trawl shelf and slope surveys because collection of diet data by observers is somewhat opportunistic and because the multispecies models do not currently use winter observer data. The AFSC diet data sampling plan does not define how many fish of each length are to be collected and examined. Therefore, these numbers were defined by the historical average from the years 2005-2009 for the shelf and slope surveys (Fig. 6.2). Bottom trawl shelf survey Bottom trawl shelf survey data were generated for each year they were historically used in the single-species assessments and every year of the future simulations, with the hauls defined by the actual historical locations and mean dates of hauls from 1982-2009 (Appendix A.1). For each shelf survey haul, the density and condition factor of each species by length and age was extracted from FEAST, along with bottom temperature. Density of fish was converted to numbers of fish in the haul using the historical mean area swept (4.773 ha), availability to the trawl (Figure 6.3a) and capture probability of fish available to the trawl (Figure 6.3b). The availability to the bottom trawl (those fish 0-3m off the bottom) was calculated for fish in the entire water column because fish are not vertically distributed in FEAST. Pollock larger than 9 cm were assumed to be either available to the bottom trawl survey or to the EIT survey, represented by the assumption that availability to the bottom trawl and to the EIT surveys sums to one for each length (Figures 6.3a and 6.3d). The numbers of fish captured by the bottom trawl survey by FEAST length and age bins were converted to 1 cm length bins and assessment age bins. This matrix of numbers of fish and condition factor by length and age per survey haul was used in all subsequent calculations. The number of length samples taken per haul (200/species) was defined by the bottom trawl shelf survey protocol (Lauth and Acuna, 2009). The numbers of fish weighed and aged per haul (4 pollock, 3 Pacific cod) and the numbers of fish weighed per haul (4 pollock, 2 arrowtooth flounder) were selected based on the actual historical data for 2004-08. Arrowtooth flounder otoliths are usually collected during the shelf and slope surveys. However, they are not needed for the single-species assessment and the sampled otoliths have only been aged for one year of the past 12. The multispecies assessment models require data on weight-at-age. It is likely that reading of arrowtooth flounder otoliths would be given a higher priority if multispecies assessment models were used to provide management advice. Therefore, the data generation process assumed that three arrowtooth per haul were weighed and aged (similar to current aging rates for Pacific cod in the shelf survey). The simulated haul samples for the shelf survey were converted to aggregated EBS data for use in the stock assessments using the same methods as are used in reality (Lauth and Acuna, 2009); data were aggregated to haul and then to stratum. The expected survey CPUE (numbers per hectare) for a given haul is computed using the density of fish, availability to trawl, capture probability of fish available to the trawl, and the historical mean area swept (4.773), i.e.:

54


,shelf, ( ) ,shelf , ( ) ,shelf ,shelf, ,4.773s h k s h k s s

y y l a l ll a

U D V P= ∑∑ (6.3)

where ,shelf , ( )s h kyU is the expected CPUE for species s in haul h (where haul h is in survey stratum

k) during the shelf survey conducted in year y, ,shelf , ( ), ,

s h ky l aD is the density of fish of species s and

age a in length bin l at the location of haul h when the shelf survey took place during year y, ,shelfslV

is the availability of fish of species s in length bin l to the shelf survey (Figure 6.3a), and ,shelfslP is

the probability of capture for fish of species s in length bin l during the shelf survey (Figure 6.3b). Haul length (

Nl ) and age (

Na ) frequencies are calculated similarly:

,shelf , ( ) ,shelf , ( ) ,shelf ,shelf, , ,4.773s h k s h k s s

y l y l a l la

N D V P= ∑ (6.4a)

,shelf , ( ) ,shelf, ( ) ,shelf ,shelf, , ,4.773s h k s h k s s

y a y l a l ll

N D V P= ∑ (6.4b)

The expected length and age frequencies for species s during year y by stratum (k) are calculated as mean haul length and age frequencies weighted by haul CPUE, i.e.:

shelf shelf( ) ( ),shelf , ,shelf , ( ) ,shelf , ( ) ,shelf , ( ), ,

1 1( ) /

n k n ks k s h k s h k s h ky l y l y y

h hN N U U

= =

= ∑ ∑ (6.5a)

shelf shelf( ) ( ),shelf , ,shelf , ( ) ,shelf , ( ) ,shelf , ( ), ,

1 1( ) /

n k n ks k s h k s h k s h ky a y a y y

h hN N U U

= =

= ∑ ∑ (6.5b)

where shelf ( )n k is the number of hauls during the shelf survey in stratum k.

Stratum expected length and age frequencies are used to calculate the expected length and age frequencies for the whole Eastern Bering Sea:

,shelf, ,shelf ,, ,

s EBS s ky l y l

kN N= ∑ (6.6a)

,shelf, ,shelf,, ,

s EBS s ky a y a

kN N= ∑ (6.6b)

The expected total numbers by stratum for the shelf survey is calculated using the equation: shelf ( )

,shelf, shelf, ,shelf, ( )shelf

1

1( )

n ks k k s h ky y

hN S U

n k =

= ∑ (6.7)

where shelf ,kS is the area of stratum k in the shelf survey. The expected total number for the entire EBS is the sum of the stratum estimates:

,shelf, ,shelf,s EBS s ky y

kN N= ∑ (6.8)

Standard errors were calculated for the numbers estimates:

55


helf

,shelf,

shelf, 2 ( )2 ,shelf , ( ) ,shelf, 2

helf helf1

( ) ( )( )( ( ) 1)

s

s ky

k n ks h k s ky ys sN

h

S U Un k n k

σ=

= −− ∑ (6.9)

,shelf, ,shelf,2

s EBS s ky yN N

kσ σ= ∑ (6.10)

where ,shelf,s kyU is the mean (across haul) CPUE for stratum k during year y.

The expected biomasses by haul from the shelf survey are calculated using the FEAST condition factor for each age, length, time, and location, the length-weight relationship, and the estimated number of fish in the haul:

,shelf, ( ) ,shelf, ( ) ,shelf , ( ) ,shelf ,shelf, , , ,4.773

ss h k s h k s s h k s sy y a l l y l a l l

a lB R L D V Pεα= ∑∑ (6.11)

where ,shelf, ( ), ,

s h ky a lR is the condition factor for animals of species s and age a in length bin l at haul h

during the shelf survey in year y (weight relative the expected weight), sα and sε are the

parameters of length-weight relationship, and lL is the mean length for a fish in length bin l. The expected CPUE in biomass is calculated for each haul from the haul biomass estimates and the mean area swept:

,shelf, ( ) ,shelf, ( ) 4.773s h k s h ky yZ B= (6.12)

where ,shelf, ( )s h kyZ is the CPUE in biomass for species s at the haul h during the shelf survey in year

y. The expected biomass by stratum, ,shelf,s k

yB , and the expected biomass for the entire EBS, ,shelf,s EBSyB

and their standard errors are:

shelfshelf, ( )

,shelf, ,shelf, ( )shelf

1( )

k n ks k s h ky y

h

SB Zn k =

= ∑ (6.13a)

shelf

,shelf,

shelf, 2 ( )2 ,shelf, ( ) ,shelf, ( ) 2

shelf shelf1

( ) ( )( )( ( ) 1)s k

y

k n ks h k s h ky yB

h

S Z Zn k n k

σ=

= −− ∑ (6.13b)

,shelf, ,shelf,s EBS s ky y

kB B= ∑ (6.14a)

,shelf, ,shelf,2

s EBS s ky yB B

kσ σ= ∑ (6.14b)

where ,shelf, ( )s h kyZ is the mean (across hauls) CPUE in biomass for species s for stratum k during the

shelf survey in year y. Expected mean length-at-age (

L ) is calculated as the mean of the length-at-age (L) samples weighted by CPUE:

56


helf helf

,shelf, ,shelf, ,shelf, ,shelf,, ,

1 1( ) /

s sn ns EBS s h s h s hy a y y a y

h hL U L U

= =

= ∑ ∑ (6.15)

where shelfn is the number if hauls in the shelf survey, ,shelf,,

s EBSy aL is the mean length-at-age for species

s during year y during the shelf survey for the entire EBS, and ,shelf,,

s hy aL is the mean length-at-age

for species s during year y during haul h. Expected mean weights-at-age are calculated for each haul:

shelf, ( )

, , ( ) , , ( ), , ,i , , ( )shelf, ( )

1,

1 h ks

ns shelf h k s shelf h k s

y a y a a i h kh kiy a

W R Ln

εα=

= ∑

(6.16)

where , , ( ),s shelf h k

y aW is the mean weight of fish of species s and age a sampled in haul h

during the shelf survey in year y, , , ( ), ,i

s shelf h ky aR is the condition factor for ith fish of species s

and age a sampled in haul h during the shelf survey in year y, , , ( )a i h kL is the length of ith fish of species s and age a sampled in haul h during the shelf survey in year y, shelf, ( )

,h k

y an is the number of animals of age a which were sampled in haul h during the shelf survey in year y. The mean weight-at-age for the entire EBS is given by

,shelf, ,shelf, ( ) ,shelf, ( ) ,shelf, ( ), ,( ) /s EBS s h k s h k s h k

y a y a y yh h

W W U U= ⋅∑ ∑ (6.17)

where ,shelf,,s EBS

y aW is the mean weight of fish of species s and age a in the shelf survey during year y.

The bottom temperature for the entire EBS for the shelf survey is defined by the mean haul bottom temperature weighted by the proportion of survey area it accounts for:

shelf, shelf,shelf, shelf, ( )

shelf shelf, shelf shelf,1 1( ) /

( ) ( )

k kn nEBS h k

y y k kh h

k k

S ST Tn k S n k S= =

=⋅∑ ∑∑ ∑

(6.18)

where shelf,EBSyT is the mean temperature during the shelf survey of year y, and shelf, ( )h k

yT is the temperature at haul h during the shelf survey in year y. Bottom trawl slope survey Bottom trawl slope survey data were generated for each year they were historically used in the single-species assessments and for each even year of the future simulations, with the hauls defined by the actual historical locations and dates of hauls in 2008 (Appendix A.2). The density and condition factor for arrowtooth flounder by length and age was extracted from the FEAST output. Density of arrowtooth flounder was converted to number of fish in the haul using the historical mean area swept (7.487 ha), availability to trawl, and capture probability of arrowtooth (Fig. 6.3c). The FEAST length and age bins were converted to 1 cm length bins. The number of length samples for arrowtooth flounder taken in each haul (300) was defined by the bottom trawl slope survey

57


methods (Hoff and Britt, 2009). Similar to the shelf survey, three arrowtooth flounder per haul were assumed to be weighed and aged. The following calculations were used to convert haul samples to aggregated EBS data for use in the stock assessments, and to follow the actual methods as outlined in Hoff and Britt (2009). The survey CPUE (numbers per hectare) for a given haul is computed using the density of fish, availability to trawl, capture probability of fish available to the trawl, and the historical mean area swept (7.487):

,slope, ( ) ,slope, ( ) ,slope ,slope, ,7.487s h k s h k s s

y y l a l ll a

U D V P= ∑∑ (6.19)

where ,slope, ( )s h kyU is the expected CPUE for species s in haul h (where haul h is in survey stratum

k) during the slope survey conducted in year y, ,slope, ( ), ,

s h ky l aD is the density of fish of species s and

age a in length bin l at the location of haul h when the slope survey takes place, ,slopeslV is the

availability of fish of species s in length bin l to the slope survey, and ,slopeslP is the probability of

capture for fish of species s in length bin l during the slope survey. Expected haul length frequencies are calculated similarly:

,slope, ( ) ,slope, ( ) ,slope ,slope, , ,7.487s h k s h k s s

y l y l a l la

N D V P= ∑ (6.20)

The expected length frequencies for species s during year y by stratum (k) are calculated as mean haul length frequencies weighted by haul CPUE:

slope slope( ) ( ),slope, ,slope, ( ) ,slope, ( ) ,slope, ( ), ,

1 1( ) /

n k n ks k s h k s h k s h ky l y l y y

h hN N U U

= =

= ∑ ∑ (6.21)

where slope ( )n k is the number of hauls during the slope survey in stratum k.

Stratum expected length frequencies are used to calculate the expected length frequencies for the whole Eastern Bering Sea:

,slope, ,slope,, ,

s EBS s ky l y l

kN N= ∑ (6.22)

The expected biomass by haul from the slope survey is calculated using the FEAST condition factor for each age, length, time, and location, the length-weight relationship, and the estimated number of fish in the haul:

,slope, ( ) ,slope, ( ) ,slope, ( ) ,slope ,slope, , , ,7.487


a lB R L D V Pεα= ∑∑ (6.23)

where ,slope, ( ), ,


during the slope survey in year y (weight relative the expected weight). The expected CPUE in biomass is calculated for each haul from the haul biomass estimates and the mean area swept:

58


,slope, ( ) ,slope, ( ) 7.487s h k s h ky yZ B= (6.24)

where ,slope, ( )s h kyZ is the CPUE for species s at the haul h during the slope survey in year y. The

expected stratum and EBS biomass estimates and their standard errors can be computed using:

slopeslope, ( )

,slope, ,slope, ( )slope

1( )

k n ks k s h ky y

h

SB Zn k =

= ∑ (6.25a)

slope

,slope,

slope, 2 ( )2 ,slope, ( ) ,slope, ( ) 2

slope slope1

( ) ( )( )( ( ) 1)s k

y

k n ks h k s h ky yB

h

S Z Zn k n k

σ=

= −− ∑ (6.25b)

,slope, ,slope,s EBS s ky y

kB B= ∑ (6.25c)

,slope, ,slope,2

s EBS s ky yB B

kσ σ= ∑ (6.25d)

where slope,kS is the area of stratum k in the slope survey. Echo-integration trawl (EIT) survey The NMFS EIT survey data were generated for each year they were historically used in the single-species assessments and for each odd year for the future simulation years. The continuous transect EIT survey was converted to the discrete FEAST grid cells by defining static survey locations for which EIT data would be generated (Appendix A.3). The density and condition factor of pollock by length and age was extracted from the FEAST output for each EIT survey station. Density of fish was converted to number of fish in the haul using the area swept (18.5822 ha), and the combined effects of availability and capture probability for the survey (Fig. 6.3d). The area swept for each survey station was defined by the width of the acoustic beam at the historical mean bottom depth of the EIT survey multiplied by the historical (1994-2010) mean total length of the EIT transects in the U.S. EEZ, divided by the number of EIT survey stations in Appendix A.3. The width of the 38 kHz acoustic beam used in the AFSC EIT surveys at the average bottom depth of the survey (125 m) is about 14 m (Patrick Ressler, AFSC, pers. comm). It was necessary to make assumptions about the availability of fish in the water column to the EIT survey, which reports fish 16 m from the surface to 3 m off the bottom because fish are not distributed vertically in FEAST. The assumption was that pollock larger than 9 cm are either available to the bottom trawl survey or available to the EIT survey, is represented by the assumption that availability to the bottom trawl survey and to the EIT survey sums to one for each length (Figs 6.3a and 6.3d). FEAST length and age bins were converted to 1 cm length bins and the bins used in the single-species stock assessments. The numbers of samples taken for pollock length, age, and weight data each year from the EIT survey were defined by the fraction of age 3+ biomass at the start of the year that was sampled from the most recent five surveys (2002, 2004, 2006, 2007, 2008; 3.82E-03 lengths, 3.63E-04 ages, 4.29E-04 weights (numbers sampled/ton)). The total number of samples taken from the EIT survey was calculated each year as the age 3+ biomass on January 1 of that year multiplied these sampling fractions.

59


The following calculations were used to convert haul samples to aggregated EBS data for use in the stock assessments, and to follow the actual methods as outlined in Honkalehto et al. (2008). The expected survey CPUE (numbers per hectare) for a given haul is computed using the density of fish, availability to EIT, capture probability of fish available to the EIT, and the historical mean area swept (18.5822 ha):

,EIT, ( ) ,EIT, ( ) ,EIT ,EIT, ,18.5822s h k s h k s s

y y l a l ll a

U D V P= ∑∑ (6.26)

where ,EIT, ( )s h k

yU is the expected CPUE for species s in haul h (where haul h is in survey stratum k) during the EIT survey conducted in year y, ,EIT, ( )

, ,s h ky l aD is the density of fish of species

s and age a in length bin l at the location of haul h when the IET survey took place during year y, ,EITs

lV is the availability of fish of species s in length bin l to the EIT survey, and ,EITslP is the

probability of capture for fish of species s in length bin l during the EIT survey. Expected haul length (

Nl ) and age (

Na ) frequencies are calculated similarly:

,EIT, ( ) ,EIT, ( ) ,EIT ,EIT, , ,18.5822s h k s h k s s

y l y l a l la

N D V P= ∑ (6.27)

,shelf , ( ) ,EIT, ( ) ,EIT ,EIT, , ,18.5822s h k s h k s s

y a y l a l ll

N D V P= ∑ (6.27)

The expected length and age frequencies for species s during year y by stratum (k) are calculated as the expected mean haul length and age frequencies weighted by haul CPUE:

,EIT, ,EIT, ( ), ,

s k s h ky l y l

hN N= ∑ (6.28a)

,EIT, ,EIT, ( ), ,

s k s h ky a y a

hN N= ∑ (6.28b)

Expected length and age frequencies for the entire EBS are calculated from expected haul length frequencies by stratum:

Ny,ls,EIT ,EBS = Ny,l

s,EIT ,h(k )

h∑ (6.29a)

Ny,as,EIT ,EBS = Ny,a

s,EIT ,h(k )

h∑ (6.29b)

The expected biomass by haul from the IET survey is calculated using the FEAST condition factor for each age, length, time, and location, the length-weight relationship, and the number of fish by haul:

,EIT, ( ) ,EIT, ( ) ,EIT, ( ) ,EIT ,EIT, , , ,18.5822


a lB R L D V Pεα= ∑∑ (6.30)

where ,EIT, ( ), ,


during the EIT survey (weight relative the expected weight).

60


The expected CPUE in biomass is calculated for each haul from the haul biomass estimates and the mean area swept:

,EIT, ( ) ,EIT, ( ) 18.522s h k s h ky yZ B= (6.31)

where ,EIT, ( )s h kyZ is the expected CPUE in biomass for species s at the haul h during the EIT survey

in year y. The estimates of the EBS biomass are:

EITEIT,EBS

,EIT,EBS ,EIT, ( )EIT

1

ns s h ky y

h

SB Zn =

= ∑ (6.32)

where EIT,EBSS is the area of the EIT survey. Consistent with the assumptions of the single-species assessments for pollock, the standard error for the EIT EBS biomass estimate is assumed to be 20% of the EIT EBS biomass estimate, and ,EITs

yn is the number of fish of species s sampled for weight during the EIT survey. The mean weights-at-age for pollock in the EBS from the EIT survey are calculated as the mean weight for each age from all samples taken from all hauls in the survey:

EIT

,EIT,EBS ,EIT, ,iEIT

1

1 ys

ns s s

y a y iiy

W R Ln

εα=

= ∑

(6.33)

where ,EIT,EBS,s

y aW is the mean weight of animals of species s and age a in the EIT survey during year

y, and ,EIT,i

syR is the condition factor for the ith fish of species s sampled during the EIT survey in

year y. Catch estimates The single-species assessments require an estimate of catch biomass for each year. The catch for year y is not known when the assessment in year y is undertaken. For consistency with actual practice, the catch for year y is taken as the actual catch from January to October 1 of year y plus an estimate about how much additional catch would occur from October 1 to December 31, i.e:

,,Oct1,

,1

sys

y s

CC

ττ

τφ=

−

(6.34)

where ,syC τ is the (estimate of) the catch biomass for species s caught by fleet τ during year y,

,,Oct1

syC τ is the catch biomass for species s caught by fleet τ during year y from January 1 to October

1, and ,s τφ is the average proportion of the catch of species s by fleet τ from January 1 to October

1. The values for ,s τφ are based on the average proportion of the total annual catch that was caught

from October 1 to December 31 during 2003-09 (Table 6.5). If the estimate from Equation 6.34 exceeded the TAC for year y, the TAC for year y was used instead. The catch biomasses for the years before the current year are assumed to be known exactly. These catch biomasses were

61


allocated to the fleets used in the single-species stock assessments (Table 6.5). Specifically, there were two fleets for EBS pollock, nine fleets for Pacific cod, and a single fleet for arrowtooth flounder. Fishery observer data Fishery observer data were generated for each year such data were available historically and used in the single-species assessments, and every year of the projection period. Only observer data up to year y-1 are used in assessment conducted during year y for consistency with reality. The total numbers caught at-age and at-length, and the mean weight-at-age in the catch (via the condition factor) were computed using FEAST output. In addition, the observer data were reported by sex for arrowtooth flounder. The numbers of fish sampled for length, age and weight were based on the fractions of catch (numbers/t) that were actually sampled by fishery observers (2007 onwards for pollock, 2008 onwards and the previous four years for Pacific cod [years vary by fleet], and 2004-08 for arrowtooth flounder) (Table 6.5).

Expected length and age frequencies sampled from the catch were converted from FEAST bins to the bins used in the single-species stock assessments. Expected length and age frequencies from the catch for the entire EBS were defined as the sum over all fleets:

,catch,EBS ,catch,, ,

s sy l y lN N τ

τ

= ∑ (6.35a)

,catch,EBS ,catch,, ,

s sy a y aN N τ

τ

= ∑ (6.35b)

where ,catch,EBS,

sy lN is the total catch-at-length for species s during year y , ,catch,

,sy lN τ is the catch-at-

length for species s and fleet τ during year y, ,catch,EBS,

sy aN is the total catch-at-age for species s during

year y , and ,catch,,

sy aN τ is the catch-at-age for species s and fleet τ during year y.

Expected mean weights at age (

W ) are calculated first for each fleet in the single-species stock assessments (

τ ): catch,

,,catch, ,catch,, , , ,catch,

1,

1 y as

ns s s

y a y a i a iiy a

W R Ln

τ

τ τ ετ α

=

= ∑ (6.36)

where ,catch,τ,s

y aW is the expected mean weight of fish of species s and age a sampled in the

catches by fleet τ during year y, ,catch,, ,i

sy aR τ

is the condition factor for ith fish of species s and

age a sampled in the catch of fleet τ during year y, ,

s

a iLε

is the length of ith fish of species

s and age a sampled from the catch by fleet τ during year y, catch,

,y an τis the number of animals

of age a which were sampled from the catches by fleet τ in year y. The expected mean weight at age of the catch for the entire EBS is given by

62


,catch, ,catch, ,catch, ,catch,, ,( ) /s EBS s s s

y a y a y yW W C Cτ τ τ

τ τ

= ∑ ∑ (6.37)

Stochastic data The expected data defined earlier are stochastic in the sense that the FEAST output from all grid cells is not used – only those data from the selected grid cells during the selected days are used to construct the biomass indices and the length- and age -compositions. However, the grid cells and days are pre-specified (Appendix A) so there is no inter-simulation uncertainty due to this source of uncertainty. In contrast, Equations 6.33 and 6.36 allow for subsamping of the catch for length and age within length, while the diet data are subsampled as well3. Although not yet implemented, the data on biomasses and numbers provided to the assessments will be log-normal with CVs chosen so that the sampling CV for the survey "matches" the actual CV. Some of the uncertainty associated with survey results is due to spatial variation. However, spatial variation under-estimates the extent of variation between the survey estimates of biomass and numbers and the outputs from single-species assessments. Rather than adding noise to the biomass and numbers estimates by haul, the “additional variation” would be added at the level the data are used in the assessment. The extent of additional variation can only be calculated once FEAST model output is available, data sets can be generated without additional variation and the single-species assessments applied. Implementation FEAST model simulation results are output in netcdf files containing several days’ data. Species are defined in FEAST by density at each grid cell and time step. Key simulated FEAST data needed for the MSE simulated surveys are then consolidated from the netcdf files into csv files using an R wrapper. The MSE code then extracts the simulated data from these csv files for each year of the model run and simulates the surveys.

3 Although this is yet to be coded.

63


Table 6.1. Simulated data to be generated from FEAST for the assessments.

Data source Data type Species Needed for assessment model(s)

Shelf survey Length frequency Pollock, cod, AF SSA, MSMt, ECOSIM Shelf survey Age frequency Pollock, cod, AF SSA, MSMt, Shelf survey Raw weight at age Pollock, cod, AF ECOSIM Shelf survey Mean weight at age Pollock, cod, AF SSA, MSMt Shelf survey Numbers estimate Pollock, cod, AF SSA, MSMt Shelf survey Biomass estimate Pollock, cod, AF SSA, MSMt, ECOSIM Shelf survey Bottom temperature AF SSA EIT survey Length frequency Pollock SSA EIT survey Age frequency Pollock SSA, MSMt EIT survey Numbers estimate Pollock SSA, MSMt EIT survey Biomass estimate Pollock SSA, MSMt, ECOSIM

Slope survey Length frequency Pollock, cod, AF SSA (AF only), ECOSIM Slope survey Raw weight at age Pollock, cod, AF ECOSIM Slope survey Numbers estimate AF SSA, MSMt Slope survey Biomass estimate Pollock, cod, AF SSA (AF only), MSMt,

ECOSIM Observer data Length frequency Pollock, cod, AF SSA Observer data Age frequency Pollock, cod SSA, MSMt Observer data Mean weight at age Pollock SSA

Fishery Total catch Pollock, cod, AF SSA, MSMt, ECOSIM

64


Table 6.2. Diet data requirements for Ecosim (blank values = zeros). Large and small phytoplankton do not consume any of the species in the model.

Predators Prey

Pollock Juveniles1

Pollock Adults1

P. Cod Juveniles1

P. Cod Adults1

Herring1

Arrowtooth Juveniles1

Arrowtooth Adults1

Squids2

Salmon returning2

Salmon outgoing2

Myctophidae2

Capelin2

Pollock Juveniles 0.01 0.05 0.01 0.03 0.24 0.54 Pollock Adults 0.09 0.31 0.22 P. Cod Juveniles 0.00 0.00 0.00 P. Cod Adults 0.00 0.01 0.00 Herring 0.00 0.00 0.01 Arrowtooth Juvenil 0.00 0.00 Arrowtooth Adults 0.00 0.00 0.00 Squids 0.00 0.00 0.00 0.01 0.25 Salmon returning 0.00 Salmon outgoing 0.03 Myctophidae 0.00 0.00 0.00 0.00 0.03 Capelin 0.00 0.00 0.00 0.01 0.01 0.03 Sandlance 0.00 0.00 0.01 0.03 0.00 0.03 Eulachon 0.00 0.00 0.02 0.03 Crab 0.00 0.00 0.00 0.16 0.00 Shrimp 0.00 0.04 0.01 0.06 0.05 0.04 OtherFEAST+ 0.21 0.06 0.88 0.16 0.00 0.46 0.03 0.05 0.25 Epifauna 0.02 0.00 0.01 0.11 0.01 0.00 GeorginaBenthos& 0.01 0.00 0.04 0.09 0.00 0.00 0.00 Jellies 0.00 0.00 0.00 Euphausiids 0.33 0.36 0.03 0.03 0.96 0.24 0.10 0.45 0.25 0.25 0.90 0.90 Copepods 0.43 0.39 0.01 0.00 0.04 0.00 0.00 0.36 0.25 0.75 0.10 0.10 Microzooplankton 0.00 0.00 0.00 Lg Phytoplankton Sm Phytoplankton Pelagic Detritus Benthic Detritus

65


Table 6.2 (Continued)

Predators Prey Sandlance2 Eulachon2 Crab3 Shrimp2 OtherFEAST2 Epifauna2 Infauna2 Jellies3 Euphausiids2 Copepods2 Microzooplankton2 Pollock Juveniles 0.00 Pollock Adults P. Cod Juveniles P. Cod Adults Herring Arrowtooth Juveniles Arrowtooth Adults Squids 0.00 Salmon returning Salmon outgoing Myctophidae 0.00 Capelin 0.00 Sandlance 0.00 Eulachon 0.00 Crab 0.00 Shrimp 0.00 OtherFEAST 0.05 0.30 0.02 0.00 0.02 Epifauna 0.23 0.00 0.02 GeorginaBenthos 0.59 0.10 0.45 0.50 0.08 Jellies Euphausiids 0.90 0.90 0.00 0.20 0.01 0.71 Copepods 0.10 0.10 0.02 0.17 0.25 0.00 Microzooplankton 0.01 0.06 0.15 0.49 Lg Phytoplankton 0.03 0.06 0.50 0.25 Sm Phytoplankton 0.01 0.10 0.25 0.70 Pelagic Detritus 0.30 Benthic Detritus 0.14 0.40 0.46 0.49 0.89 1 = data sampled from FEAST; 2 = pre-specified ‘literature values’ (not sampled from FEAST); 3 = EBS ‘literature values’ simulated from FEAST & = bivalves, polychaetes, misc. worms, and benthic microbes; + = North Pacific shrimp, Benthic Amphipods, Chaetognaths, and Mysids.

66

Table 6.3. Biological data requirements for the Ecosim assessment model, with example inputs.

Group Biomass (t/km)

P/B1 (yr-1)

Q/B2 (yr-1)

EE3

Prod/ Cons4

Unassim5 PDF6 BDF6

Pollock Juveniles 4.49* 1.45& 8.40& 0.2 0.4 0.6 Pollock Adults 18.49& 0.67& 3.17& 0.2 0.4 0.6

P. Cod Juveniles 0.19* 1.78& 8.83& 0.2 0.4 0.6 P. Cod Adults 2.46& 0.41& 2.28& 0.2 0.4 0.6

Herring 0.61& 0.32+ 3.52+ 0.2 0.4 0.6 Arrowtooth Juveniles 0.01* 0.81& 6.14& 0.2 0.4 0.6

Arrowtooth Adults 1.00& 0.18& 1.16& 0.2 0.4 0.6 Squids 3.20 10.67 0.80 0.2 0.4 0.6

Salmon returning 0.16& 1.65+ 11.60+ 0.2 0.4 0.6 Salmon outgoing 1.28+ 13.56+ 0.80 0.2 0.4 0.6

Myctophidae 0.80 3.65 0.80 0.2 0.4 0.6 Capelin 0.80 3.65 0.80 0.2 0.4 0.6

Sandlance 0.80 3.65 0.80 0.2 0.4 0.6 Eulachon 0.80 3.65 0.80 0.2 0.4 0.6

Crab 2.94& 1.30+ 3.12+ 0.2 0.1 0.9 Shrimp 0.58 2.41 0.80 0.2 0.1 0.9

OtherFEAST 27.70& 4.12 19.44 0.2 0.12 0.88 Epifauna 18.77& 4.08 0.20 0.2 0.1 0.9

GeorginaBenthos 109.29& 8.77 0.31 0.29 0.1 0.9 Jellies 0.34+ 0.88 3.00+ 0.2 0.4 0.6

Euphausiids 5.48+ 0.80 0.35 0.2 0.4 0.6 Copepods 5.79+ 26.52+ 0.80 0.2 0.4 0.6

Microzooplankton 45.00+ 36.50 0.35 0.25 0.4 0.6

Lg Phytoplankton 101.79

+ 0.80 0 0.4 0.6

Sm Phytoplankton 110.92

+ 0.80 0 0.4 0.6 Pelagic Detritus 0 0 Benthic Detritus 0 0

1 = Production/biomass ratio; 2 = consumption/biomass ratio; 3 = ecotrophic efficiency; 4 = production/consumption; 5 = unassimilated fraction of consumption used in detritus calculations; 6 = proportions of unassimilated consumption and dead animals going into either the pelagic or benthic detritus pool.

& = data sample from FEAST; + = EBS ‘literature values’ simulated from FEAST; none = pre-specified ‘literature values’ (not sampled from FEAST)


Table 6.4. Catch data requirements for the Ecosim assessment model. (X denotes a situation in which a sector catches a species) [CP denotes catcher-processor and CV catcher vessel]

Group CP trawl CP pot CP longline CV trawl CV longline CV pot Pollock Juveniles Pollock Adults X X X X X P. Cod Juveniles P. Cod Adults X X X X X Herring Arrowtooth Juveniles Arrowtooth Adults X X X X X Squids X X X Salmon returning Salmon outgoing Myctophidae X X Capelin X X Sandlance X X Eulachon X X Crab X Shrimp X X OtherFEAST Epifauna X X X X X X GeorginaBenthos X X Jellies X X Euphausiids Copepods Microzooplankton Lg Phytoplankton Sm Phytoplankton Pelagic Detritus Benthic Detritus

68


Table 6.5. Fishery observer samples (fraction of total catch measured (#/t)) taken from the simulated FEAST data. Observer data for the single-species assessments are extracted from the listed FEAST fleets.

SSA fleet splits FEAST fleets Length samples

Length / weight

samples

Age samples

Oct 1 - Dec 31 Catch

expansion (

φ) Pollock Catcher processor vessels CP_Pol_trawl 0.287289 0.031288 0.00306 0.0255

Catcher vessels* CV_Pol_trawl 0.220611 0.024026 0.00235 0.0688 Pacific cod

Trawl Jan-May CP_Pcod_trawl, CV_Pcod_trawl

0.3177 N/A

0.0346 Trawl Jun-Aug CP_Pcod_trawl, CV_Pcod_trawl

0.2516 N/A

Trawl Sept-Dec CP_Pcod_trawl, CV_Pcod_trawl

0.2019 N/A

Hook & Line Jan-May CP_Pcod_HAL, CV_Pcod_HAL

0.5626 N/A

0.2114 Hook & Line Jun-Aug CP_Pcod_HAL, CV_Pcod_HAL

0.9294 N/A

Hook & Line Sept-Dec CP_Pcod_HAL, CV_Pcod_HAL

0.9623 N/A

Pot Season Jan-May CP_Pcod_pot, CV_Pcod_pot

0.4523 N/A

0.0840 Pot Season Jun-Aug CP_Pcod_pot, CV_Pcod_pot

0.2423 N/A

Pot Season Sept-Dec CP_Pcod_pot, CV_Pcod_pot

0.2423 N/A

Arrowtooth flounder

CP_Pol_trawl, CV_Pol_trawl, CP_Pcod_trawl, CV_Pcod_trawl, CP_Pcod_HAL, CV_Pcod_HAL, CP_Pcod_pot, CV_Pcod_pot,

CP_Other_trawl, CV_Other_trawl, CP_Other_HAL, CV_Other_HAL, CP_Other_pol, CV_Other_pot

0.1417 N/A 0.0994

* Catcher vessels include those that deliver to either mothership or shore

69


Table 6.6. The seven scenarios to run in forecast for the BSIERP MSE project as decided at the MSE workshop.

Assessment model Climate model EBS cap

SSA CCMAT47 2 million t

MSMt CCMAT47 2 million t

Ecosim CCMAT47 2 million t

SSA Miroc-M 2 million t

MSMt Miroc-M 2 million t

SSA MIUB/ECHO-G 2 million t

SSA CCMAT47 Change cap

Figure 6.1. The mean proportion of the fishery catch and survey length-composition for arrowtooth flounder that is male by length based on data for 1981-2009 (years weighted by numbers in each annual sample). The sex-ratio was assumed to be 1:1 for the 18cm length bin for the slope survey owing to lack of data.

70


Figure 6.2. Stomach samples from the (a) shelf and (b) slope surveys by length and species. The values given are averages over surveys from 2005-2009.

71


Figure 6.3. (a) availability to the shelf survey for pollock (Jim Ianelli pers. comm.), Pacific cod (Nichol et al., 2007), and arrowtooth flounder (Jim Ianelli pers. comm.), (b) selectivity for pollock (Jim Ianelli, AFSC, pers. comm.), Pacific cod (Stan Kotwicki, AFSC, pers. comm.), and arrowtooth flounder (Kotwicki and Weinberg, 2005) assumed for the shelf survey, (c) availability and selectivity of arrowtooth flounder to the slope survey (Stan Kotwicki pers. comm.), and (d) combined availability and selectivity for pollock to the EIT survey (Jim Ianelli pers. comm.).

72


Chapter 7. Blended Forecasts and Short-Term Forecasts The first portion of this chapter (Blended Forecasts) is taken from the final report for B.73 Management Strategy Evaluation where it appeared as chapter 5. The short-term forecasting is original content of this final report. BLENDED FORECASTS Introduction The Scientific and Statistical Committees, SSCs, of the Regional Fishery Management Councils are required to provide recommendations for overfishing levels, OFLs, and Acceptable Biological Catches, ABCs, as well as evaluate whether a stock is subject to overfishing or is in an overfished state. For most major stocks, these recommendations are based on the outcomes of quantitative stock assessment methods, which involve fitting population dynamics models to monitoring data collected during fishing and surveys. For stocks managed by the North Pacific Fishery Management Council (NPFMC 2012, PMFC 2011), the stock assessments are based on single-species models, i.e. the impacts of time-varying predation mortality are ignored (or subsumed into the variation in the estimates of recruitment, under the assumption that the majority of predation mortality which is time-varying occurs before the age-at-recruitment). Most stock assessments involve pre-specifying the values for some of the parameters of the population dynamics model (e.g., the rate of natural mortality, M, fecundity as a function of length or age, and the survey catchability coefficient), making structural assumptions (e.g. vulnerability for a given fleet is a time-varying logistic function of length, recruitment is related to spawning stock size according to the Beverton-Holt form of the stock-recruitment relationship), choosing the data sets used when fitting the model (e.g., should fishery catch rate data be used or ignored given uncertainties regarding the relationship between catch rate and abundance), and assigning weights to the data sources which are included in the assessment. Although model fits to data may be similar, the results of stock assessments can be highly sensitive to parameter values and choices regarding model structure (e.g., Taylor and Stephens, 2013; Holsman et al. in review) In general, fisheries management advice (and hence OFLs and ABCs) is based on a single “best” model (and hence set of assumptions), and uncertainty is quantified about that model conditioned on its assumptions being correct. Typically, uncertainty is quantified using asymptotic methods, bootstrapping, or Bayesian methods (Magnusson et al., 2013). However, many sources of uncertainty are ignored when applying these methods, so the measures of uncertainty reported to managers usually underestimate the true amount of uncertainty (Ralston et al., 2011; Punt et al., 2012). The difference between the OFL and the ABC for a stock (the “buffer”) is meant to reflect the amount of scientific uncertainty. ABCs are often set so that the probability that the ABC exceeds the true OFL equals a selected value, P* (where P*< 0.5), i.e. P(ABC>OFL)=P* (Prager et al., 2003; Shertzer et al., 2008; Prager and Shertzer, 2010). However, the true probability that the ABC exceeds the OFL will be larger than P* if uncertainty is underestimated because the uncertainty associated with assumptions regarding model structure is ignored. The use of multispecies and ecosystem models for fisheries management is considered to be a key component of Ecosystem Based Fisheries Management (EBFM) (Marasco et al., 2007; Plagányi, 2007). However, similar to single-species stock assessment methods, projections based on two ecosystem models (or variants of one ecosystem model with alternative assumptions) often reflect uncertainty about model structure and assumptions regarding values for pre-specified parameters For example, Kaplan et al. (in press) evaluated the impacts of depleting forage species in the California Current ecosystem using Atlantis (Fulton et al., 2011; Horner et al., 2010) and Ecopath-with-Ecosim (Christensen and Walters, 2004; Field et al., 2006). However, the results from these two ecosystem

73


models differed markedly for some ecosystem components, to the extent that it is uncertain whether reducing forage species abundance would have a negative or positive effect on some ecosystem components. In addition, Kinzey and Punt (2009) showed that the results of a multispecies stock assessment were sensitive to the choice of the relationship between predation mortality and the density of predators and prey. For example, the multispecies models examined by Kinzey and Punt (2009) predicted that Pacific cod (Gadus macrocephalus) in the Aleutian Islands could have been increasing or decreasing prior to 1990 depending on this relationship. The selection of the model for providing management advice is often done informally or in discussion based council process that favors familiar, well tested, models but can impede rapid adoption of updated assessment models that incorporate ecosystem information. However, model uncertainty can be accounted for using “model averaging”. Model averaging involves integrating model outputs from multiple models taking account of the relative probability (or plausibility) of each model. This provides a means for accounting for the uncertainties in models themselves by incorporating multiple simulations from a variety of models. Model outputs can include projections of population size under alternative harvest control rules or catch scenarios as well as specific outputs such as OFLs and ABCs. Model averaging allows diverse, yet plausible, model outputs to collectively be used to guide management, and can provide estimates of uncertainty derived from both data fit (as is the case with singular models) as well as model structure and assumptions. It allows the uncertainty regarding which model is correct to be reflected in the outputs used for management purposes, along with the uncertainty conditioned on the model given that it is assumed to be correct. Here we provide a brief review of multi-model inference for fisheries assessment applications, focusing in particular on two alternative ways to implement model averaging for EBFM. We then use model averaging to integrate the results from three classes of model (single-species, temperature-specific single-species, temperature-specific multispecies) for three scenarios regarding future catch in the eastern Bering Sea in terms of impacts on the spawning stock biomass of walleye pollock (Gadus chalcogrammus), Pacific cod and arrowtooth flounder (Atheresthes stomias). Overview of model averaging There are two main ways to conduct model averaging: Bayesian Model Averaging (BMA) and “ensemble” forecasting. BMA requires that estimates of the posterior probability of each candidate model be available. This probability needs to be derived by fitting the model to available data. However, the probability of the model given the data cannot be derived for all models (e.g. dynamic ecosystem models such as Atlantis or the Forage/Euphausiid Abundance in Space and Time (FEAST) model) because they cannot be formally fitted to data. It is consequently impossible to apply BMA in many situations. When this is the case, “posterior probability distributions” can be approximated by “envelopes of plausibility” derived from ensemble/Monte Carlo runs of each model where each run is based on a different (yet plausible) set of parameters, with the probability assigned to each model based on expert judgment (i.e. the “Delphi method”), a process which we will term “ensemble” forecasting. Butterworth et al. (1996) proposed the following four-level scheme to assign ‘plausibility ranks’ to the hypotheses underlying alternative models which could be used to weight models when “ensemble” forecasting is conducted:

1. how strong is the basis for the hypothesis in the data for the species or region under consideration

2. how strong is the basis for the hypothesis in the data for a similar species or another region; 3. how strong is the basis for the hypothesis for any species; and 4. how strong or appropriate is the theoretical basis for the hypothesis?

For the population dynamics models typical of fisheries management, both ways of conducting model averaging fundamentally involve making projections. Each model can be projected multiple times (the outcomes will differ if there are multiple parameter choices for each model or the projections account for future stochasticity due to recruitment variability for example). The results of model averaging can

74


be summarized by the overall mean of some quantity of management or scientific interest, the spread of results, and by individual trajectories. The mean of the projections is a form of “best estimate”, but this does not take advantage of the primary aim of conducting multiple forecasts, namely to characterize uncertainty. Ianelli et al. (2011) summarize the results of projections for multiple models by illustrating intervals containing 50% and 80% of the outcomes as well as some individual trajectories to illustrate the overall uncertainty. Bayesian Model Averaging (BMA) The philosophy underlying Bayesian model averaging has been outlined by several authors (e.g. Buckland et al., 1997; Durban et al., 2005; Hoeting et al., 1999; Kass and Raftery, 1995; Raftery et al., 2005; Chimielechi and Raftery, 2011). Ideally, BMA involves fitting the range of models to the available data and computing the probability of each model given the data. This weighting of models should ideally be conducted using Bayes factors, which evaluate the credibility of a model relative to all other models (Aitkin, 1991; Kass and Raftery, 1995). However, studies have weighted alternative models, using, for example, the Deviance Information Criterion (Spiegelhalter et al., 2002), Akaike’s Information Criterion (Akaike, 1973; Burnham and Anderson, 1998), and the Bayes Information Criterion (Schwartz, 1978). The latter two weighting schemes are non-Bayesian, but are relatively easy to compute in contrast to Bayes factor and DIC which require that a Bayesian analysis be conducted, which can be computationally prohibitive even for relatively simple ecosystem models (e.g. Parslow et al., 2013). Bayes factor, DIC, AIC, and BIC can only be computed if each model is fit to the same data set. If the models are not fit to the same data set, it is necessary to weight each model using a more ad hoc approach, such as fitting the models to a subset of the data and predicting the remaining data, aka cross-validation. In this case, the weight assigned to each model would be proportional to the inverse of the mean square error associated with its predictions. Given probabilities for each model, the Bayesian model averaged forecast is constructed by conducting multiple projections for each model and generating the overall forecast by sampling projections with probability proportional to the probability of the model. Table 7.1 summarizes an application of Bayesian model averaging in which five models are used to predict the fishing mortality and spawning biomass corresponding to maximum sustainable yield (FMSY and SMSY) for George’s Bank haddock, Melanogrammus aeglefinus. The best model in Table 7.1 is RBH, but model RZBH is almost as likely. The model-averaged results are as expected closest to the best models, but the standard errors for the model-averaged results are larger than for either of the two best models. The weights assigned to each model in Table 7.1 are based on the Bayes factor. Bayes factors can be computed in this case because all of the models use the same data and the models are fairly simple. Ensemble forecasting Ensemble forecasting involves generating multiple projections of future system state under different choices for assumptions or parameter values. In principle, both structural and parameter uncertainty can be addressed through the use of multi-model ensembles. This approach is widely used in climate modeling where uncertainty is reflected in the accuracy of the approximations to the well-known and accepted physical principles of climate, and the inherent variability of the climate system. The climate system is chaotic, and the timing and phases of major and long-lasting fluctuations are largely unpredictable beyond time-horizons of a few years. Consequently, slightly different initial conditions for a climate model can lead to markedly different outcomes 40-50 years into the future. Whether including climate in population dynamics models has major impacts on the estimated future state of the populations under investigation depends on how the dynamics of the populations are linked to climate and the strength of the associated relationships. Probabilities can be assigned to model configurations (the underlying model equations and the values for its parameters) or entire model configurations can be considered implausible using hindcast

75


simulations of past conditions (e.g. Overland and Wang, 2007), although past performance is not necessarily a good indicator of success in simulating future climate (Reifen and Toumi, 2009). A’mar et al. (2009) based projections on six general circulation models which were selected for both their accuracy with respect to the historical data and their predictions with respect to future climate scenarios. Specifically, these six models were in the subset of models that replicated the spatial pattern and temporal characteristics of the first principal component of sea surface temperature (SST) in the North Pacific Ocean (the PDO) observed in the latter half of the twentieth century (A’mar et al., 2009). Fisheries examples of model averaging Model averaging in fisheries assessments are rare; the focus for fisheries management tending to be either selection of a best model or identification of harvest control rules that are robust to model selection and parameter value uncertainty (Butterworth, 2007). However, there are a few examples of where model averaging has been applied to fisheries population dynamics models and these are reviewed here. Bayesian Model Averaging BMA has been applied to account for uncertainty regarding the form of the stock-recruitment relationship (usually Ricker vs Beverton-Holt) and the error structure (autocorrelated or not, and the distribution for the residuals) (Patterson, 1999; Brodziak and Legault, 2005). BMA was used by Brandon and Wade (2006) to account for uncertainty regarding the form of the population dynamics model underlying a stock assessment (density-dependent or non-density-dependent, and whether the stock was at its environmental carrying capacity at the start of the modeled period) in an assessment of the Bering Sea-Chucki-Beaufort seas stock of bowhead whales, Balaena mysticetus. The weights assigned to each model by Brandon and Wade (2006) were based on Bayes factor and they developed their posterior distributions for each model using the sample-importance-resample algorithm, which allowed straightforward computation of the posterior probability of each model. Wilberg and Bence (2008) used Monte Carlo simulation to show that model averaging of alternative formulations for how fishery catchability changes over time performed better than using DIC to select a “best” model. Brodziak and Piner (2010) used BMA to integrate uncertainty due to the form of the stock-recruitment relationship (Ricker or Beverton-Holt), the extent of autocorrelation about the stock-recruitment relationship, and two values for the steepness of the stock-recruitment relationship for striped marlin (Tetrapturus audax) in the North Pacific. Unlike Broziak and Legault (2005), Brodziak and Piner (2010) approximated the Bayes factor using BIC. Ensemble forecasting This basic approach was used by Dick and Ralston (2009) when conducting analyses to quantify the impact of various rebuilding strategies for cowcod (Sebastes levis) off southern California. Dick and Ralston (2009) conducted forecasts for a range of assessment models, each of which was conditioned on one of a set of values for the steepness of the Beverton-Holt stock-recruitment relationship. Each projection was weighted based on a pre-specified probability distribution for steepness. Hamel (2011) conducted projections to evaluate times for Pacific Ocean Perch (Sebastes alutus) to rebuild to the proxy for the biomass at which maximum sustainable yield is achieved, BMSY for three models, given different levels of future fishing mortality and catch. Two of these models were assigned probability 0.25 and the third model was assigned a probability of 0.5. The selection of probabilities in this case was not entirely subjective because the specifications for the two models assigned probability 0.25 were developed to have approximately this probability (Hamel and Ono, 2011). Ianelli et al. (2011) evaluated the performance of management strategies for walleye pollock in the eastern Bering Sea. Recruitment was linked to predictions of SST from 82 Intergovernmental Panel on Climate Change (IPCC) models—SST, among other environmental factors, was found by Mueter et al. (2011) to be a possible factor affecting pollock recruitment. These 82 models were selected by downscaling IPCC models to the eastern Bering Sea ecosystem and using retrospective studies to identify models that perform poorly for this system (Wang et al., 2010).

76


Kolody et al. (2008), Kolody (2011), and Davies et al. (2012) developed an ‘uncertainty grid’ for assessments of swordfish (Xiphias gladius) in the Indian and Pacific Oceans, and explored structural uncertainty in a balanced factorial design. The results of the assessment were presented in terms of box plots of output statistics for each level of the factors considered. Kolody et al. (2008) explored sensitivity to stock-recruitment steepness, mixing proportions, growth rate/maturity/mortality options, the extent of variation about the stock-recruitment relationship, selectivity constraints, and data weights. Of 768 model configurations, a set of 192 model configurations considered the “most plausible ensemble” were used to summarize stock status. This ensemble was selected using three metrics: the root mean square fit to the catch rate index, the effective sample sizes for the length-frequency data, and the difference between observed and model-predicted mean catch lengths (similar to method of Francis 2011). Kolody (2011) assigned weighs to each of the factors on which the uncertainty grid was based using auxiliary information and the quality of the fits to the data, which led to some factors, such as that recruitment is related deterministically to spawning biomass, being assigned zero weight. An unusual form of ensemble modeling has been applied to calculate strike limits for the Bering-Chucki-Beaufort Seas stock of bowhead whales and the Eastern North Pacific stock of gray whales, Eschrichtius robustus. This involves calculating strike limits from two different methods and averaging them (Punt and Donovan, 2007). The philosophy underlying this approach is that each model can be wrong some of the time so averaging model results will lead to an outcome that is never very badly wrong (but is usually somewhat wrong). Application to walleye pollock, Pacific cod and arrowtooth flounder Alternative models Three classes of models formed the basis for the analysis (Table S.7.1):

1. The single species assessment models currently used by the AFSC to provide management advice for Eastern Bering Sea (EBS) walleye pollock (e.g., Ianelli et al., 2012), Pacific cod (e.g. Thompson and Lauth, 2012), and arrowtooth flounder (e.g., Spies et al., 2012). The assessments for these stocks are based on software developed specifically for those stocks coded using AD Model Builder (Fournier et al., 2012). All single species assessments have the following features in common: (a) they are fundamentally age-structured and use an annual time step. (b) estimates of annual fishing mortality rates are conditioned on the total catch (retained and discards) estimates, (c) fishery data (catch biomass and catch proportions at age) are aggregated over seasons and areas within each year, (d) proportions at age from surveys and fisheries are fitted using estimated (or assumed) multinomial sample sizes, and (e) survey indices (abundance or biomass) are modeled using lognormal assumptions and annually-specified observation errors (variances).Uncertainty in the projections based on these models reflects both parameter uncertainty, captured through MCMC sampling from the joint posterior distribution, and process error, captured through variation in recruitment about mean recruitment.

2. The Temperature-Specific Multispecies Model (MSMt; Holsman et al., in review) is an example of a “model of intermediate complexity” (Plagányi et al., 2014). The implementation of MSMt for the eastern Bering Sea includes the three focal species, models natural mortality for each species and age as the sum of a pre-specified residual natural mortality and time-varying predation mortality due to the predators included the model. Predation mortality is driven by temperature-dependent daily ration and a suitability function, which is based on observed proportions of each prey species by age in the diets of each predator species by age. Weight-at-age is also assumed to depend on temperature and varies annually. The parameters of MSMt are estimated by fitting the model to data on catch age-composition as well as survey biomass index and age-composition data. The projections of the model assume that future recruitment at age-0 is lognormal about mean recruitment. Two variants of MSMt are considered, one in which account is taken of multispecies interactions, MSMtA, and one which natural mortality is assumed to be constant over time, MSMtB. MSMtB differs from the single

77


species models used by AFSC in that weight-at-age in MSMtB depends on temperature and some other structural simplifications (e.g., constant fishery selectivity over time). Stochasticity in future projections based on MSMtA and MSMtB account only for process error in future recruitment.

Four climate scenarios are considered for MSMtA and MSMtB because temperature influences weight-at-age and the predation mortality function in MSMt. The four scenarios are: 1) future temperature is constant and equal to the mean of historical temperatures (temperature scenario 1), and future temperature in each projection year is the mean summer water column temperature predicted from a ROMS model for the Bering Sea forced by three statistically downscaled global climate models based on the IPCC A1B climate scenario (Wang et al., 2010), including: (temperature scenario 2) ECHO-G version 4, T30 resolution model (Legutke and Voss, 1999), (temperature scenario 3) CCMA model (Canadian Centre for Climate Modelling and Analysis CGCM3-t47; Flato et al., 2000, Flato and Boer 2001, Kim et al. 2002, 2003), (temperature scenario 4) MIROC 3.2 (Watanabe et al., 2011, K-1 model developers, 2004) (Fig. 5.1). A single realization of each of these three climate scenarios was used for atmospheric forcing and oceanic boundary conditions for the regional ROMS forecasts of the Bering Sea (present to 2040). Projections Each projection of the 1,000 iterations for each model involved the forecast period (2013-2039). Projections were undertaken from three future catch scenarios:

1. Catches set to the mean catch over the most recent 15 years (Table 7.2). 2. Catches set to the maximum catch over the most recent 15 years (Table 7.2) 3. No future catches of any species Results Results by model scenario Figures 7.2-7.4 shows the time-trajectories of spawning stock biomass for each of three models individually when catches are removed for each of three model types. Three sets of results are shown for the MSMtA and MSMtB models, one for each climate scenario. There is relatively little difference amongst the four climate scenarios for the MSMtA and MSMtB models, although the projections when future temperature in each projection year is based on downscaled global climate projections (rows 2-4 in Figures 7.3 and 7.4) are more variable. This is unsurprising given the low variability shown by the data in Figure 7.1. The general patterns between the two single-species models (AFSC; Figure 7.2 and MSMtB; Figure 7.4) share some qualitatively similar traits but also show some major differences. Specifically, the declines in abundance under the mean and maximum catches for pollock and Pacific cod are much greater for MSMtB. The cause of the differences between the two single-species models is unlikely to be due to temperature impacts on weight-at-age because the qualitative difference in results remains even when future temperature equals the historical mean. This is more likely due to differences in the assumptions regarding fishery selectivity in projections where MSMtB is balancing periods of selectivity shifted well to the left of maturity and the more recent selectivity trend which is more focused on older pollock, whereas the single-species model uses the assumption that the most recent 5-year average selectivity-at-age is most appropriate for projection purposes. This points out that simplifications in the MSMt models’ treatment of individual species are important to consider in evaluating projecting interactions. The results are also markedly sensitive to whether MSMt is applied in single-species or multispecies mode (Figures 7.3 and 7.4). Specifically pollock is predicted to decline and then rebuild under all catch scenarios (including zero catch) for MSMtA (multispecies mode) whereas pollock is predicted to

78


increase under zero catch and decline under mean and maximum catches for MSMtB (single-species mode). The difference between the single-species and multispecies predictions for cod is attributable primarily to the combined effect of cod and arrowtooth predation and pollock cannibalism in MSMtA. The abundance of Pacific cod is more robust between MSMtA and MSMtB, but the extent of decline in cod abundance is much greater under the multispecies version of the model (again reflecting a slight but important source of predation on age 1 cod). The trends in biomass of arrowtooth flounder are similar between MSMtA and MSMtB for the first few years of the projection period. However, unlike the MSMtB, MSMtA predicts stability or an increasing trend in arrowtooth abundance post 2030 (Figures 7.3 and 7.5). Because the increasing trend in arrowtooth abundance is not evident in MSMtB, this result is probably a consequence of temperature effects of weight-at-age combined with a slight decrease in predation by cod (and possibly pollock). Model averaged results Figure 7.5 shows results for three model classes when results are pooled over climate scenarios for the two MSMt models. The results of the projections, including those based on model averaging, are summarized quantitatively in Table 7.3 by the median and 50% and 90% intervals for spawning biomass in 2039, the last year of the projection period. As expected, model averaging across climate scenarios (assigning equal weight to each climate scenario) confirms that the impacts of the different climate scenarios on the model outcomes are not large (Figure 7.5). The widths of the 90% intervals in Figure 5.5 for the model-averaged results for MSMtA and MSMtB are not larger than those for the individual climate models, suggesting that variation in recruitment has a larger impact on uncertainty than the choice of climate model in this case. Figure 7.6 and Table 7.3 show results when the AFSC single-species assessments and the model-averaged results for MSMtA and MSMtB in Figure 7.5 are model-averaged (with equal weight assigned to all three models). In this case, the widths of the 90% intervals are wider for the model-averaged results than for the results for each individual model, reflecting that between-model variation is greater than the variation due to climate scenario (models MSMtA and MSMtB), parameter uncertainty (AFSC single-species models) and recruitment variation (all three models). Discussion Effectively capturing uncertainty is key focus for modern stock assessment science, and quantifying uncertainty in fisheries stock assessment models has been focus for stock assessment scientists for decades (e.g. Patterson, 1999; Hill et al., 2007; Magnusson et al., 2013). A full accounting for uncertainty requires adequately representing uncertainty regarding growth rates, natural mortality, the form and parameters of the stock-recruitment relationship, and how data are weighted. However, conventional approaches to quantifying uncertainty fail to capture ‘model uncertainty’, i.e. the uncertainty associated with the structural assumptions of a model. In general, single-species stock assessments make a small number of very strong assumptions (e.g. that natural mortality is independent of time) while multispecies and ecosystem models make more, but more specific assumptions (e.g. that the form of feeding functional relationship has the Holling Type II form) and often are forced to make other simplifying assumptions (e.g., constant fishery selectivity). Application of model averaging approaches (BMA or ensemble) is an appropriate way to express the full range of uncertainty, to the extent possible. The results in Figures 7.5 and 7.6 and Table 7.3 highlight the importance of different sources of uncertainty on predictions of spawning stock biomass under different catch scenarios. In particular, variability in climate scenarios contribute less to overall uncertainty than recruitment variation for the MSMtA and MSMtB models. However, model uncertainty is a more marked source of uncertainty than parameter uncertainty, recruitment variation, and the choice of climate scenario. It is, however, noteworthy that the impact of model uncertainty depends on the particular catch scenario under investigation. It is largest for the zero catch scenario, in particular given the impact of ‘release’ of Pacific cod, a major predator of pollock in the MSMtA model. The models are more consistent in their

79


predictions when the projections are based on the mean catch and most consistent for the projections based on the maximum catch where the biomass of predators and concomitant predation mortality is lowest (and thus differences between model parameterizations of predation mortality are lowest). Comparing alternative models has also raised another challenge in developing more “holistic” multispecies models. As noted above, simplifications in the multispecies model (e.g., constant fisheries selectivity) can introduce substantial differences in projections. For example, considering the estimated selectivity (Figure S.7.1) and mean body mass at age (but the same natural mortality-at-age) for pollock results in substantially different yield curves between the MSMt model and the single species model used for projections (Figure 7.7). This highlights the need to compare potentially subtle demographic characteristics when comparing multispecies projections with their single-species counterparts. Nevertheless, the different ways of modeling selectivity and body mass at age are plausible and reflect alternative hypotheses. Accounting for this source of uncertainty when conducting model averaging reflects the uncertainty due to choice made by modelers in the absence of definitive evidence in favor of one of the ways. Ralston et al. (2011) based their characterization of scientific uncertainty in the outputs of single-stocks species stock assessments by the extent of different assessment variation, where a key component of this uncertainty reflects choices made by analysts. The model forecasts were assigned equal probability in constructing the model-averaged forecasts. This was because there is no way for the hindcast and forecast skills of the three models to be compared at present. The ideal of using Bayes factor (or DIC, AIC, BIC) is infeasible in this case because although the parameters of the single-species model and MSMt are estimated by fitting them to monitoring data, each model has slightly different statistical weights and/or levels of aggregation in the data sources. In principle, each model could be weighted objectively by a cross-validation-like approach. For example, one could fit the model including data only up to 2008 and using the fitted model to predict the survey estimates of abundance for 2009, 2010, 2011, etc. given the catches that actually occurred during 2009, 2010, 2011; models that fit the observations better would obtain a higher weight. The illustrative application of this paper was based on three models. However, there are several other models that could have been included in the application. These include alternative multispecies models such as the multispecies virtual population analysis model of Jurado-Molina and Livingston (2002), and the statistical multispecies model developed by Kinzey and Punt (2009). Other models available for the Bering Sea include an Ecosim model (Aydin et al., 2007), the FEAST model, the multispecies surplus production model of Mueter and Megrey (2006), and a spatially-structured model of pollock (Hulson et al., 2013) Future work could involve evaluating the hindcast and forecast skill of projections based on a single model as well as on a model average of multiple models (c.f. Wilberg and Bence, 2008). This could involve fitting the model to a subset of the data and conducting projections. The skill of the modeling approach could then be evaluated in terms of the percentiles of the predicted distributions in which the actual observed fell. Ideally, the percentiles associated with the data should be uniformly distributed over 0-100. Large numbers of observations in the upper and lower tails of the forecast distributions would suggest that uncertainty is underestimated while no or few observations in the tails would suggest that uncertainty is overestimated. The benefits of using single models or model-averaged results could also be evaluated using simulations in which a true model is defined and data typical of an actual situation generated. This approach has been used extensively to evaluate the performance of single-species stock assessment methods, but has only been applied in a limited capacity for multispecies and ecosystem models (Kinzey (2010) being a noteworthy exception). Ultimately all approaches to applying model averaging involve subjective choices. These range from the initial choice of models to consider, along with a prior probability associated with each model. The latter is particularly a concern when many of the models are based on the same underlying philosophy. For example, the single-species assessments and MSMt, while different in several respects, make identical assumptions regarding many biological and fishery processes and cannot be considered to be totally independent. Similarly, MSMtA and MSMtB are identical except that the former allows for time-

80


varying predation mortality while the latter does not. The outcomes from this model averaging exercise are expressed in terms of time-trajectories of spawning output given a time-series of catches. However, the primary use of stock assessments is to define whether overfishing is taking place and whether the stock is in overfished stock, which, given the way fisheries management advice is provided in the US, requires a way to define the management reference points SMSY and FMSY. These reference points are well-defined for the single-species assessments (although the precision of the estimates even from single-species models can be poor). However, there are several alternative ways to define these reference points for multispecies and ecosystem models (Mofitt et al., in review). Holsman et al. (in review) illustrates how SMSY and FMSY can be calculated for a range of definitions for SMSY and FMSY. Ultimately, model averaging could be used to compute ensemble distributions for stock status relative to reference points if probabilities could be assigned to each of the definitions for SMSY and FMSY. We suggest that model uncertainty can be as large, or even exceed, many of the types of uncertainties considered routinely in stock assessments. Use of model averaging can quantify the range of outcomes from multiple models and hence better characterize uncertainty. Given that ABCs are OFLs are often reduced based on scientific uncertainty, accounting for model uncertainty can inform buffers between OFLs and ABCs and hence an improved ability to achieve fishery goals such as avoiding overfishing and preventing stocks from becoming overfished.

81


Table 7.1. Spawning stock biomass (SMSY: thousands of metric tons) and fishing mortality rate (FMSY: per year) associated with MSY for Georges Bank Atlantic cod (Gadus morhua) based on five stock-recruitment models.

Model Posterior Probability

SMSY FMSY

RBH 0.34 193.7 (26.2) 0.21 (0.03) RABH 0.15 176.1 (39.1) 0.23 (0.05) RZBH 0.33 188.7 (33.6)

0.22 (0.02)

0.16 172.7 (34.6) 0.23 (0.03) SRK 0.01 87.5 (57.4) 0.69 (0.01) Model Average 80% credibility intervals

184.7 (38.2) (135.8, 233.6)

0.23 (0.06) (0.15, 0.31)

RBH, informative recruitment priors with uncorrelated Beverton-Holt; RABH, informative recruitment priors with autocorrelated Beverton-Holt; RZBH, informative recruitment and steepness priors with uncorrelated Beverton-Holt; RZABH, informative recruitment and steepness priors with autocorrelated Beverton-Holt; SRK, informative slope at origin priors with uncorrelated Ricker (Ricker 1954, modified from Hill et al., 2007).

Table 7.2 Catches (t) used in the projections.

Stock Mean catch over 1998-2012

Maximum catch over 1998-2012

Pollock 1,226,280 1,490,900 Pacific cod 191,938 220,134 Arrowtooth flounder 13,458 17,737

82


Table 7.3 Percentiles of the distributions for the spawning stock biomass for the individual models and for the model averaged results.

a) Zero catch scenario

Model Climate Scenario

Pollock Pacific cod Arrowtooth flounder Low

5% Low 25%

Med. 50%

Up 25%

Up 95%

Low 5%

Low 25%

Med. 50%

Up 25%

Up 95%

Low 5%

Low 25%

Med. 50%

Up 25%

Up 95%

MSMtB

Average 3,671 4,566 5,281 6,208 8,039 398 498 575 662 815 275 322 365 414 498 ECHO-G 3,827 4,764 5,509 6,471 8,391 425 532 615 708 871 237 279 317 358 434 CCMA 4,093 5,089 5,876 6,906 8,959 471 591 682 785 969 192 227 258 292 355

MIROC-ESM 3,963 4,931 5,690 6,696 8,678 449 562 649 747 920 211 250 284 321 390

MSMtA

Average 1,525 2,011 2,461 3,103 4,683 258 310 354 405 497 217 265 313 368 470 ECHO-G 1,426 1,903 2,359 3,017 4,649 261 313 357 410 510 181 221 261 308 395 CCMA 1,359 1,784 2,226 2,859 4,342 275 329 376 432 531 142 174 206 242 309

MIROC-ESM 1,500 1,975 2,412 3,056 4,566 279 334 381 438 537 163 199 237 278 352

Single species 4,042 5,144 6,269 7,806 11,616 322 398 474 564 725 412 477 529 588 691

MSMtB Averaged 3,830 4,810 5,620 6,538 8,450 432 544 628 730 902 212 260 304 357 445 MSMtA Averaged 1,458 1,894 2,378 3,022 4,616 265 324 369 421 518 161 206 251 308 411

All Averaged 1,631 2,927 5,088 6,556 9,653 294 381 478 611 812 178 254 331 476 617

83


b) Mean catch scenario



5% Low 25%

Med. 50%

Up 25%

Up 95%

Low 5%

Low 25%

Med. 50%

Up 25%

Up 95%

Low 5%

Low 25%

Med. 50%

Up 25%

Up 95%

MSMtB

Average 0 1 644 1,919 3,794 0 0 33 121 279 183 229 272 319 406 ECHO-G 0 55 996 2,251 4,159 0 6 83 176 342 148 188 225 266 341 CCMA 0 340 1,447 2,663 4,737 0 46 143 240 423 110 144 175 208 271

MIROC-ESM 0 41 989 2,306 4,277 0 4 78 178 355 133 170 204 240 309

MSMtA

Average 145 1,609 2,773 4,160 7,149 0 0 2 36 139 174 239 299 376 507 ECHO-G 104 1,453 2,642 4,091 7,216 0 0 5 51 161 132 187 237 306 422 CCMA 44 1,312 2,532 3,929 7,130 0 0 14 69 181 91 134 175 229 322

MIROC-ESM 105 1,545 2,797 4,257 7,404 0 0 5 53 170 121 172 219 278 381

Single species 2,031 2,844 3,651 4,820 6,973 0 61 170 261 437 357 421 471 531 632

MSMtB Averaged 0 66 993 2,240 4,311 0 5 81 184 360 129 176 216 266 353 MSMtA Averaged 66 1,428 2,669 4,129 7,216 0 0 5 52 167 116 172 229 306 446

All Averaged 0 1,134 2,636 3,867 6,380 0 2 73 189 378 127 198 279 427 561

84


c) Maximum catch scenario



5% Low 25%

Med. 50%

Up 25%

Up 95%

Low 5%

Low 25%

Med. 50%

Up 25%

Up 95%

Low 5%

Low 25%

Med. 50%

Up 25%

Up 95%

MSMtB

Average 0 1 644 1919 3794 0 0 33 121 279 183 229 272 319 406 ECHO-G 0 55 996 2251 4159 0 6 83 176 342 148 188 225 266 341 CCMA 0 340 1447 2663 4737 0 46 143 240 423 110 144 175 208 271

MIROC-ESM 0 41 989 2306 4277 0 4 78 178 355 133 170 204 240 309

MSMtA

Average 145 1609 2773 4160 7149 0 0 2 36 139 174 239 299 376 507 ECHO-G 104 1453 2642 4091 7216 0 0 5 51 161 132 187 237 306 422 CCMA 44 1312 2532 3929 7130 0 0 14 69 181 91 134 175 229 322

MIROC-ESM 105 1545 2797 4257 7404 0 0 5 53 170 121 172 219 278 381

Single species 2031 2844 3651 4820 6973 0 61 170 261 437 357 421 471 531 632 MSMtB Averaged 0 66 993 2240 4311 0 5 81 184 360 129 176 216 266 353 MSMtA Averaged 66 1428 2669 4129 7216 0 0 5 52 167 116 172 229 306 446

All Averaged 0 1134 2636 3867 6380 0 2 73 189 378 127 198 279 427 561

85

BSIERP B.70 FEAST Final Report

1

2

Figure 7.1. The four future temperature time-series on which the MSMt projections are based. The constant 3 temperature is the average over time for the “hindcast” (dashed line). 4

86


5

Figure 7.2. Time-trajectories of spawning stock biomass for walleye pollock, Pacific cod and arrowtooth 6 flounder for three catch series when the projections are based on the AFSC single-species model. The bold 7 lines are distribution medians, the light shaded areas contain 50% of the distributions and the dark shaded 8 areas contain 90% of the distributions. 9

87


10

Figure 7.3. Time-trajectories of spawning stock biomass for walleye pollock, Pacific cod and arrowtooth 11 flounder (columns) for three catch series when the projections are based on the MSMtA model. The results 12 for each temperature scenario are shown as rows: average of hindcast values (a-c), ECHO-G (d-f), CCMA 13 (g-i), and MIROC-ESM (j-l). The bold lines are distribution medians, the light shaded areas contain 50% 14 of the distributions and the dark shaded areas contain 90% of the distributions. 15

88


16

Figure 7.4. Time-trajectories of spawning stock biomass for walleye pollock, Pacific cod and arrowtooth 17 flounder (columns) for three catch series when the projections are based on the MSMtB model. The results 18 for each temperature scenario are shown as rows: average of hindcast values (a-c), ECHO-G (d-f), CCMA 19 (g-i), and MIROC-ESM (j-l). The bold lines are distribution medians, the light shaded areas contain 50% 20 of the distributions and the dark shaded areas contain 90% of the distributions. 21

89


22

Figure 7.5. Model averaged results (over climate scenarios) for time-trajectories of spawning stock biomass 23 for walleye pollock, Pacific cod and arrowtooth flounder for three catch series. The bold lines are 24 distribution medians, the light shaded areas contain 50% of the distributions and the dark shaded areas 25 contain 90% of the distributions. 26

27

90


28

29

Figure 7.6. Model averaged results for time-trajectories of spawning stock biomass for walleye pollock, 30 Pacific cod and arrowtooth flounder for three catch series. The bold lines are distribution medians, the light 31 shaded areas contain 50% of the distributions and the dark shaded areas contain 90% of the distributions. 32

91


33

Figure 7.7. Results of simple yield per-recruit as related to spawning biomass per recruit based on selectivity 34 and mean body mass estimates used in the stand-alone pollock single species model and that used in the 35 single species version for pollock in the MSMt model. 36 37

38

0.0

0.1

0.2

0.3

20% 40% 60% 80% 100%

Yiel

d pe

r rec

ruit

Spawning biomass per recruit

SS

MSMt

92


Supplemental material 39 40 Table S.7.1. Model comparison for each of the stock-assessment models. 41

42

Model Data range

Ages (pollock, P.

cod, arrowtooth) Weight at age

Age specific mortality (M)

Survey age selectivity

Single species [15,12, 21] Fixed Fixed

Variable for pollock, fixed for P. cod and

arrowtooth

MSMtA 1979-2012 [12,12, 21] Annual varying with temp.

Annual varying with temp. and predator

biomass Fixed

MSMtB 1979-2012 [12,12, 21] Annual varying with temp.

Annual varying with temp. Fixed

43

44

93


45

Figure S.7.1. Survey age selectivities for each species from the single species model (solid line), MSMtA 46 (dashed line), and MSMtB (dotted line). 47

48

94


SHORT TERM FORECASTING 49 50 51 Introduction 52 53 While long-term forecasting is largely thought of as a strategy development/evaluation tool, the results it 54 provides are rarely applicable to day-to-day fisheries management. In contrast, seasonal or short-term 55 forecasting tied to climate or environmental conditions has long being seen as a way to reduce the 56 uncertainty inherent in managing aquatic resources in large systems (Pulwarty and Redmond, 1997). One 57 way of informing management may be tied to the expected distribution of a species for which bycatch needs 58 to be minimized (Hobday et al., 2011), another may be adjusting quotas depending on the observed 59 biophysical conditions under which productivity increases or decreases. Alternatively, information to the 60 fishing industry may be in the way of the expected spatial distribution of preferred environmental conditions 61 (or dynamic habitat) of commercial target species, which may contribute to increase fishing efficiency 62 Lewison, R. et al., 2015). 63 64 In the eastern Bering Sea, the cold pool extent is an ecosystem indicator that is tracked annualy based on 65 data from the bottom trawl shelf survey; it is defined by bottom temperatures less than 2°C. It is related to 66 ice extent in that sea ice creates a pool of cold bottom water on the eastern Bering Sea continental shelf, 67 and this cold pool has been shown to influence the latitudinal and longitudinal distribution of the groundfish 68 community, including several important commercial species (Kotwicki et al. 2005; Spencer, 2008; Meuter 69 and Litzow 2008, Stevenson and Lauth 2012). This in turn changes the spatial distribution of fishing effort 70 (Haynie and Pfeiffer 2012). The cold pool also influences water mixing and water stratification (Stabeno et 71 al. 2012). Studies have implicated warmer temperatures as the primary reason for changes in the distribution 72 of forage fish and benthic infauna community in the eastern Bering Sea shelf (Coyle 2007; Hollowed et al. 73 2012). Warming could have further impacts on the EBS from changing the timing of the spring 74 phytoplankton bloom, to favoring the northward advance of subarctic species and retreat of arctic species 75 (Stabeno et al. 2007, 2010, 2012). With this in mind we explored the potential of the Bering 10K ROMS-76 NPZD to predict the extent of the cold pool in the eastern Bering Sea shelf within 9 months. 77 78 Methods 79 80 The nine month forecasts were conducted by Al Hermann using the model Bering 10K ROMS-NPZD with 81 forcing files from the output of the CFSR (Climate Forecast System Reanalysis) forecast simulations. From 82 the forecast using the Bering 10K ROMS-NPZD we calculate the extent of the cold pool with scripts written 83 by Aydin. 84 85 Results 86 87 Figures 7.8 shows results from the observed vs estimated and predicted vs observed cold pool in the eastern 88 Bering Sea shelf. The prediction for 2014 (decreased cold pool) was close to the actual observed cold pool. 89 The results for the 2015 forecast predicts an even smaller cold pool (Figure 7.9). These results have been 90 presented to the Plan Team of the Bering Sea and Aleutian Islands as well as to the Ecosystem Committee 91 of the North Pacific Fisheries Management Council who have expressed interest. Further research on the 92 use of Bering 10K ROMS-NPZD output for the simulation of ecosystem indicators will be carried out by 93 the Fisheries and the Environment program (FATE) of NOAA. 94

95


Figure 7.8: Short-term forecast of the extent of the cold pool using Bering 10K ROMS-NPZD. The top 95 panel shows the modeled (hindcast) extent of the cold pool for 2012 (left) and the observed extent (right) 96 based on data from the annual bottom trawl survey for the eastern Bering szea shelf. The middle panel 97 shows the extent of the cold pool for January 2014 as predicted by the model (left) and observed based on 98 survey data (right). The bottom panel shows the predicted extent for January 2015. 99 100 101 102

96


103 Figure 7.9 Extent of the cold pool as estimated from from actual (blue line) and simulated (orange line) 104 survey data using the Bering 10K ROMS-NPZD model. Orange filled circles represent the predictions for 105 2014 and 2015. 106 107

108

References 109

110

Coyle, K.O., B. Konara, A. Blanchard, R.C. Highsmith, J. Carroll, M. Carroll, S.G. Denisenko, and B.I. 111 Sirenko. 2007. Potential effects of temperature on the benthic infaunal community on the 112 southeastern Bering Sea shelf: Possible impacts of climate change. Deep Sea Research II 54: 2885-113 2905. 114

Haynie, A.C. and L. Pfeiffer. 2012. Why economics matters for understanding the effects of climate change 115 on fisheries. ICES Journal of Marine Science. doi: 10.1093/icesjms/fss021 116

Hobday, A.J., hartog, J.R., Spillman, C.M., Alves, O. 2011. Seasonal forecasting og tuna habitat for 117 dynamic spatial management. Canadian Journal of Fisheries and Aquatic Sciences, 68:898-911. 118

Hollowed, A.B., S. Barbeaux, E.D. Cokelet, E. Farley, S. Kotwicki, P.H. Ressler, C. Spital, and C.D. 119 Wilson. 2012. Effects of climate variations on pelagic ocean habitats and their role in structuring 120 forage fish distributions in the Bering Sea. Deep Sea Research II 65-70:230-250. 121

Kotwicki, S., T. W. Buckley, T. Honkalehto, and G. Walters. 2005. Variation in the distribution of walleye 122 pollock (Theragra chalcogramma) with temperature and implications for seasonal migration. 123 Fishery Bulletin 103:574-587 124

Lewison, R., Hobday, A.J., Maxwell, S., Hazen, E., Hartog, J.R., Dunn, D.C., Briscoe, D., Fosette, S., 125 O’Keefe, C.E., barnes, M., Abecssis, M., Bograd, S., Bethoney, N.D., Bailey, H., Wiley, D., 126 Andrews, S., Hazen, L., Crowder, L.R. 2015. Dynamic ocean management: identifying critical 127 ingredients of dynamic approaches to ocean resource management. BioScience XX:1-13. 128 Doi:10.1093/biosci/biv018. 129

Mueter, F.J. and M.A. Litzow. 2008. Sea ice retreat alters the biogeography of the Bering Sea continental 130 shelf. Ecological Applications, 18: 309-320. 131

Pulwarty, R.S. and Redmond, K.T. 1997: Climate and Salmon Restoration in the Columbia River Basin: 132 The Role and Usability of Seasonal Forecasts. Bull. Amer. Meteor. Soc., 78, 381–397. 133 doi: http://dx.doi.org/10.1175/1520-0477(1997)078<0381:CASRIT>2.0.CO;2 134

97

http://dx.doi.org/10.1175/1520-0477(1997)078%3C0381:CASRIT%3E2.0.CO;2


Spencer, P. D. 2008. Density-independent and density-dependent factors affecting temporal 135 changes in spatial distributions of eastern Bering Sea flatfish. Fisheries Oceanography 136 17:396-410. 137

Stabeno, P.J., N. B. Kachel, S. E. Moore, J.M. Napp, M. Sigler, A. Yamaguchi, and A. Zerbini. 138 2012. Comparison of warm and cold years on the southeastern Bering Sea shelf and some 139 implications for the ecosystem. Deep Sea Research II, 65-7-: 31-45. 140

Stabeno, P.J., J. Napp, C. Mordy, and T. Whitledge, 2010. Factors influencing physical structure 141 and lower trophic levels of the eastern Bering Sea shelf in 2005: Sea ice, tides and winds. 142 Progress in Oceanography, 85: 180–196. 143

Stabeno, P.J., N.A. Bond and S.A. Salo. 2007. On the recent warming of the southeastern Bering 144 Sea shelf. Deep Sea Research II, 54: 2599-2618. 145

Stevenson, D.E. and R.R. Lauth. 2012. Latitudinal trends and temporal shifts in the catch 146 composition of bottom trawls conducted on the eastern Bering Sea shelf. Deep Sea 147 Research II 65-7-:251-259. 148

149

150

151

98


Chapter 8. Setting up a fully integrated ecosystem model (Climate to 152

fisheries) designed for MSE and hypothesis testing 153 154 155 Introduction 156 157 This chapter briefly describes the integrated ecosystem model Bering 10K ROMS-NPZD-FEAST-158 FAMINE as well as specific tasks and modifications required to couple and set up the model as:1) an 159 operating model for MSE and 2) a tool for hypothesis testing. The framework was also designed to pair 160 field work and historical data analysis with modeling of the different ecosystem components. Model 161 integration actions specific to each component are included in the corresponding sections below. 162 163

164 165 Figure 8.1 Structure of the integrated model Bering 10K ROMS-NPZD-FEAST-FAMINE for the Bering 166 Sea showing one way and two way feedback between components. 167 168 Each component of the model (climate, physical oceanography, lower trophic level, upper trophic level and 169 fishing effort allocation) was developed and validated independently, then recalibrated jointly once fully 170 coupled. The model can operate in hindcast or forecast mode; the differences in information flow between 171 these modes is shown in figures 8.2 and 8.3 respectively and refer to the type of climate input, fishing effort 172 input and allocation in FEAST and connection to the MSE routine. Differences or modifications pertaining 173 to each mode are further described in the sections describing each component. 174 175

99


176

177

178 Figure 8.2 Information flow in the Bering 10K ROMS-NPZD-FEAST-FAMINE model in hindcast (top) 179 and forecast (bottom) mode. 180

100


181 Climate input 182 The base run is a long term hindcast from 1970 to 2009. Two forcing datasets were utilized: the common 183 ocean reference experiment reanalysis (CORE; Large and Yeager, 2008) was used for the hindcast of years 184 1969-2004. For the hindcast of years 2003-2009 the climate forecast system reanalysis (CFSR; Saha et al 185 2010). The combination of this products is due to the fact that (at the time of the analysis), CORE products 186 were not available for years beyond 2005, while CFSR products were not available for years prior to 1979. 187 Output from years 2003 and 2004 (the overlapping years) from each hindcast was crosscompared in order 188 to adjust CFSR for compatibility with CORE. 189 190 Forecast simulations use output from the IPCC climate models: CGCM-t47 (low ice), ECHO-G (high ice) 191 and MIROCM (medium ice), all using emission scenario A1B; the forecast are for years 2003-2040 These 192 models were selected based on an evaluation of the atmospheric and sea ice components of 23 different 193 coupled atmosphere - ocean general circulation models, using the National Center for Climate Prediction 194 (NCEP) Reanalysis and observations of sea ice as validating data sets (Wang et al., 2010). The models were 195 also specifically evaluated in terms of their performance along the boundaries anticipated for the ROMS 196 modelling of the SE Bering Sea shelf. 197 198 Model integration specifics: the original period for the hindcast simulation covered 1970 to 2004. 199 However, there was a later request to extend the hindcast to field years. The hindcast was extended to 2009, 200 which required new extractions and processing of data from the climate forecast analysis mentioned above 201 (not previously used, new scripts were written), (2) new atmospheric and oceanographic boundary 202 conditions extracted, processed and formatted, (3) new atmospheric and oceanographic forcing files 203 extracted, processed and formatted and (4) new files extracted, processed and formatted for the fisheries. 204 205 206 Physical oceanography 207 The physical model is based on the Regional Ocean Modeling System as applied to the Northeast Pacific 208 and is an updated version (NEP5, Curchitser and Hedström) of the model described by Hermann et al. 209 (2009). This hydrodynamic model includes both ice and tides. The NEP domain stretches from Baja 210 California, Mexico to the Kamchatka Peninsula with a homogeneous resolution of ~10 km. A smaller grid 211 was developed for the Bering Sea (Bering 10K) that cuts off the southern portion of the NEP5 domain from 212 SE Alaska to Baja (Figure 2). Although the domain covers the entire Bering Sea, the modeling effort focuses 213 on the Eastern Bering Sea shelf with general performance of the Bering Sea domain assessed independently 214 from that of NEP5 (currently, NEP6 as of the latest verison, Daneilson et al 2001). Boundary conditions 215 for the CORE-based hindcast were taken form NEP5; for the CFSR-based hindcast, hourly values from the 216 CFSR model were filtered to 5 days averages (Hermann et al 2013). Simulations of the biophysical 217 dynamics for the Bering 10K grid were based on a reduced vertical resolution from 60 to 10 levels. 218 219 Model integration specifics:The use of the smaller Bering 10K grid and 10 levels was in response to the 220 need for faster run time than the ROMS-NEP5. Although ROMS-Bering 10K was based on NEP5, these 221 models were in fact developed in parallel, with code synchronizations as progress was made on NEP5. 222 Although part of project NSF 0732534, the development of the smaller grid and 10 levels was an unforeseen 223 addition that included modifying the code as needed in order to couple it to the models for the lower and 224 upper trophic levels. We note here that the initial simplified ROMS model had 5 levels, but the loss of 225 vertical resolution proved too high; 10 layers was preferred as it decreased the number of layers by a factor 226 of 6 while still retaining enough complexity and detailed vertical resolution. Another mayor modification 227 is the use of the vertical component (depth vector) in ROMS as the length vector for FEAST variables, as 228 FEAST variables are not depth specific. To enable this modification, FEAST variables as in fact coded as 229 dyes as opposed to true variables (like those in the physics and NPZD),. For the original 60 layer Bering 230 10K ROMS version the number of dyes would have been 30, but with 10 layers, the number of dyes 231

101


increased to 182. Other modifications pertaining to the oceanography are detailed in Hermann et al. (2013). 232 233 234

235 Figure 8.3.Model domains: NEP model (larger rectangle) provides boundary conditions for the BS model 236 (upper half of rectangle). Both models have ~10km resolution. 237 238 239 Lower trophic levels 240 The embedded Nutrient-Phytoplankton-Zooplankton (NPZD) model is an updated version of that 241 developed for the Gulf of Alaska by Hinckley et al. (2009). The state variabes include nitrate, ammonium, 242 iron, large and small phytoplankton, microzooplankton, small and large copepods (oceanic and shelf), 243 euphausiids (oceanic and shelf) and jellyfish. Most importantly, the model has been coupled to two 244 additional modules: one for ice biology (including nitrate, ammonium, and ice algae) and another for the 245 benthos (composed by benthic infauna and benthic detritus). This model was developed to run with ROMS-246 Bering 10K on 60 layers. In this mode, several one year and multi-year simulations were run. A slightly 247 modified version was used to run on 10 layers, also on the Bering 10K grid (Hermann et al., 2013). A full 248 schematic of the model is included in Figure 5.3 of Chapter 5. 249 250 251 Upper trophic levels: 252 The upper trophic levels (fish) are coupled to the oceanography and lower trophic levels through the Forage-253 Euphausiid Abundance in Space and Time model. FEAST has 15 fish groups and is designed around the 254 "landscape approach" for modeling fish foraging, mortality, and growth. The landscape approach (also 255

102


known as the dynamic habitat approach) treats the space of a model as a series of layers, each layer defining 256 a different spatial (dynamic) quantification of habitat. For example, a temperature layer, a prey density 257 layer, a prey size layer, and a mortality layer may be used, quantifying any given point for its "growth" or 258 "predation" potential thus determining growth and survival (dynamic state variables) of the fish. The 259 landscape approach has been successful, for example, at predicting the distance at which fish congregate 260 around a front; in a front between warm and cold water, warm or cold adapted fish will approach the front 261 from either side, stopping where gain from frontal concentrations of prey are cancelled out by thermal 262 stress. This can be a powerful tool for modeling dynamic climate scenarios in which fronts shift, break 263 down, or otherwise change over time. FEAST is largely designed for hypotheses testing and prediction of 264 fine scale seasonal distribution and abundance of the key species in the Bering Sea shelf under a suite of 265 climate scenarios. 266 267 FEAST basic structure 268 269 The 15 groups modeled in FEAST include: walleye pollock, arrowtooth flounder, Pacific cod, Pacific 270 herring, salmon (outgoing and incoming), capelin, sandlance, eulachon, myctophids, squids, shrimp, crabs 271 and epifauna (Figure 3). Of these, pollock, cod and arrowtooth flounder are the core species modeled in 272 high detail (11 age classes with 14 length classes, bioenergetics, and movement driven by prey, temperature, 273 and predation mortality. Herring, salmon, capelin, sandlance, eulachon and myctophids have length but not 274 age structure, and invertebrates (squids, shrimp, crabs and epifauna) are modeled as biomass pools. An 275 additional aggregated group, “miscellaneous zooplankton”, is also modeled as a biomass pool. Predation is 276 based on size, preference and availability. A schematic of the FEAST model and its structure is shown in 277 Chapter 5. The model includes two-way feedback between the NPZD and FEAST: NPZD provides 278 zooplankton prey fields to FEAST while FEAST provides fish predation mortality on zooplankton to the 279 NPZD model. Alternatively, instead of using fish predation from FEAST, the NPZD model can run with a 280 quadratic mortality term (keeps mortality proportional to biomass), which allows for the NPZD-FEAST to 281 be run with just one-way couplingand. Two-way coupling (or feedback) ensures the inclusion of both 282 bottom-up and top-down effects in restructuring the ecosystem. 283 284 Data integration in FEAST 285 286 FEAST makes extensice use of historical databases as well as processed results from other BSIERP 287 projects. Basic parameters used in FEAST (e.g. growth, size-based predation) were estimated from 288 historical geographically explicit time series and databases from regular surveys and programs conducted 289 in the Bering Sea shelf through the year, but primarily from May to September. These include, among 290 others, spatially explicit data collected from the early 1960’s to present on fisheries, stomach samples, 291 biomass, age, length, and distribution of groundfish. Survey methods include acoustics, bottom and 292 midwater and surface trawl nets among others. Fine model details such as bioenergetics and foraging 293 response have been developed in collaboration with field-biologists part of BEST-BSIERP. Catch data was 294 developed in coordination with project B.71 Economics. Table 8.2 shows data used for the different 295 parameters and processes in the model. Initial conditions were based on characterization of distribution in 296 warm or cold year based on bottom trawl survey data. Stock assessment numbers length at age were 297 proportioned across the shelf according to the mean cold/warm/neutral fish distribution. Xi-eta gridcells 298 were allocated to their corresponding 20x20 mi bottom trawl survey block station. Files were built for each 299 fish species, for every year from 1970 to 2009, however for the hindcast only 1970 was used. Input files 300 were needed both for the fisheries removals (as explained below) and to correct the recruitment (age-1) of 301 pollock, cod and arrwotooth flounder. Age-1 estimates from the stock assessment were used to scale the 302 numbers of age-zeroes at the beginning of the year back to stock assessment of age-1. 303 304 Two versions of FEAST were developed to satisfy, on one hand, the high level of oceanographical detail 305 needed to test hypotheses through emergent properties and, on the other hand, the high number of 306

103


simulations to conduct the MSE. The first version has 60 vertical layers and a 7 min timestep; the second 307 version retains most features –same species, spatial resolution (~10km) and fisheries but has only 5 vertical 308 layers and a 30 min. timestep. Figure 4 shows the sea bottom temperature as estimated by the 5 layer FEAST 309 model for August 15, 1999 (a cold year) and the corresponding density of pollock (number of fish/m2) 310 length 0-10cm.. The 60 layer version of the model is in its final development stage, while the 5 layer version 311 will be improved by incorporating correlations derived from the hindcast of the detailed version to speed 312 up runtime. The detailed 60 layer FEAST will be used for the base run and once validated, three forecasts 313 will be run forced with each of the climate models. The 5-layer model will be used to explore management 314 options and simulations needed to conduct the MSE. 315 316 317 Spatial distribution of fisheries 318 319 Hindcast: A retrospective analysis combining fish ticket data and the catch accounting system (performed 320 by B.72, Economics), provided the weekly spatially explicit catch forcing files used to incorporate fisheries 321 removals in the 1970-2009 hindcast. This task was non-trivial. Historical databases were used to: 1) 322 calculate and allocate foreign catches conducted pre-1992, and 2) ensure that post 1992 ticket data was 323 matched to that from the Catch Accounting System, which is the system that provides the catch time series 324 used in stock assessments. The fisheries included in the hindcast were defined and selected in close 325 collaboration with project B.72 and stock assessment authors (particularly for arrowtooth flounder since 326 arrowtooth and Kamchatka flounder used to be grouped in surveys, catches, and assessment; Jim Ianelli, 327 part of the B.73 MSE project, is the lead author of the Bering Sea pollock stock assessment). Final fisheries 328 groups were separated by sector (due to North Pacific Fisheries Management Council procedures during 329 catch allocations) and gears (with some gears aggregated based on their cumulative low catch, see Table 330 8.1). The catches were downscaled from the statistical areas (“Stat6” areas) used by the Alaska Department 331 of Fish and Game (ADFG) for the Bering Sea (spatial resolution for ticket data of 0.5 degree latitude by 1 332 degree longitude) to the Bering 10K grid (~10 km2). Each individual cell of the Bering 10K grid was 333 assigned membership to a Stat6 based on the degree of spatial overlap. Weekly Stat6 catch data were then 334 evenly distributed to the FEAST cells within each Stat6. The final catch data received for input into the 335 FEAST model were in tons per day with the weekly catch per FEAST cell divided by the number of days 336 per week (typically seven except for week 52). The number of FEAST cells per stat6 ranged from 1 to 29, 337 with mean of 20. 338 339 Table 8.1 Fisheries included in FEAST. 340

Sector: Catcher Processor Sector Catcher vessel Gear Species Gear Species Trawl Pollock Cod Cod Other Other Pots Cod Pots Cod Other Other Hook and Line Cod Hook and line Cod Other Other Gillnet Herring Trawl Seine Herring Pots Crab

341

104


342 Forecast: FAMINE (fishing allocation model in Nash Equilibrium) is an optimization model that 343 allocates fishing effort and catch each year. The FAMINE model specified separate fleets that 344 target walleye pollock, and Pacific cod further stratified by i) vessel class (i.e., catcher vessels, 345 catcher processors), and ii) gear type (i.e., trawl, hook-and-line, pot). There was also an ‘other’ 346 category that included arrowtooth and other flatfish. The spatial allocation of catch by fleet to area 347 is based on the assumption that all players (vessels) play an optimizing (i.e., maximizing or 348 minimizing) strategy so the catches are allocated under the assumption of a Nash Equilibrium 349 (NE). The decision rules are linear and static under the assumption that payoffs are linear-350 quadratic. Under these conditions, static profit-maximization occurs. Static profit-maximization 351 implies a set of Karush-Kuhn-Tucker (KKT) inequalities. Taken together, KKT conditions form a 352 Linear Complementarity Problem (LCP) and a Nash Equilibrium (each player plays a best-353 response), and every LCP is equivalent to a Quadratic Program (QP) 354 355 Model integration specifics: Data provided by the economics project was used to build weekly input files 356 in netCDF format. An important modification involves the way the model is run. Because ROMS does not 357 have general knowledge (the model does not track system totals, only cell-specific values), the model is 358 run at one week intervals. At the end of the week, species-length biomass estimates from the model at the 359 stat6 level are used to allocate fishing effort (based on the weekly catch input files) proportional to biomass. 360 This step is time consuming, as it requires to stop and write files every week in a simulation. 361 362 363 Management Strategy Evaluation 364 365 Formal Management Strategy Evaluation (MSE) can be conducted using the Bering 10K-ROMS-NPZD-366 FEAST-FAMINE model as the “operating model” and currently developed methods (stock assessments, 367 multispecies statistical model) as “assessment” models. The aim is to assess the skill of each model in 368 determining past and current states as well predicting future states. 369 370 Model integration specific: One of the multispecies models included in the MSE routine is an Ecopath 371 model for the EBS. The model developed for this component was a modified version based on the larger 372 EBS Ecopath model (Aydin et al., 2007). The underlying food web was simplified to replicate that depicted 373 by the groups in the NPZD and FEAST models and includes the fisheries as described by the economics 374 project. Figure 8.5 shows a schematic of the food web represented by the Ecopath model (also applicable 375 to NPZD-FEAST-FAMINE). Several scripts were also written to extract outputs from the Bering 10K 376 ROMS-NPZD-FEAST-FAMINE that simulate survey data from the annual summer bottom trawl survey, 377 the annual hyrdroacoustics survey, and catch surveys. This again requires a modification on how the model 378 is run, as the simulation needs to stop order to run the scripts at the end of the year to get aggregate totals 379 (as opposed ot the ceel-specific values) used in the MSE routine. The specifics of the implementation of 380 FEAST as the operating model of MSE are included in Chapter 6. Modifications in the FEAST code: the 381 most relevant modification was the addition of age structure. Originally, the FEAST model was only length-382 based, with each species having 20 length bins of 2cm each. FEAST tracks numbers, condition factor and 383 length for each species, which originally implied 180 variables. However, the inclusion of simulated stock 384 assessment routines as part of MSE imposed the need to track ages for the main three species: walleye 385 pollock, Pacific cod and Arrowtooth flounder, which escalated the number of variables to 1980. This was 386 not practical, so the number of length bins was reduced to 14 bins of 4cm, which brought down the number 387 of varables to 1380. The total number of FEAST variables in the model is 1817. 388 389

105


390 Figure 8.4 Schematic food web of the Ecopath model for the EBS included in the MSE routine using Bering 391 10K ROMS-NPZD-FEAST-FAMINE as the operating model. 392 393 394 FEAST functionality 395 396 There are multiple levels levels of functionality or availability of functions in key processes wthi FEAST 397 which are described below: 398 399

force_ration: values lower than zero do rations by functional response 400 force_vonB values higher than zero force growth following a von Bertalanffy curve. 401 feast_mixed with values 1 or zero; a value of zero means there is no mixed layer and the water column is 402

homogeneous throughout; 1 uses an algorithm to identify the mixed layer, effectively allowing for 403 stratification of the water column with respect to the mixed layer. 404

feast coupled with values of 1 or zero; a value of zero turns off feedback from FEAST to NPZD, meaning 405 there is no predation mortality from fish passed on to the NPZD component. A value of 1 denotes 406 feedback from FEAST to NPZD. Because NPZD and FEAST are coupled, there is always one way 407 feedback: the NPZD always provides zooplankton prey fields to FEAST. This mode only controls 408 the coupling from FEAST back to NPZD (enabling two-way feedback between NPZD and 409 FEAST). 410

feast_mort with values of 1 or zero; a value of zero means there is no added mortality, either M1 or M2. 411 A value of 1 denotes there is added mortality (in addition to predation, fishing mortality and 412 starvation). 413

feast_fishing with values of 1 or zero; a value of zero denotes there are no fisheries removals, a value of 414 one denotes there are. 415

feast_growth with values of 1 or zero; a value of zero turns off the fish growth, a value of 1 turns growth 416 on. 417

feast_recruitment with values of 1 or zero; a value of zero turns off the fish recruitment, a value of 1 turns 418

106


recruitment on. 419 feast_movement with values of 1 or zero; a value of zero turn fish movement off, a value of 1 turns it on, 420

allowing fish to move within the extent of the model. 421 422 Detailed Data integrated into FEAST 423 424 Table 8.2. Data integrated into the FEAST model parameters. Abbreviations used: POL walleye pollock, 425 COD Pacific cod, ATF arrowtooth flounder, HER Pacific herring, CAP capelin, EUL eulachon, SAN 426 Pacific sandlance, MYC myctophids, SAL1 salmon outgoing, SAL2 salmon returning, SHR shrimp 427 (pandaliid), SQU squids, EPI epifauna (motile) CRA crab (snowcrab) OTH other (miscellaneous 428 zooplankton e.g.: mysids, amphipods, chaetognats) COP s, mall copepods, NCA Neocalanus sp, NCAS 429 shelf-associated Neocalanus sp., NCAO oceanic Neocalanus sp. EUP euphausiids, EUPS shelf associated 430 euphausiids, EUPO oceanic euphausiid, BEN benthos (infauna, sessile). All parameter values are those 431 used for the hindcast simulation HC_V146_fpmask of FEAST. 432 433 parameter value description Data source/ integration

Integration with NPZD, splitting total quadratic mortality (mpred) in NPZD into quadratic mortality (mpred) and fish predation (fpred) mortality from FEAST

mpredCop 0.035 Daily mortality for copepods NPZD model

mpredNCa 0.025 Daily mortality for Neocalanus NPZD model

mpredEup 0.025 Daily mortality for euphausiids NPZD model

fpredCop 0.015 Daily mortality assumed to be from fish predation on small copepods

Tuning, this is added to parameters above to give original value in NPZD

fpredNCaS 0.025 Daily mortality assumed to be from fish predation on Neocalanus shelf


fpredNCaO 0.025 Daily mortality assumed to be from fish predation on Neocalanus oceanicf


fpredEupS 0.025 Daily mortality assumed to be from fish predation on Euphausiids shelf


fpredEupO 0.025 Daily mortality assumed to be from fish predation on Euphausiids shelf


Fsh_age_offset (pollock) [0,0,2,3,5,6,6,7,7,8,9]

Minimum length bin at age vector (ages 0-10)

Based on length at age data from RACEBASE and BASIS

Fsh_age_offset (P.cod) [0,0,2,4,8,9,11,12,13,14,15]



Fsh_age_offset (arrowtooth)

[0,0,0,3,4,5,6,7,7,8,9]



Size bin for aged fish, age 0 bins are half width of age 1+ bins

Fsh_Lsize pol 4 Length of size bin in cm (15 bins) Minimum to capture RACE length at age range

Fsh_Lsize cod 4 Length of size bin in cm (15 bins) Minimum to capture RACE length at age range

Fsh_Lsize atf 4 Length of size bin in cm (15 bins) Minimum to capture RACE length at age range

107


Fsh_Lsize her 2 Length of size bin in cm (20 bins) Minimum to capture RACE length at age range

Fsh_Lsize cap 2 Length of size bin in cm (20 bins) Minimum to capture RACE length at age range

Fsh_Lsize eul 2 Length of size bin in cm (20 bins) Minimum to capture RACE length at age range

Fsh_Lsize san 2 Length of size bin in cm (20 bins) Minimum to capture RACE length at age range

Fsh_Lsize myc 2 Length of size bin in cm (20 bins) Minimum to capture RACE length at age range

Fsh_Lsize sal1 2 Length of size bin in cm (20 bins) Minimum to capture RACE length at age range

Fsh_Lsize_sal2 2 Length of size bin in cm (20 bins) Minimum to capture RACE length at age range

Fsh base prey (all species) 16 Number of length based prey species available

Model structure

Vertical water column preference (not used)

Fsh_a_T pol 0.45 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_a_T cod .5 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_a_T atf .45 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_a_T herl 1 Proportion of time spent in layer No data; assumed to be all time

Fsh_a_T cap 1 Proportion of time spent in layer No data; assumed to be all time

Fsh_a_T eul 1 Proportion of time spent in layer No data; assumed to be all time

Fsh_a_T san 1 Proportion of time spent in layer No data; assumed to be all time

Fsh_a_T myc 1 Proportion of time spent in layer No data; assumed to be all time

Fsh_a_T sal1 1 Proportion of time spent in layer No data; assumed to be all time

Fsh_a_T sal2 1 Proportion of time spent in layer No data; assumed to be all time

Fsh_b_T pol .1 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_b_T cod .15 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_b_T atf .1 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_b_T herl 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_b_T cap 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_b_T eul 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_b_T san 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_b_T myc 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_b_T sal1 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_b_T sal2 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_c_T pol .45 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_c_T cod .8 Proportion of time spent in layer Survey data & pers comm Troy Buckley

108


Fsh_c_T atf .45 Proportion of time spent in layer Survey data & pers comm Troy Buckley

Fsh_c_T herl 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_c_T cap 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_c_T eul 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_c_T san 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_c_T myc 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_c_T sal1 0 Proportion of time spent in layer No data; assumed to be all time

Fsh_c_T sal2 0 Proportion of time spent in layer No data; assumed to be all time

Length-weight relationships based on Von Bertalanffy curve fitting (needs adjustment for model numbers to tons and back)

fsh_A_L pol 0.005531 Parameter a of L-W converison Baseline using all RACE/BASIS data

fsh_A_L cod 0.004118 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L atf 0.004439 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L herl 0.008594 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L cap 0.00034 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L eul 0.004439 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L san 0.005816 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L myc 0.005816 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L sal1 0.005816 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L sal2 0.005816 Parameter a of L-W converison Baseline using all RACE data

fsh_A_L pol 3.044172 Parameter b of L-W converison Baseline using all RACE/BASIS data

fsh_B_L cod 3.253258 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L atf 3.19894 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L her 3.107793 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L cap 4.2304 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L eul 3.19894 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L san 3.0294 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L myc 3.0294 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L sal1 3.0294 Parameter b of L-W converison Baseline using all RACE data

fsh_B_L sal2 3.0294 Parameter b of L-W converison Baseline using all RACE data

Functional response (initial values here can improve by fitting values)

fsh_Bv_min pol 0.01 Assumed values to find all prey all sizes

fsh_Bv_min cod 0.01 Assumed values to find all prey all sizes

109


fsh_Bv_min atf 0.01 Assumed values to find all prey all sizes

fsh_Bv_min her 0.01 Assumed values to find all prey all sizes

fsh_Bv_min cap 0.01 Assumed values to find all prey all sizes

fsh_Bv_min eul 0.01 Assumed values to find all prey all sizes

fsh_Bv_min san 0.01 Assumed values to find all prey all sizes

fsh_Bv_min myc 0.01 Assumed values to find all prey all sizes

fsh_Bv_min sal1 0.01 Assumed values to find all prey all sizes

fsh_Bv_min sal2 0.01 Assumed values to find all prey all sizes

fsh_B_Lzero pol 1000 Assumed values to find all prey all sizes

fsh_B_Lzero cod 1000 Assumed values to find all prey all sizes

fsh_B_Lzero atf 1000 Assumed values to find all prey all sizes

fsh_B_Lzero her 1000 Assumed values to find all prey all sizes

fsh_B_Lzero cap 1000 Assumed values to find all prey all sizes

fsh_B_Lzero eul 1000 Assumed values to find all prey all sizes

fsh_B_Lzero san 1000 Assumed values to find all prey all sizes

fsh_B_Lzero myc 1000 Assumed values to find all prey all sizes

fsh_B_Lzero sal1 1000 Assumed values to find all prey all sizes

fsh_B_Lzero sal2 1000 Assumed values to find all prey all sizes

fsh_B_Lone pol 30 Assumed values to find all prey all sizes

fsh_B_Lone cod 30 Assumed values to find all prey all sizes

fsh_B_Lone atf 30 Assumed values to find all prey all sizes

fsh_B_Lone her 30 Assumed values to find all prey all sizes

fsh_B_Lone cap 30 Assumed values to find all prey all sizes

fsh_B_Lone eul 30 Assumed values to find all prey all sizes

fsh_B_Lone san 30 Assumed values to find all prey all sizes

fsh_B_Lone myc 30 Assumed values to find all prey all sizes

fsh_B_Lone sal1 30 Assumed values to find all prey all sizes

fsh_B_Lone sal2 30 Assumed values to find all prey all sizes

fsh_B_Lpow pol 2 Assumed values to find all prey all sizes

fsh_B_Lpow cod 2 Assumed values to find all prey all sizes

110


fsh_B_Lpow atf 2 Assumed values to find all prey all sizes

fsh_B_Lpow her 2 Assumed values to find all prey all sizes

fsh_B_Lpow cap 2 Assumed values to find all prey all sizes

fsh_B_Lpow eul 2 Assumed values to find all prey all sizes

fsh_B_Lpow san 2 Assumed values to find all prey all sizes

fsh_B_Lpow myc 2 Assumed values to find all prey all sizes

fsh_B_Lpow sal1 2 Assumed values to find all prey all sizes

fsh_B_Lpow sal2 2 Assumed values to find all prey all sizes

fsh_A_enc pol 0.001 Assumed values to find all prey all sizes

fsh_A_enc cod 0.002 Assumed values to find all prey all sizes

fsh_A_enc atf 0.001 Assumed values to find all prey all sizes

fsh_A_enc herl 0.001 Assumed values to find all prey all sizes

fsh_A_enc cap 0.001 Assumed values to find all prey all sizes

fsh_A_enc eul 0.001 Assumed values to find all prey all sizes

fsh_A_enc san 0.001 Assumed values to find all prey all sizes

fsh_A_enc myc 0.001 Assumed values to find all prey all sizes

fsh_A_enc sal1 0.001 Assumed values to find all prey all sizes

fsh_A_enc sal2 0.001 Assumed values to find all prey all sizes

fsh_B_enc pol 2 Assumed values to find all prey all sizes

fsh_B_enc cod 2 Assumed values to find all prey all sizes

fsh_B_enc atf 2 Assumed values to find all prey all sizes

fsh_B_enc herl 2 Assumed values to find all prey all sizes

fsh_B_enc cap 2 Assumed values to find all prey all sizes

fsh_B_enc eul 2 Assumed values to find all prey all sizes

fsh_B_enc san 2 Assumed values to find all prey all sizes

fsh_B_enc myc 2 Assumed values to find all prey all sizes

fsh_B_enc sal1 2 Assumed values to find all prey all sizes

fsh_B_enc sal2 2 Assumed values to find all prey all sizes

Maximum stomach size

fsh_A_S pol 0.072024

Parameter a Length-Stomach size relationship

Estimated from REEM data stomach size and digestion rate

111


fsh_A_S cod 0.01



fsh_A_S atf 0.005



fsh_A_S herl 0.000565


From REEM data stomach size and digestion rate; assumed for all small fish

fsh_A_S cap 0.000565



fsh_A_S eul 0.000565



fsh_A_S san 0.000565



fsh_A_S myc 0.000565



fsh_A_S sal1 0.000565



fsh_A_S sal2 0.000565



fsh_B_S pol 1.2

Parameter b Length-Stomach size relationship


fsh_B_S cod 2



fsh_B_S atf 1.7



fsh_B_S herl 3.158436



fsh_B_S cap 3.158436



fsh_B_S eul 3.158436



fsh_B_S san 3.158436



fsh_B_S myc 3.158436



fsh_B_S sal1 3.158436



fsh_B_S sal2 3.158436



Standard bioenergetics consumption (Wisconsin model) parameters

Fsh_C_TM pol 15 H20 temp. at which consumption ceases Cianelli et al.

Fsh_C_TM cod 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_TM atf 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_TM her 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

112


Fsh_C_TM cap 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_TM eul 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_TM san 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_TM myc 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_TM sal1 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_TM sal2 15 H20 temp. at which consumption ceases Wisconsin model Hanson et al

Fsh_C_T0 pol 10 Temperature of max consumption rate Cianelli et al.

Fsh_C_T0 cod 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 atf 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 her 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 cap 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 eul 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 san 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 myc 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 sal1 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_T0 sal2 10 Temperature of max consumption rate Wisconsin model Hanson et al

Fsh_C_Q pol 2.6 Temperature dependendence coefficient Cianelli et al.

Fsh_C_Q cod 1.88 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q atf 5.5 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q her 4.6 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q cap 4.6 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q eul 4.6 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q san 4.6 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q myc 4.6 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q sal1 4.6 Temperature dependendence coefficient Wisconsin model Hanson et al

Fsh_C_Q sal2 4.6 Temperature dependendence coefficient Wisconsin model Hanson et al

Caloric density (joules/g ww) size based if applicable or fixed (ED = energy density)

Fsh_ED_m pol 38.75 Slope value for size-dependent ED Based on pollock; Buckley & Livingston

Fsh_ED_m cod 38.75 Slope value for size-dependent ED Based on pollock; Buckley & Livingston

Fsh_ED_m atf 38.75 Slope value for size-dependent ED Based on pollock; Buckley & Livingston

113


Fsh_ED_m her 0 Not size based No data?

Fsh_ED_m cap 0 Not size based No data?

Fsh_ED_m eul 0 Not size based No data?

Fsh_ED_m san 0 Not size based No data?

Fsh_ED_m myc 0 Not size based No data?

Fsh_ED_m sal1 0 Not size based No data?

Fsh_ED_m sal2 0 Not size based No data?

Fsh_ED_b pol 2500

Intercept value for size-dependent ED pollock Buckley & Livingston 1994 NOAA Tech Memo NMFS-AFSC-37

Fsh_ED_b cod 2500 Intercept value for size-dependent ED pollock; Buckley & Livingston 1994

Fsh_ED_b atf 2500 Intercept value for size-dependent ED pollock; Buckley & Livingston 1994

Fsh_ED_b her 4499 Not size based Assumed for average for small fish

Fsh_ED_b cap 4499 Not size based Assumed for average for small fish

Fsh_ED_b eul 4499 Not size based Assumed for average for small fish

Fsh_ED_b san 4499 Not size based Assumed for average for small fish

Fsh_ED_b myc 4499 Not size based Assumed for average for small fish

Fsh_ED_b sal1 4499 Not size based Assumed for average for small fish

Fsh_ED_b sal2 4499 Not size based Assumed for average for small fish

Respiration bioenergetics substract 1 for standard g/g/day units on B_R

Fsh_A_R pol 0.0075 Parameter of allometric funciton Cianelli et al.

Fsh_A_R cod 0.008 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R atf 0.0057 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R her 0.0195 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R cap 0.0195 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R eul 0.0195 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R san 0.0195 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R myc 0.0195 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R sal1 0.0195 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_A_R sal2 0.0195 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R pol 0.749 Parameter of allometric funciton Cianelli et al.

Fsh_B_R cod 0.828 Parameter of allometric funciton Wisconsin model Hanson et al

114


Fsh_B_R atf 0.656 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R her 0.74 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R cap 0.74 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R eul 0.74 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R san 0.74 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R myc 0.74 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R sal1 0.74 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_B_R sal2 0.74 Parameter of allometric funciton Wisconsin model Hanson et al

Fsh_F_A pol 0.15 Fraction consumed food egested Cianelli et al.

Fsh_F_A cod 0.17 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A atf 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A her 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A cap 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A eul 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A san 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A myc 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A sal1 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_F_A sal2 0.2 Fraction consumed food egested Wisconsin model Hanson et al

Fsh_U_A pol 0.11 Fraction of food excreted Cianelli et al.

Fsh_U_A cod 0.09 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A atf 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A her 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A cap 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A eul 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A san 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A myc 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A sal1 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_U_A sal2 0.111 Fraction of food excreted Wisconsin model Hanson et al

Fsh_SDA pol 0.125 Specific dynamic action Cianelli et al.

Fsh_SDA cod 0.17 Specific dynamic action Wisconsin model Hanson et al

115


Fsh_SDA atf 0.161 Specific dynamic action Wisconsin model Hanson et al

Fsh_SDA her 0.125 Specific dynamic action Wisconsin model Hanson et al

Fsh_SDA cap 0.125 Specific dynamic action Wisconsin model Hanson et al

Fsh SDA eul 0.125 Specific dynamic action Wisconsin model Hanson et al

Fsh_SDA san 0.125 Specific dynamic action Wisconsin model Hanson et al

Fsh_SDA myc 0.125 Specific dynamic action Wisconsin model Hanson et al

Fsh_SDA sal1 0.125 Specific dynamic action Wisconsin model Hanson et al

Fsh_SDA sal2 0.125 Specific dynamic action Wisconsin model Hanson et al

Fsh_R_TM pol 18 Lethal water temperature Cianelli et al.

Fsh_R_TM cod 24 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM atf 24.9 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM her 18 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM cap 18 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM eul 18 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM san 18 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM myc 18 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM sal1 18 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_TM sal2 18 Lethal water temperature Wisconsin model Hanson et al

Fsh_R_T0 pol 13 Temp. of max. respiration rate Cianelli et al.

Fsh_R_T0 cod 21 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 atf 20.9 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 her 15 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 cap 15 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 eul 15 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 san 15 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 myc 15 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 sal1 15 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_T0 sal2 15 Temp. of max. respiration rate Wisconsin model Hanson et al

Fsh_R_Q pol 2.6 Temperature dependence coefficient Cianelli et al.

Fsh_R_Q cod 1.88 Temperature dependence coefficient Wisconsin model Hanson et al

116


Fsh_R_Q atf 5.5 Temperature dependence coefficient Wisconsin model Hanson et al

Fsh_R_Q her 4.6 Temperature dependence coefficient Wisconsin model Hanson et al

Fsh_R_Q cap 4.6 Temperature dependence coefficient Wisconsin model Hanson et al

Fsh_R_Q eul 4.6 Temperature dependence coefficient Wisconsin model Hanson et al

Fsh_R_Q san 4.6 Temperature dependence coefficient Wisconsin model Hanson et al

Fsh_R_Q myc 4.6 Temperature dependence coefficient Wisconsin model Hanson et al

Fsh_R_Q sal1 4.6 Temperature dependence coefficient Wisconsin model Hanson et al

Fsh_R_Q sal2 4.6 Temperature dependence coefficient Wisconsin model Hanson et al

Additional mortality (estimates and or fitted values other than predation and/or fisheries mortsality and starvation )

Fsh_omega pol 0.1 Other mortality Assumed values

Fsh_omega cod 0.2 Other mortality Assumed values

Fsh_omega atf 0.03 Other mortality Assumed values

Fsh_omega her 0.1 Other mortality Assumed values

Fsh_omega cap 0.1 Other mortality Assumed values

Fsh_omega eul 0.1 Other mortality Assumed values

Fsh_omega san 0.1 Other mortality Assumed values

Fsh_omega myc 0.1 Other mortality Assumed values

Fsh_omega sal1 0.1 Other mortality Assumed values

Fsh_omega sal2 0.1 Other mortality Assumed values

Fsh_mu pol 1 Other mortality Assumed values

Fsh_mu cod 1 Other mortality Assumed values

Fsh_mu atf 1 Other mortality Assumed values

Fsh_mu her 1 Other mortality Assumed values

Fsh_mu cap 1 Other mortality Assumed values

Fsh_mu eul 1 Other mortality Assumed values

Fsh_mu san 1 Other mortality Assumed values

Fsh_mu myc 1 Other mortality Assumed values

Fsh_mu sal1 1 Other mortality Assumed values

Fsh_mu sal2 1 Other mortality Assumed values

Fsh_zeta (all species) 0 Other mortality Assumed values

Fsh_s_mega (all species) 0 Other mortality Assumed values

117


Growth partitioning

Fsh_g_W (all species) 1 Partition of length vs weight gain Assumed the same for all species

Recruitment functions (length based logistic curves, maturity and fecundity separate

Fsh_mat_a pol -7.8 Parameter a of logistic maturity Hinckley & various published literature

Fsh_mat_a cod -12.4 Parameter a of logistic maturity Various published literature

Fsh_mat_a atf -9.4 Parameter a of logistic maturity Various published literature

Fsh_mat_a her -12.5 Parameter a of logistic maturity Various published literature

Fsh_mat_a cap -6 Parameter a of logistic maturity Various published literature

Fsh_mat_a eul -10 Parameter a of logistic maturity Various published literature

Fsh_mat_a san -7.5 Parameter a of logistic maturity Various published literature

Fsh_mat_a myc -10 Parameter a of logistic maturity Various published literature

Fsh_mat_a sal1 0 Parameter a of logistic maturity Not in model (abundance 0)

Fsh_mat_a sal2 0 Parameter a of logistic maturity Not in model (abundance 0)

Fsh_mat_b pol 0.2 Parameter b of logistic maturity Hinckley & various published literature

Fsh_mat_b cod 0.2 Parameter b of logistic maturity Various published literature

Fsh_mat_b atf 0.2 Parameter b of logistic maturity Various published literature

Fsh_mat_b her 0.5 Parameter b of logistic maturity Various published literature

Fsh_mat_b cap 0.5 Parameter b of logistic maturity Various published literature

Fsh_mat_b eul 0.5 Parameter b of logistic maturity Various published literature

Fsh_mat_b san 0.5 Parameter b of logistic maturity Various published literature

Fsh_mat_b myc 0.5 Parameter b of logistic maturity Various published literature

Fsh_mat_b sal1 0 Parameter b of logistic maturity Not in model (abundance 0)

Fsh_mat_b sal2 0 Parameter b of logistic maturity Not in model (abundance 0)

Fsh_fec_a pol -15 Parameter a of logistic fecundity Hinckley & various published literature

Fsh_fec_a cod -13.4 Parameter a of logistic fecundity Various published literature

Fsh_fec_a atf -19.65 Parameter a of logistic fecundity Various published literature

Fsh_fec_a her -12.5 Parameter a of logistic fecundity Various published literature

Fsh_fec_a cap -6 Parameter a of logistic fecundity Various published literature

Fsh_fec_a eul -10 Parameter a of logistic fecundity Various published literature

Fsh_fec_a san -7.5 Parameter a of logistic fecundity Various published literature

Fsh_fec_a myc -10 Parameter a of logistic fecundity Various published literature

118


Fsh_fec_a sal1 0 Parameter a of logistic fecundity Not in model (abundance 0)

Fsh_fec_a sal2 0 Parameter a of logistic fecundity Not in model (abundance 0)

Fsh_fec_b pol 0.3 Parameter b of logistic fecundity Hinckley & various published literature

Fsh_fec_b cod 0.2 Parameter b of logistic fecundity Various published literature

Fsh_fec_b atf 0.3 Parameter b of logistic fecundity Various published literature

Fsh_fec_b her 0.5 Parameter b of logistic fecundity Various published literature

Fsh_fec_b cap 0.5 Parameter b of logistic fecundity Various published literature

Fsh_fec_b eul 0.5 Parameter b of logistic fecundity Various published literature

Fsh_fec_b san 0.5 Parameter b of logistic fecundity Various published literature

Fsh_fec_b myc 0.5 Parameter b of logistic fecundity Various published literature

Fsh_fec_b sal1 0 Parameter b of logistic fecundity Not in model (abundance 0)

Fsh_fec_b sal2 0 Parameter b of logistic fecundity Not in model (abundance 0)

Fsh_fec_max pol 0.2 Max. fecundity of logistic fecundity Hinckley & various published literature

Fsh_fec_max cod 0.045 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max atf 0.073058 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max her 0.1 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max cap 0.1 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max eul 0.1 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max san 0.1 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max myc 0.1 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max sal1 0.1 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_max sal2 0.1 Max. fecundity of logistic fecundity Various published literature

Fsh_fec_prop pol 0.5 Female proportion of population Assumed

Fsh_fec_prop cod 0.5 Female proportion of population Assumed

Fsh_fec_prop atf 0.5 Female proportion of population Assumed

Fsh_fec_prop her 0.5 Female proportion of population Assumed

Fsh_fec_prop cap 0.5 Female proportion of population Assumed

Fsh_fec_prop eul 0.5 Female proportion of population Assumed

Fsh_fec_prop san 0.5 Female proportion of population Assumed

Fsh_fec_prop myc 0.5 Female proportion of population Assumed

119


Fsh_fec_prop sal1 0 Assumed

Fsh_fec_prop sal2 - Assumed

Fsh_rec_prop pol 0.28384 Proportion recruits from eggs produced Assumed

Fsh_rec_prop cod 0.29 Proportion recruits from eggs produced Assumed

Fsh_rec_prop atf 1.36 Proportion recruits from eggs produced Assumed

Fsh_rec_prop her 14.192 Proportion recruits from eggs produced Assumed

Fsh_rec_prop cap 14.192 Proportion recruits from eggs produced Assumed

Fsh_rec_prop eul 14.192 Proportion recruits from eggs produced Assumed

Fsh_rec_prop san 14.192 Proportion recruits from eggs produced Assumed

Fsh_rec_prop myc 14.192 Proportion recruits from eggs produced Assumed

Fsh_rec_prop sal1 0 Proportion recruits from eggs produced Assumed

Fsh_rec_prop sal2 0 Proportion recruits from eggs produced Assumed

Fsh_z_muL pol Mean length age zero hatched BASIS

Fsh_z_muL cod Mean length age zero hatched BASIS

Fsh_z_muL atf Mean length age zero hatched BASIS

Fsh_z_muL her Mean length age zero hatched No recruitment

Fsh_z_muLcap Mean length age zero hatched No recruitment

Fsh_z_muL eul Mean length age zero hatched No recruitment

Fsh_z_muL san Mean length age zero hatched No recruitment

Fsh_z_muL myc Mean length age zero hatched No recruitment

Fsh_z_muLsal1 Mean length age zero hatched No recruitment

Fsh_z_muL sal2 Mean length age zero hatched No recruitment

Fsh_z_sdL pol Standarde dev of length age-0 hatched BASIS

Fsh_z_sdL cod Standarde dev of length age-0 hatched BASIS

Fsh_z_sdL atf Standarde dev of length age-0 hatched BASIS

Fsh_z_sdL her Standarde dev of length age-0 hatched No recruitment

Fsh_z_sdLcap Standarde dev of length age-0 hatched No recruitment

Fsh_z_sdL eul Standarde dev of length age-0 hatched No recruitment

Fsh_z_sdL san Standarde dev of length age-0 hatched No recruitment

Fsh_z_sd myc Standarde dev of length age-0 hatched No recruitment

Fsh_z_sdLsal1 Standarde dev of length age-0 hatched No recruitment

Fsh_z_sdL sal2 Standarde dev of length age-0 hatched No recruitment

Fsh_sp_sday pol First spawning day (julian) BASIS

120


Fsh_sp_sday cod First spawning day (julian) BASIS

Fsh_sp_sday atf First spawning day (julian) BASIS

Fsh_sp_sday her First spawning day (julian) No recruitment

Fsh_sp_sday cap First spawning day (julian) No recruitment

Fsh_sp_sday eul First spawning day (julian) No recruitment

Fsh_sp_sday san First spawning day (julian) No recruitment

Fsh_sp_sday myc First spawning day (julian) No recruitment

Fsh_sp_sday sal1 First spawning day (julian) No recruitment

Fsh_sp_sday sal2 First spawning day (julian) No recruitment

Fsh_sp_eday pol Last spawning day (julian) BASIS

Fsh_sp_eday cod Last spawning day (julian) BASIS

Fsh_sp_eday atf Last spawning day (julian) BASIS

Fsh_sp_eday her Last spawning day (julian) No recruitment

Fsh_sp_eday cap Last spawning day (julian) No recruitment

Fsh_sp_eday eul Last spawning day (julian) No recruitment

Fsh_sp_eday san Last spawning day (julian) No recruitment

Fsh_sp_eday myc Last spawning day (julian) No recruitment

Fsh_sp_eday sal1 Last spawning day (julian) No recruitment

Fsh_sp_eday sal2 Last spawning day (julian) No recruitment

Fsh_z_sday pol First hatching day BASIS

Fsh_z_sday cod First hatching day BASIS

Fsh_z_sday atf First hatching day BASIS

Fsh_z_sday her First hatching day No recruitment

Fsh_z_sday cap First hatching day No recruitment

Fsh_z_sday eul First hatching day No recruitment

Fsh_z_sday san First hatching day No recruitment

Fsh_z_sday myc First hatching day No recruitment

Fsh_z_sday sal1 First hatching day No recruitment

Fsh_z_sday sal2 First hatching day No recruitment

Fsh_z_eday pol Last hatching day BASIS

Fsh_z_eday cod Last hatching day BASIS

Fsh_z_eday atf Last hatching day BASIS

Fsh_z_eday her Last hatching day No recruitment

Fsh_z_eday cap Last hatching day No recruitment

Fsh_z_eday eul Last hatching day No recruitment

121


Fsh_z_eday san Last hatching day No recruitment

Fsh_z_eday myc Last hatching day No recruitment

Fsh_z_eday sal1 Last hatching day No recruitment

Fsh_z_eday sal2 Last hatching day No recruitment

Happiness (movement) max swim speed body lengths/sec (happy values are arbitrary values to standardize values)

Fsh_max_speed (all species )

1

Fsh_happy_01 (all species) -0.1

Fsh_happy_99 (all species) 0.1

434 Prey size and energetic content 435

SHR SQU EPI CRA OTH Fsh_simple_length Mean length

cm 2 10 1 2 2

Fsh_simple_JperG Cal/gWW 3000 3000 3000 3000 3000 COP NCAS NCAO EUPS EUPO BEN Fsh_zoop_len Mean length

cm 0.1 0.35 0.60 1.70 1.70 1

Fsh_zoop_JperG Cal/gWW 2500 3000 3500 4000 4000 2929 436 Fish diet parameters 437 Estimated first from food habits data and bottom trawl survey biomass, then tuned. A row with the same 438 value across means diet us based on prey abundance 439 440 Predator-prey information matrices: rows predators (1=pol, 2=cod, 3=atf, 4=her, 5=cap, 6=eul, 7=san, 441 8=myc, 9=sal1, 10=sal2); columns are prey. 442 443 pred Pol Cod Atf Her Cap Eul San Myc Sal1 Sal2 Shr Squ Epi Cra Oth 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fsh_q_G 1 1 0.1 0.01 o.4 0.4 0.4 0.1 3 2 2 1.95 4 .06 .001 .2 Fsh_q_G 2 0.2 0.3 0.2 1 1 1 .25 1 1 1 0.1 1 0.5 2 .5 Fsh_q_G 3 1.2 0.4 0.2 1 1 1 .25 1 1 1 0.1 1 0.01 0.0001 .3 Fsh_q_G 4 0.1 0.1 0.1 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.05 2 0.001 0.0001 0.2 Fsh_q_G 5 0.1 0.1 0.1 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.05 0.5 0.001 0.0001 0.2 Fsh_q_G 6 0.1 0.1 0.1 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.05 0.5 0.001 0.0001 0.2 Fsh_q_G 7 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 1 0.001 0.2 Fsh_q_G 8 0.3 0.3 0.3 2 2 2 2 2 2 2 0.05 2 0.001 0.0001 2 Fsh_q_G 9 1 1 1 2 2 2 2 2 2 2 0.05 2 0.001 0.0001 .2 Fsh_q_G 10 1 1 1 2 2 2 2 2 2 2 0.05 2 0.001 0.0001 .2 Fsh_q_G 11 0.1 0.1 0.1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_q_G 12 0.4 0.3 0.1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_q_G 13 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_q_G 14 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_q_G 15 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 444

Pred Cop NCas NCao Eups Eupo Ben Fsh_q_G 1 5 3 3 6 6 0.0001 Fsh_q_G 2 2 2 2 0.5 0.5 0.001 Fsh_q_G 3 2 2 2 2 2 0.002 Fsh_q_G 4 5 2 2 4 4 0.0001 Fsh_q_G 5 5 2 2 4 4 0.0001 Fsh_q_G 6 5 2 2 4 4 0.0001 Fsh_q_G 7 5 2 2 4 4 0.0001 Fsh_q_G 8 5 2 2 4 4 0.0001

122


Fsh_q_G 9 5 2 2 4 4 0.0001 Fsh_q_G 10 5 2 2 4 4 0.0001 Fsh_q_G 11 1 1 1 1 1 0 Fsh_q_G 12 1 1 1 1 1 0 Fsh_q_G 13 1 1 1 1 1 0 Fsh_q_G 14 1 1 1 1 1 0 Fsh_q_G 15 1 1 1 1 1 0

445 446 pred Pol Cod Atf Her Cap Eul San Myc Sal1 Sal2 Shr Squ Epi Cra Oth 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fsh_alfa_G 1 16.975 12.975 12.975 16.975 16.975 12.975 16.975 160.73 12.975 12.975 234.73 140.73 50.975 12.975 228.62 Fsh_alfa_G 2 5.426 5.426 5.426 5.426 5.426 5.426 5.426 5.426 5.426 5.426 41.372 23.502 80.135 133.02 12.505 Fsh_alfa_G 3 3 6.6279 6.6279 6.6279 6.6279 6.6279 6.6279 6.6279 6.6279 6.6279 27.298 3.4421 427.95 21.797 5.6126 Fsh_alfa_G 4 18.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 194.73 3.3176 12.975 12.975 228.62 Fsh_alfa_G 5 18.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 194.73 3.3176 12.975 12.975 228.62 Fsh_alfa_G 6 18.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 194.73 3.3176 12.975 12.975 228.62 Fsh_alfa_G 7 18.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 194.73 3.3176 12.975 12.975 228.62 Fsh_alfa_G 8 18.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 194.73 3.3176 12.975 12.975 228.62 Fsh_alfa_G 9 18.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 194.73 3.3176 12.975 12.975 228.62 Fsh_alfa_G 10 18.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 12.975 194.73 3.3176 12.975 12.975 228.62 447 448

Pred Cop NCas NCao Eups Eupo Ben Fsh_alfa_G 1 30.805 41.922 35.662 22.503 22.503 12.975 Fsh_alfa_G 2 30.805 41.922 35.662 19.322 19.322 98.658 Fsh_alfa_G 3 289.89 144.95 97.844 6.3573 6.3573 61.732 Fsh_alfa_G 4 27 35 34.662 26 26 12.975 Fsh_alfa_G 5 27 35 34.662 26 26 12.975 Fsh_alfa_G 6 27 35 34.662 26 26 12.975 Fsh_alfa_G 7 27 35 34.662 26 26 12.975 Fsh_alfa_G 8 27 35 34.662 26 26 12.975 Fsh_alfa_G 9 27 35 34.662 26 26 12.975 Fsh_alfa_G 10 27 35 34.662 26 26 12.975

449 450 pred Pol Cod Atf Her Cap Eul San Myc Sal1 Sal2 Shr Squ Epi Cra Oth 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fsh_beta_G 1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0154 0.1647 0.1647 0.0174 0.0174 0.0847 0.1647 0.0162 Fsh_beta_G 2 0.256 0.256 0.256 0.256 0.256 0.256 0.256 0.256 0.256 0.256 0.0716 0.1104 0.0484 0.0257 0.1669 Fsh_beta_G 3 0.5 0.1863 0.1863 0.1863 0.1863 0.1863 0.1863 0.1863 0.1863 0.1863 0.0813 0.659 0.009 0.1635 0.3045 Fsh_beta_G 4 0.1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0174 0.7731 0.1647 0.1647 0.0162 Fsh_beta_G 5 0.1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0174 0.7731 0.1647 0.1647 0.0162 Fsh_beta_G 6 0.1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0174 0.7731 0.1647 0.1647 0.0162 Fsh_beta_G 7 0.1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0174 0.7731 0.1647 0.1647 0.0162 Fsh_beta_G 8 0.1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0174 0.7731 0.1647 0.1647 0.0162 Fsh_beta_G 9 0.1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0174 0.7731 0.1647 0.1647 0.0162 Fsh_beta_G 10 0.1 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.1647 0.0174 0.7731 0.1647 0.1647 0.0162 451 452

Pred Cop NCas NCao Eups Eupo Ben Fsh_beta_G 1 30.805 41.922 35.662 22.503 22.503 12.975 Fsh_beta_G 2 30.805 41.922 35.662 19.322 19.322 98.658 Fsh_beta_G 3 289.89 144.95 97.844 6.3573 6.3573 61.732 Fsh_beta_G 4 27 35 34.662 26 26 12.975 Fsh_beta_G 5 27 35 34.662 26 26 12.975 Fsh_beta_G 6 27 35 34.662 26 26 12.975 Fsh_beta_G 7 27 35 34.662 26 26 12.975 Fsh_beta_G 8 27 35 34.662 26 26 12.975 Fsh_beta_G 9 27 35 34.662 26 26 12.975 Fsh_beta_G 10 27 35 34.662 26 26 12.975

453 454

123


Fisheries parameters 455 Number of gears by sector: 16 where CV stands for Catcher Vessel and CP stands for Catcher Processor; 456 PLCK=pollock, PCOD=cod, OTHR=target species other than pollock or cod, HERR=herring; TWL= trawl, 457 HAL=hook and line, POT=pots, GILL= gillnets, SEINE=seine Catchability is standardized (0 to 1 values), 458 is length-based and follows a logistic curve. Fisheries structure developed in collaboration with economics 459 division at the AFSC based on data from the Catch Accountability System and the ticket data from Alaska 460 Department of Fish and Wildlife. 461 462 Gear Pol Cod Atf Her Cap Eul San Myc Sal1 Sal2 Shr Squ Epi Cra Oth CP_PLCK_TWL 1 Fsh_catch_sel 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 1 26 30 23 19 19 19 19 19 19 19 19 19 19 19 19 Fsh_catch_99 1 61 108 75 36 36 36 36 36 36 36 36 36 36 36 36 CP_PCOD_TWL Fsh_catch_sel 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 2 26 30 23 19 19 19 19 19 19 19 19 19 19 19 19 Fsh_catch_99 2 61 108 75 36 36 36 36 36 36 36 36 36 36 36 36 CP_PCOD_HAL Fsh_catch_sel 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 3 44 40 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 3 83 101 60 60 60 60 60 60 60 60 60 60 60 60 60 CP_COD_POT Fsh_catch_sel 4 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 4 47 46 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 4 86 106 60 60 60 60 60 60 60 60 60 60 60 60 60 CP_OTHR_TWL Fsh_catch_sel 5 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 5 26 30 23 19 19 19 19 19 19 19 19 19 19 19 19 Fsh_catch_99 5 61 108 75 36 36 36 36 36 36 36 36 36 36 36 36 CP_OTHR_HAL Fsh_catch_sel 6 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 6 44 40 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 6 83 101 60 60 60 60 60 60 60 60 60 60 60 60 60 CP_OTHR_POT Fsh_catch_sel 7 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 7 47 46 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 7 86 106 60 60 60 60 60 60 60 60 60 60 60 60 60 CV_PLCK_TWL Fsh_catch_sel 8 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 8 30 33 23 23 23 23 23 23 23 23 23 23 23 23 23 Fsh_catch_99 8 63 70 84 35 35 35 35 35 35 35 35 35 35 35 35 CV_PCOD_TWL Fsh_catch_sel 9 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 9 30 33 23 23 23 23 23 23 23 23 23 23 23 23 23 Fsh_catch_99 9 63 70 84 35 35 35 35 35 35 35 35 35 35 35 35 CV_PCOD_HAL Fsh_catch_sel 10 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 10 44 43 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 10 83 101 60 60 60 60 60 60 60 60 60 60 60 60 60

124


CP_PCOD_POT Fsh_catch_sel 11 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 11 50 45 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 11 84 101 60 60 60 60 60 60 60 60 60 60 60 60 60 CV_OTHR_TWL Fsh_catch_sel 12 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 12 30 33 23 23 23 23 23 23 23 23 23 23 23 23 23 Fsh_catch_99 12 63 70 84 35 35 35 35 35 35 35 35 35 35 35 35 CV_OTHR_HAL Fsh_catch_sel 13 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 13 44 43 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 13 83 101 60 60 60 60 60 60 60 60 60 60 60 60 60 CV_OTHR_POT Fsh_catch_sel 14 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 14 50 45 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 14 84 101 60 60 60 60 60 60 60 60 60 60 60 60 60 HERR_GILL Fsh_catch_sel 15 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 15 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 15 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 HERR_SEINE Fsh_catch_sel 16 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Fsh_catch_01 16 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Fsh_catch_99 16 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 463 464 465 466

467

125


Chapter 9. End-to-End Modeling as part of an Integrated Research Program 468

in the Bering Sea 469 470 471 André E. Punt1, Ivonne Ortiz2,3, Kerim Y. Aydin3, George L. Hunt, Jr1, Francis K. Wiese4 472 473 474 475 1: School of Aquatic and Fishery Sciences, University of Washington, Seattle WA 98195-5020 476 477 2: Joint Institute for Study of the Atmosphere and Ocean, University of Washington, Seattle, WA, 98195-478 5672 479 480 3: Alaska Fisheries Science Center, NOAA Fisheries, 7600 Sand Point Way NE, Seattle, WA 98115 481 482 4: Stantec Consulting Ltd., 2515 A Street, Anchorage AK 99503 483 484

This chapter is the revised text of the paper submitted to Deep Sea Research II as part of the special issue 485 4 of the BSIERP program 486

487

488

489

490

491

492

493 494 495 496 497 498 499 500 501 502 503 504 505 506 Citation: Punt, A.E., Ortiz, I, Aydin, K., Hunt, G.L., and Wiese, F. End-to-End Modeling as part of an 507 Integrated Research Program in the Bering Sea Prepared for submission to Deep-Sea Research II, BSIERP 508 4th special issue 509

510

126


511 Abstract 512 513 Traditionally, the advice provided to fishery managers has focused on the trade-offs between short- and 514 long-term yields, and between future resource size and expected future catches. The harvest control rules 515 that are used to provide management advice consequently relate catches to stock biomass levels expressed 516 relative to reference biomass levels. There are, however, additional trade-offs. Ecosystem-based fisheries 517 management (EBFM) aims to consider fish and fisheries in their ecological context, taking into account 518 physical, biological, economic, and social factors. However, making EBFM operational remains 519 challenging. It is generally recognized that end-to-end modeling should be a key part of implementing 520 EBFM, along with harvest control rules that use information in addition to estimates of stock biomass to 521 provide recommendations for management actions. Here we outline the process for selecting among 522 alternative management strategies in an ecosystem context and summarize a Field-integrated End-To-End 523 modeling program, or FETE, intended to implement this process as part of the Bering Sea Project. A key 524 aspect of this project was that, from the start, the FETE included a management strategy evaluation 525 component to compare management strategies. Effective use of end-to-end modeling requires that the 526 models developed for a system are indeed integrated across climate drivers, lower trophic levels, fish 527 population dynamics, and fisheries and their management. We summarize the steps taken by the program 528 managers to promote integration of modeling efforts by multiple investigators and highlight the lessons 529 learned during the project that can be used to guide future use and design of end-to-end models. 530 531 Keywords: Bering Sea; End-to-end Modeling; Ecosystem Based Fisheries Management; Management 532 Strategy Evaluation 533 534 535

536

127


1. Introduction 537 538

Progress on implementing ecosystem-based fisheries management (EBFM)4 involves multiple facets, 539 including a better understanding of the processes which characterize and control ecosystems. EBFM needs 540 to be grounded by national and international legislation, which in the US is governed by the Magnuson–541 Stevens Fishery Conservation and Management Act (US Public Law 104–297). The Bering Sea Project 542 (the combined Bering Ecosystem Study, BEST, and the Bering Sea Integrated Ecosystem Research Project, 543 BSIERP) aimed to improve ecosystem understanding, and to support fisheries management in the eastern 544 Bering Sea. It employed a combination of field studies and an end-to-end ecosystem model that included 545 climate drivers, lower trophic levels and fish dynamics, which in turn could be driven by various fisheries 546 (Wiese et al., 2012). Development and successful implementation of this project was a substantial 547 undertaking that involved over a hundred principal investigators, with much of the historical data and 548 fieldwork synthesized into the modeling. The Bering Sea Project has led to a better understanding of what 549 it means to develop models for EBFM. 550 551

The primary focus of the Magnuson–Stevens Fishery Conservation and Management Act has been on 552 single-species. However, there is an increasing recognition worldwide for the need to account for factors 553 that are ignored when conducting single-species stock assessments. Likewise, there is growing recognition 554 of the need to take into account the interactions among fisheries in scientific study, as well as in 555 management decision making. This recognition has led to policy documents and statements of intent that 556 fisheries management should move to a more ecosystem-based or ecosystem-focused approach. 557

558 In 1999, the National Research Council defined EBFM as “an approach that takes into account major 559

ecosystem components and services, both structural and functional, in management of fisheries. It values 560 habitat, embraces a multispecies perspective, and is committed to understanding ecosystem processes. Its 561 goal is to achieve sustainability by appropriate fishery management”. Several authors have since proposed 562 alternative definitions for EBFM (e.g., Witherell et al., 2000; FAO, 2003; Sissenwine and Murawski, 2004; 563 McLeod et al., 2005; Murawski and Matlock, 2006; Marasco et al., 2007; Francis et al. 2007). All of these 564 definitions include reference to habitat and multi-species effects and more recently to climate impacts, and 565 impacts of management on human as well as biological communities. For example, Marasco et al. (2007) 566 provided the following definition for EBFM: “Ecosystem-based fishery management recognizes the 567 physical, biological, economic and social interactions among the affected components of the ecosystem and 568 attempts to manage fisheries to achieve a stipulated spectrum of societal goals, some of which are in 569 conflict”. This definition recognizes that socio-economic factors are core to an EBFM; this is supported by 570 recent mathematical models evaluating trade-offs among management strategies that explicitly account for 571 user responses to management regulations (e.g. Fulton, et al., 2011b). It also recognizes that management 572 takes place within a legal management framework. 573

574 Several calls for the implementation of EBFM have been made (e.g. Pikitch et al., 2004). Section 406 575

of the 1996 US Sustainable Fisheries Act provided initial guidance on inclusion of ecosystem principles in 576 management plans, and mandated the formation of the Ecosystems Advisory Panel to the National Marine 577 Fisheries Service, which reviews progress towards incorporation of ecosystem principles in Fishery 578 Management Plans. However, balancing EBFM implementation with existing mandates for single-species 579 catch limits has been challenging (see, for example, Moffitt et al., the volume). 580

581

4 Several acronyms have been proposed for ecosystem-based fisheries management (EBFM), including EAF (ecosystem approach to fisheries). Conceptually, except for EBM (ecosystem based management) and EAM (ecosystem approach to management), which often envisage management of sectors in addition to fisheries, all these definitions have the same ultimate intent albeit their implementation may be at different management levels. We use EBFM in this paper for convenience.

128


While it has been recognized that quantitative ecosystem modeling will be a necessary component of 582 EBFM, developing ecosystem models for fisheries management has been challenging, because: (1) field 583 programs for EBFM are often “add-ons” to single-species surveys resulting in limited data for 584 parameterizing ecosystem models; (2) ecosystem models, in part to ease complexity, often do not calculate 585 quantities needed for management, such as age-structured spawning stock biomass; (3) resources often do 586 not allow engagement of experts at all ecosystem levels during the course of a modeling project, possibly 587 leading to misuse or misunderstanding of results; and (4) data requirements and computational complexity 588 make it difficult to “certify” such models for management use given requirements for accuracy and the 589 reporting of uncertainty. 590

591 The Bering Sea Project included an end-to-end model that would synthesize available data, incorporate 592

new data from the parallel field program, and inform the ongoing research efforts. This project consequently 593 required co-ordination of research activities by a diverse group of principal investigators to ensure that 594 broad research goals would be achieved. Project goals included understanding biological and ecological 595 processes, exploring various hypotheses related to the dynamics of the Bering Sea Ecosystem, and 596 evaluating resource management options through a formal Management Strategy Evaluation (MSE). 597

598 The modeling project was designed to be tightly coupled to the fieldwork at all stages, with feedback 599

and synthesis occurring at all levels. It required the development of standards for the ecosystem modeling 600 efforts, and a different level of organizational guidance and regular feedback compared to ‘traditional’ 601 projects. The combined organizational, modeling, and synthesis challenges were sufficiently unique from 602 the process of “simply” constructing an end-to-end model from previously-available data that we describe 603 the project using a new term, the Field-integrated End-To-End modeling program, or FETE. 604

605 Section 2 of this paper introduces the Bering Sea Project, and the concept and key components that 606

constitute a FETE. Section 3 summarizes an approach (initially developed by Marasco et al. [2007]) for 607 constructing management systems to implement EBFM based on the MSE approach and ecosystem 608 modeling. While MSE was not the only focus of the modelling component of the project, it required the 609 integration of all components of FETE. Section 4 outlines expectations for FETE models, guidelines 610 established to ensure that the project was as statistically and ecologically rigorous as possible, and identifies 611 progress against these expectations and guidelines. Section 5 summarizes best practices and future 612 directions of integrated end-to-end modelling, i.e. what makes a successful FETE? Finally, Section 6 613 summarizes the legacy of the project. 614

615

2. The BEST, BSIERP and FETE 616 617

The development of BEST, and subsequently BSIERP, was initiated at an international planning workshop 618 held in September 2002 to examine the feasibility and value of developing a large interdisciplinary study 619 of the Bering Sea. A second planning workshop was convened in March 2003, the result of which was the 620 development of the Bering Ecosystem Study Science Plan (2004). Contemporaneous with the development 621 of the BEST Science Plan was the development of a long-term science plan for the North Pacific Research 622 Board (NPRB). Following the guidance of an ad hoc National Research Council panel which emphasized 623 the importance of large-scale integrated studies of the marine ecosystems of the eastern North Pacific, 624 similar to that being developed by BEST for the Bering Sea, NPRB developed a science and implementation 625 plan for the BSIERP in 2005. After a limited field season funded by NSF in 2007, negotiations between 626 NPRB and NSF resulted in a historic partnership for work in the Bering Sea, with NSF funding climate, 627 ocean physics and lower trophic-level studies up through zooplankton, and NPRB funding work on large 628 zooplankton through fish, seabirds, marine mammals and humans. The now combined Bering Sea Project 629 launched its first field season in 2008 and included over one hundred principal investigators covering almost 630 all disciplines of marine science (Wiese et al., 2012). 631

129


632 To aid in the development and evaluation of the modeling component in the proposals, the NPRB 633

funded an Ecosystem Modeling Committee (EMC) in 2006, consisting of scientists not funded in the 634 program, but experts in atmospheric and marine sciences, conceptual thinkers, as well as experienced 635 modelers. The EMC was charged with designing modeling selection criteria to be used in proposal review 636 and subsequent evaluations, providing advice to the funded modeling team, giving feedback to the funding 637 agencies on the effort’s progress, and helping the modelers obtain needed resources. 638

639 The resulting program, including modeling, field integration, and program review, made up the FETE. 640

Key features included: 641 1. End-to-end in scope and expertise: Core modeling efforts and expertise were built around end-642

to-end research (climate, physics, plankton, fish, other animals, and humans). Critical here was the 643 inclusion of expertise in the integration process, not merely the inclusion of “canned” results from 644 other models and domains in the finished model. 645

2. A priori and continuous integration between fieldwork and modeling: Fieldwork and modeling 646 were designed together from the start, with common end-goals in mind. Interactions between 647 researchers occurred throughout the program and were structured (workshops or meetings) to allow 648 for formal adjustments throughout the project as the field work informed the models and vice versa. 649

3. Model outputs appropriate to stakeholder goals: A priori consideration of stakeholder needs (as 650 well as feedback from them during the program) was necessary to ensure models would produce 651 adequate and useful results for management. For example, carbon is used in biogeochemical models 652 concerned with climate change, but biomass may be used when examining fish foraging behavior, 653 and numbers of fish-at-age is a key component to fisheries management. 654

4. Modularity and “competition” in model design: The structure of the FETE allowed individual 655 components to be re-examined through “competitive” modeling; i.e. extracting the simplest 656 component from the end to end model that captures the essence of or drivers of the interactions and 657 using them in alternative less complex models. 658

5. Centralized integration and steering: To achieve this integration and have project goals useful 659 to management, it was necessary to have strong project leadership, with a mandate to guide the 660 FETE both scientifically and programmatically, including overseeing changes in scope or model 661 design throughout the whole project. 662

Specific examples demonstrating how these key features were implemented in the Bering Sea Project, 663 especially with respect to management strategy evaluation, are discussed in Sections 3-5. 664

665 2.1 FETE modeling program components 666 A central component of the FETE was the Bering 10K ROMS-NPZD-FEAST-FAMINE5 model complex 667 (Fig. 1) that formed the basis for exploring the impact of fishing and climate on both ecological processes 668 and the performance of management strategies. It was used to run a 1970-2009 hindcast, and was set-up to 669 run in forecast mode using input from selected IPCC climate models that performed well for the Eastern 670 Bering Sea: CGCM-t47 (low ice), ECHOG (high ice) and MIROCM (medium ice) (Wang et al., 2010). The 671 oceanography was based on the ROMS-Bering10K (10 km resolution), a coupled ocean-sea ice model 672 whose spatial grid is a subset of the NEP5 model described and evaluated by Danielson et al. (2011), which 673 itself built on a model described by Curchitser et al. (2005) and Herman et al. (2013). The lower trophic 674 levels were modeled using a nutrient-phytoplankton-zooplankton detritus (NPZD) model coupled to the 675 ROMS-Bering10K, specifically designed to incorporate the ice dynamics of the Bering Sea, and modeled 676 nutrients, phytoplankton, copepods, euphausiids and detritus (Gibson and Spitz, 2011). Model coupling 677

5 It is important to distinguish the FETE modeling as a whole from any particular realization of the end-to-end model. A model in this group (e.g. “NPZD” or “FEAST”) is referred to by its target trophic level, and may or may not include feedback to other components depending on the particular run. FETE as a whole refers to this suite, regardless of which components are being used for a particular result.

130


included feedback from the NPZD to the ROMS-Bering10K through phytoplankton density, which affects 678 shortwave penetration (heat absorption) in the upper water column and between NPZD and the Forage 679 Euphausiid Abundance in Space and Time (FEAST) model (Aydin et al., this volume; Ortiz et a1., this 680 volume-b) (functionally the fish module for this effort), through predation. A key design feature, unusual 681 in many end-to-end models, was dynamic top-down coupling from fish to zooplankton. FEAST, thus 682 coupled to both the NPZD and the ROMS-Bering 10K, and was a multispecies bioenergetics model, with 683 consumption as a function of length-based prey selection, prey preference and availability, and predator 684 movement based on biomass gain optimization. Removals by fishery effort were based on spatially-explicit 685 historical catches for the hindcast, and on a model of fishing effort allocation for model projections 686 (FAMINE; Fishing effort Allocation Model In Nash Equilibrium). 687 688 689 3. Management Strategy Evaluation and EBFM 690

691 The Bering Sea Project used MSE to evaluate management strategies needed to achieve ecosystem 692 objectives (sensu Sainsbury et al., 2000; Fulton et al., 2007; Dichmont et al., 2008; 2013). An MSE (Smith, 693 1994; Smith et al., 1999; Goodman et al., 2002; Butterworth, 2007; Punt et al., in press) involves assessing 694 the performance of alternative candidate management strategies relative to performance measures that 695 quantify the management (and legal) goals for the managed system. Thus, an MSE involves developing and 696 parameterizing a model of the system to be managed. In the absence of data, it may also involve using 697 hypotheses for how the system may change over time (Punt et al. 2014). 698 699

An MSE (Fig. 2) aims to represent all key processes in system models and can provide performance 700 metrics that relate to a broad range of goals. In the context of the Bering Sea Project, a key process was 701 developing the scenarios regarding future climate. A concern with end-to-end models is the general inability 702 to estimate the values for their parameters using standard statistical models due to either lack of data or 703 limits of computing time (Gaichas et al. [2010, 2011] being a noteworthy exception in this regard). 704

705 Which candidate management strategies are evaluated in an MSE depends in large part on the interests 706

of the managers. Ideally, management strategies for EBFM should be based on the results of process studies, 707 monitoring of ecosystem indicators, and ecosystem models, in addition to the outcomes of single-species 708 stock assessments. In principle, management strategies for EBFM could involve monitoring a range of 709 ecosystem indicators and modifying management practices based on whether the indicators are outside of 710 acceptable limits, analogously to the types of management strategies used for single-species fisheries 711 management. Management strategies for EBFM could be based on assessment methods that include multi-712 species considerations explicitly. However, to date the control rules that would underlie such management 713 strategies have seldom been implemented or even fully defined (Moffitt et al., this volume). 714

715 To address this challenge, the FETE included a workshop with stakeholder groups to identify a 716

preliminary set of management strategies (Fig. 3). In some cases, implementing the proposed strategies 717 required modifications to the end-to-end model; these adjustments were made as the project progressed. 718 The selected management strategies were based on three types of assessment methods: Ecosim, CEATTLE 719 (the multispecies statistical model of Holsman et al., (the volume)) and the single-species assessment 720 methods currently used to provide management advice to the North Pacific Fisheries Management Council. 721 Each assessment method was linked to appropriate harvest control rules, which produced estimates of Total 722 Allowable Catches. The workshop also recommended exploring a management strategy which did not 723 implement the 2 million tonne cap on total harvest, which is currently written into regulation for the eastern 724 Bering Sea (Fig. 3). The workshop also specified management scenarios based on the impact of climate 725 change. 726

131


4. Guidelines and principles for the development of ecosystem models, and how to apply them 727 towards end-to-end modeling 728 729

The questions the EMC developed to evaluate the proposals for the modeling component of the Bering Sea 730 Project focused on what the various models were meant to produce and why, whether the outputs would be 731 useful for management and would provide measures of uncertainty, how existing and future data could be 732 integrated into the model, how the model could inform ongoing research, and whether the model could be 733 validated. The questions and their rationale are discussed below and, even though they were developed for 734 the Bering Sea Program, they provide a way to evaluate any model. 735 736 1. What is the model intended to predict? 737

This may seem like an extremely simple question. However, many models, particularly those of the 738 end-to-end variety, claim to be able to predict many types of impacts. The aim of this question was to 739 ensure that the models were designed given specific scientific and management questions, rather than 740 having the models developed and subsequently retrofitted to address questions of scientific and 741 management relevance. 742 743

The FEAST and NPZD models (effectively the biological component of the integrated model) were 744 designed as predictive models responsive to long term climate variation and geared to address two basic 745 purposes: (1) understand the underlying processes by which environmental variability affects biological 746 processes such as primary and secondary production and fish recruitment and distribution, and (2) 747 characterize the environmental effects on the distribution of fishing effort and hence the age structure 748 in fish populations and recruitment to the fishery. This involved using FEAST as the system model for 749 an MSE aimed at walleye pollock Gadus chalcogrammus, Pacific cod Gadus macrocephalus, and 750 arrowtooth flounder Atheresthes stomias. 751

752 The ROMS model was designed to enable climate factors to be explicitly represented in the 753

dynamics of the resources, while the FAMINE and MSE models were developed to represent 754 management and how management actions translate into fishing effort and hence fishing mortality. 755 756

2. What specific aspect of the prediction is anticipated to be of direct value for fisheries management? 757 Many proposals for scientific research claim that their research will be of direct use for management 758 purposes. The EMC envisaged that by explicitly stating how predictions would be used for management 759 purposes, the modeling proposal and the subsequent research would be more likely to lead to 760 predictions that would actually achieve this purpose. 761 762

Amongst the main goals was the ability to predict the responses of fish stocks and fishermen to 763 long-term climate scenarios. The high resolution of ROMS (~10km) would provide maps that would 764 allow detailed representation of fleet distributions. The full end-to-end model was geared to address 765 expected changes in potential total allowable catches and fish availability to the catcher processors and 766 catcher vessels, which have distinct spatial constraints. Each individual model had outputs which were 767 linked, such that changes in climate would feed through the simulated ecosystem to impact how 768 management strategies would be able to achieve the goals established for EBFM. 769 770

3. What measure of "accuracy" in the prediction is crucial to determining the usability of that 771 prediction to fisheries management? 772 In principle, models can make predictions of virtually any quantity. However, the estimates may be 773 very biased and/or imprecise. The EMC expected that the desired quality (or accuracy) of predictions 774 would be evaluated before the modeling was to be conducted. This was perhaps one of the most 775 challenging of the questions because establishing hard standards for model accuracy is difficult. 776 Validations are time consuming to perform and can be computationally expensive. Some types of error 777

132


are cumulative, and only emerge after multiple years into the simulation. In general, validations and 778 performance assessments do not have a set level of accuracy. Rather, they have levels of conformance 779 as measured by correlation, principal component analysis and comparisons between the observed data 780 and model output. 781 782

Even when each modeling component within the overall model (ROMS, NPZD, FEAST, FAMINE, 783 MSE) provided plans that included statistical techniques to measure variance and accuracy, the number 784 and diversity of variables in each model made it impossible to provide the desired level of accuracy for 785 each output from the integrated model. For example, even if it is possible to explain 50% or more of 786 the variance of the data used in a particular model, the cascading effect of such variability or lack of 787 accuracy on processes outside that model may be greater. For example, initial sea temperature estimates 788 in the ROMS model, considered to be within acceptable ranges in an oceanographic context, drove the 789 bioenergetics of lower and upper trophic levels towards and beyond their upper tolerance limits. 790 Moreover, it moved the location and extent of the cold pool – a key environmental factor known to 791 impact the dynamics of groundfish stocks (NPFMC, 2012) – thus changing critical temporal and spatial 792 ecosystem dynamics. 793

794 4. What alternative models are plausible competitors whose performance should be tested against the 795

model being developed? 796 All models should be recognized as simplifications of the system under consideration. The EMC 797 recognized the need for multiple alternative models so that the predictive skill of the proposed model 798 could be evaluated relative to alternative (generally less complex) models, and because it is not 799 uncommon for the predictions from ecosystem models to be very sensitive to their structure. 800 801

The EMC envisioned complementing and competing models: in particular, correlative models to 802 be developed as part of the Bering Sea Project (Mueter et al., 2011; Siddon et al., 2011, 2013a, b; Heintz 803 et al., 2013), and existing models such as MSM (Jurado-Molina et al., 2005) and the Ecopath model for 804 the Eastern Bering Sea (Aydin et al., 2007), as well as currently used single-species stock assessments. 805 Also developed were a multi-species biomass dynamics model for walleye pollock, Pacific cod, 806 arrowtooth flounder (the three main species in FEAST), and small mouthed flatfish (not in FEAST) 807 (Uchimaya et al., this volume), and a statistical model linking recruitment of walleye pollock to 808 variability in late summer sea surface temperatures and to the biomass of major predators (Mueter et 809 al. 2011). 810 811

5. How will the achieved predictive power of the model be compared against the performance of 812 plausible alternatives, and how will this guide subsequent choices about model form and 813 parameterization? 814 The quality of fishery models is generally assessed in terms of hindcast skill, i.e. the ability to replicate 815 the data used for model calibration, and this is clearly a minimum requirement for any ecosystem (or 816 other) model. Considerable effort has been dedicated to developing metrics for evaluating hindcast skill 817 for stock assessment models, including residual analysis and Bayesian methods for posterior predictive 818 checks. However, the EMC expected model performance (and model refinement) to be based on 819 forecast as well as hindcast skill. 820 821

Given the expected performance of FEAST’s forecast skill, several attributes, including those 822 linked to the stock assessment models, required calibration. The predictions, which could be compared 823 among models, included spatial aspects such as species distribution by age, as well as key regional and 824 length-specific trophic interactions (e.g., Buckley et al., this volume). 825

826 The ability to review the performance of forecasts based on the FAMINE and MSE components of 827

the integrated model was limited given lack of sufficient computational resources. However, forecast 828

133


skill could have been evaluated by running the calibrated end-to-end model to a year other than the 829 most recent year and projecting forward. Unfortunately, time constraints of the overall project, given 830 the available computational resources, precluded this. 831 832

6. What data are available to drive, calibrate, and test the model? 833 This question recognized that data are used in multiple ways in ecosystem models. The EMC envisaged 834 that some sources of data would be included in the model as “facts”. However, data in this context also 835 include values for parameters that are pre-specified based on auxiliary information. For example, when 836 applying models such as Ecosim, diet is frequently assumed to be known. All models, ecosystem or 837 otherwise, include parameters that are not known from auxiliary information but which must be 838 estimated from the monitoring data. The model fitting process should ideally involve minimizing some 839 form of objective function involving discrepancies between the observed data and model predictions. 840 However, it is computationally infeasible to fit large complex ecosystem models such as FEAST or 841 Atlantis (Fulton et al., 2011a) to monitoring data, so the model calibration process is more heuristic 842 than formal. The EMC considered model validation a key step in the modeling process and expected 843 that some of the available data would be kept away from the modelers to allow an independent test of 844 model skill. Use of this form of cross-validation is common in some modeling fields, but is relatively 845 uncommon with fisheries modeling where, given the general lack of data, all of the available 846 information is used for model calibration. 847 848

The primary sources of data for FEAST were the historical databases kept by the Alaska Fisheries 849 Science Center (NOAA) for fish age, length, weight, distribution, feeding habits and fishery catches. 850 Data for the models of the lower trophic levels and the ROMS model were based on past data, as well 851 as from moorings and process studies that were part of the Bering Sea Project. The FAMINE model 852 was driven using data on fishing effort and ice cover, whereas the MSE model used information 853 generated by FEAST. However, no current amount of field work could provide the data needed to 854 estimate all parameters and validate all levels of the end-to-end model. In hindsight, the availability 855 and consolidation of such data proved to be a bottleneck for model development, particularly for the 856 NPZD model and the process studies. 857 858

7. How will the existing data be used to quantify model fit and predictive power? 859 Evaluating model fit (hindcast skill) is a key element of single-species stock assessment, and extensive 860 terms of reference have been developed to detect violations of the ability to replicate data (e.g., PFMC, 861 2012). How to evaluate hindcast skill, however, is not as developed for multi-species models (see, 862 however, Gaichas et al., 2010, 2011), and particularly not for models that produce spatial outputs, 863 owing to spatial autocorrelation in the data available for evaluating model skill. Simple metrics (e.g., 864 all species remain in the system) have been used to evaluate model fit and hindcast skill for ecosystem 865 models, but these metrics are not nearly as sophisticated as those used for single-species stock 866 assessments. 867 868

Evaluating predictive power involves similar issues to evaluating hindcast skill, but with the 869 additional complexity that the assumptions which can be made when making future predictions need to 870 be specified and evaluated carefully. A variety of approaches were used to validate the components of 871 the end-to-end model. For example, the climate models used for the forecast were selected based on 872 performance in the Bering Sea, mainly their ability to capture ice cover and the Pacific Decadal 873 Oscillation (Wang et al., 2010). 874

875 Validation of physical characteristics (correlations between observed and model estimates) such as 876

ice cover and temperature was carried out by Danielson et al. (2011) for the 60-layer ROMS North East 877 Pacific 5 model. The smaller grid used for the Bering 10K ROMS-NPZD and Bering 10K ROMS-878 NPZD-FEAST-FAMINE model has a reduced vertical resolution from 60 to 10 levels. Hermann et al. 879

134


(2013) conducted both correlation and principal component analyses using available time series for 880 physical data, such as temperatures at mooring 2 (M2), ice extent and salinity; multivariate analysis 881 was performed using data from the Bering Sea. Herman et al. (2013) also used temperature, salinity 882 and total chlorophyll from the Alaska Fisheries Science Center’s annual Bering-Aleutian Salmon 883 International Survey (BASIS) research cruises in a multivariate analysis. Gibson and Spitz (2011) 884 conducted a sensitivity analysis of the NPZD portion of the end-to-end integrated model. Assessments 885 of fish movement and distribution patterns (Ortiz et al., this volume -a), biophysical processes (Ortiz et 886 al., this volume-b) and fish bioenergetics (Aydin et al., this volume) were also conducted. 887

888 For FEAST, historical data from 1982 to 2007 were used to estimate parameters related to the fish 889

bioenergetics (length-weight relationships and length-energy density) and the relationship found 890 between recruitment and fall condition of age-0 pollock was used to assess model performance (Aydin 891 et al., this volume). Refinements of these processes were made based on the field studies. For spatial 892 aspects, historical data were used to construct initial conditions for fish in all years from 1971 to 2010. 893 This allowed testing of single individual years. However, since only the first year uses initial conditions 894 derived from data, for multiyear runs, subsequent years could be validated using the remaining 895 historical data. 896

897 Ideally, a more holistic validation of the entire end-to-end model could have been achieved had 898

there been both cold and warm years during the field seasons encompassed by the Bering Sea Project. 899 Contrast in environmental conditions during the fieldwork years was originally envisaged in the 900 proposals that led to the Bering Sea Project. However, all field years were cold, thus precluding this 901 approach to model validation. 902

903 In general, FEAST succeeded in capturing the general growth (Aydin et al., this volume), 904

movement and distribution of fish (Ortiz et al., this volume -a), and was sensitive to cold and warm 905 years. However, the model failed to predict recruitment and survival of age-zero fish satisfactorily for 906 multi-year historical runs in which small age-structure errors could accumulate over the run, and the 907 numbers of age-1 pollock had to be nudged to their stock assessment estimated numbers at the end of 908 each model-year. 909 910

8. What pertinent future data are anticipated to become available within the time frame of the project 911 and how will these future data be used to quantify model fit and predictive power? 912 The FETE involved model development, data collection occurring in parallel, and this question was 913 developed to ensure that fieldwork and modelling were integrated. Obtaining data for the lower trophic 914 levels for cold and warm years was not feasible due to the lack of warm years during the field program 915 (Stabeno et al., 2012). Several data sets that became available during the program were integrated into 916 the modeling efforts (either for parameter estimation or to assess model performance), namely 917 improved spatial distribution of age-0 and age-1 pollock, zooplankton surveys, acoustic estimates of 918 euphausiids, winter distribution of the pollock spawning stock, seasonal energy density of juvenile 919 pollock, consumption of small, medium and large copepods by fish, and a series of data from the lower-920 trophic-level component. Several of these data sets, e.g. pollock bioenergetics, acoustic estimates of 921 euphausiid biomass, and additional oceanographic data, are now regularly updated and have become 922 part of standard surveys due to their usefulness for supporting analyses. Other data gaps have led to 923 new analyses (such as zooplankton seasonal and spatial patterns) and pilot projects (winter zooplankton 924 sampling). 925 926

9. How has it been determined that the proposed quantity and quality of data can be expected to be 927 sufficient for the intended use in tuning and testing the model? 928 This question attempted to integrate the remainder of the questions, and hence provide an overall basis 929 for evaluating the design of the modeling. Unfortunately, this question won’t be fully addressed until 930

135


the end-to-end model has been applied more extensively. 931 932 5. Discussion: Best practices and future directions 933

934 The approach for developing end-to-end models for management purposes outlined by Marasco et al. 935 (2007) is comprehensive, and, when combined with the questions developed by the EMC, should have led 936 to a process in the FETE where a set of models was selected that were relevant to the system at hand, could 937 be calibrated to existing data and tested through comparison with independent data sources, and were useful 938 for evaluating management strategies in an ecosystem context. Practice, however, often differs from theory, 939 and hence here we summarize our experience and distill what we consider best practices to facilitate 940 subsequent efforts and end-to-end modeling in general. 941

5.1 Be realistic about what can be accomplished within a given timeline 942 It is important to be realistic about the constraints due to the size and complexity of a model before work 943 starts on its development and parameterization. In the case of the Bering Sea Project, the complexity of the 944 FETE effort only became fully apparent as the project proceeded. For example, coupling the individual 945 models was a major undertaking, which, although recognized as a key task when the overall project was 946 designed, and a goal that was achieved, was an ongoing constraint on the speed of model development. As 947 such, a significant amount of effort should be spent early on fully scoping out the model needs, 948 especially in terms of integration. Most modelers are generally well aware of their individual needs and 949 are somewhat realistic about what can be done. Developing end-to-end models for actual ecosystems and 950 management, however, is a much younger endeavor, resulting in a tendency to underestimate challenges 951 and project outcomes on the basis of potential rather than reality. 952 5.2 Larger-scale software projects need logistical support on a par with fieldwork 953 Care should be taken when a project’s scale exceeds that of an individual or a small team and encompasses 954 multiple institutions. While technology scales, large-scale software development, as an activity, does not 955 (Brooks, 1995). Scientists used to working as individuals, on individual pieces of code, need to expect time 956 devoted to logistics of working with large computers at multiple institutions, transferring files, and keeping 957 source code synced. When coupling models from different disciplines and modeling teams, code is often 958 written independently and then synchronized. Software and hardware management and familiarity with 959 the structure and parameters of all components of the model are critical for achieving a working end-960 to-end model. 961 962 5.3 Clear separation of scientific versus logistics oversight 963 Rose et al. (2010) note that the challenge of interdisciplinary research is “as much of a people challenge as 964 a technical one”. In the case of the integrated modeling work, the first few years were coordinated through 965 the EMC. Their role was to guide and facilitate, but not to make final decisions. The questions designed by 966 the EMC included both scientific concerns (comparing outputs to data) and logistical concerns (time frame 967 of data). However, the EMC functioned almost entirely as a scientific review body during the initial stages 968 of the actual work on the project. Logistics were initially to be handled by the modelers collectively; while 969 a lead modeler was appointed, it was primarily in a communication/coordination role rather than as a firm 970 project leader. 971 972

As the project developed and many modelers focused on their own timelines and model developments, 973 it became clear that a modeling facilitator was needed to help maintain a unified standard and 974 expectation across projects in terms of cross-collaboration, facilitation, product delivery, priorities 975 and overall model management. Such an independent, but informed, coordinator was appointed during 976 the latter part of the project and helped to keep the overall outcome in mind whenever individual goals and 977 timelines were in conflict. A third model of how an independent group can facilitate and oversee a modeling 978 project is provided by the Gulf of Alaska Integrated Ecosystem Research Program (GOAIERP). This is a 979 much smaller project than the Bering Sea Project with a markedly smaller modeling component. In 980

136


particular, there is no attempt to develop an end-to-end model for the Gulf of Alaska at present, so the 981 logistics involved in the modelling are markedly less. However, the GOAIERP project is still considered a 982 FETE, as it includes field/modeler interactions and end-to-end expertise that connects the core fish models 983 to end-to-end processes in an integrated manner. In this case, an individual was contracted by NPRB on an 984 as–needed basis to provide guidance to the modeling group. 985

5.4 Open and frequent communication with field biologists 986 In addition to being the source of most of the data for validation, field biologists provide expert advice and 987 direction when confronted with modeling decisions for which there are apparently equally suitable options 988 or no data. Close communication with groups of field biologists also facilitates consensus building, 989 improved understanding of model structure and ultimately, acceptance of the model. In the FETE, much 990 effort was put towards facilitating frequent conversations between modelers and field teams, and the latter 991 consequently had a clear expectation that ongoing data collection would ‘feed into’ the modeling. This 992 might have been a realistic expectation if it were a simple issue of adding data to a data file and running the 993 model. However, adding data can lead to changes in the model structure because the model structure is, by 994 definition, tailored to the data. There is also a lag time between data collection, analyses and pattern/process 995 identification. While it is obviously desirable to allow data collection efforts to feed into model 996 development and parameterization, the process should not be considered routine, fast, easy or not disruptive 997 to the overall modeling process. Addressing the issue of how to integrate new data into the modeling 998 process needs to be addressed early in the project design, and the logistic constraints need to be 999 recognized. For example, new data could be used for validation purposes in the final year of a project if 1000 sufficient data are collected to parameterize the model in the first place. This issue was identified at the 1001 start of the project, but the extent of the task was not totally understood at the time. The possibility of the 1002 results of a major piece of fieldwork calling for a major change to model structure was not recognized at 1003 the time the project was designed, but rather later during development. 1004 1005 5.5 Adequacy and availability of data for model validation/testing 1006 Ideally, the existing data and the temporal and spatial coverage of the key variables in the models should 1007 match. In the FETE, many of the oceanographic and lower trophic level data available to validate the model 1008 came from point data, e.g., moorings, which provide reliable time-series but poor geographic coverage, or 1009 from oceanographic stations, spread over a large area but with no associated long-term time-series. 1010 Eventually, an effort was made to use other sources of data (such as, for example, temperatures collected 1011 during annual fishery surveys) appropriate for model validation. In addition, a series of data sources were 1012 combined to define regions of similar bio-physical characteristics that could be used for model comparison 1013 rather than relying on point sources (Ortiz et al., 2012). The existing data should also be compiled and made 1014 available in advance. For both the oceanographic and the lower trophic level modeling efforts, data and 1015 validation came late in the process, too late for the benefits of improved parameters to be included in the 1016 simulations coupling fish dynamics. Future attempts at end-to-end modeling should involve a group to 1017 identify all potential data sources, a designated entity in charge of compiling, formatting, and 1018 disseminating such datasets, and the creation of the framework by which to conduct model validation. 1019

5.6 Most work is sequential and iterative as opposed to simultaneous and independent (non iterative) 1020 All models have to be integrated and re-validated as a whole. The size of this task is highly dependent on 1021 overall model structure and level of coupling/linkage between the different model components. This is not 1022 a one-time occurrence and demands longer timelines, as response time depends on each party’s time 1023 availability and priorities, in addition to the actual difficulty of the problem itself. Therefore, even when 1024 one of the components of an end-to-end model is considered finalized, time should be allotted to 1025 support further implementation and testing of subsequent coupled versions of the integrated model. 1026 1027

In the FETE, this issue proved particularly challenging for the use of MSE, as forward projections could 1028 not commence until the remainder of the Bering 10K ROMS-NPZD-FEAST-FAMINE model had been 1029

137


developed and validated. Having an MSE component from the start of the program meant that management 1030 quantities to be extracted from the model (e.g. spawning stock biomass for fish stocks) were built into the 1031 model design from the start, rather than in an ad hoc manner afterwards. However, the first viable (hindcast) 1032 version of the fully-coupled model was finalized only after six years, so the “top-of-the-food chain” portions 1033 of the project (MSEs and Economics) ended up being much more limited in scope than intended. We 1034 propose two alternatives to address this problem: 1035

(1) Start projects of this type in multiple phases. In particular, phase 1 would involve developing the 1036 ecosystem component model which will operate together as a system model while phase 2 would 1037 involve refining the system model and also conducting the MSE. Phase 1 would involve steps such 1038 as a stakeholder workshop to identify the management strategies to evaluate and also the 1039 specification of the data that are needed to apply to selected management strategies. These steps 1040 are needed so that the biological component of the system model is structured to generate the data 1041 needed as the basis for the MSE. 1042

(2) Conduct the MSE as part of the FETE, but also develop a “simple” system model as a component 1043 of the project so that some MSE results can be obtained. It is likely that some management strategies 1044 will fail to achieve the management objectives using a simple model. It would be expected that 1045 management strategies which ‘fail’ for simple system models will also ‘fail’ for more sophisticated 1046 and realistic system models. 1047

It should be noted that there is a cost associated with developing ecosystem models to evaluate management 1048 options beyond that required to increase ecosystem understanding. For example, the management strategies 1049 to be evaluated required data on the age structure of fishery and survey catches. The original design of the 1050 FEAST model involved modeling population length- but not age-structure; including population age-1051 structure in FEAST increased the number of variables for pollock, Pacific cod and arrowtooth flounder 1052 from approximately 180 to 1386 and reduced the number of length bins from 20 to 14. The management 1053 strategy evaluations also required fisheries by sector (catcher vs catcher/processor vessels) in addition to 1054 by gear and species, thus doubling the number of modelled fisheries. Moreover, the need to manage 1055 according to total catch quotas also required the model to be stopped at regular intervals during the 1056 simulation to keep track of total catches and effort allocation, which added additional complexity to the 1057 overall project. 1058

5.7 Mismatch of required performance levels and performance measures between single discipline 1059 approaches and multidisciplinary ones 1060 When development of the fish model in BSIERP started, there was an incomplete understanding of the state 1061 of development of the oceanographic model. Later, it was noted that the oceanographic model predictions 1062 of temperature were biased by approximately 2°C. This bias was considered acceptable within an 1063 oceanographic context, but unacceptable for the bioenergetics in the fish model, and for the consequences 1064 of temperature on fish distribution. Particular emphasis should be placed on differences in required 1065 scales of results between models. For example, a 1-dimensional version of the coupled ROMS-NPZD was 1066 developed early on in the modeling for calibration to a specific data source (the M2 mooring). It was initially 1067 thought and planned that the 1-D model would be sufficient to quickly test and calibrate the fish model 1068 while it was under development. However, the combination of M2 being a poor location for fish due to 1069 productivity, and the importance of horizontal movement for calibrating fish growth, meant that the testbed 1070 had to await a 3D model, thus slowing down achievement of planned milestones. 1071 1072

Models are always a mix of mechanistic and statistical aspects. FEAST is a primarily mechanistic 1073 model with as few embedded phenomenological correlations amongst variables as possible. This pertains 1074 to (but not exclusively) the EMC's questions regarding data availability and usage. Some data were used to 1075 set up the mechanics, some data were used to test model performance (e.g. the spatial distribution of fish 1076 species by age and length), and some were used as a given process part of the system. It is important to 1077

138


distinguish between using data as "facts", and the steps or mechanics of growth and data used to evaluate 1078 performance of a synergistic property. How much a model is “steered” towards the mechanistic vs. the 1079 phenomenological gradient is a constant choice, and while some guidelines and principles are general and 1080 applicable to all ecosystem modeling, some are specific as they depend on the nature of the project. 1081 Decision making should be consistent with both the mechanistic and the phenomenological gradient 1082 throughout the entire project. Individual component performance metrics should be in line with the 1083 overall purposes of the model and not with a discipline-specific need or standard. Alternatively, if 1084 there are multiple purposes, there needs to a clear process for prioritizing those purposes. 1085

1086 The mismatch in levels of performance between single discipline and multidisciplinary work often 1087

requires a recalibration of the various components once coupled so general patterns can be captured. Further 1088 model refinement improves timing, magnitude and other attributes and decreases the need to compensate 1089 the mismatch between models. 1090 1091 5.8 Lack of familiarity with model limitations pertaining to other disciplines 1092 There is a learning curve when working with multidisciplinary models that can only be gained by experience 1093 and joint collaboration. While all the modelers involved had experience developing models within their 1094 field of expertise, most were unaware or unfamiliar with computing languages, common practices, model 1095 structure, model restrictions and expectations from the other disciplines. This resulted in serious 1096 implications for model design. For example, the fish modelers assumed that time savings could occur 1097 through coarser time steps (which couldn’t be done due to physical constraints), while the physicists 1098 assumed that the fish could be modeled with fewer state variables covering length and ages of fish (which 1099 couldn’t be done due to biological and MSE constraints). A consequence of this was much longer run times 1100 and hence increased difficulties with model development and calibration. In addition, the funded proposal 1101 was modified through discussions with the funding bodies, other researchers on the project and the EMC. 1102 Consequently, the workplan for the modelling was modified during the project development process instead 1103 of during the proposal development phase. Clear, transparent communications between all components 1104 needs to occur during proposal development and early phases of the program to avoid 1105 misunderstandings and to dispel wrong assumptions. Moreover, the relationship between realism 1106 and run times needs to be recognized during the project design stage. 1107

5.9 Coherence of final products from different funding agencies 1108 Different components of the project were completed at different times, and the early finishers were thus 1109 initially disengaged from the synthesis. Eventually, the issue was addressed by several synthesis projects 1110 being funded. The mis-match in the funding of synthesis efforts reinforces the importance of including 1111 adequate time for synthesis as well as for time for modelers to deal with requests from, and 1112 interaction with, other modelers and field biologists from all components involved in the integrated 1113 program. A program needs to start with a synthesis of the kinds of data that will be needed to address the 1114 central questions driving the program, as well as a synthesis at the end. This wrap-up synthesis requires that 1115 many if not most of the basic papers from the program are in press so that they are available to the synthesis 1116 teams. Pushing the final synthesis too early means that much of the material derived from the field and 1117 modeling program will not be available for the synthesis. 1118 1119

6. Conclusions: program legacy 1120 1121

Looking at each individual project separately, the Bering Sea Program’s modeling effort, or FETE, was 1122 extremely successful by most scientific funding standards. The oceanographic model, the NPZD model, 1123 and the fish growth/movement model, can be seen as separate 3-year modelling projects; compared to a 1124 traditional sequential approach (completing work bottom-up from physics to fish), the overall program 1125 condensed 9 years of research into 6 years. Advances were made in physical modelling of the region 1126

139


(Danielson et al., 2011, 2012; Hermann et al., 2013), measuring uncertainties in NPZD models (Gibson and 1127 Spitz, 2011), and quantifying seasonal versus interannual environmental effects on the growth, feeding 1128 rates, and survival of fish (Aydin et al., this volume), effects of prey availability and temperature on fish 1129 distribution (Ortiz et al., this volume-a), and year-round biophysical processes and their effect on fish and 1130 fisheries (Ortiz et al., this volume-b). 1131

1132 The structure of the overall Bering Sea Project, including in-depth principal investigator meetings and 1133

structured workshops between modelers and observationalists, facilitated strong connections for specific 1134 components. This is reflected by the large number of observationalist/modeler partnerships that developed 1135 during the project. Modelers have brought key results from ROMS, NPZD, and/or FEAST (such as 1136 predicted euphausiid densities) to the ongoing NSF synthesis project, fueling modelling and data analysis 1137 well beyond the scope of the original program (e.g., Sigler et al., this volume,). 1138

1139 The project has also had ramifications in the ongoing monitoring of the Bering Sea. The Alaska 1140

Fisheries Science Center is continuing the development of the FETE and is currently using it to target 1141 specific model parameter uncertainties for extended research during ongoing monitoring activities. This 1142 new, integrated activity should significantly operationalize the FETE, both as model and field components, 1143 to provide EBFM advice on an ongoing basis. Combined, these factors have the potential of creating an 1144 institutional structure that will link modeling and field work more tightly into the future. Additionally, the 1145 program has brought fisheries modeling into the developing field of high-performance computing and high-1146 performance data applications. 1147

1148 The MSE project included an initial workshop with attendance from a broad range of stakeholders and 1149

decision makers, and included the development of potential management scenarios. The end results are 1150 visible in the North Pacific Fishery Management Council’s current research priorities, which include the 1151 development of management strategy evaluations and continued production of whole-ecosystem models 1152 for integrated ecosystem assessment. 1153

1154 Every model, just like every field measurement, is in some sense “wrong”; a model, however complex, 1155

is a simplification of reality. The researcher’s challenge is to consider modeling like field research, as an 1156 ongoing, iterative process, producing new questions as well as answers. The models, as proposed, included 1157 a brief to change the very way that field research interacted with models. In that, they were highly 1158 successful; the legacy that this project left is visible today in the ongoing collaborations between researchers 1159 of the Bering Sea, stakeholders, agencies, management bodies, and the public. 1160

1161 Ultimately, the question which needs to be answered is whether it will ever be feasible to construct a 1162

FETE which follows all of the steps outlined by Marasco et al. (2007), and fully addresses the questions 1163 developed by the EMC. We believe that the Bering 10K ROMS-NPZD-FEAST-FAMINE model has 1164 already increased understanding about the Bering Sea ecosystem and its fisheries, even if it could not follow 1165 all of the steps nor fully address all of the questions. Nevertheless, the guidance provided through the work 1166 of the EMC, along with the experience gained through this project, suggests that a FETE will enhance the 1167 development and use of end-to-end models to increase understanding of ecosystems and provide useful 1168 information for both management and research prioritization. 1169

1170 The lessons learned during the development of the FETE are applicable to future model development 1171

work in the North Pacific but also in regions where similar endeavors are being undertaken such as the 1172 Benguela (Travers-Trolet et al., 2014) and the California (e.g. Fulton et al., 2011a; Kaplan et al., 2012) 1173 current systems. These lessons are particularly relevant when considering the development of permanent 1174 operational programs for EBFM, such as the Integrated Ecosystem Assessment program of NOAA (Levin 1175 et al., 2009), where it is envisioned that ecosystem models, if coupled with ongoing feedback from field 1176 researchers and management, may form an organizing principle for a core EBFM team to provide 1177

140


ecosystem-based management and research advice in an ongoing fashion. 1178 1179

7. Acknowledgements 1180 1181

This paper is derived from the authors' presentations at the Daniel Goodman Memorial Symposium (20-21 1182 March 2014), and we thank the organizers for inviting us to be part of that important occasion. The authors 1183 would like to acknowledge the central contribution of the EMC to the Bering Sea Project, and in particular 1184 the contributions of Dan Goodman, former Science Panel member of the NPRB, whose idea it was to form 1185 an EMC, and who chaired that effort during its existence. Without the framework Dan and the committee 1186 provided, this modeling effort would have been fragmented and less tuned to delivering products useful not 1187 only for scientific understanding, but also for direct application to management objectives and decision 1188 making. AEP and IO were partially supported by the North Pacific Research Board. GLH was partially 1189 supported by National Science Foundation grant number 1107250. FKW was fully supported by NPRB in 1190 his former role as Science Director and Modeling Manager for NPRB. Mike Sigler, Martin Dorn (AFSC), 1191 Chris Harvey (NWFSC), three anonymous reviewers and the guest editor are thanked for their comments 1192 on an earlier version of this paper. 1193 1194 1195 8. References 1196

1197 Aydin, K., Gaichas, S. Ortiz, I, Kinzey, D. Friday, N., 2007. A Comparison of the Bering Sea, Gulf of 1198

Alaska, and Aleutian Islands Large Marine Ecosystems Through Food Web Modeling. NOAA Tech 1199 Memo NMFS-AFSC 178. 298 pp. 1200

Aydin, K., Ortiz, I., Hermann, A.J., This volume. New approach to modeling fish bioenergetics, a case 1201 study for walleye pollock, Pacific cod and arrowtooth flounder in the Eastern Bering Sea. Deep Sea 1202 Res. II 00, 00–00. 1203

Bering Ecosystem Study (BEST) Science Plan. 2004. Fairbanks, AK: Arctic Research Consortium of the 1204 U.S. ix+82 pp. 1205

Brooks, F., 1995. The Mythical Man-Month. Addison-Wesley, New York. 1206 Buckley, T.W., Ortiz, I., Kotwicki, S., Aydin, K., This volume. Summer diet composition of walleye 1207

pollock in the eastern Bering Sea, 1987-2011, and predator-prey relationships with copepods and 1208 euphausiids. Deep-Sea Res. II 00, 00–00. 1209

Butterworth, D.S., 2007. Why a management procedure approach? Some positives and negatives. ICES J. 1210 Mar. Sci. 64, 613–617. 1211

Curchitser, E.N., Haidvogel, D.B., Hermann, A.J., Dobbins, E.L., Powell, T.M., Kaplan, A., 2005. Multi‐1212 scale modeling of the North Pacific Ocean: Assessment and analysis of simulated basin‐scale variability 1213 (1996–2003). J. Geophys. Res. 110, C11021, doi:10.1029/2005JC002902. 1214

Danielson, S., Curchitser, E., Hedstrom, K., Weingartner, T., Stabeno, P.J., 2011. On ocean and sea ice 1215 modes of variability in the Bering Sea. J. Geophys. Res. 116, C12034, 1216 http://dx.doi.org/10.1029/2011JC007389. 1217

Danielson S., Hedstrom, K., Aagaard, K., Weingartner, T., Curchitser, E., 2012. Wind- induced 1218 reorganization of the Bering shelf circulation. Geophys. Res. Lett., 39, L08601, 1219 http://dx.doi.org/10.1029/2012GL051231. 1220

Dichmont, C.M., Deng, A., Punt, A.E., Ellis, N., Venables, W.N., Kompas, T., Ye, Y., Zhou, S., Bishop, 1221 J., 2008. Beyond biological performance measures in Management Strategy Evaluation: Bringing in 1222 economics and the effects of trawling on the benthos. Fish. Res. 94, 238–250. 1223

Dichmont, C.M., Ellis, N., Bustamante, R.H., Deng, R., Rickell, S., Pascual, R., Lozano-Montes, H., 1224 Griffiths, S., 2013. Evaluating marine spatial closures with conflicting fisheries and conservation 1225 objectives. J. Appl. Ecol. 50, 1060–1070. 1226

FAO Fisheries Department. 2003. The ecosystem approach to fisheries. FAO Technical Guidelines for 1227

141


Responsible Fisheries. No. 4, Suppl. 2. Rome, FAO. 1228 Francis, R.C., Hixon, M.A., Clarke, M.E., Murawski, S.A., Ralston, S., 2007. Ten commandments for 1229

ecosystem-based fisheries sciences. Fisheries 35, 217–233. 1230 Fulton, E.A., Link, J.S., Kaplan, I.C., Savina-Rolland, M., Johnson, P., Armsworth, C., Home, P., Gorton, 1231

R., Gamble, R.J., Smith, A.D.M., Smith, D.C., 2011a. Lessons in modeling and management of marine 1232 ecosystems: the Atlantis experience. Fish and Fish. 12, 171–188. 1233

Fulton E.A., Smith, A.D.M., Smith, D.C., 2007. Alternative management strategies for Southeastern 1234 Australian Commonwealth Fisheries: Stage 2: Quantitative Management Strategy Evaluation. Report 1235 to the Australian Fisheries Management Authority and the Fisheries Research and Development 1236 Corporation. CSIRO Marine and Atmospheric Research. 1237

Fulton, E.A., Smith, A.D.M., Smith, D.C., van Putten, I.E., 2011b. Human behavior: the key source of 1238 uncertainty in fisheries management. Fish and Fish. 12, 2–17. 1239

Gaichas, S.K., Aydin, K.Y., Francis, R.C., 2010. Using food web model results to inform stock assessment 1240 estimates of mortality and production for ecosystem-based fisheries management. Can. J. Fish. Aquatic 1241 Sci. 67, 1490–1506. 1242

Gaichas, S.K., Aydin, K.Y., Francis, R.C., 2011. What drives dynamics in the Gulf of Alaska? Integrating 1243 hypotheses of species, fishing, and climate relationships using ecosystem modelling. Can. J. Fish. 1244 Aquat. Sci. 68, 1553–1578. 1245

Gibson, G.A., Spitz., Y.H., 2011. Impacts of biological parameterization, initial conditions, and 1246 environmental forcing on parameter sensitivity and uncertainty in a marine ecosystem model for the 1247 Bering Sea. J. Mar. Sys. 88, 214–231 1248

Goodman, D., Mangel, M., Parkes, G., Quinn, T., Restrepo, V., Smith, T., Stokes, K., 2002. Scientific 1249 review of the harvest strategy currently used in the BSAI and GOA groundfish fishery management 1250 plans. North Pacific Fishery Management Council. Anchorage, AK. 1251

Heintz, R.A., Siddon, E.C., Farley, Jr., E.V., Napp, J.M., 2013. Correlation between recruitment and fall 1252 condition of age-0 pollock (Theragra chalcogramma) from the eastern Bering Sea under varying 1253 climate conditions. Deep-Sea Res. II 94, 150–156. 1254

Hermann, A.J., Gibson, G.A., Bond, N.A., Curchitser, E.N., Hedstrom, K., Cheng, W., Wang, M., Stabeno, 1255 P.J., Eisner, L., Ciecel, K.D., 2013. A multivariate analysis of observed and modeled biophysical 1256 variability on the Bering Sea shelf: Multidecadal hindcasts (1970-2009) and forecasts (2010-2040). 1257 Deep Sea Res. II 94, 121–139. 1258

Holsman, K.K., Ianelli, J., Aydin, K., Punt, A.E., Moffitt, E.A., This volume. A comparison of fisheries 1259 biological reference points estimated from temperature-specific multi-species and single-species 1260 climate-enhanced stock assessment models. Deep Sea Res. II 00, 00–00. 1261

Jurado-Molina, J., Livingston, P., Ianelli, J., 2005. Incorporating predation interactions in a statistical catch-1262 at-age model for a predator-prey system in the eastern Bering Sea. Can. J. Fish. Aquat. Sci.62, 1865–1263 1873. 1264

Kaplan, I.C., Horne, P.J., Levin, P.S., 2012. Screening California Current fishery management scenarios 1265 using the Atlantis end-to-end ecosystem model. Prog. Ocean. 102, 5-18. 1266

Levin, P.S., Fogarty, M., Murawski, S.A., Fluharty, D., 2009. Integrated Ecosystem Assessments: 1267 Developing the scientific basis for ecosystem-based management of the ocean. PLoS Biology 7(1): 1268 e1000014. doi:10.1371/journal.pbio.1000014 1269

Marasco, R.J., Goodman, D., Grimes, C.B., Lawson, P.W., Punt, A.E., Quinn II, T.J., 2007. Ecosystem-1270 based fisheries management: some practical suggestions. Can. J. Fish. Aquat. Sci. 64, 928–939. 1271

McLeod, K.L., Lubchenco, J., Palumbi, S.R., Rosenburg, A.A., 2005. Scientific consensus statement on 1272 marine ecosystem-based management. Signed by 221 academic scientists and policy experts with 1273 relevant expertise. Communication Partnership for Science and the Sea (COMPASS) [online]. 1274 Available from http://compassonline.org/?q=EBM. 1275

Moffitt, E., Punt, A.E., Holsman, K., Aydin, K.Y., Ianelli, J.N., Ortiz, I., This volume. Moving towards 1276 Ecosystem Based Fisheries Management: Options for parameterizing multi-specie sharvest control 1277 rules. Deep Sea Res. II 00, 00–00. 1278

142

http://apps.webofknowledge.com.offcampus.lib.washington.edu/full_record.do?product=WOS&search_mode=GeneralSearch&qid=1&SID=3ASvIM9r5uxaeNNKWBV&page=1&doc=3



http://compassonline.org/?q=EBM


Mueter, F.J., Bond, N.A., Ianelli, J.N., Hollowed, A.B., 2011. Expected declines in recruitment of walleye 1279 pollock (Theragra chalcogramma) in the eastern Bering Sea under future climate change. ICES J. Mar. 1280 Sci. 68, 1284–1296. 1281

Murawski, S.A., Matlock, G.C., editors. 2006. Ecosystem Science Capabilities Required to Support 1282 NOAA’s Mission in the Year 2020. U.S. Dep. Commerce, NOAA Tech. Memo. NMFS-F/SPO-74. 1283

North Pacific Fishery Management Council (NPFMC). 2012. Stock Assessment and Fishery Evaluation 1284 Report for the groundfish resources of the Bering Sea / Aleutian Islands region. North Pacific Fishery 1285 Management Council 605 West 4th Ave., Suite 306 Anchorage, AK 99501. 1297pp. 1286

Ortiz, I., Wiese, F.K., Grieg, A., 2012. Marine Regions Boundary Data for the Bering Sea Shelf and Slope. 1287 UCAR/NCAR-Earth Observing Laboratory/Computing, Data, and Software Facility. Dataset. 1288 http://dx.doi.org/10.5065/D6DF6P6C. 1289

Ortiz, I., Aydin, K., Hermann, A.J., This volume-a. Modeling fish movement, a case study of walleye 1290 pollock, Pacific cod and arrowtooth flounder in the eastern Bering Sea. Deep Sea Res. II 00, 00–00. 1291

Ortiz, I., Aydin, K., Hermann, A.J., Gibson, G., This volume-b. Climate to fisheries: a vertically integrated 1292 model for the eastern Bering Sea. Deep Sea Res. II 00, 00–00. 1293

Pacific Fishery Management Council (PFMC). 2012. Terms of Reference for the Groundfish and Coastal 1294 Pelagic Species Stock Assessment and Review Process for 2013-2014. Pacific Fishery Management 1295 Council, 7700 NE Ambassador Place, Portland, OR 97220, USA. 1296

Pikitch, E.K., Santora, C., Babcock, E.A., Bakun, A., Bonfil, R., Conover, D.O., Dayton, P., Doukakis, P., 1297 Fluharty, P., Heneman, B., Houde, E.D., Link, J., Livingston, P.A., Mangel, M., McAllister, M.K., 1298 Pope, J., Sainsbury, K.J., 2004. Ecosystem-based fishery management. Science 305, 346–347. 1299

Punt, A.E., A’mar, T., Bond, N.A., Butterworth, D.S., de Moor, C.L., Oliveira, J.A.A., Haltuch, M.A., 1300 Hollowed, A.B., Szuwalski, C., 2014. Fisheries management under climate and environmental 1301 uncertainty: Control rules and performance simulation. ICES J. Mar. Sci. 71, 2208-2220. 1302

Punt, A.E., Butterworth, D.S., de Moor, C.L., De Oliveira, J.A.A. and M. Haddon. Management Strategy 1303 Evaluation: Best practices. Fish. Res. 00, 00-00; 1304

Rose, K.A., Allen, J.I., Artioli, Y., Barange, M., Blackford, J., Carlotti, F., Cropp, R., Daewell, U., Edwards, 1305 K., Flynn, K., Hill, S,L., HilleRisLambers, R., Huse, G., Mackinson, S., Megrey, B., Moll, A,. Rivkin, 1306 R., Salihoglu, B., Schrum, C., Shannon, L., Shin, Y., Smith, S.L., Smith, C., Solidoro, C., St. John, M., 1307 Zhou, M., 2010. End-to-End modeling for the analysis of marine ecosystems: challenges, issues and 1308 next steps. Mar. Coast. Fish. 2, 115–130. 1309

Sainsbury, K.J., Punt, A.E., Smith, A.D.M., 2000. Design of operational management strategies for 1310 achieving fishery ecosystem objectives. ICES J. Mar. Sci. 57, 731–741. 1311

Siddon, E.C., Duffy-Anderson, J.T., Mueter, F.J., 2011. Community-level response of fish larvae to 1312 environmental variability in the southeastern Bering Sea. Mar. Ecol. Prog. Ser. 426, 225–239. 1313

Siddon, E.C., Heintz, R.A., Mueter, F.J., 2013a. Conceptual model of energy allocation in walleye pollock 1314 (Theragra chalcogramma) from age-0 to age-1 in the southeastern Bering Sea. Deep-Sea Res. II 94, 1315 140–149. 1316

Siddon, E.C., Kristiansen, T., Mueter, F.J., Holsman, K.K., Heintz, R.A., Farley, E.V., 2013b. Spatial 1317 match-mismatch between Juvenile fish and prey provides a mechanism for recruitment variability 1318 across contrasting climate conditions in the eastern Bering Sea. PLOS ONE 8 (12), e84526. 1319

Sigler, M.F., Heintz, R.A., Hunt, G.L. Jr., Lomas, M.W., Napp, J.M., Stabeno, P.J., This volume. A Mid-1320 trophic View of Subarctic Productivity: Lipid Storage, Location Matters and Historical Context. Deep-1321 Sea Res. II. 00, 00–00. 1322

Sissenwine, M.P., Murawski, S.A., 2004. Moving beyond “intelligent tinkering”: advancing an ecosystem 1323 approach to fisheries. Mar. Ecol. Progr. Ser. 274, 291–295. 1324

Smith, A.D.M., 1994. Management strategy evaluation – the light on the hill. In Population dynamics for 1325 fisheries management, pp. 249–253. Ed. by D. A. Hancock. Australian Society for Fish Biology, Perth. 1326

Smith, A.D.M., Sainsbury, K.J., Stevens, R.A., 1999. Implementing effective fisheries management 1327 systems – management strategy evaluation and the Australian partnership approach. ICES J. Mar. Sci. 1328 56, 967–979. 1329

143







Stabeno, P.J., Kachel, N.B., Moore, S.E., Napp, J.M., Sigler, M., Yamaguchi, A., Zerbini, A.N., 2012. 1330 Comparison of warm and cold years on the southeastern Bering Sea shelf and some implications for 1331 the ecosystem. Deep Sea Res. II, 65-70, 31-45. dii:http://dx.doi.org/10.1016/j.dsr2.2012.02.020 1332

Travers-Trolet, M., Shin, Y-J., Field, J.G., 2014. An end-to-end coupled model ROMS-N2P2Z2D2-1333 OSMOSE of the southern Benguela foodweb: parameterisation, calibration and pattern-oriente 1334 validation. African J. Mar. Sci. 36, 11-29. 1335

Uchimaya, T., Kruse, G.H., Mueter, F.J., This volume. A multispecies biomass dynamics model for 1336 investigating predator-prey interactions in the Bering Sea groundfish community. Deep Sea Res. II 00, 1337 00–00. 1338

Wang, M., Overland, J.E., Bond, N.A, 2010. Climate projections for selected large marine ecosystems. J. 1339 Mar.e Sys. 79, 258–266. 1340

Wiese, F.K., Wiseman, Jr., W.J., Van Pelt, T.I., 2012. Bering Sea Linkages. Deep Sea Res. II. 65-70, 2–5. 1341 Witherell, D., Pautzke, C.P., Fluharty, D., 2000. An ecosystem-based approach for Alaska groundfish 1342

fisheries. ICES J. Mar. Sci. 57, 771–777. 1343 1344 1345 1346 1347

1348

144

http://dx.doi.org/10.1016/j.dsr2.2012.02.020


1349

1350 1351 Figure 9.1. Outline of the Bering 10K ROMS-NPZD-FEAST-FAMINE model, showing data flows across 1352 coupled modules. 1353

1354

145


1355 1356 Figure 9.2. Conceptual outline of Management Strategy Evaluation. 1357 1358 1359

146


1360 Figure 9.3. Management scenarios for Management Strategy Evaluation. 1361

1362 1363

147


Chapter 10 Conclusions 1364 1365

1366 Initial expectations were that most of the conclusions would come from the forecasts, directly addressing 1367 potential impacts of climate on fish distribution, fleet behavior and management strategy evaluation. This 1368 became untenable as getting a viable hindcast of FEAST took more than expected and there was not enough 1369 time to run the forecasts. Thus, our conclusions center on the redefined objectives and the challenges of 1370 developing an integrated model of this magnitude with economics and management objectives in mind, 1371 what we learned from the hindcast, and incorporating field data into the model. 1372 1373 The final products of this projects are thus: 1374

1a. Develop bioenergetics model of fish growth for walleye pollock, arrowtooth flounder, Pacific cod, 1375 herring, salmon, capelin, sandlance, eulachon, myctophids, squids, crabs, shrimp, and 1376 miscellaneous zooplankton. (Chapter 3) 1377

1b. Model reproductive output and cycle of fish groups. (Chapter 3) 1378 1c. Develop movement model for fish and fully couple to NPZ and ROMS (Chapter 3 and 4 and 5) 1379

2. Simulate hindcast with spatially explicit fisheries removals as estimated by project B.72 1380 Economics. (Chapter 5 and 8) 1381

3. Coding fully coupled to economics and MSE (Chapter 6 and 8) 1382 3. Select ecosystem/ production indices for multi-species model (MSMt) from Bering 10K-ROMS-1383

NPZD 2003-2039) forecasts based on three different climate models (from NSF 0732534) 1384 (Chapter 7) 1385

5. Set-up a fully coupled integrated ecosystem model (Climate to fisheries) designed for MSE (Chapter 1386 8 and 9) 1387

1388 We implemented a spatially explicit multispecies bioenergetics local model that makes two notable 1389 contributions: first the use of dynamic prey fields which enables the model to respond to local conditions 1390 and their temporal variability. Second, we modified the activity parameter to takes into account size-based 1391 energy expenditure caused by foraging and locating prey. This allows the model to distinguish the energetic 1392 costs across different sizes, even when the prey field is the same for all lengths. The model is more realistic 1393 and allows a better understanding of ontogenetic changes in energy costs of foraging. We also formulated 1394 a model for reproductive output and cycle of fish based on size, age, and energy available for egg 1395 production. This implementation requires further fine-tunning, as much of the survival of juveniles 1396 depended on food availability during the winter, which is drastically decreased in the NPZD model. To 1397 better address this, we could have developed simpler models, like a non spatial model or pseudo spatial 1398 model using static prey fields to have a better control of the prey availability during the first winter. We 1399 overcame this shortcoming by correcting the number of modelled age zeroes, to those estimated in the stock 1400 assessments. This is not ideal as the sudden increase creates an unexpected high increase in the demand of 1401 local prey fields, but it allows the model to simulate continuously through all ages and provides useful 1402 comparisons for future improvements in the mortlality of age-zeroes. The movement model combines prey 1403 availability and predation mortality to guide direction and size-specific speed of movement. We were able 1404 to quantify the contribution of these factors across multiple sizes, with predation avoidance being more 1405 relevant for the smaller sizes of pollock. In contrast, small sizes of arrowtooth respond to prey availability 1406 only, as they experience almost no risk of predation. These two factors can expleain the general distribution 1407 patterns of pollock, Pacific cod and arrowtooth flounder in the Bering Sea shelf. The movement model also 1408 predicts migrations out of the shelf, showing migration routes through the north to Russian waters. 1409 1410 The hindcast simulation was succesfuly run, showing there are distinct spatiotemporal patterns across the 1411 eastern Bering shelf and that conditions in the 1970’s were different than those experienced in more recent 1412

148


years. Hindcasts were run with and without FEAST, and from both simulations we extracted time series of 1413 ecosystem indicators used as forcing functions. ROMS-NPZD forecasts simulations were conducted under 1414 three climate scenarios, and from those, time series of ecosystem indicators were also extracted and applied 1415 to drive multispecies model. An offshoot of this objective is the implementation of short-term forecasts 1416 using ROMS-NPZD; this is currently used to estimate the size of the cold pool but will be extended in the 1417 future to other indicators. Results from this have been shown to Plan Teams and the North Paciifc Fisheries 1418 Management Council. 1419 1420 In hindsight, we summarized into one objective what was effectively a multi-step process and development 1421 of three independent models: ROMS-Bering10K, NPZ and FEAST. While the model was successfully 1422 coded and coupled to the economics and MSE, no forecasts were done and we negatively impacted those 1423 projects dependent in these forecasts and simulations. Re-evaluating project progress, we recognize 1424 forecasts could have been run without compromising model complexity by running forecasts with one-way 1425 feedback between the NPZD and FEAST model and with FISH mortality and growth turned off in FEAST. 1426 This would have allowed for the MSE and economics routines to be tested and forecasts simulated without 1427 the burden of the level of performance of FEAST. The movement function alone in FEAST, even without 1428 fish growth or mortality, would have sufficed to simulate fishing effort response to climate change (as 1429 filtered by fish distribution and ice cover) would have allowed for the MSE function under these conditions. 1430 In this case, MSE would still have estimated the quota and the multispecies assessment models could have 1431 been tested against the single species. Unfortunately, because of excessive emphasis placed on FEAST 1432 performance, this option was not considered. An important lesson here is to reconsider the extent to which 1433 product dependent projects (in this case FAMINE and MSE) can still benefit from preliminary results, or 1434 minimum results needed to pass on to the next project. In this case, the need of ROMS to perform with very 1435 little bias in the estimation of temperature, was unnecessarily carried over to the expectation on the level of 1436 performance required for FEAST to be a useful product for the subsequent projects. The first hindcasts of 1437 FEAST were in May 2012, and with the modifications described above, forecasts could have been started 1438 then. We note that this date would have been past the original schedule, however most prior delays were 1439 outside the control of the project. Here we highlight a few of the most time-consuming issues throughout 1440 the project: i) change of the extent of the model to include seabirds and mammals. During the first PI 1441 meeting, NPRB requested for seabirds and marine mammals to be included in the model. Assessing the 1442 feasibility of this took time and focus away from the actual model development. ii) Delay in purchase and 1443 set up of first cluster. iii) Inability to test NPZ at different sites: the initial output from the NPZD used for 1444 FEAST development came from constructed data, not from the linked model. Ths meant there was a unique 1445 set of conditions to test the bioenergetics of fish, which required testing under multiple and varying 1446 conditions. Although we did develop a 1D version of the bioenergetics, it proved to e insufficient to 1447 adequately model stability and behavior. iv) Model complexity and simulation time: despite the fact that 1448 FEAST does not have vertical movement, the underlying model ROMS was actually increased from 40 to 1449 60 layers, adding time to the simulations. Multiple attempts were made to simplify this, however effort 1450 towards the development of a simplified effort was interrupted as preference was given to the 60 layer 1451 model. Although ultimately we did develop a simplified ROMS, this was not a easy task and multiple trials 1452 were performed before adopting this approach. v) Lack of winter zooplankton: logistically, this was one of 1453 the most challenging aspects to overcome in order for the model to run smoothly through winter. We 1454 highlight there was little guidance here to overcome this challenge in the NPZD model, and eventhough the 1455 model runs through winter, the amount of zooplankton during this season is a very much needed 1456 improvement. Some work has been done towards this end, however the code has not been available to 1457 update the NPZD-FEAST portion. We recognize that while these issues were outside our control, we shoud 1458 have been mor realistic about how much we could make up for lost time. These aggregate delays drove 1459 FEAST well off-schedule, however all throughout, there was an expectation that time could be made up by 1460 simplifying the model further into a 2D version. This 2D version was dependent on ROMS-NPZD hindcast 1461 simulation of the model with 60 layers that has not been donducted to date. Failure to realistically assess 1462 the feasibility of building a faster version hindered the development and implementation of alternative non 1463

149


FEAST dependent products/methods for the MSE project. 1464 1465 Despite the lack of time to run the forecasts, we proved the feasibility of developing a high spatial resolution 1466 model that links climate, oceanography and lower trophic levels to fish and fisheries. The FEAST food web 1467 is comprised of 25 groups (including those in the NPZD model), with fisheries, and captures the interactions 1468 between the environment and the main forage fish in the Bering Sea. We demonstrated the feasibility of a 1469 model of this complexity to be designed to address economics and management questions. Commonly, 1470 MSE and economics studies are conducted using readily available models, having to fill-in gaps or add 1471 conversions between model outputs and MSE required inputs. Here we developed the model to produce the 1472 output in the format required to MSE, despite the added complexity or run time of the simulations. We were 1473 also able to use FEAST as a model to focus research in other areas where gaps in process understanding or 1474 lack of corroboration data were found –these stemming from either the model development itself or the 1475 analysis of the hindcast. This has initiated discussion into new lines of research, the need for new and 1476 improved surveys, the joint prioritization of lines of research and the implementation of multidisciplinary 1477 working groups targeting distinct lines of research. Lastly, the integrated model provides a framework onto 1478 which field and modeling work results from climate to fish can be incorporated into a strategic management 1479 tool, and provide time series for stock assessment models strengthening the ties between fieldwork and 1480 management. FEAST allows specific components to be evaluated quantitatively as to their impact on fish 1481 populations or fisheries, and also serves as a tool from which other tools such as ecosystem indices, 1482 environmental indices, and short term forecasts can be derived. 1483 1484

Collaborative processes and best practices 1485

1486

Here we discuss some program level issues that, while not under the control of our project, were part of the 1487 day-to-day project management. Initially, the Ecosystem Modeling Committee served to coordinate the 1488 modeling effort and was useful in identifying gaps in product delivery across modeling –particularly 1489 between the oceanography and NPZ. However, its role later on became confounded, functioning as some 1490 kind of model evaluation committee but without providing any feedback, despite statistical plans submitted 1491 by the various modeling groups and requests for specific guidelines. There was also no follow-up to issues 1492 raised during meetings that had no immediate solution and there were no records of such meetings made 1493 available. Future overarching committees should be available for questions/feedback year round and follow 1494 up on recommendations or agreements. This does not mean such committee should be qualified to address 1495 all questions, but it should if needed, point to a body or expert that can address the issues be those related 1496 to logistics, communications, statistical support, etc. 1497 1498 Data incorporation into the model assumed that field results would be readily available shortly after the 1499 field work started. Though the data collected was available, the analysis and results of such data was not. 1500 In some cases we were able to switch model development to a different component to wait for the results 1501 and modify the model accordingly, but in others we were not able to benefit from them. Future programs 1502 should differentiate between incorporation of fieldwork data and fieldwork processed results, as the first is 1503 more straightforward to incorporate than the second. 1504 1505 Other lessons learned and best practices for the overall project were addressed in the discussion and 1506 conclusion section in Chapter 9. Here we retake those lessons learned and detail specifically how these best 1507 practices –or lack there of- supported or hindered the development of FEAST, and they relate to our 1508 experience integrating field data into the model and overall participation in BSIERP, BEST Synthesis and 1509 the Alliance group at AFSC (this last is a NOAA collaborative initiative that includes PMEL, Behavioral 1510 Ecology, Recruitment Processes, Ecosystem Monitoring and Assessment, Recruitment Energetics and 1511

150


Coastal Analysis). Our intent is these will aid in the planning and development of future integrated research 1512 programs. An assessment of the benefits and shortcomings of FEAST within BSIERP is included in the 1513 section “BSIERP and Bering Sea Project Connections”. 1514 1515 1. Be realistic about what can be accomplished within a given timeline, anticipate potential impacts of 1516

missing deadlines on related projects, and have simpler alternatives in advance. In our case, our 1517 delay prompted the behavioral foraging model team to switch approaches by using empirical as 1518 opposed to predicted prey data. We note though, that they made this switch early on and were able 1519 to establish new milestones with almost no adverse impacts to their project. The MSE project kept 1520 the same approach until much later in the project, initially only cutting simulation time. The MSE 1521 team recognized that an alternative simple operating model would have provided some results from 1522 the management strategy evaluation. We also recognize we could have conducted forecasts by 1523 limiting the functionality (but not compromising the complexity) of FEAST. These alternatives 1524 should have been explored either implemented from the start or re-evaluated earlier in the process 1525

2. Larger-scale modeling projects require logistical support on par with fieldwork. This would extend 1526 support to invest in database set-up and maintenance to facilitate data use and sharing. For example 1527 the AFSC has completely redesigned and joined various datasets from BASIS, spring surveys and 1528 others into cloud-based and ArcServer applications such as ECO-DAAT (not public yet). But we 1529 note this was not the case at project onset and data from academic institutions, in general, is rarely 1530 set-up this way. 1531

3. Clear separation of scientific versus logistics oversight. A coordinator would have prioritized activities, 1532 and prevented bottlenecks in the process. Also, ideally a coordinator would have prevented 1533 excessive emphasis on FEAST performance and could have facilitated earlier discussions of 1534 alternative plans to provide some products to the economics and MSE projects. 1535

4. Open and frequent communication particularly in the form of workshops was highly productive. We 1536 attended multiple workshops, some to request help in specific areas, others to learn about findings 1537 from other projects. In all cases, these personal interactions facilitated collaboration. 1538

5. Re-evaluate data available for study. We’ve found new information or new approaches stemming from 1539 cross-collaboration can guide a novel analyses on readily available data, and recommend re-1540 exploring datasets when either new teams or people with different expertise join the research. This 1541 was the case not only in our interactions with other projects, but within our own group. For example, 1542 analyzing FEAST results prompted different analysis on the food lab data and also different 1543 analyses of the bottom trawl survey data. 1544

6. Synthesis work is sequential and iterative as opposed to simultaneous and independent (non iterative). 1545 The value of cross-collaboration and synthesis lies in joint discussions, but is very time consuming 1546 and is best in person (at least until the discussion is more focused). Time was a bigger issue with 1547 remote collaborators as impromptu meetings were hard. 1548

7. Mismatch of required performance levels and performance measures between single discipline 1549 approaches and multidisciplinary ones. Differences in spatio-temporal scales were not identified 1550 and addressed early on – this included scale of processes studied, sampling surveys, etc. as well as 1551 relevance across disciplines. Here we note that while the issues with the underlying ROMS and 1552 NPZD were mostly performance based, metrics and results were not necessarily a critical issue for 1553 the Economics and MSE overall goal. There should have been be a clear process to prioritize both. 1554 This process was not clear from the beginning and was ad hoc in BSIERP. 1555

8. Designate an entity to compile and distribute data. For BSIERP the inventory of data sources was 1556 assumed to be known and accessible to all participants of the model effort. In practice this cost time 1557 to the modelers, and prevented full integration of data in some cases. 1558

9. Address lack of familiarity with data/methods pertaining to other disciplines: despite the multiple 1559 modelling meetings in BSIERP, there was no meeting in which a summary explanation of data 1560 limitation, spatio-temporal coverage, amount of data available, etc. was explained in detail. This 1561 forstalled a baseline understanding of the data at hand and capabilities across disciplines. 1562

151


10. Coherence of final products from different funding/research agencies/groups. Despite the large amount 1563 of cross-collaboration, teams/institutions were ultimately responsible to their corresponding 1564 funding sources, which had different priorities and expectations on the final product product. This 1565 impacted the time allotted for synthesis and in particular for the integrated modeling effort, and 1566 forced the modelers to make-up for gaps expected to be covered by other projects. 1567

1568 1569

Next steps and future work 1570

1571 The six-year development of FEAST as a large-scale collaborative effort has created a fertile ground for 1572 continuing integration efforts in the Bering Sea and beyond. Going forward, the Alaska Fisheries Science 1573 Center is making FEAST a centerpiece of its strategy for developing an Integrated Ecosystem Assessment 1574 (IEA) for the Alaskan region, focusing on the Bering Sea. This effort, part of NOAA’s national IEA 1575 program (http://www.noaa.gov/iea/) will include regular updates to FEAST, but will also use the model as 1576 a focus for collaborative fieldwork from disciplines from physics through biology, economics, and social 1577 sciences. Moreover, as IEAs focus on delivering management results, this will serve as a direct conduit for 1578 bringing process-oriented fieldwork into the management arena via management strategy analyses, 1579 ecosystem indicator development, and improved prediction capabilities both in the short and long term. 1580 1581

Guidance for field work 1582

1583 The development process of FEAST was rich in interaction between field researchers and modelers. The 1584 process has been continued at the AFSC by implementing a more collaborative framework between the 1585 ecosystem and stock assessment modelers and the oceanography, zooplankton, and juvenile fish research 1586 groups part of PMEL and AFSC. Essentially, new multidisciplinary working groups address topics of 1587 interest, starting first with those that can be addressed with readily available data. New surveys are being 1588 discussed, as well as improved sampling methods, best sampling gear, required funding, and potential 1589 funding sources. Some of the detailed needs (e.g. specific bioenergetics parameters) are listed in individual 1590 chapters; a few specific overarching gaps are important to consider here. 1591 1592 First and foremost, fall blooms may play a previously unacknowledged key role in the Bering Sea 1593 ecosystem. Heintz et al. (2013) linked the condition factor in fall to recruitment survival over winter. 1594 Estimates from the FEAST hindcast show there would be enough small phytoplankton as well as other 1595 favorable conditions to fuel a zooplankton fall bloom. Furthermore, weekly diet data from pollock stomachs 1596 collected 1982-2012, show evidence of an increase in zooplankton consumption. This evidence supports 1597 the need for: 1) a pilot study in late fall – winter surveys, for zooplankton and pollock, and 2) extend 1598 sampling at moorings for Chla measurements at medium and deep depths, as subsurface blooms occur later 1599 than those at the surface. Ideally, there would be a way to standardize current zooplankton surveys 1600 conducted by AFSC to have directly comparable seasonal measurements. 1601 1602 Second, the primary focus of BSIERP on pollock has brought a higher integration of the understanding of 1603 its life history, relation to environmental factors and fisheries. A similar integration effort is intended for 1604 cod and arrowtooth within the AFSC as part of the new collaborative effort (Alliance). Importantly, focus 1605 should be placed on the pelagic/benthic connections, in particular in the relationship between cod and 1606 commercial crab. While individual projects have examined cod predation on commercial crabs, the 1607 incorporation of these dynamics into a multispecies dynamic picture is an important next step. 1608

152


1609 A third focus area for model/field collaboration is the extension of FEAST into the Northern Bering Sea. 1610 While the current model includes physics up to the Bering Strait, and produces outputs for these northern 1611 regions, the model’s plankton and fish community are specifically calibrated to the southeast Bering Sea. 1612 In particular, NPRB Project 1423 (Farley et al.) has been funded to parameterize FEAST specifically for 1613 the plankton and fish important to juvenile survival of Arctic-Yukon-Kuskokwim (AYK) region Chinook 1614 salmon, and merge this modeling with fieldwork on juvenile salmon survival. 1615 1616 The Aleutian Islands represent a final focus area for future model/field collaboration. The modeling of 1617 Aleutian stocks in FEAST (and associated analysis of field data) has not to date been funded. In particular, 1618 such a modeling effort could resolve spatial and temporal prey fields for endangered Steller sea lions in 1619 relation to local fisheries interactions. 1620 1621 Other data gaps to enhance/inform model performance: seasonal cycle and relative increase in abundance 1622 (e.g. magnitude difference between spring and fall bloom) emphasis in overall winter data (diets, 1623 zooplankton, primary production), arrowtooth settlement: dynamics, phenology; geopgraphy; total 1624 abundance estimates as opposed to site-specific density (zooplankton); egg data for pollock, cod and 1625 arrowtoorh; migration of fish (pollock and cod and arrowtooth) to Russia and Aleutian Islands; overall 1626 forage fish data. Details of data gaps can also be found in Table 8.2 1627

Ongoing and future modeling work 1628

1629 The FEAST hindcast and coupling to FAMINE and MSE served as a proof of concept for the integration 1630 of a climate to fisheries model designed to address hypothesis testing and serve management objectives. 1631 Future work will focus on conducting not only the forecasts originally planned for MSE, but also short 1632 term-forecast of ecosystem conditions and the use of ROMS-NPZ as well as FEAST to test and develop 1633 ecosystem indicators. Additionally, we believe this first incursion of fisheries into high performance 1634 computing should encourage and support extension of this approach to other regions, as well as other 1635 biological programs and areas of research. 1636 1637 In particular, the next two years of the FEAST modeling effort, leveraged by funds from the NOAA IEA 1638 and Fisheries and the Environment (FATE) programs, will focus on: 1639

1. The development of indicators for ecosystem assessment from current hindcast results. The 1640 hindcast itself has produced over a terabyte of model results, including detailed physics, plankton, 1641 and fish predictions for the 1970-present. Key indicators might include water temperature, ice 1642 meltback patterns, and plankton production. One FATE-funded project will determine the 1643 historical relationships between specific FEAST results and quantities important to stock 1644 assessments (e.g. recruitment, natural mortality, growth) for incorporation into stock assessments. 1645 A second FATE-funded project will evaluate current ecosystem indicators under different future 1646 climate conditions. 1647

2. The development of nine-month, near-term forecasts from downscaled atmospheric projections. 1648 For the past several years, nine-month climate projections have been contributed by Nick Bond to 1649 the ecosystems chapter of the North Pacific Fisheries Management Council’s annual Stock 1650 Assessment and Fisheries Evaluation (SAFE) report. This project would downscale these 1651 atmospheric projections to create nine-month forecasts of FEAST quantities, including plankton 1652 production and juvenile fish survival. The initial stage of this project, as funded by the IEA 1653 program, will be to validate such predictions from hindcast re-forecasts. 1654

3. The completion of 50-year forecasts, and associated management strategy analyses, 1655 incorporating fisher prediction and behavior. This work, originally planned as part of the BSIERP 1656 effort (FAMINE), will continue to completion under IEA funds. 1657

153


1658 1659

1660

154


Chapter 11. BSIERP and Bering Sea Project connections 1661 1662 FEAST had a key role in serving both as a framework to synthesis much of the current understanding for 1663 mechanisms and processes on the Bering Sea shelf, as well as to test management scenarios. As such, its 1664 integration activities had three main aspects: 1) to promote collaboration with field biologists and integrate 1665 as much new data/processes as possible into the fish model; 2) to ensure the setup was compatible with 1666 current stock assessments, reference points, catch allocation, and catch accounting systems, and 3) to couple 1667 the fish model to the oceanography and lower trophic models, and use forcing files from climate projections 1668 for forecasts. Early on, Al Hermann joined the FEAST team serving as a bridge between the oceanography, 1669 lower trophic levels and fish modeling. 1670 1671 An end-to-end model is a synthesis exercise in itself, benefitting from BSIERP’s requirement for linkages 1672 across projects and data sharing before publishing. We found the following to be key elements for future 1673 consideration: 1674 1675 Benefits 1676 1677

- Strong emphasis from NPRB in linkages across projects, data sharing and cross-collaboration. 1678 This aspect (as well as magnitude of funds involved) further motivated 1679

- An institution-wide support within NOAA and prioritization of cross-collaboration within AFSC 1680 and PMEL as well collaboration across both institutions. This support was at least partially brought 1681 about by NPRB’s requirement to establish as many linkages as possible across projects. 1682

- The change in philosophy and day-to-day business practices would not have been possible without 1683 BSIERP. At the core of this transition were: 1684

- Dedicated multiday-workshops to share partial results and open discussions. We believe this to be 1685 one of the most productive mechanisms to improve communication across groups, promoting cross-1686 collaboration, integration, and multidisciplinary approaches. 1687

- Additional funding from NPRB to support workshops organized a posteriori facilitated discussions, 1688 particularly within the modelling group. These workshops had not been planned as part of the 1689 proposals. 1690

- These workshops were also the primary way of consensus building for the fish modeling between 1691 modelers and field biologists, openly discussing model assumptions, limitations and developing 1692 best approach in collaboration with the field biologists who supplied data, statistical models and/or 1693 expert opinion. This applies from top level meetings like the annual BSIERP PI and Ecosystem 1694 Modeling Committee meetings, down to smaller multi-project workshops and meetings. 1695

- In particular model development was improved by direct input at diverse times from multiple 1696 projects: seasonal bioenergetics, surface trawl survey, surface trawl survey acoustics, the projects 1697 on pollock and cod distribution, functional foraging response, forage distribution + ocean 1698 conditions, fish, birds and mammals, economic-ecological model for pollock and cod, management 1699 strategy evaluation, and downscaling global climate projections. 1700

- Conversely, FEAST provided a quantitative framework to incorporate field data into a fisheries 1701 management tool, and in some cases helped focus some of the field data analyses to address specific 1702 gaps. 1703

- The ad hoc model structure of FEAST specifically addresses North Pacific Fisheries Management 1704 Council policies and target species/sectors, and was jointly development with economists, stock 1705 assessment modelers and management strategy evaluation modelers. As such the model is set-up 1706 to follow up on economics and MSE projects. 1707

- Additional funds from NPRB increased computer power by providing a computer cluster, which 1708 allowed faster testing, faster simulations, and increased the amount of simulations for the vertically 1709 integrated model (this benefitted M.5, B.72, B.73, as well as this project). 1710

155


1711

Shortcomings 1712

1713 • Requests for changes in model objectives at the beginning of the project took focus away from 1714

actual model development. During an early BSIERP PI meeting, researchers from other groups 1715 strongly opposed to the omission of marine mammals and seabirds in the model (which had been 1716 considered for inclusion but later excluded due to time and funding restrictions). Marine mammals 1717 and seabirds were not included in the end due to data constraints, and marine mammal/seabird 1718 researchers agreed, but the time spent addressing the issue was time not spent in model 1719 development. 1720

• Cumulative delays impacted the objectives and extent of both the Economics and Management 1721 Strategy Evaluation projects within the BSIERP timeline. 1722

• Dispersed data from projects in isolated institutions lacked support/unified coordination. This had 1723 a detrimental effect on the development of the NPZ model as there was no readily organized 1724 database to supply data for model corroboration. 1725

• Lag between integrated model and individual components. This will always be the case when all 1726 model components are in active development, and it means that the performance of the integrated 1727 model will not be the same as that of individual components. 1728

• Forage fish ouput for B.74 Behavioral foraging model. Originally, we thought we would be able 1729 to provide model output of forage fish location and abundance to be used as input in the behavioral 1730 model. We were not able to provide such output within their time frame, which prompted them to 1731 switch to empirical data from the Patch Dynamics group. 1732

1733 1734

Meetings with other projects 1735

1736 We actively participated in the following meetings/workshops (several of them with additional funding 1737 from NPRB): 1738 1739

- BSIERP PI meetings (2008, 2009, 2010, 2011, 2012, 2014) 1740 - Ecosystem Modeling Committee meetings (2008, 2009) 1741 - Modeling workshops 15+ participants (July 2008, Dec 2009, funded by NPRB, lead by 1742

Wiese/Aydin) 1743 - Modeling workshops 4-8 participants (FEAST-MSE-Economics, 11/12/2009, lead by Aydin) 1744 - Field fish biologists – Modelers workshops 20+ particpants ( 08/2009, 3 day, lead by Hollowed & 1745

Aydin; 02/11/2010, 1 day, fish growth organized by Ortiz & Aydin; 02/18/2010, 1 day, fish 1746 movement, organized by Ortiz and Aydin). 1747

- Monthly BSIERP PI conference calls 1748 - Weekly/biweekly ROMS-NPZ-FEAST progress check (Aydin, Ortiz, Hermann, Gibson) through 1749

2008- Mar 2010. These meetings were aimed at discussing both planning and technical modeling 1750 issues, needs, compatibility, coding, etc. Additionally and particularly during the first year Aydin 1751 and Ortiz met with Hermann and Bond to get a better understanding of the details of the 1752 oceanographic and atmospheric modeling. These discussions were later joined by the SAB. 1753

- Weekly/biweekly ROMS-NPZ-FEAST-ECON-MSE progress check (Aydin, Ortiz, Hermann, 1754 Gibson, Moffitt, Holsman) through Apr 2010 – Sep 2013. MSE joined to better integrate data 1755 outputs and formats needed for MSE project. These discussion were later joined by the SAB. 1756

156


- Ad hoc meetings with individual researchers 2009-2014 (Aydin and/or Ortiz met with (list not 1757 exhaustive) to Jeff Napp, Ed Farley, Ron Heintz, Steve Barbeaux, Nathan Bacheler, Sandy Parker-1758 Stetter, Patrick Ressler, Trot Buckley, Lisa Eisner, Mike Sigler, George Hunt, Phyllis Stabeno, 1759 Nick Bond). 1760

- Ad hoc BEST-BSIERP Synthesis project meetings either via phone or at PMEL starting 01/2012-1761 02/2014 were attended by either Ortiz or Aydin. 1762

- BEST-BSIERP Synthesis workshops attended remotely by Aydin (Bermuda 02/2012 and Friday 1763 Harbor 05/2013). 1764

- Ad hoc meetings for marine regions Ortiz met with Francis Wiese and Angie Greig, (additionally 1765 most consultations with researchers was conducted via e-mail) 2010-2013. 1766

- Ad hoc meetings for Calorie-sheds Ortiz met with Henry Huntington and other collaborators of the 1767 LTK group (2010-2013). 1768

1769 1770

Projects/ activities stemming from BSIERP Program 1771

1772 BSIERP marine regions of the Bering Sea shelf and slope 1773 Developed marine regions to facilitate data/ results comparison across projects and supplied region 1774 description and documentation, images, shapefiles for ArcGIS users, and mapped common survey stations 1775 onto regions (BASIS, summer Bottom Trawl Survey, BEST and Polar Cruises). All materials are available 1776 at http://dx.doi.org/10.5065/D6DF6P6C 1777 1778 Using HPCC techniques to power user tools for ecosystem models: the Bering Sea example 1779 Developed web based multi-user tools for visualization and analysis of large datasets (megabytes to 1780 terabytes) in advanced formats (netCDF, HDF, etc.). The project parallelized workloads across compute 1781 nodes in a High Performance Computing Cluster (HPCC). The biophysical output of the ecosystem model 1782 FEAST was used as a test case. Data analysis and files were provided to test set-up developed by AOOS. 1783 Similar tools can be viewed here http://data.aoos.org/maps/search/models-grids.php. This project differed 1784 in that model output was multivariate (50 variables), while most current tools are uni or bivariate. 1785 1786 Calorie-sheds for Subsistence harvest 1787 This collaboration was facilitated by inteactions with the Local Traditional Knowledge projects part of 1788 BSIERP. Ortiz collaborated with Henry Huntington and other members of the LTK group to address use of 1789 resources by indigenous communities.The collaboration produced a paper lead by Huntington: Huntington, 1790 H. I. Ortiz, G. Noongwook, M. Fidel, D. Childers, M. Morsef, J. Beaty. 2013. Mapping human interaction 1791 with the Bering Sea ecosystem: comparing seasonal use areas, lifetime use areas, and “calorie-sheds”. 1792 Deep-Sea Research II. 94:292-300 http://www.sciencedirect.com/science/article/pii/S0967064513001136 1793 1794 BEST Synthesis project for the Bering Sea 1795 Ortiz and Aydin attended meetings and workshops; this collaboration produced a paper lead by George 1796 Hunt and submitted to the 4th BSIERP special issue: Hunt, G.L., Ressler, P.H., De Robertis, A., Aydin, K., 1797 Gibson, G., Sigler, M.F., Ortiz, I., Lessard, E., Williams, B., Pinchuk, A., and Buckley, T. submitted. 1798 Euphausiids in the aastern Bering Sea: A synthesis of recent studies of 2 euphausiid production, 1799 consumption and population control. Deep Sea Research II 00:00-00. 1800 1801 Visuals and materials showing applications of models developed by project M.5 for fisheries management 1802 issues. 1803 These materials usually involve schematics and videos/animations, which involve collaborations between 1804 researchers to achieve the final products. For example, processed FEAST output was regridded using script 1805

157


http://data.aoos.org/maps/search/models-grids.php

http://www.sciencedirect.com/science/article/pii/S0967064513001136


by Hermann to create composite images of bottom temperature, ice cover, ice algae and fish were then 1806 reprocessed in ArcGIS and Photoshop and finally animated using Quicktime-Pro. 1807 1808

Management or policy implications 1809

1810 The original direct application of the FEAST model was to serve as framework on which to: i) test 1811 projections of fleet behavior (with B71), and ii) test various policy scenarios for management strategy 1812 evaluation under different climate projections stemming from different earth system models. While these 1813 applications have been postponed (but will be pursued) until further refinement of FEAST, the current 1814 FEAST version and hindcast do have other applications for management: 1815 1816 Ecosystem indicators and annual ecosystem assessments: FEAST provides a framework to test and develop 1817 oceanographic ecosystem indicators. These indicators are part of the annual Ecosystem Assessment for the 1818 Bering Sea and the Ecosystem Considerations Chapter for the North Pacific Fisheries Management Council. 1819 The Ecosystem Assessment provides both a baseline (from timeseries of selected indicators) and current 1820 state of the ecosystem which is taken into consideration when adjusting the recommended quotas 1821 (Acceptable Biological Catch, ABC; Total Allowable Catch, TAC), particularly for the pollock Bering Sea 1822 fishery. 1823 1824 Times series of environmental indices: similar to ecosystem indicators, but these are developed for stock 1825 assessments, either to test correlations with recruitment, or as proxies of productivity/climate variability 1826 and include similar indices to those used in the blended forecasts for MSMt described in the final report to 1827 NPRB of B.73 MSE. We note the difference between ROMS-NPZ and ROMS-NPZ-FEAST indices is that 1828 zooplankton biomasses are coupled to fish predation in FEAST, whereas in NPZ predation mortality of 1829 zooplankton is always proportional to their abundance. 1830 1831 Mid-term plans with already funded projects include: a) testing and improving three currently used 1832 ecosystem indicators under different climate projections; the three indicators include sea ice retreat index, 1833 cold pool extent, mean zooplankton biomass; b) developing ecosystem indicators for inclusion in stock 1834 assessments (FATE projects); and b) short-term -9 month- forecast of zooplankton production; this short-1835 term forecast is aimed at more directly measuring the potential for recruitment and prey available for fish 1836 stocks, with particular interest in pollock. 1837

Publications 1838

1839 Publications in peer reviewed journals 1840 1841 1. Livingston, PA, K Aydin, JL Bolt, AB Hollowed, and JM Napp. 2011. Alaskan marine fisheries 1842

management: advances and linkages to ecosystem research. In A Belgrano and W Fowler 1843 (eds.), Ecosystem-Based Management for Marine Fisheries: An Evolving Perspective. Cambridge 1844 University Press, pp 113-152. 1845

2. Huntington, H. I. Ortiz, G. Noongwook, M. Fidel, D. Childers, M. Morsef, J. Beaty. 2013. Mapping 1846 human interaction with the Bering Sea ecosystem: comparing seasonal use areas, lifetime use 1847 areas, and “calorie-sheds”. Deep-sea Research II. 94:292-300 1848 http://www.sciencedirect.com/science/article/pii/S0967064513001136 1849

3. Buckley, T.W., I. Ortiz, S. Kotwicki, K. Aydin, In review. Summer diet composition of walleye pollock 1850 in the eastern Bering Sea, 1987-2011 and predator-prey relationships with copepods and 1851

158

http://www.sciencedirect.com/science/article/pii/S0967064513001136


euphausiids. Deep-Sea Research II, 00:00-00. (BSIERP 4th special issue). 1852 4. Moffitt, E., Punt, A.E., Holsman, K., Aydin, K.Y., Ianelli, J.N. and I. Ortiz. Submitted. Moving towards 1853

Ecosystem Based Fisheries Management: options for parameterizing multi-species harvest control 1854 rules. Deep-sea Research II 00: 00-00. 1855

5. Holsman K.K. and K. Aydin. In review. Comparative methods for evaluating climate change impacts on 1856 the foraging ecology of Alaskan groundfish. Marine Ecology Progress Series 00: 00–00. 1857

6. Holsman, K.K., Ianelli, J., Aydin, K., A.E. Punt, A.E. and E.A. Moffit. Submitted. Comparative 1858 biological reference points estimated from temperature-specific multispecies and single species 1859 stock assessment models. Deep-sea Research II 00: 00-00. [Contributions by Punt and Moffitt to a 1860 project primarily funded elsewhere] 1861

7. Ianelli, J., Holsman, K., Punt, A.E. and K. Aydin. Submitted. Multi-model inference for incorporating 1862 trophic and climate uncertainty into stock assessment estimates of fishery biological reference 1863 points. Deep Sea Research II 00:00-00. 1864

8. Punt, A.E., Ortiz, I., Aydin, K.Y., Hunt Jr. G.L., Wiese, F.K. Submitted. Integrated ecosystem modeling 1865 and fisheries management strategy evaluation: where are we and what is the future? Deep Sea 1866 Research II 00:00-00. (BSIERP 4th special issue) 1867

9. Hunt, G.L., Ressler, P.H., De Robertis, A., Aydin, K., Gibson, G., Sigler, M.F., Ortiz, I., Lessard, E., 1868 Williams, B., Pinchuk, A., and Buckley, T. submitted. Euphausiids in the aastern Bering Sea: A 1869 synthesis of recent studies of 2 euphausiid production, consumption and population control. Deep 1870 Sea Research II 00:00-00. (BSIERP 4th special issue) 1871

10. Ortiz, I., K. Aydin and A. J. Hermann. In prep. Modeling fish movement, a case study of walleye 1872 pollock, Pacific cod and arrowtooth flounder in the eastern Bering Sea. 1873

11. Ortiz, I., K. Aydin, A. J. Hermann, and G. Gibson. In prep. Climate to fisheries: a vertically integrated 1874 model for the eastern Bering Sea. Deep Sea Research II 00:00-00. (BSIERP 4th special issue). 1875

12. Aydin, K., I. Ortiz, and A. J. Hermann. In prep. New approach to modeling fish bioenergetics, a case 1876 study for walleye pollock, Pacific cod and arrowtooth flounder in the Eastern Bering Sea. 1877

1878 1879 Publications non peer reviewed 1880 1881 1. Aydin K., Bond N., Curchitser E. N., Gibson M. G. A., Hedström K., Hermann A. J., Moffitt E., Murphy, 1882

J., Ortiz, I., Punt, A., and Wang M. Integrating data, fieldwork, and models into an ecosystem-level 1883 forecasting synthesis: the Forage–Euphausiid Abundance in Space and Time (FEAST) model of 1884 the Bering Sea Integrated Research Program. ICES Document CM 2010/L: 21, Nantes, 1885 France; 2010 available at http://www.ices.dk/sites/pub/CM%20Doccuments/CM-2010/L/ 1886 L2110.pdf 1887

2. Moffit, E., Ortiz, I., Punt, A. 2011. Summary and results of the management strategy evaluation 1888 workshop (Oct 27-28 2011), Seattle, WA. Technical Report for NPRB. 24p. 1889

3. Ortiz, I., Wiese, F.K., Grieg, A., 2012. Marine Regions Boundary Data for the Bering Sea Shelf and 1890 Slope. UCAR/NCAR-Earth Observing Laboratory/Computing, Data, and Software Facility. 1891 Dataset. http://dx.doi.org/10.5065/D6DF6P6C. 1892

4. Moffitt, E. Punt, A.E., ianelli, J. Aydin, K., Ortiz, I., Hlsman, K., and Dalton, M. 2014. BSIERP 1893 Management Strategy. North Pacific Research Board Final Report B73, xx p 1894

1895 1896

1897

159

http://www.ices.dk/sites/pub/CM%20Doccuments/CM-2010/L/%20L2110.pdf

http://www.ices.dk/sites/pub/CM%20Doccuments/CM-2010/L/%20L2110.pdf



Poster and oral presentations at scientific conferences or seminars 1898

1899 Presentations 1900 1901 1. 2014 April, Big Island, Hawaii. FUTURE Open Science Meeting. Kerim Aydin. Models linking climate 1902

to fish: Spatially explicit ecosystem model. 1903 2. 2014 April, Big Island, Hawaii. FUTURE Open Science Meeting. Kirstin Holsman, Kerim Aydin, Jim 1904

Ianelli and André E. Punt Using multi-species models to evaluate climate and trophic impacts on 1905 recommended harvest rates of groundfish in the Bering Sea 1906

3. 2014. April, Big Island, Hawaii. FUTURE Open Science Meeting. Kerim Aydin, Ivonne Ortiz, Albert 1907 J. Hermann, Georgina A. Gibson and André E. Punt Evaluating long-term climate predictions for 1908 the Bering Sea ecosystem using a suite of modeling approaches. 1909

4. 2014 March, Bozeman, Montana. Daniel Goodman Memorial Symposium. Ivonne Ortiz.The scientific 1910 challenges to implementing complex end-to-end ecosystem models. 1911

5. 2014 February, Honolulu, Hawaii. Bering Sea Open Sciences Meeting. Ivonne Ortiz, Kerim Aydin, Al 1912 Hermann, Georgina Gibson. The benefits of hindsight: examining results of the Bering Sea 1913 Project’s vertically-integrated modeling effort from physics to fish. 1914

6. 2014 February, Honolulu, Hawaii. Bering Sea Open Sciences Meeting. George L. Hunt, Jr., Kerim 1915 Aydin, Hongsheng Bi, Alex DeRobertis, Georgina Gibson, Alexei Pinchuk, Patrick Ressler. What 1916 controls the distribution and abundance of euphausiids over the southeastern Bering Sea shelf? 1917

7. 2014 February, Honolulu, Hawaii, ASLO Ocean Sciences Meeting. Kirstin K. Holsman, Kerim Aydin, 1918 Jim Ianelli, André Punt, Elizabeth Moffitt. Using multi-species models to predict climate-change 1919 impacts on Bering Sea (AK) Fisheries. 1920

8. 2013 October, Nanaimo, Canada. Kirstin K. Holsman, Kerim Aydin and Jim Ianelli. Using multi-species 1921 food-web and assessment models to evaluate climate change impacts on fisheries 1922

9. 2013 January, Alaska Marine Sciences Symposium, Anchorage, AK. Holsman, K.K.*, Ianelli, J.N., 1923 Aydin, K. and A.E. Punt. 2013. The influence of climate change and predation on biological 1924 reference points estimated from multispecies and single species stock assessment models. 1925

10. 2012 October, Hiroshima, Japan. PICES Annual Meeting. Ivonne Ortiz, Kerim Aydin and Albert J. 1926 Hermann. 20 species, 15 lengths: How fish move driven by happiness as defined by growth and 1927 predation. 1928

11. 2012 October, Hiroshima, Japan Stephani Zador, Kirstin Holsman, Sarah Gaichas and Kerim Aydin. 1929 Developing indicator-based ecosystem assessments for diverse marine ecosystems in Alaska. 1930

12. 2012 May, Yeosu, Korea. PICES/ICES Climate Change Effects on Fish and Fisheries. Ivonne Ortiz, 1931 Kerim Aydin and Al Hermann (Presented by Nicholas Bond). From climate to fisheries: 1932 Performance of a 40-year hindcast for the Eastern Bering Sea. 1933

13. 2012 January, Anchorage, Alaska. Alaska Marine Sciences Symposium. Ortiz, I., Aydin, K. Hermann, 1934 A. Climate to Fisheries: evaluating a 40-year hindcast of the FEAST model for the Bering Sea. 1935

14. 2011 May, Seattle, WA Ecosystem Studies of Sub-Arctic Seas. Ortiz, I., Aydin, K. Hermann, A. Forage-1936 euphausiid abundance in Space and Time: Seasonal patterns 1937

15. 2011 April, Seattle, WA Think Tank School of Aquatic and Fisheries Sciences, University of 1938 Washington. Ortiz, I., Aydin, K. and E.A. Moffitt. 2011. Modeling fish in a vertically integrated 1939 model, from climate to MSE. 1940

16. 2011 March BSIERP PI, Anchorage, Alaska. Aydin, K. Progress on FEAST and Bioeconomics. 1941 17. 2010 November, Anchorage, AK. Wakefield Symposium. Aydin, K. Hermann, A. Ortiz, I. The FEAST 1942

model for the Bering Sea. 1943 18. 2010 November Anchorage, AK. Wakefield Symposium. Ortiz, I. and Robson, B. Using Regional Food 1944

Webs to Explore Fisheries and Foragers Interactions: A Case study on Northern Fur Seals (funding 1945 for project not from BSIERP). 1946

19. 2010 October, Portland, USA PICES Annual Meeting. Albert J. Hermann, Kerim Aydin, Nicholas A. 1947

160


Bond, Wei Cheng, Enrique N. Curchitser, Georgina A. Gibson, Kate Hedstrom, Ivonne Ortiz, 1948 Muyin Wang and Phyllis J. Stabeno (Invited). Modes of biophysical variability on the Bering Sea 1949 shelf. 1950

20. 2010 October, Portland, Oregon. PICES Annual Meeting. Jackie R. King, Vera N. Agostini, Chris J. 1951 Harvey, Gordon A. McFarlane, Michael G. Foreman, James E. Overland, Nicholas A. Bond and 1952 Kerim y. Aydin Climate forcing and the California Current ecosystem 1953

21. 2010 October, Portland, Oregon PICES Annual Meeting. Kerim Aydin and Troy Buckley. An analysis 1954 of 30 years of seasonal and geographic variability in marine food webs through fish food habits 1955 and stable isotope analyses. 1956

22. 2010 September, Nantes, France. ICES Annual Science Meeting. Aydin K., Bond N., Curchitser, E.N., 1957 Gibson M. G. A., Hedström K., Hermann A. J., Moffitt E., Murphy, J., Ortiz, I., Punt, A., and Wang 1958 M. Integrating data, fieldwork, and models into an ecosystem-level forecasting synthesis: the 1959 modeling challenge of the Bering Ecosystem Study/Bering Sea Integrated Research Program 1960 (BEST-BSIERP). 1961

23. 2010 September, Nantes, France. ICES Annual Science Meeting. Ortiz, I. Spatially explicit food web 1962 models and implications on natural mortality. 1963

24. 2010 April, Sendai, Japan PICES/ICES Climate Change Effects on Fish and Fisheries talk Ivonne Ortiz, 1964 Kerim Aydin, Nicholas A. Bond, Enrique N. Curchitser, Georgina A. Gibson, Kate Hedström, 1965 Albert J. Hermann. Integrating data, fieldwork, and models into an ecosystem-level forecasting 1966 synthesis: The modeling challenge of the Bering Ecosystem Study/Bering Sea Integrated Research 1967 Program (BEST-BSIERP). 1968

25. 2010 April, Sendai, Japan PICES/ICES Climate Change Effects on Fish and Fisheries. Albert J. 1969 Hermann, Kerim Aydin, Nicholas A. Bond, Wei Cheng, Enrique N. Curchitser, Georgina A. 1970 Gibson, Kate Hedström, Ivonne Ortiz, Muyin Wang and Phyllis J. Stabeno. Simulated modes of 1971 biophysical variability on the Bering Sea shelf. 1972

26. 2010 April, Sendai, Japan PICES/ICES Climate Change Effects on Fish and Fisheries. Stephani Zador 1973 and Kerim Aydin. Patterns in a changing climate: Fine-scale analysis of arrowtooth flounder catch 1974 rates in the eastern Bering Sea reveals spatial trends in abundance and diet. 1975

27. 2010 January, Anchorage, Alaska. Alaska Marine Sciences Symposium. Aydin. K., Buckley, T. and 1976 Hunsicker, M. A web for all seasons: an analysis of 30 years of seasonal and geographic variability 1977 in marine food webs through fish food habits and stable isotope analyses. 1978

28. 2009 October, Girdwood, Alaska. BSIERP PI Meeting. Aydin, K. Dalton, M., Haynie, A., Hermann, 1979 A. Ortiz, I., Pfeiffer, L. Upper trophic levels B70, B71, B72. 1980

29. 2009 June, Ortiz, I. and Aydin, K. Zooplankton to top predator dynamics on a fine scale in the Eastern 1981 Bering Sea (FEAST). 1982

30. 2009 June, GLOBEC, Victoria, Canada. Ortiz, I. and Aydin, K. The FEAST model for the Bering Sea: 1983 Forage/Euphasiid Abundance in Space and Time. 1984

31. 2009 January, Anchorage, Alaska. Alaska Marine Sciences Symposium. Ortiz, I. FEAST: 1985 Forage/Euphasiid Abundance in Space and Time. 1986

1987 1988 Posters 1989 1990 1. 2014. February, Honolulu, Hawaii, ASLO Ocean Sciences Meeting. Ortiz, I., Aydin, K., and Hermann, 1991

A. and Gibson, G. Seasonal fish growth and mortality during cold and warm years in the eastern 1992 Bering Sea. 1993

2. 2014 February, Honolulu, Hawaii. Bering Sea Open Sciences Meeting. Kirstin K. Holsman, Kerim 1994 Aydin, Jim Ianelli, André Punt, Elizabeth Moffitt. Using multi-species models to evaluate climate 1995 and trophic impacts on recommended harvest rates of groundfish in the Bering Sea (AK). 1996

3. 2013 January, Anchorage, Alaska. Alaska Marine Sciences Symposium. Ortiz, I., Aydin, K. and 1997 Hermann, A. FEAST Forage and Euphausiid Abundance in Space and Time Assumptions, 1998

161


knowledge and gaps. 1999 4. 2012 January, Alaska Marine Sciences Symposium, Anchorage, AK. Moffitt, E.A. Punt, A.E., Ianelli, 2000

J.N., Holsman, K.K., Aydin, K. and I. Ortiz. 2012. Definition of multi-species control rules for the 2001 Bering Sea management strategy evaluation. 2012. Alaska Marine Science Symposium, 2002 Anchorage, AK. January 2012 (poster). 2003

5. 2011 January, Alaska Marine Sciences Symposium, Anchorage, AK. Huntington, H., and Ortiz, I. 2004 “Calorie-Sheds” of Subsistence Harvests: Testing the Concept in Togiak and Savoonga, Alaska 2005

6. 2011 January, Alaska Marine Sciences Symposium, Anchorage, AK. Ortiz, I., Wiese, F., Greig, A. 2006 Marine Regions of the Bering Sea shelf and slope. 2007

7. 2011 May Seattle, WA. Ecosystem Studies of Sub-Arctic Seas (ESSAS) Kerim Aydin, Ivonne Ortiz and 2008 Albert J. Hermann. Forage-euphausiid abundance in space and time: Seasonal patterns 2009

8. 2011 May Seattle, WA. Ecosystem Studies of Sub-Arctic Seas (ESSAS). Albert J. Hermann, Georgina 2010 A. Gibson, Kerim Aydin, Nicholas A. Bond, Wei Cheng, Enrique N. Curchitser, Kate Hedstrom, 2011 Ivonne Ortiz, Muyin Wang, Phyllis Stabeno, Lisa Eisner and Markus Janout Modeled and observed 2012 modes of biophysical variability on the Bering Sea shelf 2013

9. 2010 February, Portland, Oregon. Ocean Sciences Meeting.A J Hermann, K Aydin, N A Bond, W Cheng, 2014 E N Curchitser, G A Gibson, K Hedstrom, I Ortiz, M Wang, P J Stabeno. The Bering Sea Integrated 2015 Ecosystem Research Program: downscaling climate change to a subarctic region with coupled 2016 biophysical models. 2017

10. 2010 February, Portland, Oregon. Ocean Sciences Meeting. T Buckley, T TenBrink, J M Napp, K Aydin 2018 Comparing zooplankton consumed by walleye pollock and caught by nets in the eastern Bering 2019 Sea. 2020

11. 2009 January, Alaska Marine Sciences Symposium, Anchorage, Alaska. FEAST: Forage/Euphausiid 2021 Abundance in Space and Time. 2022

12. 2009 October BSIERP PI Meeting, Girdwood, Alaska. FEAST: Ortiz, I., Aydin, K., Hermann, A. 2023 Forage/Euphausiid Abundance in Space and Time (Spatial patterns in diets during cold and warm 2024 years). 2025

13. 2008 October BSIERP PI Meeting, Girdwood, Alaska. Aydin, K., Hermann, A., Ortiz, I. FEAST: 2026 Forage/Euphausiid Abundance in Space and Time (model structure) 2027

2028 2029 2030

2031

162


Outreach/workshops 2032

2033 Our outreach activities had two main target audiences: i) general public/ kids, and ii) management agencies/ 2034 fishery stakeholders. Below are the details of these efforts, some still ongoing and now part of our program’s 2035 (Resource Ecology and Ecosystem Modeling, AFSC, NOAA) activities. 2036 2037 General Public 2038 2039 Factsheets 2040 1. Hungry Fish Make a Difference (BSIERP Headlines) Ortiz, I., Aydin, K. and Hermann, A. 2041

http://nprb.org/assets/images/uploads/BSH_70_Hungry_Fish.pdf 2042 2. Subsistence Food Comes from a Vast Area ((BSIERP Headlines) Huntington, H., Ortiz, I., Noongwook, 2043

G., Fidel, M. Childers, D., Morse, M., Beaty, J., Alessa, L., Kliskey, A. 2044 http://nprb.org/assets/images/uploads/BSH_69_Subsistence_Food.pdf 2045

2046 Events 2047 3. 2011-present Nine past events of Scientist Spotlight at the Pacific Science Center. Ivonne Ortiz is a 2048

science communication fellow since 2011, developed a hands-on activity “Food webs and food 2049 chains in the Bering Sea”, and attends three events a year. Event is scheduled once a month with 2050 booths throughout the museum where children and the general public can interact one-to-one with 2051 local scientist. Each event hosts several science communication fellows who have developed a 2052 hands-on activity specifically with the Pacific Science Center. 2053

2054 Presentations in Schools (K-12) 2055 4. 2010 November St. Paul School district: Two talks to middle school classrooms; talk included excel lab 2056

with Food Lab data. 2057 2058 Teachers’ Workshop 2059 5. 2012 July, Seattle, Alaska Fisheries Science Center. Ivonne Ortiz presented at the High School Science 2060

Teachers Workshop; Climate to Fisheries talk, walk through excel lab with Food Lab data and 2061 showed hands-on activity. The theme was connectivity of the marine environment and humans, 2062 from climate to oceanography, production, fish, fisheries and management. 2063

2064 Academic/ Fishery Stakeholders Workshops 2065 6. 2014 April, Big Island, Hawaii, FUTURE Open Science Meeting. Kerim Aydin was Co-convened with 2066

Anne Hollowed, Kirstin Holsman the session:. Ecosystem projection model inter-comparison and 2067 assessment of climate change impacts on global fish and fisheries. 2068

7. 2011 October, Seattle, WA. MSE workshop organized as part of BSIERP by Moffitt, Ianelli and Punt. 2069 Ivonne Ortiz and Kerim Aydin attended workshop and gave presentations of FEAST. Fisheries 2070 scientists, ecosystem modelers, fisheries economists, industry representatives, and NPFMC 2071 representatives (Council and SSC members as well as NPFMC staff) were present and provided 2072 input related to the choice of management strategies and scenarios to consider in the MSE project 2073 and the definition of control rules for multi-species assessment. 2074

2075 Presentations to Management/ Policy Institutions 2076 8. 2014 June, Washington, DC. Congressional Briefing, Presentation by Kerim Aydin: Closing the Data 2077

Gaps: Challenges in Stock Assessment Science presentation/ question session. 2078 9. 2014 February Kerim Aydin on BSIERP modeling (FEAST and MSE). NPFMC SSC, presentation/ 2079

question session. 2080

163

http://nprb.org/assets/images/uploads/BSH_70_Hungry_Fish.pdf

http://nprb.org/assets/images/uploads/BSH_69_Subsistence_Food.pdf


10. 2013 September Presentation by Kerim Aydin on Ecosystem Based Management and FEAST. NPFMC 2081 Ecosystem Committee. 2082

11. 2013 September Presentation by Jim Ianelli on: Developing simple multi-species trophic interaction 2083 models (as mediated by FEAST) driven by climate for testing management systems. NPFMC, 2084 SSC. 2085

12. 2012 December Presentation by Jim Ianelli on an update of BSIERP activities. NPFMC SSC. 2086 13. 2012 September Presentation of FEAST by Kerim Aydin to Groundfish Plan Teams for the Bering Sea/ 2087

Aleutian Islands and the Gulf of Alaska. 2088 14. 2012 March Presentation by Jim Ianelli on the outcomes of the October 2011 MSE workshop and the 2089

preliminary results of the development of the multi-species harvest control rules. NPFMC SSC. 2090 2091

Acknowledgements 2092

2093 We would like to thank all the researchers that contributed to the development of FEAST, in particular Ron 2094 Heintz, Steve Barbeaux, Ed Farley, Patrick Ressler, and Matt Wilson. Anne Hollowed facilitated many 2095 discussions and closer collaboration with field biologists. Many thanks to our model partners, Liz Moffitt, 2096 Jim Ianelli, Michael Dalton, James Murphy, Charlotte Boyd. Kirstin Holsman for her contributions for 2097 implementing the use of ecosystem indicators in a multispecies model. Al Hermann for being a true bridge 2098 between the ROMS, NPZ and FEAST teams and models, and willingness to explain model nuisances as 2099 well as physical processes. Nick Bond and Muyin Wang did the same with atmospheric processes and 2100 climate models. Enrique Curchitser and Kate Hedstrom developed the original NEP5 ROMS. Georgina 2101 Gibson developed the NPZD model. Sarah Gaichas early on for initial development of the Ecosim model 2102 for MSE. Henry Huntington for a fruitful collaboration with the LTK group. Mike Sigler for coordinating 2103 meetings with the SAB and reviewing an earlier version of this report. Lastly, Clarence Paultzke and 2104 specially Francis Wiese for their continued support to the modeling effort despite the many challenges. 2105 2106

2107

164


Literature cited: 2108 2109 Aitkin, M. 1991. Posterior Bayes factors. Journal of the Academic Society Series B 53:111-142. 2110 Akaike, H. 1973 Information theory and an extension of the maximum likelihood principle, in: 2nd 2111

International Symposium on Information Theory. Academiai Kiado, Budapest, pp. 267-281. 2112 A’mar, Z.T., Punt, A.E. and Dorn, M.N. 2009. The evaluation of two management strategies for the Gulf 2113

of Alaska walleye pollock fishery under climate change. ICES Journal of Marine Science 66:1614-2114 1632. 2115

Aydin, K., Gaichas, S., Ortiz, I., Kinzey, D and Friday, N. 2007. A Comparison of the Bering Sea, Gulf of 2116 Alaska, and Aleutian Islands Large Marine Ecosystyems Through Food Web Modeling. NOAA 2117 Tech Memo NMFSAFSC 178, 298 pp. 2118

Brodziak, J. and C.M. Legault. 2005. Model averaging to estimate rebuilding targets for overfished stocks. 2119 Canadian Journal of Fisheries and Aquatic Sciences 62: 544–562 2120

Brodziak, J. and K. Piner. 2010. Model averaging and probable status of North Pacific striped marlin, 2121 Tetrapturus audaz. Canadian Journal of Fisheries and Aquatic Sciences 67: 793–805. 2122

Buckland, S.T., Burnham, K.P. and N.H. Augustin. 1997. Model selection: an integral part of inference. 2123 Biometrics 53: 603–618. 2124

Burnham, K.P. and D.R. Anderson. 1998 Model Selection and Inference: a Practical Information-2125 theoretical Approach. Springer Verlag, New York. 353 pp 2126

Chimielechi, R.M. and A.E. Raftery. 2011. Probabilistic visbility forecasting using Bayesian Model 2127 Averaging. Monthly Weather Review 139: 1626–16361 2128

Clement, J.L., Cooper, L.W., Grebmeier, J.M., 2004. Late winter water column and sea ice conditions in 2129 the Northern Bering Sea. Geophysical Research, 109, C03022, doi:10.1029/2003JC002047. 2130

Christensen, V., Walters, C., 2004. Ecopath with Ecosim: methods, capabilities and limitations. Ecological 2131 Modelling 172, 109–139. 2132

Danielson, S., Eisner, L., Weingartner, T., Aagaard, K. 2011. Thermal and hyaline variability over the 2133 central Bering Sea shelf: Seasonal and interannual perspectives. Continental Shelf Research 2134 31:539-554. 2135

Danielson, S., Curchitser, E., Hedstrom, K., Weingartner, T, and Stabeno, P. 2011. On ocean and sea ice 2136 modes of variability in the Bering Sea, J. Gephys. Res., 116, C12034, doi:10.1029/2011JC007389. 2137

Davies, N., Hoyle, S., Hampton, J., 2012. Stock assessment of striped marlin (Kajikia audax) in the 2138 southwest Pacific Ocean. WCPFC document WCPFC-SC8-2012/SA-WP-05. 90 pph 2139

Durban, J.W., Elston, D.A., Ellifrit, D.K., Dickson, E., Hammond, P.S. and P.M. Thompson. 2005. 2140 Multisite mark-recapture for cetaceans: population estimates with Bayesian model averaging. 2141 Marine Mammal Science 21: 80–92. 2142

Field, J.C., Francis, R.C. and K. Aydin. 2006. Top-down modelling and bottom-up dynamics: linking a 2143 fisheries-based ecosystem model with climate hypotheses in the Northern California Current. 2144 Progress in Oceanography 68: 238–270. 2145

Fissel B. Dalton, M., Felthoven, R., Garber-Yonts, B., Haynie, A., Himes-Cornell, A., Kasperski, S., Lee, 2146 J., Lew, D., Pfeiffer, L., Seung, C. 2012. Stock assessment and fishery evaluation report for the 2147 Groundfish fisheries of the Gulf of Alaska and Bering Sea/Aleutian Islands area: Economic Status 2148 of the Groundfish fisheries off Alaska, 2012. North Pacific Fisheries Management Council. 2149 Anchorage, AK. 2150

Flato, G.M., Boer, G.J., Lee, W.G., McFarlane, N.A., Ramsden, D., Reader, M.C. and A.J. Weaver. 2000. 2151 The Canadian Centre for Climate Modelling and Analysis global coupled model and its climate. 2152 Climate Dynamics 16: 451-467 2153

Flato G.M. and Boer G.J. 2001. Warming asymmetry in in climate change simulations. Geophys. Res. Lett., 2154 28, 195-198. 2155

Francis, R.I.C.C., 2011. Data weighting in statistical fisheries stock assessment models. Canadian Journal 2156 of Fisheries and Aquatic Sciences 68: 1124–1138. 2157

165


Fournier, D.A., Sibert, J.R., Majkowski, J. and J, Hampton. 1990. MULTIFA: a likelihood-based methods 2158 for estimating growth parameters and age composition from multiple length frequency data sets 2159 illustrated using data for southern bluefin tuna (Thunnus maccoyii). Canadian Journal of Fisheries 2160 and Aquatic Sciences 55: 2105–2116. 2161

Fournier, D.A., Skaug, H.J., Ancheta, J., Ianelli, J., Magnusson, A., Maunder, M.N., Nielsen, A. and J. 2162 Sibert. 2012. AD Model Builder: using automatic differentiation for statistical inference of highly 2163 parameterized complex nonlinear models. Optimization Methods and Software 27: 233–249. 2164

Fulton, E. A., A. D. M. Smith, and C. R. Johnson. 2004. Biogeochemical marine ecosystem models, I. 2165 IGBEM: a model of marine bay ecosystems. Ecological Modelling 174:267–307. 2166

Fulton, E.A., Link, J.S., Kaplan, I.C., Savina-Rolland, M., Johnson, P., Ainsworth, C., Horne, P., Gorton, 2167 R., Gamble, R.J., Smith, A.D.M., and D.C. Smith. 2011. Lessons in modelling and management of 2168 marine ecosystems: the Atlantis experience. Fish and Fisheries 12: 171–188. 2169

Grebmeier, J.M., Overland, J.E., More, S.E., Farley, E.V., Carmack, E.C., Cooper, L.W., Frey, K.E., Helle, 2170 J.H., McLaughllin, F.A., McNutt, S.L. 2006. A major ecosystem shift in the northern Bering Sea. 2171 Science, 311: 1461-1464. 2172

Hamel, O.S. 2011. Rebuilding analysis for Pacific Ocean Perch in 2011. Pacific Fishery Management 2173 Council, 7700 NE Ambassador Place, Portland, OR 97220. 15 pp. 2174

Hamel, O.S. and K. Ono. 2011. Stock Assessment of Pacific Ocean Perch in Waters off of the U.S. West 2175 Coast in 2011. Pacific Fishery Management Council, 7700 NE Ambassador Place, Portland, OR 2176 97220. 165 pp. 2177

Heintz, R.A., Siddon, E.C, Farley, Jr. E.V., Napp, J.M. 2013. Correlation between recruitment and fall 2178 condition of age-0 pollock (Theragra chalcogramma) from the eastern Bering Sea under varying 2179 climate conditions. Deep Sea Research II 94:150-156. 2180

Hermann, A. J., Gibson, G.A., Bond, N.A., Curchitser, E.N., Hedstrom, K., Cheng, W., Wang, M., Stabeno, 2181 P.J., Eisner, L. and Cieciel, K.D. 2013. A multivariate analysis of observed and modeled biophysical 2182 variability on the Bering Sea shelf: multidecadal hindcasts (1970-2009) and forecasts (2010-2040). Deep 2183 Sea Research II, doi:10.1016/j.dsr2.2013.04.007. 2184 Hermann, A.J., Gibson, G.A., Bond, N.A., Curchitser, E.N., Hedstrom, K., Cheng, W., Wang, M., and 2185

Stabeno, P.J. in review. Projected futures biophysical states of the Bering Sea. Deep Sea Research 2186 II 00:00-00. 2187

Hoeting, J.A., Madigan, D., Raftery, A.E. and C.T. Volinsky. 1999. Bayesian Model Averaging: a tutorial. 2188 Statistical Sciences 14: 382–417. 2189

Honkalehto, T., Williamson, N., Jones, D., Mccarthy, A. And D. McKelvery. 2008. Results of the Echo 2190 Integration-Trawl Survey of Walleye Pollock (Theragra chalcogramma) on the U.S. and Russian 2191 Bering Sea Shelf in June and July 2007. NOAA Technical Memorandum NMFS-AFSC-190 63 pp. 2192

Horner, P.J., Kaplan, I.C., Marshall, K.N., Levin, P.S., Harvey, C.J., Hermann, A.J. and E.A. Fulton. 2010. 2193 Design and parameterization of a spatially explicit ecosystem model of the Central California 2194 Current. NOAA Technical Memorandum NMFS-NWFSC-104, NOAA, USA. 140 ppHoff, G.R. 2195 and Britt, L.L. 2009. Results of the 2008 Eastern Bering Sea Upper Continental Slope Survey of 2196 Groundfish and Invertebrate Resources. NOAA Technical Memorandum NMFS-AFSC-197 313 2197 pp. 2198

Hollowed, A.B., Curchitser, E.N., Stock, C. A., and Zhang, C.I. 2013. Trade-offs associated with different 2199 modeling approaches for assessment of fish and shellfish responses to climate change. Climate 2200 Change 119: 111-129. 2201

Holsman K.K. and K. Aydin. In review. Comparative methods for evaluating climate change impacts on 2202 the foraging ecology of Alaskan groundfish. Marine Ecology Progress Series 00: 00–00. 2203

Holsman K.K., Ianelli, J., Aydin, K., Punt, A.E. and E.A. Moffitt. In review. Comparative biological 2204 reference points estimated from temperature-specific multispecies and single species stock 2205 assessment models. Deep Sea Research II 00:00-00 2206

2207 Hulson, P-J,F., Quinn, T.J. II, Hanselman, D.H. and J.N. Ianelli. 2013. Spatial modeling of Bering Sea 2208

166

http://tandfonline.com/doi/abs/10.1080/10556788.2011.597854




walleye pollock with integrated age-structured assessment models in a changing environment. 2209 Canadian Journal of Fisheries and Aquatic Sciences 70: 1402–1416 2210

Ianelli, J.N., Honkalehto, T., Barbeaux, S., Kotwicki, S., Aydin, K. And N. Williamson. 2012. Pg 51-155 2211 in Stock Assessment and Fishery Evaluation Report for the Groundfish Resources for the Bering 2212 Sea/Aleutian Islands Regions. North Pacific Fishery Management Council, 605 West 4th Ave, Suite 2213 306, Anchorage AK 99501.Jurado-Molina, J. and P. Livingston. 2002. Climate-forcing effects on 2214 trophically linked groundfish populations: implications for fisheries management. Canadian 2215 Journal of Fisheries and Aquatic Sciences 59: 1941–1951. 2216

Jurado-Molina, J., Livingston, P. and J. Ianelli. 2005. Incorporating predation interactions in a statistical 2217 catch-at-age model for a predator-prey system in the eastern Bering Sea. Canadian Journal of 2218 Fisheries and Aquatic Sciences 62: 1865–1873. 2219

Kaplan, I.C., Brown, C.J., Fulton, E.A., Gray, I.A., Field, J.C. and A.D.M. Smith. 2013. Impacts of 2220 depleting forage species in the California Current. Environmental Conservation, 1-14 2221 http://dx.doi.org/10.1017/S0376892913000052 ) 2222

Kass, R.E. and A.E. Raftery. 1995. Bayes factors. Journal of the American Statistical Association 90: 773–2223 795 2224

Kim, S.-J., G.M. Flato, G.J. Boer and N.A. McFarlane, 2002: A coupled climate model simulation of the 2225 Last Glacial Maximum, Part 1: transient multi-decadal response. Climate Dynamics 19: 515-537 2226 Kinzey, D., and A.E. Punt. 2009. Multispecies and single-species models of fish population 2227 dynamics: comparing parameter estimates. Natural Resource Modelling 22: 167–204 2228

Kishi, M.J., Kashiwai, M., Ware, D.M., Megrey, B.A., Eslinger, D.L., Werner, F.E., Aita, M.N., Azumaya, 2229 T., Fujii, M., Hashimoto, S., Huang, D., Iizumi, H., Ishida, Y., Kang, S.,Kantakov, G.A., Kim, H.-2230 C., Komatsu, K., Navrotsky, V.V., Smith, S.L., Tadokoro, K.,Tsuda, Yamamura, O., Yamanaka, 2231 Y., Yokouchi, K., Yoshie, N., Zhang, J., Zuenko, Y.I.,Zvalinsky, V.I., 2007. NEMURO — a lower 2232 trophic level model for the North Pacific marine ecosystem. Ecological Modelling 202, 12–25. 2233

Kolody, D. 2011. An age-, sex- and spatially-structured stock assessment of the Indian Ocean swordfish 2234 fishery 1950-2009, including special emphasis on the south-west region. IOTC Document IOTC-2235 2011-WPB-17. 75 pp. 2236

Kolody, D., Campbell, R. and N. Davies. 2008. A MULTIFAN-CL stock assessment of southwest Pacific 2237 swordfish 1952-2007. WCPFC document WCPFC-SC4-2008/SA-WP-6 (Rev1). 90 pp. 2238

Kotwicki, S., and Lauth, R.R. 2013. Detecting temporal trends and environmentally-driven changes in the 2239 spatial distribution of groundfishes and crabs on the easterb Bering Sea shelf. Deep –Sea Research 2240 II, 94: 231-243. 2241

Lauth, R.R. and E. Acuna. 2009. Results of the 2008 Eastern Bering Sea Continental Shelf Bottom Trawl 2242 Survey of Groundfish and Invertebrate Resources. NOAA Technical Memorandum NMFS-AFSC-2243 195 241 pp. 2244

Legutke S and R. Voss. 1999. ECHO-G, the Hamburg atmosphere-ocean coupled circulation model. DKRZ 2245 technical report 18, DKRZ, Hamburg. Lehodey, P., F. Chai, and J. Hampton. 2003. Modelling 2246 climate-related variability of tuna populations from a coupled ocean–biogeochemical–population 2247 dynamics model. Fisheries Oceanography 12:483–494. 2248

Livingston, P. A., Aydin, K., Boldt, J., Ianelli, J., Jurado-Molina, J. 2005. A framework for ecosystem 2249 impacts assessment using an indicator approach. ICES Journal of Marine Science, 62: 592-597. 2250

Magnusson, A., Punt, A.E. and R. Hilborn. 2013. Measuring uncertainty in fisheries stock assessment: the 2251 delta method, bootstrap, and MCMC. Fish and Fisheries 14: 325–342 Marasco, R.J., Goodman, 2252 D., Grimes, C.B., Lawson, P.W., Punt, A.E. and T.J. Quinn II. 2007. Ecosystem-based fisheries 2253 management: some practical suggestions. Canadian Journal of Fisheries and Aquatic Sciences 64: 2254 928–939. 2255

Methot, R.D. and C.R. Wetzel. 2013. Stock Synthesis: a biological and statistical framework for fish stock 2256 assessment and fishery management. Fisheries Research 142: 86–99. Moore, S.E., Waite, J.M., 2257 Friday, N.A., Honkalehto, T. 2002. Cetacean distribution and relative abundance on the central-2258 eastern and southeastern Bering Sea shelf with reference to oceanographic domains. Progress in 2259

167


Oceanopgraphy, 55: 249-261. 2260 Mueter, F.J. and B.A. Megrey. 2006. Using multispecies surplus production models to estimate ecosystem-2261

level maximum sustainable yields. Fisheries Research 81: 189–201 2262 Mueter, F., and Litzow, M.A. 2008. Sea ice retreat alters the biogeography of the Bering Sea continental 2263

shelf. EcologicaL Applications, 18: 309-320. 2264 North Pacific Fishery Management Council (NPFMC). 2011. Stock Assessment and Fishery Evaluation 2265

Report for the Groundfish Resources of the Bering Sea / Aleutian Islands Regions. North Pacific 2266 Fishery Management Council, 605 West 4th Ave, Anchorage, AK 99501 North Pacific Fishery 2267 Management Council (NPFMC). 2012. Stock Assessment and Fishery Evaluation Report for the 2268 Groundfish Resources of the Bering Sea / Aleutian Islands Regions. North Pacific Fishery 2269 Management Council, 605 West 4th Ave, Anchorage, AK 99501 Parslow, J., Cressie, N., 2270 Campbell, E.P., Jones, E. and L. Murray. 2013. Bayesian learning and predictability in a stochastic 2271 nonlinear dynamical model. Ecological Applications 23: 679–698. 2272

Patterson, K.R. 1999. Evaluating uncertainty in harvest control law catches using Bayesian Markov chain 2273 Monte Carlo virtual population analysis with adaptive rejection sampling and including structural 2274 uncertainty. Canadian Journal of Fisheries and Aquatic Sciences 56: 208–221. 2275

Plagányi, É.E. 2007. Models for an ecosystem approach to fisheries. FAO Fish. Tech. Pap. 477. Rome, FAO. 108 pp. 2276 Prager, M.H. and K.W. Shertzer. 2010. Deriving Acceptable Biological Catch from the Overfishing 2277 Limit: Implications for Assessment Models. North American Journal of Fisheries Management 30: 2278 289–294. 2279

Prager, M.H., Porch, C.E., Shertzer, K.W. and J.F. Caddy. 2003. Targets and limits for management of 2280 fisheries: a simple probability-based approach. North American Journal of Fisheries Management 2281 23: 349–361 Punt, A.E. and G. Donovan. 2007. Developing management procedures that are robust 2282 to uncertainty: Lessons from the International Whaling Commission. ICES Journal of Marine 2283 Science 64: 603–612 Punt, A.E., Siddeek, M.S.M., Garber-Yonts, B., Dalton, M., Rugolo, L., 2284 Stram, D., Turnock, B.J. and J. Zheng. 2012. Evaluating the impact of buffers to account for 2285 scientific uncertainty when setting TACs: Application to red king crab in Bristol Bay, Alaska. ICES 2286 Journal of Marine Science 69: 624–634 Raftery, A.E., Gneiting, T., Balabdaoui, F. and M. 2287 Polakowski. 2005. Using Bayesian model averaging to calibrate forecast ensembles. Monthly 2288 Weather Review 133: 1155–1174 Ralston, S., Punt, A.E., Hamel, O.S., DeVore, J. And R.J. 2289 Conser.2011. An approach to quantifying scientific uncertainty in stock assessment. Fishery 2290 Bulletin 109: 217–231 Reifen, C. and R. Toumi. 2009: Climate projections: Past performance no 2291 guarantee of future skill? Geophysical Research Letters 36: L13704 Shertzer et al 2008 2292

Shin, Y., and P. Cury. 2001. Exploring fish communitydynamics through size-dependent trophic 2293 interactions using a spatialized individual-based model. Aquatic Living Resources 14:65–80. 2294

Shin, Y.-J., Cury, P., 2004. Using an individual-based model of fish assemblages to study the response of 2295 size spectra to changes in fishing. Canadian Journal of Fisheries and Aquatic Sciences 61, 414–2296 431. 2297

Spiegelhalter, D.J., Best, N.G., Carlin, B.P. and A, van der Linde. 2002. Bayesian measures of model 2298 complexity and fit. Journal of Royal Society B 64: 583–639. 2299

Spies, I., Wilderbuer, T.K., Nichol, D.G. and K. Aydin. 2012. Arrowtooth flounder. Pp 895-965 in Stock 2300 Assessment and Fishery Evaluation Report for the Groundfish Resources for the Bering 2301 Sea/Aleutian Islands Regions. North Pacific Fishery Management Council, 605 West 4th Ave, Suite 2302 306, Anchorage AK 99501 Stabeno, P.J., Kachel, N.B., Moore, S.E., Napp, J.M., Sigler, M., 2303 Yamaguchi, A., Zerbini, A.N. 2012. Comparison of warm and cold years on the southeastern 2304 Bering Sea shelf and some implications for the ecosystem. Deep-Sea Research II. 65-70: 31-45 2305

Stabeno, P.J., Farley, E.V., Kachel, N.B., Moore, S., Mordy, C.W., Napp, J.M., Overlans, J.E., Pinchuk, 2306 A., Sigler, M.F. 2012b. A comparison of the physics of the northern and southern shelves of the 2307 eastern Bering Sea and some implications for the ecosystem. Deep Sea Research II, 65-70:14-30 2308

Stephensen, S.W. and Irons, D.B. 2003. A comparison of colonial breeding seabirds in the eastern Bering 2309 Sea and Gulf of Alaska. Marine Ornithology 31: 167-173. 2310

168


Schwartz, G. 1978. Estimating the dimension of a model. Annals of Statistics 6: 461–464. 2311 Taylor, I.G. and A. Stephens. 2013. Stock assessment of shortspine thornyhead in 2013. Pacific Fishery 2312

Management Council, 7700 NE Ambassador Place, Portland, OR 97220. 2313 Thompson, G.G. and R.R. Lauth. 2012. Assessment of the Pacific cod stock in the Eastern Bering Sea and 2314

Aleutian Islands area. pp 245-254 in Stock Assessment and Fishery Evaluation Report for the 2315 Groundfish Resources for the Bering Sea/Aleutian Islands Regions. North Pacific Fishery 2316 Management Council, 605 West 4th Ave, Suite 306, Anchorage AK 99501. Travers-Trolet, M., Y-2317 J Shin & JG Field (2014). An end-to-end coupled model ROMS-N2P2Z2D2-OSMOSE of the 2318 southern Benguela foodweb: parameterisation, calibration and pattern-oriented validation, African 2319 Journal of Marine Science, 36:1, 11-29. 2320

Wang, M., Overland, J.E., Bond, N.A., 2010. Climate projections for selected large marine ecosystems. J. 2321 Mar. Syst. 79(3-4), doi: 10.1016/j.jmarsys.2008.11.028, 258-266. 2322

Wang, M., Overland, J.E., and Stabeno, P. 2012. Future climate of the Bering and Chukchi Seas projected 2323 by global climate models. Deep-Sea Research II, 65-70: 40-57. 2324

Wilberg, M.J. and J.R. Bence. 2008. Performance of deviance information criterion model selection in 2325 statistical catch-at-age analysis. Fisheries Research 93: 212–221 Witherell, D., Pautzke, C., and 2326 Fluharty, D. 2000. An ecosystem-based approach for Alaska groundfish fisheries. ICES Journal of 2327 Marine Science, 57:771-777. 2328

Zador, S. (Ed.). 2013. Ecosystem considerations 2013. North Pacific Fishery Management Council, 605 W 2329 4th Ave Suite 306, Anchorage, AK 99501. November 18, 2013. 235pp. 2330 http://www.afsc.noaa.gov/REFM/Docs/2013/ecosystem.pdf. 2331

2332 2333

169



Appendix A 2334 2335 List of fields for which weekly output from the FEAST hindcast 1970-2009, run V146, is available at the 2336

Bering Sea Project Data Archive. 2337 2338 FIELD NAME LONG NAME 2339 temp_latlon time-averaged potential temperature 2340 salt_latlon time-averaged salinity 2341 no3_latlon time-averaged nitrate concentration 2342 nh4_latlon time-averaged ammonia concentration 2343 iron_latlon time-averaged iron concentration 2344 phs_latlon time-averaged small phytoplankton concentration 2345 phl_latlon time-averaged large phytoplankton concentration 2346 mzl_latlon time-averaged large microzooplankton concentration 2347 cop_latlon time-averaged small copepod concentration 2348 ncas_latlon time-averaged neocalanus shelf. concentration 2349 ncao_latlon time-averaged neocalanus oceanic concentration 2350 eup_latlon time-averaged euphausiid concentration 2351 lcz_latlon large crustacean zooplankton 2352 det_latlon time-averaged detritus concentration 2353 detf_latlon time-averaged fast sinking detritus concentration 2354 jel_latlon time-averaged jellyfish concentration 2355 prod_phs_latlon time-averaged primary production small phytoplankton 2356 prod_phl_latlon time-averaged primary production large phytoplankton 2357 prod_mzl_latlon time-averaged secondary production large microzooplankton 2358 prod_cop_latlon time-averaged secondary production copepods 2359 prod_nca_latlon time-averaged secondary production neocalanus 2360 prod_eup_latlon time-averaged secondary production euphausiids 2361 prod_jel_latlon time-averaged secondary production jellyfish 2362 w_latlon time-averaged vertical momentum component 2363 omega_latlon time-averaged S-coordinate vertical momentum component 2364 akt_latlon time-averaged temperature vertical diffusion coefficient 2365 akv_latlon time-averaged vertical viscosity coefficient 2366 zeta_latlon time-averaged free-surface 2367 ben_latlon benthos concentration 2368 detben_latlon benthic detritus concentration 2369 icephl_latlon time-averaged Ice algae concentration 2370 iceno3_latlon time-averaged Ice nitrate concentration 2371 icenh4_latlon time-averaged Ice Ammonium concentration 2372 hsbl_latlon time-averaged depth of oceanic surface boundary layer 2373 shflux_latlon time-averaged surface net heat flux 2374 ssflux_latlon time-averaged surface net salt flux, (E-P)*SALT 2375 latent_latlon time-averaged net latent heat flux 2376 sensible_latlon time-averaged net sensible heat flux 2377 lwrad_latlon time-averaged net longwave radiation flux 2378 swrad_latlon time-averaged solar shortwave radiation flux 2379 aice_latlon time-averaged fraction of cell covered by ice 2380 hice_latlon time-averaged average ice thickness in cell 2381 altitude altitude (depth) 2382 u_latlon time-averaged u-momentum component 2383

170


v_latlon time-averaged v-momentum component 2384 ubar_latlon time-averaged vertically integrated u-momentum component 2385 vbar_latlon time-averaged vertically integrated v-momentum component 2386 sustr_latlon time-averaged surface u-momentum stress 2387 svstr_latlon time-averaged surface v-momentum stress 2388 bustr_latlon time-averaged bottom u-momentum stress 2389 bvstr_latlon time-averaged bottom v-momentum stress 2390 uice_latlon time-averaged u-component of ice velocity 2391 vice_latlon time-averaged v-component of ice velocity 2392 pollock_average_length_age_00_latlon pollock_average_length_age_00_latlon 2393 pollock_average_length_age_01_latlon pollock_average_length_age_01_latlon 2394 pollock_average_length_age_02_latlon pollock_average_length_age_02_latlon 2395 pollock_average_length_age_03_latlon pollock_average_length_age_03_latlon 2396 pollock_average_length_age_04_latlon pollock_average_length_age_04_latlon 2397 pollock_average_length_age_05_latlon pollock_average_length_age_05_latlon 2398 pollock_average_length_age_06_latlon pollock_average_length_age_06_latlon 2399 pollock_average_length_age_07_latlon pollock_average_length_age_07_latlon 2400 pollock_average_length_age_08_latlon pollock_average_length_age_08_latlon 2401 pollock_average_length_age_09_latlon pollock_average_length_age_09_latlon 2402 pollock_average_length_age_10_latlon pollock_average_length_age_10_latlon 2403 pollock_biomass_age_00_latlon pollock_biomass_age_00_latlon 2404 pollock_biomass_age_01_latlon pollock_biomass_age_01_latlon 2405 pollock_biomass_age_02_latlon pollock_biomass_age_02_latlon 2406 pollock_biomass_age_03_latlon pollock_biomass_age_03_latlon 2407 pollock_biomass_age_04_latlon pollock_biomass_age_04_latlon 2408 pollock_biomass_age_05_latlon pollock_biomass_age_05_latlon 2409 pollock_biomass_age_06_latlon pollock_biomass_age_06_latlon 2410 pollock_biomass_age_07_latlon pollock_biomass_age_07_latlon 2411 pollock_biomass_age_08_latlon pollock_biomass_age_08_latlon 2412 pollock_biomass_age_09_latlon pollock_biomass_age_09_latlon 2413 pollock_biomass_age_10_latlon pollock_biomass_age_10_latlon 2414 pollock_condition_factor_age_00_latlon pollock_condition_factor_age_00_latlon 2415 pollock_condition_factor_age_01_latlon pollock_condition_factor_age_01_latlon 2416 pollock_condition_factor_age_02_latlon pollock_condition_factor_age_02_latlon 2417 pollock_condition_factor_age_03_latlon pollock_condition_factor_age_03_latlon 2418 pollock_condition_factor_age_04_latlon pollock_condition_factor_age_04_latlon 2419 pollock_condition_factor_age_05_latlon pollock_condition_factor_age_05_latlon 2420 pollock_condition_factor_age_06_latlon pollock_condition_factor_age_06_latlon 2421 pollock_condition_factor_age_07_latlon pollock_condition_factor_age_07_latlon 2422 pollock_condition_factor_age_08_latlon pollock_condition_factor_age_08_latlon 2423 pollock_condition_factor_age_09_latlon pollock_condition_factor_age_09_latlon 2424 pollock_condition_factor_age_10_latlon pollock_condition_factor_age_10_latlon 2425 pollock_energy_density_age_00_latlon pollock_energy_density_age_00_latlon 2426 pollock_energy_density_age_01_latlon pollock_energy_density_age_01_latlon 2427 pollock_energy_density_age_02_latlon pollock_energy_density_age_02_latlon 2428 pollock_energy_density_age_03_latlon pollock_energy_density_age_03_latlon 2429 pollock_energy_density_age_04_latlon pollock_energy_density_age_04_latlon 2430 pollock_energy_density_age_05_latlon pollock_energy_density_age_05_latlon 2431 pollock_energy_density_age_06_latlon pollock_energy_density_age_06_latlon 2432 pollock_energy_density_age_07_latlon pollock_energy_density_age_07_latlon 2433 pollock_energy_density_age_08_latlon pollock_energy_density_age_08_latlon 2434

171


pollock_energy_density_age_09_latlon pollock_energy_density_age_09_latlon 2435 pollock_energy_density_age_10_latlon pollock_energy_density_age_10_latlon 2436 pollock_numbers_age_00_latlon pollock_numbers_age_00_latlon 2437 pollock_numbers_age_01_latlon pollock_numbers_age_01_latlon 2438 pollock_numbers_age_02_latlon pollock_numbers_age_02_latlon 2439 pollock_numbers_age_03_latlon pollock_numbers_age_03_latlon 2440 pollock_numbers_age_04_latlon pollock_numbers_age_04_latlon 2441 pollock_numbers_age_05_latlon pollock_numbers_age_05_latlon 2442 pollock_numbers_age_06_latlon pollock_numbers_age_06_latlon 2443 pollock_numbers_age_07_latlon pollock_numbers_age_07_latlon 2444 pollock_numbers_age_08_latlon pollock_numbers_age_08_latlon 2445 pollock_numbers_age_09_latlon pollock_numbers_age_09_latlon 2446 pollock_numbers_age_10_latlon pollock_numbers_age_10_latlon 2447 cod_average_length_age_00_latlon cod_average_length_age_00_latlon 2448 cod_average_length_age_01_latlon cod_average_length_age_01_latlon 2449 cod_average_length_age_02_latlon cod_average_length_age_02_latlon 2450 cod_average_length_age_03_latlon cod_average_length_age_03_latlon 2451 cod_average_length_age_04_latlon cod_average_length_age_04_latlon 2452 cod_average_length_age_05_latlon cod_average_length_age_05_latlon 2453 cod_average_length_age_06_latlon cod_average_length_age_06_latlon 2454 cod_average_length_age_07_latlon cod_average_length_age_07_latlon 2455 cod_average_length_age_08_latlon cod_average_length_age_08_latlon 2456 cod_average_length_age_09_latlon cod_average_length_age_09_latlon 2457 cod_average_length_age_10_latlon cod_average_length_age_10_latlon 2458 cod_biomass_age_00_latlon cod_biomass_age_00_latlon 2459 cod_biomass_age_01_latlon cod_biomass_age_01_latlon 2460 cod_biomass_age_02_latlon cod_biomass_age_02_latlon 2461 cod_biomass_age_03_latlon cod_biomass_age_03_latlon 2462 cod_biomass_age_04_latlon cod_biomass_age_04_latlon 2463 cod_biomass_age_05_latlon cod_biomass_age_05_latlon 2464 cod_biomass_age_06_latlon cod_biomass_age_06_latlon 2465 cod_biomass_age_07_latlon cod_biomass_age_07_latlon 2466 cod_biomass_age_08_latlon cod_biomass_age_08_latlon 2467 cod_biomass_age_09_latlon cod_biomass_age_09_latlon 2468 cod_biomass_age_10_latlon cod_biomass_age_10_latlon 2469 cod_condition_factor_age_00_latlon cod_condition_factor_age_00_latlon 2470 cod_condition_factor_age_01_latlon cod_condition_factor_age_01_latlon 2471 cod_condition_factor_age_02_latlon cod_condition_factor_age_02_latlon 2472 cod_condition_factor_age_03_latlon cod_condition_factor_age_03_latlon 2473 cod_condition_factor_age_04_latlon cod_condition_factor_age_04_latlon 2474 cod_condition_factor_age_05_latlon cod_condition_factor_age_05_latlon 2475 cod_condition_factor_age_06_latlon cod_condition_factor_age_06_latlon 2476 cod_condition_factor_age_07_latlon cod_condition_factor_age_07_latlon 2477 cod_condition_factor_age_08_latlon cod_condition_factor_age_08_latlon 2478 cod_condition_factor_age_09_latlon cod_condition_factor_age_09_latlon 2479 cod_condition_factor_age_10_latlon cod_condition_factor_age_10_latlon 2480 cod_energy_density_age_00_latlon cod_energy_density_age_00_latlon 2481 cod_energy_density_age_01_latlon cod_energy_density_age_01_latlon 2482 cod_energy_density_age_02_latlon cod_energy_density_age_02_latlon 2483 cod_energy_density_age_03_latlon cod_energy_density_age_03_latlon 2484 cod_energy_density_age_04_latlon cod_energy_density_age_04_latlon 2485

172


cod_energy_density_age_05_latlon cod_energy_density_age_05_latlon 2486 cod_energy_density_age_06_latlon cod_energy_density_age_06_latlon 2487 cod_energy_density_age_07_latlon cod_energy_density_age_07_latlon 2488 cod_energy_density_age_08_latlon cod_energy_density_age_08_latlon 2489 cod_energy_density_age_09_latlon cod_energy_density_age_09_latlon 2490 cod_energy_density_age_10_latlon cod_energy_density_age_10_latlon 2491 cod_numbers_age_00_latlon cod_numbers_age_00_latlon 2492 cod_numbers_age_01_latlon cod_numbers_age_01_latlon 2493 cod_numbers_age_02_latlon cod_numbers_age_02_latlon 2494 cod_numbers_age_03_latlon cod_numbers_age_03_latlon 2495 cod_numbers_age_04_latlon cod_numbers_age_04latlon 2496 cod_numbers_age_05_latlon cod_numbers_age_05_latlon 2497 cod_numbers_age_06_latlon cod_numbers_age_06_latlon 2498 cod_numbers_age_07_latlon cod_numbers_age_07_latlon 2499 cod_numbers_age_08_latlon cod_numbers_age_08_latlon 2500 cod_numbers_age_09_latlon cod_numbers_age_09_latlon 2501 cod_numbers_age_10_latlon cod_numbers_age_10_latlon 2502 atf_average_length_age_00_latlon atf_average_length_age_00_latlon 2503 atf_average_length_age_01_latlon atf_average_length_age_01_latlon 2504 atf_average_length_age_02_latlon atf_average_length_age_02_latlon 2505 atf_average_length_age_03_latlon atf_average_length_age_03_latlon 2506 atf_average_length_age_04_latlon atf_average_length_age_04_latlon 2507 atf_average_length_age_05_latlon atf_average_length_age_05_latlon 2508 atf_average_length_age_06_latlon atf_average_length_age_06_latlon 2509 atf_average_length_age_07_latlon atf_average_length_age_07_latlon 2510 atf_average_length_age_08_latlon atf_average_length_age_08_latlon 2511 atf_average_length_age_09_latlon atf_average_length_age_09_latlon 2512 atf_average_length_age_10_latlon atf_average_length_age_10_latlon 2513 atf_biomass_age_00_latlon atf_biomass_age_00_latlon 2514 atf_biomass_age_01_latlon atf_biomass_age_01_latlon 2515 atf_biomass_age_02_latlon atf_biomass_age_02_latlon 2516 atf_biomass_age_03_latlon atf_biomass_age_03_latlon 2517 atf_biomass_age_04_latlon atf_biomass_age_04_latlon 2518 atf_biomass_age_05_latlon atf_biomass_age_05_latlon 2519 atf_biomass_age_06_latlon atf_biomass_age_06_latlon 2520 atf_biomass_age_07_latlon atf_biomass_age_07_latlon 2521 atf_biomass_age_08_latlon atf_biomass_age_08_latlon 2522 atf_biomass_age_09_latlon atf_biomass_age_09_latlon 2523 atf_biomass_age_10_latlon atf_biomass_age_10_latlon 2524 atf_condition_factor_age_00_latlon atf_condition_factor_age_00_latlon 2525 atf_condition_factor_age_01_latlon atf_condition_factor_age_01_latlon 2526 atf_condition_factor_age_02_latlon atf_condition_factor_age_02_latlon 2527 atf_condition_factor_age_03_latlon atf_condition_factor_age_03_latlon 2528 atf_condition_factor_age_04_latlon atf_condition_factor_age_04_latlon 2529 atf_condition_factor_age_05_latlon atf_condition_factor_age_05_latlon 2530 atf_condition_factor_age_06_latlon atf_condition_factor_age_06_latlon 2531 atf_condition_factor_age_07_latlon atf_condition_factor_age_07_latlon 2532 atf_condition_factor_age_08_latlon atf_condition_factor_age_08_latlon 2533 atf_condition_factor_age_09_latlon atf_condition_factor_age_09_latlon 2534 atf_condition_factor_age_10_latlon atf_condition_factor_age_10_latlon 2535 atf_energy_density_age_00_latlon atf_energy_density_age_00_latlon 2536

173


atf_energy_density_age_01_latlon atf_energy_density_age_01_latlon 2537 atf_energy_density_age_02_latlon atf_energy_density_age_02_latlon 2538 atf_energy_density_age_03_latlon atf_energy_density_age_03_latlon 2539 atf_energy_density_age_04_latlon atf_energy_density_age_04_latlon 2540 atf_energy_density_age_05_latlon atf_energy_density_age_05_latlon 2541 atf_energy_density_age_06_latlon atf_energy_density_age_06_latlon 2542 atf_energy_density_age_07_latlon atf_energy_density_age_07_latlon 2543 atf_energy_density_age_08_latlon atf_energy_density_age_08_latlon 2544 atf_energy_density_age_09_latlon atf_energy_density_age_09_latlon 2545 atf_energy_density_age_10_latlon atf_energy_density_age_10_latlon 2546 atf_numbers_age_00_latlon atf_numbers_age_00_latlon 2547 atf_numbers_age_01_latlon atf_numbers_age_01_latlon 2548 atf_numbers_age_02_latlon atf_numbers_age_02_latlon 2549 atf_numbers_age_03_latlon atf_numbers_age_03_latlon 2550 atf_numbers_age_04_latlon atf_numbers_age_04latlon 2551 atf_numbers_age_05_latlon atf_numbers_age_05_latlon 2552 atf_numbers_age_06_latlon atf_numbers_age_06_latlon 2553 atf_numbers_age_07_latlon atf_numbers_age_07_latlon 2554 atf_numbers_age_08_latlon atf_numbers_age_08_latlon 2555 atf_numbers_age_09_latlon atf_numbers_age_09_latlon 2556 atf_numbers_age_10_latlon atf_numbers_age_10_latlon 2557 adult_salmon_average_age_00_latlon adult_salmon_average_length_age_00_latlon 2558 adult_salmon_biomass_age_00_latlon adult_salmon_biomass_age_00_latlon 2559 adult_salmon_condition_factor_age_00_latlon adult_salmon_condition_factor_age_00_latlon 2560 adult_salmon_energy_density_age_00_latlon adult_salmon_energy_density_age_00_latlon 2561 adult_salmon_numbers_age_00_latlon adult_salmon_numbers_age_00_latlon 2562 capelin_average_length_age_00_latlon capelin_average_length_age_00_latlon 2563 capelin_biomass_age_00_latlon capelin_biomass_age_00_latlon 2564 capelin_condition_factor_age_00_latlon capelin_condition_factor_age_00_latlon 2565 capelin_energy_density_age_00_latlon capelin_energy_density_age_00_latlon 2566 capelin_numbers_age_00_latlon capelin_numbers_age_00_latlon 2567 eulachon_average_length_age_00_latlon eulachon_average_length_age_00_latlon 2568 eulachon_biomass_age_00_latlon eulachon_biomass_age_00_latlon 2569 eulachon_condition_factor_age_00_latlon eulachon_condition_factor_age_00_latlon 2570 eulachon_energy_density_age_00_latlon eulachon_energy_density_age_00_latlon 2571 eulachon_numbers_age_00_latlon eulachon_numbers_age_00_latlon 2572 juv_salmon_average_length_age_00_latlon juvenile_salmon_average_length_age_00_latlon 2573 juv_salmon_biomass_age_00_latlon juvenile_salmon_biomass_age_00_latlon 2574 juv_salmon_condition_factor_age_00_latlon juvenile_salmon_condition_factor_age_00_latlon 2575 juv_salmon_energy_density_age_00_latlon juvenile_salmon_energy_density_age_00_latlon 2576 juv_salmon_numbers_age_00_latlon juvenile_salmon_numbers_age_00_latlon 2577 myctophids_average_length_age_00_latlon myctophids_average_length_age_00_latlon 2578 myctophids_biomass_age_00_latlon myctophids_biomass_age_00_latlon 2579 myctophids_condition_factor_age_00_latlon myctophids_condition_factor_age_00_latlon 2580 myctophids_energy_density_age_00_latlon myctophids_energy_density_age_00_latlon 2581 myctophids_numbers_age_00_latlon myctophids_numbers_age_00_latlon 2582 pacific_herring_average_length_age_00_latlon pacific_herring_average_length_age_00_latlon 2583 pacific_herring_biomass_age_00_latlon pacific_herring_biomass_age_00_latlon 2584 pacific_herring_condition factor_age_00_latlon pacific_herring_condition factor_age_00_latlon 2585 pacific_herring_energy_density_age_00_latlon pacific_herring_energy_density_age_00_latlon 2586 pacific_herring_numbers_age_00_latlon pacific_herring_numbers_age_00_latlon 2587

174


sandlance_average_length_age_00_latlon sandlance_average_length_age_00_latlon 2588 sandlance_biomass_age_00_latlon sandlance_biomass_age_00_latlon 2589 sandlance_condition_factor_age_00_latlon sandlance_condition_factor_age_00_latlon 2590 sandlance_energy_density_age_00_latlon sandlance_energy_density_age_00_latlon 2591 sandlance_numbers_age_00_latlon sandlance_numbers_age_00_latlon 2592 eggs_01_latlon eggs_pollock_latlon 2593 eggs_02_latlon eggs_cod_latlon 2594 eggs_03_latlon eggs_arrowtooth_latlon 2595 eggs_04_latlon eggs_herring_latlon 2596 eggs_05_latlon eggs_capelin_latlon 2597 eggs_06_latlon eggs_eulachon_latlon 2598 eggs_07_latlon eggs_sandlance_latlon 2599 eggs_08_latlon eggs_myctophids_latlon 2600 eggs_09_latlon eggs_adult_salmon_latlon 2601 eggs_10_latlon eggs_juvenile_salmon_latlon 2602 epifauna_biomass_age_00_latlon epifauna_biomass_age_00_latlon 2603 ice_prop_latlon ice_prop_latlon (same as aice) 2604 misc_zoop_biomass_age_00_latlon misc_zoop_biomass_age_00_latlon 2605 shrimp_biomass_age_00_latlon shrimp_biomass_age_00_latlon 2606 snow_crab_biomass_age_00_latlon snow_crab_biomass_age_00_latlon 2607 squid_biomass_age_00_latlon squid_biomass_age_00_latlon 2608 temp300m_latlon average_temperature_in_top_300m_latlon 2609 temp_bottom_latlon temp_bottom_latlon 2610 zoop300m_01_latlon zoopbiomass_top300m_small copepods_latlon 2611 zoop300m_02_latlon zoopbiomass_top300m_Neocalanus_shelf_latlon 2612 zoop300m_03_latlon zoopbiomass_top300m_Neocalanus_ocean_latlon 2613 zoop300m_04_latlon zoopbiomass_top300m_euphausiids_shelf_latlon 2614 zoop300m_05_latlon zoopbiomass_top300m_euphausiids_ocean_latlon 2615 zoop300m_06_latlon zoop_biomass_top300m_benthos_latlon 2616 zoopFmort_01_latlon fish_predation_mortality_small_copepods_latlon 2617 zoopFmort_02_latlon fish_predation_mortality_Neocalanus_shelf_latlon 2618 zoopFmort_03_latlon fish_predation_mortality_Neocalanus_ocean_latlon 2619 zoopFmort_04_latlon fish_predation_mortality_euphausiids_shelf_latlon 2620 zoopFmort_05_latlon fish_predation_mortality_euphausiid_ocean_latlon 2621 zoopFmort_06_latlon zoop_fish_predation_mortality_benthos_latlon 2622 zoopQmort_01_latlon zoop_quadratic mortality_small_copepods_latlon 2623 zoopQmort_02_latlon zoop_quadratic_mortality_Neocalanus_shelf_latlon 2624 zoopQmort_03_latlon zoop_quadraticmortality_Neocalanus_ocean_latlon 2625 zoopQmort_04_latlon zoop_quadraticmortality_euphausiids_shelf_latlon 2626 zoopQmort_05_latlon zoop_quadraticmortality_euphausiids_ocean_latlon 2627 zoopQmort_06_latlon zoop_quadraticmortality_benthos_latlon 2628 2629

175

north pacific research board bering sea integrated ... · marine population assessment and...

Documents