AmeriFlux:Progress and emerging challenges

Beverly Law, Oregon State University, AmeriFlux Science Team Chair

AmeriFlux Science Steering GroupDennis Baldocchi, University of California, BerkeleyDavid Bowling, University of UtahJing Chen, University of TorontoKenneth Davis1, The Pennsylvania State UniversityDavid Hollinger, USDA Forest ServiceXuhui Lee, Yale UniversityHank Margolis, Laval UniversityWilliam Munger, Harvard UniversitySteven Running, University of MontanaHans Peter Schmid, Forschungszentrum KarlsruhePeter Thornton, National Center for Atmospheric ResearchShashi Verma, University of Nebraska1presenting

outline

• Overview of AmeriFlux• Status report• Recent research progress• Emerging challenge: improving climate

system modeling• What is needed to meet the challenge?

0. Overview of AmeriFlux

AmeriFlux Networkhttp://public.ornl.gov/ameriflux

• 93 active sites in 3 countries• 32 research teams• Companion Canadian network• Most major ecoregions in N. America covered

• Calibration lab runs comparisons across sites• Open-access, central data base enables global use of observations• Operational since 1996

Science objectives• Quantify exchange of carbon, water and energy

between terrestrial ecosystems and the atmosphere across a range of vegetation types, disturbance histories, and climatic conditions.

• Understand processes governing the terrestrial carbon cycle and linkages with the water, energy and nitrogen cycles.

• Produce a high-quality data base and synthesize observations across the network.

Core measurements• Fluxes of CO2, water vapor, and sensible heat

flux via eddy covariance.• Radiative fluxes and micrometeorological

conditions.• Biophysical characterization of sites (e.g.

vegetation age and type, nutrient status, carbon pool sizes, soil type).

• Membership requirements:– Address common science questions in strategic plan– Collect certain measurements year-round – Share data in standard format via common database – Participate at annual meetings– Participate in calibration and intercalibration activities– Participate in network-wide syntheses of results

• No support or funding comes with membership– e.g. for data management and submission requirements,

synthesis contributions

• Responsibilities of membership are not binding

Current AmeriFlux Structure:A Cooperative

• Structure is very inclusive– participation is broad– growth has been rapid (20 sites 10 years ago, now ~100, 350

worldwide)

• Funding structure encourages experimentation and innovation – Each funding cycle, each site must propose something “new”

or risk a poor review.

• Structure is ideal for learning how to build a network.• Structure is not ideal for maintaining a coherent, long-

term network.

Characteristics of theCooperative Structure

1. Status of AmeriFlux

AmeriFlux Productivity

0

100

200

300

400

500

600

1999 2001 2003 2005 2007

SitesReportingSite Years

Data Holdings• 548 site-years of

half-hourly data from 100 sites

• Flux measurements span 1991-2007

Publications• Investigators have

produced roughly 100 peer-reviewed publications/year for the past 5 years.

• Multi-site syntheses are increasingly common.

http://public.ornl.gov/ameriflux/viewstatus_Ameriflux.cfm

Following 3 figures courtesy of Tom Boden

AmeriFlux in a global context

• 74 AmeriFlux sites in Fluxnet dataset account for 305 site-years of data• More than 50 proposals for global synthesis papers received from site investigators. Global data not yet open-access

(AmeriFlux portion is open).• About 30% of submitted AmeriFlux site-years were rejected due to incomplete or insufficient quality data - network is

somewhat heterogeneous.

Number of sites contributing data to the La Thuile Fluxnet global synthesis data base

Length of Data Record and Sponsors for Active AmeriFlux Sites

0

2

4

6

8

10

12

14

16

18

1 2 3 4 5 6 7 8 9 10 10+

11 new sites began in 2007

Mean operating age – 5.7 yrs

10 sites operating over a decade but some of these have uncertain future funding

Number of years of operation for a given site

Nu

mb

er

of s

ites

• Major sponsors for active sites – DOE (39), None (17), USDA (15), NOAA (7), NSF(7), Universities (5), and NASA (3).• Non-centralized nature of AmeriFlux leads to instability in the network - potentially high turnover rate among sites.

Summary of status• An AmeriFlux network of decadal-scale flux measurements is emerging within a global coop of networks of "standardized" flux and biological measurements.

• Coherent network observations are central elements of many new synthesis studies. Research and publication is moving from site studies to network studies.

•There is increasing use of network data products by the carbon cycle and climate modeling communities.

• There is a recent decrease in the number of funded sites and the potential for a high turnover rate among sites.

• Some sites are providing insufficient data for synthesis activities.

• Network goals are compromised by funding sites individually.

2. Recent research highlights

- Process studies quantify and understand processes that influence ecosystem-atmosphere interactions

- Diagnoses of regional carbon budgetsconstruct large scale flux estimates

- Applications of flux network data to improving climate system modeling

Quantification of climate-ecosystem interactions: Clouds, aerosols and ecosystems

• AmeriFlux data quantifies the impact of clouds and aerosols on carbon sequestration and evapotranspiration at the land surface

• Flux observations showed aerosol and cloud effects on light quality and photosynthesis

• Importance of diffuse light effects now being incorporated in regional/global land surface models

• Similar process-oriented studies, using single or multiple sites, are ongoing and common across AmeriFlux sites.

Aspen forest, BOREAS southern study area

PPFD (mol m-2 s-1)

0 500 1000 1500 2000

Ne

t E

co

sys

tem

Ex

ch

an

ge

(

mo

l m-2

s-1

)

-40

-30

-20

-10

0

10

20

30

Cloudy daysClear days

Niyogi et al. 2004

Law et al. 2002 sink

Net

CO

2 F

lux

(m

ol m

-2 s

-1)

Diagnoses of regional carbon budgets:Regional clusters of flux towers

(Law et al. 2004, Turner et al. 2007)

(Desai et al., 2007; in press)

Several studies have now combined flux tower observations, satellite remote sensing, environmental conditions and terrestrial carbon cycle models to estimate regional fluxes.

Diagnosis of regional carbon budgets: MODIS-AmeriFlux synthesis

r = 0.855 ± 0.175

Heinsch et al., 2006

Figures courtesy of Steve Running

Improvement of a climate system model using flux network data: NCAR

CCSM/CLM example

AmeriFlux, Canadian and Amazon sites8 sites and 52 site-years of data

CarboEurope sites7 sites and 45 site-years of data

Temperate, tropical, sub-alpine, boreal and mediterranean sites

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Community Land Model (CLM) versions

CLM3.0: original code (Oleson et al. 2004)

CLMgw: prognostic ground water scheme, more infiltration, sun-shade canopy etc.

CLMgw+rsoil: new bare soil evaporation resistance

CLM3.5: diagnostic nitrogen control on Vmax, photosynthesis and stomatal conductance

CLM3.0 vs. CLM3.5 yields large differences in regional hydrologic balances.

Courtesy Reto Stoeckli, Colorado State University

Assimilation of flux measurements into terrestrial carbon cycle models for model evaluation

Site-years analyzedWLEF: 1997-2004Harvard: 1992-2003Howland: 1996-2003UMBS: 1999-2003M. Monroe: 1999-2003

TRIFFID model (used in Cox et al., 2000)

Ricciuto et al., in prep

Environmental conditions at flux

tower site

Flux tower site vegetation, soil, and disturbance characteristics

Terrestrial carbon cycle model

Bayesian parameter estimation algorithm

Model prior parameter values

(and pdfs)Modeled fluxes at tower sites

1

2

Optimized terrestrial carbon

cycle model

3

Modeled fluxes at tower sites

Optimized model parameters (and pdfs)

Flux observations

(subset)

Biometric data. Soil moisture.

Leaf area.

Flux observations (independent

set)

4

Evaluation

Evaluation

Assimilation of flux measurements into terrestrial carbon cycle models

seasonal cycle

mean annual flux

Diurnal, synoptic and seasonal cycles, and mean annual fluxes can be reproduced well.

Interannual variability in mean annual fluxes is not simulated well.

Results broadly consistent with previous efforts (e.g. Braswell et al., 2005)

3. Emerging challenge: Evaluate and improve climate prediction/projection

using climate system models

Problem: Uncertainty in the interaction of the terrestrial carbon cycle and climate

• C4MIP: comparison of 10 coupled climate/carbon models.

• Large uncertainty (225 ppm range in cumulative atmospheric CO2) in terrestrial biosphere contribution to atmospheric CO2 through 2100.

Friedlingstein et al., 2006

IPCC WG1 AR4 Summary for Policy Makers

Page 14: “Models used to date do not include uncertainties in climate-carbon cycle feedback nor do they include the full effects of changes in ice sheet flow, because a basis in published literature is lacking.”

Page 17: “Climate carbon cycle coupling is expected to add carbon dioxide to the atmosphere as the climate system warms, but the magnitude of this feedback is uncertain. This increases the uncertainty in the trajectory of carbon dioxide emissions required to achieve a particular stabilisation level of atmospheric carbon dioxide concentration.”

3rd IPCC report

AmeriFlux network record: a fundamental climate system observation

AmeriFlux network record: a fundamental climate system observation

Causal chain: CO2 fluxesCO2 mixing ratioSurface temperature

Note: A single flux measurement does not capture a global value

Challenge: Improve predictive skill in coupled climate-carbon

cycle modeling

• AmeriFlux role: Provide a CO2 flux network data product that can be used as the instrumental temperature record has been used.

Future: Poorly maintained network

hindcast forecast

Future: Well maintained network

Current

present

Observational constraints

Potential coupled climate-carbon cycle model evaluation

Terrestrial uptake of carbon (GtC yr-1)

Time

0

10

5

-5

Possible carbon cycle forecasts

Objectives:• To provide feedback to the science community on the

performance of terrestrial biogeochemistry models coupled to CLM within CCSM3

• To provide a new observation-based diagnostics package for terrestrial carbon cycling in coupled carbon-climate models

• To define, conceptually, how biogeochemistry should be evaluated in climate models

The NCAR Community Climate Systems Model (CCSM) Carbon and Land Model

intercomparison Project (C-LAMP)

Forrest Hoffman, Peter Thornton, Yen-Huei Lee, Nan Rosenbloom, Jim Randerson, Inez Fung and Steve Running

C-LAMP Model SimulationsRun Description Time period

Forcing with observed climate (Dai et al.)

1.1 Spin-up ~4000 years

1.2 Control 1798-2005

1.3 Varying climate 1948-2005

1.4 Varying climate, CO2, and N

deposition

1798-2005

1.5 Varying climate, CO2, N deposition,

and land use change

1798-2005

Land-atmosphere coupled

2.1 Spin-up ~1000 years

2.2 Control 1800-2005

2.3 Varying climate (Hadley SSTs) 1800-2005

2.4 Varying climate, CO2, and N

deposition

1800-2005

2.5 Varying climate, CO2, N deposition,

and land use change

1800-2005

C-LAMPevaluation

metrics utilize:

flux network data,

ecological inventories,

remote sensing,

atmospheric CO2,

ecosystem experiments

Table 2. Biogeochemical Model Evaluation Datasets:

Metric Metric components Sub-Score

Total Score

Status

NPP Matching EMDI Net Primary Production (NPP) observations 5 20 Complete EMDI comparison, normalized by PPT 5 Complete Correlation with MODIS (r2) 5 Complete Latitudinal profile (r2) 5 Complete

LAI Matching MODIS observations - Phase (derived separately for major biome classes) 5 20 Complete - Mean (derived separately for major biome classes) 5 Complete - Maximum (derived separately for major biome classes) 5 Complete - Growing season length (derived separately for major biome

classes) 5 Complete

CO2 Seasonal

Cycle Matching the phase and amplitude at NOAA GMD Globalview observation stations

20 Complete

Carbon Stocks Aboveground vegetation within the Amazon Basin from Saatchi

et al. (2006) 5 10 Complete

Global belowground carbon (top 30 cm) from Batjes (2005). 5 In progress

Energy & CO2 Matching eddy covariance observations from FLUXNET fluxes - Net radiation (monthly means) 5 20 In progress

- Latent heat (monthly means) 5 In progress - Sensible heat (monthly means) 5 In progress - CO2 fluxes (monthly means) 5 In progress

Transient dynamics

Beta factor for CO2 fertilization – Norby mean 10 Complete

Rate constants for litter decomposition (LiDAT from Parton) and litter mass from Holland and Post.

Not ready

El Nino anomaly 1998 (NEE and fire components) – TRANSCOM, CarbonTracker, and Van der Werf et al. fires

In progress

dNPP/dT, dNPP/PPT Not ready Total: 100

Role of AmeriFlux within the North American Carbon Program

Tower flux

Ecological inventory

Atmospheric inventory

year

month

hour

day

Tim

e S

cale

Spatial Scale

(1m)2 = 10-4ha

(1000km)2 = 108ha

(100km)2 = 106ha

(10km)2 = 104ha

(1km)2 = 102ha

Rearth

Cha

mbe

r flu

x/pl

ot d

ata

Remote sensing and terrestrial models

Diagnose, attribute, predict.

4. What is needed to meet this challenge?

Observational constraints

Terrestrial uptake of carbon (GtC yr-1)

Time

0

10

5

-5

hindcast forecast

present

Possible carbon cycle forecasts

CurrentFuture: Well maintained network

Vision

1. Sustain and enhance a core set of long-term, high-quality flux measurement sites.

2. Continue mechanistic research to improve model structure and identify important climate-ecosystem interactions

3. Conduct network design studies– How many sites are needed? – What mix of shorter vs. longer term sites is optimal?– How long are the required time series?– What complementary data are needed at each site?

4. Sustain and enhance an easily-accessed, homogeneous, data base

Needs

Summary of status• An AmeriFlux network of decadal-scale flux measurements is emerging within a global coop of networks of "standardized" flux and biological measurements.

• Coherent network observations are central elements of many new synthesis studies. Research and publication is moving from site studies to network studies.

•There is increasing use of network data products by the carbon cycle and climate modeling communities.

• There is a recent decrease in the number of funded sites and the potential for a high turnover rate among sites.

• Some sites are providing insufficient data for synthesis activities.

• Network goals are compromised by funding sites individually.

RecommendationsThe AmeriFlux science steering group recommends

sustaining and enhancing a core network of long-term, high-quality flux measurement sites to address the increasing need for syntheses of multi-site, long-term data records.

A coherent network of sites has added value that exceeds a collection of individual, short-term studies. A mechanism should be developed to recognize this added value when

questions of funding arise.

A stable core network will provide a critical contribution to our ability to predict future climate by enabling the

development and evaluation of coupled carbon-climate models and earth systems analysis models.

Additional materials in case of questions.Delete when posting talk online.

Relationship between AmeriFlux and NEON: Complementary, intersecting efforts

AmeriFlux NEON

Scientific endeavor including infrastructure, human resources (PIs and their collective expertise), and well-established international ties

Infrastructure project

Wild and managed lands Wildlands focus

More than a decade of data and data management

Not yet funded

Committed to DOE/CCSP and NACP objectives

Broad ecological research agenda

Role of AmeriFlux within the North American Carbon Program

Flux tower upscaling

Atmospheric inventory

Ecological inventory

Strengths Potent source of mechanistic understanding. Excellent temporal precision and resolution.

Absolute net carbon balance at global scales.

Absolute above ground carbon balance at plot scales.

Weaknesses Difficult to cover large areas. Potential systematic errors.

Poor spatial resolution.

Difficult to determine mechanisms.

Difficult to cover large areas.

Poor temporal resolution.

Topics that can be addressed with integration of AmeriFlux data and models

• Where and when will forests be vulnerable to fires, and how do changes in forest processes affect climate?

• How would biofuel harvesting impact forest functioning and C sequestration?

• How will changes in water availability and population impact water availability to crops and forested watersheds that serve urban areas?

• What are potential interactions between future climate scenarios, and carbon, water, and nitrogen cycling?

Observed interannual variability: Only local processes? Probably not.

Gap-filled fluxes from the 6 midwestern flux tower sites.

Interannual variability of similar plant functional types appears to be coherent.

Similar processes, linked to climate, influencing sites as far as several hundred kilometers apart in a similar way?

LC = wetland; WC, MMSF, UMBS = mature hardwood; Syl = mixed old growth; WLEF = mixed

Contributions of AmeriFlux Research to DOE Climate Change Science Program Elements

• Climate forcing – carbon cycle, atmospheric water vapor

• Climate change prediction• Responses of ecosystems to climate change• Climate mitigation