Lyon Lanerolle (NOAA CSDL)and the

Estuarine Hypoxia COMT team

The Estuarine Hypoxia Component of the Coastal Ocean Modeling Testbed

(COMT)

A Community Coastal and Ocean Modeling Testbed (COMT)

to Improve Understanding and OperaConal Forecasts of Extreme Events and Chronic Environmental CondiCons AffecCng the U.S.

Five Teams:

1)Chesapeake Bay Estuarine Hypoxia ForecasCng2) IntegraCon of West Coast OperaConal Coastal & Ocean Models3) Puerto Rico/US Virgin Islands InundaCon & Wave ForecasCng4) Northern Gulf of Mexico Ecological ForecasCng5) Cyberinfrastructure

VIMS: Marjy Friedrichs (lead PI) Carl Friedrichs (VIMS- ‐PI) Ike Irby (funded student) Aaron Bever (consultant) Jian Shen (collaborator) Cathy Feng (collaborator)

NOAA- C‐ SDL: Lyon Lanerolle (NOAA- ‐PI) Frank Aikman (collaborator)

WHOI: Malcolm Scully (WHOI- ‐PI) UMCES: Raleigh Hood (UMCES- ‐PI)

Hao Wang (funded student)Wen Long (collaborator) Jeremy Testa (collaborator)

The Estuarine Hypoxia COMT Team

Estuarine Hypoxia ObjecCveTo assess the readiness and maturity of a suite of

exisCng coastal ecological community models for determining past, present and future hypoxia events within the Chesapeake Bay, in order to accelerate the transiKon of hypoxia model formulaKons and products from “academic research” to “operaKonal centers”

Chesapeake Bay EH centers include:• NOAA/NOS/CO- ‐OPS• Chesapeake Bay Ecological PredicCon System (CBEPS)• EPA Chesapeake Bay Program (CBP)

Estuarine Hypoxia Goal

Compare mulKple models within the Estuarine Hypoxia Testbed in order to improve exisCng:

1. CBOFS short- ‐term forecasKng: by incorporaCng new oxygen and physical model enhancements into the exisCng operaConal NOAA CO- ‐OPS Chesapeake Bay OperaConal Forecast System (CBOFS) for evaluaCon during their next update

2. CBEPS short- ‐term forecasKng: by developing a 24/7 predicCve capacity for nowcasCng/forecasCng of oxygen/hypoxic volume as part of CBEPS at the NOAA CBPO and UMCES

3. CBP scenario- ‐based forecasKng: Apply the official CBP nutrient reducCon strategies to COMT models to determine whether they perform similarly to the regulatory CBP model in terms of predicCng the effect of reduced nutrients on hypoxia

• StaCsCcally compare output from four Chesapeake Bay models:– three ROMS models, varying biological complexity

(ChesNENA, ChesROMS- ‐BGC, ROMS- ‐RCA)– biologically sophisCcated CBP regulatory model

(CH3D- ICM)‐

• How well do they reproduce the mean and seasonal variability of:– temperature, salinity, straCficaCon, dissolved oxygen (DO),

chlorophyll- a‐ , and nitrate

Model Comparisons via Chesapeake Testbed

IH4

MAT0078 MAT0016 IH2IH6

IH1

IH5

IH3T

F2.4

H-3H-2

H-1

WE4.1 CB7.2E C-3

C-2

C-1LE4.3N WE4.2N WE4.2 OP-3 CB6.4

CB7.3E OP-2

OP-1

ON-3

OC-3

CB6.2

H-1ACB6.3

CB7.2

CS-3

WT8.3

WT8.2

WT8.1

WT7.1 CB3.3W CB3.3E

WT6.1

WT4.1

PAT0176 WT5.1

WT2.1WT1.1

WE4.4 CB7.3

WE4.3

TF5.4 TF5.5 TF5.5ANTF5.5AS

TF5.5ATF5.6

TF5.3

TF5.2

TF4.4

TF4.2

TF3.3

TF3.2

TF3.1TF3.1

W

TF3.0

TF3.1D

TF2.1

TF2.3 TF2.2

TF1.7

TF1.6CB4.2W

TF1.4

TF1.5

TF1.3TF1.2

WXT0001

TF1.0

LE5.5

LE5.3LE5.4

LE5.2

LE5.1 LE4.3S WE4.2SRET5.2N LE4.3

LE4.2 LE4.2S

LE4.1

LE3.7

LE3.3 LE3.3A

LE3.4

LE3.6

LE3.2

LE3.1

LE2.3

LE2.2

LE1.4LE1.3

CB4.3W

RET1.1

LE1.1

ET9.1

ET8.1

ET7.1

CCM0069 TRQ0088

ET6.2

ET6.1

ET5.2

ET5.1

ET5.0ET4.2

ET4.1

ET3.1

ET2.3

ET2.2

CB2.1

ET1.1

ET2.1

EE3.5

EE3.4

EE3.2

EE3.1

EE3.0

EE2.2

EE2.1

EE1.1

CB7.4 CB8.1

CB7.1

CB6.1

CB5.5

CB5.4

CB5.3

CB5.2

CB5.1W CB5.1

CB4.4

CB3.2

CB3.1

CB2.2

CB1.1

CB1.0

DER0015

TF5.6A

TF5.2A

TF5.0J

TF5.0A

TF4.4ATF4.1A

TF4.0P

TF4.0M

TF3.2A

TF3.1C

TF3.1A TF3.1E

RET5.2

RET5.1

RET4.3

RET4.2

RET4.1 RET4.3S

RET3.2

RET3.1 RET3.1N

RET2.4RET2.3

RET2.2

RET2.1

LE5.5B

LE5.2S

LE3.6N

ET10.1

CB8.1E

CB7.4N

CB7.1N

CB5.4W

CB4.3E

CB4.2E

CB4.1WCB4.1E

CB3.3C

XFB1986

XDJ9007

XAK7810

WIW0141

TRQ0146

SEN0008

RET5.2S

RET3.1S

PXT0972

PXT0809

POT1830

CAC0031POT1596

POT1595

POT1472

POT1471

POT1184

POK0087

JON018

4 GWN0115PAT0285

NPA0165

MON0528

MON0269

MON0155

MON0020

HOK0005

GUN0476

GUN0258

FRG0018 NOM0007

CON0180

CON0005

ANT0203

CAC0148

BPC0035

ANT0366

MDR0028 WT3.1

LE1.2

EE3.3

TF3.1B

LE5.5A

LE5.2N

LE4.2N

LE3.6S

LE3.2S

LE3.2NCB7.1S

CB4.3C

CB4.2C

CB4.1C

XGG8251

XCI4078

RET5.1A

RET4.3N

PIS0033

MNK0146

GUN0125

CJB0005

BXK0031

ANT0044

MARYLAND

-

Legend

Fall Line0 5 10 15 20

Miles

DELAWARE

VIRGINIA

See St. Mary's

Inset

See Elizabeth

Inset

JamesRiver

York R.

Rappahannock

River

Potomac

River

Nan

ticok

eR.

Patuxent

River

Choptank

Rive

r

Cheste

r R.

Patapsco R.Susque

hann

a R.

See DC Inset

Maryland and Virginia Water

Quality Monitoring Stations

Compare simulaCons to

observaCons at 10 main stem staCons for ~16 cruises in

2004 and 2005

UnbiasedRMSD

Bias

1

1

Overestimates Observed Std Dev

Underestimates Observed Std Dev

Ove

rest

imat

es

Obs

erve

d M

ean

Und

eres

timat

es

Obs

erve

d M

ean

Model skill same as skill of mean of

observations

Model Skill Assessment via Target Diagrams

CH3D - ICM

ChesNENA

ChesROMS - BGC

ROMS - RCA

Bottom Temperature Bottom Salinity StratificationBottom Dissolved OxygenSurface Chlorophyll Surface NitrateBias

Unbiased RMSD

1

1

Overall skill of all four models (temporal + spatial variability):• are highest in terms of Temperature• are similar to each other in terms of T, S, stratification and DO• are different in terms of chlorophyll and nitrate

Regardless of complexity, models achieve similar skill scores in terms of seasonal variability of T, S, straCficaCon and oxygen

All models reproduce DO beder than variables that are typically thought to be primary influences on DO (straCficaCon, chlorophyll, nitrate)o This is because seasonal DO variability is sensiCve to T (solubility effect),

and the models reproduce T very wello Modeled DO simulaCons may be very sensiCve to any future increases in

Bay temperature

Oxygen forecasCng is possible with simple biological formulaCon

Model Comparisons via Chesapeake Testbed

Estuarine Hypoxia Goal

Compare mulKple models within the Estuarine Hypoxia Testbed in order to improve exisCng:

1. CBOFS short- ‐term forecasKng: by incorporaCng new oxygen and physical model enhancements into the exisCng operaConal NOAA CO- ‐OPS Chesapeake Bay OperaConal Forecast System (CBOFS) for evaluaCon during their next update

2. CBEPS short- ‐term forecasKng: by developing a 24/7 operaConal capacity for nowcasCng/forecasCng of oxygen/hypoxic volume as part of CBEPS at the NOAA CBPO and UMCES

3. CBP scenario- ‐based forecasKng: Apply the official CBP nutrient reducCon strategies to COMT models to determine whether they perform similarly to the regulatory CBP model in terms of predicCng the effect of reduced nutrients on hypoxia

CBOFS• CBOFS based on Regional Ocean Modeling

System (ROMS)• Grid generated in segments and pasted

seamlessly using Delf 3D- ‐RGFGRID generator

• Bathymetry: NOS soundings cut- off ‐ at 2m depth

• Init Conds: NOAA T, S climatology for lower Bay and CBP profiles for upper Bay

• Rivers: discharge = USGS; T, S = CBP• Outer Bdy Conds: T, S = NOAA climatology• Outer Bdy Tides: Cdal harmonic

consCtuents for WL and barotropic currents from ADCIRC database

• No sediment, precipitaCon, wei n g/drying or data assimilaCon

CBOFS Model Grid

CBOFS – Model Output Archive LocaCons

Lyon – what would you think about pasCng a screen grab or two from CBOFS to put in here, so folks can see what’s available currently online, and so you can explain that the goal is to add oxygen tothe list of other variables currently available?

Archive water elevaCons, 3D currents, T and S at all of the above locaCons

CBOFS – Model Outputs

Goal 1: CBOFS short- ‐term forecasKng Incorporate new oxygen and physical model enhancements into the exisCng operaConal NOAA CO- ‐OPS CBOFS for evaluaCon during their next update

Progress to Date:

•Staying in touch with NOS/CO- ‐OPS on their salinity improvements•COMT colleagues have recommended updated model opCons (advecCon scheme, TKE parameter, etc…)

Ongoing Work:

•Re- ‐run CBOFS (2.5y) with new model opCons and updated code•Compare mulCple physical simulaCons; assess model skill relaCve to other COMT models•Incorporate “best” constant biology DO model into CBOFS (Year 2?)•Compare mulCple DO simulaCons; assess model skill•Finalize CBOFS code and have it ready for NOS/CO- O‐ PS next update

CBOFS

Estuarine Hypoxia Goal

Compare mulKple models within the Estuarine Hypoxia Testbed in order to improve exisCng:

1. CBOFS short- ‐term forecasKng: by incorporaCng new oxygen and physical model enhancements into the exisCng operaConal NOAA CO- ‐OPS Chesapeake Bay OperaConal Forecast System (CBOFS) for evaluaCon during their next update

2. CBEPS short- ‐term forecasKng: by developing a 24/7 predicCve capacity for nowcasCng/forecasCng of oxygen/hypoxic volume as part of CBEPS at the NOAA CBPO and UMCES

3. CBP scenario- ‐based forecasKng: Apply the official CBP nutrient reducCon strategies to COMT models to determine whether they perform similarly to the regulatory CBP model in terms of predicCng the effect of reduced nutrients on hypoxia

Chesapeake Bay Ecological PredicCon System (CBEPS) and Model Framework

•Coupled hydrodynamic/biogeochemical model (ChesROMS) running “operaConally” at UMCES (formally supported by NOAA/NCBO)

• Nowcasts = real Cme USGS river discharge; Forecasts = assume river flows persist for 3 days

• Atmospheric forcing for 3- d‐ ay forecasts from the North American Meteorological Model

•Simple seasonal climatologies/flow for biogeochemical boundary condiCons

•Baywide nowcasts & 3 day forecasts of T and S are generated daily and posted

•Baywide ecological nowcasts & 3 day forecasts of Sea Nedl es and Vibrio are generated daily, based on T, S logisCcal regression models (Vibrio not posted)

CBEPS NowcasCng/ForecasCng Sea Nedles:hdp://chesapeakebay.noaa.gov/forecasCng- ‐sea- ‐nedles

• Maps generated daily and posted on website

• Nowcasts and 3- d‐ ay forecasts

• Sea Surface Temperature

20 Jan 2014

CBEPS NowcasCng/ForecasCng Sea Nedles:hdp://chesapeakebay.noaa.gov/forecasCng- ‐sea- ‐nedles

• Maps generated daily and posted on website

• Nowcasts and 3- d‐ ay forecasts

• Sea Surface Temperature

• Sea Surface Salinity

20 Jan 2014

CBEPS NowcasCng/ForecasCng Sea Nedles:hdp://chesapeakebay.noaa.gov/forecasCng- ‐sea- ‐nedles

• Maps generated daily and posted on website

• Nowcasts and 3- d‐ ay forecasts

• Sea Surface Temperature

• Sea Surface Salinity• Sea NeVles

20 Jan 2014

CBEPS NowcasCng/ForecasCng Sea Nedles:hdp://chesapeakebay.noaa.gov/forecasCng- ‐sea- ‐nedles

• Maps generated daily and posted on website

• Nowcasts and 3- d‐ ay forecasts

• Sea Surface Temperature

• Sea Surface Salinity• Sea Nedles• Vibrio (not posted)

20 Jan 2014

Goal 2: CBEPS short- ‐term forecasKng Develop a 24/7 predicCve capacity for nowcasCng/forecasCng of oxygen/hypoxic volume at NOAA CBPO & UMCES

Progress to Date:

•Currently running 3 day forecasts “operaConally” at UMCES for Chesapeake- ‐ biogeochemistry (based on “old” version of ChesROMS)•Developed “sea nedles” forecast – transiConed to a 24/7 demonstraCve productat NOAA through CBOFS (Success!) – sCll need to carry out skill assessment• Other organisms act similarly, e.g. Vibrio

Ongoing Work:

• Add simple DO formulaCon to list of variables forecasted• Update ChesROMS physics• UlCmately merge features of two ROMS- ‐BGC models: ChesNENA and

ChesROMS

Chesapeake Bay Ecological PredicCon System (CBEPS)

Estuarine Hypoxia Goal

Compare mulKple models within the Estuarine Hypoxia Testbed in order to improve exisCng:

1. CBOFS short- ‐term forecasKng: by incorporaCng new oxygen and physical model enhancements into the exisCng operaConal NOAA CO- ‐OPS Chesapeake Bay OperaConal Forecast System (CBOFS) for evaluaCon during their next update

2. CBEPS short- ‐term forecasKng: by developing a 24/7 operaConal capacity for nowcasCng/forecasCng of oxygen/hypoxic volume as part of CBEPS at the NOAA CBPO and UMCES

3. CBP scenario- ‐based forecasKng: Apply the official CBP nutrient reducCon strategies to COMT ROMS models to determine whether they perform similarly to the regulatory CBP model in terms of predicCng the effect of reduced nutrients on hypoxia (Year 2)

CBP Scenario- ‐based ForecasCngProgress to Date:

• Developed methodology to use an alternate hydrodynamic+biogeochemical model to reproduce Water Quality Standards for CBP

• EPA Chesapeake Bay Program folks are enthusiasCc about our proposedeffort to assess confidence/uncertainCes in their regulatory model

Future Work:

•Run alternate model(s) with CBP’s nutrient reducCon scenarios.•Apply CBP protocol to both sets of model scenarios•For each model, idenCfy when/where Bay will meet required “water

quality standards”•How do the model results diverge? Where/when are the greatestuncertainCes in the TMDLs computed from these model results?

NOAA CBOFS nowcasts/forecasts UMCES CBEPS nowcasts/forecasts CBP scenario- ‐based forecasts

The Estuarine Hypoxia COMT model skill comparisons are improving: