Update on the Estuarine Hypoxia Component
of the SURA Super-Regional Modeling Testbed
Aaron Bever1, Marjorie Friedrichs1, Carl Friedrichs1
1Virginia Institute of Marine Science
and the Estuarine Hypoxia Testbed Team
Outline of Presentation:
• Intro: Participants/Motivation/Methods
• Results: Salinity and Stratification
• Results: Dissolved Oxygen
• Conclusions/Future Work/Upcoming Workshop
Presented at CBP Modeling Subcommittee Meeting
Annapolis, MD, January 12, 2011
Estuarine Hypoxia – Some Key Results from 1st Six Months of Project
• Hydrodynamic base runs for 2004 were completed and compared for five 3-D
models; CH3D did best (slightly).
• All models underestimate stratification (both magnitude and variability).
• Absolute salinity results notably improve with greater horizontal resolution and
with improved riverine forcing.
• Ability to reproduce observed stratification is not obviously sensitive to changes in
riverine forcing or in horizontal or vertical resolution.
• Initial results from a first round Hypoxia run for 2004 completed and assessed; a
very simple, physics-dominated DO model works about as well as ICM over a single
annual cycle.
• The annual cycle in DO variations is reproduced better than the annual cycle in
stratification and is surprisingly insensitive to variability in stratification.
• Bottom DO and hypoxic volume is most sensitive to seasonal variability in wind.
Seasonal changes in temperature (including respiration) are secondary. Seasonal
changes in discharge are unimportant(!).
Update on the Estuarine Hypoxia Component
of the SURA Super-Regional Modeling Testbed
Aaron Bever1, Marjorie Friedrichs1, Carl Friedrichs1
1Virginia Institute of Marine Science
and the Estuarine Hypoxia Testbed Team
Outline of Presentation:
• Intro: Participants/Motivation/Methods
• Results: Salinity and Stratification
• Results: Dissolved Oxygen
• Conclusions/Future Work/Upcoming Workshop
Presented at CBP Modeling Subcommittee Meeting
Annapolis, MD, January 12, 2011
Funded by NOAA/IOOS through SURA (Southeastern Universities Research
Association). Initially one year of funding, began June 2010
Part of a larger NOAA/IOOS/SURA larger “Super-Regional Testbed to Improve
Models of Environmental Processes on the U.S. Atlantic and Gulf of Mexico Coasts”
Pilot projects in the larger “Super-Regional Testbed” will address three chronic
issues of high relevance within the super region:
• Coastal Inundation
• Estuarine Hypoxia
• Shelf Hypoxia
Estuarine Hypoxia Testbed
Funded by NOAA/IOOS through SURA (Southeastern Universities Research
Association). Initially one year of funding, start date June 2010
Motivation:
January 2010 NOAA Funding Opportunity Announcement: “FY2010 Integrated
Ocean Observing System Community Modeling Environment To Support a SuperRegional Test Bed”:
“The program priorities for this funding opportunity are to conduct a super-regional
testbed demonstration of the community modeling environment by transitioning
models, tools, toolkits and other capabilities to a Federal operational facility to
improve the understanding, prediction, and mitigation of the consequences of
extreme events and chronic conditions.”
Super-Regional Modeling Testbed Estuarine Hypoxia Team
• Carl Friedrichs (VIMS) – Team Leader
Federal partners
•
•
•
•
David Green (NOAA-NWS) – Transition to operations at NWS
Lyon Lanerole (NOAA-CSDL) – Transition to operations at CSDL; CBOFS2
Lewis Linker (EPA), Carl Cerco (USACE) – Transition to operations at EPA; CH3D, CE-ICM
Doug Wilson (NOAA-NCBO) – Integration w/observing systems at NCBO/IOOS
Non-federal partners
•
•
•
•
•
•
•
•
Marjorie Friedrichs (VIMS) – Metric development and model skill assessment
Ming Li (UMCES) – ROMS hydrodynamics in CB
Wen Long, Raleigh Hood (UMCES) – ChesROMS with NPZD water quality model
Scott Peckham (UC-Boulder) – Running multiple ROMS models on a single HPC cluster
Malcolm Scully (ODU) – ChesROMS with 1 term oxygen respiration model
Kevin Sellner (CRC) – Academic-agency liason; facilitator for model comparison
Jian Shen (VIMS) – SELFE, FVCOM, EFDC models in CB
John Wilkin, Julia Levin (Rutgers) – ROMS-Espresso + 7 other MAB hydrodynamic models
Motivation (cont.): A change in community thinking toward multiple models
2006:
“…workshop participants felt that … from a regulatory and practical point of view
multiple models are not appropriate for the CBP… If there are multiple models…
then court cases are invited. …Development of multiple models… is often
impractical and potentially counterproductive…”
Final report of the Chesapeake Bay Scientific and Technical Advisory Committee’s workshop: Modeling in the Chesapeake Bay
Program: 2010 and Beyond
2010:
“Use of a multiple model consensus, where models use different approaches and
assumptions, provides one way to better inform management decisions. This
approach is well known for its application to forecasting hurricane paths and
predicting the effects of carbon dioxide fluctuations on climate…”
Committee on Environment and Natural Resources. 2010. Scientific Assessment of Hypoxia in U.S. Coastal Waters. Interagency Working Group
on Harmful Algal Blooms, Hypoxia, and Human Health of the Joint Subcommittee on Ocean Science and Technology. Washington, DC.
Motivation (cont.): Existing paradigm is that stratification must be modeled
well to predict hypoxia well.
Mid-summer oxygen differences across the pycnocline versus salinity difference for
mid-Cheseapeake Bay stations, 1949-1980 (Flemer et al. 1983; Boicourt et al. 1992).
Methods: 5 Hydrodynamic models
3 flavors of ROMS
ChesROMS
(R. Hood, UMCES)
ROMS = Regional Ocean Modeling System
CBOFS2
(L. Lanerolle, NOAA-CSDL)
UMCESroms
(Ming Li, UMCES)
100x150 = 15000 grid cells
332x291 = 96612 grid cells
Low Res: 80x120 = 9600 grid cells
High Res: 160x240 = 38400 grid cells
Methods: 5 Hydrodynamic models (cont.)
EFDC
(Jian Shen, VIMS)
CH3D
(Chesapeake Bay Program)
18000 wet grid cells
54000 wet grid cells
Methods: Dissolved Oxygen Models
Three dissolved DO models considered so far:
(1) 1-term DO model in ChesROMS: single
respiration term (temperature dependent;
not dependent on nutrients; M. Scully)
(2) Regression of DO to T & S (W. Long)
(3) CE-QUAL-ICM Multi-component model (CBP)
Methods: S and DO from Up to 32 CBP station locations
Mean Summer 2009
Dissolved Oxygen [mg/L]
= CBP stations used in
this model-data comparison
http://www.eco-check.org/
Explanation of Target diagram
(modified from M. Friedrichs)
Update on the Estuarine Hypoxia Component
of the SURA Super-Regional Modeling Testbed
Aaron Bever1, Marjorie Friedrichs1, Carl Friedrichs1
1Virginia Institute of Marine Science
and the Estuarine Hypoxia Testbed Team
Outline of Presentation:
• Intro: Participants/Motivation/Methods
• Results: Salinity and Stratification
• Results: Dissolved Oxygen
• Conclusions/Future Work/Upcoming Workshop
Presented at CBP Modeling Subcommittee Meeting
Annapolis, MD, January 12, 2011
Bottom Salinity for All Stations:
bias [psu]
unbiased
RMSD
[psu]
CH3D reproduces mean salinity best
(by M. Friedrichs)
Target Diagrams for S at the Pycnocline for 23 EPA monitoring Stations During 2004
Bias
Unbiased
RMSD
(by A. Bever)
Salinity at 32 stations shaded by station salinity
Comparison of UMCES ROMS (low 2D resolution) and UMCES ROMS HR (high 2D resolution):
Increased horizontal resolution does improve predictions of absolute salinity.
(by A. Bever)
CBOFS2 Sensitivity Tests:
Improved river bc’s do improve salinity; vertical resolution does not.
(by A. Bever)
Stratification for All Stations:
bias [psu/m]
unbiased
RMSD
[psu/m]
Stratification is a challenge; CH3D and EFDC reproduce
seasonal/spatial variability slightly better.
(by M. Friedrichs)
Target Diagrams for |∂S/∂z| at Pycnocline for 23 EPA monitoring Stations During 2004
(by A. Bever)
Stratification at 32 stations shaded by station salinity
Comparison of UMCES ROMS (low 2D resolution) and UMCES ROMS HR (high 2D resolution):
Increased horizontal resolution does not improve predictions of stratification.
(by A. Bever)
CBOFS2 Stratification Sensitivity Tests:
Improved river bc’s & vertical resolution do not improve
stratification.
(by A. Bever)
Update on the Estuarine Hypoxia Component
of the SURA Super-Regional Modeling Testbed
Aaron Bever1, Marjorie Friedrichs1, Carl Friedrichs1
1Virginia Institute of Marine Science
and the Estuarine Hypoxia Testbed Team
Outline of Presentation:
• Intro: Participants/Motivation/Methods
• Results: Salinity and Stratification
• Results: Dissolved Oxygen
• Conclusions/Future Work/Upcoming Workshop
Presented at CBP Modeling Subcommittee Meeting
Annapolis, MD, January 12, 2011
Results: Dissolved Oxygen
bias [mg/L]
unbiased
RMSD
[mg/L]
(by M. Friedrichs)
1. Models reproduce seasonal/spatial variability of oxygen better
than stratification
2. Simple 1-term DO model gives lower RMSD than ICM
Observed and Modeled Top-to-Bottom DS and Bottom DO in Central Chesapeake Bay
(by M. Scully)
Results: Bottom DO (mg/L)
1-term DO model
mean station DO [mg/L]
mean station DO [mg/L]
ICM
1-term DO model underestimates high DO and
overestimates low DO: high not high enough, low not low
enough
(by A. Bever)
Results: Bottom DO (mg/L)
1-term DO model
latitude
latitude
ICM
1-term DO model underestimates DO in south and
overestimates in north; need more respiration in north
and less in south
(by A. Bever)
Results: Bottom DO (mg/L)
1-term DO model
ICM
Data pooled by month highlights underestimate of spatial
variability by 1-term DO model.
(by A. Bever)
Results: Hypoxic Volume – Model and Station-Based Estimates
(by A. Bever)
Results: Hypoxic Volume Sensitivity Tests (1-term model)
(by M. Scully)
Update on the Estuarine Hypoxia Component
of the SURA Super-Regional Modeling Testbed
Aaron Bever1, Marjorie Friedrichs1, Carl Friedrichs1
1Virginia Institute of Marine Science
and the Estuarine Hypoxia Testbed Team
Outline of Presentation:
• Intro: Participants/Motivation/Methods
• Results: Salinity and Stratification
• Results: Dissolved Oxygen
• Conclusions/Future Work/Upcoming Workshop
Presented at CBP Modeling Subcommittee Meeting
Annapolis, MD, January 12, 2011
Estuarine Hypoxia – Some Key Results from 1st Six Months of Project
• Hydrodynamic base runs for 2004 were completed and compared for five 3-D
models; CH3D did best (slightly).
• All models underestimate stratification (both magnitude and variability).
• Absolute salinity results notably improve with greater horizontal resolution and
with improved riverine forcing.
• Ability to reproduce observed stratification is not obviously sensitive to changes in
riverine forcing or in horizontal or vertical resolution.
• Initial results from a first round Hypoxia run for 2004 completed and assessed; a
very simple, physics-dominated DO model works about as well as ICM over a single
annual cycle.
• The annual cycle in DO variations is reproduced better than the annual cycle in
stratification and is surprisingly insensitive to variability in stratification.
• Bottom DO and hypoxic volume is most sensitive to seasonal variability in wind.
Seasonal changes in temperature (including respiration) are secondary. Seasonal
changes in discharge are unimportant(!).
Proposed 2nd Year Efforts
Use multiple hydrodynamic and hypoxia models to now-cast Chesapeake Bay (CB) water column salinity (S), stratification
(dS/dz) and dissolved oxygen (DO) every 2 weeks before observations are released by EPA monitoring cruises. S and dS/dz
could be predicted using EFDC, SELFE, and at least 2 types of ROMS. DO could be predicted with at least 3 types of
hypoxia models: (i) one-term DO respiration, (ii) statistical DO as a function of S & T, and (iii) one or more NPZD
formulations.
Use the testbed (including cyberinfrastructure web tools and displays) to compare and evaluate the skill of these 4 or
more hydrodynamic now-casts and 3 or more hypoxia now-casts every ~2 weeks (i.e., as soon as EPA data is released).
Multiple skill metrics would be utilized, such as Target analysis of S, dS/dz and DO, and examination of total hypoxic
volume.
Recode EPA’s CE-QUAL-ICM ecosystem/hypoxia model formulation (the formulation presently used by EPA to set TMDLs)
within ROMS to allow more carefully constrained direct comparisons of diverse hypoxia formulations. Include CE-QUALICM in the biweekly model prediction and comparison if possible.
Work with our testbed co-PIs at NOAA CSDL to truly transition the one-term DO respiration and statistical DO
formulations into the CBOFS2 version of ROMS and perform pre-operational and/or beta-public-releases of operational
forecasts of CB hypoxia (forced by wind forecasts) via NOAA CSDL and NCEP.
To allow broad community access to the above models, complete installation of plug-and-play executable versions of the
above hypoxia formulations in a community-accessible CSDMS HPC cluster in conjunction with ~ 5 different CB model
grids (ChesROMS, CBOFS2, UMCES-Li, CHIMP, VIMS-Shen) plus multiple sources for forcing (e.g., multiple wind models;
climatology vs. MAB model output).
Hydrodynamic Model Planning Meeting
A joint CCMP, CSDMS, CBP and SURA-Testbed hydrodynamic modeling workshop will be
convened at the Patuxent Wildlife Center, June 9-10, 2011.
Purpose is to review the state-of-the-art in coastal and estuarine hydrodynamic modeling and
compare the strengths and weaknesses of different types of models in terms of their grid
structures and their ability to simulate physical properties such as temperature and salinity
variability and stratification.
A strong emphasis will be placed on how well these models perform in Chesapeake Bay and
how differences might impact simulation of water quality parameters such as light, nutrients,
chlorophyll and oxygen concentrations.
The overarching goals of this workshop will be to 1) review, summarize and finalize the results
from the SURA-Testbed model intercomparison project and 2) provide input to the
Chesapeake Bay modeling subcommittee that can be used to inform selection of a future
hydrodynamic model for assessing water quality and management impacts.