Presentation to Whiting School of Engineering Class November 8, 2005

1. Upgrades of the Chesapeake Bay Airshed, Watershed, and Water Quality/Sediment Transport/Filter Feeder Models Presentation to Whiting School of Engineering Class November 8, 2005 Lewis Linker Chesapeake Bay Program Office

2. The purpose of models is insight, not numbers. Richard Hammond

3. Overview: • Overview/Key Points. • Tools for assessing the magnitude and effects of the nutrient and sediment loads. - Watershed Model Phase 5 - Water Quality/Sediment Transport/Filter Feeder Model - CMAQAirshed Model.

4. Overview/Key Points: Why We Are Upgrading • the Modeling Tools • What we’ve been asked to do is to achieve water quality standards for DO, chlorophyll, and clarity . • Our Tributary Strategies when fully implemented are estimated reduce nutrients by about one half and sediment by about one third from the zenith f load levels in the mid-1980s. • • The Tributary Strategy nutrient and sediment reductions are estimated to achieve the DO and chlorophyll water quality standards, but not the clarity standard, so more needs to be done.

5. Overview/Key Points: Why We Are Upgrading • the Modeling Tools • There are four sources of tidal suspended sediment: 1) fall lines of major rivers (AFL), 2) local watershed inputs (BFL), 3) shore erosion, and 4) sediment resuspension. Assessment of all four sources will be improved with the new tools available for the Reevaluation. • There’s a gap between the clarity-SAV water quality standard of 185,000 acres and the estimated SAV acres after full implementation of the Tributary Strategies (~138,000 acres). • • We’ve likely gone about as far as we can with watershed reductions and if any further reductions from this source are contemplated, they will likely be small.

6. Overview/Key Points: Why We Are Upgrading • the Modeling Tools • Shoreline sediment inputs are about equal to watershed inputs. We have yet to explore reductions from shoreline management, but reductions from this source are expected to be modest. • Filter feeders, particularly oysters, improve water clarity, enhance SAV restoration, and a decision on the a role filter feeders play in the clarity/SAV water quality standard will be needed. • Sediment resuspension in tidal waters is a wild card that will be quantified for the first time with our assessment tools. This will allow us to complete a suspended sediment budget from all sources for the first time, and will provide a sound foundation for sediment allocation decision-making.

7. Overview of the Assessment Tools: Current Modeling Structure A Regression Model of 15 monitoring sites over 10 simulation years. Changes in air quality management simulated with the Regional Acid Deposition Model (RADM) with a domain covering the Eastern states and limited grid capabilities Watershed Model Phase 4.3 94 model segments, 9 land uses, 20 calibration sites, 10 simulation years, fixed annual land use Chesapeake Bay Water Quality Model Hydrodynamic Model, Sediment Benthic Model, and Submerged Aquatic Vegetation, 10 simulation years, 13,000 model cells

8. Overview of the Assessment Tools: New Modeling Structure for the Reevaluation Nitrate and ammonia deposition from improved Daily Nitrate and Ammonium Concentration Models using 35 monitoring stations over 18 simulation years. Adjustments to deposition from Models-3/Community Multi-scale Air Quality (CMAQ) Modeling System Phase 5 Watershed Model Year-to-year changes in land use and BMPs; 899 segments; 24 land uses; 296 calibration stations; 18 simulation years; sophisticated calibration procedures; calibration demonstrably better in quality and scale Chesapeake Bay Estuary Model Detailed sediment input; Wave model for resuspension, Full sediment transport; Filter feeder simulation; Simulation of Potomac algal blooms; 54,000 model cells; 18 simulation years

9. Assessing the Magnitude and Effects of Sediment Loads: The Watershed Model Phase 5 Watershed Model Phase 5 refinements will improve assessments of: • Sediment loads from the fall lines (AFL) of major rivers. • Local watershed (BFL) sediment loads.

10. The Phase 5 Model Will Improve Assessments of Loads From the Fall Lines of Major Rivers: • The fall lines are well characterized by monitoring and modeling of sediment loads. Watershed erosion of sediment is generally well understood with areas of active investigation in legacy sediment, sediment scour from developed impervious watersheds, and stream channel restoration. • Phase 5 is expected to improve the simulation of fall line sediment loads with more spatial detail, more land uses of high sediment loads such as construction sites, better BMP characterization, and other improvements.

11. Phase 5 Increases Calibration Points 26 Fold (stations * time) Phase 4.3 Calibration Phase 5 Calibration Calibration sites = 20 Watersheds = 94 Land uses = 9 Simulation Years = 10 CB Watershed Calibration sites = 237 Watersheds = 684 Land uses = 24 Simulation Years = 18 Extended Network Calibration sites = 296 Watersheds = 899

12. The Mechanisms of Sediment Simulation are Improved in Phase 5 Edge of Field BMP Factor 1. Sediment processes are simulated on the land surface resulting in an Edge-Of-Field sediment load. More land use types are simulated in Phase 5. 2. A time series of Best Management Practices (BMPs) is applied based on available data. 4. A delivery factor based on the land use distance from the stream is applied (see below), resulting in the Edge-Of-Stream load. Land Acre Factor Edge of Stream Delivery Factor 3. A time series of land use acreage factors is applied. 5. Processes of deposition and scour are simulated in the stream, resulting in concentrations that can be compared to observations. In Stream Concentrations

13. Phase 4.3 had three Potomac calibration stations and none below the fall line. Phase 5 has 41 calibration stations with 13 below the fall line. Monocacy Chain Bridge Point of Rocks Conococheague Hancock Wills Creek Shenandoah North Fork South Fork Anacostia Quantico

14. Potomac Fall Line – Phase 4.3

15. Potomac Fall Line – Phase 5.0

16. Watershed Model (Phase 5) estimated sediment loads for the Potomac River compared to fall line monitoring estimated sediment loads Estimator

17. Susquehanna Above the Dams – Phase 4.3

18. Susquehanna Above the Dams – Phase 5.0

19. The Phase 5 Model Will Improve Assessments of Loads From Local Watershed (BFL) Sediment Inputs: • The BFL loads are generally less well characterized by monitoring and modeling than the major fall lines, but are expected to improve considerably with the refined Phase 5 Watershed Model augmented by increased local watershed monitoring, and shallow water monitoring.

20. Phase 5 Below Fall Line Model Segments at the Mouths of the Patuxent and Potomac Rivers. Phase 4.3 BFL model segments = 50 Phase 4.3 BFL calibration stations = 2 Phase 5 BFL model segments = 391 Phase 5 BFL calibration stations = 45

21. The Water Quality/Sediment Transport/Filter Feeder Model, Otherwise Known as the Bay Model, Will Improve Assessments of Tidal Sediment Loads Bay Model refinements simulating sediment transport will improve assessments of: • Shore erosion loads. • Resuspended sediment loads. • - Sediment reductions due to the effect of filter feeders. • - Water Quality/Sediment Transport/Filter Feeder Model refinements contingent on funding.

22. The Bay Model Provides Improved Shoreline Erosion Inputs: • A single estimate of shoreline erosion was used for the entire Bay (based on Ibison et al., date) for the 2003 Allocation analysis. • The Water Quality/Sediment Transport/Filter Feeder Model will improve on this with detailed estimates of erosion, on the scale of tens of kilometers, in Maryland and Virginia.

23. Shoreline Sediment Inputs are Estimated To Be About Equal to Watershed Inputs: Estimated Tidal Sediment Inputs for 1990 Shoreline inputs Watershed

24. The Bay Model Provides Simulation of Wave Resuspension Sediment Loads: • For the first time wave resuspension will be included in the Water Quality/Sediment Transport Model/Filter Feeder Model. The wave resuspension module uses at wind directions from the eight cardinal directions ( N, NE, E, etc.) and the fetch, or the length of water that the wind is able to transmit energy to wave generation and then in turn to sediment resuspension in shallow waters. • The time step is hourly for the entire simulation period of 1985 to 2000. • Simulation of wave simulation will allow us to complete a full suspended sediment budget for tidal “sediment sheds” accounting for fall line and local watershed, shoreline, and resuspended sediment sources.

25. Observed and simulated hourly wave heights at Poplar Island between October 27 to November 10, 1995. The wave simulation is applied to every shallow water cell in the WQSTM.

26. The Bay Model Provides Simulation of the Effect Oyster and Menhaden Filter Feeders Have in Reducing Suspended Sediment Loads The Water Quality/Sediment Transport/Filter Feeder Model will incorporate of oyster and menhaden filter feeders as well as four other generalized filter feeder groups to provide assessments of: • the synergy between health living resources and improved clarity conditions, such as healthy SAV beds reducing sediment resuspension. • the contribution filter feeders play in reducing suspended sediment and algae, and improving water clarity in shallow waters.

27. The Bay Model will allow a substantial reassessment of the table below which shows nonachievement of the clarity water quality standard in most of the Bay even under full application of the Tributary Strategies.

28. Other Refinements of the Water Quality/ Sediment Transport/Filter Feeder Model: • Sediment resuspension from currents, tidal and residual with improved spatial detail using 57,000 model cells compared to 13,000 model cells in previous version. • Improved clarity simulation (color, sediment, & algal components) for better simulation SAV and benthic algae. • If funded, refined suspended sediment reductions due to menhaden, oysters, and other filter feeders. Oysters have been demonstrated to have a substantial benefit in reducing shallow water suspended material.

29. A decision will need to be made on the role filter feeders play in attaining the SAV/clarity water quality standard

30. CMAQ Airshed Model • Replaces Regional Acid Deposition Model (RADM). • Provides estimates of nitrogen deposition resulting from changes in precursor emissions from utility, mobile, and industrial sources due to management actions or growth. • Provides estimates of the influence of source loads from one region on deposition in other regions.

31. How the Atmospheric Deposition Simulation Works From monitoring stations we get: - Daily precipitation volumes - Weekly NH4+ concentrations - Weekly NO3- concentrations We apply a regression model (J Lynch & J Grimm) to the monitoring station wetfall data to get spatially detailed daily ammonia and nitrate deposition. For the detailed spatial scales of the Phase 5 we refined spatial and temporal variations in wet deposition. • Phase 4 Watershed Model • 15 NADP/NTN monitoring stations • 1984-1992 • Phase 5 Watershed Model • 29 NADP/NTN monitoring stations • 6 AirMoN monitoring stations • 1984-2001

32. Regression Model Estimated Atmospheric Deposition NH4+ Wet-fall Concentration (mg/L) May 1, 1998 NH4+ Wet-fall Deposition (kg/ha) May 1, 1998 Estimates produced by applying daily ammonium concentration model to grids of estimated daily precipitation from the National Weather Service Climate Prediction Center’s U.S. Daily Precipitation Analyses.

33. Regression Model Estimated Atmospheric Deposition NH4+ Wet Deposition (kg/ha) Mean annual (1985-2001) NO3- Wet Deposition (kg/ha) Mean annual (1985-2001) Estimates produced by applying daily ammonium and nitrate concentration model to grids of estimated daily precipitation from the National Weather Service Climate Prediction Center’s U.S. Daily Precipitation Analyses.

34. How the Atmospheric Deposition Simulation Works • Community Multi-scale Air Quality (CMAQ) Model: • Replaces Regional Acid Deposition Model (RADM). • Provides estimates of wet:dry for nitrate and ammonium providing dry and total deposition and completing the base daily deposition data set. • In scenario mode, CMAQ provides estimates of nitrogen deposition resulting from changes in emissions from utility, mobile, and industrial sources due to management actions or growth. The base deposition determined by regression is adjusted by a reduction ratio deposition determined by CMAQ. • CMAQ estimates the influence of source loads from one region on deposition in other regions.

35. CMAQ Model Capacity Goes Beyond Nitrogen Deposition

36. 26,000 (tons) Base Case Clear Skies Chesapeake Bay Watershed Principle N Airshed NOx Emissions from Electricity Generation Sources within the Chesapeake Bay Nitrogen Airshed Projected in 2020: Base Case vs. Clear Skies Source: EPA OAR/OAP/CAMD 3/4/03

37. CMAC Scenarios Supporting the Reevaluation • Current funding allows for about ten CMAQ scenarios. The following list of key air scenarios for the Reevaluation have been identified by the Modeling Subcommittee: • 2015 CAIR (all current national air regulations included) 1/06. • 2020 CAIR (all current national air regulations included) 1/06. • 2020 CAIR with additional, aggressive utility or electric generating units (EGU) controls 7/06. • 2030 CAIR with aggressive EGU controls 4/06. • 2020 CAIR with additional, aggressive EGU, industry, and mobile source controls (to approximate a Limit of Technology future) 7/06. • Allocation of Bay State responsibility to watershed deposition for PA, VA, MD, DE, WV, NY 9/06 (each State requires a separate scenario).

38. Progress and Implications of Model Upgrades Summary: • Completing the sediment allocation will support healthy benthic and SAV communities in tidal waters and improve water quality in the watershed. • Completing the sediment allocation will be in some ways more difficult than the nutrient allocation: • we have less experience in managing tidal sediment. • we’ll need to be strategic as some suspended sediment loads may be beneficial. • legacy sediment is known to be a problem with lag times from decades to centuries. • we need to improve our understanding of the synergy between healthy benthic/SAV living resources and suspended sediment.

39. Progress and Implications of Model Upgrades Summary: • There are four sources of tidal suspended sediment, 1) fall line loads, 2) below fall line coastal plain loads, 3) shoreline erosion loads, and 4) resuspension loads. • The Watershed Model Phase 5 and the Water Quality/Sediment Transport/Filter Feeder Model will allow a complete tidal sediment budget estimate for all “sediment-sheds” from the four sediment sources. • There’s a gap between the clarity-SAV water quality standard of 185,000 acres and the estimated SAV acres after full implementation of the Tributary Strategies (~138,000 acres).

40. Progress and Implications of Model Upgrades Summary: • A decision on the a role filter feeders play in the clarity/SAV water quality standard is needed. • The CMAQ Model will improve our simulation of nitrogen deposition in the watershed and will allow a deeper integration of air and water programs trough its “one atmosphere” simulation approach providing nitrogen deposition, acid rain deposition, visibility, ozone, and PM 2.5 simulation in the same model.

41. Background and Documentation: • http://www.chesapeakebay.net/modsc.htm Under Publications tab is extensive documentation of all CBP models. • http://www.chesapeakebay.net/modsc.htm Under Current Projects and Info. tab are links to the community models of the watershed and estuary.

42. Specific Questions: • (1)     Development and update of the model • -          Who is the actual developer organization of the model? • A: Watershed Model – CBPO; Bay Model – Carl Cerco, ERDC; Airshed Model – Robin Dennis, EPA/NOAA. • -          How often is the model updated? • A: About every five years for each of the models. • -          How much does it cost and who covers such cost? • A: As an example the Bay Model costs about $3.5M with costs split about 50/50 among Federal and State partners. • (2)     Model linkage • -          Is the input of such loads as nitrogen and phosphorus from the airshed model and/or the watershed model to the water quality model done in each time step to each grid, or is it done on a monthly basis? • A: The time steps are small, down to seconds for the Airshed Model, hours for the Watershed Model and about 5 minutes for the Water Quality Model. Linkage among all is on a daily basis.

43. Specific Questions: • (3)     Outline of models • -          Do you have an overview diagram showing altogether the 24 state variables composing models and detailed flows of such elements as nitrogen, phosphorus and silica to link them? A: See The 2002 Chesapeake Bay Eutrophication Model on web site. • (4)     Model verification • -          Please show the model verification of each variable made at the time of the Chesapeake 2000 Agreement. We checked “water quality model and associated documentation” carried on your website but could not find the information we are looking for. A: See Shoreline Sensitivity Scenario SENS153 -Final Calibration with Refined SAV Simulation on web site. • -          In the model, do you take into account the respective inputs/outputs of carbon, nitrogen, phosphorus and silicate in the transitional process starting with inflow from rainfall, inflow from rivers, movements in the seabed – i.e. sinking from seawater into sediment/release from sediment back into seawater – and all the way through the boundary of the Bay? A: Yes. Full sediment diagenesis is included. • -          What does flux from bottom mud to seawater near the seabed generate? Do you have any other elements involved than ammonia, nitrate-nitrite or phosphate? What seasonal changes do they make? A: Large seasonal changes in PO4, NH4, & NO3 flux are simulated. • -          Is DO consumption from seawater to bottom mud near the seabed generated? What seasonal changes does that make? A: SOD is simulated and the seasonal changes are large.

44. Specific Questions: • (5)     Future prediction with model • -          Your year of prediction seems to be 2010. Do you have comparative calculation results of the reference year (i.e. the year for which the model recreated the conditions) vs. 2010? A: We have the Base run of 1985 to 1994. • -          In terms of water quality, do you use the calculation results of the model “as is” for the target values of 2010? Or are you trying to achieve the values set forth in the Clean Water Act? A: Our aim is to achieve DO, chlorophyll, and clarity that living resources need to thrive in the Chesapeake. • -          In making future predictions, what items and parameters are changed from the calculations of the reference year? In the airshed model and the watershed model, we understand that you have made calculations for that point of time in future and have changed the resultant outputs of such loads as nitrogen, phosphorus and other elements accordingly. Have you conducted any additional exercises with these models? A: Many questions, little time….. • -          If water quality is improved, light in the water will likely increase, which in turn will increase SAV. How have you dealt with this phenomenon in your prediction? A: SAV is fully simulated. • -          We understand that when loads from rivers are reduced, water quality of the Bay will improve, contributing to smaller deposits of organic suspended matter on the seabed. As a result, the amounts of ammonia, nitrate-nitrite and phosphate released from the seabed will be reduced and so will DO consumption. These changes should be duly taken into account to make an accurate future prediction of water quality. How are these matters reflected in your model? A: Sediment diagenesis is fully simulated.

45. Specific Questions: • (6)     Elementary Flux by Model • -          Do you have calculation results of annual flux of such elements as carbon, nitrogen, phosphorus and silica obtained from the results of calculations of the reference year? Also, do you have drawings? • A: Some flux calculations and graphics are available in the 2002 Chesapeake Bay Eutrophication Model Document. • Regarding the annual material circulation in the entire Bay obtained from the model, please describe the roles and functions of Oyster Reef and SAV. In particular, do you make a material assessment of the Bay as a tideland and SAV habitat?A: Oysters and SAV are directly simulated and coupled into the water quality simulation so they have a direct influence on water quality.