Regional Model Comparison: Mixing/Vertical Stratification in Chesapeake Bay
Session Date: June 1st 2010
Session Time: 11:47
Session Lead: Wen Long
Session Co-Lead(s): Ping Wang, Kevin Sellner
Session Abstract: A number of modelers from the region are currently conducting runs of multiple Chesapeake Bay hydrodynamic models using common data input for the period 2003-2005. The goal is to compare model output of vertical distributions of salinity and temperature versus field observations for specific stations/locations and times (well mixed late winter, stratified summer, and possibly mixed fall). The results will be presented at the May CheMS'10 meeting in a special 1/2 day session to identify models best approximating vertical distributions of salt and temperature, as these are critical to maximizing future biogeochemical modeling estimates of dissolved oxygen in the stratified bay, the governing factor in summer estimates of nutrient fluxes, productivity, and living resource distributions. Possible models in the suite include three versions of ROMS, CH3D, SELFE, and ELCIRC, with others possible in the future. Results will also inform future discussions of single model implementation for application in the bay or ensemble modeling to best approximate ambient conditions.
Presentations:
| Time | Title |
| 8:30 | Chesapeake Bay hydrodynamic models: Needs for ecological forecasting - Kevin Sellner - Chesapeake Research Consortium and Chesapeake Community Modeling Program |
| 8:50 | Salinity and Temperature Simulations for the Chesapeake Bay by the CH3D Hydrodynamic Model - Ping Wang - UMCES/CBPO |
| 9:10 | Identifying potential problems associated with numerical simulations of hydrodynamics in the Chesapeake Bay - Jian Shen - Virginia Institute of Marine Science |
| 9:30 | Evaluation of the NOAA/NOS Chesapeake Bay Operational Forecast System upgrade (CBOFS2) temperature and salinity predictions - Lyon Lanerolle - National Oceanic and Atmospheric Administration - National Ocean Service |
| 9:50 | Temperature and Salinity Predictions of the Chesapeake Bay Regional Ocean Modeling System (ChesROMS): Evaluation with In Situ Data and Intercomparison with Other Models - Wen Long - Univ of Maryland Center for Environmental Science |
| 10:30 | Group Discussion - Group |
Post-Session Review: This session was a focused study on the inter-comparison of the results of 4 models of the Chesapeake Bay with observational data. The over-arching goal was improving our understanding and predictive skills of salinity and temperature stratification structure of the Bay. The models examined during this session were CH3D (Wang, EPA Chesapeake Bay Program), EFDC (Shen, Virginia Institute of Marine Science), ChesROMS (Long et al., Univ of Maryland) and CBOFS2 (Lanerolle et al., NOAA). The comparison study and session presentations were coordinated to ensure consistent comparisons were made across the models using identical datasets. The later half of the session consisted of a 2 hour discussion among session attendees. The following items have been identified as important issues to follow up:
- We will conduct synthetic tests with unified forcing fields, not withstanding the differences between grid, bathymetry and structures of the models. The synthetic tests require that wind forcing, lateral flows and river forcings etc are all the same for all models.
- In addition to our initial focus on comparing model results and data in the main stem of the Bay, we will also compare results at the shallow flanks and shoals, as well as mid-reach of tributaries.
- The bathymetry of our models is different than the measurements for the Chesapeake Bay Program stations. Thus a sensitivity test of bathymetry for the model is needed. Most of our model bathymetry is from the NOAA Coastal Relief Model database and has not assimilated station measurements from the Chesapeake Bay Program. There is a need to assimilate the Bay Program depth measurements into model bathymetry in order to better resolve distinct features of the Bay, e.g. the deep channel along the central axis of the bay and shoal areas on the flanks.
- In order to better understand the difference between turbulence schemes in the models, we will also compare vertical profiles of TKE (turbulent kinetic energy), tracer and momentum eddy viscosity fields.
- Rebeca Murphy from JHU showed some new matrices for quantifying stratification, we can adopt her approach. (Rich Patchen and Lyon Lanerolle at NOAA will follow up on this aspect).
- It has been determined that we will need to include ground water flow and lateral inflows in the models. The CH3D model has basic lateral fresh flow loading to the Bay. There is a need to pursue a ground water flow dataset or model results (e.g. from Ward Sanford et al. at USGS) and incorporating lateral flow in models other than CH3D.
