MM5: We presently use the NCAR/Penn State fifth-generation mesoscale regional weather model (MM5) (Dudhia et al., 2003) to compute the surface forcing fields needed to drive FVCOM.  MM5 features non-hydrostatic dynamics, terrain-following sigma-coordinate, variable domain and spatial resolution, multiple grid nesting, 4-D data assimilation, and several planetary boundary layer (PBL) modules to represent turbulent mixing over the ground and ocean (Grell et al., 1994), and uses NCEP weather model fields as initial and boundary conditions with two-way nesting capability.  We have modified MM5 to construct a surface weather hindcast and forecast system for fishery studies in the GoM/GB (Chen et al., 2005a). This model (called GoM-MM5) is configured with a “regional” domain (covering the Scotian Shelf/GoM/GB/New England Shelf) and a “local” domain (New England Shelf/ Mass Bay) with horizontal grid spacing of 9 and 3 km respectively, and 31 sigma levels in the vertical with finer resolution in the PBL.  The NB/MHB-MM5 was set up as a subgrid of GoM-MM5, with a horizontal resolution of 1 km covering both bays and calibrated using observed winds recorded on 6 NOAA buoys in this region (Fig. 2).

To improve the model-based surface wind stress and heat flux estimates, GoM-MM5 uses the COARE 3 bulk algorithm (Fairall et al, 1996, 2003) for the air-sea fluxes, insolation and cloud cover data from the International Satellite Cloud Climatology Project for the radiative fluxes, and all coastal NDBC and C-MAN surface weather data available in the local domain are incorporated through 4-D data assimilation.

NB/MHB-MM5 is presently in operational use to forecast surface weather conditions over the eastern NES region, providing a 3-day forecast of surface winds, heat flux and precipitation/evaporation for research of the impact of the Brayton Pt. Power Plant on the MHB ecosystem. With assimilation of in-situ surface wind and other data collected in and around NB/MHB, this high-resolution “coastal” MM5 accurately reproduces the sea breeze which dominates the surface wind field variability in this region. We are working on replacing MM5 with the new state-of-the-art community mesoscale Weather Research Forecast model (WRF) in 2006, which will allow horizontal resolution down to 1 km.  The new NCEP North American weather Model (NAM) incorporates the WRF core model physics and provides a horizontal resolution of 8-12 km.  This advance combined with the use of WRF in our GoM/NB/MHB system should improve the accuracy of the surface fields used in this project.

FVCOM: FVCOM is a prognostic, unstructured-grid, finite-volume, free-surface, 3-D primitive equation coastal ocean circulation model coded for full parallelization (Chen et al., 2003; Chen et al. 2004, Cowles, 2004). In common with other coastal models, FVCOM uses the modified Mellor and Yamada level 2.5 (MY-2.5) and Smagorinsky turbulent closure schemes for vertical and horizontal mixing, respectively (Mellor and Yamada, 1982; Galperin et al., 1988; Smagorinsky, 1963), and a sigma coordinate to follow bottom topography. The General Ocean Turbulent Model (GOTM) developed by Burchard’s research group in Germany (Burchard et al., 1999; Burchard, 2002) has been added to FVCOM to provide optional vertical turbulent closure schemes. Unlike existing coastal finite-difference and finite-element models, FVCOM is solved numerically by calculation of fluxes resulting from discretization of the integral form of the governing equations on an unstructured triangular grid. This approach combines the best features of finite-element methods (grid flexibility) and finite-difference methods (numerical efficiency and code simplicity) and provides a much better numerical representation of momentum, mass, salt, heat, and tracer conservation. The ability of FVCOM to accurately solve scalar conservation equations in addition to the topological flexibility provided by an unstructured grid and the simplicity of the coding structure makes FVCOM ideally suited for interdisciplinary applications in the integrated NB/MHB system (Chen et al., 2005b-c; Ji et al., 2005a-c).

FVCOM can be run in hind-, now-, or forecast mode using realistic forcing. To improve the accuracy of these simulations, in-situ data can be assimilated using Kalman filter methods. In collaboration with P. Malanotte-Rizzoli and her MIT colleagues, we have implemented an Ensemble Kalman Filter (EnKF) (Zang and Malanottee-Rizzoli, 2003) into FVCOM. We plan to use the Ensemble Kalman Filter for hind- and now-casting in the proposed work. FVCOM can also be run in prognostic mode with idealized forcing to investigate specific processes and “what if” scenarios concerning changes in tidal currents and circulation associated with proposed dredging operations and changes in coastlines and bathymetry due to large storms as well as the study of river/estuary/bog systems where there is significant groundwater. Examples can be viewed on http://codfish.smast.umassd.edu or http://fvcom.smast.umassd.edu.

MHB/NB FVCOM is configured with an unstructured triangular mesh covering the entire MHB/NB region (Fig. 3). The computational domain is bounded in the inner shelf with an open boundary line running from Rhode Island Sound to Buzzard Bay. Our initial experiments show that in order to resolve current separation due to strong tidal flows in MHB, a model must have a minimum resolution of ~50 m or less (Zhao et al., 2006). To resolve the thermal plume due to the hotwater injection from the power plant, a model must have a minimum resolution of 10 m or less. Our present MHB/NB model grid has a resolution of 10-20 m in upper NB and GB to fully resolve small- and meso-scale eddies in PR and GB. 31-sigma levels are used in the vertical, which corresponds to a vertical resolution of less than 1 m in the deepest parts. FVCOM also includes a mass-conservative wet/dry point treatment method that can simulate the flooding/drying process over intertidal wetlands.

FVCOM Biological Module: Various ecosystem models have been implemented in FVCOM, including NPZ, NPZD, NPZDB, and water quality models. To make FVCOM more flexible for use in ecosystem studies, we have built a unstructured grid generalized FORTRAN 95 biological module (GBM) into FVCOM to allow users to select either a pre-built biological model (such as NPZ, NPZD, etc) or construct their own biological model using the pre-defined pool of biological variables and parameterization functions. GBM includes seven groups: (1) nutrients, (2) autotrophy, (3) heterotrophy, (4) detritus, (5) DOM, (6) bacteria, and (7) auxiliary. A biological model can be constructed using “function pointers” to select both model structures and parameterization functions. GBM can be run simultaneously together with FVCOM with parallelization (“online” mode) or driven separately by FVCOM output (“offline” mode).  This module acts like a platform that allows us to examine the relative importance of different physical and biological processes under well-calibrated physical fields. Validation tests are presently being conducted to use GBM to re-build NPZ (Franks and Chen, 1996, 2001) and multi-species NPZD models (Ji et al., 2005a-c) for use in GB/GoM applications (see Tian and Chen, 2005 for details).

One of this application GBM was to use it to build an ecosystem model for the study of relative influences of natural (river discharges) and anthropogenic (hot-water plume) on the occurrence of the low-DO zone (or hypoxia) in the MHB/NB system. Hypoxia is the ecosystem result from imbalance between dissolved oxygen replenishment and demand. Excessive nutrient loading fuels phytoplankton growth and DO produced through phytoplankton photosynthesis in the euphotic layer is partly lost to the atmosphere due to air-sea exchange and utilized by heterotrophic processes among the food web. The DO concentration in shallow areas is also influenced by the subsequent remineralization of the synthesized organic matter in the deeper layer and in the sediment.  An ecosystem model suitable for the study of hypoxia should include the key physical (air-sea fluxes, advection, and mixing) and biological/biogeochemical processes occurring in NB/MHB.  There are two existing biological models for NB. The first one is the food web model developed by Kremer and Nixon (1978) and the other is the water quality model (WQM) implemented in FVCOM for MHB studies (Chen et al., 2004). The KN model focuses on the food web dynamics without explicit inclusion of DO, so that it cannot be applied directly to study hypoxia. The FVCOM WQM was built based on the EPA water-quality analysis simulation program (WASP) by including the benthic flux from sediment resuspension via sedimentation processes (Amborse et al., 1993, Zheng et al., 2004). This is a typical eutrophication model consisting of eight water-quality state variables: 1) ammonia (NH3); 2) nitrate and nitrite (NO2 and NO3); 3) inorganic phosphorus (OPO4); 4) organic nitrogen (ON); 5) organic phosphorus (OP); 6) phytoplankton (PHYT); 7) carbonaceous biochemical oxygen demand (CBOD); and 8) dissolved oxygen (DO). The benthic and sediment suspension processes are incorporated into the water model by adding a benthic layer and a sediment pool at the seabed. This biological/chemical model incorporates basic transformation processes including photosynthesis, uptake, respiration, nitrification, denitrification, benthic flux, sediment suspension, and external loads, etc. This WQM, which is similar to the water quality model developed by Spaulding et al. (1999) for NB/MHB, however, does not contain heterotrophic and bacterial processes, so that it can not capture realistic food web dynamics in NB/MHB.

We have used the FVCOM GBM to construct an ecosystem model for the study of hypoxia and phytoplankton dynamics in NB/MHB (Fig. 4).  This model consists of 3 nutrients [NH4+, NO3- and Si(OH)4], 2 phytoplankton (small and large sizes), 2 zooplankton (mirco and meso sizes), bacteria, detritus, DOM and DO. NB is a nitrogen-limited system in which growth and seasonal succession of diatoms are regulated by silicate (Kremer and Nixon, 1978; Oviatt et al., 1995; Townsend and Thomas, 2001, 2002). NH4+ is a key index of the nutrient supply through the internal regeneration process. Dividing the phytoplankton into 2 species helps us distinguish diatoms (silicate-dependent) and flagellates (non-silicate-dependent) that show distinct functions in the energy flux among the food web. Microzooplankton feed mostly on small phytoplankton, whereas meso-zooplankton mainly consumes diatoms. They have different efficiencies in energy flow from the pelagic to the benthic system: fecal pellets of microzooplankton remain suspended in the water column, while those of meso-zooplankton sink toward the bottom. Detritus consist of dead phytoplankton and zooplankton, fecal pellets or aggregate, which sinks in part into the sediment and taken directly by bacteria. Carnivores (such as ctenophores, fish larvae and menhaden) are key predators regulating zooplankton abundance in NB (Kremer and Nixon, 1978). Since our focus is on nutrient-sustained phytoplankton production and remineralization that leads to hypoxia, we combine predators into a single term that simulates predation pressure on zooplankton. Unlike WQM, we introduce DOM to represent all dissolved organic matter that is remineralized mainly by bacteria.  Based on a 1-D model experiment, Kremer and Nixon (1978) suggested that benthos (shellfish) respiration and sediment oxygen demand are two potential candidates that are related to rapid consumption of DO in NB. These two processes are taken in account in the proposed biological model.  DO constitutes the major component of the model in which DO is replenished and produced mainly by re-aeration at the air-sea interface and phytoplankton photosynthesis and consumed by heterotrophic respiration (micro- and meso-zooplankton, predators, benthic shellfish and bacteria). The model also includes the DO demand due to denitrification and sediment decomposition (Seitzinger et al., 1984; Rudnick and Doering, 1985).