MODELING THE PHYSICAL, OPTICAL AND BIOLOGICAL PROPERTIES OF CHESAPEAKE BAY

Abstract

This thesis describes a relatively simple biogeochemical model that I developed and coupled with a three-dimensional circulation model of the Chesapeake Bay. To improve the performance of the physical model I attempted to assimilate high-resolution salinity data using a Newtonian relaxation scheme. In general, the simple assimilation scheme leads to visibly improved density structures in the Bay. However, the injection of high-resolution salinity data produces transient gravitational readjustment, which can have a significant impact on the biogeochemical properties and processes in the estuary. Therefore, this approach cannot be directly applied in biogeochemical modeling studies. Instead, I show that adjusting the salinity at open-ocean boundaries is also able to improve the density structure of the inner estuary.

To obtain a relatively simple but effective way to model light attenuation variability in the coupled physical-biological model, I adopted a simple, non-spectral empirical approach. Surface water quality data and light measurements from the Chesapeake Bay Program were used to determine the absorption coefficients in a linear regression relationship. The resulting model between light attenuation coefficient (Kd) and water quality concentrations (chlorophyll, TSS and salinity as a proxy for CDOM) gives generally good estimates of Kd in most parts of Chesapeake Bay. I also discuss the feasibility and caveats of using Kd converted from Secchi depth in the empirical method.

To develop the relatively simple biogeochemical model for Chesapeake Bay, I adopted a simple NPZD-type biological model and added in necessary additional components and simple parameterizations of the important processes for estuarine applications. The coupled model is then run under very different conditions: a dry year (1995) and a very wet year (1996). Observations of DIN, chlorophyll, total suspended solids (TSS), dissolved oxygen (DO), and light attenuation coefficient (Kd) obtained from Chesapeake Bay Program are used to validate the model. I demonstrate that this simple biological model is capable of reproducing the major features in nutrient, phytoplankton, DO, TSS and Kd distributions in a complex ecosystem like Chesapeake Bay, and the model is robust enough to generate reasonable results under both wet and dry conditions. Sensitivity studies on selected parameters are also reported.