A simple empirical optical model for simulating light attenuation variability in a partially mixed estuary

Representation of the subsurface light field is a crucial component of pelagic ecosystem and water quality models. Modeling the light field in estuaries is a particularly complicated problem due to the significant influence of high concentrations of dissolved and particulate matter that are derived from both terrestrial and estuarine sources. The goal of this study was to develop a relatively simple but effective way to model tight attenuation variability in a turbid estuary (Chesapeake Bay, United States) in a coupled physical-biological model. We adopted a simple, nonspectral empirical approach. Surface water quality data (salinity was used as a proxy of chromophoric dissolved organic matter [CDOM]) and light measurements from the Chesapeake Bay Program were used to determine the absorption coefficients in a linear attenuation model using regression methods. This model predicts K-c (specific attenuation due to phytoplankton/chlorophyll a [chl a]), K-t (specific attenuation due to total suspended solids), and K-s (a function of specific attenuation coefficients of CDOM in relation to salinity). The Bay-wide fitted relation between the tight attenuation coefficient and water quality concentrations gives generally good estimates of total light attenuation, K-d. The direct inclusion of salinity in the relationship has one disadvantage: it can yield negative values for K-d at high salinities. We developed two separate models for two different salinity regimes. This approach, in addition to solving the negative K-d problem, also accounts for some changes in specific light absorption by chl a, seston (nonphytoplankton particulate matter), and CDOM that apparently occur in different salinity regimes in Chesapeake Bay. The resulting model predicts the statistical characteristics (i.e., the mean and variance) of K-d quite accurately in most regions of Chesapeake Bay. We also discuss in this paper the feasibility and caveats of using K-d converted from Secchi depth in the empirical method.