A probability analysis of salinity conditions in the St. Sebastian River estuary, taking into account the effects of freshwater discharge over a period from May 1992 to March 1996, was used to determine the likelihood (probability) that salinities, as represented by daily mean specific- conductance values, will fall below a given threshold. The effects of freshwater discharge on salinities were evaluated with a simple volumetric model fitted to time series of measured specific conductance, by using nonlinear optimization techniques. Specific-conductance values for two depths at monitored sites represent stratified flow which results from differences in salt concentration between freshwater and saltwater. Layering of freshwater and saltwater is assumed, and the model is applied independently to each layer with the assumption that the water within the layer is well mixed. The model of specific conductance as a function of discharge (a salinity response model) was combined with a model of residual variation to produce a total probability model. Flow distributions and model residuals were integrated to produce a salinity distribution and determine differences in salinity probabilities as a result of changes in water-management practices.
Two possible management alternatives were analyzed: stormwater detention (reducing the peak rate of discharge but not reducing the overall flow volume) and stormwater retention (reducing peak discharges without later release). Detention of freshwater discharges increased the probability of specific- conductance values falling below a given limit (20,000 microsiemens per centimeter) for all sites but one. The retention of freshwater input to the system decreased the likelihood of falling below a selected limit of specific conductance at all sites. For limits of specific conductance (1,000 microsiemens per centimeter or 20,000 microsiemens per centimeter, depending on the site), the predicted days of occurrence below a limit decreased ranging from 17 to 68 percent of the predicted days of occurrence for unregulated flow.
The primary finding to be drawn from the discharge-salinity analysis is that an empirical-response model alone does not provide adequate information to assess the response of the system to changes in flow regime. Whether a given level of discharge can produce a given response on a given day is not as important as the probability of that response on a given day and over a period of many days. A deterministic model of the St. Sebastian River estuary based only on discharge would predict that retention of discharge peaks should increase the average salinity conditions in the St. Sebastian River estuary. The probabilistic model produces a very different response indicating that salinity can decrease by a power of three as discharges increase, and that random factors can predominate and control salinity until discharges increase sufficiently to flush the entire system of saltwater.