A Numerical Simulation of Residual Circulation in Tampa Bay. Part II: Lagrangian Residence Time

Lagrangian retention and flushing are examined by advecting neutrally buoyant point particles within a circulation field generated by a numerical ocean model of Tampa Bay. Large temporal variations in Lagrangian residence time are found under realistic changes in boundary conditions. Two 90-day time periods are examined. The first (P1) is characterized by low freshwater inflow and weak baroclinic circulation. The second (P2) has high freshwater inflow and strong baroclinic circulation. At the beginning of both time periods, 686,400 particles are released uniformly throughout the bay. Issues relating to particle distribution and flushing are examined at three different spatial scales: (1) at the scale of the entire bay, (2) the four major regions within the bay, and (3) at the scale of individual model grid cells. Two simple theoretical models for the particle number over time, N(t), are fit to the particle counts from the ocean model. The theoretical models are shown to represent N(t) reasonably well when considering the entire bay, allowing for straightforward calculation of baywide residence times: 156 days for P1 and 36 days for P2. However, the accuracy of these simple models decreases with decreasing spatial scale. This is likely due to the fact that particles may exit, reenter, or redistribute from one region to another in any sequence. The smaller the domain under consideration, the more this exchange process dominates. Therefore, definitions of residence time need to be modified for “non-local” situations. After choosing a reasonable definition, and removal of the tidal and synoptic signals, the residence times at each grid cell in P1 is found to vary spatially from a few days to 90 days, the limit of the calculation, with an average residence time of 53 days. For P2, the overall spatial pattern is more homogeneous, and the residence times have an average value of 26 days.