The Phase II model application included the same configurations as Phase I at the north boundary of the Lake, which are designed to increase exchange between Lake Jesup and the St. Johns River.   All six cases included the existing Channel A connecting  the St. Johns River with  lake Jesup.  In test Case 1 Channels B and C were included as additional connections to the St. Johns River at the north boundary of the Lake.  Case 1 conditions also included closing of Government Cut.  In Case 2 both Channel C and Government Cut were closed, leaving Channels A and B open.  Case 3 included   Channel C open and Government Cut open.  In Case 4, Channel C was closed, but Government Cut remained open leaving Channels A and B to handle exchange with the St. Johns River.  Case 5 included Channel A and C as open, Government Cut closed, and Channel B as closed.    Case 6 was designed to simulate conditions that would exist if the entire State Route 46 Causeway was removed.  In this case, with the exception of Channel A, exchange with the St. Johns River occurred across a shallow area of wetlands and flood plain having no distinctive channels.

         Boundary conditions used to force the 3-D model included time series data of water elevations at the north boundary at Channel A and at Elder Spring.  In the Phase II studies of river-induced flushing an adjusted water elevation time series provided by the U.S. Geological Survey (USGS) was added at the entrance of Channel C. Under Cases 1 through 5 the average elevation of this time series was increased by 3 cm above the average elevation and applied at Channel A to provide hydraulic forcing of River Flows into the model. Under Case 6 it was assumed that the water elevation across the north boundary of the Lake would not be substantially different since this would be a broad shallow wetland area having no distinctive channels after removal of the SR 46 Causeway.  Thus, the USGS water elevation time series was applied at the location of Channel C without the 3 cm adjustment.

      Initial results of the six tests cases predicted that a component of the hydrologic balance of Lake Jesup was due to water volume exiting the model at the southwest open boundary cell located at Elder Spring.  Results of these early runs were presented in a project summary on September 21, 2000. Based on discussions with District Scientists who are familiar with the watershed features of Lake Jesup, it was concluded that the hydrologic balance of Lake Jesup does not include any significant “backflow” into the surrounding watershed.  Thus, in the final set of model runs for the six test cases exit of water from the southwest boundary was eliminated by “masking” the computational model cells around the open boundary cell at Elder Spring. This resulted in the primary hydrologic balance of the model to occur between watershed inflows, exchange across the north boundary of  Lake Jesup, and evaporation and rainfall as specified in the meteorological sub-model.  The balance was then reflected in the approximate balance between gross inflow and outflow at the north boundary of the model domain.

         Model simulations included numerical Lagrangian drifters specified in the model to track flushing rates and simulations of current velocity to predict circulation patterns.  Drifters were launched from 50 locations within the Lake in the first set of runs under Phase II. A second set of model runs of all six cases included multiple launches of 4 drifters at the north boundary of the model to track the intrusion of water from the St. Johns River into the Lake and examine flushing by river inflows in the upper Lake.

          Results of the model simulation show that all of the six cases examined would provide flushing of the upper compartment of Lake Jesup. Flushing in the upper Lake is attributed to strong volume exchange with the St. Johns River. The gross volume exchanged between the River and Lake was predicted to be in the range of 50 to 150 million cubic meters over a 100-day simulation period.  However the net inflow from the St. Johns River into the lake was predicted to be relatively small and ranged from 5 million to 15 million cubic meters. In each test case the net inflow was predicted to be smaller than the gross volume exchange by an order of magnitude.  The net predicted river inflow ranges from 15% to 47% of the total net Lake volume increase that occurred during the model tests. Thus, it is concluded that the Lake remained strongly influenced by watershed inflows under all cases that were tested.

       The flushing times and patterns among the cases were generally similar. In addition to flushing of the upper Lake within 30 to 40 days after the model launched, a secondary area in the central Lake was flushed within 65 to 81 days.  The size of the second central Lake area flushed at a slower rate  varied considerably.  The largest area of flushing was predicted to occur under Cases 2 and 6.  The total area of the lake flushed and the rate of flushing of the central Lake were similar under Cases 2 and 6, which can be rated as providing the best overall benefits with respect to flushing among all model test cases.  From the results of the model tests it is concluded that introducing river water into the Lake provided rapid flushing of the upper Lake and moderate to good flushing of the central Lake. However, it is likely that the river water moved into Lake Jesup and formed a front that confined the lower Lake and prevented flushing of the lower compartments of the Lake.   To further distinguish among the cases predicted to be most beneficial it is recommended that a year-long model simulation be conducted for Cases 2 and 6. The results of these two year-long simulations should be compared to the results of a year-long simulation of existing conditions. The major goal of these final runs would be to compare flushing among the various cases during the wet season to flushing patterns during the transition to the dry season when water levels and the volume of Lake Jesup decreases.