rivers were omitted from the model setup. However, if a more thorough study of the 

 entire bay is needed, river discharges to the bay should be represented. 



Calibration of l\/lodel 



During the process of estabUshing a numerical model to represent a given study area, 

 calibration is performed to ensure it adequately predicts hydrodynamic conditions. 

 Accuracy of a model is determined by the accuracy of the boundary and forcing 

 conditions, representation of the geometry of the study area (i.e., bathymetry and land- 

 and-water interface), and to a lesser extent, by the values of certain parameters, 

 principally the bottom-friction coefficient. A satisfactory comparison between 

 calculations and measurements in the cahbration procedure gives confidence that the 

 model adequately simulates hydrodynamic processes. Cahbration exercises were 

 conducted via hindcast simulations of September 1997, for which measurements of the 

 current and water level were made as part of this project. 



Procedure 



Cahbration was performed primarily through adjusting the bottom-friction 

 coefficient, which was specified globally throughout the model domain. For the DMS 

 apphcation, no attempt was made to modify the value of the friction coefficient according 

 to location. Simulations were performed by varying the bottom-friction coefficient 

 between 0.0020 and 0.0050 to determine a first estimate. A simulation was conducted for 

 each coefficient, and its accuracy was determined by comparing calculated water-surface 

 elevations to those measured at the Port Lavaca water-level station. 



With this estimate serving as a basis, the model was run over a range of values with a 

 narrower range of coefficients specified for optimizing the calculated water-surface 

 elevations. Coefficients chosen for this stage ranged from 0.0030 to 0.0050, and the 

 appropriateness or accuracy of each coefficient was again determined by comparing 

 calculated water-surface elevations to those measured at the Port Lavaca station. 



Finally, the model was run with only two coefficients, 0.0035 and 0.0040, and then- 

 accuracy was judged by comparing calculated velocities to velocities measured in the 

 MSC and GIWW. The optimum value of the global bottom friction coefficient was 

 found to be 0.0040. 



Calibration period conditions 



The cahbration procediu-e previously discussed was conducted over a 14-day period 

 beginning on YD 243 (31 August 1997) at 0000 Greenwich Mean Time (GMT) and 

 endmg on YD 257 (13 September) at 0000 GMT. The model was run with a 4-s time- 

 step. Uniform, time-varying wind velocities were apphed over the entire model domain, 

 and water level forcing was specified at the open gulf boundary. Once the optimum 

 bottom friction coefficient value was determined, tiie model was run for the 31 -day 

 period begmning on YD 243 at 0000 GMT and endmg on YD 274 (1 October) at 

 0000 GMT. 



Coastal wind speeds measured at the East Matagorda station (Figure 22) for the 

 period YD 246 through 263 show a periodicity with maximum wind speed ranging from 

 about 5 to 8 m/sec. The hour at which these peak winds occur ranges from 1900 to 

 0000 hr GMT, or 1300 to 1800 hr Centtal Standard Time (CST), respectively. 



Chapter 3 Circulation Modeling 



27 



