Model Calibration. This task will provide the necessary model calibration to ensure that the Water and Wastewater models are representative of the behavior of the system on a day-to-day basis and to ensure model outputs are as accurate as possible.
Model Calibration. There shall be steady state and transient calibrations of the models. The steady-state calibration shall be performed to predevelopment conditions as defined in section 3.1.5. The mean absolute error or root mean squared error between measured hydraulic-head and simulated hydraulic head shall be less than 10 percent of the measured hydraulic-head drop across the model area for each model layer, and better if possible. The error shall not be spatially biased (For example, not by areas with more control points than other areas). Final calibration results shall report the mean absolute error, root mean squared error and the mean error (Xxxxxxxx and Xxxxxxxx, 1992, p. 238-241). The difference between the total simulated inflow and the total simulated outflow (that is, the water balance) shall be less than one percent and ideally less than 0.1 percent for each model layer within each county. Initial parameters for the models shall be derived from the data generated during the development of the conceptual models. Parameters adjusted during calibration (for example, recharge, hydraulic conductivity, and vertical hydraulic conductivity) shall be within defensible limits within the framework of the conceptual model such that the resulting model has realistic values and realistic spatial distributions of parameters. Any changes to model parameters must be thoroughly documented in the final report. If unrealistic hydrologic parameters must be used to calibrate the model or the model cannot be calibrated to the above calibration criterion for matching hydraulic head or the error on the water balance, the project manager shall meet with TWDB staff to discuss how to proceed with the model. The TWDB does not want over-calibrated models. The transient model shall start with the steady-state model for the first stress period and stop at the end of the most recent year with calibration data. Stress periods may be of variable length according to the density of information on pumping and recharge, but the stress periods for the transient historical calibration period shall not be greater than one year. Particular attention shall be paid to accurately representing water levels and fluxes during times of drought and in areas with large drawdowns. Mean absolute error or root mean squared error between measured hydraulic head and simulated hydraulic head should be less than 10 percent of the maximum hydraulic-head drop across the model area and better, if possible, for specified...
Model Calibration. The DWR established high water benchmarks during the spring of 2005 during Lewiston Dam releases of 7,000 cfs and 4,500 cfs. Benchmarks were located at 102 locations between Lewiston Dam and the North Fork Trinity River over the course of approximately 40 miles. After flows receded, the DWR surveyed each benchmark to determine the actual water surface elevation observed during the high flow releases. Mainstem USGS gage data were analyzed to determine the flow at each benchmark at the time it was established. Main channel Xxxxxxx’x roughness values in each model were adjusted over a range from 0.030 to 0.040, and overbank Xxxxxxx’x roughness values typically varied between 0.080 and 0.200, to match DWR surveyed water surface elevations within 0.5 feet for the 4,500 and 7,000 cfs discharge. Figure 2 shows the model calibration results with computed versus measured water surface elevations within the Xxxxxxx City/Indian Creek study reach. Figure 2 also identifies the location of tributaries that provide accretion flow to this reach (i.e., Indian Creek, Xxxxxx Creek and Reading Creek). In May of 2006, the DWR again established benchmarks during the TRRP fishery flow releases of 10,000 cfs. These benchmarks have yet to be surveyed. However, at the time of the 10,000 cfs release, the HEC-RAS model developed by DWR was run at the same flows as were actually occurring at the time, with very little deviation (less than 6 inches in observed locations) between predicted and actual water surface elevations. Therefore, the TRRP believes that the HEC-RAS model for the No Action alternative may be used to accurately predict water surface elevations at flows in the Xxxxxxx City to within 6 inches.
Model Calibration i. The Consultant shall prepare an approach for the model calibration and provide the City with a list of data requirements. The model will be calibrated using data provided by the City.
Model Calibration
