DWR_3066022 - Laserfiche WebLink

Western Dam Engineering Technical Note May 2017 16 o Runoff rates and timing are most sensitive to watercourse routing parameters.  Water infrastructure – Water infrastructure (i.e. diversions, reservoirs, etc.) and their operational strategies can significantly alter watershed runoff volumes, rates and timing and may need to be considered as part of a calibration process. It should also be noted that stochastic methodologies and simulations, like Monte Carlo, are becoming more common and accessible for hydrologic evaluations associated with dams. These evaluations produce and model a suite of simulations (on the order of thousands) by varying input parameters based on specified probabilities of occurrence within a reasonable range of potential values. The results of these simulations provide confidence intervals for parameters of interest (e.g., peak runoff rate, etc.), which can be an extremely useful part of verification and calibration evaluations. Ultimately, the purpose of a verification and associated calibration process is to provide a reasonable and appropriate model for a range of conditions. As such, hydrologic models should not be calibrated to agree exceedingly well with only a single condition or scenario, as this could render the model unacceptable and/or inappropriate for other conditions or scenarios. For this reason, model validations are performed to provide a basis for adjusting and testing calibrated models for additional conditions and scenarios. Model Validation and Verification General Discussion and Overview Model validations are evaluated based on an initially calibrated model and comparisons between model results and observed or estimated data resulting from independent or additional conditions and scenarios. For example, the gage discharge data shown in Figure 7 represent a time series record. The entire record is not used as part of initial model calibrations so that a portion of the data set can be “set aside” and used to independently validate model results for the initially calibrated model. These “independent” data sets differ based on available data and the nature of corresponding calibrations:  Calibrations – Independent data are based on measured or observed (i.e., gage) data and could include additional individual precipitation or stream discharge events, derived frequency events Figure 6. Hydrologic Model Calibration Process