Submitted to: ASAE Annual International Meeting
Publication Type: Proceedings
Publication Acceptance Date: 6/15/1998
Publication Date: N/A
Citation: Interpretive Summary: Typically, watershed models use precipitation data developed from 24-hr totals and a synthetic pattern of rainfall to describe how rainfall intensities vary within a storm. However, newer models of watershed hydrology and water quality require short-time increment precipitation data of the order of minutes that have the same characteristics as storms for selected areas. The data for such model needs either do not exist, or do not have a sufficiently long record length. A computer model was developed that generates a time sequence of storm durations and storm depths. The model was tested to determine the best mode of operation of the model when data are supplied to it on a monthly basis. The results show the model works well. The model is useful as a first step in simulating the short-time increment data that watershed models require. It also can be incorporated into weather generator models that synthesize long records of weather variables such as temperature, solar radiation, etc. It has the potential for use in global climate models to study different climate-change scenarios.
Technical Abstract: A storm generator was developed using an empirical and Monte Carlo approach. Storms were identified in a record of precipitation data by determining the critical duration that separates bursts of rainfall from one another. Empirical distributions of storm durations and subsequent storm depths were developed. These were used as input to a model that samples the observed distributions and generates a sequence of storm outputs through time. The study investigated the effects of incorporating several types of interpolation schemes between months and between storm durations. Data time-discretized in this manner were the bases for characterization of data for the model. A simple interpolation between storm-depth distributions and no month-to-month interpolation gave good results for modeling both frequency distributions and monthly and period total rainfall. The model has potential for use in simulating within-storm mintensities by providing a time series of storm depths and durations that can be subsequently disaggregated. The storm simulator can also be incorporated into a weather generator given proper parameters.