A Procedure to Identify and Rank Rainfall/Runoff Phenomena for the Evaluation of Urban Stormwater Models
Computer models are needed to predict the effects of changes in land-use and climate within an urban stormwater drainage catchment. As with any computer model result, it is important to clearly state the reliability of the calculations and the modeling assumptions that were used (James, 2005). A common practice that is used to demonstrate the reliability of an urban stormwater model is to compare calculated results to data that are obtained from field measurements within the catchment (often a split-test calibration/verification procedure is employed). This practice is applicable to the analysis of existing drainage catchments using historical weather data but the difficulty with predicting the effects of changes in land-use or climate is that field data on the changed system are not available. Therefore, the predictive capabilities of the model must be validated, which includes establishing the uncertainties in the calculations (O’Connell and Todini, 1996; Ewan and Parkin, 1996). Quantification of these uncertainties is needed so that informed decisions can be made regarding the implementation of a land-use change or how best to mitigate the risk associated with climate change scenarios.
This paper is only available in PDF Format:
View full text PDF