Effects of spatial and temporal variability of rainfall on urban hydrological models

Abstract

"The working hypothesis is that as fewer gauges are used to interpolate a rain event there will be more error. Preliminary results indicate that high rain events produce much more error, around 300% relative error. It is also seen that the bigger the watershed the bigger the errors – this is mainly due to the fact that larger storms are associated with a larger variance and reducing the number of gauges used reduced the information provided to the model. Future work will focus on calibrating the model with field data and developing robust mechanism in order to properly choose which gauges to keep and what type of averaging and interpolating conditions should be used. "