Abstract

Urban rainfall-runoff models were the first hydrologic simulation models that computed runoff separately for different land cover classes, namely pervious and impervious areas. The two contributions were added and routed together through sewers and surface drainage systems. This method of simulating storm runoff has been successfully applied to a larger number of land use classes, namely impervious, barren or low vegetation, intermediate height vegetation, forests, wetlands and water covered areas.

The problem encountered was the definition of parameter values, the most important being the effective saturated conductivity of the soil and the overland flow resistance. For urban areas, these values are well known for impervious areas. Thus only the values for the pervious areas needed to be optimised. However, for non urban areas, or for urban areas with a greater diversity of land cover classes, these parameters (and others) needed to be determined for each land cover class. This can only be accomplished on watersheds where there are at least as many streamflow gauges as there are land cover classes and where there is a diversity of the mix of classes.

Remotely sensed data offers in each gauged area the detailed real-time information on watershed conditions and rainfall amounts necessary to solve this problem. This information is of particular importance in flood forecasting because the hydrologic processes that most directly affect flash floods occur at the land surface and are greatly influenced by the terrain, or land cover characteristics of the watershed. Weather radar now provides detailed real-time rainfall data at frequent intervals. The timing and areal distribution of rainfall also directly affect the flood hydrographs. It thus appears that data management systems and hydrologic models designed to make maximum use of this remotely sensed data would offer improved flash flood forecasting.

This paper describes the application of the WATFLOOD data management system and the SIMPLE rainfall-runoff model to a heavily urbanised watershed that exceeds a size that can be readily modelled by a conventional urban runoff model. The problems associated with parameter optimisation and the estimation of rainfall distributions is discussed.

This paper is only available in PDF Format:

  View full text PDF