LARGE-SCALE DISTRIBUTION MODELING AND THE UTILITY OF DETAILED GROUND DATA

A large-scale distribution function model was used to investigate the effect of differing parameter mapping schemes on the quality of hydrol ogical predictions. Precipitation was mapped over a large forested cat chment area (163 km(2)) using both one-dimensional linear and three-di mensional non-linear interpolation schemes. Lumped stream flow predict ions were found to be particularly sensitive to the different precipit ation maps, with the three-dimensional map predicting 12% higher mean annual precipitation, resulting in 36% higher modelled stream flow ove r a three-year period. However, spatial predictions of stream flow app eared worse when derived from the three-dimensional map, which is cons idered the better of the two precipitation maps. This implies uncertai nty in either the model's response to precipitation or the precipitati on mapping process (the 12% precipitation difference was strongly dete rmined by a single, short term gauge). Leaf area index (LAI) was mappe d using both remote sensing and species based methods. The two LAI map s had similar lumped mean values but exhibited significant spatial dif ferences. The resulting lumped predictions of stream flow did not vary . This suggests a linear response of water balance to LAI in the non-w ater-limited conditions of the study area, and de-emphasizes the impor tance of quantifying relative spatial variations in LAI. Topographic m aps were created for a small experimental subcatchment(15 ha) using bo th air photographic interpretation and ground survey. The two maps dif fered markedly and lead to significantly different spatial predictions of runoff generation, but nearly identical predicted hydrographs. Thu s, at scales of small basins, accurate topographic mapping is suggeste d to be of little importance in distribution function modelling becaus e models are unable to make use of complex spatial data.Predictions of water yield can be very sensitive (in the case of precipitation) or i nsensitive (in the case of small-scale topography) to changes in spati al parameterization. In either case, increased complexity in spatial p arameterization does not necessarily result in better, or more certain prediction of hydrological response. (C) 1998 John Wiley & Sons, Ltd.