Effects of spatial variability on the rainfall runoff process in a small loess catchment

Numerous field experiments have revealed that hydrological processes and parameters can show considerable spatial variability. When modelling the rainfall runoff process, the influence of this variability on the runoff should be assessed. For regionalization purposes, it is an important question if and how this variability is reflected in the behaviour on a larger scale. A quasi-three-dimensional, process-oriented model is used to understand the influence of spatial variability on the runoff behaviour. The study area, "Neuenburger Pfad", is a small loess catchment in southwestern Germany. Because of low infiltration capacity of the loess soil, runoff is produced by infiltration excess overland flow. Three examples are presented which elaborate on the effects of different types of spatial variability. The first example shows that an agricultural road plays a crucial role in the runoff behaviour of the investigated catchment. The second example studies the role of the spatial distribution of soil parameters. It demonstrates the importance of structured variability in heterogeneous fields. Purely random fields result in hydrographs similar to the homogeneous case, whereas fields with structured variability yield very different runoff. The third example is concerned with the effects of different soil moisture interpolations, which serve as initial state for the runoff model. The examples presented show that spatial variability can have a dominant influence on the rainfall runoff behaviour. For sound modelling, a close inspection of the particular catchment characteristics is necessary. Structured variability should be included in hydrological models because of the larger influence compared to stochastic variability. To understand the extent and the significance of catchment variability, an interdiciplinary teamwork should be fruitful. (C) 1998 Published by Elsevier Science B.V. All rights reserved.