Can soil moisture be mapped onto the terrain?

Soil moisture heterogeneity has an effect on the rainfall-runoff characteristics of a landscape. The aggregate effect on the mean water balance over an area can be quantified successfully using models such as the PDM (Moore, 1986) and TOPMODEL (Beven and Kirkby, 1979). These rainfall-runoff models have been embedded in the large-scale land surface schemes used in meteorological models. However, there is also a requirement (e.g. model validation) to identify the spatial structure of the fine-scale soil moisture heterogeneity that makes up these aggregate models. In some types of landscape, this will be dictated by topography, in others by soil characteristics, or by a combination of both. A method to distribute area-average soil moisture according to the likely effect of local topography is presented and tested. The heterogeneity of the soil moisture is described by the Xinanxiang distribution (Zhao et al., 1980), commonly used to describe the natural spatial heterogeneity of the landscape. This distribution is then mapped onto the terrain using a topographic index to locate the wettest and driest areas. Soil moisture data from the Wye catchment in Wales and from the Pang catchment in Berkshire, England, are used to test the method. It is found that soil moisture data from the Wye catchment follow the topographic index reasonably well, whereas data from the quick-draining, chalky Pang catchment do not. The conclusion that topographic index is a useful indicator only in some landscapes applies equally to using this mapping method and those models that use topographic index directly.