Spatial modelling of soil water drainage rates : a case study validated on a small catchment
PublisherUniversity of Reading
MetadataShow full item record
Traditionally, soil scientists have ignored the spatial dimension concentrating upon the physics of soil water movement at points/sites. The development of geographic information systems (GIS) provides an opportunity to develop soil-water models taking into account differences (spatially and temporally) in land cover and soil type. This thesis has attempted to integrate traditional soil water modelling within a GIS for estimating soil water drainage rates for a typical water catchment with spatially varied soil type and land cover. The Campbell soil model has been adopted for its simplicity and accuracy for estimating soil water infiltration/distribution. However, Campbell's model fails to take into account sufficiently the role of vegetation in infiltration/distribution. The Meteorological Office Rainfall and Evaporation Calculation System provides effective ways of deriving potential evapotranspiration under various vegetated surfaces. The integration of these two models has resulted in the development of the Soil Water Drainage Model which is more accurate and capable of incorporating spatial soil and vegetation data. The model was developed and tested for the south-west Reading study area. The model's behaviour under various conditions was examined and its performance evaluated. It has found that a bare soil surface has a significant water drainage difference from a vegetated one and soil type affects soil drainage rate, especially when the soil is bare or abnormal climatic events occur. It has also shown that grassland reduces greatly the quantity of soil water loss as drainage and most drainage occurred in winter months and early spring. The accuracy of the model's prediction for estimating soil water drainage was validated on the small Winterbourne catchment (45km2) near Newbury, based on the catchment water mass equilibrium. The water drainage rates predicted by the model are reliable and consistent with the observed river discharge data. However, the results suggest a number of ways in which the methodology could be improved.