Use of soil survey data to improve simulation of water movement in soils.

Soil water retention and hydraulic conductivity functions are crucial input data for models that simulate water movement in soils. When these functions have to be generated for areas of land, the intended application of the results of the modelling determines the level of generalization at which the problem should be addressed. In the research described in this thesis a soil map at a scale corresponding to the identified level of generalization, is used as the basic document from which soil hydraulic functions are derived for an area of land. The hydraulic functions of the major pedological horizons distinguished during the soil survey are measured. Pedological differences do not necessarily correspond with soil hydraulic differences. This results in a limited number of hydraulic "building blocks" with a significantly different soil physical behaviour. Transforming the major pedological horizons into "building blocks" provides the information to transform the soil map into a map of soil physical units.Direct measurement of the hydraulic functions is cumbersome and costly. As an alternative, the existing data base of measured hydraulic functions is analysed and pedotransfer functions are developed. The use of pedotransfer functions is found to be a cost-effective method of translating the basic soil data recorded during soil survey into hydraulic functions.A concept of functional criteria is introduced for evaluating differences in the hydraulic functions measured in the major pedological horizons distinguished and for evaluating different methods of generating these functions. Functional criteria are practical aspects of soil behaviour calculated with the hydraulic functions as input. Hydraulic functions are not an aim in themselves, but serve as input data for simulation models. Therefore, in this study the evaluation of differences in these functions is based on the evaluation of differences in calculated functional criteria and not on a statistical comparison of the functions themselves.Finally, aspects of spatial and temporal variability are investigated. The scaling technique is successfully applied to quantify the complex spatial variability in measured hydraulic functions in a distribution function of scale factors. However, using the results of scaling to calculate the variability in model output results in a conservative estimate of this variability. Model output is also affected by temporal variability in meteorological and water table data. Meteorological and water table data from a 30-year period are used to calculate moisture deficits and trafficability. The influence of temporal variability is reflected by presenting graphs that show the probability of the occurrence of moisture deficits and adequate trafficability