ABSTRACT
In the past decade, interest in computer models that can be used for quantitative assessment of the impact of environmental change on runoff, water availability and quality has grown rapidly. These models are expected to produce reliable results for large regions that range from countries to continents. Modelling of hydrological and environmental processes on such scales requires spatially distributed data of differing nature. Data requirements and model structure strongly depend on the purpose of the model and on the manner in which the developers want to describe the relevant processes. From the policy-maker’s point of view, such models should use and produce variables that are related to socio-economic development. If these models are used to assess the impact of environmental change, such as climatic change or land use change, the natural scientist that develops the model prefers to describe the natural processes on a physical basis. The reason is that when limiting natural conditions, such as precipitation or temperature, change, the reliability of results produced by empirical models will be reduced. Apart from our limited knowledge of many relevant processes, it is for practical reasons not possible to get all the data necessary to develop a model that will satisfy all interested parties. Therefore, such models will typically be based on limited data resources, and will use simplified descriptions of complex natural, economic and social processes.
