Model parametrization and data needs

The data crisis in scenario studies embodies topics such as (i) the relevance of existing data for current and forthcoming models; (ii) deciding on the use ofexisting soil data or sampling new data ; (iii) the sensivity of models to several sources of uncertainty in model inputs and consequences for data sampling; (iv) pragmatic approaches to the data crisis; and (v) the presentation of uncertain model results. In all these topics, the magnitude of the data crisis with respect to any parameter is a function of the spatial and temporal variability of this parameter, and of the sensivity the simulation model shows to this variability. The five topics are addressed using case studies and literature rev iews for illustrative purposes. Conclusions drawn with respect to the above topics, are : (i) soil data have merely an identification function when used with conceptual models, but have an estimation function too with implemented models; (ii) existing soil data may not be representing soil bodies unbiasedly, in which case probability sampling in combination with simulation modelling is a proposed method to select representative data based on soil behaviour; (iii) an uncertainty analysis of a model is a useful method to obtain sampling priorities; (iv) Pragmatic solutions to the data cris is exist in the field of exogenous model inputs, model initialisation and obtaining values for process parameters; (v) the presentation of results of scenario studies in terms of probabilities enables the incorporation of the effects of several sources of uncertainty and hence is a powerful method to be further developed.