Sensitivity testing of a model for exploring water resources utilisation and management options

This paper investigates the model sensitivities to input parameter values in a Biophysical Toolbox for integrated catchment assessment and management of land and water resources. The toolbox was developed for application in the highland regions of northern Thailand. It incorporates the IHACRES rainfall-runoff model, a crop model (CATCHCROP), and an erosion model (a Universal Soil Loss Equation [USLE], modified for application to northern Thailand). In developing the individual models in the Biophysical Toolbox, emphasis was placed on limiting model complexity. The toolbox was developed and tested using data from the Mae Chaem catchment in northern Thailand. Due to the short duration and sparse distribution of the available data, the complexity of the model structure is constrained to consider only the key processes of interest. Despite the relative simplicity of the individual models, linkages between the models encompassed in the Biophysical Toolbox increase the complexity of the modelling system. This paper explores sensitivities in the Biophysical Toolbox to the parameters of CATCHCROP as this component has the greatest potential for propagating errors though the Toolbox. A simple sensitivity analysis was undertaken, whereby parameter values were perturbed to ascertain the effect of these perturbations on output indicators. The hydrologic component showed strong non-linearity to the infiltration parameters of the CATCHCROP model, although this did not greatly impact estimates of total annual discharge. Additionally, a number of CATCHCROP parameters did not greatly impact some output indicators of the toolbox. While the CATCHCROP model is relatively parsimonious when compared with many crop models, there are still 19 model parameters that must be determined from the field or prescribed from the literature. There is potential for the model to be further simplified although more extensive model testing is required to ensure that such simplifications would not adversely impact on the utility of the toolbox.