Abstract for presentation at Spatial Sciences Institute International Biennial Conference

Hydrological modelling is a domain within which considerable uncertainty is inherent. This is caused by a number of fairly typical factors: sparse calibration data, imperfect knowledge of the processes being modelled, and high variability in natural phenomena such as climate. Somewhat surprisingly, however, the study of uncertainty in hydrological models is not well advanced. Moreover, the uncertainty that is generally referred to by those in the hydrological modelling community is associated with model structure and/or calibration coefficients -- little consideration is given to the quality of the input data and the impacts that these may have on model outputs. Of particular interest in this paper are inputs that are spatial in nature such as the size of a catchment, soil type, and landcover. Related to all of this is a lack of consideration about how to communicate any of this uncertainty to consumers of model results.
In this paper we will provide a general overview of hydrological models – the goals of the models, the inputs required, the scale at which they operate, and their internal functions. We will then identify typical inputs that are spatially based and explain how uncertainty on them can be described and communicated to users with an emphasis on advanced visualisation techniques. We will then apply simulation analysis to a catchment scale hydrologic model (2Csalt) to demonstrate the sensitivity of the model to spatial uncertainty and the consequences of not considering spatial uncertainty. Finally, we will suggest means by which the uncertainty of data outputs can be presented visually in order to improve understanding of the implications of spatial uncertainty on land management and policy decisions.