How good should a river basin model be to assess the impact of climate change on river flooding for a specific geographical area? The determination of such an appropriate model should reveal which physical processes should be incorporated and which data and mathematical process descriptions should be used at which spatial and temporal scales. A procedure for determining an appropriate model has been developed and applied to the above mentioned specific problem for the Meuse river in France, Belgium and the Netherlands. The model appropriateness procedure consists of three steps. First, the dominant processes and associated key variables are identified. Second, statistical analyses with respect to the key variables are performed, which result in appropriate spatial and temporal scales for each key variable and relationships between key variable scales and the output variable. These latter relationships are used to combine the appropriate scales to one appropriate model scale. In the third step, mathematical process descriptions consistent with these model scales are selected. The resulting appropriate components have been implemented in an existing modelling framework to obtain the appropriate model. Two additional models were constructed to assess the sensitivity of the results to model complexity. The appropriate spatial model scale turned out to be around 10 km with a daily time step. The model results became somewhat better with increasing model complexity. The general trend with climate change (doubling of the CO2 concentration) is a small decrease (5 %) of the average discharge and a small increase (5-10 %) of discharge variability and extreme discharges. It was found that the uncertainties in extreme discharges with climate change are large and that those due to precipitation and extrapolation errors are the most important ones.
|Title of host publication||Integrated assessment ande decision support. Proceedings of IEMSs 2002, Vol. I|
|Editors||A.E. Rizzoli, A.J. Jakeman|
|Place of Publication||Lugano, Switzerland|
|Publisher||University of Lugano|
|Publication status||Published - 24 Jun 2002|