A new method improving the simulation of streamflow in climate models.

Climate models perform poorly in terms of their prediction of streamflow. They tend to deal with the land surface hydrology using either an extremely simplified bucket model or a highly parameterized Soil Vegetation Atmosphere Transfer (SVAT) scheme. Hydrologists, on the other hand, have developed a full spectrum of surface hydrology models, from simple empirical to more complex, physically based schemes, which perform quite well at predicting streamflow. Here a surface hydrology model of moderate complexity (7 parameters) is used in conjunction with a regional climate model to predict streamflow. This approach improves the climate models streamflow prediction considerably, even though much of the discrepancy can be attributed to the prediction of rainfall by the climate model.

Figure 3: Runoff estimates over the FIFE site, 1987-88.