In all past general circulation model (GCM) land surface model components, runoff has not been considered to be very important. It has generally been treated simply as an excess of water that magically disappears from the local water balance. In real catchments, of course, runoff does not disappear but may have an effect on the hydrology and energy balance of areas downslope or downstream. Far more effort, computer time, and parameters have been devoted to formulating the controls on the local energy fluxes of latent and sensible heat than the controls on runoff production. There are several good reasons why runoff production requires more attention, and this lack of attention is now starting to be redressed in the development of so-called macroscale hydrological models
Runoff in many environments is a major part of the water balance and in areas where availability of water is a critical control on latent heat fluxes, then estimating correctly the partitioning of the water balance into that part that is runoff and that part that is available for évapotranspiration may be crucial. This may be a more difficult problem than estimating an areal average évapotranspiration flux since, as is clear from the discussion above, runoff generation has an important spatial dimension due to dependencies on patterns of rainfall inputs, patterns of soil characteristics, and the effects of topography. An important reason why it might be important to improve the runoff generation algorithms in GCMs is that there are long-term discharge measurements available for model evaluation in some catchments at a range of scales and in a wide range of climatic conditions. There is also an increasing recognition that transfers across the grid elements of a GCM, by either surface or regional groundwater flows, may contribute to the controls on patterns of inputs of freshwater to the oceans. At least in areas subject to seasonal flooding (such as the Amazon, Nile, Niger, and other large river basins), such transfers may also control the magnitude of latent heat fluxes over extensive areas.
Macroscale representations of runoff generation are still at an early stage and, given the dependencies on complex spatial heterogeneities and antecedent conditions, it is still not clear as to what an appropriate strategy will be for the formulation and parameter identification of a large-scale model.
Was this article helpful?