Abstract
The spatial and temporal distribution of erosion and sediment yields is an issue which is of interest to many people including environmental managers and policy makers. As our climate changes we need to be able to forecast which areas will be prone to erosion in order to implement suitable management policies.
Erosion is the result of a complex web of environmental interactions. SEDWEB is a dynamic model which aims to produce scenarios of future erosion and sediment yields. The model combines both deterministic and stochastic approaches. Many processes cannot be modelled in an entirely deterministic manner as systems are generally not closed and parameter values are not known exactly (particularly at the global level when much data is poor or unavailable). In these cases a stochastic approach is adopted for both the uncertainty in parameter values and to generate synthetic climate data.
The model initially grows a vegetation cover that can be supported by the climate of that area. The climate data consists of global rasterized grids at a resolution of 0.5 degrees by 0.5 degrees, of temperature, precipitation, cloud cover (Leemans & Cramer, 1991) and potential evapotranspiration based on the Priestly-Taylor formula (1972). From the monthly rainfall values in this grid daily rainfall distributions are calculated based on a mixed gamma distribution. Vegetation is grown until it reaches equilibrium with this climate at which point runoff is then calculated based on a soil storage threshold. The magnitude of this threshold is affected by the soil amount and the humus produced by the decay of the grown vegetation. This information allows the climatic soil erosion potential (CSEP) index to be calculated (Kirkby & Cox, 1995) as an indicator of the 'natural' erosion potential of an area based on its climate.
The effect of relief on erosion and sediment yields is simulated within SEDWEB using a high resolution global digital elevation model (5' * 5') to calculate distributions of slope and runoff contributing to area. This information can then be combined with the CSEP framework to incorporate the impacts of topography.
As the model has an explicit dependence on climate it is possible to create scenarios of future erosion under changing climate based on GCM output for differing carbon dioxide and sulphide levels. The resulting erosion forecasts are output as information surfaces which can then be used to provide a management tool.
The model code is written in C and embedded within the GRASS GIS. This allows data to be passed between neighbouring cells and the various data surfaces with cell attributes modified as a result of model operations. The advantage of this is that all data is held in one database in a common format accessible to all sections of the model. In the case of global models where a large dataset is needed this is particularly advantageous. A further advantage is the ability to call on GRASS functions for standard procedures such as flow routing and to provide display facilities for the data. Unlike loose or tight coupled models, embedded models allow the modeller to concentrate on the physical processes themselves rather than on the processes of spatial data management.
References
Kirkby, M.J. and Cox, N.J. 1995. "A climatic index for soil erosion potential (CSEP) including seasonal and vegetation factors", Catena, 25 (1-4), 333-352.
Leemans, R. and Cramer, W.P. 1991. The IIASA database for mean monthly values of temperature, precipitation and cloudiness on a global terrestrial grid. RR-91-18 International Institute for Applied Systems Analysis. Austria: Laxenberg.
Priestley, C.H.B. and Taylor, R.J. 1972. "On the assessment of surface heat flux and evaporation using large scale parameters", Monthly Weather Review, 100 (2), 81-92.