气候变化领域集成服务门户

Vegetation is known to have strong influence on evapotranspiration (ET), a major component of terrestrial water balance. Yet hydrological models often describe ET by methods unable to include the variability of vegetation characteristics in their predictions. To take advantage of the increasing availability of high-resolution open GIS data on land use, vegetation and soil characteristics in the boreal zone, a modular, spatially distributed model for predicting ET and other hydrological processes from grid cell to catchment level is presented and validated. An improved approach to upscale stomatal conductance to canopy scale using information on plant type (conifer/deciduous) and stand leaf-area index (LAI) is proposed by coupling a common leaf-scale stomatal conductance model with a simple canopy radiation transfer scheme. Further, a generic parametrization for vegetation-related hydrological processes for Nordic boreal forests is derived based on literature and data from a boreal FluxNet site. With the generic parametrization, the model was shown to reproduce daily ET measured using an eddy-covariance technique well at 10 conifer-dominated Nordic forests whose LAI ranged from 0.2 to 6.8 m(2) m(-2). Topography, soil and vegetation properties at 21 small boreal headwater catchments in Finland were derived from open GIS data at 16 m x 16 m grid size to upscale water balance from stand to catchment level. The predictions of annual ET and specific discharge were successful in all catchments, located from 60 to 68 degrees N, and daily discharge was also reasonably well predicted by calibrating only one parameter against discharge measurements. The role of vegetation heterogeneity in soil moisture and partitioning of ET was demonstrated. The proposed framework can support, for example, forest trafficability forecasting and predicting impacts of climate change and forest management on stand and catchment water balance. With appropriate parametrization it can be generalized outside the boreal coniferous forests.