BOOK REVIEWS

Abstract

A distributed eco-hydrological model based on soil—vegetation—atmosphere transfer processes is applied to estimate actual evapotranspiration (ET) and gross primary production (GPP) over the Wuding River basin, Loess Plateau, China, based on digital elevation model, vegetation and soil information between 2000 and 2003 over three grid sizes: 250 m, 1 km and 8 km. The spatial patterns of annual ET and GPP are related to precipitation variability and land-use/cover conditions. The grid size is shown to affect the spatial patterns of annual ET and GPP, the effect on GPP being more significant than that on ET. Geostatistical and regression analyses demonstrate that precipitation and vegetation influence the scaling effect of ET and GPP in a complex way. When precipitation is high, the scaling effect of ET is more dependent on precipitation. The scaling effect of ET and GPP from 1-km to 8-km grid size is much larger than that from 250-m to 1-km grid size, showing the 1-km grid size to be a feasible choice for simulation of their spatial patterns. Although the annual GPP is more sensitive to the grid size than annual ET, both daily ET and daily GPP averaged over the whole basin seem to be insensitive to the grid size, illustrating that the coarse grid size can be used to simulate spatially-averaged variables without losing much accuracy.