Spatial precipitation and evapotranspiration in the typical steppe of Inner Mongolia, China – A model based approach using MODIS data

We present a simple method to derive spatial precipitation (P) and evapotranspiration (ET) for the typical steppe of the Xilin river catchment at 1 km and 8-day resolution during the main vegetation period (23 April to 28 August) of 2006. The hydrological model BROOK90 was parameterised from eddy covariance measurements. The daily model input data, precipitation, minimum (Tamin) and maximum air temperature (Tamax), were derived by manipulating MODIS leaf area index (LAI) and surface temperature data. P was estimated based on a linear regression of P measured at several sites against the mean gain of the MODIS LAI of surrounding 3 × 3 pixels areas (R2 = 0.76). Tamin and Tamax were derived using a relationship between measured Tamin and Tamax and MODIS surface temperatures (R2 = 0.92 and R2 = 0.88, respectively). The mean precipitation was 145 mm; it varied between 52 mm in the north-western region and 239 mm in the eastern region. In spring, the modelled ET was low (<0.8 mm d−1); evaporation dominated over transpiration and spatial differences were small. At the end of June, the mean ET reached its maximum (2 mm d−1) and spatial differences were pronounced. From July on, transpiration dominated over declining evaporation, and spatial differences decreased in August.

► We model spatial precipitation from leaf area index gain in a typical steppe.
► We model spatial evapotranspiration (ET) with BROOK90 in 1 km, 8-day resolution.
► After model parameterisation no auxiliary ground based data is needed.
► More studies on ET could be conducted in others grasslands applying this method.
► Livestock husbandry of semi-arid grasslands could benefit from this method.