Sensitivities of terrestrial water cycle simulations to the variations of precipitation and air temperature in China

The quality of simulated soil hydrological variables (i.e., soil moisture, evapotranspiration, and runoff) is largely dependent on the accuracy of meteorological forcing data, especially precipitation and air temperature. This issue is quantitatively addressed here by running the Community Land Model (CLM3.5) over China from 1993 to 2002 using the reanalysis-based precipitation and air temperature and in situ observations in the meteorological forcing data set. Compared to the in situ measured soil moisture data, the CLM3.5 simulation can generally capture the spatial and seasonal variations of soil moisture but produces too-wet soil in northeastern and eastern China and too-dry soil in northwestern China. This deficiency is significantly reduced when the in situ measured precipitation data are used to drive the model. An index is also constructed to quantify the sensitivities of soil hydrological variables to variations of precipitation and air temperature. The highest sensitivity of surface hydrological variables to precipitation appears over semiarid regions, while the sensitivity to air temperature for different variables varies regionally (semiarid regions for runoff and soil moisture and humid regions for evapotranspiration (ET)). Over semiarid regions, precipitation and air temperature are equally important to the simulations of soil hydrological variables. Over humid regions, in contrast, ET is more dependent on air temperature than on precipitation, while soil moisture and runoff are less affected by air temperature.