Dependence on land surface model of the time scale of soil moisture variability

Most of the studies on soil moisture variability and its relationship to seasonal surface climate are based on the use of numerical models. The impact of the formulation for the land surface processes on the results from these studies, however, is not well addressed. In this paper, we investigate the dependence of the time scale of soil moisture variability and its relationship to surface air temperature on the treatment of land surface processes. We compare simulations by four atmospheric general circulation models (AGCMs) that use land surface models of different degrees of complexity. We focus on the simulation of soil moisture and surface air temperature in northern hemisphere midlatitudes (30°N-50°N) during warm seasons. It is found that the simulated time scale of soil moisture is substantially different among the AGCM simulations, with the average e-folding time scale ranging from 1.5 months to 5 months. In addition, the magnitude of the simultaneous correlation between soil moisture and surface air temperature also differs. Additional analysis suggests that the difference among the AGCMs in the characteristic time scale of soil moisture anomalies and their impact on the surface air temperature is mainly due to the respective model’s soil moisture field capacity, with other details in the parameterization scheme playing a secondary role.