Contribution of soil moisture feedback to hydroclimatic variability

Abstract. While a variety of model experiments and analyses of observations have explored the impact of soil moisture variation on climate, it is not yet clear how large or detectable soil moisture feedback is across spatial and temporal scales. Here, we study the impact of dynamic versus climatological soil moisture in the GISS GCM ModelE (with prescribed sea-surface temperatures) on the variance and on the spatial and temporal correlation scale of hydrologically relevant climate variables (evaporation, precipitation, temperature, cloud cover) over the land surface. We also confirm that synoptic variations in soil moisture have a substantial impact on the mean climate state, because of the nonlinearity of the dependence of evapotranspiration on soil moisture. We find that including dynamic soil moisture increases the interannual variability of seasonal (summer and fall) and annual temperature, precipitation, and cloudiness. Dynamic soil moisture tends to decrease the correlation length scale of seasonal (warm-season) to annual land temperature fluctuations and increase that of precipitation. Dynamic soil moisture increases the persistence of temperature anomalies from spring to summer and from summer to fall, and makes the correlation between land precipitation and temperature fluctuations substantially more negative. Global observation sets that allow determination of the spacetime correlation of variables such as temperature, precipitation, and cloud cover could provide empirical measures of the strength of soil moisture feedback, given that the feedback strength varies widely among models.