Soil moisture conditions directly influence the net surface energy balance and determine the partitioning of the surface heat flux into sensible and latent contributions, which in turn control the evolution of the soil moisture distribution. There have been studies of the importance of soil moisture anomalies for episodes of drought (Atlas and Wolfson, 1993) and flooding (Beljaars et al., 1996; Giorgi et al., 1996), and the impact of initial soil moisture conditions on mid-latitude weather (Betts et al., 1996; Schär et al., 1999). Results from other GCM studies (e.g., Milly and Dunne, 1994; Bonan, 1996; Ducharne et al., 1996) and regional and global water budgets analyses (e.g., Brubaker et al., 1993; Brubaker and Entekhabi, 1996) have deepened our appreciation of the importance of land-surface hydrology in the regional and global energy and water exchanges. In relation to climate change, such mechanisms are relevant since they might lead to, or intensify, a reduction in summer soil moisture in mid-and high latitude semi-arid regions under doubled CO₂ conditions (Wetherald and Manabe, 1999). Most of these studies reported some impact of soil conditions upon land precipitation during episodes of convective activity, and there is observational evidence from lagged correlation analysis between soil moisture conditions and subsequent precipitation over Illinois that this mechanism is active in mid-latitudes (Findell and Eltahir, 1997). The formulation of surface runoff and baseflow has been calculated to have an indirect but strong impact on the surface energy balance (Koster and Milly, 1997).

The feedback mechanisms between soil moisture conditions and precipitation are particularly relevant to climate change studies since they may interact with, and determine the response to, larger-scale changes in atmospheric circulation, precipitation and soil moisture anomalies. The modelling of soil moisture-climate interactions is complicated by the range of time-scales involved, as soil moisture profiles can have a �memory� of many months, and the interaction of vertical soil moisture transfers with the larger-scale horizontal hydrology. Work is continuing to improve the realism of vertical water transfers, the effect of soil water on evapotranspiration rates, and the parametrization of sub-grid scale variability in land hydrological components (e.g., Avissar and Schmidt, 1998; Wood et al., 1998). To date, there have been few attempts to describe the effects of within-grid horizontal transfers of water, but there has been success in connecting river routing schemes to GCMs (Dümenil et al., 1997; see also Chapter 8, Section 8.5.4.2). Development in this area has lagged significantly behind that of vegetation canopy processes, despite the fact that the former are critical to a land-surface scheme�s overall performance.

IPCC Homepage