Understanding land-surface influences on moist convection using observations and climate model experiments

 

Authors

Ian N. Williams — Lawrence Berkeley National Laboratory
Margaret S. Torn — Lawrence Berkeley National Laboratory

Category

Boundary layer structure, including land-atmosphere interactions and turbulence

Description

Surface energy partitioning often plays an important role in conditioning the atmosphere for deep convection. Soil moisture and vegetation state can influence this partitioning, and thus may contribute to convective triggering. We explored this process using observations from the U.S. Southern Great Plains (SGP) Atmospheric Radiation Measurement (ARM) Climate Research Facility site, and a single-column NCAR Community Earth System Model (CESM). We compared simulations having default land-model parameters to those having observationally constrained parameters with increased stomatal conductance and soil resistance to evaporation. The constrained parameters increased transpiration, reduced soil evaporation, and led to higher latent heat flux at times of large-scale moisture divergence. This resulted in a shallower boundary layer (PBL) and lower cloud base height during summer drought. The default land-model produced a PBL that was drier and more stable to moist convection than did the modified land model, with fewer diurnal deep convection events compared to either the modified land-model experiment or the observations. The model results suggest that transpiration helps trigger moist convection by moistening the PBL and lowering the lifted condensation level (LCL). Predictions of LCL, cloud base height, convective available potential energy (CAPE), and precipitation were improved in the experiment compared to the default model, indicating that summer precipitation prediction can be improved by correcting the contribution of transpiration to evapotranspiration in some land models.