Impacts of Clouds and Convectively Induced Atmospheric Heterogeneities on the Surface Energy Balance in the Energy Exascale Earth System Land Model

Convective processes in the boundary layer, resolved in cloud-scale models, drive the formation of horizontal convective rolls, cells, and cumulus clouds that lead to heterogeneities in the meteorological inputs to land surface models. The effect of these atmospheric heterogeneities (in wind speed, incoming solar radiation, temperature, and specific humidity) on land model prediction is not well understood. In this study the DOE Energy Exascale Earth System Model (E3SM) was used to quantify the impact of atmospheric heterogeneity on modeled surface fluxes and energy balance. The E3SM land model was run with 3 km meteorology (referred to as the experiments) for a variety of cases during summer months over the US Southern Great Plains, including boundary layer rolls, cells, and cumulus cases. These experiments were paired with control experiments having meteorological input coarsened (averaged) to 30 km resolution. The high-resolution meteorological inputs were obtained from separate hindcasts of the 3 km Weather Research and Forecasting (WRF) model. Apart from the expected large impact of variations in incident solar radiation at the surface, wind speed variations emerged as the next most important; wind speed variations enhanced surface fluxes by up to 10 w/m2 per 1 m/s increase in wind speed in the experiments with boundary layer rolls, cells, and clouds. To estimate the aggregate impact that atmospheric heterogeneities could have in global climate models, the predicted surface fluxes from both the experiment and control were averaged to 30 km. Differences in 30 km flux averages are attributed to nonlinear relationships between meteorological variables, primarily incoming solar and wind speed, and surface fluxes. The differences also vary depending on the initial background state of land surface vegetation and soil moisture. Atmospheric heterogeneities cause minimal impacts to latent heat fluxes in areas with limited soil moisture or vegetation stress. The study demonstrates the significance of convectively induced atmospheric heterogeneities for land surface modeling. The lack of inclusion of these atmospheric heterogeneities can at times lead to significant differences – up to 80 W/m² – in modeling of the surface energy balance on scales relevant to climate prediction.