Climate impacts of parameterizing subgrid variation and partitioning of land surface heat fluxes to the atmosphere with the NCAR CESM1.2

Abstract

All current global climate models (GCMs) utilize only grid-averaged surface heat fluxes to drive the atmosphere, and thus their subgrid horizontal variations and partitioning are absent. This can result in many simulation biases. To address this shortcoming, a novel parameterization scheme considering the subgrid variations of the sensible and latent heat fluxes to the atmosphere and the associated partitioning is developed and implemented into the National Center for Atmospheric Research (NCAR) Climate Earth System Model 1.2 (CESM1.2). Compared to the default model, in addition to the improved boreal summer precipitation simulation over eastern China and the coastal areas of the Bay of Bengal, the long-standing overestimations of precipitation on the southern and eastern margins of the Tibetan Plateau (TP) in most GCMs are alleviated. The improved precipitation simulation on the southern margin of the TP is from suppressed large-scale precipitation, while that on the eastern edge of the TP is due to decreased convective precipitation. Moisture advection is blocked toward the southern edge of the TP, and the anomaly of anticyclonic moisture transport over northern China extends westward, suppressing local convection on the eastern edge of the TP. The altered large-scale circulation in the lower atmosphere resulting from anomalous heating and cooling in the planetary boundary layer is responsible for the change in moisture transport. The performance of other key variables (e.g., surface energy fluxes, clouds and 2 m temperature) is also evaluated thoroughly using the default CESM1.2, the new scheme and the scheme stochastically allocating the subgrid surface heat fluxes to the atmosphere (i.e., without subgrid partitioning included). This study highlights the importance of subgrid surface energy variations and partitioning to the atmosphere in simulating the hydrological and energy cycles in GCMs.