Re-visiting the High Temperature and Low Moisture Model Bias in the Central United States

Simulating summer season precipitation and temperature in the central United States has been a tremendous challenge for both global and regional climate models. One outstanding issue is the often reported low bias in precipitation and the high bias in near-surface temperature, not only in coarse-resolution models with parameterized convection but also in high-resolution convection-permitting models. The coexistence of warm-dry biases is seemingly associated with a feedback loop involving convection, soil moisture, solar radiation, surface heat/moisture transport and near-surface temperature. A considerable amount of effort has been expended to understand and mitigate this model deficiency in the research and operational community, but thus far little progress has been made.

Motivated by recent advances in model physics developments and newly available high-resolution reanalysis, we revisited this long-standing problem via multiple sets of seasonal (April-August) convection-permitting simulations with the Weather Research and Forecasting (WRF) model. The simulations consist of (i) 4-km-resolution simulations without convective parameterization, (ii) 4-km-resolution simulations with two different scale-aware convective parameterizations, and (iii) 1-km-resolution simulations. The hourly 31-km-resolution ERA5 data are used to provide initial and boundary conditions for all of the simulations.

The objectives are 1) to evaluate the skill of the convection-permitting model with newly improved physics at simulating summer precipitation and temperature climatology, 2) to assess the possibly added value of increasing resolution from 4 km to 1 km, and 3) to examine the value of scale-aware convective parameterization at convection-permitting grid spacing. Our preliminary results show the use of a new suit of physics improves the WRF performance significantly, with much reduced bias in both daily high-low temperature and precipitation as compared to previously published results. A detailed analysis of the simulations is in progress. The results will be presented in the conference.