Addressing a Warm/Dry Bias over Central North America with Improved Boundary Layer and Land Surface Physics and Data Assimilation

Representing shallow cumulus in numerical weather prediction and climate models is a significant challenge. Misrepresenting these subgrid-scale clouds can result in large errors in the downwelling shortwave radiative flux at surface, resulting in large errors in the surface temperature that results in feedbacks into the accuracy of the thermodynamic fields in the planetary boundary layer. The NOAA hourly-updated Rapid Refresh model (RAP) and its high-resolution cousin (HRRR) have been modified over the last few years to improve the treatment of shallow cumulus and its impact on the downwelling shortwave radiation. Other changes have also contributed to improved accuracy for cloud/boundary-layer evolution, including land-surface treatment for irrigation and refined cloud assimilation. Thermodynamic profiles retrieved from a Atmospheric Emitted Radiance Interferometer (AERI) and horizontal wind profiles from a Doppler lidar, both of which have temporal resolutions better than 5 minutes, are used to evaluate the diurnal evolution of the temperature, humidity and winds in the planetary boundary layer in three versions of the RAP model. The earliest version of the model does not include any treatment for subgrid-scale clouds, and consequently has significant biases in all three fields over the diurnal cycle. The more recent versions of the model have different treatments for subgrid scale clouds, and have smaller biases. The treatment of subgrid scale clouds continues to be a work in progress, and these results demonstrate the impact that has been achieved so far.