Assessing the Impact of Changes in Land Surface Conditions on WRF Predictions in Arid Regions

A thorough evaluation of the Weather Research and Forecasting (WRF) Model is conducted over the United Arab Emirates, for the period September 2017-August 2018. Two simulations are performed: one with the default model settings (control run), and another one (experiment) with an improved representation of soil texture and land use land cover (LULC). The model predictions are evaluated against observations at 35 weather stations, radiosonde profiles at the coastal Abu Dhabi International Airport, and surface fluxes from eddy-covariance measurements at the inland city of Al Ain. It is found that WRF's cold temperature bias, also present in the forcing data and seen almost exclusively at night, is reduced when the surface and soil properties are updated, by as much as 3.5K. This arises from the expansion of the urban areas, and the replacement of loamy regions with sand, which has a higher thermal inertia. However, the model continues to overestimate the strength of the near-surface wind at all stations and seasons, typically by 0.5-1.5ms(-1). It is concluded that the albedo of barren/sparsely vegetated regions in WRF (0.380) is higher than that inferred from eddy-covariance observations (0.340), which can also explain the referred cold bias. At the Abu Dhabi site, even though soil texture and LULC are not changed, there is a small but positive effect on the predicted vertical profiles of temperature, humidity, and horizontal wind speed, mostly between 950 and 750 hPa, possibly because of differences in vertical mixing.