Alleviation of Low Humidity Biases in WRF-ARW and UFS-SRW Model Simulations over Complex Terrain

A low-level low humidity bias typically results in WRF-ARW, GFS-FV3, and FV3-LAM simulations over regions of complex, mountainous terrain. This bias often has secondary effects that increase errors of forecast temperature and precipitation. We explore the nature and probable causes of this dry bias. We note that the dry bias is particularly acute outside of the United States where conventional observations are less plentiful. We then explore various methods for alleviating the dry bias in the Tomorrow.io Comprehensive Bespoke Atmospheric Model (CBAM) framework. The implemented methods include the adaptation of the Gridpoint Statistical Interpolation (GSI) soil nudging. This method, employed for the HRRR and RAP operational forecasts, adjusts the soil temperature and moisture based on increments of the lowest-level temperature and water vapor fields. Notable changes occur in the soil fields that produce additional adjustments in the low-level humidity and temperature forecast fields.