Evaluation of WRF Model Performance in Forecasting Supercooled Liquid Water Across the Uinta Mountains of Utah

This study evaluates the ability of a high resolution (2-km horizontal resolution) Weather Research and Forecast (WRF) model driven with North American Mesoscale (NAM) output to simulate cloud liquid water (LW) over the Uinta Mountains of Utah. Radiometer observations of LW (a critical ingredient for effective cloud seeding), are compared with modeled cloud conditions to better understand the ability of an operational WRF configuration to provide insight into impending cloud seeding conditions. The LW was measured on the southern and northern slopes of the Uinta range during the winters of 2019 and 2022, respectively. The WRF model’s performance in forecasting LW was assessed by comparing it with the observed radiometer data. Results indicate that the model accurately predicted LW during major storms but exhibited discrepancies in magnitude, overestimating values on the southern slope and underestimating them on the northern slope. Statistical analysis revealed a bias score of 0.83 for 2019 and 1.06 for 2022, with a probability of detection of 0.74 and 0.70, respectively. This study demonstrates the value of the WRF model for supporting cloud-seeding operations while emphasizing the need for additional radiometer measurements to evaluate LW forecasts under very limited data availability. Because Utah’s complex terrain and varying mountain orientations create strong spatial variability in LW, expanded observations are crucial for validating model output and improving forecasts through better selection of cloud-microphysics and boundary-layer schemes. While the findings provide useful guidance for optimizing precipitation-enhancement strategies in Utah’s mountainous regions, WRF performance must be further assessed using longer-term LW datasets, as current forecasts show both positive and negative biases across the Uinta Range.