Toward Improving Short-Range Fog Prediction in Data-Denied Areas Using the Air Force Weather Agency Mesoscale Ensemble

Abstract

This work develops and tests the viability of a new framework for producing short-range (20 h) probabilistic fog predictions using post-processing of a 4-km, 10-member Weather Research and Forecasting (WRF) ensemble configured to closely match the Air Force Weather Agency Mesoscale Ensemble Forecast System. The raw WRF predictions produce excessive forecasts of zero cloud water, mainly caused by a negative relative humidity bias, which is largely traced to a warm overnight bias. Post-processing mitigates these systematic errors by leveraging traits of a joint parameter space in the predictions to modify individual ensemble members not predicting fog on their own. The method is generally most effective when the space is defined with a moisture parameter and a low-level stability parameter. Cross-validation shows the method adds significant overnight skill to predictions in valley and coastal regions compared to the raw WRF forecasts, with modest skill increases after sunrise. Post-processing does not improve the highly skillful raw WRF predictions at the mountain test sites. Since the framework addresses only systematic WRF deficiencies and identifies parameter pairs with a clear, non-site-specific physical mechanism of predictive usefulness, it is transferable without the need for recalibration, and therefore does not require any observational record to employ.