Improvements in modeling California's Tule Fog

Despite our increasing understanding of the relevant physical processes, forecasting topographically confined cold pools and their associated meteorological phenomena (e.g., fog and freezing rain) remains a challenging problem. In particular, the difficulty of modeling the Tule fog of California's Central Valley led us to question assumptions inherent in Planetary Boundary Layer (PBL) parameterizations in the WRF model, specifically the popular Yonsei University (YSU) scheme. It is known that this PBL treatment approximates entrainment – an important fog maintenance process – via the surface fluxes (among other things). While this assumption may work well for non-foggy situations, in the foggy PBL the surface heat source is normally much smaller than the radiative flux divergence at the top of a fog layer, which is typically ignored in the entrainment estimation. Furthermore, the prescribed YSU diffusivity profile was specifically designed for a surface heat source.

As an alternative, we propose that a reversed diffusivity profile may be more appropriate in cases of fog. We have modified the YSU PBL accordingly to detect and handle foggy boundary layers, with the goal of simulating fog events more realistically without damaging model performance in other areas and times. Thus far results have been promising, showing drastic improvements in surface visibility forecasts for Tule fog but also subtle improvements in the marine boundary layer.