Towards Improving Representation of Sub-Grid-Scale Clouds in WRF and Application to Aircraft Icing Forecasts

While it appears that the new combination of RAP/HRRR model and Thompson et al (2008) microphysics produces a superior prediction of explicit supercooled liquid water (SLW; i.e., aircraft icing) than the older RUC model and Thompson et al (2004) microphysics, there remains a negative bias in the prediction of icing clouds.  This is of great importance to the FAA-funded Terminal Area Icing Weather Information for NextGen (TAIWIN) project.  Recent research by Cintineo et al (2014) reveals that none of four tested WRF microphysics schemes (Thompson, Morrison, Milbrandt, WSM6) properly captured sufficient clouds with temperatures that are very frequently found to contain icing between zero and ‑20°C.  The fact that four different schemes all have moderately large negative bias of this type of cloud system leads us to believe the root cause is not the microphysics scheme alone.  Rather, the root cause could be a combination of factors including model resolution, turbulent diffusion, vertical mixing, and treatment of clouds at scales smaller than model spacing.

To address this under-prediction of clouds in WRF, a new application of a Sundqvist et al (1989) cloud fraction scheme has been created.  The new treatment produces superior skill with regard to solar radiation reaching the ground, primarily due to the increased representation of missing clouds.  Also, by accounting for partially cloudy grid boxes, this leads directly to more correct radiation fluxes that promote explicit cloud growth by the microphysics scheme (due to longwave radiation cooling of the partly cloudy layers and altered solar energy passing through clear skies to the ground).  The new treatment also has promise as a simple post-processor for application to direct icing forecasts as a supplement to a weather model's explicit supercooled liquid water content, particularly in intermittent cloud conditions.  Results of the new cloud fraction scheme applied to HRRR-model forecasts and verified against icing pilot reports (PIREPs) will be shown.

This research is in response to requirements and funding by the Federal Aviation Administration (FAA).  The views expressed are those of the authors and do not necessarily represent the official policy or position of the FAA.