The Forecast Icing Product was developed, using the output from the Rapid Update Cycle (RUC) numerical weather prediction model, under funding from the Federal Aviation Administration's Aviation Weather Research Program. Using data from research aircraft, the vertical structure of sub-freezing clouds containing ice crystals, supercooled liquid water (SLW), or mixed-phase conditions were examined for the presence or absence of SLD (supercooled large drops). Many of the SLD-producing clouds had multiple SLW layers which were isolated from the boundary layer by thin isothermal layers. These non-boundary layer clouds that formed SLD tended to have low drop concentrations (< 100 cm3), relatively warm cloud layer top temperatures (> -14oC) and modest amounts of SLW (> 0.2 gm-3). When boundary-layer clouds with higher drop concentrations formed SLD, warm cloud layer top temperatures were also present but much higher liquid water contents (> 0.6 gm-3) were required. Additional analyses using National Weather Service soundings and their associated surface precipitation reports were also used to confirm the research aircraft observations.

The SLD algorithm begins by using the FIP three-dimensional cloud field and RUC soundings and condensate. Cloud top temperature for each cloud layer and total condensate within and above the boundary layer are identified. When the cloud layer top temperature and total condensate exceed defined thresholds, a potential for SLD to form is forecast. Once large droplets are formed, they are assumed to grow through collision and coalescence as they fall through lower portions of the cloud.

This paper presents a summary of the background research behind the algorithm, the FIP SLD algorithm in detail and some verification results.