Improved moisture and PBL initialization in the RUC using METAR data

Hourly surface observations provided in meteorological aviation reports (METARs) are critical for nowcasting by severe weather and aviation forecasts. The Rapid Update Cycle (RUC) uses these observations more than other operational NWP model, but still does not yet assimilate many of the METAR parameters considered most important by forecasters. A considerable effort has been underway at FSL since early 2003 to improve this aspect of the RUC analysis, and we have gained confidence that METAR observations can likely be used much more fully than is currently done even in the operational RUC.

The vertical correlation of surface parameter errors has long been a dilemma for assimilation of surface observations. While surface observations are relatively dense, there are almost no high-frequency observations of the vertical profile above the surface. Thus, a technique has been developed to use the background planetary boundary layer depth to provide an estimate of the depth over which the observation-minus-background values of temperature, moisture, and wind are likely to be applicable. This technique has been in real-time testing of the RUC for over a year, showing an overall improvement in temperature and dewpoint forecast statistical accuracy, and improvements in convective available ppotential energy (CAPE) forecasts for convection forecasts. This was shown most notably for the 10 November 2002 tornado outbreak case. Further revisions are currently in testing to improve the robustness of this technique. This improvement is important given the sensitivity of the CAPE estimate to small changes in moisture and temperature profiles.

Results will be shown both for overall statistical improvement and important case studies using this PBL-based technique for assimilation of METAR observations.