Recent development of an exploratory C&V forecast system for aviation applications in the continental U.S. has utilized expert system methodology to merge numerical and observational inputs in the synthesis of current analyses and forecasts out to nine hours. Its products (posted at www.rap.ucar.edu/projects/cvis/index.html) have yielded encouraging early results and useful insight into needs for future development. This paper describes current and emerging work associated with use of surface and satellite data, application of numerical and observations-based forecast models, forecast synthesis through expert systems techniques, statistical verification, and interactive display techniques in development of the system.
The current system utilizes expert system techniques to manage the dynamically weighted integration of a growing number of analysis, forecast and verification components, as outlined below:
• ASOS reports and GOES imagery are used to produce a U.S. national analysis of current ceiling, visibility and flight category conditions updated four times per hour on a 20 km grid. Broader use of GOES products, NRL cloud classification routines, and NEXRAD data are planned for use in gap-filling and incorporation in other analysis processes.
• The forecast function is currently built around a dynamically weighted combination of forecast fields from the RUC, Eta and persistence. Incorporation of Eta model forecasts, statistical forecasts (utilizing climatology, current observations and NWP model results), and use of the high-resolution COBEL boundary layer column model at selected sites are in development.
• Verification and evaluation of forecast results are logged and available on an ongoing basis through the NOAA/FSL Real-Time Verification System (RTVS) posted at www-ad.fsl.noaa.gov/fvb/rtvs/index.html. Expanded use of ongoing verification results, improvement in the expert system performance feedback loop that helps guide forecast component weighting, and focused diagnostic case studies of critical C&V events will improve fundamental system design as well as capability to examine and track forecast performance.
The paper will highlight current system methodology, analysis and forecast performance, and system development efforts in progress.
Supplementary URL: