Terminal Aerodrome Forecast Verification

Terminal aerodrome forecasts (TAFs) are essential aviation weather forecast tools.  Issued every 6 to 8 hours, they provide aviators with a 24- to 30-hour forecast of wind direction and speed, surface visibility, types of visibility obscuring phenomena, and cloud layer coverage and heights.  The U.S. Air Force operates a number of bases within the continental United States (CONUS) that are in close proximity (within several miles) of major civilian airports.  TAFs for the civilian airports are issued by National Weather Service (NWS) Weather Forecast Offices (WFOs), which employ forecasters who generally have considerable local forecasting experience, while TAFs for Air Force bases are issued by one of three U.S. Air Force Operational Weather Squadrons (OWS), which employ relatively inexperienced forecasters.  As numerical weather prediction (NWP) model performance steadily improves, it is useful to objectively quantify the value added by human weather forecasters in TAF production.  In this study, we analyze forecast accuracy for different TAF forecast hours at four pairs of Air Force bases and civlian airports in close proximity to each other:  McChord Field and Olympia Regional Airport, both near Tacoma, Washington; Offutt Air Force Base (AFB) and Eppley Airfield, both near Omaha, Nebraska; Wright-Patterson AFB and James M. Cox Dayton International Airport, both near Dayton, Ohio; and MacDill AFB and Tampa International Airport, both near Tampa, Florida.  Metrics compared include categorical errors in cloud ceiling height, surface visibility, and crosswind component, as well as precipitation type.  Skill scores for TAFs are computed against persistence, climatology, and NWP output from the National Centers for Environmental Prediction (NCEP) 5-km resolution North American Model (NAM) CONUS nest model.