Background
Adverse ceiling and visibility (C&V) conditions create serious flight safety hazards for general aviation, as well as costly reductions in traffic flow efficiency for commercial operations. On a yearly basis, C&V hazards result in ~46 general aviation accidents, ~75 deaths and losses of ~$156M. C&V is second only to convective weather as a cause of U.S. flight delays. RAL research funded by the FAA Aviation Weather Research Program (AWRP) directly addresses both safety and efficiency concerns through the development of improved C&V weather information and decision support systems. The National Ceiling and Visibility (NCV) Product Development Team, comprised of researchers from RAL, NOAA’s Global Systems Division and the Naval Research Laboratory at Monterey (NRL), creates improved national-scale analyses and forecasts and new tools to present this information to pilots, weather briefing staff and others. P. Herzegh leads the NCV team, and G. Wiener heads the engineering effort.
While the NCV effort is focused on enroute safety in the nation's airspace, the Terminal Ceiling and Visibility (TCV) Product Development Team, works to improve safety and efficiency in the airport terminal area. This effort, under the overall leadership of MIT Lincoln Laboratory, develops methods and automated systems to provide detailed forecasts of C&V to help reduce flight delays at major, high-traffic air terminals in the northeast U.S. W. Wilson leads the R&D work at RAL. Though their principal domains of responsibility differ, the NCV and TCV programs collaborate closely.
Current Activities
NCV Analysis Product
Nearest-neighbor interpolation of real-time observations from the ~1700 operational CONUS C&V reporting sites are selectively combined with satellite data and high-resolution terrain data to formulate a continuously-updated analysis of ceiling, visibility, flight category and terrain obscuration. The product is conveyed to test users in real-time in both graphical and gridded form. A similar approach is taken in the development of a prototype Alaska analysis product, which at present is less fully developed than its continental U.S. counterpart.
NCV Forecast Product
This work is built on the premise that five or more independent forecast modules (numerical models, NWS forecaster guidance products and observation-based methods) can be adaptively combined to yield an integrated forecast that is significantly more skillful than any of its input modules. The system has been developed and is being tested. A second effort is focused toward development of a forecast input module based on the use of data mining to establish C&V forecast rulesets tailored to individual sites. This data mining module is in use at fifty-one U.S. sites.
TCV Statistical Forecasting
The goal of TCV statistical forecasting effort is to generate statistical forecast models of near-term changes in low ceilings. Initial efforts have concentrated on the use of surface observations as the predictors. The training archive contains data from 1977-2004. Data quality analysis is a major issue, and a substantial effort has been devoted to a tight data quality analysis, which is conditioned situationally, seasonally, and diurnally. The statistical analysis is based on extensions of the non-linear regression techniques that were used successfully for TCV’s earlier San Francisco Marine Stratus forecasting system. The extensions are designed to enhance the automation of the analysis and to screen the substantially greater number of potential predictors.
C&V Field Studies in the Northeast
Field studies of the physical processes associated with the low ceiling and visibility events are an important complement to the development of the analysis and forecast systems cited above. Toward this end NCV, TCV and the RAL’s Winter Weather project jointly operate an instrumented field site to obtain observational data to support diagnostic case studies focusing on the factors controlling low ceiling and reduced visibility conditions in the northeastern (NE) United States. The C&V field site is located on the premises of the Brookhaven National Laboratory (BNL) in east-central Long Island.
Results/Accomplishments
Overview
Figure 2: Conceptual view of the relationship between the NCV system (red) and the NWS forecast preparation process (blue). NCV analysis and forecast grids flow to NWS as initial fields for the TAF and GFA forecast process. Following forecaster input, modified grids populate the National Digital Forecast Database. Derived warning information flows to end users (green).
In FY05 the National Weather Service (NWS), the Joint Planning and Development Office, and the NCV team determined that NCV’s forecast systems are viable technologies for the operational production of analysis and forecast grids that will flow into the NWS forecast process (Fig. 2). Further, these systems were found to be key elements in meeting the need for additional terminal area forecasts without staffing increases. The NCV grids will populate the aviation parameters within the NWS’ National Digital Forecast Database under forecaster oversight, and will provide forecasters with first-guess guidance at air terminal forecast sites across the CONUS.
NCV Analysis Product
The NCV real-time analysis grids for ceiling, visibility and flight category were shown to be accurate for local conditions and have good reliability in representing the most probable conditions between reporting sites (Fig. 1). Incorporation of high-resolution terrain data has allowed realistic representation of terrain obscuration, a major hazard under low C&V conditions. A concise, localized graphical product can now be formulated for low-cost transmission to the general aviation cockpit to guide a pilot’s hazard avoidance and escape efforts.
NCV Forecast Product
Adaptive selection within the NCV system yields value-added 2-10h gridded forecasts of ceiling, visibility, flight category and terrain obscuration, with skill typically exceeding that of the input forecast modules. The system presents the opportunity to produce a range of forecasts from aggressive to conservative, with the probability of detecting hazardous conditions being traded off against the probability of a false alarm. The NCV-developed data mining forecast module frequently outperforms other modules, adding skill to the overall system.
Operational Readiness
The FAA Quality Assurance Product Development Team assessed the performance of the NCV continental U.S. analysis product in FY05 with favorable results. The product was graduated to “experimental” status, and is ready for a structured operational trial. The NCV gridded ceiling forecast and analysis products will undergo evaluation by NWS Eastern Region for use in aiding forecaster production of NWS aviation grids and terminal forecasts at selected sites.
NCV Forecast System Development
Figure 1. NCV real-time ceiling analysis showing clear conditions (cyan) through the midwest and lowered ceilings in the northeast, northwest and Texas. Significant terrain obscuration (white) is shown in the northwest.
Herzegh, Wiener, M. Tryhane and R. Bateman collaborated in a key redesign of the continental U.S. forecast system methodology in FY05. The revised system incorporates new use of two NWS MOS (model output statistics) products at up to ~1400 sites, RUC model data and persistence at ~1700 sites, and data mining ruleset forecasts at 51 sites. Optimal selection of forecast inputs is based upon recent past performance stratified by forecast module, station, time of day, time of year, forecast lead time and forecast variable. Key supporting functions provide new Java display capabilities, post-analysis replay, and flexible forecast verification tools. Wiener, J. Cowie, A. Braeckel and B. Weekley executed the substantial NCV system re-engineering required to implement the modified forecast and display systems. Early results of the redesign point to a significant improvement in performance. MOS product types, data mining ruleset sites and other features will be further expanded in FY06.
TCV Statistical Forecasting
W. Wilson completed a sophisticated data quality analysis system and applied it to several data sets. The resulting data provide substantially tighter models than were possible using the original data. A new statistical development, synergy analysis, creates composite predictors that have highly improved statistical skill and provides automation of a process that previously was accomplished manually through time-consuming consultation with forecast experts.
Field Studies
Figure 3:Left: Strong inversion capping an marine stratus layer observed at the Brookhaven field site. Right: Weakened inversion and drier air aloft that contributed to the rapid dissipation of the stratus layer.
R. Tardif conducted an analysis of conditions observed during low ceiling events at the BNL site using high-resolution soundings obtained in May 2005. Key factors for a stratus intrusion case were the evolving low-level flow associated with a coastal mesoscale circulation and the absence of a significant moisture contrast between the cloud layer and the lower free atmosphere above cloud top. Advection of drier air aloft and conditions conducive to the entrainment of the dry air across the cloud top interface caused rapid dissipation of stratus (Fig. 3).