Convective Weather Forecasting: Aviation Weather Research and Development
Figure 1. NCWF6 images generated with the NCWF Viewer depicting a) current convection (hot colors) with 2 hour performance polygon (pink contour), b) base interest with 1 hour probability forecast (purple shades) and c) base interest with 6 hour probability forecast.
The research and development efforts of RAL's Convective Weather Group are aimed at improving short-term (0 – 6 hour) thunderstorm forecasting and bridging the gap in skill between observation-driven expert systems and numerical weather prediction. This work is particularly relevant to aviation; a reliable short-term prediction of the likelihood of convection and its characteristics (e.g., location, extent, organization and vertical development) is critical in the strategic planning of air traffic. To this end, the FAA Aviation Weather Research Program has long providing support for the development of a new automated forecast system, which generates 1 – 6 hours probabilistic forecasts of convection that is potentially hazardous to aviation.
FY06 Accomplishments:
A new system, called the National Convective Weather Forecast product (NCWF6), was completed in March 2006 and underwent extensive testing and enhancement throughout the summer. The NCWF6 was designed to blend observation-based probabilistic forecasts (based on observed storm motion and trend information) with recalibrated probabilistic forecasts of convective precipitation derived from Rapid Update Cycle (RUC) model output. The blending is currently done using a weighted average, where the weights are assigned based on past statistical performance of the two individual systems and are a function of the forecast lead time. The resulting forecasts can be displayed in real-time and evaluated in an operational setting (Figure 1). The data are also being made available to research groups that are developing automated decision support tools requiring probabilistic forecasts. Verification statistics are generated in real-time and early analysis indicates that the merged product has a better Critical Success Index (CSI) score than the individual components at lead times of 4 hours (Figure 2), but with biases that are still unsatisfactorily high (not shown). New techniques for reducing this bias are currently under development
FY07 Plans:
Figure 2. CSI skill scores for August 2006 as a function of forecast lead-time for observation-based forecasts (red), calibrated RUC forecasts (green) and merged forecasts (blue) for the 0.05 probability level.
Efforts will focus on development of a new forecast product called “Consolidated Storm Prediction for Aviation”. This new tool will combine the best aspects of the 16 existing products that currently serve the aviation community. RAL staff will collaborate closely with sponsors and developers of the current products in creating an appropriate infrastructure to support consolidation of various components of those products. Members of the aviation community will participate in the initial design review to be held in October, and their input will be captured in a plan to the FAA detailing specific FY07 activities. Long-term plans for the consolidated product must also be responsive to the Joint Planning and Development Office's vision for the Next Generation Air Transport System. Initial efforts at RAL will focus on achieving by 2012 an initial operating capability over the CONUS which provides forecasts out to 4 hours with updates every 15 minutes. The goal for 2025 is to create a storm analysis and forecasting capability that will be global and provide forecasts out to 24 hours.