Real Time Aviation Forecasting at WSI
On the numerical modeling side, the WRF model is run in-house over several domains at different resolutions for different forecast periods four to eight times per day. A 36 km grid is run over the North Atlantic Ocean (ATLANTIC) four times daily, a 14 km grid is run over Europe (EUROPE) four times daily, and a 12 km grid is run over the Continental US (CONUS) eight times daily. All the forecasts are run on a cluster of 40 IBM dual processor 3.06 and 3.20 GHz machines. A 48 hour forecast over the CONUS takes 100 minutes. In contrast, a 24 hour forecast over the ATLANTIC (including 1/3 of the eastern US) takes 15 minutes. The model output is evaluated by forecasters at WSI and it is also used as input to a suite of post-processing algorithms that generate forecasts for severe weather, ceiling height and visibility, and atmospheric turbulence.
Gridded ceiling and visibility forecasts are generated using output from the WRF model as input to existing ceiling and visibility algorithms. The algorithms include the Stoelinga-Warner algorithm that relies on light attenuation from various hydrometeor types and an FSL algorithm that relies on relative humidity and dew-point depression.
Gridded turbulence forecasts are generated from WRF output using an ensemble of turbulence forecasting algorithms developed at WSI. The 24 hr turbulence forecasts from the ATLANTIC forecasts are used primarily to support commercial trans-Atlantic flights over a region where high-resolution turbulence forecasts are not otherwise available. The ensemble consists of modified versions of five existing popular turbulence algorithms plus their mean plus their maximum. Despite the simplicity of this first-generation turbulence forecasting system, it provides considerable skill over random guessing and compares favorably to the considerably more sophisticated Graphical Turbulence Guidance (GTG) forecasts based off the RUC model.
Finally, a variety of terminal centric forecasts, including TAFS, are created for both domestic and international locations at several WSI facilities. Production of these forecasts utilize a combination of sophisticated automation techniques combined with efficient human editing and rely extensively on real time verification and monitoring.
More detailed description of the WSI real-time aviation forecasting activities as well as performance statistics as well as plans for future activities will be discussed.