Initial validation of a Convective Weather Avoidance Model (CWAM) in departure airspace

Abstract

The Convective Weather Avoidance Model (CWAM) translates gridded, deterministic weather observations and forecasts into Weather Avoidance Fields (WAF). The WAF gives the probability, at each point in the grid, that a pilot will choose to deviate around convective weather at that location. CWAM have been developed and validated for en route, high altitude, level flight [1,2], low altitude level flight [3], and for descending arrivals [4]. A heuristic CWAM for departures was also developed and deployed as part of the Route Availability Planning Tool (RAPT) prototype development in New York and Chicago [5,6]. This paper presents an evaluation of the departure CWAM that is currently deployed as part of RAPT, based on an analysis of departure traffic in the Chicago terminal area during convective weather events. The validation of the CWAM in departure airspace is challenging, since it requires the inference of pilot intent from flight track and weather data. This task is particularly difficult in terminal airspace, because of the high degree of flexibility that pilots and air traffic controllers have in vectoring aircraft around storms and through the airspace. We describe our approach in solving 3 problems: • Automation of weather avoidance detection in terminal airspace • Validation of the WAF for departure flights between 20 and 120 km from the airport • Interpretation of results to identify potential sources of error in the CWAM and weather avoidance strategies employed in departure airspace To validate WAF, we automatically classify trajectory avoidances and intersections of WAF contours, determine the WAF probability associated with the weather that is avoided and intersected for each flight, and compare the observed probability of weather avoidance to the avoidance probability predicted by the WAF. The validation is based on the analysis of several test days from the summer of 2010 in the Chicago O'Hare departure airspace (Figure 1). In addition to the calibration of WAF avoidance probabilities, this paper will present an analysis of WAF-based deviation prediction errors, and observed weather avoiding tactics in the departure airspace.