摘要: This analysis investigates which weather variables are correlated with arriving pilots' convective cell penetration/deviation behavior in the DFW terminal airspace. The report indicates that it is possible to train a statistical c... 展开 This analysis investigates which weather variables are correlated with arriving pilots' convective cell penetration/deviation behavior in the DFW terminal airspace. The report indicates that it is possible to train a statistical classifier that characterizes the probability that pilots will penetrate or deviate around airspace occupied by thunderstorms. The analysis outlines several desirable characteristics of a probability-of-deviation classifier. The classifiers trained and tested in this study were able to correctly classify more than 80 percent of the storm cell encounters in an independent data set. Those classifiers could be used as a starting point for evaluating the utility of probability-of-deviation maps in automated decision aid tool-particularly for those regions of the TRACON more than 20-30 km from the airport. The analysis does not find any correlation between the weather variables and the pilots' penetration/deviation behavior near the destination airports. The vast majority of encounters near the airport in this study resulted in penetrations. Finally, arriving aircraft in this data set were more likely to penetrate storms when the were following another aircraft, more than 15 minutes behind where they ought to be based on the nominal flying time scheduled for the trip, or when they were flying after dark. 收起