A Comparison of Exhaust Condensation Trail Forecast Algorithms at Low Relative Humidity

Abstract

The Schrader and Schumann contrail forecast algorithms and a third algorithm are evaluated under low relative humidity conditions using a dataset of asynoptic atmospheric soundings and 318 coincident ground-based aircraft and contrail observations collected near Dayton, Ohio. Aircraft were positively identified and their flight altitudes were determined using either Federal Aviation Administration flight logs or data collected with the Wright?Patterson Air Force Base air traffic control radar. The sounding data were used with assumed aircraft performance parameters to prepare nowcasts of contrail critical temperatures. The nowcasts were examined subjectively by comparing the distribution of correct and incorrect forecasts as a function of the difference between critical temperature and ambient temperature. Objective evaluation against the contrail observations also was done using several common statistical measures. The third algorithm produced critical temperatures that were systematically too cold, leading to forecasts with low skill in comparison with unbiased random forecasts. The Schumann and Schrader algorithms were skillful in comparison with unbiased random forecasts and performed similarly to one another. The distribution of critical temperature deviations from ambient temperature for incorrect forecasts made with the Schumann and Schrader algorithms (but not the third algorithm) can be explained by potential errors in the input data. All results were statistically significant.