Statistical Determination of Vertical Resolution Requirements for Real-Time Wake-Vortex Prediction

Wake-vortex interaction is one of the primary drivers for aircraft separation minimums. Real-time wake-vortex prediction algorithms could enable following distances to be reduced, improving the efficiency of the air transport system. This study uses statistical analysis to determine the necessary vertical resolution of meteorological parameters required to enable real-time wake-vortex models to predict wake-vortex position and strength using two separate analyses. The first analysis sought to characterize the accuracy of the predictions of real-time wake-vortex models for 1 m (highest) and 100 m (lowest) resolution by comparing the predictions to measured wake vortices. Using NASA's AVOSS Prediction Algorithm Suite v4.19 and two data sets collected at Memphis and Dallas-Fort Worth, the results indicate that the range of vortex circulation root-mean-square (RMS) error is [1.5-32.6] m(2)/s, the range of vortex Y-position RMS error is [0.7-134.5] in, and the range of vortex Z-position error is [1.2-10.1] m. The Y-position error had the largest variation with the maximum deviation exceeding 25 m. The second analysis sought to characterize the precision of the real-time wake-vortex model predictions as the resolution of the input meteorological files was reduced from 1 m to 100 m. Normalized RMS values (by 1 m resolution) were calculated for resolutions of 5, 15, 20, 25, 30, 40, and 100 in averaged over the first 60 s of data available. For predicted circulation, differences were found between medians at input resolutions 5-20 m vs 25-100 m; for V- and Z-position, differences were found between medians 100 m and all others.