Abstract
Throughout the month of June 1965 tetroons ballasted for 150–300 m altitude were released and radar-tracked in New York City and environs. The study evaluates the skill with which data from both a dense mesoscale network of surface wind observations and a less dense network of balloon-derived wind observations in the planetary boundary layer can be used to reconstruct the tetroon trajectories.
Root-mean-square errors in predicting 2- and 4-hr tetroon positions from surface-wind-derived trajectories are reduced by the addition of a vector correction to account for vertical wind shear; this correction also randomizes the direction of the errors. Corrected surface wind trajectories, when compared with the tetroon trajectories, are slightly better than those computed from the balloon-derived winds.
The best results obtained yield rms 2-hr prediction errors of 15 km; the median error of this distribution expressed as a percentage of the range of each tetroon flight was 36%.
