Optimizing Dropsonde Location to Improve Estimates of a Wind Profile at a Remote Location

Observed wind profiles are often used as proxies for wind profiles nearby in space and time. In this study, we have used various model forecast fields as predictors within a multi-variate linear regression model, optimized to estimate the expected magnitude of the wind error at various directions from the target wind profile. We will discuss how the location of a proxy sounding can be chosen to provide the best regression-based prediction of the wind profile at another location. We will also discuss how knowledge of atmospheric fields between the sounding and another location of interest, such as is available from model forecasts, can be used to better predict the magnitude of the differences between the two locations.