The Accuracy of Vertical Air Velocities from Doppler Radar Data

Abstract

Eight methods of calculating vertical air velocity in a column are compared. Each method requires some or all of the following data: horizontal divergence, vertical precipitation velocity, hydrometeor terminal fall speed, and vertical air velocity boundary conditions. Some or all of these quantities are commonly deduced or specified by a researcher during the analysis of Doppler radar data. The responses of the methods to different magnitudes and behaviors of errors in the input data are examined with a Monte Carlo method. The experiments are conducted with both random and systematic errors. Two idealized kinds of systematic errors are considered: bias error (a constant error through the column) and trend error (an error of constant magnitude that changes sign at the midpoint of the mass of the column). The performances of the methods are mapped over error space. A researcher, knowing approximately the characteristics of the errors of a particular set of Doppler radar data, can use the results of these experiments to select a method based on how well it performs in the presence of random errors and how well its performance holds up as bias and trend errors increase.