Rosenstiel School of Marine and Atmospheric Sciences, University of Miami, Miami, FloridaJialin Lin
NOAA–CIRES Climate Diagnostics Center, University of Colorado, Boulder, Colorado
A simple new analysis method for large single-Doppler radar datasets is presented, using data from several tropical field experiments. A cylindrical grid is chosen, to respect both the geophysical importance of altitude and the radar importance of range and azimuth. Horizontal and temporal fine structure are sacrificed, by compiling data as hourly histograms in 12 × 24 × 36 spatial grid cells of 15° azimuth × 8 km horizontal range × 500 m height, respectively. Mean Doppler radial velocity in each region is automatically unfolded (dealiased) using a simple histogram method, and fed into a velocity–azimuth display (VAD) analysis. The result is a set of hourly horizontal wind and wind divergence profiles, with associated error estimates, for circles of different radii centered on the radar.
These divergence profiles contain useful heating profile information in many weather situations, not just occasional cases of uniform widespread rainfall. Consistency of independent estimates for concentric circles, continuity from hour to hour, and good mass balance indicate high-quality results in one 48-h example sequence shown, from the East Pacific Investigations of Climate (EPIC 2001) experiment. Linear regression of divergence profiles versus reflectivity-estimated surface rain rates is used to illustrate the dominant systematic pattern: convective rain with low-level wind convergence evolves into stratiform rain with middle-level convergence, on a characteristic time scale of several hours. Absolute estimates of moisture convergence per unit of Z–R calculated rainfall vary strongly among experiments, in ways that appear to indicate reflectivity calibration errors. This indicates that Doppler data may offer a useful and unique bulk constraint on rainfall estimation by radar.