Automated Detection of Boundary Layer Depth Using Dual-Polarization Radar Observations

Dual-polarization WSR-88D radars have been running operationally throughout the U.S. since 2013, providing us with a wealth of information about objects in the atmosphere. The advent of this technology has revolutionized our uses of radar data by helping us identify hailstones, better differentiate snow vs. rain, and even track bird migration. As it turns out, dual-polarization WSR-88D radar data can also provide information on the height of the convective boundary layer (CBL), which is useful for predicting severe weather, pollution concentrations, and precipitation amounts. Traditionally, boundary layer depth has been observed using rawinsonde data from balloon launches twice daily (0000 and 1200 UTC) at 97 locations in the U.S. Since only one of these launches occurs during daylight hours, there is only one measurement of the CBL per day. With 159 WSR-88D radars scanning the atmosphere every 5-10 minutes, we can obtain estimates of the CBL depth at spatial and temporal resolutions never before possible.

Differential reflectivity (ZDR) has a local minimum at the top of the CBL due to Bragg scatter caused by small-scale, isotropic turbulence (ZDR values near zero) combined with the scatter from biota (large ZDR values). To automatically detect the local minimum in ZDR we calculate quasi-vertical profiles (QVPs) of ZDR. Radar observations at each range gate are averaged azimuthally over 360° and converted to height above ground level to create a height profile of the mean ZDR. Then, to find the local minima at each observation time, a 1-D continuous wavelet transform (CWT) using a Ricker wavelet is performed on each QVP of ZDR. The Ricker wavelet is advantageous because it can detect the peaks and troughs of a noisy signal. Then, a Kalman Filter is applied to the CWT-identified local minima and a CBL depth estimate based on ZDR multiplied by ZDR variance to find the true location of the CBL top. Results indicate that a 1-D CWT followed by the Kalman filter works well for automatically detecting the CBL height from QVPs of ZDR for a variety of days and conditions. The outcome of this work is to use dual-polarization WSR-88D data to provide regular observations of CBL depth and to develop a climatology of CBL depths at 50 radar locations for the years 2014 and 2022.