Reflecivity produced using standard processing (top) and range oversampling with adaptive pseudowhitening processing (bottom). The reflecitivity from range oversampling processing was produced by using only 50% of the data, which would indicate the potential for faster update times with improved data quality.
We have recently completed the operational implementation of range oversampling on the NWRT PAR.
A real-time implementation of range oversampling is an exciting accomplishment for the NWRT PAR project. But it's even more special for me, since it was the topic of my doctoral dissertation. Back in the late 90s, I developed the foundation for these techniques and evaluated them theoretically and via simulations. A few years later, we managed to modify our research radar to collect range-oversampled data for off-line processing. This was a huge accomplishment, as we were able to validate our theoretical work with real data. However, not until recently has the computational power available to us been sufficient for a real-time implementation of range oversampling processing.
For the past 3 years, we’ve been developing and improving the software/DSP infrastructure of the NWRT PAR that allows seamless expandability to accommodate complex signal processing functions. Also, we’ve been researching ways to improve the performance of the range oversampling techniques and ways to solve non-trivial practical issues in order to fit it with the rest of the signal processing functions. This work will help pave the road for an eventual implementation of range oversampling on the WSR-88D.
Our preliminary look at the data suggests that we’re getting our “free lunch” and a huge pay-off for the invested time and resources. As can be seen in these images, we are able to get lower errors of estimates (i.e., smoother fields) with 50% shorter dwell times. That is, we can reduce all our scan update times in half and at the same time improve the data quality!
We have already modified the scanning strategies used for PARISE 2010 to reduce update times by 50% and are planning to use range oversampling as our default processing mode from now on.
For a tutorial on range oversampling click here.
More recent work is documented in these two AMS Radar Conference papers from 2005 and 2007, these Journal of Oceanic and Atmospheric Technology papers from 2009, 2011, and 2012, and this AMS Annual Meeting paper from 2010.
We recently received a second award from the National Science Foundation (NSF) for our research project "Understanding the Relationship Between Tornadoes and Debris Through Observed and Simulated Radar Data."
This fall, I had the honor and privilege to teach an OLLI class with my friend and colleague Jami Boettcher. "NEXRAD Weather Radar: How it Works and What Those Images Tell Us" kept us busy for 5 weeks this fall.
Our paper "Bootstrap Dual-Polarimetric Spectral Density Estimator" made the cover of the April 2017 issue of the IEEE Transactions on Geoscience and Remote Sensing journal.
I have accepted to serve as an associate editor for the American Meteorological Society’s Journal of Atmospheric and Oceanic Technology.
I have been chosen as the winner of the 2016 OU College of Atmospheric and Geographic Sciences Dean’s Award for Outstanding Service.