Research Tools: NEXRAD
Research to Operations
NSSL scientists helped develop the Weather Surveillance Radar - 1988 Doppler (WSR-88D) radars, also known as NEXt-generation RADar (NEXRAD). Since the first Doppler weather radar became operational in Norman in 1974, NSSL has worked to extend its functionality and effectiveness, and proved to the NOAA National Weather Service (NWS) that Doppler weather radar was a crucial forecasting tool. The NWS now has a network of 158 NEXRADs.
NSSL continues to develop and improve algorithms that detect and notify forecasters of hail, severe thunderstorms, tornadic circulations, downbursts and gust fronts. Research to operations activities also include developing and improving signal processing techniques.
Clean AP
Description of work to be performed: Leveraging on work done for the MPAR project, develop ground clutter mitigation techniques for the dual-polarimetric WSR-88D RDA. For stationary ground clutter (e.g., terrain, buildings), we propose to combine detection and filtering into a single algorithm (CLEAN-AP) that has been endorsed by the TAC because it outperforms the current WSR‑88D filtering (GMAP) and detection (CMD) schemes. Taking advantage of the dual-polarimetric upgrade for the WSR-88D fleet, we also propose to explore the use of polarimetric spectral densities as a means to discriminate between meteorological and non-meteorological returns, including biological targets.
Objective: Radar returns from the ground, known as ground clutter, can contaminate weather signals, often resulting in severely biased meteorological estimates. A ground clutter filter is implemented on the WSR-88D (GMAP) with the goal of eliminating ground clutter returns and providing unbiased meteorological estimates. However, the WSR-88D fleet continues to have data-quality issues with ground clutter mitigation using the current scheme. If not effectively removed, these contaminants artificially inflate quantitative precipitation estimates (QPE) and obscure Doppler-velocity signatures of weather.
Range and Velocity Ambiguity Resolution
Description of work to be performed: A complete range-and-velocity ambiguity mitigation solution for the WSR-88D, which involves the design of scanning strategies that exploit both the existing phase coding algorithm (SZ-2) and the proposed Staggered PRT, is planned for future enhancements of the WSR-88D RDA. NSSL proposes to transfer a complete staggered-PRT algorithm that includes dual polarization and ground clutter suppression using the CLEAN-AP filter. Additionally, NSSL proposes to improve the performance of SZ-2 algorithm on the polarimetric WSR-88D fleet by optimizing the algorithm adaptable parameters.
Objective: A complete range-and-velocity ambiguity mitigation solution would result in less obscuration (“purple haze”) and better depiction of Doppler velocity fields. Efforts in this area are expected to culminate in significantly improved data quality for the polarimetric WSR-88D radar. The increased data quality will result in an improved ability to detect severe weather, flash floods, winter storms, and provide aviation forecasts.
Online Noise Estimation
Description of work to be performed: Leveraging on work done for the MPAR project, we propose to develop a radial-by-radial noise-power estimation techniques for the dual-polarimetric WSR-88D RDA.
Objective: The objective of the technique is to produce a noise estimate at each antenna position. Thus, noise is estimated from samples that contain both signals of interest and noise. By providing more accurate noise power value, more accurate products will be produced in case of weak signals. Also, investigations have shown that in most cases, the “legacy” supplied noise value is higher than the actual one; thus, increase in radar coverage is another likely benefit.
Coherency-based Thresholding
Description of work to be performed: This technique has been developed at NSSL and its benefits have been demonstrated. The algorithm has been already been implemented in the ORDA. Support in testing and validating the technique will be provided.
Objective: The technique is aimed at improving the detection of weak signals. Thus, the objective is to mitigate the loss in sensitivity due to the 3.5 SNR drop resulting from the dual-polarization upgrade of the WSR-88D fleet. It has been shown that the technique is capable of achieving this goal.
DP Signal Processing Enhancements
Description of work to be performed: In the guide to the NWS and its contractor on how to integrate computations of dual polarimetric variables into the architecture of the WSR-88D RDA, Dusan Zrnic provided a series of recommendations to follow the initial dual-polarization deployment. Proposed enhancements include improvements to the mitigation of range and velocity ambiguities, the recovery of lost sensitivity, and the efficient computation of polarimetric variables. We propose to follow up on the recommendations outlined in this report and address other data-quality issues that may arise after the initial deployment of dual polarization on the WSR-88D fleet.
Objective: To mitigate risk, signal processing recommendations for the initial deployment of dual polarization on the WSR-88D fleet were purposely designed to involve minimal changes with respect to the legacy, single-polarization processing pipeline. As such, many opportunities for further data-quality improvements exist. An area of particular concern is the variance of dual-polarization estimates, which affect the performance of algorithms, such as quantitative precipitation estimation (QPE) and hydrometeor classification (HCA).
Range Oversampling
Description of work to be performed: Recently, we implemented an adaptive algorithm based on range oversampling on the NWRT PAR as part of the MPAR project. With this implementation, we were able to define new scanning strategies that produce data with similar quality in half the time. We propose to extend this technique for operational implementation on the dual-polarized WSR-88D radars.
Objective: In a recent survey by the ROC, 62% of forecasters indicated that they would like to get meteorological data at faster rates, especially for the observation of fast-evolving phenomena such as tornadoes. The conventional trade-off involves sacrificing either spatial coverage or data precision to obtain faster scans. However, with range oversampling it is possible to add a new dimension to this trade-off: signal processing. That is, with range oversampling it is possible to obtain low-variance data without sacrificing update time or spatial coverage. This is particularly important for the polarimetric variables, which are needed with higher precision than is possible using legacy, single-polarization dwell times.