The WSR-88D Chaff Detection Algorithm Utilizing a Support Vector Machine based on Human Truthing

Military or other-purpose chaff, which consists of many small, metallic strands, can be a nuisance for weather radars. The WSR-88D radar network, used jointly by the National Oceanic and Atmospheric Administration (NOAA), the Federal Aviation Administration (FAA), the Department of Defense (DOD), and many other consumers, is widely affected by chaff exercises and/or releases on a daily basis. While National Weather Service (NWS) forecasters tend to be able to distinguish chaff from weather using polarimetric moment estimates, air traffic controllers do not have the luxury of polarimetric data at their workstations. In cases of mixed chaff and convection, this can lead to difficulty in route planning and avoidance. In order to address this issue, a WSR-88D Chaff Detection Algorithm (CDA) has been developed and is scheduled for inclusion in Build 20 of the WSR-88D Open Radar Product Generator (ORPG). The CDA is based on a new Hydrometeor Classification Algorithm (HCA) class, a series of image processing techniques, and a Support Vector Machine (SVM) based on human truthing of millions of chaff, weather, and clutter resolution volumes. The human truthing, new HCA class, and image processing techniques are briefly covered, with a focus on the SVM methodology and implementation. A series of cases processed in the ORPG are presented, including case successes and failures. Implementation strategies are shown, and expectations for future use within the FAA are discussed.