Core Science Keynote: Findings from Phased-Array Radar Innovative Sensing Experiments

The weather industry is exploring phased array radar (PAR) technology as a potential replacement system for the Weather Surveillance Weather Radar – 1988 Doppler (WSR-88D) network. While the WSR-88D network has the capability to provide “dynamic” volume coverage patterns like Supplemental Adaptive Intra-volume Low-levels Scan (SAILS), the volume scan update time is limited by mechanical scanning. While using SAILS low-level scans are more frequent, the overall scan time is significantly increased. Electronic scanning allows the radar operator to scan storms adaptively without extending the volume scan time. Rather the volume scan time is reduced by 75%.

The SPY-1A panel installed at the National Weather Radar Testbed in Norman, Oklahoma during the early 2000’s through ~2016 provided the opportunity to study how adaptive, rapid-scanning of storms impacts the depiction and evolution of severe storms, as well as the depiction of severe storm resulting from the assimilation of these data into convection-allowing models. Importantly, the usability of these data by humans was explored through Phased Array Radar Innovative Sensing Experiments (PARISE) conducted by the NOAA National Severe Storms Laboratory within the NOAA Hazardous Weather Testbed.

The overarching objective of the Phased Array Radar Innovative Sensing Experiment (PARISE) is to understand what effects higher-temporal resolution volumetric radar data may have on National Weather Service (NWS) forecasters’ warning decision processes (e.g., Heinselman et al. 2012; Heinselman et al. 2015; Bowden et al. 2015; Bowden and Heinselman 2016; Wilson et al. 2017). Using a retrospective walk through approach, these experiments studied similarities and differences in what forecasters’ were seeing, thinking, and doing when working potentially severe storm cases using phased array radar vs legacy WSR-88D data in simulated real time. Also studied was the resulting warning performance in terms of probability of detection, false alarm rate, and warning lead time for severe, tornadic, and non-tornadic events. In addition to warning performance, later studies examined: 1) how temporal resolution impacts forecasters’ cognitive workload and if and when data overload exists; 2) impacts of update time on data interrogation techniques and ability to apply their conceptual models, 3) how rapid-scan data could be integrated into operations, and 4) training that would be useful in transitioning rapid-update data to operations. This presentation will summarize the findings from these studies and provide ideas on how to progress this research using the recently installed dual-polarized Advanced Technology Demonstrator.