Developing a Mosaicked Gust Front Detection Algorithm for TRACONS with Multiple TDWRs

Gust front detection is an important Initial Operational Capability (IOC) of the Integrated Terminal Weather System (ITWS). The Machine Intelligent Gust Front Algorithm (MIGFA) being deployed for ITWS uses multi-dimensional, knowledge-based signal processing techniques to detect and track gust fronts in Terminal Doppler Weather Radar (TDWR) data. Although MIGFA represents a significant improvement over previous algorithms, MIGFA's detection capability is hampered by sensor limitations inherent in single Doppler radar processing. Gust front Doppler velocity convergence signatures often vanish as gust fronts become radially aligned during overhead passage, making uninterrupted tracking difficult. A study was conducted to determine the number of gust fronts that were dropped by MIGFA as they crossed the Dallas/Ft. Worth (DFW) airport due to radial alignment. Results showed that detections for 21 of the 52 fronts were lost before the fronts impacted the DFW runways. However, half of these fronts remained detected by MIGFA operating on data from another nearby TDWR.

The current IOC ITWS approach for gust front product generation at large TRACON sites with multiple TDWR coverage is not optimal. At such sites, gust front detections are generated independently by separate MIGFAs running off of each TDWR. A separate ITWS gust front product mapping algorithm then utilizes a rule base to divide up the TRACON coverage area and pre-determine which radar's gust front detections are to be given priority for display in each sub-region. This product-level fusion has been found to be problematic, as it is difficult to pre-determine a radar priority scheme that is optimal for all possible gust front geometries, leading to fragmented and confusing displays. A more desirable approach would be to perform data fusion from all available TDWRs at an earlier stage, and then use the merged data as the basis for gust front feature extraction, resulting in generation of a single set of gust front detections for the TRACON coverage area. Some additional challenges arise with this approach, such as difficulties with merging asynchronous radar data of moving gust fronts, and performing wind shift analyses on the mosaicked gust front product. Despite these challenges, this data-level fusion approach should significantly enhance overall gust front detection capability, improve safety, and reduce confusion among ITWS users in TRACONs with multiple TDWRs.

Several analysis tasks are being performed to validate the proposed approach. A test suite of cases having gust front detections that were dropped by one MIGFA due to radial alignment, but maintained by MIGFA output from a second TDWR during airport impact is presently being assembled. Methods of data fusion and subsequent gust front feature extraction and wind analysis are being investigated and will lead to development of a suitable real-time algorithm. Preliminary results of these investigations will be presented.