This research uses image classification and machine learning methods on radar reflectivity mosaics to segment, classify, and track quasi-linear convective systems (QLCSs) in the United States for a 22-yr period. An algorithm is trained and validated using radar-derived spatial and intensity information from thousands of manually labeled QLCS and non-QLCS event slices. The algorithm is then used to automate the identification and tracking of over 3000 QLCSs with high accuracy, affording the first, systematic, long-term climatology of QLCSs. Convective regions determined by the procedure to be QLCSs are used as foci for spatiotemporal filtering of observed severe thunderstorm reports; this permits an estimation of the number of severe storm hazards due to this morphology. Results reveal that nearly 32% of MCSs are classified as QLCSs. On average, 139 QLCSs occur annually, with most of these events clustered from April through August in the eastern Great Plains and central/lower Mississippi and Ohio River Valleys. QLCSs are responsible for a spatiotemporally variable proportion of severe hazard reports, with a maximum in QLCS-report attribution (30%–42%) in the western Ohio and central Mississippi River Valleys. Over 21% of tornadoes, 28% of severe winds, and 10% of severe hail reports are due to QLCSs across the central and eastern United States. The proportion of QLCS-affiliated tornado and severe wind reports maximize during the overnight and cool season, with more than 50% of tornadoes and wind reports in some locations due to QLCSs. This research illustrates the utility of automated storm-mode classification systems in generating extensive, systematic climatologies of phenomena, reducing the need for time-consuming and spatiotemporal-limiting methods where investigators manually assign morphological classifications.
This research quantifies the spatiotemporal statistics of composite radar reflectivity in the vicinity of severe thunderstorm reports. By using over 20 years (1996–2017) of data and 500,000 severe thunderstorm reports, this study presents the most comprehensive analysis of the mesoscale presentation of radar reflectivity composites during severe weather events to date. We first present probability matched mean composites of approximately 5,000 radar images centred on tornado reports that contain one of three types of manually‐labelled convective storm modes—namely, (a) quasi‐linear convective system (QLCS); (b) cellular; or (c) tropical system. Next, we generate composites for tornado report data stratified by EF‐scale and for four temporal periods during which notable severe weather events took place. The data are then stratified by hazard, region, season, and time of day. The results show marked spatiotemporal and intra‐hazard variability in radar presentation. In general, cellular convection is favoured in the Great Plains of the United States, whereas QLCS convection is favoured in the Southeast United States. Night and cool‐season subsets showed a preference for QLCS convection, whereas day and warm‐season subsets showed a preference for cellular convection. These results agree well with the existing literature and suggest that the data extraction and organization approach is sound. Because of this, these data will be useful for future image classification studies in climate and atmospheric sciences—particularly those involving storm mode classification.
more » « less- Award ID(s):
- 1637225
- NSF-PAR ID:
- 10379707
- Publisher / Repository:
- Wiley Blackwell (John Wiley & Sons)
- Date Published:
- Journal Name:
- International Journal of Climatology
- Volume:
- 41
- Issue:
- S1
- ISSN:
- 0899-8418
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract -
Abstract The Weather Surveillance Radar-1988 Doppler (WSR-88D) network has undergone several improvements in the last decade with the upgrade to dual-polarization capabilities and the ability for forecasters to rescan the lowest levels of the atmosphere more frequently through the use of Supplemental Adaptive Intra-volume Scanning (SAILS). SAILS reduces the revisit period for scanning the lowest 1 km of the atmosphere but comes at the cost of a longer delay between scans at higher altitudes. This study quantifies how often radar volume coverage patterns (VCPs) and all available SAILS options are used during the issuance of 148 882 severe thunderstorm and 18 263 tornado warnings, and near 10 474 tornado, 58 934 hail, and 127 575 wind reports in the dual-polarization radar era. A large majority of warnings and storm reports were measured with a VCP providing denser low-level sampling coverage. More frequent low-level updates were employed near tornado warnings and reports compared to severe thunderstorm warnings and hail or wind hazards. Warnings issued near a radar providing three extra low-level scans (SAILSx3) were more likely to be verified by a hazard with a positive lead time than warnings with fewer low-level scans. However, extra low-level scans were more frequently used in environments supporting organized convection as shown using watches issued by the Storm Prediction Center. In recent years, the number of midlevel radar elevation scans is declining per hour, which can adversely affect the tracking of convective polarimetric signatures, like Z DR columns, which were found above the lowest elevation angle in over 99% of cases examined.more » « less
-
more » « less
Abstract Many studies have aimed to identify novel storm characteristics that are indicative of current or future severe weather potential using a combination of ground-based radar observations and severe reports. However, this is often done on a small scale using limited case studies on the order of tens to hundreds of storms due to how time-intensive this process is. Herein, we introduce the GridRad-Severe dataset, a database including ∼100 severe weather days per year and upward of 1.3 million objectively tracked storms from 2010 to 2019. Composite radar volumes spanning objectively determined, report-centered domains are created for each selected day using the GridRad compositing technique, with dates objectively determined using report thresholds defined to capture the highest-end severe weather days from each year, evenly distributed across all severe report types (tornadoes, severe hail, and severe wind). Spatiotemporal domain bounds for each event are objectively determined to encompass both the majority of reports and the time of convection initiation. Severe weather reports are matched to storms that are objectively tracked using the radar data, so the evolution of the storm cells and their severe weather production can be evaluated. Herein, we apply storm mode (single-cell, multicell, or mesoscale convective system storms) and right-moving supercell classification techniques to the dataset, and revisit various questions about severe storms and their bulk characteristics posed and evaluated in past work. Additional applications of this dataset are reviewed for possible future studies.
-
more » « less
This research assesses the utility and validity of using simulated radar reflectivity to detect potential changes in linear and nonlinear mesoscale convective system (MCS) occurrence in the Midwest United States between the early and late 21st century using convection‐permitting climate simulation output. These data include a control run and a pseudo‐global warming (PGW) run that is based on RCP 8.5. First, using a novel segmentation, classification, and tracking procedure, MCS tracks are extracted from observed and simulated radar reflectivity. Next, a comparison between observed and the control run MCS statistics is performed, which finds a negative summertime bias that agrees with previous work. Using a convolutional neural network to perform probabilistic predictions, the MCS dataset is further stratified into highly organized, quasi‐linear convective systems (QLCSs)—which can include bow echoes, squall lines, and line echo wave patterns—and generally less‐organized, non‐QLCS events. The morphologically stratified data reveal that the negative MCS bias in this region is largely driven by too few QLCSs. Although comparisons between the control run and a PGW run suggest that all MCS events are less common in the future (including QLCS and non‐QLCS events), these changes are not spatially significant, whereas the biases between the control run and observations are spatially significant. A discussion on the importance and challenges of simulating QLCSs in convection‐permitting climate model runs is provided. Finally, potential avenues of exploration are suggested related to the aforementioned issues.
-
more » « less
Abstract The quantification of storm updrafts remains unavailable for operational forecasting despite their inherent importance to convection and its associated severe weather hazards. Updraft proxies, like overshooting top area from satellite images, have been linked to severe weather hazards but only relate to a limited portion of the total storm updraft. This study investigates if a machine learning model, namely, U-Nets, can skillfully retrieve maximum vertical velocity and its areal extent from three-dimensional gridded radar reflectivity alone. The machine learning model is trained using simulated radar reflectivity and vertical velocity from the National Severe Storm Laboratory’s convection permitting Warn-on-Forecast System (WoFS). A parametric regression technique using the sinh–arcsinh–normal distribution is adapted to run with U-Nets, allowing for both deterministic and probabilistic predictions of maximum vertical velocity. The best models after hyperparameter search provided less than 50% root mean squared error, a coefficient of determination greater than 0.65, and an intersection over union (IoU) of more than 0.45 on the independent test set composed of WoFS data. Beyond the WoFS analysis, a case study was conducted using real radar data and corresponding dual-Doppler analyses of vertical velocity within a supercell. The U-Net consistently underestimates the dual-Doppler updraft speed estimates by 50%. Meanwhile, the area of the 5 and 10 m s−1updraft cores shows an IoU of 0.25. While the above statistics are not exceptional, the machine learning model enables quick distillation of 3D radar data that is related to the maximum vertical velocity, which could be useful in assessing a storm’s severe potential.
Significance Statement All convective storm hazards (tornadoes, hail, heavy rain, straight line winds) can be related to a storm’s updraft. Yet, there is no direct measurement of updraft speed or area available for forecasters to make their warning decisions from. This paper addresses the lack of observational data by providing a machine learning solution that skillfully estimates the maximum updraft speed within storms from only the radar reflectivity 3D structure. After further vetting the machine learning solutions on additional real-world examples, the estimated storm updrafts will hopefully provide forecasters with an added tool to help diagnose a storm’s hazard potential more accurately.
