Radar imagery from 6:56 p.m. shows a close-up of the Mayfest supercell centered west of Benbrook, Texas. The pink and darkest red colors represent radar indications of large hail with this storm. The storm impacted the Mayfest festival at 7:10 p.m.

Credit: National Weather Service

An image from the work of Ph.D. student David Gagne using machine learning algorithms to predict hail a day in advance using the Storm Scale Ensemble Forecast (SSEF) that is run each year by CAPS as part of NOAA's Hazardous Weather Testbed. The top two panels show raw output of variables in the model associated with hail. The bottom panel shows results from the random forest machine learning algorithm, which is used to calibrate the raw model output and dramatically improve the forecast of hail. The blue circles are locations where hail exceeding 50 mm (2 inches) in diameter was observed, and the shading is the probability of hail exceeding 50 mm in diameter occurring within 50 miles of a given point. The machine learning method correctly identifies the area of highest hail threat 24 hours in advance while producing very few false alarms.

Credit: David Gagne, University of Oklahoma

This image show radar-indicated swaths of hail from the May 20, 2013, storms in the left panel, and, in the right panel, swaths of hail predicted by the ensemble. The color of the shading indicates the maximum hail size (the scale is in millimeters, but the breaks between colors at 25, 50 and 75 mm correspond closely to hail diameters of 1, 2 and 3 inches). The inset image in the left-hand panel is a picture of hail that fell during these storms, submitted by a member of the public to local news station KFOR. The algorithm being used to generate the image is called MESH (Maximum Estimated Size of Hail) -- this algorithm is used in severe weather operations by NWS. It uses a combination of radar information and measurements of temperature at different levels within the atmosphere to derive an estimate of maximum hail size at the surface.

Credit: University of Oklahoma