{"id":38136,"date":"2020-09-02T19:24:55","date_gmt":"2020-09-02T19:24:55","guid":{"rendered":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/?p=38136"},"modified":"2020-09-02T19:24:55","modified_gmt":"2020-09-02T19:24:55","slug":"using-goes-abi-and-deep-learning-to-nowcast-lightning","status":"publish","type":"post","link":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/archives\/38136","title":{"rendered":"Using GOES ABI and deep learning to nowcast lightning"},"content":{"rendered":"

NOAA and CIMSS are developing a product that uses a deep-learning model to recognize complex patterns in weather satellite imagery to predict the probability of lightning in the short term. Deep learning is a branch of machine learning based on artificial neural networks, which have the ability to automatically learn targeted features in the data by approximating how humans learn.<\/span><\/p>\n

A convolutional neural network (CNN) was trained on over 23,000 images of GOES-16 Advanced Baseline Imager (ABI) data to predict the probability of lightning, within any given ABI pixel, in the next 60 minutes. The CNN was trained using 118 days of data collected between May and August of 2018. Images of GOES Geostationary Lightning Mapper (GLM) flash-extent density (created with<\/span> glmtools<\/span><\/a>) were used as the source of lightning observations. Note that GLM is an optical sensor that observes both in-cloud and cloud-to-ground lightning.<\/span><\/p>\n

The CNN currently uses four ABI channels: band 2 (0.64-\u00b5m), band 5 (1.6-\u00b5m), band 13 (10.3-\u00b5m), and band 15 (12.3-\u00b5m). Bands 2 and 5 are only utilized under sunlit conditions. Utilization of additional channels and time sequences of images is under investigation. The model uses a semantic image segmentation architecture to assign the probability of lighting in the next 60 minutes to each pixel in the image. The model is very computationally efficient, only needing 30 seconds to process the ABI CONUS domain and 3 seconds to process an ABI mesoscale domain using multithreading on a 40-CPU linux server.<\/span><\/p>\n

Currently, the model only utilizes satellite radiances. Thus, it can be applied to nearly any spatial domain covered by the ABI or an ABI-like sensor (e.g. AHI). Based on near-real-time testing, the model routinely nowcasts lightning initiation with 10-30 minutes of lead-time. We expect the skill and lead-time will increase as new predictors (e.g. more ABI fields, NWP, radar where available) are added to the model.<\/span><\/p>\n

Below are a sampling of recent examples. The base images are 0.64-\u00b5m reflectance, with GLM-derived flash-extent density overlaid as filled semi-transparent polygons. The flash-extent density is the accumulated number of flashes within the previous 5 minutes. The CNN-derived probabilities are displayed as contours at select levels (<\/span>near-real-time output<\/span><\/a> is available through RealEarth).<\/span><\/p>\n

The overall objective is to improve lightning nowcasts in support of aviation, mariners, and outdoor events\/activities. Beyond improving the CNN, our work will focus on packaging the output into actionable information for forecasters and other decision makers.<\/span><\/p>\n

A cold front in Iowa<\/h2>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/IA_pltg_20200828.mp4<\/a><\/video><\/div>\n

 <\/p>\n

Thunderstorm development on sea-breeze boundaries in Florida and the Bahamas<\/h2>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/FL_pltg_20200827.mp4<\/a><\/video><\/div>\n

 <\/p>\n

Diurnally and orographically forced storms in the Southwest U.S. and Rocky Mountains<\/h2>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/GLMFED_CH02_20200821.mp4<\/a><\/video><\/div>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/Rockies_pltg_20200827.mp4<\/a><\/video><\/div>\n

 <\/p>\n

Storms in central Oklahoma, on the edge of Hurricane Laura\u2019s cloud shield<\/h2>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/OK_pltg_20200827.mp4<\/a><\/video><\/div>\n

 <\/p>\n

A couple of examples over the Northeast U.S.<\/h2>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/northeast_pltg_20200825.mp4<\/a><\/video><\/div>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/northeast_pltg_20200824.mp4<\/a><\/video><\/div>\n

 <\/p>\n

A boundary of convection on the southern bank of Lake Ontario<\/h2>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/NYPA_pltg_20200827.mp4<\/a><\/video><\/div>\n

 <\/p>\n

Storms in a warm sector in IL\/IN\/OH, perhaps along an outflow boundary<\/h2>\n
https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/IN_pltg_20200828.mp4<\/a><\/video><\/div>\n

 <\/p>\n

Southeast U.S. offshore region<\/h2>\n

The background image in this example transitions from 10.3-\u00b5m brightness temperature to 0.64-\u00b5m reflectance, while the flash-extent density enhancement also changes, in an attempt to enhance contrast.<\/p>\n

https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-content\/uploads\/sites\/5\/2020\/09\/offshore_pltg_20200829.mp4<\/a><\/video><\/div>\n","protected":false},"excerpt":{"rendered":"

NOAA and CIMSS are developing a product that uses a deep-learning model to recognize complex patterns in weather satellite imagery to predict the probability of lightning in the short term. Deep learning is a branch of machine learning based on artificial neural networks, which have the ability to automatically learn targeted features in the data […]<\/p>\n","protected":false},"author":14,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[10,74,30,59],"tags":[90,84,91,87],"class_list":["post-38136","post","type-post","status-publish","format-standard","hentry","category-general-interpretation","category-goes-16","category-lightning","category-probsevere","tag-abi","tag-goes-16","tag-lightning","tag-probsevere"],"acf":[],"_links":{"self":[{"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/posts\/38136","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/users\/14"}],"replies":[{"embeddable":true,"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/comments?post=38136"}],"version-history":[{"count":19,"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/posts\/38136\/revisions"}],"predecessor-version":[{"id":38166,"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/posts\/38136\/revisions\/38166"}],"wp:attachment":[{"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/media?parent=38136"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/categories?post=38136"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/cimss.ssec.wisc.edu\/satellite-blog\/wp-json\/wp\/v2\/tags?post=38136"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}