Amanda Burke - April 8

Convective Meteorology (Mesoscale Dynamics) Detecting Above-Anvil Cirrus Plumes on a Pixel Scale Using Deep Learning Amanda Burke Friday, April 8 03:30 PM Online In satellite imagery, above-anvil cirrus plumes (plumes for short) are the strongest indicator of potentially significant severe weather, appearing on average 30 minutes before severe weather is

Start

April 8, 2022 - 3:30 pm

End

April 8, 2022 - 4:30 pm

Convective Meteorology (Mesoscale Dynamics)

Detecting Above-Anvil Cirrus Plumes on a Pixel Scale Using Deep Learning

Amanda Burke

Friday, April 8

03:30 PM

Online

In satellite imagery, above-anvil cirrus plumes (plumes for short) are the strongest indicator of potentially significant severe weather, appearing on average 30 minutes before severe weather is reported. Real-time plume identification could provide forecasters with information of the convective environment in areas where radar coverage is sparse. One current drawback of plume classification for severe weather prediction is that identifying plumes by hand can be challenging. Not only are the features relatively small in satellite images, but specialized knowledge can be necessary for ideal classifications. A solution to quickly identify plumes over large domains is through training deep learning (DL) models to output plume classifications based on expert drawn labels. With multiple timesteps of satellite imagery as inputs, a UNET (type of DL model) can learn spatio-temporal patterns in data and output predictions on a pixel-scale in a few minutes.  For this talk I will show preliminary findings of plume classification using infrared-only data for cases between 2019-2020 over the contiguous United States.