Title: Identifying Lightning Processes in ERA5 Soundings with Deep Learning

Abstract: Atmospheric environments favorable for lightning and convection are commonly represented by proxies or parameterizations based on expert knowledge such as CAPE, wind shears, charge separation, or combinations thereof. Recent developments in the field of machine learning, high resolution reanalyses, and accurate lightning observations open possibilities for identifying tailored proxies without prior expert knowledge. To identify vertical profiles favorable for lightning, a deep neural network links ERA5 vertical profiles of cloud physics, mass field variables and wind to lightning location data from the Austrian Lightning Detection and Information System (ALDIS), which has been transformed to a binary target variable labelling the ERA5 cells as cells with lightning activity and cells without lightning activity. The ERA5 parameters are taken on model levels beyond the tropopause forming an input layer of approx. 670 features. The data of 2010-2018 serve as training/validation. On independent test data, 2019, the deep network outperforms a reference with features based on meteorological expertise. SHAP values highlight the atmospheric processes learned by the network which identifies cloud ice and snow content in the upper and mid-troposphere as very relevant features. As these patterns correspond to the separation of charge in thunderstorm cloud, the deep learning model can serve as physically meaningful description of lightning. Depending on the region, the neural network also exploits the vertical wind or mass profiles to correctly classify cells with lightning activity.