Identifying lightning processes in ERA5 soundings with deep learning
Open Access
- 26 February 2025
- journal article
- research article
- Published by Copernicus GmbH in Geoscientific Model Development
- Vol. 18 (4) , 1141-1153
Abstract
Atmospheric environments favorable for lightning and convection are commonly represented by proxies or parameterizations based on expert knowledge such as convective available potential energy (CAPE), wind shear, charge separation, or combinations thereof. Recent developments in the field of high-resolution reanalyses, accurate lightning observations, machine learning (ML), and explainable artificial intelligence (XAI) open possibilities for identifying tailored proxies without prior expert knowledge. This study utilizes a deep neural network trained to match temporally and vertically well-resolved ERA5 soundings of cloud physics, mass-field, and wind-field variables with lightning observations from the Austrian Lightning Information & Detection System (ALDIS). The ML model only receives the raw model atmosphere data as inputs, without incorporating any expert parameters or proxies derived from the model levels. Using and adapting appropriate XAI methods, it is then demonstrated how the inner workings of this well-performing deep learning model can be uncovered to identify physically meaningful patterns within the ERA5 soundings that describe lightning processes. The ERA5 parameters are taken on model levels beyond the tropopause, forming an input layer of approx. 670 features, and the lightning data are transformed to a binary target variable labeling the spatio-temporal ERA5 grid cells as cells with lightning activity and cells without lightning activity. Scaled Shapley additive explanations (SHAP) values are introduced to highlight the atmospheric processes learned by the neural network and show that the model identifies cloud ice and snow content in the upper troposphere and mid-troposphere as very relevant features. As these patterns correspond to the separation of charge in thunderstorm clouds, the deep learning model can serve as a physically meaningful description of lightning. The scaled SHAP values also reveal that, depending on the location, the model additionally learns to correctly classify cells with lightning activity by exploiting mass-field or wind-field variables. This approach also showcases how XAI can be used to accelerate knowledge discovery in areas where expert knowledge is still scarce.Keywords
All Related Versions
Funding Information
- Austrian Science Fund (P 31836)
- Österreichische Forschungsförderungsgesellschaft (872656)
This publication has 57 references indexed in Scilit:
- Interpretation of black box for short-term predictions of pre-monsoon cumulonimbus cloud events over KolkataJournal of Data, Information and Management, 2022
- Modelling deep convective activity using lightning clusters and machine learningInternational Journal of Climatology, 2021
- Contrasting lightning projection using the lightning potential index adapted in a convection-permitting regional climate modelClimate Dynamics, 2021
- Winter Positive Cloud‐to‐Ground Lightning Flashes Observed by LMA in JapanIEEJ Transactions on Electrical and Electronic Engineering, 2021
- The ERA5 global reanalysisQuarterly Journal of the Royal Meteorological Society, 2020
- Performance of a simple reanalysis proxy for U.S. cloud‐to‐ground lightningInternational Journal of Climatology, 2019
- On the estimation of the lightning incidence to offshore wind farmsElectric Power Systems Research, 2018
- Gridded lightning climatology from TRMM-LIS and OTD: Dataset descriptionAtmospheric Research, 2014
- Mesoscale Meteorology in Midlatitudes2010
- Lightning injury: A reviewBurns, 2008