Detection of bow echoes in kilometer-scale forecasts using a convolutional neural network

Arnaud Mounier (1) , Laure Raynaud (1) , Lucie Rottner (1) , Matthieu Plu (1) , Philippe Arbogast (2) , Michaël Kreitz (3) , Léo Mignan (4) , Benoît Touzé (2)

1 CNRM - Centre national de recherches météorologiques
2 DIROP/PG - Département Prévision Générale
3 ENM - Ecole Nationale de la Météorologie
4 Direction Interrégionale Ouest

Bow echoes (BEs) are bow-shaped lines of convective cells that are often associated with swaths of damaging straight-line winds and small tornadoes. This paper describes a convolutional neural network (CNN) able to detect BEs directly from French kilometer-scale model outputs in order to facilitate and accelerate the operational forecasting of BEs. The detections are only based on the maximum pseudo-reflectivity field predictor (pseudo because expressed in mm • h −1 and not in dBZ). A pre-processing of the training database is carried out in order to reduce imbalance issues between the two classes (inside or outside bow echoes). A CNN sensitivity analysis against a set of hyperparameters is done. The selected CNN configuration has a hit rate of 86% and a false alarm rate of 39%. The strengths and weaknesses of this CNN are then emphasized with an objectoriented evaluation. The BE largest pseudo-reflectivities are correctly detected by the CNN which tends to underestimate the size of BEs. Detected BE objects have wind gusts similar to the hand labeled BE. Most of the time, false alarm objects and missed objects are rather small (e.g., < 1500 km²). Based on a cooperation with forecasters, synthesis plots are proposed that summarize the BE detections in French kilometer-scale models. A subjective evaluation of the CNN performances is also reported. The overall positive feedback from forecasters is in good agreement with the object-oriented evaluation. Forecasters perceive these products as relevant and potentially useful to handle the large amount of available data from numerical weather prediction models.

Domaines: Océan, Atmosphère; Machine Learning [stat.ML]

Arnaud Mounier, Laure Raynaud, Lucie Rottner, Matthieu Plu, Philippe Arbogast, et al.. Detection of bow echoes in kilometer-scale forecasts using a convolutional neural network. Artificial Intelligence for the Earth Systems, 2022, ⟨10.1175/AIES-D-21-0010.1⟩. ⟨meteo-03811214⟩