Detection of Bow Echoes in Kilometer-Scale Forecasts Using a Convolutional Neural Network

Abstract

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 pseudoreflectivity field predictor (“pseudo” because it is expressed in mm h−1 and not in dBZ). A preprocessing 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 object-oriented evaluation. The BE largest pseudoreflectivities 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.

Rain Scanner Radar and Optical Flow Combination For Early Identification of The Incoming Bow Echo Heavy Precipitation

“…The proposed method using a combination of three parameters (circularities, eccentricity, and orientation) is presented in this study. This method is different from other previous research [4,5,6] since it uses a combination of three parameters from the shape measurement of an object image to detect bow echo. In this section, the limitation of this study, as well as future works, are discussed.…”

“…A different approach is done by Fanlin and Jinyi [5] which measured the contour to find the curve of the bow echo. Another method used pseudo reflectivity combined with the Convolutional Neural Network to detect the bow echo [6].…”

A derecho climatology (2004–2021) in the United States based on machine learning identification of bow echoes

“…Perhaps most closely related to this study is Mounier et al (2022), who used a U-Net to detect bow echoes in simulated radar reflectivity images from a forecast model. A U-Net is an appropriate choice for the segmentation of bow echoes because merging multi-resolution information is crucial for identifying the feature.…”