A Branched Convolutional Neural Network for Forecasting the Occurrence of Hazes in Paris using Meteorological Maps with Different Characteristic Scales

Abstract A deep convolutional neural network has been developed to forecast the occurrence of the low visibility events or hazes in Paris. It is trained by using multi-decadal daily regional maps of various meteorological and hydrological variables as the input features and surface visibility observations as the targets. To better preserve the characteristic spatial-scale information of different input features during the training, two branched architectures have recently been developed. These new architectures have improved the performance of the network, producing impressive scores in validation and also in an evaluation using the data of 2021 and 2022 that have not been used in the training and validation.