Search

This paper aims to explore machine learning techniques for post‐processing high‐resolution Numerical Weather Prediction (NWP) products for the early detection of convection. Data from the Arome Ensemble Prediction System and satellite observations from the Rapidly Developing Thunderstorm (RDT) product by Météo‐France are used to train a recurrent neural network model to predict areas of total convection and moderate convection. The learning task is formulated as a binary classification problem using a long short‐term memory (LSTM) network architecture. Results from the LSTM model are compared with an object‐based probabilistic approach to forecast convection using metrics such as a receiver operating characteristics (ROC) curve, the Brier score and reliability. Results indicate that the LSTM model performs similarly to the object‐based probabilistic benchmark when classifying moderate convection areas and shows improved skill when classifying areas of total convective. Finally, the LSTM model results are presented within an air traffic management context to showcase the potential use of machine learning models within an operational application. [ABSTRACT FROM AUTHOR]

Convective weather is a major source of disruption to air traffic operations responsible for roughly one third of en-route delay in the network. Understanding how weather impact air traffic flows is an important first step in improving Air Traffic Flow Management operations during convective weather. In this research we exploit advances in machine learning to find the spatial correlations between convective weather and regulated air traffic flows. Our approach utilises historical satellite observations and traffic data from the summer of 2018 and 2019. Three deep learning model architectures are trained to detect traffic hotspots that are subject to regulation due to weather conditions. We also establish a baseline indicator to assess prediction quality. The weather and traffic data are transformed into images using a 2D grid covering a substantial portion of Europe, with a high-resolution spatial resolution. The temporal dimension is considered through hourly or 15-minute time windows, depending on the model architecture. By integrating this information, our models are able to capture relevant correlations effectively. Results indicate a clear correlation between weather and regulated traffic and reveal that our machine learning model is able to learn from the spatial correlations in the data and thus characterised the geographical areas where traffic volumes regulated due to convective weather are expected to appear during Air Traffic Flow Management practice. • Historical satellite observations and traffic data are used to train a CNN to detect ATFM weather hotspots. • Accounting for the spatial distribution of storms enhances the identification of ATFM hotspots during convective weather. • Data driven models can reduce the cognitive workload to solve demand-capacity imbalances. [ABSTRACT FROM AUTHOR]