Deep neural network adaptation to different environmental contexts: A case study of ozone forecast

Many environmental variables, in particular, related to air or water quality, are measured in a limited number of points and often for a limited time span. This forbids the development of accurate models for those locations due to an insufficient number of data and poses the question of whether a model developed for another measurement station can be reliably applied. Such a question is particularly critical when the model is constituted by a neural network, i.e., by an approach fully based on local data. In this context, the paper discusses the results of the application of a model to forecast ozone concentrations trained on stations with various characteristics in different environmental settings.