Forecasting the Daily Temperature Extremes from Radiosonde Measurements Using Neural Networks

This study investigates the potential of direct prediction of daily extremes (maximum and minimum) of 2-m temperature from a radiosonde measurement using neural networks (NNs). The analysis is based on 3800 daily profiles measured in the period 2004-2019. Various setups of dense sequential NNs are trained to predict the daily extremes at different lead times. The purpose of the analysis was not to develop a model that would be better than operational numerical weather prediction models but primarily to investigate the capabilities of neural networks for such forecasts. Specifically, our goal was to understand how neural network-based models use different types of input data and how network design and its complexity affect their behavior. The data utilization and behavior of the network depend on whether the NNs are used to do short-term or long-term forecasts - this is why the analysis was performed for a wide range of forecast lead times ranging from 0 to 500 days into the future. The analysis of very simple NNs, consisting of only a few neurons, used to predict same-day extremes highlighted how the nonlinear behavior of the NN increases with the number of neurons. It also showed how different training realizations of the same network could result in different behaviors of the NN. The behavior in the part of the predictor phase space with the highest density of training cases was usually quite similar for all training realizations, while the behavior elsewhere was more variable and more frequently exhibited unusual nonlinearities. We also analyzed more complex NN setups that were used for the short- and long-term forecasts of temperature extremes. Besides the profile measurements, some setups used additional predictors such as the previous-day measurements and climatological values of extremes. The behavior of the setups was also analyzed via two XAI methods, which help determine which input parameters have a more significant influence on the forecasted value.