Predicting Tmax from latest radiosonde observations

In February 2021, in the midst of the second epidemic wave, Doruntina Hoxha joined our atmospheric research group at UL FMF. Welcome, Doruntina! Doruntina will be working on the use of machine learning in subseasonal prediction. Her first task was to evaluate the impact of nighttime radiosonde measurements of temperature and relative humidity on the T_max and T_mi prediction at the same station for the same day (day 0), the next day (day 1) etc. In the distant future, the neural networks might be able to predict the weather evolution directly from the measurements, without the NWP.