Uso de redes neuronales para la identificación de las situaciones sinópticas de Font

Abstract

Synoptic classifications serve as powerful tools for understanding the atmospheric dynamics that influence climatic variability in a specific region. Inocencio Font's classification, comprising 23 synoptic situations, summarizes the dominant circulation patterns over the Iberian Peninsula and the Balearic Islands. Since 2020, the Techniques and Applications for Forecasting Area (ATAP, by its Spanish acronym) at Aemet has subjectively classified each day's synoptic situation according to the most representative Font's situation. Automating this task could facilitate expanding the classification to encompass broader time periods, including medium and long-range forecasting, reanalysis or climate projections. To this end, this study presents the development of an automatic classifier comprising 10 artificial neural networks. These networks utilize mean sea level pressure and geopotential at 500hPa as input variables, emulating ATAP’s procedure. The results show the assigned synoptic patterns reflect accurately the expected weather based on Font’s classification.