News
Aemet refines seasonal forecasts to anticipate droughts and wildfires
Aemet and CSIC have applied neural networks to Copernicus seasonal models to achieve greater territorial detail. The products are now available on the Plataforma Estatal de Servicios Climáticos and are aimed at monitoring droughts and wildfire risk.
Copernicus, the European Union’s Earth observation programme, provides information and services on climate, the atmosphere, the oceans, and the environment. Its seasonal models underpin a methodology that Aemet has integrated into the Plataforma Estatal de Servicios Climáticos (PESC) alongside CSIC.
These products can now be accessed in the Visor de Predicción Estacional on the platform, together with other forecasts and data on their validation. The tool aims to give a more detailed picture of climate differences between various regions of Spain.
How forecasts are regionalised
The methodology transforms large-scale forecasts into regionalised information using super-resolution techniques based on convolutional neural networks. This is an application of machine learning that increases the spatial resolution of forecasts.
The Servicio de Cambio Climático de Copernicus funded the project through the National Collaboration Programme (NCP). Its goal is to improve climate services related to drought and wildfires in Spain.
Applications for droughts and wildfires
Aemet plans to gradually incorporate the regionalised forecasts into other PESC services. Priorities include drought monitoring and assessing meteorological wildfire risk.
Medium-term climate information is aimed at public administrations and businesses, as well as other users who need to anticipate changes in variables such as temperature or precipitation.