Advancing drought forecasting in Spain: Integration of meteorological indices and Random Forest algorythm for future projections
Year: 2026
Type: article
Abstract: Drought is a serious environmental issue that negatively impacts water resources, agricultural production, ecosystems, and economic activities as a result of prolonged periods of low precipitation. In particular, the depletion of water resources and difficulties in accessing water pose significant threats to societies. In this context, developing effective forecasting systems in regions at risk of drought is critical for managing water resources more efficiently and taking timely measures. This study examines the potential of integrating various drought indices and machine learning techniques to improve the accuracy of meteorological drought predictions. Using data from 54 meteorological stations in Spain for the 1973–2023 period, drought analyses were conducted based on the standardized precipitation index (SPI), standardized precipitation evapotranspiration index (SPEI), and reconnaissance drought index (RDI). Future drought predictions were made using the Random Forest (RF) algorithm. The RF algorithm successfully analyzed historical climate data to understand the temporal and spatial dynamics of drought occurrences. Additionally, a newly developed drought mapping approach demonstrated that short-term droughts are more prevalent in northern Spain compared to the southern regions. The findings highlight the likelihood of increased drought severity in specific areas and its potential impacts on agricultural production and water management. This study serves as a crucial guide for policymakers aiming to develop drought management strategies and contributes to effective planning to mitigate future drought impacts. Furthermore, the developed software is provided as open source alongside the article.
Source: Időjárás
Authors: Gözde Nur Akşan, Fatih Dikbaş
Language: English
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7165167832