Machine Learning Approaches for Predicting Reference Evapotranspiration: A Comparative Study Using Ground and Gridded Climate Data in Fes Region

World Water Policy, Год журнала: 2025, Номер unknown

Опубликована: Янв. 22, 2025

ABSTRACT Climate data are essential for agricultural planning and water resource management; however, their availability is limited in numerous regions of Africa. Gridded climate present a potential solution, yet, accuracy estimating reference evapotranspiration (ET o ) remains uncertain. This study aims to evaluate the performance gridded comparison ground‐based observations predicting ET Fes region Morocco. Two machine learning (ML) models, random forest (RF) long short‐term memory (LSTM), were trained tested on 10 years from both (AgERA5) ground (in situ) observation sources assess predictive capabilities. The results demonstrated that RF outperformed LSTM under fewer input parameter configurations, achieving R 2 > 0.70, while exhibited superior across all configurations 0.95. However, AgERA5 consistently underestimated compared observations. underestimation highlights need bias correction improve reliability. Addressing these limitations would allow datasets support better irrigation scheduling, enhance use efficiency, reduce crop stress with access localized data. demonstrates combining ML bridge gaps, emphasizing importance improving dataset practical applications management.

Язык: Английский

Hydrogeochemical process, multivariate statistical, geospatial and index approach for evaluation of groundwater quality for irrigation purposes at Visakhapatnam region, Southern India

Environmental Geochemistry and Health, Год журнала: 2025, Номер 47(5)

Опубликована: Апрель 3, 2025

Язык: Английский