Multi-Model Ensemble Machine Learning Approaches to Project Climatic Scenarios in a River Basin in the Pyrenees

This study employs machine learning algorithms to construct Multi Model Ensembles (MMEs) based on Regional Climate Models (RCMs) within the Esca River basin in the Pyrenees. RCMs are ranked comprehensively based on their performance in simulating precipitation (pr), minimum temperature (tmin), and m...

Full description

Saved in:
Bibliographic Details
Published in:Earth systems and environment 2024-12, Vol.8 (4), p.1159-1177
Main Authors: Bilbao-Barrenetxea, Nerea, Martínez-España, Raquel, Jimeno-Sáez, Patricia, Faria, Sergio Henrique, Senent-Aparicio, Javier
Format: Article
Language:English
Subjects:
Algorithms
Annual precipitation
Bias
Climate
Climate change
Climate Change/Climate Change Impacts
Climate models
Decision making
Earth and Environmental Science
Earth Sciences
Earth System Sciences
Environmental Science and Engineering
General circulation models
Geography
Hydrology
Impact analysis
Learning algorithms
Machine learning
Monitoring/Environmental Analysis
Original Article
Precipitation
River basins
Rivers
Simulation
Stream discharge
Stream flow
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This study employs machine learning algorithms to construct Multi Model Ensembles (MMEs) based on Regional Climate Models (RCMs) within the Esca River basin in the Pyrenees. RCMs are ranked comprehensively based on their performance in simulating precipitation (pr), minimum temperature (tmin), and maximum temperature (tmax), revealing variability across seasons and influenced by the General Circulation Model (GCM) driving each RCM. The top-ranked approach is used to determine the optimal number of RCMs for MME construction, resulting in the selection of seven RCMs. Analysis of MME results demonstrates significant improvements in precipitation on both annual and seasonal scales, while temperature-related enhancements are more subtle at the seasonal level. The effectiveness of the ML–MME technique is highlighted by its impact on hydrological representation using a Temez model, yielding outcomes comparable to climate observations and surpassing results from Simple Ensemble Means (SEMs). The methodology is extended to climate projections under the RCP8.5 scenario, generating more realistic information for precipitation, temperature, and streamflow compared to SEM, thus reducing uncertainty and aiding informed decision-making in hydrological modeling at the basin scale. This study underscores the potential of ML–MME techniques in advancing climate projection accuracy and enhancing the reliability of data for basin-scale impact analyses.
ISSN:2509-9426
2509-9434
DOI:10.1007/s41748-024-00408-x