Optimizing Climate Change Projections: A Comparative Analysis of Model Ensembles in the Rio Grande Basin in the SW US.

Abstract Submission: ABSTRACT
New Mexico, located in the Southwestern United States, is among the region’s most vulnerable to climate change. Historically characterized by its arid desert climate, the state now faces increasingly unpredictable weather patterns. Projections suggest prolonged droughts and reduced precipitation during the wet season, posing significant challenges for water management, particularly in the Rio Grande Basin (RGB), which supports major cities and vital reservoirs such as Elephant Butte and Caballo. This study examines future climate scenarios in the RGB, stretching from southern Colorado to El Paso, Texas, using state-of-the-art projections from the CMIP6 dataset, sourced via NEX-GDDP. We focus on three key climate pathways: SSP1-2.6, SSP2-4.5, and SSP4-8.5. Fifteen General Circulation Models (GCMs) were selected based on their historical performance and were bias-corrected using the Empirical Quantile Mapping (EQM) method. Three multi-model ensemble approaches were employed to enhance the accuracy of climate forecasts: Arithmetic Averaging, Bayesian Model Averaging, and an Optimization method leveraging genetic algorithms. Our findings indicate that Bayesian Model Averaging and the genetic algorithm-based Optimization method outperform Arithmetic Averaging in generating reliable multi-model ensembles for climate projections. This research not only advances climate modeling methodologies but also provides critical insights into water resources management for one of the most climate-sensitive regions in the US.
Keywords: Climate Change projections, General Circulation Models, Southwest USA, Rio Grande Basin, Multi-model ensemble, Genetic Algorithms

Learning Objectives/Expected Outcome (Optional) :