Logo image
Three Generations of NARCliM: Evaluation of Precipitation, Temperature and Their Extremes Over the CORDEX Australasia Domain
Journal article   Open access

Three Generations of NARCliM: Evaluation of Precipitation, Temperature and Their Extremes Over the CORDEX Australasia Domain

Fei Ji, Moutassem El Rafel, Giovanni Di Virgillio, Jatin Kala, Jason P. Evans, Julia Andrys, Eugene Tam, Stephen White, Dipayan Choudhury, Rishav Goyal, …
International Journal of Climatology, Early View
2026
pdf
NARCliM7.31 MBDownloadView
Open Access CC BY-NC-ND V4.0

Abstract

dynamical downscaling ET-SCI model evaluation NARCliM precipitation and temperature extremes
The NSW and Australian Regional Climate Modelling Version 2.0 (NARCliM2.0) is the latest generation of regional climate modelling project, building on the success of NARCliM versions 1.0 and 1.5. Evaluating and comparing the three generations is crucial to ensure scientific credibility, improve model skill, enhance understanding of climate processes, and provide reliable information for climate risk assessment and adaptation planning. This study presents the first comprehensive evaluation and comparison of NARCliM2.0 in simulating precipitation, temperature, and their associated extremes for Australia, identifying key improvements over its predecessors. We assess the skill of individual regional climate models (RCMs) in simulating mean and extreme climate, rather than focusing solely on the ensemble mean as per previous studies. Results show NARCliM2.0 substantially improves simulations of multi-model means of mean maximum temperature, precipitation and related extremes (e.g., TXx, TXge35, CDD, R99p), with notably reduced biases, particularly in southeast Australia. These improvements are largely due to advancements in RCM design, including enhanced model physics and higher resolution, resulting in lower domain-averaged absolute biases and greater value for climate impact assessments. However, NARCliM 2.0 shows limited improvement in simulating mean minimum temperature and annual minimum temperature (TNn), reflecting ongoing challenges in representing night-time processes. Overall, these findings provide valuable insights for future downscaling projects, emphasizing the importance of careful RCM configuration, model independence and continued development.

Details

Logo image