A systematic review of climate downscaling and extremes in Coupled Model Intercomparison Project 6 (CMIP6)

Abstract

The Coupled Model Intercomparison Project Phase 6 (CMIP6) will help enhance the understanding of future climate trends and extreme weather affecting agriculture and society. However, the accuracy of regional climate assessments is limited by low-resolution datasets and model uncertainties, necessitating downscaling and bias correction for better projections. Relevant literature was sourced from Scopus and assessed using the Protocol Search Appraisal Synthesis Analysis Report (PSALSAR) involving Population, Intervention, Comparator, Outcome, and Context (PICOC) for protocol identification and Preferred Reporting Items for Systematic Review (PRISMA) for article scrutiny. This review examined 278 articles, of which 39 % focused on extreme climate analysis, 31 % on performance evaluation, 13 % on downscaling with extremes, 11 % on downscaling, and 7 % on evaluation with extremes. Publications surged since 2020, with a 36.3 % increase by 2023. Approximately 68 % of the publications were in Asia, often evaluating model performance using statistical metrics such as root mean square error (RMSE), correlation coefficient (CC), and percent bias (PBIAS), with station data as a reference. This review further explores how global climate projections, specifically surface temperature increases across Shared Socioeconomic Pathways (SSPs): 1.8 °C (SSP1-2.6), 2.7 °C (SSP2-4.5), 3.6 °C (SSP3-7.0), and 4.4 °C (SSP5-8.5). Annual precipitation is projected to increase in both frequency and intensity, particularly in far future (2071–2100). Some studies have highlighted that the minimum temperature is expected to show greater increase, than maximum temperature across regions, leading to frequent extreme weather events. In contrast, the extreme occurrences are heterogeneous, with some regions expecting increasing trends and others decreasing, depending on their geographical condition. These findings highlight the significance of downscaling climate models, because their accuracy relies on geographical features, grid resolution, and reference datasets. Precise, finer-scale projections are needed for location-specific extreme weather predictions, aiding agricultural planning, and risk management.