Inconsistency in historical simulations and future projections of temperature and rainfall: A comparison of CMIP5 and CMIP6 models over Southeast Asia

Abstract

The objective of this research was to assess the difference in historical simulations and future projections of rainfall and temperature of CMIP5 (RCP4.5 and 8.5) and CMIP6 (SSP2–4.5 and 5–8.5) models over Southeast Asia (SEA). Monthly historical rainfall and temperature estimations of 13 global climate models common to both CMIPs were evaluated to assess their capability to reproduce the spatial distribution and seasonality of European Reanalysis (ERA) rainfall and temperature. Models were used to determine uncertainty with spatiotemporal variability of rainfall and temperature projections. The CMIP6 GCMs did not appear to perform better than the older CMIP5 in SEA unlike other parts of the globe, except for rainfall. The CMIP6 models showed Kling-Gupta Efficiency (KGE) values in the range of -0.48-0.6, 0.21-0.85 and 0.66-0.91 in simulating historical rainfall, maximum temperature and minimum temperature compared to 0.13-0.46, 0.3-0.86 and 0.42-0.92 for CMIP5. The improvement in CMIP6 models in SEA was in the low uncertainty in ensemble simulation. The projections of CMIP5 and CMIP6 showed a relatively smaller increase in temperature with the CMIP6 ensemble when compared to CMIP5 models, while rainfall appeared to decrease. The geographical distribution of the changes indicated a greater increase in temperature in the cooler region than in the warmer region. In contrast, there was increase in rainfall in the wetter region and a smaller improvement in the drier region. This indicates increased homogeneity in temperature spatial variability, but more heterogeneity in rainfall, in the SEA region under climate warming scenarios.