Multimodal climate change prediction in a monsoon climate

Abstract

The uncertainty in the climate projection arising from various climate models is very common, and averaging such results poses a risk of underestimation or sometimes overestimation of impact in magnitude and frequency. Further, the performance of various climate models in monsoon degrades drastically due to the skewed nature. Under these circumstances, the performance of the climate model in the monsoon and non-monsoon periods is critical for accurate assessment. A multimodal approach has been used in the present work to quantify the uncertainty involved in the climate model using reliability ensemble averaging (REA). Based on AR6 of IPCC, the ensemble of 26 global climate models (GCMs) was used to evaluate the model performance and possible change in seasonal precipitation in four cities with distinct climate conditions, namely, Coimbatore, Rajkot, Udaipur, and Siliguri. The results show that non-monsoon and monsoon rainfall are expected to increase in all the regions. Most of the models perform poorly in simulating monsoon climate, especially in the monsoon period and are highly inconsistent spatially. The study also finds that the model performance is largely linked to the ratio of natural variability and mean. HIGHLIGHTS The paper discusses multimodal performance in climate change assessment in a monsoon-fed climate where the precipitation pattern is skewed.; Model bias for each climate model was quantified on a seasonal scale using Reliability Ensemble Averaging.; The performance of the GCMs largely depends on the ratio of natural variability to mean at regional level.; A reliable estimate of the change in climate variables is made using REA in CMIP6 framework for four climate scenarios, namely, SSP126, SSP245, SSP370, and SSP585.;

Keywords

• ar6
• climate change
• multimodal analysis
• reliability ensemble averaging
• uncertainty