Researchers at the Indian Institute of Technology-Madras (IIT-M) have suggested the use of a multimodal approach to quantify climate model uncertainty. The approach offers reliable climate change estimates and quantifies uncertainty among models crucial for informed climate planning and management.
Globally, researchers in climate change rely on the reports published by the Intergovernmental Panel on Climate Change (IPCC) through its three working groups to develop climate models. The global climate models (GCMs) are primary tools for the accurate prediction and demonstration of possible future impacts of climate change to policymakers. However, these models cannot be used for the regional assessment of climate.
Several researchers have, hence, advocated the use of the multimodal climate data to quantify uncertainty. They point to uncertainties in future projections, natural variability, model uncertainty, and aerosol forcing complexities for this. The two widely used models are Bayesin model averaging and reliability ensemble averaging (REA), said S. Mohan, professor in the Department of Civil Engineering, who is leading the research.
The REA model is used to address critical challenges in climate modelling in monsoon-prone regions such as India, offering valuable insights for climate change projections, he said. Research scholar Akash Sinha is also part of the research.
The REA method harnesses model bias in replicating present-day climate and projecting variations, allowing researchers to calculate the weightage for each model. “The approach provides more dependable estimates of climate change variables and effectively quantifies uncertainty among diverse models, which is pivotal for making informed decisions in climate-related planning and management,” Mr. Mohan said.
Based on the latest IPCC report, the ensemble of 26 global climate models was used to evaluate the possible change in seasonal precipitation in Coimbatore, Rajkot, Udaipur, and Siliguri, cities with distinct climate conditions.
The study found that the performance of the models of the GCMs for the skewed distribution of rainfall is very poor. The REA method, however, predicted the most reliable estimate of the change in climatic variables and quantified the uncertainty among various models. The study has been published in the peer-reviewed Journal of Water and Climate Change.
