To Predict a Storm

Since the past decade or so, the climate everywhere has been see-sawing like nothing anybody has seen before. Extreme climate seems to be the law of the land because of climate change. Do we stand a chance to Mother Nature? Is there any solution to this crisis?

The World Meteorological Organization (WMO) and the United Nations Environment Programme (UNEP) established the Intergovernmental Panel on Climate Change (IPCC) in 1988. The IPCC, through its three working groups, publishes its report on a regular basis. Based on these reports, several climate centres around the globe develop climate models for various scenarios, known as Global Climate Models (GCMs). GCMs are the primary tool for accurate prediction and demonstration of the possible future impacts of climate change to policymakers. 

GCMs, however, have some drawbacks. The spatial resolution of the output of GCMs is quite large and cannot be used for the regional assessment of the climate of a small region. These models also show biased representations of observed time series on a spatial scale. The reliability of GCMs falls with finer temporal scales, and the uncertainty in the far-end future projection is also high due to natural variability, model uncertainty, and uncertainty in natural and anthropogenic aerosol forcing.

Therefore, several researchers have advocated the use of multimodal climate data to quantify uncertainty. The two widely used methods in multimodal climate assessment are Bayesian model averaging and reliability ensemble averaging (REA). 

REA is the preferred method as the weightage for each model is calculated based on model bias in replicating the present-day climate and the variations in the projection of different models.

One of the major challenges in using climate models is in a monsoon climate, as the model performance degrades drastically due to its skewed behaviour. Therefore, the quantification of the uncertainty in various climate models in the monsoon and non-monsoon periods, especially in a country like India, is essential for proper planning and management.

In this study, Prof. S Mohan and Mr. Akash Sinha from the Department of Civil Engineering, Indian Institute of Technology Madras, Chennai, India, have taken a multimodal approach to quantify the uncertainty of the climate model using REA.

Based on the latest IPCC report, i.e. the Sixth Assessment Report (AR6) report, the ensemble of 26 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.

It was found that the performance of the GCMs for the skewed distribution of rainfall is very poor, resulting in the broader projection band. Hence, the REA method used in the study predicts the most reliable estimate of the change in climatic variables and quantifies the uncertainty among various models.

It was also found that the performance of GCMs is largely linked to the ratio of natural variability to the mean of the climatic variable. This observation can be used to select the number of GCMs for climate studies for other regions. This study could be extended to other parameters, such as standard deviation, correlation, etc., in the model bias.

Dr. V. Jothiprakash from the Department of Civil Engineering, Indian Institute of Technology Bombay, Mumbai, India, gave his analysis of the work done by the authors with the following comments: “Climate change – the phenomena behind many natural water-related disasters around the world. Researchers all over the world are quantifying the effect of climate change, especially on rainfall, temperature, and runoff. The hydrology and monsoon in India have a large spatial and temporal variation starting from the highest rainfall region in the world (Mawsynram in Meghalaya, North-eastern India) to a meagre rainfall in Thar desert areas. Because of large variations in the climate over a large area, multi-modals were used to quantify the climate change projections. Several researchers reported climate change in India during monsoon months, but very few reported climate changes during monsoon and non-monsoon, over the country. Averaging the climate projects over a large space and temporally over monsoon and non-monsoon results is highly inconsistent, this is very much reported in the paper. This is a very useful paper to quantify the uncertainty estimation in climate change projections. The study uses the reliability ensemble averaging (REA) method to quantify the uncertainty involved in the climate model, in this study 26 global climate models (GCMs) were used and analyzed to get a better clarity of climate change over the Indian Sub-continent. The paper discusses in detail the climate change projections over four cities having varied climates namely, Coimbatore, Udaipur, Rajkot, and Siliguri. There is a good scope to extend the work to the entire country and study climate change and its effects and uncertainty during monsoon and non-monsoon separately.”

Article by Akshay Anantharaman
Click here for the original link to the paper


Leave a Reply

Your email address will not be published. Required fields are marked *