A new artificial intelligence-backed algorithm has been created to help build the predictability of Indian Summer Monsoons (ISMR). This predictor discovery algorithm (PDA) uses a single ocean-related variable to facilitate a more skillful forecasting of the ISMR and make it possible to plan out economic decisions such as agricultural ones 18 months in advance. Scientists from the Institute of Advanced Study in Science and Technology (IASST), Guwahati, Indian Institute of Tropical Meteorology (IITM), Pune, plus Cotton University, Guwahati, have come to the conclusion that sea surface temperature (SST) is not adequately suitable for estimation of long-term forecasting of the ISMR. This discovery was only made when the endeavor was done using the predictor discovery algorithm.
The team formed the PDA which was able to project the ocean thermocline depth (D20) over the entire tropical region between 1871 and 2020, mapping this on to the correlation between ISMR and D20 over the same duration. Through this, the potential skill of ISMR was found to be as high as 0. 87 eighteen months before the onset of the ISMR season. The use of a machine-learning based ISMR prediction model was then utilized to make eighteen month lead forecasts of the ISMR between 1980 to 2011, the skill of which was found to be 0. 65.
Successful execution of this model was based on the machine-learning algorithm’s capabilities to develop relationships between ISMR and tropical thermocline patterns based on 45 climate simulations and observations undertaken between 1871 and 1974. Through this, the team was able to make educated predictions on the ISMR, and while the potential skill of the ISMR at this lead is 0. 87, there is still room for improvement in the model.
The ministry of science and technology made this announcement, and this success could help further the ability of scientists and economists to create effective models for long-term forecasting of the ISMR, especially with aid from artificial intelligence. As for the company mentioned in the article, the Institute of Advanced Study in Science and Technology (IASST) is an autonomous institute of department of science and technology (DST) located in Guwahati. The institute carries out research in science and technology and is part of the Public University system in India. As for the people mentioned in the article, Devabrat Sharma, Santu Das, Subodh K. Saha and B. N. Goswami are scientists who are part of the research team that created the new algorithm. Furthermore, they were also the ones who were able to make the successful 18-month lead forecasts of ISMR.