Next year, India’s annual summer monsoon forecast may be made by a supercomputer running a dynamical model.
The Secretary, Ministry of Earth Sciences (MoES), Madhavan Rajeevan, said the dynamical model, being tested at the Indian Institute of Tropical Meteorology, (IITM) Pune for a decade was “ready for operational purposes next year.” A dynamical monsoon model works by simulating the weather on powerful computers and extrapolating it over particular timeframes.
Though this method is normally effective in forecasting weather over a few days, using it to forecast the annual monsoon over 3 or 4 months has proved difficult.
While such models have been used for research purposes for long, the India Meteorological Department (IMD) has never integrated them into its operational forecast. “We hope to be able to launch it next year though discussions are still ongoing,” Mr. Rajeevan said, on the sidelines of an Earth Sciences Ministry foundation day event on Wednesday. The IMD, a division of MoES, is the official national weather forecaster. Mr. Rajeevan clarified that statistical models would still be in use next year. The IMD Director-General, L.S. Rathore, said, “We are ready to use the dynamical model, but this doesn’t mean one is abandoned for the other. Both have their role and we must use what’s best.”
The IMD relies on an ensemble model, a statistical technique that uses an average of six meteorological values correlated to the monsoon such as sea surface temperatures in the Pacific, and North Atlantic sea level pressure. These values are derived from century-old meteorological data linked to the historical performance of the monsoon.
Prediction failures
This traditional approach in recent decades has failed to predict monsoon failures — in 2002 and 2004 for instance — leading to calls by meteorologists for a new, modern forecasting system.
Though the dynamical model, called the Coupled Forecast System version 2, has so far achieved only 60 per cent accuracy in forecasting the monsoon, Mr. Rajeevan says it is good enough for now. “No doubt it needs to improve and the aim is to make that 77 per cent but we have to start somewhere,” he added. A confidence boost came when the dynamical model and the ensemble technique correctly signalled a drought in 2015.
While the IMD has for some years put out the dynamical model’s forecast along with the traditional one, its plans to give prominence to the dynamical model signals a new approach. This is a precursor to giving monsoon predictions over India’s 36 sub-divisions rather than only four broad geographic regions that encompass them. A dynamical approach can also be more easily tuned to account for rapidly changing global weather conditions.