Last week, officials at the India Meteorological Department (IMD) said that rainfall during the monsoon months of June, July, August and September would be less than what it anticipated in April. From a prediction that it would be 98% in April, the downsizing to 95% in June could in many ways be construed as a bold admission of the agency’s meteorological clairvoyance. The fact is that the Met department usually gives as optimistic an outlook of the weather even as ominous signs stare in its face. In 2009, it said rainfall would be normal and only admitted of a drought when it had firmly set in.
Given India’s widely variant rainfall, a 3% reduction—while certainly not apocalyptic—is still cause for discomfiture. IMD continues to rely on statistical models, which rely on a historical database of weather phenomena over India, to make predictions. It says that global dynamical models, which simulate existing meteorological conditions and crunch the data in supercomputers, in short the state of the art, aren’t good enough to be applied for forecasting the Indian monsoon.
Fair enough, but why then hasn’t it yet operationalized a programme, called Monsoon Mission, whose raison d´être is to develop dynamical models for India, especially when IMD itself admits that statistical models are passé?
This year,?too, it has revised its April estimate downwards less because of its faith in existing models and more due to a lack of predictive indicators in the Pacific and the Indian Ocean and a hunch that monsoon trends that have played out in the past, will play out this year, too. IMD’s prescience may well turn out to be right, but it would be more out of fortuity than conviction.
The department can’t be entirely blamed for the lack of good dynamical models as the equipment it needs, including automatic weather stations, Doppler weather radars and meteorological balloons are in short supply, and stuck in the complicated, constricting arteries of bureaucracy.
Unless these newer models are employed, IMD’s statistical tinkering will not matter to India’s sowing fields—60% of which are unirrigated—and where farmers continue to rely more on experience and prayer.
Weak models and weaker predictions are not about fewer rain drops alone. In India, this has a chain of economic consequences. Agriculture being the weakest link in the overall demand equation, lower rainfall means the government has to take remedial measures to pump prime the economy on a war footing or else 1 percentage point of growth has to be shaved off. If this is known in advance, the right steps can be taken at the right time.
Monsoon prediction: statistical dance or accurate model? Tell us at email@example.com