Getting monsoon forecasts right is essential for Asia's third-biggest economy, of nearly $2 trillion. Agriculture, employing more than half India's population of 1.2 billion, contributes about 15 percent to this figure.
India's 235 million farmers rely on monsoon rains from June to September to water more than half the farm area in the world's second-biggest producer of rice, wheat and sugar, as only about 45 percent of arable land has irrigation facilities.
"About half-a-dozen Indian, and almost an equal number of foreign, weather offices have joined hands to help us bring more precision in forecasting," L. S. Rathore told Reuters in an interview.
"We are hopeful of more accuracy in the monsoon forecast by 2017 with this collaborative research initiative," he added.
The IMD, which initially tied up with Britain and the United States, has also teamed up with the weather offices of Japan, Australia, and South Korea.
The department aims to deploy more radars and observatories on the ground and completely overhaul its mathematical model as the two main thrusts of its new programme, Rathore said.
The National Centre for Ocean Information Services, the Indian Institute of Tropical Meteorology, and the National Centre for Medium Range Weather Forecasting are among the Indian bodies involved in the collaboration.
The weather office now forecasts monsoon rains based on its long-term rainfall data for the past 130 years. It also considers ocean temperatures and atmospheric pressure.
There are two annual forecasts, one in April and the other in June, after the monsoon rains cover half the country.
The improved model would help the IMD issue four to five monsoon forecasts instead of two, and provide region-specific information as well, Rathore said.
"Almost all premier institutions of the world and India are involved in the exercise to take a look at the whole gamut of the meteorological process to sharpen our forecast," he said.
India has been able to accurately forecast monsoons, but finds it hard to predict extreme weather events such as drought, he said, adding that the new effort would end that handicap.
The IMD failed to foresee the worst drought in nearly four decades in 2009. As a result, rice output fell and India had to import sugar, sending global prices to a 30-year high.
"Precision in the monsoon forecast can raise farm production by 10 percent to 15 percent in rainfed areas," said Ravinder Singh, head of agricultural physics at the Indian Agricultural Research Institute.
Accurate forecasts would eliminate such sharp output swings as farmers would choose drought-resistant crop varieties and the government would make timely contingency plans, Rathore said.
The Indian weather office has twice gone wrong in its forecast during the eight years since 2005, in 2009 and 2007.
It predicted an average monsoon year in 2009 but the season ended with a rainfall deficit of 22 percent, making for the worst monsoon in nearly four decades.
In 2007, in contrast to the forecast of below average rainfall, the 2007 monsoon turned out to be above average.