At the meeting, Indian scientists recommended upgrading the country's supercomputer facilities, and developing or adopting appropriate high-resolution forecast models and data systems to improve short-range weather prediction and long-range monsoon prediction.
Other key recommendations from the meeting include boosting scientific capacity, fine-tuning existing models for dependable location-specific forecasts, and maintaining interaction with people using the services, especially farmers.
Demands for better forecasts have increased during the past decade, particularly from the agriculture sector, which forms the backbone of the India's economy. More than 70 per cent of the country's population earns a living through agriculture and needs accurate weather forecasts to plan when to sow seeds, irrigate and fertilise their fields, and harvest their crops.
"India has missed the global revolution in weather prediction," according to Jagdish Shukla, president of the US-based Institute of Global Environment and Society, who said India is ten years behind developed countries in terms of the accuracy of its prediction models, while its computing power is one-hundredth that in advanced centres.
Compared with the world's best computing models that give three-day weather forecasts with 80 per cent accuracy, Indian forecasts are accurate just 50 per cent of the time, said Shukla.
He added that the smallest area that India can accurately predict weather for is 100 kilometres by 100 kilometres; forecasts in industrialised nations can narrow the area of prediction down to one-hundredth of this size.
The existing system is severely hampered by a shortage of both qualified personnel and appropriate technology.
Indian researchers pointed out that moves to update computing equipment should be matched with increasing the number of skilled scientists to analyse the data and develop models.
This will take time, however. With most science students opting for the more lucrative information technology sector, fewer are pursuing careers in meteorology.
"It will take a minimum of five years to train a meteorologist if we begin today and another 10-15 years to meet the [staffing needs]," said Ramamurthy.
Stuart Bell, head of numerical modelling at the UK Meteorological Office, told the conference that India should invest in resources and computing capability to reduce the shortfall in trained scientists.