PUNE: The statistical models used by the India Meteorological Department (IMD) had failed to predict all the three droughts in India in the last decade. Though statistical models will still be used for monsoon forecast, the ministry of earth sciences is putting more emphasis on dynamic models.

M Rajeevan of National Atmospheric Research Laboratory said, "the failure to predict the 2009 drought has raised many serious issues. On the other hand, the state-of-the art coupled ocean atmospheric models have sho-wed improved skills in predicting inter annual variability of Indian summer monsoon rainfall."

He was speaking at the golden jubilee conference of Indian Institute of Climate Change (IITM), Pune, on 'opportunities and challenges in monsoon prediction in changing climate'. Since 2011, the IITM has used the coupled model for monsoon forecast.

Better weather forecast needs data from all parts of the globe. "In every part of the world, farmers are saying that the climate is not as it used to be. Hence, traditional knowledge is also failing. For better prediction of weather, we need observations from all countries. We need super computers of even higher capacities. We need to have knowledge about how to translate scientific progress into concrete applications," said Michel Jarraud, secretary general, World Meteorological Organisation.