This year's south-west monsoon will be normal, says the India Meteorological Department in a forecast issued recently. But that is also what the IMD predicted last April for the 2009 monsoon, which then turned out to be a severe drought. This time the odds favour the IMD forecast being correct. Atmospheric scientists usually define a normal monsoon as one where the nationwide rainfall from June to September is within 10 per cent of the long period average for the season; a rainfall deficit of more than 10 per cent is taken as a drought. By that definition, the south-west monsoon has been normal 70 per cent of the time over the last 130-odd years. There has been a drought in only about 16 per cent of the years. There have been droughts in two successive years on just three occasions. Moreover, thus far whenever a drought involved a rainfall deficit of 20 per cent or more, as happened last year, the subsequent monsoon has invariably been normal.
India was the first country in the world to embark on operational seasonal forecasting. The first such forecast for the south-west monsoon was issued over a century ago, on June 4, 1886, drawing on the inverse relationship between Himalayan snow cover and monsoon rainfall. Since then, for its operational seasonal forecasts, the IMD has relied on various empirical models that took into account statistical relationships between various atmospheric and oceanic parameters and the monsoon. But published research that looked at the forecasts issued from 1932 to 2004 came to a dispiriting conclusion. There has not been, the scientists noted, “any improvement over the years, in spite of the continuing attempts to revise the operational models based on rigorous and objective statistical methods.” Besides, the IMD has not been able to predict a drought in its seasonal forecasts. Now, the Ministry of Earth Sciences has proposed launching a National Mission on Monsoon that will seek to develop reliable dynamical models for seasonal forecasting of the rainy season. Such models try to reproduce the complex interplay of processes that go on in the ocean and the atmosphere. The existing dynamical models, developed principally in the U.S, Europe, and Japan, have been good at forecasting the progress of the monsoon a few days in advance and shown considerable skill at predictions for up to 10 days. But they perform poorly in seasonal forecasts and fail to accurately reproduce the monsoon's year-to-year variability. Improving the models will require better understanding of the monsoon's links with processes that happen elsewhere. Then the models must be suitably modified to incorporate the new knowledge. All this is going to take both time and effort. But such an undertaking could offer the best hope for better seasonal forecasts.