New Delhi, April 16: The Met office sprang two surprises today — one rather nasty.
For the summer of 2003, it predicted below-normal rainfall for the second year in succession.
But the Indian Meteorological Department (IMD)’s annual monsoon forecast is a scientific achievement in itself since it has been made 40 days in advance. Till now, the forecast has been issued on May 25.
With last year’s drought amplifying concerns about monsoon behaviour, weather scientists have adopted a new forecasting method that allows an early forecast and leaves room for a mid-course correction.
The new long-range forecast says rainfall for the country as a whole in 2003 is likely to be 96 per cent of the long-period average. Rainfall between 90 and 97 per cent of the long-period average is classified as below normal.
In another deviation from custom, the IMD has also predicted a 39 per cent probability of below-normal rainfall this year, and a 21 per cent probability of a drought. This is the first time the Met office has assigned probability values to the behaviour of the monsoon.
“We wanted to have a model that can predict drought or no drought with reasonable accuracy,” said Dr Ranjan Kelkar, director-general of the IMD. When tested against past performance of the monsoon, the new model successfully predicted drought eight out of nine times.
Everyone, except the scientists who have created the model, would be hoping that its strike rate drops to eight out of 10 — that this year’s prediction would be wrong. Atal Bihari Vajpayee would be hoping a little more than others, perhaps.
A string of states is going to polls late this year, some — like Rajasthan and Madhya Pradesh — in drought-prone regions, which are still recovering from last year’s blow.
The drought led to a shrinking of agriculture, pulling the overall economic growth rate down: the forecast for the year just ended is only 4.4 per cent.
The unexpected drought last year had jolted the IMD and accelerated the search for better prediction techniques. The long-range forecast issued today is based on a new forecasting method devised by 25 IMD scientists in New Delhi and Pune.
It uses eight weather parameters — such as the Eurasian snow cover, Pacific Ocean temperatures and wind patterns — to predict the monsoon.
The IMD has abandoned the earlier 16-parameter model it had used since 1988. When it worked well, there was little motivation to replace it.
“It’s like not wanting to change a cricket team that is winning,” said Dr Madhavan Rajeevan, a scientist of IMD, Pune. But since 1999, the 16-parameter model has consistently delivered wrong forecasts, with actual rainfall 7 to 12 per cent below the predicted values.
The search for more reliable forecasts led the scientists to a set of eight parameters for which data is required only until March. This allows the IMD to issue the monsoon forecast in mid-April.
The new method will also enable the IMD to issue a forecast update around mid-July, making slight modifications if required. Until now, this was not possible. The long-range forecast covers the entire country and says nothing about regional variations.