| A man and his bike hitch a ride on a cyclevan after torrential rain flooded Calcutta in late September. File picture
New Delhi, Sept. 30: A new monsoon forecasting strategy adopted by scientists appears to have failed in its very first year with rainfall exceeding normal despite a 7 per cent below-normal prediction.
Data from the Indian Meteorological Department show that the cumulative monsoon rainfall from June 1 through September 26 was excess or normal in 30 of 36 meteorological subdivisions and deficient only in six.
The IMD had this year abandoned its decades-old but trouble-ridden method and used a new technique to predict that the monsoon rainfall would be 93 per cent of the average. But observations show that the actual rainfall until September 26 was 91.5 cm — about five per cent excess over the normal average of 87.5 per cent.
“It appears the new model has not worked. We are now trying to analyse why this has happened,” Madhavan Rajeevan, the director of the Climate Research Centre at the IMD told The Telegraph.
“There was excess rainfall in the southern peninsula and deficient rainfall in the northwest.”
The inaccurate forecast has surprised the weather science community because the new “ensemble-based” prediction method had been validated through scientific peer review before it was put into operation this year.
“We are still comfortable with the science behind the model,” Rajeevan said. “What we’re trying to understand is what happened this year.”
But the model performed somewhat better in predicting — qualitatively, if not quantitatively — the plight of northwest India.
The IMD long-range forecast in June had predicted 10 per cent deficient rainfall in the northwest.
The northwest remained largely dry with drought-like conditions in some regions: 40 per cent deficient rainfall in western Uttar Pradesh, 28 per cent below normal in Punjab and 33 per cent deficient in Haryana.
Ironically, the old model had predicted one per cent excess rainfall, closer to the actual 5 per cent excess. “But the result of a single year cannot determine the fate of any model,” a scientist said.
Scientists had decided to abandon the old model, in use since the late 1980s, because its predictions had proved unreliable in recent years.
Both models make use of statistical correlation between distant weather parameters — temperature, pressure and wind — and rainfall to predict the long-range behaviour of the monsoon.
The new forecast model had thrown up a below normal figure for the south. But rainfall has been bountiful — excess rain ranging from 32 per cent in Kerala to 40 per cent in parts of Karnataka to 101 per cent in Rayalaseema.