The Hindu Business Line : S-W monsoon likely to be below normal

APRIL indeed is turning out to be a cruel month. Not just due to summer having gripped the country with all its fury— but because the India Meteorological Department (IMD) has in its long range forecast painted a picture of "below normal rainfall" in the South-West Monsoon season, from June - September this year, a five per cent, margin of error, notwithstanding.

According to the 8-parameter probabilistic model, the 2003 South-West Monsoon Rainfall shows a total 60 per cent probability of dismal rainfall - including a 21 per cent probability of drought and a 39 per cent probability of below normal rainfall.

According to this model, there is a 14 per cent probability of near normal rainfall, 23 per cent probability of above normal rainfall and three per cent probability of excess rainfall.

Announcing the forecast, about 40 days before the traditional announcement on May 25, IMD's Director-General, Dr R.R.Kelkar, told mediapersons that the new forecast models provided them with a longer lead time.

According to the long-range forecast, based on the newly-adopted 8-Parameter Power Regression Model, the South-West monsoon rainfall in 2003 for the country was likely to be 96 per cent of the Long Period Average (LPA), thus falling in the "below normal rainfall" category. The model had an error margin of five per cent.

"Another new 10-parameter model, which uses data up to June, will make it possible to generate a quantitative forecast up to mid-July. By that time, the nature of advance of the monsoon would be known and any disturbing trends in the El Nino phenomenon would be apparent," he said.

Unwilling to give a date for the onset of monsoon, he, however, added that since July rainfall was critical for agriculture - a new model was built for forecast of rainfall in this month. The model, based on 38 years of data, has an error of nine per cent, "because the review period is small, one month," he clarified. IMD will issue a forecast for July along with a long range forecast update in mid-July 2003.

Dr Kelkar said that last year's forecast had an error of 20 per cent. He admitted that statistical models had an inherent weakness. "Statistical monsoon prediction models are based on strong correlations of the monsoon rainfall with certain antecedent atmospheric, oceanic and land parameters. These correlations can never be 100 per cent. Another weakness is that statistical models assume that correlations remain constant in future, while they may and in fact do, change with time and lose their significance."

Meanwhile, elaborating on the five new Long Range Forecast models developed by the IMD for the South-West Monsoon in the country, he said that, the 8-parameter Power Regression Model, requiring data up to March, would replace the existing 16-parameter model, which required data up to May.

Article E-Mail :: Comment :: Syndication

Stories in this Section
S-W monsoon likely to be below normal

AirTel cuts rates

Are foreign lines being wooed to bid for SCI?

HLL posts 8.22 pc rise in first-quarter net

Truckers strike may cast its shadow on Maruti

`Britannia gains most from World Cup promos'

As strike begins to bite — Truckers invited for talks today