The projected monsoon rainfall during these four months for the country as a whole is defined as normal as it is within the plus or minus 10 per cent of its long-period average, ie, 88cm. He said that as the El Nino was at its peak in the previous year and was now gradually waning since December, 1997, the second half of the monsoon would be better than the first half. Kelkar discounted the views of analysts that Indian monsoons are adversely affected by the El Nino factor. He said that Indian monsoons do not have a one-to-one correspondence with El Nino, though deficient rainfall has been observed in some El Nino years. About 16 factors influence Indian monsoons. This year out of the 16 factors, nine are found to be favourable as against eight last year.

The nine factors which are found to be favourable this year are the 50 hPa (20km) northern hemisphere wind pattern in winter, 500 hPa (6km) wind pattern in April, temperatures ruling central India in May, temperatures ruling the eastern coast in March, northern hemisphere surface pressure anomaly in January-April, Argentina pressure in April, North India's surface temperature in winter, the declining El Nino effect of the previous yearand the North Indian temperature in March. Amongst the unfavourable factors are the Eurasian snow cover in December, southern oscillation index in spring, Indian Ocean equitorial pressure in January-May, the current year's El Nino effect, Himalayan snow cover in January-March and Darwin pressure in spring.

The IMD's operational long-range prediction model uses 16 regional and global land-ocean-atmosphere parametres, which are physically related to the Indian monsoon rainfall. Each model parametre is defined in terms of observations over a specific location and period antecedent to the monsoon. The estimated model error is of the order of plus/minus 4 per cent. Additionally, the monsoon rainfall is influenced by concurrent events occurring during the monsoon season itself which cannot be fully anticipated at the time of generating the long-range forecast in May. The 16-parametre model of IMD has two distinct components. In the parametric model, the configuration of favourable and unfavourable signals fromthe 16 land-ocean-atmosphere parametres are analysed to arrive at a qualitative inference of the performance of the monsoon rainfall. The power regression model, which is essentially a statistical model, uses the numerical values of the 16 parametres and yields a quantitative estimate of the rainfall in terms of the percentage of its long-period average.