In its first status report on the country's most-awaited yearly weather phenomenon, the nodal agency for long range forecast in the country predicted the monsoon rains to be 95 per cent of the long period average (LPA) with an error margin of five per cent. In other words, India can expect to get 84-cm rainfall, which is 95 per cent of 89 cm, the normal overall long period average for the country.

Basing its predictions on a brand new system involving newly-adopted statistical forecast models, the India Meteorological Department (IMD) report augurs well for the agriculture, navigation, hydel power and other water-based economic activities. But the point to remember is this is neither the last word (as the forecast is slated to be updated in June), nor is it of much value for prior planning for most of these sectors. The more useful information on the expected onset of the monsoon and the distribution of rainfall would be available much later, when it is already too late for remedial actions. In any case, the impact of the monsoon on agricultural growth has been ebbing over the years.

The annual June-September monsoon is vital to the economic health of the $885 billion economy because it provides the main source of water for agriculture, which generates more than a fifth of the country's gross domestic product. Farming and related activity provide a livelihood to more than two-thirds of India's 1.1 billion people, and good rains usually spur rural spending on a wide range of industrial products from soap to motorcycles. Bountiful rain helped the economy to expand at an average of over 8.5 percent over the past three years -- behind only China among major economies.

The Crisis

The Rabi season's contribution to the total production is at present roughly equivalent to that of the Kharif season. The last two years' experience is a case in point. Agricultural growth was an attractive 6 per cent in 2005-06, when the monsoon rainfall was 99 per cent. The growth dropped to 2.7 per cent in 2006-07, when the monsoon rainfall was even better at 100 per cent. This shows that while there is a requirement for a long-term forecast, steps must be taken to reinforce and give a face-lift to the monsoon prediction capability so that the region-wise and time-wise rainfall pattern could be visualised much earlier.

From that point of view, it is audible that the IMD has this year taken a couple of new initiatives. First, it has evolved and used two entirely novel prediction models even though these are statistical models that have in the past tended to slip up at crucial times. While the first model, using five global weather-related parameters, has been used for making the present long-range forecast, the second one-based on six parameters-will be used to refine this prediction and project rainfall in the critical month of July, as also in the four geographical regions of the country.

Secondly, it has initiated transition from statistical models to a dynamic forecast system in vogue in many other countries. No doubt, the development of such a system is a laborious task, especially when it comes to foreseeing a highly complex phenomenon like the monsoon, but the IMD seems to be inching closer to achieving that feat. As such, it has put together an experimental forecast based on the dynamic system which has also pointed to a normal monsoon. But, this system still needs to be tested further and validated before it can be put into service. For that matter, the new statistical models being used this year, too, have yet to be adequately validated. After all, the past models were also introduced with much hope in 2003 but had to be discarded because they failed to give correct predictions in two of the four years they were in use. Therefore, the IMD really needs to redouble its efforts to fine-tune the dynamic methodology to raise the credibility of its forecasts.