You seem to have a fascination for models on diverse subjects. Does your background have something to do with this?
I'm basically a chemical technologist. I have co-authored a book on polymer science -- a rather popular book, I would say. As far as my personal background is concerned, I believe one's background or expertise is not a hindrance in branching out to other areas. Take the area of chemical compounds. In the 1940s, the doubling time for the discovery of compounds used to be 10 years. That has now come down to a year and a half. If you now have 30 million known chemical compounds, in a year, there will be 50 million. In such a situation, it is difficult to say you've specialised in any one thing.

I think this century is essentially for those who will be able to integrate different kinds of knowledge. I'm never happy being called a specialist and the secret of my inspiration lies in my holistic perspective on science.

Are you suggesting that expertise is being discounted in an effort to integrate the sciences?
I don't think the old classical concept of expertise in a particular area is valid any more. In India, 55 per cent of what you think is already there, the person who provides the leadership only provides the remaining 45 per cent. He brings in an understandable, cohesive focus, and he classifies it in a manner that already exists. So it's possible for someone with a little background to do anything.

How did you get interested in the monsoon model and how did you go about constructing the model?
When I became secretary in the department of science and technology, I was interested in the Indian Meteorological Department, which had very capable scientists. I found that 90 per cent of the things were already there. It was really a question of somebody who was not conditioned by the same mind-set taking the lead. So I got involved, I interacted with a lot of people to find out how all their information could be converted into a model.

I found that work on physical assessment of the monsoon process had already started. Some 200-odd parameters, which seemed to affect the monsoon process, had been identified. After largescale screening, we found there were certain parameters that were not totally independent, but seemed to effect the monsoon process as a whole. That's how we got the 16 parameters. People used to think a mild summer would be followed by a bad monsoon. Certainly, that particular parameter had some effect, but it's not as if every parameter has the same effect on the monsoon. So we formulated a model where each of the 16 parameters were given a certain weightage. And this is what allows a precise forecast.

A common criticism of monsoon forecasts has been that they are inaccurate. How do you respond to this?
The criticism is mostly aimed at short-range forecasts. The monsoon model talks of forecasts for 10 days and beyond -- after 10 days, it's a long-range forecast. For short-term forecasts one has to depend on synoptic data -- radar and satellite pictures. But in a country the size of India, the scene changes very fast and we don't have the kind of comprehensive radar coverage that is used abroad to provide a very different report.

The medium-range forecast is tricky and requires a really fast computer with a large memory and database. Medium-range forecast is still in the experimental stage in India. Short-range forecasts are much better, but still leave a lot to be desired. Long-range forecasts, where the monsoon model comes in, is the one that is receiving the maximum accolades because it has been fairly accurate. What prompted you to bring out a model on population growth? What are its fundamental assumptions?
While people have written lots of books on science, development and population, I was always curious about the interconnection between them. If we explore these connections deep enough, we can come up with some action possibilities. Of course, one has to keep in mind the reality: how do you face it, how are you going to influence the future of generations to come? That is what interests me.

The first thing I did while constructing the population model was to take into account all the elements that affected population growth. I found that the death rate started declining after 1941. In 1991, we reached a single-digit figure of 9.8 deaths per 1,000. But the number of live births per 1,000 also dropped from 36.9 in 1971 to 33.9 per 1,000 in 1981 to 29.9 in 1991.

My projection is that the birth rate will come down to 21 in 2001 and the death rate to 8.1 forecast that the natural growth rate six-and-a-half years from now would be around 1.3 per cent; the annual exponential growth rate (1991-2001) will be around 1.42 per cent, a steep decline from the present 2.14 per cent.

This means that India has already started moving swiftly to the final stage of Demographic transition. With a birth rate of 21 per 1,000, a death rate of 8 per 1,000 per year and a natural increase of 13 per 1,000 (birth rate- death rate), India will reach the threshold of the Net Reproduction Rate of 1 within a decade from now. At this time the bench mark figure will appear mark figure will appear and India's population will begin to move towards stabilisation at the billion plus mark.

You have based your model on census data that many people feel is inaccurate.
Honestly, I don't know how accurate the census data is. But there are some important parameters that everyone assumes are correct, such as the birth rate, death rate or the infant mortality rate. I don't think anybody really disagrees with these figures because the census is the only organisation across the country that generates such data. I have based my work on these four parameters, which, to me, are very important. I don't think lack of accuracy in the census data affects the accuracy of my model.

Isn't this contradictory?
I will reiterate that inaccuracy of census data does not affect my population model because when you are dealing with a large set of figures, minor deviations or inaccuracies do not affect your calculations. Moreover, the 1981 census is far more accurate than earlier census reports.

Can you quantify that it is actually the middle class that is contributing to population growth? The general notion holds the lower social strata responsible.
I must admit my model has tried to quantify only aggregates. I've not tried to find out the respective income groups that have been responsible for population growth. However, I feel this notion may be right to a certain extent. There is a growing feeling that the expanding middle class in India may have registered a significantly higher rate of population growth in the last 20 years.

What are the main development concerns in population?
Once we understand growth trends, we can plan. No development can be possible without population control. Modern technology and biotechnology is adequate to meet population concerns. Even with the minimum level of inputs, it is possible for us to curb population growth in India. Access to services is extremely important and the government needs to introduce incentive schemes to couples who plan their families.

If what you say is true, why has it been so difficult to check India's growing population?
By minimum input -- in this case, the stabilisation of population growth -- I don't mean small effort. A small but critical input can have significant impact. I am not suggesting we have not made significant inputs, but we've always worked under the assumption that we have to bring down the rate of population growth. The first thing we must do is stabilise population growth. Only then we can talk of curbing rate of growth.

I see you have a bias so far as numbers are concerned. You seem to be blinded by numbers?
I don't think I have a bias towards understanding of the population problem. In fact, I do not even suffer from the usual bias of a demographer. But I certainly feel that every developing country must make a conscious effort to curb disturbing trends in population growth.