THE monsoon rainfall for the country as a whole has a fair degree of correlation with the total grain output (TGO). This should be so because while the vagaries of the monsoon affect the rain-fed kharif crop directly, the reservoir levels and soil moistu re conditions for the rabi crop are intimately linked to the performance of the southwest monsoon. Although the area under irrigation has increased four-fold since the 1950s, it is still only 36 per cent of the total cultivated area. The percentages of i rrigated area for foodgrains, oilseeds and pulses are only 40, 26 and 13 respectively. Hence rain-fed agriculture is still the dominant factor in crop productivity, and improving productivity in rain-fed areas has a direct impact on the country's economy .
Having said this, it must be pointed out that this correlation is only at a gross level because the vulnerability or sensitivity of individual crops to variations in rainfall varies considerably. The percentage variation in the TGO for 1 per cent deviati on of rainfall from the mean (elasticity) has increased from 3 to 5 per cent in the post-Green Revolution period. (This increase, according to agricultural experts, becomes apparent if the normative growth rate is factored out from the TGO.) That is, cro ps have become more vulnerable to fluctuations in rainfall. One of the factors is the impact of agricultural practices that have come to stay in the post-Green Revolution period, such as the use of high-yielding varieties, the increased use of chemical f ertilizers and pesticides and higher water consumption. In fact, the 10 per cent window for "normal" rainfall was based on this empirical rainfall-TGO relationship seen in the 1950s. From the perspective of the present increased elasticity, it is being a rgued that the window needs to be narrowed.
It stands to reason that the elasticity should be higher in semi-arid regions with low rainfall levels and vice versa. Also, in general the elasticity of kharif crops is significantly higher though over-exploitation of ground water could lead to higher e lasticity for rabi crops as well in the future. While the elasticity of wheat is low (understandably so because 75 per cent of the wheat crop is under irrigation), those of rice and pulses have shown an upward trend and that of oilseeds has shown a decli ne (largely as a result of a fall in productivity). The increase of rice's vulnerability to rainfall can also be understood in terms of greater water requirement. The proportion of rice output in the total kharif output has also increased. There is also the region-wise variation of sensitivity of foodgrains to rainfall as well as regional variability of rainfall itself. It is these various offsetting factors that continue to give a gross correlation of total rainfall to the TGO.
Given this gross correlation, while an all-India forecast is perhaps of some use to decision-makers, from a farmer's point of view it is of little use. For a proper application of agricultural strategies that can help the farmer, what is required is a fo recast of the spatial and temporal variations in rainfall at a regional level. What a farmer needs is a forecast a few days ahead of the onset of the monsoon - to make proper decisions on when to start the sowing operations - as well as indications of pr olonged wet and dry spells, which could have adverse impacts such as moisture stress, water-logging and loss due to pests and diseases. Unfortunately, the state-of-the-art model in monsoon forecast is still far removed from being able to make such foreca sts, particularly in the case of semi-arid regions where rainfall variability is high. Even with a skewed spatial distribution of rainfall, one can have a normal all-India rainfall. Conversely, a deficit all-India rainfall can still result in normal rain fall in a given region. For example, the correlation of rainfall distribution in a semi-arid region like Anantapur with all-India monsoon rainfall is not high.
So the forecast of a deficit all-India rainfall can be misleading and cause unnecessary concern among farmers. Indeed, this has happened to the groundnut farmers of Karnataka in the wake of a deficit rainfall (93 per cent) forecast by the Centre for Math ematical Modelling and Computer Simulation (CMMACS) of the Council of Scientific and Industrial Research (CSIR) for this year in contrast to the India Meteorological Department's (IMD) forecast of a normal monsoon (99 per cent). The incident is a pointer to the imperative need for close interaction of scientists with farmers to assist them with scientifically evolved farming strategies based on models of rainfall variability, so that they do not end up taking inappropriate measures.
In this respect, some research work has been done by Sulochana Gadgil and her colleagues at the Centre for Atmospheric and Oceanic Sciences of the Indian Institute of Science (IISc), Bangalore. An interesting concept developed by the team is that of 'Coh erent Rainfall Zones' for each State or region where rainfall characteristics are similar and well-correlated. Interestingly, these zones seem to be at complete variance with the meteorological sub-divisions and the agro-meteorological sub-divisions deli neated by the IMD. Used in conjunction with 'agricultural zones' - areas where historically similar cropping patterns are followed - this concept would help define an optimum agricultural strategy if rainfall forecasts, particularly of the wet and dry sp ells, can be made for these regions. Working with the farmers of rain-fed groundnut crops in the semi-arid regions of Karnataka, the group has been able to evolve farming strategies tailored to the rainfall variability in the region, to achieve optimum p roductivity.
The IISc team chose Pavagada and Sira taluks of Tumkur district, close to the border with Andhra Pradesh, near Anantapur. Although, based on their experience with the rainfall pattern in the region, farmers would have evolved appropriate strategies for t raditional crops, they are yet to come to terms with the high-yielding TMV-2 variety of groundnut, which they have used for the last 20 years. Inputs from the IISc scientists seem to have helped combat the adverse impact of the highly variable monsoon to a considerable extent.
The scientists used a mathematical model developed in the United States, called PNUTGRO, to correlate monsoon variability with peanut production. Basically, PNUTGRO correlates moisture in the soil with the peanut crop. From the data on rainfall distribut ion, they found that PNUTGRO could simulate reasonably well the peanut production with TMV-2 at the Anantapur Agriculture Research Station. Modifying this model to incorporate factors relating to pests and diseases (which too are related to moisture leve ls), they found that it was possible to derive Anantapur district's actual peanut production in different years. This successful experiment formed the basis for evolving farming strategies, which are now being implemented by the groundnut farmers of Pava gada and Sira with good results.
