Dealing with drought

In the absence of water resource planning, management and conservation measures, modern methods of forecasting monsoon will be of little use to drought-prone regions.

R. RAMACHANDRAN

IN meteorological terms, a drought is "a sustained, regionally extensive, deficiency in precipitation". All other definitions are related to the effect or impact of below normal precipitation on water resources, agriculture and social and economic activities; hence the terms hydrological drought and agricultural drought. In quantitative terms, the definitions could vary among countries and regions. In India, the definition for "meteorological drought" adopted by the Indian Meteorological Department (IMD) is a situation when the deficiency of rainfall at a meteorological sub-division level is 25 per cent or more of the long-term average (LTA) of that sub-division for a given period. The drought is considered "moderate", if the deficiency is between 26 and 50 per cent, and "severe" if it is more than 50 per cent.

Based on this definition, the National Commission on Agriculture has given the following broad classifications:

* Hydrological drought: prolonged meteorological drought resulting in depletion of surface water from reservoirs, lakes, streams, rivers, cessation of spring flow and fall in groundwater levels causing severe shortage of water for livestock and human needs;

* Agricultural drought: when soil moisture and rainfall are inadequate during the crop growing season to support healthy crop growth to maturity, which situation causes extreme crop stress and wilting. It is defined as a period of four consecutive weeks (of severe meteorological drought) with a rainfall deficiency of more than 50 per cent of the LTA or with a weekly rainfall of 5 cm or less during the period from mid-May to mid-October (the kharif season) when 80 per cent of the country's total crop is planted, or six such consecutive weeks during the rest of the year.

It is interesting to look at the rainfall data for the last couple of years in the regions affected by the current drought - west and east Rajasthan, Gujarat region and Saurashtra, Kutch and Diu, Telengana, Rayalaseema and coastal Andhra Pradesh (see table). In all these regions, the deficiency in the southwest monsoon, which accounts for nearly 90 per cent of the rainfall and hence is the main source of water, was itself not all that severe. However, the monsoon rainfall last year was below normal in all the regions. In fact, the deficiency of monsoon rainfall in regions of Andhra Pradesh was not much at all.

Though the immediate post-monsoon (October-December) rainfall was above normal in all the regions except Andhra Pradesh, the winter (January-February) and pre-monsoon (March-May) rainfall during this year has been severely deficient. In the Gujarat region, the rainfall deficiency has been prolonged and whatever little impact the above-normal post-monsoon rainfall may have had has been undone by the highly deficient winter and pre-monsoon rainfall. However, in regions of Andhra Pradesh, post-monsoon rainfall - whose LTAs are actually substantial unlike in other regions - was deficient and so has been the pre-monsoon rainfall this year. But winter rainfall was above normal though this accounts only for a small fraction of the annual rainfall. In any case, in Andhra Pradesh too, the overall deficiency in rainfall has not been alarming enough to be the sole reason for the current drought.

From the perspective of these data, therefore, in the meteorological sense, one may say that it is a case of "severe meteorological drought" only in the Kutch region. Indeed, rainfall deficiency is much less than what it was in the worst drought years of 1987 and 1974. But the ground situation is that the drought is much more widespread in the regions of Gujarat and Rajasthan, which are experiencing a severe "hydrological drought", if not a "meteorological drought". Some quarters of the officialdom and the media have, however, used the rainfall data alone to deny the drought conditions.

The current drought actually is even more serious than if it was a case of severe meteorological drought leading to a severe hydrological drought. It points to the fact that withdrawal of water from the various resources of the region - which get recharged only by rainfall - has been unsustainably high in recent times leading to a drastic depletion of these resources. The situation is a reflection of a complete failure in water management and a lack of planning in these water-scarce regions.

This hydrological drought may result in a severe agricultural drought as well if the monsoon turns out to be deficient this year too for these regions. Preliminary data on some of the relevant meteorological factors up to April do seem to be unfavourable for a good monsoon and even statistically, given the string of 11 "normal" monsoons, the probability for a below-normal monsoon for the country as a whole is high. And, given that these arid and semi-arid regions of western India are drought-prone, deficient rainfall in these regions is indeed likely again this year.

Long-range rainfall forecast - in terms of spatial and temporal distribution during the monsoon season - therefore assumes importance. The IMD does have a 16-parameter statistical model that relates the seasonal rainfall for the country as a whole to the values of 16 meteorological parameters during the months up to April, and this is the forecast model that has claimed success in prediction of rainfall behaviour since 1988. However, its prediction is of the average over the entire country. But historically, the average for the country has been normal two-thirds of the time because inter-annual variability of rainfall in high precipitation regions such as northeastern India is small and that in low precipitation regions such as Gujarat, Saurashtra and Rajasthan is very high. That is, even if there is a drought situation in the arid and semi-arid regions of western and northwestern India, the monsoon would more often than not be normal.

The model's utility as an operational tool for region-wise agricultural and resource planning purposes, particularly in the drought-prone regions, is, therefore, minimal. What is required is not such a prediction in the gross but over a sub-division or cluster of sub-regions with homogeneous rainfall. But such a forecast model, which can be operationalised, does not exist as of date although experimental models have been many. The most notable among them is the one developed by B. Parthasarathy and associates at the Indian Institute of Tropical Meteorology (IITM), which uses groupings of contiguous sub-divisions to give five homogenous rainfall regions. The belief is that over a single homogenous region, the predictability is vastly increased.

However, last year, for the first time the IMD used its 16-parameter model to give long-range forecast for three homogeneous regions. Unlike Parthasarathy's more refined grouping, IMD's grouping harks back to 1935. From then until 1988, when the new 16-parameter model came into being, regional forecasts, using methods that used smaller number of parameters, for these three broadly homogeneous regions were given. This was discontinued after the arrival of the new model. The separate forecasts has now been revived using the apparently more sophisticated new "power regression model".

However, last year's forecast was quite off the mark even in the gross, let alone at the regional level. Against a countrywide forecast of 108 per cent of the LTA, the actual rainfall was only 96 per cent of the LTA. But it was yet a normal rainfall because of the wide 10 per cent window in the definition of "normal". At the regional level, northwestern India received 94 per cent of the LTA against a prediction of 111 per cent, peninsular India received 90 per cent against a prediction of 114 per cent and northeastern India 104 per cent as against a prediction of 98 per cent. The forecast for the current year, issued on May 25, has once again attempted to give regional forecasts. But four of the 16 parameters have been changed in an apparent bid to rectify last year's major failure of the model. The rainfall for the country as a whole has been placed at 99 per cent of the LTA and for the three broadly homogeneous regions of northwestern, peninsular and northeastern India at 102 per cent, 98 per cent and 100 per cent of the LTA respectively.

However, as against the above forecast by the IMD, the CSIR's (Council of Scientific and Industrial Research) Centre for Mathematical Modelling and Computer Simulation (CMMACS) in Bangalore has predicted a below-normal monsoon for this year. It is based on a neutral network long-range forecast model that has been recently developed by P. Goswami and associates. Based on a 123-year data set, they were able to predict the 1999 rainfall accurately and forecast a below-normal monsoon for this year. They believe that the model is workable for sub-divisional forecasts as well including the drought-prone areas of Rajasthan and Gujarat.

Long-term forecast in these arid and semi-arid areas is rendered difficult by the high variability of rainfall from year to year, which is as high as 40 per cent. Indeed, D.R. Sikka, former director of the IITM, who has recently published a detailed report on monsoon drought in India, is of the view that the definition of drought should be region-specific and linked to the rainfall variability. With a definition of 25 per cent deficiency, such a high inherent variability implies that the probability of the occurrence of "meteorological drought" in these regions is very high. He also believes that long-range forecasting of monsoon rains at sub-divisional scale may not be feasible because of this high variability. In his analysis, even within a sub-division, the rainfall varies greatly from district to district and, within a district, from block to block, implying a largely convection-driven precipitation in these sub-divisions, which will not be amenable to forecast.

Even if the IMD's prediction for the regions that northwestern India is likely to get reasonably good rainfall is to be believed, it does not assure that a drought will not occur. Last year too, northwestern India received 94 per cent of the LTA and yet there was a drought. This only underscores the point that regional forecasts too will not serve to prevent droughts unless appropriate "hydrological" measures are effected.

The upshot of these arguments is that monsoon forecast, even over homogeneous regions, is of little use to these semi-arid regions and a drought should always be expected. Therefore, long-term water resource planning, water management and water conservation measures assume greater importance in preventing hydrological and agricultural droughts. The fact that even without a severe meteorological drought there is widespread hydrological drought points to a signal failure in this regard. In the Indian context, water being a State subject, the issues of planning and water management become all the more complex.

WHILE there is a National Water Policy (1987), which gives the highest priority to water for drinking purposes, in terms of water management policies and long-term planning, there is a tendency to put too much stress on one type of technology and there is a huge preoccupation with large dams within the government, points out Shekhar Singh of the Indian Institute of Public Administration (IIPA), who has served in the advisory capacity on several government committees on water management and planning. Even small hydel schemes have now been moved to the Ministry of Non-conventional Sources from the Ministry of Water Resources because no one is interested in them, he says.

Indeed, in the wake of the present drought there is renewed pressure from various quarters to complete the Sardar Sarovar Project (SSP). It is true that the Indira Gandhi Canal has improved the conditions in east Rajasthan greatly. But, given the present estimates of water availability in the SSP, the Narmada canal is unlikely to reach water to Kutch. The sugarcane growing areas along the path of the canal will consume enormous amounts of the water. Water conservation measures, demand management, allocation of water for different uses and involvement of communities and local participation in water management have never been given adequate importance. There is no system in place to meet crisis situations by planning for the worst-case scenario, Shekhar Singh points out.

Almost 90 per cent of the drinking water needs are met from groundwater but only 5 per cent of the total groundwater extracted is needed for domestic use. Irrigation accounts for 85 per cent, and the remaining 10 per cent is used by other sectors, including industries. The rapid increase in groundwater-based irrigation using tubewells that are more than 150 metres deep and high-power pumps in recent times is one of the many reasons for groundwater depletion and the drop in the groundwater table. The area covered by groundwater-based irrigation has increased from 6.5 million hectares (mha) in 1950-51 to 40 mha. An annual decline in the water table of up to 2 m is considered normal and can cope with even a deficient rainfall the following year. A decline of up to 4 m is a cause for concern and above 4 m is a stress situation. The Central Ground Water Board (CGWB) describes the current situation as a stress situation. However, there are reports that unregulated overdraft has resulted in water tables dropping 10 to 20 m in several areas in Gujarat.

Technology development in deep drilling and pumping methods (even up to 250 m), on the one hand, and populist policies of free or low-cost electricity for irrigation purposes, on the other, have contributed greatly to this massive exploitation of groundwater for irrigation purposes. Coupled with the highly anomalous policy that vests rights to groundwater with the owner of the land, there is no legal limit on the volume of water that a landowner can extract. Unregulated exploitation by rich farmers with large land holdings, who can afford large pumps, leave the poor and the marginal farmers without access to even basic water needs. Landlords become waterlords, resulting in inequity and social conflicts in rural communities.

Of late, several industries have begun to exploit groundwater in an unregulated manner, says Shekhar Singh. The industries are located where groundwater is indicated and groundwater gets consumed within the industrial premises for different purposes, including dilution of effluent waste. This rapid groundwater depletion has not only reduced water availability but also affected its quality in terms of excess fluoride and arsenic content and brackishness. This has made State Public Health Engineering Departments (PHEDs) abandon low-cost hand pump-based systems and instead undertake costly and complicated piped water schemes that exploit whatever little surface water that exists in these areas.

There is clearly a need to regulate groundwater extraction so that an adequate share of it is used for drinking water and irrigation purposes. Groundwater is pollution-free, is not brackish and is not lost due to evaporation even in arid regions. But nearly 90 per cent of groundwater exploitation occurs through private hands. A model Bill to regulate and control the development of groundwater has been circulated to all the States since 1970. But, except Maharashtra, no State has enacted this legislation. Gujarat, which faces severe problem of non-availability of fresh water, does have some legislation in place to regulate groundwater use. But recurrent droughts indicate that it is not enough. Implementing and monitoring groundwater use need to be strengthened. Similarly, there have been no moves to notify certain areas where groundwater extraction for private use is forbidden.

The basic problem is, therefore, one of a rapid localised overdraft of groundwater in excess of recharge of subterranean aquifers. Rainwater is the basic source of this recharge potential. This had to be conserved through proper conservation measures and recharging means and techniques such as check dams, percolation ponds and rainwater harvesting. There is a provision in the Central funding to States for adopting recharge technologies and measures.

According to S.K. Sharma of the CGWB, there is a massive programme on paper for recharging and several location-specific artificial recharging methods have been evolved. For instance, some successful techniques have been demonstrated in the drought-prone regions of Gujarat and Rajasthan. But these measures have not been replicated elsewhere in the State, and the money allocated gets diverted to other schemes. The PHEDs invest in costly surface water supply instead of community and location-specific conservation measures. Corruption in purchase of equipment for such projects is a major factor, particularly during drought situations, according to Shekhar Singh.

THE National Drinking Water Mission launched in 1986 as part of Sam Pitroda's concept of Technology Missions, which was renamed in 1991 as the Rajiv Gandhi National Drinking Water Mission (RGNDWM), has the primary goal of providing safe drinking water to all rural habitations in the next five years. The failure of this mission at least in problem areas has come to the fore with recurrent droughts in arid and semi-arid regions. The problem, according to Shekhar Singh, who was part of the mission's review committee in 1996, is in the typically bureaucratic preoccupation with numbers and targets of handpumps, piped water schemes installed and habitations covered, rather than ensuring their viability and sustainability. What has resulted is what the mission approach wanted to avoid - namely, bureaucratisation. Now, there is a separate department to manage the mission with little coordination with other Ministries and departments.

What has happened, according to Shekhar Singh, is that these techno-fixes have resulted in the conservation ethos in the local communities being slowly eroded and traditional catchment, conservation and groundwater harvesting methods such as 'tankas' , 'kundis' (tanks), 'phads' (system tanks), 'bawdis', 'jahalaraos' and wells being abandoned in favour of piped supplies that exploit the meagre surface water resources. The RGNDWM projects are funded to an extent of 75 per cent by the Centre and 25 per cent by the States and 20 per cent of the funds are earmarked for recharging measures. Shekhar Singh says that while money for the tubewells gets picked up by the States, the funds for harvesting or recharging projects seldom get used. He feels it necessary to dovetail the mission schemes with traditional systems and empower as well as fund directly local panchayats, which would have a sense of ownership and responsibility to ensure a proper operation and maintenance of these systems.

Rainwater harvesting, which depends only on the average annual rainfall in the region, has been particularly ignored as a conservation measure in these arid regions on the apparent argument that it has little role in arid (25 per cent of the country) regions and semi-arid regions (48 per cent of the country) with an annual rainfall of less than 50 cm. It is well established that Israel, with an annual rainfall of about 23 cm, has a sustainable system based on rainwater harvesting. Also in the desert regions of Australia, with an annual rainfall of 7.5 cm, run-off has been brought down to 4.5 cm. Consider a place like Jaisalmer, which has an annual average rainfall of 10 cm. Over an area of 1 ha or 10,000 square metres, the total amount of water received is one million litres. Even with a low collection efficiency of 0.25, about 10 to 15 five-member families can meet their water needs at about 15 litres per capita a day. There are traditional and modern improved methods for efficient rainwater harvesting, which can be adopted for location-specific environments and needs. Given the increasing scarcity of water and frequency of calamities, rainwater harvesting may prove to be the only feasible solution in the rural areas.

Technology-based water resource management measures such as desalination through reverse osmosis (RO) membranes, defluoridisation and arsenic removal technologies, water prospecting through terrain resistivity measurements face the other major problem of lack of technical expertise and manpower. Besides the CSIR's pilot demonstration units, the desalination programme has been largely contracted out to private companies based on imported equipment. These apparently have tended to close shop and abandon the equipment once they recover their investment. The CSIR technology has not spread because of a lack of domestic manufacturers of the thin film composite (TFC) membranes on a large scale except Bharat Heavy Electricals Limited (BHEL) which has now put up a small plant. This, and a similar Defence Research and Development Organisation (DRDO) technology, were designed for brackishness up to 10,000 ppm of total dissolved solids (TDS) as compared to imported membranes, which accept higher concentrations of even up to 35,000 ppm of sea water.

Satellite-based remote sensing technology too has been brought to bear upon water resource management. The National Natural Source Management System (NNRMS) of the Department of Space (DoS) collaborates with the Central Water Commission (CWC), the Water Resources Ministry, the CGWB and the Ministry of Rural Development to monitor water availability as well as to delineate potential geological sites for groundwater prospecting.

Remote-sensed data is also being collected for the Ministry of Rural Development's Integrated Mission for Sustainable Development (IMSD) and for the RGNDWM. But the implementation of these plans often depends on the initiatives taken by the respective District Collectors, says D.P. Rao, Director, National Remote Sensing Agency (NRSA). According to him, wherever the plans have been implemented and check dams and other water collection and recharge means established, there has been a marked improvement in water availability. In the overall water resource management for the country, satellite data can serve only as indicators; they need to be supplemented by ground-based surveys and exploration to identify exploitable sources precisely. The NNRMS/DoS also has its own standing committee on water resources. But fructification of these efforts have been slow because of inaction on the ground level by State governments.

SATELLITE-BASED remote sensing also offers a quick and effective means of drought surveillance. One way to minimise the impact of droughts is to monitor continually the conditions in drought-prone regions, particularly in the wake of poor rainfall, and provide early warning to prevent or minimise the effects of an impending hydrological and agricultural drought. The NRSA has put in place the National Agricultural Drought Assessment and Monitoring System (NADAMS) developed for the Department of Agriculture and Cooperation. The system uses the Advanced Very High Resolution Radiometer (AVHRR) satellite of the National Oceanic and Atmospheric Administration (NOAA). This satellite has a large swath (about 2,700 km) and daily coverage as against the 22-day revisit cycle of the Indian Remote Sensing satellite (IRS) series. The NOAA, however, has a coarser resolution as compared to the IRS satellites. Also, the IRS' Wide Field Sensor (WiFS) has a higher dynamic range than the NOAA. The NRSA has evolved methodology that incorporates the IRS' WiFS data (with 188 m resolution) and supplements the NOAA data in drought monitoring under the NADAMS.

This drought monitoring and early warning project is being carried out jointly with the CWC, the CGWB and the IMD. Under this project, which has now been operationalised, vegetation index (V.I.) maps, which are based on the concept that vegetation vigour, is an indication of water availability or lack thereof, are prepared to use the NOAA data. This involves vegetation known to absorb strongly in the visible light but little in the near-infrared. The difference of visible and near-infrared reflectance represents photosynthetically active vegetation, which information is used to construct a V.I. Moisture stress in vegetation, resulting from prolonged rainfall deficiency, is reflected by the lowering of the V.I. Combined with a surface water index, the V.I. provides a clue to the onset of hydrological and agricultural droughts. The NADAMS is currently providing crop and seasonal condition reports at district and sub-district levels through the kharif season in 11 agriculturally important and drought-prone States.

In the context of the current drought, D.P. Rao stated that by October 1999, it was clear that plants were under stress owing to lack of water in several districts of Andhra Pradesh, Rajasthan and Gujarat. The NADAMS had accordingly sent drought warnings to State governments. Since these regions do not receive rain from the northeast monsoon, the situation could only get worse. This is a contingency which could have been foreseen but unfortunately appropriate actions were not taken by the State governments.

Indeed, warnings of impending drought were always there - from a meteorological perspective, from an agricultural perspective and from a water management perspective; from people on the ground as well as from satellites in the sky. For such drought-prone regions, the authorities ought to have been prepared for a crisis situation every year instead of adopting fire-fighting methods year after year. The fact that in spite of a string of good monsoons the country is faced with a calamitous situation calls for a public enquiry, and responsibility needs to be fixed, says Shekhar Singh. Otherwise, this year's misery and loss of lives will soon be forgotten.