‘Hand over the Ministry of Agriculture to a politician from the arid Deccan plateau; someone who has known hunger, distress and seasonal migration. And make a political commitment to revamp agricultural science and markets in these regions.’ This was the reply from an eminent scientist, in 1991, when Prime Minister Narasimha Rao asked him what he could do ‘to transform dryland agriculture and bring prosperity to poor villages spread over ten states.’ In the words of this eminent scientist, the meeting ended on a sad and helpless note. ‘Our agricultural research system¹ has become too rigid and set in its ways. Even if markets change, we in the sciences may not change. We are blinded by our green revolution euphoria. We can no longer see reality.’ Little has changed since then. Highlighting the bleakness is the ICAR’s Vision 2020 – the absence of the word ‘drought’ and an exclusively crop-based strategy to address rainfed agriculture!²

This essay explores the inability of agricultural S&T to understand and respond to the demands of dryland agriculture. Though there are several problems that are generic to agricultural science and technology, some of these are specific obstacles to the generation and utilization of relevant knowledge for the resource-constrained agriculture of the drylands. If scientific research is to make a significant contribution to dryland agriculture, there must be (i) a broader and more inclusive definition of knowledge and issues relevant to livelihoods in the drylands, and (ii) a new paradigm of institutional learning and change, where science and technology will become part of a wider coalition of actors, inputs and processes. To ensure sustainable rural livelihoods, the public sector, in partnership with other actors where feasible, has to launch a new knowledge initiative for the drylands along with other investments in infrastructure, natural resource management and human resources development. This is as much a political challenge as a technological one.

 

Historically, formal scientific research for dryland agriculture was organised to serve the colonial interest in dryland crops of commercial value – cotton, indigo, sugarcane, wheat, lac, and so on. Whether it was experimental research in the country or transfer of technology (especially varieties and new implements) from the West to India, these efforts were conducted by individual researchers or in experimental farms funded by the government or by Commodity Boards using a specific fund (based on a cess levied) for research. Some of the early successes in rainfed rice, sugarcane and cotton came from these experimental farms. All these different research organizations under the provincial governments, commodity boards, quasi-public/private farms, etc. were taken over and consolidated within the ICAR in 1965.

Contrary to popular perception, dryland agriculture in India even today is marked by a variety of actors/organizations and processes. In the public sector, the ICAR institutes and the State Agricultural Universities conduct research for the drylands. The ICAR has specific research organizations – the Central Arid Zone Research Institute (CAZRI), the Central Research Institute for Dryland Agriculture (CRIDA), and parts of the Indian Grasslands and Fodder Research Institute (IGFRI), Central Soil Salinity Research Institute (CSSRI) mandated to conduct research on specific crops and problems in the drylands. The International Crops Research Institute for the Semi Arid Tropics (ICRISAT) is located in India, and several donor-funded programmes focus on research for dryland agriculture (such as the Aga Khan Rural Support Programme in Gujarat, or DFID’s programmes in the western and eastern regions).

Private and cooperative sectors are also active in R&D in the dryland farming tracts, especially in sectors such as private plant breeding and seed production. Over 40 private firms breed and sell certified seeds of dryland crops (especially cotton, pulses, oilseeds, maize and sorghum). Some of the cooperatives have been very successful in dryland production systems like milk (Amul and other state dairy cooperatives), sugarcane, grapes, and other fruits. The cooperatives, more adept at collecting and adapting technologies and information for their own farming community members (than in conducting research/technical experiments), have proven how different ways of communication and delivering information can enable better technology adoption.

 

Farmer-to-farmer extension is a major source of technological information as well as new technologies (seeds/crop varieties, pesticides, crop protection practices, irrigation practices etc.) in the drylands. Traditional wisdom and local cultural knowledge seems to be critical for survival in the drylands.

Despite the involvement of several national and international players and organisations, the overall performance of formal R&D for dryland agriculture has been far from satisfactory. R&D investment has been low, and the use of that investment (as we shall discuss here) has been highly ineffective for dryland agriculture. Measures such as research intensity (R&D investment in rupees per unit of commodity/crop produced) are misleading. They show that dryland crops (pulses and oilseeds, and some coarse cereals) fare better than irrigated crops (especially wheat, rice, and sugarcane) (Kerr, 1996). But this is clearly because of the massive production and productivity differences between dryland crops and irrigated crops. Nor do these measures give us an adequate picture of the research effort that needs to go into dryland agriculture. Moreover, measuring research intensity adds to the myth that increased investment within existing research organizations and paradigms is in itself a solution.

 

The arid and semi-arid tracts in India contribute over 45% of the national agricultural production. Rainfed agriculture in 100 million hectares of arable land is the manifest opposite of the prosperous irrigated cereal-based green revolution tracts. It is accepted that the green revolution has bypassed millions living in the drylands, surviving on one rainfed crop a year – which is bound to fail or is not taken at all if the monsoon is poor, erratic, delayed or absent (ICAR, 1998).

The ICAR’s classification of the country into 20 agro-climatic zones includes rainfall, soil type and length of growing period as the defining features of each zone. Using rainfall and access to irrigation as defining features of dryland agriculture, all agro-climatic zones with the exception of five zones, are either completely or partially dryland agriculture zones. The need for a different typology to develop dryland/rainfed agriculture using S&T inputs becomes obvious considering the range of complexities associated with dryland agriculture that extend beyond agro-climatic conditions and include socio-economic realities of these tracts.

 

Dryland agriculture is marked by risk and coping strategies to avoid risk. These strategies include diverse crops and cropping systems, dependence on livestock and other non-farm rural income, limited cash crop production (where assured markets exist), common property based livelihoods or access to livelihoods, low input use and technology adoption. Household income in a typical arid/semi-arid village is subject to a high degree of seasonal variation. Seasonal out-migration, lasting four to six months every year, is routine. These villages are also marked by relatively high proportion of scheduled caste and scheduled tribe households (traditionally disadvantaged to access modern input intensive agricultural technology), low or negligible rural infrastructure investments and services, high incidence of morbidity and malnutrition, illiteracy and limited access to education, all revealing multiple dimensions of poverty. In terms of crop production the rainfed cropping systems also reveal very high inter-farm variation within the same village.³ 

Thus, dryland agriculture in India is marked by several constraints (technological and non-technological) and for a full understanding of these, one needs to go beyond the narrow technical criteria usually adopted to define drylands.

India’s experience with generating technological solutions for technological problems reinforces the definition of dryland agriculture in narrow technical terms. The green revolution is premised on the definition of hunger and malnutrition as food production problems. Thus the national goal of food self-sufficiency is translated into an applied research strategy to produce technological solutions. Risk proofing through assured irrigation, chemicals and fertilizers, and high yielding varieties was the hallmark of the green revolution technologies. Applied research within the key disciplines now focuses almost exclusively on technology generation to assure this green revolution paradigm – albeit within the narrow range of controlled conditions.

Agricultural research addresses the problems of dryland agriculture assuming that the same norms of irrigated agriculture can be applied here. Attempts to understand dryland agriculture and solutions offered are always technology centred – crop-based (varieties), input-based and irrigation based (water use efficiency, water saving technologies, etc.).

 

Several scholars lament the inadequate attention paid within the crop sciences to the rainfed crops – especially the more stress tolerant and nutritious coarse cereals and minor millets/pulses/oilseeds. Even in cases where such attention to coarse cereals is given, varieties are screened in multi-location trials and assessed using average yield/productivity criteria from test plots across the country. This means that varieties released are often not ones that survive the extreme stress levels of the drylands. It also means that the local adaptation levels of the varieties are not a concern when assessing and releasing crop varieties – though inter-farm variability in yields even within one village is a major issue in the drylands.

Farmers in the dry tracts look for higher yields, but as one among many other traits like grain to fodder ratio, crop duration, seed quality, drought tolerance and pest resistance. The rainfed rural economy has adapted to the inherent instability in crop yields by cultivating diverse crops, engaging in several livestock based and other non-farm options. Often fodder yield is considered more crucial than grain yield – especially in the dry villages subject to severe seasonal stress.⁴ Plant breeding as a discipline has limited capacity to understand the varietal selection and livelihood options in the drylands.

 

Even soil science research is assessed using the criterion of yield enhancement; the knowledge inputs and conservation/resource management services from soil science notwithstanding. Given the lack of assured irrigation, fertilizer use and other chemical/pesticide application is limited in the drylands. Besides lack of information about use and availability, there is also a question of access to these inputs (including access to credit and repayment capacity estimated against incremental income even if the technology is for medium or long term natural resource management). Agricultural science knows precious little about the choices that farmers make for land management in the drylands – even share cropping (in villages where entire families migrate for over four-six months every year) or crop-livestock relationships are analysed using the same technical and economic criteria that are used in irrigated areas.

Agricultural science has limited capacity to conduct experiments along the scales that determine input use and resource management in a dryland farm. Watershed development activities have shown that households in a village can be made aware of the layout of a watershed and availability of groundwater and appropriateness/impacts of input use at different points in the watershed. But the soil sciences lack experimental designs (and partnerships with local stakeholders) to integrate technical information to assess soil quality (capacity for nitrification, porosity, and other structural features), along the watershed contours.

As a rule, agricultural research is capable only of plot cultures or plot level on-station/on-farm experiments (even for ritualistic fertilizer trials).⁵ Even in dryland R&D organizations with some on-farm research in addition to the on-station research (CRIDA for instance), the emphasis is on taking on-station research design or results to the farm. There is no attempt to learn with the farmer or to understand local topographic or edaphic factors. The lack of communication and learning across disciplines within the agricultural sciences is an affliction in dryland agricultural R&D too. Communication among sub-disciplines of the same discipline (say the sub-disciplines of soil science) is minimal or non-existent.⁶ 

 

Dryland agriculture is defined by lack of assured irrigation; yet scientific research for the drylands is governed by the norms of irrigated agriculture. R&D assumes that enhancing irrigation in the drylands will solve all the problems. The availability of water and the diversity of water and land use patterns (be it the absence of irrigation in the arid Rajasthan tracts that produce high grade mehndi for export, the tank irrigated rice production systems in South India, or the fast-depleting-groundwater-irrigated groundnut and sugarcane systems of Gujarat/Maharashtra/western Madhya Pradesh) ought to have conveyed the message to agricultural science that irrigation and green revolution type production enhancement technologies are not the answer to dryland agriculture. Even if all the irrigation potential is used, 40% of India’s food requirements will still need to come from the dryland tracts.

The recent spate of farmer’s suicides in these areas proves how the green revolution package dependent on high levels of water use, purchased inputs, and reliable markets does not even comprehend the limited resource availability and high inter and intra-seasonal variability that characterize dryland agriculture. Sadly, rural development schemes and private investments that encourage the exploitation of groundwater for irrigation of water intensive cash crops in the drylands, rely on ‘technology packages’ specially developed by agricultural research organizations, and fed into these schemes by subject matter specialists.

Water management, especially integrated water resources management is the answer for the dryland tracts. Successful cases of NGO or other civil society led watershed management programmes reveal how farmers and other water users in a watershed can agree about and effectively manage their water resources and crop/livestock production systems through appropriate processes, investment strategies, benefit sharing mechanisms, conservation technologies, production technologies, etc. The distinguishing feature here is the absence of a single point ‘technological’ solution as found in irrigated crop production. Here technology, be it irrigation development or plant protection methods, becomes part of a more comprehensive package of processes and inputs.

 

The agricultural sciences in India are so organized as to insulate its disciplines from learning lessons from the field in collaboration with (a) other disciplines –social sciences, natural resources research, animal husbandry, among others, or (b) other non-research partners – civil society organizations,line departments, private firms, financial organizations and farmers organizations. Opportunities for interdisciplinary research and working in collaboration with other partners in the drylands declined rapidly following the consolidation and centralization of public sector agricultural research – especially given the norms for evaluation of research based on peer reviewed research publications.

Dryland agricultural research is not accountable to a local constituency – be it research in the public sector or private sector. Only the cooperatives seem to have made a mark in agricultural knowledge systems in the drylands. Will the future of dryland agricultural research bring new organizational formats and new partners? The emerging message is one of inadequacy of the existing paradigm of R&D to address the knowledge and development needs of the drylands.

 

The approach of delivering technological solutions to development problems that are chiselled down into narrow technical problems (using convenient criteria) is universal. In the case of dryland agriculture, even the scientific solution itself seems to be limited by the narrowly defined ‘technical’ criteria (of rainfall, soil type or growing period) leaving out any reference to other ecological (status of water table/land degradation – erosion, salinity) or socio-economic criteria. Conventional neo-classical agricultural economics and extension science in the research organizations is structured in such a way as to impede the entry of new paradigms that bring better understanding of social contexts and have sustained the conventional linear R&D models and technology generation functions in R&D organizations. At this current juncture, dryland agriculture demands that the social sciences lead agricultural R&D by helping the technical sciences understand rural realities in these tracts.

There are several examples of agricultural and rural innovation (defined as the socially embedded utilization of technologies and other knowledge) that offer important lessons for the agricultural sciences in the drylands. Agricultural research should:

(a) Shift mandates from research to innovation: Given the complexity of production conditions in dryland agriculture, the emphasis has to increasingly shift towards utilization of knowledge in economically and socially productive ways. To ensure successful innovation, the researchers and research organizations should be encouraged to find appropriate coalitions to work in, with local accountabilities and mandates designed to serve local contexts. For them research output or even the choice of research problem should become one among several steps in a flexible innovation programme, to achieve the desired development outcome.

(b) Improve its capacity to learn from local contexts: The shift of emphasis towards innovation presupposes a change in the attitude of researchers and research organisations towards local knowledge. They should have the humility to learn from local contexts before attempting to intervene in them.

Agricultural science needs to understand that agriculture is one among several sources of livelihoods for rural households in the drylands. Unless these other livelihoods options, their relationship with agriculture, local contexts and systems are understood, and a diverse range of stakeholders involved in innovation, technologies and institutional arrangements necessary for development cannot be identified and enabled. The attitude of learning from local contexts also should make the researchers mindful of the multiple impacts of their actions on the local environment.

(c) Build new relationships and collaborations with other actors: All the cases of successful innovation and development evident in the drylands (cheap drip irrigation systems, integrated pest management, watershed development, etc.) reveal several actors working synergistically as an interested and committed coalition, improvising and nuancing their interventions with a deep respect for local knowledge – whether it is from input dealers, industry, extension staff/ departments, or illiterate risk averse farmers.

The foregoing discussion clearly shows that there is an urgent need to change the rules/norms governing the generation and application of knowledge in the drylands. The paradigm of linear knowledge flows and technological determinism has to be changed and the new norms/institutions for a learning community need to be laid out. More funds or new organizations can help dryland agriculture only if accompanied by such a change in approach and in the rules of the game.

 

The answer presently favoured is to privatize everything – presumably because reforming failed public sector is considered a hopeless task. There are powerful forces at work, which attempt to transfer research and extension from public to private/voluntary agencies. While accepting that increasing collaboration among public, private and civil society actors is important, the key role of the public sector in research and extension needs to be re-emphasised. In fact, the public sector has to play several roles besides generating technology. It has to play a dynamic monitoring and planning role to get these collaborations to work in a resource poor dry-land context. It also has to ensure that knowledge is used as part of a wider package of inputs and processes for sustainable development in drylands, with a great deal of flexibility in ways of working and choice of partners in different contexts.

 

Footnotes:

1. The Indian agricultural research system is composed of (i) the Indian Council of Agricultural Research (ICAR) with its National Institutes, Central Research Institutes and National Research Centres and Coordinated Programmes, (ii) the State Agricultural Universities – now 32 in number, (iii) a few general universities and some institutes under the Ministry of Commerce, or the Ministry of Science and Technology, in the public sector, several private/corporate sector input (seeds, fertilizers, agro-machinery, chemicals and pesticides) producing and marketing firms, and some voluntary organizations (though there are thousands, only a few are dedicated to conducting research for/in agriculture and rural issues).

2. The document does mention that ‘(D)espite occasional droughts, famines have been averted.’ ICAR, 2000, p. 6.

3. In India, ‘interfarm variation in yields in predominantly rainfed villages is often greater than 50%, compared to 20-30% in irrigated areas’ (result quoted in Kerr, 1996, p.60).

4. In such villages entire families migrate for months, leaving the aged and invalid to care for a cow or a few goats using limited stored dry fodder.

5. Within soil science (with several sub-disciplines), it is the sub-discipline ‘soil fertility and fertilizer’ that makes for nearly half of the total number of soil science publications (Raina et al, 2006).

6. For example, an experiment in ‘soil fertility and fertilizer’ to assess ideal green manure and fertilizer combinations for a dryland crop (bajra or sorghum), is of little use if it does not learn from soil physics, pedology, soil survey and mapping, to account for the physical, biological and agro-ecological features of the soil resources in the region, and the history of land use in the farm/village.

 

References:

ICAR, NATP Main Document, Indian Council of Agricultural Research, New Delhi, 1998.

ICAR, Vision 2020, ICAR, New Delhi, 2000.

J.M. Kerr, ‘Sustainable Development of Rainfed Agriculture in India,’ EPTD Discussion Paper No. 20, IFPRI, Washington, DC, 1996.

R.S. Raina, S. Sangar, R.V. Sulaiman and A.J. Hall, ‘The Soil Sciences in India: Policy Lessons for Agricultural Innovation’, Research Policy 35(5), 2006, 691-714.