6.1. Operational needs
6.2. Specific Research Topics
6.3. The Watershed Experiments
The detailed implementation of these multidisciplinary programmes is not a subject for this paper. It is evident that the many participants in such NRM programmes must form consortia to work towards a common aim, as is now being done. Any scientist who has worked in such consortia is well aware of the effort, time and money needed to ensure proper planning, coordination and execution, and this aspect must not be dismissed lightly. The managerial and organizational aspects can easily ruin a large programme if they are not well designed for efficient running, careful planning, proper evaluation and final accountability (Tinker, 1994a). It is argued in the main report of this paper that another layer of management is not wanted, and this is certainly correct. However, control and coordination responsibilities must be located in one of the existing levels for each consortium, and this must be known to all participants. Some mechanism is also needed to ensure overall contact and coordination because of the extent of interconnection in the different parts of NRM work.
There is a strong consensus of opinion in the CGIAR Centres and in their many collaborating organisations that progress in NRM research is only possible if there is a close working relationship between the many disciplines involved. In particular the biophysical scientists and the socioeconomists engaged in research must collaborate at all stages of the research process, including problem identification, research design and execution, and evaluation of the results. The exchange of views must be a two-way process, in which the understanding and evaluation of the problem is fully shared. Biophysical solutions that are not economically and socially viable must be regarded as incomplete or inadequate solutions. Where biophysical solutions can only become viable in practice if policy changes are made, these results must be fed into the design of policy research. Thus biophysical and socioeconomic scientists must participate fully as members of the multidisciplinary INRM research consortia being developed to improve the use and conservation of natural resources.
A broader approach to sustainable agriculture and NRM research, including attention to all environmental factors, must be multi-faceted as well as multidisciplinary, and several different modes of research organisation will have to be used. It is not possible to be prescriptive, because of the variety of problems, and their location-specific and country-specific nature. There are, however, some general points that can be made. The first is that work must be truly multidisciplinary, as is argued at greater length above.
It is vital that the work is done in a fully participatory way with the farming communities. This has been well argued by Greenland et al., (1994) and others, and only needs endorsement here. Contact with farmers is needed to ensure that their real problems are well understood and that proposed solutions are relevant to their purposes. Secondly, the detailed operations of farmers must be analyzed to ensure that all possible ideas and opportunities are exploited in the research.
The linkage with local organisations, particularly NARSs, NGOs and universities is equally essential. It is simply impossible for CGIAR Centres to do all the necessary work themselves. Local organisations can contribute vital local knowledge, contacts with farmers, and on-the-ground presence to the research, as well as their professional skills.
It is envisaged that work on NRM will be organised in various ways. Firstly there will be projects in the core and complementary programmes of Centres, linked with particular issues, and probably often with an adaptive or applied aspect. Secondly, there will be work in Systemwide programmes, such as have been developing over the past few years within the CGIAR. Existing ecoregional programmes contain much NRM research, and the various elements of the Zschortau plan for Soils, Water and Nutrient Management research are being developed in a similar way (DSE/IBSRAM 1995). Thirdly, we are suggesting that integrated multidisciplinary experiments on watersheds should be used as the basis for a more specific approach to sustainable agriculture, on a small number of major sites. This triple level of experimentation should provide a solid and flexible system to meet all the significant problems of NRM in relation to soils and water.
Soil and water problems will continue to appear in the mainstream research of the Centres, and will be dealt with as part of these programmes. The specific ecoregional programmes that are now developing must often contain SWNM elements, as described above, and will need targeted soil and water programmes. It is recommended that, wherever possible, these projects should be classified or organised in relation to the watershed model discussed here, thus using it as a framework. That includes identifying the work by terrain type, by main processes and by outputs and impacts (in the sense of Figure 1). This will allow research efforts with a similar conceptual base to be grouped and reviewed together, so that general results and principles, successful and unsuccessful technologies can be identified and compared, despite the many site-specific differences that will certainly appear.
However, an underlying weakness in this use of the watershed approach, despite its advantages, is that sets of watershed data will be combined from different times and places, and are unlikely to be compatible. It is important that in a few sites a full watershed experiment should be set up, so that a coherent data set is available for a full analysis, and possibly a watershed model. Ideally, any agronomic or other intervention aimed at improved production in a particular land use unit can be tested for impacts on soil movement, river flow and water; quality within the same watershed, in different terrain types and land use units. The environmental impacts and hence the sustainability of the agronomic changes can therefore be proven in this new type of experimentation.
It is important that all this work should be done for clearly identified reasons, and with the best available concepts and techniques. Several soils and water research topics are listed below, which have present priority because of their potential for poverty alleviation, or for reducing impacts on the environment and enhancing sustainability, or because scientific developments suggest that these areas will become essential research subjects before long. This last may imply the strengthening of subjects in which CGIAR Centres may have no special comparative advantage now, but where it will be needed in the future.
A further rationale for addressing such needed research comes from World Bank experience. The Bank's evaluation unit (OED 1989, p. xix) in assessing required effort for increasing the relevance of economic and sector work in countries where resource degradation is judged to be a matter of critical importance (such as Nepal, Niger, Pakistan, Philippines, Sudan and Uzbekistan, to mention a few) concluded as follows: "it would be necessary to: (a) examine the national and, in some cases, the international dimensions of resource issues, the intersectoral linkages, tradeoffs and conflicts, and the interrelationship with macro policies; (b) assess what the government, non-government organisation and external aid agencies are doing to address the issues; (c) identify and evaluate a set of policy options to build on what is already being done; (d) discuss the institutional constraints frankly; and (e) suggest how the information base necessary to improve policies for resource management on a broad scale can be strengthened." (emphasis added). The CGIAR System can thus provide a direct contribution to bettering NRM globally by being a key provider of, and facilitator of the provision of, this information.
There are many aspects of soils and water research in which CGIAR Centres are well equipped and staffed, and which will be used to the full in the future escalation of SWNM research. In others it may be advisable for the System to take stock. If our suggestions for a stronger stress on watersheds and hydrology are taken up, skills in these subjects will be needed. It would be best to obtain these largely from existing centres of expertise in the first instance, until the need for in-house skills in the Centres can be judged. The topics listed here are traditional, in the sense that research on them has been continuing for many years. They are therefore all of central importance to the use of soils, and in all there are new developments that need to be taken up by all CGIAR Centres with appropriate interests.
(a) Soil organic matter (SOM) levels and properties are crucial to soil structure and fertility. These include the rate at which mineral nutrients are formed by SOM breakdown, the retention of some nutrient ions, and the improvement of soil structure and resilience. The faunal and microbiological processes that lead to stabilised SOM, the ways in which the properties of this may vary, its interaction with the mineral parts of the soil, and its rate of breakdown need more work, particularly in the tropical environment. This is essential for sustainability of soil use. The use of models to predict the behaviour of organic matter in different conditions and depths in the soil is developing rapidly. A world-wide network on soil organic matter studies and models (SOMNET) (Gregory and Ingram, 1996) has just been launched with CGIAR staff involvement. Socioeconomic work on the costs and opportunities of possible sources of organic matter that could be added to the soil, including gender aspects and alternative uses, adds to the research agenda.
(b) Soil biological relations also include the microbial and faunal relationships around roots, including plant growth promoting bacteria, soil-borne diseases, symbiotic and non-symbiotic N fixation, and mycorrhizal function (Tinker and Barraclough, 1988; Lynch, 1990). Research on these subjects is important for plant nutrition, soil-borne diseases and the possibility of biological disease control. The activities of soil fauna such as earthworms and termites are also important, and need more detailed study in terms of their function in the soil. Soil biology is undergoing a general resurgence of interest at present, and good collaborators in centres of excellence in the developed world should be available. It is a specialised, but extremely important, part of the general study of biodiversity.
(c) More cost-effective and socially acceptable methods of combating and predicting erosion are highly desirable, because of direct impacts upon the productivity of eroding fields, and the environmental damage caused downslope and down stream. The chance of finding genuinely new methods of control may not be high, but it is important that methods of predicting erosion in particular conditions are improved, so that the consequences of other changes in farming systems or technologies can be estimated at an early stage. Further understanding of rainfall erosivity is also needed. This is essential for sustainability of soil use, as erosion is generally a particularly irreversible process. Knowledge of the correct economic and policy circumstances in which technical interventions should be used are however vital for their success.
(d) New and improved ways of managing water for greater efficiency are essential, especially for rainfed agriculture in the semi-arid tropics, where it is often found that only a small proportion of the rainfall is actually transpired by the crop, for reasons explained earlier in this paper. This requires research on better management of crop and soil for maximum infiltration and water storage, and elimination of other constraints to rapid growth. Combined nutrient and water deficits need particular study to determine efficient ways of detecting and dealing with both. There is now a better understanding of how stomata function, and that they are influenced by chemical signals from the root, so that understanding of transpiration rate is improved.
(e) Better models of the movement of water, salt and other solutes such as plant nutrients are needed. These are used for controlling waterlogging and salinity in irrigated areas, and also for predicting water-borne disease risks. The models for nutrient movement is mainly to understand the losses by leaching under high-rainfall conditions, but are also relevant to eutrophication of watercourses. It is likely that most of this modelling work can be done by universities and other centres in the more-developed world, but adaptive research on their effective use would be necessary by local organisations.
More acceptable use of water also depends upon a better understanding of the decisions and incentives faced by water users, especially in public irrigation schemes and in the private exploitation of aquifers. Research on water pricing and allocation, water users associations and other socioeconomic aspects is therefore required.
(f) Better methods of managing soil nutrient fertility are essential for sustained productivity increases, because of nutrient export in produce and the continued shortening of the fallowing component of shifting cultivation cycles. This calls for the maximising of nutrient recycling, and the use of additional organic manures and inorganic fertilisers in controlled amounts tailored to the circumstances of the individual farmer and the different field situations. The decisions of the farmer will be strongly influenced by fertiliser/farm product price ratios, which depend on many market and policy-related issues. Where soils are already intensely depleted, the use of "one-off heavy applications of nutrients may be deserving of careful economic assessment and could turn out to be cost-effective in some circumstances (Sanchez and Izac, 1995)
(g) The improved and more efficient use of nutrients that is required needs applied studies on the rate of movement and reaction of nutrients in soils and of the dynamics of development of crop root systems, to determine how long applied nutrients remain potentially available for crop growth. It also calls for continued research on nutrient turnover in the soil through the microbial biomass.
(h) Acid soils cover large areas of the humid and subhumid tropics. The traditional method of curing this, by the application of liming materials, is often not practicable due to the absence of local sources of lime, and the consequently prohibitive costs. Thus savanna soils in Latin America can be used for pastures, but their productivity could be greatly increased if the acidity effect could be removed. This can be partly achieved by maintaining high soil organic matter levels, which sequesters the soil aluminium ions in acid soils, and partly by introducing pasture, crop and treecrop varieties that are tolerant to acidity and economically feasible. More research on both aspects is needed.
(i) The general subject of modelling is continuing to grow rapidly in importance, as computers become ever more powerful and simulation modelling is applied to a steadily growing list of topics. It may be that some CGIAR Centres are less well advanced than they should be in these applications, despite excellent work in particular Centres. This is, of course, essential for hydrological work, as has been emphasised in this paper, but it is also becoming standard procedure for the study of most soil and agricultural processes. This includes models for soil chemical and microbiological reactions, crop growth, transport in the soil profile and many others, including farm/household and community decisions and processes. Models can be misused, but ultimately they are the only way to determine how our understanding of component processes can predict the behaviour of complex systems.
(j) One area where there is most clearly a role for international centres and researchers is in developing better measurement of the status and trends of natural resources being managed under different regimes, to be better able to document trends and extents of change. This is being addressed in at least one initiative, namely that on Land Quality Indicators, underway between the World Bank, CIAT, and others. One of the areas where our knowledge is most deficient is in this domain of measurement concerning NRM and environmental consequences. A particular measurement theme still yet under-emphasised (notwithstanding such IFPRI Environment and Production Technology Division work as by Rosegrant and Evenson (1995)) is the careful assessment of changes in total factor productivity (TFP) at various levels of agricultural systems. Crosson and Anderson (1993) have argued strongly that TFP trends must serve as the main guideposts to policy work in NRM in general and NRM research in particular.
The NARS, which are to be assisted by the CGIAR System to become more effective in natural resource management, are likely to face particular difficulties in operationalizing such measures. There is little consensus within academic and practitioner circles on such measurements, which should themselves be linked in some systematic way to models of causation, so that the measures are not arbitrary but rather can contribute to an improved theoretical understanding of the complex dynamics within natural resource domains. One way of improving this linkage is by setting up a monitoring programme as an integral part of long-term field experiments, so that the change in soil or other parameters can be linked to performance and output. Environmental monitoring at all scales is a rapidly growing activity that is essential in maintaining environmental quality (Tinker, 1994b).
(k) One theme on which further socioeconomic research should be done concerns local government and decentralisation issues affecting NRM. There are strong advocates pro and con with regard to devolution of responsibility for soil, water and forest resources to local government bodies and/or user groups. Some consider this the only salvation, while others fear it will accelerate the degradation of ecosystems. Examples can be cited to support either view. What is not known are the frequencies of alternative outcomes or, more important, what conditions or objectives make either approach more likely to produce desirable (or undesirable) outcomes for NRM. A major research initiative on decentralisation launched at the World Bank (by H.P Binswanger and others) should also help shed light on these questions, several aspects of which are also already being addressed in IFPRI's research programme on "Property Rights and Communal Action."
(l) Some different suggestions concern how useful research might be done for the purposes of promoting better, more sustainable natural resource use. The International Association for the Study of Common Property (IASCP) is an exemplary organisation for bringing together people from a variety of disciplines and with different national and institutional backgrounds (including the CGIAR Centres).
In particular there appears to be a need to bring together socioeconomists working on soil degradation issues to identify priorities more sharply, in close coordination with biophysical scientists. The Zschortau Plan (DSE/IBSRAM, 1995) has set up a potential programme that is strongly focused on soil degradation mechanisms. The socioeconomic aspects of soil degradation are of comparable importance to the biophysical aspects, and it is important that socioeconomists get together to formulate their plans. There appears to be a need for a structure such as a Network for the Economics of Soil Degradation for socioeconomists, but working in close linkage with the biophysical groupings, to produce a coordinated approach to these multidisciplinary problems.
(m) A large amount of biophysical research has been done, in various contexts, with the aim of increasing crop and other enterprise productivity in a sustainable way. This is particularly relevant to the need to move from shifting cultivation systems to permanent cropping. It is desirable to further review this body of work, and to determine in which conditions it is believed that the problem has been solved in a biophysical sense. If the solution has not been adopted, the question should be asked whether farmers find it unattractive and it is not taken up for socioeconomic reasons. If so, the further question should be asked whether other and more attractive biophysical solutions have been sought, or whether socioeconomic interventions have been considered. This study, which requires no further field work in the first instance, will help to extract maximum value from work already done.
The watershed approach can be an organising principle for research prioritization and management on land use, soil or water; it can be used to organise large integrated experiments; it can be a planning tool for the development of watersheds; it can be an ongoing system for long-term management of a watershed. The present study is engaged only with research, but the subsequent use of the watershed approach for development or management would increase the value of research done in this way. This will allow a large-scale linkage of research results with policy and management, if the latter are also seen in a watershed context.
To summarise, the benefits of the approach are:
(a) A complete oversight over the whole water resource in a defined area, with the possibility of planning water allocations between different uses.(b) Clear information on externalities, in the sense of off-site effects, so that cause and effect, costs and benefits of possible interventions can be compared.
(c) The definition of relatively uniform land units, so that agricultural, or other land use advice can be precisely targeted, and applied in other watersheds.
(d) Present land allocations and uses can be tested to determine if they are technically and economically acceptable and sustainable.
(e) A rational framework for research planning is provided, that can follow effects back to their spatial origins.
(f) An ability to see the socioeconomic conditions, the infrastructural situation and the flows of inputs and outputs across the whole watershed.
(g) A better structure for dealing with biodiversity and other environmental questions.
A watershed programme needs an international level approach. The complexity of a fully integrated study, the variety of disciplines involved, and the difficult technical issues that will almost inevitably be raised by the many interdisciplinary contact points, will demand staff of high calibre and with good backup. It will be essential to have good contacts and collaboration with universities and other centres of excellence from the more-developed world. The development of the necessary models for expressing the interactions and functions of a watershed do likewise. The organisation of all the adaptive and applied work in such a programme is itself a considerable challenge and an important research task. There may also be scope for manipulative experiments in these ecosystem-scale projects (Rasmussen et al., 1993).
A full watershed experiment will be a considerable investment, and which will need to be continued for a number of years to give full value - for example, it would be optimistic to believe that sustainability can be proven in a period of five years. The work done in Kenya over the past 50 years shows what can be done, but it also emphasises the importance of consistent long-term support for the work.
An experiment of this type would require a powerful consortium of Centres, NARSs and other national or regional organisations, and centres of excellence in more-developed countries. Consequently only a small number of them can be set up, and they would have the status of CGIAR flagship sites. Their size, high cost and visibility demand that site-selection, planning and organisation should be fully thought through and meticulous, and that the assignment of responsibilities should be absolutely clear. Similar comments apply, of course, to all SWNM work in the Ecoregional and Systemwide programmes.
Within a watershed there may be several land uses or agricultural systems. It may not be necessary to do research on all of them, but there will be a need to have basic expertise about each one to allow it to be included in the overall approach. There will be a need for hydrologists, agronomists, soil scientists, economists, social scientists, livestock scientists and foresters. This disciplinary complexity does, of course, carry dangers of fragmentation, because researchers will tend to see their problems through disciplinary lenses, and may drift away from the main objectives. There has to be frequent and repeated contact between the various types of scientist, and the responsibility for overall management has to be clearly identified and enforced. The managing of a full watershed experiment is not to be undertaken lightly, but the added value from successful single-site collaboration could be great.
Equally, close linkage with the farmer is essential in soil/water natural resources work. For example, the detailed treatment of the land may be of great importance for water relations, and the exact cropping regime is vital for soil protection, in addition to the many socioeconomic issues that arise. It is possible that, over a large spatial extent, this participatory link to the working farmers, knowledgeable as they are about the management of their natural resources, may best be made through NGOs (Malena, 1995). The need for strong local involvement of scientists and farmers is essential to ensure appropriate solutions to real problems, and to ensure that all off-site problems are detected and quantified (Rasmussen and Meinzen-Dick, 1995).
Watersheds are of many sizes, and it is impractical to mount a concentrated research effort on more than a moderate spatial scale. The best approach is probably to select a very few typical sub-watersheds for intensive experimentation, and then ensure that essential data such as streamflows and pollutant concentrations are measured across the whole of the selected watershed. Socioeconomic data will almost certainly also have to be measured widely. The ultimate aim is to have sufficient data to model the watershed as a whole. A programme at this level is a sound base for extrapolation of results, and for assessing off-site costs of any proposed improvement. More complex programmes, such as those described above, could be approached as experience grew.
It may be asked if this is practicable in less-developed countries, and how it will benefit the farmer. It will not produce immediate benefit in the short term, but the approach is undoubtedly the way forward for improved land-use management. It is necessary to develop forward looking strategic research, from which the next generation of applied and adaptive research will draw its inspiration. The alternative is constant ad hoc experimentation addressing now this single issue, now that, but without coherence.
The logical user-level product from modelling is a Decision Support System, which could be considered as part of a watershed programme. CIAT (1994) considers that such systems will be useful in reconciling divergent interests amongst stakeholders, and is already working on one.
