Details of the proposed research program

96 J. Dumanski, C. Pieri Agriculture, Ecosystems and Environment 81 2000 93–102 develop, model, test and refine the LQIs Pieri et al., 1995; World Bank, 1997b. This normally will result in a small, select group of LQIs that relate closely to the normally limited AEZs in a given country. When the LQIs are developed and interpreted, they can either be reported by the AEZs, or the results can be aggregated to sub-national and national levels by weighting the LQIs according to their contribution towards some national objectives, such as calculating national wealth, rural poverty reduction, achieving food security, etc. 2.2. Objectives for the LQI research program Major gaps in knowledge on the impacts of human interventions in the landscape were identified in the various international workshops leading to develop- ment of the LQI research plan. These are: • links between land quality and poverty; • links between land quality and land management, i.e. management practices currently being used by farmers; • lack of geo-referenced land management informa- tion; • lack of time series data and system resilience infor- mation; • rural-urban land management and land use planning interactions; • impacts of macro-level policies on land decisions and land quality. While all the knowledge gaps are important, the first three are the highest priorities, and these are the basis for the research program. This results in development of the following research objective: “Identifying and testing cause-effect relationships among land quality, land management, and poverty is the primary research objective for the LQI program.”

3. Details of the proposed research program

A minimal number of recommended Core LQIs were identified using criteria and guidelines from ear- lier workshops Dumanski, 1994; OECD, SCOPE and other indicator initiatives, and from the regional LQI workshops held in 1995 and 1996 in Cali, Nairobi, Rome and Washington. Emphasis was on making better use of available data and information, and integrating these with new information management technologies such as remote sensing, geographic information systems GIS, relational data base man- agement systems, the internet, etc. The Core LQIs may be developed from direct mea- surement remote sensing, census, etc, or estimated using well tested, scientifically sound procedures. Emphasis should be on analyses to identify driving forces of land use change and the impacts of the change. Interpretation of the LQIs should not be done in isolation, but within the context of what is happen- ing with land managementland use in the countries in question. Recommended contextual information to assist in interpretation of the LQIs include social and institutional information tenure, gender, educational levels, etc, farm household economics , climate, and land use as available from the census. Characteristics of Core LQIs are: • measurable in space, i.e. over the landscape and in all countries; • reflect change over recognizable time periods 5–10 years; • are a function of independent variables; • are quantifiable and usually dimensionless ratio es- timates. 3.1. Core LQIs recommended for short term development This section provides short descriptions and the cri- teria for each of the core LQIs being recommended. The major research issues are identified, along with some information on data sources and procedures for assessment and interpretation. The first two LQIs are already developed, and full instructions are available on: www.ciesin.org 3.1.1. Nutrient balance and nutrient depletion This LQI describes nutrient stocks and flows as re- lated to different land management systems in specific AEZs and specific countries. It is based on several years of research and experience in East Africa and Europe Smaling et al., 1996. The major research issue is nutrient cycling: • nutrient mining, nutrient imbalance and decreas- ing yields caused by removal of harvested products grains and residue, erosion and leaching, primar- J. Dumanski, C. Pieri Agriculture, Ecosystems and Environment 81 2000 93–102 97 ily in developing countries with low rates of fertil- izer use; • nutrient loading and environmental degradation phosphorus, nitrates, etc caused by excess fertil- izer and manure application in developed countries, but increasingly also in developing countries with high fertilizer subsidies such as China. The research process is similar to environmental ac- counting. It involves establishment of nutrient balance sheets with losses and additions as estimated from nutrient removal through crop harvesting, erosion, etc., compared to nutrient additions due to fertilizers, organic inputs, recharging of the nutrient supply due to legume rotations, deep rooting systems, natural recharging due to atmospheric fixation, etc. Smaling et al. 1996 have demonstrated how this LQI can be developed for several hierarchical scales, including the farm level, region AEZ level, and national level. Actual measurements of nutrient cycling are rarely available, but Smaling et al. 1996 have demonstrated that reliable estimates of nutrient balance can be devel- oped using crop yield statistics from national census, nutrient additions from fertilizer trade statistics, and various biophysical models for denitrification, leach- ing, erosion, etc; 3.1.2. Yield gap Three closely related statistics are proposed for this issue: • yield trends direction, rate of change for the major food crops in specific AEZs; • production risk change in the expected annual vari- ability in yield for the major food crops in specific AEZs; • yield gap current yields compared to potential farm level, economically feasible yields for the major food crops in specific AEZs. These are useful indicators because they are so easily understood, easily converted into economic terms, and they are useful for monitoring both project and program performance. However, they have value as LQIs only if changes in yield are clearly related to land management in specific AEZs and for spe- cific management systems. There are many possible cause-effect relationships with yield, e.g. marketing, climate change, subsidies, land management, etc., and the indicator can be misleading without compre- hensive analyses of the driving forces causing yield change. Knowledge of farming systems, marketing, the policy environment and other contextual informa- tion, as well as cause-effect relationships of current land management on yield trends and variability are necessary. The key research issues are: • to what extent are changes in land quality resulting in corresponding changes in crop yield and produc- tion risk; • how can reliable estimates of yield gaps be devel- oped for developing countries, and what are the management options to improve the yield gap; • are there practical biological and economic thresh- olds yield and variability to ensure sustainable production systems. Yield gap analyses relates to residual yield opportu- nities that could be realized with improved cultivars, fertilization, and land management. Alternatively, it reflects the extent to which the biological production systems are currently being pushed, realizing that if pushed beyond a biological threshold, the systems will likely fail. Knowledge of this threshold is strategically important for policy, program and management plan- ning in high production environments, since the mar- gin for error decreases as the biological production threshold is approached. Concurrently, the impact of error at these production levels is likely to be signifi- cant. Yield data are available from the FAOSTAT data base, as well as other sources, such as special surveys, safety net and farm subsidy programs, international marketing programs, etc. Alternately, crop growth models can be used to simulate long-term average yields and annual yield variability to a lesser extent, providing that the models have been thoroughly tested and found to be reliable. Also, remote sensing data can be used to estimate level and variability of crop yields, or to provide agro-meteorological inputs for crop growth models. A major concern is that, national yield statistics are normally a ratio of national production divided by area harvested. Such estimates can be highly biased, since experience has shown that expansion of cultivation on marginal lands, for example, can depress ‘national average’ yields at the same time that yields in more favorable areas are increasing due to increased fertil- izer use, improved soil conservation, adoption of im- proved cultivars, etc. Ideally, yield data for the major 98 J. Dumanski, C. Pieri Agriculture, Ecosystems and Environment 81 2000 93–102 food crops should be geographically referenced dis- aggregated on an AEZ basis, and then analyzed in re- lation to the agricultural production systems and land management practices being used by farmers. 3.1.3. Agricultural land use intensity and land use diversity These two LQIs are intended to assess the general impacts of current land management at the farm level. Assessing land use intensity and land use diversity provides information on trends towards or away from sustainable land management. Land use intensity is intended to estimate the im- pacts of agricultural intensification and increased frequency of cultivation on land quality. Normally, intensification involves increased cropping, shift to- wards more value-added production, and increased amounts and frequency of inputs. Such changes, without concurrent adjustments in land management practices, often result in nutrient mining, soil ero- sion and other forms of land degradation. However, agricultural intensification, if properly implemented, can also result in improved land quality, as shown in Kenya, Brazil, and other areas El-Swaify, 1999. The difference is the specific management practices adopted by farmers during the transition towards intensification often based on expected increased profits from the more intensive production systems. Land use diversity agro-diversity is the degree of diversification of production systems over the land- scape, including livestock and agro-forestry systems; it is the antithesis of mono-cropping. Agro-diversity is often practiced by farmers as part of their risk man- agement strategy, i.e. to guard against crop failure and financial collapse, but it is also a useful indicator of flexibility and resilience in regional farming sys- tems, and their capacity to absorb shocks or respond to opportunities, e.g. Kenya English et al., 1994. Agro-diversity is assessed on the number, kind and complexity of components in the farming systems, but it can also be approximated from national statistics by the number and kinds of crops per growing season, the extent and frequency of rotations, presence or absence of livestock in the production systems, etc. The key research issues are: • is current land management contributing to in- creased land degradation or improving land quality; • are current agricultural management practices contributing to improved global environmental management. A major concern with these LQIs is that data on current land management practices are generally not available except from special surveys, and various surrogates will have to be developed. Some of these are already available in the literature, such as land use intensity based on crops per growing season, extent and frequency of rotations; cultivation intensity, etc.; ratio of cultivated land to cultivable land; area and fre- quency of soil enhancing to soil depleting crops; ra- tio of mono-cropping to mixed cropping, etc. These are based on data available from the census and from agricultural and household surveys. Time series infor- mation on crop rotations, etc. can be generated from remote sensing Huffman et al., 2000. 3.1.4. Land cover This LQI is intended to estimate the extent, dura- tion, and time of vegetative cover on the land surface during major periods of erosive events. It is also in- tended to measure land cover change over time, and the correlation with land use pressures. This LQI will be a surrogate for estimating ma- jor land processes such as erosion, desertification, deforestation, etc. In combination with land use in- tensity and agro-diversity, it contributes to increased understanding on the issue of agriculture and envi- ronmental sustainability. The land cover LQI extent, duration, time series, including critical erosive peri- ods will require application of remote sensing data, supplemented by ground truthing. The key research issues are: • is current ground cover adequate to protect against land degradation during critical erosion periods; • how is the kind, extent and duration of land cover changing over time; • what pressures are causing change in land cover. Archived remote sensing coverages are already available in many cases to initiate testing and develop- ment. Also, some of the required indices, such as the normalized difference vegetation index NDVI and revised NDVI, are already available Ashbindu Singh, UNEP, personal communication. Indices for specific areas will have to be mapped, and then compared overlain with mapped estimates of erosion risk, as obtained from biophysical models such as EPIC and others. Such analyses, calculated over to time with J. Dumanski, C. Pieri Agriculture, Ecosystems and Environment 81 2000 93–102 99 historic data or simulated using best estimates of land use and weather data, will provide information on the degree and probability of erosion risk. Land cover kind, extent, and duration can be in- terpreted as a surrogate for land degradation. Such indicators are important for planning and evaluating national and regional programs in soil conservation and agricultural production, but increasingly they are required for estimating the potential impacts of im- proved land management on issues of global signif- icance, such as carbon sequestration, desertification, and biodiversity. Increasing the biomass and extent of land cover often increases soil organic matter and soil water storage, and this enhances yield potential and decreases production risk. At the same time, however, this process sequesters increasing amounts of atmo- spheric carbon in the soil, thereby providing global environmental benefits by mitigating climate change, controlling desertification, and often increasing the quality of wildlife habitat. 3.2. Core LQIs recommended for longer term development sub-national AEZ program These are LQIs, where the theoretical basis for de- velopment is still being developed, e.g. soil quality, or where reliable data to report on the LQI are generally not available, e.g. land degradation. These LQIs are designated for the second phase of the LQI research plan, they require a longer term commitment, and they will be more expensive to develop. The recommended LQIs are: 3.2.1. Soil quality The scientific basis for this LQI is that soil quality reflects and is measured by the conditions which make the soil a living body. There are currently several na- tional and international working groups trying to de- velop a relevant and effective indicator of soil quality, but successes are mixed. The evidence indicates that traditional soil chemical, physical and taxonomic data by themselves are not useful to describe soil qual- ity, but a more definitive indicator may be developed from integrating these data with soil biology and by monitoring the soil food chain. Soil life functions also require favorable chemical and physical conditions, but the latter are not suitable indicators on their own. In the meantime, a useful surrogate may be developed from data on soil organic matter, particularly, the so-called dynamic or microbial carbon pool bacterial and fungal carbon. Recent research indicates that soil quality and soil agro-diversity are closely linked and could be evaluated by characterizing the resilience of the main ‘functional groups’ in the soil, namely macrofauna earthworms, termites, nematodes, mi- crosymbionts mychorrhiza, N-fixers, and microbial biomass fungi, protists, bacteria Swift, 1997. The key research issues are: • do current land management practices maintain, en- hance or reduce the capacity of soil organic matter to perform the above functions; • do current land management practices maintain the biological life and biodiversity of the soil, and thus enhance the environmental resilience of the soil for maintenance of global life support functions. Soil organic matter SOM is the best surrogate for soil health, since SOM reflects the residue of plant and soil organisms that have lived and died in the soil. The basic functions of SOM are development and maintenance of soil structure, nutrient and or- ganic carbon storage, and maintenance of biological activity. Thus, SOM is the primary source for and a temporary sink for plant nutrients and soil organic carbon in agroecosystems, maintains soil tilth, aids in the infiltration of air and water, promotes water storage, reduces erosion, and controls the efficacy and fate of applied pesticides. Specific indicators will have to be developed to characterize the sensitivity of organic matter for these functions. This LQI would need to be linked closely with land use intensity, agro-diversity, and land cover because of the implications and possible impacts for improved land management. 3.2.2. Land degradation Land degradation erosion, salinization, com- paction, organic matter loss, etc has been studied for many years, and there is a strong theoretical base for several State LQIs to describe the processes. The major deficiency, however, is the serious lack of sys- tematic and reliable data for reporting on the extent and impacts of these indicators, although there are innumerable small research studies in many parts of the world all of which provide some information on the current status of land degradation for small areas. This information is being assembled through 100 J. Dumanski, C. Pieri Agriculture, Ecosystems and Environment 81 2000 93–102 programs such as WOCAT and ISRIC, but progress is slow. The key issues are: • do current land management practices contribute to loss of agricultural production potential; • do current land management practices contribute to off-site environmental damage and reduced envi- ronmental resilience, as land becomes degraded or is allowed to erode due to mismanagement. Systematic monitoring and application of these indicators has not been instituted in any country ex- cept for the USA. Because land degradation is site specific, monitoring systems have to be highly com- prehensive and cover all land areas, and these rapidly become exceedingly expensive. Alternate means of assessing land degradation are necessary. A consid- erably less expensive approach is approximated in Canada by building on the agricultural census with a few strategic questions on current land management practices Dumanski et al., 1994. Improved infor- mation on land management at the farm level thus become available with each normal census reporting period. In addition, some promising new technologies for estimating the extent, duration and impacts of land degradation, involving the use of biophysical models such as EPIC, some remote sensing procedures, or a combination of these with ground surveys, are becom- ing available Ingram and Gregory, 1996. Research is needed, however, to develop cost effective procedures for application at local project and national scales, and aggregation of results for policy and program application. 3.2.3. Agro-biodiversity Agro-biodiversity is a new concept, and much work is still required to develop this to the point where practical indicators could be developed for monitor- ing and evaluation. By far the majority of the earth’s land surface has already been modified by human interventions, so an understanding of land use sys- tems, the dynamics of land use change, the driving forces behind land use change, and the resiliency of land use systems are fundamental to better managing biodiversity issues in agriculture. The primary research issues are: • how to better integrate biodiversity in the process of intensifying agriculture Srivastava et al., 1996, and manage the gene pools used in crop and animal production; • how to manage and protect the diversity and healthy living conditions for soil microorganisms, realis- ing their roles in biologic nitrogen fixation, etc., but also as sources for antibiotics, etc.; • how to manage the coexistence of natural species, particularly wildlife and water fowl, in agricultural areas. It is often argued that intensive agriculture is the principle cause of habitat destruction and biodiversity loss, but the alternative, agricultural extensification, exacerbates habitat decline and species loss Smith, 1996. This is because extensification normally en- tails expansion of cultivated areas on marginal lands, which usually are the prime areas of natural habitat. At the same time, agricultural systems, particularly those based on traditional agriculture or systems in transition that retain some traditional plant varieties and cultiva- tion practices, are an important conservatory of genetic materials important to agriculture. Agricultural sys- tems, such as mixed systems with integrated crop and livestock production and those involving agro-forestry, utilize and manage a wide genetic pool for production. However, the adoption of high-input, ‘modern’ farm- ing, with reliance on monoculture, mechanization, and pesticides and other agri-chemicals, normally dimin- ishes biodiversity and destroys many beneficial insects and soil microorganisms. Agricultural intensification, however, should not be considered simply as the adoption of modern farming systems. Some new technologies such as integrated pest management, conservation tillage, and agro- forestry actually expand the genetic resources used in agriculture. The proper role of agro-biodiversity is to achieve the balance required to ensure the health and sustain- ability of agroecosystems, and protect the flexibility and resilience of farming systems againt outside stresses and periodic shocks such as pests, diseases and droughts Smith, 1996. Maintaining a wide genetic base in agriculture is essential to ensure the resilience of farming systems, because this provides greater options for alternative production systems, and it is the basis for improved risk management and rapid response management. Ensuring the resiliency of these systems is an important condition to sustain- ability. J. Dumanski, C. Pieri Agriculture, Ecosystems and Environment 81 2000 93–102 101 Table 1 Proposed indicators for agro-biodiversity Indicators Surrogates Natural habitat loss Encroachment by agricultural production systems Habitat fragmentation Encroachment of agriculture in an uncoordinated manner Species loss even when natural habitat is intact Pollution; sedimentation; excessive harvesting, etc Decline of biodiversity of crop species on farms Adoption of monocropping; reduction of mixed cropping Decline of biodiversity within species Adoption of modern varieties; expansion of intellectual property rights Although, data with which to assess agro-biodiversity are universally scarce, various surrogates for this can be developed using information on land use dynamics. Smith 1996 proposes some preliminary indicators and surrogates Table 1. Some indicators are specific to managing natural habitats, while others will be similar to those identified under land use intensity, land use diversity, and land cover see previous section. 3.3. Core LQIs recommended for development by liaison with other authoritative groups This activity is mostly a matter of liaison with international and national agencies already working on indicators, and selecting those indicators which best meet the requirements of the LQI program. An associated major activity will be to identify reliable sources of data, and making these available through the internet. The LQIs included in this category are: 3.3.1. Water quality Primary interest is on water supply and quality related to land use change, e.g. estuarine, in-land waters, irrigation water quality, etc.; link with water quality indicators already being developed under the OECD indicators program, and others. 3.3.2. Forest land quality Primary interest is on land quality related to enviro- nmental management and forest production, specifically land quality changes related in forest management; link with the international forestry indi- cators process. 3.3.3. Rangeland quality Primary interest is on rangeland quality related to change in land use intensity and diversity, and the role of livestock in mixed production systems; link with range monitoring programs being developed under the convention to combat desertification, and various re- mote sensing programs. 3.3.4. Land contaminationpollution Primary interest is on type, extent and impacts of land contamination by human interventions; inter- ested in heavy metals, organic contaminants, radioac- tive contamination, etc.; link with agencies collating available information.

4. Conclusions and recommendations