Toward Improved Indicators to Measure Desertification and Monitor the Implementation of the Desertification Convention

Hartmut Krugmann

Introduction

This paper briefly examines earlier work on desertification assessment and indicators. It reviews the framework of the Desertification Convention and what it says about the use of indicators in desertification monitoring and reporting, with a view to characterizing the kinds of indicators needed to monitor its implementation. A number of general characteristics of possible desertification indicators are listed and an example of the classification of potential local indicators are provided from research conducted by the author in Rombo Location in Kajiado District, Kenya.

Indicators and Plan of Action to Combat Desertification

The severity and widespread occurrence of the problem of desertification was formally recognized for the first time in 1977 at the United Nations Conference on Desertification (UNCOD). This meeting conceived a Plan of Action to Combat Desertification (PACD) which was pursued between 1979 and 1991. The importance of assessing and monitoring states and processes of desertification at local or higher levels was recognized under the PACD and pursued in the 1980s by the Food and Agriculture Organization (FAO) and the United Nations Environment Programme (UNEP), who developed a provisional methodology for assessment and mapping of desertification (PMAMD). The methodology has been tested and adapted through two pilot projects (Kharin 1990).

One included a field test of the methodology in two districts of Western Mali which defined four types of land degradation processes: vegetation cover, wind erosion, water erosion, and soil compaction. For each, FAO/UNEP criteria were adapted to fit the local conditions. Methods for desertification mapping based on remote sensing were then applied to generate larger-scale maps on the present state of desertification, desertification rates, and desertification risks.

The other project was carried out by the Government of Kenya (GOK) Department of Resource Surveys and Remote Sensing (DRSRS), and UNEP. It focused on two field areas: the Lake Baringo and Marsabit Districts in Kenya (Ottichilo et al. 1990), and concluded that most of the indicators proposed by the FAO/UNEP Provisional Methodology could only be used at the local, or pilot level, because the costs of using the indicators and methods for assessment and mapping of desertification at the regional and national levels would be prohibitive, and the process time-consuming. The study had used a combination of remote sensing techniques and field surveys to collect data on selected desertification indicators. Detailed data were evaluated for use at the local level. Selected data elements and other data were used in a geographic information system (GIS) to develop generalized models for application at regional and national levels, including separate models for water erosion, wind erosion, range carrying capacity, vegetation degradation, and human population.

Recommendations of the Baringo-Marsabit study included: a) the use of remote sensing techniques as cost-effective, rapid, and also conducive to periodic data acquisition, for national-level assessment; and b) the inclusion of socioeconomic data in any assessment of desertification. Table 1 reproduces the list of physical, biological and socioeconomic desertification factors selected from the sets of indicators proposed by the FAO/UNEP Provisional Methodology. Table 2 presents the list of composite indicators which the authors constructed from the selected desertification factors, and also indicates the level of application of the indicators (local or national).

One immediate conclusion can be drawn from the FAO/UNEP projects: all of the chosen indicators are scientific - generated externally, for local or national use. Locally generated grassroots indicators are not considered at all; nor is the possibility of involving local people in the collection of scientific or grassroots data at the local level to reduce costs and time demands.

The Desertification Convention and Indicators

The 1994 Convention on Desertification is to be implemented through various national and subregional action programs. The Convention text outlines the content of these action programs and describes the institutional mechanisms that are to be used to develop and implement the action programs (see articles 10-15 of main Convention text as well as articles 8-13 of the Regional Implementation Annex for Africa). The Convention's approach to combating desertification (and mitigating the effects of drought) emphasizes the importance of local action by local people and communities. Clearly, the war against desertification is won or lost at the local level. At the same time, the Convention recognizes explicitly that local action requires an enabling environment (facilitating legislation, long-term policies and action) at higher (national and international) levels (Krugmann 1994).

The Convention also recognizes that effective local action will not be possible unless local rural people and communities:

· have greater control and responsibility over their local resources;

· are able to command a greater range and level of resources (to be more "resourceful"); and

· are able to participate in and influence higher-level decision-making processes by which they are affected.

Throughout the Convention text, explicit statements about the importance of poverty reduction and the need for local participation in all desertification control activities are found. The essential elements of the enabling environment are also obtained in the Convention.

At the national and subnational (provincial, district, location, sublocation, etc.) levels, this includes:

1. More democratic (participatory) and decentralized political and administrative structures - Their purpose should be to devolve authority over natural resource management to government levels that are as close as possible to those (local) people who directly depend on the natural resources for their livelihood.

2. Appropriate land and resource tenure and ownership systems and policies - These should reflect the existing sociocultural diversity across local settings, allow statutory laws to build on (rather than undermine) existing local customary rules, and provide security in resource tenure for local land users, who can then reinvest profits locally and, in particular, make long-term investments in land improvement.

3. Appropriate economic policies, structural adjustment approaches, marketing structures, and trade patterns and mechanisms - These should improve terms of trade for local livestock keepers and farmers vis-à-vis larger-scale markets, and allow local livestock keepers and farmers to retain a greater proportion of the marketing margin for local (re)investment, including land resource conservation measures.

4. Encouragement of local capacity for self-help and, beyond that, provision of technical and financial support to build the capacity of representative and participatory local community institutions.

5. Adaptation of education systems and curricula to give greater weight to traditional local knowledge and discovery of ways to combine it with modern scientific knowledge in natural resource use and management.

At the international level, the enabling environment includes aspects such as:

1. Enabling world trade patterns - To allow access to Northern markets, facilitate diversification of the domestic economic base, and complement development assistance objectives (rather than undermine them).

2. Gradual reduction in foreign debt obligations, linked, if possible, to progress in creating national enabling environments for more sustainable local resource use and livelihood.

3. Appropriate global agreements (like the Desertification Convention itself) to harmonize international environmental, economic and other policies and to coordinate national actions that have global impacts.

Having summarized the framework for action under the Convention, let us now turn to the role of indicators in measuring and monitoring desertification processes as well as the implementation of action programs.

The Convention document lists indicators related to the collection, analysis and exchange of relevant data and information for the systematic observation of land degradation in affected areas (Article 16 (c)). More specifically it mentions the development of "integrated sets of physical, biological, social and economic indicators." In the Regional Implementation Annex for Africa there is also a call for the establishment of "pertinent, quantifiable and readily verifiable indicators to ensure the assessment and evaluation of national action programs, which encompass actions in the short, medium and long terms, and of the implementation of such programs" (Article 9 (d)).

Article 26 of the Desertification Convention, on the communication of information, requires affected Parties to the Convention to report periodically on measures and strategies used to implement the Convention, and specifically on national action programs and their implementation. It is left to the Conference of the Parties to determine the format of such periodic reports. Yet it is fair to say that desertification indicators will likely be central to whatever reporting format is agreed upon. The time up to the ratification of the Convention (perhaps in 1997) and until the first meeting of the Conference of the Parties should therefore be used to examine, test, and begin to experiment with different indicators in order to come to a better understanding of what sets of indicators would be most appropriate.

The Convention text clearly establishes the need to look for multidisciplinary and integrated sets of indicators. The importance of different time horizons for indicators (and their sensitivity to change over time) is acknowledged implicitly: actions are to be taken in different time frames and this will require the identification of different indicators through which the effects of these actions can be monitored.

Beyond that, additional qualities of indicators can be inferred from the framework for action underlying the Convention, as outlined above. Perhaps most importantly, close attention to local level environmental change and action in combating desertification, the importance of tapping relevant local knowledge, and the need to rely on both local and scientific knowledge, and bring the two together - aspects that are all spelled out in the Convention - call for the local application of indicators that are generated either locally (i.e., grassroots indicators) or externally (likely scientific indicators). (The distinction between "local" and "external" is based on Gupta's Generation of Knowledge matrix: Appendix, Table A2.) Some combination of the two, or what I refer to below as "hybrid indicators," may also be possible. Nevertheless, it is grassroots indicators based on local environmental change that have usually been neglected or entirely ignored in developing indicator sets and designing monitoring and reporting systems.

Grassroots Indicators and "Hybrid" Indicators

In this section, I characterize attributes and qualities of hybrid indicator systems by using examples from the Rombo Location, Kajiado District, which is located in the southeast corner of Kenya's Maasailand. I will also touch upon the issue of how such systems might be made operational when there are tight constraints in terms of practicality, cost, and rapid implementation.

General Characteristics of Desertification Indicators

Several characteristics of relevant desertification indicators can be delineated. First, they tend to occur in a hierarchy or cascading series of levels including micro-, meso-, and macroindicators. These levels reflect perspectives, experiences, processes, and actions at different levels. For example, government officials at different levels (national, provincial, district, location, etc.) may be making decisions that affect local desertification processes or the implementation of local projects intended to combat desertification. Or, price changes of livestock products or agricultural produce in national markets may influence farm gate prices and hence levels of cash or family labor surplus, some of which could be invested into land conservation measures.

Indicators are also dynamic, signaling and reflecting change in variables over a certain period of time. For example, changing tree or grass cover or changing tree or grass species composition in a given area over a period of a decade may reflect or signal processes of resource degradation due to competing land uses such as transhumant pastoralism and rain-fed or irrigated agriculture. Or, rising numbers of landless people could be an indicator of land privatization in drylands some years back. An interesting example of an indicator whose dynamic quality leads to conclusions that are quite different from those derived from a more static picture is the landlord-to-tenant ratio in irrigated agriculture identified in the Rombo area. This ratio is presently about 20:80 (irrigated), which signals environmental sustainability problems, since short-term tenants normally have no interest in land conservation investments yielding long-term returns. Nor do Maasai landlords encourage such investments on the part of tenants, as they do not want the tenants to become so permanently installed as to eventually claim rights to the land. It is interesting to note, however, that the landlord-to-tenant ratio has been rising over the past 10 years - at least for irrigated agriculture in the Rombo area - from 5:95 in 1984 to 20:80 in 1994, indicating rising interest among the Maasai to practise irrigated farming themselves. This trend appears to augur well for the sustainability of irrigated farming in the area. It also indicates great cultural change among the Maasai who, traditionally, have been nomadic pastoralists reluctant to get involved in agriculture.

It should also be recognized that indicators are specific to given ecological, cultural, social or economic contexts, or to gender or age class. For example, tree flowering just before the rains come: while the phenomenon may occur in many different locations, the particular trees in flower may be specific to a particular area. The average number of wives a man has is culturally specific. In Maasai society, this number seems to have gone down, hand in hand with complex socio-economic changes involving the penetration of the market economy, diversification of the livestock economy into agriculture and other investments, education, conversion to Christian beliefs, and many other factors. This may be an indicator on the degree to which traditional social organization based on principles of solidarity and sharing has been changing, with far-reaching consequences for the degree of sustainability of land use. Along similar lines, the decreasing average number of families living in residential units or manyattas indicates increasing social fragmentation.

Indicators can be both quantitative and qualitative. Quantitative indicators (such as market prices for, or yield of, agricultural produce) are easier to measure and aggregate. On the other hand, qualitative indicators (such as the new phenomenon of Maasai men going to the butcher to buy meat) are often better able to capture the complexity of changing situations.

Indicators may be considered "direct" or "indirect," although this can be a false dichotomy. There is a wide range in the degree to which variables of a system signal a process of land degradation, or indicate the effects of action taken to control desertification more or less directly. For example, the appearance of gullies is a direct indicator of soil erosion, and hence land degradation. The decreasing price of charcoal in the Rombo area is a somewhat more indirect reflection of increasing environmental degradation: increasing rates of land clearing for agriculture on the group ranch increase the wood supply for charcoal making, which in turn puts downward pressure on the charcoal price (on the informal market, as charcoal making is illegal). An even more indirect measure of land degradation (or its potential) may be the rising average landlord-to-tenant ratio in irrigated agriculture noted above.

The selection of indicators naturally will depend on the purpose for which they are to be used. In connection with desertification assessment and the monitoring of the implementation of the Desertification Convention, environmental indicators may be broadly categorized as either descriptive or performance-oriented (World Bank 1994). Descriptive indicators illustrate the status of the environment, or the process of environmental change over time. Most, if not all, of the examples of indicators mentioned so far in this paper fall into this category. In general, measuring and monitoring of desertification processes will involve the use of descriptive indicators. Performance indicators, on the other hand, are measured against some benchmark, physical threshold or normative policy goal that may be related to sustainability. Such indicators become important in the context of the implementation of national action plans (NAPs), and local programs and projects that may be undertaken to fight desertification. Clear definition of policy objectives for factors like resource use efficiency, tenure security, emission levels, or equality in resource use and access, will be necessary to construct performance indicators to measure progress towards the specified target.

To use the example of landlords and tenants in the Rombo area, suppose a project were formulated providing incentives for landlords to farm themselves, rather than leasing land out to tenants. A target of 100% landlord-farmers could be set as a goal, and the landlord-to-tenant ratio, which previously characterized the land-tenure dimension of the process of environmental degradation, could be turned into a performance indicator measuring the effects of the project.

Indicators as Parts of National and Local Reporting Systems

National reporting under the Desertification Convention will likely be a matter of considerable experimentation and debate on the kinds of indicator and monitoring systems best suited for the purpose. One could think of a hybrid indicators system consisting of a (minimum) set of descriptive indicators covering all the drylands in the country on the one hand, and of special sets of performance-oriented local indicators to monitor the impacts of particular antidesertification projects on the other.

The general (minimum) set of indicators would be monitored continuously and would reflect the cascading series of policy action levels that are contributing to the enabling environment, from the international and national level down to the local. At subnational levels, this set of indicators could cover the full extent of drylands in a country, depending on availability of resources. There would be opportunities for including grassroots indicators and involving local communities in monitoring at the local level.

Locally specified sets of indicators for particular local initiatives would be designed according to the nature and purpose of the initiatives. Monitoring of these special local sets would begin at the start of the projects and continue as long as necessary to observe their impact. Since projects would be designed and implemented in a participatory fashion, there would be ample opportunity for community participation in selecting and monitoring indicators, including grassroots indicators.

For obvious reasons, it is important to select indicators that are easy and inexpensive to measure or monitor. A viable system of national reporting to the Conference of the Parties of the Desertification Convention, once ratified, will require minimizing the degree of effort and expense involved in monitoring desertification processes or the effects of antidesertification programs. Available financial resources are always limited: the Desertification Convention is certainly no exception. It is also important to keep an appropriate balance between the costs and efforts of monitoring vis-à-vis the costs and efforts invested in the implementation itself. (The monitoring of elephant movements in East Africa is an example of a perceived imbalance between monitoring and implementation efforts. It has been charged that far greater resources were used to monitor elephants than to do something to improve the prospects of elephant conservation or sustainable utilization.)

Too much monitoring of action and its effects at the expense of action itself may be counterproductive. It is best to choose simple indicators that are easily measured or monitored, and can be used in different local situations with little effort of adaptation. Care should be taken, however, not to fall into the trap of using faulty or irrelevant indicators in the name of economizing or available resources. For example, the market price of a good or service (environmental or otherwise) may be easy to monitor, but market failure and other effects may obscure what exactly the price indicates or reflects. The price of charcoal in the Rombo area, for example, certainly does not internalize environmental costs. Quite the opposite: the faster land is cleared, the lower the price.

In this connection, grassroots indicators, or locally adapted scientific or "hybrid" indicators, present advantages. They can be monitored by local communities. Drawing on numerous local eyes and ears will reduce the cost of monitoring the implementation of the Desertification Convention to governments. Such decentralization of monitoring activity is also consistent with the principle of maximum local participation in all efforts of desertification control.

Classification of Land Degradation Indicators from Rombo Location

In this section, some of the desertification-related indicators which might be considered at the local level in a place like Rombo are discussed. The actual classification system is presented in Table 3. Most, or all, of the listed indicators emerged from in-depth discussions with a variety of local resource users - pastoralists, cultivators of rain-fed and irrigated land, landlords and tenants, traders of livestock or agricultural produce, small-scale businessmen, and combinations thereof. But before listing the indicators, a brief description of Rombo Location and a summary of the major factors facilitating desertification is in order.

Table 3. Classfication of indicators from Rombo location, Kajiado District, Kenya. General Specific Examples Ecological indicators Reduction in, or disappearance of, particular tree species Olmokatan tree - The bark is used as a catalyst for local brewing, as a traditional medicine for stomach problems (wormicide) "to clean the system", and as an appetizer.

Olorien tree - The trunk is used to make charcoal pieces to clean calabashes (rubbing of inside surface). Charcoal residues also serve to preserve milk.

Oiti tree - A hard wood to make sticks for walking and building purposes.

Orkonil tree - The roots are used by the Maasai in soup and tea making, as an appetizer, and a kidney cleaner.

Oliloriti tree - The bark is boiled to make a beverage (milk and sugar added), the hardwood is used for sticks, etc; and, it is also used as an herbal medicine against stomach problems.

Osokonoi tree - The bark is used as a medicine to cure stomach problems, also for chest pain and sore throat (tonsils). Reduction or disappearance of (nutritious) grass species Entimonyoa grass

Erikaru grass Flowering of tree a few days before the rains come Oiti tree

Olmokotan tree Tree leaves change colour just before the rains arrive Leaves of Olmomonyi tree turn dark Appearance spreading of gullies Formation of "wicked winds" Visible vertical vortices or small tornadoes through the sucking up of dust Climatic indicators Cloud formations indicate impending rains Elders look at stars to predict rains Reduction in average rainfall Variability in rainfall Variability in beginning or end of rainy season Prolonged absence of rainfall (drought) Land use indicators Elephants "harvest" maize No water for livestock downstream from irrigation schemes Economic indicators Price indicators: relative prices in marketing chain (e.g., farmgate vs. wholesale) Local prices of crops, livestock; terms of trade between the two or between rural and urban dwellers; prices of water, charcoal. Income indicators (difficult?) Wealth and asset indicators: e.g. cattle holdings per family Time or money spent by families on fetching firewood and water Indicators of economic diversification Number of people practicing agriculture or livestock management. Where are profits (re) invested? Government support and extension services Social indicators Number of families per homestead (residential unit) Inequality indicators (distribution of income and assets per family) Number of wives per family Form of bride wealth (cows, money) What proportion of children go to school? Proportion of children that looks after cows. Livestock transfers and associateships Who bears the risks - giver or receiver? Changing gender roles Institutional indicators Resource tenure indicators: landlord-to-tenant ratio Status of land subdivision

Access to water Existence of cooperatives or local interest groups Cultural indicators Proportion of Maasai who cultivate Maasai going to the butcher to buy meat Who selects wives? Changing proportion of ethnic groups in the area Political indicators Local power relationships Local effects of multiparty politics (e.g., land sub-division necessary for political reasons)

Rombo Location is situated in the southeastern corner of the Loitokitok Division of Kajiado District, in the immediate vicinity of the Tanzanian border and Mt. Kilimanjaro. The area is part of Kenyan Maasailand. My field work covered only two sublocations of Rombo Location, namely the Rombo and Njukini Sublocations. This area comprises the Rombo Group Ranch, held under group title, as well as farmed land (rain-fed or irrigated) held under individual private title. The group ranch and the individual ranches of plots were created in the late 1960s to early 1970s, when Maasailand was adjudicated. The Rombo Group Ranch has an area of 38 365 hectares of mostly semi-arid land and a population in the range of 20 000 to 25 000 (3398 registered members).

There are a number of trends and issues which contribute to land degradation in the area. First, there is a growing encapsulation of the extensive livestock economy that used to be the sole basis of livelihood for Maasai nomadic pastoralism. Over the past few decades, critical dry-season grazing areas - highlands and mountain slopes around Mt. Kilimanjaro, elevated higher-rainfall areas to the East, and a variety of water springs in the lower lands fed by Mt. Kilimanjaro - have been removed from the extensive livestock economy, as a result of land adjudication (privatization), rain-fed and irrigated (sedentary) cultivation on the individual private ranches and plots that were created, and the establishment of the Tsavo and Amboseli National Parks. Access to grazing resources and water sources, especially in the dry season, and lack of sufficient herd mobility due to spreading agriculture (even on the group ranch) and population growth, have become serious problems. The livestock economy seems to be fast losing its viability as the sole basis of livelihood for the Maasai. Increasing diversification into agriculture (primarily maize, beans and irrigated horticulture) and other investments can be observed among the Maasai.

One also finds significant resource conflicts between agriculture, livestock and wildlife that have to do with available land, grazing, and water resources in this region. These conflicts arise from the expansion of agriculture on the group ranch, livestock crossing into Tsavo National Park to graze during the dry season, and wildlife entering the group ranch and private farms to feed on livestock and "harvest," or destroy, agricultural fields and trees, primarily in the dry season. There are also significant conflicts among irrigators within and across irrigation canals.

A third cause of land degradation has been the widespread custom among Maasai land owners not to farm themselves but to lease out their agricultural land to tenants - either Wachagga from across the border or members of Kenyan ethnic groups, primarily Kikuyu and Kamba. As a consequence, there are many more tenants than landlords farming in the area, cultivating the land on the basis of short-term (one year) informal lease arrangements. Short-term perspectives and lack of land conservation are fueled by both the landlords, who often fear claims by (Kenyan) tenants to land they have farmed for some time, and by the tenants, who "mine" the land for lack of incentives to invest in longer-term land improvement. Also, the tenants' profits from agriculture are not reinvested locally, but tend to flow out of the area.

Rain-fed and irrigation cultivators are exploited by outside brokers and middlemen, and this has unsustainable implications for the natural resource base. The local road, transport and communication infrastructure is very bad and the prime reason for the ongoing exploitation. The proportion of the marketing margin that is appropriated locally is small (farmgate prices are much lower than the wholesale or retail prices in Mombasa or Nairobi). To date, the lack of local institutional cohesion (see below) has thwarted attempts to form local farming cooperatives and other pressure groups to break the stranglehold of outside brokers and exporters of horticultural produce.

The group ranch had decided to subdivide, giving each registered member (and some of his sons) a share. The experience in other parts of Kajiado Districts shows that subdivision leads to the selling of shares (and landlessness) among those poorer members of the group ranch who end up with a dry nonviable piece of land or default on bank loans for which the piece of land has been used as a collateral. Such occurrences partly account for the strong social stratification and institutional fragmentation in the area. This is also a result of the following factors: competing land uses; the existing cultural diversity and differences of an ethnic mix of indigenous Maasai on the one hand and nearly as many immigrants from other parts of Kenya on the other; the daily and longer-term influx of Tanzanian tenant farmers; and a penchant for politics among the local elite. This situation has made reaching a consensus and organizing community action to address major issues facing the local population difficult.

Finally, population growth may be diminishing in Kenya overall, but is still strong in this region. The Maasai have very large families: 4-5 wives and 20-30 children in total are not rare. The combination of all these factors and others which may appear as research continues, underlie land degradation and desertification processes in the location.

As Table 3 suggests, a classification of possible local indicators of land degradation, as derived from discussions with local people in the Rombo area, can be divided into categories including: ecological, climatic, land-use, economic, social, institutional, cultural, and political. In conclusion, let me stress that this table and its examples should be considered a "work in progress," as at the time of writing, this research project was still underway.

Conclusion

This paper has attempted to illuminate the characteristics of desertification indicators and the need for improved, and possibly combined, "hybrid" indicator systems through examples from research undertaken in Rombo Location, Kenya.

It is apparent that improved indicators (including grassroots indicators) cannot provide all the answers either. Just as the implementation of the national desertification action programs requires actions and mechanisms at higher levels (an enabling environment), so will it be necessary and useful to identify and construct indicators that reflect higher-level perspectives and experiences. The hierarchy of levels of policy and legislative action may, therefore, correspond to a cascading series of indicators. The real challenge will be to evolve hybrid indicator systems bringing together the different relevant perspectives and knowledge systems - local and (sub)national, "bottom up" and "top down," traditional and scientific - to name just some of the many possibly opposed, but potentially complementary, sets of indicators.