The overall purpose of the project is to promote sustainable livelihoods for the poor. There are many possible routes to this end, but the method proposed here derives from the fact that predominantly the poor of the world depend directly on natural resources, through cultivation, herding, collecting or hunting for their livelihoods. Therefore, for the livelihoods to be sustainable, the natural resources must be sustained. The overall project approach is to conduct a number of separate country studies focusing on selected sites representing a single ecosystem type or family. From these studies we can proceed to derive commonalities and differences on which to base conclusions and policy recommendations. "Ecosystem type" here refers to a set of ecosystems which share certain key defining criteria. This guidebook concentrates on the example of arid and semi-arid lands (ASALs) in Africa . Within the selected set of ecosystems or biome, and over millennia, people have traditionally evolved ways of life and stores of knowledge that enable them successfully and sustainably to provide for their livelihood needs. (We may note here that many traditional societies have located themselves so as to be able to exploit several neighboring ecosystems or ecotones.) More recently, but quite independently, global scientific knowledge and understanding of these same ecosystems has grown. Yet the two sets of knowledge, local and global, have been poorly integrated, and have not combined to inform a manageable set of policy alternatives.

We can gain much by studying adaptive strategies of people to a set of ecosystems, deducing comparative knowledge that will be of use at the local, national and international levels. Within a set of similar ecosystems, common problems and challenges can be identified to which the inhabitants or users have had to devise solutions. The art of selection of sites in which to study these strategies is important - we must ensure that the different sites have enough in common to be compared meaningfully with each other, and yet display enough differences of characteristics to ensure representation across a range, and for comparisons and differences to emerge. The definition of the ecosystem "envelope" of the project is therefore to hold constant a major variable. We do not yet know enough about adaptive strategies to enable us, for example, to make meaningful comparisons between strategies in tropical forests and those in arid lands.

ASALs provide an example of an ecosystem family. The term "arid lands" refers to areas prone to frequent and prolonged droughts and receiving up to 350 mm mean annual precipitation. "Semi-arid lands" are areas receiving from 350 mm to 700 mm precipitation, in which rain-fed agriculture combined with pastoralism is possible. The purpose of this distinction is not to create an agricultural-pastoral dichotomy, but to encourage, where possible, the selection within each country of two case studies representing points sufficiently distant on the agro-pastoralist continuum. In some areas, typified by Burkina Faso, the distinctions are clearly observable. For example, the Fulani specialized pastoralists occupy the drier northern zone, often herding cattle on behalf of Mossi agriculturists in the more humid zone.

ASALs constitute an important set of ecosystems globally, accounting for over a third of the land area and a seventh of the world's population (Ahmad and Kassas, 1987: 4). In Africa, perhaps a tenth of the continent's population, and a much higher proportion of those considered poor, derives all or most of their livelihoods directly from these lands. The fragility of ecosystems in these lands and their declining pastoral and agricultural productivity have been well documented for decades. Livelihoods in these areas are also highly vulnerable to shocks and stresses, including those arising from increases in human and animal populations, from increased intensity of use or changes in patterns of use, and from ecosystem change and declining natural productivity.

However, over time the peoples themselves who live in ASALs have evolved practices and ways of life which, in past times, enabled them to live in an environment characterized by unpredictability and variation.

Compounding this underlying threat to the balance between the natural resources and the livelihoods of the peoples dependent on them, is a reduction in access caused by changes in designated land-use through competing uses (e.g. Berhanu 1995, p.5). These changes include establishment of national parks - many of which are in ASALs - irrigation and resettlement schemes, commercial cattle and game ranches, private farms and refugee settlements. As if this were not enough, ASALs have also often been areas heavily affected by physical insecurity, which itself has undermined productivity.

In Africa, ASALs frequently have been regarded as peripheral or marginal in national politics, policies and political geography, and have often been characterized as "problem areas". Centrally-planned development, although not without its successes, has often failed. ASALs have been characterized by a lack of clear and consistent central policy, or even by "hostile" policies based on perceptions of "traditional" ASAL farming and pastoral systems as inefficient, unproductive, unsustainable, and destined inevitably to be undermined by the "modern" economy. The rate of project implementation is often extremely low. In many cases ASALs have become net importers of food and of an increasing proportion of the means of livelihoods, through the export of labor - usually very poorly paid because of poor educational attainments and low skills - and through the importation of food and relief efforts. Under these circumstances it is not unusual to find that development objectives, which sometimes in the past embraced ambitions such as to turn ASALs into major sources of national meat supply or even the bread-basket of the country, have often become reduced to the modest objective of ensuring merely that people can survive there. Sometimes they are characterized by an absence of policy, and an implicit policy assumption that over the long term they cannot form the basis of sustainable livelihoods (cf. Mutiso, 1995:37-40).

In the case of the IISD project, selection of different country sites illustrating typical points along the range of variation, resulted in a broad typology with South Africa at one end and Afars at the other. The South African example illustrated relatively dense, top-down planned settlement, a relatively high level of government provision of services (roads, irrigation, electricity), high dependency on migrant wage labor for household incomes (about 50 per cent); and low dependency on livestock. The Afars example showed an area with minimal government involvement (but planned irrigation that would alienate traditional grazing areas) low population density, and low integration with the wider economy. It was possible to discern the outlines of a hypothesis that would indicate possible paths of future historical progression from one end of the scale to the other.

This approach could be applied to other sets of ecosystems within the same problematic of adaptive strategies for sustainable livelihoods. For sustainable livelihoods based on different ecosystems, but challenged by similar threats, and people's organizing responses to combat development threats, see the Thailand tropical forest example described by Janet Durno (1995).

Whatever ecosystem family is selected, it will be useful as a preliminary step to compile a brief, non-technical and insightful summary review of the "state of the art" of knowledge on that ecosystem family, as found in contemporary literature. In this, it is most important to realize that human societies and economies are part of the ecosystem. Directly and indirectly, they not only depend on ecosystems, but through their activities and interactions they help to modify and change them.