The Sahel since the great famines of the 70’s and 80’s (Alpha Gado, 1993) has been the focus and the stake of intense debates in terms of research and development policies. As described by Warren (2005) and Herrmann & Hutchinson (2005), only a multi-dimensional approach will allow to go beyond the shortcuts and simplifications that have been overused to describe this rapidly evolving environment. Actually, the main characteristic of the Sahelian part of West Africa is the intricate complexity of the social, environmental and economic dimensions that differentially affect the rural populations of this region.Such an assertion may be applied to many rural societies in the world. However, in the Sahel, this complexity is enhanced by the extremely harsh environmental conditions on the one hand (Gentil et al., 1993; Milleville & Serpantié, 1994; Hiernaux, 2001; Abdoulaye, 2002; Walther, 2004; Yamba, 2004; Fauquet, 2005) and by the tremendous effects that social relationships have on the evolution of the systems of activities and the related farming systems of the Sahelian villages (Balandier, 1955; Luxereau & Roussel, 1997; Olivier de Sardan et al., 1995; Raynaut et al., 1997). Even though the agro-ecological and economic constraints have been widely characterized, the ways in which they affect the rural population in practice still have to be determined. These ways are by all means social because it is the social organizations that locally determine how the Sahelian village populations provide access to natural resources and other wealth-producing assets (Bonnal et al., 1997).Therefore, addressing the complexity of this intrication between social, agroecological and economic dimensions is a necessary step towards a better understanding of how economic and biophysical conditions interact with social structures and behaviors to drive changes. Evaluating and estimating the nature and the relative importance of these dimensions and the ways in which they condition the evolution of the population implies investigating the pattern on which they are intricate at different levels: individuals, family, village and beyond. But the problem with such an approach is that a real dialogue between the related disciplines has yet to be built. We fully assume the citation from Raynaut (2007): "if one wants to understand the transformation processes already taking place in the Sahelo-Sudanese societies and even in all the rural societies in transition, [one has to] work with a complex model that rejects linear or univocal interpretations and that permits to describe how natural constraints, market pressures, technical practices and social and symbolic relationships are combined into a global system of interactions”. A further step beyond such a conceptual model as suggested by Raynaut is quantitative modeling. Actually, a modeling approach appears unavoidable if one wants to explore the differential consequences of these intricate relationships on local dynamics. Building such a model is a challenging task that cannot be achieved at the level of a single thesis. We therefore propose to restrain the study by focusing on the most prominent levels and factors that may condition this complex system and its evolutions and to focus the research on the economic activities and the different factors that affect them, because they define the differential living conditions of the population. This approach is facilitated by recent modeling tools, e.g. Agent-based models, which are designed to allow a dialogue between different scientific approaches (Rouchier & Requier-Desjardins, 1998).For the Sahelian rural population, the village and its "terroir", as well as the family are still the most important levels for the management of economic activities. However, strong centrifuge forces affect these levels (Olivier de Sardan, 2003b; Yamba, 2004), implying that the individual is also a relevant scale for the study of these societies. Therefore, we develop an approach that combines the three scales, i.e. villages made of families and individuals who have differing access to economic assets. Such a model may encounter many inconsistencies especially between social and anthropological sources if its fundamental structure is only based on literature. A game-theory approach leads to omit the problem of the intrication of the concerned relationships and dimensions. Therefore, we stand on a field-based approach where field results support the local consistency of the model and guide the selection among the concepts, informations and data from the literature. Such an approach provides more confidence on the practical relevance of the results but also helps to favor the dialogue between the different concerned dimensions. To extend the geographical domain of application of the model, the research is carried out on three study sites that constitute contrasted situations of the Sahel. Logistic, temporal and institutional contingencies impose the selection of the three sites within the Sahelian region of the Republic of Niger.The overall aim of the thesis is to better understand the socioeconomic dynamics of the rural societies of the Nigerien Sahel by developing a modeling approach addressing the complexity of the relationships between the population and the factors that affect it, i.e. the economic and biophysical environment, the social structure and external interventions. Most specifically, the research focuses on the differentiation factors that discriminates the components of the population facing environmental and social constraints. The main contribution of this thesis is to consider that these factors occur at the family level, yet assuming that decisions are taken at the individual level. The key-steps are thereby to decode the components of this social space, to formalize them through a dynamic model that can explore the complexity of the Nigerien villages' situations and to provide insights to research and development issues.1. Consequently, the first specific objective is the building of this multidisciplinary methodology, with an ABM model as the main medium (Amblard & Phan, 2006). We then propose to test this methodology by applying it on several research issues that have strong implications for the development in this region, thereby fulfilling the mandate of the research and development program that hosted us.2. It is well known in the research community working on Sahelian Niger that the everyday access of the villagers to production assets in the Sahelian Niger is controlled by social factors (Raynaut, 1988; Luxereau & Roussel, 1997; Rain, 1999; Olivier de Sardan, 2003b; Yamba, 2004). But the relative influence of such elements still has to be investigated. Our second specific objective is therefore to simulate and estimate the impacts of several social differentiation factors on the dynamics of individuals, families and the population and on the distribution of production activities. The impact of the biophysical environment on this evolution is assessed by applying the model to three contrasted areas of Sahelian Niger.3. Our third specific objective is thereafter to use the model to analyze the villagers' responses to development interventions according to the specific agro-ecological and socio-economic characteristics of the study sites. These two development incentives are inorganic fertilizers availability and inventory credit. They have been selected for analysis because they are emblematic of development strategies in the country and are even considered as the most promising development techniques by many institutional development operators. Moreover, they are the most prominent proposals the institutions associated in the project try to promote in the different investigated sites.4. Finally, we propose to confront our hypotheses on a longer time perspective. Raynaut (2007) presents the changes that affect Sahelian societies as a consequence of modifications of farming systems. The origin of such modifications is seen as external, namely demographic growth and economic globalization. However, Sahelian societies quickly evolve and social evolutions have to be considered. We hypothesize that social changes may be the catalytic factor and the medium through which these farming system transitions operate. Our fourth and last specific objective is therefore to estimate the impacts of transitions in family structure on the evolutions of the populations, their systems of activities and the related farming systems.To reach the objectives, two connected challenges have to be dealt with:· A field challenge: The challenge is to decode the complexity of the relationships that condition the access to economic and production assets among villagers (Lavigne-Delville, 1999). Indeed, we assume as a working hypothesis that social relationships have a strong impact on individual, family and village evolutions. People are not only producers, but their production activity is part of a system conditioned by social relationships. This implies characterizing the heterogeneity and identifying the discrimination factors between villagers in their access to production means and the use of the related income.· A methodological challenge: it is about building a behavioral model for several contrasting villages of the Nigerien Sahel that is capable of combining information from different disciplines. Agent-based modeling is an efficient and flexible approach for integrating different types of information and data8 coming from various sources and for simulating the interaction between different aspects of a complex reality in a dynamic way.3 Scientific tools and disciplinesAs noticed in the preceding section, there is no a priori restriction on the disciplines called upon in this study as far as they concern the same object, the rural populations socially and spatially living on a regular basis in the terroir. 4 Outlines of the thesisThis thesis is structured in six sections.The first chapter groups all the elements that concern the studied field and the associated problematic. We first present a global description of the Sahelian Niger characteristics and evolutions. In particular, we present the main production and economic activities that are practiced in this part of the country. We reposition this description along its historical pathway, together with a brief historical reconstitution of the farming systems and the related social and family relationships. All these sections provide the elements needed to build the different concepts and hypotheses we used and developed in this thesis.The second chapter concerns the field materials and methods. The three sites we investigated are presented here. We then present the field investigation tools we used from which we derived the variables used for conceptualizing these rural societies and for building the rules that support their functioning. It is followed by the results of the perception-based zonation carried out in two of the three sites.Chapter 3 is dedicated to the modeling choices and concepts. The implementation of the SimSahel Agent-based model, which resulted from these choices, observations and hypotheses, is then presented. We first justify our choice of the Agent-based model as the most relevant model type for such societies and issues. We then present the way in which we use the ABM tool, by establishing the different assumptions, hypotheses and concepts that are necessary for such a model. The model implementation is thereafter presented. Finally, we present several results concerning a sensitivity test on several selected parameters and a confidence building exercise to test the model validity and robustness for further uses.The subsequent steps address the three earlier-mentioned research questions that are central to this thesis. These questions are elaborated according to investigation results and have been translated into scenarios implemented in the SimSahel model.Chapter 4 concerns the impact of family organization on the access to production assets and the sustainability of the rural population. Two scenarios are considered, one corresponding to the patriarchal family archetype, the other one being a model of a mononuclear family. This present-day version of the social dimension allows exploring a hypothesis regarding the historical dimension of social family evolutions.Chapter 5 introduces development interventions through two emblematic actions proposed to a present-day village. These two actions are on the one hand support to inorganic fertilizers, a tool that mainly seeks to maximize crop productivity, and on the other hand an inventory credit called "warrantage" as a potential food-security enhancement tool. The concept of reputation is introduced in the model and several scenarios are proposed and analyzed.In the sixth and last chapter, we consider a longer-term perspective by analyzing the effects of family transition processes on farming system evolutions. This is based on the hypothesis that family organization changes may be the transmission belt of these transitions of systems of activity and thereby of farming systems. It allows reinvestigating both social and geographical dialectics that are discussed in the previous chapters from a more historical perspective, as suggested in chapter 3., Populations, systèmes agraires et transitions sociales au Sahel Nigérien : Une approche par modélisation Multi-agent