Despite the fact that dryland savanna ecosystems provide a host of essential ecosystem goods and services to both humans and livestock, they are often confronted with dangerously vacillating levels of locally available natural resources to support rural livelihood strategies in the face of increasing anthropological influences and global climate change impacts. This points to the vital roles which the socio-cultural and bio-physical environment sub-systems play to ensure the stability of the complex socio-ecological system (SES). Several attempts have been made in the past to focus more on scientifically-based means of investigation than including contributions of the local resource users to better understand and harness SES. Notwithstanding, local ecological knowledge (LEK), which is an effective, investigative tool for understanding interactions between the ecological and social sub-systems of complex SES, has recently received increasing attention in studies on the effects of climate and land use changes on the availability and utilization of natural resources in communal rangelands. Surprisingly, little is still known when it comes to LEK of forage resources, particularly in the West African Sudanian Savannas.

The overarching goal of this study was to investigate local agro-pastoralists’ knowledge on forage resources used by cattle, goats and sheep and how they adapt their rangeland management strategies to vegetation dynamics. I hypothesized that LEK can potentially provide insight into reasons how and why forage resources are overexploited, and into management strategies to conserve or restore them. The study encapsulates three major empirical components: (i) LEK distributional patterns in forage resources utilization (Chapter 4), (ii) local valuation criteria for forage resources (Chapter 5), and (iii) local perceptions on forage species diversity, abundance trends, habitats distribution and ecological drivers to forage species changing trends over the past few years via the ‘lenses’ of local agro-pastoralists (Chapter 6). Using a stratified random sampling approach, I sampled sixteen villages across three dominant socio-linguistic groups and a steep climatic aridity gradient in both Ghana (seven villages) and Burkina Faso (nine villages) to address the aforementioned empirical components of the study. Although individual ethnobotanical interviews were chosen over focused or group discussions to extract the bulk of independent primary ethnobotanical data from local agro-pastoralists,

I also complemented the data collection process with personal observations and ethnobotanical walks for purposes of triangulation.

For the distributional patterns of LEK on forage resources utilization among local agro-pastoralists, this thesis examines socio-cultural and environmental variables which specifically affect various components of LEK dynamics (be it LEK on herbaceous, woody and crop-related forage plants), reflecting their capacity to recollect and list vernacular names of forage species. This section also looks at the local climatic variability implications for LEK accumulation on forage plants (Chapter 4). Furthermore, this study addresses the aspect of local valuation criteria for forage resources by agro-pastoralists. Here, I elicited LEK on forage resources by asking them to cite and rank specific forage species mostly considered to be palatable for various domestic livestock and at different seasons. I also asked agro-pastoralists’ to provide underlying reasons for their rankings to gather more information on explicit valuation criteria for available forage resources. This anthropological dataset was matched with ecological information obtained from rangeland vegetation sampling using 20m x 50m per plot for the woody vegetation at different topographic positions (Chapter 5). Regarding local perceptions on forage species diversity, abundance trends, habitat distributions and ecological drivers, local agro-pastoralists were interviewed to specifically answer questions relating to above-stated ecological variables (Chapter 6).

To disentangle the effects of socio-cultural and environmental variables on LEK accumulation and local explicit valuation criteria, I employed various statistical approaches such as exploratory data analyses with IBM SPSS v. 22 as well as generalized linear mixed-effects models (GLMM) with rigorous model selection procedures using the Akaike Information Criterion (AIC) with R software (Chapters 4 & 5). I also used non-metric multi-dimensional scaling (nMDS) to visualize similarities or dissimilarities of LEK distributional patterns as well as a two-way non-parametric permutational multivariate analysis of variance (PERMANOVA) using PC-ORD v.5 for purposes of triangulation (Chapter 4). Also, I used ANTHROPAC 1.0 software to calculate cognitive salience index (CSI) of both anthropological and ecological datasets. In Chapter 6, the forage species diversity metrics were estimated and other ecological variables done

using descriptive statistics, bivariate correlation analysis and also performed CSI calculations.

The results of this study reveal that those who resided in villages with moist environmental conditions seem to generally exhibit superior LEK on forage resources in terms of citation of many forage species than those who situate in dry rural communities. This evidence also is true for the ability of local agro-pastoralists to provide various underlying reasons for ranking of cited forage species. It was also evident that local agro-pastoralists possessed extensive knowledge and understanding of the habitat distribution, abundance trends and effects of predominant ecological drivers in the study region. The findings of this study, therefore, contribute to the on-going scholarly debates on how LEK-oriented research is crucially important, and the need to incorporate it into the scientific approach to enhance the functional understanding of ethno-ecologically useful natural resources for sustainable development and livelihood improvement.