The role of unplanned human settlements in land cover change: A case study of the Chyulu Hills squatters - Kenya

Investigations in this thesis advance our understanding on the role of unplanned human settlements in land cover change. Despite its complexity, the fundamental causes of land cover change are well known, and comprise biophysical, economic, policy/institutional and demographic factors. Of these, population growth and migration are implicated as the most significant. In future almost all land cover change will occur in rural areas which are experiencing rapid increases in population and migration, already a critical concern for conservation programs worldwide. Squatters are immigrants who illegally establish residence in areas deemed unsettled or with poorly-defined tenure. Their land use decisions tend to be risk-minimization strategies, influenced by their socio-economic characteristics and lack of long-term stakes in the land they occupy. This reduces their motivation to invest in and conserve resources, and they often become exceptional land degraders. In order to reduce their impact on landscapes and minimise the conflicts associated with their settlements, there is need for targeted research on their role in land cover change. I contribute knowledge to this problem using the Chyulu Hills as a case study. First, I reviewed current knowledge on migration and land cover change and the influential structural constructs that guide research in this field. Based on this review, I evaluated the dynamics of squatter migration to the Chyulu Hills, and the history of conflicts of squatters with the protected areas. Combining community surveys with statistical analysis, I revealed the proximate and underlying causes of the migration. I found that squatter movements to Chyulu Hills were widespread and long term, and that squatters are drawn from all over Kenya. Household socioeconomic and policy-related factors such as provision of schools, land tenure security and issuance of free land had a strong influence on squatter migration. Land and settlement policies were also implicated in migration and squatting. Most significantly, I established a high prevalence of secondary migration, where squatters migrated repeatedly to maximise their livelihood opportunities. The responsiveness of squatters to policy stimuli highlighted its potential to regulate or influence squatter mobility in rural landscapes. Next, I combined remote sensing/GIS tools, community surveys and landscape metrics to evaluate the rates and patterns of land cover change in the Chyulu Hills between 1967 and 1999, a period of high squatter migration. I found rapid changes in land cover over the 32 years of squatter settlements. The total proportion of native vegetation (dense forests, open forests and shrublands) decreased from 91% to 36%. In the same period, dense forests declined by 8.7% while cultivation expanded by approximately 40%. I explained these changes in the context of policy and management of land and the protected areas of Tsavo and Chyulu Hills. Squatter problems were related to land alienation during the British colonization of Kenya. I found preferential patterns of village establishment in or near dense forests. Landscape fragmentation confirmed the spontaneous nature of squatter settlements, and it is not clear yet how these transformations might affect the wildlife that uses it as a dispersal area. The community knowledge of the patterns and processes of land cover change, assessed using an adapted semi-quantitative tool known as proportional piling, agreed well with quantitative GIS measures of land cover change. Following this, I tested alternative postulates of the role of biophysical, household socioeconomic, access and institutional factors in the variation in remnant vegetation in villages in Chyulu Hills. Using Generalised Linear Models, model averaging and hierarchical partitioning analyses, I selected the best set of models within a 95% confidence interval, and ranked the relative importance of the explanatory factors in the variation in remnant vegetation. I found that squatter presence and their associated characteristics of short stays in villages, larger households and younger household heads were the most significant factors in remnant vegetation variations. This was supported by results of hierarchical partitioning. I then analyzed the policy on management of squatters and protected areas. For the most part, squatter management is ad hoc and reactionary, biased towards urban squatters, and largely limited to site and services improvements. I argued that this arose from a multiplicity of governance institutions and policies for natural resource management with overlapping mandates, rigid conservation management paradigms, and lack of provisions for participatory engagement of all stakeholders. I proposed the promotion of enabling policy environments that embrace participatory opportunities supported by building squatter human and social capital. Long-term land use planning should consider alternative livelihood sources that are less dependent on land as viable opportunities for reducing landscape changes and conflicts associated with squatters in rural areas. I recommend further research that integrates a broader spectrum of exogenous and contextual factors such as national policies, droughts and pandemics such as HIV/AIDS that influence household socioeconomics in rural land cover change. I also propose predictive studies that build potential future scenarios and that utilise finer temporal and spatial resolutions