COS 286-5 - Predictive modeling of forest areas at risk for degradation due to unregulated charcoal extraction in Ghana

Abstract: Processes governing land degradation in Ghana are inextricably intertwined with regional-scale variations in the intensity of agricultural mechanization, local livelihood strategies, and policies relating to granting of mining leases. Studies in northern Ghana show that increased mechanization has facilitated farm expansion into forested landscapes. Bound by land tenure and availability, farm expansion facilitated by mechanization can lead to the loss of economically and culturally important trees. Losses of sustainable non-timber forest produce (NTFP) can then have a direct bearing on marginalizing women who have relied on NTFP harvests for supplementing incomes. Resultantly, livelihood strategies often shift to potentially environmentally unsustainable subsistence activities such as charcoal production. While charcoal production has historically been a mainstream supplementary livelihood strategy across the region and is generally considered a sustainable activity when local communities are involved and local customs are followed, abrupt shifts in livelihood patterns may affect forest degradation in several complex ways. One of the major factors determining forest degradation in Central Ghana is the experience of producers wherein newer entrants to the activity can potentially cause more harm than full-time producers. Shifting patterns of agricultural mechanization and mining leases may cause the creation of floating workers that may be relegated to exploring livelihood strategies that may have a deleterious effect on land resources. Degraded lands may then be either unavailable for cultivation, be rendered unfit for agriculture, and forest regeneration, or can be invaded by exotic species, and eventually lead to negative impacts on livelihoods and increased conflict.
In comparison to the task of detecting land degradation due to artisanal mining activities in the region, mapping charcoal operations has been difficult given the small spatial of operations that preclude the use of all but the highest-resolution data. In this talk, we will present a 'user-in-the-loop' predictive risk-mapping strategy that aims at optimizing efforts at managing and mitigating unregulated charcoal operations in the Savannah region of Ghana. Eventually, we aim to train collaborators in generating ‘risk maps’ that are predicated on utilizing observational and anecdotal spatiotemporal correlates of charcoal production to generate probabilistic risk maps of locations conducive to charcoal production.