Satellite Study Reveals Alarming Forest Loss in Cameroon’s Santchou Wildlife Reserve

Deep in the western highlands of Cameroon, a protected area that should be a sanctuary for primates, pangolins, and forest antelope is quietly disappearing. A new study of the Santchou Wildlife Reserve, a 7,000-hectare pocket of Congo-Guinean forest ringed by eight villages, has combined more than two decades of satellite imagery with labor-intensive field inventories to paint the most detailed picture yet of how this reserve is being transformed. The findings, published in Discover Conservation, are stark: between 2000 and 2022, mature high-altitude forest collapsed from two-thirds of the reserve to little more than a third, while agriculture, human settlement, and illegal logging carved the landscape into a fragmented mosaic. With an annual degradation rate of 7.5 percent and a deforestation rate of roughly 1 percent, the reserve is losing forest far faster than Cameroon as a whole, and projections suggest the pressure will only intensify by mid-century.

The research team, led by Anaelle Brunda Djiaha and Marlène Ngansop Tounkam of the University of Douala together with Philippes Mbevo Fendoung of the National Advanced School of Public Works in Yaoundé, built their analysis on Landsat satellite scenes captured in the dry-season months of February 2000, March 2010, December 2014, and January 2022. Choosing dry-season imagery was a deliberate technical decision: cloud cover in Cameroon’s wet season can obscure the ground and corrupt classification algorithms, so the researchers favored dates when visibility was near optimal. Each scene was assembled into multispectral composites, radiometrically corrected to convert raw digital numbers into surface reflectance, and classified using the Maximum Likelihood algorithm, a supervised method that assigns each pixel to the land cover class it statistically most resembles. Five classes emerged: mature highland forest, medium-aged secondary forest, shrub savannah, agrosystems, and built-up areas.

The rigor of this classification was validated on two fronts. Confusion matrices built from thousands of reference points, ranging from 3,215 samples in 2000 to more than 31,000 in 2022, yielded Kappa coefficients between roughly 0.93 and 0.99 across all four dates, well above the 0.85 threshold generally accepted as indicating strong agreement between classified and reference data. On the ground, the team surveyed more than 110 GPS-truth points over 20 days of fieldwork, correcting misclassified pixel blocks and confirming which land cover categories were genuinely present. This dual validation matters because the entire downstream analysis, from change-detection matrices to carbon accounting, rests on the accuracy of those classifications.

The temporal story that emerges is dramatic. In 2000, high-altitude forest covered 66 percent of the reserve, roughly 6,199 hectares, with secondary forest at 13 percent, savannah at 15 percent, and agriculture and settlement each at 3 percent. By 2010 the mature forest share had slipped to 63 percent, but the steepest decline came in just four years: between 2010 and 2014, highland forest plummeted from 5,932 to 3,693 hectares, an annual loss of more than 570 hectares, while medium-aged secondary forest nearly doubled as degraded and regenerating land replaced old-growth canopy. By 2022, mature forest covered only 38 percent of the reserve, agrosystems had surged from 6 to 14 percent, and built-up areas had climbed to 5 percent, more than doubling from 136 to 456 hectares in just eight years. Transition matrices confirm the scale of conversion: over the full 22-year period, more than 6,000 hectares of land shifted out of secondary forest trajectories and over 1,250 hectares of mature forest were converted to other uses.

To understand what is driving this transformation, the researchers combined remote sensing with field observation and interviews. The verdict was unambiguous: agriculture accounts for 46 percent of degradation, human occupation for 34 percent, illegal logging for 14 percent, and bushfires for the remaining 6 percent. The reserve is surrounded by villages inhabited primarily by Mbo’o and Bamiléké communities whose livelihoods depend on cash crops such as cocoa, coffee, and oil palm as well as food crops like maize and cassava. A 20-kilometer secondary road linking Foumban to Bale has intensified access and anthropization, while selective artisanal sawmilling and slash-and-burn clearing accelerate the fragmentation. Elephants and panthers have already vanished from the reserve; the remaining wildlife, including primates, pangolins, aulacodes, and monitor lizards, now survives in an increasingly perforated habitat.

Beyond mapping change, the study quantified what this forest still stores in climate terms. Using established allometric equations for Cameroon’s tropical moist forests, the team measured every tree in 30 square plots of 30 by 30 meters distributed across the vegetation types, converting diameter and wood density into aboveground biomass, then into carbon using the standard 0.47 conversion fraction. The reserve’s overall carbon rate came to 0.874 tonnes of carbon per hectare, with medium-aged secondary forests standing out as the strongest carbon sinks at 0.534 tC/ha, followed by highland forest at 0.280 tC/ha. Agrosystems stored a negligible 0.002 tC/ha. The authors note candidly that these values are well below the 30 to 50 tC/ha typically reported for tropical secondary forests, attributing the gap to advanced degradation, plot representativeness, and extrapolation uncertainty, and they cross-validated their spatial carbon maps with an R-squared of 0.78 and a margin of error of roughly plus or minus 15 percent.

Perhaps the most consequential result is spatial: by overlaying the 2000 and 2022 forest classifications and subtracting what remained, the team identified 3,703 hectares, 53 percent of the entire reserve, as suitable for reforestation. These are zones where forest existed at the start of the century and no longer does, prime candidates for planting native species or assisted natural regeneration. If restored, the researchers estimate, this area could hold a carbon storage capacity of about 0.83 tC/ha, a meaningful contribution to both biodiversity recovery and climate mitigation in a region where protected areas are increasingly recognized as critical carbon reservoirs.

Looking forward, the team ran a CA-Markov cellular automata model in Idrisi Selva software, using transition probabilities derived from the historical record to simulate land cover in 2050. The projections, which the authors carefully frame as exploratory scenarios rather than predictions, suggest agricultural land expanding by 30 percent, built-up areas by 40 percent, and agroecosystems by 25 percent, primarily at the expense of remaining primary forest and savannah. The model rests on a stationarity assumption, that past trends will persist, which the authors acknowledge is vulnerable to disruption by policy shifts, land reform, or economic shocks, and they call for sensitivity analyses and hindcasting validation in future work. Even with those caveats, the trajectory is clear: without intervention, demographic and economic pressure will continue converting the reserve into farmland and settlements.

The policy implications extend well beyond one Cameroonian reserve. The authors argue that reversing these trends requires a comprehensive package: promoting sustainable agricultural practices, launching reforestation in the 3,703 hectares of priority zones, and, critically, involving local communities through training and participatory natural resource management. They point to evidence that REDD+ initiatives succeed best when payments for ecosystem services are coupled with local development programs such as agroforestry, and they stress the need for stricter land use regulations and the integration of conservation objectives into public policy. For a reserve whose very existence as a protected area is being tested by the communities that surround it, the path forward depends on aligning economic incentives with ecological survival, ensuring that Santchou’s remaining forests, and the carbon and biodiversity they harbor, are still standing in 2050.

Subject of Research: Forest cover dynamics, carbon stock assessment, and 2050 land use projections for the Santchou Wildlife Reserve in western Cameroon

Article Title: Forest dynamics and carbon stocks in the Santchou wildlife reserve from 2000 to 2022 and projections for 2050

Article References: Djiaha, A. B., Tounkam, M. N., & Fendoung, P. M. (2026). Forest dynamics and carbon stocks in the Santchou wildlife reserve from 2000 to 2022 and projections for 2050. Discover Conservation, 3(1), Article 28. https://doi.org/10.1007/s44353-026-00097-9

Image Credits: AI Generated

DOI: 10.1007/s44353-026-00097-9

Keywords: deforestation, forest degradation, carbon stocks, remote sensing, Landsat, Cameroon, protected areas, CA-Markov model, reforestation, REDD+, biodiversity conservation, land use change