Deforestation in Gola Forest Region, Sierra Leone: Geospatial Evidence and a Rice Farmer's Expected Utility Analysis

John Christian Abu-Kpawoh, Purdue University


Global land cover change constitutes the single major threat to ecological systems (Dale, 1997). These changes result from burning (Lavorel et al, 2007), agricultural expansion (Angelsen and Kaimowitz, 1999) and other activities triggered by socio-economic wants. Gola Rainforest National Park (GRNP) in Sierra Leone is under such threat from rice farmers. A 1926-gazetted demarcation established the park boundaries using stone landmarks, which was re-drawn in 1956 due to conflicting interests between the local communities and the British Colonial government. The increasing pressure from arable land-insecure population today has caused these gazetted landmarks to be overrun, compromising the efforts of the government and international organizations to protect the largest tract of forest in Sierra Leone. The incentive-driven interaction of competing land uses lead to inevitable land use change decisions in and around the park (Barbier, 1997), where the park is contextually a tract of forest cover surrounded by agricultural land. Evidence to date shows that land uses around the park have compromised the park’s quality as a sustainable forest reserve in the last twenty-five years (1991 – 2016). A major conservation problem relates to forest-edge communities need for arable land to practice their dominant slash-and-burn agriculture. We conduct geospatial analysis from Landsat images and show that there is land cover conversion in the region for the period 1991 – 2016. We hypothesize that rice farmers have driven this conversion. These farmers are risking penalties stipulated in the 2012 NPAA Act to convert forests to agricultural land. Using Erdas Imagine and ArcGIS Pro, we classify two Landsat images (January 1991 and April 2016). We obtain overall accuracy of 88% and 95% for the 1991 and 2016 classifications respectively, which contains four land cover classes, including water, forest, other vegetation/secondary vegetation, and bare land. The quantitative result of this classification suggests that the community forests in the Gola region (seven Gola chiefdoms) are losing forest cover at an annual rate of 4.18%, while the Park is increasing its forest cover at a rate of 0.03% per year. Even though the park appears to increase its forest cover, classification details reveal encroachment to just within the park from the boundary, compromising the quality and integrity of the park. Adapting the theoretical poaching model in Messer (2010), we model the farmer’s decision to engage in a profitable illegal farming (deforestation), as a function of expected returns and the cost of punishment. In this context, the farmer decides on the amount of hectares to illegally cultivate subject to a time constraint, and taking the risk of being caught as given, the latter being function of government law enforcement. We develop two cost studies, one for legal and one for illegal rice cultivation, and obtain the revenue parameters. Revenue parameters are validated through reports of the Planning, Evaluation, Monitoring and Statistics Division (PEMSD) of Sierra Leone’s Ministry of Agriculture, Forestry & Food Security. Our simulation results suggest that increasing the amount of fines imposed and/or lengthy imprisonment is not a feasible or realistic policy to dis-incentivize deforestation and illegal farming. The probability of being caught committing crime and the fine have to be both unrealistically large to deter a risk neutral farmer. Another policy relates to reducing the net-benefit from illegal farming, and increasing the risk of being caught. We find that a policy that increases the risk is most effective in reducing illegal farming with minimal policy change. We then compare the potential of these policies in reducing the incentives to illegally farm with the current policy consisting solely of law enforcement. We expect the results to be relevant to the authorities in Sierra Leone and other countries facing the responsibility of safeguarding natural resources and farmer subsistence while facing extremely limited financial resources and failing institutions.




Atallah, Purdue University.

Subject Area

Environmental economics|Natural Resource Management|Remote sensing

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server