Adding Fuel to the Fire; the Effect of Political Unrest on Forest Burning in Sub-Saharan Africa

Abstract

The connection between forest fires and political unrest has been noted throughout the literature (see Thomas, 2012,) and in a number of countries including; Malawi, Madagascar, Kenya and Ethiopia (Kull, 2002, Hoffmann et al., 2009, Cochrane, 2009 and Kuhlken, 1999), but never before has it been quantified. Political unrest can often manifest itself in the form of protest or armed conflict, these may involve fire setting. This investigation evaluates and visualises the correlation between these two parameters by linking the Moderate Resolution Imaging Spectroradiometer (MODIS) burned area product with World Bank political unrest indicators and the Armed Conflict Location and Events Dataset (ACLED) across sub-Saharan Africa. The investigation uses spatial techniques to analyse these data; the ACLED points were buffered by 50, 75 and 100km, these buffers were then used to extract the burned area data surrounding the points. These data were used to calculate what proportion of burnt area within that country burnt within the defined distances of a political unrest event; these outputs were then plotted against the World Bank political unrest data for the years 2002 – 2010, no data was available for 2003. This relationship was statistically tested using a Spearman’s rank correlation. Outliers were accounted for using other data sets including the Global Road Inventory Project (GRIP) road data, Global Land Cover 2000 (GLC 2000), and BioClim climate data. The results of this investigation show that there is a statistically significant relationship between political unrest and forest fire for all of the years being analysed (p = .000 - .017). This result is of great importance as currently no fire or forest risk models incorporate political unrest as a parameter. It also has implications for forest protection programmes, as if corruption and unrest are not addressed these schemes will never reach their full potential.