A global analysis on impact of conflicts on long-term greening trends

English. Armed conflicts were and are still shaping the global terrestrial land surface and can have severe direct and indirect impacts on societies, economies and the environment. This study attempts to assess the impact of armed conflicts on vegetation on global and regional scale. Four aims were formulated to scrutinize the impact of conflicts based on 1) globally distributed conflicts involving events with more than 25 deaths, 2) types of conflicts, namely: state-based violence, non-state violence and one-sided violence, 3) regional examples: Rwanda and Afghanistan and 4) levels of conflict severity by death number: greater than 0, 25 and 100. The long-term impacts were analysed by calculating linear Normalised Difference Vegetation Index (NDVI) trends in locations hit by conflicts with respect to the incident date, using an ordinary least square model. To exclude climate influence on the vegetation, over the study period of 1982-2015, the global monthly 4km climate dataset TerraClimate was used to predict NDVI in a multiple linear regression for the periods after the conflicts. It was found that conflicts have neither positive nor negative significant overall impact on vegetation on a global scale (aim 1). Between the types of conflicts (aim 2) and the death number thresholds (aim 4) no significant differences could be identified. The Rwanda results, in contrast to the Afghanistan results (aim 3) showed a disproportional amount of negative NDVI slopes, but again the slopes where not significant. In conclusion, the analysis resulted only in insignificantly small trend changes, which leads to the assumption that on this scale, conflicts have no overall strong impacts on greening.

English. Introduction Armed conflicts are shaping the global land surface and can have severe direct and indirect impacts on societies, economies and the environment. This study looks at the impact of armed conflicts on vegetation on a global and regional scale. To analyse these impacts, satellites are often used, since they are observing areas which can be dangerous to access in war times, some of the datasets reach back over 30 years and they are often covering the whole globe. Four aims are formulated to analyse the impact of conflicts. For aim conflicts worldwide, were more than 25 people died (based on media reports), are considered. The second aim is based on different types of conflicts, which are defined by the Uppsala Conflict Data Program (UCDP). For the third aim the Rwanda genocide in 1994 is compared to the Afghanistan-US (and allies) War (2001 – 2008). In the fourth aim three different levels of conflict severity, defined by the number of deaths per conflict is used. Methods Several datasets for conflicts, vegetation and climate, which are all available for free online (see full thesis document for sources), are used in this research. The long-term impacts are analysed by calculating trends of the change in vegetation in locations hit by conflicts. For this, global satellite information over the period of 34 years (1982-2015) is used. A so-called vegetation index (VI), representing the greenness of plants, is used. It is based on the absorption and reflectance of light by leaves. To exclude the climate influence on the vegetation, over this long study period, a global climate dataset (TerraClimate) is used to compare the climate to the vegetation. The datasets are statistically compared with a correlation analysis. Results & Discussion Looking at the (linear) trends overall, neither positive nor negative significant impact of conflicts on vegetation, on a global scale (aim 1, 2 and 4), was found. The results of the analysis of Rwanda, in contrast to the results for Afghanistan (aim 3), show a disproportionally high number of negative trends, but again the slopes are not significant. Looking at the averages for the time before the conflicts and after the conflicts, it was found that the vegetation (greening) increased overall. Conclusions The analysis resulted in, on average, very small trend changes, which leads to the conclusion that on this scale, conflicts have no overall measurable impacts on greening. However, in some regions an abundance of negative or positive trends were found.