Incidence and impacts of war-driven mammal declines in African ecosystems

Abstract

Warfare has been common in Africa, and the civilian and military movements, artillery use, lawlessness, poverty, and changes in conservation investment that accompany armed conflict can affect wildlife either positively or negatively. I provide the first quantitative synthesis of war’s impact on wildlife, and consider the aftermath of war-driven mammal declines for ecological communities.

In Chapter One I evaluate war’s impact on Africa’s large mammals. I pair literature-based density estimates for 36 species from 126 protected areas and 19 countries with conflict locations, and nine other ecological and socio-economic predictors which might explain mammal declines. For 1946–2010 (and 1989–2010) mammals declined more often where conflict was more frequent, and conflict frequency was the best predictor of mammal population trajectories.

In Chapter Two, I measure the impacts of war-driven mammal declines on vegetation structure in Mozambique’s Gorongosa National Park. Using historical (1977) and modern satellite (2012) satellite imagery, I document a one-third increase in tree cover across the 3670 km2 park following the near-extirpation of large herbivores during the country’s civil war. The results accord with those from smaller-scale experiments showing that the effect of eliminating grazers (fewer trees) is trumped by that of removing browsers (more trees).

Chapter Three is a global meta-analysis that finds the strength of large mammalian herbivore impacts (through competition, facilitation, and habitat modification) on other animals are stronger in areas of lower net primary productivity; there plants are slower to recover from direct effects of herbivory, which mediate herbivore impacts on animals. We can expect mammal declines (war-driven and otherwise) to have their greatest community-wide impacts in these low-productivity habitats.

Finally, in Chapter Four, I consider habitat selection in Gorongosa’s post-war, recovering populations of three congeneric antelope: bushbuck, nyala, and kudu. Understanding species’ habitat preferences allows ecosystem management for their benefit. Using hourly antelope movement data and GIS-based habitat classifications, I show that all three species preferred areas closer to termite mounds, plus certain fire regimes and vegetation structure. The strength of selection for each of these factors scaled with species’ body size, making this a rare demonstration of landscape-scale consequences of allometry.