Warfare-induced mammal population declines in Southwestern Africa are mediated by species life history, habitat type and hunter preferences

Civil wars often coincide with global biodiversity hotspots and have plagued the everyday reality of many countries throughout human history. However, how do civil wars affect wildlife populations? Are these impacts the same in savannah and forest environments? How persistent are the post-war consequences on wildlife populations within and outside conflict zones? Long-term monitoring programs in war zones, which could answer these questions, are virtually nonexistent, not least due to the risks researchers are exposed to. In this context, only a few methodologies can provide data on wild populations during war conflicts. We used local ecological knowledge to assess the main consequences of a prolonged civil war (1975–2002) in Southwestern Africa on forest and savannah mammals. The post-war abundance in 20 of 26 (77%) mammal species considered in this study was lower in open savannah compared to the closed-canopy forest environments, with some species experiencing a decline of up to 80% of their pre-war baseline abundance. Large-bodied mammals were preferred targets and had been overhunted, but as their populations became increasingly depleted, the size structure of prey species gradually shifted towards smaller-bodied species. Finally, we present a general flow diagram of how civil wars in low-governance countries can have both positive and negative impacts on native wildlife populations at different scales of space and time.

. Graphic depiction of the ordinal population abundance scale for different mammal species, consistently used as a visual aid during local interviews, which ranged from 0 (when the local population was conspicuously "absent") to 4 (when the population was "highly" abundant). Table   Table A1. Model average results. Estimated values indicate the coefficient associated with the variable listed on the left. This represents the estimated amount by which the odds (log x) of each response variable would increase if each explanatory variable were one unit higher. Standard errors are an average estimate of how much any given response variable would fluctuate if the study were re-run identically, but with new data.
Z-values indicate the degree to which explanatory variables exert a significant effect.
Pr(>|z|) are listed as two-tailed p-values that correspond to z-values following a standard normal distribution. Significance levels as following: ns P > 0.05; * P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001. Text S1. Qualitative results in relation to overall hunting pressure at Quiçama during the intermittent cease fire intervals between the beginning and the end of the 27-yr war period. We also include reasons for shifts in abundance of game mammal species, causes of species declines, and suggest remediation efforts.

Response
Fire-weapon disturbance: When we mentioned the overall impact of the civil war on wildlife, it is important to emphasise that if disturbance is restricted to only exchange of fire between combatants, but no hunting, the animals may be initially frightened, but once they perceive that they are not threatened, they typically remained where they were.
However, if some wild animals were slaughtered, they can perceive the non-human smell of blood, and thereby may move elsewhere.
Wildlife migration: The pronounced reduction in the local fauna is therefore due not only to the slaughter of local animal populations but also because this affects migration patterns. In particular, large carnivores migrate in search of prey, especially lion, which feeds on large prey. The animals that first migrated from open savannahs to forest areas of southern Quiçama and more distant regions were elephants and buffaloes, but those from the forest in northern Quiçama later left the study area, and some began to attack humans, thereby making it riskier to hunt them.

Refugees:
When talking about the impact of civil war on the environment, it is important to consider a wide spatial scale because while, on the one hand, there was out-migration of refugees and soldiers focused on killing residents in some areas, on the other hand areas of low confrontation received those refugees, who contributed to a significant mortality in the local fauna. Weapons: In addition to military personnel and local residents who worked in local civil defence, any civilian could barter weapons and ammunition with militants in exchange for natural resources. As weapons and ammunition became cheap, hunters would indulge in much less selective shooting. Also, given their lack of knowledge of the proper use of personal artillery, and since automatic rifles are not the most suitable weapons for game hunting, many animals were lethally injured but were not subsequently retrieved, thereby greatly contributing to incidental mortality and animal carcasses that were presumably consumed by scavengers. The most common weapons used were AK-47, G3 and mouser, but heavy weapons (e.g.RPG7) were also used to kill elephants.
Local demographics induced by the war: Little peace: In the 1990s it was still possible to find a few herds of red buffalo with about 70 animals. However, after the 1992 ceasefire, access to automatic weapons became even easier, and hunting pressure intensified, so buffalo herds were reduced to 20 individuals each. Currently, buffaloes are either solitary or live in small groups in Central and South of Quiçama. During the years of conflict, hunting was carried out closer to residential areas, but during the ceasefires and the post-war period (mainly after land-mines were disabled) the area became safer and hunters began to hunt farther away.
Current situation and future prospects: Hunting activities are still intense today at Quiçama, and in recent years, due to the economic crisis, it has further increased. Activities such as community-based game management could ensure wildlife population recovery as well as bring economic and social benefits to the local population. Also, the ecotourism potential of Quiçama remains very high but unexploited. Its proximity to the capital city of Luanda and highly photogenic and diverse habitats are ideal criteria for a successful ecotourism experience. Educational and health investments are also important to enhance local welfare and quality of life. Without these basic requirements, it remains an enormous challenge to even consider the possibility of sustainable game hunting.