Ecological drivers of ungulates Wildlife-Livestock interactions in Africa
The African continent needs to reconcile the exponential growth of its human populations and the preservation of its conservation areas in the years to come (Chape et al., 1$$&; Wittemyer et al., 1$$%a). Sub-Saharan Africa’s population was !.!0 billion in 1$1$ (World Bank) and could reach 1.!1 billion by 1$&$ (Ezeh et al., 1$1$). Africa is a continent where more than 0 million km² of land is protected (Chape et al., 1$$&) and where a wide diversity of large herbivores and large carnivores still exist (Fritz, De Garine-Wichatitsky, et al., !55#; Fritz & Loison, 1$$#). It is also a continent where competition for space is frequently accompanied by habitat fragmentation. Indeed, the strong growth of cultivated areas, the intensification of livestock activities and the expansion of urban areas are exacerbating pressures on W/L interfaces located at the periphery of protected areas (Craigie et al., 1$!$). Combined with the consequences of climate change, which are projected to intensify in the years to come (WMO, 1$1$), this anthropogenic pressure poses a significant threat to the sustainability of African ecosystems and particularly at the W/L interfaces (Vicente et al., 1$1!).
Many interfaces, interactions occur in the context of anthropized resources (e.g., plantations ecosystems), natural resources (e.g., wildlife) and a mix between natural and anthropized resources. In African W/L interfaces, natural resource preference can be considered as one of the key predictors of W/L interactions. Domestic and wild herbivores, for example, have a phylogenetic proximity that induced similar resource requirements and thus, potential resource competition. Even if dietary niche partitioning can condition resource competition between animal species, large herbivores consume several shared forage resources. Indeed, most of the literature provides evidence for potential competition via dietary overlap (Breebaart et al., 1$$1) and emphasizes that dietary overlap is greater for animal species with similar body size (e.g., gut capacity, bite size, food intake rate, and feeding site selection) (Kartzinel et al., 1$!&). Domestic animal species such as cattle, for instance, are more likely to compete with similar size ruminant grazers such as African buffalo than browsers (e.g., kudu), non-ruminant grazers (e.g., wildebeest), very large herbivores (e.g., elephant), or small herbivores (e.g., duikers). However, empirical studies on foraging between herbivores of different body sizes, and the nature of the interspecific interactions between these herbivores does not always match theoretical expectations (Stears & Shrader, 1$1$). Larger grazers can compete with smaller grazers by reducing food availability, especially during the dry season in tropical areas (Arsenault & Owen-Smith, 1$$1), or through more long-term negative effects via habitat modification (Prins, 1$$$). Conversely, small herbivores can potentially outcompete larger herbivores by reducing the availability of high-quality forage in areas where resources can be scarce according to season (Illius & Gordon, !5%3). If competition influences W/L interactions, facilitation can also play a role in mechanisms that may increase or decrease the likelihood of direct or indirect interactions. Some wildlife species such as the wildebeest can select areas grazed by livestock contrary to the buffalo that tend to avoid such areas (Tyrrell et al., 1$!3a). Facilitation may lead animal species to share forage patches and use these patches successively to allow post-grazing regeneration (Odadi et al., 1$!!a) or high-quality grass regrowth thus positively easing forage accessibility for smaller grazers (Western & Gichohi, !552).
When natural resources (e.g., pasture, surface water) are abundant and widely distributed, free ranging wild animals tend to avoid areas frequented by livestock due to direct competition, as described above, or simply because of fear of humans (Connolly et al., 1$1!; Riginos et al., 1$!1). However, the behavioral response by wildlife to the proximity of strongly anthropized areas differs according to animal species and locations. In savanna W/L interfaces found in Zimbabwe for instance, African buffalo completely avoid communal areas whereas elephants sometimes enter these territories for crop raiding (Guerbois et al., 1$!1). These differences in behavior imply a diversity in frequency, temporality, and location of interactions between wildlife and livestock. On the contrary, when natural resources are limited, localized areas where resources can be found become potentially favorable to the aggregation of many animal species. In arid and semi-arid areas, animal species spatial distribution is constraint by the location and availability of this natural resource (Ogutu, Reid, et al., 1$!0). Seasonal variability encourages resource-driven patterns and intensify potential interactions when the resource becomes scarce due to the limited distance that animal species can travel and therefore, their ability to reach other resource points potentially still available (Valls-Fox et al., 1$!%). However, this analysis must be nuanced. When natural resources are sparsely distributed, animal species or group of animal species that have a limited movement radius around their core home range can use different resource points that are too far apart for them to interact (Borchering et al., 1$!3).
The risk of pathogen transmission favored by the increase in contact between different animal species
Direct and indirect interactions between different animal species potentially result in the transmission of pathogens from wildlife to livestock, and from livestock to wildlife in space and time (Nugent, 1$!!). Most pathogen transmission events remain undetected at the W/L interface and when they are detected, it is difficult to assess when and where exactly they have happened accurately (Voyles et al., 1$!&).
In pathogen transmission event, wildlife can act as maintenance hosts for diseases, exacerbating the circulation of pathogens and their circulation within W/L interfaces (Bengis, Kock, & Fischer, 1$$1). Livestock can be directly impacted through increased mortality and reduced productivity that, in turn, can affect human societies via economic losses associated with cost of control, loss of trade, decreased market values and food insecurity (Dehove et al., 1$!1). Animal species respective roles and relative importance are extremely difficult to quantify and disentangle in a system associating wildlife and livestock (Lefevre et al., 1$!$). This diversity of species included in host communities implies complex and dynamic mechanisms of W/L interactions who are dependent on seasonal dynamics, strains circulations, virulence according to the inter-specific contacts, intra and inter-specific contact rates and frequency of contacts (Miguel, 1$!1a).
On the other end, pathogen transmissions have the capacity to directly impact wildlife by disturbing whole species’ health, affecting biodiversity, engendering changes in animal behaviors and population compositions or, in the most extreme cases, causing community collapse with multiple extinctions (Williams et al., 1$$1). Inventory of known livestock pathogens revealed that 33% are capable of infecting multiple host species, including wildlife (Cleaveland et al., 1$$!). This potential of pathogen transmission from livestock to wildlife can be amplified as we know that W/L interfaces are dynamic and bidirectional with pathogens circulating freely within and between wildlife and livestock species (Bengis, Kock, & Fischer, 1$$1). Indeed, most wild animal species are sympatric using shared resources (e.g., pasture, surface water), are in direct interaction with similar vectors and are exposed to human negative effects on their habitats (e.g., shifts in farming practices, land use changes, deforestation, encroachment into pristine habitats) (Perry et al., 1$!2).
Successive pathogen transmission events between sympatric hosts can as well occur over time (de Garine-Wichatitsky et al., 1$!2) via large scale movements (i.e., regional, continental, and intercontinental movements). For instance, in the Great Limpopo TFCA, pathogen and subsequent disease transmission risks across this W/L interface exist and can occur both ways threatening cattle and wildlife population (Caron et al., 1$!#). These patterns of pathogen circulation reinforce the risk of potential transmissions to humans as a great number of these circulating pathogens are zoonoses (Jones et al., 1$$%). Humans have evolved in proximity of animals for a long time, especially with domestic animals with whom they exchange pathogens frequently through diverse transmission modes (e.g., contact via animal husbandry activities, animal consumption) (Cleaveland et al., 1$$!). Of course, direct, and indirect interactions between humans and wildlife are less common but exist. Wildlife represents a direct source of pathogens which can lead to pathogen jumps and result in disease developments within the host. For these phenomena to occur, livestock species can perfectly play the role of “bridge” between humans and wildlife (Caron et al., 1$!&).
The consequences of pathogen transmission favored by W/L livestock interactions for the currently globalized human societies can be terrible for humans (in terms of health, political, social, and economic aspects), as the on-going severe acute respiratory syndrome coronavirus 1 (SARS-CoV-1) pandemic is demonstrating us (at the time of writing this manuscript), but also for the animal well-being in the case of some pathogens. Until recently, wildlife wasn’t considered as impacted by pathogens and researchers, managers as well as the general public were not as aware and equipped to apprehend and to measure such risks (de Garine-Wichatitsky et al., 1$!0). Therefore, our ability to model W/L interactions and the associated potential disease transmissions at the W/L interface is crucial in order to better apprehend the functioning of these complex ecosystems, thus improving our capacities to counter the constant increase of these phenomena more efficiently (Gibb et al., 1$1$).
Rural Zimbabwe, a land of contrasts and upheaval
Access to, the use of, and the global apprehension of natural resources have changed over the years in Zimbabwe, thus shaping the structures and organizations of its contemporary W/L interfaces. Three representative periods of natural resource conservation and utilization can be identified in that regard, namely the precolonial period, the colonial period, and the post-independence period (Muboko & Murindagomo, 1$!0).
History of biodiversity conservation and the relationship with wildlife in Zimbabwe (Pre-colonial and colonial periods)
In Precolonial Zimbabwe (before !%5$), societies, large and small, were mainly farming communities practicing agriculture and pastoralism in a subsistence-oriented economy that included trades. Several multi-ethnic empires succeeded one another according to periods of expansion and isolation. In this period of time, access to natural resources was mostly governed by traditional beliefs, taboos, and customs (Chenje et al., !55%; Kwashirai, 1$$3). Specific areas were believed to be hosts to some spiritual forces and thus were considered as sacred sites where visiting, hunting, collecting fruits, extracting firewood, and any other natural products were prohibited activities (Chemhuru & Masaka, 1$!$). While the sacredness of some sites is difficult to establish and formally prove, the fact that such myths helped protect natural environments, as some areas remained intact, is clear. Traditional societies enforced wildlife conservation by discouraging indiscriminate killing of animals, as such acts were punishable by the spirits conveyed through control mechanisms of traditional taboos, totems, and customs (Kwashirai, 1$$3). Historically in Zimbabwe, intensive cattle productions were clustered along the edges of the Matabeleland plateau close from the line of maximum possible extension of the tsetse fly (genus Glossina), vector of human and animal trypanosomiases (Garlake, !53%). This settlement configuration allowed herding extension in additional areas during the time of the year when the fly was not dangerous for humans and livestock. Seasonal transhumance was necessary in order to extend the grazing range and supposedly to alleviate pressures such as lack of political control of the plateau, demands on the land by population or cereal agriculture, exhausting of grazing areas condensed on the plateau by intensive herding practices, and seasonal loss of nutrition of the plateau grasses (Garlake, !53%). In the !5th century, the population densities were far less than what they are today (23 inhabitants per km² in 1$!%, source: The World Bank), the human habitats were more fragmented across the land, and land-tenure system functioned as a mechanism of social control (Cousins et al., !551).
During the colonial period (from !%5$ to !5%$), colonialism gradually disrupted the traditional structuration of societies by introducing Christianity, mercantilism, and capitalism, thus favorizing the emergence of new identities, new commodities, new languages, new ideologies, new political and economic outlooks, and new relationship to the environment. The British colonial administration progressively introduced protective and command type natural resource and wildlife conservation legislations to preserve a once plentiful wildlife population which had been severely endangered by the great rinderpest epidemic of !%5#-!%53 (Onselen, !531) as well as by intensive exploitation by slave traders, hunter explorers, prospectors, and adventurers (Child, 1$$%). In !515, the Game and Fish Preservation Act gave the governor of colonial Zimbabwe the ability to control the exploitation of wildlife and resulted in the creation of several natural reserves that correspond, for some of them, as the current National Parks (e.g., Hwange National Park) (Bond & Cumming, 1$$#). Wildlife populations increased drastically as a result of such law and began, at the same, to threaten human settlements and commercial cattle ranching by competing for grazing and harboring pests and diseases (G. Child, 1$$%). In !52$, the Land Apportionment Act (Jennings, !52&) divided the entirety of the land into European settler’s areas and African native reserves. The application of this act gradually led to the emergence of a landholding structure where only 0,%$$ large-scale “European-colonial” commercial farmers occupied !!.1 million hectares of land amongst the most fertile while one million communal-area families occupied only !#.2 million hectares located in marginal agricultural areas (Chenje et al., !55%). In !53&, the Parks and Wildlife Act gave responsibility for wildlife to the private landowner (Murombedzi, 1$!$). The purpose was to protect wildlife populations within protected areas due to the deterioration of migration routes, to reduce wildlife management costs, and to reinforce the colonial government authorities in involving the private sector to the wildlife management to avoid personal interests (Duffy, 1$$$). The Parks and Wildlife Act was not extended to communal areas and local populations but to designated administrative authorities responsible to manage communal lands. This uneven system resulted in disastrous disequilibrium. Natural resources in the “European-colonial”.
In a context of increased contact between wild and domestic animal species at interfaces: the choice of two focal species
The African buffalo (Syncerus caffer) and cattle (Bos taurus & Bos indicus) (Figure !.2) are keystone species for conservation and production systems in W/L interfaces in Zimbabwe. The African buffalo is one of the “Big Five” (P. H. Williams et al., 1$$$) and is an important member of the ungulate guild who shapes habitat heterogeneity in and outside protected areas where the human presence is low (Estes, 1$!1). Cattle, in subsistence farming communities, provide draught power, source of protein, cash incomes, safety net and social status (Baudron et al., 1$!1). Buffalo and cattle are both grazer ungulates, close phylogenetically, sharing common resources (i.e., forage and water) (Hoffmann, 1$$1), and are thus likely to overlap in range and compete for resources, particularly in environments where natural resources are spatially segregated (e.g., savannas) (Odadi et al., 1$!!). Both species rely on their behavior and the management of the land use by humans to cope with constrained access to natural resources (e.g., access to artificial water, forage intake by the herder) (Kaszta et al., 1$!%). Their shared use of space increases the likelihood of direct and indirect interactions which, in turn, promotes the risk of pathogen transmission (Caron et al., 1$!2), a threat to farmers and biodiversity conservation (Caron et al., 1$!2). In the context of this thesis, we have decided to focus on those two particular species in order to deploy our methodology and follow up on several previous studies (Miguel, 1$!1; Perrotton, 1$!&; Valls Fox, 1$!&) that have focused on these two animal species and the different problematics associated with their interactions.
Intrinsic behaviors of one of the subspecies (Syncerus caffer caffer)
The African buffalo (Syncerus caffer) is a ruminant mammal belonging to the Bovidae family (Figure !.0) and is the existing largest and most massive of the African bovids. The African buffalo is currently considered as a single species despite displaying important morphological variations such as body size, weight, fur color, horn shape, and length according to geographical locations. Four subspecies form the entire African buffalo population: Cape buffalo (Syncerus caffer caffer), forest buffalo (Syncerus caffer nanus), West African savanna buffalo (Syncerus caffer brachyceros) and Central African savanna buffalo (Syncerus caffer aequinoctialis) (East, !555). The population density of these subspecies is unevenly distributed throughout the African continent. The highest population densities are found along the African Rift, in East Africa (Figure !.&).
In this thesis, we are focusing our interest on the Cape buffalo as this specific subspecies is the only representative of the Syncerus caffer species in southern Africa, and more particularly in Zimbabwe, our defined study area. The social structure of the Cape buffalo has been closely studied, even if most of the studies are often descriptive (Prins, !55#; Ryan et al., 1$$#; Sinclair, !533). Cape buffalo live in large herds containing &$ to &$$ animals (Cornélis et al., 1$!0) and a number of smaller social groups made up of several females and their most recent offspring (up to two years of age). The herd structure tends to maintain cohesion in order to provide protection for weakened individuals even if complex fusion-fission dynamics exist within herds (Wielgus, 1$1$). Bachelor groups containing as many as a dozen or so males, along with groups of similarly aged juveniles are also found within the herd substructure. Adult males either associate with a female group or distance themselves apart from the herd in a small unit of similar older males. In some instances, old males can be solitary, living away from the herd from which they originate (Grzimek, !55$). Where and when large and rich pastures are present, temporary aggregations of 1,$$$ to 2,$$$ buffalo can potentially form from several smaller herds (Kingdon, 1$!&). However, contrary to the smaller herds, these large groups lack social cohesion and occur occasionally. Cape buffalo tend to be non-migratory, usually inhabiting an exclusive home range that can vary in size (Figure !.0) and are specific to one herd (Nowak & Walker, !555). In southern and eastern Africa though, Cape buffalo herds can periodically subdivide due to fission (splitting) – fusion (merging) dynamics but within the herd’s usual home-range (Prins, !55#; Ryan et al., 1$$#).
Intrinsic behaviors of the species and its relations with the surrounding environments
Cattle (Bos taurus or Bos indicus) are ruminant mammals belonging to the Bovidae family (Figure !.#) and domesticated by human societies. They are considered escaped or released domestic animals because if not well contained by adequate fences or herder’s directives, cattle tend to form feral herds and wander into native vegetation wherever suitable food is available (Findley, !53#). A cattle’s herd is structured according to a dominance hierarchy where each individual yield to those above in the hierarchy. The hierarchy’s status is hereditary and as a result, calves adopt their mother’s status. Females protect their respective calves by chasing anything that threatens them and is not reluctant to share parental care within the ensemble of the herd. Dominant males maintain this status until defeated by younger males in challenges. (Patent & Munoz, !552). Cattle’s herd home-ranges vary greatly (Figure !.#) according to geographical locations and their respective socio-ecological organizations and seasonal climate fluctuations (Moyo et al., 1$!2). In some W/L interfaces, where restrictions of access due to seasonal crop productions apply, cattle can be encouraged by the herder to range away from communal land into natural reserves or National Parks despite not having an official granted access (Valls-Fox et al., 1$!%).
Cattle usually feed on grasses, stems, and other herbaceous plant material present in pastures not maintained by humans or in open agricultural fields and consume about 3$kg of grass in an %-hour day in average (Ng, 1$$!). Cattle can modify native vegetation by browsing, crushing, and trampling, and in areas with high human density, they can severely impact natural systems, causing erosion, introduction of non-native grasses and herbaceous plants, destruction of riparian habitats, as well as overgrazing (Moyo et al., 1$!2). In addition to their grazing activities, cattle need to drink every day, and in some instances, are totally dependent on the water resources made available by humans (e.g., borehole, dip tank) in areas where the availability of surface water is spatially and seasonally sparce.
Table of contents :
Chapter !: General introduction
!.! The human/domestic animal/wildlife interface: places of interaction, places of emergence
!.!.! Current trends (Anthropocene / Global changes) and wildlife-livestock interface
!.!.1 Ecological drivers of ungulates Wildlife-Livestock interactions in Africa
!.!.2 Anthropogenic drivers of Wildlife-Livestock interactions
!.!.0 The risk of pathogen transmission favored by the increase in contact between different animal species
!. » Rural Zimbabwe, a land of contrasts and upheaval
!.1.! History of biodiversity conservation and the relationship with wildlife in Zimbabwe (Precolonial and colonial periods)
!.1.1 Complex interactions between different actors in contemporary Zimbabwe. Social, economic, and political contexts (post-colonial period)
!.1.2 A changing climate that upset an already fragilized balance at the W/L interface
!.# In a context of increased contact between wild and domestic animal species at interfaces: the choice of two focal species
!.2.! The African buffalo (Syncerus caffer)
!.2.!.! Intrinsic behaviors of one of the subspecies (Syncerus caffer caffer)
!.2.!.1 Direct links of the Syncerus caffer caffer with its immediate environment
!.2.1 Domesticated cattle (Bos taurus & Bos indicus)
!.2.1.! Intrinsic behaviors of the species and its relations with the surrounding environments
!.2.1.2 A domestic animal species dependent on herders’ decision rules
!.% Research questions & objectives of the thesis
!.0.! The origin of the thesis research’s questions
!.0.1 Three main research questions
!.0.2 The structure of the manuscript
Chapter « : Study sites and data used « .( Preamble
« .! Three study sites, three W/L interfaces
1.!.! The interface of Hwange/Dete
1.!.!.! Geographical, climatic, vegetation and hydrology characteristics
1.!.!.1 Current ecological context
1.!.1 The interface of Gonarezhou/Malipati and Kruger/Pesvi
1.!.1.! Geographical, climatic, vegetation and hydrology characteristics
1.!.1.1 Current ecological context « . » Data used in this thesis
1.1.! Telemetry data
1.1.!.! Capture of animals and installation of GPS collars
1.1.!.1 The utilization of pre-processed telemetry data
1.1.1 Remote sensing data
1.1.2 In-situ data and empirical knowledge « .# Chapter summary
Chapter #: Literature review of the environmental drivers influencing the buffalo and cattle movements in space and time
#.! The article
2.!.1 Review Article Methodology
2.!.2 Environmental Drivers Influencing the Movements of Buffalo and Cattle and the Satellite Remote Sensing Tools to Characterize them
#. » Chapter summary
Chapter %: Characterization of the study sites’ environmental variables via satellite remote sensing %.
%.! A three steps classification methodology
%. » Characterizing the surface water
0.1.! Methodological approach
0.1.1 Results and descriptions
%.# Discriminating the agricultural areas
0.2.! Methodological approach
0.2.1 Results and descriptions
%.% Producing the final landcover maps
0.0.! Methodological approach
0.0.1 Results and descriptions
%.’ Chapter summary
Chapter ‘: A spatialized mechanistic animal movement model based on collective movements of self-propelled individuals
‘.! Overview of the animal movement ecology
‘. » Mathematical models follow two paradigms
&.1.! Statistical models
&.1.1 Mechanistic models
‘.# The choice of a mechanistic model based on collective movements of self-propelled individuals
&.2.! The individual versus the collective
&.2.1 The synthesis of the two, when individuals influence the collective
‘.% Spatializing the model and combining it with SRS data
&.0.! The notion of space in animal movement modelling
&.0.1 The language “Ocelet”
Chapter ): Combined used of remote sensing and spatial modelling: When surface water impacts buffalo (Syncerus caffer caffer) in savanna environments
#.!.1 Material & method
#.1.! A movement model that also consider the landcover
#.1.1 Application in Gonarezhou/Malipati and Kruger/Pesvi ).# Chapter summary
Chapter +: Spatial modelling of contacts between wildlife and livestock in Southern Africa
+.! The article
+. » Chapter summary
Chapter *: General Discussions & perspectives
*.! Summary of the objectives and findings