Facing old and new risks in arid environments: The case of pastoral communities in Northern Kenya

Pastoralism is an important form of livelihood in Kenya, particularly in northern Kenya. While pastoralists have always faced risks such as political marginalization, harsh climate conditions and violent conflict, pastoral communities are increasingly exposed to new risks such as wildlife conservancies as well as large-scale wind and oil projects. The growing climate security literature has provided some insights into how changing rainfall patterns are affecting pastoralist conflicts, but we know little about the compound nature of multiple risks. This paper seeks to narrow this knowledge gap. It aims to better understand new and old risks and their combined impact on pastoral communities in the counties of Turkana, Samburu and Marsabit. The study is based on a comprehensive and structured review of the scientific literature. The findings show that all pastoral communities in northern Kenya face the old risks while the newer ones are county-specific. In Turkana, potential oil spills threaten land and water resources upon which pastoralist communities depend. In Samburu, wildlife con-servancies have changed the land-use system, and in Marsabit a pastoral community is exposed to noise and visible emissions from a wind park. Common to these risks are that they: (1) are caused by top-down governance processes with little to no community involvement; (2) reduce the mobility and access to water and pasture of (some) pastoral communities; (3) change the security situation. A rethinking of wildlife conservancies and energy projects is needed that involves pastoral communities from the outset and makes them the primary beneficiaries of any measure implemented in northern Kenya.


Introduction
Across the world, nomadic pastoral communities face a wide range of risks and challenges. These include privatization of land, fragmentation of rangelands, violent conflict, climate and environmental change, development projects, relocation and sedentarization programmes as well as economic and social marginalization [1][2][3]. A common outcome of these developments is that they limit the mobility that the pastoral communities depend on for their livelihood [4]. For example, in the Tibetan Plateau the combination of climate change and privatization of land "may have dramatic negative consequences for pastoralists". In Norway, national policies releases and newspaper articles. Based on the reviewed literature, the risks addressed in this article were chosen inductively. We use a common definition of risk which identifies it as the outcome of (1) the likelihood of an event to occur and (2) its potential impact [26]. We define "old risks" as those that pastoral communities have been facing for at least three decades. "New risks" are those that have emerged more recently. While private wildlife conservancies have existed before, community conservancies "began to mushroom" [25] after the Northern Rangelands Trust (NRT) as the key conservancy actor was founded in 2004 [27]. Construction of the Lake Turkana Wind Power (LTWP) park started in 2014 and the first electricity was produced in 2018 [28]. The discovery of oil in Turkana dates back to 2012 [29].
Northern Kenya and specifically the counties of Turkana (oil), Marsabit (wind) and Samburu (conservancies) were chosen as the study region because these counties share some common attributes but also some distinct features which allow for a productive comparison and the identification of risks that pastoral communities in Kenya and other countries in East Africa face now or in the future. Because of the dry climate, rainfed agriculture is very limited in northern Kenya (Fig 1). Fishing is practiced in Lake Turkana but the main livelihood in northern Kenya is pastoralism. The main pastoral groups are the Turkana, Samburu, Borana, Somali, Pokot, Ilchamus, and Rendille. Each group has its own language and some of the groups are in violent conflict with each other which poses an obstacle to development [19,20].

Marginalization
We define marginalization as the process in which a group is pushed to or kept at the margins of a society [see also 32]. The result is that this group cannot fully participate in society and feels discriminated and neglected. Studies suggest that this is certainly the case among pastoral communities in northern Kenya [19,20,33,34]. This perception is reinforced by the feeling that the government does not protect pastoral communities against risks such as climate change and violent conflict [e.g. 35].
Marginalization and discrimination can be based on different attributes of a group [36]. The three attributes most relevant here are ethnicity, geographic location, and livelihood. The Turkana and Samburu are minor ethnic groups in Kenya. Only about 2% of the Kenyan population are Turkana and about 0,6% are Samburu [37]. While the process of devolution, started in 2010, has shifted more political power and financial capacities to the county level, Turkana, Samburu and Marsabit are still among the least developed in Kenya in terms of per capita income, education and health [ Table 1, see also 38,39,40]. This shows that the marginalization which began in colonial times has persisted until today [20]. The three counties are also being marginalized because of their peripherical geographic location and distance to the central government in Nairobi in the south-central part of the country [41]. In its Development Strategy for Northern Kenya and other Arid Lands, the Government of Kenya notes in 2013 "[t]he defining feature of Northern Kenya is its separation from the rest of the country, which manifests itself in both physical and psychological ways." While in the same document this is identified as the "primary policy challenge", the central government has shown little interest in the country's north, likely because of its low tax base and insignificant contribution to the gross county product [see also 42,43]. The third attribute that contributes to the marginalization of the population in the north of the country is its livelihood. The narratives used in official government documents range from portraying "mobile pastoralism [being] the most appropriate livelihood and land use system [in ASAL]" to "mobile pastoralism was irrational, unproductive and environmentally destructive" [44]. In 2022 a State Department for the ASALS and Regional Development (SDARD) was established. Among the key objectives of the SDARD is the "socio-cultural integration" of ASAL [45].
In all three counties and particularly in Samburu, the per capita income is significantly lower than the average in Kenya (Table 1). Turkana has the lowest level of primary education and Marsabit the highest under-five mortality rate. Several national and international

Climate change
Kenya is prone to extreme weather events and has been exposed to drought in recent decades . Studies project that mean annual temperatures in East Africa will continue to rise [61], with higher mean annual precipitation [62] and a complete deglaciation of Mount Kenya before 2030 [63]. Depending on the spatial and temporal scale, the model used, and consideration of various factors such as temperature or precipitation, drought characteristics are projected to increase or decrease [16,64,65]. Overall, it can be said that different climatic factors have become more variable and further changes are expected in the future, resulting in new challenges to which people must adapt [e.g. 55,66]. These changes have a particular impact on pastoralists in Kenya, who are mostly nomadic and depend almost entirely on livestock for their livelihoods and thus primarily rely on rainfed pasture and water for their livestock [54]. For example, one study examined the impacts of these changes on the Turkana, where climate change is leading to dwindling pastures, reduced water availability due to less rainfall, longer dry spells and, thus, decreased land use options, resulting in less water and grass available for livestock and, consequently, inadequate animal nutrition [67]. In addition to climate change, land use change and sociopolitical and institutional changes are leading to the loss, fragmentation and degradation of environmental resources upon which pastoralists rely [68]. These changes lead to resource competition and their large-scale migration in search of water and pasture [54].
As a result, livestock numbers are also in decline, leading the Turkana to look for alternatives such as firewood and charcoal burning, the sale of local brew, and fishing in Lake Turkana, while some turn to agropastoralism [67,69]. In response to the new challenges and environmental shocks and stresses, pastoralists in northern as well as in montane Kenya have increasingly turned to camel herding and farming to better cope with unexpected environmental changes, making this a climate change adaptation strategy to improve their climate resilience [55,70]. Pollini and Galaty [71] list three main strategies by which pastoralists adapt: intensification (increasing productivity per hectare by changing land use systems), extensification (pastoralists expanding their territory into unused areas or neighbouring communities), and diversification (in which pastoralism is complemented by other elements such as farming, tourism, or wage jobs, usually linked to migration). The latter occurs more frequently among pastoralists in East Africa [72]. Diversification into conservation and tourism is mainly used by pastoralists in wildlife-rich rangelands, which are important alternative sources of income during droughts [73]. Other activities include sand harvesting, used, for example, by young schooled people in Laikipia North subcounty in Kenya, to replace or complement pastoralism [74]. In Samburu, some pastoralists increased their herds, while others used income diversification as a strategy or switched to completely non-herding options such as casual labour, tourism, charcoal or murram sales, family remits or government assistance [75]. In lower eastern Kenya, irrigated maize and in situ water conservation measures are seen as the most effective adaptation strategies [76]. In general, irrigation is a climate-smart practice, which goes hand in hand with the shift from pastoralism to agropastoralism in dryland communities [77]. To increase water availability for irrigable farming and livestock, agro-pastoral dams and irrigation schemes have been built [78]. Other adaptations practiced in Marsabit, Turkana and Laikipia County include livestock migration or higher livestock mobility to find new pasture and water resources, forage production, engagement in livestock markets, and improved livestock breeds or more diverse herd composition [12,79,80]. Another longterm strategy applied by Turkana pastoralists is sending children to school to improve their education, with the plan that this long-term investment will eventually pay off in higher incomes [12]. However, not all alternatives are suitable for guaranteeing secure long-term livelihoods; moreover, some are partly unsustainable, such as selling charcoal, which results in deforestation and other negative long-term effects, and fishing, which leads to overfishing, receding water levels and competition for Lake Turkana [67].

Violent conflict
As the previous section has shown, climate change is affecting the availability of water and pasture; these, in turn, are at the centre of violent conflicts between pastoralist communities [19,81,82]. Hence, climate change has at least an indirect impact on these conflicts. Conflicts between pastoral groups mostly occur in the form of livestock raids in which one group attacks the other to acquire livestock or gain control over water and pasture [24]. In Kenya, raids mostly take place in the north of the country in the counties of Turkana, Samburu and Marsabit (Fig 1). The source of the violent events shown in Fig 1 is the ACLED (Armed Conflict Location & Event Data Project) database. ACLED relies on trained data experts and a variety of sources (e. g. newspapers, reports) to collect the data. Almost all violent events shown in Fig 1 were battles, defined as "a violent interaction between two politically organized armed groups at a particular time and location." The majority of these battles in northern Kenya are livestock raids [83].
There is no agreement on the exact links between climate and environmental conditions, resources, and raids, but several studies provide useful insights. For instance, Adem et al. [84] found that drought conditions and resource scarcity are a motivation for Turkana pastoralists to raid other groups and access pasture. Similar findings were made by Schilling et al. [19]. Seter et al. [85] reviewed cases of conflict between pastoral and farming communities in East and West Africa and noted that drought played a role in two of the three cases located in Kenya [see also 86]. In contrast, others have argued that wetter periods are more violent because of tactical factors. During the rainy season, livestock is healthier and, hence, more likely to survive the journey from the place of the raid back to the attackers' camp. Furthermore, higher vegetation provides more cover for raiders who are also less occupied with keeping their own herd alive [82,87]. A possible explanation for these seemingly opposing explanations can be found in the "Resource Abundance and Scarcity Threshold" (RAST) hypothesis: "In regular years with sufficient rain, raiding is mostly conducted before and during the rainy seasons [. . .] [b]ut when rains partly or completely fail and a certain threshold of resource scarcity is reached, raids are conducted despite less fortunate raiding conditions" [24]. In contrast to wet season raids, the primary objective of raids during drought conditions is to gain or secure control over scarce pasture and water resources rather than acquiring livestock. It remains unclear, however, exactly where the threshold between resource abundance and resource scarcity as a driver of one or the other type of raid lies [88].
Unlike the link between climate conditions and raids, the effects of violent conflict on pastoral communities are less disputed among scholars and generally seem much clearer. The raids have direct and indirect effects on the communities involved. Among the direct effects are higher rates of mortality, physical injuries-for example, caused by gunshot wounds or when people fall while fleeing attacks-and loss of livestock and access to water and pasture. This further impacts the psychological health of communities, who experience trauma, grief, and a general feeling of insecurity. These direct effects cause indirect ones, including hunger, displacement and poverty [22,23]. These are not only the results of attacks but also due to the inefficient use of resources that general insecurity causes. For example, there is underutilized pasture in the border region between the Turkana and Pokot because both groups are too afraid of attacks to access it [19]. The feeling of insecurity also undermines the social contract between pastoral communities and the state [35]. As the communities feel unprotected by the state against attacks by other pastoral groups within and outside of Kenya, trust by pastoral communities in state authorities tends to be low [20].

Wildlife conservancies
The number of conservancies has been growing strongly over the past two decades in Kenya. There are 167 conservancies covering 6.35 million hectares of land, equal to 11% of the country's land mass [89]. According to the conservancies' website, the aim is "transforming communities while safeguarding our iconic wildlife" [89]. Hence, supporting local communities and protecting wildlife in demarcated areas of land are at the heart of the conservancies. These can be managed by individuals as private conservancies or collectively as community conservancies [90]. The largest concentration of conservancies can be found in northern Kenya in the counties of Samburu, Marsabit and Isiolo (Fig 1). Melako, Namunyak and Sera are some of the largest conservancies [91]. The key conservancy actor is the Northern Rangelands Trust (NRT), which is the largest non-governmental organisation (NGO) governing conservancies in Kenya and which receives support from international donors [90,92]. In northern Kenya, the NRT and the conservancy model play an ambivalent role. On the one hand, conservancies offer income opportunities through tourism as well as improving health and education infrastructure [90,93]. On the other hand, conservancies pose a risk to some pastoralists as the conservancies limit herd movement and change the land-use and grazing system [92,94]. While most conservancies are not fenced, core areas are defined where grazing is not permitted [25]. These areas, as well as the conservancies in general, are protected by armed rangers. The establishment of their own security measures is the second key risk posed by the NRT to some pastoral groups as the security forces influence the conflict and raiding dynamics in and around Samburu [94]. Schetter et al. [25] even assert that the "NRT became the engine for a dynamic re-organization of force and the exercise of direct violence". Taking over key government functions, including the provision of security, education and health services, the authors argue that the NRT has established "a state within the state" [25].

Wind park in Marsabit.
The Lake Turkana Wind Power (LTWP) park is the largest wind power project on the African continent [95]. It has a total capacity of 310 Megawatt (MW) and consists of 365 wind turbines which were installed by June 2017 [96]. Because of delays in the power line, it took until September 2018 for the wind park to be connected to the national electricity grid [28]. The village closest to the wind park is Sarima, with about 1,180 people [cited in 97]. Loiyangalani is the closest town with a population of about 35,713 inhabitants, located about 40 km away from the wind park [98]. The people living in the proximity of the wind park are mostly Turkana. Other groups in the region include Samburu and Rendille. The area has a history of violent conflict fought between the three groups over livestock, water and land [e.g. 23]. The county of Marsabit, where the wind park is located, is among the most marginalized in Kenya in terms of income per capita, education and health, and road infrastructure [ Table 1, 99,100].
The project was planned and implemented top-down with few opportunities for the local population to influence it [21]. Hashimshony Yaffe and Segal-Klein [41] point out that the wind park shows how "the state's goals [were prioritized] over the heritage of the local residents" and that the project was implemented "without an inclusive, sustainable vision for the specific region". Among the risks posed by the project to the population of Sarima are relocation, less access to land, a potentially aggravating effect on existing violent conflicts and negative effects on health. In 2015, Sarima was relocated about 2 km to make room for a feeder road to the wind park. Some community members received compensation of about $123 to $151 and the project implementers built a fence around the village in its new location [21,97]. This measure was appreciated by members of the Sarima community and the relocation did not create major problems, but concerns were raised that the village would be moved again [21]. The wind park covers an area of about 162 km 2 but is not fenced off, so the pastoralists can still move with their herds between the turbines. Nevertheless, some access to land is lost as the headquarters and some other facilities are fenced. The risk of an aggravating effect of the wind park on existing conflicts relates to the issue of employment. Especially during its construction phase, the project increased tensions between Turkana and Samburu over who would get more employment opportunities for unskilled labour. In addition, even once the park was built and most of the unskilled labourers had lost their jobs, opportunities to work at the wind park-for instance, as security guards-were disputed between Turkana and Samburu [21]. One can note regarding the health risks of the wind park that some turbines are located only a few hundred meters from the village. While no studies exist on the health effects of the shadows cast and noise emission from the rotors, research from other countries indicates that wind parks can negatively affect the psychological health of the surrounding population [101][102][103]. In addition to the risks, the wind park also brought some benefits to the community of Sarima. Apart from (mostly temporary) employment opportunities, the presence of the wind park's security personnel provides some additional security for the Sarima against attacks by the Samburu. Furthermore, the accessibility of Sarima was improved by a new road. A borehole was drilled for the community of Sarima, but the desalination unit broke and had not been repaired at the time Schilling et al. [21] conducted their research. While the wind park produces significant amounts of electricity, the village of Sarima is not connected to the national grid and is hence without access to electricity.

Oil exploitation in Turkana.
In 2012, significant oil reserves were discovered in Turkana, with the main basin alone containing more than 750 million barrels of oil [29]. Like the county of Marsabit, Turkana is highly underdeveloped in terms of education, infrastructure, health facilities and income opportunities (Table 1). Its level of violence is greater than Marsabit's because of frequent raids, especially between the Turkana and Pokot (see section 2.3). The major risk posed by the extraction of oil for pastoralism-the main livelihood in Turkana-is competition for scarce water resources and the potential pollution of water and land. To extract one barrel of oil, three to four times the amount of water is needed [104]. The major oil company, Tullow, expects to produce 100,000 barrels of oil per day [105]. Oil production has not yet started but, once it does, the risks will unfold as the oil company and local communities will compete for the same resources, although with very different capabilities. The oil company is able to drill very deep to access groundwater while the communities do not have this option. Thus far, the oil company has given some local communities access to water by filling water tanks, creating dependence on the company by the community. Once the production of oil starts and the planned oil pipeline is completed, oil spills and chemicals used for the extraction of oil pose a significant risk to water and pasture resources that are needed by pastoral communities for their livelihood [21,106]. For instance, studies from Nigeria show how detrimental oils spills are to soil, groundwater and vegetation [107]. In Turkana, there is also a higher risk of an escalation of conflict between local communities and the oil company, as community members have repeatedly blocked access roads or attacked oil sites to emphasize their demands for employment opportunities and water. Tullow hires Kenya Police Reservists (KPRs), who are community members who are permitted to carry firearms [108]. While the presence of the KPRs increases the security of the oil sites, the KPRs leave other areas more vulnerable to attacks [21,109]. When comparing the wind park with the oil extraction, one can note that the community of Sarima does not really benefit from the park while the communities in Turkana are likely to be on the losing end, at least when oil production commences [31,90]. The expected start of oil production has been moved several times and Tullow has not given a new date [105]. In early 2023, Tullow announced that the company will invest 1.24 billion KSh (10 million US$) to scale up the development of oil projects in Kenya [110].

Discussion of risks and vulnerability
Pastoralist communities in northern Kenya face multiple interacting risks (Fig 2). Compared to other groups in Kenya, pastoralists are particularly vulnerable to these risks. Vulnerability is commonly understood as a function of (1) exposure to a certain risk or hazard, (2) the sensitivity in terms of the affected resource that is critical for the livelihood and (3) the adaptive capacity in terms of income, health, and education level [111]. The previous sections have shown that all pastoral communities in northern Kenya are exposed to old risks (marginalization, climate change and conflict) while the new risks (wildlife conservancies and energy projects) unfold in specific locations. The communities are sensitive to both old and new risks because pastoralism depends on pasture, water and mobility which are all impacted by the risks. The high degree of political and economic marginalization, in combination with the violent conflicts undermines the adaptive capacity of the pastoral communities.
Among the old risks, the violent raids can be identified as the greatest risk. Among the new risks, one can note that wildlife conservancies and the exploitation of oil and wind resources are very different but they also share several commonalities. Firstly, they are all planned and implemented top-down with few to no opportunities for pastoral communities to influence key decisions. Secondly, they limit the access of pastoralists to water and land, as well as their general mobility. Thirdly, the security situation changes with the introduction of new security providers, be it armed rangers in Samburu, KPRs in Turkana or security personnel in Marsabit. Whilst the concentration of security forces in and around the conservancies, oil sites and the wind park generally make these places less attractive for potential raids, it may make other sites more vulnerable to attacks, especially where communal security forces such as KPRs have been absorbed by powerful external actors such as NRT, Tullow or the wind park operators. Some differences between the three types of projects can also be noted. Conservancies offer the highest potential for pastoral communities to benefit as they are (at least officially) a primary target group. However, existing studies suggest that the conservancies create winners and losers among pastoral groups while influencing the conflict dynamics between them. Oil exploitation produces the most significant externalities as the chemicals used to extract the oil and the oil itself may pollute large areas of land and quantities of water that are the basis for pastoral livelihoods.

Conclusions
Pastoral communities in northern Kenya are not only stuck "between a rock and hard place," as a report by UNDP [112] suggests. They face new and old risks from multiple sides. Their combination causes tremendous suffering and undermines the livelihood of pastoralism in general. This should not be misunderstood as a portrayal of pastoral communities as passive and helpless victims. Pastoralism is accepted as a highly adaptable and suitable livelihood for the harsh and variable climate conditions of northern Kenya, and the different violent and non-violent types of resistance used by pastoral communities show their agency. Nevertheless, if the Kenyan government is genuinely interested in the well-being of pastoral communities in the country's north, it needs to acknowledge and support this type of livelihood. This implies addressing the greatest risks-violent conflicts-first. Promising approaches are measures to improve community relations between fighting communities (such as Pokot and Turkana) and offering young men attractive alternatives to becoming raiders. This can only be done with massive investment in educational infrastructure and the creation of formal employment opportunities. These in turn cannot be provided by the oil or wind energy industries as they are not labour-intensive. Furthermore, the Kenyan state cannot allow external actors to provide security to its population. That is solely the government's responsibility, which it needs to fulfil.
In addition, strict regulations for environmental protection need to be developed and implemented to protect the communities against oil spills. Lastly, regardless of the type of project that is implemented in a pastoral area, pastoral communities need to be the primary beneficiaries. The largest share of the profits of any project needs to be used to compensate local communities, for instance for lost access to land and improving their health and education infrastructure. With respect to the exploitation of oil and wind, the Natural Resources Benefit Sharing Bill is a promising first step. The bill proposes that 40% of the revenues that the county receives from resource extraction should be used for communal projects [113]. However, the bill has not been passed yet and its application will be a challenging task. Given the spectrum and magnitude of risks, it is important to place pastoral communities at the centre of any project. This would not only reduce the likelihood of local resistance against projects but it would also offer the chance for risks to become opportunities.