Conservation from the inside-out: Winning space and a place for wildlife in working landscapes

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. People and Nature published by John Wiley & Sons Ltd on behalf of British Ecological Society The art of land doctoring is being practiced with vigor, but the science of land health is yet to be born. Aldo Leopold: A Sand County Almanac.


| THE E VOLUTI ON OF CON S ERVATI ON
The scope of conservation has expanded continuously in the modern era, from natural resource management in the 18th century, to regulated hunting and protected area set-asides in the 19th century, and the conservation of species, biotic communities, ecosystems and biodiversity in the 20th century (Watson, Dudley, Segan, & Hockings, 2014). The expanded scope reflects changing human sensibilities, views and uses of nature (Nash, 1989;Thomas, 1983) and, more recently, the aspirations of modern states and the international community (Adams, 2005;Sitarz, 1993;United Nations, 1992;Watson et al., 2014;Western & Pearl, 1989;Western, Waithaka, & Kamanga, 2015;WWF, 1980). The philosophical and ethical foundations of conservation have in turn deepened from a largely utilitarian creed to encompass recreational, romantic, spiritual, educational, scientific, intrinsic and other values (Callicott, 1990;Meine, Soulé, & Noss, 2006;Vucetich, Bruskotter, & Nelson, 2015).
Changes in societal perceptions are echoed by the global perspective of the World Parks Congress (WPC) over the last half-century.
The vision for national parks at the 1972 WPC projected the traditional view of areas set aside to protect natural wonders and wildlife for human recreation and enjoyment. The vision expanded to include parks for sustainable development at the 1982 WPC, and human well-being at the 2014 congress (McNeely, 1993;WPC, 2014). In the process six categories of internationally accredited protected areas have been recognized, ranging from Category I: strict nature reserves for scientific and wilderness protection, to Category V: land and seascapes where human uses have produced areas of distinctive aesthetic, ecological and cultural value, and Category VI: areas managed mainly for the sustainable use of natural ecosystems (Dudley & Stolton, 2008).
Although the WPC has raised the target for terrestrial coverage from 10% in the 1980s to 20% for 2020 (WPC, 2014), the coverage falls far short of the area needed to sustain biodiversity for several reasons. First, historically, most protected areas were set aside for scenic, recreational and aesthetic reasons and for specific wildlife attractions, rather than for biodiversity or conserving ecosystem viability (Bennett et al., 2009;Fynn & Bonyongo, 2011;Jenkins, Pimm, & Joppa, 2013;Western & Gichohi, 1993). Second, most protected areas were set aside in lands marginal for development, not to conserve biodiversity (Joppa & Pfaff, 2009;Pressey & Bottrill, 2008;Venter et al., 2017). Third, most protected areas are too small to avoid a loss of species due to insularization, habitat fragmentation and ecological disruption (Newmark, 2008). Fourth, over one-third of land user protected areas is degrading due to human pressures (Jones et al., 2018). Finally, the prospects of retrofitting protected areas to maximize biodiversity coverage are limited, given existing land tenure and uses (Pressey, Visconti, & Ferraro, 2015). Calls for allocating half the earth to protected areas, primarily free of human activity (Dinerstein et al., 2017;Wilson, 2016), face insurmountable political obstacles and fail to redress the systemic shortcomings of existing protected areas (Büscher & Brockington, 2017). The alternative is to consider finding space in lands used for human uses compatible with conserving biodiversity and maintaining the ecological health of the land.

| CON S ERVATI ON IN THE H UMAN RE ALM
The prospects of conserving biodiversity within the human realm by sustaining the health of the land and its capacity for renewal has long been recognized as an evolutionary possibility and ecological necessity (Leopold, 1949). Leopold, in calling for a land ethic, noted that the art of land doctoring is being practiced with vigour, but the science of land health is yet to be born. The land ethic is gaining recognition as being essential for halting global losses of biodiversity (Kremen & Merenlender, 2018). The principles of the World Conservation Strategy (WWF, 1980), Agenda 21 (Sitarz, 1993), the Convention on Biological Diversity (United Nations, 1992) and the Millennium Ecosystem Assessment (Assessment, 2003) all underscore the importance of land health for sustainable development.
Yet, land health also depends on cultural values and governance institutions (Brockington et al., 2019;Ostrom, 2009). Many traditional societies, referred to as biocultures by Nicolay Vavilov (Nabhan, 2012), evolved husbandry practices and cultures that have sustained the health of the land for generations in the face of environmental perturbations and climate change.
Conserving biodiversity in landscapes transformed by farming, ranching and other human uses depends on wildlife having access to their land and in their lives on terms favourable to them. The options for doing so include land sharing, land saving and land sparing (Fischer et al., 2014). The mixed-use and implicit coexistence in land sharing move biodiversity conservation beyond the sharp distinction between natural and human-modified landscapes implicit in protected areas to a land sparing-sharing continuum based 8. The human-centered approach reinforces land health and spatial connectivity and encourages multi-level and distributed governance arrangements embedded in large regional and national jurisdictions.

K E Y W O R D S
coexistence, community-based conservation, conservation governance, landscape conservation, natural resource management, pastoralism, wildlife, working landscapes on the intensity of uses (Phalan, 2018;Western & Pearl, 1989). The scale ranges from heavily transformed and intensively used lands such as monoculture farmlands with little biodiversity to extensive uses in less modified landscapes such as grass-fed ranching, agroecological farming, renewable forestry practices, recreational uses of wildlands rich in biodiversity and extensive mobile pastoralism (Kremen & Merenlender, 2018). Land sparing-sharing models can contribute significantly to biodiversity conservation, though not necessarily from a livelihood and social perspective (Phalan, 2018).
The expansion of nature conservation from the protection of habitats and species to sustaining biodiversity launched by the World Conservation Strategy in 1980 calls for large-scale planning across a broad range of land uses, users, jurisdictions and agencies (WWF, 1980). Biodiversity conservation in the human realm, especially at the scale needed to conserve species in the face of projected climate and land use changes in the 21st century (Newbold, 2018), must also fit in with other land uses and minimize or offset the socio-economic losses incurred (Donaldson, Wilson, & Maclean, 2017).
If scaling up biodiversity conservation has great potential, it also faces severe obstacles (Curtin, 2015;Western & Pearl, 1989).
Community-based conservation, largely focused on ecosystem-level conservation, has shown some success in conserving wildlife and biodiversity in rural landscapes and in improving socio-economic development (Naidoo et al., 2016;Oldekop, Holmes, Harris, & Evans, 2016;Shahabuddin & Rao, 2010). Scaling up biodiversity conservation to a large landscape level calls for yet wider networks and jurisdictions cutting across varied land uses, user interests and jurisdictional boundaries (Arts et al., 2017;Sayer et al., 2013;Scarlett & McKinney, 2016).
Several such landscape approaches have emerged over the last few decades, ranging from the Yellowstone to Yukon Conservation Initiative (Merrill, 2005), Colorado River Initiative (Adler, 2007), Malpai Borderlands Group (Curtin, 2002), Kavango Zambezi Transfrontier Conservation Area (Cumming, 2008) and the Kenya-Tanzania Borderlands Conservation Initiative (BCI). 1 In many cases, the scale needed to secure sufficient space for biodiversity to accommodate large-scale migrations of species traverses several national boundaries. Many of these initiatives have struggled to reach their landscape-scale goals due to a limitation in economic resources for biodiversity conservation.
The tools for encouraging landscape biodiversity conservation are primarily economic in nature and include (a) conservation leases and easements, (b) payment for ecosystem services, (c) cost offsets, and (d) consumptive and non-consumptive uses of wildlife (Bedelian & Ogutu, 2017;Naidoo et al., 2016;Nelson et al., 2010).
New economic accounting methods for valuing natural capital and ecological services, such the Total Economic Value (Costanza et al., 1997) and the System of Environmental-Economic Accounts (Jasch, 2003), emphasize the multiple benefits of conserving biodiversity. These tools complement the efforts of parks to expand the reach of nature conservation and alleviate human threats to biodiversity.
Large herbivores and carnivores have proven especially hard to conserve outside protected areas given a history of overhunting, displacement, the large spaces and intact habitats they need to survive and the dangers and competition they pose to people (Tilman et al., 2017). Smith, Smith, Lyons, and Payne (2018) show that a strong down-sizing of species has characterized the human impact on ecosystem structure and function over at least the past 125,000 years and is likely to continue to do so far into the future.
Despite these obstacles, there is considerable scope for conserving large herbivores and carnivores in the pastoral areas of Africa and rangelands around the world, far more than in protected areas (Western & Pearl, 1989). The East African savannas reflect the long co-evolution of humans and wildlife that survived both the Pleistocene megafaunal extinctions and colonial-era decimation of wildlife in other places (Adams & McShane, 1992). The reasons are due in part to the cultural and use values of wildlife (Western, 2019), the seasonal mobility of pastoralists and the milk-based pastoralist economy. If classified by the IUCN criteria, the pastoral-dominated savannas would be recognized as the equivalent of Category V landscapes where human uses have produced areas of distinctive aesthetic, ecological and cultural values, and Category VI areas managed mainly for the sustainable use of natural ecosystems.
In this article, we focus on wildlife and biodiversity conservation in the East African savannas but also draw broader inferences for the rangelands which cover 25% of the earth's surface and working landscapes more generally. Successfully conserving wildlife in the savannas depends on how herding people manage their lands (Groom & Western, 2013) and finding places for wildlife which fit within the livelihood needs and cultural values of communities (de Pinho & Ellis, 2009). We argue that the space and mobility for sustaining large mammals can be secured indirectly through an approach we term conservation from the 'inside-out'.
By inside-out we mean drawing on husbandry and conservation practices used to maintain the productivity and resilience of pastoralism or other land uses that directly or indirectly maintain large free-ranging wildlife movements in the process. Though overlapping, inside-out conservation is distinguished from bottom-up approaches to wildlife conservation in using primary livelihood considerations to win space for wildlife indirectly rather than through direct incentive-based approaches.
Ultimately, the inside-out approach aligns biodiversity conservation to land health and sustainable husbandry practices, thus widening the scope for nature conservation to the majority of the earth's surface without necessarily relying on large capital investments or tackling biodiversity as the primary focus. Here, we use a case study that looks at the adaptation of traditional pastoral practices in East Africa to the fast-changing global age, and by doing so offers prospects for sustaining land health, improving livelihoods and conserving biodiversity.

| PA S TOR AL AND WILDLIFE L ANDSC APE S
Maasai pastoralists occupy much of the 100,000 km 2 area that spans the Kenya-Tanzania borderlands and the Rift Valley ( Figure 1). In this region, bimodal rainfall varies from 250 to 1,700 mm across altitudes ranging from 600 m on the floor of the Rift Valley to over 5,000 m on Kilimanjaro. The range of biomes cutting across the rainfall and altitudinal gradients make the Kenya-Tanzania borderlands one of the richest biodiversity regions in Africa and mammalian diversity centres on Earth (Jenkins et al., 2013). The region also has some 16 national parks and reserves stretching from Serengeti and Maasai Mara in the West to Tsavo and Mkomazi in the East (Figure 1). The continuous patchwork of grasslands, bushlands and woodlands reaching across this region also support the greatest abundance of pastoral livestock and wildlife in Eastern Africa .
The Maasai people are made up of 13 politically semi-autonomous sections, each sharing the same clan and age-set groupings. The overlapping clans and age-sets move seasonally between wet and dry season grazing areas, largely in synchrony with wildlife migrations (Western & Nightingale, 2005) by using reciprocal social ties that span the seasonal ranges and foster connections to adjoining Maasai sections (Spear & Waller, 1993).
The seasonally coordinated movements between wet and dry season ranges define the community of users and sustain their livestock herds except in extreme years when herders often move to distant refuges to evade droughts. Drought refuges are typically in the wetter regions of the landscape, such as highland pastures, wetlands and areas receiving recent rains (Fynn, Murray-Hudson, Dhliwayo, & Scholte, 2015). The drought movements expand the scale of use from the annual range of a few thousand square kilometres to tens of thousands, negotiated though their reciprocal arrangements (Western & Finch, 1986).
The conservation of the grazing commons to ensure herd productivity and resilience is deeply rooted in Maasai governance and herding practices (Spear & Waller, 1993). A family's survival and well-being are bound to the welfare of its livestock and the availability of pasture and water. The Maasai nevertheless have no word for conservation. The link between rainfall, pasture production, herd productivity, family welfare and the maintenance of commons resources is, instead, incorporated in the concept of erematare. Akin F I G U R E 1 Map of the conservation and protected areas network in the Kenya-Tanzania borderland, highlighting the location of our case study within the region. Our case study looks at three scales: Shompole and Olkiramatian conservancies in the centre of the map, the South Rift at the landscape level and the regional level across the whole map. The red box indicates the extent of Figure 3. Data are from the World Database of Protected Areas (www.prote ctedp lanet.org), South Rift Association of Landowners (www.soralo.org), Maasai Mara Wildlife Conservancies Association (www.marac onser vanci es.org) and BigLife Foundation (www.bigli fe.org) to Vavilov's biocultures and Leopold's land ethic, eramatare is an ethos embedded in husbandry practices, cultural customs and the governance of Maasai society. Eramatare linkages stretch across the landscape through social networks, giving families access to the resources needed to sustain them through the seasons and in times of drought. Eramatare also extends to wildlife, which holds many values and uses among the Maasai, including food, clothing, medicine, ornamentation, utensils, clan symbols and aesthetic appeal (Western, 2019). Many species, including eland and buffalo, are regarded as 'second cattle' and are used as a standby food source in times of drought (Western, 1997).

| H UMAN -DRIVEN CHANG E S IN E A S T A FRI C A
Starting in the 1940s, concerns over the impact of population growth, land transformation and poaching of wildlife led to the creation of national parks and reserves presently covering 15% of Tanzania and 8% of Kenya. By the 1970s recognition that national parks are too small to avoid extinctions and insufficient to protect the wildlife migratory routes led to policies for engaging communities in wildlife conservation through tourism and hunting revenues (Western & Pearl, 1989). In the ensuing decades the policy shift Despite the spread of community-based conservation in Africa's rangelands, its success hinges on the future of pastoral economies and cultures. Pastoralism, the dominant form of land use in the sub-arable savannas, accounts for over 90% of the large mammal biomass in Kenya . Despite the economic potential of tourism and sport hunting, wildlife revenues contribute to only a small portion of household income (Kristjanson & Trench, 2009).
The strong cultural identity and social bonds rooted in livestock also make it likely that pastoralism will remain the primary form of livelihood even where wildlife revenues are a significant portion of family incomes.
Pastoralism nevertheless faces many of the same threats as wildlife, namely the loss of space and mobility; land use changes and land degradation; a loss of livestock production; and decreasing resilience to droughts (Boone, 2005;Hobbs et al., 2008). The loss of ecological adaptability to environmental perturbations is further compounded by a breakdown in the traditional governance institutions rooted in social reciprocity which regulate pasture use and minimize local risks of drought, disease and other hazards (Mwangi & Ostrom, 2009). The breakdown stems from several factors arising from pre-and post-colonial government policies, including government-mandated regulations of the rangelands, the creation of wildlife parks and reserves, forced eviction of pastoral communities for agricultural development, sedentarization and subdivision policies for rangelands and the replacing and breakdown of traditional decision-making practices (Mwangi & Ostrom, 2009). Because of these common threats, redressing the threats to pastoralism and pastoral lands indirectly alleviates the same threats to wildlife. Further, conserving the cultural and governance practices underpinning the mobility and resilience of pastoral herds also reduces human-wildlife conflict and the prospects for continued coexistence (Western, 2019).
Here we look at an example of how two adjacent Maasai communities in Kenya's South Rift (Figure 1) are drawing on both traditional and contemporary knowledge of husbandry and governance practices to sustain pastoralism and conserve wildlife. We then look at how the expansion of such ecosystem-level efforts can benefit from increased scales of governance and management to not only build livestock production and resilience but scale up to large landscapes to conserve biodiversity.

| The benefits of landscape scale to pastoralists and wildlife
Among pastoral communities free-ranging movements give livestock access to resources over large regions where forage and water sources are patchy and ephemeral. Mobility allows herders to track the richest pastures through the season and minimizes exposure to drought, disease, local pasture degradation and perturbations in semi-arid areas where rainfall is highly variable in time and space (Ash, Gross, & Smith, 2004;Boone, 2005;Hobbs et al., 2008;Wang et al., 2006;Western, 1982;Western & Finch, 1986). These ecological benefits of large-scale mobility in areas where resource availability is highly stochastic are reflected in the energy bonus of improved digestive efficiency, growth rates and milk yields for pastoral livestock and wildlife alike (Illius & O'Connor, 2000;Owen-Smith, 2004;Wang et al., 2006). In the case of pastoralists, the scale of use is tied to the scale of social networks allowing free and safe movement (Western, 2019). Large social networks have in turn been shown to increase the benefit to individuals through a 'return to scale' across a range of societies from hunter-gatherers to agrarian and urban economies (Hamilton, Milne, Walker, & Brown, 2007;West, 2017).
The ecological benefits of mobility in offsetting stochastic resource availability apply widely regardless of habitat and landscape variability. The benefits become far more beneficial, however, in landscapes varying in elevation, climate, soils, hydrology, habitat and plant nutrition. In such varied landscapes, the functional heterogeneity at scale adds the productivity, diversity and resilience of large herbivores (Fynn, Augustine, Peel, & Garine-Wichatitsky, 2016;Owen-Smith, 2002, 2004. In the case of the African savannas, wildlife survival, abundance and resilience to seasonal flux and drought are also scale-dependent, and like livestock, they depend on the ecological benefits accruing across large functionally heterogeneous landscapes (Fryxell et al., 2005;Owen-Smith, 2004;Western & Gichohi, 1993). Species such as elephants, lions, wild dogs, giraffe and migratory wildebeest, zebra and gazelle in the Kenya-Tanzania borderlands cover thousands of square kilometres in the course of seasonal movements (Dolrenry, Stenglein, Hazzah, Lutz, & Frank, 2014;Fryxell et al., 2005;Mose, Nguyen-Huu, Western, Auger, & Nyandwi, 2013;Osipova et al., 2018). In that species richness, habitat diversity and ecosystem integrity all increase with landscape heterogeneity (Peterson, Allen, & Holling, 1998; Figure 3), conserving metapopulations of landscape species-species using a large geographic area which includes a wide variety of other species-conserves biological diversity and integrity in the process. Large open landscapes also avoid the compression effects of large mammals on biodiversity (Western & Maitumo, 2004) and rangeland health (Hobbs et al., 2008;Western, Mose, Worden, & Maitumo, 2015) through seasonal movements and population dispersal .
Scale and mobility in increasing functional heterogeneity also facilitate coexistence between livestock and wildlife by increasing foraging options, minimizing competition (Fynn et al., 2016;Tyrrell, Russell, & Western, 2017) and expanding the scope for the spatial and temporal separation of wildlife and livestock. Temporal separation in turn reduces pathogen transmission, crop-raiding and livestock depredation (Valls-Fox et al., 2018;Western, 2019).
Finally, scale and mobility increase the structural and functional heterogeneity of landscapes through disturbance effects created by the differential grazing and browsing impacts (Fuhlendorf et al., 2016;Fynn et al., 2016) important to conserving the richness and integrity of ecosystems. Examples of patch dynamics caused by the seasonal movements of livestock and wildlife to optimize foraging include the interaction of elephants and livestock in creating a shifting mosaic of habitats (Asner, Vaughn, Smit, & Levick, 2016), the intensity, timing and scale of fires influencing biodiversity and forage quality (Fynn et al., 2016;Morrison et al., 2018) and abandoned settlements in creating nutrient hotspots and habitat succession (Muchiru, Western, & Reid, 2009;Vuorio, Muchiru, Reid, & Ogutu, 2014). The expanded scope for coexistence linked to scale also favours ecotourism, hunting and recreation activities that diversify pastoral economies and sustains ecological services such as carbon sequestration, nutrient cycling, and water supplies at local, national and global scales (Yahdjian, Sala, & Havstad, 2015).
Expanding the scale of landscape management from an inside-out approach is driven by the vested interests of pastoralists and agro-pastoralists in sustaining their livelihood productivity and resilience and, incidentally, avoids the negative impacts of fragmentation of rangelands (Groom & Western, 2013;Hobbs et al., 2008).

Conservation approached from the self-interests of herders and
their social networks shift the focus from top-down outside-driven programs to culturally embedded and community-based approaches to conservation (Figure 2).

F I G U R E 2 A diagrammatic
representation of conservation from the inside-out for rangeland and pastoral systems. Successful resource governance of the pastoral rangelands depends on enabling conditions (Arts et al., 2017) linked to the ecology of scale and social networks (Cumming, Olsson, Chapin, & Holling, 2013). Scale benefits pastoralists through increased mobility, herd production, resilience and the maintenance of land health and, in the process, conserves biodiversity and ecosystem services. The benefits and implications of the ecology of scale can act across multiple scales (ecosystem, landscape, region) Social-ecological system theory (SES) and landscape governance theory recognize that successful management of common property resources depends on managing the commons through social networks and ruled-based institutions with coherence at the appropriate social and ecological scales, the capacity for adaptive management through devolved management rights and marketable landscapes (Figure 2; Cumming et al., 2013;Ostrom, 2007;Reid, Fernández-Giménez, & Galvin, 2014).
Our case study demonstrates that fundamental principles of SES theory and landscape governance theory are integral to but fading in traditional Maasai communities. The traditional grazing practices and land use practices, social networks and governance arrangements are changing but remain central in sustaining natural resource management from an ecosystem to landscape and regional level.

| Ecosystem scale
The pastoral lands of Kajiado and Narok counties in Kenya cover Among other functions, the group ranch committees regulate the movement of livestock and seasonal pastures (Russell et al., 2018;Tyrrell et al., 2017;Western, 2019) and arbitrate herding arrangements within and between clans and with adjoining group ranches and Maasai sections (ibid.). The traditional late-season pastures double up as wildlife conservancies attracting tourism enterprises, including lodges, cultural visits and nature walks (J. Kamanga, pers. comm.). The group ranch committees also oversee livestock markets and guide development activities through land planning and zoning to ensure complementarity in land and resource use.
The regulation of livestock movements among community members ensures optimum use of pastures, the conservation of late-season grass reserves and minimum conflict with wildlife over pastures and water (Tyrrell et al., 2017). The governance practices afford community members access to grazing grounds, a variety of habitats for sustaining herd productivity and the mobility required to capitalize on patchy rains, avoid both disease outbreaks and competition with wildlife (Russell et al., 2018). The ability of herders to live alongside wildlife and benefit through tourism enterprises and traditional values (Roque de Pinho, 2009) depends on their rich knowledge of wild herbivores and carnivores and their skills in averting conflict (Western, 2019).
The role of traditional deployment of community scouts termed ele'enore is to gather the information the community needs to weigh its grazing options and reach collective decisions on herd deployment, watering regimes and the avoidance of serious conflicts with wildlife (Western, 2019). Olkiramatian, drawing on the ele'enore tradition, set up a Lale'enok Resource Centre which trains and deploys resources assessors to gather a wide variety of ecological, social and market data relevant to collective herd management, land planning, resource management and market access. The resource assessors use automated data collection platforms to enter and analyse information for rapid dissemination and decision-making Tyrrell et al., 2017).
The mix of traditional and contemporary knowledge, herding practices and governance arrangements maintains the seasonal scale of livestock movements on Shompole and Olkiramatian, ensuring that wildlife and livestock can benefit from seasonal migrations and adaptability to droughts Tyrrell et al., 2017). In contrast to the sharp wildlife declines in pastoral areas which have been subdivided (Groom & Western, 2013;Ogutu et al., 2016;Said et al.,), wildlife populations on Shompole and Olkiramatian have been more resilient.

| Landscape-scale
The success of the Olkiramatian and Shompole conservation efforts encouraged the African Conservation Centre (ACC) to create a landowner network connecting Amboseli National Park and Maasai Mara National Reserve across the Rift Valley, aimed at forging a new tourist link and conserving the biodiversity-rich landscape (Figures 1 and 3). The governance of SORALO's membership builds on the existing social networks and reciprocal grazing arrangements connecting adjacent Maasai clans and sections. SORALO is registered as a land trust and like Shompole and Olkiramatian, incorporates traditional and contemporary governance institutions and practices. The SORALO board includes representatives of regional group ranch clusters and draws on strong community inputs and the political influence it has with county and national agencies.
The geographic reach and group ranch network SORALO covers scales up Shompole and Olkiramatian ecosystem-level governance to take advantage of the expanded benefits accruing from a landscape scale, including access to drought refuges, the conservation of watersheds and ecosystem services they depend on beyond their boundaries, the protection of wildlife corridors, joint wildlife scouting operations, ecotourism planning and access to country, national government and NGO services.

| Regional-scale
The Kenya-Tanzania borderlands are globally important as a vertebrate biodiversity hotspot and regionally as a centre of plant diversity accounting for a quarter of all the plants recorded in Kenya and Tanzania (Ministry of Environment Natural Resources and Regional Development Authorities, 2015; Figure 3). The borderlands also support large populations of endangered and threatened species, including one of the largest free-ranging elephant populations, richest carnivore assemblages in Africa, and among the largest remaining large herbivore migrations world-wide (Harris, Thirgood, Hopcraft, Cromsigt, & Berger, 2009;Jenkins et al., 2013). Coupled with sixteen protected areas and a $1.5 billion annual tourism industry, the borderlands have attracted international attention as one of the earth's last great natural wonders.
Despite the attention and large well-equipped government ranger forces, wildlife has declined sharply in national parks both sides of the border in the last few decades (Craigie et al., 2010;Western, Russell, & Cuthil, 2009). The declines stem from the same threats confronting the future of free-ranging pastoralists at an ecosystem and group ranch level; land conversion, subdivision, settlement and growing pressures on the open rangelands. Species dependent on large-scale migrations to access seasonal forage and sustain large populations are especially vulnerable to land compression and fragmentation. Both the Tarangire and Nairobi National Park herbivore populations are threatened by a lack of collaboration and incentives at the regional scale (Newmark, 2008;Western & Gichohi, 1993).
In the case of elephants, the main cause of the decline in the 1970s and 1980s arose from a ten-fold rise in ivory prices in the world market creating a surge in poaching and compression of remaining herds into the relative safety of national parks Stiles, 2004). The compression reduced woody F I G U R E 3 5 × 5 km grid maps of the Kenya-Tanzania borderlands (border -black line) and major lakes (light blue) showing: (a) Habitat heterogeneity based on the coefficient of variation of Enhanced Vegetation Index (EVI) from Tuanmu and Jetz (2015). The accompanying graph shows heterogeneity increasing steeply with area coverage from a central pixel to a peak at 15,000 km 2 . The accompanying graphs to maps (b) and (c)

| Inside-out conservation as a way forward
Conservation in working landscapes connected from small land management entities to large multiscale landscapes holds untapped potential for alleviating the insularization and other human impacts on protected areas as well as conserving biodiversity in rural landscapes. Global warming is adding urgency to conserving species beyond protected areas as their ranges shift in response to climate change (Thomas et al., 2004). The threats call for new approaches to conserving biodiversity in the rural landscape.
Solutions range from creating buffer zones for protected areas, sparing and land sharing, this approach is centered on people rather than wildlife (Phalan, 2018) and allows for a range of land use options along the sharing to sparing continuum. As our case study shows, the productivity and resilience of pastoral communities are linked to the ecological and functional benefits of scale, heterogeneity and land health, which if sustained, make space for large free-ranging wildlife populations and biodiversity. While the inside-out approach is applicable to rangelands, it has relevance to other large complex systems that benefit from increasing the scale of management to larger landscapes, such as marine fisheries (Curtin, 2015).
Finding such space also depends on finding a place for biodiversity within a community's aspirations and economic options. To succeed, biodiversity conservation in rural lands must expand to a landscape scale and cut across national, jurisdictional, institutional and cultural boundaries, and create network connections between them (Scarlett & McKinney, 2016). The feasibility of using an inside-out approach to conserve large open spaces becomes more feasible when linked to existing theories connecting landscapes among multiple jurisdictions using a polycentric and devolved approach to governance (Ostrom, 2007), landscape theory based on sustainable practices (Arts et al., 2017;Curtin, 2015) and a sound ecological understanding using functional heterogeneity theory (Fuhlendorf, Fynn, McGranahan, & Twidwell, 2017;Fynn et al., 2016).
Of particular use is the SES framework (McGinnis & Ostrom, 2014;Ostrom, 2007Ostrom, , 2009, which allows for the analysis of relationships among multiple levels of complex SESs at different spatial and temporal scales. It also helps us understand how specific parts of the systems are related and interact. Applying the SES framework to natural resource management reveals several important elements affect the likelihood of users' self-organizing to sustainably manage a resource, including: (a) the existence of institutions that work at the correct social and ecological scales; (b) governance that is multi-scale and multi-level, (c) communities who have clear devolved autonomy over resource management; (d) communities who see an importance of the resource, and see a benefit (through economic and non-economic values) from their natural resources; (e) strong social norms of collaborative governance and management (Cumming et al., 2013;Ostrom, 2007Ostrom, , 2009Reid et al., 2014).
Our case study shows that these conditions are achievable for the management of pasture (and subsequently wildlife), and when they are conceptually twinned with a capacity for effective landscape governance (see Arts et al., 2017 for landscape governance as a capacity; Figure 1) and locally appropriate ecological principles, the results are mutual benefits to people, livestock and wildlife.
Biodiversity and livelihoods dependent on land health converge in the rangelands more than in most other biomes. The convergence stems from two formerly disparate objectives. First, the ambit of protected areas for conserving biodiversity has widened from the strict protectionism of IUCN Category I parks to include Category VI areas accommodating a range of compatible human values and uses.
Second, land users, including pastoralists, private commercial ranchers, fishing, hunting and forestry communities are finding common ground in conserving healthy, open landscape for their inclusive values. These include not only primary livelihoods but also other uses and values as nature sensibilities expand and international commitments to sustainable development deepen. These include a sense of place, a healthy environment and the range of ecological services it provides (Arts et al., 2017;Sayer et al., 2013).
The inside-out approach to winning space and a place for wildlife and biodiversity is place-based and draws on local knowledge and informal governance arrangements. It avoids the bottom-up wildlife-centric approaches which are driven from the outside and that alienate communities working on the land for a living. Instead, conservation approached from the inside outward promotes a human-centered approach that reinforces land health and spatial connectivity. It also encourages a shift from colonial and post-colonial central government command-and-control policies that have hampered wildlife and natural resource management towards devolved rights and ownership emerging with political and economic liberalization trends around the world.
Expanding conservation benefits from the inside-out is especially pertinent to conserving large herbivores and carnivores which need large open landscapes, pose considerable threats and costs to rural communities and are the most vulnerable of all species to farming, ranching and resource extraction (Tilman et al., 2017). Given that the downsizing of large herbivores and carnivores will have the biggest impact on ecosystem structure and function in coming decades, finding places for such landscape species will avert trophic cascade effects and the ecological impoverishment arising from the loss of keystone species (Dirzo et al., 2014;Smith et al., 2018).
Unlike much of East Africa's privately protected areas, conservancies and Wildlife Management Areas which 'buy' tolerance through tourism, trophy hunting or other land leases, the inside-out approach to conserving large open landscapes and wildlife expands self-interests and shared community livelihoods, social networks and cultural values. In the case of the Maasai, eramatare linkages at the individual levels between family and herds, and at the community level for land access and land health through common pool governance arrangements, these connections promote both traditional intrinsic and instrumental values for wildlife.
The linkages also reduce conflict and increase the prospects for coexistence with wildlife. Nevertheless, governments, NGOs and other conservation agencies have a large supporting role to play in offsetting the opportunity costs of wildlife and mitigating conflict as subsistence economies erode with population growth and market economies.
Collaboration at large scales and across human-dominated landscapes faces enormous challenges and at best will only complement and not supplant the need for protected areas and other conservation tools. Yet, as Leopold recognized, the extension of ethical values to include the health of the land is an evolutionary possibility and ecological necessity. By drawing on lessons from cultural institutions, principles and practices which have maintained the commons for generations, combined with ecological, political and social sciences, we see the possibility of landscape doctoring evolving into the science of land health.

ACK N OWLED G EM ENTS
We wish to thank the Liz Claiborne Art Ortenberg Foundation, Cincinnati Zoo and Jeanne Musgrove for their support. We thank Olkiramatian, Shompole and other South Rift communities for their support and collaboration. We thank Philip Stickler for helping with the graphics. Thank you to Jane Wambui and Gregory Akall for proof-reading the Swahili translation.

CO N FLI C T O F I NTE R E S T
J.K., G.W. and S.R. work for SOLARO and D.W. works for the African Conservation Centre. All authors have engaged in conservation within the region outlined in the case study.

AUTH O R S ' CO NTR I B UTI O N S
D.W., S.R. and P.T. developed the original concept for this manuscript. D.W. led the writing. P.T. and P.B. produced the figures and analysed the data and assisted D.W. with the drafting of the manuscript alongside S.R., G.W. and J.K. All authors revised the manuscript critically for important intellectual content, approved the version to be published and agree to be accountable for the aspects of the work that they conducted.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data in this manuscript are avalaible through their referenced sources.