A social-ecological typology of rangelands based on rainfall variability and farming type

Journal of Arid Environments xxx (2017) 1e9

Contents lists available at ScienceDirect

Journal of Arid Environments journal homepage: www.elsevier.com/locate/jaridenv

A social-ecological typology of rangelands based on rainfall variability and farming type John-Oliver Engler*, David J. Abson, Robert Feller, Jan Hanspach, Henrik von Wehrden Faculty of Sustainability, Leuphana University of Lüneburg, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 April 2017 Received in revised form 10 July 2017 Accepted 13 September 2017 Available online xxx

We present a social-ecological typlogy for the world's rangelands that integrates the much debated nonequilibrium concept from ecology with socio-economic characteristics of rangeland systems. We propose that, as a first approximation, the socio-economic properties can be adequately captured and differentiated by the distinction between the two main types of rangeland farming systems worldwide: subsistence and commercial farming. The resulting typology has four categories, which are ‘commercial equilibrium’, ‘commercial non-equilibrium’, ‘subsistence equilibrium’ and ‘subsistence non-equilibrium’. We provide and discuss examples for each category. Moreover, we point out how this typology might help to understand and address some of the problems related to unsustainable rangeland management. Finally, we provide and discuss a global map of rangelands that illustrates the geographic distribution of all four rangeland types. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Range management Non-equilibrium theory Commercial agriculture Subsistence agriculture Biophysical degradation Social-ecological systems

A substantial part of Earth's habitable land mass is covered by rangelands, i.e. land that is not covered by ice, rocks or water. As of 2009, roughly 52% of global meat production took place in grassland-based systems of South and Central America as well as Africa. The commercial livestock sector was recently estimated to employ at least 1.3 billion people globally (Thornton, 2010), estimates of the number of livelihoods directly depending on rangelands vary from 600 million (Thornton et al., 2006) to 2 billion (Reynolds et al., 2007). Rangelands thus support, directly and indirectly, millions of people. Moreover, rangeland ecosystems provide a multitude of ecosystem services with considerable economic value (Goldstein et al., 2011, Havstad et al., 2007; MEA 2005; Tanaka et al., 2011), are home to numerous species of herbivores and plants (Olff et al., 2002), and are thought to have considerable potential in climate change mitigation (Lal, 2003, 2004; FAO, 2009). Many of these rangeland areas feature arid or semi-arid climates with high inter-annual variability of rainfall (Olbrich, 2012). The main e and often the only economically viable (Quaas et al., 2007) e land use is livestock farming. Even though exact numbers are

hard to come by, it is safe to say that a very substantial part of the world's livestock is supported by such dryland rangelands1 (e.g. FAO, 2009; Thornton, 2010). Unsustainable management of rangelands, particularly overgrazing, is a major problem in many parts of the globe, especially those that classify as dryland rangelands (Safriel et al., 2005; SADC, 2009; von Wehrden et al., 2012). Many rangelands are vulnerable to interdependent biophysical (e.g. soil degradation, invasive species) and socio-economic (e.g. demographic shifts, market price fluctuations) changes, which unchecked, have potentially negative influence on both human livelihoods and the ecology of these systems. While there is an increasing awareness that rangelands are complex social-ecological systems (Vetter, 2005; Fox et al., 2009; Briske et al., 2011), much of the discussions around rangeland classification have evolved from disciplinary debates focusing on a single aspect of rangelands as illustrated by the non-equilibriumeequilibrium debate in ecology (Illius and O'Connor, 1999; Briske et al., 2003; Vetter, 2005). While there are some conceptual frameworks that can be used to classify rangelands based on both the ecological and socioeconomic properties of these systems, they tend to be difficult to

* Corresponding author. Quantitative Methods of Sustainability Science Group, Leuphana University of Lüneburg, Scharnhorststr. 1, D-21335 Lüneburg, Germany. E-mail address: [email protected] (J.-O. Engler).

1 According to the 1994 U.N. Convention to Combat Desertification, ‘dryland rangelands’ are rangelands where annual precipitation is less than two thirds of potential evapotranspiration.

1. Introduction

https://doi.org/10.1016/j.jaridenv.2017.09.009 0140-1963/© 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

2

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9

apply over large spatial extents. For example, the Integrated Social, Economic, and Ecologic Conceptual Sustainable Rangelands Framework (Fox et al., 2009) or Ostrom (2009) General Framework for Analyzing Sustainability of Social-Ecological Systems dissect socio-ecological systems into constituting layers or ‘tiers’ describing states and interactions within the system. Fox et al. (2009) successfully demonstrated how these tiers and their constituents interact with each other in rangelands. However, their approach requires considerable data input, and we assume that this might be one of the reasons why it has not been widely used so far. Ostrom (2009) approach is less quantitative in nature and has been applied to analyze Tibetan pastoral systems (Wang et al., 2014). Yet, neither approach provides a general typology of rangelands, nor have they been used so far to construct one. However, such a rangeland typology has been advocated by some researchers recently for orientation in the general debate about rangeland health and sustainability (e.g. Vetter, 2005; Briske et al., 2011). We think that a general typology of global rangelands might indeed be a useful tool to inform sustainable rangeland use. Different system characteristics lead to different outcomes, and what might be a successful policy in one system may be less effective, or even counterproductive in another. A negative externality can, for example, be tackled by a Pigouvian tax in a commercial farming system, but usually not in a subsistence system, because farmers’ access to and use of markets is fundamentally different in the two systems. Moreover, the tight coupling of ecological and economic factors in rangelands suggests further policy differentiation is required in terms of both social and ecological system properties. Here, we take up this demand for a general typology of rangelands and propose a relatively simple approach towards a socialecological classification that might be used to inform rangeland management in a way that accounts for their ecological and socioeconomic properties. Our proposed rangeland typology is based on two key system properties; the ecologically driven and empirically well-supported non-equilibrium concept for rangelands (Ellis and Swift, 1988; von Wehrden et al., 2012) and a socio-economic categorization based on whether farming is subsistence or commercial (cf. Hardin, 2004). We show how these two system properties can be combined to construct an easily applicable rangeland typology, with four categories, that can inform sustainable rangeland management. We illustrate and discuss our framework with examples of rangeland systems from around the world, and provide a global map of rangelands according to our typology, which can be seen as a complement to the “anthropogenic biomes” approach by Ellis and Ramankutty (2008), with a focus on sustainable rangeland use. 2. A social-ecological classification of rangelands Beginning with key concepts, we set up, explain and discuss our social-ecological classification for rangelands in the following. Subsequently, we provide examples for each resulting category. Finally, we provide a global map using our classification, and discuss our findings. 2.1. Key concepts High inter-annual precipitation variability is the central climatic characteristic of (semi-) arid rangelands world-wide (Olbrich, 2012). Ecologically, there is a correlation between the absolute sum of precipitation and its inter-annual coefficient of variation (Cv) with net primary productivity of above-ground biomass (ANPP). In fact, precipitation variability may be more important for ANPP than absolute sum of precipitation (Knapp et al., 2002). High

inter-annual variability in precipitation leads to highly variable grass production creating, in turn, income uncertainty for farmers, since what a farmer can produce in any given year depends directly on what the land can support (Quaas et al., 2007). This close relationship of ecology and economics has made (semi-) arid rangelands a prime object of study for ecological economics (e.g. Olbrich, 2012; Jacoby et al., 2014). Biophysical degradation commonly refers to the worsening of one or several biological or physical parameters with respect to several consecutive data collections at different points in time. A unified or commonly accepted rule when an author refers to a rangeland as being in a “degraded” state seems to be lacking, so there is a certain degree of subjectivity involved. The non-equilibrium theory of rangelands (Ellis and Swift, 1988) links precipitation variability with ecological dynamics. The key tenet of non-equilibrium theory is that precipitation variability is the main driver of rangeland dynamics in arid and semi-arid rangelands.2 With its emphasis on abiotic rather than biotic factors as primary drivers of vegetation and livestock dynamics, the nonequilibrium concept constitutes a controversial shift of paradigm in ecology (Vetter, 2005). Particularly, the non-equilibrium concept proposes a Cv value of 33% as threshold value, above which nonequilibrium conditions hold (Ellis and Swift, 1988). The theory predicts that equilibrium conditions (Cv ;< ;33%) generally favor biophysical degradation, because of less frequent droughts, which, by causing animal die-offs, allow the system to recover from grazing pressure. A recent global meta-analysis of rangelands studies found widespread use of the non-equilibrium concept and strong support for the validity of the 33% Cv threshold (von Wehrden et al., 2012). We therefore adopt this threshold to clearly distinguish between rangelands with equilibrium dynamics and rangelands with non-equilibrium dynamics. Rangeland systems can be further differentiated by two socioeconomic system types that occur in both equilibrium and nonequilibrium rangelands: commercial and subsistence farming. Commercial livestock farming refers to the rearing of animals on private land for production of, among others, meat, eggs, dairy products and skins for the exclusive purpose of selling them at markets to make a profit (Cambridge Dictionaries Online, 2014). Commercial farmers generally have good access to well-developed markets for selling their produce, but also for the acquisition of farming infrastructure such as irrigation systems, extra fodder, chemical fertilizer, land, and financial capital or insurance (e.g. Ingenillem et al., 2014). Commercial farming mostly takes place in higher income countries, with China, India, Pakistan and Namibia being notable exceptions, and accounts for 53% of the world's agricultural GDP (World Bank, 2009). In high income countries, particularly in the United States, there has been a trend in recent decades towards shifting production capacities from extensive rangeland farming to more intensive production systems (Food and Water Watch, 2010). Nevertheless, globally there remains a continuum of commercial farming practices from extensive to intensive grazing, often determined by the biophysical constraints of the particular rangeland in which the grazing occurs. Subsistence livestock farming or communal livestock farming refers to the rearing of animals such as goats, sheep, cattle, yaks or camels mainly for the personal use of their milk, eggs, meat or other animal products. Typically, subsistence farmers try to sell or exchange some of their produce at local markets to supplement their livelihoods when possible, but most or all of the produce is regularly consumed by the farmer and their family (Waters, 2007).

2 In humid rangelands, selectivity by herbivores is another important driver of plant composition and rangeland dynamics.

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9

3

Fig. 1. Our proposed typology for rangelands (red box in the bottom right), with the horizontal axis classifying rangelands from equilibrium to non-equilibrium conditions (i.e. low Cv to high Cv values) and the vertical axis according to farming system (i.e. subsistence to commercial). We provide climatic/biophysical (top) as well as socio-economic (left) characteristics and outcomes common to the respective sub-systems. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Moreover, due to geographical remoteness, political instability, lack of infrastructure or otherwise high transaction costs, subsistence farmers often do not have access to national or global agricultural markets (Waters, 2007). Today, subsistence livestock farming is wide-spread in large parts of rural Africa, Latin America and Asia. Oftentimes, subsistence farming occurs on commonly-owned or managed land. Especially in central and western Asia, India, east and southwest Africa as well as northern Eurasia, nomadic and semi-nomadic3 herding are still frequent forms of subsistence livestock farming, where people migrate along with their animals from place to place, temporarily staying where they find sufficient fodder and water for their animals. 2.2. The framework We propose a two-axis matrix of rangelands, which combines the ecological dimension with the socio-economic dimension of livestock farming in rangelands. In doing so, we address the need to explicitly incorporate the role of humans in shaping the dynamics of rangelands (Vetter, 2005; Berger et al., 2013; von Wehrden et al., 2015). In our view, humans take two roles: they are part of the ecosystem as well as ecosystem users. In particular, we use the nonequilibrium concept in our framework to distinguish between equilibrium and non-equilibrium rangelands. This means that the ecological axis in the framework ultimately maps precipitation variability. The socio-economic dimension is mapped along an axis featuring commercial and subsistence farming as poles (Fig. 1). The framework's setup creates a new typology of global rangelands, and proposes that unsustainable use of rangelands may occur regardless of management regime, i.e. whether farming is done commercially or for subsistence.

3 In contrast to nomadic herders, semi-nomadic herders have a base camp where they grow crops during periodic settlements.

Hence, following our concept, we expect four major regimes for the world's rangelands according to their respective ecological and socio-economic context, examples of which we discuss in the following section. 2.3. Illustrating the framework with examples According to our classification from Fig. 1, the four major rangeland regimes for livestock farming are (counter-clockwise from top left corner): (1) commercial equilibrium (CE), (2) subsistence equilibrium (SE), (3) subsistence non-equilibrium (SNE), and (4) commercial non-equilibrium (CNE). We discuss examples for each type of rangeland in the following. 2.3.1. Commercial equilibrium (CE) rangeland: Northern China Rangelands in northern Chinese provinces of Xinjiang, Gansu, Neimongol (Inner Mongolia) and Heilongjian are our example of a CE system. While some authors have classified the system as nonequilibrium (cf. Han et al., 2008), recent findings suggest the system rather be classified as equilibrium (von Wehrden et al., 2012: Fig. 3). Generally, annual rainfall decreases from the northeast to the northwest, so that vegetation is either grass steppe or shrub steppe, depending on exact location. The system of commercial farming in China was established in the late 1970s under Deng Xiaoping's program “Socialism with Chinese characteristics” when the agricultural sector was decollectivized and privatized. Price controls were released and the Household Responsibility System (HRS) was put in place, which allowed considerable freedom regarding all aspects of rangeland management at the household level. However, the HRS explicitly encouraged increased livestock numbers (Sheehy et al., 2006). Due to this history, Chinese farms are small compared to commercial livestock farms in other countries (Han et al., 2008), and this fact seems to be responsible for ambiguous classifications of the system in the literature (cf. Hardin, 2004; and Fig. 2 below).

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

4

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9

Fig. 2. Global map of rangeland systems according to our typology. The map is based on Hardin's classification of global agriculture (Hardin, 2004), von Wehrden et al. (2012) global precipitation map and the global map of rangelands by the Society for Range Management (Launchbaugh and Strand, 2015). Countries explicitly discussed in this paper are highlighted in red. We employed the Robinson Projection to produce the map. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Yet, while there may be a substantial part of subsistence farming left, we argue that the system is shifting, and that livestock farming has predominantly been undertaken on a fully commercial scale for at least a decade. As part of this process, the Chinese government increasingly seeks to ‘super-size’ farms to augment productivity (China Economic Review, 2013). Today, China is the world's largest producer of sheep and goats and the fourth largest of cattle (FAOSTAT, 2013). Biophysical degradation widely affects rangelands in China's north (Zhao et al., 2007; Han et al., 2008). According to Han et al. (2008),4 60% of Inner Mongolian, 65% of Xinjiang and 50% of Gansu rangelands are affected by degradation, which can largely be attributed to over-grazing, even though the failed conversion of the arid or semi-arid rangelands into croplands also seems to have contributed to these numbers (Sheehy et al., 2006; Han et al., 2008). There have been noteworthy large-scale governmental efforts to fight biophysical degradation. Restoration measures include, but are not limited to, hand sowing of forage plants, irrigation, relocation of local population and fencing (cf. Sheehy et al., 2006; Han et al., 2008). These measures are costly, their long-term benefits still unclear, and Chinese political peculiarities may limit transferability to other regions. Other examples for a CE system are extensive cattle farms in the U.S. mid-western states such as Montana, South Dakota and Wyoming (Fleischner, 1994), or the Brazilian Cerrado region (Klink and Machado, 2005).

2.3.2. Subsistence equilibrium (SE) rangeland: Northern and Central Mongolia The Northern and Central Mongolia biome is equilibrium mountain steppe, and its vegetation is generally dominated by grasses, forbs and shrubs (Fernandez-Gimenez and Allen-Diaz, 1999). Most of the land area in Mongolia is unsuitable for cropping and animal husbandry has been the major human land use for

4

Sheehy et al. (2006) report similar numbers.

the past centuries, possibly even millennia (Fernandez-Gimenez and Allen-Diaz, 1999; Harris, 2013). Nomadic or semi-nomadic herding is still largely practiced country-wide and three quarters of the total land area is allotted to grazing (Harris, 2013). Despite the fast rise of the mining sector, it is estimated that about one quarter to one third of Mongolia's total population of 3 million are engaged in some aspect of pastoralism (Endicott, 2012; World Bank, 2013). The most important livestock are goats, sheep, horses, camels and cattle. Following the collapse of socialism in the early 1990s, herds have been privatized and numbers of cashmere goats have risen substantially, the wool of which is the only source of cash to farmers. Yet, the overall land use system of family (semi-) nomadic herding with very little or no contact to any sort of market has remained intact. A typical Mongolian pastoralist family sustains itself on several hundred animals, which are moved about four times a year to a different location (Fernandez-Gimenez and AllenDiaz, 1999). In the SE regime of Northern and Central Mongolia, there is relatively stable forage provision due to less variable precipitation, resulting in high grazing pressure, which favors biophysical degradation. Consequently, in agreement with nonequilibrium theory, biophysical degradation has been observed in Northern and Central Mongolian rangelands (von Wehrden et al., 2015). Other SE systems can be found, for example, in the semi-arid regions of southern India (Harrup et al., 2016) and in the majority of Ethiopia (Bachewe, 2009).

2.3.3. Subsistence non-equilibrium (SNE) rangeland: Central South Mongolia The Central South region of Mongolia is covered by the Gobi desert, which features non-equilibrium desert steppe (von Wehrden et al., 2015). Vegetation types present are the same as in the north, but diversity, abundance and net primary productivity decreases with increasing aridity from north to south (FernandezGimenez and Allen-Diaz, 1999; Sasaki et al., 2008). The socioeconomic background is identical with the situation in Northern

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9

Mongolia. In contrast, there is relatively little or no biophysical degradation in the grazed rangelands of the Gobi desert (von Wehrden et al., 2009), except around water points and key resource provision points (Sasaki et al., 2008). In the SNE rangelands of south Mongolia, herd sizes are dramatically reduced in years of drought, because risk-management strategies such as provision of supplementary fodder or large-scale transportation of animals are not commonly available to farmers. For example, in the severe 2000e2002 series of droughts and very cold winters (‘Zuds’), there was a decline of about 50% in livestock numbers in the South Gobi province (Central South Mongolia) bringing livestock numbers down to the levels observed during the then still centrally-planned economy of the late 1980s. Other examples for the SNE type of rangeland are parts of Kenya (Boone and Wang, 2007), Syria (Geerken and Ilaiwi, 2004), Senegal (Miehe et al., 2010) and northern Namibia (Zeidler et al., 2004). 2.3.4. Commercial non-equilibrium (CNE) rangeland: Namibia Namibian commercial cattle farms may be considered a prime example of a CNE system, since Cv values are above 33% throughout the commercial cattle farming region in the Central North of Namibia. Vegetation is characterized by coexistence of grasses and bushes, of which only the grasses are palatable for cattle (Ingenillem et al., 2014). Primary productivity of grasses is highly seasonal and limited by recent precipitation (Olbrich, 2012). The system in its current configuration has existed since Namibian independence from South Africa in 1990. The main commercial cattle farming region in Namibia's central north covers some 31.5 million hectares (Mendelsohn, 2006: 10), features a high inter-annual variation in precipitation throughout, and can thus be clearly characterized as non-equilibrium (cf. Fatichi et al., 2012; von Wehrden et al., 2012). Consequently, single farms are very large, on average well above 7000 ;ha, with an average herd size of only a €rtner, few hundred animals, at times even less (Engler and Baumga 2015). Despite its non-equilibrium classification, a special form of biophysical degradation e bush encroachment where palatable grasses are gradually replaced by bushes unpalatable to cattlee has been found to be a serious issue (Olbrich et al., 2014; Ingenillem et al., 2014). Because of the availability of various risk management strategies, such as provision of supplementary feed when needed, herd sizes remain unaffected by droughts. Hence, from an ecological point of view the system behaves as if it were equilibrium. This may be used to address the question raised by Ingenillem et al. (2014) who found in a sample of 399 Namibian commercial cattle farmers that an increase in rainfall actually comes with a decrease in pasture quality.5 Using our framework, this does not come as a surprise: an increase in absolute precipitation means e all other things being equal e a gradual move towards equilibrium conditions, where biophysical degradation is to be expected according to the non-equilibrium theory. To some extent this shift in system characteristics is supported by the commercial nature of the farming systems that help to maintain stocking densities even in non-equilibrium conditions. Other examples for CNE systems that are externally stabilized include Australian drylands where rangelands are irrigated and chemically fertilized on a large scale (Abson et al. 2017), the U.S.

5 Pasture quality was operationalized as deviation of actual bush cover from the one indicated as ‘optimal’ by the respective farmers in a paper-pencil questionnaire. 6 Here, we have to note that Californian agriculture is geographically typified as ‘Mediterranean agricultural system’, a defining characteristic of which is the coexistence of subsistence and commercial farming within the same system. It is thus a hybrid between the two, but we would still rather categorize Californian cattle and sheep ranching e which takes mostly place in the San Joaquin Valley in central California e as commercial.

5

southwest6 (USDA, 2016) and South Africa (du Toit et al. 2013). 2.4. Mapping global rangelands We provide a global map that shows the world's commercial and subsistence rangelands along with areas featuring nonequilibrium conditions in Fig. 2. To produce the map, we combined Hardin's work on global agricultural systems (Hardin, 2004) with the global map of rangelands provided by the Society for Range Management (Launchbaugh and Strand, 2015), which is based on the potential vegetation approach. That is, we understand rangelands in a broad sense as all land areas not covered by water bodies, forests or rocks and ice. Finally, we added von Wehrden et al. (2012) global map of (non-)equilibrium precipitation characteristics as a third layer. All mapping was performed using ArcGIS 10.2.1 (ESRI, 2014). A unique mix of human land use has developed in Mediterranean climates around the world, often involving, among other characteristics, a mixture of limited commercial and subsistence use (Mørch, 1999). We have refrained here from introducing Mediterranean rangelands as a further category, because we maintain that the use of rangelands potentially in that category is predominantly commercial, even though subsistence land use may be present. The map shows no subsistence rangeland farming in Australia and North America, except for an area in Central Mexico. On both of these continents, there are substantial shares of non-equilibrium rangelands. In South America, Central Asia and Africa, we find all four categories of our typology. In Europe, there are almost no nonequilibrium rangelands, and there is some mixture of commercial and subsistence use around the Mediterranean Sea. As discussed above, we would classify the rangeland use in Northern China as commercial based on reported trends in recent years (see above), and also based on potential future development. However, available sources still classify this region as area with subsistence land use (Hardin, 2004). While we have adopted Hardin's classification to create our map, we note that it is important to acknowledge the highly-dynamic change in land use in this area, and maintain that an exact classification will probably remain difficult due to the continuous transformation of these rangeland systems. Rangelands cover 71.684 million km2 of Earth's surface, which amounts to 54.7% of its total land mass excluding Antarctica (see Table 1). Of these rangelands, 57.842 million km2 are used by humans (80.7% of total rangeland area, 44.1% of total land mass). We find that human rangeland use is predominantly for subsistence (31.557 million km2, 54.6% of rangelands), which is roughly one quarter of Earth's total land mass excluding Antarctica (Table 1, middle panel). 45.5%, or 26.265 million km2, of rangelands are commercially used, which amounts to 20.0% of total land mass. The bottom panel of Table 1 shows that 51.5% of human-used rangelands feature equilibrium dynamics and 29.2% feature nonequilibrium dynamics. Among commercial rangelands, equilibrium conditions are more than twice as frequent as nonequilibrium conditions (18.442 million km2 CE compared to 7.823 million km2 CNE, upper panel of Table 1), whereas the subsistence rangelands are more evenly distributed regarding ecological dynamics (18.490 million km2 SE compared to 13.087 million km2 SNE). While CE and SE rangelands equally share half of the world's rangelands (18.4 million km2 each), SNE rangelands are almost twice as extensive as CNE rangelands (13.1 million km2 compared to 7.8 million km2). Our GIS analysis revealed areas classified as rangelands by the Society for Range Management, which are however not subject to human subsistence or commercial use (13.842 million km2), and which amount to 19.3% of all rangelands or 10.6% of Earth's land

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

6

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9

Table 1 Global rangeland use according to our social-ecological typology: percentages are provided relative to total rangeland area and relative to total land mass excluding Antarctica. Rangeland type

Area [million km2]

Pct. of rangelandsa

Pct. of total land massb

Commercial/equilibrium (CE) Commercial/non-equilibrium (CNE) Subsistence/equilibrium (SE) Subsistence/non-equilibrium (SNE) Commercial Subsistence Human-used equilibrium rangelands Human-used non-equilibrium rangelands

18.442 7.823 18.490 13.087 26.265 31.577 36.932 20.910

25.7 10.9 25.8 18.3 45.4 54.6 51.5 29.2

14.1 6.0 14.1 10.0 20.0 24.1 28.2 16.0

Total rangeland area with human use Total rangeland area w/o human use

57.842 13.842

80.7 19.3

44.1 10.6

Total global rangeland area

71.684

100.0

54.7

a

We calculated total rangeland area using the map of global rangelands provided by the Society of Range Management (Launchbaugh and Strand, 2015). b We refer to total land mass as Earth's land mass excluding Antarctica and have computed percentages accordingly. In Robinson Projection, this land mass amounts to 131.05 million km2.

mass. Most notably, these areas include the Great Western and Eastern Erg and Hoggar areas in Algeria, the Libyan Desert in Libya and Egypt, the Rub al-Chali and Nefud Deserts on the Arabian Peninsula, northern Canada's tundra and large parts of central Western Australia. 2.5. Discussion It has been pointed out repeatedly that uncritical adoption of the non-equilibrium concept and premature derivation of management implications from it might in fact foster unsustainable rangeland management (Illius and O'Connor, 1999; Vetter, 2005; von Wehrden et al., 2012). Following up on this argument, we have presented a framework here that extends the non-equilibrium concept and frames degradation of grazed non-equilibrium rangelands, which cannot be explained by non-equilibrium theory alone, as consequence of stabilizing the system through external inputs, such as supplementary feed, chemical fertilizers or irrigation. Using the example of Namibia, we have proposed that the system in turn might behave ecologically as if it were equilibrium, even though it would classify as non-equilibrium. This mechanism might explain the reported degradation in grazed non-equilibrium rangelands such as Namibia, Syria and Australia. Our framework is a ‘gross oversimplification’ (Ellis and Ramankutty, 2008: 440) of complex social-ecological systems. Yet, it may actually be richer than it first seems. While the proposed typology is categorical in nature, there are secondary continuous dimensions to each axis that could ultimately allow for further and more nuanced differentiation. For instance, the Cv gradient along the ecological axis could imply a gradient in terms of variability of net primary productivity (biomass variability). Likewise, possible secondary dimensions along the socio-economic axis could be the scale of farming operations, which usually goes from small-scale subsistence farms to large-scale commercial farming, usually both in terms of livestock numbers and area involved. Economic variability also increases along this axis in a sense that, in contrast to subsistence farms, commercial farms are closely connected to markets which makes them susceptible to fluctuations in input, output and factor prices. Lastly, there is also a temporal dimension along both axis, because climate change might increase or decrease long-term precipitation variability, which in turn might have repercussions on what type of farming is predominantly practiced, or if farming can be practiced at all. There is no one-size-fits-all solution to rangeland management. Particularly, the question of rangeland management needs to be put into socio-economic context, and what works for one system might

be inadequate for another one, depending on where the system is on the commercial-subsistence dimension, the ecological dimension, and also depending on the history of local land use. As Ostrom (2009: 421) stated, “Simple blueprint policies do not work.” In this context, the main benefit of our proposed classification for rangelands is that it adds a socio-economic component to the discussion about rangeland health, which seems to have often been neglected in past discussions as pointed out by Vetter (2005). It is inadequate to view humans merely as ‘disturbance’ to the ‘natural’ system. Much rather, the socio-economic background of a system should always be taken into account when thinking about remedies for rangeland degradation. Here we highlight some of the challenges these four different rangeland types face. Our intention is not to suggest particular policies for a given rangeland type, but rather to point out that different sorts of interventions are likely to work differently in different social-ecological systems. Important contributing factors to unsustainable rangeland management in commercial farming systems, particularly in the CNE type, are negative externalities and market failure. They contribute to over-fertilization of soils, excessive use of pesticides and loss of biodiversity, to name only a few. Tegtmeier and Duffy (2004) have estimated external costs in U.S. commercial livestock farming at more than $700 million annually. In commercial farming systems with their access to market infrastructure, one option of tackling externalities is institutional measures such as internalizing externalities by imposing taxes, creating a market for pollution permits (e.g. Hanley et al., 2007), or by raising consumer awareness for such externalities. Generally, these mechanisms would work regardless of whether a system is CE or CNE. However, one might think of designing direct state regulatory mechanisms regarding, for example, the use of fertilizer in CNE systems and the stocking density in CE systems as a way to control degradation from over-use of rangelands. Another more market-based solution would be to internalize external costs incurred by the use of certain means of production by having a tax tariff that takes into account (i.e. internalizes) the actual external costs. The obvious problem with this approach is of course that external costs have to be quantified, which is not a straightforward task, as exemplarily illustrated by the debate about the right carbon tax rate (cf. Yohe et al., 2007). The sometimes cumbersome legislation process might add yet another intricacy. In comparison with CNE systems, we think there may be an additional need to account for the ‘natural’ tendency of rangelands to degrade over time in CE systems. One management approach would involve state regulations regarding stocking densities, which would have to account for the inter-annual rainfall variability at farm locations. In commercial systems, we think there

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9

is in general the possibility to aim at sustainable rangeland management through the use of market mechanisms such as taxes or subsidies. In addition, non-market measures, such as government regulations might help meet the goal of long-term sustainable rangeland use in CE systems. Unsustainable use of rangelands in SE and SNE farming systems probably requires different approaches. As fleshed out previously, there is often little or no access to markets and no privately owned, but common land (cf. Fig. 1). Pasture land is very often a common pool resource, which is potentially prone to over-use (Lloyd, 1980 [1833], Hardin, 1968). Particularly SNE systems seem to be vulnerable to state interventions, as illustrated by the example of Tibet (Næss, 2013; Wang et al., 2014). In the 1990s and early 2000s, the Chinese government decided to sedentarize nomads in the Tibetan Autonomous Region in an effort to fight overgrazing and to ensure a better enforceability of state regulations with the ultimate aim to augment productivity (Yamaguchi, 2011). However, this lead to localized high grazing pressure at settlement sites with no possibility for the nomads to follow their usual rotational grazing routine (Miehe et al. 2008). The outcome e certainly unintended by the Chinese administration e was a largely degraded system, undesirable from both, conservation and human well-being point of view. It seems that, because of the ever-changing local precipitation conditions, SNE systems work best when undisturbed by state legislations as these are simply too slow or too remote to adequately react. Instead, for subsistence systems in general, it has been argued that an investment in local farmers' education, and the creation of a local farmers' network to establish a discussion platform and stronger sense of community to enable more efficient self-organization might be suitable policy options (Ostrom, 2009). Providing access to financial capital through micro-credits and micro-insurance (cf. Siegel et al., 2001) as a form of risk management could also be a viable strategy. These policy options are very local in nature and focus on the ability to ‘self-organize’. Hence, following this logic, the options for SE and SNE systems all seem very context-dependent. For example, micro-credits and microinsurances could explicitly take into account rainfall variability as parameter, and the amount of cash and insurance provided to farmers could be adapted to the current state of affairs. Yet another special case illustrating the need for a case-by-case approach is (semi-) nomadic subsistence use of rangelands. In these systems, herding practices may depend on a multitude of further geographical and social-ecological factors that may foster peculiar practices. For example, Syrian nomads have been reported to have water delivered by trucks to their campsites for irrigation (Geerken and Ilaiwi, 2004). Moreover, other factors such as clan affiliation and longstanding cultural traditions may be the organizing principle rather than political legislation. Furthermore, our global social-ecological rangeland map provides an alternative perspective on global human land use as compared to Ellis and Ramankutty (2008) work on anthropogenic biomes, and focuses on human rangeland use. Ellis and Ramankutty (2008) report a total global rangeland area of 39.7 million km2 (Ellis and Ramankutty (2008): supplemental information), which is considerably less than our result of 71.7 million km2. This difference can mainly be explained by the fact that Ellis and Ramankutty classified the North American and Russian tundra as well as barren regions into a separate category called “wildlands”. Different labeling here largely explains the difference in areas between what Ellis and Ramankutty (2008) report and our result, based on the Society for Range Management's classification, since “wildlands” and “rangelands” in Ellis' and Ramankutty's framework together amount to a total of 69.1 million km2. A further source of difference goes back to the Society for Range Management's rather broad classification of rangelands and non-rangelands, which is based on

7

the ecological concept of potential vegetation. While this entails the categorization of areas as rangelands that are in fact predominantly used as croplands, it does not impact our framework as such, and reflects the fuzzy boundary between these two land use categories. The proposed framework must not be misinterpreted as ‘binary reduction of a colorful world towards a black-and-white view’ (Retzer, 2006: 163). Rather, it should be seen as an attempt towards a richer understanding of rangelands as complex social-ecological systems, and may be used as a less nuanced and less contextspecific complement to Ostrom (2009) or Fox et al. (2009). As such, it could provide a starting point to achieve a visualization of the different kinds of dynamics in rangelands, which depend on ecological and socio-economic factors. Within this research agenda, this could be an important milestone towards a ‘complete’ map of global rangelands in terms of their social-ecologic state to inform sustainable rangeland management. 3. Conclusions We see two major take-home messages. (1) Sustainable rangeland management requires integration of ecological and socioeconomic factors. The framework presented may be a helpful and easily accessible tool towards a better understanding of rangelands as complex social-ecological systems. Its four categories may help inform rangeland users as well as policy makers about the dynamics of the system they are using and managing, in order to conserve it for future generations. (2) Artificial stabilization of nonequilibrium rangelands should be handled with great care. It is likely to lead into mid- or long-term biophysical degradation, and to make livestock farming utterly obnoxious, or only feasible at high costs. Actual large-scale restoration efforts in China show that it might actually be possible to reverse even very severe cases of rangeland degradation, but only at very high costs. It is furthermore unclear whether such large-scale restoration would be feasible outside the peculiar setting of China. A thoughtful management approach taking into account what we know from ecology and the social sciences will be needed, particularly, the presence of externalities and market failures in agricultural systems should be acknowledged, for they are key to solutions when thinking about institutional design. Acknowledgments We thank Julia Bandel for her work on earlier versions of the map shown in Fig. 2. J.-O. Engler greatly acknowledges financial support in line with the grant “Biodiversity and productivity in rangelands and their management” from Leuphana University. References Abson, D., Sherren, K., Fischer, J., 2017. Chapter 14: the resilience of Australian agricultural landscapes characterized by land sparing vs. land sharing. In: Gardner, S., Ramsden, S., Hails, R. (Eds.), Agricultural Resilience: Perspectives from Ecology and Economics, Ecological Reviews (British Ecological Society). Cambridge University Press (in press). Bachewe, F., 2009. The State of Subsistence Agriculture in Ethiopia: Sources of Output Growth and Agricultural Inefficiency. Ph.D. dissertation, retrieved from the University of Minnesota Digital Conservancy. http://hdl.handle.net/11299/ 54278. on 06/15/2017. Berger, J., Buuveibaatar, B., Mishra, C., 2013. Globalization of the cashmere market and the decline of large mammals in Central Asia. Conserv. Biol. 2 (7), 679e689. Boone, R.B., Wang, G., 2007. Cattle dynamics in African grazing systems under variable dynamics. J. Arid Environ. 70, 495e513. Briske, D.D., Fuhlendorf, S.D., Smeins, F.E., 2003. Vegetation dynamics on rangelands: a critique of the current paradigms. J. Appl. Ecol. 40, 601e614. Briske, D.D., Sayre, N.F., Huntsinger, L., Fernandez-Gimenez, M., Budd, B., Derner, J.D., 2011. Origin, persistence and resolution of the rotational grazing debate: integrating human dimensions into rangeland research. Rangel. Ecol. Manag. 64 (4), 325e334.

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

8

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9

Cambridge Dictionaries Online, 2014. Entry on ‘commercial Farming’. Cambridge University Press, UK. http://dictionary.cambridge.org/de/worterbuch/businessenglisch/commercial-farming. retrieved on 04/28/2015. China Economic Review, 2013. Beijing Is Delicately Super-sizing the Countries Farms, 07/15/2013, retrieved on 03/05/2015 from. http://www. chinaeconomicreview.com/beijing-land-reform-Brazil-rural-farms-Wulimingindustrial-farming. Ellis, E.C., Ramankutty, N., 2008. Putting people in the map: anthropogenic biomes of the world. Front. Ecol. Environ. 6 (8), 439e447. Ellis, J.E., Swift, D.M., 1988. Stability of African pastoral ecosystems: alternate paradigms and implications for development. J. Rangel. Manag. 41, 450e459. Endicott, E., 2012. A History of Land Use in Mongolia. The Thirteenth Century to the Present. Palgrave Macmillan. Engler, J.-O., Baumg€ artner, S., 2015. Model choice and size distribution: a Bayequentist approach. Am. J. Agric. Econ. 97 (3), 978e997. ESRI, 2014. ArcGIS Desktop: Release 10.2.1. Environmental Systems Research Institute, Redlands, CA. FAO [Food and Agriculture Organization of the United Nations], 2009. Review of evidence on drylands pastoral systems and climate change. In: Neely, C., Bunning, S., Wilkes, A. (Eds.), FAO Land and Water Discussion Paper 8 (Rome). FAOSTAT [Food and Agriculture Organization of the United Nations Statistics Division], 2013. FAO Statistical Yearbook. retrieved from. http://faostat3.fao.org/ home/E. on 03/27/2015. Fatichi, S., Ivanov, V.Y., Caporali, E., 2012. Investigating inter-annual variability of precipitation at the global scale: is there a connection with seasonality? J. Clim. 25, 5512e5523. Fernandez-Gimenez, M.E., Allen-Diaz, B., 1999. Testing a non-equilibrium model of rangeland vegetation dynamics in Mongolia. J. Appl. Ecol. 36, 871e885. Fleischner, T.L., 1994. Ecological costs of livestock grazing in Western North America. Conserv. Biol. 8 (3), 629e644. Food & Water Watch, 2010. Factory Farm Nation: How America Turned its Livestock Farms into Factories. report November 2010, retrieved from. www. foodandwaterwatch.org. on 02/10/2015. Fox, W.E., et al., 2009. An integrated social, economic, and ecologicalconceptual (ISEEC) framework for considering rangeland sustainability. Soc. Nat. Resour. 22 (7), 593e606. Geerken, R., Ilaiwi, M., 2004. Assessment of rangeland degradation and development of a strategy for rehabilitation. Remote Sens. Environ. 90, 490e504.  pez-Hoffman, L., Nabhan, G.P., Knight, R.L., Goldstein, J.H., Presnall, C.K., Lo Ruyle, G.B., Toombs, T.P., 2011. Beef and beyond: paying for ecosystem services on Western US rangelands. Rangelands 33, 4e12. Han, J.G., Zhang, Y.J., Wang, C.J., Bai, W.M., Wang, Y.R., Han, G.D., Li, L.H., 2008. Rangeland degradation and restoration management in China. Rangel. J. 30, 233e239. Hanley, N., Shogren, J.F., White, B., 2007. Environmental Economics in Theory and Practice, second ed. Palgrave Macmillan, New York, NY. Hardin, G., 1968. The tragedy of the commons. Science 162 (3859), 1243e1248. Hardin, D., 2004. World Regional Maps. Longwood University, Farmville, VA. http:// longwood.edu/staff/hardinds/Maps/Mapindex.htm. retrieved on 03/25/2015. Harris, C.D., 2013. Mongolia. retrieved from. In: Encyclopædia Britannica. http:// www.britannica.com/EBchecked/topic/389335/Mongolia/27446/Agricultureforestry-and-fishing. on 03/03/2015. Harrup, L.E., Laban, S., Purse, B.V., Reddy, Y.K., Reddy, Y.N., Byregowda, S.M., Kumar, N., Purushotham, K.M., Kowalli, S., Prasad, M., Prasad, G., Bettis, A.A., De Keyser, R., Logan, J., Garros, C., Gopurenko, D., Bellis, G., Labuschagne, K., Mathieu, B., Carpenter, S., 2016. DNA barcoding and surveillance sampling strategies for Culicoides biting midges (Diptera: Ceratopogonidae) in southern India. Parasites Vectors 9, 461. Havstad, K.M., Peters, D.P.C., Skaggs, R., Brown, J., Bestelmeyer, B., Fredrickson, E., Herrick, J., Wright, J., 2007. Ecological services to and from rangelands of the United States. Ecol. Econ. 64, 261e268. Illius, A.W., O'Connor, T.G., 1999. On the relevance of non-equilibrium concepts to arid and semi-arid grazing systems. Ecol. Appl. 9, 798e813. €rtner, S., 2014. Determinants and Interactions of Ingenillem, J., Merz, J., Baumga Sustainability and Risk Management of Commercial Cattle Farmers in Namibia. University of Lüneburg Working Paper Series in Economics 304, July 2014. €rtner, S., Frank, K., 2014. How do indiJacoby, O., Quaas, M.F., Müller, B., Baumga vidual farmers' objectives influence the evaluation of rangeland management strategies under a variable climate? J. Appl. Ecol. 51, 483e491. Klink, C.A., Machado, R.B., 2005. Conservation of the Brazilian Cerrado. Conserv. Biol. 19 (3), 707e713. Knapp, A.K., Fay, P.A., Blair, J.M., Collins, S.L., Smith, M.D., Carlisle, J.D., Harper, C.W., Danner, B.T., Lett, M.S., McCarron, J.K., 2002. Rainfall variability, carbon cycling, and plant species diversity in a mesic grassland. Science 298, 2202e2205. Lal, R., 2003. Global potential of soil carbon sequestration to mitigate the greenhouse effect. Crit. Rev. Plant Sci. 22 (2), 151e184. Lal, R., 2004. Soil carbon sequestration impacts on global climate change and food security. Science 304, 1623e1627. Launchbaugh, K., Strand, E., 2015. Global Map of Rangelands. Society for Range Management map retrieved from. http://www.webpages.uidaho.edu/what-isrange/rangelands_map.htm. on 01/08/2015. Lloyd, W.F., 1980. [1833], W.F. Lloyd on the checks of population. Popul. Dev. Rev. 6 (3), 473e496. Mendelsohn, J., 2006. Farming Systems in Namibia, Research and Information Services in Namibia (Windhoek, Namibia).

MEA [Millennium Ecosystem Assessment], 2005. Ecosystems and Human Wellbeing: Synthesis. Island Press, Washington, DC. Miehe, G., Miehe, S., Kaiser, K., Jianquan, L., Zhao, X., 2008. Status and dynamics of the Kobresia pygmaea ecosystem on the Tibetan plateau. AMBIO A J. Hum. Environ. 37 (4), 272e279. Miehe, S., Kluge, J., von Wehrden, H., Retzer, V., 2010. Long-term degradation of Sahelian rangeland detected by 27 years of field study in Senegal. J. Appl. Ecol. 47, 692e700. Mørch, F.C., 1999. Mediterranean agriculture e an agro-ecological strategy. Dan. J. Geogr. 1, 143e156. Næss, M.W., 2013. Climate change, risk management and the end of Nomadic pastoralism. Int. J. Sustain. Dev. World Ecol. 20 (2), 123e133. Olbrich, R., 2012. Environmental Risk and Sustainability: the Case of Livestock Farming in Semi-arid Rangelands. Ph. D. dissertation. Leuphana University of Lüneburg. available online at: http://www.leuphana.de/ub. €rtner, S., 2014. Personal norms of sustainability and Olbrich, R., Quaas, M.F., Baumga farm management behavior. Sustainability 6 (8), 4990e5017. Olff, H., Ritchie, M.E., Prins, H.H.T., 2002. Global environmental controls of diversity in large herbivores. Nature 415, 901e904. Ostrom, E., 2009. A general framework for analyzing sustainability of socialecological systems. Science 325, 419e422. €rtner, S., Becker, C., Frank, K., Müller, B., 2007. Uncertainty and Quaas, M.F., Baumga sustainability in the management of rangelands. Ecol. Econ. 62 (2), 251e266. Retzer, V., 2006. Impacts of grazing and rainfall variability on the dynamics of a Sahelian rangeland revisited (Hein, 2006) e new insights from old data. J. Arid Environ. 67, 157e164. Reynolds, J.F., et al., 2007. Global desertification: building a science for dryland development. Science 316, 847e851. SADC [Swiss Agency for Development and Cooperation], 2009. Benefits of Sustainable Land Management, UNCCD, World Overview of Conservation Approaches and Technologies. FAO, Center for Development and Environment. Safriel, U., Adeel, Z., Niemeijer, D., Puidefabreagas, J., White, R., Lal, R., Winslow, M., Ziedler, J., Prince, S., Archer, E., King, C., 2005. Drylands. Chapter 22. In: Hassan, R., Scholes, R., Ash, N. (Eds.), Ecosystems and Human Well-being: Current State and Trends, Millennium Ecosystem Assessment Series, vol. 1. Island Press, Washington D.C. Sasaki, T., Okayasu, T., Jamsran, U., Takeuchi, K., 2008. Threshold changes in vegetation along a grazing gradient in Mongolian rangelands. J. Ecol. 96, 145e154. Sheehy, D.P., Thorpe, J., Kyrichuck, B., 2006. Rangeland, livestock and herders revisited in the northern pastoral region of China. USDA Forrest Serv. Proc. RMRS-P-39 62e82. Siegel, P.B., Alwang, J., Canagarajah, S., 2001. Viewing Micro-insurance as a Social Risk Management Instrument. Social Protection Discussion Paper Series No. 0116. Social Protection Unit, Human Development Network, The World Bank. Tanaka, J.A., Brunson, M., Torell, L.A., 2011. A social and economic assessment of rangeland conservation practices. In: Briske, David D. (Ed.), Conservation Benefits of Rangeland Practices: Assessment, Recommendations, and Knowledge Gaps. USDA-Natural Resource Conservation Service, Washington, DC, USA, pp. 371e422 available online at: http://www.nrcs.usda.gov/Internet/FSE_ DOCUMENTS/stelprdb1045804.pdf. (Accessed 25 March 2015). Tegtmeier, E.M., Duffy, M.D., 2004. External costs of agricultural production in the United States. Int. J. Agric. Sustain. 2 (1), 1e20. Thornton, P.K., 2010. Livestock production: recent trends, future prospects. Phil. Trans. R. Soc. B 365, 2853e2867. Thornton, P.K., et al., 2006. Mapping Climate Vulnerability and Poverty in Africa. ILRI, Nairobi, Kenya see. http://www.dfid.gov.uk/research/mapping-climate.pdf. du Toit, C.J.L., Meissner, H.H., van Niekerk, W.A., 2013. Direct methane and nitrous oxide emissions of South African dairy and beef cattle. South Afr. J. Anim. Sci. 43 (3), 320e339. USDA [United States Department of Agriculture], 2016. Overview of the United States Cattle Industry, National Agricultural Statistics Service. Agricultural Statistics Board retrieved from. http://usda.mannlib.cornell.edu/usda/current/ USCatSup/USCatSup-06-24-2016.pdf. on 06/15/2017. Vetter, S., 2005. Rangelands at equilibrium and at non-equilibrium: recent developments in the debate. J. Arid Environ. 62, 321e341. Wang, Y., Wang, J., Li, S., Qin, D., 2014. Vulnerability of the Tibetan pastoral systems to climate and global change. Ecol. Soc. 19 (4), 8. https://doi.org/10.5751/ES06803-190408. Waters, T., 2007. The Persistence of Subsistence Agriculture: Life beneath the Level of the Marketplace. Lexington Books, Lanham, MD. von Wehrden, H., Wesche, K., Miehe, G., 2009. Plant communities of the southern Mongolian Gobi. Phytocoenologia 39, 331e376. von Wehrden, H., Hanspach, J., Kaczensky, P., Fischer, J., Wesche, K., 2012. Global assessment of the non-equilibrium concept in rangelands. Ecol. Appl. 22 (2), 393e399. von Wehrden, H., Wesche, K., Chuluunkhuyag, O., Fust, P., 2015. Correlation of trends in cashmere production and declines of large wild mammals: response to Berger et al. (2013). Conserv. Biol. 29 (1), 286e289. World Bank, 2009. Minding the Stock: Bringing Public Policy to Bear on Livestock Sector Development. Report no. 44010-GLB. Washington, DC. World Bank, 2013. Mongolia: Portable Solar Power for Nomadic Herders retrieved from. http://www.worldbank.org/en/results/2013/04/08/portable-solar-powerfor-nomadic-herders. on 03/03/2015. Yamaguchi, T., 2011. Transition of mountain pastoralism: an agrodiversity analysis of the livestock population and herding strategies in southeast Tibet, China.

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009

J.-O. Engler et al. / Journal of Arid Environments xxx (2017) 1e9 Hum. Ecol. 39 (2), 141e154. Yohe, G.W., et al., 2007. Executive summary. In: Parry, M.L., et al. (Eds.), Perspectives on Climate Change and Sustainability, Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press.

9

Zeidler, J., Hanrahan, S., Scholes, M., 2004. Determining termite diversity in arid Namibian rangelands e a comparison of sampling methods. Afr. Zool. 39 (2), 285e292. Zhao, W.Y., Li, J.L., Qi, J.G., 2007. Changes in vegetation diversity and structure in response to heavy grazing pressure in the northern Tianshan Mountains, China. J. Arid Environ. 68, 465e479.

Please cite this article in press as: Engler, J.-O., et al., A social-ecological typology of rangelands based on rainfall variability and farming type, Journal of Arid Environments (2017), https://doi.org/10.1016/j.jaridenv.2017.09.009