Alpine Botany

, Volume 127, Issue 1, pp 1–15 | Cite as

A global inventory of mountains for bio-geographical applications

  • Christian Körner
  • Walter Jetz
  • Jens Paulsen
  • Davnah Payne
  • Katrin Rudmann-Maurer
  • Eva M. Spehn
Open Access
Original Article

Abstract

Mountains are hotspots of biodiversity. Yet, evaluating their importance in global biodiversity inventories requires the adoption of a pertinent definition of mountains. Here, we first compare the well-established WCMC and GMBA definitions, which both use geographical information systems. We show that the WCMC approach arrives at twice the global mountain area and much higher human population numbers than the GMBA one, which is explained by the inclusion of (mostly) low latitude hill country below 600 m elevation. We then present an inventory of the world’s mountains based on the GMBA definition. In this inventory, each of the 1003 entries corresponds to a polygon drawn around a mountain or a mountain range and includes the name of the delineated object, the area of mountainous terrain it covers stratified into different bioclimatic belts (all at 2.5′ resolution), and demographic information. Taken together, the 1003 polygons cover 13.8 Mio km2 of mountain terrain, of which 3.3 Mio km2 are in the alpine and nival belts. This corresponds to 83.7% of the global mountain area sensu GMBA, and 94% of the alpine/nival area. The 386 Mio people inhabiting mountainous terrain within polygons represent 75% of the people globally inhabiting mountains sensu GMBA. This inventory offers a robust framework for the integration of mountain biota in regional and larger scale biodiversity assessments, for biogeography, bioclimatology, macroecology, and conservation research, and for the exploration of a multitude of socio-ecological and climate change-related research questions in mountain biota, including the potential pressure on alpine ecosystems.

Keywords

Alpine Biodiversity Climate Elevation Montane Map of life 

Introduction

Mountains are remarkable in many ways, and in particular in the exceptional biodiversity they host. Likely because of the rapid change in climatic conditions with elevation across very short geographical distances and because of strong contrasts in life conditions with varying exposure and topography, mountains support an estimated one-third of terrestrial species diversity (Körner 2004) and host half of all 34 global biodiversity hotspots (Chape et al. 2008). The global alpine terrain alone hosts about twice the number of flowering plant species expected based solely on the area it covers (Körner et al. 2011). Additionally, because of the biogeographic isolation under which mountain ecosystems have evolved, these biomes also harbor high numbers of endemic species (Vetaas and Grytnes 2002; Barthlott et al. 2005; Tang et al. 2006). The richness of plant species observed in mountains is particularly important, in that a diverse vegetation is important for securing soils on steep slopes, thereby contributing to the protection of landscapes and populations against natural hazards and the impact of extreme events. Mountain vegetation further provides food, fiber, and fodder and mountains in their entirety supply clean water to half of humankind (Messerli and Ives 1997; Egan and Price 2016), attract tourists, and often host cultural heritage sites and landscapes of outstanding species richness and beauty. Because of their geodiversity, mountain ecosystems have served as refuge for organisms during past climatic changes and are predicted to provide protection for flora and fauna also under forthcoming climatic change (Scherrer and Körner 2011). Accordingly, mountain ecosystems represent areas of prime conservation value (Messerli and Ives 1997; Hamilton 2002; Harmon and Worboys 2004; Körner and Ohsawa 2006).

The global occurrence of mountains across all latitudes and the steep, small-scale environmental gradients that characterize them provides exciting research opportunities (Körner 2000). Additionally, as large environmental and biodiversity data sets become more readily available, new analyses become possible on a macro-ecological scale (Jetz et al. 2012). Because of their steep climatic gradients, mountains are also sentinels of change in a rapidly warming world (La Sorte and Jetz 2010; Pauli et al. 2012) and, thereby, offer unique ‘experiments by nature’ for studying the mechanisms driving the evolution and maintenance of biodiversity (Janzen 1967; Brown 1971; Körner 2000, 2001). Yet, comparative research in mountain biodiversity and conservation planning has been constrained by the lack of a standardized delineation and environmental characterization of the world’s mountains. The spatial placement of inventories mobilized from the literature and the determination of expected species occurrences following expert-based bioclimatic associations have turned out to be particularly challenging (Jetz et al. 2012, http://www.mol.org). For the world’s islands, both an existing inventory (Dahl 1991) and a more recent bioclimatic and physical characterization (Weigelt et al. 2013) have enabled fascinating research, including the comparative analyses of island flora and fauna and of their phylogenetic structure (Weigelt et al. 2016).

One major reason for the absence of an inventory of the mountains of the world is a long-lasting controversy about the arguably difficult definition of mountains (e.g., Smith and Mark 2003; Byers et al. 2013). This controversy results in part from the fact that different mountain definitions have been proposed for different contexts of use and even for different countries. Accordingly, a definition tailored for applications related to life conditions for humans, plants, and animals may not serve the purpose of applications related to hydrological (catchments) and climatological (atmospheric circulation) phenomena, mountaineering and tourism in general, mining, or political boundaries (Körner and Ohsawa 2006). Yet, regardless of the context of use, most people would agree that to define mountains, a distinction must be made between elevation (and its implications such as reduced barometric pressure and the associated reduction in atmospheric temperature), and steepness as a measure of inclination and gravitational forcing (Fig. 1). Among the geographic information system (GIS)-based approaches commonly applied to formulate a mountain definition, the one by Kapos et al. (2000) constrains mountains by a combination of elevation and ruggedness, whereas the one by Körner et al. (2011) constrains mountains by ruggedness of terrain only, irrespective of elevation. These two approaches define whether a point (a defined grid area) of the earth’s surface is mountainous or not, but they do not define mountain territory as geographical regions. The former, herein called WCMC, was developed at the United Nations Environmental Programme World Conservation Monitoring Center (WCMC), used as part of the process of creating a map of the world’s mountain forests (Blyth et al. 2002), and subsequently adopted by other international organisations. The latter, herein called GMBA, was developed by the Global Mountain Biodiversity Assessment network for mountain biodiversity and biogeography research (e.g., Paulsen and Körner 2014) and is being increasingly adopted in diverse scientific communities from natural sciences to the humanities (e.g., Green and Stein 2015).
Fig. 1

A mountain concept best differentiates between elevation as such versus topographic ruggedness. Left the village of Reine at sea level on the Norwegian Lofoten Islands, surrounded by very steep, but low elevation mountains. Right close to 2000 m asl plains in Wyoming (USA, Green River; photos by UM Weber). The landscape on the left resembles mountain terrain, the one on the right does not

In this paper, we offer a first version of a much needed inventory of the mountains of the world. This inventory rests on the GMBA definition of mountains, which, based on a quantitative comparison between the WCMC and GMBA definitions that we herein provide, appears as the most pertinent in a biogeographic context. Each entry corresponds to a polygon drawn around a single mountain or a mountain range and includes a local name as provided in major published maps and atlases, geographic coordinates, the total polygon area, and the area of mountainous terrain within polygon, subdivided into bioclimatic belts (sensu GMBA). Additionally, and to illustrate the broad utility of this inventory, we also report on human population densities in mountainous areas. The exact boundaries of any of the polygons have no bearing on the mountain statistics we provide (i.e., covered area of mountain terrain), as these strictly follow the GMBA definition of mountains.

Materials and methods

Summary of WCMC and GMBA mountain definitions

In the WCMC method, Kapos et al. (2000) define six classes of land area that belong to mountain terrain using the GTOPO30 digital elevation model, the WGS84 original projection, 2.5′ pixels, and local elevation ranges in meters. All area above 2500 m asl is considered to be mountainous irrespective of ruggedness, whereas land below 300 m asl is non-mountainous by default, whatever its inclination. Land between 300 and 2500 m is attributed to mountain terrain based on ruggedness, with requirements of a minimum elevation contrast of 300 m over a horizontal distance of 7 km between 300 and 1000 m asl; a ≥ 5° slope between the highest and lowest points within a grid cell or a minimum elevation contrast of 300 m over a horizontal distance of 7 km between 1000 and 1500 m; and a ≥ 2° slope between the highest and lowest points within a grid cell between 1500 and 2500 m.

In the GMBA method, Körner et al. (2011) make no distinction by elevation, but apply a minimum 200 m elevational amplitude among 3 × 3 = 9 grid points of 30″ in 2.5′ pixels. For a 2.5′ pixel to be defined as rugged (i.e., ‘mountainous’), the difference between the lowest and highest of the 9 points must exceed 200 m. A 2.5′ pixel in a grid corresponds to an area of 4.6 × 4.6 km near the equator. This size decreases as latitude increases. The maximum horizontal distance between any two of the 9 points within a 2.5′ pixel that shows the largest vertical distance ranges from 0.9 to 3 km, depending on latitude and relief.

Following McVicar and Körner (2012), the term ‘elevation’ is used herein for the vertical distance of terrain from sea level, while ‘altitude’ applies to atmospheric conditions. Note also that the term ‘ruggedness’ does not imply ‘steepness’. ‘Steepness’ is commonly defined as a feature of a single slope at a resolution of, for instance, 10–100 m, which is a scale that is far too small to be handled in a global geostatistics, even in a narrow geographical grid. Thus, both the WCMC and GMBA approaches use ruggedness of terrain instead as a measure of the vertical amplitude within a pre-defined space. This corresponds to the geometrical slope between the lowest and the highest grid point in a specific area.

Comparison of mountain definitions

Differences between the WCMC and GMBA mountain coverage were evaluated based on a comparison of the global area of land outside Antarctica defined as mountainous according to either approach. The values obtained for each definition were summed for eight mega-regions (mostly continents): North and South America, Africa, Europe, Asia, Australia, Oceania, and Greenland. The countries belonging to central America were attributed to North America. Herein, the terms “mountainous” and “mountain” are used for both the WCMC and the GMBA approaches despite the differences between the two definitions.

Delineation of mountain polygons (POLY)

Polygons were delineated by hand in ArcGIS10.0 (http://www.esri.com). The process started with the localization of mountains and mountain ranges in atlases and online resources. The sources selected in 2013 when the work started consisted in the world atlas of Knaurs (1999), a regional inventory of the mountains of Asia by Gurung (1999), and additional online maps at comparable resolution and occasionally in local language that were used in case of doubts (see supplementary material and https://ilias.unibe.ch/goto_ilias3_unibe_cat_1000515.html). A polygon was then drawn around a given mountain or mountain range if it (1) covered a minimum of 20 adjacent pixels of rugged terrain sensu GMBA on the GIS map (at a 2.5′ resolution, this represents approximately 420 km2 at the equator and smaller areas at higher latitudes; see above and Körner et al. 2011 for the definition of ruggedness), (2) had a name specified in the chosen resources (but see the discussion for a few exceptions), and (3) did not consist of scattered entities separated by extensive areas of lowland terrain. The drawings did not account for political borders or exact continental boundaries. The resulting polygons became either large if they included entire mountain ranges to which all the above criteria applied, or they remained small if they included only individual mountains. Although the primary objective was to include as much rugged terrain (sensu GMBA) as possible, the polygons inevitably included some non-rugged terrain and occasionally excluded less conspicuous, smaller scale rugged terrain that a strictly grid-based statistical procedure would identify as ‘mountainous’. Based on primary feedback from GMBA network members and to simplify world-wide comparisons, we then also grouped individual polygons belonging to large mountain systems (e.g., all polygons belonging to the Rocky Mountains or to the Hindukush-Himalaya). When possible (e.g., Interior Mountains, Hindukush-Himalaya), individual polygons were simply aggregated, in which case the mountain area covered by the resulting large polygons was strictly the sum of the mountain area covered by the individual polygons. When large mountain systems obviously included inconspicuous mountain terrain that was not comprised in any individual polygon, the large polygons also included this additional mountain terrain. In these cases (e.g., Cordillera de Los Andes, Qilian Shan), the mountain area covered by the large polygons was more than the sum of the mountain area covered by the individual polygons. These very large, customized mountain polygons simply serve to facilitate the localization of larger, sub-continental mountain systems for specific applications. Accordingly, they are provided exclusively online (www.mountainbiodiversity.org, https://ilias.unibe.ch/goto_ilias3_unibe_cat_1000515.html) and are neither included in the final polygon count discussed in the result section nor in any of the calculations.

Climatic belts

Irrespective of the way mountains are defined, a global comparison of mountains should not employ elevation above sea level, because different life conditions can be encountered at identical elevations across latitudes, but also between front ranges and the interior of given mountain systems (mass elevation effect). Hence, the only way to arrive at meaningful comparisons is to apply climatic criteria, here referred to as climatic (thermal) belts.

Climatic belts were defined by first applying an algorithm that allows dividing mountain terrain in land above the potential climatic treeline (alpine, including what is often termed sub-nival and nival) and below (montane), and subsequently stratifying the land below the potential treeline in belts ranging from upper montane with recurrent freezing to frost-free tropical low montane and rugged lowland terrain (Körner et al. 2011; Körner 2012). Calculations were based on publicly available climatic data only (WorldClim, http://www.worldclim.org). The algorithm, developed and tested across all latitudes, predicts the potential climatic treeline at elevations where the seasonal mean temperature is 6.4 °C and the local growing season consists of at least 94 days, each with a mean temperature of ≥0.9 °C (Paulsen and Körner 2014). By additionally accounting for snowpack and drought, the algorithm provides a potential treeline, irrespective of whether trees are present at or absent from where they could still grow, which in turn implies that the reference line becomes independent of local peculiarities of tree distribution due to disturbances such as fire, logging, pastoralism, avalanches, erosion. The treeline algorithm was applied to the rugged terrain in both the GMBA and the POLY approaches.

Another climate classification frequently used in climate and climate change research, as well as in physical geography, hydrology, agriculture, and biology is that of Köppen–Geiger (Kottek et al. 2006), which distinguishes five climatic zones that are subsequently subdivided based on precipitation and air temperature. Because this classification is meaningful primarily on a global scale, as opposed to the regional scale at which the polygons are drawn, we applied it to the global mountainous terrain sensu GMBA only.

Estimation of total polygon area, area of mountainous terrain, and area by climatic belt

For each polygon, we quantified the total area and the area of mountainous terrain and stratified the latter by climatic belts. As those calculations strictly followed the GMBA definition of mountains, the exact shape and extent of any particular polygon were irrelevant. At the 2.5′ resolution applied (corresponding to 21 km2 at low latitudes), the accuracy of such calculations is limited when an area of interest is small because edge effects become large relative to the size of the area. Hence, results are expected to be best for large polygons. Additional problems such as the elevational (and thus climatic) amplitude within a 2.5′ grid (e.g., a narrow valley passing through an otherwise predominantly high-elevation pixel) are intrinsic to gridded data.

Polygons were then assigned to mega-regions (see above) and both the total polygon area and the area of mountainous terrain, subdivided into climatic belts, were summarized for each of these mega-regions. The decision to ignore political borders and exact continental boundaries that was made when drawing the polygons was also followed when assigning them to mega-regions. Polygons ranging across the boundaries of two mega-regions (e.g., Greater Caucasus between Asia and Europe) or interfacing with two mega-regions (e.g., Serrania de San Blas in North America interfaces with a grid cell belonging to South America) were attributed twice (i.e., once to each mega-region). Accordingly, the number of polygons used for area calculations did not correspond exactly to the number of uniquely drawn objects.

Human population statistics

Anthropogenic pressure associated with human settlements in and below the montane belts, land use, land transformation, and other human activities may endanger mountain biota more than climate change in many regions of the world (Körner and Ohsawa 2006; Spehn et al. 2006). Here, we provide human population data for the GMBA mountain areas and for the individual mountain polygons to help draw inferences about human impact on mountain biota, including alpine terrain. To quantify human population in mountains, we used the FAO gridded global human population density data and overlaid it with the gridded GMBA and POLY mountain data. Similar quantifications were performed for the WCMC mountain area by Huddelston et al. (2003), with a new release by FAO Mountain Partnership in 2015 (FAO 2015). The FAO data consisted of LandScan (see http://web.ornl.gov/sci/landscan/landscan_documentation.shtml#01) and FAOSTAT data year 2000–2012 (provided by FAO/CDE, pers. com. A. Vita). LandScan data were adjusted by FAO for each country and standardized to the data available on FAOSTAT before being reallocated at pixel level following the LandScan distribution. Because we subdivided mountain terrain into thermal belts, the world’s human population could also be assigned to climatic niches and to climatic envelopes in any non-mountainous terrain.

Results

Comparison of the WCMC and GMBA mountain definitions

The WCMC- and GMBA-based estimates of mountain terrain differ by a factor of about two (Table 1; Fig. 2). While neglecting the admittedly small fraction of often steep but low elevation mountains (<300 m asl, red in Fig. 2, left image in Fig. 1), WCMC includes vast areas of intra-mountain and foreland terrain as well as hill country. For instance, WCMC defines 94% of Switzerland as mountains, including almost all cropland and most urban areas that are commonly not considered mountainous, and approximately 7 of the 8.5 Mio Swiss people appear as mountain inhabitants (see below). The GMBA approach remains quite inclusive by attributing 66% of Switzerland to mountains. This in turn leads to an estimated ‘mountain population’ of approximately 2 Mio people living mainly in the large valleys inside the Alps but not in the metropolitan areas that are included by the WCMC definition. Similarly, almost all of Turkey is attributed to mountains in the WCMC approach, whereas the GMBA one excludes major cropland and urban areas in hill country (Fig. 2). Furthermore, WCMC in contrast to GMBA attributes flat fractions of high-elevation tablelands (>2500 m asl) such as the Mongolia highlands and, at higher elevation, the Tibetan plateau and the Andean Altiplano, to the mountain category. Regardless of the method adopted to delineate mountains, the area of land defined as mountainous is largest at around 600 m of elevation (cut off at 300 m asl in the WCMC approach) and approaches zero above 6000 m asl (Fig. 3).
Table 1

Global mountain terrain, separated by mega-regions, according to the WCMC and the GMBA definitions

Mega-region

Total land area [103 km2]

WCMC mountain area [103 km2]

(% of total land area)

GMBA mountain area [103 km2]

(% of total land area)

Africa

29,900

4255 (14.2)

1191 (4.0)

Asia

44,385

17,386 (39.2)

8883 (20.0)

Australia

7677

357 (4.6)

126 (1.6)

Europe

9780

1862 (19.0)

916 (9.4)

Greenland

2120

836 (39.4)

92 (4.3)

North America

21,683

6300 (29.1)

2902 (13.4)

Oceania

424

209 (49.2)

175 (41.1)

South America

17,755

4034 (22.7)

2149 (12.2)

Total

133,724

35,238 (26.4)

16,434 (12.3)

Fig. 2

Four examples showing the differences between the WCMC and the GMBA mountain definitions. a Asia Minor (Anatolia, Turkey), b Northern Scandinavia (the Norwegian coast), c the Tibetan Plateau, d the European Alps. Color code: green the WCMC and GMBA definition apply; yellow the WCMC definition only applies; gray neither the WCMC nor the GMBA definition apply; red only the GMBA definition applies (mountain areas <300 m asl)

Fig. 3

The global mountain area at different elevations as assessed by WCMC (solid line), GMBA (dashed line), and POLY (dotted line). The cut edge on the left is due to the 300 m asl lower limit of terrain accepted as mountainous by WCMC, irrespective of ruggedness

Global mountain terrain as captured by mountain polygons and climatic belts

This first version of the GMBA mountain inventory consists of 1003 polygons (Fig. 4, see supplementary material), which can be visualized and queried for spatial an biodiversity information online at http://www.mountainbiodiversity.org/. Yet as 14 of these polygons appear in two mega-regions simultaneously (see above) calculations were performed with 1017 objects (c.f. supplementary material). The GMBA mountain polygons cover 19.3% or nearly 26 Mio km2 of the global land area outside Antarctica and 83.7% of the total global mountain area estimate based on the GMBA mountain definition (Table 2). Specifically, the polygons capture between 85.8 and 88.8% of the terrain defined as mountainous (sensu GMBA) in Asia, Europe, North America, and Greenland, 91% of the mountainous terrain of South America, around 60% of the mountainous terrain of Africa and Australia, and around 37% of the mountainous terrain of Oceania (Table 2). With more than 14 Mio km2 of polygon area, of which 7.6 Mio. km2 fall in the mountain category, Asia hosts by far the greatest fraction of the global mountain terrain (Table 2). North America and South America follow with ca. 2.5 and 2 Mio km2 of mountain terrain, respectively, and Africa and Europe with approximately 0.7 and 0.8 Mio km2, respectively. Small fractions of less than 0.1 Mio km2 are located in Australia, Greenland, and Oceania. Because of their smooth shape, the polygons inevitably consist of both mountainous and non-mountainous terrain. Currently, 53.3%, or 13.8 Mio km2, of the total polygon area (26 Mio km2) consists of mountain terrain sensu GMBA.
Fig. 4

The global distribution of mountain polygons. Statistics shown in Tables 2 and 3 represent all 1003 polygons. For certain applications, groups of polygons (e.g., the various polygons identified in the Rocky Mountains) can be merged to larger polygons (e.g., one large Rocky Mountains polygon). See electronic supplementary material and https://ilias.unibe.ch/goto_ilias3_unibe_cat_1000515.html for a detailed list of polygons

Table 2

Areal statistics of polygons (POLY, 2.5′, resolution) and area of mountain terrain within polygons for mega-regions outside Antarctica

Mega-region

Number of polygons

Total land area [103 km2]

Total polygon area

[103 km2]

(% of total land area)

Mountain area within polygons [103 km2]

(% of total polygon area)

Mountain area outside polygons

[103 km2]

Mountain area within polygons in  % of global GMBA mountain area

Africa

70

29,900

1721 (5.8)

706 (41.0)

485

59.3

Asia

454

44,385

14,045 (31.6)

7600 (54.1)

1283

85.6

Australia

7

7677

181 (2.4)

74.8 (41.3)

50.9

59.5

Europe

58

9780

1464 (15.0)

785 (53.6)

131

85.7

Greenland

2

2120

217 (10.3)

81.8 (37.6)

10.4

88.8

North America

268

21,683

4756 (22.0)

2491 (52.3)

411

85.8

Oceania

60

424

96 (22.7)

64.8 (67.2)

110

37.1

South America

98

17,755

3342 (18.8)

1956 (58.5)

192.6

91.0

Total

1013

133,724

25,832 (19.3)

13,760 (53.3)

2674

83.7

The rugged area within polygons is defined by the GMBA mountain definition. For methodological reasons (see main text), the calculations are based on a total number of 1017 polygons

At the level of individual climatic belts, the largest areas of mountainous terrain are found in the lowest, ‘no-freezing’ belt in Africa and South America, in the lower montane belt in Asia, Australia, and Europe, in the upper montane belt in North America and Oceania, and in the nival belt in Greenland (Table 3). Globally, 21.5% of mountain terrain occurs above the potential climatic treeline and 78.5% occurs below. This corresponds to 2.64 versus 9.62% of the entire terrestrial area outside Antarctica (Körner et al. 2011)
Table 3

Areal statistics for the world’s mountain terrain within polygons (POLY, 2.5′ resolution) separated by major climatic belts and mega-regions

 

Moutain area within polygons in km2 (% of total mountain area within polygons per mega-region)

Mega-region

Nivala

Upper alpine

Lower alpine

Upper montane

Lower montane

Warmb

No freezingc

Africa

43 (<0.01)

85 (0.01)

963 (0.1)

16,184 (2.3)

125,564 (17.8)

84,257 (11.9)

479,161 (67.8)

Asia

196,031 (2.6)

420,559 (5.5)

1,356,970 (17.9)

1,637,965 (21.6)

1,765,122 (23.2)

729,388 (9.6)

1,494,421 (19.7)

Australia

0

0

1010 (1.4)

22,078 (29.5)

38,521 (51.5)

11,386 (15.2)

1752 (2.3)

Europe

58,401 (7.4)

34,845 (4.4)

63,721 (8.1)

233,953 (29.8)

367,623 (46.9)

10,768 (1.4)

15,368 (2.0)

Greenland

71,000 (86.8)

7668 (9.4)

3030 (3.7)

146 (0.2)

0

0

0

North America

106,870 (4.3)

167,467 (6.7)

424,972 (17.1)

756,635 (30.4)

519,487 (20.9)

141,961 (5.7)

373,450 (15.0)

Oceania

450 (0.7)

241 (0.4)

3371 (5.2)

31,088 (48.0)

16,762 (25.9)

0

12,842 (19.8)

South America

27,952 (1.4)

70,351 (3.6)

292,615 (15.0)

372,769 (19.0)

298,660 (15.2)

62,824 (3.2)

831,109 (42.5)

The climatic belts follow the definitions by Körner et al. (2011) and are anchored at the potential climatic treeline

aSeason length ≤10 days

bWarm mountain zone, but with possible freezing

cMountain terrain with perpetual warm conditions, no freezing

Human populations in mountains

Mountain polygons as delineated here host a human population of 1.2 billion people, of which 386 million are actually living in mountains sensu GMBA. This corresponds to a density of 28 people km−2 in the mountainous terrain of polygons as compared to 67 people km−2 in the non-mountainous terrain (i.e., the mountain forelands). Of the 386 million ‘mountain people’, 233 million are inhabiting the mountains of Asia, between ca. 53 and 54 million are inhabiting the mountains of Africa and South America, respectively, and ca. 23 million are living in the mountains of each North America and Europe (Table 4). In comparison, the purely GIS-based WCMC and GMBA mountain definitions lead to an estimated 915 (FAO 2015) and 511 million people inhabiting mountains, respectively. Hence, when considering only the strictly mountainous terrain within polygons, the human population estimates reach 75.5% of those obtained at a global scale based on the GMBA definition.
Table 4

Human population statistics for the world’s mountain terrain (GMBA definition and POLY, 2.5′ resolution) separated by mega-regions

Mega-regiona

Total human population [103]

Total human population in total polygon area

[103]

Human population in mountain terrain [103]

(% of total human population)b

Human population densities in mountainsc

(totald)

GMBA

POLY

GMBA

POLY

Africa

1,230,000

151,900

83,100 (6.7)

52,700 (34.7)

69.8

74. (88.2)

Asia

4,497,000

734,100

302,200 (6.7)

233,000 (31.7)

34.0

30.6 (52.3)

Australia

22,530

597

264 (1.2)

73.4 (12.3)

2.1

1.0 (3.3)

Europe

541,300

69,178

29,360 (5.4)

23,300 (33.7)

32.1

29.7 (47.2)

North America

548,400

107,500

37,920 (6.9)

23,100 (21.6)

13.1

9.3 (22.6)

Oceania

8088

322

855 (10.6)

174 (53.9)

4.9

2.7 (3.3)

South America

451,100

135,400

57,490 (12.7)

53,800 (39.7)

26.8

27.4 (40.5)

Total

7,298,420

1,199,000

511,200 (7.0)

386,000 (32.2)

31.1

28.0 (46.4)

aBecause of the low total human population in Greenland (approximately 56,000 inhabitants) and the resulting large uncertainty associated with the estimated total human population in total polygon area, Greenland is not shown in this table

bFor POLY, the total human population corresponds to the total population in total polygon area

cThe human population density in mountains is calculated by dividing the human population in mountain terrain by the mountain area as defined by GMBA (Table 1) and as estimated within polygons (Table 2)

dThe total human population density is calculated by dividing the total human population in total polygon area by the total polygon area in each mega-region (Table 2)

At the 2.5′ resolution of our survey, about half of the people inhabiting mountains sensu GMBA are living in the entirely frost-free (tropical) lowest climatic belt (47% based on the GMBA definition and 44% based on POLY), roughly the other half (ca. 49% based on the GMBA definition and 50% based on POLY) are living in the lower (predominantly warm) montane belt, ca. 4 and 5% of all GMBA and POLY ‘mountain people’, respectively, are living in the uppermost climatic belt below the potential treeline (upper montane), and less than 1% are statistically expected to live in even colder rugged terrain (lower/upper alpine to nival), possibly at high latitude such as in North Siberia, Northern Scandinavia, and Alaska (Table 5). The latter percentage likely results from the inclusion of warmer terrain (e.g., mountain valleys with steep vertical gradients) in otherwise predominantly ‘cool’ (i.e., high-elevation) pixels, which happens as an artefact of using a 2.5′ grid resolution. There is an over-proportional fraction of human ‘mountain populations’ in the ca. 16% of all mountains of the terrestrial surface outside Antarctica that are not covered by the polygons (125 million people). These are located in low elevation rangeland (tropical and subtropical) that still matches the ruggedness criterion, but is too inconspicuous to be captured by a mountain polygon.
Table 5

Human population statistics for the world’s mountain terrain (GMBA definition and POLY, 2.5′ resolution) separated by major climatic belts

Climatic belt

Mountain area [103 km2]

Human population in mountain terrain [103]

(% of total populationa)

Human population densities in mountainsb

(totalc)

GMBA

POLY

GMBA

POLY

GMBA

POLY

Nival

526

461

236 (96.0)

236 (96.0)

0.4

0.5 (0.3)

Upper alpine

743

701

604 (71.3)

605 (72.5)

0.8

0.9 (0.8)

Lower alpine

2255

2147

2927 (59.2)

2900 (64.5)

1.3

1.4 (1.1)

Upper montane

3367

3071

18,485 (16.2)

17,809 (60.9)

5.5

5.8 (5.5)

Lower montane

3733

3132

132,800 (12.9)

113,500 (38.7)

35.6

36.2 (49.5)

Warm

1338

1041

107,830 (6.7)

81,890 (26.2)

80.6

78.7 (127)

No freezing

4473

3208

248,300 (5.5)

169,028 (30.3)

55.5

52.7 (87.0)

Total

16,435

13,760

511,164 (7.0)

385,960 (32.2)

31.1

28.0 (46.4)

aFor POLY, the total human population corresponds to the total population in total polygon area

bThe human population density in mountains is calculated by dividing the human population in mountain terrain by the mountain area

cThe total human population density in polygons is calculated by dividing the total human population in total polygon by the total area within polygons

When applying the Köppen–Geiger climate classification to all land outside Antarctica instead of the climatic belts described above, we estimate that ‘mountain people’ (sensu GMBA) represent 4.4% (133 million) of the 3 billion people living in equatorial low elevation life conditions, 9.2% (333 million) of the 3.6 billion people living in the warm temperate life zones with varying seasonality of moisture supply, and 5.5% (35million) of the 0.6 billion people living in cool temperate climates with cold winters (snow, Table 6). Densities vary from close to zero up to 77–114 km−2 in the warm temperate monsoon regions and are mostly below 20 people km−2 in the cooler parts of the world.
Table 6

Human population statistics for the world’s mountain terrain (GMBA definition) separated into climate types (acronyms in brackets) as defined by Köppen and Geiger (Kottek et al. 2006)

Köppen–Geiger climate type

Total land area [103 km2]

Mountain area [103 km2]

(% of total land area)

Total human population [103]

Human population in mountain terrain [103]

(% of total population)

Human population densities

Totala

Mountainb

Equatorial fully humid rainforest (Af)

6720

1030 (15.3)

327,000

34,800 (10.6)

48.7

33.8

Equatorial monsoon (Am)

4620

408 (8.8)

418,000

23,900 (5.7)

90.6

58.5

Equatorial savannah with dry summer (As)

1110

55.9 (5.1)

82,700

2680 (3.2)

74.8

48.0

Equatorial savannah with dry winter (Aw)

22,000

1070 (4.8)

1,870,000

61,300 (3.3)

85.1

57.5

Hot arid steppe (BSh)

7390

222 (3.0)

178,000

4500 (2.5)

24.1

20.3

Cold arid steppe (BSk)

4640

1080 (23.2)

43,600

2370 (5.4)

9.4

2.2

Hot arid desert (BWh)

8960

116 (1.3)

102,000

2090 (2.0)

11.4

18.1

Cold arid desert (BWk)

1000

102 (10.2)

6170

933 (15.1)

6.2

9.2

Warm temperate perpetually humid with hot summer (Cfa)

10,100

739 (7.3)

1,070,000

84,000 (7.9)

105

114

Warm temperate perpetually humid with warm summer (Cfb)

4770

1200 (25.1)

398,000

43,400 (10.9)

83.4

36.1

Warm temperate perpetually humid with cool summer and cold winter (Cfc)

215

159 (73.8)

1780

1250 (70.4)

8.3

7.9

Warm temperate with dry and hot summer (Csa)

4490

657 (14.7)

425,000

34,300 (8.1)

94.8

52.2

Warm temperate with dry and warm summer (Csb)

1370

452 (32.9)

72,200

12,500 (17.3)

52.6

27.6

Warm temperate with dry and cool summer and cold winter (Csc)

48.3

34.1 (70.5)

46.3

39.2 (84.7)

1.0

1.2

Warm temperate with dry winter and hot summer (Cwa)

8130

1040 (12.8)

1,380,000

79,900 (5.8)

169

76.9

Warm temperate with dry winter and warm summer (Cwb)

2250

1120 (49.7)

248,000

74,800 (30.2)

110

66.9

Warm temperate with dry winter and cool summer and cold winter (Cwc)

124

79 (63.6)

5290

3250 (61.6)

42.6

41.2

Snow climate perpetually humid with hot summer (Dfa)

2120

18.9 (0.9)

67,400

362 (0.5)

31.8

19.1

Snow climate perpetually humid with warm summer (Dfb)

8730

571 (6.5)

215,000

8560 (4.0)

24.6

15.0

Snow climate perpetually humid with cool summer and cold winter (Dfc)

11,700

1280 (10.9)

13,300

1840 (13.8)

1.1

1.4

Snow climate perpetually humid, extremely continental (Dfd)

634

26 (4.1)

316

0.73 (0.2)

0.5

0

Snow climate with dry and hot summer (Dsa)

207

37.4 (18.1)

9570

964 (10.1)

46.3

25.8

Snow climate with dry and warm summer (Dsb)

465

250 (53.8)

8480

3620 (42.7)

18.2

14.5

Snow climate with dry and cool summer and cold winter (Dsc)

63.1

51.3 (81.2)

124

111 (89.7)

2

2.2

Snow climate with dry winter and hot summer (Dwa)

985

68.3 (6.9)

262,000

8310 (3.2)

266

122

Snow climate with dry winter and warm summer (Dwb)

1710

266 (15.6)

56,900

8820 (15.5)

33.4

33.2

Snow climate with dry winter and cool summer and cold winter (Dwc)

2650

906 (34.1)

9710

2840 (29.3)

3.7

3.1

Snow climate with dry winter, extremely continental (Dwd)

99.2

24.3 (24.5)

15.4

1.14 (7.4)

0.2

0

Bare ground with very short snow-free period (EFb)

1360

328 (24.2)

75.4

66.2 (87.8)

0.1

0.2

Cold and dry desert (EFp)

6.49

5.45 (83.7)

38.1

37.9 (99.6)

5.9

7.0

Perpetual snow or ice (EFp)

2950

494 (16.7)

236

226 (96)

0.1

0.5

Tundra climate with drought in snow-free period (ETd)

3220

486 (15.1)

28,600

7560 (26.4)

8.9

15.5

Humid tundra climate (ETf)

8860

2070 (23.4)

3150

1910 (60.7)

0.4

0.9

Total

134,000

16,400 (12.3)

7,300,000

511,000

54.6

31.6

aThe total human population density is calculated by dividing the total human population by the total land area

bThe human population density in mountains is calculated by dividing the human population in mountain terrain by the mountain area as defined by GMBA

Discussion

Based on a quantitative comparison of mountain definitions that include ruggedness, elevation, and climate, we argue that the combination of ruggedness and climate is the most pertinent approach for biological questions, including biodiversity assessments and forest inventories. The mountain inventory we then provide based on this approach facilitates the access to regionalized mountain statistics.

Global mountain statistics

Both GIS-based approaches (WCMC and GMBA) rely on specific assumptions and serve as ‘conventions’ that can help in establishing standardized protocols. The GMBA 200 m ruggedness threshold across the 9 neighbouring 30″ grid points for instance resulted from test runs, exploring the smallest elevation amplitude that still captures what might be considered a mountain rather than a hill. Tests runs for the Alps revealed that a 200 m ruggedness threshold causes a clear distinction between the Alps and the forelands, with almost all narrow, low elevation valleys within the Alps still belonging to the mountain category (because the valley floors are commonly <2 km wide). In Switzerland in general, the 200 m threshold allows for the Swiss plateau or Swiss midlands, with all major cities and cropland, to fall outside the GMBA mountain definition. By the WCMC criteria, this same area largely falls into the mountain category, as do all plateaus at high elevation (e.g., Tibetan Plateau, yellow area in Fig. 2, bottom left), whereas most coastal mountains are disregarded (e.g., Norway, red area in Fig. 2, top right). Hence, the major part of the discrepancy between the WCMC and GMBA approach is located far below 1000 m asl in the undulating, low elevation hill country, outside the mountains, with a peak near 300 m asl, and does presumably not result from the inclusion or exclusion of the high plateaus (the small hump near 5000 m asl in Fig. 3), because the really flat parts of those plateaus are quite small. These distinctions are important in view of the human population statistics, as many people considered to be ‘mountain people’ in the WCMC approach (FAO 2015) are living on land that causes a divergence in the WCMC and the GMBA estimates at very low elevation (see examples in Fig. 2).

It is worth noting that a very large fraction (the majority) of the world’s mountains falls in a low elevation category by all three mountain concepts, with comparatively high temperatures, particularly at low latitude. Furthermore, neither of the applied definitions uses low temperatures alone as a criterion for mountains, thereby preventing the attribution of vast Arctic and Antarctic lowland terrain to mountainous areas.

Mountain polygons

In contrast to a grid-based assessment of all terrestrial area, mountain polygons cannot provide an exhaustive representation of mountainous terrain. Mountain polygons are arbitrary shapes that inevitably include some low elevation or non-mountainous terrain and fail to account for certain (including isolated) mountains that would be recognized by a grid-based statistical procedure. Hence, only the mountainous terrain within polygons should be used to calculate mountain statistics, and not the total polygon area. With a coverage of approximately 13.8 Mio km2, the polygons capture most of the 16.5 Mio km2 of global mountain terrain estimated with the GMBA approach. This suggests that individual polygons and our inventory as a whole capture the existing mountain terrain sufficiently well for locating mountains and mountain ranges and for comparative purposes. The missing 2.7 Mio km2 mostly consists of mountain terrain that is too scattered to be included in discrete polygons. Many of the missing, smaller mountains would probably not appear as distinct mountains on a physical world map and may also remain unnamed even though the absence of name is not always a matter of size. Larger rugged areas, for which no name could be found in atlases or online, were identified with codes for the purpose of this work. The largest discrepancies between GMBA and POLY were observed in Africa and Australia and may partly be attributed to the greater difficulties of delineating mountain terrain in areas where rugged terrain is more scattered or less conspicuous such as on the old Gondwana land shields.

The 1003 polygons of this first inventory are neither perfect nor comprehensive and refinements, including the use of a higher resolution Digital Elevation Model, will be needed in subsequent releases to achieve better coverage and higher accuracy. The three major criticisms to this first release are (1) the absence of some specific mountains or ranges, (2) the lumping of certain mountains, and (3) the inclusion of non-rugged terrain in individual polygons. The latter is intrinsic to the polygon delineation concept but does not affect the area of mountain terrain within polygons, which is strictly defined by GIS-based algorithms. The second criticism on lumping individual mountains is the most challenging to address, as the arguments for and against doing so can be numerous and controversial. User feedback will help improve this first version of the GMBA mountain inventory to possibly overcome the shortcomings that might result from the decisions we made in the delineation process.

Climatological stratification of mountain terrain

Since from a climatic point of view, it makes a considerable difference whether a mountain defined by ruggedness is in the Arctic, the temperate zone, or near the equator, a climatic stratification of land area is essential to identify ecologically comparable land cover units across the mountains of the world. Since available climatic layers (e.g., from WORLDCLIM) can be overlaid onto mountain topography, categorizations into any climatic belt are possible. For a climatic stratification to make sense and become globally useful, it should ideally capture conditions that reflect established biogeographic categories such as terrain above (alpine) and below (montane) the climatic treeline. As shown previously, the potential high-elevation tree limit is found at a globally common isotherm for the growing season (Körner 2012; Körner and Paulsen 2004; Körner et al. 2011) and, therefore, offers a bioclimatic reference, against which all other climatic belts in mountains can be positioned. Measures of temperature that include the dormant period in extratropical mountains, such as mean annual temperature (Jobbágy and Jackson 2000) or the mean temperature for the warmest month only—the often quoted 10 °C July isotherm for the northern temperate zone (Köppen 1919; Jarvis et al. 1989; Ohsawa 1990; Malyshev 1993), may fit either by chance or for certain latitudes or regions, but they do not capture global patterns (Hardy et al. 1998; Körner 1998).

Geophysical mountain statistics thus need to address two different issues: the first one is related to the structure of the elevated land surface, its ruggedness, and its exposure to the forces of gravity (the topographic nature of mountains), and the other one is related to the climatic life conditions (the climatic nature of mountains). At a given elevation, topography is independent of latitude, whereas climate is not. Topographically similar mountain terrain can be found in the Arctic and at the equator. Hence, a mountain concept that can be used for global comparisons must combine topography with a coherent climatological characterization as either one alone is insufficient and absolute elevation (m asl) alone is not ecologically meaningful either. Yet, there is a caveat, in that climatic data obtained from weather stations (2 m air temperature) do not represent the actual climatic conditions experienced within low stature vegetation and by soil biota (Scherrer and Körner 2011). Hence, assessing the actual life conditions of small plants, ground dwelling animals, and soil microbes will continue to require ground truth data.

Human populations in mountains

Given the significance of land use for biodiversity, it is interesting to estimate to which extent mountains are inhabited by humans. Using the combined ruggedness and climatic stratification of the world’s mountains (sensu GMBA), we estimated human population densities in specific climatic belts within polygons and globally. Among the various observations we made, three were particularly surprising: (1) the comparatively high human population density in mountainous areas, given that much of the rugged terrain is very steep, very high, forested, or represents dangerous terrain; (2) the 53 million inhabitants in the relatively small mountain areas within polygons of Africa, indicating an over-proportional significance of mountain livelihoods; and (3) the approximately 23 million people inhabiting mountains in Europe, which indicate that the immediate forelands and the interior valleys are densely populated. Taken together, the human population data clearly evidence that by far the majority of ‘mountain people’ are inhabiting low elevation rugged terrain in tropical and warm temperate regions, and face challenges associated with steep slopes rather than with low temperatures. Above the inhabited areas, land use might nevertheless be very intense.

Because the fraction of mountainous land differs between mountain definitions, human population estimates for specific regions differ as well. Accordingly, the values we present differ substantially from estimates provided elsewhere (e.g., EEA report 2010; FAO 2015), which are based on the WCMC definition of mountains or variants of it. However, based on the estimates we obtain for human populations in Swiss mountains for example, we believe that the GMBA definition of mountains leads to meaningful results. Additional variation in the reported continent-based estimates may come from methodological assumptions specific to our calculations and from the division of the world we apply, which differs from the United Nations Statistics Division “M.49” standard used in FAO 2015 primarily, in that Australia is kept separate from Oceania, and Greenland is kept separate from North America.

In our opinion, the availability of human population data and the resulting climate co-defined livelihood statistics are a goldmine for the mountain biology and ecology community. These data open up the avenue for many interesting socio-ecological research questions, including the estimation of ecologically important regional demands for natural resources such as fire wood or grazing land, of current and future pressure on medicinal plants, and of pressure from hunting and tourism. Known caveats associated with these data include the fact that (1) the global statistics ignore political boundaries, which means that a regional match with national inventory data is not possible, (2) the 2.5′ grid (4.6 × 4.6 km at the equator) is quite coarse, which means that any rugged 2.5′ pixel can, and often does, include non-rugged land fractions suitable for settlement; (3) the data cannot reliably capture land use intensity, particularly pastoralism, which is not necessarily directly correlated with population densities; and (4) the data are likely imprecise in specific locations when none of the data sources is accurate. Specific drawbacks of the light-emission information used in generating the LandScan include their low reliability in certain locations such as high-elevation mountain resorts, where light emissions are often disproportionally high compared to the number of inhabitants. These caveats remain despite regional adjustments but the data still allow the estimation of potential human impact on mountain ecosystems and the identification of regions of high and low risk for biodiversity losses.

Conclusions

The field of mountain biodiversity research has long been suffering from the lack of a standardized inventory of the mountains of the world for robust comparisons of biogeographic characteristics and biodiversity data across mountain ranges and biomes. Here, we offer such an inventory in which each of the 1003 entries corresponds to a polygon drawn around a mountain or a mountain range sensu GMBA. With a total of 13.8 Mio km2 of mountain terrain, the polygons cover 83.7% of the global mountain area estimated using the GMBA approach, which is a purely grid-based (geostatistical) survey of mountains on the terrestrial surface outside of Antarctica. With this coverage, we are confident that the inventory is representative and offers a useful tool for biodiversity researchers. However, to improve its representativeness and accuracy beyond applying a higher resolution Digital Elevation Model, we encourage users to explore and manipulate the GIS files available online (https://ilias.unibe.ch/goto_ilias3_unibe_cat_1000515.html) and provide their feedback and suggestions in view of the online release of subsequent versions. By offering for each polygon a subdivision of the mountainous terrain into climatic belts above and below the potential bioclimatic tree limit, we enable biodiversity experts to place their assessments in a meaningful climatological context, irrespective of elevation and latitude. The human population density data in turn are a useful resource for biodiversity risk assessment and conservation planning.

The polygon-based mountain inventory of the world described here is a unique information layer that is now available on the online ‘Mountain Portal’ of GMBA (www.mountainbiodiversity.org). Through collaboration with the Map of Life (http://mol.org) and EarthEnv projects (http://www.earthenv.org), the mountain polygon information can already be visualized together with an increasingly detailed set of additional topographic variable and a growing range of biodiversity information provided through Map of Life. Exploring the biodiversity information at the level of individual climate belt within polygons will become possible in the near future.

Defining mountains and life zones within mountains will remain a challenging task, given that different applications call for different definitions and assumptions. Here, we suggest that the most useful approach for biological questions, including biodiversity assessments and forest inventories, is the combination of a ruggedness term with a bioclimatic stratification. We further argue against approaches based on elevation above sea level, as life is driven by climate rather than by elevation. Organisms, including humans, that live at identical elevations can indeed be exposed to rather different climates across the globe. We hope that the mountain inventory we present and the additional information on climate and human populations we provide can serve as a reference for biodiversity assessment and comparative ecology of mountains.

Notes

Acknowledgements

We thank Alessia Vita and her team at the Mountain Partnership Secretariat of FAO for the provisioning of the FAO human population data and their great help with these data. We also thank Nadine Brinkmann for her help with delineating the polygons during her GMBA internship, Susanna Riedl for her contribution in illustrating results, and the Swiss National Science Foundation for financial support to GMBA (Grants 31FI30_159677, 31FI30_152940, and 31FI30_135726/1).

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Declaration of authorship

CK and ES initiated the comparisons of mountain definitions. KR was responsible for the polygon delineation. JP processed the polygons and population data in GIS and applied the various mountain definitions and climatic belt modeling. WJ established the conceptual links to the Map of Life project. CK drafted the text, with substantial contributions by KR, ES, and DP. All authors contributed to final editing.

Supplementary material

35_2016_182_MOESM1_ESM.xlsx (152 kb)
Supplementary material 1 (XLSX 152 kb)

References

  1. Barthlott W, Mutke J, Rafiqpoor D, Kier G, Kreft H (2005) Global centers of vascular plant diversity. Nova Act Lc 92(342):61–83Google Scholar
  2. Blyth S, Groombridge B, Lysenko I, Miles L, Newton A (2002) Mountain watch: environmental change and sustainable development in mountains. UNEP-WCMC, CambridgeGoogle Scholar
  3. Brown JH (1971) Mammals on mountaintops: non-equilibrium insular biogeography. Am Nat 105:467. doi:10.1086/282738 CrossRefGoogle Scholar
  4. Byers AC, Price LW, Price MF (2013) An introduction to mountains. In: Price MF, Byers AC, Friend DA, Kohler T, Price LW (eds) Mountain geography: physical and human dimensions. University of California Press, Berkeley, pp 1–10Google Scholar
  5. Chape S, Spalding MD, Jenkins MD (eds) (2008) The world’s protected areas. UNEP-World Conservation Monitoring Centre, CambridgeGoogle Scholar
  6. Dahl AL (1991) Island directory. In: UNEP regional seas directories and bibliographies 35: 573Google Scholar
  7. Egan P, Price MF (2016) Mountain ecosystem services and climate change—a global overview of potential threats and strategies for adaptation. UNESCO, ParisGoogle Scholar
  8. European Environmental Agency (2010) Europe’s ecological backbone: recognizing the true value of our mountains. EEA Report 6Google Scholar
  9. FAO (2015) Mapping the vulnerability of mountain peoples to food insecurity. In: Romeo R, Vita A, Testolin R, Hofer T (eds), RomeGoogle Scholar
  10. Green K, Stein JA (2015) Modeling the thermal zones and biodiversity on the high mountains of Meganesia: the importance of local differences. Arct Antarct Alp Res 47:671–680. doi:10.1657/AAAR0014-083 CrossRefGoogle Scholar
  11. Gurung H (1999) Mountains of Asia—a regional inventory. International Centre for Integrated Mountain Development, KathmanduGoogle Scholar
  12. Hamilton LS (2002) Why mountains matter. World Conserv (IUCN) 33–1:4–5Google Scholar
  13. Hardy DR, Vuille M, Braun C, Keimig F, Bradley RS (1998) Annual and daily meteorological cycles at high altitude on a tropical mountain. Bull Am Meteorol Soc 79:1899–1913. doi:10.1175/1520-0477(1998)079<1899:AADMCA>2.0.CO;2
  14. Harmon D, Worboys GL (2004) Managing mountain protected areas: challenges and responses for the 21st century. Himadoc, ICIMOD, Kathmandu. doi:10.1659/0276-4741%282005%29025%5B0387%3AMMPACA%5D2.0.CO%3B2
  15. Huddleston B, Ataman E, de Salvo P, Zanetti M, Bloise M, Bel J, Franceschini G, Fe d’Ostiani L (2003) Towards a GIS-based analysis of mountain environments and populations. FAO, RomeGoogle Scholar
  16. Janzen DH (1967) Why mountain passes are higher in the tropics. Am Nat 101:233–249CrossRefGoogle Scholar
  17. Jarvis PG, Grace J, Huntchings N, Monteith JL, Shuttleworth WJ, Fowler D, Corlett J, Thomas J, Grace J (1989) Tree lines. Philos T Roy Soc B 324(1223):233–245. doi:10.1098/rstb.1989.0046 CrossRefGoogle Scholar
  18. Jetz W, McPherson J, Guralnick RP (2012) Integrating biodiversity distribution knowledge: toward a global map of life. Trends Ecol Evol 27(3):151–159. doi:10.1016/j.tree.2011.09.007 CrossRefPubMedGoogle Scholar
  19. Jobbágy EG, Jackson RB (2000) Global controls of forest line elevation in the northern and southern hemispheres. Global Ecol Biogeogr 9:253–268. doi:10.1046/j.1365-2699.2000.00162.x CrossRefGoogle Scholar
  20. Kapos V, Rhind J, Edwards M, Price MF, Ravilious C (2000) Developing a map of the world’s mountain forests. In: Price MF, Butt N (eds) Forests in sustainable mountain development: a report for 2000. CAB International, Wallingford, pp 4–9. doi:10.1007/1-4020-3508-X_52 CrossRefGoogle Scholar
  21. Knaurs (1999). Knaurs grosser Weltatlas, AugsburgGoogle Scholar
  22. Köppen W (1919) Baumgrenze und Lufttemperatur. Petermanns Geogr Mitt 65:201–203Google Scholar
  23. Körner C (1998) A re-assessment of high elevation treeline positions and their explanation. Oecologia 115:445–459. doi:10.1007/s004420050540 CrossRefGoogle Scholar
  24. Körner C (2000) Why are there global gradients in species richness? Mountains might hold the answer. Trends Ecol Evol 15:513–514. doi:10.1016/S0169-5347(00)02004-8 CrossRefGoogle Scholar
  25. Körner C (2001) One of nature’s most innovative laboratories. Trends Plant Sci 6:7–8. doi:10.1016/S1360-1385(00)01832-X CrossRefPubMedGoogle Scholar
  26. Körner C (2004) Mountain biodiversity, its causes and function. Ambio, Special Report 13:11–17. doi:10.1111/j.1365-2699.2003.01043.x
  27. Körner C (2012) Global mountain statistics based on treeline elevation. In: Körner C (ed) Alpine treelines. Springer, Basel, pp 57–62. doi:10.1007/978-3-0348-0396-0 CrossRefGoogle Scholar
  28. Körner C, Ohsawa M (2006) Mountain systems. In: Hassan R, Scholes R, Ash N (eds) Ecosystems and human well-being: current state and trends, 1. Island Press, Washington, DC, pp 681–716Google Scholar
  29. Körner C, Paulsen J (2004) A world-wide study of high altitude treeline temperatures. J Biogeogr 31:713–732. doi:10.1111/j.1365-2699.2003.01043.x CrossRefGoogle Scholar
  30. Körner C, Paulsen J, Spehn EM (2011) A definition of mountains and their bioclimatic belts for global comparisons of biodiversity data. Alp Bot 121:73–78. doi:10.1007/s00035-011-0094-4 (publication), doi:10.7892/boris.83486 (data)
  31. Kottek M, Grieser J, Beck C, Rudolf B, Rubel F (2006) World map of the Köppen–Geiger climate classification updated. Meteorol Z 15(3):259–263. doi:10.1127/0941-2948/2006/0130 CrossRefGoogle Scholar
  32. La Sorte FA, Jetz W (2010) Projected range contractions of montane biodiversity under global warming. Proc R Soc B Biol Sci 277:3401–3410. doi:10.1098/rspb.2010.0612 CrossRefGoogle Scholar
  33. Malyshev L (1993) Levels of the upper forest boundary in northern Asia. Vegetatio 109:175–186. doi:10.1007/BF00044749 CrossRefGoogle Scholar
  34. McVicar TR, Körner C (2012) On the use of elevation, altitude, and height in the ecological and climatological literature. Oecologia 171:335–337. doi:10.1007/s00442-012-2416-7 CrossRefPubMedGoogle Scholar
  35. Messerli B, Ives JD (1997) Mountains of the world: a global priority. Parthenon, New York. doi:10.1002/(SICI)1099-145X(200003/04)11:2<197::AID-LDR390>3.0.CO;2-U
  36. Ohsawa M (1990) An interpretation of latitudinal patterns of forest limits in South and East-Asian mountains. J Ecol 78:326–339. doi:10.2307/2261115 CrossRefGoogle Scholar
  37. Pauli H, Gottfried M, Dullinger S, Abdaladze O, Akhalkatsi M, Benito Alonso JL, Coldea G, Dick J, Erschbamer B, Fernandez Calzado R, Ghosn D, Holten JI, Kanka R, Kazakis G, Kollar J, Larsson P, Moiseev P, Moiseev D, Molau U, Mesa Molero, Nagy L, Pelino G, Puscas M, Rossi G, Stanisci A, Syverhuset AO, Theurilla JP, Tomaselli M, Unterluggauer P, Villar L, Vittoz P, Grabherr G (2012) Recent plant diversity changes on Europe’s mountain summits. Science 226(6079):353–355. doi:10.1126/science.1219033 CrossRefGoogle Scholar
  38. Paulsen J, Körner C (2014) A climate-based model to predict potential treeline position around the globe. Alp Bot 124:1–12. doi:10.1007/s00035-014-0124-0 CrossRefGoogle Scholar
  39. Scherrer D, Körner C (2011) Topographically controlled thermal-habitat differentiation buffers alpine plant diversity against climate warming. J Biogeogr 38:406–416. doi:10.1111/j.1365-2699.2010.02407.x CrossRefGoogle Scholar
  40. Smith B, Mark D (2003) Do mountains exist? Towards an ontology of landforms. Environ Plan B 30(3):411–427. doi:10.1068/b12821 CrossRefGoogle Scholar
  41. Spehn E, Liberman M, Körner C (eds) (2006) Land use change and mountain biodiversity. CRC Publishers, Boca RatonGoogle Scholar
  42. Tang Z, Wang Z, Zheng C, Fang J (2006) Biodiversity in China’s mountains. Front Ecol Environ 4(7):347–352. doi:10.1890/1540-9295(2006)004[0347:BICM]2.0.CO;2
  43. Vetaas OR, Grytnes JA (2002) Distribution of vascular plant species richness and endemic richness along the Himalayan elevation gradient in Nepal. Global Ecol Biogeogr 11(4):291–301. doi:10.1046/j.1466-822X.2002.00297.x CrossRefGoogle Scholar
  44. Weigelt P, Jetz W, Kreft H (2013) Bioclimatic and physical characterization of the world’s islands. Proc Nat Acad Sci 110:15307–15312. doi:10.1073/pnas.1306309110 CrossRefPubMedPubMedCentralGoogle Scholar
  45. Weigelt P, Steinbauer MJ, Sarmento Cabral J, Kreft H (2016) Late Quaternary climate change shapes island biodiversity. Nature 532:99–102. doi:10.1038/nature17443 CrossRefPubMedGoogle Scholar

Copyright information

© The Author(s) 2016

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors and Affiliations

  • Christian Körner
    • 1
  • Walter Jetz
    • 2
    • 3
  • Jens Paulsen
    • 1
  • Davnah Payne
    • 4
  • Katrin Rudmann-Maurer
    • 1
  • Eva M. Spehn
    • 1
    • 4
  1. 1.Department of Environmental Sciences, Institute of BotanyUniversity of BaselBaselSwitzerland
  2. 2.Ecology and Evolutionary BiologyYale UniversityNew HavenUSA
  3. 3.Department of Life SciencesImperial College LondonBerkshireUK
  4. 4.Global Mountain Biodiversity Assessment, Institute of Plant SciencesUniversity of BernBernSwitzerland

Personalised recommendations