Hostname: page-component-7c8c6479df-27gpq Total loading time: 0 Render date: 2024-03-18T03:32:57.044Z Has data issue: false hasContentIssue false

Delimiting tropical mountain ecoregions for conservation

Published online by Cambridge University Press:  12 May 2011

PHILIP J. PLATTS*
Affiliation:
York Institute for Tropical Ecosystem Dynamics (KITE), Environment Department, University of York, Heslington, York YO10 5DD, UK Conservation Science Group, Zoology Department, University of Cambridge, Cambridge CB2 3EJ, UK
NEIL D. BURGESS
Affiliation:
Centre for Macroecology, Evolution and Climate, Department of Biology, Copenhagen University, Universitetsparken 15, DK-2100, Copenhagen, Denmark and WWF-US Conservation Science Program, 1259 24th Street NW, Washington DC, USA
ROY E. GEREAU
Affiliation:
Missouri Botanical Garden, PO Box 299, Saint Louis, MO 63166-0299, USA
JON C. LOVETT
Affiliation:
Twente Centre for Studies in Technology and Sustainable Development (CSTM), University of Twente, Postbus 217, 7500 AE, Enschede, Netherlands
ANDREW R. MARSHALL
Affiliation:
York Institute for Tropical Ecosystem Dynamics (KITE), Environment Department, University of York, Heslington, York YO10 5DD, UK Flamingo Land Ltd, Kirby Misperton, Malton, North Yorkshire YO17 6UX, UK
COLIN J. McCLEAN
Affiliation:
York Institute for Tropical Ecosystem Dynamics (KITE), Environment Department, University of York, Heslington, York YO10 5DD, UK
PETRI K.E. PELLIKKA
Affiliation:
Department of Geography, University of Helsinki, PO Box 64, 00014 Helsinki, Finland
RUTH D. SWETNAM
Affiliation:
Conservation Science Group, Zoology Department, University of Cambridge, Cambridge CB2 3EJ, UK
ROB MARCHANT
Affiliation:
York Institute for Tropical Ecosystem Dynamics (KITE), Environment Department, University of York, Heslington, York YO10 5DD, UK
*
*Correspondence: Philip Platts Tel: +44 1904 434780 e-mail: philip.platts@rocketmail.com
Rights & Permissions [Opens in a new window]

Summary

Ecological regions aggregate habitats with similar biophysical characteristics within well-defined boundaries, providing spatially consistent platforms for monitoring, managing and forecasting the health of interrelated ecosystems. A major obstacle to the implementation of this approach is imprecise and inconsistent boundary placement. For globally important mountain regions such as the Eastern Arc (Tanzania and Kenya), where qualitative definitions of biophysical affinity are well established, rule-based methods for landform classification provide a straightforward solution to ambiguities in region extent. The method presented in this paper encompasses the majority of both contemporary and estimated preclearance forest cover within strict topographical limits. Many of the species here tentatively considered ‘near-endemic’ could be reclassified as strictly endemic according to the derived boundaries. LandScan and census data show population density inside the ecoregion to be higher than in rural lowlands, and lowland settlement to be most probable within 30 km. This definition should help to align landscape scale conservation strategies in the Eastern Arc and promote new research in areas of predicted, but as yet undocumented, biological importance. Similar methods could work well in other regions where mountain extent is poorly resolved. Spatial data accompany the online version of this article.

Type
Papers
Copyright
Copyright © Foundation for Environmental Conservation 2011

INTRODUCTION

Ecological regions (ecoregions) are widely employed as units for conservation priority setting (Olson & Dinerstein Reference Olson and Dinerstein2002; Burgess et al. Reference Burgess, Hales, Ricketts and Dinerstein2006). By aggregating ecosystems with similar biophysical characteristics, interdependences and spheres of influence within ecoregions, assessment and management can target the system as a whole and avoid any disconnect between science and policy. This holistic approach is more faithful to environment-ecosystem interconnections, including the role of human populations, but by definition imposes strict and static divisions on complex and ever changing landscapes. Careful delineation is therefore crucial. Landform, geology, climate, vegetation and evolutionary history, as well as cultural and political considerations, are all important (McMahon et al. Reference McMahon, Wiken and Gauthier2004). In practice, different criteria favour different objectives and no single solution is optimal for all applications in all regions (Olson et al. Reference Olson, Dinerstein, Wikramanayake, Burgess, Powell, Underwood, D'Amico, Itoua, Strand, Morrison, Loucks, Allnutt, Ricketts, Kura, Lamoreux, Wettengel, Hedao and Kassem2001).

Some biogeographical units such as mountains, wetlands and islands appear by their nature to be clearly defined in space. In the case of mountains, however, there is no universally accepted method for marking the transition to lowlands, and thus no consistent way of defining precisely the geographical limits of a mountainous region (Gerrard Reference Gerrard1990). The essence of the problem, as in any ecoregion, is that environmental gradients are continuous (from sea level to mountain top) and so any spatial dichotomy is necessarily subjective. Freely available digital elevation data, together with improvements in desktop mapping software, have brought advances in the development of a systematic process by which to define and study mountains. Practitioners can now experiment with different ways of bounding their region of interest, so that qualitative definitions of biological affinity may where appropriate progress to spatially explicit rule-based algorithms, increasing the prominence of such regions and their specific management challenges on the political stage (Browne et al. Reference Browne, Fox and Funnell2004; McMahon et al. Reference McMahon, Wiken and Gauthier2004).

As assessment mechanisms for biodiversity and ecosystem services move towards implementation within international frameworks (Larigauderie & Mooney Reference Larigauderie and Mooney2010), there is a requirement for rule-based definitions that can resolve ambiguities in the placement of ecoregion boundaries. In the present study, using the Eastern Arc of Tanzania and Kenya as an example, we discuss some of the challenges and uncertainties in defining tropical mountain ecoregions. This chain of 13 block-faulted massifs harbours one of the world's most important concentrations of biodiversity across a series of fragile sites (Brooks et al. Reference Brooks, Mittermeier, Mittermeier, da Fonseca, Rylands, Konstant, Flick, Pilgrim, Oldfield, Magin and Hilton-Taylor2002; Mittermeier et al. Reference Mittermeier, Robles Gil, Hoffmann, Pilgrim, Brooks, Mittermeier, Lamoreux and da Fonseca2004). The ‘Eastern Arc Mountains [sic] Forests’ have recently been proposed for UNESCO World Heritage status (http://whc.unesco.org/), and will soon be subject to increased international attention as the United Nations REDD Programme (Reducing Emission from Deforestation and Forest Degradation in Developing Countries) is piloted in Tanzania (Burgess et al. Reference Burgess, Bahane, Clairs, Danielsen, Dalsgaard, Funder, Hagelberg, Harrison, Haule, Kabalimu, Kilahama, Kilawe, Lewis, Lovett, Lyatuu, Marshall, Meshack, Miles, Milledge, Munishi, Nashanda, Shirima, Swetnam, Willcock, Williams and Zahabu2010). We propose for the first time a rigorous set of topographical limits for these mountain habitats, using as a starting point the global mountain topology developed by the UNEP World Conservation Monitoring Centre (UNEP-WCMC; Kapos et al. Reference Kapos, Rhind, Edwards, Price, Ravilious, Price and Butt2000). Increasing the spatial resolution from 1 km to 90 m, we rescale terrain parameters using empirical and remotely sensed data on the distribution of endemic plant species and mountain forests respectively, before bounding the ecoregion according to topographic prominence and established biogeographical distributions.

A qualitative definition for the Eastern Arc

The Eastern Arc Mountains were first described as distinct from surrounding Afromontane habitats in the 1980s due to the exceptional proportion of rare species (Lovett Reference Lovett1985). The distinction is qualitatively explained according to key environmental characteristics thought to underlie the high levels of endemism (Fig 1a): first, the great age of the Precambrian crystalline and metamorphic substrate compared to geologically more recent volcanoes such as Kilimanjaro and Meru (>30 million versus <2 million years old; Schlüter Reference Schlüter1997); and second, the relative consistency of rain-bearing winds from the Indian Ocean compared to more variable climatic regimes within the great lake catchments of Nyasa/Malawi, Tanganyika and Victoria. Accordingly, the chain extends roughly 750 km from the Taita Hills in south-east Kenya to the Udzungwa Mountains in south-central Tanzania; volcanoes to the north are excluded (different substrate), as are highlands south of the Makambako Gap (geologically younger and rainfall patterns not directly influenced by the Indian Ocean).

Figure 1 Global mountain typology defined by UNEP-WCMC at 1 km resolution. (a) East Africa, showing divisions in Tanzanian forest on the basis of geology and climate (from Lovett Reference Lovett1990): Coastal, Eastern Arc and Northern forests are all under the direct climatic influence of the Indian Ocean (rather than the Great Lakes), but only the Eastern Arc is on igneous and metamorphic rock. (b) Zoomed perspective of the 13 Eastern Arc ranges (blocs).

We identified the 13 most commonly-cited ranges (blocs; Fig. 1b). Sometimes omitted are: the Taita Hills (the only bloc north of the Kenya-Tanzania border; see for example Burgess et al. Reference Burgess, Mwakalila, Madoffe, Ricketts, Olwero, Swetnam, Mbilinyi, Marchant, Mtalo, White, Munishi, Marshall, Malimbwi, Jambiya, Fisher, Kajembe, Morse-Jones, Kulindwa, Green and Balmford2009); the Nguu (sometimes combined with Nguru; see Newmark Reference Newmark2002); and Malundwe Mountain, owing to its small extent and position within a chain of low elevation hills (see Stanley & Olson Reference Stanley and Olson2005). An occasional addition is Image, but this mountain is more usually included in Udzungwa (Mbilinyi et al. Reference Mbilinyi, Malimbwi, Shemwetta, Zahabu, Katani and Kashaigili2006).

Mountain limits

The lack of a systematic basis for boundary placement in the Eastern Arc has meant that the precise spatial extents depicted have varied from study to study, hindering information exchange and multilateral action in response to new data or changes on the ground. Mountain limits have been defined variously by hand, using elevational contours, forest or reserve boundaries, or a combination of these factors (for example see Olson et al. Reference Olson, Dinerstein, Wikramanayake, Burgess, Powell, Underwood, D'Amico, Itoua, Strand, Morrison, Loucks, Allnutt, Ricketts, Kura, Lamoreux, Wettengel, Hedao and Kassem2001; Doggart et al. Reference Doggart, Perkin, Kiure, Fjeldså, Poynton and Burgess2006). Many depictions can be traced back to Lovett (Reference Lovett1992): drawn from contours on a Shell Road Map, these boundaries were intended as diagrammatic and heuristic; however a number of derivatives are now in common usage, conspicuous in their repetition of shortfalls in the original. A 1000 m cut-off for the Taita Hills, for instance, depicts bloc-extent far to the west of the mountains proper, but excludes important sites for conservation in the east.

Elevation offers a simple, intuitive, but often insufficient method for delimiting mountain regions (Messerli & Ives Reference Messerli and Ives1997). With respect to plant endemism in the Eastern Arc, one convention is to impose a uniform lower limit of 500 m (see for example government reports and the World Heritage application), intended to distinguish montane habitats from the nearby coastal forests (Lovett et al. Reference Lovett, Hansen, Hørlyck, Burgess and Clarke2000). We find this definition to be too strict in the east, excluding patches of forest from the lower slopes of six mountain blocs (Platts et al. Reference Platts, Ahrends, Gereau, McClean, Lovett, Marshall, Pellikka, Mulligan, Fanning and Marchant2010). Conversely, because the East African interior is characterized by a highland plateau, almost all non-mountainous land on the western side exceeds elevations of 500 m. Similarly for herpetofauna, species turnover is highest at elevations of c. 400 m in the north-east, but c. 800 m in the south (Poynton et al. Reference Poynton, Loader, Sherratt and Clarke2007).

A global mountain typology

Effective mountain delineation across all aspects requires consideration of not just elevation, but also steepness of slope and terrain roughness (Gerrard Reference Gerrard1990). Implementing these criteria on a global scale, UNEP-WCMC developed a map of the world's mountains and mountain forests by classifying 1 km2 land parcels according to elevation, slope and local relief (Kapos et al. Reference Kapos, Rhind, Edwards, Price, Ravilious, Price and Butt2000; Blyth et al. Reference Blyth, Groombridge, Lysenko, Miles and Newton2002). The classification consists of six elevational bands, with terrain constraints strictest at low elevations (Fig. 1). Whilst this typology effectively represents the Eastern Arc on a world map, it is too broadly defined to identify all features important at the site level or to distinguish between adjacent ranges (for example South Pare versus West Usambara; Fig. 1b). Kapos et al. (Reference Kapos, Rhind, Edwards, Price, Ravilious, Price and Butt2000) suggested that for quantitative applications at sub-national scales their map could be verified and potentially refined using higher-resolution vegetation data, together with an appropriate measure of relative relief; we test these criteria in the context of the Eastern Arc Mountains.

MATERIALS AND METHODS

Regional refinements to the global typology

Elevational zonation

We adhered mainly to globally defined elevational classes (Fig. 1), with the exception of the lowest mountain category, which was applied without the 300 m lower limit. As in other regions where mountains occur along a coastline (see for example Nordregio 2004), marked topography and associated forest vegetation extend almost to sea level. For higher elevations, the hypsographic curve for eastern Tanzania and south-east Kenya (cumulative height versus relative area) exhibits similarities with the breakpoints defined by UNEP-WCMC: frequencies are highest for land at elevations of 1000–1500 m, with a further change in trend at elevations c. 2500 m. The unqualified inclusion of all land above 2500 m elevation is perhaps inappropriate for regions with large highland plateaus (see Meybeck et al. Reference Meybeck, Green and Vorosmarty2001), but is of little consequence here where only mountain summits exceed this elevation.

Terrain parameters

The spatial resolution of the entire typology was increased from 1 km to 90 m. Shuttle Radar Topography Mission elevation data (Jarvis et al. Reference Jarvis, Reuter, Nelson and Guevara2008) were extracted for the 56 half-degree grid squares that intersect with the Eastern Arc (Fig. 1) and resampled from a three arc-second geographic projection to a 90 m equal-area conic projection (Africa Albers). Slope constraints were increased from 5° to 10° (class 2) and from 2° to 5° (class 3) to correct for the greater topographic heterogeneity captured per unit area at this finer resolution.

At the global scale, the local elevation range parameter (hereafter, LERRADIUS) is intended to capture ‘older mountains of regional significance’ (Kapos et al. Reference Kapos, Rhind, Edwards, Price, Ravilious, Price and Butt2000). Eastern Arc blocs consist entirely of such old weathered slopes, but in some cases are relatively small and, particularly in the east, rise abruptly from surrounding plains. A parameterization of the LER better suited to regional-scale analysis was therefore sought. With a view to maximizing the inclusion of forest habitats and endemic plant records within a minimal mountain area, we tested LER radii in the range 250 m to 2 km and thresholds in the range 5% to 30% (of the radius). In exploratory analyses, larger radii led to excessive buffering around the base of the mountains or, with stricter thresholds, the omittance of outlying peaks; smaller radii returned patterns akin to the slope parameter, which is roughly equivalent to LER90m.

The final step in the derivation of each variant of this regional typology was to apply a spatial filter, such that grid squares adopted the majority mountain class within a 500 m radius. This resolved fine-scale anomalies in raster grids, aiding the transition to vector format.

Matching to mountain vegetation

The decision of which parameterization was best suited to the Eastern Arc was guided by overlaying high-resolution (1 ha) vegetation data. Indigenous broadleaved forests in the Taita bloc were identified from SPOT multi-spectral satellite images (Clark & Pellikka Reference Clark, Pellikka, Röder and Hill2009). Tanzanian forests were based on data acquired from the Tanzanian Ministry of Natural Resources and Tourism (MNRT 1997; updated with later imagery from 2000 onwards by the Remote Sensing and GIS Laboratory, Sokoine University of Agriculture, Tanzania). We extracted all forest classified as submontane, montane or upper-montane, and additionally considered any lowland forest contiguous with a submontane patch, thus removing the elevational bias of vegetation classification (see Pócs Reference Pócs1976; Lovett & Pócs Reference Lovett and Pócs1993; Hall et al. Reference Hall, Burgess, Lovett, Mbilinyi and Gereau2009). These estimates were then corrected according to Mbilinyi et al. (Reference Mbilinyi, Malimbwi, Shemwetta, Zahabu, Katani and Kashaigili2006), Conservation International (2008), Marshall et al. (Reference Marshall, Jorgensbye, Rovero, Platts, White and Lovett2010) and local knowledge (see for example http://celp.org.uk/projects/tzforeco/). In total, we identified a forest area of 4388 km2.

Given that c. 70% of forest cover has been lost since the beginning of the 20th century (S. Willcock, unpublished data 2010) and that forests on some eastern aspects might regenerate further downslope if it were not for land-use barriers and frequent burning, it was appropriate to also consider estimates of ‘palaeoecological’ extent (Hall et al. Reference Hall, Burgess, Lovett, Mbilinyi and Gereau2009; Tanzania only). Hereafter, these estimates are termed ‘preclearance’ rather than palaeoecological. Forest would have extended beyond the base of the mountains during interglacial periods (for example Holocene and Eemian) and consisted of different vegetation assemblages in cooler, potentially drier environments of lower CO2 concentration associated with the last glacial period (c. 10 000–114 000 yrs BP; Elenga et al. Reference Elenga, Peyron, Bonnefille, Jolly, Cheddadi, Guiot, Andrieu, Bottema, Buchet, de Beaulieu, Hamilton, Maley, Marchant, Perez-Obiol, Reille, Riollet, Scott, Straka, Taylor, Van Campo, Vincens, Laarif and Jonson2000; Vincens et al. Reference Vincens, Garcin and Buchet2007).

Lastly, we overlaid point distributions for 378 species of vascular plants strictly endemic to the Eastern Arc (according to a 500 m lower limit). Over 2000 spatially referenced herbarium specimen records were available (http://www.tropicos.org/), representing 77% of all strict-endemics and spanning all plant growth forms. To investigate the efficacy of the 500 m lower limit for endemism, and to avoid bias resulting from such a convention, we further considered the distributions of 53 ‘near-endemic’ plant species (455 records, some below 500 m), documented only within the half-degree cells (Fig. 1b) and whose known distributions do not extend to coastal forests, Neogene Volcanoes or the Lake Nyasa Highlands. Based on locality information supplied with these records, we assumed a spatial accuracy of one arc-minute (<2 km).

Bounding the chosen regional typology

Amalgamating classified features

First, adjacent mountain classes were dissolved such that each spatially distinct feature was represented by a single polygon. With the exception of small isolated fragments (<1 km2 and >1 km from nearest neighbour), these features were buffered by 1 km and then simplified and smoothed using bend reduction and polynomial approximation tools in ArcGIS Desktop 9.3.1 (exponential kernel, tolerance = 1 km). This method ensured the preservation of all mountainous features on a 1 km grid (a popular format for regional mapping) and yielded topologically simpler boundaries that additionally enclosed unclassified habitats benefiting from close proximity to complex relief. Following Blyth et al. (Reference Blyth, Groombridge, Lysenko, Miles and Newton2002), isolated inner basins and plateaus were filled if less than 25 km2.

Mountain selection based on relative relief

The distinction between a mountain and a hill is largely semantic, with equivalent relative relief considered mountainous in one region whilst merely hilly in another (OED [Oxford English Dictionary] 1989). It was however appropriate to make such a distinction in the Eastern Arc so that minor relief not commonly perceived as mountainous could be systematically removed. Thompson (Reference Thompson1964) suggested a topographic prominence of 2000 ft to be a good rule of thumb, which we rounded down to the metric equivalent of 600 m. A cluster of features, bounded as above, was considered ‘mountainous’ if attaining an altitude of at least 600 m relative to adjacent elevations; ‘hilly’ clusters (<600 m prominence) were included in the ecoregion only if known to support natural forest vegetation and/or strictly endemic plant species.

RESULTS

Sensitivity to local elevation range

The biggest changes in overall mountain area resulted from varying the LER threshold according to fixed percentages of its defining radius (Table 1). The lowest threshold (5%) was too permissive, merging adjacent range-boundaries and including minor undulations throughout the lowlands (Fig. 2). Higher thresholds were more effective in terms of forest and endemic plant records captured per unit area, but if too high (30%) then lower slopes and mountain foothills were curtailed. A 20% threshold provided the best compromise between forest inclusion and range distinction. With the threshold percentage held constant, changing the LER radius affected polygon complexity but had little impact on overall mountain area. Small radii tracked closely fine-scale changes in relief; large radii better represented broad-scale trends in the landscape. Both LER500m and LER1km worked well, but we preferred the former because slightly more forest and endemism records were included, and because the latter excluded lower slopes in south-west Mahenge (inspection of three-dimensional elevation).

Table 1 Mountain area and forest sites included as a result of different calibrations of the local elevation range parameter (LER radius/threshold, mountain classes 1–2). Analyses conducted within Eastern Arc half-degree grid squares using 90 m elevation data and slope thresholds of 5° (class 2) and 10° (class 3); otherwise mountain classification follows UNEP-WCMC, but with no lower limit in class 1. *Mountain typology from which Eastern Arc boundaries were derived (Appendix 2, available online at http://www.journals.cambridge.org/enc2011001).

Figure 2 Sensitivity of the regional mountain typology to local elevation range (LER, radius = 500 m). Upper pane: elevation in the Uluguru bloc (×5 vertical exaggeration). Lower pane: variations in extent resulting from different LER thresholds (% of radius).

At this level (LER500m, 20% threshold), 94% of the estimated forest area (both present-day and preclearance) was classified as mountainous. The remaining 6% was largely within 1 km of a classified feature and so would be captured by the ecoregion boundary (Table 1). One exception was open-canopy forest in the lowest parts of Matundu (south-east Udzungwa), which was omitted by even the most permissive LER threshold. Respectively, 99% and 93% of records for strictly and nearly endemic plant species originate from areas classified as mountainous by the chosen typology.

Boundary placement

Polygons corresponding to the south-east slopes of Mt. Kilimanjaro were spatially distinct and so straightforward to remove (Fig. 3). In the south, the Eastern Arc is distinguished from adjacent highlands by climatic influence and vegetation type rather than landform. Consequently, the Udzungwa bloc had to be divided manually from moorlands south-east of the Makambako Gap: we followed the Mpanga River upstream from the Kilombero Valley towards the southern perimeter of Mufindi Scarp Forest Reserve, and then traced this west to dissect the polygon fully.

Figure 3 Boundary placement in the northern blocs. Features identified as ‘mountainous’ by the chosen typology were aggregated within simplified boundaries and distinguished as mountains or hills/escarpments depending on their prominence relative to adjacent elevations. Italicized hill names indicate inclusion on the basis of plant endemism.

Within the Eastern Arc chain, the majority of range boundaries were clearly distinguished (Fig. 4). Exceptions occurred for those blocs separated by steep-sided river valleys (Ukaguru versus Rubeho versus Udzungwa), in which cases the divisions were imposed post hoc by overlaying the respective river paths (Mkondoa and Great Ruaha Rivers). The most problematic bloc to delineate from surrounding terrain was Malundwe, a 1259 m peak within the Mikumi Hills (Fig. 2). The explicit delineation of this mountain and its small area of forest (<3 km2) was important for consistency with previous studies, but parameter combinations strict enough to isolate it resulted in the oversimplification or omission of marked topography elsewhere (for example LER2km with 30% threshold). Since the Mikumi Hills in their entirety are rarely considered part of the Eastern Arc chain, we distinguished the small Malundwe peak and its forest patch using a 900 m elevation contour, buffered and simplified as for the other blocs (Fig. 4). An alternative delineation that retains the Mikumi relief is provided (Appendix 1, available online at http://www.journals.cambridge.org/enc2011001).

Figure 4 Ecoregion boundaries, overlaid with forest distributions and protection status. Mid-elevation plateaus in Udzungwa are an option for inclusion (grasslands/heathlands but no natural forest). The majority of ‘near-endemic’ plant species have been collected within 10 or 20 km of the ecoregion boundary. Density of rural persons is highest within 30 km.

The proposed boundaries enclose a total area of c. 48 000 km2 (52 000 km2 including plateaus). Elevations range from 121 m to 2636 m above mean sea level, with both extremes corresponding to the highly prominent Uluguru range (Table 2). The 600 m elevation cut-off for relative relief proved appropriate, with only five ‘hilly’ features warranting inclusion on the basis of forest cover and/or plant endemism: two in Matundu (south-east Udzungwa), one west of the Nguu Range, and two in the Taita Hills. The last support locally endemic plant species: Encephalartos kisambo on the Maungu Hills and Monadenium guentheri on Maktau Hill, east and west of the main Taita bloc, respectively (Fig. 3). Handeni Hill, 50 km equidistant from West Usambara and Nguu, might have been included on similar grounds, but these forests are more usually associated with coastal vegetation.

Table 2 Summary of the Eastern Arc ecoregion, detailed by mountain bloc. Preclearance forest follows Hall et al. (Reference Hall, Burgess, Lovett, Mbilinyi and Gereau2009). Per cent gazetted is according to UNEP-WCMC (2009). Human populations are based on LandScan (2006) estimates, corrected according to the protected area data (no people live in National Parks or Game Reserves) and ward-level household surveys from the Tanzanian census (NBS, 2002; Appendix 3, available online at http://www.journals.cambridge.org/enc2011001).

The boundaries capture over 99% of all present-day and preclearance forest (Table 2), the lowest parts of Matundu being the only exception (beyond the south-central perimeter of Kilombero Nature Reserve). This area could be included post hoc (Appendix 1, available online at http://www.journals.cambridge.org/enc2011001). Non-woody species collected from mid-elevation plateaus in south-west Udzungwa (for example Oldenlandia oxycoccoides) might be considered strict-endemics by elevational criteria, but are not included by either the global or regional mountain typologies (slopes shallower than 2° and 5°, 1 km and 90 m resolutions, respectively). At both scales, steep plateau margins are considered mountainous; thus, depending on the application, Udzungwa plateaus could be included post hoc (Fig. 4). Alternatively, dry western margins could be removed, restricting Udzungwa to just those forested slopes benefiting from high orographic rainfall.

Of the 53 near-endemic plant species, at least 30 have been recorded exclusively within the topographical limits defined here (up to 40 if we allow for spatial error in collection localities). Unless additional specimen data suggest otherwise, these species could be revised to strictly endemic. Most other near-endemics have been collected within 10 km of the Eastern Arc boundary, and all within 20 km. Magombera Forest, for example, occupies a lowland position 6 km east of the Udzungwa bloc. Vegetation is predominantly of the kind common in coastal forests, but the presence of some characteristic Eastern Arc flora (for example Dialium holtzii and Isoberlinia scheffleri) and fauna (for example Udzungwa Red Colobus monkey Procolobus gordonorum) suggests past connectivity to higher elevation forests to the west. Such species persist in lowland areas like Magombera owing to their proximity to the coast rather than altitude or slope per se; instead of a qualitative definition of near-endemism, the vast majority could be effectively defined as occurring only within the Eastern Arc perimeter and up to 10 km or 20 km beyond.

Protected areas

Less than 10% of the Eastern Arc remains forested, compared with 37% preclearance cover. Notwithstanding spatial errors and misclassifications in the protected area and land cover data, we estimated that 75% of present-day forest lies within the protected area network (Table 2, Fig. 4). A further 50 km2 corresponds to forest reserves proposed but not yet gazetted, mainly in South Pare, East Usambara and Udzungwa. Other areas are traditionally managed or in private estates, and so are not represented in the database of protected areas used here (UNEP-WCMC 2009). In Kenya, forests are gazetted only within Kasigau Forest Reserve. Forests in the Tanzanian blocs are more widely protected by the state, consisting of both forest reserves (1909 km2) and nature reserves (785 km2) (Nilo and Amani in East Usambara, Kilombero in Udzungwa, and Uluguru). A further 586 km2 of Eastern Arc forest lies within three national parks, namely Udzungwa Mountains (north-east Udzungwa), Mikumi (all of Malundwe) and Mkomazi (north-east South Pare).

Forest estimates may contain some woodland, up to an eighth of the total, accounting for the lack of protection in some blocs (for example North Pare). The forest area within reserves may also be overestimated due to fragmentation resulting from mixed forest cultivation plots and long-lived fire-maintained grasslands (Finch & Marchant Reference Finch and Marchant2011). Conversely, forest regrowth in depopulated areas outside reserves is likely to be underrepresented.

Human populations

At the time of the 2002 Tanzania census, we estimate that the Eastern Arc had a population of c. 2.2 million people (2.35 million people if including the Udzungwa plateaus) and that a further 5.4 million people lived within 30 km (Table 2). Mean and median population densities are around 15% and 250% above the national averages for Tanzania, respectively (Fig. 5). Median density (more representative of rural communities than the mean) decreases with increasing distance from the bloc perimeters, up to a distance of 30 km (Figs 4 and 5). Major towns situated within the boundaries include Same (South Pare), Korogwe (West Usambara), Kilosa (Ukaguru), Mpwapwa (Rubeho) and Iringa (Udzungwa), with Morogoro and Ifakara just outside the Uluguru and Udzungwa blocs, respectively. According to these estimates, population densities are highest in West Usambara and North Pare, followed by East Usambara and Uluguru (Table 2).

Figure 5 Human population density versus distance to the Eastern Arc (includes Udzungwa plateaus). (a) Peaks in mean density correspond to towns and cities with populations exceeding 100 000. (b) Median density better portrays the distribution of rural persons in relation to the mountain resource (0–30 km).

DISCUSSION

Ecoregions are useful tools for making explicit the strong biogeographical affinities that exist within many regions of the world. Where consistently defined, they provide a spatial platform for monitoring, managing and forecasting the health of the constituent ecosystems, as well as the people reliant upon them for natural resources. Communities living in and around tropical mountains benefit from lower ambient temperatures, access to forest products, clean potable water, improved agricultural potential and fewer vectors for disease (for example mosquitoes). Advantages such as these are evident from the decrease in rural population density with increasing distance from the Eastern Arc boundary, a trend also affected by restrictions on human settlement in adjacent game reserves. Large-scale studies suggest a general pattern of human pressures in biologically important regions (Cincotta et al. Reference Cincotta, Wisnewski and Engelman2000; Balmford et al. Reference Balmford, Moore, Brooks, Burgess, Hansen, Williams and Rahbek2001; but see Joppa et al. Reference Joppa, Loarie and Pimm2009), driven by real or perceived benefits or by extraction frontiers, such as mining or logging, which open up previously remote environments for human settlement (Joppa et al. Reference Joppa, Loarie and Nelson2010; Scholte & de Groot Reference Scholte and de Groot2010).

In common with much of tropical Africa, the population of Tanzania has increased dramatically over the last half century, from 10 million people in 1960 to 42 million people in 2008 (http://data.worldbank.org/), with growth rates highest in urban areas and mountain foothills. Kenya has experienced a similar population boom, from eight million to 39 million people over the same period. Population growth is exerting increasing pressure on water supplies, energy production and land for agriculture. Demand for timber, poles and charcoal is also increasing as stocks in more accessible Miombo woodlands and coastal forests continue to diminish under pressure from urban centres such as Dar es Salaam (Ahrends et al. Reference Ahrends, Burgess, Milledge, Bulling, Fisher, Smart, Clarke, Mhoro and Lewis2010). Increased resource demand and intensified land use have direct implications for forest health and local livelihoods, as well as more diffuse impacts such as biodiversity loss and the release of sequestered carbon into the atmosphere (Kohler et al. Reference Kohler, Giger, Hurni, Ott, Wiesmann, von Dach and Maselli2010).

Global typologies derived from digital elevation data provide a welcome platform for large-scale studies of mountain environments and the people they support (Huddleston et al. Reference Huddleston, Ataman, de Salvo, Zanetti, Bloise, Bel, Franceschini and Fè d'Ostiani2003). At more local scales, regional biogeography, naming conventions and micro-relief gain importance, and so targeted case studies are required to obtain relevant delineations of region extent. Here, we provide a consistent topographical foundation for delimiting the Eastern Arc Mountains, one of the world's most important ecoregions for conservation (Olson & Dinerstein Reference Olson and Dinerstein2002). The approach we present is generic and could be readily calibrated for application to other mountain regions.

Mountain ecosystems, like all biological assemblages, are moving targets: the products of ancient evolutionary processes, recent climatic conditions and, in the case of the Eastern Arc, ongoing disturbance by humans, fire and large herbivores such as elephants. Ecoregion extent inferred directly from climate, vegetation and/or land use therefore requires frequent revision in response to new data or changes in those variables. Moreover, if considered independently of surrounding habitats and without historical perspective, contemporary snapshots of forest mosaics and local climates have restricted potential to aid understanding of the system as a whole (Fjeldså & Lovett Reference Fjeldså and Lovett1997). In cases where mountainous relief is central to the historical affinities under consideration, boundaries defined by topographical means are in our view preferable. Here, they provide a geographically coherent framework for monitoring, which is likely to incorporate potentially important but as yet undocumented sites for conservation.

When the Eastern Arc was first proposed as phytogeographically distinct in the 1980s, the biological importance of many of the blocs was unconfirmed. Forests in Rubeho, for instance, having only recently received funding for botanical surveys, have long been undervalued in terms of conservation priority (Doggart et al. Reference Doggart, Perkin, Kiure, Fjeldså, Poynton and Burgess2006). Similarly, recent focus in Nguru has revealed a number of species new to science, particularly amongst herpetofauna (Menegon et al. Reference Menegon, Doggart and Owen2008). Nguu remains largely unsurveyed, but recent field visits and bioclimatic modelling suggest it too could be rich in species (Platts et al. Reference Platts, Ahrends, Gereau, McClean, Lovett, Marshall, Pellikka, Mulligan, Fanning and Marchant2010). The majority of forest in eastern parts of Nguru and Nguu are within reserves, but their western outliers remain ungazetted, leaving them open to degradation.

The size of the ecoregion is here defined to be 48 000–52 000 km2, depending on the inclusion of mid-elevation plateaus in Udzungwa. This is higher than a previous estimate of 37 000 km2 (Tanzania only) published by the Forestry and Beekeeping Division (Mbilinyi et al. Reference Mbilinyi, Malimbwi, Shemwetta, Zahabu, Katani and Kashaigili2006), which imposes a lower limit of 500 m elevation, omits the western margins of Udzungwa and includes fewer outlying peaks. In Kenya, we identify the main Taita bloc, plus Mount Kasigau and the Sagalla Hills to the south and east, as well as two lower elevation hills (Maungu and Maktau) known to support locally endemic plant species. The explicit inclusion of these outliers is especially important given the plight of other Taita forests, now restricted to a few remnant patches (Pellikka et al. Reference Pellikka, Lötjönen, Sijander and Lens2009), threatening the persistence of many rare species (Rogo & Oguge Reference Rogo and Oguge2000). Although furthest from the main bloc, Kasigau forests are relatively undisturbed and so may provide refugia for important flora and fauna (for example Taita White-eye Zosterops silvanus; Mulwa et al. Reference Mulwa, Bennun, Ogol and Lens2007). Elsewhere, outlying features were identified by the regional mountain typology, but omitted from the current definition because of low relative elevation and lack of data on the presence/absence of characteristic Eastern Arc vegetation. Because new data may yet justify their inclusion, we provide the spatial extents of all identified features (Appendix 1, available online at http://www.journals.cambridge.org/enc2011001).

Although these boundaries were not placed directly according to forest distributions, we did use vegetation data to indicate appropriate terrain parameters. Our estimation of current forest cover at over 4300 km2 is more than some previous studies, but less than the 5700 km2 of natural forest proposed by Newmark (Reference Newmark2002). Other sources put the figure closer to 3500 km2 (Mbilinyi et al. Reference Mbilinyi, Malimbwi, Shemwetta, Zahabu, Katani and Kashaigili2006; Burgess et al. Reference Burgess, Butynski, Cordeiro, Doggart, Fjeldså, Howell, Kilahama, Loader, Lovett, Mbilinyi, Menegon, Moyer, Nashanda, Perkin, Rovero, Stanley and Stuart2007 and references therein), but are similar to Newmark's estimate if woodlands are included. The land cover data used here have been iteratively improved during a series of workshops in Tanzania through the Valuing the Arc Programme (http://valuingthearc.org/), and were further corrected for the current application by reference to forest change estimates, government reports and our own field notes. Unless ground-truthed, all such estimates are subject to uncertainty, especially as regards the distinction between closed woodland and deciduous/degraded forest, which can be easily confused in remotely sensed images. Levels of disturbance and fragmentation in the forest interior are also difficult to assess remotely.

To reduce our reliance on these data, which at best provide a current snapshot of forest distribution, we further considered extrapolations of historical forest cover (Hall et al. Reference Hall, Burgess, Lovett, Mbilinyi and Gereau2009). Such estimates are not directly applicable as ecoregion boundaries as sediment cores are sparsely distributed and an evidence-based assessment of historical extent is only starting to become possible (Mumbi et al. Reference Mumbi, Marchant, Hooghiemstra and Wooller2008; Finch et al. Reference Finch, Leng and Marchant2009; Finch & Marchant Reference Finch and Marchant2011), but they do provide a broad indication of preclearance cover. The fact that our boundaries fully enclose these kinds of estimates suggests that they are well placed to withstand future changes in distribution, at least over timescales relevant for management. If the boundaries were to require adjustment in the future, this could be addressed multilaterally according to the framework presented here.

Genetic evidence from a range of focal taxa shows that Eastern Arc vegetation has persisted over many millions of years. Phylogenetic analysis of the tree Macaranga capensis indicates long-term separation of populations on different mountain blocs, but also past connectivity (A. S. Jump, personal communication 2010); species presently restricted to montane areas may have once encroached on tropical lowlands. Strong affinities have also been observed between the flora and fauna of the Eastern Arc and those of west and central Africa, suggesting remnants of a pan-African forest belt (Couvreur et al. Reference Couvreur, Chatrou, Sosef and Richardson2008). More locally, the Lake Nyasa Highlands, Mount Kilimanjaro and coastal forests all contain plant taxa otherwise restricted to the Eastern Arc, as do the Shimba Hills in south-east Kenya, again indicating past connections or possibly long-distance dispersal.

Because of these historical overlaps in composition, the dichotomy of Eastern Arc and coastal vegetation can sometimes be contentious (Burgess & Clarke Reference Burgess and Clarke2000). An academic reason for maintaining some form of distinction between mountain and coastal vegetation in East Africa is that, since 2004, they have been classified as belonging to different biodiversity hotspots (Eastern Afromontane versus Coastal Forests of East Africa; Mittermeier et al. Reference Mittermeier, Robles Gil, Hoffmann, Pilgrim, Brooks, Mittermeier, Lamoreux and da Fonseca2004). In addition, the Tanzanian mountains contain central government reserves administered by the catchment forest office, whereas the districts administer coastal forest reserves. From a human perspective, mountain people face different challenges to those living in towns or rural lowlands (Kreutzmann Reference Kreutzmann2001); thus ensuring their social welfare and right to the traditional use of natural resources, whilst also fulfilling conservation objectives, demands a specific focus on the mountain region.

The 500 m threshold for plant endemism, although a pragmatic response to the need for spatially consistent mountain limits, does not account for geographical differences in baseline elevation, nor is it consistent with other taxonomic groups (compare with Poynton et al. Reference Poynton, Loader, Sherratt and Clarke2007). Using the altitudinal range of forest within a given mountain bloc is an appealing alternative, but one that is particularly sensitive to recent patterns of deforestation. The spatial limits presented here might be an appropriate basis for a new set of endemism criteria. Our results suggest that well over half of the plant species here tentatively considered ‘near-endemic’ would be classified as strictly endemic according to the derived ecoregion boundary, and that all could be consistently defined according to an inclusion zone of up to 20 km. Moreover, in a preliminary retabulation of plant endemism, which compares the recorded elevations of specimens with the altitudinal limits of the corresponding mountain blocs (as detailed in Table 2), nearly all ‘near-endemics’ can be considered strictly endemic. We encourage similar tests for other taxonomic groups (spatial data available online at http://www.journals.cambridge.org/enc2011001).

In providing a spatially explicit definition for this area of endemism, we hope to prompt research and conservation in lesser-studied parts of the Eastern Arc, which could be biologically and politically undervalued due to spatial bias in the data (Platts et al. Reference Platts, Ahrends, Gereau, McClean, Lovett, Marshall, Pellikka, Mulligan, Fanning and Marchant2010; Ahrends et al. Reference Ahrends, Burgess, Gereau, Marchant, Bulling, Lovett, Platts, Kindemba, Owen, Fanning and Rahbek2011). The boundaries are also relevant for reforestation strategies, particularly ahead of the United Nations REDD pilot (http://www.un-redd.org/) in Tanzania.

ACKNOWLEDGEMENTS

Our thanks to providers of data to the Missouri Botanical Garden's TROPICOS database, local experts for advice on land cover classification, and for the comments of two anonymous reviewers. This work was funded by the Marie-Curie programme of the European 6th Framework (MEXT-CT-2004-517098), and by grants from the Academy of Finland, Leverhulme Trust, Packard Foundation and Royal Society.

References

Ahrends, A., Burgess, N.D., Milledge, S.A.H., Bulling, M.T., Fisher, B., Smart, J.C.R., Clarke, G.P., Mhoro, B.E. & Lewis, S.L. (2010) Predictable waves of sequential forest degradation and biodiversity loss spreading from an African city. Proceedings of the National Academy of Sciences USA 107 (33): 1455614561.CrossRefGoogle ScholarPubMed
Ahrends, A., Burgess, N.D., Gereau, R.E., Marchant, R., Bulling, M.T., Lovett, J.C., Platts, P.J., Kindemba, V.W. Owen, N., Fanning, E. & Rahbek, C. (2011) Funding begets biodiversity. Diversity and Distributions 17 (2): 191200.CrossRefGoogle Scholar
Balmford, A., Moore, J.L., Brooks, T., Burgess, N., Hansen, L.A., Williams, P. & Rahbek, C. (2001) Conservation conflicts across Africa. Science 291 (5513): 26162619.CrossRefGoogle ScholarPubMed
Blyth, S., Groombridge, B., Lysenko, I., Miles, L. & Newton, A. (2002) Mountain Watch. Cambridge, UK: UNEP World Conservation Monitoring Centre.Google Scholar
Brooks, T.M., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Oldfield, S., Magin, G. & Hilton-Taylor, C. (2002) Habitat loss and extinction in the hotspots of biodiversity. Conservation Biology 16 (4): 909923.CrossRefGoogle Scholar
Browne, T., Fox, R. & Funnell, D. (2004) The ‘invisible’ mountains. Using GIS to examine the extent of mountain terrain in south Africa. Mountain Research and Development 24 (1): 2834.CrossRefGoogle Scholar
Burgess, N.D. & Clarke, G.P. (2000) Coastal Forests of Eastern Africa. Gland, Switzerland and Cambridge, UK: IUCN.Google Scholar
Burgess, N.D., Hales, J.D., Ricketts, T.H. & Dinerstein, E. (2006) Factoring species, non-species values and threats into biodiversity prioritisation across the ecoregions of Africa and its islands. Biological Conservation 127 (4): 383401.CrossRefGoogle Scholar
Burgess, N.D., Butynski, T.M., Cordeiro, N.J., Doggart, N.H., Fjeldså, J., Howell, K.M., Kilahama, F.B., Loader, S.P., Lovett, J.C., Mbilinyi, B., Menegon, M., Moyer, D.C., Nashanda, E., Perkin, A., Rovero, F., Stanley, W.T. & Stuart, S.N. (2007) The biological importance of the Eastern Arc Mountains of Tanzania and Kenya. Biological Conservation 134 (2): 209231.CrossRefGoogle Scholar
Burgess, N., Mwakalila, S., Madoffe, S., Ricketts, T., Olwero, N., Swetnam, R., Mbilinyi, B., Marchant, R., Mtalo, F., White, S., Munishi, P., Marshall, A., Malimbwi, R., Jambiya, G., Fisher, B., Kajembe, G., Morse-Jones, S., Kulindwa, K., Green, J. & Balmford, A. (2009) Valuing the Arc. A programme to map and value ecosystem services in Tanzania. Mountain Research Initiative Newsletter 18 (3): 1821.Google Scholar
Burgess, N.D., Bahane, B., Clairs, T., Danielsen, F., Dalsgaard, S., Funder, M., Hagelberg, N., Harrison, P., Haule, C., Kabalimu, K., Kilahama, F., Kilawe, E., Lewis, S.L., Lovett, J.C., Lyatuu, G., Marshall, A.R., Meshack, C., Miles, L., Milledge, S.A.H., Munishi, P.K.T., Nashanda, E., Shirima, D., Swetnam, R.D., Willcock, S., Williams, A. & Zahabu, E. (2010) Getting ready for REDD+ in Tanzania: a case study of progress and challenges. Oryx 44 (3): 339351.CrossRefGoogle Scholar
Clark, B.J.F. & Pellikka, P.K.E. (2009) Landscape analysis using multiscale segmentation and object-orientated classification. In: Recent Advances in Remote Sensing and Geoinformation Processing for Land Degradation Assessment, ed. Röder, A. & Hill, J., pp. 323342. London, UK: Taylor & Francis Group.Google Scholar
Cincotta, R.P., Wisnewski, J. & Engelman, R. (2000) Human population in the biodiversity hotspots. Nature 404 (6781): 990992.CrossRefGoogle ScholarPubMed
Conservation International (2008) Forest cover and change in Eastern Arc mountain and coastal forests, 1990–2000 [www document]. URL http://www.cepf.net/where_we_work/regions/africa/eastern_arc_coastal_forests/Google Scholar
Couvreur, T.L.P., Chatrou, L.W., Sosef, M.S.M. & Richardson, J.E. (2008) Molecular phylogenetics reveal multiple tertiary vicariance origins of the African rain forest trees. BMC Biology 6: 10.CrossRefGoogle ScholarPubMed
Doggart, N., Perkin, A., Kiure, J., Fjeldså, J., Poynton, J. & Burgess, N. (2006) Changing places: how the results of new field work in the Rubeho Mountains influence conservation priorities in the Eastern Arc Mountains of Tanzania. African Journal of Ecology 44 (2): 134144.CrossRefGoogle Scholar
Elenga, H., Peyron, O., Bonnefille, R., Jolly, D., Cheddadi, R., Guiot, J., Andrieu, V., Bottema, S., Buchet, G., de Beaulieu, J.L., Hamilton, A.C., Maley, J., Marchant, R., Perez-Obiol, R., Reille, M., Riollet, G., Scott, L., Straka, H., Taylor, D., Van Campo, E., Vincens, A., Laarif, F. & Jonson, H. (2000) Pollen-based biome reconstruction for southern Europe and Africa 18,000 yr BP. Journal of Biogeography 27 (3): 621634.CrossRefGoogle Scholar
Finch, J.M., Leng, M.J. & Marchant, R. (2009) Late Quaternary vegetation dynamics in a biodiversity hotspot, the Uluguru Mountains of Tanzania. Quaternary Research 72 (1), 111122.CrossRefGoogle Scholar
Finch, J.M. & Marchant, R. (2011) A palaeoecological investigation into the role of fire and human activity in the development of montane grasslands in East Africa. Vegetation History and Archaeobotany (in press).CrossRefGoogle Scholar
Fjeldså, J. & Lovett, J.C. (1997) Geographical patterns of old and young species in African forest biota: the significance of specific montane areas as evolutionary centres. Biodiversity and Conservation 6 (3): 325346.CrossRefGoogle Scholar
Gerrard, A.J. (1990) Mountain Environments: an Examination of the Physical Geography of Mountains. London, UK: Belhaven Press.Google Scholar
Hall, J., Burgess, N.D., Lovett, J., Mbilinyi, B. & Gereau, R.E. (2009) Conservation implications of deforestation across an elevational gradient in the Eastern Arc Mountains, Tanzania. Biological Conservation 142 (11): 25102521.CrossRefGoogle Scholar
Huddleston, B., Ataman, E., de Salvo, P., Zanetti, M., Bloise, M., Bel, J., Franceschini, G. & Fè d'Ostiani, L. (2003) Towards a GIS-based Analysis of mountain environments and populations. Environment and Natural Resources Working Paper No. 10. Food and Agriculture Organization of the United Nations, Rome, Italy.Google Scholar
Jarvis, A., Reuter, H.I., Nelson, A. & Guevara, E. (2008) Hole-filled seamless SRTM data. Version 4. International Centre for Tropical Agriculture (CIAT). URL http://srtm.csi.cgiar.org/Google Scholar
Joppa, L.N., Loarie, S.R. & Nelson, A. (2010) Measuring population growth around tropical protected areas: current issues and solutions. Tropical Conservation Science 3 (2): 117121.CrossRefGoogle Scholar
Joppa, L.N., Loarie, S.R. & Pimm, S.L. (2009) On population growth near protected areas. Plos One 4 (1): e4279.CrossRefGoogle ScholarPubMed
Kapos, V., Rhind, J., Edwards, M., Price, M.F. & Ravilious, C. (2000) Developing a map of the world's mountain forests. In: Forests in Sustainable Mountain Development: a State of Knowledge Report for 2000, ed. Price, M.F. & Butt, N., pp. 49. Wallingford, UK: CAB International.CrossRefGoogle Scholar
Kohler, T., Giger, M., Hurni, H., Ott, C., Wiesmann, U., von Dach, S.W. & Maselli, D. (2010) Mountains and climate change: a global concern. Mountain Research and Development 30 (1): 5355.CrossRefGoogle Scholar
Kreutzmann, H. (2001) Development indicators for mountain regions. Mountain Research and Development 21 (2): 132139.CrossRefGoogle Scholar
LandScan (2006) High resolution global population data set. Copyright UT-Battelle, LLC, operator of Oak Ridge National Laboratory under contract No. DE-AC05–00OR22725 with the United States Department of Energy. URL http://www.ornl.gov/sci/landscan/Google Scholar
Larigauderie, A. & Mooney, H.A. (2010) The intergovernmental science-policy platform on biodiversity and ecosystem services: moving a step closer to an IPCC-like mechanism for biodiversity. Current Opinion in Environmental Sustainability 2 (1–2): 914.CrossRefGoogle Scholar
Lovett, J.C. (1985) Moist forests of eastern Tanzania. Swara 8 (5): 89.Google Scholar
Lovett, J.C. (1990) Classification and status of the moist forests of Tanzania. Mitteilungen aus dam Institut für Allgemeine Botanik Hamburg (Notifications from the Institute for General Botany Hamburg) 23a: 287300.Google Scholar
Lovett, J.C. (1992) Classification and affinities of the Eastern Arc moist forests of Tanzania. Ph.D. thesis, University of Wales, Bangor, UK.CrossRefGoogle Scholar
Lovett, J.C. & Pócs, T. (1993) Assessment of the condition of the Catchment Forest Reserves, a botanical appraisal. Catchment Forest Report [www document]. URL http://www.easternarc.or.tz/Google Scholar
Lovett, J.C., Hansen, J.R. & Hørlyck, V. (2000) Comparison with Eastern Arc forests. In: Coastal Forests of Eastern Africa, ed. Burgess, N.D. & Clarke, G.P., pp. 115125. Gland, Switzerland and Cambridge, UK: IUCN.Google Scholar
Marshall, A.R., Jorgensbye, H.I.O., Rovero, F., Platts, P.J., White, P.C.L. & Lovett, J.C. (2010) The species area relationship and confounding variables in a threatened monkey community. American Journal of Primatology 72 (4): 325336.CrossRefGoogle Scholar
Mbilinyi, B.P., Malimbwi, R.E., Shemwetta, D.T.K., Songorwa, Zahabu, E., Katani, J.Z. & Kashaigili, J. (2006) Forest Area Baseline for the Eastern Arc Mountains. Dar es Salaam, The United Republic of Tanzania: Conservation and Management of the Eastern Arc Mountain Forests, Forestry and Beekeeping Division.Google Scholar
McMahon, G., Wiken, E.B. & Gauthier, D.A. (2004) Toward a scientifically rigorous basis for developing mapped ecological regions. Environmental Management 34: S111S124.CrossRefGoogle Scholar
Menegon, M., Doggart, N. & Owen, N. (2008) The Nguru mountains of Tanzania, an outstanding hotspot of herpetofaunal diversity. Acta Herpetologica 3 (2): 107127.Google Scholar
Meybeck, M., Green, P. & Vorosmarty, C. (2001) A new typology for mountains and other relief classes: an application to global continental water resources and population distribution. Mountain Research and Development 21 (1): 3445.CrossRefGoogle Scholar
Messerli, B. & Ives, J.D.E. (1997) Mountains of the World: A Global Priority. New York, NY, USA: Parthenon.Google Scholar
Mittermeier, R.A., Robles Gil, P., Hoffmann, M., Pilgrim, J., Brooks, T., Mittermeier, C.G., Lamoreux, J. & da Fonseca, G.A.B. (2004) Hotspots Revisited. Mexico City, Mexico: CEMEX.Google Scholar
MNRT (1997) National Reconnaissance Level Land Use and Natural Resources Mapping Project. Dar es Salaam, The United Republic of Tanzania: Ministry of Natural Resources and Tourism.Google Scholar
Mulwa, R.K., Bennun, L.A., Ogol, C. & Lens, L. (2007) Population status and distribution of Taita White-eye Zosterops silvanus in the fragmented forests of Taita Hills and Mount Kasigau, Kenya. Bird Conservation International 17 (2): 141150.CrossRefGoogle Scholar
Mumbi, C.T., Marchant, R., Hooghiemstra, H. & Wooller, M.J. (2008) Late Quaternary vegetation reconstruction from the Eastern Arc Mountains, Tanzania. Quaternary Research 69 (2): 326341.CrossRefGoogle Scholar
NBS (2002) Tanzania Population and Housing Census 2002. Dar es Salaam, The United Republic of Tanzania: National Bureau of Statistics.Google Scholar
Newmark, W.D. (2002) Conserving Biodiversity in East African Forests: a Study of the Eastern Arc Mountains. Berlin, Germany: Springer.CrossRefGoogle Scholar
Nordregio (2004) Mountain Areas in Europe: Analysis of Mountain Areas in EU Member States, Acceding and Other European Countries. Stockholm, Sweden: Nordic Centre for Spatial Development.Google Scholar
OED (1989) Oxford English Dictionary. Second Edition. Oxford, UK: Oxford University Press.Google Scholar
Olson, D.M. & Dinerstein, E. (2002) The Global 200: priority ecoregions for global conservation. Annals of the Missouri Botanical Garden 89: 199224.CrossRefGoogle Scholar
Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N., Underwood, E.C., D'Amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, C.J., Allnutt, T.F., Ricketts, T.H., Kura, Y., Lamoreux, J.F., Wettengel, W.W., Hedao, P. & Kassem, K.R. (2001) Terrestrial ecoregions of the worlds: a new map of life on Earth. Bioscience 51 (11): 933938.CrossRefGoogle Scholar
Pellikka, P.K.E., Lötjönen, M., Sijander, M. & Lens, L. (2009) Airborne remote sensing of spatiotemporal change (1955–2004) in indigenous and exotic forest cover in the Taita Hills, Kenya. International Journal of Applied Earth Observation and Geoinformation 11 (4): 221232.CrossRefGoogle Scholar
Platts, P.J., Ahrends, A., Gereau, R.E., McClean, C.J., Lovett, J.C., Marshall, A.R., Pellikka, P.K.E., Mulligan, M., Fanning, E. & Marchant, R. (2010) Can distribution models help refine inventory-based estimates of conservation priority? A case study in the Eastern Arc forests of Tanzania and Kenya. Diversity and Distributions 16 (4): 628642.CrossRefGoogle Scholar
Pócs, T. (1976) Vegetation mapping in the Uluguru mountains (Tanzania, East Africa). Boissiera 24: 477498.Google Scholar
Poynton, J.C., Loader, S.P., Sherratt, E. & Clarke, B.T. (2007) Amphibian diversity in east African biodiversity hotspots: altitudinal and latitudinal patterns. Biodiversity and Conservation 16 (4): 11031118.CrossRefGoogle Scholar
Rogo, L. & Oguge, N. (2000) The Taita Hills forest remnants: a disappearing world heritage. Ambio 29 (8): 522523.CrossRefGoogle Scholar
Schlüter, T. (1997) Geology of East Africa. In: Contributions to the Regional Geography of the Earth (translated from German). Berlin, Germany: Borntraeger.Google Scholar
Scholte, P. & de Groot, W.T. (2010) From debate to insight: three models of immigration to protected areas. Conservation Biology 24 (2): 630632.CrossRefGoogle ScholarPubMed
Stanley, W.T. & Olson, L.E. (2005) Phylogeny, phylogeography, and geographic variation of Sylvisorex howelli (Soricidae), an endemic shrew of the Eastern Arc Mountains, Tanzania. Journal of Zoology 266: 341354.CrossRefGoogle Scholar
Thompson, W.F. (1964) How and why to distinguish between mountains and hills. Professional Geographer 16 (6): 68.CrossRefGoogle Scholar
UNEP-WCMC (2009) The World Database on Protected Areas. Cambridge, UK: UNEP-WCMC. URL http://www.wdpa.org/Google Scholar
Vincens, A., Garcin, Y. & Buchet, G. (2007) Influence of rainfall seasonality on African lowland vegetation during the Late Quaternary: pollen evidence from Lake Masoko, Tanzania. Journal of Biogeography 34 (7): 12741288.CrossRefGoogle Scholar
Figure 0

Figure 1 Global mountain typology defined by UNEP-WCMC at 1 km resolution. (a) East Africa, showing divisions in Tanzanian forest on the basis of geology and climate (from Lovett 1990): Coastal, Eastern Arc and Northern forests are all under the direct climatic influence of the Indian Ocean (rather than the Great Lakes), but only the Eastern Arc is on igneous and metamorphic rock. (b) Zoomed perspective of the 13 Eastern Arc ranges (blocs).

Figure 1

Table 1 Mountain area and forest sites included as a result of different calibrations of the local elevation range parameter (LER radius/threshold, mountain classes 1–2). Analyses conducted within Eastern Arc half-degree grid squares using 90 m elevation data and slope thresholds of 5° (class 2) and 10° (class 3); otherwise mountain classification follows UNEP-WCMC, but with no lower limit in class 1. *Mountain typology from which Eastern Arc boundaries were derived (Appendix 2, available online at http://www.journals.cambridge.org/enc2011001).

Figure 2

Figure 2 Sensitivity of the regional mountain typology to local elevation range (LER, radius = 500 m). Upper pane: elevation in the Uluguru bloc (×5 vertical exaggeration). Lower pane: variations in extent resulting from different LER thresholds (% of radius).

Figure 3

Figure 3 Boundary placement in the northern blocs. Features identified as ‘mountainous’ by the chosen typology were aggregated within simplified boundaries and distinguished as mountains or hills/escarpments depending on their prominence relative to adjacent elevations. Italicized hill names indicate inclusion on the basis of plant endemism.

Figure 4

Figure 4 Ecoregion boundaries, overlaid with forest distributions and protection status. Mid-elevation plateaus in Udzungwa are an option for inclusion (grasslands/heathlands but no natural forest). The majority of ‘near-endemic’ plant species have been collected within 10 or 20 km of the ecoregion boundary. Density of rural persons is highest within 30 km.

Figure 5

Table 2 Summary of the Eastern Arc ecoregion, detailed by mountain bloc. Preclearance forest follows Hall et al. (2009). Per cent gazetted is according to UNEP-WCMC (2009). Human populations are based on LandScan (2006) estimates, corrected according to the protected area data (no people live in National Parks or Game Reserves) and ward-level household surveys from the Tanzanian census (NBS, 2002; Appendix 3, available online at http://www.journals.cambridge.org/enc2011001).

Figure 6

Figure 5 Human population density versus distance to the Eastern Arc (includes Udzungwa plateaus). (a) Peaks in mean density correspond to towns and cities with populations exceeding 100 000. (b) Median density better portrays the distribution of rural persons in relation to the mountain resource (0–30 km).

Supplementary material: File

Platts Supplementary Appendix

Platts Supplementary Appendix

Download Platts Supplementary Appendix(File)
File 2.3 MB
Supplementary material: File

Platts Supplementary Data 1

Platts Supplementary Data 1

Download Platts Supplementary Data 1(File)
File 636.9 KB
Supplementary material: File

Platts Supplementary Data 2

Platts Supplementary Data 2

Download Platts Supplementary Data 2(File)
File 615 KB
Supplementary material: File

Platts Supplementary Data 3

Platts Supplementary Data 3

Download Platts Supplementary Data 3(File)
File 9.3 MB