Biodivers Conserv DOI 10.1007/s10531-012-0393-0 ORIGINAL PAPER Bat diversity in the lowland forests of the Heart of Borneo Matthew J. Struebig • Monika Bo_zek • Jan Hildebrand Stephen J. Rossiter • David J. W. Lane • Received: 8 June 2012 / Accepted: 18 October 2012 Ó Springer Science+Business Media Dordrecht 2012 Abstract Borneo’s rainforests are renowned for their high levels of biodiversity, yet information on the distribution and structure of this diversity is lacking, particularly for less charismatic taxonomic groups. We quantified bat diversity across ten sites within a contiguous tract of largely undisturbed rainforest in the Heart of Borneo (HoB) transboundary conservation area. Using comparative analyses of 1,362 bat captures from six sites in Brunei Darussalam, together with data from four additional sites in neighbouring territories, we show that the main differences in bat assemblage composition between sites were driven by the abundances of a few cave-roosting species. Beta diversity (distance decay) was notably low and non-significant. Bat assemblage structure in these undisturbed palaeotropical forests is therefore relatively homogenous in the absence of environmental gradients. By adding 15 bat species to the Brunei national inventory, we confirm the area of north Borneo to be species-diverse and therefore a priority for conservation efforts. However, we also highlight that coastal forest to be included in a recent extension to the HoB hosts bat assemblages with the fewest species and lowest densities. We maintain that extending the HoB in Brunei to include a more diverse portfolio of habitat types is still Electronic supplementary material The online version of this article (doi:10.1007/s10531-012-0393-0) contains supplementary material, which is available to authorized users. M. J. Struebig (&) School of Anthropology and Conservation, Durrell Institute of Conservation and Ecology, University of Kent, Canterbury CT2 7NR, UK e-mail: m.j.struebig@kent.ac.uk; m.struebig@qmul.ac.uk M. J. Struebig
J. Hildebrand
S. J. Rossiter School of Biological and Chemical Sciences, Queen Mary University of London, London E1 4NS, UK M. Bo_zek Genome Centre, John Vane Science Centre, Charterhouse Square, Queen Mary University of London, London EC1M 6BQ, UK D. J. W. Lane Department of Biology, Universiti Brunei Darussalam, Jl. Tungku Link, Gadong BE1410, Brunei Darussalam 123 Biodivers Conserv warranted on the grounds of maximising botanical diversity and habitat area, as long as it does not detract attention from interior forests that support higher vertebrate diversity. Keywords Chiroptera
Brunei
Beta diversity
Distance decay
Tropical forest
Conservation Introduction The tropical rainforests of Borneo are globally recognised for their high levels of species diversity and endemism, forming part of the Sundaic hotspot of Southeast Asia (Sodhi et al. 2004). At least 288 terrestrial mammals (Payne et al. 2000) are reported, of which 44 are endemic (Mackinnon et al. 1996). Mammal diversity and endemism appear to match patterns of other taxonomic groups, with diversity greatest in the north of the island within the states of northern East Kalimantan (Indonesia), Sabah and eastern Sarawak (Malaysia), and the sultanate of Brunei Darussalam (Catullo et al. 2008; Mackinnon et al. 1996). Although this pattern might in part be due to historical biases in reporting, it has also received support from recent island-wide analyses of species distributions (Beck et al. 2011; Meijaard and Nijman 2003; Raes et al. 2009). Unfortunately, the region of northern Borneo is also undergoing high levels of habitat loss; unsustainable logging practices and clearance of rainforest for plantation agriculture are of particular conservation concern (Fitzherbert et al. 2008). In a bid to safeguard Borneo’s biodiversity the governments of Brunei, Indonesia and Malaysia recently signed a declaration to conserve 22 million ha of continuous habitat within a transboundary conservation area known as the Heart of Borneo (HoB) (WWF 2011). The proposed area was designed to maintain the forests of greatest landscape connectivity, ecosystem service, and importance for the island’s charismatic megafauna. To meet these targets more effectively there have been recent calls to extend the area to include additional lowland forests along Borneo’s northern coast. However, very little is known of the diversity and distributions of much of the wildlife within the HoB, and the proposed extension, and this is particularly true for less charismatic groups such as small mammals and bats. Bats are a highly diverse animal group and, like other vertebrates, form a centre of richness in the Indo-Malayan region (Findley 1993). In the palaeotropics bats form a major component of forest fauna, and in Borneo comprise at least 93 species representing a third of terrestrial mammals (Payne et al. 2000). Palaeotropical bat assemblages are typically dominated by insectivorous species of the families Hipposideridae and Rhinolophidae, as well as the Vespertilionidae subfamilies Kerivoulinae and Murininae (Furey et al. 2010; Kingston et al. 2003). These species are highly adapted for foraging in forest-interior habitats, and form clear and predictable ensembles based on their ecomorphological traits (e.g. wing loading, aspect ratio and echolocation signal design; Kingston et al. 2003) and roosting ecology (e.g. foliage, tree cavities, caves; Struebig et al. 2008). However, such specialised characters and lifestyles mean that many forest-interior bat species are, in consequence, particularly susceptible to habitat disturbance. Bat assemblages undergo structural changes following habitat degradation (Furey et al. 2010), and lose species following forest fragmentation (Struebig et al. 2008, 2011) and conversion (Phommexay et al. 2011). Should these trends continue, up to forty percent of Southeast Asia’s bats are predicted to become regionally extinct by 2100 (Lane et al. 2006). 123 Biodivers Conserv Some of the first research on palaeotropical bat assemblages was undertaken on Borneo (Francis 1990, 1994), but there has been remarkably little research on the island since. Of the few studies that have been undertaken, most have focused on single site inventories. For example, bat diversity estimates are highest for Kabili-Sepilok Forest Reserve in Sabah (41 species; Francis 1990, 1994) and Ulu Temburong National Park in Brunei (35 species; Kofron 2002). Unsurprisingly, diversity estimates are greater where sampling was extended to multiple sites within an area (e.g. 44 species in Kinabalu National Park; Yasuma and Andau 2000). Nevertheless, aside from site inventories we still know very little about how bat assemblages are structured over the forested landscape of Borneo, and elsewhere in the palaeotropics. As part of a broader survey programme on Borneo, we undertook bat surveys in Brunei Darussalam at sites within a 380,000 ha tract of mostly undisturbed forest, covering the main geological formations in north Borneo and elsewhere in the HoB. Historical bat survey records in this area are limited to fruit bats and easily-captured insectivorous species (Kofron 2002; Struebig et al. 2010; Yasuma and Abdullah 1997). Sampling in this region provided us with the opportunity to study assemblage patterns in a near-pristine ecosystem, while also informing the conservation mission of the HoB initiative. Species composition can be randomly aggregated according to neutral variation in dispersal and ecological drift, and/or governed by variation in other factors such as local habitat conditions (Chase 2005). Therefore, to separate these potential influences, we examined spatial turnover in bat assemblages between sites in the undisturbed forest by testing for decay in assemblage similarity with geographic distance. Distance decay describes the tendency for assemblages to be spatially correlated, whereby nearby sites are biologically more similar than those further away (Soininen et al. 2007). If assemblage similarity is dictated by neutral historical influences, and less so by environmental heterogeneity (such as that presented by habitat differences between sites), then a decay in similarity with distance is expected. Methods Study sites Bat sampling was undertaken at six forested sites in Brunei that represent the main geological formations of north Borneo. Sites were selected to improve existing inventories (at Andulau, Tasek Merimbun and Ulu Temburong) or to provide new data from un-surveyed areas (Peradayan, Bukit Teraja and Sungai Ingei) (Fig. 1; Table 1). All sites are connected by continuous lowland rainforest and bat surveys were undertaken in the dominant vegetation formation, mixed dipterocarp forest (MDF). Under-surveyed sites were identified by mapping all bat species locality records from the literature as well as from specimen records from Brunei Museums department (Kofron 2002; Struebig et al. 2010; Yasuma and Abdullah 1997; Fig. 1). While study sites were all of the same broad forest type (MDF), they varied substantially in their geology, and hence the potential for supporting bat species dependent on formations such as caves (Table 1). Two of the forests were associated with nutrient poor soils with a notable absence of cave or boulder formations. The first was Andulau Forest Reserve, which is one of the last remaining areas of a unique forest assemblage growing on yellow-sandy soils, and which is considered botanically ‘hyperdiverse’ by Ashton (2010). We surveyed an undisturbed area in the vicinity of Ashton’s original botanical plots, described in Sukri et al. 123 Biodivers Conserv KS 0 5 10 20 km SABAH MB LB SA RA WA DV SOUTH CHINA SEA K WEST KAL. CENTRAL KAL. EAST KAL. Bandar Seri Begawan SOUTH KAL. Tutong Bangar Kuala Belait Andulau Peradayan Merimbun Temburong Labi Teraja Ingei Fig. 1 Locality records of bat species in Brunei Darussalam, north Borneo, according to Brunei Museums records, the literature, and this study. Point size is proportional to the number of bat species reported from a site (i.e. site richness; range 1–30 species). All 61 localities (310 records) are at least 1 km apart. The six sampling sites from this study are indicated in large bold type, and four additional sites used for beta diversity analyses are shown in the inset (DV Danum Valley, KS Kabili-Sepilok, LB Loagan Bunut, MB Maliau Basin). Forest cover according to Miettinen et al. (2011) is indicated by grey shading (lowland, grey and upland, dark grey); major rivers are shown just for Brunei. The location of Brunei Darussalam (dark grey shading) is shown relative to the Heart of Borneo area (light grey shading) and other Borneo states in the inset (2011). The second site was within Tasek Merimbun Heritage Park in Tutong district, which encompasses 7,800 ha. Bat surveys here were undertaken along low-lying fringes of otherwise poorly drained peat soils. The fringes encompass a substantial area and are characterised by clay lithosols rather than the sandy soils described by Ashton (2010). The remaining four sites were geologically distinct from Andulau and Merimbun. Ulu Temburong National Park encompasses hilly terrain of approximately 50,000 ha in the Batu Apoi Forest Reserve in Temburong district, and is characterised by dark grey shale with occasional layers of sandstone and limestone (Sukri et al. 2011). We sampled bats 123 Biodivers Conserv Table 1 Protection status, geological features and bat assemblage characteristics of sites surveyed in Brunei Darussalam during 2009 and 2010 Site, protection status Soils Rock strata Trap nights na Catch rateb Sc Ed1/D 1. Andulau, forest reserve 4°390 N, 114°310 E Infertile yellow sandy latosols Soft sandstone/ clay 50 108 2.2 20 0.72 2. Tasek Merimbun, heritage park 4°350 N, 114°400 E Clay lithosols Thick clay/shale 50 119 2.4 15 0.67 3. Bukit Teraja, protection forest 4°18 N, 114°25 E Clay lithosols Thick clay/shale 11 124 11.3 15 0.15 4. Sungai Ingei, conservation forest 4°900 N, 114°420 E Yellow podsols Soft sandstone 20 531 26.6 22* 0.32 5. Peradayan, forest reserve 4°450 N, 115°100 E Shallow immature podsols Hard sandstone ridge outcrops 16 319 19.9 20 0.15 6. Ulu Temburong, national park 4°320 N, 115°120 E Yellow clay latosols Shale and clay 59 141 2.4 22 0.52 Geological information is based on Ashton (2010) and Ashton et al. (2003) a Number of individual bats captured in harp traps b Mean number of bats captured per trap c Total number of species captured in harp traps. Further species were captured in mist nets at some sites (see Table 2 and Supplementary Materials for full inventories) d Simpson evenness at a common sample size of 96 individuals derived by rarefaction. Values are calculated by the formula E = (1/D)/S when D is Simpson diversity and S is richness (both rarefied to common sample size) along the east ridge of Kuala Belalong Field Studies Centre, which has the largest bat inventory recorded from Brunei (Kofron 2002; Struebig et al. 2010). Peradayan Forest Reserve is near to Ulu Temburong, but is characterised by podsol soils, hard sandstone ridges and boulder formations. Although the reserve has been selectively logged in parts, the majority remains intact. Within the Peradayan Reserve bats were sampled at Bukit Patoi, an unlogged area of 1,027 ha. Sungei Ingei Conservation Forest remains a remote outpost of Brunei at the headwaters of the Belait river. This site is characterised by soft sandstone with some boulder formations leading to steep hills near the Sarawak border. Bat surveys were undertaken along national border patrol trails within 5 km of a temporary base-camp established by Brunei Museums department. The sixth site, Bukit Teraja, lies within 6800 ha of protected forest in Belait district and, unlike the previous three sites, is characterised by thick clay and shale, with some boulder formations. Surveys were undertaken along border patrol trails near the Teraja waterfall. Bat capture and identification Bats were captured in four-bank harp traps (Muzeum Zoologicum Bogoriense, Cibenong, Indonesia) using a standardised protocol successfully employed on Borneo and elsewhere in Southeast Asia (Kingston et al. 2003; Rossiter et al. 2012; Struebig et al. 2010, 2011). Traps were positioned across trails and streams each night, approximately 50 m apart, and checked in the evening and following morning within 2 h following dusk and dawn. Each trap was set for one night and then moved to a new position the following day to minimise trap shyness and to give a fully standardised trapping unit across nights. Between three and 123 Biodivers Conserv seven traps were set each night depending on the logistical demands of the survey. In Southeast Asia harp traps are most effective at capturing the insectivorous bats that forage in the forest-interior (sensu Kingston et al. 2003), and so this group of bats were the subject of subsequent analyses. Mist nets (70 denier nylon, four shelves, 30 mm mesh size, Ecotone, Gdynia, Poland) were set over rivers or streams at two sites (Ingei and Merimbun) to supplement bat inventories, but capture data were not included in analyses because of well-documented variation in capture success between traps and nets (Kingston et al. 2003). Bat surveys were undertaken in the dryer seasons of 2009 and 2010 (June to October). Because periods of heavy rain can adversely affect catch rates (Kingston, 2009), trapping in such periods was avoided. Collected bats were held individually in cloth bags and identified from external characters using a key based on Payne et al. (2000). Standard external data on sex, age and reproductive status as well as morphological measurements (weight; forearm length; tibia length), were obtained before marking each bat on the wing using a 3 mm diameter biopsy punch (Stiefel Laboratories, High Wycombe, UK) so that recaptures could be recognised and removed from analyses. We limited specimen collection to only vouchers of taxa new to the Brunei list or those of taxonomic uncertainty. In total 30 specimens were collected (typically only one or two individuals of a species), stored in 70 % ethanol and deposited in the faunal collection of Brunei Museums department. All other bats were released at the capture point unharmed within 12 h. Live trapping procedures and handling of bats followed guidelines of the American Society of Mammalogists (Sikes et al. 2011). Site level (alpha) diversity Because assemblage analyses can be strongly affected by sample size differences between sites, bat diversity at each location was estimated using rarefied species accumulation curves of site abundance data. To account for variation in bat density between sites, we undertook sample-based rarefaction rescaled to individuals in EstimateS (Colwell 2009), and observed site species richness was compared at the lowest common sample size (n = 96). Richness was also compared at a higher common sample size by predicting richness up to the level of the most abundant site (Ingei; n = 472). We used the multinomial model introduced by Shen et al. (2003), implemented in the software SPADE (Chao and Shen 2003–2005), whereby the number of new species recorded after additional sample effort is predicted. Although there are many other estimation or prediction methods available, they frequently become unstable and perform poorly when extrapolating to high sample sizes (Shen et al. 2003). The Shen predictor is relatively stable under these circumstances and has performed favourably when tested against other estimators using harp trap data from elsewhere in Southeast Asia (Kingston 2009). Moreover, we undertook preliminary analyses predicting to a lower sample size and the standard errors associated with the richness predictions were similar. To quantify assemblage evenness, we used the reciprocal form of the Simpson index corrected for species richness (E1/D = (1/D)/S), which is a robust measure of diversity weighted towards the influence of dominant species in the sample. High values therefore indicate even species abundances irrespective of the number of species. This metric was also rarefied to a common sample size in order to reliably compare evenness between sites. Site similarity and distance decay (beta diversity) Similarity indices are prone to the same sampling related biases as estimates of alpha diversity, and so to minimise this problem we used frequency-based indices that better 123 Biodivers Conserv account for variation in sample sizes between sites. The Morisita–Horn similarity index is based on the Simpson concentration and so is sensitive to the dominant species of assemblages, which results in undetected species (i.e. those expected to be rare) having little effect on the metric (Chao et al. 2008). Using this index had the added advantage that we were also able to calculate similarity across all sites using a recently developed probabilistic approach (Chao et al. 2008). Chao et al.’s version of the Morisita index estimates similarity for multiple assemblages based on shared information between any two assemblages. Differences in the magnitude of this similarity were assessed using 95 % confidence intervals based on 200 bootstrap replicates. We also repeated pair-wise analyses using a recently modified version of the Jaccard index, which extends this measure of similarity to use abundance data and account for missing species (Chao et al. 2005). Pairwise similarity matrices were calculated in EstimateS and visualised by presenting data from inventories as an unrooted tree produced by neighbour joining in the software Phylip v3.69 (http://www.phylip.com/). In order to improve our ability to detect distance decay, we also obtained and included data from four other undisturbed lowland sites in the neighbouring territories of Sabah and Sarawak (Fig. 1). These were Danum Valley Conservation Area (Kingston and Hodgkison 1994), Kabili-Sepilok Forest Reserve (Francis 1990), Loagan Bunut National Park (Gumal et al. 2008) and Maliau Basin Conservation Area (Struebig, Bernard and Turner, Unpublished Data). Crucial to our analyses, all of these inventories were undertaken in relatively undisturbed areas, in the same contiguous forest as the Brunei sites, and used comparable sampling protocols. To test for distance decay we assessed correlation between the matrices of pair-wise similarity values and those of geographic distances using a Mantel test in Genalex (Peakall and Smouse 2006) with 999 permutations. We also re-tested these distance-matrix correlations using the non-parametric RELATE test in Primer version 5 (Clarke and Warwick 2001), which does not assume linearity. Results Assemblage composition We captured a total of 1,362 individual bats of 35 species at the six Brunei study sites, the majority of which (1,249 bats of 27 species) were forest-interior insectivorous species (Tables 1, 2). The representation of bat families in assemblages varied over the Brunei forested landscape, with vespertilionids of the subfamilies Kerivoulinae and Murininae typically well represented at most sites, and other families (Hipposideridae and Rhinolophidae) being more patchily represented across sites (Table 2; Fig. 2). Nonetheless in terms of total captures, hipposiderid and rhinolophid species that predominantly roost in caves were the most abundant taxa (894 individuals of ten species over the six sites), and hipposiderids in particular contributed the most to high capture rates and low evenness at Ingei, Peradayan and Teraja (Table 1; Fig. 2). Conversely, bats from these families were often present but comparatively rare in the forest of Andulau and Merimbun (Table 2; Fig. 2). Dominant species of this ensemble included Hipposideros cervinus (abundant at Ingei, Peradayan, Temburong and Teraja), and Rhinolophus creaghi (highly abundant at Ingei, and representing the first record of this species in Brunei). Rare cave-roosting species included H. ater and Megaderma spasma, both of which were only captured once, at Temburong and Peradayan respectively. 123 123 Table 2 Bats species reported from the six survey sites in Brunei Darussalam, including the total number of species and individuals captured in this study Family, speciesa Foraging strategyd Roosting ecologye Aethelops aequalis (alecto) Bf T Balionycteris maculata Bf Chironax melanocephalus Cynopterus brachyotis Andulau Merimbun Teraja Ingei Peradayan Temburong T 5 8 1 8 3 2 Bf T 1 Bf T 3 2 Cynopterus minutus Bf T Dyacopterus spadiceus Cf C X Eonycteris major Cf C X Eonycteris spelaea Cf C Macroglossus minimus Bf T X Megaerops ecaudatus Bf T X X X Megaerops wetmorei Bf T X X X Penthetor lucasi Bf/Cf C Pteropus vampyrus Cf T Rousettus amplexicaudatus Cf C Rousettus spinalatus* Cf C Emballonura alecto Ei T X X Emballonura monticola Ei T 1 13 PTEROPODIDAE (fruit bats) X 1 X X X X X X X X X EMBALLONURIDAE (sheath-tailed bats) MOLLOSIDAE (free-tailed bats) Oi C Oi C Ni T/C Ni T Ni C X X X MEGADERMATIDAE (false vampires) Megaderma spasma 1 NYCTERIDAE (hollow-faced bats) Nycteris tragata (javanica) 1 1 1 HIPPOSIDERIDAE (roundleaf bats) Hipposideros ater* 1 Biodivers Conserv Cheiromeles torquatus Mops mops Family, speciesa Foraging strategyd Roosting ecologye Hipposideros bicolor Ni C Hipposideros cervinus Ni C Hipposideros cineraceus Ni C Hipposideros diadema Ni C Hipposideros doriae Ni T Hipposideros dyacorum Ni C Hipposideros galeritus Ni C Hipposideros ridleyi Ni T Rhinolophus borneensis Ni C Rhinolophus creaghi* Ni C Andulau Merimbun Teraja Ingei Peradayan Temburong 91 5 1 2 199 266 27 X 8 X 4 11 2 3 2 15 4 3 6 11 5 2 2 9 5 52 5 13 RHINOLOPHIDAE (horseshoe bats) 2 167 Rhinolophus philippinensis Ni C Rhinolophus sedulus Ni T 13 5 1 X 2 10 Rhinolophus trifoliatus Ni T 6 14 1 1 4 3 Kerivoula hardwickii Ni T 6 24 1 1 3 1 Kerivoula intermediab Ni T 17 5 1 8 10 Kerivoula lenisc* Ni T 5 5 Kerivoula minutab* Ni T 4 8 1 15 2 4 Kerivoula papillosac Ni T 9 19 9 4 Kerivoula pellucida* Ni T 3 2 3 1 3 Kerivoula sp.c* Ni T 1 5 10 9 4 Kerivoula whiteheadi Ni T X X VERPERTILIONIDAE (evening bats) Kerivoulinae (woolly bats) 123 Phoniscus atrox * Ni T Phoniscus jagorii* Ni T 1 Ni T 1 1 2 1 Murininae (tube-nosed bats) Murina aenea* 1 6 21 1 Biodivers Conserv Table 2 continued 123 Table 2 continued Family, speciesa Foraging strategyd Roosting ecologye Murina cyclotis Ni T Murina rozendaali* Ni T 1 Murina suilla* Ni T 7 Glischropus tylopus Ei T 5 Myotis ater* Ei C/T Andulau Merimbun Teraja X 4 Ingei Peradayan Temburong 1 5 1 4 1 4 28 Vespertilioninae (evening bats) Myotis muricola Ei T Myotis ridleyi Ei T Pipistrellus tenuis Ei C/T Hypsugo (Pipistrellus) vordermanni Ei C/T Total bats species reported: 2 1 1 X 1 X 1 2 X 25 20 X X 23 25 17 39 Additional species captured in mist-nets in this study or reported elsewhere in the literature are indicated by X. New additions to the Brunei Darussalam bat species inventory are indicated by asterisks. Previous nomenclature for species are indicated by parentheses a Nomenclature follows Simmons (2005). Older species names used in Payne et al. (2000) are listed in parentheses following the species name b Kerivoula intermedia and minuta are difficult to distinguish in the hand and were therefore separated by body mass (K. intermedia [ 2.75 g; range: 2.75–4.0 g; K. minuta \ 2.5 g; range: 2.0–2.5 g) c Bats historically defined as Kerivoula papillosa exhibit extensive morphological variation on Borneo and elsewhere in Southeast Asia, and so the concept of K. papillosa includes multiple species (Khan et al. 2010). We defined three taxa separated by forearm length, examples of which were later confirmed as distinct taxa by genetic barcodes. K. lenis \ 40 mm; K. papillosa 40-44 mm; Kerivoula spp. [ 45 mm. The molecular systematics of this species complex is under ongoing investigation d Foraging strategies assigned based on wing morphology (Kingston et al. 2003; Struebig et al., 2010): Bf, frugivorous or nectarivorous species that forage in clutter below the forest canopy; Of, frugivorous or nectarivorous species that forage in open areas and over large distances; Ei, insectivorous species that forage in partially cluttered edges and/or canopy edges; Oi, insectivorous species that forage in open areas and over large distances; Ni, forest-interior insectivorous species that typically forage in narrow-spaces or clutter (although they may commute some distance to foraging sites). Foraging strategies are used in combination with roosting ecology to assign species to ensembles Biodivers Conserv e The main roost structures reported in the literature and by personal observations for each bat species (Kofron, 2002; Payne et al. 2000; Struebig et al., 2010; Yasuma and Andau, 2000): T, ephemeral roosts in trees, hollows or other foliage; C, caves, boulders, houses. Roosting ecology is used in combination with foraging strategies to assign species to ensembles. When two categories are listed the first the predominant observation from the literature Biodivers Conserv a b Andulau c Merimbun Ingei Murininae Vespertilioninae Kerivoulinae Rhinolophidae Pteropodidae Hipposideridae Other 0.0 d 1.0 2.0 3.0 4.0 0.0 e Peradayan 1.0 2.0 3.0 4.0 0.0 f Teraja 1.0 2.0 3.0 4.0 2.0 3.0 4.0 Temburong Murininae Vespertilioninae Kerivoulinae Rhinolophidae Pteropodidae Hipposideridae Other 0.0 1.0 2.0 3.0 4.0 0.0 1.0 2.0 3.0 4.0 0.0 1.0 Capture rate (abundance/trap nights) Fig. 2 Abundance per unit effort of the main bat families (Rhinolophidae, Pteropodidae, Hipposideridae) and sub-families (Murininae, Kerivoulinae, Vespertilioninae) captured in harp-trap inventories of the six Brunei sites. Open bars indicate that capture rates exceeded 4.0 bats per trap Tree cavity/foliage-roosting species represented a lower proportion of overall captures (375 bats; 18 species), but were more uniformly distributed across sites (Table 2; Fig. 2). Dominant species of this ensemble included Kerivoula papillosa, K. minuta and K. intermedia, which were well represented at most sites surveyed. Rare species included Murina rozendaali and Phoniscus jagorii (both singletons, recorded from a single capture each in Andulau for the first time in Brunei), as well as P. atrox, Murina aenea and H. doriae, which were recorded by only two captures each. Pooling forest-interior bat capture data from the six sites generated a species accumulation curve that had clearly passed the point of inflection and approached an asymptote, suggesting that the majority of bat species of this ensemble had been captured. (Fig. 3a). Alpha diversity When species richness estimates were rarefied down to a common sample size, Andulau and Temburong were ranked as sites with the most bat species (Fig. 3b, c), hosting significantly more species than Ingei, Peradayan and Teraja. However, when accounting for species not yet recorded in inventories (using the Shen multinomial predictor), Andulau was outranked by Ingei, Temburong and Teraja, supporting significantly fewer species than the latter two sites (Fig. 3d). Contrary to expectations, Merimbun supported the fewest predicted bat species overall (12-13 species; Fig. 3d), but together with Andulau and Temburong exhibited high species evenness (Table 1). Notably, the Teraja inventory 123 Biodivers Conserv Species a b Pooled rarefied species accumulation curve for Brunei 35 35 30 30 25 25 20 20 15 15 10 10 5 5 Andulau Peradayan Merimbun Ingei Temburong Teraja 0 0 0 200 400 600 800 1000 1200 0 100 Individuals c Species Rarefied species accumulation curves for Brunei sites d Rarefied species richness (n=96) 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 200 300 400 Individuals Predicted species richness (n=472) 0 Andulau Merimbun Ingei Peradayan Teraja Temburong Site Andulau Merimbun Ingei Peradayan Teraja Temburong Site Fig. 3 Species accumulation and richness of forest-interior insectivorous bats sampled at six forest sites in Brunei. Accumulation curve and singletons curve for all sites combined (a), and accumulation curves for sites independently (b) are derived by sample-based rarefaction re-scaled to individuals. Comparisons of richness are made at the lowest common sample size (c), and also at the sample size of Sungai Ingei, the site with the greatest number of individuals captured (d). Error bars (c, d) indicate the 95 % confidence limits of mean diversity derived from rarefaction (c, sensu Colwell 2009) or prediction (d, sensu Shen et al. 2003) at n individuals. Values falling outside of the confidence limits of other sites are significantly different. Open circles on c and d indicate the observed richness (i.e. non-rarefied) of all captured forest-interior bats exhibited low richness and evenness (Table 1), resulting in a species accumulation curve that intersected those of other sites, thus indicating that more species may actually be present than our sampling uncovered and that diversity comparisons with this curve should be interpreted with caution. Beta diversity Overall similarity in bat species composition was low to moderate (Chao-Morisita similarity amongst Brunei sites = 0.422 ± 0.014 SE; all ten sites, 0.529 ± 0.012) indicating that at least some species were poorly represented across all of the sites studied. Pair-wise dissimilarity coefficients revealed two patterns of interest reflected by branches in our unrooted tree that were highly differentiated from each other in terms of assemblage composition (Fig. 4). The two nutrient poor forests sampled in Brunei (Andulau and 123 Biodivers Conserv Ingei Danum Sepilok Peradayan Teraja Maliau Andulau Merimbun Bunut Temburong Fig. 4 Unrooted tree showing similarity of forest-interior insectivorous bat assemblages at 10 forested sites in northern Borneo (Brunei, Sabah, and Sarawak) based on the Morisita similarity index. Two main branch clusters are evident: the nutrient poor forests of Andulau and Merimbun, as well as Bunut, host similar assemblages (dominated by tree/cavity-roosting bat species), as do Ingei, Teraja and Peradayan in Brunei, and Danum and Sepilok in Sabah (dominated by cave-roosting bat species). Temburong, the most species diverse assemblage overall, is intermediate between the two clusters. The six Brunei sites sampled during this study are in bold black type; additional sites from Sabah and Sarawak are shown in grey italics Merimbun) supported similar bat assemblages to each other and also to that of Bunut, another nutrient poor forest area, reflecting the even abundance distribution of tree cavity/ foliage-roosting species of the vespertilionid subfamilies Kerivoulinae and Murininae (Fig. 2a, b). A second branch was represented by the Brunei sites Ingei, Peradayan and Teraja, as well as Danum and Sepilok in Sabah; all dominated by the cave-roosting species, H. cervinus. In fact Peradayan and Teraja, which were characterised by boulder formations, supported near-identical bat assemblages (Morisita similarity = 0.99). The Temburong assemblage was intermediate between these two groupings, hosting the majority of species from the other sites but at more even abundances. When the nutrient poor sites were removed from analysis, overall similarity amongst sites was significantly greater (for Brunei, 0.687 (CI: 0.631–0.743); including other sites, 0.656 (CI: 0.622–0.689). Mantel and RELATE tests did not reveal evidence of distance decay, suggesting that assemblage structure was poorly influenced by neutral processes at the spatial scale of our study. Assemblage dissimilarity was not significantly correlated with geographic distance whether determined by Morisita coefficients (R2MANTEL = 0.018, P = 0.177; RhoRELATE = -0.233, P = 0.083) or Chao-Jaccard coefficients (R2MANTEL = 0.039, P = 0.116; RhoRELATE = -0.207, P = 0.058). This finding was consistent when geographic distances were logtransformed and similarity coefficients calculated on square-root transformed abundance data to reduce the skew of highly abundant species. Discussion Our bat surveys in 380,000 ha of largely undisturbed Bornean rainforest revealed substantial diversity and subtle differences in assemblage composition between sites. We recorded 35 bat species at six sites in Brunei, which added 15 more species to the national species list (see Supplementary Material). Of these 35 species, 27 were insectivorous forest-interior specialists. A species accumulation curve summarising assemblage data 123 Biodivers Conserv from these six sites suggested that the landscape-scale inventory for forest-interior bats was near-complete. Indeed the Borneo-wide total for this foraging group is in the region of 35 species (Struebig et al. 2010). The presence of four singletons in the Brunei inventory (Fig. 3a), and the presence of three taxa reported as singletons in the four inventories used for distance decay analyses, indicates that forest-interior bat species remaining to be uncovered in the Brunei region are likely to be rare. On the other hand, there are a substantial number of vespertilionids known from other forested areas on Borneo that were not well represented in the Brunei inventory (particularly pipistrelles, Tylonycteris and Hesperoptenus species). These bats, of which there at least 20 reported from Borneo, are known to forage around edges, canopies and open spaces, and are typically difficult to capture with harp traps. Therefore underrepresentation in our inventories does not necessarily infer rarity. In order to sample these species effectively future surveys would benefit from acoustic monitoring techniques and mist-netting over water bodies. The pattern of bat assemblage composition across north Borneo’s forested landscape was mostly driven by distinct assemblages at a small number of sites. Relatively depauperate bat assemblages in nutrient-poor forests contributed substantially to the signal of landscape-wide beta diversity, and when these sites were removed from analyses the similarity in bat species composition across the forested landscape was notably greater. While Andulau and Merimbun were predicted to support moderate numbers of forest-interior bat species (13–18 according to Shen multinomial predictions), the observed species composition was typically a subset of otherwise richer sites further inland. Two tree cavity/foliage-roosting species (Murina rozendaali and Phoniscus jagorii) found at the nutrient poor sites during our study were notably rarely captured and are only known from a handful of records on Borneo (Struebig et al. 2010), but we believe it likely that further surveys would reveal these species in forests elsewhere in the region. On the other hand, locality records from this study and elsewhere on Borneo (Struebig et al. 2010) suggest that another species (Kerivoula whiteheadi) could be largely restricted to such habitat, as all confirmed records are from forests on peat or sandy soils. The main differences in species composition between sites were driven by the presence or absence of dominant cave-roosting species (e.g. Hipposideros cervinus). The proximity of a bat survey to major roosting sites can have a strong bearing on the assemblage composition described, with the dominance of cave-roosting species detectable several kilometres away from a large bat roost (Struebig et al. 2009). Once these species are recognised and accounted for in assemblage analyses—for example by transforming species abundances or by removing selected species (McCune and Grace 2002)—bat assemblage composition in undisturbed palaeotropical forests, at least in Borneo, would appear relatively homogenous in the absence of environmental gradients. Our inability to detect a significant pattern of distance decay between assemblage similarity and geographic distance supports the idea that, at the spatial scale studied here, the neutral processes of ecological drift and differential dispersal capabilities do not have a major role in shaping bat assemblages in the undisturbed forests on Borneo. This is in contrast with the findings from Bornean moths (Beck and Khen 2007) and butterflies (Cleary and Genner 2006), for which assemblage composition appears to be partly influenced by geographical distance. Notably, much of the distance decay detected by these studies was observed over short geographic distances (\20 km), which were not well covered in our study. However, in a previous study of assemblage similarity over comparable geographic distances in peninsular Malaysia, bat assemblages also exhibited little differentiation (Struebig et al. 2011). Together with the results of this study this finding suggests that the dispersal limitations expected to shape beta diversity are not strong enough to significantly influence palaeotropical bat assemblages in continuous habitat, at 123 Biodivers Conserv least across the distances examined (ca. 500 km). It has recently been suggested that low beta diversity (and hence weak distance decay) could be generally expected for tropical taxa, particularly those in homogeneous rainforest habitats with weak gradients in climatic, altitudinal and geological characteristics (Novotny et al. 2007). The homogeneity of bat species composition in undisturbed continuous habitat indicates that historical processes can be largely discounted from shaping assemblages, and thus structural differences between assemblages may be reliably attributed to other processes such as those associated with habitat disturbance. This is particularly relevant when identifying undisturbed ‘controls’ for studying the impact of environmental change on this animal group. Our results therefore confirm the validity of recent studies that have compared bat assemblages in disturbed habitats with those in such undisturbed controls (e.g. Furey et al. 2010; Phommexay et al. 2011; Struebig et al. 2008, 2011). The north of Borneo is a species-diverse region and therefore a priority area for conservation efforts. Our findings lend strong support for improving the conservation status of Brunei’s inland forests, which otherwise remain inadequately protected. The inland forests of Brunei and eastern Sarawak support some of the most diverse bat assemblages on Borneo and are comparable in assemblage structure to well-known forest sites in Sabah (e.g. Danum Valley, Maliau Basin, as shown here). For example, the forests of Sungai Ingei host a particularly diverse bat community that includes several species not yet reported from elsewhere in Brunei and only patchily distributed on Borneo (Rhinolophus creaghi, R. acuminatus, and Rousettus spinalatus). On a national level this area is of high conservation priority based on faunal richness. In addition, homogeneity of bat assemblage structure over the extensive forest area in this study suggests substantial interconnectedness among sites, thus indicating that a large forest area should be conserved to maintain natural ecological processes. Although our species inventory for one of the inland sites (Teraja) was notably low, the low evenness and lack of inflection on the species accumulation curve suggests that substantially more species remain to be uncovered and so bat diversity is likely to be comparable to other inland sites. Conversely, we also show that the largely coastal, nutrient-poor forests of Brunei (Andulau and Merimbun) and Sarawak (Bunut) host relatively depauperate bat assemblages. These forests support a number of rare and threatened species (Murina rozendaali, M. aenea and Hipposideros ridleyi; all classified as Vulnerable (IUCN 2011), but overall bat diversity is low with bat species composition similar to peat swamps and heath forests elsewhere on Borneo (Struebig et al. 2006). Our findings are particularly timely given calls to extend the HoB area to the coastal region of Brunei so that botanically ‘hyperdiverse’ forests can be better represented in the conservation area. These forests are clearly a rare habitat-type: they support the highest levels of tree diversity found in the palaeotropics, appear to be unique to the north and west of Borneo (Ashton 2010), and, following decades of clearance and fragmentation, they now have their largest remnants in Brunei and eastern Sarawak. However, we show that botanical hyperdiversity does not necessarily support high levels of faunal richness. It remains to be seen whether the patterns of bat diversity we describe in these forests is matched by similar trends in other animal groups, though we note that the accessibility of coastal habitats makes them particularly susceptible to disturbance and high hunting pressure. Indeed, other botanically hyperdiverse parks are being rapidly depleted of vertebrates (e.g. Lambir National Park in Sarawak, Harrison 2011). Broadening the geographic scope of the HoB is clearly warranted to improve botanical representation and the protection of a larger habitat area. However such proposals should not detract attention from intact interior forests that are associated with high levels of faunal diversity. 123 Biodivers Conserv Acknowledgments The bulk of this work was supported by a Universiti Brunei Darussalam (UBD) grant awarded to MJS whilst a Research Fellow at the university, with additional support from the Queen Mary University of London (QMUL) Expedition Fund to JH. MJS is particularly grateful to Kamariah Abu Salim and the UBD Research Committee for approving the research. During this project we were grateful to the help of many individuals from UBD, Brunei Museums, Forestry Department and local communities. We are particularly grateful to Hj Saidin Bin Salleh, Director of Forestry and Bantong Bin Antaran, Director of Brunei Museums, for granting us permission to access research sites, and to Siti Norhayatty Morni for facilitating access. Special thanks go to Joe Charles, Ang Bee Biaw and Samhan Nyawa, organisers of the Sungai Ingei Faunal Expedition (2010–2012), which was sponsored by Standard Chartered Bank and supported by the Ministry of Industry & Primary Resources, UBD and the World Wide Fund for Nature. Finally we would like to thank Farah Anie, Ulmar Grafe, Caroline Sch}oner and Michael Sch}oner, staff of Tasek Merimbun and Kuala Belalong Field Studies Centre for helping with and facilitating fieldwork, as well as the students of the QMUL Tropical Ecology field-courses, for assisting MJS at Tasek Merimbun. References Ashton P (2010) Conservation of Borneo biodiversity: do small lowland parks have a role, or are big inland sanctuaries sufficient? Brunei as an example. Biodivers Conserv 19:343–356 Ashton PS, Kamariah KS, Said IM (2003) A field guide to the forest trees of Brunei Darussalam and the northwest Borneo hotspot. Universiti Brunei Darussalam, Bandar Seri Begawan Beck J, Khen CV (2007) Beta-diversity of geometrid moths from northern Borneo: effects of habitat, time and space. J Anim Ecol 76:230–237 Beck J, Schwanghart W, Khen CV, Holloway JD (2011) Predicting geometrid moth diversity in the Heart of Borneo. Insect Conserv Divers 4:173–183 Catullo G, Masi M, Falcucci A, Maiorano L, Rondinini C, Boitani L (2008) A gap analysis of Southeast Asian mammals based on habitat suitability models. Biol Conserv 141:2730–2744 Chao A, Shen TJ (2003–2005) Program SPADE (Species Prediction And Diversity Estimation). Program and user’s guide published at http://chao.stat.nthu.edu.tw. Accessed 10 May 2011 Chao A, Chazdon RL, Colwell RK, Shen TJ (2005) A new statistical approach for assessing similarity of species composition with incidence and abundance data. Ecol Lett 8:148–159 Chao A, Jost L, Chiang SC, Jiang J-H, Chazdon RL (2008) A two-stage probabilistic approach to multiplecommunity similarity indices. Biometrics 64:1178–1186 Chase JM (2005) Towards a really unified theory for metacommunities. Funct Ecol 19:182–186 Clarke KR, Warwick RM (2001) Change in marine communities: an approach to statistical analysis and interpretation. Primer-E Ltd, Plymouth Cleary DFR, Genner MJ (2006) Diversity patterns of Bornean butterfly assemblages. Biodivers Conserv 15:517–538 Colwell RK (2009) EstimateS: statistical estimation of species richness and shared species from samples. Version 8.2. Available at http://purl.oclc.org/estimates. Accessed July 2011. Available from purl.oclc.org/estimates Findley JS (1993) Bats: a community perspective. Cambridge University Press, Cambridge Fitzherbert EB, Struebig M, Morel A et al (2008) How will oil palm expansion affect biodiversity? Trends Ecol Evol 23:538–545 Francis CM (1990) Trophic structure of bat communities in the understorey of lowland dipterocarp rain forest in Malaysia. J Trop Ecol 6:421–431 Francis CM (1994) Vertical stratification of fruit bats (Pteropodidae) in lowland dipterocarp rainforest in Malaysia. J Trop Ecol 10:523–530 Furey NM, Mackie IJ, Racey P (2010) Bat diversity in Vietnamese limestone karst areas and the implications of forest degradation. Biodivers Conserv 19:1821–1838 Gumal MT, Hon J, Kong D (2008) Birds and bats of Loagan Bunut National Park Sarawak, Malaysia. Wildlife Conservation Society, UNDP/GEF Peat Swamp Forest Technical Series No. 11 (MAL/99/G31), Kepong Harrison R (2011) Emptying the forest: hunting and the extirpation of wildlife from tropical nature reserves. Bioscience 61(11):919–924 IUCN (2011) IUCN Red list of threatened species. Version 2011.2. World Conservation Union (IUCN), Gland, Switzerland. www.iucnredlist.org. Downloaded on 08 February 2012 Khan FAA, Solari S, Swier VJ, Larsen PA, Abdullah M, Baker RJ (2010) Systematics of Malaysian woolly bats (Vespertilionidae: Kerivoula) inferred from mitochondrial, nuclear, karyotypic, and morphological datasets. J Mammal 91(5):1058–1072 Kingston T (2009) Analysis of species diversity of bat assemblages. In: Kunz TH, Parsons S (eds) Ecological and behavioral methods for the study of bats, 2nd edn. Johns Hopkins University Press, Washington, pp 195–215 123 Biodivers Conserv Kingston T, Hodgkison R (1994) Malaysian fruit bat project ‘94. Unpublished report to the Royal Geographical Society, London Kingston T, Francis CM, Zubaid A, Kunz TH (2003) Species richness in an insectivorous bat assemblage from Malaysia. J Trop Ecol 19:67–79 Kofron CP (2002) The bats of Brunei Darussalam, Borneo. Mammalia 66:259–274 Lane DJW, Kingston T, Lee BPY-H (2006) Dramatic decline in bat species richness in Singapore, with implications for Southeast Asia. Biol Conserv 131:584–593 Mackinnon K, Hatta G, Halim H, Mangalik A (1996) The ecology of Kalmantan Indonesian Borneo. Periplus Editions, Singapore McCune B, Grace JB (2002) Analysis if ecological communities. MJM Software Design, Oregon Meijaard E, Nijman V (2003) Primate hotspots on Borneo: predictive value for general biodiversity and the effects of taxonomy. Conserv Biol 17:725–732 Miettinen J, Shi C, Tan WJ, Liew SC (2011) 2010 land cover map of insular Southeast Asia in 250-m spatial resolution. Remote Sens Lett 3(1):11–20 Novotny V, Miller SE, Hulcr J et al (2007) Low beta diversity of herbivorous insects in tropical forests. Nature 448:692–695 Payne J, Francis CM, Phillips K, Kartikasari SN (2000) Mamalia di Kalimantan, Sabah. Sarawak dan Brunei Darussalam, Wildlife Conservation Society Indonesia Program, Jakarta Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes 6:288–295 Phommexay P, Satasook C, Bates P, Pearch M, Bumrungsri S (2011) The impact of rubber plantations on the diversity and activity of understorey insectivorous bats in southern Thailand. Biodivers Conserv 20:1441–1456 Raes N, Roos MC, Slik JWF, Van Loon EE, Ter Steege H (2009) Botanical richness and endemicity patterns of Borneo derived from species distribution models. Ecography 32:180–192 Rossiter SJ, Zubaid A, Mohd-Adnan A et al (2012) Social organization and genetic structure: insights from codistributed bat populations. Mol Ecol 21:647–661 Shen TJ, Chao A, Lin JF (2003) Predicting the number of new species in further taxonomic sampling. Ecology 84:798–804 Sikes RS, Gannon WL, Animal Care and Use Committee of the American Society of Mammalogists (2011) Guidelines of the American Society of Mammalogists for the use of wild mammals in research. J Mamm 92:235–253 Simmons N (2005) Chiroptera. In: Wilson D. E. andReeder D. M. (eds), Mammal species of the world: a taxonomic and geographic reference, 3rd edn. John Hopkins University Press, Baltimore, pp 312–529 Sodhi NS, Koh LP, Brook BW, Ng PKL (2004) Southeast Asian biodiversity: an impending disaster. Trends Ecol Evol 19:654–659 Soininen J, McDonald R, Hillebrand H (2007) The distance decay of similarity in ecological communities. Ecography 30:3–12 Struebig MJ, Galdikas BMF, Suatma (2006) Bat diversity in oligotrophic forests of southern Borneo. Oryx 40:447–455 Struebig MJ, Kingston T, Zubaid A, Mohd-Adnan A, Rossiter SJ (2008) Conservation value of forest fragments to Palaeotropical bats. Biol Conserv 141:2112–2126 Struebig MJ, Kingston T, Zubaid A et al (2009) Conservation importance of limestone karst outcrops for Palaeotropical bats in a fragmented landscape. Biol Conserv 142:2089–2096 Struebig MJ, Christy L, Pio D, Meijaard E (2010) Bats of Borneo: diversity, distributions and representation in protected areas. Biodivers Conserv 19:449–469 Struebig M, Kingston T, Petit E et al (2011) Parallel declines of species and genetic diversity in tropical forest fragments. Ecol Lett 14:582–590 Sukri RS, Wahab RA, Salim KA, Burslem DFRP (2011) Habitat associations and community structure of dipterocarps in response to environment and soil conditions in Brunei Darussalam, northwest Borneo. Biotropica. doi: 10.1111/j.1744-7429.2011.00837.x WWF (2011) Business Solutions: Delivering the Heart of Borneo Declaration. WWF Green Network Report, p 82 Yasuma S, Abdullah MA (1997) An invitation to the mammals of Brunei Darussalam: part II. outdoor information. Japan International Cooperation Agency and Forestry Department of Brunei Darussalam, Bandar Seri Begawan Yasuma S, Andau M (2000) Mammals of Sabah. JICA-Sabah Wildlife Department, Malaysia, Kota Kinabalu 123