In the face of the current global extinction crisis (Bradshaw et al., Reference Bradshaw, Sodhi and Brook2009; Ceballos et al., Reference Ceballos, Ehrlich and Dirzo2017), it is vital to establish protected areas that can serve as biological reservoirs for species and ecological processes (Chape et al., Reference Chape, Harrison, Spalding and Lysenko2005; Jenkins & Joppa, Reference Jenkins and Joppa2009; Le Saout et al., Reference Le Saout, Hoffmann, Shi, Hughes, Bernard and Brooks2013; Venter et al., Reference Venter, Fuller, Segan, Carwardine, Brooks and Butchart2014). Governments, who are primarily responsible for creating and maintaining protected areas, report their progress by reference to Aichi Target 11 of the Convention on Biological Diversity (CBD). However, with the majority of biodiversity concentrated in the tropics, the pressure on developing nations across the equatorial belt to fulfil their CBD obligations is overwhelming (Chandra & Idrisova, Reference Chandra and Idrisova2011; Harrop & Pritchard, Reference Harrop and Pritchard2011; Mallari et al., Reference Mallari, Collar, McGowan and Marsden2016; Jones et al., Reference Jones, Venter, Fuller, Allan, Maxwell, Negret and Watson2018). Indonesia is a so-called megadiverse country, with particularly high biological richness on the islands of Borneo, Sumatra and Java, and is considered one of the top three biodiversity hotspots globally because of high levels of both endemism and habitat loss (Mittermeier & Mittermeier, Reference Mittermeier and Mittermeier1997; Myers et al., Reference Myers, Mittermeier, Mittermeier, da Fonseca and Kent2000). Conserving biodiversity on Java is a particular challenge, with a human population of c. 145 million and a population density of > 1,110 people/km2 (Badan Pusat Statistik, 2016). Biological endemism is concentrated on the island's volcanic mountains, particularly in the western part where the equatorial climate delivers high levels of rain to the forests that cover the slopes of c. 20 mountains (Whitten et al., Reference Whitten, Soeriaatmadja and Afiff1997). Forest cover in the area declined by c. 40% during 1988–2000 and then stabilized at c. 5,200 km2. The remaining forests are highly fragmented and in need of protection (Higginbottom et al., Reference Higginbottom, Collar, Symeonakis and Marsden2019).
The mountains in western Java harbour numerous endemic and Sundaic species (Whitten et al., Reference Whitten, Soeriaatmadja and Afiff1997). The easternmost of these mountains, Gunung Slamet, in western Central Java province, shares many taxa with the mountains further west but also possesses a degree of endemism of its own, such as a distinctive subspecies of the Javan laughingthrush Garrulax rufifrons slamatensis (Collar & van Balen, Reference Collar and van Balen2013). Its forests, like those in other mountain areas in Central Java, transition from tropical in the west to monsoon in the east (Widhiono, Reference Widhiono2015), and comprise lowland, montane and subalpine types (Soemarno & Girmansyah, Reference Soemarno, Girmansyah, Maryanto, Noerdjito and Partomihardjo2012). The biological richness of Gunung Slamet was first described during Indonesia's colonial period. Since 2000, there have been studies of ferns (Praptosuwiryo, Reference Praptosuwiryo2013), termites (Pribadi et al., Reference Pribadi, Raffiudin and Harahap2011), butterflies (Widhiono, Reference Widhiono2015), bees and wasps (Widhiono et al., Reference Widhiono, Sudiana and Sucianto2016), birds (Widodo, Reference Widodo2010, Reference Widodo, Maryanto, Noerdjito and Partomihardjo2012), the Javan surili Presbytis comata (Setiawan et al., Reference Setiawan, Djuwantoko, Bintari, Kusuma, Pudyatmoko and Imron2007), Javan gibbon Hylobates moloch (Setiawan et al., Reference Setiawan, Nugroho, Wibisono, Ikawati and Sugardjito2012; Wahyuni & Nasution, Reference Wahyuni and Nasution2016) and Javan leopard Panthera pardus melas (Wibisono et al., Reference Wibisono, Wahyudi, Wilianto, Pinondang, Primajati, Liswanto and Linkie2018). In 2012 the Indonesian Institute of Sciences published a compilation of studies on the mountain's geology, climatology and biodiversity (Maryanto et al., Reference Maryanto, Noerdjito and Partomihardjo2012).
When Indonesia's protected area system underwent a review in the early 1980s, 150 km2 of natural habitat at 1,000–3,418 m altitude on Gunung Slamet were recommended for protection as a nature reserve (MacKinnon, Reference MacKinnon1982; MacKinnon et al., Reference MacKinnon, Smiet and Artha1982; Nijman & Sözer, Reference Nijman and Sözer1996; Sözer et al., Reference Sözer, Nijman, van Balen, Setiawan, Prawiradilaga and Subijanto1997). However, no action was taken on this recommendation and forest areas on the mountain, which serve as hydrological protection forest for five rivers flowing north and five flowing south, are currently managed by the State Forest Management Agency (Perum Perhutani). The absence of formal protection for its biodiversity has left Gunung Slamet threatened by encroachment of human activities (Setiawan et al., Reference Setiawan, Nugroho, Wibisono, Ikawati and Sugardjito2012). As part of a wider initiative to document the biodiversity of western Java's montane forests and to make recommendations for implementing a protected area status, we undertook field research on Gunung Slamet in 2018, to assess its biological diversity and identify priority conservation areas. Here, we present data on the distribution and relative abundance of key faunal taxa, and on forest cover, forest loss and threats such as bird trapping.
Gunung Slamet, an active stratovolcano in Purbalingga Regency in Central Java (Fig. 1), is the second highest mountain on Java, at 3,432 m (Maryanto et al., Reference Maryanto, Noerdjito and Partomihardjo2012). It receives an annual rainfall of c. 4,500 mm, which is among the highest precipitation levels in Indonesia. The rainy season is October–May/June, and the dry season July–September. Daytime temperatures are 20–28 °C (Widhiono, Reference Widhiono2015). The volcano last erupted in 2014, and previously in 2009, with some forest burnt as a result. We worked at six study sites on accessible trails leading up to the summit, at altitudes of 970–2,512 m (Table 1, Fig. 1), avoiding the most heavily used tourist trail and the site of a recent forest fire on the eastern side of the mountain, and the restricted area surrounding the proposed Baturraden geothermal plant on the western side.
We carried out fieldwork during September–December 2018, using three survey methods: camera traps, walking transects along trails to record birds, and searches for amphibians, complemented by vegetation surveys along transect lines. We positioned 20 camera traps along trails, c. 150–200 m apart, at locations with signs of activities of terrestrial mammals such as scratch marks, rubs, scats and animal trails (Fig. 1). The camera traps were fastened to trees, c. 0.5 m above the ground and set to operate continuously for 3 days at each site, recording photographs and video footage, for a total of 360 camera-days. Our camera setup did not target small mammals such as rodents and bats. To estimate bird encounter rates, we walked a mean of 10 transects of variable length and duration along trails (including those where we set up camera traps) at all sites apart from Ketenger 1, noting bird species seen and/or heard, number of individuals per group and time of day. We generally walked transects between 6.00 and 9.00, at a speed of c. 1 km/h. We surveyed for amphibians along streams and in wetter areas surrounding the camp and trails for c. 3 hours after dusk for 3 days per site, recording amphibian presence and habitat type with photographs. We surveyed vegetation in plots with a radius of 10 m, every 200 m along transect lines. At each site we measured the girth at breast height of the three largest trees, and recorded presence of palms and other indicator species, ground cover, and evidence of human disturbance such as cut mist-nets, tree stumps and hunter/trapper camps.
We aggregated bird encounter rates at each site, as mean number of groups per hour. Species not encountered during transect walks, but that appeared in the complete species list for all sites, were included as zero counts. We estimated bird species richness at each site using the Chao 1 estimator over species presence on transects (Chiu et al., Reference Chiu, Wang, Walther and Chao2014). We tested for differences in encounter rate between sites, and between threatened and non-threatened species, using Kruskal–Wallis and Wilcoxon two sample tests, respectively, in R 4.0.0 (R Core Team, 2020). To identify mammals from the camera-trap captures, several authors provided preliminary identifications that were then compared. In cases of discrepancy we consulted experts, and we omitted the records if uncertainty remained about species identification (6% of images of mammals). We estimated mammal encounter rates from camera-trap data by calculating the proportion of cameras with records of each species of the total number of cameras that had recorded mammal species, thus avoiding some of the issues associated with identifying independent observation events (Wearn & Glover-Kapfer, Reference Wearn and Glover-Kapfer2017).
Identifying priority areas for conservation
To prioritize areas of particular significance for the protection of biodiversity, we built species distribution models for species of conservation concern, using an approach suitable for small numbers of occurrence points (Breiner et al., Reference Breiner, Guisan, Bergamini and Nobis2015). For birds categorized as threatened or near threatened on the IUCN Red List (i.e. categories Critically Endangered, Endangered, Vulnerable, Near Threatened), individual models were built for species with > 10 records and then summed to produce a generalized habitat suitability map. For mammals, all species of conservation concern (globally threatened, on Indonesia's national protection list, or endemic; Ministry of Environment and Forestry, 2018b) were modelled together as a single group. All species have different ecological niches, but as the objective was to prioritize areas for conservation more broadly, we interpreted the model results to show those areas of habitat most critical for mammals of conservation concern (there is evidence to suggest threatened species can be effective surrogates for establishing conservation priorities; Tognelli, Reference Tognelli2005; Drummond et al., Reference Drummond, Wilson, Meijaard, Watts, Dennis, Christy and Possingham2010). Predictors, at 30 m resolution, were topographic (elevation, slope, roughness and topographic wetness index) and habitat-related (percentage of tree cover in 2018, and normalized difference vegetation index, NDVI). Elevation and derived topographic metrics were calculated from the ASTER digital elevation model (Aster DEM, 2001). The topographic wetness index provides an indication of upstream flow accumulation and is correlated with soil characteristics (Quinn et al., Reference Quinn, Beven and Lamb1995; Raduła et al., Reference Raduła, Szymura and Szymura2018); the index was created with the dynatopmodel package in R (Metcalfe et al., Reference Metcalfe, Beven and Freer2018). We extracted tree cover from Hansen et al. (Reference Hansen, Potapov, Moore, Hancher, Turubanova and Tyukavina2013), and NDVI from two merged Landsat 8 images (taken on 22 October and 7 November 2019) with almost no cloud cover (USGS, 2019).
We used species occurrence points from transect surveys, occasional sightings and camera traps. Additional points were taken from a survey by SvB on west Gunung Slamet in June 2018 as part of an environmental impact assessment in the area of a proposed, but presently abandoned, geothermal plant. To avoid sampling bias, we restricted random background points to within a 2.5 km radius of presence points (Kramer-Schadt et al., Reference Kramer-Schadt, Niedballa, Pilgrim, Schröder, Lindenborn and Reinfelder2013; Merow et al., Reference Merow, Smith and Silander2013), given a lack of other biological records of similar taxa that would have allowed us to implement more complex methods (Phillips et al., Reference Phillips, Dudík, Elith, Graham, Lehmann, Leathwick and Ferrier2009). We followed modelling methods described by Breiner et al. (Reference Breiner, Guisan, Bergamini and Nobis2015). For each species or group, we made ensembles of all combinations of bivariate models using Maxent 3.4.1 (Phillips & Dudík, Reference Phillips and Dudík2008), with each individual model weighted by Somer's D, a rescaled area under the receiver operating characteristic curve (AUC) that gives prominence to higher values. To validate the models, we implemented 10 runs of 2-fold cross validation, obtaining an overall mean AUC for each species. The mean AUC of the 10 runs of each bivariate model was then used to weight a final full ensemble model of all bivariate combinations using all the occurrence points. We created models using the dismo, ecospat and raster packages in R (Hijmans et al., Reference Hijmans, Phillips, Leathwick and Elith2016; Di Cola et al., Reference Di Cola, Broennimann, Petitpierre, Breiner, D'Amen and Randin2017; Hijmans, Reference Hijmans2019).
Forest loss, threats and conservation status
We calculated forest loss during 2001–2018 using a global forest cover dataset based on 30 m resolution Landsat images (Hansen et al., Reference Hansen, Potapov, Moore, Hancher, Turubanova and Tyukavina2013). For the Tropical domain, this classification obtained an overall accuracy of 99.5% and a precision of 87.0% for forest loss (Hansen et al., Reference Hansen, Potapov, Moore, Hancher, Turubanova and Tyukavina2013). We performed the analysis on the area surrounding Gunung Slamet at altitudes > 500 m, per 500 m altitudinal band, per year.
To examine human use of forest trails, we counted and categorized people recorded by camera traps as either recreational users (e.g. hikers wearing trainers and backpacks) or local people believed to be exploiting forest resources (e.g. wearing flip-flops and synthetic sacking bags). We excluded from these counts members of the research team recorded by cameras. We used a G test (Zar, Reference Zar1999) to evaluate whether certain sites were used more or less frequently by recreational or resource users. We obtained additional data on the use of forest resources and bird trapping from a bird market survey in the town of Purwokerto (20 km from the summit of Gunung Slamet), and through informal conversations with members of the local community, including our guides on the field trips.
We observed 99 bird species on the transect walks, of which 13 are of conservation concern globally (two Critically Endangered, one Endangered, four Vulnerable and six Near Threatened; Supplementary Table 1a), and 21 are included in Indonesia's national protection list (Ministry of Environment and Forestry, 2018b). Twenty-five species were restricted to the Java and Bali Forest Endemic Bird Area (Stattersfield et al., Reference Stattersfield, Crosby, Long and Wege1998). Species richness varied from 45 species at Kaliwadas 2 to 67 at Kaliwadas 1, both on the north side (Table 1). Estimated species richness was highest at Ketenger 2 with 79 ± SE 9 species (Table 1). In terms of species composition, sites on the same flanks (north or south) and geographically closest were most similar. However, when we used encounter rates to gauge similarity between sites, the highest site on the south side, Ketenger 3, was more similar to two other sites on the north side than to its nearest site, Ketenger 2 (Fig. 2). Encounter rates per site were low for species at the upper end of their altitudinal distribution (e.g. the blue whistling thrush Myophonus caeruleus, 0.006 groups/h), and highest for the Javan tesia Tesia superciliaris (3.75, 4.58 and 5.66 groups/h at Kaliwadas 2, Kaliwadas 1 and Guci, respectively). Median encounter rates were significantly lower at Ketenger 2 compared to three other sites: Guci, Kaliwadas 1 and Kaliwadas 2 (Kruskal–Wallis; χ 2 = 15.9, df = 4, P = 0.003), and significantly lower for species of conservation concern than others (Wilcoxon test; W = 2924, P = 0.010). Maximum encounter rates for threatened or near threatened species did not coincide at the same sites, with all sites having a maximum value for at least one species. Threatened species encountered on both sides of the mountain were the endemic subspecies of the Javan laughingthrush G. rufifrons slamatensis, with a maximum encounter rate of 0.43 ± SE 0.08 groups/h, Javan cochoa Cochoa azurea (0.48 ± SE 0.10 groups/h), Javan trogon Apalharpactes reinwardtii (0.20 ± SE 0.08) and Javan hawk-eagle Nisaetus bartelsi (0.15 ± SE 0.05 groups/h). For security reasons we are not reporting the exact locations where we encountered threatened species (Collar et al., Reference Collar, Eaton and Sykes2017).
Thirteen mammal species were captured by 51 of 112 camera traps (total effort = 10,358 h; Table 1, Supplementary Table 1b). A further two species, the Vulnerable Javan lutung Trachypithecus auratus and Endangered Javan gibbon Hylobates moloch, were only recorded on transect walks. Of the 15 mammal species recorded, five are globally threatened (one Critically Endangered, two Endangered, two Vulnerable) and nine are on the Indonesian protection list (Ministry of Environment and Forestry, 2018b; Supplementary Table 1b). Of the mammals known to occur on Gunung Slamet, we failed to record the Javan mongoose Herpestes javanica, wild boar Sus scrofa, dhole Cuon alpinus (Maryanto et al., Reference Maryanto, Noerdjito and Partomihardjo2012) and mountain weasel Mustela lutreolina (Meiri et al., Reference Meiri, Duckworth and Meijaard2007). However, we recorded three species not previously known to occur there: the Sunda porcupine Hystrix javanica, Sunda pangolin Manis javanica and banded linsang Prionodon linsang. The maximum number of species recorded by a single camera trap was four, and > 70% of camera traps with mammal sightings recorded just one species. Mammal species richness per site was 5–9, with the maximum at Kaliwadas 2 on the north side (Table 1). Threatened species caught on camera traps were the Javan leopard Panthera pardus melas, Sunda pangolin and Javan surili. The Javan stink badger Mydaus javanensis, Javan treeshrew Tupaia javanica and three-striped ground squirrel Lariscus insignis were recorded at four sites each, with the Javan stink badger the most frequently recorded species (11 cameras), followed by the Javan leopard (10) and Asian palm civet Paradoxurus hermaphroditus (10). Encounter rates (proportion of cameras that recorded a particular species) for mammals ranged from 0.04 for the Sunda porcupine to 0.22 for the Javan stink badger (Supplementary Table 1b).
We found 17 species of reptiles and amphibians, all of which are categorized as Least Concern (IUCN, 2019), including two endemic species, the pearly tree frog Nyctixalus margaritifer and Java flying frog Rhacophorus margaritifer (Supplementary Table 1c).
Priority zones for birds and mammals on Gunung Slamet
We included a total of 154 records of five bird species, and 74 records of 10 mammal species in the species distribution models (Supplementary Table 2). Evaluation results for the ensembles of small models were generally good, with the overall mean AUC of all species 0.93 ± SE 0.005. Individual mean AUC values of the 10 runs ranged from 0.88 ± SE 0.012 to 0.95 ± SE 0.004. The summed species distribution model for birds showed broad congruence with the model for mammals, with most suitable habitat consisting of a broad band of forest around the whole mountain at an altitude > 1,000 m, but covering a larger area on the western slopes. High values of habitat suitability for birds extended to slightly lower altitudes than for mammals, especially on the southern slopes at 1,000–2,000 m (Fig. 3).
Forest loss and other threats
In total, 13.7 km2 of forest (1.7% of total) were lost from Gunung Slamet during 2001–2018, with most loss occurring below 2,000 m (Figs 1 & 4). A large, continuous block of forest covers Gunung Slamet at higher altitudes and reaches lower elevations on the south and south-east slopes (Fig. 1). At lower altitudes, the forest is severely fragmented. A core area, over c. 800 m in elevation, contains 288 km2 of continuous forest. The presence of large trees with girths > 4 m indicated high quality of forest habitat, but at all but one site we observed cut stumps in > 30% of vegetation plots (Table 1).
People (excluding the research team) were recorded by 63% of cameras. Potential resource users were recorded by > 50% of cameras, and recreational users by 22%. Recreational and resource users followed similar trails, but were recorded less at the Guci site. However, there was no significant association between site and user type (G test; G = 9.11, df = 5, P = 0.105). An additional trail on the eastern side of Gunung Slamet, not surveyed here, provides access for large numbers of visitors (Mittermeier et al., Reference Mittermeier, Oliveros, Haryoko, Irham and Moyle2014).
Several local people, including our guides and porters, admitted to trapping birds on Gunung Slamet, although all did so opportunistically rather than as their main source of income. The main trapping techniques were snares, stick traps and mist-nets, sometimes with tape lures or by stringing a snake from a tree so that its movement attracts birds. Guides informed us that trappers could use the extensive network of trails to access most, if not all, of the mountain's forest areas, except those used by large numbers of tourists. At 40% of vegetation plots across all sites, the nearest main trail had a secondary trail leading from it (Table 1). According to local people, access to the restricted area around the geothermal site was more difficult, but still possible for bird trappers. On 6 September 2018 at the live bird market in Purworkerto, we observed the indigo flycatcher Eumyias indigo, mountain warbler Phylloscopus trivirgatus, hill blue-flycatcher Cyornis banyumas and several other species that reportedly had been caught locally on the mountain.
The biological importance of Gunung Slamet
Gunung Slamet is a stronghold of Java's unique montane fauna. During 6 weeks of fieldwork, we recorded 13 bird species of global conservation concern, 28 birds with restricted ranges (within the Java and Bali Forest Endemic Bird Area), five threatened mammal species, six endemic mammals and two endemic frogs. Thirty of these species are included in the Indonesian government protection list (Ministry of Environment and Forestry, 2018b) and four species legislated as priorities for conservation strategies: the Javan hawk-eagle (very high priority) and Javan leopard, Javan gibbon and Javan surili (high priority; Ministry of Forestry, 2008).
The presence, and in some cases abundance, of certain species is noteworthy. We observed G. rufifrons slamatensis, the subspecies of the Javan laughingthrush endemic to Gunung Slamet, at several sites on the mountain. It had not been recorded in the wild since 1925 (Collar & van Balen, Reference Collar and van Balen2013; Mittermeier et al., Reference Mittermeier, Oliveros, Haryoko, Irham and Moyle2014), but was present in bird markets (Nijman et al., Reference Nijman, Ardiansyah, Hendrik, Imron and Nekaris2020; Species 360 ZIMS, 2020). The Javan hawk-eagle was widespread despite increasing pressure from trappers (Eaton et al., Reference Eaton, Shepherd, Rheindt, Harris, van Balen, Wilcove and Collar2015), as were the Javan cochoa and Javan trogon. The latter, previously known only from West Java (Collar & van Balen, Reference Collar and van Balen2002) has extended its range considerably, with Gunung Slamet as its new easternmost limit. The Javan flameback Chrysocolaptes strictus has also rarely been recorded in Central Java (Mees, Reference Mees1996), and ours appears to be the first confirmed record there of the Javan scops-owl Otus angelinae (König & Weick, Reference König and Weick2008). The presence of the Sunda grasshopper warbler Locustella montis on Gunung Slamet suggests another range expansion, c. 100 km to the west of the species’ previously known range (Rozendaal, Reference Rozendaal1989). These records highlight the key position of Gunung Slamet as a meeting point of the eastern and western Javan avifaunas. Notable absences from our surveys, and also from a 2013 survey (Mittermeier et al., Reference Mittermeier, Oliveros, Haryoko, Irham and Moyle2014), were the Javan green magpie Cissa thalassina, known from a single record on Gunung Slamet from 1917 (van Balen et al., Reference van Balen, Eaton and Rheindt2013), and Javan oriole Oriolus cruentus. The Javan green magpie was probably not recorded as a result of considerable trapping pressure on this species (van Balen et al., Reference van Balen, Eaton and Rheindt2013; Nijman et al., Reference Nijman, Sari, Siriwat, Sigaud and Nekaris2017). The oriole, however, although generally inconspicuous and easily overlooked, was recorded by our survey team at four sites on two other mountains in West Java during the same field season (A.R. Junaid & G.C. Aprianto, unpubl. data). Other species that were unexpectedly rare included the white-bellied fantail Rhipidura euryura and chestnut-backed scimitar-babbler Pomatorhinus montanus, possibly also indicative of heavy trapping pressure.
Amongst the threatened mammals, the Javan gibbon, Javan surili and Javan leopard were relatively common. We expected the mountain to harbour significant populations of these species, given that > 280 km2 of forest remain, with likely population densities in West and Central Java of 1.70 and 5.96 and individuals/km2 for the Javan gibbon (Setiawan et al., Reference Setiawan, Nugroho, Wibisono, Ikawati and Sugardjito2012) and Javan surili (Setiawan et al., Reference Setiawan, Wibisono, Nugroho, Agustin, Imron, Pudyatmoko and Djuwantoko2010), respectively, and 10–12 individuals/100 km2 for the Javan leopard (Rahman et al., Reference Rahman, Rianti, Muhiban, Muhtarom, Rahmat, Santosa and Aulagnier2018). By contrast, the heavily trapped Sunda pangolin was recorded only once, although pangolins require higher camera-trap effort or specific targeting to achieve higher detections (Khwaja et al., Reference Khwaja, Buchan, Wearn, Bahaa-el-din, Bantlin and Bernard2019).
Threats and opportunities on Gunung Slamet
Forests on the lower slopes of Gunung Slamet are fragmented, with agriculture encroaching up to higher altitudes on the northern side. Although forest loss is less at higher elevations, > 160 km2 of montane forest are estimated to have been lost during 2000–2012 across Java (Margono et al., Reference Margono, Potapov, Turubanova, Stolle and Hansen2014). Nevertheless, the forest on Gunung Slamet's upper slopes is still relatively intact, with large trees persisting at the study sites despite evidence of small-scale logging. This forest forms a 280 km2 continuous block above c. 800 m altitude, and our species distribution models indicate this offers the most suitable habitat for threatened species.
Bird trappers operate via an extensive network of trails across the mountain. There are no quotas for the capture of wild birds in Central Java, and commercial harvest is illegal (Ministry of Environment and Forestry, 2018a, 2020). Camera-trap footage revealed significant traffic of resource users in the forest. Given the lack of formal protection and the access provided by the trail network, trapping pressure is probably high and persistent (Marshall et al., Reference Marshall, Collar, Lees, Moss, Yuda and Marsden2020). Eleven specimens of G. rufifrons slamatensis were recently acquired by breeding centres in Indonesia from markets, five in 2017, and six in 2018 (Species 360 ZIMS, 2020). Hikers also use the 3–4 main trails to the mountain's summit, and the recreational park and waterfall at Ketenger on the lower slopes. Their presence may benefit the mountain's wildlife (Hakim et al., Reference Hakim, Rahardi and Rachmansyah2018), deterring trappers and providing income to local guides, many of whom nevertheless trap birds to supplement their income. However, Gunung Slamet is one of the main destinations for mountain hiking in Central Java, with thousands of people ascending the summit each year, and further regulation beyond ticketing may be required (Republika, 2010).
Although we believe our findings are comprehensive considering the methods used, more species would probably be recorded with longer surveys across other seasons. We conducted our surveys during the dry season, for logistical reasons, and species detectability may vary between seasons. We did not cover the westernmost slopes of Gunung Slamet, for logistical reasons and because of access restrictions around the geothermal site. These restrictions may inhibit trappers, but if construction of the plant is abandoned, which appears likely, the new road to the site will facilitate access for resource users.
Improving the protection status of Gunung Slamet
The current status of Gunung Slamet is a so-called protection forest (hutan lindung), managed by Perum Perhutani, the state forest enterprise (Supplementary Table 3). Protection forests are managed to regulate water systems, prevent flooding, control erosion, and maintain soil fertility, but are not classified as protected areas under Indonesian government regulations (CIFOR, 2003). Two small protected areas are present at Guci (2 ha) and Telogo Dringo (48.5 ha), on the north and west flanks, respectively, but these are far from the main blocks of montane forest (UNEP–WCMC & IUCN, 2019; Fig. 1, Supplementary Table 3). However, our findings and previous studies highlight the significance of the mountain for biodiversity (Setiawan et al., Reference Setiawan, Nugroho, Wibisono, Ikawati and Sugardjito2012), including as a key site for the Javan hawk-eagle (Sözer et al., Reference Sözer, Nijman, van Balen, Setiawan, Prawiradilaga and Subijanto1997). We therefore recommend the designation of a higher protection status for Gunung Slamet, in line with regional priorities for forest conservation (Sodhi et al., Reference Sodhi, Koh, Clements, Wanger, Hill and Hamer2010).
One option is to designate Gunung Slamet as a protected area within IUCN categories I‒IV, under national management. A Nature Reserve under Indonesian legislation would confer strict protection (Basjarudin, Reference Basjarudin1971), and our findings suggest that the mountain's levels of biological diversity and endemism meet the criteria for this designation (Indonesian Government, 2011). However, both resource and recreational use could be in conflict with this designation. Several villages in the districts of Tegal, Pemalang, Pubalingga, Banjaran and Banyumas source water from Gunung Slamet (Gunawan, Reference Gunawan, Maryanto, Noerdjito and Partomihardjo2012). A more accommodating protected area category is National Park (Indonesian Government, 2011), which has the objective of protecting life-support systems and species diversity while also allowing the sustainable use of natural resources. A third option is the designation of Essential Ecosystem within High Conservation Areas, managed by Perum Perhutani locally, including the identification of priority areas for birds and mammals, such as those areas highlighted by the species distribution models (Ditjen KSDAE, 2018). Essential Ecosystems aim to protect biodiversity and maintain ecosystem services outside conservation areas (Direktorat Bina Pengelolaan Ekosistem Esensia, 2019), but have yet to be fully regulated by the Ministry of Environment and Forestry (Supplementary Table 3). Gunung Slamet is a vital refuge for montane biodiversity in Central Java and therefore should receive greater protection in the form of one of these three designations.
We thank the student volunteers, Arya and Hafiz, and our local guides, for their help with fieldwork; Faris Muladi and Hannah Khwaja for help with identification of mammal images; the Ministry of Environment and Forestry and Perum Perhutani for facilitating research and fieldwork; Rainforest Trust, Chester Zoo and the EAZA Silent Forest Campaign for funding; and Vincent Nijman and an anonymous reviewer for their constructive comments. We dedicate this paper to the memory of Tony Whitten.
Field surveys: ARJ, assisted by GCA, RS, FK, CD, SJM; support with bird identification: BvB; data analysis: CD; writing: CD, NJC, SJM; editing: all authors.
Conflicts of interest
This research complied with the Oryx guidelines on ethical standards. Fieldwork did not involve collection of any specimens. No ethical standards for camera trapping currently exist in Indonesia, however, all images have been stored securely and are treated as strictly confidential. The project complies with ethical standards of Manchester Metropolitan University and was carried out with permissions obtained from relevant authorities in Indonesia (Memorandum of Understanding with Kementerian Lingkungan Hidup Dan Kehutanan: S. 37/KSDAE/PIKA/ILSA.0/1/2018).