Many tropical biosphere reserves face the problem of overexploiting biological resources, which is a major cause of biodiversity loss. We used an integrative approach, combining camera trapping and bat trapping in diverse habitats over multiple surveys in Yangambi Biosphere Reserve, Democratic Republic of the Congo. We investigated whether species-level traits of volant and nonvolant mammals, as well as key features of their habitats, can predict effects on species detectability and occupancy, and thus on community-level species richness. We applied single-season occupancy models using multi-species data for camera trap and bat trap datasets. We found a positive effect of denser woodlands on occupancy estimates over 6480 camera-days. Average adult body weight and whether species were group-living negatively influenced species detection estimates. Species richness and body size assemblages approximated marginal rainforest diversity. This included small-bodied, generally high-hunting-resilient arboreal and terrestrial species. During 324 trap-nights, site location was the main determinant of habitat occupancy at the community level. Occupancy estimates were not, however, directly linked to variables measured in different stands. This may be a consequence of habitat patterns at larger spatial scales, which require additional data to clarify. Species-level characteristics that might influence detection estimates were not, however, included in the parsimonious output of the model used. Our findings indicate specific mechanisms triggering mammalian impoverishment above a critical size that may drive ecological changes. Thanks to certain life-history traits, bats can compensate for some of the lost services previously provided by large mammals, thus contributing to the maintenance of biodiversity dynamics.
Research Article
Mammal depletion processes in a UNESCO biosphere reserve: an integrative analysis of population viability
https://doi.org/10.21203/rs.3.rs-2746604/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Aerodynamics
Conservation practice
Democratic Republic of the Congo
Mobile link organisms
Occupancy patterns
Biosphere reserves (BRs) have been designated by the United Nations Educational, Scientific and Cultural Organization (UNESCO)’s Man and the Biosphere Programme (MAB) since the 1970s as exemplary social-ecological systems for improving human-environment relationships in conservation management (Kratzer and Ammering 2019). Despite the global networking of BRs, insufficient funding for long-term biomonitoring, particularly in low-income countries, has hampered the achievement of the designated goals. Consequently, various mechanisms to support sustainable management, including the MAB program (Van Cuong et al. 2017) and the Lima Action Plan (Reed 2016; Kratzer and Ammering 2019), are failing to solve the biodiversity crisis. In an attempt to overcome this shortcoming, the Aichi Mechanism of the Convention on Biological Diversity (CBD) issued a set of indicators and policies to measure and ultimately prevent the extinction of plant and animal populations (Rovero and Ahumada 2017). These metrics do not, however, adequately capture the practice and scale of hunting, with no reference to wildlife hunted but not trafficked (Benítez-López et al. 2017; Driscoll et al. 2018). Consequently, there is no scientific consensus on operational indicators of biodiversity status (Scholes and Biggs 2005). The Global Biodiversity Outlook (GBO-5), released in September 2020, highlights the collective global failure to meet the CBD targets (Kumar 2020). The lack of baseline data prevents the spatial and temporal analyses and trend predictions required to guide conservation actions. This is an obvious constraint in addressing conservation challenges, such as how intrinsic species traits, richness, composition, mobility, and habitat use patterns may affect ecosystem dynamics.
As BRs are not members of the International Union for Conservation of Nature (IUCN), regulations governing their resource use are ambiguous. Increasing anthropogenic pressures are affecting BRs in unanticipated ways, causing the local decline of many large vertebrate species (Davidson et al. 2009). For instance, larger, social, diurnal species are probably more vulnerable to decline than smaller species because they are considered more important by hunters (Laméris et al. 2020). There is a growing consensus that intrinsic factors are only predictive of increased risk of extinction for species weighing more than 3 kg (Cardillo et al. 2005). Below this critical weight, extrinsic, non-genetically heritable factors, such as the size of the geographic range as well as the density of nearby human populations may be the determining factor for the risk of becoming extinct for a given species (Matthews et al. 2011). Forests neighbouring human settlements with an important population may host edge (marginal rainforest) diversity within the rainforests, including primarily small, rapidly reproducing mammal species (i.e., species with high resilience to hunting). In contrast, forests situated more distantly from human settlements may harbour more important species (deep rainforest) diversity, including large, more slowly reproducing species (i.e., species with low hunting resilience). This pattern in the structure of the standing biomass of mammals along a human density gradient was recently described in Central Africa (Fa et al. 2015).
While some declining species may be replaced by species with similar ecological functions, this is much less the case for a significant number of species or functional groups (Wilkie et al. 2011). The loss of keystone species (ecosystem engineers) can disrupt functional relationship and bring about major ecological shifts (Mills et al. 1993; Wright and Jones 2006). In this context, so-called mobile link organisms, such as volant mammals (i.e., bats), are able to restore some of the functions previously performed by large mammals (Lundberg and Moberg 2003; Couvreur et al. 2004; Kremen et al. 2007). Indeed, there are two suborders of bats, namely the Yinpterochiroptera, which includes the phytophagous bats and their close relative, the insectivorous echolocating bats (i.e., the Rhinolophoide superfamily), and the Yangochiroptera, which includes other echolocating bats (Teeling et al. 2005; Almeida et al. 2021). As these two suborders exhibit a wide range of life-history traits and an evolutionarily diverse species status. Through the combining of their foraging and commuting activities, bats are known to be key in enhancing forest resilience and restoration by connecting habitats and ecosystems as seed dispersers, pollinators, and pest controllers that would otherwise remain disconnected (Sritongchuay et al. 2019). Some morphological parameters of bats (e.g., wing and body size) are important factors to infer their flight performance and to estimate their ability to circulate in different environments. Bat species characterised by a low wing aspect ratio tend to fly slower, but with a superior maneuverability, which allows them to travel in a cluttered environment (Norberg and Rayner 1987). On the other hand, bat species with greater wing loading reduces flight maneuverability. Consequently, such species will be capable of flying faster, higher and longer in open areas (Elangovan et al. 2004; Schunk et al. 2017). Thus, the interspecific variation in the morphological wing characteristics in bats impacts the scope of suitable habitats. Hence, these differences allow different bat species to cope with different local ecological constraints and interspecific competition resulting from resource use.
Coexistence requires that ecologically similar species differentiate their niches and partition resources, such as food, habitat and/or activity time. Although these mechanisms are difficult to assess under natural conditions, food allocation is the most important dimension in many mammal communities and is linked to morphological differences, such as body size (Rosenzweig 1966; Davies et al. 2007). This is especially apparent in sympatric species, where the ratio of largest to smallest body size varies by an average of 1.3 (Emmons 1980). In theory, these predictions indicate the level of difference necessary for two species to co-occupy different niches but at the same level of a food web (Hutchinson 1959). Because niche partitioning is based on food size, spatial and temporal resource partitioning may also be important between species of similar foraging behaviour have different abilities to compete (Di Bitetti et al. 2010; Davis et al. 2018). For example, squirrels in Central African rainforests show differences in forest use, with vertical forest strata in the same area dividing certain species into tree-dwelling (mainly frugivores) and ground-dwelling for the majority of insectivorous species (Emmons 1980). It is clear that the organisation of biotic elements in African rainforests is highly specialised, with forest composition, plant species diversity and habitat structure changes at smaller spatial scales (Adamescu et al. 2018). Therefore, it is essential to understand what factors drive spatial patterns according to which wildlife occur, to preserve and manage species in a spatial context.
Here, we combined a suite of species-level traits of volant and nonvolant mammals with key rainforest attributes likely to covary with spatial patterns of habitat use. We focused on species-specific life-history traits of nonvolant mammals (activity patterns, habitat use strategies, standing body mass, and social behaviour), and flight ability index (aspect ratio, maximum range speed, and wing loading) of bats. We aimed to investigate whether the effect of species-level traits of volant and nonvolant mammals, and forest stand-level characteristics, can predict detectability and habitat occupancy, and thus community-level species richness at a finer spatial scale. We hypothesised that in mammal-poor rainforests, body size is a major predictor influencing community-level detection estimates of nonvolant mammals. The key factor influencing volant mammal detection estimates may not be clearly expressed at the multi-species level, combining phytophagous and insectivorous bat ensembles, while habitat structure only affects nonvolant mammal occupancy at finer spatial scales.
Study area and protocol design
The Democratic Republic of the Congo (DRC) hosts three BRs (Lufira, Luki and Yangambi). These were established without adaptive partnerships between local communities and management bodies, making their management problematic (Sikuzani et al. 2020). The concept of a biosphere reserve or Man and Biosphere Reserve has not yet translated into legal texts, resulting in an unclear legal status of BRs as protected areas (Koy et al. 2019). This study focuses on the Yangambi Biosphere Reserve (YAB), an area in what had been the Yangambi Floristic Reserve since 1939 and was later declared a MAB reserve in 1977. There is no official management plan and the boundaries of the YAB remain controversial, with land use remains a challenging issue (Kipute et al. 2021). The YAB covers an area of about 2250 km² centered at 0.771°N, 24.527°E, including a proposed core area of about 375 km² (Fig. 1). The proposed zoning plan – as a spatial framework of three interrelated zones (one or more core areas, a buffer zone surrounding the core areas, and a transition zone) for achieving the three interconnected functions of a BR (Kratzer and Ammering 2019) – based solely on a floristic inventory was never clearly implemented. A continuum of closed evergreen forest dominates Yangambi region on the right bank of the Congo River, with scattered human-dominated areas (Van de Perre et al. 2018). The climate is characterised by two dry seasons, December-March and June-July, alternating with two wet seasons, April-May and August-November (van Vliet et al. 2019). Annual rainfall ranges from 1400 to 2000 mm, and the average temperature is around 26°C.
The sampling protocol is based on detection and non-detection data and is applicable to multiple taxa and contexts. The study design emphasised the status of bats as mobile link organisms that can facilitate within-community variation in species-specific traits such as mobility, home range, and habitat requirements (Kremen et al. 2007; Devarajan et al. 2020). The sampling design based on the assumption that the communities to be sampled are distributed over a spatial scale. Using different sampling methods in the same area illustrates the proper consideration of at finer spatial scales. We used camera traps for terrestrial and arboreal mammals and trapping techniques for volant mammals to account for the different spatial use patterns of forest mammals. To better study the forests adjacent to the YAB’s core area, the buffer zone was monitored longer than the transition zone. The transition zone was only surveyed on the ground to update a list of possible species. We monitored bats in both the buffer and transition zones with equal survey effort, which overlapped in time and space with the camera surveys. We used species-level attributes that could have a direct impact on detection estimates, e.g., nonvolant species-specific life-history traits (activity patterns, habitat use strategies, standing body mass, and social behaviour) and flight ability index (aspect ratio, Maximum range speed, wing loading) of volant mammals. Morphological trait estimators of uncaptured bats (i.e., species previously recorded in the same area) were compiled from the literature (Norberg and Rayner 1987; Patterson and Webala 2012). We described the forest characteristics where the traps were set. Building on published work (van Vliet et al. 2019), we outlined the number of hunters actively involved in the bushmeat market chain to support a broader interpretation of current trends in the standing body mass of mammal in the YAB.
Sampling And Covariates
We developed a comprehensive sampling design that may be representative for volant and nonvolant, arboreal and ground-dwelling mammals, including replication of sites. The bat surveys were based on approaches developed for Neotropical bat surveys (Farneda et al. 2015; Rocha et al. 2016; Ferreira et al. 2017), while camera design follows approaches developed in Rwanda (Moore et al. 2020). To capture local variation, the forest description, volant and nonvolant mammal surveys were conducted using a design based on a total of 36 plots (20 × 5 m), with 18 plots in the buffer zone and 18 in the transition zone. At each site, we recorded forest clutteredness and potential foraging/visitation area indicators. We recorded stem density, including lianas with diameter at breast height (dbh) ≥ 5 cm, and basal area of woody stems with dbh ≥ 10 cm (i.e., the sum of the cross-sectional area at 130 cm, or above the buttress, of all live trees). We determined the straight-line distance between each plot and the nearest waterways by combining public domain digital map data (RGC 2010) with our own field waypoints using QGIS 3.6.3 (QGIS Development Team 2019).
Each site (i.e., forest stand) was monitored by camera for six consecutive months at ground level, and for a further six consecutive months in the arboreal substrate at the same location (December 2019 to December 2020), concurrently with the bat surveys. One camera site corresponds to one plot. The ground cameras were placed at various heights ranging from 0.6 to 1 m (mean = 0.72 m ± 0.13 SD). Height variations were due to either the slope of the terrain or the need to record smaller to larger species dwelling on/near the ground. Arboreal cameras were placed below the canopy between 11 and 19 m (mean = 14.5 m ± 3.0 SD), with height variation depending on climbability. One camera was deployed per site, totalling 18 cameras for each substrate over six months, with monthly checks to ensure cameras had sufficient battery life and memory card space. Spatial schemes of camera distribution were designed to meet occupancy independence assumptions and exceed estimated home ranges of most small-bodied species, such as arboreal rodents (Kingdon 2015). The camera models (Bushnell Trophy Cam Essential HD and Victure HC200) were randomly selected for each site, both ground and arboreal, and were set to record a 32-second video each time they were triggered, with no delay between videos. We used expert knowledge to annotate camera recordings and then identified them using reliable sources (Kingdon et al. 2013a, b; Kingdon and Hoffmann 2013). The survey effort for each camera was set to 30 continuous recording days, i.e., one sampling occasion, henceforth event. A total of 6480 camera-days were carried out, i.e., 18 cameras × 30 days × (6 ground events + 6 arboreal events). A camera-day is defined as one camera used for one day.The bat sampling schedule covered the late rainy season and the early dry season. Each site was surveyed to ensure that sampling was continuous and not interrupted by heavy rains. We simultaneously used one set of six traps per site, namely four nylon mist nets (Ecotone, Poland) of three different mesh sizes (16 × 16 mm; 19 × 19 mm; 30 × 30 mm), shelves (5; 5; 4), dimensions (9 × 3 m; 9 × 3 m; 9 × 3.2 m), and two identical custom-made harp traps (Prince of Songkla University, Thailand) modified from (Francis 1989), iron and aluminum poles, four-bank 150 × 150 cm capture area, total weight 9 kg. The traps were randomly selected at each sampling point and placed near waterways and flyways. The height of the mist nets depended on available flyways and varied from 1 m to 7 m above the ground. To reduce the effects of time of night, observer, and trap shyness (Hoyle et al. 2001; MacKenzie and Royle 2005), we used a rotating survey scheme between sites. To fully standardise trapping effort, surveys were repeated at the same site seven days apart during different time periods. Because we were interested in studying bat flight activity, we used temporal events, as part of the night replication. We fixed the sampling time for the fly-out period so that the sampling duration (a 4-hour period corresponds to one event) was the same for the three replicate surveys, namely early evening (6:00pm-10:00pm), middle of the night (10:00pm-2:00am), and early morning (2:00am-6:00am) in Central Africa Time (UTC + 2:00). During each event, traps were checked at least 30 minutes apart.The total bat survey effort was 324 trap-nights (i.e., 6 traps × 3 consecutive nights × 3 replicates × 6 sites), with a trap-night defined as the simultaneous opening of a set of six traps for 4 hours after sunset. Captured bats were individually placed in cloth bags (30 × 40 cm) before measurements of taxonomically relevant external characters. We measured ear, tail (if possible) and tibia lengths with a caliper (± 0.01 mm), forearm and wingspan lengths with a ruler (± 0.1 mm), and body mass using spring balances (30g/0.2g, 50g/0.5g, 100g/1g, 300g/2g Pesola, Ecotone, Poland). We qualitatively differentiated juveniles from adults by transilluminating the wing of an individual with a headlamp and visualizing the cartilaginous zone of the long phalanges (Kunz and Anthony 1982). Individuals with a gap at the fourth metacarpal-phalangeal joint were defined as juveniles. Scrotal position was an additional indicator to distinguish adult males from juveniles. Reproductive status of adult females was categorised as pregnant (based on gentle palpation of the abdomen), lactating (enlarged nipples without surrounding fur and milk secretion when gently squeezed), or no reproductive signs (Lučan et al. 2014). Within 2 hours, captured bats were released unharmed at the capture point after being photographed for possible comparison with illustrated keys. We used reliable sources for bat identification in the DRC (Patterson and Webala 2012; Cakenberghe et al. 2017; Monadjem et al. 2020).
Index Estimates
Bat capture data were pooled for each bat trap site. Forest characteristics were averaged at the site scale into a single dataset. We calculated bat wing aspect ratio (AR), wing loading (WL) and predicted maximum range speed (Vmr) as potential species covariates according to the following allometric equations (Norberg and Rayner 1987; Rhodes 2002; Elangovan et al. 2004): AR = B2/S with B in metres and S in square metres; WL = M × g/S with M in kg, gravitational acceleration (g) equals 9.81 metres per square second, and WL is expressed in Newtons per square metre (N.m−²); Vmr = 8.71 × M0.423 × B− 0.498 × S− 0.144, with Vmr in metres per second (m/s).
The camera data recorded in the transition zone during 540 camera-days (18 cameras × 30 days) were not included in the analyses. Instead, they were used to update the list of potentially available species in the area. We combined ground and arboreal camera data from buffer zone and forest characteristics for each site into a single dataset. At the site level, we analysed the camera and bat detection datasets separately. We evaluated the distribution of camera success using quarterly pooled data (i.e., 18 cameras × 90 days), while keeping the bat trap data per the event design. We evaluated event-specific detection rates using the trap success index: (detections/trap-nights) × 100 (Miño et al. 2007). We replaced trap-night with camera-day for the three-event base to adapt the equation to camera data. We used the Kruskal-Wallis H-test to examine whether the camera trap success over four 3-month periods and the bat trap success over three event fell from the same distribution (Ostertagová et al. 2014). Once significant results were obtained (p < 0.05), Dunn's Kruskal-Wallis multiple comparisons with Holm's p-adjusted method and a correction to control for experimentwise error rates (\(\widehat{\epsilon }\)2) were applied to accurately determine the difference of trap success among events (Gelman and Tuerlinckx 2000).
Occupancy Modelling And Species Richness Estimates
We analysed repeated presence/absence data from a collection of spatial unit site over time event. This model was based on the assumption that non-detection (or non-capture) can be distinguished from absence through repeated sampling, and that using pooled data for all species observed during sampling can improve species-specific occupancy estimates (Zipkin et al. 2010). Indeed, the models built included two sub-models, namely an occupancy sub-model and a detection sub-model to estimate detection probability to account for imperfect species detection and sparse recording (Mackenzie et al. 2002; Royle and Kéry 2007). As this modelling approach requires a fixed list of possible species (S), we generated a species occurrence matrix containing the ‘true’ observation state at each site. We included previous unpublished data in the analyses to control for potential false-negative detections, where species may be present but not recorded (Miller et al. 2011). Each species observed at least once in both terrestrial and arboreal substrates was assigned a detection history of 1, otherwise 0 for unobserved but previously recorded species (as well as in the transition zone, specifically for nonvolant mammals). To control for false-positive errors in the estimates, we accounted for the fact that some nonvolant mammals may not be detectable on either ground or arboreal cameras. Thus, during the ground camera trapping (months one to six), arboreal species had a history of NA detection. The same was true during arboreal camera trapping (months seven to twelve), where ground-dwelling species had a history of NA detection.
We created separate matrices for mammal species characteristics and habitat descriptors, with each row of data relating to a particular site. Only average adult body mass and camera position height were scaled to a fixed range (0 to 1) before running the models to facilitate model convergence. Values were back-transformed into their original units when discussing model predictions. We fitted single-season occupancy models with multi-species data (camera records and bat captures separately) using the RPresence package (version 2.12.27). We included species covariates for detection probability and habitat covariates for occupancy probability (Table 1, Supplementary Information Table S1, and Table 2). We estimated psi or ψj: the occupancy probability at a given station j (i.e., the proportion of species present at each site), and pi: the detection probability for each species i (i.e., the probability of detecting species i given that it is present). We selected the most parsimonious model with a change in the corrected Akaike information criterion score (i.e., ΔAICc < 2) from a set of candidate models with the lowest AICc (Burnham et al. 2011). We checked whether the 95% confidence interval (95%CI) for the regression coefficient (β) included zero to evaluate the influence of covariates on model parameters. Estimates of species richness (SR) values were derived from occupancy probabilities (Mackenzie et al. 2002):
| Model component | Description | |
|---|---|---|
| Covariates of camera trap dataset | ||
| Detection (p) | Activitya,b | Activity patterns, either 1 (diurnal) or 0 (nocturnal) species |
| Groupa,b | Social behaviour, either 1 (group-living), or 0 (solitary or pair-living) | |
| Massa | Average adult body mass (kg) | |
| Strataa,b | Forest strata use patterns, i.e., A (arboreal), G (ground), B (both) | |
| Occupancy (ψ) | BAb | Basal area of trees with dbh ≥ 10 cm measured on a 20 x 5 m plot (m².are − 1) |
| CHb | Camera height, i.e., the height above ground level at which each camera was placed (m) | |
| LRb | Liana cover ratio, i.e., the proportion of lianas (≥ 5 cm) to total non-liana stems (≥ 5 cm) at 20 x 5 m plot | |
| ORb | Forest stand growth ratio, i.e., the proportion of old-growth stems (≥ 30 cm) to undergrowth stems (< 30 cm to ≥ 5 cm) at 20 x 5 m plot | |
| Siteb | Site location, i.e., a 100 m² area where forest characteristics were recorded (i.e., plot) and cameras were placed at ground and tree level. There is a possible co-variation effect with other uncontrolled conditions around the investigated forest stands. | |
| STb | Stem density for all trees and lianas (dbh ≥ 5cm) at each site, scaled to m² level (stems.m− 2). | |
| WDb | Straight-line distance between each site and the nearest water (km) | |
| Covariates of bat trap dataset | ||
| Detection (p) | ARa,b | Wing aspect ratio |
| Vmrb | Maximum range speed of bat species (m/s) | |
| WLa,b | Wing loading (N.m− 2) | |
| Occupancy (ψ) | BAb | Basal area of trees with dbh ≥ 10 cm measured at a set of six plots of 20 x 5 m |
| ORb | Forest stand growth ratio, i.e., the proportion of old-growth stems (≥ 30 cm) to undergrowth at a set of six plots | |
| Siteb | Site location, i.e., a set of six 100 m² points where both forest characteristics were measured and six bat traps were set. There is a possible co-variation effect with other uncontrolled conditions around the investigated forest stands | |
| STb | Stem density for all trees and lianas (dbh ≥ 5cm) at a set of six 100 m² points averaged at 100 m² level (stems.are − 1) | |
| WDb | Average straight-line distance from bat site to the nearest water source (km) | |
| aValue from the literature, e.g., body mass range (Kingdon et al. 2013a, b; Kingdon and Hoffmann 2013) and wing morphology of non-captured species (Norberg and Rayner 1987; Patterson and Webala 2012; Cakenberghe et al. 2017) | ||
| bOur own measurements or observations | ||
| Model component | K | AICc | ∆AICc | wi | βψ(SE) | βp(SE) | |
|---|---|---|---|---|---|---|---|
| (a) | ψ(ST)p(mass + group) | 5 | 1182.544 | 0 | 0.46 | 9.06(3.16) | -39.08(7.53), -0.51(0.26) |
| (b) | ψ(site)p(.) | 3 | 274.743 | 0 | 0.261 | -1.92(0.97) | . |
| ψ(site)p(Vmr) | 4 | 276.131 | 1.388 | 0.131 | -1.90(0.97) | 0.12(0.14) | |
| ψ(site)p(AR) | 4 | 276.275 | 1.532 | 0.122 | -1.91(0.95) | 0.13(0.16) | |
| ψ(site)p(WL) | 4 | 276.670 | 1.927 | 0.1 | -1.91(0.97) | 0.01(0.03) | |
| A dot indicates no variable | |||||||
Where S: total number of possible species, ψij: unconditional occupancy of each station j for each species i. The models were fitted in the R computing environment Rx64 4.0.4 (R Core Team 2021).
Descriptive results
The sampled forest stands were located at varying distances from the main watercourses, averaging 1.009 km (SD = 0.786 km). Mean stem density across all sites was 1214 stems ha− 1 (SD = 162 stems ha− 1, dbh ≥ 5 cm), and mean basal area was 38 m² ha− 1 (SD = 23 m² ha− 1, dbh ≥ 10 cm, Supplementary Information Table S2). We analysed 465 detections of mammal species recorded by ground and arboreal cameras on 6480 camera-days. Excluding detections of bats, dormice, and rats, we counted 19 of at least 45 species potentially available in the area (Table 3a, b, c). The number of available species varied depending on whether they were arboreal (63%), arboreal and terrestrial (45%), and terrestrial only (27%), or small (63%), medium (32%), and large (5%). The ratio of day- to night-active species was 0.27 and 95% of the species were solitary. Quarterly camera-trapping success showed similar trends in detection partitioning between ground and arboreal cameras, H(3) = 7.65, \(\widehat{\epsilon }\)2 = 0.11, CI 95% (0.05, 1), N = 72, p = 0.054 (Supplementary Information Appendix A). The range of standing body sizes of the species commonly recorded at the sites varied from large > 10 kg, e.g., Cephalophus dorsalis, medium about 3–10 kg, e.g., blue duiker, and small < 3 kg, e.g., Galagoides thomasi (Table 3).
| Mammals | Mass | Activity | Group | Detection | IUCN | |
|---|---|---|---|---|---|---|
| (a) Arboreal substrate | ||||||
| Anomalurus beecrofti | Beecroft’s anomalure | 0.66 | 0 | 0 | 5 | LC |
| Anomalurus derbianus | Lord Derby's anomalure | 1.1 | 0 | 0 | 0 | LC |
| Anomalurus pusillus | Lesser anomalure | 0.3 | 0 | 0 | 1 | LC |
| Cercopithecus mitis | Blue monkey | 12 | 1 | 1 | 0 | LC |
| Galagoides thomasi | Thomas’s dwarf galago | 0.1 | 0 | 0 | 54 | LC |
| Gliridaea | Glirids | 0.08 | 0 | 0 | 6 | LC |
| Idiurus zenkerib | Zenker’s pygmy anomalure | 0.02 | 0 | 1 | 1 | LC |
| Lophocebus albigena | Grey-cheeked mangabey | 7 | 1 | 1 | 0 | LC |
| Perodicticus edwardsi | Central potto | 1.6 | 0 | 0 | 1 | LC |
| Piliocolobus langi | Kisangani red colobus | 9 | 1 | 1 | 0 | EN |
| (b) both (arboreal and ground-level) substrates | ||||||
| Allochrocebus lhoesti | L'Hoest's monkey | 8.5 | 1 | 1 | 0 | VU |
| Aonyx congicus | Congo clawless otter | 25 | 1 | 0 | 0 | NT |
| Cercopithecus ascanius | Red-tailed monkey | 4 | 1 | 1 | 0 | LC |
| Cercopithecus neglectus | De Brazza’s monkey | 8 | 1 | 1 | 0 | LC |
| Chiropteraa | Bats | 0.01 | 0 | 1 | 123 | LC |
| Dendrohyrax dorsalis emini | Western tree hyrax | 4.5 | 0 | 0 | 0 | LC |
| Funisciurus spp. | African striped squirrels | 0.3 | 1 | 0 | 49 | LC |
| Galagoides demidoff | Demidoff’s dwarf galago | 0.06 | 0 | 0 | 11 | LC |
| Genetta maculata | Large-spotted genet | 3.2 | 0 | 0 | 5 | LC |
| Genetta servalina | Servaline genet | 2.7 | 0 | 0 | 0 | LC |
| Genetta victoriae | Giant genet | 3.5 | 0 | 0 | 1 | LC |
| Heliosciurus rufobrachium | Red-legged sun squirrel | 0.42 | 1 | 0 | 45 | LC |
| Muridaea | Murids | 0.05 | 0 | 0 | 83 | LC |
| Nandinia binotata | African palm civet | 3.2 | 0 | 0 | 1 | LC |
| Pan troglodytes | Robust (common) chimpanzee | 35 | 1 | 1 | 0 | EN |
| Papio anubis | Olive baboon | 29 | 1 | 1 | 0 | LC |
| Paraxerus boehmi | Böhm’s bush Squirrel | 0.1 | 1 | 0 | 0 | LC |
| Phataginus tricuspis | White-bellied pangolin | 3 | 0 | 0 | 6 | EN |
| Protoxerus stangeri | African giant squirrel | 0.8 | 1 | 0 | 46 | LC |
| (c) Ground-level substrate | ||||||
| Atherurus africanusc | Brush-tailed porcupine | 4 | 0 | 1 | 7 | LC |
| Bdeogale nigripes | Black-legged mongoose | 3.5 | 0 | 0 | 1 | LC |
| Cephalophus dorsalis | Bay duiker | 22 | 0 | 0 | 1 | NT |
| Civettictis civetta | African civet | 20 | 0 | 0 | 0 | LC |
| Cricetomys emini | Giant pouched rats | 1.4 | 0 | 0 | 13 | LC |
| Crossarchus alexandri | Alexander’s cusimanse | 2 | 1 | 1 | 0 | LC |
| Herpestes ichneumon | Ichneumon mongoose | 4 | 1 | 0 | 0 | LC |
| Hyemoschus aquaticus | Water chevrotain | 12 | 0 | 0 | 0 | LC |
| Orycteropus afer | Aardvark | 82 | 0 | 0 | 0 | LC |
| Panthera pardus | Leopard | 60 | 0 | 0 | 0 | VU |
| Philantomba monticola | Blue duiker | 9 | 1 | 0 | 13 | LC |
| Potamochoerus porcusc | Red river hog | 115 | 0 | 1 | 1 | LC |
| Rhynchocyon cirneic | Chequered giant sengi | 0.44 | 1 | 0 | 3 | LC |
| Smutsia gigantea | Giant pangolin | 35 | 0 | 0 | 0 | EN |
| Thryonomys swinderianus | Greater (marsh) cane rat | 5.2 | 0 | 1 | 0 | LC |
| Tragelaphus spekiic | Sitatunga | 119 | 0 | 1 | 1 | LC |
| (d) Volant species | AR | WL | Vmr | |||
| Afronycteris nana | Banana serotine | 6.4(0.9) | 3.6(0.4) | 3.1 | 2 | LC |
| Casinycteris argynnis | Short-palated fruit bat | 6.5(0.6) | 9.9(0.9) | 4.9(0.2) | 3 | LC |
| Doryrhina cyclops | Cyclops roundleaf bat | 6.2(0.5) | 9.3(1.7) | 4.8(0.3) | 2 | LC |
| Eidolon helvum | African straw-coloured fruit bat | 8.7(0.3) | 24.9(1.8) | 6.9(0.2) | 3 | NT |
| Epomops franqueti | Franquet's epauletted fruit bat | 6.7(0.3) | 16.3(1.2) | 6.2(0.3) | 11 | LC |
| Glauconycteris cf. beatrix | Beatrix's bat | 6.6 | 6.6 | 4 | 1 | LC |
| Glauconycteris humeralis | Allen's spotted bat | 5.7 | 3.1 | 3.1 | 1 | DD |
| Hipposideros beatus | Benito roundleaf bat | 6.7(0.7) | 5.8(0.3) | 3.8(0.1) | 7 | LC |
| Hipposideros caffer | Sundevall's roundleaf bat | 7(0.3) | 5.5(0.3) | 3.7(0.1) | 3 | LC |
| Hipposideros fuliginosus | Sooty roundleaf bat | 6.3 | 7.7 | 4.4 | 1 | LC |
| Hypsignathus monstrosus | Hammer-headed fruit bat | 7.2(0.5) | 19.4(1.5) | 6.6(0.3) | 3 | LC |
| Macronycteris gigas | Giant roundleaf bat | 7.7(0.8) | 16.1(0.1) | 5.9(0.1) | 2 | LC |
| Megaloglossus woermanni | Woermann's fruit bat | 5(0.3) | 7.8(1.3) | 4.7(0.7) | 5 | LC |
| Mimetillus moloneyid | Moloney's mimic bat | 6.7 | 18 | 6 | 0 | LC |
| Myonycteris torquata | Little collared fruit bat | 6.1(0.5) | 10.9(0.8) | 5.2(0.2) | 24 | LC |
| Myotis bocagii | Bocage’s mouse-eared bat | 6.8(0.7) | 6.2(1.3) | 3.9(0.4) | 3 | LC |
| Nycteris cf. arge | Bate's slit-faced bat | 5 | 3.6 | 3.4 | 1 | LC |
| Nycticeinops crassulus | Broad-headed serotine | 4.4 | 5.1 | 4 | 1 | LC |
| Scotonycteris bergmansi | Bergmans's fruit bat | 7.5(0.8) | 13.3(2.6) | 5.3(0.6) | 3 | NT |
| aData compiled at a higher taxonomic level because these taxa could not be reliably identified | ||||||
| bAnimal netted during the bat surveys | ||||||
| cCamera trap detections from the transition zone only | ||||||
| dThis species has been reported in a previous field note, the morphometric characteristics are from the literature (Norberg and Rayner 1987) | ||||||
| Note: Species with 0 detection and/or superscript (b), (c), (d) were modelled as not available in the detection history if not observed in the buffer zone or if this bat was not captured in both the transition and buffer zones of the Yangambi Biosphere Reserve | ||||||
Bat data were collected between November 2019 and April 2020. We recorded 76 bats during 324 trap-nights, including seven species of phytophagous bats (frugivores and nectarivores) and 11 species of insectivores. The sex ratio was 1.26 males/females for phytophagous and 1 male/female of insectivorous bats. The pregnancy rate was 26% of adult female phytophagous vs. 8% for insectivorous bats, while the lactation rate was 11% for phytophagous vs. 42% for insectivores. There were few juvenile in phytophagous bats (8%) and none in insectivores. Since taxonomy based on external morphological characters did not rule out morphologically identical cryptic species, descriptive parameters for each bat species are presented in Supplementary Information Table S1. Morphological characteristics converged to define three groups based on theoretical flight abilities, including slow-flying bats, which are generally insectivorous species (Mass < 12 g, wing loading < 7 N.m−², and maximum range speed < 15 km.h− 1, e.g., Hipposideros beatus), intermediate-speed bats (mass < 30 g, wing loading < 15 N.m−², maximum range speed < 20 km.h− 1, e.g., Hipposideros fuliginosus, Megaloglossus woermanni) and fast-flying bats (mass > 75 g, wing loading > 15 N.m−², maximum range speed > 20 km.h− 1, e.g., Macronycteris gigas, Eidolon helvum (Table 3d). Bat trap success distributions (both taxonomic groups) indicated that at least one event differed from others, H(2) = 9.24, \(\widehat{\epsilon }\)2 = 0.54, 95% CI (0.38, 1), N = 18, p = 0.009. Dunn's pairwise test showed that trap success at early evening (6:00pm-10:00pm) differed significantly from trap success at early morning (2:00am-6:00am, p-Holmadj = 0.008, Supplementary Information Appendix B). This means that bat capture was more successful at early evening than at early morning.
Species Detectability And Habitat Occupancy
Using all available ground and arboreal camera data, the most parsimonious model ψ(ST)p(mass + group) included a positive effect of stem density on occupancy probability (βψ = 9.06, SE 3.16, Table 2, Supplementary Information Table S3a). An average occupancy probability of 0.303 (SE 0.073) was estimated on this basis. Furthermore, species occurrence counts (N) at 18 camera sites were clustered into three body size groups to assess species distribution patterns. The median number of occurrences of small-bodied species corresponded to the sites with a high occupancy probabilities relative to the estimated overall mean (Supplementary Information Appendix C). Despite this, there was no change in the overall distribution of species occurrences across the study area, H(2) = 0.54, \(\widehat{\epsilon }\)2 = 0.004, 95% CI (0.0004, 1), N = 131, p = 0.76. The details of all camera recordings showed a low probability of detection, which correlated negatively with an increase in standing body mass (βp = -39.08, SE 7.53) and with species living in groups (βp = -0.51, SE 0.26, Table 2, Supplementary Information Table S3a). Average detection probability was thus 0.059 (SE 0.009, range: 0-0.152, Fig. 2). Although the median detections in the arboreal species was higher than the estimated overall mean, the detection probability distributions were statistically different between species according spatial use patterns, H(2) = 9.36, \(\widehat{\epsilon }\)2 = 0.21, 95% CI (0.12, 1), N = 45, p = 0.009. Thus, arboreal species had higher detection probabilities than ground-dwelling species (p-Holmadj = 0.007, Supplementary Information Appendix D).
Based on the model with the lowest AICc score, ψ(site)p(.) associated with the bat trapping data, site was found to be an important factor for habitat occupancy (βψ = -1.923, SE 0.966). On this basis we found an estimated average occupancy probability of 0.849 (SE 0.078). However, these trends in stand level occupancy suggest a possible co-variation effect with other uncontrolled conditions at larger spatial scales. However, this model did not account for covariates that could affect bat detection (Table 2, Supplementary Information Table S3b). The average detection probability was 0.162 (SE 0.024). For all models that sufficiently converged, detection estimates and correlation coefficients varied slightly (Table 2). This indicates that there are small differences between flight-related morphological traits and lifestyle within at the bat community level. Nevertheless, model rankings still hold regardless of the model chosen selected.
Regarding occupancy probability at each surveyed site, relationships between estimated species richness and number of camera detections per site were similar. Estimated species richness increased with the number of records, regardless of whether ground and arboreal detections were combined (Fig. 3a), ground only (Fig. 3b) or tree only (Fig. 3c). The methods which combined both ground and arboreal camera records throughout the study period provided more accurate estimates, with an average species richness of about 14 species (range: 7 to 21 species detected at a site, including bats, glirids and murids of high taxonomic rank). Yet, these trends represent less than half of the at least 45 species likely to occur. Bat surveys were planned in such a way that the maximum estimated species richness was achieved at the mean capture number (Fig. 3d). This tendency is evidenced by the average of 16 species (range: 7 to 19 species estimated per site) compared to the expected minimum of 19 species. This indicated that precision in estimating species richness depends on both sampling design and statistical modelling.
General findings
The study area was a typical old-growth rainforest described throughout the Congo Basin (Lewis et al. 2013), expected to support diverse mammal species and composition. Due to the temporal and spatial scale of this study, field records alone do not enable us to provide a clear explanation for the absence of most large-bodied species, and most species with elevated global conservation status. Although the two sampling efforts in the buffer and transition zones are not comparable, these observations seem consistent given the large number of active hunters living in and around the buffer zone and involved in the bushmeat market chain (Fig. 1a). The camera-trap records evidence tended to support the existence of small-bodied, fast reproducing mammals inhabiting the buffer zone and large-bodied, slow-reproducing taxa occurring in the transition zone. During the 2017–2018 surveys, Cricetomys emini, Atherurus africanus, Protoxerus stangeri, Dendrohyrax dorsalis emini and Philantomba monticola were the main mammal carcasses recorded by approximately 80% of hunters from the buffer zone. Cephalophus dorsalis, Cercopithecus ascanius ascanius and Potamochoerus porcus were the main mammals hunted in the transition during the same survey period (van Vliet et al. 2018).
In addition to the recorded endangered white-bellied pangolin (Phataginus tricuspis), the presence of the endemic endangered Kisangani red colobus (also known as Lang’s red colobus or Piliocolobus langi) has been reported in the literature for which the YAB is the only protected area (Linder et al. 2021). Research conducted in 2019, gathering information from traditional hunters, revealed that it has been more than 10 years since individuals of Kisangani red colobus were killed in the YAB (Kaisala et al. 2020). However, due to poor wildlife documentation in the YAB, the few existing reports are based on bushmeat surveys (van Vliet et al. 2018, 2019) providing little measurable on-the-ground indicators. These studies often led to ambiguity in the taxonomic description and provenance of traded species, overlooking animals that are not trafficked but hunted for hunter’s own consumption (Driscoll et al. 2018). It is therefore extremely difficult to match this type of information to what is actually documented in a specific area, as it is often distorted along the bushmeat transaction chain. As is the case elsewhere in the Afrotropics, this finding is akin to the consequence of a failed conservation management policy. Removing species from the wild at rates that cannot be compensated for by reproduction may increasingly impact resident fauna.
The MAB programme alone cannot bridge this theoretical-practical gap in fulfilling BR functions in low-income countries, where conservation programmes are sector-specific and funder-driven. Given that each BR undergoes periodic ten-year reviews (Kratzer 2018), several possible scenarios exist for how BR management policies could improve human-environment interactions. Building on this study, one prospect is to interweave different protected area statuses into a single large area to catalyse multi-stakeholder initiatives on social issues and ecosystem services in conservation management policies (Longato et al. 2021). The BRs are not spatially exclusive. This means that the buffer and/or transition zone may include areas with other international designations, such as Ramsar Sites (Bridgewater and Kim 2021; Kingsford et al. 2021) and/or World Heritage Sites (Sayer et al. 2000; Rodwell 2012). Such a mechanism for allocating land can help to create a coherent local network of protected areas. This may provide an important incentive for new approaches to management and conservation, which should be tailored to local cultural, economic and social dimensions.
Population Viability And Habitat Use
We analysed detections of volant and nonvolant mammals and the characteristics of habitat in which they occurred. Although available species with arboreal use patterns were more frequently recorded, the dominance of small-bodied and solitary species influenced the range of estimates at the community level. These findings suggest that increased human activity associated with hunting, and even the presence of the traps themselves, may alter activity patterns of animals, making them less detectable (Moore et al. 2021). To some extent, these patterns of community structuring may rely on an underlying effect of extirpating large, low-hunting-resilient species, from which only opportunistic high-hunting-resilient species persisted (Fa et al. 2015). Furthermore, a higher number of animal occurrences was recorded in the sites with a denser stem structure. Mammal occurrence in terrestrial ecosystems relies on the perception of environmental cues through severe trade-offs between foraging efficiency, risk avoidance, and resting or rumination sites (Ehlers Smith et al. 2017; Carreira et al. 2020). The lack of suitable habitat availability in the changing forest biome is likely to be a major limiting factor for the population of specialist species (Tscharntke et al. 2005). Species richness and composition are therefore shaped by habitat characteristics, including forest structure (DeWalt et al. 2003), such that a land-use matrix would be underutilised by many nonvolant forest mammals.Compared to field-reported observations from tropical forests in Borneo (Struebig et al. 2012), Malaysia (Francis 1994) and Thailand (Stewart et al. 2014), our bat trap-based estimates of insectivorous and phytophagous bat population parameters provided representative information. The inter-species estimates of bats occupancy support a possible co-variation effect with other uncontrolled conditions at larger spatial scales. In a community-level study, it may be difficult to disentangle such species-level differences in occupancy patterns, particularly given differences between species in body size, ability to receive food cues and foraging tactics (Lomáscolo and Schaefer 2010; Hodgkison et al. 2013). Due to the abundance of records of phytophagous bats, their species-specific traits are likely influenced the community-level occupancy estimates. This may be due to a combination of a set of species-level traits and environmental requirements, including variations in ambient temperature and cloudiness, which are known to influence bat flight activity (Thomas and Jacobs 2013). Such patterns of space use raise the question of whether territorial habits exist, and how they relate to home range size and species-specific endogenous traits. In fact, habitat occupancy in the phytophagous bats of the lowland rainforest can be competitive between species, since territoriality is directly related to site fidelity and species solitariness (Bonaccorso et al. 2005). More strikingly, species with low to moderate flight performance and smaller home ranges, such as Megaloglossus woermanni, have mean home ranges of about 147 ha (Weber et al. 2009). Where habitat degradation is significant, competition for vital resources may increase, leading territorial species to forage in multiple patches. This may force optimal foragers to leave and move to the next patch (Brown 1988). To some extent, this may trigger a latent process of depletion of the local bat fauna, or even of wildlife as a whole. However, few studies have adequately addressed the nature and magnitude of the response of bats to a range of extrinsic factors, and such assessments are still poorly documented in the Afrotropics.
When estimating population parameters, bat trap-based estimates can have analytical drawbacks. Our modelled occupancy estimates circumvented this disadvantage by using multi-species data, combining current and historical data to account for possible species in area. This study highlights the importance of using multi-species data, combining methods for monitoring large and small, flying and non-flying, terrestrial and arboreal mammal species and their habitats to evaluate population viability. Using such an approach provides more accurate, measurable field evidence in an area subject to latent processes of mammal depletion. Nonvolant mammals have retracted their range into denser forest stands, where the population parameters of the small-bodied species influence all community-level estimates. As bats can fly, they occur randomly in any forest stands, so that their capture-based detections are not associated with any morphological trait estimator. Given a series of aspects of the bat life-history, such as larger home ranges, powerful flight regimes for most species, less restrictive diets and ubiquitous distribution in the Afrotropical mainland, we can postulate that the habitat use patterns in bats can be equated with a proxy for the functioning of rainforest ecosystems. We expect that the use of these methods will be strongly encouraged to stimulate integrative studies of the biodiversity, which is undergoing acute changes at global, regional and local levels.
Acknowledgments
We thank the Center for International Forestry Research and the Faculty of Science of the University of Kisangani for providing facilities under the 11th European Development Fund grant. We also thank the International Foundation for Science, L’Ecole Régionale Postuniversitaire d’Aménagement et de Gestion intégrés des Forêts et Territoires Tropicaux through the Belgian government fund, and IDEA WILD, a 501(c)(3) non-profit organization https://ideawild.org, for supporting this research. The authors thank the Institut National pour l’Etude et la Recherche Agronomique de Yangambi for cooperation during the fieldwork.
Funding
This work was supported by the 11th European Development Fund (reference: FORETS project), the International Foundation for Science (grant number: D/6071-1), and the Belgian government (reference: YPS project).
Competing interests
The authors declare no conflict of interest.
Ethical approval
The Scientific Committee of the University of Kisangani approved the research protocol following biosafety standards, animal welfare principles and nature conservation rules (permit FS/D/KGB/007/2018). Institutional animal care and use committees’ approval was not required for the non-invasive observations. Live capture procedures and bat handling complied with the guidelines of the Animal Care and Use Committee of the American Society of Mammalogists (Sikes 2016).
Author contributions
CM, AL, EV and GCG designed the study and developed the methodology. CM collected the data. CM, PB, TMB and VVC identified the species. CM and JFM analysed the data. CM, FVdP and JFM wrote the original draft. AL, CM, EV, FVdP, GCG, JFM, PB, TMB and VVC reviewed and edited the manuscript. All authors approved the final version of the manuscript.
Data availability
The datasets generated and/or analysed as well as R code can be made available for research purposes upon request to the corresponding authors.
- Adamescu GS, Plumptre AJ, Abernethy KA, et al (2018) Annual cycles are the most common reproductive strategy in African tropical tree communities. Biotropica 50:418–430. https://doi.org/10.1111/btp.12561
- Almeida FC, Amador LI, Giannini NP (2021) Explosive radiation at the origin of Old World fruit bats (Chiroptera, Pteropodidae). Org Divers Evol 21:231–243. https://doi.org/10.1007/s13127-021-00480-5
- Benítez-López A, Alkemade R, Schipper AM, et al (2017) The impact of hunting on tropical mammal and bird populations. Science (80- ) 356:180–183. https://doi.org/10.1126/science.aaj1891
- Bonaccorso FJ, Winkelmann JR, Byrnes DGP (2005) Home range, territoriality, and flight time budgets in the Black-Bellied fruit bat, Melonycteris melanops (Pteropodidae). J Mammal 86:931–936. https://doi.org/10.1644/1545-1542(2005)86[931:HRTAFT]2.0.CO;2
- Bridgewater P, Kim RE (2021) The Ramsar Convention on Wetlands at 50. Nat Ecol Evol 5:268–270. https://doi.org/10.1038/s41559-021-01392-5
- Brown JS (1988) Patch use as an indicator of habitat preference, predation risk, and competition. Behav Ecol Sociobiol 22:37–47. https://doi.org/10.1007/BF00395696
- Burnham KP, Anderson DR, Huyvaert KP (2011) AIC model selection and multimodel inference in behavioral ecology: some background, observations, and comparisons. Behav Ecol Sociobiol 65:23–35. https://doi.org/10.1007/s00265-010-1029-6
- Cakenberghe VV, Gembu Tungaluna G-C, Musaba Akawa P, et al (2017) The bats of the Congo and of Rwanda and Burundi revisited (Mammalia: Chiroptera). Eur J Taxon 1–327. https://doi.org/10.5852/ejt.2017.382
- Cardillo M, Mace MG, Jones KE, et al (2005) Multiple Causes of High Extinction Risk in Large Mammal Species. Science (80- ) 309:1239–1241. https://doi.org/10.1126/science.1116030
- Carreira DC, Brodie JF, Mendes CP, et al (2020) A question of size and fear: competition and predation risk perception among frugivores and predators. J Mammal 101:648–657. https://doi.org/10.1093/jmammal/gyaa034
- Couvreur M, Christiaen B, Verheyen K, Hermy M (2004) Large herbivores as mobile links between isolated nature reserves through adhesive seed dispersal. Appl Veg Sci 7:229–236. https://doi.org/10.1111/j.1654-109X.2004.tb00614.x
- Davidson AD, Hamilton MJ, Boyer AG, et al (2009) Multiple ecological pathways to extinction in mammals. Proc Natl Acad Sci 106:10702–10705. https://doi.org/10.1073/pnas.0901956106
- Davies JT, Meiri S, Barraclough TG, Gittleman JL (2007) Species co-existence and character divergence across carnivores. Ecol Lett 10:146–152. https://doi.org/10.1111/j.1461-0248.2006.01005.x
- Davis CL, Rich LN, Farris ZJ, et al (2018) Ecological correlates of the spatial co‐occurrence of sympatric mammalian carnivores worldwide. Ecol Lett 21:1401–1412. https://doi.org/10.1111/ele.13124
- Devarajan K, Morelli TL, Tenan S (2020) Multi‐species occupancy models: review, roadmap, and recommendations. Ecography (Cop) 43:1612–1624. https://doi.org/10.1111/ecog.04957
- DeWalt SJ, Maliakal SK, Denslow JS (2003) Changes in vegetation structure and composition along a tropical forest chronosequence: implications for wildlife. For Ecol Manage 182:139–151. https://doi.org/10.1016/S0378-1127(03)00029-X
- Di Bitetti MS, De Angelo CD, Di Blanco YE, Paviolo A (2010) Niche partitioning and species coexistence in a Neotropical felid assemblage. Acta Oecologica 36:403–412. https://doi.org/10.1016/j.actao.2010.04.001
- Driscoll DA, Bland LM, Bryan BA, et al (2018) A biodiversity-crisis hierarchy to evaluate and refine conservation indicators. Nat Ecol Evol 2:775–781. https://doi.org/10.1038/s41559-018-0504-8
- Ehlers Smith YC, Ehlers Smith DA, Ramesh T, Downs CT (2017) The importance of microhabitat structure in maintaining forest mammal diversity in a mixed land-use mosaic. Biodivers Conserv 26:2361–2382. https://doi.org/10.1007/s10531-017-1360-6
- Elangovan V, Raghuram H, Yuvana Satya Priya E, Marimuthu G (2004) Wing Morphology and Flight Performance in Rousettus Leschenaulti. J Mammal 85:806–812. https://doi.org/10.1644/BWG-131
- Emmons LH (1980) Ecology and Resource Partitioning among Nine Species of African Rain Forest Squirrels. Ecol Monogr 50:31–54. https://doi.org/10.2307/2937245
- Fa JE, Olivero J, Real R, et al (2015) Disentangling the relative effects of bushmeat availability on human nutrition in central Africa. Sci Rep 5:8168. https://doi.org/10.1038/srep08168
- Farneda FZ, Rocha R, López-Baucells A, et al (2015) Trait-related responses to habitat fragmentation in Amazonian bats. J Appl Ecol 52:1381–1391. https://doi.org/10.1111/1365-2664.12490
- Ferreira DF, Rocha R, López-Baucells A, et al (2017) Season-modulated responses of Neotropical bats to forest fragmentation. Ecol Evol 7:4059–4071. https://doi.org/10.1002/ece3.3005
- Francis CM (1989) A Comparison of Mist Nets and Two Designs of Harp Traps for Capturing Bats. J Mammal 70:865–870. https://doi.org/10.2307/1381730
- Francis CM (1994) Vertical stratification of fruit bats (Pteropodidae) in lowland dipterocarp rainforest in Malaysia. J Trop Ecol 10:523–530. https://doi.org/10.1017/S0266467400008191
- Gelman A, Tuerlinckx F (2000) Type S error rates for classical and Bayesian single and multiple comparison procedures. Comput Stat 15:373–390. https://doi.org/10.1007/s001800000040
- Hodgkison R, Ayasse M, Häberlein C, et al (2013) Fruit bats and bat fruits: The evolution of fruit scent in relation to the foraging behaviour of bats in the New and Old World tropics. Funct Ecol 27:1075–1084. https://doi.org/10.1111/1365-2435.12101
- Hoyle SD, Pople AR, Toop GJ (2001) Mark-recapture may reveal more about ecology than about population trends: Demography of a threatened ghost bat (Macroderma gigas) population. Austral Ecol 26:80–92. https://doi.org/10.1046/j.1442-9993.2001.01092.x
- Hutchinson GE (1959) Homage to Santa Rosalia or Why Are There So Many Kinds of Animals? Am Nat 93:145–159. https://doi.org/10.1086/282070
- Kaisala D, Falay P, Mande C, Hart JA (2020) Rediscovery of the Kisangani Red Colobus Distribution , taxonomic and phenotypic variation , ecology , threats , and recommendations for its conservation. In: Searching for Bonobo in Congo. https://www.bonoboincongo.com/. Accessed 17 Jul 2022
- Kingdon J (2015) The Kingdon field guide to African Mammals, 2nd edn. Bloomsbury Publishing Plc, London
- Kingdon J, Happold DCD, Butynski T, et al (2013a) Mammals of africa: Rodents, Hares and Rabbies, Volume III. Bloomsbury Publishing, London
- Kingdon J, Happold DCD, Butynski T, et al (2013b) Mammals of africa: Hippopotamuses, Pigs, Deers, Giraffe and Bovids, Volume VI. Bloomsbury Publishing, London
- Kingdon J, Hoffmann M (2013) Mammals of Africa. Volume V: Carnivores, Pangolins, Equids and Rhinoceroses, 2nd edn. Bloomsbury Publishing Plc, London
- Kingsford RT, Bino G, Finlayson CM, et al (2021) Ramsar Wetlands of International Importance–Improving Conservation Outcomes. Front Environ Sci 9:1–6. https://doi.org/10.3389/fenvs.2021.643367
- Kipute DD, Mampeta SW, Kahindo J-MM, et al (2021) Complexity of zoning management in biosphere reserves: the case of the Yangambi Biosphere Reserve in DRC. Int For Rev 23:55–67. https://doi.org/10.1505/146554821832140358
- Koy JK, Ngonga AMM, Wardell DA (2019) Moving beyond the illusion of participation in the governance of Yangambi Biosphere Reserve (Tshopo Province, Democratic Republic of Congo). Nat Conserv 33:33–54. https://doi.org/10.3897/natureconservation.33.30781
- Kratzer A (2018) Biosphere Reserves research: a bibliometric analysis. ecomont (Journal Prot Mt Areas Res Manag 10:36–49. https://doi.org/10.1553/eco.mont-10-2s36
- Kratzer A, Ammering U (2019) Rural innovations in biosphere reserves – A social network approach. J Rural Stud 71:144–155. https://doi.org/10.1016/j.jrurstud.2019.01.001
- Kremen C, Williams NM, Aizen MA, et al (2007) Pollination and other ecosystem services produced by mobile organisms: a conceptual framework for the effects of land-use change. Ecol Lett 10:299–314. https://doi.org/10.1111/j.1461-0248.2007.01018.x
- Kumar A (2020) CBD’s Global Biodiversity Outlook 5: Final Assessment of Aichi Biodiversity Targets and Beyond. Asian Biotechnol Dev Rev 22:5–20
- Kunz TH, Anthony ELP (1982) Age Estimation and Post-Natal Growth in the Bat Myotis lucifugus. J Mammal 63:23–32. https://doi.org/10.2307/1380667
- Laméris DW, Tagg N, Kuenbou JK, et al (2020) Drivers affecting mammal community structure and functional diversity under varied conservation efforts in a tropical rainforest in Cameroon. Anim Conserv 23:182–191. https://doi.org/10.1111/acv.12526
- Lewis SL, Sonké B, Sunderland T, et al (2013) Above-ground biomass and structure of 260 African tropical forests. Philos Trans R Soc B Biol Sci 368:20120295. https://doi.org/10.1098/rstb.2012.0295
- Linder JM, Cronin DT, Ting N, et al (2021) Red colobus (Piliocolobus) conservation action plan 2021-2026. IUCN, International Union for Conservation of Nature
- Lomáscolo SB, Schaefer HM (2010) Signal convergence in fruits: A result of selection by frugivores? J Evol Biol 23:614–624. https://doi.org/10.1111/j.1420-9101.2010.01931.x
- Longato D, Cortinovis C, Albert C, Geneletti D (2021) Practical applications of ecosystem services in spatial planning: Lessons learned from a systematic literature review. Environ Sci Policy 119:72–84. https://doi.org/10.1016/j.envsci.2021.02.001
- Lučan RK, Bartonička T, Benda P, et al (2014) Reproductive seasonality of the Egyptian fruit bat ( Rousettus aegyptiacus ) at the northern limits of its distribution. J Mammal 95:1036–1042. https://doi.org/10.1644/14-MAMM-A-035
- Lundberg J, Moberg F (2003) Mobile Link Organisms and Ecosystem Functioning: Implications for Ecosystem Resilience and Management. Ecosystems 6:0087–0098. https://doi.org/10.1007/s10021-002-0150-4
- Mackenzie DI, Nichols JD, Lachman G, et al (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology 83:2248–2255. https://doi.org/doi.org/10.1890/0012-9658(2002)083[2248:ESORWD]2.0.CO;2
- MacKenzie DI, Royle JA (2005) Designing occupancy studies: general advice and allocating survey effort. J Appl Ecol 42:1105–1114. https://doi.org/10.1111/j.1365-2664.2005.01098.x
- Matthews LJ, Arnold C, Machanda Z, Nunn CL (2011) Primate extinction risk and historical patterns of speciation and extinction in relation to body mass. Proc R Soc B Biol Sci 278:1256–1263. https://doi.org/10.1098/rspb.2010.1489
- Miller DA, Nichols JD, McClintock BT, et al (2011) Improving occupancy estimation when two types of observational error occur: non-detection and species misidentification. Ecology 92:1422–1428. https://doi.org/10.1890/10-1396.1
- Mills LS, Soulé ME, Doak DF (1993) The keystone-species concept in ecology and conservation. Bioscience 43:219–224. https://doi.org/https://doi.org/10.2307/1312122
- Miño MH, Cavia R, Villafañe IEG, et al (2007) Seasonal abundance and distribution among habitats of small rodents on poultry farms. A contribution for their control. Int J Pest Manag 53:311–316. https://doi.org/10.1080/09670870601105949
- Monadjem A, Taylor PJ, Cotterill FPD (Woody), Schoeman MC (2020) Bats of Southern and Central Africa: A biogeographic and taxonomic synthesis, second edition, 2nd edn. Wits University Press, Johannesburg
- Moore JF, Pine WE, Mulindahabi F, et al (2020) Comparison of species richness and detection between line transects, ground camera traps, and arboreal camera traps. Anim Conserv 23:561–572. https://doi.org/10.1111/acv.12569
- Moore JF, Uzabaho E, Musana A, et al (2021) What is the effect of poaching activity on wildlife species? Ecol Appl 283. https://doi.org/10.1002/eap.2397
- Norberg UM, Rayner JM V. (1987) Ecological morphology and flight in bats (Mammalia; Chiroptera): wing adaptations, flight performance, foraging strategy and echolocation. Philos Trans R Soc London B, Biol Sci 316:335–427. https://doi.org/10.1098/rstb.1987.0030
- Ostertagová E, Ostertag O, Kováč J (2014) Methodology and Application of the Kruskal-Wallis Test. Appl Mech Mater 611:115–120. https://doi.org/10.4028/www.scientific.net/AMM.611.115
- Patterson BD, Webala PW (2012) Keys to the Bats (Mammalia: Chiroptera) of East Africa. Fieldiana Life Earth Sci 6:1–60. https://doi.org/10.3158/2158-5520-12.6.1
- QGIS Development Team (2019) QGIS Geographic Information System. Open Source Geospatial Foundation Project
- R Core Team (2021) R: A Language and Environment for Statistical Computing. R Found Stat Comput https://www.R-project.org
- Reed MG (2016) Conservation (In)Action: Renewing the Relevance of UNESCO Biosphere Reserves. Conserv Lett 9:448–456. https://doi.org/10.1111/conl.12275
- RGC (2010) Référentiel Géographique Commun de la République Démocratique du Congo (RGC). Inst Géographique du Congo
- Rhodes MP (2002) Assessment of sources of variance and patterns of overlap in microchiropteran wing morphology in southeast Queensland, Australia. Can J Zool 80:450–460. https://doi.org/10.1139/z02-029
- Rocha R, López-Baucells A, Farneda FZ, et al (2016) Consequences of a large-scale fragmentation experiment for Neotropical bats: disentangling the relative importance of local and landscape-scale effects. Landsc Ecol 32:31–45. https://doi.org/10.1007/s10980-016-0425-3
- Rodwell D (2012) The Unesco World Heritage Convention, 1972–2012: Reflections and Directions. Hist Environ Policy Pract 3:64–85. https://doi.org/10.1179/1756750512Z.0000000004
- Rosenzweig ML (1966) Community Structure in Sympatric Carnivora. J Mammal 47:602–612. https://doi.org/10.2307/1377891
- Rovero F, Ahumada J (2017) The Tropical Ecology, Assessment and Monitoring (TEAM) Network: An early warning system for tropical rain forests. Sci Total Environ 574:914–923. https://doi.org/10.1016/j.scitotenv.2016.09.146
- Royle JA, Kéry M (2007) A Bayesian state-space formulation of dynamic occupancy models. Ecology 88:1813–1823. https://doi.org/10.1890/06-0669.1
- Sayer J, Ishwaran N, Thorsell J, Sigaty T (2000) Tropical Forest Biodiversity and the World Heritage Convention. AMBIO A J Hum Environ 29:302–309. https://doi.org/10.1579/0044-7447-29.6.302
- Scholes RJ, Biggs R (2005) A biodiversity intactness index. Nature 434:45–49. https://doi.org/10.1038/nature03289
- Schunk C, Swartz SM, Breuer KS (2017) The influence of aspect ratio and stroke pattern on force generation of a bat-inspired membrane wing. Interface Focus 7:20160083. https://doi.org/10.1098/rsfs.2016.0083
- Sikes RS (2016) 2016 Guidelines of the American Society of Mammalogists for the use of wild mammals in research and education: J Mammal 97:663–688. https://doi.org/10.1093/jmammal/gyw078
- Sikuzani YU, Muteya HK, Bogaert J (2020) Miombo woodland, an ecosystem at risk of disappearance in the Lufira Biosphere Reserve (Upper Katanga, DR Congo)? A 39-years analysis based on Landsat images. Glob Ecol Conserv 24:e01333. https://doi.org/10.1016/j.gecco.2020.e01333
- Sritongchuay T, Hughes AC, Bumrungsri S (2019) The role of bats in pollination networks is influenced by landscape structure. Glob Ecol Conserv 20:e00702. https://doi.org/10.1016/j.gecco.2019.e00702
- Stewart AB, Makowsky R, Dudash MR (2014) Differences in foraging times between two feeding guilds within Old World fruit bats (Pteropodidae) in southern Thailand. J Trop Ecol 30:249–257. https://doi.org/10.1017/S0266467414000042
- Struebig MJ, Bożek M, Hildebrand J, et al (2012) Bat diversity in the lowland forests of the Heart of Borneo. Biodivers Conserv 21:3711–3727. https://doi.org/10.1007/s10531-012-0393-0
- Teeling EC, Springer MS, Madsen O, et al (2005) A Molecular Phylogeny for Bats Illuminates Biogeography and the Fossil Record. Science (80- ) 307:580–584. https://doi.org/10.1126/science.1105113
- Thomas AJ, Jacobs DS (2013) Factors Influencing the Emergence Times of sympatric Insectivorous Bat Species. Acta Chiropterologica 15:121–132. https://doi.org/10.3161/150811013X667920
- Tscharntke T, Klein AM, Kruess A, et al (2005) Landscape perspectives on agricultural intensification and biodiversity - ecosystem service management. Ecol Lett 8:857–874. https://doi.org/10.1111/j.1461-0248.2005.00782.x
- Van Cuong C, Dart P, Hockings M (2017) Biosphere reserves: Attributes for success. J Environ Manage 188:9–17. https://doi.org/10.1016/j.jenvman.2016.11.069
- Van de Perre F, Willig MR, Presley SJ, et al (2018) Reconciling biodiversity and carbon stock conservation in an Afrotropical forest landscape. Sci Adv 4:eaar6603. https://doi.org/10.1126/sciadv.aar6603
- van Vliet N, Muhindo J, Nyumu KJ, et al (2018) Mammal Depletion Processes as Evidenced From Spatially Explicit and Temporal Local Ecological Knowledge. Trop Conserv Sci 11:194008291879949. https://doi.org/10.1177/1940082918799494
- van Vliet N, Muhindo J, Nyumu KJ, Nasi R (2019) From the Forest to the Dish: A Comprehensive Study of the Wildmeat Value Chain in Yangambi, Democratic Republic of Congo. Front Ecol Evol 7:132. https://doi.org/10.3389/fevo.2019.00132
- Weber N, Kalko EK V, Fahr J (2009) A First Assessment of Home Range and Foraging Behaviour of the African Long-Tongued Bat Megaloglossus woermanni (Chiroptera: Pteropodidae) in a Heterogeneous Landscape within the Lama Forest Reserve, Benin. Acta Chiropterologica 11:317. https://doi.org/10.3161/150811009X485558
- Wilkie DS, Bennett EL, Peres CA, Cunningham AA (2011) The empty forest revisited. Ann N Y Acad Sci 1223:120–128. https://doi.org/10.1111/j.1749-6632.2010.05908.x
- Wright JP, Jones CG (2006) The concept of organisms as ecosystem engineers ten years on: Progress, limitations, and challenges. Bioscience 56:203–209. https://doi.org/10.1641/0006-3568
- Zipkin EF, Andrew Royle J, Dawson DK, Bates S (2010) Multi-species occurrence models to evaluate the effects of conservation and management actions. Biol Conserv 143:479–484. https://doi.org/10.1016/j.biocon.2009.11.016
No competing interests reported.