Though tropical forests cover less than 10% of the Earth’s land surface (Bradshaw et al., 2009), they are incredibly diverse and support at least two-thirds of the world’s biodiversity (Raven, 1988). Current research estimates that tropical forests will have one of the highest rates of natural vegetation loss in the near future (Smith, et al., 2023). Protecting tropical forests is the central thrust of several national and international biodiversity conservation strategies (CBD, 2015; Panfil & Harvey, 2015). Tropical forest provides substantial local, regional, and global benefits through economic goods and ecosystem services. As they have greater species diversity than other ecosystems, tropical forest biodiversity losses are likely greater (Rosenzweig, 1992). Mammals in tropical forests are a priority group as they represent a rich and functionally diverse component of this biome even as they face threats such as hunting, habitat loss, and habitat fragmentation (Ceballos et al., 2005; Schipper et al., 2008; Visconti et al., 2011; Ahumada et al., 2011). Tropical mammals are among the most threatened species globally, yet their status remains challenging to establish due to the inherent difficulty in their observation (Linkie et al., 2007). It is important to consider that their long-term trends may not be well-established, and the ‘most threatened’ designation assumes certain knowledge about their status. Information on distribution, habitat utilization, and threat response are critical to developing conservation strategies (Espartosa et al., 2011).

Mammals play a crucial role in maintaining ecosystem functions, including forest structures, nutrient cycling, carbon storage, seed dispersal and trophic management (Brodie et al., 2009; Jansen et al., 2010; Estes et al., 2011; Sobral et al., 2017). However, the secretive and cryptic nature of forest-dwelling mammals hinders data collection and the development of conservation measures. Statistical techniques, such as occupancy modeling, which considers and corrects imperfect detection, are essential to generate evidence for environmental policies (MacKenzie et al., 2002; Brodie et al., 2015; Guillera-Arroita, 2017). Track and sign surveys, direct observations, camera trapping, and interviews with local communities and hunters currently detect mammalian fauna (Ramesh and Downs, 2014; De Bondi et al., 2010; Aiyudurai et al., 2010). Numerous studies indicate that camera traps are an effective method to survey mammals, including rare, elusive, and nocturnal species (Silveira et al., 2003; Tobler et al., 2008; Harmsen et al., 2010; Espartosa et al., 2011; O’Connell et al., 2011). The use of camera traps to collect data on the distribution and abundance of mammals can provide a wealth of information (Rowcliffe and Carbone, 2008; Tobler et al., 2009; Pettorelli et al., 2010; Kinnaird and O’Brien, 2012; Ramesh et al., 2012; Ramesh et al., 2013). Sign surveys (Carrillo et al., 2000; Naughton-Treves et al., 2003; Harvey et al., 2006; Bali et al., 2007; Parry et al., 2007; Norris et al., 2008) are increasingly being used to survey tropical forest mammals. In tropical forests, suitable substrates for footprint impressions are not always present, and weather conditions lead to biases in sign surveys (Dirzo and Miranda, 1990; Norris et al., 2008).

To reconcile conservation and development, it is increasingly important to comprehend how human-modified landscapes in the tropics affect mammal communities (Boron et al., 2019). In tropical rainforests, habitat loss, hunting, land-use change, and deforestation threaten biodiversity conservation (Schipper et al., 2008; Visconti et al., 2011; Barlow et al., 2016; Hansen et al., 2008; Laurance and Luarance, 1999). Agriculture and tree plantations are rapidly increasing across the tropics to meet rising human demand for food, timber, and fiber (Hansen et al., 2013; Abood et al., 2015; Haddad et al., 2015; Riitters et al., 2016). These developments are lowering the area of forests and fragmenting or reducing wildlife habitat (Laurance et al., 2002; Fischer and Lindenmayer, 2007; Laurance et al., 2011). In addition, the problem of hunting has worsened due to expanding human populations, easier access to remote forests, and the use of modern hunting techniques and weapons. According to recent estimates, hunting pressure and land-use changes have caused mammalian populations to decline by more than 80% and 30%, respectively (Almeida-Rocha et al., 2017; Benítez-López et al., 2017). The native populations of local communities in northeast India have more in common with Southeast Asian cultures than peninsular India, and hunting is common (Datta, 2007). Hunting is common in Arunachal Pradesh, and it has decimated wildlife populations in the state (Datta, 2002; Hilaluddin and Ghose, 2005; Mishra et al., 2006).

Several surveys and documentation have been carried out in Arunachal Pradesh on large and small mammals (Athreya et al., 1997; Chowdhury, 1997; Datta, 1998; Datta et al., 2008a; Mishra et al., 2006; Gopi et al., 2010; Gopi et al., 2012; Selvan, 2013; Dasgupta et al., 2014; Selvan et al., 2014b; Dasgupta et al., 2015; Adhikarimayum and Gopi, 2018), Malayan sun bear (Sethy and Chauhan, 2012) and red panda (Kakati, 1996). Most wildlife surveys in Arunachal have been restricted to low and mid-elevation forests and have focused on rare species (Katti et al., 1990; Athreya and Johnsingh, 1995; Selvan et al., 2013; Roy et al., 2015). Very few surveys or ecological research in the state have produced significant discoveries and information (Datta and Goyal, 1997; Selvan et al., 2014a) of new species like leaf deer (Datta et al., 2003), Chinese goral (Mishra et al., 2004) and Arunachal macaque (Sinha et al., 2005).

In this context, we initiated the first systematic camera trapping study to evaluate the assemblage of terrestrial mammals in Mehao WLS and to understand factors affecting detection probability and occupancy. Mehao WLS is located in Arunachal Pradesh in the eastern Himalayas, which has been recognized as a global biodiversity hotspot (Myers et al., 2000). The rugged terrain and dense forest cover of Mehao WLS harbor 60 species of mammals (Chakraborty and Sen, 1991). Idu Mishmi and Adi are the main communities living around Mehao WLS. The Idu Mishmi community has social taboos restricting hunting and forest use, which may have contributed to protecting biodiversity in the area (Nijhawan and Mihu, 2020). Few studies have been conducted on mammals in Mehao WLS (Katti et al., 1990; Chakraborty and Sen, 1991; Sarma et al., 2015; Ahmad et al., 2023). The ones that have been carried out focus primarily on the status and distribution of hoolock gibbon (Chetry et al., 2010; Roy et al., 2015), while a write-up on Mehao WLS was published by Sinha (1984).

Our specific objectives for this research were (1) To determine the species richness and relative detection rate of mammals in and around Mehao WLS and (2) To investigate the factors influencing the detection probability and occupancy of mammalian fauna. The findings of this study are of paramount importance for managing and conserving rare and elusive species in and around Mehao WLS.

The current study was conducted in Mehao WLS, Lower Dibang Valley district of Arunachal Pradesh ( Figure 1 ). The sanctuary was established in 1980 and is home to numerous species of wildlife, including tigers, leopards, hoolock gibbons, tree shrews, wild dogs, and clouded leopards. Mehao WLS lies between 28˚ 05’ and 28˚ 15’ north latitude and 93˚ 30’ and 95˚ 45’ east longitudes and is spread over an area of 281.5 sq km. The terrain in Mehao WLS is plain in the southern section and hilly in the northern sections, with an elevation range of 400-3,568m above mean sea level. The valleys are narrow around river courses. The sanctuary is dotted with small flat areas at various elevations that provide habitat for various wildlife species. All the streams flow into the Dibang River, which later empties into the Lohit River—an important tributary of the Brahmaputra River. Several seasonal streams drain into these rivers, which flow rapidly during the monsoon.

Champion and Seth (1968) state that the study area falls within 8B/CI East Himalayan Sub-tropical Wet Hill Forests. The forest type changes with altitude in Mehao WLS. It includes tropical evergreen forest (900m amsl), subtropical and temperate forest (900m-1,800m amsl), temperate broad leave forest (1,800m-2,800m amsl), and temperate conifer forest (2,800m-3,500m amsl). The forest is dominated by Terminalia myriocarpa, Messua ferrea, Dillenia indica, Castanopsis indica, and Albizia lucida while primary shrub and herb species include Clerodendrum viscosum, Tephrosia candida, Maesa indica, and Solanum khasianum. Bamboo species in the area include Dendrocalamus giganteus, Phyllostachys bambusoides, Bambusa pallida, Dendrocalamus hamiltonii, and Melocanna baccifera. Mehao WLS is also home to Mishmi teeta (Coptis teeta), an extremely valuable medicinal herb and threatened holoparasitic plant, Sapria himalayana (Ahmad et al., 2020). The sanctuary is warmer at lower altitudes and cooler at higher altitudes. In the summer, the average temperature at a lower altitude is around 38°C and falls to 4°C- 12°C in the winter. At higher altitudes, the summer temperature is around 12°C, and it experiences snowfall in the winter with temperatures around -3°C.

Numerous Idu Mishmi villages along the sanctuary’s periphery include Koronu, Injuno, Abango, Balek, Simari, and Tiwarigaon. Strong taboos exist among the Idu Mishmi, prohibiting hunting gibbons and other wild animals. This has probably contributed to the abundance of wildlife in the region.

The field survey was conducted between November 2017 to February 2018. Reconnaissance was carried out during the initial months of the study period to develop adequate knowledge and understanding of the area while also collecting baseline data for further intensive sampling. An effort was made to deploy camera trap units randomly throughout the study area. A total of 58 camera traps were deployed with only one per 3 sq km grid, thus covering an area of 174 sq km. Cuddeback C1 camera traps were used. They have a trigger speed of 1.4 seconds, a detection range of 100 feet and were programmed to recover as fast as possible (FAP) after being triggered. Along animal trails, cameras were placed 20 cm above the ground and left active for 24 hours to capture images of all mammalian species encountered. Each grid contained one camera trap site with an average distance of 1.2 to 1.5 km between each trap. Then, we calculated Global Moran’s I (Moran, 1950) using the Spatial Autocorrelation tool in ArcGIS (version 10.3) to determine whether the spatially rarified occurrence data assumed the random distribution after applying spatial filtering (version 10.3). Each camera was installed for 50 days. Thus, there were a total of 2,900 trap nights. Trap sites were placed throughout the study area.

Visual encounters or sightings are rare in the study area due to many species’ elusive and nocturnal nature. In addition, several species found in the study area cannot be studied using standard photographic images for capture-recapture estimation techniques as individuals lack unique identifiable features. In the monsoon season, bridges often get washed away, and frequent landslides impede field research. Droppings, scats, and pellets are often degraded due to incessant rain in the monsoon. Thus, there is limited scope for conventional research methods in this landscape to monitor the abundance of mammals.

During this research, we could not inspect some camera units during the winter and pre-monsoon periods, as the access routes were inaccessible after heavy snowfall and rainfall. Extensive fieldwork could not be conducted in the northern part of Mehao WLS as the area remained inaccessible.

This study documented several rare and elusive species in Mehao WLS ( Supplementary Figure 2 ). During the camera trap survey and direct sightings, 27 species of mammals were recorded. However, our analysis of six distinct mammalian species revealed novel information with regard to the conservation value of the study area. We documented that Mehao WLS supports rare and threatened species such as clouded leopard, spotted linsang, marbled cat, hoolock gibbon, sambar, mainland serow, etc. The analysis revealed that mammalian fauna in this area responded to ecological and anthropogenic factors in various ways, but the responses were species-specific, as predicted. Species responded differently to different covariates, implying that management actions must be prioritized based on the relative importance of species in terms of conservation needs. Habitat and anthropogenic disturbance include habitat restoration and protection, invasive species control, and biodiversity conservation. For anthropogenic disturbance, mitigation efforts may focus on reducing human-wildlife conflicts, sustainable land use planning, wildlife-friendly infrastructure design, and regulatory measures to limit harmful activities. These management actions can help mitigate the negative impacts of habitat loss and anthropogenic disturbance on wildlife species and promote ecosystem conservation and sustainable management. We did not conduct sampling outside the sanctuary, but the species in this study may also be present in agricultural and human-occupied landscapes. --

Including species in ecological models with covariates relies significantly on their detection rates in camera trap data, reflecting their presence and behavior within a studied habitat. In our analysis, the Asiatic golden cat, leopard cat, Himalayan black bear, yellow-throated marten, masked palm civet, Mishmi takin, and wild boar were regrettably omitted due to their notably lower capture rates. While these species undoubtedly contribute to the region’s biodiversity, their infrequent appearance in camera trap records posed a challenge in establishing reliable relationships with the covariates under investigation. Focusing on species with higher capture rates ensures robustness in our model, allowing for a more comprehensive understanding of the ecological dynamics between covariates and the more frequently observed wildlife within this ecosystem.

Given our predictions, the detection probability of northern red muntjac showed positive significance to the presence of mithun as compared to mainland serow, as the former is more of generalist species (Paudel & Kindlmann, 2012). However, this study has observed that northern red muntjac are close to human settlement areas with gentle slopes (Paudel and Kindlmann, 2012), and mithun is a semi-domesticated bovid species that prefers to live near human settlements. Our occupancy survey revealed that northern red muntjac primarily choose dense forests with low-growing vegetation, such as bamboo-mixed forest for foraging and rest (Teng et al., 2004; Odden and Wegge, 2007).

Mainland serow showed a significant association with water bodies. Due to the absence of dense understory vegetation on steep slopes, the serow is primarily found in the mountain’s deeply dissected rocky gorges and ravines, which also provide shelter (Wu and Zhang, 2004). There are steep hills and deep valleys with multiple rocky streams and some perennial rivers surrounded by dense forests and patches of grass. This provides a suitable habitat for serow, which feeds on vegetation growing near water bodies and provides cover from predators such as black bears and humans. Serows also use dense forests near water habitats to maintain their body temperature (Aryal, 2009).

The Asiatic brush-tailed porcupine is a lesser-known rodent species in northeast India (Talukdar et al., 2019; Molur, 2020). This is the first record of this species from Mehao WLS. They prefer the plain area with a gentle slope in the dense forest of Mehao WLS. They have moderate temperature tolerance, and the dense forest near waterbodies provides cover and moisture to maintain their body temperature. The species was recorded from other forests in Arunachal Pradesh, including Namdapha NP, Pakke WLS, and Changlang district (Agarwal, 2000; Datta et al., 2008a, Datta et al., 2008b).

However, the Asiatic golden cat is one the least studied species in tropical Asia (Yongdrup et al., 2019), and it is thought to be rare due to limited sightings (Grassman et al., 2005). It is a solitary hunter and forest-dependent species whose range is threatened by significant habitat loss and fragmentation (McCarthy et al., 2015). Over time, land use has changed forest cover, and human pressure on natural resources has increased at lower elevations. The species has also been recorded in the high-altitude ranges from India to Bhutan (Bashir et al., 2011; Jigme, 2011; Dhendup, 2016). During the survey, we captured four different morphs of the Asiatic golden cat in Mehao WLS.

We found that yellow-throated marten prefers elevated areas. Forest habitat quality is correlated with elevation, especially since there is relatively less anthropogenic disturbance at higher altitudes, as urban areas, agricultural fields, and orchards are generally found at lower altitudes (Lee et al., 2021). The presence of yellow-throated martens decreased in areas with rugged terrain because they tend to avoid shrub-covered regions, given their preference for an arboreal lifestyle (Duckworth, 1997; Appel and Khatiwada, 2014). The masked palm civet prefers forest habitats in lowland areas with ample amounts of water, where they feed on berries, seeds, molluscs, small mammals, amphibians, insects, and arthropods (Zhou et al., 2008; Matsuo and Ochiai, 2009; Belden et al., 2014). Though there is anthropogenic pressure, this species still survives in lowland areas as it has a degree of tolerance to habitat alteration (Zhou et al., 2008; Semiadi et al., 2016).

The Idu Mishmi community has intricate animistic beliefs linked to the forests and the landscape, which ensures prudent resource management. The practice of social taboos by local people can be harnessed by conservation projects so that local people appreciate and willingly engage in community-based conservation projects. Current scientific literature reveals that habitat fragmentation and loss of habitat due to anthropogenic pressure (like expansion of agricultural land, built-up area, and construction of highways) leads to local threats to these species. Such anthropogenic pressures may also affect species distribution and abundance in the landscape (Aiyadurai et al., 2010).

Mehao WLS is an important landscape for mammalian species due to its diverse topography, vegetation, and climatic conditions. Our study provides quantitative insights for multi-species conservation decision-making as it accounts for significant ecological variables. The modeling of multi-species occupancy provided data to predict mammalian responses to land-use changes. These insights are important for decision-making by forest officials. This would benefit various mammalian species and help formulate an informed land-use policy.

This study also provides crucial ecological data for various mammalian species. It exemplifies a simple method to estimate the distribution and habitat occupancy of rare and elusive mammals in Mehao WLS. Due to the varied nature of species targeted in our study, we recognize that the data available for analysis may differ significantly. This disparity can introduce limitations to the comprehensiveness of our study, as it may not provide sufficient insights into the behaviours and patterns of certain species. To overcome this limitation, it is essential to conduct additional sampling efforts across multiple seasons. This approach is necessary to ensure a more comprehensive and robust analysis, particularly when surveying rare and elusive species. However, microhabitats are essential for managing and conserving mammalian species. Such data is crucial for designing and implementing species-specific conservation initiatives and effectively managing protected areas. To date, there has been little supporting information on mammalian species in Mehao WLS (Ahmad et al., 2023), and it is important to carry out a long-term study on mammals in the landscape (Katti et al., 1990).

The effectiveness of patrolling along with other management and conservation interventions will be effective with continued monitoring that can reveal the presence of various species, identify threats, and monitor the prevalence of invasive species. The cameras deployed in this study remained operational throughout the day, and the data indicates that the arboreal nature of certain species may limit the number of photographic records. Despite this possible limitation, it is important to initiate conservation efforts for various species according to available data, which indicates that the species occur in small numbers and require greater protection. Data analysis of mammalian species throughout their range is critical for effectively and efficiently conserving these species in a continuously changing landscape.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The animal study was approved by Wildlife Institute of India, Dehradun. The study was conducted in accordance with the local legislation and institutional requirements.

AA carried out field work, analyzed the data, and led the writing of the manuscript. GG conceived ideas and designed methodology. Both authors contributed to the article and approved the submitted version.