Opportunity for Thailand's forgotten tigers: assessment of the Indochinese tiger Panthera tigris corbetti and its prey with camera-trap surveys

Abstract Dramatic population declines threaten the Endangered Indochinese tiger Panthera tigris corbetti with extinction. Thailand now plays a critical role in its conservation, as there are few known breeding populations in other range countries. Thailand's Dong Phayayen-Khao Yai Forest Complex is recognized as an important tiger recovery site, but it remains poorly studied. Here, we present results from the first camera-trap study focused on tigers and implemented across all protected areas in this landscape. Our goal was to assess tiger and prey populations across the five protected areas of this forest complex, reviewing discernible patterns in rates of detection. We conducted camera-trap surveys opportunistically during 2008–2017. We recorded 1,726 detections of tigers in 79,909 camera-trap nights. Among these were at least 16 adults and six cubs/juveniles from four breeding females. Detection rates of both tigers and potential prey species varied considerably between protected areas over the study period. Our findings suggest heterogeneity in tiger distribution across this relatively continuous landscape, potentially influenced by distribution of key prey species. This study indicates that the Dong Phayayen-Khao Yai Forest Complex is one of the few remaining breeding locations of the Indochinese tiger. Despite limitations posed by our study design, our findings have catalysed increased research and conservation interest in this globally important population at a critical time for tiger conservation in South-east Asia.


Introduction
T he tiger Panthera tigris has suffered catastrophic declines in its population (%) and habitat (%) over the past century (Nowell & Jackson, ; Goodrich et al., ; Wolf & Ripple, ). Evidence suggests only  source sites (i.e. sites with breeding populations that have the potential to support future recovery of the tiger over a larger area) remain across the species' range, totalling , km  (.% of current range; Walston et al., ). Habitat loss has been particularly acute in South and South-east Asia, with a % reduction from  to  (Sanderson et al., ) and an estimated forest loss of , km  in priority tiger conservation landscapes from  to  (Joshi et al., ).
The Indochinese tiger Panthera tigris corbetti is one of six extant tiger subspecies and is categorized as Endangered on the IUCN Red List (Lynam & Nowell, ; Goodrich et al., ). It was historically distributed throughout most of mainland South-east Asia (Luo et al., , ) across Cambodia, Lao, Myanmar, southern China, Thailand and Viet Nam (Lynam, ). Evidence suggests three range countries (Cambodia, Lao and Viet Nam) have lost viable populations, and the Indochinese subspecies may qualify for Critically Endangered status (Lynam & Nowell, ). Despite previous evidence of a viable breeding population in Nam Et Phou Loey National Protected Area in Lao (Johnson et al., ; Vongkhamheng, ), recent evidence suggests tigers may have been extirpated from the country (Rasphone et al., ). Tigers are probably extinct in Cambodia, prompting plans for reintroduction (Gray et al., ), and in Viet Nam there have been no confirmed tiger records in .  years (Lynam & Nowell, ). A paucity of reliable population data in current range countries has obscured these declines (Lynam & Nowell, ), and information on remaining populations is needed urgently.
It is possible that the only remaining source sites for the Indochinese tiger are in Myanmar and Thailand. However, studies in key landscapes in Myanmar have documented low and potentially declining numbers (Lynam et al., ; Rao et al., ; Naing et al., ; Moo et al., ) reinforcing the importance of Thailand as the tiger's last stronghold in the region. In Thailand's  action plan, the national tiger population was estimated to be - individuals (Pisdamkam et al., ). A recent, updated government report included landscape-specific population estimates of at least - individuals (DNP, ), with potentially only two viable populations, in the Western Forest Complex (, km  ) and the Dong Phayayen-Khao Yai Forest Complex (, km  ) in eastern Thailand.
Although a number of tiger-focused studies have been conducted in other parts of Thailand, including ongoing monitoring in the Western Forest Complex (Duangchantrasiri et al., ), data from the Dong Phayayen-Khao Yai Forest Complex are limited. Information on tigers there has originated primarily from general assessments of faunal communities or other carnivores, or from interviews and personal communications (Lynam, ; Kanwatanakid et al., ; Lynam et al., ; Jenks et al., ). Evidence suggests that tigers may have been extirpated in Khao Yai National Park, but almost no information is available from other areas in this forest complex. To our knowledge, there have been no studies focusing on tigers across this forest complex in its entirety. Comprehensive studies on prey species, an important factor for tiger distribution and persistence (Karanth & Stith, ; Karanth et al., ), are also lacking.
Given catastrophic population and range declines elsewhere in Thailand and South-east Asia, knowledge of the tiger population of the Dong Phayayen-Khao Yai Forest Complex is of national, regional and global importance. Here, we describe results from the first camera-trap study focused on tigers and implemented across all protected areas in this landscape, conducted during -. We aimed to assess tiger and prey populations and to identify any patterns in detection frequencies of tigers and prey species amongst protected areas. Our findings provide baseline information for tigers and their prey, and also document potentially important information on other mammal species of research and conservation interest.

Study area
The Dong Phayayen-Khao Yai Forest Complex lies c.  km north-east of Bangkok (Fig. ). To the east it partially borders the international boundary between Thailand and north-west Cambodia. The terrain is hilly, with altitudes The complex contains all major forest types characteristic of eastern Thailand, but is primarily covered by mixed evergreen and mixed dipterocarp/deciduous primary and secondary forest. It also contains grassland/scrub areas, some of which are anthropogenic. These forests have been influenced to varying degrees by a complex history of human presence and exploitation, including logging, settlements, agriculture and other activities (Lynam et al., ). Currently, the complex is surrounded almost completely by a human-dominated matrix of villages, farmland and infrastructure.

Methods
We conducted camera-trap surveys during March -February . The study design was opportunistic because of limited resources, and data collection for the five protected areas varied in spatial and temporal extent (Supplementary Fig. , Supplementary Material ), precluding analysis within an occupancy framework. We placed camera traps in locations suitable for tigers, to maximize detections. Such locations included geographical or topographic features (e.g. ridges, river valleys) and access roads or trails likely to be used regularly by tigers (Karanth, ; Karanth & Nichols, ). We also used tiger track and sign (e.g. pugmarks, scats), and presence of prey species, to identify prospective camera locations.
We considered consecutive detections of a species at one camera station to be independent if they occurred after .  minutes (O'Brien et al., ). Individual tigers were given an alphanumeric identifier to compile detection histories. Tigers not conclusively identified were marked as unknown. We calculated detection rates of tigers and prey as number of detections per  camera-trap nights, with cumulative rates reported for each protected area across survey years. Although such indices do not reliably indicate abundance (Jennelle et al., ; Sollmann et al., ), we also carried out a comparative analysis of photographic capture rates for tigers and prey for all five protected areas (Supplementary Material ).

Results
Camera traps were active for a total of , camera-trap nights at  locations. Survey effort varied significantly across protected areas. Thap Lan National Park (, camera-trap nights) and Pang Sida National Park (, camera-trap nights) accounted for c. .% (,) of total camera-trap nights and c. .% (n = ) of stations. Survey effort by protected area and year is summarized in Table .
Surveys recorded , independent detections of tigers during the study period (Table ). Tigers were documented in three of the five protected areas (Thap Lan National Park, Pang Sida National Park and Dong Yai Wildlife Sanctuary), with Thap Lan National Park and Pang Sida National Park accounting for . % of detections (, and  detections, respectively). Tigers were detected in Dong Yai Wildlife Sanctuary only in  (seven detections). Tigers were not detected in Khao Yai National Park and Ta Phraya National Park. Detection rates in Thap Lan National Park were higher than in Pang Sida National Park with cumulative means of . (range .-.) and . (range .-.) detections per  camera-trap nights, respectively. In total, at least  adults were documented: seven females, seven males and - partially identified adults whose sex could not be confirmed ( Table ) Surveys documented successful breeding in  and , with six cubs/juveniles from four adult females. One litter of two juveniles were photographed without their mother (who could thus not be identified). One cub (C), first documented in , appeared to be independent from its mother by .
We documented six potential prey species: gaur Bos gaurus, banteng Bos javanicus, Chinese serow Capricornis milneedwardsii, northern red muntjac Muntiacus vaginalis, sambar Rusa unicolor and wild boar Sus scrofa. We considered these species potential tiger prey based on information from Thailand and elsewhere within the tiger's range (Karanth et al., ; Sunquist, ; Steinmetz et al., ). All but one potential prey species (banteng) were documented in all five protected areas.
Mean cumulative detection rates of sambar (Supplementary Table , Supplementary Material ) were considerably higher in Thap Lan National Park (. detections per  camera-trap nights) than in other protected areas (.-.), whereas detection rates of other prey species were comparatively lower in this Park. Mean cumulative detection rates for wild boar were highest in Dong Yai Wildlife Sanctuary (. detections per  camera-trap nights) and Pang Sida National Park (.). Sambar and wild boar were generally detected more frequently than other prey species.
Although tigers were the primary focus of surveys, we also documented a number of other species (Supplementary Table , Supplementary Material ), with  detections of other felids, including the Asiatic golden cat Catopuma temminckii ( detections), mainland clouded leopard Neofelis nebulosa (), marbled cat Pardofelis marmorata () and leopard cat Prionailurus bengalensis (). We did not detect leopards Panthera pardus. We documented  mammal species in total, including one Critically Endangered, five Endangered,  Vulnerable, three Near Threatened and  categorized as Least Concern.

Discussion
This study provides insights into tigers and their prey in the understudied Dong Phayayen-Khao Yai Forest Complex. Most of our detections of tigers were in Thap Lan and Pang Sida National Parks, potentially a result of larger survey effort (, and , camera-trap nights, respectively, of a total of ,). This was the result of our opportunistic study design, which prioritized survey areas based on potential or confirmed tiger presence. Nonetheless, the absence of detections of tigers or their sign from two of the five protected areas, despite reasonable survey effort, suggests higher tiger abundance in these two Parks than elsewhere in this forest complex. Tiger presence across the complex appears to be heterogeneous, but to an unknown degree. Our records from Dong Yai Wildlife Sanctuary are from an area just outside the formerly known extant range of P. tigris (Goodrich et al., ). The lack of tiger detections from Khao Yai National Park is consistent with speculation that tigers have been extirpated from this protected area (Lynam et al., ; Jenks et al., ), although our survey effort and coverage in this Park was relatively low (, camera-trap nights).
Although the number of tigers we documented in the complex is not a population estimate, our results suggest the population may be larger than previously assumed (Lynam, ), and also document the long-term persistence of a number of individuals in this area ( Supplementary  Fig. , Supplementary Material ). To our knowledge, the photographs of tiger cubs we obtained are the first confirmed records of successful breeding in the forest complex since at least  (Lynam et al., , ; Jenks et al., ) and confirm that the site supports a breeding population. Breeding and subsequent dispersal could potentially result in expansion into Khao Yai National Park, and contribute to overall population recovery.
The presence of prey is important for tiger distribution, density and persistence (Karanth & Stith, ; Karanth et al., ), as noted by studies elsewhere in Thailand  However, a dedicated prey study is required to determine the extent to which tigers in these parks rely on these species. Low prey detection rates in Ta Phraya National Park and Dong Yai Wildlife Sanctuary could explain the absence of tiger detections in these two areas.
We did not detect leopards, which, given that they have similar behavioural patterns to tigers and can tolerate some degree of spatial overlap (Karanth & Sunquist, ; Andheria et al., ), suggests they may be absent from the forest complex. The Indochinese leopard Panthera pardus delacouri has not been detected recently in other parts of South-east Asia, suggesting a decline in its population and range (Rostro-García et al., ). Abundance and diversity of suitable prey are important for the co-existence of tigers and leopards (Karanth & Sunquist, ; Andheria et al., ). Historical overhunting of prey in the forest complex could have driven competitive exclusion of leopards by tigers or other carnivores (Harihar et al., ; Volmer et al., ). Direct hunting by humans may have also driven population declines. However, given the paucity of reliable historical data, the reasons for the absence of the leopard in the Dong Phayeyen-Khao Yai Forest Complex remain unconfirmed.
Our data could not be used to estimate tiger occupancy or population size because the study design would violate key assumptions of the appropriate methods (Harmsen et al., ; Welsh et al., ). Methodologically rigorous study designs should be employed wherever possible in monitoring wildlife populations, but if resources are constrained an opportunistic study design may be appropriate (Harihar et al., ; Stein et al., ; Johnson et al., ). Although conclusions that can be drawn from such studies are limited, they can contribute important insights into species presence in poorly studied areas (Stein et al., ; Jenks et al., ).
At the start of this study, tigers were believed to have disappeared from Khao Yai National Park (Lynam, ; Lynam et al., ; Jenks et al., ), information was lacking for other areas and resources were limited. In these circumstances, an opportunistic study design was suitable to address our fundamental research question, specifically, to confirm tiger presence. Early findings suggested tigers were present in the area, which enabled us to secure further funding and improved access to resources such as camera traps that were later used for tiger density and population estimates. Additional funding also enabled investments in law enforcement, patrol-based monitoring and community outreach programmes. To build on this work, we recommend additional analyses to model relationships between tigers, prey, threats and habitat required for spatial prioritization of protection and recovery interventions.
Our study provides insight into what is probably one of the most important extant tiger populations remaining in mainland South-east Asia. A comprehensive investigation of the tiger in other understudied sites in the region is urgently needed to generate a more accurate picture of their status. To recover and double the population of wild tigers (Global Tiger Initiative, ; Harihar et al., ), additional resources will need to be allocated to implement robust monitoring in sites where tigers remain.
To our knowledge, our work is the first to assess the tiger population across the Dong Phayayen-Khao Yai Forest Complex and suggests this region is important for the Indochinese tiger, which has lost most of its range in South-east Asia. Our findings establish this forest complex as home to one of the few remaining breeding populations of Indochinese tigers, demonstrate the long-term persistence of some individuals, and suggest heterogeneous tiger presence across the five protected areas, potentially influenced by distribution of prey species. Our initial results have catalysed increased research and conservation investment in this landscape at a critical time for tiger conservation in South-east Asia.