Mortality records (1979–2011) shed light on threats to Asian Elephants Elephas maximus Linnaeus, 1758 (Mammalia: Proboscidea: Elephantidae) in Nilgiris, southern India

1,2,3 Department of Ecology and Environmental Sciences, Pondicherry University, Kalapet, Pondicherry 605014, India 2 2550 rue Saint Exupery, 38420 Le Versoud, France 4 Sigur Nature Trust, Chadapatti, Mavinhalla P. O. Nilgiris, Tamil Nadu 643223, India 5 Additional Principal Chief Conservator of Forests, Tamil Nadu Forest Academy, Coimbatore, Tamil Nadu 641003, India 6 Wildlife Protection Society of India (WPSI), S-25 Panchsheel Park, New Delhi 110017, India 1 pdavidar@gmail.com (corresponding author), 2 rossetclement@yahoo.com, 3 pratheeshmammen@gmail.com, 4 jp.puyravaud@gmail.com, 5 srivastavaraj3@yahoo.com, 6 belinda@wpsi-india.org


INTRODUCTION
The Asian Elephant Elephas maximus is under threat of extinction from habitat loss and poaching (Lomolino & Channell 1995;Choudhury et al. 2008) and now survives in fragmented habitats which decrease its long-term viability (Leimgruber et al. 2003). India is estimated to host approximately 60% of all Asian Elephants (Riddle 2010). The largest population, estimated to be around 6000 elephants, is in the Nilgiris-Eastern Ghats (NEG) region of southern India, which is around 15000km 2 in area (Baskaran 2013).
Ivory poaching, habitat loss and fragmentation are major threats to elephant populations in India (Leimgruber et al. 2003;Choudhury et al. 2008), and while ivory poaching has recently declined, its impact on sex ratios still lingers (Baskaran 2013). Armbruster et al. (1999) showed that there is a lag period before a slowly-declining species such as the Asian Elephant can be driven to extinction. Therefore, it is crucial to identify and ameliorate threats before they become serious. In this study we examined mortality records of 291 Asian elephants from the Mudumalai Tiger Reserve (MTR), the Nilgiri North Division (NND) and Nilgiri South Division (NSD), which are contiguous forests and form part of the Nilgiris-Eastern Ghats elephant range in southern India. The database covered a period of 33 years , which included a period of intensive ivory poaching in the 80s, 90s and early 2000s (Daniel et al. 1987;Prasad 2000), probably causing about 45-68 % of male elephant (tusker) deaths in Tamil Nadu (Sukumar 1989).
Illegal killings for the ivory trade are a source of elephant mortality, and the MIKE program (Monitoring the Illegal Killing of Elephants) was established to assess trends in illegal killings in range states to help CITES (Convention on International Trade in Endangered Species, (http://www.cites.org) policy on ivory trade (Burn et al. 2011). Moreover this data being from forests contiguous to two MIKE sites (Burn et al. 2011) could augment understanding threats to this population.
We looked at: (1) patterns and causes of elephant mortality over time, by age class, gender and across areas with different levels of protection; and (2) humancaused deaths over time and in reserves with differing levels of protection. We tested the null hypothesis that causes of death would not differ over time, with age, gender, and levels of protection.
MTR is located along the west of the Nilgiris District, and is contiguous with the Reserved Forests (RF) of NND towards the east and the northern slopes of the upper Nilgiri Plateau and has an average elevation of ~9 50m. The higher altitude RFs (> 2000m) towards the west and south such as Upper Bhavani, Korakundah, Kundah, and Naduvattam fall under the NSD and have a matrix of montane evergreen forests, grasslands and plantations (Fig. 1).
There were 183 records from TNFD, 82 from WPSI and nine from NWEA. Seventeen records were common to TNFD and WPSI.
The records included information on the date when the animal was discovered, its sex, estimated age and age category (adult, sub adult, juvenile or calf), the range/beat/division in which it was found, and possible causes of death. For some sites verification was done using range office records.
The causes of death were described in official records as 'one female elephant of 32 years died at this place suffered injuries. Post-mortem done. Death due to internal haemorrhage. Bullet found' etc. In 248 instances, the records noted the case as 'wildlife offense' and listed as 'poaching', 'disease', 'electrocution', 'hit by vehicle', 'others', 'unknown' and minor categories such as 'fell off rock', 'injured'. From the wildlife offense listed and the reason for the death as per official veterinarian causes (Table 2). Both sexes were equally susceptible to accidents (χ2= 0.13, df=1, ns) and diseases (χ2= 1.68, df=1, ns).

Comparison of PAs with differing levels of protection
The number of records over the decades significantly differed between MTR and NND (χ2=15.69, df=2, p = 0.0004). Deaths increased in MTR over time and peaked in NND between 1990 and 2000 (Fig. 3). Adult deaths did not significantly differ between MTR and NND (χ2=0.31, df=1, ns), although NND supported fewer elephants ( Table 3). Levels of poaching deaths were significantly higher in NND than MTR (χ2 = 7.45, df=1, p = 0.006, Table 3). There were regional differences in the causes of mortality with unknown causes and poaching dominating in MTR, poaching in Gudalur and the Sigur plateau and accidents/poaching in Coonoor/Kallar (χ2=19-49, df=3, p = 0.0002, Table 4).

DISCUSSION
Our study suggests that elephant populations in these reserves are still at risk despite decreases in the levels of poaching and human-caused deaths over the years. Overall, survivorship is low: more so for adult males than females, and the recently assessed sex ratios of adults is still skewed towards females, a legacy of the ivory poaching in the past (Baskaran 2013). The proportion of illegal killings (0.22) is higher than that reported for Asian Elephants (0.15; Burn et al. 2011), which might increase further if the causes for the majority of recent deaths could be established. The adult females in the population despite facing less threat from humans have lower survivorship than female timber elephants in Myanmar (Clubb et al. 2008).
The Nilgiris-Eastern Ghats region holds the largest population of Asian Elephants in the wild, and is critical for the long term survival of the species. This range lies within a matrix of human dominated landscapes where high human densities, expansion of settlements and intensive agriculture are serious threats. Maintaining the viability of this population in the midst of inhospitable landscapes is a challenge for conservation.
The reserved forests of the NND have proportionately higher number of human-caused deaths probably because of higher intensity of anthropogenic pressures (Silori & Mishra 2001;Davidar et al. 2007Davidar et al. , 2010Baskaran et al. 2012) and lower levels of protection. These forests which connect the Western and Eastern Ghats, and provide a passage to higher elevation ecosystems, are important range lands for elephants (Davidar et al. 2012). Radio collared elephants used MTR as well as the reserved forests as part of their range (Desai 1991;Desai & Baskaran 1996), and are seasonal visitors in higher elevation montane shola forests and grasslands of the NSD and the Mukurthi National Park (Davidar et al. 2012). Therefore, these habitats need to be protected stringently and connectivity ensured.
Ivory poaching remains a general and widespread threat to elephants in India (Wildlife Protection Society of India: http://www.wpsi-india.org/projects/elephant_ poaching.php, viewed 21 July 2013) and vigilance is required see that it does not recur in this region. Other threats could be cryptic and difficult to detect, such as poisoning of wildlife which is widespread in India (Richards 2011), and ingestion of plastic waste. In the recent decade 24% of elephant deaths were attributed to intestinal compaction. Plastics are ubiquitous in the Nilgiri District despite a ban, and are widely dispersed by tourists and tourism operators (What ails the Nilgiris, http://www.keralatourismwatch.org/node/130, viewed 23 July 2013). Plastics cause digestive impaction in livestock (Remi-Adewunmi et al. 2004), and wildlife (Beck & Barros 1991;Moser & Lee 1992;Jacobsen et al. 2010;Kumar & Dhar 2013).
The limitations of the study were that the analyses were confined to recorded deaths whereas many deaths particularly those of younger elephants, and deaths in more remote areas, would not have been detected. Similarly, difficulties in ageing elephants could bias survivorship. In conclusion, detection of deaths need to be improved through regular patrols throughout the reserves; database on elephant deaths should be systematised; forensic capabilities upgraded and elephant populations monitored.