Animal-vehicle collision on the roads is a major cause of mortality of a wide range of animal taxa both within and around protected areas. This study has been conducted in the National Highway 715 (new) covering a continuous stretch of 64 km that passes through Kaziranga National Park (KNP) of Assam (India). The area falls between the boundary of KNP on its north and North Karbi Anglong Wildlife sanctuary on the south. The survey concentrated on the mortality study of four groups of vertebrates viz., amphibians, reptiles, birds, and mammals resulting from collisions with vehicles from October 2016 through September, 2017. A total of 6036 individual roadkills were registered, belonging to 53 species, 23 other taxa and 30 families of vertebrates, with herpetofauna being the most affected group followed by birds and mammals. The study evaluated seasonal variation in the overall roadkill pattern with highest mortality in the monsoon season 38.27% (n = 2310) and with peak casualties starting with the onset of rainfall (February and March) and during monsoons (July and August). The amphibian mortality was also found to be highest during the monsoon with 43.28% (n = 1575) of kills, as compared to the other three groups. NH-715 (new), therefore serves as a challenging passage for the animals, forming a major barrier for the faunal component of the Kaziranga-Karbi Anglong landscape. This study thus tried to reflect the often overlooked issue of roads and highways in terms of direct mortality of animals due to traffic and thereby can be helpful in understanding the seriousness of the situation and identifying prospective measures to be taken for sustainable coexistence of both animals and human.

Herpetofauna, Kaziranga National Park, mortality, National Highway, road ecology, vehicular collision

India has the second largest road system in the world according to the National Highway Authority of India (NHAI) covering over 5.89 million kilometres with all signs pointing to an explosion of expansion in the upcoming years (Indian Road Industry Report 2020). The state of Assam has a total road length of 47,936 km including 3908.5 km of National Highways, 3134.36 km of State Highways, 413.03 km of major district roads and 37 030 km of rural roads (India Brand Equity Foundation 2019). In India, the total length of roads and the number of automobiles has increased tremendously. The total extent of roads in India has increased more than 11 times during the past six decades from 1951 to 2015 (from 3.99–46.71 lakh kilometre), which was 4.2% in 1950’s (ROADS-Statistical Year Book 2018). The number of registered motor vehicles has also been growing at a compound annual growth rate (CAGR) of 10.16% per year over the last five years. Considering this vast network of roads, it is crucial to assess the magnitude of wildlife mortality due to road traffic (Glista et al. 2008).

Roadways are obviously integral to commuting and transportation, but are certainly known to pose some detrimental effects on the flora and fauna surrounding them (Mazumdar and Gogoi 2010). They may affect animal populations in different ways. The most prominent effect of such linear structure is mortality through vehicular collisions (Das et al. 2007; Grilo et al. 2018; Jeganathan et al. 2018). Other more complex consequences identified are habitat modification and fragmentation (Carr and Fahrig 2001), leading to population isolation, changes in animal distribution and movement patterns (Desai and Baskaran 1998), increased inbreeding, decrease in population size and high possibilities of local extinction (Quinn and Hasting 1987). Another corollary effect of road is the volume of space they occupy (Trombulak and Frissell 2000), contamination of roadside habitats due to automobile exhaustions (Beeby and Richmond 1987) and increased avoidance behaviour, thus acting as a barrier to gene flow (Mader 1984). The effect of such infrastructures is felt by mammals (Baskaran and Boominathan 2010), birds (Robertson 1930), reptiles (Das et al. 2007), amphibians (Seshadri et al. 2009) and macro invertebrate fauna (Haskell 2000) as well. Thus, road fragments the habitats, and with the growing demand for more networks, animals are increasingly forced to cross roads to perform their routine necessities and are often killed by vehicles (Hourdequin 2000).

Among the various threats posed to wildlife, collisions with vehicles are becoming a major concern for many species (Bager and Rosa 2011). Road attributes like width, length and condition of the pavement, directly amplify the rate of vehicular collisions. Setting up roads and measures without proper structural designs, leads to an increased rate of vehicular collisions for animals (Oxley et al. 1974). Road density and their constructional works can harm and alter biodiversity at local, regional and landscape scales, the effect of which can sometimes remain unnoticed for decades (Findlay and Houlahan 1997). The International Union for Conservation of Nature (IUCN) also considers roads and railways in their list of threats for various wild animals. Mortality due to vehicular collision for any sustaining population may not exhibit the need of immediate conservation concerns, but it is presumed that small, isolated, declining, threatened populations and species are also affected by road mortality (Mumme et al. 2000). Species-specific behaviour in response to the road environment also guides the risk of vehicular collisions (Erritzoe et al. 2003). Many species use the road for daily activities like foraging, nesting, predation, scavenging, shelter, which can increase their vulnerability to road mortality (Fulton et al. 2008). Road habitats may also act as a habitat sink or ecological trap for birds and small mammals (Mumme et al. 2000).

Factors like roadkill rates, traffic traits and landscape attributes are effective in determining the spatial location of roadkills (Ramp et al 2005), but their temporal distribution for predicting seasonal patterns have been addressed less often. Explanations for seasonal variation in roadkill have been correlated to the breeding and foraging behaviour of the species (Erritzoe et al. 2003). The temporal roadkill patterns of small mammals, birds, and lizards are also linked to their phenology (D’Amico et al. 2015). In Amphibia, seasonal peaks in population sizes (Rosa and Bager 2012) and migration (Langen et al. 2009) have also been related to temporal roadkill patterns. Environmental variations and seasonal life-history traits (D’Amico et al. 2015; McCardle and Fontenot 2016) can also cause temporal or seasonal roadkill peaks.

Our goal in this study was to evaluate the magnitude of mortality due to collisions with vehicles, of vertebrate fauna, specifically amphibians, reptiles, birds, and mammals in the highway stretch that passes through Kaziranga-Karbi Angling landscape complex in North-Eastern India, from October 2016 to September 2017. We also tried to evaluate the seasonal variation and pattern of roadkill distribution among the vertebrate groups in the study area. Similar studies have been carried out in India, but no detailed study for this high diversity hotspot is known yet (Islam and Saikia 2014).

The higher incidence of roadkill mortality in the study area is evidence of heavy loss of vertebrate wildlife species in the National highway passing through an important protected area, i.e. KNP of Assam. Similar types of study conducted across the globe also reached comparable findings (Glista et al. 2008; Baskaran and Boominathan 2010; D’Amico et al. 2015). The percentage rank accumulation pattern of roadkill here shows stabilisation with the number of survey days for both fresh and old category of roadkill, thus depicting appropriate survey effort for the concerned study period. However, it is found that only 70 higher roadkill days accounted for 80% of the roadkill statistics in the NH 715 (new) that passes through KNP. This, in turn, signifies the low average number of casualties in most of the survey days (> 70 days), thereby marking the problem to be manageable, if high risk days are identified and proper measures enforced for controlling animal mortality.

The present study accounts for the highest number of amphibian roadkills, (60.29%) and thus concurs with the view of Sundar (2004) that amphibians are more vulnerable to collisions with vehicles because they cross roads slowly and are not easily noticeable. Hence drivers tend to disregard them, and this is exacerbated by their activity pattern and population structure (Hels and Buchwald 2001). However, due to lack of existing data on the abundance and ecological structure of amphibia inhabiting the road habitats, in earlier studies in North East India, it hinders us from making any comparative assessment of our work. Hence, we, for the first time are providing quantitative data on roadkill mortalities of amphibia from this region. Also, our findings are in conjunction with that of Vijayakumar et al. (2001) from Anamalai hills, that Duttaphrynus melanostictus are dominant among amphibian roadkills. This could be attributed to their high abundance in this area, since they are cosmopolitan in distribution (Dutta 1997) and are known to be abundant in disturbed habitats (Inger et al. 1984). Amphibians gather near street lampposts and vehicle head lights to feast on insects and also show high human commensalism (Daniels 2005), hence could be more prone to roadkills. This study recorded a high number of amphibian mortality during the wet season, marking a sharp increase with the onset of monsoon (February – March) and during monsoon (July – August), but ending with a tremendous decline in the dry season. This bimodal pattern of roadkill observed corresponds to the seasonality of the climate of this region, since this period corresponds to the rainiest season in the study area with rainfall peaks during these months. This could be linked to their breeding pattern and behaviour, as it drives them to aggregate near water bodies (Hels and Buchwald 2001) and thus augments with the findings of Smith and Dodd (2003), in their roadkill study in Florida, where most amphibian roadkills were related to water levels. Therefore, if traffic intensity continues to increase, the increasing roadkill rates may eventually reduce the population to a level where its reproductive output will be too small to reach the carrying capacities of the breeding pools, which in turn may drive the population to a level where demographic stochastic processes become important for the survival of the population (Hels and Buchwald 2001).

There are plenty of reptile roadkill studies from other parts of the country (Seshadri et al. 2009; Baskaran and Boominathan 2010) but focusing on this part of Northeast India only Das et al. (2007) has reflected the gravity of this problem and provided ample evidence of roadkills. Their study recorded 68 individuals of reptiles of which 89.7% were snakes followed by lizards 10.2%, and Boiga gokool was the most highly encountered reptile. Whereas, our study revealed a total of 1281 individual reptiles of which 79.70% were snakes and 20.29% were lizards, with Calotes versicolor revealed to be the most affected among lizards and Amphiesma stolatum among snakes, hence marking an ultimate increase in the number of lizard kills. There are several factors responsible for such a disparity between the studies. For example, changes in traffic velocity, since the study was conducted more than a decade ago, and differences in sampling period and effort. Being, poikilothermic in nature (Porter 1972), reptiles have been consistently reported to be severely affected by road traffic (Trombulak and Frissell 2000) and the present observations on roadkill from NH-715, supports this. Dodd et al. (1989) and Das et al. (2007) postulated that reptiles are more tempted by roads, in order to maintain their body temperature overnight, as the road surface remain warmer than the nearby areas. Also, the movement of reptiles, particularly snakes, is impeded on road surfaces, and hence increases their risk of mortality (Roe et al. 2006). The current findings of very high numbers of lizard (Calotes versicolor) compared to snakes, could be attributed to the higher number of canopy gaps between both sides of the road, thus enhancing their activity near the road. In addition, most lizard kills were observed in locations with high canopy gaps. However, this finding remains in disparity with the findings of Dodd et al. (1989) and Das et al. (2007). Bernardino and Dalrymphe (1992) found a substantial increase in the number of snake mortalities during the dry season (37% of total kills) as compared to mortalities in the wet season but our study reflects a more or less equal distribution of snake mortality throughout the year. This could be because of the high breeding activity of most of the snakes during the summer and the rainy season (Chittaragi and Hosetti 2014), and mortality during the dry season could have occurred because of increased vehicular influx of visitors to the park, since increase in traffic volume in conjunction with movement rates, makes the animals more prone to roadkill (Chittaragi and Hosetti 2014). Also, unlike mammals, water does not act as a limiting factor for reptile movement during the floods.

Birds are attracted to roads as a location of resource availability, notably food. (Rytwinski and Fahrig 2012). The road attracts predator populations towards particular small mammals and carrion, insects and worms washed out on to roads, and snakes that are attracted to the heat (Erritzoe et al. 2003). Other resources found near or on the roads are grit and salt (Erritzoe et al. 2003), puddles that serve as a water source (Hodson 1962), telephone and power lines that serve as perches (Robertson 1930) and road fencing that offer breeding sites and shelter (Mead 1997). Hence an enormous number of birds fall victims to vehicular collision, while they concentrate on these resources available along roads. A study in Mudumalai Tiger reserve by Bhaskaran and Boominathan (2010) found that birds were least affected by traffic and comprised 7% of the total kills. Also, it was augmented that birds are less susceptible to vehicular hits because of their ability to fly away quickly. However, the present study recorded 9.87% (n = 596) of avian mortality and found them to be highly susceptible to roadkill due to combination of a variety of factors. Birds thus, mostly succumb to ecological traps, since they descend on the road to feed on carcasses of other road killed animals, small insects, grains spread on the road by fringe farmers, grits and small sand particles, specifically by some species (S. S. personal observation) and to prey on amphibians and reptiles available on the paved road. It was also observed that nocturnal birds like owlets were vulnerable to roadkill, colliding frequently with fast-moving vehicles in this area of study because of their low-level flight and predatory behaviour (Boves 2007). Here, at certain areas where there has been higher bird mortality, both the sides of the highway are at a much lower elevation, and birds, when flying from the lowlands, get hit by vehicles while crossing the highway. The fact that birds were more affected during the wet season than the dry season could be because of the increase in herpetofaunal movement on the road, depicting the high availability of prey and also the increase in carcass availability.

This study recorded the mortality of an endangered carnivore, one Indian Leopard Panthera pardus and three Small Indian Civet Viverricula indica. These figures seem to be very small compared to other taxa, but such loss is intolerable considering their low population density. Several similar studies across India have reported a high roadkill of large cats (Gruisen 1998a; Bhaskaran and Boominathan 2010). Various studies in mammal roadkill in India conducted in many protected areas have also documented the deaths of many species of conservation concern (Rajvanshi et al. 2001). This study also records the roadkill of a primate Rhesus Macaque macaca mullata. Although the mortality rate is lower, the number is still significant. Rhesus macaques were observed crossing the highway almost daily at certain locations, but mostly nearer to human settlements. Although they are very intelligent and highly adaptive primates, they still fall victim to roadkill. Most of the mammal roadkills recorded in this study are nocturnal (Bandicoot Rat, Indian Leopard, Small Indian Civet, House Mouse, and various unidentified Rat and Bat Sp.) species that could have been killed while crossing the roads, as they get blinded by the vehicle’s headlights (Bhaskaran and Boominathan 2010). Also rat and mouse mortality seemed to be higher in areas near agricultural patches and at the time of rearing of paddy (S. S. personal observation).

Also, a substantial majority of the animal carcasses, 1.92% (n = 121) remained unidentified and categorised as “uncategorised”, since they could only be identified up to order, due to their extreme decomposed state. This, in turn, opens up the scope for further detailed studies regarding more precise identification of roadkills.

In conclusion, it could be said that mortality due to collisions with vehicles has been identified as a major conservation issue, but one that is very challenging to address. Ecologists have been trying to diarize the estimation of road-kills for a long time (Stoner 1925), but their impacts are difficult to quantify and requires systematic studies (Smith and Dodd 2003). However due to a variety of factors like searcher efficiency (Kline and Swann 1998), scavenger bias (Boves 2007) or actual cause of death (Kerlinger and Lein 1988), these figures may remain underestimated. Likewise, the figures can also be over-estimated, considering only carcasses are being studied (Hernandez 1988). Therefore, a more detailed study of the same is vital. . This study is prefatory in nature, and further detailed survey of roadkills in relation to species occupancy, their abundance and behaviour will help in understanding the problem at a broader level. Nevertheless, our work clearly indicates the perilous concerns of this issue, revealing a very high number of annual roadkills.. The major factor contributing to these roadkills is the high speed of the vehicles. Thus, reduction of speed should be managed along with proper mitigation designs for the safe movement and existence of all the animals.

Linear infrastructures are an integral part of our daily system and are a major root of developmental activities. But development must always run in parallel with our naturally functioning ecosystem, since our sustenance depends upon the sustenance of nature. Hence, we need not put a stop to development but rather incorporate proper and eco-friendly designs and innovations in tandem with it, in order to reduce vehicular collisions in intrusions within any protected area. It is thereby, important to quantify the magnitude and the effect of vehicular traffic on faunal groups, which would help conserve them, as various infrastructure projects, including roads and highways, are being planned to cater to the country’s growing needs.

We extend our sincere thanks to the Department of Zoology, Gauhati University, Guwahati, for their help and support. We also acknowledge the Forest Department Kaziranga National Park and PCCF Wildlife & CWLW Assam (WL/FG.31/Pt/2017) for granting us permission to carry out our work. We thank all our field assistants and resource persons, without whose help this work would not have been completed.