Demographic and genetic structure of a severely fragmented population of the endangered hog deer (Axis porcinus) in the Indo-Burma biodiversity hotspot

40.05 km2. The characteristic floating meadows, locally known as phumdis, covers 26.41 km2 and the remaining extent of 14.09 km2 consists of open water, dry lands and hills [17]. The floating meadows of the park are of varying thickness, and they support small populations of the hog deer and Eld’s deer (Rucervus eldii). The change in the hydrology of the park due to the construction of the Ithai barrage in the downstream of Manipur River, in the early 1980s, has resulted in a rapid change in the Lake ecosystem and the natural process of meadow for- mation [17]. The meadows, that used to settle down during the lean season on the Lake bed and get replenished with soil and nourishment, are now continuously floating due to perma- nent flooding of the Park. Consequently, the thickness of the phumdis is decreasing at a rapid rate, which is detrimental, as the thickness has to be sufficient to support the weight of the wild ungulates such as hog deer and Eld’s deer, which are found sympatrically in the park. The park experiences three distinct seasons, summer, winter and monsoon [18]. The float- ing meadows supports around 185 species of herbs, grasses and sedges of which Phragmites karka, Leersia hexandra, Capillipedium spp., Zizania latifolia and Hemarthria compressa con- tribute to the main food plants of the ungulates in the park [6, 19]. Field methods The entire park was divided into 1 km × 1 km grids to examine the distribution pattern and habitat use of hog deer. In each grid, based on the phumdi thickness and the presence of open

water, plots of 50 m x 2 m were laid on transects of length 500 m (n = 66 in 2006; n = 75 in 2007; n = 72 in 2008), were laid and hog deer faecal pellet groups within each transect were recorded. Habitat parameters were recorded in circular plots of radius 2 m along each transect at 50 m intervals. The habitat parameters collected include ground cover (vegetation, litter, water and phumdi), substratum character (water/phumdi/soil), phumdi thickness, water depth, total number of faecal pellet groups, signs of associated species, ocular visibility, distance from uplands/hills, distance to water for drinking/wallowing and disturbance factors (fishing, graz- ing and vegetable collection). A 0.5 m × 0.5 m quadrat was laid within each plot to record the number of each plant spe- cies present. On the basis of the distribution of the faecal pellet groups, the different areas of the park were classified as high, medium and low-density areas. Boat surveys were carried out at different study sites in the park to identify specific locations that were suitable for construct- ing machans (temporary bamboo watch towers) for estimating and monitoring the hog deer population. This exercise provided baseline information for constructing machans and also helped classify the habitat into thick and thin phumdis. No machan was constructed in those blocks in which no faecal pellets were recorded. In total, twenty-two machans were con- structed with an average height of 7 meters to facilitate the monitoring activity in the dense vegetation of the park.

The point count method [20, 21] was used to estimate the abundance of hog deer in a 20 km2 area of the park excluding the water bodies and thin phumdis. Eight machans were con- structed in the high pellet density areas and seven machans in the medium- and low-pellet- density areas. The counting of animals was carried out in the month of March from the mach- ans using 8 × 40 prismatic binoculars when the visibility in the park was in its maximum due to the phenological trait of the grasses in the area. It was done between 0530 and 0900 hours and between 1500 and 1730 hours for five consecutive days using a double-observer method developed to estimate detection probabilities and to increase the detection accuracy [22, 23]. During each point count, a designated “primary observer” indicates to the “secondary observer” all the animals detected, who then records all the detections. The observers take the primary and secondary roles alternately during the course of a survey. This approach permits Demographic and genetic structure of hog deer in the Indo-Burma biodiversity hotspot PLOS ONE | https://doi.org/10.1371/journal.pone.0210382 February 6, 2020 3 / 16