Characterisation of breeding habitat of Grizzled Giant Squirrel Ratufa macroura (Mammalia: Sciuridae) in Chinnar Wildlife Sanctuary, Western Ghats, India

The Grizzled Giant Squirrel (GGS) Ratufa macroura (Pennant, 1769) is a ‘Near Threatened’ and endemic giant squirrel distributed in southern India and Sri Lanka. In India, the species is distributed in more than 10 locations between Cauvery Wildlife Sanctuary in Karnataka in the north and Srivilliputhur Grizzled Squirrel Wildlife Sanctuary in Tamil Nadu in the south. A study was conducted in the riparian habitats of Chinnar Wildlife Sanctuary in Kerala to understand the habitat characteristics, including the drey site use of GGS. The vegetation of the GGS habitat was studied using the quadrat method, and the dreys were counted using the transects. A total of 95 species of trees were identified from the riverine vegetation, and the GGS used 36 species of trees for drey construction. Most of the dreys were found on Mangifera indica, Terminalia arjuna, Ficus microcarpa, Diospyros ebenum, and Pongamia pinnata. However, the GGS may prefer trees such as Mitragyna parviflora, Diospyros ebenum, Ficus microcarpa, Albizia procera, Acacia nilotica, and Acacia leucophloea for drey construction. The study also highlights the usage of large trees with extensive crown by the GGS for various activities such as feeding, resting, moving, and nesting, thus signifying the necessity for protecting the remaining riverine habitat at Chinnar Wildlife Sanctuary to ensure the long-term conservation of GGS. We recommend an urgent restoration by restocking with already existing, native tree species of the riverine habitat due to the extremely poor regeneration of trees in the riverine habitat that support the only population of the GGS in Kerala.


INTRODUCTION
The Grizzled Giant Squirrel (GGS) is endemic to southern India (Kerala, Karnataka, and Tamil Nadu) and Sri Lanka (Image 1). GGS shows one of the most important paradigms of isolated populations. In India, it is known to occur in severely fragmented locations, and a few with connections including Srivilliputhur Grizzled Squirrel Wildlife Sanctuary (Joshua & Johnsingh 1994), Theni Forest Division (Babu et al. 2013), Palani Hills (Davidar 1989), Anamalai Tiger Reserve (Kumar et al. 2002), Sirumalai (Sathasivam et al. 2008), Thiruvannamalai Forest Division (Babu & Kalaimani 2014), Hosur Forest Division (Baskaran et al. 2011), Athur & Dharmapuri (Paulraj 1991;Paulraj & Kasinathan 1993), Pakkamalai Reserve Forest, Gingee (Vimalraj et al. 2018), Cauvery Wildlife Sanctuary-Shivanasamudra Falls and Mekedatu on the Cauvery river basin (Karthikeyan et al. 1992;Kumara & Singh 2006); and Chinnar Wildlife Sanctuary (Chinnar WS) in Kerala (Ramachandran 1989;1993, Senthilkumar et al. 2007; . The GGS has three subspecies, with one present in southern India while all the three subspecies present in Sri Lanka (Ellerman 1961;Moore & Tate 1965;Phillips 1984;Corbet & Hill 1992;Menon 2014). Ratufa macroura dandolena is the smallest among the three races of GGS globally and is seen in southern India and Sri Lanka. In comparison, the other two races, Ratufa macroura macroura and Ratufa macroura melanochra, are endemic to Sri Lanka. The home range of the GGS is between 0.197 ha and 0.611 ha (Joshua 1992). There are less than 500 mature individuals of GGS in India (Joshua et al. 2008;Goonatilake 2019). However, a recent study estimated the population of the GGS and found a considerably low population in Chinnar WS than the previous estimate .
Though the Chinnar WS has an extent of 90.44 km 2 , the distribution of GGS is confined to a narrow stretch of riparian vegetation along the Chinnar and Pambar rivers and their tributaries. Either side of this riparian vegetation is surrounded by scrub jungle to dry deciduous forests not used by GGS. The actual extent of the riverine habitat preferred by GGS comes to only 2 km 2 which is around 2% of the total area of the sanctuary. The previous studies on GGS in Chinnar WS (Ramachandran 1989(Ramachandran , 1993Senthilkumar et al. 2007; revealed that the animal's habitat is patchy in distribution and limited by the treeless areas in Chinnar WS (Ramachandran 1993). However, in Srivilliputhur WS and Sri Lanka, the GGS exploits the plantations of mango, coconut, and tamarind (Joshua 1992;Phillips 1984). The significant conservation challenges being faced by the GGS in Chinnar are increased predation risk due to opening up of the canopy due to natural and anthropogenic effects (Thomas et al. 2017), hybridisation between GGS and Indian Giant Squirrel , disturbance in the habitat because of tourism and road kills (Ramachandran 1993).
Although there are some studies on the habitat and nesting behaviour of this species (Joshua & Johnsingh 1994;Senthilkumar et al. 2007), a detailed study on the drey site usage of the species is not available. The information about drey site usage will be helpful for the long-term conservation of GGS.

Study area
The study was conducted between April 2013 to May 2014 in the Chinnar WS, Kerala, southern India. It is located between  in the Kerala part of southern Western Ghats, in Idukki district (Figure 1). Though the Chinnar WS supports the only known population of GGS in Kerala, the GGS is seen in the Anamalai Tiger Reserve, in Tamil Nadu too, which is adjacent to the Chinnar WS. The terrain of Chinnar is undulating, with altitudes ranging 440-2,372 m. Chinnar supports different vegetation J TT types such as southern tropical thorn forest (scrub jungle), southern dry mixed deciduous forest (dry deciduous forest), southern moist mixed deciduous forest (moist deciduous forest), tropical riparian fringing forest (riparian forest), southern montane wet temperate forest (shola forests), and southern montane wet grassland (grasslands) (Anonymous 2012; Champion & Seth 1968). The dominant vegetation among these is the dry deciduous forest followed by scrub jungle, mainly found in the plains and lower altitudes. The dry deciduous and scrub jungle together constitute about 70% of the total forest area in Chinnar (Thomas et al. 2017). However, the GGS in Chinnar WS are primarily seen only in the riverine forests along the Chinnar and Pambar rivers and tributaries which account for only about 2% of the Chinnar WS (Ramachandran 1993;.

Sampling of the Grizzled Giant Squirrel dreys
Eight, 1,000 m long transects were laid randomly after the reconnaissance survey done in the riparian habitats of Chinnar WS. These transects were walked twice a month for 12 months and recorded details of all the sighted dreys such as the number of dreys, tree species, tree height, and drey height.

Vegetation sampling
A total of 100, 10 x 10 m, quadrats were sampled for studying the vegetation in the riverine habitats in Chinnar WS. The 100m 2 (10 x 10 m) quadrats were laid at every 100m on five transects in the riverine habitats. In each quadrat, all trees with >10 cm GBH were enumerated, where the name of each tree species, the height of the tree in meters and girth at breast height in meters were recorded. The vegetation characters of tree species were quantified by calculating the following eight parameters as detailed below (after Pascal 1988).

Statistical analysis
We calculated the selectivity index (Ivlev 1961) to find out the relationship between the vegetation and the drey site preference by the GGS. We also performed a linear regression model to find out the relationship between the height of the tree species and the drey height.

Tree species composition and diversity in riparian habitat at Chinnar Wildlife Sanctuary
A total of 95 tree species were recorded from the riparian habitats of Chinnar WS. The ten most dominant  @-arranged in the descending order of the IVI index value | D-density (trees/ha) | RD-relative density | A-abundance | PF-percentage frequency | RF-relative frequency | BA-basal area (m 2 /ha.) | RBA-relative basal area | IVI-important value index | RIVI-relative important value index. tree species in the GGS habitat were Pongamia pinnata, Terminalia arjuna, Mangifera indica, Pterocarpus marsupium, Alphonsea sclerocarpa, Ficus benghalensis, Syzygium cumini, Ficus microcarpa, Sapindus tetraphylla, and Spondias pinnata ( Table 1). The riparian habitat of the GGS is also characterised by a tree density of 862 trees per hectare and a tree basal area of 7.13 m 2 /ha.

Characterising of the drey site occurrence of Grizzled Giant Squirrels
The GGSs were found to be using about 36 trees in Chinnar WS for drey construction (Table 2), and a total of 144 dreys were recorded. The following five species were found holding 54.86 % of the total dreys. The five species were Mangifera indica (n= 19) Terminalia arjuna (n= 18), Ficus microcarpa (n= 16), Diospyros ebenum (n= 14), and Pongamia pinnata (n= 12). The correlation studies between the important value index (IVI) of the trees and drey numbers showed (Figure 2) that there is no correlation between the dominant trees and the drey selection. This also suggests that the choice of nesting trees by the GGS is not random, and going by the ranks of the Ivlev index indicate that the GGS may have a preference for the trees such as Mitragyna

Regeneration of the trees in the riverine habitat in Chinnar WS
The regeneration of the trees used by the GGS for drey construction was extremely low in Chinnar WS (Figure 3), as evidenced by the absence of individuals in the lower girth classes for most tree species. Ideally, the girth class distribution of the tree species in an undisturbed forest should have been showing an inverse 'J' pattern (Pascal 1988). In contrast, at Chinnar riverine patch, the plants with lower girth classes were more or less completely absent for most of the trees except Pongamia pinnata.

Relationship between the tree height and the drey height
The linear regression model analysis clearly showed a strong correlation between the height of the tree and the height at which the drey was constructed (R= 0.9483, P <0.0001) (Figure 4). It can also be deduced from this graph that the GGSs showed more significant variation in the height of the trees used to construct the dreys, which varied from 5 m to 30 m (Figure 4), with most of the drey height being between 15 to 20 m. However, it is interesting to note that the tree height influences the drey height in the respective habitat (Figure 5a,b).

DISCUSSION
The GGS were found using large, mature trees for the drey construction in the riverine habitat at Chinnar WS. The dreys were built just below the canopy of the trees. GGS's usually prefer significantly larger trees with greater girth and taller trees with multiple branches for drey construction (Senthilkumar et al. 2007). The selection of mature trees with greater canopy continuity could facilitate easy movement to and from the drey in all directions, equip the animal by providing a significant advantage to escape from predators and to move to other parts of the home range for foraging and other activities. The extent to which GGS is using a habitat depends on the composition of tree species and structural attributes of the forests canopy, predominantly the canopy continuum. Discontinuous forests are known to restrict their movement and dispersal because of their arboreal habit. Most of the arboreal dwellers and GGS prefer those habitats that provide dense canopy cover and higher canopy height (Baskaran et al. 2011;Nagarajan et al. 2011).
The observations in this study corroborate the findings of previous studies that giant squirrels prefer areas with canopy connectivity to live and build their dreys (Baskaran et al. 2011). The first branching height of the tree increases with its total height; hence the drey would have to be higher on taller trees. This may be helping the animal to escape from predation and effective utilisation of its entire home range for resources.
The GGS in Chinnar WS was found to construct globular dreys using leaves and twigs. One pair of GGS makes multiple dreys within their home range, similar to other giant squirrels (Prater 1971;Srinivas et al. 2008). The GGS was found to construct multiple dreys at a time. The construction and use of multiple dreys might provide conveniently placed insulated nesting places throughout the territory or reduce drey predation (Borges 2015). Drey rotation may also help to avoid extremes of weather conditions like temperature and rainfall. The rotation of the drey will also reduce ectoparasite load.
In the Sitanadi WS, 77.68% of the dreys of giant squirrels were found on deciduous trees, while 22.32 % were located on the evergreen trees (Kanoje 2008). However, in the present study, 73.61 % of nesting trees