Counting Squirrels in Indian Forests

Although the ecology of squirrels has been extensively studied, most past work is characterized by the failure to account for detection and heavy reliance on indices rather than directly measuring abundance. This has involved acoustic and visual surveys, sign surveys (tracks, middens and dreys) and capture-recapture sampling (trapping rates) methods are adopted to estimate squirrel abundance. Such field studies assume the detection probability of the species to be equal in all sites, leading to incorrect estimates of true abundance.

Authors Devcharan Jathanna, N. Samba Kumar and K. Ullas Karanth demonstrate the use of line transect sampling involving 1996 km of walk effort to estimate densities of the Indian giant squirrel (Ratufa indica) in six sites in Southern India. These are the highlights of their study published in journal Current Science in 2008 (download paper).

Study sites

The study was conducted in six sites in Karnataka – Muthodi and Lakkavalli in Bhadra Wildlife Sanctuary in 1998, Nalkeri and Sunkadakatte in Nagarahole National Park in 2000 and in Bandipur in 1999. A dry deciduous forest in the southern part of Bhadra Tiger Reserve, Muthodi is covered by teak plantations and receives an annual rainfall of 2000–2540 mm. Lakkavalli is a moist and dry deciduous forest in the northern part of Bhadra. Along the western border of Nagarahole National Park is Nalkeri, a moist deciduous forest with teak plantations, receiving an annual rainfall of 1500 mm along the western border and 900 mm in the east. Sunkadakatte in Nagarahole abutts the Kabini reservoir is dominated by dry deciduous forest, with some areas supporting moist deciduous woodland. The driest site, the Bandipur Tiger Reserve has patches of moist deciduous forests within extensive dry deciduous stretches.

Methods

• The selection of six sites allowed adequate coverage of different habitat types. Transects were walked between 0615 h-0830 h and 1545 h -1800 h and the cluster size and sighting distance to squirrels were recorded. The density of the great Indian squirrel was determined by estimating cluster densities and multiplying it by the estimated cluster sizes.
• The walk effort (distance) covered during the survey was 384 km in Muthodi, 344 km in Lakkavalli, 288 km in Sunkadakatte, 504 km in Nalkeri and 476 km in Bandipur.

The results

• The encounter rates varied across the different sites ranging from 0.18 in Bandipur to 0.78/km in Muthodi.
• The estimated probabilities of detection varied across site ranging from were 0.52 in Bandipur to 0.60 in Muthodi.
• The density of the giant Indian squirrel was estimated to be 2.37 squirrels/km2 in Bandipur, 4.5 squirrels/km2 in Nalkeri, 4.8 squirrels/km2 in Sunkadakatte, 10.1 squirrels/km2 in Muthodi and 12.2 squirrels/km2 in Lakkavalli.

Conservation Implications

The authors find variation in squirrel densities across six sites, with Bhadra recording higher densities compared to Nagarahole and Bandipur. The study demonstrates the importance of accounting for detection and distance sampling to measure the giant Indian squirrel densities in forested habitats. Line transects, although effort-intensive are very useful to carry out long-term monitoring of squirrels and other species.