Chapter 5
5.1 Determination of Multiplier
Density estimation was carried out based on distance data of giant sightings collected during the survey. Of the total 160 giant squirrel sightings, 90 were recorded from Bhimashankar Range 1 alone. Apart from higher abundance in Bhimashankar, the sampling intensity was also much higher in this range compared to the survey in the remaining areas. It was therefore decided to carry out the density analysis separately in Bhima1 Range and the rest of the study area.
Giant squirrels are much more active during the morning hours and evening hours (peak hours) with a lull in the middle of the day (offpeak hours). Therefore sampling should ideally be carried out in the peak hours when the giant squirrels are most active. However, in the present survey, it was not possible to restrict sampling only to peak hours and survey was carried out at peak as well as offpeak hours depending on the trail location and the logistics of reaching the trail.
A basic assumption of line transect surveys is that all squirrels are “available” for sampling. This is the case during the peak hours. However, during the offpeak hours some giant squirrels retire to their nests and are “not available” for sampling. This is the case only for a percentage of squirrels. This is contrary to the basic assumption that all squirrels are available for survey. Conventional distance sampling assumes that g(0)=1 on the sampling line. However, in the present situation g(0) is less than 1 in offpeak hours because of some squirrels not being available for sampling. A correction factor needs to be applied to the density estimate known as a multiplier. The multiplier is usually determined by a separate experiment.
Since the sightings made at Bhimashankar Range 1 were made in the same location and we had substantial number of sightings these were used to determine the multiplier.
We divided the sightings made at Bhimashankar into sightings made during peak hours and those made during offpeak hours. Peak hours were considered as the time before 11 am and after 4 pm during winter. In summer peak hours were considered as the time before 10 am and after 5 pm. The interval in between was considered as offpeak hours (Renee Borges, pers. comm.). The length of each trail was also divided into length sampled during peak hours and length during offpeak hours. In this manner the entire data for Bhimashankar Range 1 was segregated into two sets consisting of data collected during peak hours and data collected during offpeak hours. The encounter rates were determined for both these data sets as follows:
Re_{peal} = Number of sightings during peak hours
Effort during peak hours
Ru_{ffpeak} = Number of sightings made during offpeak hours
Effort during offpeak hours
Table 5.1: Encounter Rates (n/l) during peak and offpeak hours in Bhimashankar Sanctuary
Description

Total

Peak Hours

Offpeak hours

No. of trails

54

37

37

No. of Encounters

100

52

38

Effort

186.78

78.31

108.47

Encounter Rate


0.8812

0.5059

Standard Error


0.216

0.165

The encounter rates at peak and offpeak hours would have been the same if the giant squirrels were active throughout the day. However since they are less active during the offpeak the encounter rates during these hours is lower than the peak hours. The ratio of the two encounter rates reflects the ratio of giant squirrels that is available during offpeak hours to those available during peak hours. Therefore the multiplier is calculated as:
M = Ru_{ffpeak} = 0.5059 = 0.5741
Re_{peal} 0.8812
5.2 Density of Giant Squirrels in the Study Area
The analysis was carried out in Distance 6.0 Release 2. As stated earlier we defined two sets of data, each with their own sets of trails as shown in Table __. The offpeak trails need to be assigned the above multiplier so that the density estimate based on these trails is increased to reflect the number of giant squirrels that is not available for sampling. However the present version of Distance allows assignment of multipliers only at the Global layer and not at the transect layer. To circumvent this limitation the effort for offpeak lengths of all transects was multiplied by the above multiplier. This has the effect of reducing the transect length of the trails during offpeak hours thereby increasing the density to account for the unavailable giant squirrels.
The density estimation parameters for Bhima 1 are given in Table 5.1 and the density estimation parameters for the rest of the study area is given in Table 5.2.
Table 5.2: Density of Giant Squirrel in Bhimashankar Range 1


Value



Total

Peak

Offpeak

Number of Trails

63

55

53

Encounters

86

52

38

Number of Giant Squirrels

97

61

40

Effort

186.8

78.3

108.5

Effort (corrected)

140.6

78.3

62.3

Max strip half width (m)

42 m

Truncated strip half width (m)

30 m

Average cluster size

1.126

Selected Model

Uniform with polynomial correction

Density

15.89/sq.km. (11.0022.74)

%CV

18.29 %

Table 5.3: Density of Giant Squirrel in Rest of Study Area (Other than Bhimashankar)
Item


Value



Total

Peak

Offpeak

Number of Trails

191

123

146

Encounters

43

29

24

Number of Giant Squirrels

55

31

24

Effort

538.6

206.3

332.3

Effort (corrected)

397.1

206.3

190.8

Max strip half width (m)

55 m

Truncated strip half width (m)

40 m

Mean cluster size

1.037

Model

Uniform with cosine correction

Density

2.92/sq.km. (1.954.43)

%CV

15.82%

The density estimate for Bhimashankar is quite high at 15.89 /km^{2}while the average density of giant squirrel in the rest of the study area is 2.92/km^{2}.
Estimation of number of giant squirrel requires estimation of giant squirrel habitat in the sampled area. This is a complex exercise and involves interpretation of satellite imagery. It was not possible within the time and resources available to us. Hence estimation of giant squirrel numbers was not carried out. However this is a feasible exercise and may be attempted in the future.
