Analysis of Adult Female Tiger Home Range Spatial Attributes.



Although data on 11 tigers were collected in 1992-1994 by means of radiotracking, this report is to analyze one animal home range attributes. This is due to lack of digitized data on spatial features, the background information to be made usable one has to extract it from various sources - topographic maps, forest survey database, air photos and satellite imagery, etc.

Such features, or attributes, as a) elevation b) slope c) aspect are taken directly or derived from a digital terrain model. Other features are distance to closest stream, to closest pathway, and forest cover type (which is instead of land use, as there're 100% of wild forest).

1. Animal home range and analysis area extent.

Female adult tiger Natasha had been studied for about 2 years before the database on locations was processed. This made up 130 points for 1992-1994. Currently radiotracking allows to obtain new locations of this animal (1995 - 1996 database).

Using adaptive kernel method (Calhome, 1994) contours of home range were plotted for 50 and 95% of locations (Fig. 1) Then an arbitrary rectangle was selected inside the home range as a basis for 89 locations, or 70%. This reduction was done in order to have statistically uniform boundaries of the studied area.

2. Elevation.

The elevations in the area are 100 - 1200 m above sea level.

Elevations of 89 Natasha locations were extracted from a DEM, with average 384 m, sn-1 = 159 m, above sea level. For comparison 89 random points were scattered upon the same, with average elevation 440 m, sn-1 = 158 m. The mean error for locations me is 17 m. The Mann - Whitney test (n > 20) transformed into normally distributed Z-statistics, where Z value is .01, a = .05, no significant difference stated.

3. Slope analysis.

Slopes (in degrees) were derived from DEM. The maximum value of slope was 55 degrees for the area, zero the minimum.

Average slope for locations of tiger was found to be 8.8 degrees, sn-1 = 8.5, me = .9 degrees, compared to 14.7 degrees average, 8.7 standard deviation and .9 mean error for random points (Tab. 1).

Z-value is -2.19, significant difference stated.

Table 1.

Summary statistics for elevation and slope.

Type Average sn-1 me Z
Elevations, meters        
384 159 17 .01
440 159 17  
Slopes, degrees        
8.8 8.5 0.9 -2.19
14.7 8.7 0.9  
4. Aspect analysis

Aspects for the area were also derived from the DEM, making it possible to split the rectangle into 9 sectors of different aspect. Then statistics was calculated on the availability and use of each aspect class (Tab. 2):

Table 2.

Aspect preference of the tiger.

Aspect Available, % Used, % Ratio
Flat  14.0 38.2 2.73
N 8.4 7.9 .94 > .72
NE 8.1 5.6 .69
E 10.5 6.5 .66
SE 12.6 5.6 .44
S 12.2 11.2 .92 > .72
SW 11.3 7.9 .70
W 12.4 7.9 .64
NW 10.5 9.0 .86 > .72
As there is a big influence of flat at the expense of other classes, total amount of locations fallen into flat (34) was subtracted from the rest and other 8 classes were expected to be in equal proportions if there was no preference. Thus, the value of .72 would be average for ratio of used/available in this case.

x2 value is 18.35, df = 8, a =.05, significant difference stated.

4. Closest stream and lane average.

Streams can be assumed equally important for an animal, because they do not differ much in the studied area. Thus, there was only one category of streams when the mean distance to the closest stream was calculated for actual locations and for random points. The same is true about lanes where one class is considered significant inside the rectangle.

As 95 % of points are no further than 900 m from some stream, any value larger than that can be taken as upper limit for the scope of points to calculate mean. The average distance to the closest stream for locations was 280 m, while for random points it was 372 m.

Z-value is .75, a =.05, no difference significant.

Inside the buffer zone of 1000 m around the lanes there were found 47 locations of 89, and 28 random points of 89, with average distance to closest lane for these points 367 and 559 m, respectively.

Z-value is 1.77, a =.05, no significant difference.

5. Forest cover analysis.

The satellite imagery classification of the studied area served for the forest cover type identification. Again the criterion of availability/use was applied, while classes of forest were assigned on the dominant species (Tab. 3):

Class Available, % Used, % Ratio
Oak 35.2 27.0 .77
Birch 26.4 20.2 .77
Mixed val. deciduous 13.1 19.1 1.46
Korean pine 11.9 22.5 1.89
Larch 6.3 3.4 .54
Fir 4.2 4.5 1.07
Rocks 2 3.4 1.70
x2value is 8.55, df = 4, a =.05, x2 < 9.49, no difference significant.