The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008 
273 
mainly in the montane zone, but included even top of the moun 
tain (3742 m). 
3.2 Japanese serow 
3.2.1 The best logistic regression model (GLM) 
The best logistic regression model was considered to be pre 
dicted by the square root of altitude and slope without interac 
tion, derived from the following predictive equation: 
log (p/( 1 -/?))=-5.91785430+0.040241 36*jc,+0.26478759*x 2 , 
(Equation 3) 
where xj is slope (°), x 2 is square root of altitude (m 2 ) and p is 
the probability of Japanese serow’s occurrences. 
3.2.2 Accuracy assessment and comparison of models 
Table 2 shows the results of accuracy assessment by Kappa sta 
tistics. In general the score of indices was better at an optimized 
probability than at a threshold of probability of 0.5 except a 
case of GLM for the test data. For the Japanese serow’s predic 
tion, GARP was the best algorithm to fit the distribution; all in 
dices were perfect at a threshold of a probability of 0.5. 
3.2.3 The estimated population size 
Figure 1 (b) shows the potential suitable habitat patches for 
Japanese serow predicted by GARP, which was the most 
accurate among three algorithms. The patches were grouped 
into two local populations: the area of interest and others. 
Population size was estimated as indicated in Table 3. As for 
the area of interest, the predictive map showed a large 
connected habitat (1581 km 2 ) which covered Tanzawa region, 
Hakone region and part of Fuji region and the estimated 
population size was under 1581. This large patch in Tanzawa, 
Hakone, and Fuji ranged from 250 to 1800 m asl. The habitat in 
Fuji region was not predicted in the higher altitude range of Mt. 
Fuji (1800 to 3742 m) but it seemed to surround Mt. Fuji and 
was concentrated around the lower part of the mountain. The 
slope steepness ranged from 16 to 30 °in general, but included 
some parts with a gentle slope of 0 to 15 °. The patches in 
South Alps were of a size of about 537 km 2 and the estimated 
population size was under 537. Despite their patchy appearance, 
the habitats were not distant from each other. The predicted 
suitable habitat surrounded the mountains in South Alps region. 
Also here it showed serow’s presence in slopes ranging from 16 
to 30 0 and in lower altitude areas from 250 to 600 m. The 
boundaries of Fuji-Hakone-Izu National Park and Tanzawa 
Quasi-national Park were overlayed with environmental 
predictors used for Japanese serow in GARP. This revealed that 
the Hakone National Park and Tanzawa Quasi-national Park 
were fully covered by suitable habitat, but that only a small part 
of the Fuji National Park was suitable habitat. In general, the 
potential suitable habitat seemed located along the highways. 
However, the suitable habitat was distant from the paths and 
stone steps. 
4. DISCUSSION 
The population estimations were based on area in km' and 
population density. The results revealed that there are 5~9 bears 
in Mt. Ashitaka, 51-102 bears in Fuji and Tanzawa regions, 
160-320 in South Alps region, 4-8 in Mt. Kenashi, 4-8 in Izu 
Peninsula, and 6-11 in Hakone. In fact Izu Peninsula was the 
place where the bear was extinct (Japan Wildlife Research Cen 
ter, 1980; Ohba and Mochizuki, 2001) between in 1980 and in 
2003. Similarly, in past it is believed that the bear inhabited 
Hakone but now it is extinct (Kanagawa Prefectual Museum of 
Natural History, 2003). Since the predictive model mapped the 
bear’s potential suitable habitat, it is reasonable that it shows 
the area of extinction. The fragmentation pattern in the map in 
dicates the causes of extinction. First, the habitat in Izu Penin 
sula is too patchy and isolated from other local population. Ha 
kone is not geographically distant to the other local populations, 
such as Fuji and Tanzawa local populations (Kanagawa Prefec 
tual Museum of Natural History, 2003), but there are wide 
roads and highways among them. It suggests that these roads 
might be the main barriers which blocked the bear’s passage 
between Hakone local population and others. Several bear’s 
road kills are reported in the study area (Okumura et al., 2003). 
Therefore, the creation of green corridors for bear is recom 
mended. For Hakone and Izu Peninsula habitat patches, esti 
mated population is ranked as “in serious danger of extinction”. 
The South Alps local population is ranked “endangered”. Mt. 
Ashitaka and Mt. Kenashi are considered to be part of Fuji local 
population. There is concern about isolation of these sub-local 
populations, especially the one in Mt. Ashitaka (Ohba and Mo 
chizuki, 2001). The outcome of this research reveals that there 
is serious danger of extinction for the sub-local populations in 
Mt. Ashitaka and Mt. Kenashi. The results also confirms con 
cern for isolation in Mt. Ashitaka (Ohba and Mochizuki, 2001). 
Also it is not a part of the National Park. Therefore, it is rec 
ommended to include this area in an appropriate zoning plan. 
However, the local population of Mt. Kenashi seems to have a 
connection with adjacent Fuji local population. For the area of 
interest, the predictive map showed an existing linear corridor 
which connects Fuji and Tanzawa local population. Also this 
patch is ranked “in serious danger of extinction”. Since the 
patch connected by the corridor has a low potential population, 
it is suggested that, even after designing ecological networks, 
the size of the habitat may still not be sufficient. Also, this lin 
ear linkage is not part of National Park or Quasi-National Park. 
It is recommended to protect this linkage by creating a new 
zoning plan. Given that the sum of the populations of Kenashi 
and Fuji and Tanzawa still ranks ‘in serious danger of extinc 
tion’ it is recommended to extend the corridor further to link up 
the larger population of South Alps. 
For Japanese serow, in Fuji and Tanzawa regions population 
was estimated less than 1581, and in South Alps less than 537. 
From the assessment criteria, the number of individuals for both 
local populations is considered sufficient. Serow’s predictive 
map showed that in South Alps the local population occurred in 
lower altitudes than in the Fuji-Hakone-Tanzawa local popula 
tions. In general the predicted suitable habitat is located along 
the highways and distant from paths and stone steps. The habi 
tat selection of geographic proximity of highways and lower al 
titude may suggest the results of competition with Sika deer. 
Also when considering prey-predator relationship between Asi 
atic black bear and Japanese serow (Huygens et al., 2003), the 
serow may select the habitat where bear’s presence is avoidable. 
5. CONCLUSIONS 
Based on the most accurate predictive map for each species, po 
tentially suitable habitat was identified as the ‘core areas’ 
which may include ecological networks, if necessary. Then the 
population has been estimated in order to assess the needs for 
ecological networks. In the case of Asiatic black bear, the habi 
tat patch in the area of interest, Fuji region and Tanzawa region,