d to 
las 
for 
ially 
nce. 
ere 
ler 
for 
le, 
iffer 
  
(1) Compared to the area of the Kushiro Wetland 
shown in Figure 6, potential nesting areas in Figure 
7 are small, scattered and separated from with 
each other; and 
(2) Most potential nesting areas are located along the 
edge of the wetland which is outside of the 
protected areas. 
One exception is found in a circled area of the lower left 
part of Figure 7. Although this area is part of the 
Kushiro Wetland, the natural conditions are considered 
different from the rest of the wetland because the road 
constructed north of it has changed the water flow 
pattern and, as the result, made the area inappropriate 
for nesting sites (Masatomi, H., personal 
communication, 1996). This difference is clearly shown 
in Figure 8 in which the locations of all identified nests 
from 1979 to 1994 except 1981-84 and 1986-89 are 
plotted. There existed no nest in these years in the 
same circled area as Figure 7. The model developed in 
this study was not able to distinguish this area from the 
rest of the wetland. The land cover data of the wetland 
derived from satellite images by Masatomi, et al. (1990) 
does not show any difference associated with the 
conditions, either. 
The areas for potential nesting is only about 5 % of the 
total Kushiro Wetland, and there is not much space left 
in the Kushiro Wetland for nesting. In addition, the 
potential nesting areas are located in the peripheral 
areas of the wetland where accessibility to predators is 
high. 
6. DISCUSSIONS 
Since a perfect model for crane nesting site selection is 
unknown and will be difficult to develop, the model 
employed in this study is based on a simple set of rules 
using available data of geographic features. And there 
is no other data to evaluate the model and result of this 
study. 
However, even though the maximum crane population 
of the Kushiro Wetland is unknown, experts believe that 
many cranes appear to build their nests in less than 
ideal locations (Masatomi, 1993). In the other parts of 
eastern Hokkaido, the extent of breeding areas has 
also been gradually expanding recently, and due to the 
reduction of breeding areas caused by wetland 
development, there have been an increasing number of 
new nesting sites under undesirable conditions 
(Masatomi, et al., 1994; 1995). This tendency can be 
explained by the small potential nesting areas 
scattered along the edge of the Kushiro Wetland which 
were derived in this study. 
The significance of small potential nesting in Figure 7 
heeds to be assessed with the accuracy of the original 
map and existing nest location data sets. The horizontal 
accuracy of ground features digitized from 1:25,000 
topographic maps is less than 18.5 m R.M.S.E. based 
on the fact that the accuracy of original topographic 
maps is better than 17.5 m R.M.S.E. and the digitization 
error is estimated at 6.25 m R.M.S.E. (Murakami, 1995). 
The accuracy of the existing nest location data is 
estimated at not more than 100 m. If this value covers 
95 % (20) of the error, the root mean square error can 
be estimated at 50 m. This error of nest location is not 
small enough compared to the buffer size derived in 
this study. Since there are a number of potential nesting 
areas smaller than 50 m, it would be necessary to 
incorporate this error component in the model 
development process for more appropriate 
interpretation of the result. 
7. CONCLUSIONS 
GIS technology was applied to analyze nesting sites of 
red-crowned cranes in the Kushiro Wetland in Hokkaido, 
Japan. A simple model for crane nesting site selection 
was developed using existing nest location and 
geographic features data. Potential nesting areas 
located by applying this model were small and existed 
only along the edge of the wetland. The result was 
consistent with the understanding of experts on the 
cranes whose breeding environment is becoming 
undesirable. It is expected that GIS will be also useful 
for the subsequent analyses required in the second and 
final stage of this research to find new nesting areas 
other than eastern Hokkaido. 
The GIS approach helped objectively visualize the 
circumstances of the Kushiro Wetland for the crane 
breeding areas, which have been pointed out by crane 
experts in subjective manners. 
Since the accuracy of nest location data is not high 
enough compared to the scale of spatial analysis 
required to develop a nesting site selection model, an 
error analysis method needs to be developed and 
incorporated in the future potential nesting site 
analyses. 
ACKNOWLEDGEMENTS 
This study was supported partly by the Environment 
Agency of the Japanese Government. Professor 
Hiroyuki Masatomi of the Hokkaido College of the 
Senshu University gave us useful comments and 
suggestions for this study. Mr. Masahiko Yoshii of the 
National Park Wildlife Office of Northeastern Hokkaido 
of the Environment Agency kindly provided the 
1:200,000 scale nest location maps for this project. The 
authors wish to thank Mr. Masayoshi Maeshima and Mr. 
Kei Sato for their experienced map digitizing work and 
Arc/Info operations in this project. Mr. Masao Kanai, Mr. 
Ryozo Onozuka and Miss Rie Ando also helped greatly 
for the analyses and document figures. 
577 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996