4 2004 
  
winter 
ze, with 
lowest 
(Figure 
| ranges 
0. 
gher in 
ze from 
of the 
almost 
ober to 
nan that 
1out the 
of the 
> higher 
ure 9). 
ner range of DI 
y range of ID1 
x range of ID2 
  
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
0.15 + 
| | |—9— Summer range of ID] 
| | —— Summer range of ID4] 
| | -9— Winter range of ID3 | 
LA Winter range of IDA. | 
NOVI 
eo 
3 
0.05 | 
000 | 
    
A 
N 
$20.90 QV eue qu que 
SS NN Qs QV wv 
OP IY 
i N M f a y 
S e ¥ WY VSN ANA GV of of 
Date 
  
Figure 7. Seasonal changes in NDVI values 
of the summer and winter ranges in Dornogobi. 
( 
[0 Winter range of IDI 
LA: Winter range of ID2 
E 
Relative NDVI 
  
S QD US uS OUI eA qv UIS do uude IN SS PN 
; VEN 
Date 
  
Figure 8. Relative NDVI in the summer and winter ranges, 
Omnogobi. 
020 f E 
| 0 0. 2 "9 
016 | ; = 8 o 
| £ vip 
012 | Q $9 o. 6 n° 
1—9— Summer range of IDà 
* O0 Summer range of ID4 
—#— Winter range of ID3 
L:--Ar-- Wänter range of 1D4_| 
  
Ce NADP UN uuu de^ duo dud demo 
eS Y Wow Jo” oval AV qv s 9 
  
Date 
Figure 9. Relative NDVI in the summer and winter ranges, 
Dornogobi. 
4. DISCUSSION 
This is the first trial to show the migration routes of Mongolian 
gazelles by satellite tracking. Their cumulative moving 
distances were more than 1000 km. Satellite tracking proved 
their moving ability and provided details of their "migration 
routes with location data. These data can be used to analyse and 
understand the reasons of their long-distance migrations and for 
conservation of their habitat. Mongolian gazelles are gregarious 
and often form large groups, some as large as 80,000 animals, 
during spring and autumn migrations (Lhagvasuren & Milner- 
Gulland 1997; Jiang et al. 1998). In fact, the tracked gazelles 
belonged to large herds of hundreds of animals when they were 
captured. We also found larger herds in their home range the 
following summer. Therefore, it is likely that several thousand 
gazelles moved together with the four gazelles being tracked. 
421 
NDVI values changed seasonally. It was high from June to 
August and sharply declined to around zero from late November 
in Omnogobi and Dornogobi. It suggests that the plants leaves 
were dead or grounds were covered by snow in winter. Since 
migration of gazelles to the winter range also started in late 
November, gazelles might have moved to avoid deep snow 
accumulation. 
Changes in NDVI values between summer and winter ranges 
corresponded with seasonal migrations of gazelles in Omnogobi. 
NDVI values in the summer range were higher than those of 
annual range in summer, but lower in winter. In contrast, NDVI 
values in the winter range were lower than those of annual 
range in summer, but higher from November to December. This 
sift explains that gazelles selected the area where the vegetation 
was more abundant and migrated seasonally. 
However, the trends of NDVI values between summer and 
winter ranges were different in Dornogobi. NDVI values were 
higher in winter range than both in summer and annual ranges 
thorough the year, although NDVI values in summer range were 
higher in summer and lower in winter than annual range. This 
means that there are other factors than NDVI to explain 
seasonal migrations of gazelles in this area. 
Leimgruber et al. (2001) divided gazelle habitat into three 
categories, winter area, summer area, and calving area, and 
pointed out the seasonal switching of NDVI values between 
calving and winter areas in eastern Mongolia. In the present 
study, however, similar switching of NDVI values was shown 
without separating calving areas from summer range in 
Omnogobi. On the other hand, NDVI values were higher in the 
winter range than in the summer range throughout the year in 
Dornogobi. To evaluate NDVI as an indicator of the gazelle 
habitat, it is important to understand what brings these regional 
differences. 
In Mongolia, precipitation increases from southwest to 
northeast. Omnogobi is located in arider area than Dornogobi 
and eastern Mongolia, and the vegetation in the winter ranges of 
gazelles seems very poor in Omnogobi, because NDVI values 
were around 0.1 even in summer and lower than those in 
Dornogobi. In such areas, places where plant biomass is greater 
may be good places for gazelles, and good places could be 
recognized by higher NDVI value. In more humid areas such as 
central and eastern Mongolia, however, several vegetation types, 
for example short grasslands, tall grasslands and shrub lands, 
occur and NDVI may not directly indicate habitat qualities for 
gazelles. 
NDVI could be one of good indicators of gazelle habitat. 
However, it is important to evaluate the effective extent and the 
limitation of NDVI as an indicator. Besides, tracked gazelles 
sometimes moved along mountains and a railroad. Such 
topographic and artificial factors might affect their habitat 
selection. Further researches on migration routes, the vegetation 
survey in the field, and analysis of the habitat selection of 
gazelles are needed to conserve Mongolian gazelle. 
5. CONCLUSIONS 
We described the migration routes of Mongolian gazelles 
throughout a year for the first time, using the satellite tracking. 
The gazelles seemed to migrate seasonally, depending on the 
seasonal change of habitat quality between summer and winter