The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
1236
Bnmr
U/AI/ElSNCrVt [fwn)
Figure 1. The band model of SPOT5 data spectral characteristic
The uplift of Tibetan Plateau changes the atmospheric
circulation in mid-latitude westerly, forming the Plateau
monsoon climate. The temperature is low due to the limitation
of altitude. Therefore the average temperature is below zero in
most area. The average temperature in the south is higher than
that in the north at the same altitude. The annual average
temperature is gradually reduced along the southeast to the
northwest. The annual precipitation in Tibetan Plateau is
relatively less and the rainfall mainly concentrated in May to
September, accounting for 80%~95% of the annual
precipitation. Winter is dry season and the precipitation
occupies 5-20% of the annual. There are more thunderstorms
and hailstone in the rainy season. In addition, Tibetan Plateau is
the largest ice, snow and cold area in mid-low latitude region of
the world. There is 4.7 X 10 4 Km 2 covered by glacier accounting
for 80% glacier area in China, and frozen soil covers 150><10 4
Km 2 . The glaciers and snow lines change regularly, while the
distribution of frozen soil is significant different. Continental
glaciers are distributed in the central and northern plateau,
while oceanic glaciers are distributed in the southeast and
southern plateau. The permafrost area is general in Kailash area
which has continuous permafrost zone, The rest area is island
and seasonal permafrost (Liu et al., 1997).
The uplift of Tibetan Plateau also significantly affects the soil
forming process and its geographic distribution. It has alpine
soil which is the most comprehensive agrotype in China. The
regular zonal distribution reflects that hydrothermal conditions
gradually decrease from southeast to northwest, biological
effect becomes weaker and soil development degree tends to
original. The natural vertical zones are from the edge of the
plateau to the hinterland within plateau. Along with location
difference, the natural vertical zones have different base bands,
their structure becomes simple, and the number of the zones
decreases. In addition, the south of East Himalayas is one of the
richest rainfall regions due to the impact of southwest monsoon,,
therefore forms a vertical distributing series based on yellow
latosol and yellow lateritie red soil. The southeast Tibet
mountain gorge area mainly has yellow brown soil. It forms a
vertical distributing series of which the sub-alpine grassland
soil is the lowest level in north mid-Himalayas as well as
upstream and middle stream area of Brahmaputra. The deep
valleys of east Traverse-Mountains is dry and hot, therefore
form a soil vertical distributing series of which brown soil is the
bottom. In the eastern part of north-Tibet, the climate is clammy
and therefore it develops as alpine meadow soil. On the plateau
surface, the climate in south-eastern is cold and dry therefore it
forms as alpine steppe soil. The central plateau develops alpine
desert grassland soil. In northwest, the weather becomes drier
and colder therefore the accumulation of the humus becomes
weaker. The bottom layer is permafrost soil and therefore forms
alpine desert soil (Zhang et al. 1982).
Qixiangcuo area which lies in northern Tibet and northwest of
Naqu region was selected as the case study area. It covers the
area from N32° 11' to N32°41' in latitude and from E89°38' to
E90°17' in longitude and belongs to the shuang lake special
region which is an administrative region established in 1976 to
develop the depopulated area in northern Tibet. This area has a
vast territory and its residents live extremely scattered.
Qixiangcuo area which is gentle with open grassland locates at
Qiangtan Plateau lake basin. Its terrain is higher in the north
than the south, and mostly are dry-cold and semi-desert
meadows. The average altitude is about 4800m. The Kunlun
Mountains, Tanggula Mountains, Kekexili Mountains,
Dongbule Mountains etc. are located in this area. There are a lot
of lakes and Qixiangcuo is one of the famous lakes (Zhang et al.
1982; Wang et al., 2005).
3. RS DATA PROCESSING AND INTERPRETATION
3.1 Image data
A SPOT5 multispectral image with 10m resolution in March,
2004 and a panchromatic image with 2.5m resolution in
November, 2003 were selected as the case study data for this
interpretation task.
3.2 Image processing
3.2.1 Geometric correction
In order to match the two images in the interpretation process,
the geometric correction was adjusted first. Both of the original
images didn’t have precise geometric correction, therefore the
correction process was followed as below. First, the images
were corrected separately using a reference image to unify them
to the same geographic coordinate. Second, the two images
were proofed by the precise geometric correction. In practical
operation, a precisely corrected ETM image of September, 1999
was used as the geographic reference image, polynomial model
(cubic) was selected as the geometric correction model, and the
nearest neighbourhood interpolation was used in resampling to
correct two images separately. The correction results show that
the correction errors of two images are controlled within one
pixel. In the geometric registration, the same model used in the
geometric correction was used as registration model, and the
nearest neighbourhood interpolation was used in resampling,
the registration error was also controlled in one pixel (Meng,
2000).
3.2.2 Image synthesis