INVESTIGATION OF DEGRADED GRASSLAND USING REMOTE SENSING TECHNIQUES INJIANSHE 
TOWNSHIP AREA, GUOLUO PREFECTURE, QINGHAI PROVINCE, CHINA 
Genong Yu', Huamao Zhou, Liangxiu Zen, Hongju Zhan, 
Jun Tu, Dingchang Liu, Yafeng Zhong 
Chengdu Subcentre of Agricultural Remote Sensing, Ministry of Agriculture, 
CHENGDU 610066, P. R. CHINA. 
Nicholas Hodgson 
Natural Resourves Institute, Central Avenue, Chatham Mantime, 
Kent ME4 4TB, United Kingdom. 
Dejun Shi 
Project Management Oflice, Qinghai Bureau of’ Animal Husbandry, XINING, P.R. CHINA 
Commission VIL Working Group 3 
KEY WORDS: Remote sensing, GIS, Degraded grassland, Classification system. 
ABSTRACT 
It is of great scientific and practical significance to find out and monitor degraded grassland in terms of its extent and severeness in time using 
remote sensing and GIS. A pilot project was set up and cared out under Qinghai Livestock. Development Project, P. R. Chin/European 
Union, in order to identify the potentials of 3S techniques, ie. remote Sensing(RS), geographical information System(GIS), and. global 
positioning System (GPS), on investigating, monitoring and managing of degraded grassland. The test area is at Jianshe township, Dareng 
county, Guoluo prefecture, Qinghai province, China, between 98.900E, 33.350N and 99.750E, 33.900N (WGS8A). Two scenes of Landsat TM 
image covering the same area were used, late autumn and summer. Some adjustments of conventional grassland classification system were 
made to meet the interpretability of satellite remotely sensed images. Basic adjustment principles are compatibility with. general. grassland 
classification in China, representation of herbage and phytomass, interpretability of remotely sensed data and feasibility of grassland mapping, 
and emphasis of degraded grassland. Resulted grassland classification scheme includes 135 basic categories at three-level. The method tor 
interpretation of grassland resources using Landsat TM image is visual geo-correlated interpretation. Final maps and infonnation is refined in 
digital form and linked to GIS databases. Preliminary results show that suminer grazing plot is also severely atlected besides winter plots, 
visual interpretation of grassland in 1:100,000 using Landsat TM image is feasible, and best season for interpretation of image is summer, 
  
  
when phytomass in full wing. 
1. BACKGROUND 
It is a pilot project supported financially by European Union and 
supervised by the headquarters of Qinghai Livestock Development 
Project, the Bureau of Animal Husbandry, Qinghai Province, China. 
The objective is to find out the potentials of 3S, i.e. remote Sensing 
(RS) geographical information Systems (GIS), and global 
positioning System (GPS), on investigating, monitoring and 
managing of degraded grassland The project chose Jianshe 
township, Dareng county, Guoluo prefecture, Qinghai province, 
China. Chengdu Subcentre of Agricultural Remote Sensing (CSARS) 
joined the project on Landsat TM scene procurement, interpretation, 
and digital mapping. 
The project set the following requirements on the Landsat TM scene 
specification, 
(1) The season aspect of the TM scene was supposed optimally to be 
in the late autumn or the start of the winter, or within the period 
between late September and the end of November. The best time is 
in 1996, or as recent as possible. 
(2) The cloud cover of the study area, Jianshe township, should be 
below 10% and haze values be kept to a minimum either. 
  
! E-mail of the primary author: genong, yu(?usa. net 
(3) The scale of the hardcopy of the scene and its mapping is 
1: 100,00). 
(4) The unsupervised classification by means of visual interpretation 
should include as many discrete land/vegetaton wuts, identified on 
the image, as possible. 
(5) An interpreted transparent overlay, a digital map in the format of 
Maplnio, and all hardcopies and topographical maps should be 
provided to the project, or the Bureau of Qinghai Animal 
Husbandry. 
2. STUDYAREA 
The project defined the study area by an outer box with bottom left 
to top nght coordinates: 98.90°E, 33.35°N and 99.75°E, 33.90°N 
(Longitude/Latitude World Geodetic System 1984 Spheroid - 
WGS84). If the available scenes are restricted, however, the 
minimum area of coverage would be defined by the coordinate box: 
98.983°E, 33.429°N and 99.691°E, 33.826°N. The total area is about 
4800 square kilometers. 
The average altitude is over 4000 meters above the sea level and 
some peaks are about 5100 meters. Climate is of semi-humid alpine 
with distinguishable tow seasons: cold and warm. The cold season is 
between October and April, of gales and blizzards, and the wann 
300 International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
  
season 
averag 
month 
Januar 
9.8°C, 
(Regio 
Grassl: 
$000 n 
rhodio 
primari 
kobresi 
betwee 
slopes, 
alpine 1 
meters, 
and grc 
saxifrag 
and lov 
meters 
Commi 
Average 
Arnumal 
goat, ya 
3. 
Type of 
economi 
(Departi 
classific: 
by prof 
phytoecc 
1960s, 
classificz 
The basi 
classifica 
(1) Com] 
The fin 
inheritan 
China. ( 
developn 
1996). I 
vegetatioi 
(2) Herba 
The cla 
manager 
unport f; 
herbage, 
grassland 
(3) Inte 
grassland 
Classified 
topograph 
A key ele 
grassland 
Landsat T 
interpretat 
since it is 
be easily i 
(4) Degen 
Degenerat 
project try