Geographical information system is used to manage all the
data and the information with facilities to fulfil editing, area
summary, interactive information retrieving, and updating.
3. CASE STUDIES
3. 1 Cultivated Land and Plantation Change
There has been a decrease in agricultural land and paddy
field, due to urbanization, routine construction, and shifting
of use since the Reform and Open Policy of 1979. In
response to the need for the consistent, reliable information
supporting the macroeconomical policy-making and
furthering the reform and development of Chinese economy,
the Ministry of Agriculture initiated the project, Dynamical
Remote Sensing Survey of Cultivated Land and Plantation
Change (DRSS), in 1992.
The study area was determined both by the requirements of
the nationwide program on the same purpose and the
representativity of cultivated land, paddy field, and their
change. Therefore, the study area is at the central part of
the Purple Basin, Sichuan province, China, between E103
» -107 » and N29 » -31 » , covering an area of about 35,
000 square kilometres, which is one of the most productive
areas , especially rice.
The DRSS project was designed to obtain information using
Landsat TM images of separate dates: 1992, the current
distribution of features; 1991, the look-back. Another set of
information for comparison is the results of the First National
Inventory of Land Use carried through in the early 1980s,
which represent the status in 1981/1982. The investigation
will be repeated annually. To guarantee the availability of
data, a much bigger study area was used, in case there is a
need for shifting arising from the covering of clouds or other
reason.
The classification system includes seven categories at first
level and seven subcategories at second level. See Table
1
First Level Second Level
Code | Name Code | Name
1 Cultivated Land 11 | Paddy field
2 Settlements 21 City and
and Township township
22 Rural
settlements
3 Communication 31 | Railway
32 | Highway
33 | Civil airport
Development
4
Zone
5 Isolated mills
and mines
6 Irrigation and
Hydro-works
7 Other
Table 1. Land Use Classification System
812
The minimum polygon is greater than or equal to four square
millimetres on the map of 1:100, 000, i. e. 4 hectares on
ground, and the minimum width of linear targets required is
greater than or equal to 10 metres on ground.
The width of linear targets and dispersed isolated rural
settlements is not obtainable on the Landsat TM image and
will be learned from ground survey, large scale airphotos,
and available data. For linear targets, different methods
are used in accord with the subcategories. Width of
railways is definitive on single or double lined routine.
Highways are roughly classified to three levels and the width
at each level is almost the same. These will be figured out
from available documents. Paths and ridges cannot be seen
on the satellite image, but they do take up a large portion in
the mapped cultivated land and should be subtracted from
the cultivated land. Enough samples selected within a
stratified random sampling design in which strata are
defined with landform are investigated to find out the
compensation coefficients. Although this is laborious, the
coefficients are obtained, fortunately, mainly from the
available results of the recent nationwide land use
investigation on the scale of 1:10,000 in lowland and 1:50,
000 in highland. For rural settlements, there are many tiny
spots visible on the TM image but not considerable in
mapping. A similar method to that of paths is adopted to
make compensation for the incomplete of the information.
The number of spots’ are counted on enlarged Landsat TM
image or airphotos and the area data are achieved by
multiplying an average area of individual rural house or
homestead. Generally the area of each homestead mainly
differs on various type of landform that decided the
construction. Therefore, the strata for samples are defined
in the light of landscape regionalization.
Once the compilation for area features had been drawn and
checked both on a map and on a sampled field, the
boundaries were digitized, edited, and topologically built
using ARC/INFO geographical information system. A
statistical software with the direct access of the dBASE
formatted data specifically developed for the project was
used to sum up the area data incorporating the coefficients
and the width of the linear targets.
Figure 2 shows the diversities of land change ratio over
different counties. Urbanization results in the decrease of
paddy field and dry cultivated land. This is also obvious in
Figure 3. It was resulted that the change ratios, to some
extent, depend on the national policy and there exist
terrestrial discrepancies (Zen et al 1994).
3. 2 Resources and Environment Monitoring
Inventory and monitoring of resources and environmental
background is heavily demanded for regional development,
especially for the development of the west part of Sichuan
province which covers over the half of the provincial land.
There, the diversity of landscape with steep valley and high
mountain makes the inventory and monitoring more difficult
and inaccessible. Remote sensing technique seems an ideal
tool for this purpose. Therefore, the Resource and
Environment Monitoring (REM) was carried out nationwide
as the key project of the Eighth-Five Year Plan by a
cooperative team consisting of members from the Ministry of
Agriculture and Chinese Academy of Sciences, which will be
completed in this year.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
The m
of rer
land,
enviro
tempe
sophis
admini
The re
GIS, v
admini
agricul
sparse
A star
compa
code s
save s
the sa
system
Co
Tat
