Environmental Engineering Reference
In-Depth Information
a rather fuzzy type of land use, based on the official Chinese
statistical yearbooks (Suzhou City Statistical Bureau, 2001).
These images were classified into a dozen land use categories
used to derive the transition matrices indicating the amount of
each land use which was converted to any other during the two
periods in question: 1990 to 1995 and 1995 to 2000:
T
→
T
+
1
and
T
+
1
→
T
+
2. From these images, land parcel data was
extracted and then converted to vector data sets, complemented
by data associated with topography, geomorphology, vegetation,
precipitation and temperature used as the ancillary data in the
interpretation process. Further details on Landsat classification
are given in Liu, Liu and Deng (2002). The method adopted here
to extract dynamic changes in the vector land use datasets was
based on post-classification image comparison complemented
by field sampling to ensure quality control in the resulting
classifications. Control was executed by checking the identities
and the boundaries of sample land use patches with manual
adjustment to decrease the incidence of major errors. The TM
images in 1995 were used to interpret the dynamic change
vector data by comparing with the vector data derived from
interpreting TM images in 1990, while the same method was
applied for the period 1995 to 2000. In fact over both time
periods comparing 1990 with 2000 data, the overall classification
accuracyofmeasuredareasofalllandtypesisabout97%(Liu,Liu
and Deng, 2002), which gives us a high level of confidence in the
extracted change data and its allocation into land use categories.
Themain changes of land use in the study area took place with
the construction land and the paddy fields. In Table 24.1, we have
extracted the changes from an aggregated set of classes into urban
and rural construction (which we take to be urban/household
unit development in this context). The total construction vol-
ume was significantly reduced during 1995 to 2000 (40.7%)
and in particular urban construction (new development) was
dramatically trimmed down to around 55.4%. In contrast, rural
construction was reduced 19.5% in 1995 to 2000 compared to the
period 1990 to 1995. This conversion confirms that the policy of
protecting agriculture land and curbing the overheated economy
enacted in 1995 has had a notable impact on the desakota process
through subsequent land use changes and patterns. Moreover,
there has been a sharp increase in land use being converted from
paddy fields and a consequent drop in conversion from drylands
between the first and second time periods. This indicates that
the preferred land supply - drylands - for urban development,
has been severely diminished, which in turn has forced people
to take more productive farmland - paddy fields. Although strict
measures for controlling investment and urban expansion are
noticeable, it has been hard for government to discourage rural
residents from building more spacious houses due to increasing
affluence. As conversion from paddy fields is the largest cate-
gory in both periods, we can also examine the extent to which
paddy fields are converted to other land uses which is shown in
Table 24.2. Urban and rural construction still dominates taking
some 77% and 79% of paddy field land in the two respective
time periods with factory and transportation uses taking 3.5%
and 4.7%. Therefore the main focus of the ABM in this chapter
is to simulate the transition from rural (paddy fields) to urban
growth (including urban construction, rural construction and
special construction - factory and transportation).
24.4
The Suzhou Urban
Growth Agent Model
24.4.1
Themodel design
The formal structure of the Suzhou Urban Growth Agent Model
(SUGAM) is primarily derived from the Wuxian urban growth
agent model developed by Xie, Batty and Zhao (2007). There
aretwotypesofagentsconsideredinSUGAM:thetownship
TABLE 24.2
Percentage conversion of paddy fields to other land
uses over the two macro-time periods.
TABLE 24.1
Percentage conversion of land use to urban and
rural construction over the two macro-time periods.
Land use cover type
1990-1995
1995-2000
Land use cover type
1990-1995
1995-2000
Lost hectares of paddy fields
24362
18546
Converted hectares
of urban and rural
construction
27860(
=
16504
+
11356)
Dense forest
0.12
0.00
16514(
=
7374
+
9140)
Shrub forest
0.03
0.00
Shrub and loose
forest
0.30
0.65
Sparse forest
0.18
0.00
Orchard
0.14
0.00
Other forest
including orchards
0.06
1.12
Dense grassland
0.05
0.00
River
0.12
0.00
Highly-covered
grassland
0.00
0.36
Lake
2.66
0.00
Lake, reservoir, and
pond
0.41
1.56
Reservoir and pond
14.02
16.33
Shoal
0.01
0.00
Shoal
0.00
0.16
Urban construction
39.33
35.31
Hill and plain paddy
field
67.71
88.69
Rural construction
38.10
43.66
Large factory and transportation
3.51
4.70
Hill and plain
drylands
31.65
5.58
Plain dry land
1.72
0.00
Search WWH ::
Custom Search