Information Technology Reference
In-Depth Information
G(y)
denotes the total growth of the whole study area till year
y
,
L
d
(y)
represents the
total growth of only project
d
till year
y
. The assignment of
L
d
(y)
should be determined
from a top-down social-economic model. Eq.14 also offers a link between local
growth and global development. It is a feedback between top-down and bottom-up
decision-making.
L
i
(y)=h(L
i
(t) ) y=1, 2,...m; t=1, 2, ...,n; n>m
(15)
Eq.15 establishes a transition from
L
i
(t) to L
i
(y)
. In the previous researches of CA
application, a linear function is applied, i.e.
t=
is assumed to be a constant,
which means equal growth rate. In reality, function
h
could be a non-linear function of
iteration number
t
, which can be tested experimentally through visual exploration.
*y.
Here
3
Implementation
3.1 Case Study
Wuhan is the largest mega city in central China. In1999, it had around 4 million non-
agricultural population, 4 times more than that of 1949. During the last 5 decades,
Wuhan underwent rapid urban growth from 3000 ha of built-up area in 1949 to 3,0151
ha in 2000. As a result, Wuhan is a fresh and typical case for understanding the dy-
namic process of Chinese cities.
With the assistance of topographic maps of 1993 and SPOT Pan/Xs images of
2000, we found that land cover change in the period 1993-2000 was dominated by
strong spatial agglomeration of a few large-scale projects, which take over 60% of
total change. As a consequence, the understanding of local growth process of each
project is highly crucial to that of whole study area. Here,
Zukou
car manufacturing
center, the largest project, is taken as a case study for testing the methodology pro-
posed. The influential factors include major roads, minor roads, master planning,
physical constraints from water body. The cell size in this research is 100x100 m
2
.
3.2 CA Simulation
The validation of parameters has been proven difficult for urban CA modelling
[6, 16]
in particular when factors and parameters considered are voluminous. Here, we think
that manual test is much quicker and also more interpretable, which is based on the
modeler's reasonable understanding of urban growth process and visual exploration of
model outputs. The impact of each factor or parameter is assessed by changing its
value and holding the others constant. In this case study, the major parameters include
