ISPRS, Vol.34, Part 2W2, “Dynamic and Multi-Dimensional GIS”, Bangkok, May 23-25, 2001 
91 
GIS 
oordinates of 
ill points 
DCS 
sans: one is 
tion of every 
sess of mining 
ind movement 
j DEM can be 
ition, they are 
because the 
)llect, and use 
:ted by I DCS, 
software, and 
her. Based on 
can be done. 
Structure 
le of the data 
egrated data 
es of mining 
proposed as 
d DEM 
oal seam and 
' DEM or TIN. 
ace, rock and 
I according to 
e overlapped 
points In this 
be computed 
gh. 
¡tree and TIN 
| vector data 
ar 3D space 
; over goaf is 
i voxel can be 
nd surface is 
iifferent data 
structures are overlapped by spatial coordinates of control points. 
With this integrated data structure, the goaf can be represented 
more precisely by practical surveying data. The difficulty is how 
to compute the position of rock voxel, because the mechanical 
model is complex. A simplified method is to express the goaf and 
land surface use DEM firstly and compute the height of low and 
high face of rock on all girds, and then the rock is expressed by 
its low and high face as a even body. 
4.3.3 3D simulation based on CA model 
How to simulate the process of mining subsidence in a read 3D 
space is a very difficult problem, and there is not a satisfied 
solution so far. With the development and applications of Cellular 
Automata (CA) model, we think 3D CA model based on 3D 
raster and Octree perhaps can be used to simulate the process. 
In this model, the situation of each cell is determined by a 
moving rule based on its former situation and properties of 
neighbors. So coal seam, goaf, rock can used different moving 
rules to determine their situation, and then subsiding land 
surface represented by DEM of TIN can be overlapped on the 
3D CA. Because it is a new field and much work should be done 
further, we propose it as a future-studying theme. 
4.3 Integration of GIS and Forecasting Model 
No matter which kind of integrated data structure was adopted, 
the generation of DEM or TIN of subsiding land based on 
elevation of sample points is necessary. Two methods can be 
used to determine the elevation, one is by data from IDCS 
proposed in the former, the other is to compute the subsidence 
quantity of each points by forecasting model and then elevation 
can be got. For the latter, integration of GIS and forecasting 
model is key. Four means as follows can be used to integrate 
spatial analytical mode and GIS. 
(1) the loose integration scheme by data transmission between 
GIS and forecasting model, in which parameters needed is got 
from GIS and transmitted to spatial model to compute the 
elevation, and results are input to GIS, and then DEM or TIN can 
be generated. 
(2) external seamless integration based common interface, in 
which user can interact with a common interface and GIS and 
spatial model are conducted by messages, events and methods 
from interface. 
(3) developing forecasting module in GIS by secondary 
development, for example, using Map Basic of Maplnfo or 
AVENUE of ArcView to develop the processional model. 
(4) including GIS functional module into forecasting model. 
Fig.7 is the illustration to the four methods. All the four methods 
can be used according practice, and the former two are used 
mainly at present. 
(a) Loose integration 
(c) Including model in GIS 
(b) External seamless integration (<j) Including GIS in model 
Fig.7 Integration methods of GIS with forecasting model 
5 CONCLUSIONS 
Mining subsiding land is one of the most serious environmental 
damages in mining areas, and it should be regulated and treated 
by land reclamation and ecology reconstruction. It is useful and 
important to manage the subsiding land based on GIS, and 
some important issues including data collection, 3D 
representation and simulation, 3D data structure used, and so 
on. In this paper, an IDCS based on GPS, DPS and GIS is 
proposed to collect the 3D and dynamic information about 
subsidence, and some useful data structures for 3D 
representation and simulation including TIN or DEM, integrated 
data structure are put forward, and the integration of GIS and 
forecasting model is discussed at last. It is proved effective to 
use those techniques and methods to solve practical problems. 
Of course, there are much work should be done in the future, for 
example, the 3D CA model, simulation of dynamic process of 
mining subsidence, and so on. We would like to do further 
studies in this field. 
6 REFERENCES 
Du P, Guo D & Sheng Y, 1999, Data structures and visualization 
in 3D-GIS taking into account the properties and applications in 
mines. ISPRS, 32, 4W12, "dynamic and multi-dimensional GIS", 
Beijing, October 4-6 
He G, Yang L & Ling G et al, 1994, Mining Subsidence Science. 
Xuzhou, China University of Mining & Technology Press 
Chen Y & Guo D, 1998, Data structure of 3D MGIS, Mine 
Surveying, 1998(2): 11-14 
Li D & Li Q, 1997, Study on a hygrid data structure in 3D GIS,