) is proposed
it is uniquely
is the only
f the object.
spatial and
relationships
d nonspatial
of object. In
e shared by
e amount of
lowever, only
plemented in
s and wells,
sly. How to
omena with
ges is not
, the three
e model.
la Model
, Wherein an
id defined by
fier. All the
e linked to it
ral reference
les to these
ral reference
| data and
e object has
which places
set up two
of the object
on the object
another one
-igure 5 (b)).
| information
1aged by an
ent system
e object, i.e.,
nation and
well as the
stored in the
ined by an
management
TION
sed so far
objects in a
nportant, to
' concern of
can not only
emporal and
ent, but also
the stage of
s with GISs.
Xcts can be
tion and the
lavior of the
sved.
The development process of the environment is
described by environmental models, tracing the
major factors for development. Suppose that the
objects can be constructed based on these major
factors, we have enough reasons to state that, in
such a case, the environmental models are better to
be included in the objects to describe the
development processes of the objects. Therefore,
there are two strategies here to combine the spatial,
temporal and attribute characteristics of an object
with its development processes (in the form of
environmental models) into a single object: one is
to organize objects according to their geometric,
temporal and nonspatio-temporal characteristics,
building operation on them based upon
environmental models. Another one is to organize
the objects according to the requirements of
environmental models, assigning the spatio-
temporal references to them. As the first case is not
easily to integrate the environmental models, the
second strategy provides the possibilities to fully
embed environmental modelling into a spatio-
temporal GIS data model.
6. CASE STUDY
One of the Dutch barrier islands, Ameland, has
been selected as a test area for research. At certain
locations of the island, severe (marine) erosion
happens, while at other places accretion or
accumulation occurs. To be able to predict which
process will dominate for a certain location in future,
it is necessary to know the various processes, the
interaction of these processes and the governing
factors concerning the sediment transport on the
land-sea interface. Such information is also quite
important for optimizing the coastal defense works,
e.g. beach nourishment or planting grass, which
require high investments.
The morphodynamically active areas will be
selected as testing places. On Ameland the
geomorphologic processes can be distinguished
through the interpretation of remote sensing data in
various landscape units, particularly, the shoreface,
beach, dunes, saltmarshes and tidal flats.
We choose the landscape units as our mapping
units. In this way the data can be structured
according to environmental units, which make that
the integration of environmental modelling will be
relatively easy.
We use three levels to represent the geomorphology
of this island. The first level is the landscape units
as foreshore, beach and foredune, which are the
concepts used by geomorphologist and each of
them can be described by three factors as forms,
process and materials. These three factors can be
considered as objects at second level. The spatio-
temporal data to describe these factors are stored
as objects in the third level. The environmental
models are set up as another class of objects,
described by their environmental factors such as
sea, wave, wind and vegetation, etc. These factors
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
can also be classified as objects with spatio-
temporal references. The environmental models are
embedded into the description of the landscape
units through the process associated with these
units. Such structure is shown in Figure 6.
7. CONCLUSION
Several aspects concerning how to design data
models for coupling environmental models with
GISs are discussed in this paper, i.e. (1) whether an
application-oriented or a systemdesign-oriented
perspective should be taken for designing the
spatial model; (2) how the objects should be
constructed (geometric-based, temporal-based or
attribute-based), if an object-oriented approach is
chosen; (3) should the loosely-coupled or tightly-
coupled method be chosen for the physical
implementation of the conceptual model; and finally
(4) how environmental models can be integrated
into objects. A general spatio-temporal data model
is proposed and a case study is given as an
example of full coupling of environmental modeling
in GIS. It is hoped that the discussions in this paper
will theoretically contribute to the data model design
for the integration of environmental models within
GISs, and to the development of object-oriented
concepts in GIS data models.
References
Booch, G., 1993. Object-Oriented Analysis and
Design with applications, Second edition, The
Benjamin/Cummings Publishing Company, Inc.
Cheng Tao, Van Zuidam, R.A. and Kainz, W. 1995.
A Unified Spatio-Temporal Data Model For 4-D GIS,
Proceedings of GIS/LIS'95 Annual Conference and
Exposition, Nov. 12-16, Nashville, USA, pp. 967 -
976.
Cremers, N.H.D.T., Wouters, W.J.C., Roo, A.P.J.
de, Wesseling, C.G., Burrough, P.A., and Van
Deursen, W.P.A., 1995. Integratie GIS en Dynamic
Modellen “Definitiestudie”, Survey Department,
Directorate-General for Public Works and Water
Management (MD-GAG-NR. 9503), -pp.34.
Molenaar, M. 1995. Spatial Concepts as
Implemented in GIS, Geographic Information
Material for a Post-Graduate Course, Vol.1: Spatial
Information, Edited by Frank, A.U., Department of
Geoinformation, Technical University Vienna, pp.
91-154.
Peugeut, D.J., 1995. It's about time: A Conceptual
Framework for the Representation of Temporal
Dynamics in Geographic Information Systems,
Temporal Data in Geographic Information Systems,
complied by Frank, A.U. Kuhn, W. and Haunold, P.,
pp.149-170.
