anbul 2004
ures of the
/ector data
t of spatial
inside the
topological
writes the
diminfo,
netadata m
ecome a de
major GIS
id quadtree
Each index
ugh R-tree
capacity to
).
ia types for
d to further
| on object-
al features’
uction of a
used as the
:xtract, and
resentation
types, its
's and any
cannot be
maintained
)ase.
objects the
h-precision
f a second
)oints on a
create own
te dates or
led to meet
e attributes
c if a time
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
interval cannot be determined as accurate as the data type could
store (Plabst, 2001).
create or replace type POINT IN TIME TY as object (
year NUMBER (4,0),
month NUMBER (2,0),
day NUMBER (2,0),
hour NUMBER (2,0));
create or replace type TIME TY as object (
start POINT IN TIME TY,
end" ^POINT.IN: TIME TY,
start reliability NUMBER (1,0),
end reliability NUMBER (1,0));
3. OBJECT-RELATIONAL APPLICATION
DEVELOPMENT
In terms of Bavarian Forest National Park, object-relational
database technology can be used to model different
applications, e.g. the spatio-temporal process of deadwood
spreading based on its influencing factors or real time deer
tracking.
For the development of bark beetle outbreak some major
influencing factors have been proved to be crucial.
After windthrows climatological aspects are considered to be
decisive to trigger off a bark beetle calamity. The year's first
day at 20 degress above zero, the number of consecutive days at
that temperature level and the last day of snow coverage are
important to describe the beetles’ spring swarming potential and
the likelihood to breed twice a year (Nationalpark Bayerischer
Wald, 2001). These climate measurements affect the cambial
temperature of trees and the temperature sum which on their
part affect the beetles’ development conditions. The climate
parameters are surveyed up to three times a day at 14 weather
stations located in the national park region and kept in a
relational database model. We can create object types based on
the structure of these relational tables for any weather station:
create or replace type WALDHAEUSER TY as object (
year NUMBER (4),
month NUMBER (2),
day NUMBER (2),
t max NUMBER (3,1),
t min NUMBER (3,1),
t kw NUMBER (3,1),
SnOW cover NUMBER (3));
Based on this type an object view is created and OIDs are
assigned to the climate datasets of the corresponding weather
station per year:
create or replace view WALDHAEUSER . 1993 OV
of WALDHAEUSER TY
with object identifier (year, month, day) as
select year, month, day,t max, t min, t kw, snow. cover
from WALDHAEUSER
where year = 1993;
The columns of the relational base tables (weather stations) are
now accessible as row objects through their corresponding
object views. Any detail tables of the underlying. relational
model could now be simulated by further object views with
references that could point to the row objects of the particular
upper object view, to allow accessing the data in either way.
uA
uA
The next step is to create a superior type CLIMATE TY, that
keeps references to the row objects of all weather stations. The
data types are all bound not to object tables but to object views
as shown beforehand:
create or replace type CLIMATE, TY as object (
waldhaeuser. per. year ref WALDHAEUSER, TY,
lusen per year ref LUSEN TY,
taferlruck | per. year ref TAFERLRUCK TY);
To aggregate deadwood spreading influencing factors per year
in a main table (DEADWOOD), that addresses all
corresponding data stored in any existing (relational) database
model of the GIS-platform, table types are created based on
abstract data types, that represent each factor like in this first
example, CLIMATE TY:
create type CLIMATE VALUES NT
as table of CLIMATE TY;
create table DEADWOOD (
year NUMBER (4),
geom MDSYS.SDO GEOMETRY,
climate CLIMATE VALUES NT,
stand STAND VALUES NT,
soil SOIL. VALUES. TY,
object info XMLTYPE)
nested table climate store as CLIMATE VALUES NT TAB,
nested table stand store as STAND. VALUES NT. TAB,
nested table soil store as SOIL VALUES NT TAB;
time geom climate stand object info
NUMBER (4) | SDO, GEOMETRY | CLIMATE VALUES NT | STAND. VALUES, NT XMLTYPE
o tom menn tem prem] Sm intl d
SA T
1994 29 p rel E
c mne user af]
1995 SE Ft
1996 y O 1
Figure 6. Main table for analyzing bark beetle spreading
Deadwood polygons per year are derived from CIR images that
are acquired during annual flight campaigns each summer. As
they document the increase of deadwood areas since the
preceding year - so to speak they capture the bark beetles? work
with one-year delay, they have to be correlated with the climate
values of the previous year. This is achieved by referencing
climate row objects of object views of the preceding year. The
deadwood polygon geometries are stored in a spatial column.
As for data concerning stand structure information like species
and age of trees resp. soil types that are important to model the
humidity penetration of the ground, one can follow suit
whereby this kind of data is less dynamic or even quasi-static.
Any further information in respect of the situation of infestation
that cannot be structured to be stored as attributes can be kept
and querried directly inside the database in terms of document-
resp. data-centric XML (Lee, 2003).
Currently rejuvenation data gathered on the described
monitoring areas from 1998 to 2000 is modeled and integrated
to be analyzed in combination with existing inventory data of
1996 and 2000 using GIS and database oriented spatial
functions. In addition data mining techniques are explored to be
applied to find patterns in data that prove hypothesis how
rejuvenescence will develop in near future.
