parameters related to watershed conditions (see tab.
6). All factors determine water quality which can be
described by the trophic state as one expression of the
complex situation in a lake ecosystem.
Information about these factors are collected by local
authorities, extracted form existing databases or maps
and calculated from remote sensing data.
In this study special emphasis is put on factors, which
can be assessed by means of remote sensing and
thematic map analysis. Information from remote
sensing data give a spatial overview of the situation in
the study area for each date.
Information on these parameters are organized within a
GIS for further analysis of spatial relations.
Remote |Thematic| Other
sensing | maps | Sources
|. Local lake factors
Areal descriptors X X
Bathymetry X X
Limnology X
Genesis X
Point source pollution X X X
Il. Watershed factors
Morphology X X
Drainage system X X
Surface cover X X
(Hydro-) Geology X X
Hydrodynamics X
Nonpoint source (x) X
pollution
Tab. 6: Sources of information for factors influencing
lake properties
5.1 Local lake factors
These are parameters which describe the lakes as
individual objects (see tab.6). The local characteristics
of the lakes are predominately defined by their genesis.
The study area (25 km X 20 km) includes 53 lakes
which were selected by a minimum size of 50000 m°.
Table 7 shows the local factors, which were assessed
in this study. Except for water depth, they were
determined by remote sensing. Based on these factors,
more complex parameters such as lake shore
development (Schwoerbel, 1993) can be derived.
Shore development is the ratio between perimeter of a
lake and the perimeter of a circle containing the same
area. The parameters were stored as thematic
information for each lake within a GIS.
53 Lakes Mean Min Max| Stdev.
Area [ha] 94,9 5,4| 405,0 89,7
Perimeter [m] 7500 1350| 20100 5280
Shore 2,2 1.3 3.7 0,6
development
Maximum depth
(incomplete data) [m]
15,5 2 68 14,1
Tab.7 Summary statistics of determined local lake
factors
The parameter water depth is of special importance
because it is closely related to lake genesis. Depth has
to be collected from maps, other databases, or has to
be measured. In the ideal case full bathymetric
information exist which allow morphological analysis of
the lake basin and determination of statistical
parameters. In the study area such information only
exist for a few lakes and were stored as a separate
data layer in the GIS. For most of the lakes only
maximal depth was included in the database.
Limnological data about the lake stratigraphy and data
about sources of point pollution give additional
information about the conditions in the ecosystems.
Such information have not yet been available for this
study.
5.2 Watershed factors
For the characterization of these factors remote
sensing is a valuable method because of its ability to
access spatial and temporal variations for large areas.
Table 6 summarizes the parameters which are
considered in this study. Morphology indicates the state
of drainage evolution in an area and determines the
surface catchments. The study area is characterized by
low relief and young drainage development. Further
analysis of these parameter will be based on a digital
elevation model which is provided by the Survey of the
State of Brandenburg. Its height accuracy amounts to
2-3 m.
For the differentiation of surface cover types remote
sensing data were classified using a multitemporal
approach. The high spatial resolution panchromatic
dataset which is available for June 1997 was
incorporated into the multispectral dataset of each
scene of the 1997 period. For merging, an IHS-
transformation was used where the intensity
component was substituted by the panchromatic band
before inverse transformation to the RGB-system. This
resolution enhanced multitemporal dataset was
prepared for a supervised maximum likelihood
classification of the main types of surface cover (13
classes — see tab. 8).
The benefits of the multitemporal classification are the
possibility of discrimination between cultivated fields
and meadows and the differentiation between forest
types. The settlements were classified separately within
the high resolution data by analyzing the cooccurrence
matrices and local histograms of the gray values with
an evidence based classifier. However, the settlement
classification had to be refined by visual interpretation
for certain structures where forest and bare soils were
confused with settlements. The final result was
integrated in the overall classification.
The relation between lake water properties and
surrounding landuse was determined by calculating the
percentage of landuse classes in an area of 250 m
around each lake as an expression of the direct
influence of landuse on the lake system.
134 International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998
Cla
Wa
| She
[Cor
Mix
| con
[ Dec
| Me:
Culf
veg
Cull
| veg
Cul
| cov
Bar
Seti
*durini
Tab.
Many
fores
pred
zone
withii
area
build
lakes
perci
these
the lc
Addit
inter
well ;
inforr
them
5.3 P
