LDER
n of
areas.
tion is
nthos.
aim of
or the
within
ieten.
detail-
s sind
lichen
(il der
tralen
kurz-
IS can
proce-
exible
jle for
r inte-
rophic
and in
s 5.60
ect is
d the
> very
sedi-
wind
h and
)hytes
e bay,
phora
scens,
Fucus
in the
on the
3eisel,
carpus
spec., Pilayella litoralis, Ceramium spec.) cover more or
less closed the submersed vegetation.
For the studies some special test sites were selected at
the south-eastern bight reach coast of the island Rügen.
They include different shore types. The main sediment is
sand, in front of active cliffs the ground is litter with boul-
ders and rubbles or marl plates. This conditions allow the
growth of the most macrophyte communities of the bay.
3. DATABASE
In the last years a lot of different data, images and maps
were collected about the test area to find the best combi-
nation of information for the evaluation of macrophytes
with remote sensing and GIS.
The basis for interpretation and classification was
scanned color aerial photographs in transparency format,
taken by an photogrammetric camera in 1992 and 1994.
In addition color negative photographs were made with
the amateur camera Rollei Metric 6006 at a Cessna in
1994 and 1995 in the scale of 1:5000.
Investigations were made to evaluate the best time period
for taking the photos representing sufficiently the
underwater vegetation. It is important to make a com-
promise on visible growth of vegetation and the water
transparency. At the Greifswalder Bodden the best period
to take photos of macrophytes is between the end of April
and early in May with low amount of chlorophyll and
suspended particles in the waterbody.
Besides the aerial photographs satellite images were
used. It was found that there were no Landsat scenes at
the right time, without clouds or ice. So only a Landsat
TM scene from July 1989, a SPOT (P) scene from July 89
and March 1995 could be used. This scenes fit best to
the discussed requirements.
The extensive map material was digitized: topographical
maps in the scale 1 :10000 and hydrographic maps with
bathymetric information in the scale 1 :10000.
Further information about distribution of macrophytes,
mussels, sediment, geomorphology and others were
taken from the literature (Geisel, 1986; Scabell & Jóns-
son, 1984).
Shots with an underwater video camera were made at the
, same time of the photo flights.
The geometric link of all data is guaranteed through
Gauss-Krueger-Coordinates. Through this uniform geo-
graphical reference the image data can be tied inde-
pendently of their scale and the collected data can be
proved integrately in a GIS.
4. METHODS
4.1 Ground Truth Processing
At the same time of the photo flights aquatic studies were
realized. From a rubber dinghy the vegetation was
recorded with a SVHS underwater video camera and
additional parameters were measured, like water depth,
position, transparency of the water column. Plant and
sediment samples were taken, determined and related to
the videos and the imagery. The details of macrophytes
obtained by the samples are not usually recognizable on
the videos. The turbider the water is or the quicker the
camera skims over the ground, the more difficult the
evaluation of the macrophytobenthos is. However the
underwater videos document continuously the vegetation
borders and the horizontal and vertical structure of
populations that is not possible with a simple sampling.
The recordings were repeated every 3 weeks in one year
to monitor the structural and phenological changes of the
macrophytobenthos within a vegetation period.
Based on the videographic recording the vegetation was
grouped into typical macrophytic societies according to
their appearance (Fig. 1). This grouping corresponds well
with the sociological units of the macrophytes evaluated
by Geisel in 1986 (Geisel, 1986).
Nb. | Macrophyte communities Depth
1 Enteromorpha, Cladophora, (Pilayella) |0.0-1.0 m
2 Potamogeton, (Ceramium, Laomedia) |1.0-3.0 m
3 Zostera 2.0-4.0 m
4 Furcellaria, Polysiphonia 2.0-5.0 m
5 Fucus, (Polysiphonia, Furcellaria) 0.5-2.5 m
6 Chorda, (Potamogeton) 0.5-1.5 m
143
Fig. 1: Typical plant societies in the study area
Based on the results of the under water videos the
training areas for the supervised classification were
chosen.
For geocoding of all data of the ground truth processing
and for a later rectification of imagery control points were
signalized and measured by differential GPS. Classical
measurements of some points by electronical tachyme-
ters supported the surveying by GPS. A high number of
potential control points is present on the land surface. In
the water 'natural' control points are very rarely to find.
Therefore the stakes of the permanent weir-baskets were
used as control points.
4.2 Digital Image Analysis
The aerial photographs were digitized and rectified. The
geocoding was difficulty because of large water areas in
the images and the relative small number of control points
in the water area.
After the geocoding the 1:5000 scaled images were
mosaiced. In that way bad edge regions of the images
like sunglitter or distortions could be eliminated, the
radiometric appearance will be more homogeneously and
the expense of the following processes decrease. The
images taken by the photogrammetric camera were not
mosaiced because of the high amount of working space
needed at the computer.
Imagery was spectral classified with the Maximum-Likeli-
hood method according to the founded macrophytic
societies. The satellite images also were spectrally ana-
lyzed. Only the first 3 bands of the Landsat scene were
used for the classification. A merge of Landsat and Spot
increased the geometrical resolution. The Spot scene
from 1995 was used to produce a binary mask with the
regions of the macrophytes.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
