cases the vegetation states obviously depend only upon
annual variations.
Due to the annual variations of temperature multitempo-
ral analysis of thermal data can only be made by means
of temperature differences. For this purpose the radiation
signals measured by the TIR scanner were first transfor-
med into temperatures values (brightness temperature of
the black body) and then referenced to an average valu-
ewhich is valid during the acquisition time. The relative
difference to this average mean value was classified into
six categories (Fig. 8a,b). The results of the thermal data
analyses can be compared with the vitality features and
give additional indications to the plant state. As a rule
one can conclude that, the cooler a vegetation covered
area is compared to its surroundings, the more dense
and more vital it is. Concerning the alfalfa fields the diffe-
rent soil treatment measures have also to be considered.
Based on the analysis of statistical distribution parame-
ters, e.g. Pearson's coefficient of variation, conclusions
on the homogenous development of the primary seeded
areas can be derived. To this end both the variances of
the vitality features, referred to the corresponding area,
and the variances of the thermal and SAR data were
computed. Areas 9,10,11 and 14 turned out to be espe-
cially inhomogenous. Corresponding fieldwork confirmed
that just those areas suffered indeed from greater zones
of deficiency or, in case of the pine reforestations, where
sea buckthorn and robinia settled.
The SAR signatures shown in the plot for 1994 represent
mean values per area of the relative backscatter coeffi-
cients which were obtained by the airborne SAR
TRAVERS during the PRIRODA aircraft campaign 1994
(Marek et al., 1994). The SAR sensor operates at a wave-
length of 10cm (S-band). From an alitude of 6000m it
takes a swath width of 20km realizing a spatial resolution
of approximately 20m by 20m. Unlike the thermal data
the measured backscatter coefficients show only a slight
correlation to the vitality coefficients. A strong relations-
hip however is manifested to geometric plant features like
plant growth height. For example, area 4 with weak
5450
5452
5454
5718)
5714
5714}
45712
54
sda
Legende
mn 0-1 em1-2 292-3 e223-4 4-5
l
2km
Bezug Gau&-Kruger.
Ellipsoid Bessel
Fig. 9: Unsupervised classification of radar signatures of
the recultivated areas of the Welzow-Süd dump
774
backscattering was covered by plants with high vitality
values but with plant heights of only 6-10cm. On the
other hand, area 1 and 2 showed plants reaching heights
of 40 to 50cm. The same is valid for the reforested areas.
The highest backscatter values can be found on older
trees with heights of 3m and more. Therefore a good
discrimination is given for the unvegetated areas, which
have been assigned to class 0-1 with lowest backscatter
values. Based on that, additional conclusions, which
have been independently derived compared to those
ones on the vitality in the VIS region, can be drawn to
describe the development of vegetation during the recul-
tivation process. As a final step these results are
transformed into a GIS where complex analysis with data
from other sources can be made.
5 SUMMARY
Based on the presented complex analyses of remotely
sensed data acquired at different times, monitoring tasks
in ecological impact regions, like post mining landsca-
pes, can be performed. The normalization of data and
separation of the required information from other distur-
bing influences has turned out to be a basic problem.
This can be overcome only by the use of multisensor
data containing all wavelength regions which are
available to remote sensing and by integration of ground
truth based auxillary information.
The work presented here have been partially sponsored
by the German Space Administration (DARA) and the
Federal Ministry for Education, Science, Research and
Technology (BMBF).
6 REFERENCES
Ahern, J. F, Leckie; D. J., Drieman, J. A.,1993. Seasonal
Changes in Relative C-Band Backscatter of Northern
Forest Cover Types. IEEE Transactions on geo-
science and remote sensing, Vol.31, No.3, pp. 668-
680.
Clevers, J.G.P.W., van Leeuwen, H.J.C., Verhoef, W.,
1992. Estimating APAR by Means of Vegetation In-
dices: A Sensitivity Analysis. In: International
Archives of Photogrammetry and Remote Sensing,
Washington, D.C., Vol. XXIX, Part B7, pp. 691-698.
Marek, K.-H., 1992. Application of remote sensing data
for the investigation of environmental degradation si-
tes in East Germany. In: Proceedings of the Central
Symposium of the ‘International Space Year’ Confe-
rence, Munich, Germany, Vol Il, pp. 465-469.
Marek, K.-H., Weichelt, H., Kutuza, B., et al., 1994. Multi-
Sensor Airborne Campaigns PRIRODA 1992-96 for
Environmental Monitoring. In: Proceedings of the First
Internat. Airborne Remote Sensing Conference,
Strasbourg, France, Vol |, pp. 317-328
Qi, J., Chehbouni, A., Huete, A. R., et al., 1994. Modified
Soil Adjusted Vegetation Index (MSAVI). Remote
Sensing of Environment, 48, pp. 119-126.
Weichelt, H., Herr, W., 1987. Zur Vorverarbeitung mul-
tispektraler Daten. Vermessungstechnik, 33, pp. 270-
272:
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
