act 
ee 
ast 
zil- 
4d 
be 
ns 
in- 
ler 
jer 
en 
ble 
tic 
hs 
ire 
on 
ite 
om 
ing 
Iso 
  
Fig. 1: Overview of the opencast lignite mining region of Lusatia and investigation areas: 1 - Schlabendorf, 
2 - Welzow-Süd, 3 - Nochten (Landsat-TM, band 5, August, 9, 1992), scale approx. 1 : 450 000 
2 VEGETATION AND LAND USE CHANGES IN 
OPENCAST MINE SURROUNDINGS 
The investigation area of Nochten (see frame 3 in Fig. 1), 
which was named after the active opencast mine, is si- 
tuated in the south-eastern region of Lower Lusatia 
extending about 40km in length and 20km in width. For 
this area a monitoring system is developed using tools of 
digital image processing and GIS, and whose main com- 
ponent consists of annually quantitative and qualitative 
analyses of multispectral satellite imagery. The investi- 
gations have focused on 33.000 hectares full covered 
biotopes mapped by field survey. Defined as habitats for 
plants, animals and biotic communities, they reflect both 
the complex site specific conditions and the anthropoge- 
nic influences. Within the investigation area the latter 
ones have resulted from progessive mining activities and, 
in connection to that, the effect of spatially extending 
ground water lowering. Features of vegetation like the 
predominating species composition create the criteria for 
mapping and categorization into 25 main categories. 
Special aspects in monitoring systems to be developed 
are directed to the detection of vegetation changes 
brought about by the mining industry. In order to deter- 
mine the influence of ground water lowering on the 
habitat development, it is necessary to derive the general 
long-term trend of vegetation development in the whole 
region which is influenced by the annual course of pre- 
cipitation and temperatures. For estimating this trend 
769 
Landsat-TM imagery, available since 1988 rather perma- 
nently, have been used. Additional three years field 
investigations made on reference biotopes, which are 
located beyond the impact zone of the opencast mines, 
shall determine deviations between these biotopes com- 
pared to those ones within the impact zone. 
Since vegetation period 1995 research work on this issue 
have been carried out. The following results presented 
here are concerned with vegetation development 
observed at four acquisition dates during 1988 and 1995. 
For the analysis on wooded areas, which could be 
assumed to be completely covered by vegetation, the 
Normalized Difference Vegetation Index (NDVI) turned 
out to be a suitable measure. Its application on Landsat- 
TM data has been well proved for a long time. The 
biotope categories — nado-callunetea and sedo- 
scleranthetea which can be anticipated with greater parts 
of uncovered soil were treated with the Modified Soil 
Adjusted Vegetation Index (MSAVI2) introduced by Qi et 
al. (1994). First statistical analyses on single biotopes 
have confirmed the conclusions derived by Qi et al., that 
MSAVI2 has a greater dynamic range, thus leading to 
always higher average values compared to the NDVI 
ones for the month considered (August) More special 
investigations on terrestrial reference areas with different 
vegetation cover degrees are planned. The processing of 
the NDVI values derived from pine forest biotopes have 
been realized by scaling the values into nine classes. 
This resulted in spectral feature distributions 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996