© since 1995, Sigrid Hehl, Eckart Lange, IRL-Institut, ETHZürich


Visualizing virtual brown coal landscapes by merging GIS andRemote Sensing Data


Ministry for the Environment, Nature Protection, and SpatialPlanning, Brandenburg, Germany
Lausitzer und Mitteldeutsche Bergbauverwaltungsgesellschaft(LMBV)
Lausitzer Braunkohle AG (LAUBAG)

Geographically the East German brown coal deposits which are thelargest in Germany are located in the Halle / Leipzig / Bitterfeldregion and in the Lausitz region near the Polish border. In theLausitz region the brown coal seam reaches an average thickness of 12m. The seam is covered by an overlay shelf of 40 m - 107 m inthickness which can be removed by gigantic excavators. With 650 m inlength the excavators used in the Lausitz region are the largestmovable man-made structures on earth. Up to 60 m in depth can beremoved in one step.
Because of the decreasing demand for brown coal most of the surfacemining activities will not be continued in the future.

Effects on the Environment
Large parts of former East Germany are shaped through brown coalsurface mining activities. The most dramatic effects on theenvironment are related to groundwater, air quality, and the visualquality of the landscape.
Naturally the landscape is an agricultural landscape with largeforested sections. During the four phases of the surface miningprocess (exploration, development, production, and reclamation) therelatively flat landscape is transformed into a landscape ofcompletely different character. In the case of the pilot study'Jänschwalde' the zone of active mining is moving north leavingan area of approximately 24 km2 in the south to be reclaimed. In thissection there are two very large depressions (90 m relative to thesurroundings) of 4 km respectively 2 km in length and 800 m in width[place here fig. 1]. Extensive pumping of groundwater of upto 170 m3 per minute prevent the holes from being flooded. Thereclamation involves massive terrain alterations for stability andsafety reasons. Because of the highly acid soil conditions (pH 3 orless is possible) after the terrain modelling also soil improvementmeasures are essential as a prerequisite for further land use likeforestry or agriculture. Reclamation of such large areas take a longtime. By the year 2010 the smaller one of the two depressions is tobe filled in and the larger one to be flooded by the natural rise ofthe groundwater level. In 2030 the whole area will be reclaimed.

Major Goals of the Pilot Study
Major goals of the pilot study 'Dynamic visual simulation -reclamation surface mining site Jänschwalde' (HEHL-LANGE &LANGE 1996) are to explore ways how existing data (terrrain data,satellite imagery) can be utilized and how landscape change can bevisualized in three dimensions over time. The visualizations shall beused as the basis for the design of the new landscape and as a meansof communication among the numerous involved institutions.

Our Approach
GIS - systems are mostly 2D oriented, whereas the strength ofvisualization systems is due to the capability of dealing with the3rd dimension. In the case of the brown coal pilot study only theterrain data which is needed in the mining and reclamation process,is captured in three dimensions. Most other data is in 2D format.Automatic procedures for generating 3D objects from 2D data, i.e.having the 3D modelling based on numeric data from the GIS database,can bridge the gap between the mainly 2D GIS systems and 3Dvisualization systems (HOINKES & LANGE 1995). In the presentedexample (using Polytrim software from CLR University of Toronto)3D-objects like trees and buildings can be generated directly throughlargely automated procedures. In order to perform this taskattributed 2D data is needed, e.g. containing the height of abuilding structure, which is used as the basis for the creation ofthe 3D-objects which are then set on the terrain. Using thistechnique a unification of various (partly) existing data sourceslike photogrammetric data acquired by the mining companies, satelliteimagery, and also an integration of 2D reclamation plans and dataderived from topographic maps into a virtual landscape model isachieved.
The virtual landscape consists of a digital terrain model at 10 mresolution with a draped satellite image and 3D-Objects like singletrees, forests and buildings. Image data was available from RussianKFA-1000 (5-7 m resolution) and Russian KWR - 1000 (max. 1.5 mresolution, B/W) satellites as well as LANDSAT TM (30 m resolution)imagery. Most appropriate were the KFA-1000 images. Although thecolor range lies in the near infrared range, the appearance wasimproved through filtering.

Visualizing Change
Visualization is also the common denominator for incorporatingthe factor time into the model through the representation of thedifferent development stages between 1940 - 1995 - 2010 - 2030.Additionally two design alternatives for the future are worked out.Through actively designing vistas by altering the terrain and thevegetation scheme instead of plain afforestations the potential forrecreation is improved.
As there is no satellite imagery available dating that far back orahead the existing scene was registered and overlaid with digitizedland use information covering the past and the future. Using imageprocessing techniques geotypical textures were applied according tothe different time phases.
As a result several animations (fly-through, driving car, pedestrian)are produced which allow the comparison of the different time phasesand different design alternatives. The animations are recorded onvideo at a rate of 50 frames / second totalling 10 minutes ofanimated sequences corresponding to 40 GB of calculated images.

The visualizations which can be generated based on the virtualmodel are an essential means of communication among the variousinvolved experts from different disciplines and the general public.By providing a common basis for discussion the decision makingprocess can be optimized (see e.g. LANGE 1994).
Compared with an interactive presentation using a computer, video isa relatively static medium where the observer has to follow apredefined animation path. In order to increase the degree of realismin future applications high resolution aerial orthophotography shouldbe used (see e.g. KERSTEN & O'SULLIVAN 1996).
Although the presented results where appreciated highly by theinvolved ministry and the coal mining companies there is a fairdanger that other than this study not much more will happen in thefuture. The size of these institutions and the widely distributedresponsibilities could possibly lead to an inability of theseinstitutions towards an incorporation of new planning techniques.Effectively applying these 3D visualization techniques would requirea different understanding of the planning process.



Link to some more figures(AGIT '97-Posters)


Back to Eckart Lange


Back to Sigrid Hehl-Lange