Seismic methods proved to be powerful tools in oil and natural gas exploration and
production projects for a long time. Resolution and accuracy of imaging were improved
by more and more sophisticated algorithms and methods. Three workflows using leading
edge technologies will be presented that address different problems in geothermal
projects.

1. CRS Processing on sparse seismic datasets to improve imaging of low fold seismics
2. Quantified Fault Characterisation using 3D seismics to provide an accurate fault
inventory
3. Seismic Facies Classification to identify lithology from seismic patterns

The Common Reflection Surface (CRS) processing replaces standard DMO processing and
results in improved imaging of dipping events, a better overall signal-to-noise ratio
and event continuity. It proved to be beneficial for imaging fault patterns. In
addition, the method works well on old, low fold datasets, for which many modern
methods like prestack depth migration fail. Therefore it is an ideal processing
approach for seismic data used for identifing geothermal projects.

A method is described how to predict fault attribute in well scale from seismic data.
The basis of this approach is a high resolution coherency processing with advanced
algorithms. Only this ensures the best possible fault inventory which is essential
for such predictions. The manually or automatically picked fault data is converted
into fault attribute maps which are then tied to well information by geostatistical
methods.

Neural networks are used to classify seismic patterns into facies classes. This can
be done target-oriented using a map based algorithm or, in 3D, for the entire volume.
The facies classes are then calibrated to petrophysical meanings by well information.

14473 Potsdam, Telegrafenberg
Germany