Subsurface Deformations due to Induced Seismicity

INSIDE: Using a combined approach of InSAR, GNSS and levelling data to monitor deformations in the subsurface caused by induced seismicity from geothermal drillings in the southern Munich area.
Time period of the project:

- September 2019 - August 2022


Main goals of the project:

- Monitoring induced seismicity and associated subsurface deformation connected to geothermal drillings.

- Development of a recommended approach for a seismic-geodetic monitoring and modelling system for geothermal energy projects.

- Development of a ‘semi real-time’ monitoring approach based on Sentinel-1 acquisitions.


Geothermal energy has been on the rise for some time now as an environmentally friendly alternative to current fossil fuel-based solutions for heating and warm water production. It involves the extraction of groundwater, the exchange of heat and the subsequent reinjection of the water into the subsurface. Although these geothermal wells operate at relatively great depths, the continuous extraction and injection of groundwater are sources of induced seismic activities in surrounding area of the wells.

In the framework of the INSIDE (Induced Seismicity and Soil Deformation as Aspects of Interference in the Operation of Geothermal Plants in the Southern German Molasse Basin) project, we deal with the seismic and geodetic monitoring of these induced seismic events in the southern region of Munich. The goal of the project is not only to determine their magnitude and the connected surface deformations, but also to develop a recommended approach of a combined seismic-geodetic monitoring and modelling system for geothermal energy operations. The group of participants in the project comprises two KIT institutes: the Geodätisches Institut Karlsruhe (GIK) and the Institut für Angewandte Geowissenschaften (AGW); and three industrial partners: Innovative Energie für Pullach GmbH, Stadtwerke München GmbH, and Erdwerk GmbH.

The seismic monitoring and modelling is carried out by the AGW using seismic and noise measuring equipment installed on the surface and in the wells. For the geodetic measurements, we at the GIK are using a combination of levelling, GNSS and InSAR data. Altogether four permanent stations are set up at key locations, equipped with a collocated GNSS receiver, a passive or active SAR corner reflector and a levelling benchmark. Not only do these stations serve as data collection points at places where precisely localized measurements are essential, but they are also connection points between the three data sources.

Sentinel-1 and TerraSAR-X acquisitions are processed to extract the deformation history of points with stable phase information in the area of interest. In order to split the line-of-sight deformations into horizontal and vertical components, we utilize and combine data from both ascending and descending tracks. In the case of the Sentinel-1 data, the region of interest is covered by two adjacent tracks in both the ascending and the descending directions, the combination of which can lead to an increased number of PS points due to the difference in the incidence angles.

In order to further increase the number of extracted scatterers with useful information, apart from the Persistent Scatterer (PS) approach, the Distributed Scatterer (DS) processing is also applied to the dataset. In contrast to the point-like Persistent Scatterers, Distributed Scatterers comprise of ensembles of elementary backscatteres that each contribute to the radar echo, but as a whole, produce a stable long-term phase signal. Depending on the topography of the region of interest, the combination of the two approaches may lead to an increased number of useful scatterers, compared to only using the PS processing.

As the frame of the project covers the time period until mid-2022, it is necessary to continuously update the initial results of the InSAR processing in order to achieve ‘real-time’ monitoring as much as possible using new radar acquisitions. Because the reprocessing of the complete stack every time a new acquisition is added can be overly time consuming, one focus of the project is to develop an efficient approach for introducing new scenes into the already processed data stack.

M.Sc. Bence Ambrus