Geoelectric properties: temporal change as failure indicator

Landslides are one of the major natural threats to human lives, settlements and infrastructure, causing enormous human suffering and property losses. As resumed by the SafeLand 7th FP European Project, Europe had experienced the second highest number of fatalities and the highest economic losses caused by landslides compared to other continents: 16,000 people have lost their lives because of landslides and the material losses amounted to over USD 1700 M during the 20th century. Furthermore, the number of people affected by landslides is much larger than reported: in Italy, while about 500 people have been killed by landslides over the past 25 years, the total number of persons concerned is 50 times that number. The best way to restrain such high losses on property and lives is through effective land-use planning, based especially on a good knowledge of the landslide susceptibility, hazard and risk within specific areas as a part of mitigation. However, this ideal approach is, due to several natural, historical or political reasons in many places impossible. E.g., many human settlements and infrastructure lines have already existed in landslide-prone areas or on dormant landslide bodies decades before the establishment of detailed hazard zone maps. In most cases it is not possible to resettle people living in such areas. Consequently it is the goal of stakeholders to guarantee a safe daily life of the people concerned. For this purpose, reliable early warning systems are needed. The most commonly used early warning parameters are pore pressure and displacement. However recent research has shown that other parameters exist, which might give indications on impending triggering even a longer time before an actual displacement is measureable. This is exactly where this project wants to contribute. The main goal of the TEMPEL project is to evaluate temporal changes of geoelectrical properties of the subsurface as possible indicator of future failure of high risk landslides; such additional indicators would be beneficial to any effective early warning system. To arrive at that goal, the TEMPEL project will follow a two-fold approach. First it will improve the technology of geoelectric data acquisition and data inversion for monitoring applications significantly. This work is based on the experience gathered and on the GEOMON 4D technology developed at the Geological Survey of Austria within the last 8 years and of the expert knowledge of Jung-Ho Kim, South Korean specialist for 4D time lapse data inversion. The combined expertise of both groups is expected to generate significant benefits for TEMPEL project. On the other hand long-term monitoring studies, involving recording of geoelectric, displacement and hydrological data with a high sample interval (at least once per hour) on several landslides with different characteristics will be performed to allow a complex correlation of permanently recorded data sets of geoelectric parameters with displacement and hydrological data. Finally a comprehensive evaluation of the practical applicability of the geoelectric technology for long-time landslide monitoring and early warning will be performed.

The Geological Survey of Austria/Department of Geophysics conducted the research project TEMPEL in the years 2011 to 2014. The main objective of the project was the evaluation of the geoelectrical method in terms of a monitoring tool, for a better understanding of dynamic processes at active landslides. For this purpose, several appropriate locations were equipped with a geoelectrical monitoring system. Because of the importance of additional data (displacement, groundwater table, etc.) locations with already existing monitoring infrastructure were preferred. However, for a sufficient database at some sites the installation of additional monitoring systems, e.g. for displacement, soil humidity, soil temperature, ground water table, was necessary. The interpretation of the large amount of collected data (geoelectrical as well as additional data) should be the basis for an improved knowledge about the overall status of the particular landslide body and respectively its change during phases of reactivation. For the implementation of this objective, some technical improvements of the existing geoelectrical monitoring system (Geomon4D, power supply, etc.) as well as significant developments in the field of data processing (quality control, filtering, inversion) were necessary. Several international cooperation partners supported the project concerning infrastructure, logistics and data exchange. For the improvement of existing processing routines, which were required for long term monitoring data, a close collaboration with the Korea Institute of Geoscience and Mineral Resources (KIGAM) was intended. For this purpose, each project year included a 3 months working stay of Dr. Jung- Ho Kim, an expert in this field of research, who worked as a visiting researcher at the Geological Survey of Austria. He developed an innovative 4D inversion software enabling detailed analysis of resistivity changes of the subsurface for long periods, which is essential for the interpretation of geoelectrical monitoring data. In the course of the project, geoelectrical monitoring was performed in total at 10 different landslide locations. Depending on local conditions (evolution of the landslide, data quality, infrastructure, etc.) the monitoring periods vary between several months and four years. Based on our data we can state that in some cases, even in the period of four years, there is no certainty to record a sufficient number of reactivation phases of the landslide, to ensure high interpretation reliability. This fact shows the importance of long periods of data collection. At least five monitoring sites will be continued in follow up projects to increase the significance of the interpretation. However, based on the recorded monitoring data we can conclude that geoelectrical monitoring data contribute significantly to an improved interpretation of ongoing processes in landslide areas. The data clearly showed, that almost in all cases water infiltration was the controlling factor for displacement, but the infiltration process itself was highly dependent on local subsurface properties, preconditions (water saturation), the intensity of precipitation (can be also snowmelt) and seasonal variations. Based on the development of the new inversion software, for the first time clear differences in the infiltration processes could be proven with geoelectrical monitoring data, even if comparable water input to the subsurface appears. At the same time, the monitored displacement data showed an obvious dependence on the characteristic of water infiltration. Therefore, the results of the geoelectrical monitoring could give a rough estimation of the conditions which cause an increased probability of high displacement rates. In contrast to commonly used monitoring methods (inclinometer, tensiometer, GPS), which provide only point information, the geoelectrical method gives spatially distributed information on the subsurface. Therefore, geoelectrical monitoring in combination with conventional methods turned out to be a very useful tool to contribute to an improved interpretation of landslide triggering processes.