Statistical Analysis and Simulation of Conduit Dimensions within Karstic Systems

Yves Frantz and Pauline Collon and Philippe Renard and Sophie Viseur. ( 2019 )
in: {{IAMG}} 2019, {International Association of Mathematical Geosciences}

Abstract

Even if French mainland seismicity is known to be very low to medium, it's still monitored in order to access to a high level fundamental understanding. This understanding is mainly used to anticipate earthquake related risk in probabilistic seismic risk assessment (PSHA) and has a direct impact on the French prevention based policy regarding this hazard. The development of the monitoring network both in terms of quantity and quality had led to a large amount of data summarised in several seismic catalogues. Nevertheless, these huge database are not exhaustive and capture a lot of uncertainties from various sources. Time, magnitude and location of seismic events are particularly affected by these uncertainties. In this study, we propose to process seismic catalogues in order to estimate their uncertainties and extract consistent subset of data considering specific working hypotheses. Because of sensors sensitivity, French seismic network, as others, is not able to record all the earthquakes. This is particularly true for small events. That's why seismic hazard assessment often deals with cut-off magnitude and completeness of seismic catalogue. This cut-off magnitude describes the magnitude from which all the happened events are assumed to be recorded. Restricting data to earthquakes with magnitudes greater than the cut-off one can be seen as a loss of information. However, this rejected information is biased whereas each earthquake in the cut-off catalogue is considered as an unbiased information. In most cases, this cut-off magnitude is arbitrary chosen and the period of observation is systemically assumed to be reliable. Nevertheless, one can understand that, due to seismic network improvements, the period of observation may have an important impact on the cut-off magnitude finding strategy. We propose a methodology to find the best pair [cut-off magnitude \textendash{} period of observation] among a French earthquake catalogue. The study also considers the large heterogeneities on seismic level (from very low to medium) and on spatial distribution of seismic network by using a seismotectonic zoning in order to enhance the stationarity of sub catalogues. Results consistency is discussed regarding the classical scale laws used to characterize seismic activities (Gutenberg and Richter law, inter-event times and correlation dimensions).

Download / Links

BibTeX Reference

@INPROCEEDINGS{FrantzIAMG2019,
    author = { Frantz, Yves and Collon, Pauline and Renard, Philippe and Viseur, Sophie },
     title = { Statistical Analysis and Simulation of Conduit Dimensions within Karstic Systems },
     month = { "aug" },
 booktitle = { {{IAMG}} 2019 },
      year = { 2019 },
organization = { {International Association of Mathematical Geosciences} },
       url = { https://hal.univ-lorraine.fr/hal-02186230 },
  abstract = { Even if French mainland seismicity is known to be very low to medium, it's still monitored in order to access to a high level fundamental understanding. This understanding is mainly used to anticipate earthquake related risk in probabilistic seismic risk assessment (PSHA) and has a direct impact on the French prevention based policy regarding this hazard. The development of the monitoring network both in terms of quantity and quality had led to a large amount of data summarised in several seismic catalogues. Nevertheless, these huge database are not exhaustive and capture a lot of uncertainties from various sources. Time, magnitude and location of seismic events are particularly affected by these uncertainties. In this study, we propose to process seismic catalogues in order to estimate their uncertainties and extract consistent subset of data considering specific working hypotheses. Because of sensors sensitivity, French seismic network, as others, is not able to record all the earthquakes. This is particularly true for small events. That's why seismic hazard assessment often deals with cut-off magnitude and completeness of seismic catalogue. This cut-off magnitude describes the magnitude from which all the happened events are assumed to be recorded. Restricting data to earthquakes with magnitudes greater than the cut-off one can be seen as a loss of information. However, this rejected information is biased whereas each earthquake in the cut-off catalogue is considered as an unbiased information. In most cases, this cut-off magnitude is arbitrary chosen and the period of observation is systemically assumed to be reliable. Nevertheless, one can understand that, due to seismic network improvements, the period of observation may have an important impact on the cut-off magnitude finding strategy. We propose a methodology to find the best pair [cut-off magnitude \textendash{} period of observation] among a French earthquake catalogue. The study also considers the large heterogeneities on seismic level (from very low to medium) and on spatial distribution of seismic network by using a seismotectonic zoning in order to enhance the stationarity of sub catalogues. Results consistency is discussed regarding the classical scale laws used to characterize seismic activities (Gutenberg and Richter law, inter-event times and correlation dimensions). }
}