the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Deciphering past earthquakes from paleoseismic records – The Paleoseismic EArthquake CHronologies code (PEACH, version 1)
Bruno Pace
Francesco Visini
Joanna Faure Walker
Oona Scotti
Abstract. A key challenge in paleoseismology is constraining the timing and occurrence of past earthquakes to create an earthquake history along faults that can be used for testing or building fault-based seismic hazard assessments. We present a new methodological approach and accompanying code (Paleoseismic EArthquake CHronologies, PEACH) to meet this challenge. By using integration of multi-site paleoseismic records through probabilistic modelling of the event times and an unconditioned correlation, PEACH improves the objectivity of constraining paleoearthquake chronologies along faults, including highly populated records and poorly dated events. Our approach reduces uncertainties in event times and allows increased resolution of the trench records. By extension, the approach can potentially reduce the uncertainties in the estimation of parameters for seismic hazard assessment such as the earthquake recurrence times and coefficient of variation. We test and discuss this methodology in two well-studied cases: the Paganica Fault in Italy and the Wasatch Fault System in the United States.
Octavi Gómez-Novell et al.
Status: open (until 07 Jul 2023)
-
AC1: 'Comment on gmd-2023-71 (error in Eq.1 of the preprint)', Octavi Gomez-Novell, 18 May 2023
reply
We do this comment to highlight and correct a small error we detected on the formulation of equation 1 of the preprint. This equation defines the minprom parameter, used as a prominence threshold to detect probability peaks in the mean distribution (see section 3.2. step III of the preprint). Note that the error is not present on the code provided, but is just restricted to the preprint file.
The mentioned error relies on the ½ division of min P and maxP and comes from a previous version of the equation we worked on before the final version, which did not allow to detect all the peaks. As it is written right now, the equation is mathematically equivalent to minP/maxP. This is not correct because the minprom is used in the mean distribution curve and therefore its value should be at least ½ smaller. In the code, this value is set to ¼ of the minP/maxP to account for PDFs that overlap and therefore generate peaks with prominences smaller than the ½ height. That is, equation 4 should be:
minprom =0.25 ∗ (minP/maxP)
With this, the text explaining the equation (lines 187-191) should be adapted to accommodate this:
“The target parameter to define the minprom is the minimum peak probability (minP) of the event PDFs, namely the PDF with the widest uncertainty of them all (Fig. 4a). Because the mean distribution is normalized, the half probability value is also normalized using the maximum probability of all PDFs (maxP; Fig. 4a). This normalized quotient is then divided by 4 to account not only for the fact that the minprom is detected in the mean distribution (i.e., whose probabilities are ½), but also for the instances in which two different PDFs overlap significantly so that they generate a flattened mean distribution with peaks that have prominences smaller than ½ their maximum probability. See Eq. (1):”
We apologize to the readers for any confusion.
Best regards,
The authors.
Citation: https://doi.org/10.5194/gmd-2023-71-AC1
Octavi Gómez-Novell et al.
Model code and software
Data for "Deciphering past earthquakes from paleoseismic records – The Paleoseismic EArthquake CHronologies code (PEACH, version 1)" Octavi Gómez-Novell, Bruno Pace, Francesco Visini, Joanna Faure Walker, and Oona Scotti https://doi.org/10.5281/zenodo.7783290
Octavi Gómez-Novell et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
179 | 50 | 7 | 236 | 1 | 0 |
- HTML: 179
- PDF: 50
- XML: 7
- Total: 236
- BibTeX: 1
- EndNote: 0
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1