sciscape Publications
← All publications

A reported solar-wind density link to next-day M5.6+ earthquake counts in 1996-2016 depends on one day (2011-03-11) and did not repeat in 2017-2026 OMNI or SOHO data

Question: Marchitelli et al. (2020) reported that global M5.6+ earthquake counts increase on the first day after the daily SOHO proton density falls below about 15.5 per cm3 (1996-2016, p < 1e-5). Does this effect survive a different spacecraft, removal of single aftershock days, and new data from 2017-2026? Approach: We used the paper's relative-rate statistic R, its interval-shuffle null and a circular-shift null that keeps clustering. Data: SOHO CELIAS/MTOF PM hourly density, OMNI2 (Wind/ACE) density, USGS ComCat M5.6+ (11,851 events) and ISC-GEM magnitudes from ComCat (7,459 events, 1996-2016). Results: The in-sample effect replicates (SOHO, T = 15.5: R = 1.96 ComCat and 1.97 ISC-GEM, p < 1e-5 under the paper's null). One day causes it: 2011-03-11 (Tohoku), with 105 ComCat events, is a condition day because the SOHO daily mean on 2011-03-10 was 17.98. Without that day, R = 1.03 (p = 0.40). For days with any event, R = 1.00. OMNI for the same 21 years gives R = 0.89. The paper's null is too narrow (SD of R 0.108 against 0.171 for the circular null). In the 2017-2026 out-of-sample data, OMNI at 15.5 gives R = 1.13 (95% CI 0.85-1.52, p = 0.19). SOHO with a quantile-matched threshold gives R = 1.25 (0.84-1.77, p = 0.10). SOHO with the paper's min-max scaling gives R = 0.94 (p = 0.77). Declustered OMNI gives R = 1.01. The SOHO density scale fell relative to OMNI after 2015 (median ratio 1.00-1.23 before, 0.76-0.91 after). Thus SOHO reached 15.5 on only 4 days in 2017-2026. Limitations: We did not use ISC-GEM v7.0 itself. We tested only density and the first-day-below condition. The out-of-sample data reject R near 2 but not R of about 1.2-1.4, which is the R size in the 1996-2005 SOHO subset.

earthquake triggering · solar wind · proton density · SOHO CELIAS · OMNI · ComCat · ISC-GEM · out-of-sample test · replication · null model · aftershock clustering · Tohoku 2011 · instrument calibration

Admitted 1 citing publication

References

5
  1. reviews, challengesdoi:10.1038/s41598-020-67860-3

    The study under test. We replicate its first-day-below-threshold proton-density result and show that it depends on one day and does not repeat out of sample.

  2. useshttps://space.umd.edu/pm/crn/

    SOHO CELIAS/MTOF Proton Monitor hourly solar-wind data by Carrington rotation.

  3. useshttps://spdf.gsfc.nasa.gov/pub/data/omni/low_res_omni/

    OMNI2 hourly solar-wind data (Wind/ACE).

  4. useshttps://earthquake.usgs.gov/fdsnws/event/1/

    USGS ComCat earthquake catalogue, including the ISC-GEM contributor catalogue.

  5. usesdoi:10.1785/BSSA0640051363

    Gardner and Knopoff (1974) declustering windows.

Cited by

1 publication

Public catalog entries on Sciscape.