the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Jun 2026
Global controls on aftershock productivity: roles of magnitude and tectonic setting
Abstract. Aftershock productivity varies substantially among large earthquakes, yet the relative roles of magnitude, tectonic environment, and source properties remain incompletely resolved at the global scale. Here we perform a comprehensive, data-driven analysis of aftershock productivity for all Mw ≥ 7.0 mainshocks worldwide since 1976, using a homogeneous workflow based on USGS/NEIC catalogs, PB2002 plate-boundary geometries, and uniformly defined aftershock windows (r ≤ 200 km, t ≤ 30 days, M ≥ 4.5). Mainshock events are classified into three tectonic domains—subduction, other plate boundaries, and intraplate/crustal—allowing systematic comparison of productivity patterns across geodynamic settings. Across all tectonic classes, aftershock productivity scales linearly with mainshock magnitude in log-space, with nearly identical slopes (b ≈ 0.76–0.81), indicating a universal magnitude-controlled triggering mechanism. In contrast, intercepts differ significantly among tectonic environments: subduction earthquakes are systematically the most productive, followed by other plate-boundary events and intraplate earthquakes. Depth and distance to the nearest plate boundary exert weaker but physically interpretable secondary influences, with productivity decreasing for deeper ruptures and for intraplate events located far from plate boundaries. Residual analysis and multivariate correlations confirm that magnitude and tectonic setting explain the dominant share of global variability, whereas depth and distance provide secondary, context-dependent modulation. Together, these results establish the first unified, physically interpretable global framework of aftershock productivity, quantifying both universal magnitude-driven scaling and environment-specific productivity offsets. The findings provide actionable constraints for global ETAS parameterization, operational aftershock forecasting, and seismic-hazard models for future great earthquakes.
- Preprint
(3524 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 09 Sep 2026)
- CC1: 'Comment on egusphere-2026-877', Dr. Ranjit Das, 01 Jul 2026 reply
-
RC1: 'Comment on egusphere-2026-877', Anonymous Referee #1, 29 Jul 2026
reply
The manuscript analyses aftershock productivity of global Mw≥7 earthquakes and its dependence on mainshock magnitude, tectonic setting, depth, and distance to plate boundary. The topic is interesting and the manuscript is clearly written.
However, I have several major concerns:
- The author claims that this is the first study of this kind. However, Dascher-Cousineau et al. (2020) have done an analysis with a very similar scope. Their work needs to be acknowledged, and the results of the present work needs to put in relation with their findings.
- The author assumes a global magnitude of completeness of 4.5. He claims that previous studies have found this threshold to be valid, without citing those studies. Because this choice is of critical importance, I suggest the author provides clear evidence for the adequacy of this choice, and conducts a sensitivity analysis of the results on the completeness magnitude.
- The effect of short-term aftershock incompleteness (STAI) is not addressed, although it can have major consequences for counting events after a large earthquake. I suggest using the a-positive estimator (van der Elst and Page, 2023) to count aftershocks to account for this issue.
- Generally, the possibility that certain observed effects might be an artifact of incompleteness is not addressed. Many of the observations, for example, a decline in productivity with depth, could also be a consequence of increasing completeness magnitude with depth. It is important to acknowledge and address these possibilities.
- It is not distinguished between primary and secondary aftershocks. The author acknowledges this. However, what is not acknowledged is that varying b-values across tectonic settings could lead to varying aftershock numbers, even if the direct aftershock productivity is independent of tectonic setting. Schorlemmer et al. (2005) found that thrust events (which are most prevalent in subduction zones) have low b-values. So, even if direct aftershock number depends solely on mainshock magnitude, the lower b-value could lead to larger aftershock cascade in subduction zones. I suggest that this effect is separated from the effects studied in this article.
- Many claims made throughout the manuscript are not backed by the reported statistics. E.g., According to the a-values and reported uncertainties thereof, the there is no significant difference between estimated productivity offset of different tectonic regimes. The LOWESS trend lines are given without any uncertainties, and some claims about them are made without any assessment of whether this trend could also be due to random fluctuations due to finite sample size.
- In general, I appreciate that many statistical tests were performed. However, I strongly suggest that the author reorganizes the manuscript in such a way that it becomes clear which results support which conclusion, and only conclusions backed by statistically significant results remain in the manuscript.
Below are line-by-line comments:
Line 18: “(b ≈ 0.76–0.81)” if this is the productivity slope, it shouldn’t be called b. Confusion with the Gutenberg-Richter b-value.
Line 26: “first” is not correct. Dascher-Cousineau did a very similar study.
Lines 59-68: Dascher-Cousineau et al. conducted a global study. Their work is highly relevant and should be discussed in this introduction and results compared to theirs.
Line 75: “However, a systematic, quantitative comparison of aftershock productivity across tectonic environments at the global scale has been lacking”, This is not true. See previous comment.
Line 89: “near-global completeness above Mw 4.5”. to be confirmed – I have seen much higher estimates.
Line 93: “operational and observation-based sense” What does this mean?
Lines 96 – 110: Several potential problems. With this definition of aftershock counts, sequences with more than one M≥7 events might create overlapping data. Also, If primary and secondary aftershocks are not distinguished, what is being captured might be differences in b-value that are known to vary with tectonic regime. E.g., lower b-value leads to larger primary aftershocks, which trigger more secondary aftershocks, without productivity actually varying.
Lines 134-136: “Previous assessments of global catalogs indicate that events of this size are reliably detected worldwide since the mid-1970s, including oceanic and remote regions”. Please cite those studies. Nandan et al. (2021) found an Mc of 5.0 for global seismicity. Given that incompleteness can lead to a multitude of artifacts, it is crucial that a rather conservative estimate is chosen.
Line 136: “As a result, the completeness of the aftershock catalog is not expected to vary significantly over time at the selected magnitude threshold.” There is a very well-known effect of short-term incompleteness, that affects the completeness of a catalog, especially after large events.
Line 144: “Previous global analyses” please cite!
Line 149: “three partially overlapping time periods (1976–1990, 1991–2005, and 2006–2024)” these do not seem to be overlapping, please clarify.
Lines 151-153: “This confirms..” This only confirms that completeness above M4.5 has not changed considerably over time. Short-term incompleteness can still cause major biases, just that these biases remain relatively constant over time. Also, completeness may vary regionally, but this variation could persist over time.
Line 185: “the objective of this study is to characterize the full cascading productivity” see comment about b-value.
Line 200: “of a preceding larger mainshock” What if they occur within the aftershock window of a preceding smaller mainshock?
Line 204: “each aftershock sequence is uniquely associated with a single causative mainshock” It could happen that another M≥7 aftershock happens after more than 30 days. In that case, it aftershocks of the first main event would be falsely associated with the second one.
Eq. 3: Parameters a and b are typically used for the magnitude-frequency distribution. Productivity is typically expressed as K * 10^(alpha * M). I suggest adhering to the typical notation to avoid confusion.
Line 273: “Given the limited sample size and the intrinsic variability of aftershock sequences, p-values are interpreted cautiously” What does this mean? The idea of a p-value is to quantify how significant a result is, given the sample size.
Figure 1: A lot of events are classified as “intraplate/crustal events”, but seem to be quite close to subducion boundaries (e.g., Alaska, Southern Chile). How sensitive is this classification to the distance threshold (200km)?
Figure 2: In the legend, please indicate the estimated fit (e.g., a and b, including uncertainty).
Figure 3: Include uncertainties in the comparison.
Line 313: “with limited scatter at the highest magnitudes (Mw ≥ 8.5)” I suppose it just looks that way due to fewer events having large magnitudes. I suggest removing this remark or providing quantitative proof.
Line 314: “broadly similar slopes (b ≈ 0.76–0.81)”. Please indicate which event type has which slope and what the associated uncertainties are.
Line 315: “while intercepts differ”. According to Table 1, a values are within each others standard error range. Please acknowledge that the difference is not significant.
Lines 320-322: Productivity decrease with depth could also be due to incompleteness increasing with depth.
Line 223-325: This is expected due to the definition of the event classes.
Line 235: “Only the intraplate class shows a weak but systematic decrease in productivity at intermediate distances”. Without unceratinties, it is impossible to judge whether this decrease is systematic.
Line 326: “This suggests that continental intraplate stress regimes may influence rupture fragmentation and secondary triggering”. Please clarify how this follows.
General minor comment: Since it is a single-author paper, I believe the text should be written in “I”-form, instead of “we”-form.
References:
Dascher‐Cousineau, K., Brodsky, E. E., Lay, T., & Goebel, T. H. (2020). What controls variations in aftershock productivity?. Journal of Geophysical Research: Solid Earth, 125(2), e2019JB018111.
Nandan, S., Ram, S. K., Ouillon, G., & Sornette, D. (2021). Is seismicity operating at a critical point?. Physical Review Letters, 126(12), 128501.
Van Der Elst, N. J., & Page, M. T. (2023). a‐positive: A robust estimator of the earthquake rate in incomplete or saturated catalogs. Journal of Geophysical Research: Solid Earth, 128(10), e2023JB027089.
Schorlemmer, D., Wiemer, S., & Wyss, M. (2005). Variations in earthquake-size distribution across different stress regimes. Nature, 437(7058), 539-542.
Citation: https://doi.org/10.5194/egusphere-2026-877-RC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 77 | 30 | 6 | 113 | 3 | 4 |
- HTML: 77
- PDF: 30
- XML: 6
- Total: 113
- BibTeX: 3
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
1. The manuscript uses the M scale throughout the analysis . Recent studies have questioned the use of the M scale for earthquakes below M 7.5, arguing that the extension of the Hanks and Kanamori (1979) formulation to smaller magnitudes is based on Equation (1) of Purcaru and Berckhemer (1978) which was originally developed for earthquakes up to Ms 7.0 (Das et al., 2025; Das and Das, 2026). Furthermore, Matsu'ura (2025) discussed theoretical limitations of the M scale. These studies are not considered in the manuscript. Since earthquake magnitude is the primary variable controlling the statistical analyses presented in this study, the authors should justify their use of the M scale and discuss whether alternative magnitude formulations, such as the Mwg (Das magnitude) scale, would influence the reported relationships and conclusions.
2. A revised magnitude scale (Mwg or Das magnitude scale) has recently been proposed to address the reported limitations of the M scale using a global earthquake dataset (Das et al., 2025). The proposed scale has been shown to exhibit a closer relationship with radiated seismic energy and existing magnitude scales. Although Gasperini and Lolli (2025) criticized the Mwg scale, more recent studies have challenged several of their main conclusions (Paul et al, 2025; Das et al, 2025; Das and Das 2026). Since earthquake magnitude is the fundamental parameter in this study, the authors are encouraged to compare the results obtained using both the Mw and Mwg scales and discuss whether the choice of magnitude scale affects the regression models, statistical relationships, and overall conclusions.
References
1) https://www.nature.com/articles/s41598-025-06642-1
2) https://link.springer.com/article/10.1186/s40623-025-02278-7
3) https://link.springer.com/article/10.1007/s40098-024-01147-6
4) https://link.springer.com/article/10.1007/s12046-025-02769-z
5) https://www.sciencedirect.com/science/article/pii/0040195178901774