Pith. sign in

REVIEW 4 major objections 6 minor 4 references

A Multiscaling Fingerprint of Earthquake Diffusion in Seismic Swarms

T0 review · 4 major / 6 minor · reviewed 2026-07-14 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Earthquake swarms share a multiscaling signature that separates their migration from ordinary tectonic sequences.

desk verdict Solid empirical multiscaling contrast for ten consensus swarms vs Landers, but the “fingerprint” claim is one tectonic baseline short of a class diagnostic. read the letter →

arxiv 2607.11142 v1 pith:46T2E6VA submitted 2026-07-13 physics.geo-ph stat.AP

Cataldo Godano , Giuseppe Petrillo This is my paper
classification physics.geo-phstat.AP
keywords earthquakeswarmsanomalousdiffusionmultiscalingspectrumintereventdistancesseismicmigrationSouthernCaliforniatectonicsequences
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Seismic swarms migrate in space, but ordinary single-exponent diffusion models have not captured how that migration works. This paper identifies ten persistent swarms in the relocated Southern California catalogue and tracks the statistical moments of distances between events as time lag grows. In every swarm the multiscaling spectrum is roughly linear for negative moments (short separations) yet systematically nonlinear for positive moments (long separations), meaning small and large interevent distances obey different effective scaling laws—strong anomalous diffusion. The Landers tectonic sequence instead saturates at high moments, as if epicenters remain spatially caged. The authors therefore present the multiscaling spectrum itself as a quantitative fingerprint that can distinguish swarm-style migration from conventional tectonic sequences, giving catalogue-based work a clearer statistical handle on fluid-driven or transient processes.

What carries the argument

The multiscaling spectrum ε(q), defined as the scaling exponents of the temporal moments of interevent epicentral distances, M_q(θ) ~ θ^{ε(q)}. Linearity of ε(q) would imply a single diffusion exponent; the observed break between a linear negative-q branch and a nonlinear positive-q branch is the object that establishes strong anomalous diffusion and the swarm-versus-tectonic contrast.

What would settle it

Compute ε(q) for several other large, well-relocated tectonic mainshock–aftershock sequences of comparable size; if those sequences also show a nonlinear positive-q branch without high-order saturation, the claimed fingerprint fails as a diagnostic boundary between swarms and conventional tectonic sequences.

Watch

Extended reading notes

Core claim

All ten persistent earthquake swarms display the same multiscaling spectrum ε(q): approximately linear for negative moments and systematically nonlinear for positive moments, so that short and long interevent distances evolve under different effective scaling laws and swarm migration is strong anomalous diffusion. By contrast, the Landers tectonic sequence shows high-order saturation of ε(q), consistent with spatially bounded diffusion. The multiscaling spectrum therefore supplies a quantitative fingerprint capable of distinguishing swarm migration from conventional tectonic earthquake sequences.

Load-bearing premise

The contrast that makes the spectrum a class fingerprint rests on treating the single Landers sequence as a fair stand-in for ordinary tectonic aftershock sequences in general.

Editorial extensions

If this is right

  • Multiscaling spectra can serve as a catalogue-based diagnostic to classify clustered seismicity as swarm-like versus mainshock–aftershock.
  • Swarm migration cannot be reduced to one diffusion exponent and must be treated as heterogeneous, multi-scale transport.
  • Negative-order moments mainly track compact pairs while positive-order moments track extremes, so the two branches separately constrain short-range and long-range migration.
  • Differences in high-order curvature among swarms may later be linked to distinct driving or geometric regimes once more sequences are measured.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the fingerprint is stable, automated pipelines could flag swarm-like migration from catalogue geometry alone without waiting for full moment-release statistics.
  • The same moment-scaling test might separate industrial induced-seismicity swarms from tectonic aftershocks in monitoring settings.
  • Repeating the analysis on three-dimensional hypocenters rather than epicenters would test whether Landers-style caging is only a fault-plane projection effect.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The manuscript identifies ten persistent seismic swarms in the relocated Southern California catalogue via a consensus DBSCAN-like clustering procedure and analyzes their spatial migration with the multiscaling spectrum ε(q) of interevent-distance moments M_q(θ). All ten swarms show an approximately linear ε(q) for negative q and a systematically nonlinear positive-q branch, interpreted as strong anomalous diffusion in which small and large separations evolve with different effective exponents. By contrast, the Landers mainshock–aftershock sequence exhibits high-order saturation of ε(q), attributed to spatially bounded (caged) diffusion along the fault projection. The authors conclude that ε(q) supplies a quantitative fingerprint distinguishing swarm migration from conventional tectonic sequences.

Significance. If the reported contrast generalizes, the multiscaling spectrum would be a useful, catalog-based diagnostic for swarm versus tectonic migration, complementing moment-release shape criteria and classical single-exponent diffusion models. Strengths include: (i) a clearly stated moment definition and shared temporal fitting window across q (avoiding artificial curvature); (ii) AICc-based selection of the positive-branch polynomial; (iii) open Zenodo code for swarm identification and multiscaling; and (iv) a reproducible consensus clustering workflow that reduces dependence on a single (τ, r0, ε) choice. The qualitative split between negative- and positive-order branches is visible across the ten swarm panels and is of genuine interest for anomalous-diffusion studies of seismicity.

major comments (4)
  1. Abstract, §3 (Fig. 3), and Conclusions: the claim that ε(q) is a fingerprint “capable of distinguishing swarm migration from conventional tectonic earthquake sequences” rests on a single tectonic baseline (Landers). Saturation may reflect Landers’ long, quasi-linear fault geometry rather than a generic tectonic signature. At least one or two additional large SCEDC mainshock–aftershock sequences (e.g., Hector Mine, Ridgecrest, or other N≳200 tectonic clusters) should be processed with the same pipeline (shared temporal interval, AICc protocol, r_ij cut) so that the nonlinear-vs-saturation contrast can be tested as a class boundary rather than a sequence-specific difference.
  2. §2 Methods: only 10 of 95 persistent swarms (those with N>200) enter the multiscaling analysis. The central “all analyzed swarms” claim is therefore conditioned on a size cut that may select the most spatially extended or longest-lived systems. The manuscript should either (a) show that smaller swarms with adequate pair counts in the admissible θ windows yield the same qualitative ε(q) shape, or (b) explicitly restrict the fingerprint claim to large, persistent swarms and discuss possible size bias.
  3. §3 and Fig. 2, Swarm 043: this sequence is flagged as a “notable exception” yet is still counted among the ten that exhibit the common signature. Its negative-order slope (H_−≈0.73) and positive-branch shape differ markedly from the other panels. Either reclassify 043 (and restate “all” claims accordingly) or provide a quantitative criterion for when a spectrum is considered consistent with the swarm fingerprint versus Landers-like saturation.
  4. §3, selection of the common temporal scaling interval: candidate windows must span ≥0.7 decades, ≥6 bins, and ≥50 pairs/bin, with the “best overall linear scaling” chosen for a representative set of q. Because the same interval is then imposed on all q, the procedure can still favor intervals that look linear at low |q| while leaving high-q curvature under-constrained. Report sensitivity of ε(q) (especially the positive branch and AICc degree) to alternative admissible windows, or show that the nonlinear-vs-linear split is stable across the admissible set.
minor comments (6)
  1. Eq. (1) and surrounding text: the CV formula is typeset as “Cy — bu”; restore standard notation c_v = σ/μ and define symbols cleanly.
  2. Eq. (2)–(5): the space–time metric, membership probabilities p_i and p_ij, and Mo conversion are incompletely or garbled in the text (missing powers, incomplete formulas). Provide complete, self-contained expressions.
  3. Figure 2: independent vertical scales make cross-swarm comparison of H_− and high-q curvature difficult; consider a common y-range or a summary table of H_−, AICc degree, and ε(q) at fixed q (e.g., q=2, 5, 10).
  4. Plain Language Summary and Key Points: useful, but “classical tectonic earthquake sequence” should be qualified as “the Landers sequence” until more baselines are shown.
  5. References: Passarelli et al. (2026) and Ross et al. (2025) appear with overlapping titles/DOIs in the list; verify bibliographic entries.
  6. Open Research: Zenodo DOI is given with a space (“zenodo. 21297044”); fix the link for reproducibility.

Circularity Check

1 steps flagged · score 1.0 of 10

Empirical multiscaling spectra from catalog moments; mild self-citation of method and Landers baseline is not load-bearing for the swarm fingerprint claim.

  1. self citation load bearing [§3 Multiscaling in Earthquake Diffusion; Fig. 3 and surrounding text on Landers]
    "A natural framework for investigating these heterogeneous transport processes is provided by the multiscaling formalism introduced by Godano and Pingue (2005). ... Such a result confirms the one already obtained by Godano and Pingue (2005) for a whole California catalogue."

    The multiscaling spectrum ε(q) and the interpretation of high-q saturation as caged/bounded diffusion are taken from the authors’ prior work and reapplied; Landers is presented as confirming that earlier self-result. This is mild methodological self-citation, not a definitional reduction of the new swarm fingerprint (which is measured independently on the ten swarms). Not load-bearing for the swarm-vs-tectonic contrast claim.

full rationale

The paper’s central claim is an observed pattern: for ten consensus-selected swarms, ε(q) is approximately linear for q < 0 and systematically nonlinear for positive q, while Landers saturates at high q. ε(q) is defined operationally from the data via M_q(θ) = ⟨r^q⟩_θ ∼ θ^{ε(q)} (Eqs. 6–7), with a common temporal fitting window and AICc polynomial degree for the positive branch; nothing in that pipeline forces the negative-linear / positive-nonlinear shape by construction, nor equates the swarm spectra to the Landers spectrum. Swarm membership is obtained from an external-style consensus (Passarelli et al. centroid/skewness/kurtosis criteria plus multi-run DBSCAN co-membership thresholds p_i, p_ij ≥ 0.7), independent of ε(q). Self-citations (Godano & Pingue 2005 multiscaling formalism and prior California caged-diffusion result; Godano & Petrillo completeness CV) supply the analysis toolkit and a confirmatory Landers baseline, but do not define or force the swarm signature. No fitted parameter is renamed a prediction, no uniqueness theorem is imported, and no ansatz is smuggled that makes the fingerprint tautological. The Landers-as-sole-tectonic-baseline issue is a generalization/correctness concern, not circularity. Score 1 only for the non-load-bearing self-citation of the multiscaling method and Landers confirmation.

Assumptions & free parameters 8 free parameters · 6 assumptions · 1 invented entities

The paper is empirical multiscaling applied to catalog clusters. Load-bearing inputs are standard diffusion/multifractal formalism, catalog completeness and clustering thresholds (many hand-chosen), and the interpretive leap that nonlinear ε(q) fingerprints swarms versus tectonic sequences. No new physical particle or force is introduced; free parameters are analysis cutoffs and fit choices that can change which events and which temporal windows enter ε(q).

free parameters (8)
  • completeness CV threshold c_v
    Authors adopt c_v=0.9 (more conservative than 0.73 from simulations) to set m_c=2.0; this hand choice controls which events enter clustering and moments.
  • DBSCAN-like clustering grid (τ, r0, ε)
    27 realizations over τ∈{0.03,0.05,0.10} yr, r0∈{0.5,1.0,1.5} km (text shows related grid), ε∈{0.8,1.0,1.2}; neighborhood definition is parameter-dependent.
  • swarm moment-release criteria t_c, S, K
    t_c>0.5, S<10, K<125 taken from Passarelli et al. synthetic ETAS guidance; these thresholds define which clusters are swarm-like.
  • consensus membership cutoffs p_i, p_ij
    Persistent swarms require p_i≥0.7 and co-membership p_ij≥0.7; cutoffs are chosen, not derived.
  • minimum sizes (30 per cluster; 200 for multiscaling)
    Only clusters ≥30 events retained; only 10 swarms with >200 events analyzed—selection that shapes the claimed common signature.
  • r_ij exclusion and temporal binning
    Pairs with r_ij<0.05 km dropped; logarithmic bins with factor 1.3; admissible windows need ≥6 bins, ≥0.7 decades, ≥50 pairs—analysis design parameters.
  • shared temporal scaling interval selection
    One interval per swarm chosen for best overall linear scaling of log M_q vs log θ across representative q; this choice can influence apparent linearity/nonlinearity of ε(q).
  • positive-branch polynomial degree via AICc
    Linear/quadratic/cubic selected by AICc (prefer lower degree if ΔAICc<2); empirical description of nonlinearity, not a fixed physical model.
assumptions (6)
  • domain assumption Interevent-distance moments scale as M_q(θ)∼θ^{ε(q)}; linear ε(q)=qH means single-exponent diffusion, nonlinear ε(q) means strong anomalous diffusion.
    Core multiscaling framework from Castiglione et al. (1999) and Godano & Pingue (2005), Methods §3.
  • domain assumption Multifractal representation ε(q)=min_h[qh+d−D(h)] with embedding d=2 for epicenters is an appropriate phenomenological description.
    Invoked after Eq. 11; interpretive, not tested against alternatives.
  • domain assumption Magnitude distribution above completeness is exponential (Gutenberg–Richter), so CV→1 identifies m_c.
    Used in Methods for m_c=2.0 via CV method.
  • domain assumption Passarelli-style centroid/skewness/kurtosis criteria plus consensus graph connectivity isolate physically meaningful persistent swarms.
    Swarm catalog construction depends on these classification axioms.
  • domain assumption High-order saturation of ε(q) indicates spatially bounded ('caged') diffusion along a fault projection.
    Interpretation of Landers spectrum (Fig. 3), analogized to glassy caging.
  • ad hoc to paper A common temporal fit window for all q avoids artificial curvature better than q-dependent windows.
    Explicit methodological choice in §3; reasonable but not uniquely justified.
invented entities (1)
  • multiscaling fingerprint of swarm migration
    purpose: Name the claimed common nonlinear ε(q) signature as a diagnostic distinguishing swarms from tectonic sequences.
    Descriptive label for an empirical spectral shape, not a new physical mediator; independent evidence is only the ten-swarm sample plus Landers contrast within this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Multiscaling Fingerprint of Earthquake Diffusion in Seismic Swarms." pith.science (2026). https://pith.science/paper/46T2E6VA

@misc{pith2026260711142,
  author       = {Pith},
  title        = {Pith review of: A Multiscaling Fingerprint of Earthquake Diffusion in Seismic Swarms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/46T2E6VA}},
  note         = {Machine review of arXiv:2607.11142}
}
abstract

Seismic swarms are commonly associated with fluid migration and other transient processes, yet their spatial migration remains difficult to quantify using conventional diffusion models. Here we analyze ten persistent earthquake swarms identified within the relocated Southern California earthquake catalogue using a consensus clustering approach. We characterize their migration through the multiscaling spectrum $\varepsilon(q)$ obtained from the temporal evolution of the moments of interevent distances. All analyzed swarms exhibit a common scaling signature: the spectrum is approximately linear for negative moments but departs systematically from a single linear behavior for positive moments, indicating that small and large interevent distances evolve with different effective scaling laws. In contrast, the Landers tectonic sequence displays a high-order saturation of $\varepsilon(q)$, consistent with spatially bounded diffusion. These results reveal that earthquake swarms are characterized by strong anomalous diffusion and suggest that the multiscaling spectrum provides a quantitative fingerprint capable of distinguishing swarm migration from conventional tectonic earthquake sequences.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

4 extracted references · 1 canonical work pages

  1. [1]

    Bosl, W., & Nur, A. (2002). Aftershocks and pore fluid diffusion following the 1992 landers earthquake. Journal of Geophysical Research: Solid Earth, 107(B12), ESE-17. Castiglione, P., Mazzino, A., Muratore-Ginanneschi, P., & Vulpiani, A. (1999). On strong anomalous diffusion. Physica D: Nonlinear Phenomena, 184(1), 75-93. Convertito, V., Godano, C., Petr...

    2002

  2. [2]

    doi: 10.1111/j.1365-246X.2004.02463.x Hauksson, E., Yang, W., & Shearer, P. M. (2012). Waveform relocated earthquake catalog for southern california (1981 to june 2011). Bulletin of the Seismologi- cal Society of America, 102(5), 2239-2244. Hill, D. P. (1977). A model for earthquake swarms. Journal of Geophysical Research, 82(8), 1347-1352. doi: 10.1029 /...

    work page Pith review arXiv doi:10.1111/j.1365-246x.2004.02463.x 2004

  3. [3]

    Petrillo, G., Kumazawa, T., Napolitano, F., Capuano, P., & Zhuang, J. (2024). Fluids-triggered swarm sequence supported by a nonstationary epidemic-like description of seismicity. Seismological Research Letters, 95(6), 3207-3220. doi: 10.1785 /0220240056 Reasenberg, P. A., & Simpson, R. W. (1992). Response of regional seismicity to the static stress chang...

    2024

  4. [4]

    Uchida, N., & Beroza, G. C. (2024). Repeating earthquakes. Annual Review of Earth and Planetary Sciences, 52, 123-145. Vidale, J. E., & Shearer, P. M. (2006). A data-driven catalog of U.S. West Coast earthquake swarms: Characteristics and implications. Bulletin of the Seismo- logical Society of America, 96(5), 1652-1660. doi: 10.1785 /0120050241 Weeks, E....

    doi:10.1103/physrevlett 2024

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.