| 21 Apr 2026
The Mechanisms Behind Triggering the 2021–2022 South Alboran Seismic Swarm: Constrains from Combined Catalogues, Relocation and Spatiotemporal Analysis
Abstract. The Alboran system reflects the interplay of slow Nubia–Iberia convergence, inherited structures, and ongoing lithospheric attenuation. Earthquake occurrence mainly tracks active crustal fault networks that partition the oblique plate motion into strike-slip and extension. The 2021–2022 seismic swarm in the western Alboran Sea represents an exceptional episode characterized by complex temporal-spatial evolution, primarily influenced by fluid-driven processes interacting with inherited fault systems. We analyzed arrivals of approximately 7,000 seismic events recorded by Spanish (IGN) and Moroccan (CNRST) seismic networks. Both bulletins have been individually and jointly processed using the double-difference algorithm (HypoDD) and a regionally optimized velocity model. This approach significantly improved hypocentral precision, reducing event scatter and delineating a clearly defined, near-vertical seismic conduit-oriented NW–SE. Spatiotemporal analyses revealed distinct episodes of seismic migration, consistent with episodic fluid overpressure pulses, confirmed by diffusivity values (1.2–13.9 m2/s) characteristic of fluid-controlled swarms. Focal mechanisms predominantly indicated strike-slip motion, aligning with the regional transtensional tectonics and pinpointing the unrecognized Ras Tarf Fault as the primary seismogenic structure likely linked to the Al-Idrissi Fault System (AIFS). Integration with vertical and horizontal shear-wave velocity models (VSH and VSV) highlighted velocity anomalies at depths of 30–70 km, suggesting the presence of partially serpentinized mantle wedges above a delaminating slab segment, further supporting fluid involvement. Our results emphasize the critical interplay between deep lithospheric fluids, inherited fault structures, and regional tectonic stress, providing a comprehensive framework for understanding the 2021–2022 swarm dynamics which could improve the seismic hazard assessment in the region.
This is my formal review of the paper “The Mechanisms Behind Triggering the 2021-2022 South Alboran Seismic Swarm: Constraints from Combined Catalogues, Relocation
and Spatiotemporal Analysis” by Hamza Akka et al, submitted to Egugeophere.
The authors study the 2021-2022 seismic swarm that occurred in the Alboran sea by using seismic data deriving from Spain and Morocco seismic networks.
My overall assessment of the manuscript is not positive. In my opinion, the integration of seismic data from the Spanish and Moroccan seismic arrays represents the main strength of the study. The combined dataset significantly improves azimuthal coverage—a critical factor for offshore seismicity—and provides a more robust constraint on hypocentral locations compared with previous studies based on single-network catalogues. This effort is valuable and potentially important for understanding the swarm evolution. However, several aspects of the manuscript require substantial revision. Although the authors emphasize the development of a merged catalogue, many analyses throughout the paper still rely on the individual network datasets (e.g., focal mechanisms and some statistical analyses). I strongly suggest focusing the study primarily on the merged dataset and limiting the use of single-network catalogues to an initial quantitative comparison. Repeating the same analyses on lower-quality catalogues introduces confusion and unnecessarily complicates the manuscript. Furthermore, I am not fully convinced by the interpretation that fluid diffusion is the primary mechanism controlling the swarm evolution. While fluids can certainly play an important role in seismic swarm activity, the evidence presented here does not appear sufficiently robust to support this conclusion.
For these reasons, I recommend major revisions. My detailed comments are provided below.
DATA AND METHOD
The integration of seismic picks from the Spanish and Moroccan networks is a key contribution of this work and deserves a much more detailed description.
Specifically, the manuscript should clarify:
I suggest limiting the comparison between single-network and merged catalogues to the Hypoinverse locations only, while restricting the hypoDD relocation analysis to the merged dataset.
For both the single-network and merged catalogues, the authors should provide histograms of:
Figure 3 could therefore be reorganized to display:
Since the authors use hypoDD for relocation, I also suggest to compute waveform cross-correlations in order to improve differential travel-time measurements and further enhance location accuracy.
SPATIAL DISTRIBUTION AND SEISMICITY PATTERN
I suggest making a figure with only the final hypoDD locations that will be the starting point for the successive analysis. Figure 4 should display only the spatial distribution of such a catalog.
STATISTICAL ANALYSIS
Please better clarify why the merged catalogue consists of 5330 events while the CNRST and IGN catalogs are made of 3483 and 3720 events respectively. Do you perform a new earthquake detection using the continuous recordings of both the arrays?
I suggest using and to discuss hypoDD results only on the merged dataset, specifying how you compute the location errors (SVD, bootstrap?)
SPATIO TEMPORAL EVOLUTION
Even in this chapter, I suggest carrying out the analysis on the merged dataset. It does not make sense discussing the cluster geometry with locations suffering from high azimuthal gaps, even for the ISN catalog that covers a wide time span.
FLUID DIFFUSIVITY
Discarding figure 6 and 7 derived from single network catalogs and focusing on figure 8, I am not able to observe any diffusion pattern. It is also difficult to understand why a diffusivity plot is reported by using a declustered catalog in panel c). It is difficult for me to agree with the sentence “Taken together, the combined catalogue RT plot confirms that the 2021–2022 seismic swarm evolved as a series of fluid-assisted migration episodes” as reported in lines” 484-485.
This part should be substantially revised. I think that this diffusivity plot demonstrated that the swarm is modulated by other factors rather than fluid diffusion. When seismicity is driven by fluid diffusion, the positions of earthquakes on the RT plot are bounded by the diffusive parabolic front and by the parabolic backfront, closest to the time axis. The back front, which separates the region where diffusion effects persist from the region where pressure relaxation has already occurred, is considered a diagnostic feature of diffusion-driven seismicity (see Parotidis, M. et al., 2004. Back front of seismicity induced after termination of borehole fluid injection, GRL, doi:10.1029/2003GL018987). The authors may also test whether diffusion patterns become more evident when the RT plots are referenced to the origin times and hypocentral locations of the major seismic bursts identified within the sequence. In addition, I strongly encourage the authors to perform a Coulomb static stress transfer analysis between the major bursts using the relocated catalogue. Such an analysis could help determine whether the migration pattern is primarily fluid-driven or instead controlled by stress transfer from moderate earthquakes, such as the Mw 5.2 event of December 2021. More generally, additional analyses would be necessary to robustly assess the role of fluids in the swarm evolution. For example a local earthquake tomography to study the spatial and temporal evolution of Vp/Vs (see for example Pezzo, G., et al, (2018). Pore pressure pulse drove the 2012 Emilia (Italy) series of earthquakes. GRL, 45, 682–690. https://doi.org/10.1002/2017GL076110) or the Vp/Vs derived from clustered seismicity (e.g. Lin, G., & Shearer, P. Estimating Local Vp/Vs Ratios within Similar Earthquake Clusters. BSSA. https://doi.org/10.1785/0120060115, 2007)
FOCAL MECHANISMS
The manuscript does not adequately describe the methodology used to compute focal mechanisms. Based on the frequency–magnitude distribution of the merged catalogue (Figure 4), approximately 200 events with M > 2.5 are available. This subset should be exploited for focal mechanism analysis, potentially providing important insights into the regional stress field and faulting style.
I do not understand why the authors rely on focal mechanisms derived from the IGN catalogue, especially considering that many hypocentral depths are fixed at 5 km (Figure 9). Furthermore, the focal mechanisms shown in Figure 9 are predominantly strike-slip solutions. However, when projected onto vertical cross-sections, beach-ball representations cannot appear as they do in map view. The compressional and dilatational quadrants intersect the section plane differently depending on the strike of the cross-section (see the “pscoupe” module in GMT).
Again, if the authors intend to derive seismotectonic interpretations from focal mechanisms and hypocentral distributions, these analyses should be based on the merged and relocated dataset. Large azimuthal gaps may lead to poorly constrained focal-plane solutions.
Overall, I do not find Figure 9 particularly informative in its current form.
LITHOSHERIC STRUCTURE AND SWARM
I find some incongruence between tomograms of VSH and VSV (Figure 10) and the description of anomalies in relationship to seismicity (lines 545-555). I have the impression, especially looking at the velocity maps in Figure 10, that seismicity falls inside a band of slightly high velocity (color with a tendency to orange)
When uppermantle structure is invoked to delineate the plate geometry in this sector of Mediterranean region, I suggest take into account the tomographic model derived from teleseismic events (Monna, S. et al. (2013), New insights from seismic tomography on the complex geodynamic evolution of two adjacent domains: Gulf of Cadiz and Alboran Sea, JGR, doi:10.1029/2012JB009607)
TECTONICS ARCHITECTURE AND FAULT SYSTEM DYNAMICS
This paragraph may act as an introduction of the fault system. This does not help with the comprehension of the paper. In this section one should read how the new data and the new results add to the study of the faults system.
CONCLUSION
The sentence “an initial downward migration” (line 782) demonstrates that swarms are probably not fluid driven and other mechanisms are useful to explain it. If fluids are involved and diffusions take place an upward migration could be more realistic.