Review article: Deterministic seismic hazard assessment of the area comprised between west Gulf of Cádiz and east Alboran Sea
Abstract. The convergence zone of the NE-SW complex comprising the Gulf of Cádiz and the Alboran Sea, at the Eurasian-Nubian plate boundary, is frequently affected by seismic activity, caused by submarine long-strike strike-slip fault systems and arcuate fold-thrust systems found in the region. This has resulted in moderate to high magnitude earthquakes, including tsunamigenic earthquakes, and the area has also experienced tsunamis due to major earthquakes and gravitational landslides. This study carries out a Seismic Hazard Analysis, for bedrock conditions, of the marine area between the W of the Gulf of Cádiz and the E of the Alboran Sea in the Ibero-Maghrebian region, based on a deterministic approach, considering areal seismogenic sources. For the estimation of the seismic hazard, a Visual Basic script based on Excel has been used, which has been improved. The results obtained show that the most probable Peak Ground Acceleration (PGA) ranges from 0.2 to 0.4 g, although it can reach up to 1.0 g in certain areas. These results highlight the need for a detailed study of the distribution of seismic hazard in submarine areas, given the significant values of accelerations that can occur. This work is the first comprehensive deterministic seismic hazard assessment carried out in the Ibero-Maghrebian region and aims to take a first step to promote seismic hazard studies in marine areas, whose results can provide relevant information given the implications of earthquakes in the genesis of other natural hazards such as tsunamis and submarine landslides.
Adrián José Rosario Beltré et al.
Status: open (extended)
RC1: 'Comment on nhess-2023-52', João Fonseca, 04 May 2023
- AC1: 'Reply on RC1', Adrián José Rosario Beltré, 11 May 2023 reply
Adrián José Rosario Beltré et al.
Adrián José Rosario Beltré et al.
Viewed (geographical distribution)
Beltré et al. submitted a manuscript entitled “Review article: Deterministic seismic hazard assessment of the area comprised between west Gulf of Cádiz and east Alboran Sea” presenting the results of a deterministic seismic hazard assessment (DSHA) for the Cadiz Gulf / Alboran Sea region and benchmarking their results against previous studies for the emerged parts of their study area given the lack of previous studies for the submerged regions. The manuscript discusses at length the advantages and limitations of probabilistic (PSHA) versus deterministic assessments, and similarly for zoned versus non-zoned deterministic studies. Another aspect that deserves particular attention is the use of a Digital Terrain Model (DTM) to allow the correct estimate of the distance from source to site.
Although the manuscript presents itself as a review paper, it does not correspond in the opinion of this reviewer to the requirements of such qualification, namely “[to] summarize the status of knowledge and [to] outline future directions of research”. The remainder of this review will address the relevance of the manuscript as a potential research paper.
The manuscript adopts an in-depth description of the details of the work done that makes it read more like a technical report than a scientific paper, where the relevant aspects should be highlighted and additional details synthetized in tables or supplementary material. With the profusion of details, the reader looses the purpose of the manuscript: To present a new methodological approach? To advocate the need for hazard assessment of submerged regions? To put forward a product ready to be used for the design of infrastructures? All of the above? Perhaps because of the profusion of apparent goals, the manuscript fails, in the opinion of this reviewer, to be convincing for any of them.
When advocating for DSHA instead of PSHA, the authors seem to reinstate a dated controversy of the 1990s in the United States, after which vast clarifications of concepts and domains of applicability were obtained that were not discussed. While addressing a relatively untouched topic which is seabed hazard assessment, the authors fail to highlight the growing exposure of submarine structures and associated risk, which would be an interesting point in favour of the study. In what concerns the applicability of the results for practical purposes, the non-standard aspects of the approach would require much better justification for it to be considered fit for use.
To illustrate the latter aspect: the authors advocate the adoption of a zoned DSHA method according to Wang and Huang (2014), relying heavily on that single reference – in fact, from which the EXCEL code to perform the computations was obtained - to support that not-so-standard choice. To differentiate zoned from non-zoned DSHA, the authors state that “the application of the DSHA is not zoned, if it is done on a homogeneous seismicity region, with a single seismic source of global influence” (line 394). However, it is well known that one of the strongest arguments in favour of DSHA, and one that makes it adopted for the design of critical structures, is that it is reality-based, taking into account the geological structures surrounding the site and their seismogenic characteristics. DSHA has at its core the concept of maximum credible earthquake (MCE) at a particular geological structure or set of well identified geological structures, and usually avoids the use of area sources altogether. In the manuscript, the authors describe in details the relevant geologic structures of the region of interest, but at the time of computing the hazard these structures do not seem to be taken in consideration, the MCE being considered equal to the MMax of each area source in previous (probabilistic!) studies.
Even more concerning, it was not possible for this reviewer to understand how the distance from source to site was computed prior to its insertion in the GMPEs: after describing several second-order corrections to the distance, the authors mention that those corrections were applied to “the minimum distance calculated with the original script” (line 618, referring to a pre-existing EXCEL script that was adapted for the calculations). Were the authors assuming that a fault capable of producing the MCE for each area zone existed at the minimum distance from the site to each area source? This would certainly increase the hazard unwarrantedly.
Another non-standard approach that the authors take, still following Wang and Huang (2014), is the consideration of “non-control area sources” in addition to the “control area source” (i.e., the source producing the maximum maximum credible earthquake). In this case, the authors remark that the hazard is higher at sites near the “confluence” of source zones, an observation that in and by itself should raise concerns: if the pattern of the hazard is driven by the source zone geometry, something must be wrong with the methodology (actually, this limitation is even clearer in Wang and Huang (2014)’s hazard assessment for Taiwan, which shows a “checkered” pattern when “non-control sources” are considered, with no similarity whatever to the results obtained when non-control sources are not used (their figures 5 and 6); it should be pointed out that Wang and Huang (2014) was published in a journal which is very remote from mainstream seismological research, which may account for a less rigorous review process for that paper; it is a tell-tale detail that no mainstream hazard publication has refered Wang and Huang (2014)). The authors report a 445% (!) increase of the hazard at Cadiz (84% percentile) when non-control area sources are added to the control area source, and fail to identify that discrepancy as anomalous.
Because fundamental aspects like these were not documented to a point that this reviewer could understand, I am forced to consider the results unfit for use.
In addition to the serious reservations listed about, it must be mentioned that the manuscript contains inaccuracies easily corrected had the manuscript bee n previously subjected to any level of review prior to submission. Just to mention a few:
Line 93: “As shown by geodetic data and geodynamic studies that have been carried out for decades (e.g., Montessus de Ballore, 1894; Pastor, 1927; Munuera, 1963) tectonic models of NW-SE to WNW-ESE oblique convergence between the plates (Reilly et al., 1992; Herraiz et al., 2000) present displacements of 2 to 5 mm/yr”
Pre-plate tectonics studies can hardly have any relevance to the discussion rates of convergence between plates.
Line 212: “Seismic hazard is the probability that an earthquake occurs in a given geographic area, within a given time period, and with a ground motion intensity exceeding a given threshold (McGuire, 2004)”
This passage indicates a lack of familiarity with the concept of seismic hazard that is not adequate in a manuscript on seismic hazard assessment. Seismic hazard concerns probability of exceedance of a level of ground motion at a site over a period of time, not probability of occurrence of an earthquake in a geographic area.
line 387: “… the probability density of the 50th percentile, or average value, …”
The 50% percentile of a pdf corresponds to the median, not the “average” (in other passages. e.g., line 423, it is called “mean”, also wrongly), and even without that imprecision the sentence is devoid of meaning (the median is a value of the random variable, not of the probability density).
In view of the above remarks, I regret that I have to recommend that the manuscript should not be accepted for publication in NHESS. Should the authors consider rewriting it to try a subsequent submission, the attached pdf includes extensive comments on the passages that need improvement.