Reply to “Comment on ‘High-Definition Mapping of the Gutenberg–Richter b-Value and Its Relevance: A Case Study in Italy’ by M. Taroni, J. Zhuang, and W. Marzocchi” by Laura Gulia, Paolo Gasperini, and Stefan Wiemer

Taroni et al. (2021; hereafter TZM21) proposed a method to perform a spatial b-value mapping based on the weighted-likelihood estimation and applied this method to the Italian region as a tutorial example. In the accompanying comment, Gulia et al. (2021; hereafter GGW21) did not challenge the TZM21’s method, but they argued that the catalog used by TZM21 is contaminated by quarry blasts, introducing a bias that may impact any seismotectonic or hazard interpretations. Although in TZM21 the application to the Italian territory was only a tutorial example and we purposely did not make any thorough discussion on the meaning of the results in terms of seismotectonic or seismic hazards (that would have required many more analyses), we acknowledge the potential role of the quarry blasts, and we add some further analysis here. We thank GGW21 for giving us this opportunity. Here, removing the part of the catalog contaminated by quarry blasts and applying the same analysis as in TZM21, we obtain results that are very similar to the ones reported in TZM21; specifically, only one region that is characterized by low natural seismicity rate shows a marked effect of the quarry blasts on the b-value.

SEISMICITY of the RUSSIAN PART of EAST EUROPEAN PLATFORM and ADJACENT TERRITORIES in 2015

It is reported that 41 stationary seismic stations, 2 arrays, and 7 temporary seismic stations, located in the area of Novovoronezh and Kursk nuclear stations, monitored seismicity of the Russian territory of the East European Platform (EEP) in 2015. The registration capabilities of the seismic network at the EEP as a whole were estimated based on the average station noise level and the equation for the energy decay of seismic phases. Zones with the best capabilities have been allocated. A feature of seismicity in 2015 is the manifestation of earthquakes of moderate magnitudes (ML=2.7–3.9) in the peripheral regions (in the southwest, west, and northwest) and in zones associated with paleorift structures: in the southwest – with the Dnieper Donetsk and in the northeast – with the Kirov-Kazhim and Soligalich (Central Russian) aulacogenes. The results of the macroseismic survey are given for the earthquake in Poltava on February 2, 2015, with M=3.7; focal mechanisms of two earthquakes (03.02.2015 and 12.06.2015) are constructed. According to the data of the Latvian Center, an earthquake was recorded in the region of Lithuania bordering the Kaliningrad region. Weaker natural seismicity with ML≤2.5 was recorded in Karelia and the regions bordering with Finland, near the Kandalaksha Bay, near the Khibiny, and Lovozersky massifs on the Kola Peninsula, and on the territory of the Voronezh crystalline massif.

Toward False Event Detection and Quarry Blast versus Earthquake Discrimination in an Operational Setting Using Semiautomated Machine Learning

Small-magnitude earthquakes shed light on the spatial and magnitude distribution of natural seismicity, as well as its rate and occurrence, especially in stable continental regions where natural seismicity remains difficult to explain under slow strain-rate conditions. However, capturing them in catalogs is strongly hindered by signal-to-noise ratio issues, resulting in high rates of false and man-made events also being detected. Accurate and robust discrimination of these events is critical for optimally detecting small earthquakes. This requires uncovering recurrent salient features that can rapidly distinguish first false events from real events, then earthquakes from man-made events (mainly quarry blasts), despite high signal variability and noise content. In this study, we combined the complementary strengths of human and interpretable rule-based machine-learning algorithms for solving this classification problem. We used human expert knowledge to co-create two reliable machine-learning classifiers through human-assisted selection of classification features and review of events with uncertain classifier predictions. The two classifiers are integrated into the SeisComP3 operational monitoring system. The first one discards false events from the set of events obtained with a low short-term average/long-term average threshold; the second one labels the remaining events as either earthquakes or quarry blasts. When run in an operational setting, the first classifier correctly detected more than 99% of false events and just over 93% of earthquakes; the second classifier correctly labeled 95% of quarry blasts and 96% of earthquakes. After a manual review of the second classifier low-confidence outputs, the final catalog contained fewer than 2% of misclassified events. These results confirm that machine learning strengthens the quality of earthquake catalogs and that the performance of machine-learning classifiers can be improved through human expertise. Our study promotes a broader implication of hybrid intelligence monitoring within seismological observatories.

Induced Seismicity Completeness Analysis for Improved Data Mining

The study of induced seismicity at sites of fluid injection is paramount to assess the seismic response of the earth’s crust and to mitigate the potential seismic risk. However statistical analysis is limited to events above the completeness magnitude mc, which estimation may significantly vary depending on the employed method. To avoid potential biases and optimize the data exploitable for analysis, a better understanding of completeness, detection capacity and censored data characteristics is needed. We apply various methods previously developed for natural seismicity on 16 underground stimulation experiments. We verify that different techniques yield different mc values and we suggest using the 90% quantile of the mc distribution obtained from high-resolution mapping, with mc defined from the mode of local magnitude frequency distributions (MFD). We show that this distribution can be described by an asymmetrical Laplace distribution and the bulk MFD by an asymmetric Laplace mixture model. We obtain an averaged Gutenberg-Richter parameter b=1.03±0.48 and a detection parameter k=3.18±1.97 from mapping, with values subject to high uncertainties across stimulations. We transfer Bayesian mc mapping developed for natural seismicity to the context of induced seismicity, here adapted to local three-dimensional seismicity clouds. We obtain the new prior parameterization mc,pred=1.64log10(d3)−1.83, with d3 the distance to the 3rd nearest seismic station. The potential use of censored data and of mc prediction is finally discussed in terms of data mining to improve the monitoring, modeling and managing of induced seismicity.

Analysis of the seismic and paraseismic resistance of the VUJE X=70 technology base

Vibration caused by natural seismicity and paraseismicity used in manufacturing facilities can affect the quality of production activity. The vibration velocity of concrete slabs which are impacted by seismic action can be determined by frequency response spectrum. The paper presents a case study of a real industrial assessment.

Riesgos naturales y conservación de la arquitectura defensiva de tierra: aproximación a los daños causados por seísmos en la Alcazaba de Almería y en la muralla de La Hoya

Natural Risks and Conservation of Rammed-earth Defensive Architecture: Approach to the Damages Caused by Earthquakes in the Alcazaba of Almeria and the Wall of La HoyaThe aim of this work is to present the results obtained in the framework of the PREFORTI Project. The particular case analysed is the damages caused by earthquakes in the medieval fortifications of Almeria and the consequences on their conservation. Almeria is a zone of important seismic activity. This particularity has caused many problems to conserve its Islamic military architecture. Within this work, we include the most important earthquakes that have affected this heritage since the late fifteenth century and the constructions damaged –to a greater or lesser extent– by them, including their specific damages and an approach to their state of conservation. Due to its importance within the military constructions, we present the case study of the Alcazaba of Almeria and the wall of La Hoya, focusing on the second one. These paradigmatic constructions were widely affected by the earthquakes since the late fifteenth century. Beyond the analysis of the damages caused by the earthquakes and the different historical restorations to mitigate them, we include the emergency measures proposed in the framework of the PREFORTI Project for its better conservation against the different risks, as well as the microzonation mapping of the natural seismicity risk for the section of the wall of La Hoya in order to delve in the study of the vulnerability of this cultural asset against this natural risk, as a mechanism for its better preventive conservation.

Bachatskiy induced earthquake on June 18, 2013, ML=6.1, I0=7 (Kuzbass)

The Bachatsky earthquake of 18 June 2013 and a seismic activation of the same name coal strip mine, started several years before the earthquake and still persists today, have been studied using temporal local seismic arrays in the area. It was found experimentally that the seismic process area is closely connected to open workings, and the earthquakes are extend-ed from the working bed to a depth of 4-5 km. Adjacent to the mine depression sedimentary rocks were activated. The technogenic seismic regime is continuous and not stationary: intervals of background seismicity with relatively weak and seldom events are disturbed by bursts of activity with a rise in the magnitude of stronger earthquakes and frequency of occurrence of weak events. The seismic activation may last for 1–3 months. During the last five years, four seismic activations have been recorded, three of which were generated by large earthquakes of 09.02.2012, ML4.3; 04.03.2013, ML3.9; 18.06.2013, ML6.1. The last one was completed by a series of perceptible earthquakes with local magnitude of 3.0–3.5. The focal mechanism of the Bachatsky earthquake is a thrust fault with one of the motion planes corresponding to the anthropogenic impact. The earthquake flow forms a single process in the space with the b-value of the Gutenberg-Richter relationship different from the natural seismicity. The studied induced seismicity does not correspond to the structural regularities of natural seismicity in the Altai-Sayan mountain area. The findings prove that the Bachatsky earthquake and associated activation can be considered as man-made events.

Natural Seismicity in and around the Rome Trough, Eastern Kentucky, from a Temporary Seismic Network

The Rome trough is a northeast-trending graben system extending from eastern Kentucky northeastward across West Virginia and Pennsylvania into southern New York. The oil and gas potential of a formation deep in the trough, the Rogersville shale, which is ∼1 km above Precambrian basement, is being tested in eastern Kentucky. Because induced seismicity can occur from fracking formations in close proximity to basement, a temporary seismic network was deployed along the trend of the Rome trough from June 2015 through May 2019 to characterize natural seismicity. Using empirical noise models and theoretical Brune sources, minimum detectable magnitudes, Mmin, were estimated in the study area. The temporary stations reduced Mmin by an estimated 0.3–0.8 magnitude units in the vicinity of wastewater-injection wells and deep oil and gas wells testing the Rogersville shale. The first 3 yr of seismicity detected and located in the study area has been compiled. Consistent with the long-term seismicity patterns in the Advanced National Seismic System Comprehensive Catalog, very few earthquakes occurred in the crust beneath the Rome trough—only three events were recorded—where the temporary network was most sensitive. None of these events appear to have been associated with Rogersville shale oil and gas test wells. Outside of the trough boundary faults, earthquakes are diffusely distributed in zones extending into southern Ohio to the north, and into the eastern Tennessee seismic zone to the south. The orientations of P axes from the seven first-motion focal mechanisms determined in this study are nearly parallel with both the trend of the Rome trough and with the orientation of maximum horizontal compressive stress in the region. This apparent alignment between the regional stress field and the strikes of faults in the trough at seismogenic depths may explain the relative lack of earthquake activity in the trough compared with the surrounding crust to the north and south.