MHMarijan Herak

Marijan Herak

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Marijan Herak has a total of 204 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    The Baranja earthquakes of 1922 and 1924

    by Marijan Herak, Davorka Herak

    In regions of low seismicity, such as Baranja in northeastern Croatia, seismic hazard assessments rely heavily on the detailed characterization of the few largest known earthquakes. This study focuses on the two strongest historical earthquakes in the area, macroseismic data from Bosnia and Herzegovina, Croatia, Hungary, and Serbia. The number of intensity observations for the earthquake of 1922 was expanded from 106 to 278, whereas the previously macroseismically not analysed event of 1924 is decribed by 14 data points. Using a modified Kövesligethy–Jánosi model that accounts for intensity anisotropy in the epicentral area, we inverted the macroseismic fields to relocate the epicentres and estimate focal depths and magnitudes. Both events were relocated near the village of Zmajevac, within the Bansko Brdo tectonic unit, close to its boundary with the Drava depression. The 1922 epicentre moved 13 km north-northeast of the original location in the which occurred on 24 November 1922 and 12 August 1924. We re-evaluated these events using newly collected Croatian Earthquake Catalogue (CEC), while the 1924 epicentre shifted 22 km westward. Revised moment magnitudes are Mwm 5.3 and Mwm 4.4 for the 1922 and 1924 events, respectively. Estimated focal depths are shallower than previously listed: 11 km and 8 km, compared to the 18 km and 14 km in CEC. These results indicate that significant seismicity in Baranja is confined to the Bansko Brdo unit, with no evidence of strong earthquakes or faults with sufficient seismogenic potential in the Drava Depression or Northern Baranja–Bačka units. This has important implications for regional seismic hazard estimates. Furthermore, we find no instrumental support for the largest catalogued aftershock of the 1922 event and propose its removal. Finally, we interpret the 1924 earthquake as a late aftershock of the 1922 mainshock, suggesting a dependent relationship between the two.


  • Open Access

    Bayesian Reassessment of Seismic Moment Tensors and Their Uncertainties in the Adriatic Sea Region

    by Jinyin Hu, Hrvoje TkalčiĆ, Thanh-Son Phạm, Marijan Herak, Iva Dasović, Marija Mustać Brčić

    The determination of seismic moment tensor (MT) parameters is subject to uncertainties from data noise and structural error due to the imperfect Earth model, which is rarely considered in regional earthquake catalogs. In this study, we apply a hierarchical Bayesian MT inversion with uncertainty quantification to seven moderate-earthquakes Mw 4.5–5.5 in the Adriatic Sea region. The event collection includes three in mainland Croatia: the 2020 Mw 5.4 Zagreb earthquake and its Mw 4.9 aftershock, and the Mw 5.0 foreshock of the 2020 Petrinja earthquake, two events in the offshore Adriatic Sea: the 2021 Mw 5.2 central Adriatic earthquake, the 2024 Mw 4.6 southern Adriatic earthquake, and two in Italy: the 2022 Mw 5.5 Costa Marchigiana-Pesarese earthquake, and the 2023 Mw 4.9 earthquake in Marradi (Tuscany). The inversion output features the source depth and the posterior distributions of the MT parameters, enabling the uncertainty quantification. Comparing our results with regional routine catalogs highlights the improvement in source determination, particularly in confidence of non-double-couple components when incorporating the data and structural uncertainties. The refined source mechanisms could be useful for understanding the complex geological settings, assessing the hazard potential, and further improving the regional earthquake catalogs in the Adriatic Sea region.


  • A Comparative Study of Building Damage in Ston, Croatia, Caused by the Earthquakes of 1850 and 1996

    by Marijan Herak, Davorka Herak

    The small medieval city of Ston in southern Dalmatia (Croatia) has been hit by several strong earthquakes since mid-nineteenth century. The two most important are the ones from the years of 1850 and 1996. Although various aspects of the 1996 event have been well studied so far, the earthquake of 1850 has only recently been macroseismically analyzed. It turned out that the macroseismic epicenters, focal depths, and the epicentral intensities of the two events are virtually the same. As the categories of damage to buildings in Ston caused by the 1996 event were available from previous studies, we here combine details on damage and property ownership from reports in 1850 with the cadastral records from 1837. This allowed us to geolocate most of the damaged houses and thus directly compare the spatial distribution of damage from the two earthquakes. Although the building identification was not straight-forward and unambiguous due to unknown history of each building during the 13 yr between the cadastral survey and the earthquake, the overall damage distributions of both events are found to be similar. They both show the largest damage confined to the plain terrain below the Bartolomija hill characterized by a sedimentary cover 10–30 m thick and expected ground-motion amplification by factors of 3–5. Minimal damage for both events is observed on the hillslopes of Bartolomija in the northern part of the city, where the bedrock is shallow or outcropping. To our knowledge, this observation of the shaking effects for two strong similar earthquakes in a city that has changed little in the 146 yr between them is the only one of its kind in Croatia. It confirms consistency of spatial distribution of earthquake ground-motion amplification for comparable input earthquake motion.


  • Open Access

    Croatian catalogue and database of focal mechanism solutions, characteristic mechanisms, and stress field properties in the Dinarides and the surrounding regions

    by Marijan Herak

    This report presents the CroFMS catalogue and database of focal mechanism solutions (FMS). It is based on the first-motion polarity data for earthquakes in Croatia and the neighbouring regions, collected over last four decades. The current catalogue version contains FMS for 410 earthquakes that were computed consistently, using the same programs, velocity models, and sets of parameters. Its content is described in terms of distributions of the time and origin of polarity picks, magnitudes, epicentral distances of stations, the quality marks, the phases used, azimuthal gap and distance to the closest station. It was also shown that solutions from the CroFMS catalogue and those independently obtained from the moment tensor inversions are generally consistent.Comparison of observed FMS with the latest European model of active fault sources (EFSM20) revealed several inconsistencies, most notably in the Zagreb and Petrinja epicentral areas, the greater Rijeka region, the Central Adriatic archipelago, and in the zone of surface traces of the thrust fronts of the East-Bosnian–Durmitor and Drina–Ivanjica nappes in Bosnia and Herzegovina.Spatial averaging of FMS over the whole study area produced maps of characteristic focal mechanisms, which may be found useful in the seismic hazard assessment in the areas where active faults are not identified or characterised.A formal stress inversion of the focal mechanism solutions produced a map of maximum horizontal stress orientations across the study area, revealing localized lateral variations that had not been identified in previous studies.


  • Open Access

    Investigation of the Upper Crustal Structure in the NW Dinarides Using Local Earthquake Tomography

    by Gregor Rajh, Josip Stipčević, Mladen Živčić, Marijan Herak, Andrej Gosar, the AlpArray Working Group

    The area of the NW Dinarides lies in the northeastern corner of the Adriatic microplate – a promontory of the African plate – which plays a major role in the collision processes occurring in the peri-Adriatic region. Taking advantage of the greatly increased amount of data due to the modernization of the Slovenian seismological network, improved cross-border data exchange, and the tempoprary deployments within the large pan-european AlpArray project, we performed 3-D simultaneous hypocenter–velocity inversions with routinely picked arrival times of P- and S-waves from 502 local earthquakes. The resolution analysis showed that the resulting 3-D P-velocity and vP/vS model could be successfully recovered at least up to a depth of 15 km. Our results are consistent with previous studies, but provide better resolution in the upper crust and, for the first time, a uniform and simultaneously computed vP/vS model for the entire study area. The final 3-D models show two distinct velocity anomalies with a narrow transition zone in between. The velocity anomaly in the west of our study area reflects the thrust structure that formed due to the underthrusting of the Adriatic microplate, while the anomaly in the east marks the transition to the Pannonian basin. The recovered velocity distribution also appears to correlate spatially with seismicity, which is tied to the Moho topography at depth. The seismicity relocated with the computed models is on average shallower and has better determined locations than the seismicity located with the 1-D velocity model used in the daily earthquake analysis.


  • The Earthquake of 13 April 1850 near Ston, Croatia: Macroseismic Analyses

    by Davorka Herak, Marijan Herak, Iva Vrkić

    The earthquake that occurred near the city of Ston, Croatia, on 13 April 1850 is, together with the one from 1996, the strongest known event in the northwest part of the Dubrovnik epicentral area. This is the region with the highest seismic hazard in Croatia with a rich history of damaging earthquakes. Although listed in the relevant catalogs, this earthquake has never been addressed by a dedicated study. Herewith, we present analyses of a wealth of newly found material related to the damage and postearthquake actions of the authorities of the Province of Dalmatia, then a part of the Austrian Empire. We were able to estimate intensity at five localities, with a further six where the data were sufficient only to constrain the minimum intensity value. By far, most of the data refer to Ston and Dubrovnik. Intensity data points were inverted for the source parameters by two different methods, each of which yielded similar results. The focus is macroseismically located about 7 km east‐southeast from Ston, at a depth of 9 km. Estimated epicentral intensity of 8.2 on the European macroseismic scale is equivalent to macroseismic local magnitude MmL=6.0 or the moment magnitude Mmw=5.9⁠. The location of focus and the epicentral intensity are practically identical to those of the Ston–Slano earthquake of 1996. This is why we propose that these two earthquakes share the same composite seismogenic source consisting of a set of imbricated mostly reverse faults related to the basal thrust of the Dalmatian tectonic unit. The reliable location and quantification of the 1850 earthquake should contribute to a better understanding of the active dynamics of the set of large seismogenic faults in the Dubrovnik epicentral area.


  • Conversion between the local magnitude (ML) and the moment magnitude (Mw) for earthquakes in the Croatian Earthquake Catalogue

    by Marijan Herak

    Based on 153 earthquakes (1959–2020) listed in the Croatian Earthquake Catalogue, a conversion relation was obtained between the local magnitude ML,CR and the corresponding moment magnitude Mw as reported by the global and regional agencies. As errors were present in both variables the York regression was used. The best fit line is given by: MwL = (–0.106 ± 0.122) + (1.002 ± 0.027) ML,CR (coefficient of determination R2 = 0.90). The earthquakes considered occurred in Croatia and the neighbouring regions, and their local magnitudes ML,CR ranged between 3.5 and 6.5. Residual analysis suggests that an artificial positive magnitude shift of up to 0.3 magnitude units may have occurred in the early 1980s, when Wiechert mechanical seismographs were replaced by the instruments with velocity proportional recordings without proper recalibration of the magnitude formula. The slope of the regression close to 1.0 indicates that on the average the faults’ aspect ratio (width/length) is about 1/2.