Review of Manuscript

The manuscript is written in a high standard and is well structured. The scientific quality is high and the study makes a valuable contribution to fault-based seismic hazard assessment. I therefore recommend acceptance and publication of the paper, subject only to very minor technical corrections outlined below.

Further, to a lack of specific expertise in some methodological aspects, I am unable to provide a detailed technical evaluation of the applied methods. Consequently, this review focuses primarily on the scientific content, conceptual framing and quality of the writing.

Comments

Introduction

1) The introduction is very well written and no changes are suggested regarding writing quality. However, indeed the target study area is Italy, however recent earthquakes such as the 2023 Türkiye–Syria earthquake sequence represent a highly relevant example of multi-fault rupture behavior. Including this event and related studies in the references particularly in the context of multi-fault rupture scenarios discussed in the introduction would strengthen the relevance of the manuscript. Several studies addressing such rupture complexity should now be available.

2) The introduction provides a solid overview of fault-based PSHA approaches; however, the scope would benefit from a clearer conceptual distinction between traditional (area-source-based) PSHA and fault-based PSHA. For example, are there specific uncertainties that are better addressed by one approach compared to the other? It would be useful to include quantitative examples or clearly stated metrics (e.g. rates) illustrating the benefits and limitations of fault-based PSHA relative to traditional approaches.

Typos and minor issues

Line 115: Its probably meant to be “coseismic”

Table 3: In the title: should be “background” (check spelling)

Figures 7–12: Figure resolution is poor. The resolution limits readability even with a bit of zooming into the document.

The manuscript demonstrates a strong command of the relevant literature and provides a coherent narrative linking fault-based PSHA, multi-fault rupture scenarios, and seismicity rate modelling in the southern Apennines. The Discussion, in particular, is detailed and carefully reasoned, with appropriate acknowledgement of model limitations.

However, introduction and discussion sections tend to be descriptive rather than argumentative, and the manuscript frequently presents results without fully developing their broader scientific or methodological implications. As a consequence, the overall contribution risks being perceived as a well-executed application rather than a study that significantly advances PSHA practice. Therefore, a major revision is recommended



Major comments

1 - The Introduction outlines well-known limitations of catalogue-based PSHA and motivates the use of fault-based and multi-fault rupture approaches. However, the Discussion does not clearly close the loop by explicitly demonstrating how the presented results resolve, reduce, or reframe the issues raised in the Introduction. Yet, the manuscript stops short of articulating what new insight this study provides for that debate, beyond confirming that multi-fault ruptures perform better in this case study.

2 - The Discussion correctly acknowledges that SHERIFS is not a physics-based model and therefore cannot assess rupture compatibility or stress interactions. While this caution is appropriate, it also weakens the strength of the conclusions regarding the feasibility of large multi-fault events (e.g., Mw 7.6). The manuscript would benefit from a clearer distinction between statistical consistency with observations and physical plausibility of rupture scenarios. At present, these two concepts are sometimes conflated, which may lead readers to overinterpret the robustness of the conclusions.

3 - Section 4.5 proposes explicit weighting of rupture sets and modelling branches for PSHA logic trees. While this is one of the most practically relevant outcomes of the study, the Discussion does not sufficiently address: the transferability of these weights beyond the study area, whether the proposed weighting scheme is meant as a general methodological template or as a region-specific recommendation. This ambiguity reduces the impact of what could otherwise be a key contribution to hazard modelling practice.

4 - The distinction between SubArea 1 and SubArea 2 is well motivated and clearly described. However, the underperformance in SubArea 2 is largely attributed to data limitations and simplified fault representations, which raises the question of whether the poorer model performance reflects tectonic behaviour or model-input inadequacy. This distinction should be more explicitly discussed to avoid misinterpretation of the results.

5 - The conclusion that background seismicity has a limited impact is well supported for the chosen buffer configuration. However, the Discussion itself acknowledges that this result is sensitive to buffer size and expert judgment. As written, the conclusions regarding background seismicity risk are interpreted as more general than justified by the tested configurations.

 “A fault-based application to model seismicity rates for seismic hazard assessment in the southern Apennines (Italy)” did the studies on multi-fault rupture possibilities, especially for tectonic region with strike-slip and normal faults, are very important to understand and evaluate the earthquake hazards. The follows are some questions which is helpful for the readers to understand the manuscript. 

1. The study adopts a relaxed segmentation framework for multi-fault rupture modeling, but the explicit criteria for defining fault segment boundaries (e.g., based on geometric discontinuities, kinematic differences, or tectonic domain divisions) are not clearly stated. Could you elaborate on the segmentation principles and provide quantitative justifications for the division of the 35 seismogenic faults?
2. For the slip rate parameters assigned to each fault (derived from DISS 3.3.1 and Valentini et al., 2017), there is a lack of analysis on the consistency between slip rates of adjacent faults in the same tectonic domain. Please clarify how slip rate variations across the fault system are reconciled with regional geodynamic constraints.
3. The introduction lacks detailed information on the regional tectonic background, which is essential to justify the strike-slip and normal faulting characteristics of the fault system in the southern Apennines.
4. For Figure 1 in Page 3, the slip rates of strike-slip faults lack consistency or continuity. Areas without mapped fault traces may correspond to segments where blind faults are developed. I suggest the atuhors either incorporate this consideration into the segmentation model or discuss it in detail in the Discussion section.
5. The three rupture scenarios (Set_0, Set_1, Set_2) are defined using different distance thresholds (5 km and 10 km) for rupture propagation, but the rationale for selecting these specific thresholds is not sufficiently justified. Could you provide references or quantitative analysis to support the choice of distance criteria for multi-fault rupture connectivity?
6. The study uses Wells & Coppersmith (1994), Leonard (2010), and Thingbaijam et al. (2017) scaling relationships for magnitude estimation, but it does not address whether these global or broad-scale models are fully applicable to the tectonic setting of the southern Apennines. Please supplement a discussion on the regional adaptability of the selected scaling laws, including potential adjustments for local tectonic characteristics.
7. Line 155, for the multi-fault ruptures, the difficulty in this area is not only the rupture combinations of the segments with the same rupture characters, but the rupture combinations of the segments with strike-slip and the ones with normal-slip. These combinations are hard to judge for multi-segment ruptures. For instance, the multi-fault rupture scenario involving Faults 18, 19, and 20 is likely to be geologically feasible. I suggest the authors to add these rupture combinations. For instance, the multi-fault rupture scenario involving Faults 18, 19, and 20 is likely to be geologically feasible. Such rupture combinations are fully mechanically explicable (Oglesby, 2005) and have been observed in numerous regions worldwide (e.g., the 1995 M2 Multi-Segment Nuweiba Earthquake), as stated by Oglesby (2005), and are also incorporated into the multi-segment rupture scenarios for the SHERIFS modeling framework (e.g., Cheng et al., 2025, NHESS).

Reference:

Oglesby. 2005. The Dynamics of Strike-Slip Step-Overs with Linking Dip-Slip Faults. Bulletin of the Seismological Society of America (2005) 95 (5): 1604–1622.

Li et al. 2025. Supershear Rupture of the 1995 Mw7.2 Multi-Segment Nuweiba Earthquake in the Gulf of Aqaba. Geophysical Research Letters,  doi.org/10.1029/2025GL117448.

Cheng et al. 2025. Modeling seismic hazard and landslide occurrence probabilities in northwestern Yunnan, China: exploring complex fault systems with multi-segment rupturing in a block rotational tectonic zone. Natural Hazards and Earth System Sciences, 25, 857-877.

1. The non-mainshock slip (NMS) ratio is employed as an indicator of model quality, yet this study fails to discuss how the NMS ratio varies with fault kinematics (i.e., strike-slip vs. normal faults) across the two tectonic sub-areas. The NMS ratios of normal fault systems tend to be higher than those of strike-slip fault systems, because the extensional stress environment can activate a greater number of small-scale faults compared with the shear stress environment. It is recommended that an analysis of NMS differences between different fault types, as well as their corresponding geological interpretations, be supplemented.