General Evaluation

This manuscript addresses a highly relevant and practical problem: evaluating strong ground motions in Izmir and implications for future earthquake simulations using GMPEs, site response analyses (SRA), and comparison with Turkish Earthquake Code (TEC) design spectra. The study includes an extensive dataset, thoughtful site-specific calibration, and a compelling case for re-evaluating seismic design spectra for alluvial zones in Izmir. The work is timely and important, particularly in the context of recent seismic events like the 2020 Samos Earthquake.

Although the manuscript requires several clarifications and presentation improvements, these revisions do not affect the core methodology or conclusions. Therefore, I recommend acceptance after minor revision.

General Comments:

These are not major scientific flaws, but addressing them will improve the clarity, quality, and impact of the paper:

1. Clarify Novelty and Contribution

A short paragraph clearly stating how this study builds upon or extends those provious works (e.g., broader dataset, new scenario, comparison with TEC, etc.) would be helpful.

1. Improve Figure Quality

Some plots (e.g., Figures 7–13) suffer from low resolution, small font sizes, and unclear legends. It is recommended to:

• Increase resolution to 300 dpi
• Use consistent color schemes and line types
• Clearly label axes and add proper legends
• Language Polishing

Minor grammatical errors and awkward phrasing (e.g., “data riched event”) should be corrected throughout the manuscript. A careful proofreading or light professional editing is encouraged.

1. Enhance Discussion of Uncertainty

While the GMPE selection and SRA analyses are rigorous, a brief paragraph discussing epistemic uncertainties (e.g., in soil properties, GMPE selection) would strengthen the credibility of the conclusions.

1. Cite Recent Hazard Models or Studies for İzmir

Since the RADIUS project dates back to 1997, it would be useful to mention whether newer seismic source models or hazard assessments (e.g., AFAD, SHARE) have been compared or considered.

1. I recommend moving Figure 1 and the related tectonic context from the Introduction into a separate section (e.g., “Seismotectonic Background”). Given its importance for ground motion characteristics and scenario development, the seismogenic framework deserves clearer, dedicated presentation rather than being embedded in the general introduction.

Detailed comments by section

Section 1.1 (Scope and Aim):

1. I recommend relocating Figure 1 and the associated tectonic context into a new subsection (e.g., “Seismotectonic Background”), as this content forms the geological basis for the entire study.
2. The introductory paragraphs are overly general. I suggest revising them to emphasize specific regional seismic characteristics (e.g., fault segmentation, recurrence, basin amplification) supported by references.
3. References such as Emre et al. (2018) and McKenzie (1978) should be more clearly contextualized—why they are relevant and how they support the regional hazard description.
4. Minor language edits are needed (e.g., “data-riched” → “data-rich”; restructure awkward phrases for clarity).

Section 1.2 (Geological and Geotechnical Settings of İzmir Bay):

1. This section provides useful context; however, the distinction between geological and geotechnical properties is vague. Consider separating lithological description (e.g., rock types, basin structure) from engineering properties (e.g., Vs30, soil classes, amplification potential).
2. Language could be improved for precision:

“Loosely consolidated sediments can exacerbate ground shaking” → “Loose alluvial soils with low stiffness can significantly amplify seismic waves, particularly in the 0.5–2 s period range.”

Section 2: GMPE Dataset and Model Comparison

1. The inclusion of 33 real earthquake events with source and PGA data is commendable. However, please clarify how site conditions were accounted for (e.g., Vs30, site class) in the GMPE comparisons. Were all sites considered as rock, soil, or site-adjusted?
2. Table 1 is rich, but consider:

• Reordering events chronologically or grouping by distance/magnitude for better pattern recognition.
• Adding a column for site class or Vs30 (if available).

1. Table 2 (GMPEs) is well-structured, but the rationale for including only NGA-West1 and NGA-West2 models should be briefly explained. Are there any Turkish/regional GMPEs that were excluded, and if so, why?
2. Equation formatting (RMSE) is clear, but please define variables (e.g., N, i) inline or as subscript to improve readability.
3. Figure 3 (scatter plots):

• R² values are useful, but consider adding residual plots or at least comment on residual trends (e.g., underprediction for short distances or large PGA?).
• Improve axis labeling (units missing or too small).

1. The error metrics in Table 3 clearly identify CB14 and BSSA14 as the most appropriate models. Still, a brief sentence on whether residual bias was directionally consistent across magnitude or distance ranges would strengthen the justification.

Section 3: Site Response Validation

1. The use of the 2020 Samos earthquake for model calibration is appropriate. However, the phrase “data-riched event” should be corrected to “data-rich event.
2. The selection of stations (3513, 3519, 3522) is logical and well-supported. Still, please clarify the criteria for choosing station 3514 as the outcrop reference. Is it based on rock classification by Vs30? A sentence confirming this would help.
3. Tables 4 and Figure 5 provide essential input data, but a clear indication of soil class (e.g., ZC, ZD, ZE) and correlation with TEC site categories would enhance interpretation.
4. The geotechnical profile descriptions and the use of modulus reduction/damping curves (Figure 6) are acceptable. However, it would be beneficial to state whether site-specific laboratory curves were available or if literature-based curves (Seed & Idriss, Vucetic & Dobry) were adopted uniformly.
5. Amplification results (SRA/Outcrop) are well presented (Figures 7–9), but legends and axes in the plots should be improved for clarity (e.g., distinguish observed vs. modeled more clearly).
6. The statement that amplifications match well between 0.5–1.5 s is important (lines 218–220) and should be connected to predominant building periods in İzmir (e.g., 6–10 story RC buildings) more explicitly.
7. It would strengthen the validation to include a quantitative metric (e.g., goodness-of-fit score, misfit index, or confidence range) to support the visual matching of recorded and simulated response spectra.

Section 4: Target Spectrum and Future Scenario Analysis

1. The inclusion of near-fault effects and rupture directivity (e.g., Somerville et al., 1997) is a strong contribution. However, this discussion (lines 234–246) could benefit from a short numerical explanation or quantification of how much this effect modifies the spectrum (e.g., amplification factor at 1–2 s).
2. Figure 10 clearly compares different spectra, but the legend could be improved for readability (e.g., use thicker lines for "Target Spectrum" and "TEC/ZO" to differentiate them more clearly).
3. The authors selected 11 records (Table 5), but the scaling procedure could be described more precisely. Please clarify: Whether spectral matching or simple scale factor was used.  Whether each record meets compatibility criteria (e.g., ASCE 7–22 or Eurocode 8).
4. Figures 11 and 12 present the scaled records for different sites. However, it would strengthen the section to include statistical metrics (e.g., mean ± σ, envelope exceedance rate, number of records exceeding the target at critical periods).
5. Figure 13 effectively compares median responses to TEC and revised spectrum. However, consider labeling the TEC soil class boundaries (ZD, ZE) more explicitly in the figure or caption for clearer comparison.
6. Table 6 shows valuable correction factors (F1/TEC vs. F1/SAR), but the implications should be briefly discussed: e.g., how these high ratios (e.g., 4.76 at Bornova) indicate potential underestimation of seismic demand in current TEC-based designs.

Section 5: Summary and Conclusions

1. The section successfully summarizes the methodology and major findings. However, the conclusions would benefit from more explicit quantitative insights (e.g., amplification ratios, return periods, percent deviations) rather than mostly qualitative descriptions.
2. The statement that “TEC factors should be corrected by at least 2.50 times” (lines 310–312) is significant. I recommend the authors provide a brief justification or numerical basis here (e.g., comparison table or reference to Figure 13 or Table 6).
3. The authors mention “buildings between 3–10 stories” (line 313), which is important in practice. However, linking this to code performance levels (e.g., life safety, collapse prevention) would enhance the engineering relevance.
4. The last sentence (lines 315–316) mentions that future studies will explore basin effects. While this is appreciated, I suggest specifying how the presented study lays the groundwork for this (e.g., validation of deep soil profiles, limitations of 1D analysis).
5. Overall, this section would benefit from a clearer separation between key conclusions and recommendations. Consider formatting them as two separate bullet lists (e.g., “Main Findings” vs. “Recommendations for Future Work”).

Final Recommendation: Minor Revision

The manuscript presents a valuable, technically sound, and well-structured contribution to seismic hazard evaluation and future earthquake simulation in İzmir. With minor improvements in figure clarity, language quality, and justification of novelty, the paper will be ready for publication.

This manuscript tries to evaluate ground motion characteristics and potential ground motions by future earthquake in Izmir city, Turkey. As Turkey is prone to large earthquakes, these evaluations are considered very important. Overall procedure seems probably reasonable, but as is already pointed out by Referee 1 the manuscript needs more explanation/indication and improvement of figure quality. The comments from Referee 1 are right on point, and particularly, I strongly agree with comments 1, 2, 4, 7, 10, 12, 13, 14, 20, 25, 27, 31. In addition, I would like to indicate four major comments. Therefore, I judge this manuscript major revision.

Major comments:

1. In Section 2, the author seeks for the most appropriate GMPEs for Izmir dataset considering PGA data.  However, in Section 4, response spectra are evaluated using the selected GMPEs.  Why don't the author evaluate the appropriateness of GMPEs using response spectra? It is necessary to show the validity of use of PGA for the evaluation of GMPEs.

2. It is difficult to understand how the author actually evaluates the amplification of three stations in Izmir in Section 3, because the description on the method lacks details, the resolution of figures is very low, and the figures do not have appropriate legends and/or captions. As far as I guess from the manuscript, the author obtains the SRA results by inputting ground motion recorded at 3514 site to underground structure model shown in Table 4 and Figure 5. However, underground structure models of 3513, 3522, and 3519 sites shown in Table 4 and Figure 5 do not reach bedrock, and therefore, I am wondering at which depth the bedrock ground motion is input, which very much affects site amplification.

3. The detailed analytic procedure of Section 4 is also vague. Similar to Section 3, the manuscript just names software for analysis and lacks detailed methodology.

4. Appropriate indication and acknowledgements to data resources need to be added. Particularly, data listed in Table 5 includes data from several observation networks or institutes. Acknowledgements to continuous efforts to obtain data support such networks and institutes.

Minor comments:

- Table 2: Style of fonts needs to be consistent in the table.

- p.8, line 166: "The city covers large areas of alluvial soil condition" seems grammatically incorrect. "The city is covered with large areas of alluvial soil condition" may be better.

- p.9, line 174: "data's" seems grammatically incorrect. "data" may be better.

- Figure 6: It is better to set the same size for top and bottom panels.

- Table 5: Style of the table needs improvements (distorted aspect ratio of fonts, inconsistent for centering or right-aligning, etc).