Dear Authors,

Please find my review report in the attachment.

Best regards.

This manuscript addresses a practically significant problem: refining the accuracy of global gravity field models in coastal areas using high-resolution topography and bathymetry data via the Residual Terrain Model (RTM) forward modeling method. The methodology is sound, the experimental design is robust, the data is substantial, and the results clearly demonstrate the effectiveness of the proposed approach. The study makes a clear contribution to improving gravity field models in coastal and rugged terrains. The paper is well-structured and clearly written. I recommend acceptance after addressing the following minor points.

(1) Accuracy of Figure References:

The text contains inaccurate figure reference. For instance, Section 3.4 refers to "Fig. 2" in the context of illustrating computation points, but based on the context, this should likely be "Fig 3". Please conduct a thorough check to ensure all figure and table citations are correct throughout the manuscript.

(2) Quantifying the Contribution of the Mass Center Correction:

The proposed mass center offset correction for the RET method is a valuable improvement. However, its quantitative impact and necessity are not currently demonstrated. Providing a simple comparative result would significantly strengthen the argument for its inclusion and help readers appreciate its contribution.

(3) Justification for Using the XGM2019e Model:

The XGM2019e model provided a version up to d/o 5399 (about 2” resolution), which theoretically contains higher-resolution signal. Please justify the choice of using only the d/o 2159 model for your modeling, or demonstrate whether the XGM/RTM-GA can achieve performance superior to that of the d/o 5399 XGM2019e model.

(4) Discussion on Computational Efficiency and Relation to Existing Models:

The method developed in this study is targeted at the challenging coastal areas. However, for land areas, high-resolution gravity field models like the SRTM2Gravity model by Hirt et al. (2019) already exist and could potentially reduce computational burden. While the current manuscript is complete, I would be interested in the potential for a hybrid approach in the future: leveraging existing models over land and focusing the RTM forward modeling presented here primarily on the coastal transition zone. A discussion on this prospect would be valuable.

(5) Validation Data Coverage and Diversity:

I note that Fig. 2 shows a notable absence of validation points in the coast area (about 20km). Could you comment on the availability of other datasets that could potentially validate the model in this critical zone? Additionally, while the NGS99 dataset is robust, incorporating additional independent validation data (such as shipborne data) would further strengthen the reliability and generalizability of the manuscript.