Surface Protection Technologies for Earthen Sites in the 21st Century: Hotspots, Evolution, and Future Trends in Digitalization, Intelligence, and Sustainability

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript contains multiple redundancies and should be revised for greater precision and conciseness. Repetitive content affects overall readability, and acronyms should be defined only once, at their first appearance.

The English structure and clarity also require attention, particularly in passages such as Page 3, Lines 135–137, where phrasing is awkward and redundant.

The section describing the GPT-assisted analysis lacks sufficient methodological transparency. The authors should clearly define the evaluation criteria for GPT outputs, specify how accuracy was assessed, and clarify the respective roles of the authors in the validation process.

The paper relies excessively on visual elements (18 figures in total) without offering adequate in-depth analysis or critical interpretation of the results.

The manuscript’s current length (45 pages) is excessive for a review article and undermines reader engagement. A more concise and focused presentation is strongly recommended.

Author Response

We sincerely thank you for your valuable comments and suggestions, which have significantly helped us improve the quality of our manuscript. Below is a detailed point-by-point response to your comments:

Comment 1:
The manuscript contains multiple redundancies and should be revised for greater precision and conciseness. Repetitive content affects overall readability, and acronyms should be defined only once, at their first appearance.

Response 1:
Thank you for this helpful suggestion, which we fully agree with. We have systematically removed redundant content throughout the manuscript to enhance precision and improve readability. For instance, overlapping discussions on the significance of earthen sites in both the Introduction and Discussion sections have been consolidated, retaining only the essential background information. All acronyms are now defined at their first appearance—for example, “LiDAR” is defined as “Light Detection and Ranging” on Page 2, Line 64—and not redefined afterward.

Comment 2:
The English structure and clarity also require attention, particularly in passages such as Page 3, Lines 135–137, where phrasing is awkward and redundant.

Response 2:
We appreciate your pointing this out and fully agree. We have revised all instances of awkward phrasing, redundancy, and unclear sentence structures throughout the manuscript to improve clarity and fluency. To ensure linguistic accuracy and alignment with academic standards, the revised manuscript has been professionally edited by a native English-language scientific editing service. This included grammar, syntax, and terminology review, aiming to ensure consistency and clarity across the entire paper. We hope the revised version meets your expectations.

Comment 3:
The section describing the GPT-assisted analysis lacks sufficient methodological transparency. The authors should clearly define the evaluation criteria for GPT outputs, specify how accuracy was assessed, and clarify the respective roles of the authors in the validation process.

Response 3:
Thank you for highlighting this important issue. Following your advice, we have revised Section 2.2 ("Research Method") on Page 8, Lines 337–344, to include detailed information on the transparency and validation process of GPT-assisted semantic mining. Specifically, we implemented a three-tier validation protocol to ensure analytical accuracy:

1. Evaluation Criteria: Thematic outputs from GPT were benchmarked against domain-specific core literature, requiring a match rate of no less than 90% and correct use of technical terms.
2. Accuracy Assessment: Two independent researchers cross-validated GPT-generated keywords with results from manual content analysis, calculating an error rate (<5%), and categorizing error types (e.g., term omission, semantic mismatch).
3. Author Involvement: The authors manually refined and aligned GPT outputs that were not contextually accurate. In cases where terminology interpretation deviated from established domain usage, authoritative literature was incorporated as supplemental training material, enabling multiple iterations of deep reasoning until the outputs aligned with the contextual logic of earthen site surface conservation. This ensured that all GPT results met the predefined evaluation criteria.

Comment 4:
The paper relies excessively on visual elements (18 figures in total) without offering adequate in-depth analysis or critical interpretation of the results.

Response 4:
We sincerely appreciate your insightful feedback and agree with your assessment. Accordingly, we have conducted a thorough critical analysis of the results presented in the figures. For example, Figure 10 (Hotspot Journal Visualization of Earthen Site Surface Protection, Page 18) is now accompanied by a more nuanced interpretation that shifts the focus from general hotspot areas to the evolving influence of specialized journals.
Similarly, for Figure 13 (Keyword Co-Occurrence Cluster Analysis, Page 21), we now provide a deeper discussion of inter-cluster dynamics—for instance, how “climate change” (Cluster 2) interacts with “sustainability” (Cluster 3) in shaping innovative conservation strategies, rather than merely listing thematic clusters.
This critical approach is also reflected in our discussion of the well-known Chinese site of the Western Xia Mausoleum. On Page 31, Lines 1018–1023, we added the following analysis:
As a representative earthen site in a semi-arid zone, the Western Xia Mausoleum has implemented a multi-indicator monitoring system based on remote sensing (RS), geographic information systems (GIS), and the Internet of Things (IoT), validating the multiscale applicability of LiDAR and GIS technologies. However, from a critical standpoint, while the system integrates multiple technologies, it lacks interoperability between datasets, community participation, and ecological impact assessment. This highlights a common limitation in China’s digital conservation of historical sites: the focus on technological accumulation at the expense of systemic integration. These challenges underscore the need for a “technology–ecology–community” triadic framework in future conservation efforts.

Comment 5:
The manuscript’s current length (45 pages) is excessive for a review article and undermines reader engagement. A more concise and focused presentation is strongly recommended.

Response 5:
We appreciate your feedback regarding the manuscript’s length. The current length is the result of deliberate planning to enhance the paper’s uniqueness and scholarly contribution. Nonetheless, we understand the concern that a 45-page review may impact reader engagement. We would like to clarify that this manuscript presents a comprehensive overview of the technological evolution in earthen site surface protection over the past 25 years (2000–2025), progressing from point-based monitoring to spatial modeling and eventually intelligent integration.
It also details key technologies (e.g., improvements in LiDAR precision, the application of BIM in structural protection) and their iterative relationships to clarify phase-specific advancements and avoid oversimplification. Moreover, in order to articulate our proposed “Digitization–Intelligence–Sustainability” framework, we extensively elaborated on interdisciplinary linkages (e.g., alignment with SDG 11.4, the role of blockchain in digital certification). These connections required adequate contextual explanation, as they represent the manuscript’s core contributions. We hope this justification addresses your concern.

Please let us know if any further revisions are required. We once again thank you for your time and constructive feedback.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The key contribution of this paper lies in its comprehensive bibliometric and semantic analysis of surface protection technologies for earthen sites from 2000 to 2025. This paper advances the field by transforming fragmented studies into a systematized, technology-forward roadmap for preserving earthen heritage. It shifts the paradigm from static protection to dynamic, intelligent, and sustainable conservation.

These are some of my comments to further improve the paper since it is a bibliometric paper:

The article includes a large number of references across multiple countries and technologies, drawing from both foundational and recent studies. Nonetheless, the citation style could be improved for clarity in some tables, and a few references lack full citation details in the current draft. For eg:

• Table 1 While the table lists various studies, the limitations are without full source identification. For readers to validate or explore these references, a clear mapping between the numbered citations and full bibliographic entries is needed.
• Table 3  the reader must search manually for each number in the References list. It would improve clarity if at least the authors’ names and years were also shown in the table or noted in the caption for easier lookup. Put all the references concerned in each phase.

References should be checked for completeness. In the current version, some appear to be missing standard components.

Some references in the bibliography list use inconsistent formatting styles (e.g., inconsistent use of italics for journal names, missing publisher information, or inconsistent date placements).

In the introduction and conclusion, GPT-assisted semantic mining was repeatability explained.

Perhaps for clarify can separate them into subsection of (1) novelty of GPT use, (2) implications for methodology, and (3) generalizability to other fields.

Overall, this is a well-executed and valuable contribution to the literature on heritage conservation, especially in its interdisciplinary approach linking digital, intelligent, and sustainable paradigms.

Author Response

We sincerely appreciate your detailed feedback, which has provided clear guidance for our revisions. The specific changes made in response to your comments are as follows:

Comment 1:
Table 1: While the table lists various studies, the limitations are without full source identification. For readers to validate or explore these references, a clear mapping between the numbered citations and full bibliographic entries is needed.

Response 1:
Thank you very much for your valuable suggestion. In response to this comment, we have revised Table 1 (Page 25) by adding complete reference numbers corresponding to each cited study and its stated limitations. This allows readers to directly cross-reference with the full bibliographic entries in the reference list for further validation or exploration.

Comment 2:
Table 3: The reader must search manually for each number in the References list. It would improve clarity if at least the authors’ names and years were also shown in the table or noted in the caption for easier lookup. Put all the references concerned in each phase.

Response 2:
Thank you for your constructive suggestion. We have revised Table 3 (Page 28) by including the authors’ names and publication years for each key study cited within the table. Furthermore, all references associated with each phase have been summarized in the table caption, eliminating the need for readers to manually search through the reference list.

Comment 3:
References should be checked for completeness. In the current version, some appear to be missing standard components. Some references in the bibliography list use inconsistent formatting styles (e.g., inconsistent use of italics for journal names, missing publisher information, or inconsistent date placements).

Response 3:
We sincerely thank you for this important observation. All references have been carefully reviewed and revised to ensure completeness and consistency. Missing components have been supplemented, and formatting has been standardized in accordance with the journal’s citation style requirements.

Comment 4:
In the introduction and conclusion, GPT-assisted semantic mining was repeatedly explained. Perhaps for clarity, it can be separated into subsections of (1) novelty of GPT use, (2) implications for methodology, and (3) generalizability to other fields.

Response 4:
Thank you very much for this insightful suggestion, which has helped us improve the clarity and structure of the paper. In response, we have restructured the explanation of GPT-assisted semantic mining as follows:

1. Novelty of GPT use:
   On Page 8, Lines 346–349, we have added a detailed explanation of the methodological innovations of GPT-assisted semantic mining. These include:
   (1) Overcoming the limitations of traditional quantitative tools by enabling deep semantic association analysis beyond data aggregation;
   (2) Constructing a closed-loop system of "manual reading – GPT extraction – error calibration" (with an error rate <5%) to mitigate semantic deviations in machine-only analyses;
   (3) Efficiently integrating multi-source data to improve analytical efficiency and support rapid identification of core themes and trends.
2. Implications for methodological framework:
   On Page 25, Lines 843–848, we have elaborated on how GPT-assisted semantic mining contributes to the methodological foundation of the study in three ways:
   (1) Addressing the weakness of conventional metrics that emphasize data over logic, and revealing the intrinsic drivers of the “digitization–intelligence–sustainability” fusion through semantic linkage;
   (2) Bridging disciplinary gaps—for example, using GPT to explore the association between meteorological data and material performance studies, thereby identifying conceptual disconnects and constructing an interdisciplinary degradation prediction model;
   (3) Extracting and quantifying core findings across literature, which supports and validates the proposed three-phase evolution framework (point monitoring – spatial modeling – intelligent integration), thus enhancing the robustness of our conclusions.
3. Generalizability to other fields:
   On Page 38, Lines 1163–1168, we have added content discussing the broader applicability of this method. Specifically, GPT-assisted semantic mining can be adapted to:
   – Urban planning, by analyzing the correlation between sustainable development goals and technological integration to address multi-technology coordination bottlenecks;
   – Rural revitalization, by identifying patterns in the integration of industry and cultural heritage to support region-specific development strategies;
   – Interdisciplinary research, by identifying theoretical gaps and technological deficiencies, thus promoting a shift from fragmented studies to coherent systemic frameworks.

Please let us know if any further clarifications are required. We are once again grateful for your careful review and constructive suggestions, which have significantly contributed to the improvement of our manuscript.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This review paper synthesizes the most recent academic production on conservation of earthen sites from a technological angle. It does so by analyzing bibliographic /-metric data from the two most credible databases with the support of equally mainstream visualization software. The reasoning starts by observing the great variety of architectural structures worldwide, their conservation status, and the universal guidelines for their sustainability. The research framework is based on AI-supported selection, meaning database content was critically segmented (and manually confirmed). One main advantage of this paper is the global scope of academic production and the identification of trends, rather than merely describing output categories. Especially bursts in citation numbers, in addition to publication, reflect research density. The geographical inferences on publication volume and country-based research confirms findings in other fields (one visual detail: on Figure 4, the United Arab Emirates are incorrectly identified in West Africa). From a bibliographic standpoint, the evolution of keywords over time illustrates a more complex approach to conservation of earthen structure surfaces and currently almost always includes technological aspects, in addition to climate change/SDG or policy angles, regardless of the specific geography or tangible. Outlooks on the future are taken as increasingly integrated with AI innovation, aiming at prevention more than restoration. It is a very comprehensive study that goes beyond several other reviews, with are tendentially of a more descriptive nature. In contrast, this paper offers a methodological exercise on bibliometric datamining and offers insights for the immediate future.

Author Response

We sincerely appreciate your recognition of the comprehensiveness and methodological contributions of our research. We are also truly grateful for your positive evaluation of our work, which has greatly encouraged us to further improve this manuscript.

Comment 1:
On Figure 4, the United Arab Emirates are incorrectly identified in West Africa.

Response 1:
We sincerely apologize for this oversight and thank you for pointing it out. Based on your valuable feedback, we have revised Figure 4 (Global Publication Landscape in the Field of Earthen Site Surface Conservation, Page 11). The location previously mislabeled as “United Arab Emirates” has now been corrected to “Ghana” in accordance with the accurate world map geography. We are grateful for your careful review and attention to detail.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Most of my comments have been addressed. However, a few issues remain: each acronym should be defined only once at its first mention: for example, "light detection and ranging (LiDAR)" and "Internet of Things (IoT)" are unnecessarily redefined multiple times throughout the text. And I still find the paper overly long, which affects its readability.

Author Response

We sincerely appreciate the valuable feedback, which has significantly helped improve the manuscript. Below are the detailed revisions made in response to the comments:

Comments 1:[Each acronym should be defined only once at its first mention: for example, "light detection and ranging (LiDAR)" and "Internet of Things (IoT)" are unnecessarily redefined multiple times throughout the text.]

Response 1: [We sincerely appreciate your valuable feedback. We carefully reviewed the article to ensure that all abbreviations are defined for the first time they appear. For instance, "LiDAR" is defined as "Light Detection and Ranging", and "IoT" is defined as "Internet of Things". There will be no further redefinitions thereafter. Thank you very much for your suggestion.]

Comments 2:[ I still find the paper overly long, which affects its readability.]

Response 2:[We are extremely grateful for your valuable suggestions. We have removed the repetitive and redundant content. This is specifically manifested in, for example, the repetition of the three-stage division (exploration period, theoretical construction period, technological innovation period), and the detailed listing of "the definition and international cases of earthen ruins (such as Babylon, Memphis Pyramids, etc.)" in the Introduction. We have focused on the core mainline of "the digitalization, intelligence, and sustainability integration of surface protection technologies for earthen ruins in the 21st century", enhancing the compactness and readability of the text. We sincerely hope that the revised article will be accepted by you.]