Grotto temple heritage in China embodies both cultural-artistic value and archaeological-scientific significance. As a quintessential medium integrating stone artifacts with painted murals, their conservation practices are undergoing a paradigm shift from “emergency restoration” [1] to “preventive conservation,” [2] urgently requiring the development of systematic research frameworks and interdisciplinary support. These grottoes not only materialize ancient religious beliefs, engineering craftsmanship, and societal civilization but also serve as pivotal specimens for deciphering historical climate evolution, craft dissemination routes, and cross-regional cultural interactions. However, their immovable nature and unique geological conditions subject them to long-term coupling effects from multiple stressors in the cave micro-environment, including temperature-humidity fluctuations [3], salt migration [4], pollutant infiltration [5], and biological colonization [6]. These induce comprehensive deterioration phenomena such as powdering, pigment flaking, and biogenic crust formation [7]-[9], severely threatening the structural safety [10], material integrity [11], and artistic value of cultural relics.

With the escalation of conservation needs, the field of grotto heritage preservation has developed an interdisciplinary research framework, and the volume of related literature continues to grow. However, existing research still faces a significant “fragmentation” bottleneck:

(1) Research perspectives predominantly focus on single-disease mechanisms or localized environmental factors [12][13], lacking a macro-level grasp of the holistic research framework encompassing “environmental drivers—technology application—conservation practices”;

(2) The application of interdisciplinary technologies (e.g., monitoring sensors, microbial regulation, numerical simulation) remains fragmented [14]-[16], with the evolutionary patterns and core hot spots of technical methodologies within the field yet to be clearly defined;

(3) Predictive and regulatory capabilities are weak [17]. In the face of micro-environmental evolution and cultural heritage responses potentially triggered by conservation engineering interventions (e.g., installing rain shelters, adjusting visitor flow routes) or future environmental changes (e.g., climate change, increasing tourism pressure), effective predictive tools and evaluation methods are still lacking, resulting in a lack of foresight in conservation decision-making.

Traditional review methods struggle to efficiently integrate massive volumes of literature and visually present the developmental dynamics of a field. As a core tool in bibliometrics, CiteSpace overcomes these bottlenecks with its unique advantages in analyzing research hot spot evolution, interdisciplinary pathways, collaboration networks among core authors or institutions, and literature clustering [18]. To address the mismatch between the “fragmented” nature of grotto conservation research and the demand for a paradigm shift toward preventive conservation—where traditional literature reviews fall short in synthesizing the macro-level research landscape—this review adopts a phased retrieval strategy based on the China National Knowledge Infrastructure (CNKI) database: beginning with broad-topic searches (e.g. “grottoes,” “grottoes + conservation/restoration”) to map the overall scope of the field, followed by precise composite queries focusing on core topics (e.g. “deterioration mechanism,” “digital conservation”). The disciplinary coverage includes archaeology, cultural heritage conservation, architecture, materials science, environmental science, digital technology, and other cross-disciplinary fields, and the literature screening is limited to academic publications such as journal articles, conference papers, and theses. By analyzing domestic and international literature on grotto conservation over the years, this review systematically reveals the temporal evolution of the field from empirical-qualitative to scientific-quantitative approaches, along with shifting research hot spots across five core directions (e.g. deterioration mechanisms, micro-environment monitoring).

Figure 1. Proportion of various disciplines in Chinese grotto research.

Figure 2. Publication trends in Chinese grotto studies.

The research plan retrieved the term “grottoes” from the CNKI database, yielding a total of 10,300 documents spanning 74 years from 1951 to 2025. The content primarily covers eight disciplinary fields, with archaeological research accounting for 38.5% (Figure 1). In terms of publication trends, around 1999, the number of publications began to increase sharply year by year (Figure 2). This surge is attributed to the launch of the “Dunhuang Conservation Plan,” which established a systematic and comprehensive conservation philosophy along with large-scale investments, thereby promoting the development of grotto research.

Using CiteSpace, a cluster analysis was conducted on the 10,300 documents, with the average silhouette value (S-value) and clustering modularity value (Q-value) serving as core indicators for evaluating clustering effectiveness. An S-value > 0.7 indicates high credibility, while a Q-value >0.3 suggests a significant clustering structure in the network. In this study, S = 0.9302 and Q = 0.694, demonstrating highly convincing clustering results. The analysis categorized the keywords from the documents into five clusters (Figure 3), with the content coverage of each cluster detailed in the accompanying table.

Figure 3. Keyword cluster analysis of Chinese grottoes research.

Cluster #0 exhibits a significant correlation with Cluster #2, which is manifested in viewing grottoes as carriers of the integration between Buddhism and political power. Gu Yanfang [19] revealed grottoes as a combination of religious rituals and political symbols through analyzing the Buddhist worship scene in the Huangfu Cave; Yang Yang [20] elucidated the localization process of Buddhist culture using the Yungang Grottoes as an example. Scholar Chang Qing’s research approach is particularly systematic. By treating the grottoes at Longmen [21] and in Shaanxi [22] as “stone evidence” [21], he spent years deepening his studies from stylistic analysis to political iconography [23][24], uncovering how medieval imperial authority manipulated Buddhist narratives through cave excavation and statue destruction, thereby establishing the research paradigm of grottoes as “Buddhist political media.”

Furthermore, Cluster #0 and Cluster #4 are inherently connected due to their shared category of “cultural heritage.” The research by Wei Qi [25] is pivotal in this shift. Through his analysis of the royal statues and Sinicization style of the Longmen Grottoes, he characterized them as “a treasure of national carving art,” thereby pioneering the conceptual leap of the research object from the physical “grotto” to the cultural “heritage.”

Table 1. Content covered by keyword clustering of Chinese grottoes.

As shown in Table 1, the 82 publications related to Cluster 0 and Cluster 1 can be categorized into the following four research directions:

Art History and Iconography Studies. This direction systematically examines the themes, styles, and chronology of visual materials such as cave sculptures and murals to establish “core value carriers” for conservation practices. For example, Li Jinyang [26] interprets the imperial legitimacy embodied in the “Dual Seated Buddhas” statues of the Yungang Grottoes, emphasizing that their preservation status directly correlates with the continuity of critical historical information. Li Jingjie [27] clarifies the origins of the iconographic compositions and pigments in the Nirvana Sutra transformation scenes at Dunhuang, providing an original historical and technical basis for mural restoration.

(2) Conservation Science and Technology, focusing on the deterioration mechanisms and intervention techniques for cave sites, adopting a research approach that integrates “experimentation, monitoring, and engineering.” Liu Xiangfeng [28] identified the critical conditions for macroscopic crack formation in Yungang sandstone through freeze-thaw cycle tests. Meng Tianhua [29] applied terahertz nondestructive testing technology to predict the onset of hollow detachment in painted layers, advancing the shift from “emergency repairs” to “preventive conservation.” Li Zuixiong [30] developed PS-C grouting materials that significantly improved the strength and salt resistance of crack repairs, successfully applied in reinforcement projects at the Yungang Grottoes.

(3) Digitalization and Information Applications, primarily utilizing technologies such as 3D laser scanning and digital twins to establish a closed-loop digital workflow of “high-precision acquisition—scientific archiving—public dissemination.” Related studies create permanent digital archives for cave sites, providing essential data support for physical conservation. For instance, Zhou Wensheng [31] created a “digital twin” for Cave 12 of the Yungang Grottoes using mobile laser scanning, offering a precise reference for construction reinforcement. Huang Zeyu [32] employed digital twin technology to enable real-time monitoring of environmental and structural changes within the caves.

(4) Cultural Exchange and Mutual Learning among Civilizations. This direction explores the reinterpretation of the cultural value of cave sites in contemporary contexts, emphasizing their interaction with cultural tourism integration, public education, and social identity.

Cluster 2 and Cluster 3 form a profound connection through the “visual translation of Buddhist thought” and the “adaptation of imagery for dissemination.” Representative studies demonstrate that the stylistic evolution of cave art vividly reflects the localization process of Buddhism: Chen Shaofeng [33] noted that from the Northern Dynasties to the Sui Dynasty, the sculptural techniques shifted from “shallow stepped relief” to fluid and natural forms, aligning with the need for the popularization of Buddhist rituals and doctrines; Zhao Liya [34] categorized the Buddha statues in the Kizil Caves into three styles—“Indian-Kizil-Han,” clearly illustrating the cultural fusion path during Buddhism’s eastward transmission; the study [20] systematically outlines the stylistic sequence of Yungang sculptures, transitioning from early Western Regions styles to the sinicized “broad robes and wide belts” imagery of the middle period, and finally to late-period styles incorporating folk elements, thereby bridging the connection between Cluster 2 and Cluster 3. These studies collectively establish a comprehensive interpretive framework of “caves as vessels—sculptures conveying doctrines—styles manifesting fusion,” deepening the systematic understanding of how Buddhism in the Northern Wei period utilized visual arts to achieve localization and ethnic integration.

As shown in Figure 4 (Top 25 burst terms generated based on CiteSpace), the research hotspots and frontier evolution in this field exhibit distinct characteristics. In terms of burst strength, the top three keywords are “Longmen Grottoes,” “cultural heritage,” and “dynasty.” Among these, “Longmen Grottoes” primarily showed high activity during 2008-2009, with research focusing on protection and restoration technologies, particularly addressing engineering challenges such as water leakage control [35], grouting materials [36], and the repair of collapsed niches [37]. This trend stems from both the prominent natural deterioration faced by the Longmen Grottoes as an open-air stone heritage site and the heightened conservation demands following its designation as a national 5A-level scenic area in 2007 [38]. The related achievements mark a significant shift in grotto conservation from empirical maintenance to scientific and systematic restoration [39]. “Cultural heritage” was mainly prominent during 2023-2025, driven by dual forces: first, national cultural strategies such as the “14th Five-Year Plan” elevating cultural heritage protection to a strategic level [40]; second, the maturation of digital technologies like 3D scanning, VR, and digital twins, which have facilitated high-precision archiving, virtual restoration [32], and immersive dissemination [41], significantly expanding its pathways and impact in preservation and transmission. “Dynasty” primarily surged during 2018-2019, with academic attention centered on Northern Dynasties grotto art, delving into the interplay between grotto excavation and Buddhist thought, as well as political contexts during the Northern Wei, Northern Qi, and Northern Zhou periods [42][43]. Research themes included the spread of the “Decline of the Dharma” concept and scripture engraving [43], as well as the origins of sacred tree imagery [44]. This trend may have been influenced by significant archaeological discoveries at the time (e.g., the Ye City ruins [43][45]) and the promotion of related academic conferences and projects, reflecting a research orientation that interprets the evolution of Buddhist art from a dynastic perspective.

In terms of sustained keyword popularity, “backlight,” “Silk Road,” and “water Guanyin” have maintained research attention for approximately four years, while “condensation water” has the longest burst period (five years), indicating the academic community’s continued emphasis on this topic. Condensation water, as a significant moisture source leading to the weathering of stone cultural relics, has been systematically studied since 2008. Zhu Hua [46]et al. were among the first to analyze its formation mechanism and its impact on cave weathering, sparking initial academic focus. By 2012, significant progress had been made in related research regarding observation, mechanisms, and prevention: Wan Li [47] proposed quantitative observation methods and revealed its temporal dynamic characteristics; Zhang Ao [48] clarified the chemical weathering effect of condensation water coupled with CO₂ on carbonate rocks through simulation experiments; Ren Jianguang [49] explored the inhibitory mechanisms of cave construction features on condensation water; Huang Jizhong [50] emphasized the critical role of condensation water in water-related hazards through investigations into moisture sources. These achievements collectively promoted the sustained deepening of this topic. Subsequently, research on condensation water further advanced toward technological development. For instance, Ma Ce [51] applied infrared thermal imaging technology in 2022 to reveal the dynamic formation patterns and infiltration mechanisms of condensation water during rainfall, expanding the application pathways of non-destructive monitoring methods.

Figure 4. Keyword burst graph of Chinese grotto studies.

In the CNKI database, using the search terms “grottoes” & “protection” or “grottoes” & “restoration,” a total of 209 articles were retrieved. Through systematic review, the relevant research can be categorized into four main directions: “Research on the Deterioration Mechanisms of Grotto Cultural Relics,” “Micro-environment Monitoring and Assessment Techniques for Grottoes,” “Research on Homo sapiens Activity Interference,” and “Digital Protection Theories and Technologies,” accounting for 31.4%, 20.0%, 11.4%, and 37.1% respectively (Figure 5).

Figure 5. Proportion of research directions in the conservation and restoration of Chinese grottoes.

Although China has made significant progress in both theoretical and practical aspects of grotto cultural heritage conservation research, particularly accumulating abundant achievements in deterioration mechanisms, micro-environment monitoring, and digital preservation, a systematic review and bibliometric analysis using Citespace still reveal several critical gaps and urgent research needs that require attention in current studies. These are specifically reflected in the following aspects:

(1) The research perspective remains “fragmented,” lacking systematic and holistic integration.

Current studies predominantly focus on single disease types, localized environmental factors, or specific technological approaches. While these have yielded in-depth insights at the micro level, they have failed to effectively establish a systematic research framework for rock mass deterioration under the coupled effects of “environmental drivers—biological responses—anthropogenic disturbance.” The deterioration of grottoes often results from nonlinear interactions among multiple factors:

Taking the study of the North Grottoes Temple as an example [119], the findings clearly reveal the multi-layered and dynamic interactions between environmental and biological factors. The research confirms that the increase in relative humidity inside the caves (e.g., after the construction of protective shelters) directly promotes the proliferation and metabolic activity of insects and microbial communities on sandstone surfaces. The organic acids secreted by microorganisms accelerate the dissolution of sandstone cement, while the biofilms they form alter the physical properties of the rock surface, collectively establishing a “humidity-driven—bioactivation—accelerated corrosion” coupling cycle. Furthermore, water infiltration and soluble salt migration not only cause physical crystallization damage but also create microhabitats for halophilic microorganisms. Microbial metabolism, in turn, modifies the local chemical environment, influencing the crystallization process and destructive effects of salts, forming a chain reaction intertwined with water-salt transport and biological activity. Notably, external interventions aimed at blocking rainwater (e.g., constructing rain shelters) alter the existing ventilation and drying processes within the caves, potentially leading to microenvironmental reorganization, disrupting the original microbial ecological balance, and triggering new dominant microbial populations and unknown bioerosion risks. This case highlights that analyzing and intervening by isolating environmental and biological factors will make it difficult to comprehensively assess disease mechanisms and may even lead to unintended conservation-induced damage.

Taking the Wanqing Temple Grottoes as an example [17], the study specifically reveals the dynamic coupling between environmental and biological factors: The primary joints and unloading fractures in the grotto rock mass provide initial growth space and water channels for plant roots. For instance, the roots of cypress trees on the southern wall of Cave No. 1 grow downward along fractures, with their physical wedging action continuously expanding the fissures, ultimately leading to the fragmentation of the “V”-shaped rock mass and forming a typical cycle of “fracture—root expansion—rock deterioration.” This mechanism further triggers a systemic amplification effect: root penetration alters local water movement, potentially activating the migration of deep soluble salts and creating a humid environment for microbial proliferation. The organic acids and other substances produced by microbial metabolism then weaken the cementation around the fractures, forming a chain-reaction network from macroscopic mechanical damage to microscopic chemical corrosion.

As a complete “geology-ecology-culture” complex, the deterioration process of grottoes is often the result of the interaction of multiple factors. However, current research has not yet established a systematic methodology for the holistic protection of grotto cultural relics, lacking integrated studies across scales and multiple processes. Therefore, based on the findings from the aforementioned literature, the interaction mechanisms among these factors are inferred as shown in Figure 6 below.

Figure 6. “Environmental drivers-Biological responses-anthropogenic as disturbance” multi-factor coupling mechanism diagram of disease and deterioration in grotto relics.

(2) Interdisciplinary integration remains superficial, lacking deep technical pathways for convergence. Although multidisciplinary technologies such as conservation science, materials science, microbiology, and geographic information science have been introduced into the field of cave conservation, their application still exhibits a “technology stacking” characteristic, lacking deep integration at both theoretical and methodological levels. For instance, the coupling mechanism between microbial communities and rock weathering processes, as well as the application of intelligent algorithms in disease prediction, have yet to form a systematic technical framework and evaluation standards.

(3) Despite the diversity of monitoring technologies, there is a deficiency in intelligence and predictive capabilities. Although current environmental and disease monitoring methods are increasingly varied, they predominantly rely on “post-event responses” and lack the ability to predict and regulate the future state of grottoes. Most monitoring systems have yet to achieve data-driven intelligent diagnosis, risk warning, and adaptive regulation, particularly in the face of dynamic external stressors such as climate change and tourism pressure. There is a lack of long-term trend analysis and decision support tools based on big data.

(4) The transformation of digital preservation outcomes and the development of standardization are lagging. Despite the widespread application of technologies such as 3D modeling, virtual restoration, and digital twins, their practical effectiveness in conservation projects, data sharing mechanisms, and technical standard systems remain inadequate. How digital outcomes can effectively support the entire process of “diagnosis-intervention-evaluation” for physical cultural heritage has not yet formed a closed loop.

Future research should focus on breaking through the current “fragmented” limitations, promoting a paradigm shift in cave temple conservation from “dispersed exploration” to a systematic approach characterized by “holistic correlation,” and gradually establishing a comprehensive technical system integrating “mechanism elucidation—intelligent prediction—precise regulation—collaborative operation and maintenance.” Key tasks include: establishing a systematic research framework for multi-factor coupling to unravel the complex mechanisms of cave deterioration; bridging interdisciplinary methodological chains to develop integrated technological pathways for processes such as microbial activity and rock weathering; upgrading monitoring systems to develop intelligent disease diagnosis, risk warning, and adaptive regulation platforms based on big data and artificial intelligence; promoting the standardization, sharing, and engineering application of digital conservation outcomes, and establishing their role throughout the entire process of “diagnosis—intervention—evaluation.” Through systematic integration and path innovation, the ultimate goal is to drive a fundamental transformation in cave temple conservation toward a “preventive, precise, and sustainable” model.

Based on a systematic review of domestic and international research literature on the conservation of cave temples and a visual analysis using CiteSpace, this review employs CiteSpace to reveal the developmental trajectory of the field from empirical and qualitative approaches to scientific and quantitative methods, the evolution of research hotspots, and interdisciplinary characteristics. The study indicates that cave temple conservation in China has established a research framework centered on “deterioration mechanisms—microenvironment monitoring—digital conservation—human impact factors,” achieving significant progress in material development, disease diagnosis, and 3D reconstruction. Keyword burst analysis further demonstrates that research focus has gradually expanded from early single engineering technical issues to systemic topics such as the recognition of cultural heritage value, digital inheritance, and preventive conservation.

However, current research still exhibits a pronounced “fragmentation” tendency, with most findings lacking a holistic grasp of the multifactorial coupling mechanisms involving “environment–organism–human interactions”; interdisciplinary integration remains at the level of technical stacking, without achieving deep theoretical and methodological consolidation; although monitoring systems are increasingly refined, the capabilities for intelligent diagnosis, prediction, and regulation remain weak; digital preservation outcomes also lag in standardization, sharing, and engineering transformation.

In the future, research on the conservation of cave temple cultural relics should further emphasize a holistic perspective and systems thinking, promoting a paradigm shift from “fragmented interventions” to “integrated regulation.” Key tasks include: establishing a systematic research framework for multi-factor coupling, deepening interdisciplinary exploration in areas such as microorganism-rock mass and artifact interactions; developing intelligent monitoring and early warning platforms based on artificial intelligence to enhance the foresight and precision of conservation efforts; promoting the standardization of digital outcomes and the closed-loop application in engineering, achieving full-process support from “data-decision-intervention-evaluation.” Only through deep interdisciplinary integration and innovation in technological pathways can the conservation of cave temple cultural relics achieve a fundamental transformation from “emergency restoration” to “preventive, precise, and sustainable” approaches, providing solid scientific and technological support for the perpetual preservation and revitalization of China’s cave temple heritage.