Introduction

Immovable cultural relics (ICRs) are remnants and traces left by humans during the process of historical development that have spatial locations and ranges. They are closely tied to the surrounding natural and social environments, reflecting the social activities, social relationships, and ideologies of humans, as well as their utilization and transformation of nature during different historical periods, with significant historical, scientific and artistic value (Editorial Board of the Encyclopedia of China, 2009). The establishment of an ICR database can offer modern means for cultural relic management, provide data support for comprehensive analysis on cultural relics, and a scientific basis for archeological decision-making, thereby playing a positive role in promoting the protection, research, and management of cultural relics (Yang et al. 1997; Schlader, 2002). The design of the data model is crucial and challenging for constructing of an ICR database, as it constrains the management and analytical capabilities, and determines the utility of ICR databases (Tilley, 2004 and 2020; Zhu, 2011; De Roo et al. 2014; Hu et al. 2018).

Spatial data is an important part of ICR data. According to whether to store spatial data explicitly, the existing ICR databases are divided into two categories: traditional databases and spatial databases. An ICR database founded on traditional database technology typically utilizes relational databases to store basic information, connotation, protection, and environmental information regarding ICRs (Guo et al. 2007). Existing ICR databases vary in the range and nature of ICR attributes included and the relational elements that structure the database. Most databases store essential information of cultural relics, such as name, type, administrative region, preservation status, historical period, cultural nature, geographical environment, longitude, latitude, elevation, and area, as well as information on the excavation period, excavation area, protection level, time when the protection level was determined, and protection recommendations (Teng, 1998; Valente and Cozzolino, 2019). In addition, in response to the unique information available for different types of ICRs, for some databases, corresponding tables or fields for storage have been designed. For example, for archeological sites, information related to discovered traces and artefacts, excavated animal bones, unearthed plant seeds and sporopollenin is stored (Teng, 1998). For ancient tombs, details of size, orientation, shape and structure and tomb occupant are recorded (Jin et al. 2018; Liebens, 2003). For historical buildings, information relating to architectural style, designer, structure, dimensions, materials, and use is retained (Spiridon et al. 2016).

To better address the spatial information requirements for cultural relic information management and to improve the analysis of the relationship between cultural relics and their environments, as well as risk assessment with regard to cultural relics, ICR databases based on spatial database technology have been developed. In terms of information management of cultural relics for digital documentation of cultural heritages, vector data of historical monuments, such as district boundaries, roads, settlements, villages and cultural heritages, were stored in the database so that they could be queried. A network dataset was created to solve the shortest distance to the cultural heritage site from any point (Yakar and Doğan, 2018). Musa et al. (2020) developed a GIS-based digital archive for heritage buildings and organized heritage buildings, rivers, roads, cadasters, land use and buffer zones as vector data and basic maps as raster images. Xiao et al. (1999) developed an archeological GIS application system for the Yangtze River delta. Vacca et al. (2018) designed a geodatabase for architectural and cultural heritages, including data about the metrical dimensions, construction techniques, materials, physical-mechanical and energy performances of historical buildings derived from multiple disciplines. The Norwegian cultural heritage database has managed information of archeological sites and monuments at two levels: archeological sites and objects (Berg, 2007). In addition, documents such as photographs, inventories, measured drawings, and 3D models were integrated by using hyperlinks in the database (Yakar and Doğan, 2018; Musa et al. 2020; Vacca et al. 2018).

In terms of the spatial analysis of cultural relic information, Gao et al. (2009) established a database of archeological sites, including county site data, topographic data, drainage maps, administrative maps, and remote sensing images, and then introduced spatial analysis methods to reveal spatial-temporal shifts, spatial structures, micro-geomorphological features of the sites, and relationships between the sites and rivers. Databases, including those composed of ancient site data and environmental data, such as regions, rivers, lakes, landforms, geology, soil, erosion, digital elevation models (DEMs) data and their derivatives, were established; in addition, ancient sites were analyzed in terms of elevation, distance from the nearest river, slope, and aspect to examine the spatial and temporal distributions of ancient sites (Zhu et al. 2021), and to provide a powerful way to understand the relationships between archeological sites and landscapes. For example, the distribution of rock-cut tombs is closely related to the prevalence of tuffs and sandstone, as well as rivers (Gencheva, 2023).

With regard to the risk assessment of cultural relics, Agapiou et al. (2015) created a spatial hazard database that includes natural hazards, such as landslides, erosion, salinity and neotectonic activity, as well as anthropogenic hazards, such as urban sprawl, modern road networks, drainage networks and fires; and assessed the overall cultural heritage risks in any particular district on the basis of hazard and archeological databases. Themistocleous et al. (2016) analyzed potential causative factors for natural hazards at heritage sites, and created a geodatabase of heritage at risk. Vulnerable sites were identified, and the impacts of geohazards on cultural heritage sites were further analyzed. Combey et al. (2021) established a database that includes information on archeological sites, active earthquake faults, and historical earthquakes. The database provides support for archaeoseismological investigation and cultural heritage protection.

The data models in the above studies are project-specific, and are often constructed organically to fit the expected findings and tailored towards a specific research objective (De Roo et al. 2015). To develop a universal data model for ICR, which can be applied for multiple objectives, Tennant (2007) presented a data model consisting of six datasets, representing data related specifically to the archeological organizations; survey and site boundaries; objects associated with buildings; individual artefacts or scatters of artefacts; natural environments; and transportation facilities. Meyer et al. (2007) designed a data model that represents cultural heritage data in four categories, namely, temporal data, spatial data, archeological objects and documents, which can help to formulate queries related to periods and places. Milner (2015) proposed a data model that represents basic elements of the archeological field, such as feature classes and tables, and their interactions with relationship classes. This model helps facilitate further analysis, such as accessing multiple layers of information from the relationship classes and spatial relationships. De Roo et al. (2016) proposed a data model, known as the Archeological DAta Model (ADAM), which describes diverse archeological objects via the three concepts of nodes, attributes and relationships. The ADAM enables integration for spatial, semantic and management data. Extensions have been proposed to adapt the infrastructure for spatial information in Europe (INSPIRE) model for use in cultural heritage, with the aim of enhancing the disaster resilience of cultural heritage (Fernández-Freire et al. 2014; Chiabrando et al. 2018). The INSPIRE is intended to easily share and use spatial data. It contains some themes, such as Protected Sites and Natural Risk Zone, that are related to cultural heritage, but lacks the specificity of cultural heritage. The Protected Sites theme was extended through three main aspects of cultural heritage: the legal protection framework, the cultural entities under protection and the documentation associated to them. The Natural Risk Zone theme was extended through adding attributes to the foreseen classes represented the risk and hazard problems, and parameters information useful to assess a risk on cultural heritage.

In summary, previous studies on ICR data models have involved information on various aspects, such as ICR itself, surrounding environments, and themes from different application perspectives. However, the comprehensiveness and completeness of the information expressed by any single data model are insufficient. It is necessary to consider both attribute data and spatial data while optimizing the universality of the data model. To address the above issues, in the present study, a geographic scene-based spatiotemporal data model of ICRs is designed on the basis of the comprehensive characteristics and application needs of ICR information. By using the concept of “geographical scene” to describe distinct relationships between entities, the model achieves a comprehensive and effective organization of ICR information. It supports interactive queries and spatiotemporal analysis of both ICR itself and its related environmental information. Through this approach, the management, scientific research, development and utilization of ICR can be promoted.

Design considerations and methods

Design considerations

Given the insufficiency of comprehensiveness and completeness of the information expressed by the existing ICR data model, in this paper three goals for the design of the ICR spatiotemporal data model are proposed, as follows.

Objective 1: Comprehensive consideration of the information on ICRs. (i) Information on the ICR itself, appendages, and surrounding natural and social environments of ICRs, should be considered. (ii) It is necessary to consider the common information shared by all types of ICRs, as well as the specific information unique to different types of ICRs.

Objective 2: Complete consideration of the information on ICRs. (i) The associations between information on ICRs, including social, cultural, political, economic, personal, and event information, should be considered (Zhu, 2011). (ii) ICR information concerning different disciplines and fields should be considered to meet various application needs.

Objective 3: Consideration of the information on ICR from a spatial perspective. (i) Spatial features, such as the location and shape of the ICR, should be considered. (ii) Spatial relationships between ICR and surrounding environmental elements and related things, including natural elements, such as topography, hydrology and climate (Cook et al. 2019; Nguyen et al. 2016), as well as social elements, such as settlements, political districts, and traffic, along with related movable cultural relics, people and events, should be considered. (iii) ICR hazard risk assessment caused by natural environment.

A geographical scene is a synthesis of various natural and human elements that interact in a certain spatiotemporal time range. It is a comprehensive representation of the real world that supports the advantages of multiscale nesting, dynamic and static coupling and the expression of multielement interactions (Lü et al. 2018), which can meet the three goals of ICR data model design. The design concept for the scene-based spatiotemporal data model of ICRs is shown in Fig. 1.

Fig. 1: Scene-based design concept. Scenes are composed of spatiotemporal objects.
figure 1

Based on the concept of scene, ICR, natural environment, social environment, and related elements in the real world are abstracted as spatiotemporal objects and scene.

Full size image

Methods and tools

In this study, the characteristics and application requirements of ICR information are combined, the basic process of computer software engineering is followed, and the design and experimental verification of an immovable cultural relic data model are carried out. The method of data model design in this paper includes entity-relationship (ER) model and spatial ER model; The software tools used for data model implementation are Geodatabase and PostGIS.

The process of data model design begins with the conceptual design, which is the most important stage; logical design; and physical design (Chen, 1976). The conceptual data model is used to model the information world and is the first level of abstraction from the real world to the information world (Mylopoulos, 1992). A conceptual data model uses logical concepts that are easier for most users to understand. The conceptual models can be represented through the ER model, which consists of three elements: entities, attributes and relationships (Chen, 1976; Teorey et al. 1986; Storey, 1991). Entities are “things” or “objects” that have an independent physical or conceptual existence. Attributes are characteristics of entities or relationships. Relationships are interactions or connections of entities with each other. There are three kinds of relationships: one–one, many–one and many–many. An ER diagram is a graphical representation of the ER model. Entities are represented as rectangles, with the entity names in the rectangles. Attributes are represented as ellipses, with attribute names in the ellipses, which are connected by lines to rectangles representing entities.

The spatial ER model is an improvement on the ER model, and it can better express the spatial characteristics of the entity. A spatial ER graph extends the normal ER graph via pictogram (Calkins, 1996; Shekhar et al. 1999). Spatial entities are represented as rectangles with the geometric symbols of points, lines and polygons on the top left. Spatiotemporal grid objects are represented as rectangles, with the interior divided into square grids. Spatial relationships (including topological, orientational, and metric relationships) are implied between any two spatial entities.

On the basis of the ER diagram and the spatial ER diagram, to express the temporal characteristics of spatial entities, in this study, simplified spatiotemporal ER diagram symbols with reference to the spatiotemporal object modeling methods proposed by Zhang (2003) and Hu et al. (2018) are designed. The symbol “T” at the top left corner of the rectangle, along with its combination with the geometry symbol, indicates that this entity is a spacetime entity, meaning a spatial entity whose spatial or attribute characteristics vary over time, as shown in Fig. 2.

Fig. 2: Symbols in the spatiotemporal ER diagram.
figure 2

The spatiotemporal E-R model extended on the basis of the traditional E-R model is used to express the data structure of geographic information systems with spatiotemporal characteristics. Entities are represented by rectangles and are categorized into generic entities and spatiotemporal entities. Attributes are shown as ellipses connected to entities, describing their specific characteristics. Relationships are depicted as diamonds, linking two or more entities to represent their associations. Spatiotemporal entities are further divided by geometry type into points, lines, and polygons, indicating spatial data with temporal dimensions. Spatiotemporal raster objects are illustrated as grids marked with a “T” symbol, highlighting their temporal attributes. The “T” symbol denotes temporality in entities.

Full size image

The geodatabase model defines a generic model for geographic information. This generic model can be used to define and work with a wide variety of different user- or application-specific models. The fundamental elements of the geodatabase data model used in this study include feature datasets, raster datasets, feature classes, object classes, relationship classes, geometric networks, domains, and validation rules (MacDonald, 1999; Zeiler, 1999; Longley et al. 2015; ESRI, 2024). Built in ESRI’s (Environmental Systems Research Institute) ArcGIS product, Geodatabase is equipped with systematic supporting software tools such as ArcCatalog, ArcMap and ArcToolbox, etc., which can easily support the collection, management, spatial analysis and visualization of ICR information, as shown in Fig. 3.

Fig. 3: Instrumentality of Geodatabase.
figure 3

Geodatabase is highly instrumental. Geodatabase is deeply integrated with ArcMap, supporting visual editing, management, and analysis of spatial and attribute data through feature datasets and toolboxes.

Full size image

PostGIS is an extension of the open-source object-relational database system PostgreSQL, allowing geometric objects (such as points, lines, and polygons) to be stored within the PostgreSQL database. PostGIS enables users to customize data types, making it more efficient for storing complex information. It also offers a rich set of geospatial functions, capable of performing tasks ranging from simple geometric operations to complex spatial analyses. Furthermore, PostGIS integrates seamlessly with tools like pgAdmin and is easily compatible with other open-source geospatial software such as Geoserver and QGIS. In this study, the main components of PostGIS used include tables for storing geometric and raster data types, schemes for organizing database objects and logical structures, as well as primary and foreign keys to represent relationships between tables.

ICR data model design

Entities and relationships design

On the basis of the concept of scene, the information on ICR is abstracted into entities, such as scene, ICR ontology, movable cultural relic (MCR), environmental element (EnvElem), person, and event. Scene entities are composed of ICR ontology, environmental element, person, and event entities, which collectively form natural and social scenes with ICR ontology at their core. In addition, an ICR ontology has interrelated appendages, environmental description information, protection and utilization information, hazard information and documents, which are abstracted as entities, such as appendage, environmental description (EnvDesc), protection information (ProtInfo), hazard, HazardRiskZone and document, respectively. These entities also have complex relationships, including composition relationships, inheritance relationships, and associative relationships. Figure 4 shows the entities and entity relationships in the ICR conceptual data model.

Fig. 4: Design of entities and relationships.
figure 4

This figure shows the entities and relationships design centered on ICR, covering ICR ontology, environmental elements, person, events, management protection information and hazard risk zone, as well as introducing multiple raster environmental factors.

Full size image

Different types of ICR have both common and unique attributes. Therefore, six entities have been defined: ancient sites, ancient tombs, historical buildings, grotto temples, stone carvings, and modern historical sites. These entities are designed to inherit the attributes of the entity of ICR ontology and add their own unique attributes to the design. Although ancient buildings and modern and contemporary buildings, which are two types of ICR, differ in their temporal spans, both are essentially architectures with some unique attributes. Therefore, the two building types are merged and abstracted into the entity of “historical buildings”.

Entities of environmental elements are used to abstract natural and social environments. The environmental elements that greatly impact ancient human activities were selected and abstracted into several entities, i.e., terrain, precipitation, temperature, mountains, rivers, admin regions, settlements, roads, products, faultZone and coastalZone. These entities inherit the attributes of the entity of the environmental element and add unique attributes on the basis of the factors related to the production, distribution, and evolution of ICR associated with each entity. Cultural relics (CR) entities include ICR ontology entities and MCR entities, which inherit the properties of the CR entity, while also having uniquely designed attributes for each. There are complex associative relationships between the entities, as shown in Table 1.

Table 1 Relationships between entities.
Full size table

Entity attributes design

On the basis of an analysis of the information content as well as the entities and entity relationships of ICRs, and considering the basic principles of database design, the attributes of each entity are constructed. In particular, the designated common attributes of the ICR are shown in Table 2, in the table, * indicates multiple attributes, and T.* represents multiple attributes that vary over time, that is, multiple attributes exist in different periods.

Table 2 Design of common attributes of the ICR ontology entity.
Full size table

The different types of ICR have all the attributes of the ICR ontology. Thus, corresponding unique attributes are designed on the basis of unique information. The unique attributes of the different types of ICRs are shown in Table 3.

Table 3 Design of unique attributes of different types of ICRs.
Full size table

An environmental element entity is an entity with spatial characteristics. It contains spatial data and attribute data, and its common attributes mainly include the name, location and type of the entity. The unique attributes of environmental elements can be customized according to the different application purposes.

This study primarily considers the major natural hazard types faced by ICRs: meteorological and hydrological hazards, geological hazards, and marine hazards (Liang et al. 2023; Nguyen et al. 2016). Hazard risk assessment involves three key elements: the hazard-bearing body, hazard-causing factors, and the hazard-inducing environment. In the context of ICR hazard assessment, the hazard-bearing body corresponds to the ICR itself, with attributes such as structure, material, and preservation status included in the common ICR attributes, as these are directly related to the impact of the hazard. Hazard-causing factors are the natural or human conditions that trigger hazards, such as heavy rainfall, floods, earthquakes, and tsunamis. The specific information of hazard-causing factors is represented through the unique attributes of objects related to these factors within the environmental elements, such as attributes like flow rate and discharge for rivers in the case of floods. Therefore, hazard-causing factors do not correspond to specific entities in the conceptual model. The hazard-inducing environment refers to the natural environment in which the ICR is located, including the geology, climate, and hydrology of the area. This is represented through the entities and attribute information of environmental elements.

The attribute design of hazard entity, hazard risk zone entity and other entities is shown in Table 4.

Table 4 Attribute design of other entities.
Full size table

Results and discussion

Experimental validation

Construction of the case database

In order to verify the conceptual data model designed in this paper, it is necessary to go through the logical data model and physical data model stages. The logical data model phase transforms the conceptual model into the database structure elements of an object database or a relational database. For example, entities in the conceptual model are converted to objects in an object database or tables in a relational database, relationships between entities are converted to relationships between objects in an object database, or primary keys, foreign keys, or new database tables in a relational database. A physical model is a physical implementation of a logical model that involves specific database management system choices. In the choice of physical model specific database management system, both open-source and proprietary systems are available, and are widely used in academic research. This paper conducts experimental validation and comparison based on Geodatabase and the open-source database PostGIS, demonstrating that the conceptual data model designed in this study can be effectively applied across various specific physical models.

Geodatabase allows the definition of specific models based on the common Geodatabase framework, which provides a wide range of templates for users to implement GIS projects in different fields. PostGIS extends the PostgreSQL database as an open-source spatial database with high accessibility without paying licensing fees. This paper will use both databases in the experimental verification part. On the basis of the designed conceptual data model of ICRs, according to the definition of basic elements in the geodatabase, the entities and relationships in the conceptual model are converted to a geodatabase. Entities with spatial characteristics are converted to feature classes, and a group of feature class entities of the same type form a feature dataset. Entities without spatial characteristics are converted to object classes. The relationships between entities are converted to relationship classes; In PostGIS, entities with spatial characteristics are converted into tables with a spatial column called “geometry,” which is used to store the spatial geometry type of the entity. Entities without spatial characteristics are converted into regular tables. The conversion of relationship classes is achieved by defining foreign key constraints between entities.

Using examples from the Third National Cultural Relics Census, cultural relic excavation reports, cultural relic records, cultural relic atlases, cultural relic research findings and other data (National cultural heritage administration of China, 2007; Editorial Board of the Encyclopaedia of China, 2009; Liyang Museum, 2023; Shi et al. 2023). These materials are generally authoritative. Most data are reliable, and a few such as dating data may have precision issues. The authority of these data sources and the reliability of the data can be specified in the remarks attribute of relevant entities. On the basis of the designed conceptual data model and logical data model of ICRs, we construct an example spatial database for ICRs based on the Geodatabase and PostGIS.

The sample database based on Geodatabase includes 4 feature datasets, 15 feature classes, 6 object classes, 11 relationship classes, and 3 raster datasets. The database structure is shown on the left side of Fig. 5. The sample database based on PostGIS includes 3 Schema and 25 tables. The database structure is shown on the right side of Fig. 5.

Fig. 5: Example of the structure of geodatabase-based and PostGIS-based spatial database.
figure 5

The left side represents the Geodatabase-based structure, while the right side shows the PostGIS-based structure.

Full size image

Case: ICR information management

This case study demonstrates the storage, management, and query capabilities of ICR information in order to validate the comprehensiveness and completeness of the ICR information represented by the model in this study.

The example data include different types of ICR, such as ancient sites and ancient tombs. The common attributes of all types of ICRs are shown on the left of Figs. 6 and 7, as well as unique attributes are shown on the right of Figs. 6 and 7. When studying a certain type of ICR individually, it is necessary to consider both common and unique attributes of this type of cultural relic. Taking the type of ancient site as an example, all its attributes are shown at the bottom of Figs. 6 and 7.

Fig. 6: An example of common and unique attributes of ICRs (Based on Geodatabase).
figure 6

The common attributes of all types of ICR include basic information such as common name, type, age, location, and other fundamental details. The specific attributes of different types of ICR are more personalized characteristics that play a role on a smaller research scale. When studying a particular type of ICR, the common attributes and specific attributes together constitute its complete set of attributes.

Full size image
Fig. 7: Example of common and unique attributes of ICR (Based on PostGIS).
figure 7

Similar to the Geodatabase implementation, the common attributes of all types of ICR include basic information such as common name, type, age, location, and other fundamental details. The specific attributes of different types of ICR reflect more personalized characteristics that play a role in more specific research. When studying a particular type of ICR, the combination of common attributes and specific attributes forms the complete set of attributes, with the only difference being the software platform used for data storage and management—PostGIS in this case.

Full size image

This model comprehensively represents various entities and their relationships with ICR information, enabling support for information queries of ICR entities and their associated entities. Taking Ming Xiaoling Mausoleum and Ming Dingling Mausoleum as examples, Fig. 8 depicts entities and entity relationships in the real world.

Fig. 8: Example of ICR ontology and its associated objects.
figure 8

The example showcases the ICR objcets, Ming Xiaoling and Ming Dingling, along with their associated objects and relationships. For instance, Zixia Lake is located southwest of Ming Xiaoling, Zhu Yuanzhang is buried in Ming Xiaoling, and the Li Zicheng’s burning of the Ming imperial tombs event occurred at Ming Xiaoling.

Full size image

Figure 9 illustrates the functionality of managing ICR information via the geodatabase, displaying the entity attribute information, entity relationship information, and the capability of entity linkage queries.

Fig. 9: An example of querying and browsing the associated information of ICR.
figure 9

(i) Identifying the ICR of the Ming Xiaoling Mausoleum, obtaining related entities such as people, events, and environmental elements, and displaying a detailed table of entity attributes. (ii) Similarly, identifying the Ming Dingling Mausoleum reveals more detailed entity and entity relationship content. (iii) Displaying the linkage table established through ArcMap for the Ming Xiaoling Mausoleum to enable interactive queries with people entities and event entities.

Full size image

(i) Display of entity attribute information. The ICR feature data from the sample database are loaded into the ArcMap software. Use the identification tool to click on the Ming Xiaoling immovable cultural relics feature data, which displays the attribute information of this feature data, as well as its associated features and objects, as shown in the upper left of Fig. 9.

(ii) Display of entity relationship information. It is possible to display, within the geodatabase database, all features and objects associated with a certain feature through relationship classes, such as environmental factors, appendages, persons, and events associated with the immovable cultural relic ontology. Detailed information about these features and objects can be viewed. Taking Ming Xiaoling as an example, its entity relationships are shown in the upper right of Fig. 9.

(iii) Entity linkage query. By using the relationship associations in the immovable cultural relics attribute table, it is possible to query the person entity related to Ming Xiaoling, retrieving detailed attribute information about individuals, such as Yuanzhang Zhu and Xiuying Ma. Furthermore, by leveraging the relationships between person entities and event entities, it is possible to further link and query detailed attribute information about related events, as shown at the bottom of Fig. 9.

Figure 10 demonstrates that, using PostGIS, the linkage query between ICR, person and event entities can be realized by establishing foreign key association query based on PostGIS database. Among them, the upper part of Fig. 10 shows the linkage query of ICR (Ming Xiaoling Mausoleum) and its related person entities; The lower part of Fig. 10 shows the linkage query between ICR (Ming Xiaoling Mausoleum and Ming Dingling Mausoleum) and event entities.

Fig. 10: Example of linkage query of ICR based on PostGIS.
figure 10

The upper part shows an example of a linked query result for Ming Xiaoling’s common name and all attributes of the associated person entity. The lower part shows an example of a linked query result for the common names of both Ming Xiaoling and Ming Dingling, along with all attributes of the associated event entity.

Full size image

This case shows the storage of temporal and multi-valued attribute designs in a conceptual model. In entity attribute design, The * symbol indicates that the attribute is multi-valued, meaning it can hold multiple values. The T.* symbol represents multi-valued attributes that change over time, with each attribute existing in different time periods.

Based on Geodatabase, for T.* attributes, we need to create an association table between feature classes and object classes with temporal attributes. This table will store the feature name, attribute name, start and end times, and attribute data. Additionally, create a feature dataset to store all the relationship classes of temporal attributes. For multi-valued attributes, use a specific delimiter to separate multiple values within the attribute data.

Based on PostGIS, it allows the declaration of custom composite types to represent T.* attributes. The T attribute field in the table will consist of three values: the common attribute name and the start and end times. For multi-valued attributes, store them using an array data type.

This case uses the example of a person’s aliases changing over time to demonstrate the storage of temporal multivalued attributes in different physical models, based on Geodatabase and PostGIS, see Fig. 11 for an example.

Fig. 11: Example of multivalued time series representation.
figure 11

The left part shows the T.* attributes of a person and ICR in Geodatabase, where they are recorded in the same table and linked to the ICR entity. The right part shows, in the red box, the time periods during which different aliases of the person exist in PostGIS.

Full size image

Case: ICR information analysis

This case study demonstrates the spatiotemporal analysis and capability of expressing ICR information, thereby validating the ability of the model proposed in this study to support analyses of spatial patterns, temporal evolutionary processes, and relationships between ICR material entities and their environment, all examples are shown in Fig. 12.

Fig. 12: This diagram illustrates the multidimensional scene analysis of ICR information.
figure 12

a Temporal distribution analysis; b, c spatial distribution of ICRs and the spatial relationship analysis between them; d, e visualization of regional and local ancient site scenes, respectively; f impact analysis of a single environmental factor.

Full size image

(i) Analysis of the era distribution of ICRs. Statistical analysis can be performed on the era distribution of any type of ICR feature in the database. For example, for the ancient site features in the example geodatabase, the era distribution is analyzed via the age attributes from the feature attribute table.

(ii) Analysis of the spatial distribution of ICR. On the one hand, the spatial distribution characteristics of any type of ICR feature in the database can be analyzed. For example, kernel density analysis is performed on ancient sites in the example database. The figure shows that the distribution of ancient sites forms three high-density areas and four sub-density areas. On the other hand, the spatial relationships of different types of ICR are analyzed. By investigating the distances between the spatial distributions of different types of ICR, we can determine whether these distances affect their respective distributions - for example, by overlaying environmental factors data onto the data of ancient sites, ancient tombs, historical buildings, and other features in the sample database, and analyzing the spatial distribution and clustering characteristics of the archeological sites under the influence of environmental factors.

(iii) Visualization of regional scenes and cultural relics scenes. For regional scenes, the DEMs, remote sensing images, administrative region features, and ICR point features, as well as related persons, and events and other objects in the example geodatabase are overlaid onto ArcScene software, and their elevations are set to visually display the spatiotemporal scenes of the study area. This scene includes information on surrounding environmental entities (such as mountains and rivers) related to cultural relic entities and integrates object class scene features, such as persons and events. In addition, to display the spatiotemporal scenes within the range of a certain archeological site, the feature layers of the ancient site and its surrounding environment in the example database are selected to show remote sensing image raster data or other scene data, such as 3D models of cultural relics.

(iv) Comprehensive analysis of ICR and environmental factors. This analysis explores how individual environmental factors affect the point feature distribution of ICR, and investigates the combined effects of various environmental factors. For example, by overlaying the features of ancient sites with single environmental factors, such as water systems, contour lines, and mountain slopes, their spatial distribution pattern characteristics can be analyzed and revealed. In addition, by integrating different environmental factors, a site prediction model can be established to predict the distribution of ICR under multiple environmental factors.

Discussion

This paper designs a spatiotemporal data model for ICRs. The model framework is comprehensive, enabling the complete storage, management, and querying of ICR information through database implementations based on Geodatabase and PostGIS.

DeRoo et al. (2015) pointed out that existing data models are often project-specific, tailored to meet particular research goals. Meyer et al. (2007) designed a data model that only considers the ICR and Document entities within spatiotemporal data. Milner (2015) proposed a data model that takes into account both cultural heritage entity feature classes and relationship classes. Tennant (2007) introduced a data model that, in addition to data related to cultural heritage entities, also stores some natural environment and transportation infrastructure data. In comparison to these models, the data model presented in this paper provides a comprehensive representation of the attribute and spatial information of entities such as ICR ontology, surrounding environmental objects, people, events, and appendages. It is a common data model that can be applied to multiple objectives. Moreover, the data model designed in this paper exhibits strong flexibility and scalability in specific applications. The flexibility is mainly reflected in its ability to conduct comprehensive analyses of the common attributes of all ICR entities. The scalability is demonstrated by the ability to extend the specific attributes of different types of ICR entities according to actual application needs. In addition to ICR and environmental element entities, it can also be further expanded to include hazard entities and store predicted hazard data (Cook et al. 2019).

In terms of the unique attributes of ICR, this study only investigates their overall characteristics. The design can be further refined according to the research and application requirements for in-depth exploration and understanding of the unique attributes of different types of cultural relics. For example, cave sites and city sites, which are both ancient sites, have great differences in the description and application of their unique information. The unique attributes are further refined and designed as a simple example in Table 5.

Table 5 Unique attributes of subtypes of ancient sites and their application examples.
Full size table

In terms of data model implementation, the experiments in Section 4.1 demonstrate that both the commercial software Geodatabase and the open-source software PostGIS can effectively implement the ICR conceptual model designed in this paper, supporting the entities, attributes, and relationships as designed. Each has its own strengths and weaknesses.

Geodatabase is built into ESRI’s ArcGIS products and supported by a comprehensive suite of tools, including ArcMap, ArcCatalog, and ArcScene. This integration, along with its graphical user interface, greatly enhances ease of use, allowing users to easily create ICR databases, manage data, perform spatial analysis, and visualize results through ArcGIS products. Geodatabase uses feature classes and relationship classes to define the entities and relationships in the conceptual model, and it has stronger object modeling capabilities compared to PostGIS. However, Geodatabase does not directly support the storage of multivalued or temporal attributes, requiring special handling methods—such as adding additional related tables in this study to manage complex T* attributes. Moreover, one of Geodatabase’s advantages over PostGIS lies in its ability to model behavior. Through domains, it can restrict field values, such as limiting the height or age of ancient buildings, and by defining rules and topological relationships, it can more accurately describe the spatial relationships between ICR and environmental objects. Additionally, by using geometric networks, it can reconstruct ancient transportation networks and simulate human migration activities. These capabilities are crucial for ICR modeling and are promising directions for future research.

PostGIS, as a spatial extension of the PostgreSQL relational database, is better suited for storing and processing geospatial data than Geodatabase, though it lacks a built-in graphical interface. It relies on external tools such as QGIS or pgAdmin for data management and visualization. PostGIS implements the entities and relationships in the conceptual model through tables with spatial columns and primary/foreign keys. PostGIS also allows more flexible storage of T* attribute data and more complex spatial analysis through SQL queries. For feature classes with multiple geometry types, PostGIS allows the geometry field in a single table to store different types of geometric data (such as points, lines, and polygons), which enables the representation of various spatial objects in the same field, enhancing the flexibility of spatial representation and analysis of complex objects compared to Geodatabase.

Furthermore, the reliability of information on ICRs is a crucial factor. The reliability hinges not only on the data sources, but also on the methods and technologies employed for data collection (such as the precision issues involved in dating ICRs using Carbon-14 dating and stratigraphic dating), the timeliness, integrity and consistency of data. Comprehensive consideration of the reliability information on ICRs is a complex issue that requires further in-depth research, as well as the inclusion of entities or attributes related to data reliability in the data model.

Conclusions

To address the insufficient consideration of the comprehensiveness and completeness of information in ICR data models, in this study, a scene-based spatiotemporal data model of ICR is designed from the perspective of geographic scenes and considers the associations between ICR and their appendages, natural environment, and social environment. An example database of ICR is constructed on the basis of the geodatabase, and the effectiveness and advancement of the data model designed in this study are validated. The advantages of this data model are reflected in two main aspects: technology and application.

The technological advantages of the data model designed in this study are: (i) The entities are comprehensive, fine-grained, and spatiotemporally oriented, covering the main information of all aspects of managing and researching ICR. In addition, the section with spatiotemporal characteristics from the material entity information and environmental information of ICR is abstracted as spatiotemporal entities. (ii) The entity attributes are comprehensive and spatiotemporally oriented and fully consider the common and unique information of different types of ICR as well as the temporal and spatial information of ICR. (iii) The entity relationships are comprehensive and fully consider both the semantic and spatial relationships between the ICR material entity and other entities.

The application advantages of the data model designed in this study are: (i) The completeness of ICR information and the flexibility of the model enable the comprehensive analysis of all types of ICR and the specialized analysis of one or several types of ICR. (ii) Spatiotemporal object characteristics can support the combination of spatiotemporal analysis and archeological typology analysis to highlight the characteristics and patterns of ICR and the phenomena which they reveal from the perspective of spatial distribution and evolution. (iii) The information on the surrounding environment of ICR is elevated to a level of importance similar to that of the material entity information of cultural relics and is no longer treated as background information. Additionally, the geographic information related to ICR is comprehensively modeled, fully considering the basic geographic information and thematic features of cultural relics. This approach can support the associated query and analysis of ICR information, as well as the analysis of the relationship between ICR and the environment.