A spatial digital twin framework for urban built environments: case studies in Victoria, Australia

Abstract

With the fast-growing population and expansion of urban areas, spatial information plays a vital role in addressing location-specific problems in global challenges such as climate change, equitable land access, social inclusion, infrastructure resilience and environmental protection. There is a significant need for design and development of an integrated approach that leverages emerging and innovative spatial technologies to tackle these pressing issues. In this context, this paper presents the design, implementation, and operational deployment of a spatial digital twin (SDT) framework for urban built environments, demonstrated through three working case studies in Victoria, Australia. The framework integrates heterogeneous datasets from multiple institutional custodians—Vicmap, local councils, Land Use Victoria, Melbourne Water, and industry partners—through a cloud-based loosely-coupled architecture underpinned by open geospatial consortium (OGC) standards (OGC API Features/Maps/3D Tiles, SensorThings API), CityGML and Industry Foundation Classes (IFC). The platform uses a PostgreSQL/PostGIS database to store spatial data, Apache Kafka and Message Queuing Telemetry Transport (MQTT) to manage real-time data streams, and a React–CesiumJS web interface for 3D visualisation. It is deployed on the NeCTAR Research Cloud. Case studies include the Development Envelope Control (DEC) system for automated rule-based generation of compliant building envelopes, PedDesign for simulating pedestrian movement in dense urban environments, and a decision-support tool for stormwater capacity in urban catchments. The paper discusses the technical and institutional challenges of realising SDTs and outlines future pathways for integrating SDTs into wider future-city strategies.

1 Introduction

Accelerated urbanisation and population growth pose an unprecedented pressure on built and natural environments. Cities should deliver housing, transport, and utility services while addressing climate change, resource shortages, and Sustainable Development Goals (SDGs) (Pandey and Ghosh, 2023). Spatial information is crucial in addressing urban challenges since it connects location-based information about land, buildings, infrastructure, and environment to decision-making. Within this context, spatial digital twins (SDTs) have emerged as a powerful paradigm for urban planning (Batty, 2024) and infrastructure management (Shirowzhan et al., 2020). Digital Twin Consortium defines a digital twin as a virtual representation of real-world entities and processes, synchronized at a specified frequency and fidelity (Olcott and Mullen, 2020). In this definition, digital twins transform businesses by accelerating holistic understanding and decision making, use real-time and historical data to represent past and present and simulate possible futures, and are tailored to specific use cases. As a specific type of digital twins, SDTs are dynamic, location-based digital representation of real-world assets, systems, and environments (ANZLIC, 2019). SDTs continuously integrate live spatial and sensor data to mirror the current state of built and natural environments. This spatial foundation makes SDTs particularly useful in smart city initiatives, enabling planners to analyse interconnected infrastructure networks, environmental conditions, and social dynamics an integrated digital environment, providing the capability to simulate changes in situ. Therefore, SDTs are emerging as a key enabler of spatially enabled cities and societies. Recent research shows that SDTs can integrate 2D, 3D and 4D (time) data, harmonise multiple formats and provide web-based data access (Rajabifard et al., 2022).

SDTs serve as hubs of real-time urban data and analytics, continuously fed by city-wide sensors and systems–for example, traffic monitors, utility networks, and environmental sensors–enabling officials to visualize current conditions and predict future impacts (Mazzetto, 2024). Planners can simulate “what-if” scenarios (e.g., new transit routes, infrastructure upgrades, zoning changes) in the virtual model before deployment. Such predictive modelling reduces costly errors and optimizes resource allocation, accelerating evidence-based decision-making (Metcalfe et al., 2024). It also supports predictive maintenance of critical infrastructure by forecasting equipment failures and scheduling interventions in advance. SDTs’ intuitive interfaces enhance transparency and public engagement: stakeholders–including city planners, technical experts, and citizens–can collaboratively visualize proposed developments, making planning processes more democratic and collaborative. Together, these capabilities support more sustainable, resilient, and inclusive urban development.

For instance, SDTs implemented in Victoria and New South Wales states of Australia have significantly improved access to spatial data and enhanced the visualisation of urban environments for government, industry, and the public (Department of Transport and Planning, 2025; Department of Customer Service, 2025). However, these SDTs are limited in their interoperability and analytical capabilities, as they are predominantly designed as visualisation platforms rather than fully integrated decision-support systems. Their current architectures often restrict seamless data exchange across domains and constrain the ability to perform advanced analytics, simulation, and predictive modelling.

Moreover, much of the existing research has focused on developing SDTs customized to a specific application area, such as transport, utilities, energy, or environmental monitoring (Mazzetto, 2024). While these domain-specific SDTs demonstrate clear value within their respective contexts, they do not adequately address the interconnected and systemic nature of urban challenges, where decisions in one sector can have impacts on others. As a result, there is limited research on the design and implementation of SDTs that can support multi-domain applications, particularly in the context of integrated urban planning and infrastructure resilience. To address this gap, this paper presents the design, implementation, and deployment of SDT framework that enables interoperability, advanced analytics, and cross-domain integration. The contribution is three-fold: (i) a concrete reference architecture that harmonises 2D, 3D, and real-time streaming data through open OGC standards; (ii) a comprehensive integration model that allows domain-specific analytical tools—such as the Development Envelope Control, PedDesign, and stormwater decision-support tools reported here—to be composed within a single spatial environment; and (iii) an operational deployment on the NeCTAR Research Cloud that demonstrates the framework across three working case studies in Victoria, Australia. By moving beyond visualisation toward analytics-driven decision support, this research contributes a scalable and extensible SDT model for evidence-based urban planning and resilience assessment of built environments.

The remainder of this paper is organized as follows. In Section 2, we first review current literature on applications of SDTs in urban planning and infrastructure monitoring. Section 3 is dedicated to the methodology for developing SDTs, in which we will describe the system architecture of SDT developed in this study. Section 4 describes case studies related to urban planning and infrastructure from Victoria, Australia. This is followed by Section 5 which discusses the benefits of SDTs for urban built environments as well as current challenges in adoption of SDTs in the current urban systems. Finally, the paper concludes in Section 6 by outlining main contributions of this study as well as recommendations for future research.

2 Literature review

Over the last years, there has been significant research and development on adoption of SDTs to manage cities in real-time and forecast the impacts of planning decisions. SDTs include integrated 3D urban models enriched with real-time data, which can help planners simulate various urban development scenarios and environmental changes prior to implementation. SDTs enable scenario-based urban planning and evidence-driven design. Planners can test how changes to the built environment, such as new infrastructure, housing developments, and green spaces, can impact important aspects of urban environments such as traffic flows, pollution and public safety. For instance, an SDT platform can provide the capability to assess the impact of a new transit line on city traffic and how different zoning regulations can influence urban growth patterns. Recent studies emphasize that SDTs support participatory planning as they offer scenario analysis for investment prioritization and allow enhanced public engagement by visualizing long-term development outcomes in an accessible way (Zhu and Jin, 2025). Many SDT implementations include interactive dashboards and visually immersive interfaces so that diverse stakeholders–from technical experts to decision-makers and citizens–can explore planning scenarios and understand their implications (Zhu and Jin, 2025).

Key technologies enabling SDTs include 3D geospatial information, Building Information Modelling (BIM) integration, sensor networks for real-time urban data, and artificial intelligence/machine learning (AI/ML) for predictive analytics. The integration of BIM with IoT is particularly important since detailed BIM models contribute rich spatial and functional data about buildings and infrastructure, which, when combined with real-time IoT sensor data, allow accurate simulations of urban dynamics (Mazzetto, 2024). More broadly, recent built-environment research has emphasised the need to connect data-driven frameworks with performance indicators, sustainability outcomes, and decision-making processes. For example, Mattar et al. (2026) demonstrate how project success factors, institutional capacity, and material-performance indicators can be integrated to support more sustainable and resilient construction outcomes. This reinforces the importance of SDT frameworks that move beyond visualisation and support performance-oriented decision-making across planning, infrastructure, and built-environment management. Importantly, SDTs enable a level of detail and responsiveness in urban planning that was unattainable in traditional static spatial systems.

Some SDT have been implemented around the world. The first example is Virtual Singapore, which is a city-wide SDT creating a dynamic 3D platform for urban planning and decision support (Liu P. et al., 2023). Its key features include high-resolution 3D modeling of the entire city, real-time data integration (e.g., live traffic, climate sensors), and simulation tools for urban environmental scenarios. This platform integrates real-time sensor inputs with analytics and environmental simulation tools, allowing planners to explore solutions to complex urban problems, such as land use conflicts and transport issues, in a virtual environment (Walker, 2023). A significant benefit of Virtual Singapore’s deployment is improved decision-making capabilities for city officials and stakeholders, though it also underscored the need for robust data governance to manage the massive data streams involved. Another case in Europe is Helsinki 3D+, which is an open-access SDT of Helsinki that provides a detailed 3D city model for planners and the public (Calzati, 2023). Its key features are open data sharing, web-based 3D visualization, and integration with planning tools (e.g., for shadow casting, noise analysis). Helsinki’s approach highlights the value of transparency and citizen participation (Kopponen et al., 2024); however, it also faces the ongoing task of keeping the digital model accurate and updated, requiring continuous data maintenance and community input.

In Middle East, there are two significant SDT initiatives. First, an SDT focusing on improving urban planning, traffic management, and public service delivery has been developed in Riyadh, Saudi Arabia (Swayne, 2024). By integrating 3D models with real-time data such as traffic and utility meter data, Riyadh’s SDT enables detailed scenario analyses such as assessing how new highways can mitigate congestion and how different policies could reduce energy consumption. Another one is Dubai’s Digital Twin Initiative, which is a government-driven SDT program aiming to transform Dubai into a smart city by integrating digital twin technology into urban governance. Its features include detailed 3D city models combined with artificial intelligence and IoT data feeds to support urban planning, infrastructure management and emergency response (Lavigne, 2024). Dubai’s initiative demonstrates strong political will to leverage SDTs for holistic city management.

Furthermore, SDTs play a critical role in monitoring and managing city infrastructure systems. Urban infrastructure–including transportation networks, bridges, roads, utilities (water, electricity, gas), and buildings–benefits from the real-time data and predictive analytics that SDTs provide. Traditional infrastructure management often faces inefficiencies, poor real-time visibility, and high maintenance costs (Liu C. et al., 2023). SDTs offer a data-centric and proactive approach to infrastructure health monitoring and operations (Liu C. et al., 2023). By continuously using spatial data from sensors and integrating it with physics-based models of infrastructure, SDTs can detect anomalies, assess structural integrity, and predict future issues before they manifest as failures. Therefore, SDTs have become vital for infrastructure monitoring and management, enabling cities to shift from reactive maintenance to proactive and predictive maintenance. A prominent application area is structural health monitoring (SHM) of civil structures such as bridges, tunnels, and high-rise buildings. Researchers have developed digital twin frameworks that combine real-time sensor measurements with detailed engineering models of the structure (Nasim et al., 2026).

SDTs are also being applied to transportation networks and public utilities to improve reliability and efficiency (Javaherian Pour et al., 2025; Wu et al., 2025). City transportation departments, for example, use SDTs of road networks and traffic systems to monitor congestion in real time and test traffic management strategies virtually. By feeding live traffic sensor and GPS data into a calibrated city traffic model, operators can simulate interventions such as dynamic signal timing and rerouting during incidents, and then implement the optimal solution. This approach has been used in different cities including Singapore and Barcelona, where SDTs help optimize traffic control and reduce congestion (Zhu and Jin, 2025).

For utilities, SDTs enable continuous monitoring and more efficient resource distribution. For example, integrating smart meter data and SCADA system feeds into a utility network twin provides operators a live view of the system’s state. SDTs can identify the location of leaks and faults early and simulate the impact of switching equipment and rerouting flows. One study developed a digital twin for an urban water management system that featured a live control dashboard: it would automatically adjust pumps and valves based on sensor data to mitigate flooding risks, while simultaneously using historical data to inform predictive maintenance of the pumping equipment (Aheleroff et al., 2021). This capability allowed real-time control minimizing immediate flood threats and long-term optimization of infrastructure upkeep.

At a wider city scale, critical infrastructure systems are spatially interconnected, and SDTs are important to manage these interdependencies. Failures in one infrastructure, such as the power grid, can cascade into others such as traffic lights, water pumps, and so on. SDTs, by integrating spatial data about multiple infrastructure systems, help in understanding and monitoring such cascading effects and improving infrastructure resilience (Zhu and Jin, 2025).

3 Proposed approach for design and implementation of SDT

The design and implementation of the SDT framework in this study constitute a complex systems engineering task that requires expertise across multiple domains, including geospatial information systems, software architecture, cloud computing, big data engineering, analytics, simulation, visualisation, cybersecurity, and platform governance. While SDT platforms may differ in their specific design objectives, operational contexts, and application domains, they generally share a set of core components that form the foundation of any scalable and interoperable SDT ecosystem. Rather than presenting a generic SDT architecture, this section describes the implemented framework developed and deployed in this study. The framework operationalises SDT capabilities through a modular architecture in which data ingestion, spatial data management, 3D data transformation, real-time streaming, analytical services, visualisation, user management, and API management are implemented as separate but interoperable components. Figure 1 illustrates the conceptual architecture design, showing backbone components and their interactions.

FIGURE 1

The operationalisation of the framework follows a modular architecture that enables domain-specific tools to be integrated into the SDT without requiring each application to rebuild the underlying data and visualisation infrastructure. In the current implementation, heterogeneous datasets are ingested into a PostgreSQL/PostGIS spatial data store, streamed through Apache Kafka and MQTT where near-real-time data are required, exposed through OGC-compliant APIs and REST services, and visualised through a React and CesiumJS web client deployed on the NeCTAR Research Cloud.

3.1 Data management

At the centre of the SDT architecture lies data management, which underpins all analytical, visual, and decision-support capabilities. In the implemented framework, a dedicated data management layer orchestrates the ingestion, harmonisation, storage, and access of heterogeneous spatial data streams. This layer supports the integration of 2D spatial datasets, 3D city and building models, simulation outputs, and live sensor feeds within a shared spatial environment.

3.1.1 Spatial data

Spatial Data Management addresses the handling of conventional geospatial datasets, including vector and raster data, geodatabases, and associated metadata. In the present implementation these datasets originate from multiple institutional custodians rather than a single administrative entity: Vicmap (Department of Transport and Planning) for the statewide cadastre, addresses, and road network; local councils for planning-scheme overlays; Land Use Victoria for property identifiers; Melbourne Water and the Fishermans Bend Development Authority for drainage and catchment data; and the Australian Bureau of Statistics for demographic reference layers. Datasets may be hosted within the SDT infrastructure or, preferably, accessed dynamically from external data providers. Direct integration with external spatial data servers through open and interoperable standards enables real-time synchronisation and reduces data duplication.

The framework adopts a layered stack of open standards: OGC API–Features and OGC API–Maps for vector and map services, OGC 3D Tiles for scene delivery, and the OGC SensorThings API for sensor telemetry; CityGML is used as the semantic reference for urban objects and IFC is supported for building-level details. Metadata is described using the ISO 19115/19,139 series. Object identifiers are preserved natively from each source (e.g., Vicmap Property IDs, council building IDs, sensor URIs), which preserves institutional provenance, avoids collisions during integration, and retains round-trip linkage to the authoritative custodian. Leveraging these standards helps mitigate data silos, supports cross-organisational collaboration, and ensures the maintenance of a single, authoritative source of truth across the SDT ecosystem.

3.1.2 3D data

Three-dimensional datasets are fundamental to SDTs, as they enable realistic representation of the built and natural environment and support advanced spatial analysis. The 3D Data Management component ensures that diverse 3D data formats—BIM/IFC at building scale, LiDAR point clouds, photogrammetric meshes, and CityGML city models at LOD1–LOD3 — and publishes them to the client as OGC 3D Tiles for streaming visualisation through CesiumJS. Typical operational resolutions are sub-metre for building footprints and BIM-derived envelopes, and 1 m for the digital surface model where available.

Compared to 2D data, 3D datasets capture vertical relationships and spatial complexity more effectively. However, this richness introduces challenges related to data interoperability, semantic consistency, and information loss during format conversion. The platform addresses these through translator modules that map source schemas onto the CityGML/IFC conceptual models while retaining native identifiers, producing semantically interpretable data structures suitable for rendering, querying, and analytical operations.

3.1.3 Live data

Live Data Management enables the continuous ingestion and processing of near-real-time information from a wide range of external sources. These may include traffic updates, weather feeds, emergency alerts, social media streams, and volunteered geographic information (VGI). Such data streams are essential for maintaining situational awareness and supporting time-critical decision-making within the SDT.

This component relies on a set of connectors that monitor, synchronise, and harmonise incoming data feeds with varying formats and update frequencies. Sensor telemetry is ingested through the OGC SensorThings API and lightweight device channels use MQTT; every incoming event is time-stamped at millisecond resolution to enable replay and time-sliced querying. Natural language processing (NLP), natural language understanding (NLU), and GeoAI utilities are applied to clean, interpret, and spatially contextualise unstructured or semi-structured information, enabling its integration into SDT analytics and dashboards. In the current deployment these utilities are invoked at ingestion time; broader AI/ML predictive analytics remain a forward work item (Section 5).

3.2 Real-time data streaming

In addition to live data feeds, SDTs must support high-throughput, low-latency real-time data streams, typically generated by sources such as IoT sensor networks, CCTV systems, vehicle telemetry, and operational transaction systems. In the present deployment the streaming engine is built on Apache Kafka (as the high-throughput backbone for event durability and replay) complemented by MQTT brokers for constrained-device telemetry; this dual-protocol design covers both server-grade feeds and low-power field sensors under a unified event model.

Beyond real-time processing, the platform buffers and persists streaming data to enable historical analysis, pattern recognition, and the training of predictive models, thereby extending SDT capabilities from reactive monitoring to proactive forecasting and scenario evaluation.

3.3 Models and analytical capabilities

Analytical, predictive, and decision-support functions in SDTs are realised through a combination of system-level models and extensible computational components. Built-in models are typically designed to address domain-specific tasks and may include simulations for urban flooding, bushfire spread, pedestrian dynamics, or transport network performance. These models are often computationally intensive and rely on tightly coupled datasets and optimised execution environments.

To enhance extensibility, SDTs should also support the integration of third-party tools and analytical libraries through software development kits (SDKs) and application programming interfaces (APIs). This approach allows advanced artificial intelligence (AI), machine learning, and simulation frameworks to be embedded within the SDT environment, significantly expanding its analytical scope.

In parallel, user-contributed plug-ins provide a flexible mechanism for community-driven innovation. By enabling users to contribute scripts and algorithms, implemented in languages such as Python, R, and Java, the SDT platform can evolve continuously while capturing domain expertise distributed across its user base.

3.4 User and access management

SDTs must operate as collaborative, multi-user platforms supporting stakeholders from different organisations and disciplinary backgrounds. A comprehensive User Management framework is therefore essential to manage identities, resources, and collaborative workflows within the SDT environment.

Role-Based Access Control forms the foundation of secure resource management, enabling fine-grained control over permissions for datasets, models, and plug-ins. Through configurable access policies, users can grant or restrict create, read, update, and delete privileges at both individual and group levels, ensuring data security while facilitating cross-organisational collaboration.

3.5 Open ecosystem and API management

SDT is designed as an open, extensible ecosystem rather than a closed platform. A dedicated API Management layer exposes a consistent and governed surface to external applications and models. In the present implementation this surface comprises OGC API–Features, OGC API–Maps and OGC API – 3D Tiles for spatial services; the OGC SensorThings API for IoT streams; and a set of REST endpoints that cover tool lifecycle, scene binding, and analytical service invocation. The layer enforces authentication, authorisation, and quota control, balances computational loads between the SDT core and peripheral applications, and provides the integration surface against which user-contributed plug-ins (see Section 3.3) are built.

4 Case studies

In this section, we will explain the case studies that showcase applications of SDT in urban built environments. The three case studies demonstrate how the proposed SDT framework is operationalised through different analytical modules. Each case study connects a specific decision-support task to the shared SDT data environment, while using different input datasets, modelling assumptions, and analytical procedures. These case studies show that the framework is not limited to visualisation, but supports rule-based analysis, simulation integration, and scenario-based decision support for urban built environments.

4.1 Development envelope control

FIGURE 2

DEC consumes the Vicmap Property cadastre (metre-level), building footprints (sub-metre), road network, and the setback/height/floor-space controls parsed from the Victoria Planning Provisions. Rules are treated as hard deterministic constraints on the generated envelope; no stochastic behaviour is assumed. Formal ground-truth validation against planner-issued envelopes has not yet been performed; a benchmarking exercise with partner councils is planned as part of the next development phase.

This tool streamlines compliance by automatically applying planning rules and providing immediate feedback. Through interactive editing, planners can experiment with different setback scenarios and evaluate microclimatic effects (e.g., solar access) before finalising designs. Because DEC operates within the SDT platform, envelopes are visualised alongside existing buildings, infrastructure and environmental datasets, fostering holistic decision making.

4.2 PedDesign: simulating pedestrian movement

FIGURE 3

This application uses a tram-platform BIM (sub-metre) together with MassMotion agent trajectories supplied by ARUP at per-second resolution, including agent counts, gate configurations, and entry/exit flows. Model assumptions are inherited from MassMotion: agents follow least-cost path-finding with local collision avoidance, demand is defined exogenously (counts and schedules) rather than derived from a demand model, and distancing breaches are computed geometrically over simulated trajectories. Validation to date is bounded by the validation already established for MassMotion itself. Site-specific validation against observed pedestrian counts, together with sensitivity analysis on agent-count and distancing-threshold assumptions, is planned in collaboration with ARUP.

4.3 Water-sensitive urban design

FIGURE 4

In the Water-Sensitive Urban Design case study, the SDT framework was operationalized to support scenario-based stormwater-capacity assessment. The data management layer integrated catchment polygons, drainage-network data, road centerlines, parcel boundaries, and surface-elevation information. The 3D data layer provided the spatial context for visualizing catchments, proposed road modifications, drainage assets, and vulnerable areas within the SDT environment. The analytical layer implemented a capacity-based stormwater assessment model that estimated runoff and drainage requirements based on catchment geometry, impervious surface assumptions, road-network modifications, and design-storm thresholds. The user-interface layer enabled planners to draw or modify road segments, test alternative land-use and infrastructure scenarios, and compare stormwater-capacity implications across near-, mid-, and long-term adaptation options. The outputs included capacity indicators, spatially located vulnerability areas, and comparative scenario results to support water-sensitive planning decisions.

The pilot consumes catchment polygons and drainage-network data supplied by Melbourne Water, Vicmap road centrelines, a digital surface model (1 m where available), and Fishermans Bend parcel boundaries. In its current form the tool applies a capacity-based simplification rather than a coupled hydrodynamic simulation: it estimates runoff and drainage need from catchment geometry, impervious fraction, and design-storm thresholds. A full hydrological coupling—integrating Melbourne Water’s hydrodynamic models and calibrating outputs against historical flood events at Fishermans Bend—is the next development phase.

5 Discussion

The developed SDT framework provides significant benefits for the management of urban built environments by providing a dynamic, integrated digital representation of physical assets, spatial relationships, and urban processes. By combining geospatial data, building and infrastructure models, real-time sensor information, and analytical engines within a single environment, SDT enables us to move beyond static mapping toward dynamic and predictive intelligence in urban areas. This integrated view supports evidence-based decision-making by allowing different stakeholders, such as planners, engineers, and policymakers, to understand how urban systems interact spatially and temporally, and to assess the implications of interventions. One of the key advantages of the SDT framework is its ability to enhance urban planning and development processes. Through 3D and 4D representations of the built environment, planners can evaluate land-use scenarios, building compliance, and environmental impacts with greater accuracy than traditional 2D systems. This capability shortens approval timelines, and improves coordination among multiple stakeholders, including government authorities, developers, and communities.

The developed SDT also deliver significant benefits for infrastructure and asset management by linking spatial data with asset attributes, condition data, and performance indicators. This allows infrastructure owners and operators to monitor assets over their entire lifecycle, identify emerging risks, and implement predictive maintenance strategies. For complex urban environments, particularly those with dense underground utilities, the developed SDT improves visibility of spatial and legal relationships between assets, reducing the likelihood of service disruptions, safety incidents, and costly construction conflicts.

The developed SDT improves urban governance, transparency, and stakeholder engagement by offering a shared digital environment that can be accessed by multiple actors. Intuitive visualisation and analytics tools support clearer communication of complex spatial information, fostering collaboration across agencies and increasing public trust in planning and infrastructure decisions. As a scalable and extensible digital foundation, SDTs also future-proofs cities by enabling the integration of emerging technologies, analytical methods, and data sources, ensuring that urban management systems remain responsive to growing complexity and evolving societal needs.

While the developed SDT provides significant benefits, several technical and organisational challenges surfaced during implementation. On the backend, the central difficulty was reconciling very different update cadences and provenance rules between legacy council datasets and real-time sensor feeds, and transforming 3D models of varying origin (LiDAR-derived, BIM/IFC, photogrammetric meshes) into a uniform tiled format without loss of semantic content. Cross-tenant data-sharing rules differ between state, local, and private custodians, which required explicit RBAC policies and auditable data-access pathways. On the frontend, the main engineering challenges were streaming and rendering large tiled datasets at interactive frame rates, keeping the analytical state of each tool temporally consistent with the shared base scene, and providing a coherent user experience across tools authored by different teams—addressed through shared component libraries, and level-of-detail tuning in CesiumJS. Ensuring interoperability across the disparate data sources required coordinated adoption of common standards (OGC API services, CityGML, IFC, SensorThings API, ISO 19115 metadata) and the preservation of source-native identifiers throughout the platform.

The current evaluation of the SDT framework is primarily demonstrative, focusing on whether the proposed architecture can integrate heterogeneous spatial datasets, support interactive 3D visualization, and host domain-specific analytical tools within a shared operational environment. A full quantitative benchmarking exercise was outside the scope of this initial implementation; however, future validation will be structured around both platform-level and application-level metrics. At the platform level, performance evaluation will include data-ingestion latency, API response time, 3D tile loading time, rendering frame rate, memory consumption, and scalability under increasing numbers of users, data layers, and streaming events. At the application level, validation will assess the accuracy and reliability of analytical outputs generated by each case-study module. This includes benchmarking DEC-generated envelopes against planner-issued development envelopes, comparing PedDesign simulation outputs with observed pedestrian counts and movement patterns, and calibrating stormwater-capacity outputs against hydrological models and historical flood-event records.

Introducing SDTs into current workflows also requires significant organisational change. Traditional planning and infrastructure-management processes need to be re-engineered to incorporate SDT insights, and inter-institutional coordination remains critical. For example, sharing data between a water utility and a transportation agency through a common SDT can encounter bureaucratic and legal hurdles, so coordination among government agencies, private sector, academia, and the public is needed for SDTs to succeed. Another challenge is that building and operating an SDT is expensive; many cities, especially in developing regions, face budget constraints and struggle to justify these expenses against other needs, which risks widening the digital divide in urban innovation.

Ensuring public engagement is both a challenge and a necessity for SDTs. Lack of community buy-in can limit the data available and even generate political pushback against smart-city programs. Ethical considerations also arise regarding how SDT analytics inform urban policy: if an SDT analysis suggests interventions that favour certain neighbourhoods over others, a fair and transparent decision-making process must be in place.

6 Conclusion

This paper presented a standards-based SDT framework for urban built environments and demonstrated its operational deployment through three applied case studies in Victoria, Australia. The key contribution of the study is the translation of SDT principles into an implementation-oriented framework that integrates heterogeneous 2D, 3D, semantic, and near-real-time datasets with modular analytical and decision-support tools. The framework combines spatial data management, 3D data streaming, live-data ingestion, analytical services, API-based interaction, and web-based visualisation within a single operational environment.

The three case studies demonstrate the extensibility of the framework across different urban domains. The DEC tool shows how planning rules can be translated into spatially explicit building-envelope outputs; PedDesign shows how pedestrian simulation outputs can be integrated with 3D built-environment context; and the Water-Sensitive Urban Design tool shows how stormwater-capacity assessment can support scenario-based infrastructure planning. These examples demonstrate that SDTs can move beyond visualisation platforms toward analytical decision-support environments for urban planning and infrastructure resilience. Despite the demonstrated benefits, several challenges remain. Achieving full technical interoperability across data standards, platforms, and institutional systems continues to be a major barrier, particularly in complex institutional contexts. Real-time data ingestion and processing at a city scale also pose significant computational and governance challenges, while institutional adoption is often constrained by organisational silos, legacy systems, and skills gaps. In addition, limited public awareness and trust in SDTs may hinder their broader uptake and societal impact.

Future work will focus on systematic quantitative validation, including platform-performance benchmarking, scalability testing, comparison of analytical outputs with observed or authoritative datasets, and further integration with institutional workflows. These activities will strengthen the framework’s reliability and support its broader adoption in urban built-environment decision-making.

References

• 1

  AheleroffS.XuX.ZhongR. Y.LuY. (2021). Digital twin as a service (DTaaS) in industry 4.0: an architecture reference model. Adv. Eng. Inf.47, 101225. 10.1016/j.aei.2020.101225

  • CrossRef
  • Google Scholar

• 2

  ANZLIC (2019). Principles for spatially enabled digital twins of the built and natural environment in Australia. Prepr. ANZLIC Barton, Aust, 1–25. Available online at: https://www.anzlic.gov.au/sites/default/files/files/principles_for_spatially_enabled_digital_twins_of_the_built_and_natural_.pdf (Accessed December 10, 2025).

  • Google Scholar

• 3

  BattyM. (2024). Digital twins in city planning. Nat. Comput. Sci.4, 192–199. 10.1038/s43588-024-00606-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 4

  CalzatiS. (2023). No longer hype, not yet mainstream? Recalibrating city digital twins’ expectations and reality: a case study perspective. Front. Big Data6, 1236397. 10.3389/fdata.2023.1236397

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 5

  Department of Customer Service (2025). NSW Spatial Digital Twin. Available online at: https://nsw.digitaltwin.terria.io/ (Accessed November 08, 2025).

  • Google Scholar

• 6

  Department of Transport and Planning (2025). Digital Twin Victoria. Available online at: https://digitaltwin.vic.gov.au/ (Accessed September 08, 2025).

  • Google Scholar

• 7

  Javaherian PourE.AtazadehB.RajabifardA.SabriS.NorrisD. (2025). A graph data model for CityGML utility network ADE: a case study on water utilities. ISPRS Int. J. Geo-Information14 (12), 493. 10.3390/ijgi14120493

  • CrossRef
  • Google Scholar

• 8

  KopponenA.HahtoA.VillmanT.KettunenP.MikkonenT.RossiM. (2024). “Personalised public services powered by AI: the citizen digital twin approach,” in Research Handbook on Public Management and Artificial Intelligence (Cheltenham, United Kingdom: Edward Elgar Publishing Limited).

  • Google Scholar

• 9

  LavigneH. (2024). “DUBAI digital twin,” in Arcgis Storymaps. Available online at: https://storymaps.arcgis.com/stories/4ae54ecb2d8e4640b491de1fc319cffc (Accessed December 06, 2024).

  • Google Scholar

• 10

  LiuC.ZhangP.XuX. (2023). Literature review of digital twin technologies for civil infrastructure. J. Infrastructure Intell. Resil.2(3):100050. 10.1016/j.iintel.2023.100050

  • CrossRef
  • Google Scholar

• 11

  LiuP.ZhaoT.LuoJ.LeiB.FreiM.MillerC.et al (2023). Towards human-centric digital twins: leveraging computer vision and graph models to predict outdoor comfort. Sustain. Cities Soc.93, 104480. 10.1016/j.scs.2023.104480

  • CrossRef
  • Google Scholar

• 12

  MattarC.ArairoW.KhatibM.AzarA. (2026). Integrating sustainable material performance and project success factors: insights from the lebanese construction sector. Front. Built. Environ.11, 1734576. 10.3389/fbuil.2025.1734576

  • CrossRef
  • Google Scholar

• 13

  MazzettoS. (2024). A review of urban digital twins integration, challenges, and future directions in smart city development. Sustainability16 (19), 8337. 10.3390/su16198337

  • CrossRef
  • Google Scholar

• 14

  MetcalfeJ.EllulC.MorleyJ.StoterJ. (2024). Characterizing the role of geospatial science in digital twins. ISPRS Int. J. Geo-Information13 (9), 320. 10.3390/ijgi13090320

  • CrossRef
  • Google Scholar

• 15

  NasimM.RajabifardA.ChenY.SamaliB. (2026). A demonstration of a digital twin framework for structural health monitoring: application to bridge infrastructures. J. Infrastructure Intell. Resil.5 (1), 100184. 10.1016/j.iintel.2025.100184

  • CrossRef
  • Google Scholar

• 16

  OlcottS.MullenC. (2020). Digital twin consortium defines digital twin. Digit. Twin Consort.3.

  • Google Scholar

• 17

  PandeyB.GhoshA. (2023). Urban ecosystem services and climate change: a dynamic interplay. Front. Sustain. Cities5, 1281430. 10.3389/frsc.2023.1281430

  • CrossRef
  • Google Scholar

• 18

  RajabifardA.SabriS.ChenY. (2022). “Digital Twin for the next Generation of Urban Land Administration and 3D Spatial Planning,” in FIG Congress 2022, Volunteering for the future - Geospatial excellence for a better living. Warsaw, Poland, 11–15 September 2022. Available online at: https://www.fig.net/resources/proceedings/fig_proceedings/fig2022/papers/ts02g/TS02G_rajabifard_sabri_et_al_11729.pdf.

  • Google Scholar

• 19

  ShirowzhanS.TanW.SepasgozarS. M. E. (2020). Digital twin and CyberGIS for improving connectivity and measuring the impact of infrastructure construction planning in smart cities. ISPRS Int. J. Geo-Information9 (4), 240. Preprint, MDPI. 10.3390/ijgi9040240

  • CrossRef
  • Google Scholar

• 20

  SwayneM. (2024). Naver starts digital twin project for smart cities in Saudi Arabia. Digit. Twins, News. Available online at: https://digitaltwininsider.com/2024/08/09/naver-starts-digital-twin-project-for-smart-cities-in-saudi-arabia/August 9 (Accessed December 12, 2025).

  • Google Scholar

• 21

  WalkerA. (2023). Singapore’s Digital Twin—From Science Fiction to Hi-Tech Reality, 4. (Geneva, Switzerland: International Federation of Consulting Engineers), 2023.

  • Google Scholar

• 22

  WuD.ZhengA.YuW.CaoH.LingQ.LiuJ.et al (2025). Digital twin technology in transportation infrastructure: a comprehensive survey of current applications, challenges, and future directions. Appl. Sci.15 (4), 1911. 10.3390/app15041911

  • CrossRef
  • Google Scholar

• 23

  ZhuM.JinJ. (2025). Data‐driven urban digital twins and critical infrastructure under climate change: a review of frameworks and applications. Urban Plan.10, 10109. 10.17645/up.10109

  • CrossRef
  • Google Scholar