The role of digital twins in mitigating urban heat islands

As cities continue to expand, the need for innovative, sustainable solutions to urban heat island (UHI) effects becomes more urgent. Digital twins offer a promising approach to addressing UHI challenges such as higher energy consumption, poorer air quality and adverse health impacts. This article gives examples of their effectiveness in urban planning – from a better understanding of how heat behaves supported by machine learning models, to improved transparency and community engagement – and presents real-world case studies highlighting how cities can leverage this technology for sustainable urban development.

Urban heat island (UHI) effects occur when cities become significantly warmer than their surrounding rural areas due to human activity, dense construction and limited green spaces. These effects pose significant challenges to the health and well-being of city dwellers, particularly as urbanization continues to grow. With two-thirds of the global population projected to live in cities by 2050, addressing UHI challenges is critical for creating more livable and climate-resilient urban environments.

One of the primary factors driving UHI is urban morphology: the design, layout and materials used in city infrastructure. Dense environments with tall buildings, narrow streets and heat-retaining materials intensify UHI conditions. For example, ‘street canyons’ – created when buildings are much taller than the width of streets – trap heat and reduce airflow, worsening UHI effects. The lack of green spaces further exacerbates this problem by limiting natural cooling opportunities. In contrast, cities that incorporate green infrastructure, like trees, parks and green roofs, can significantly reduce surface temperatures, improve thermal comfort and create more livable urban environments.

Urban form vs environmental performance

Digital twins (DTs) offer a promising approach for integrating these nature-based solutions into urban planning to mitigate the UHI effect. DTs are dynamic virtual models that replicate real-world environments and are continuously updated with real-time data, enabling urban planners to simulate different strategies for reducing urban temperatures. Planners can test the impact on UHI formation of adding green spaces, altering building materials or redesigning streets before making actual changes. By providing a live, interactive view of city systems, DTs allow for more informed, data-driven decisions. They also offer insights into the relationship between urban form and environmental performance, such as how building placement affects airflow and solar energy potential – both of which influence UHI intensity.

One innovative use of digital twin technology involves 3D city modelling and environmental simulations. With the integration of geospatial data like Lidar and satellite imagery, planners can create detailed models that simulate the impact of different urban forms on temperature distribution. These models help planners assess how changes to building height, orientation and material properties affect heat retention in specific areas. By visualizing factors like solar radiation and wind patterns, DTs become a powerful tool for predicting the outcomes of planning decisions and optimizing strategies for better urban cooling.

Machine learning models

Research has shown the effectiveness of digital twins in urban planning. For example, machine learning models within digital thermal twins allow planners to understand how heat behaves in urban environments and predict how different layouts or materials might affect temperature. The 3D visualizations provided by DTs give planners an opportunity to see how buildings, streets and green spaces interact, either to trap heat or improve airflow, which helps cool down urban areas. As a result, DTs offer cities the ability to make smarter, more climate-responsive planning decisions.

Another valuable application of digital twins is their ability to involve the public in urban planning. Interactive models allow community members to see how proposed changes could affect their neighbourhoods and provide feedback, ensuring that planning decisions reflect the needs and concerns of residents. This engagement promotes transparency and collaboration in the urban planning process.

Integration challenges

However, integrating digital twins into urban planning still presents some challenges. Significant hurdles include data quality, real-time sensor integration, and the costs associated with maintaining DT systems. Additionally, digital twins require a steady stream of reliable data, which can be difficult to collect consistently across large urban areas. Nevertheless, as technology improves and more cities adopt DTs, these tools will likely become essential for tackling environmental challenges like the UHI effect.

The two case studies below highlight how cities can leverage digital twin technology in urban planning processes to combat UHI effects and support sustainable urban development.

Case study: Enschede, the Netherlands

Enschede, a city in Twente in the eastern region of the Netherlands, has been dealing with the impacts of urban heat islands due to its increasing urbanization and the associated rise in impervious surfaces. The city is particularly vulnerable to the UHI effect during heatwaves, which have become more frequent and intense in recent years.

The challenge in this case was to develop a tool for addressing UHI through data-driven decision-making. This meant creating a 3D environment that could visualize the modelling of different weather parameters, along with the capability to model the thermal comfort of a modified, built environment. It was not enough to simply visualize the pattern of thermal comfort using physiological equivalent temperature (PET) in different complex combinations of weather parameters (such as wind flow intensity, wind direction, humidity and air temperature). The aim of the project was also to see the direct effect of modifying the built environment on the immediate neighbourhood.

In collaboration with the University of Twente, the entire project was developed on open-source software and methodologies. Open-source 3D information from the national AHN-4 Lidar point cloud dataset was used to form the base conditions of the 3D model, and the weather information was taken from the EPW Typical Meteorological Year 5.2 global dataset, which used the Twente airfield weather sensor. This allowed for the modelling of weather conditions with up to one hour of accuracy in any given typical month.

Using data extracts from the two equinoxes and two solstices, the pattern of thermal comfort was calculated for the existing built form. Then, the 3D environment itself was modified to introduce tall buildings while eliminating small buildings in a given neighbourhood. The thermal comfort patterns were again calculated on different days, with the equinoxes forming the average conditions while the solstices formed the most extreme ones.

As a result, Enschede’s digital twin system is a detailed 3D model of the city that simulates patterns of thermal comfort using a modified PET indicator. This model allows city planners to visualize how different parts of the city react to changes in the built form in terms of thermal comfort on the ground. It also enables them to assess how various interventions can be planned to offset the predicted thermal hotspots created by introducing new buildings. The process of identifying developed hot zones and cool zones due to newly built forms involves using simulations to assess how changes in urban structures, such as the construction of new buildings, impact the distribution of heat in a city (Figure2).

The goal, of course, remains to replace the global weather dataset with real-time sensor information to create an even more accurate weather model to use as the basis for decision-making. However, until the network of sensors can be strategically planned and placed, the global dataset suffices to give an overall idea of the thermal comfort trends to be expected.

Case study: Wuppertal, Germany

Wuppertal, a city situated in a valley along the Wupper river in Germany, faces unique UHI challenges due to its topography and dense urban fabric. The limited airflow along the river exacerbates the UHI effect by trapping heat in the lower parts of the city. With projections indicating a significant increase in the number of hot days by 2060, effective planning is essential to mitigate UHI effects.

A project was launched to develop an interactive platform for visualizing various scenarios and their potential impacts on decision-making in Wuppertal. This resulted in the Digital Twin-Based Planning Support System (DT-PSS) to simulate the effects on UHI formation of various urban planning scenarios, such as increasing green spaces, constructing new buildings, increasing population density or modifying layouts. The DT-PSS framework integrated past data collection and analysis, the creation of a 3D city model using real-time temperature data, and the evaluation of scenarios to predict UHI formation. Machine learning, real-time data and 3D city modelling were combined to create an innovative simulation system. A random forest model was trained using nine key predictor variables, including Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI) and population density. These variables were selected for their strong statistical correlation with temperature.

This predictive model was then combined with real-time temperature data collected from ground sensors in Wuppertal, allowing continuous updates to the digital twin. The integration of real-time data enabled the system to simulate temperature variations in response to different planning scenarios. The trained model was incorporated into a 3D city model of Wuppertal and imported into the Unreal Engine (UE) platform, which facilitates advanced visualization and interactive simulations. This allows users to explore the impact on UHI formation of interventions such as altering building density or adding green spaces (Figures 3 and 4).

To ensure practical applicability, the system was designed with input from urban planners and municipal officials. A workshop was conducted where stakeholders used virtual reality (VR) glasses to interact with the model, explore ‘what-if’ scenarios and provide feedback on the system’s usability and relevance, making it more responsive to their needs.

A rough mobile version of the DT-PSS has also been developed to enhance accessibility and usability for urban planners and stakeholders. The advantages of the mobile version can be extended to citizens, as it empowers them to engage actively in the urban planning process. Providing an intuitive interface makes it possible for citizens to access and interact with the digital twin, allowing them to visualize how different urban development scenarios may affect their lives and neighbourhoods. This capability helps transparency and collaboration between planners and residents, enabling citizens to provide input and feedback on proposed interventions. Moreover, the mobile platform encourages a sense of involvement in local planning decisions, as citizens can participate in assessing the immediate effects of environmental changes and advocate for strategies that prioritize their community’s well-being. Overall, the mobile DT-PSS enhances public engagement, making urban planning more inclusive and responsive to the needs of the community.

Acknowledgements

The authors would like to express their sincere gratitude to Dr Pirouz Nourian and Dr Christine Pohl Benjamin Bleske for their valuable contributions to the joint supervision and collaboration.

Further reading

• Cárdenas, I., Koeva, M., Davey, C. & Nourian, P., (2024). Urban digital twin-based solution using geospatial information for solid waste management. Sustainable Cities and Society
• Campoverde C, Koeva M, Persello C, Maslov K, Jiao W, Petrova-Antonova D. Automatic Building Roof Plane Extraction in Urban Environments for 3D City Modelling Using Remote Sensing Data. Remote Sensing. (2024); 16(8):1386. https://doi.org/10.3390/rs16081386
• Cárdenas, I., Koeva, M., Davey, C. & Nourian, P., (2024). Solid Waste in the Virtual World: A Digital Twinning Approach for Waste Collection Planning. Springer, p. 61-74 (Lecture Notes in Geoinformation and Cartography; vol. XIII)
• Kumalasari, D., Koeva, M., Vahdatikhaki, F., Petrova Antonova, D., & Kuffer, M. (2023). Planning Walkable Cities: Generative Design Approach towards Digital Twin Implementation. Remote Sensing, 15(4), 1088
• La Guardia, M., & Koeva, M. (2023). Towards Digital Twinning on the Web: Heterogeneous 3D Data Fusion Based on Open-Source Structure. Remote Sensing, 15(3), 721.
• Kumalasari D, Koeva M, Vahdatikhaki F, Petrova Antonova D, Kuffer M. Planning Walkable Cities: Generative Design Approach towards Digital Twin Implementation. Remote Sensing. 2023; 15(4):1088. https://doi.org/10.3390/rs15041088
• Ying, Y., Koeva, M., Kuffer, M., & Zevenbergen, J. (2023). Toward 3D Property Valuation—A Review of Urban 3D Modelling Methods for Digital Twin Creation. ISPRS International Journal of Geo-information, 12(1). https://doi.org/10.3390/ijgi12010002
• The 3D sharable model created using Cesium: https://ion.cesium.com/stories/viewer/?id=d2667679-81e9-4069-9e97-db9c98711f68
• Functionality of the DT-PSS: https://youtu.be/n_TXYSCh74U
• Mobile version of the DT-PSS: https://youtu.be/Z2yDePHFJVE