New Research Publication on the role of AI and native ecosystems in Urban rehabilitation

FROM GREY TO GREEN: MODELLING EUROPEAN INDUSTRIAL CITIES THROUGH NATIVE ECOSYSTEMS AND DIGITAL LANDSCAPE INNOVATION

Authors: Precious Ovat, Faithman Ovat

Keywords: Native Ecosystems, Nature-based solutions, urban landscapes, sustainability, digital, European cities

Abstract

Native ecosystems predate human settlements, and native plants have always maintained their existence through interactions with animals and microorganisms. As European industrial cities aim for sustainability, recent research highlights the pivotal role of urban green spaces in biodiversity conservation and ecosystem resilience. Our paper examined how native ecosystems when integrated with digital innovation, can help redefine urban landscapes and restore ecological function. We aimed to ascertain (a) if natural rehabilitation methods and technology can develop a comprehensive strategy for reviving the landscape of degraded cities and urban areas that have undergone industrial decline, rapid urbanization, or climatic stress and (b) how we can use artificial intelligence to enhance our understanding of native environmental patterns to optimize landscape interventions. The methodology involved a comparative case study analysis of papers that discussed the role of technology in preserving native ecosystems and the application of Artificial Intelligence (AI) in landscape architecture. The conceptual framework focused on ecological urbanism, nature-based solutions, and digital landscape innovation. Our findings suggest that nature-based rehabilitation methods, such as community gardens, urban living labs, and policy-led greening strategies, have been instrumental in ecological restoration and social regeneration, especially in post-industrial cities. While AI was not widely applied in current case studies, its potential for ecosystem monitoring, afforestation planning, pollution modelling, and predictive analytics is evident. We conclude that native ecosystems, supported by AI-driven planning and digital tools, can offer replicable, performance-tested, and sustainable strategies for post-industrial urban transformation. Future studies should focus on developing context-specific AI applications and enabling supportive policy frameworks to integrate digital and ecological approaches in urban planning.

1.1       Introduction

1.1.1    The effect of Industrialization on European cities

In the 18th century, Europe experienced a rise in industrialization, with England and Scotland at the epicenter; this process shaped the spatial development, human population, and economy of the cities where it occurred, pushing away from an agricultural and towards an industrial economy [1]. However, degradation of their natural environments was the price for the economic boom and expanded job market which they enjoyed [1]. Also, the rural-urban migration to industrial cities during that era created a negative impact on the environment, through pollution from industrial activities and greenhouse gas emissions, resulting from long-distance vehicular transportation of goods [2]. Additionally, the intense impact of climate change aggravated the vulnerabilities of their environment while also highlighting the limitations of traditional grey infrastructure in adapting to contemporary urban needs [3]. Subsequently, when Europe shifted towards globalization and the industrial era was brought to a halt, these cities were left with land and infrastructure that needed repurposing [1].

1.1.2    The "Grey to green" concept

In the management and planning of cities, “grey” refers to built-up up urban areas while “green” refers to areas of vegetative cover [4]. The interplay of grey and green spaces could be valuable for planning urban growth and ensuring sustainability in urban areas [4]. Green spaces in urban areas have been shown to confer benefits including improved quality of human life, enhanced thermal comfort, strengthened food security, and mitigation of the impacts of climate change [5]. Thus, we discuss the concept of "Grey to Green" as a transition from construction-dominant urban and landscape design solutions towards nature-based vegetative strategies. A good example of this transition occurred in Mirafiori, where greening projects became integral to the regeneration of post-industrial areas, improving social spaces [2].

Figure 1: Map of Mirafiori Sud urban living lab, from proGIreg

1.1.3    Native Ecosystems

By definition, a native ecosystem is one in which plants, animals, and microorganisms coexisted prior to human disturbance. Essential species, such as dominant photosynthesizing plants, top carnivores, significant decomposers, and nitrogen-fixers, are necessary for a native ecosystem to sustain itself and operate independently [6]. Native ecosystems possess distinctive characteristics, and the substitution of native flora with non-native species can significantly disrupt urban food networks. Moreover, native plants adapt better to local soil and climate, support a greater diversity of local wildlife, and require less maintenance than non-native options [7]. Crucially, a shift to green infrastructure is not merely aesthetic or ecological, as it requires functionality that targets urban resilience, public health, biodiversity restoration, and spatial justice. It would involve the successful integration of permeable surfaces, biodiverse green spaces, green roofs, and restored water systems to replace or overlay grey systems [8]. It is reasonable for us to infer that NBS are well-suited for a grey to green transition in cities.

1.1.4    Digital Landscape Innovations

Digital landscape innovations are transforming how degraded urban landscapes undergo rehabilitation and redesign. For example, digital twins (DT) can be used to replicate digital models of real-time physical systems through simulations that enhance urban analytics and city planning [9], and so it can be used alongside native ecosystems in post-industrial European cities to support participatory design and continuous monitoring, ensuring green interventions are resilient, sustainable, and responsive to changing climates and social needs. Teutscher, D et al. (2025) presented a novel urban-focused DT that can be used to produce advanced simulations for pollution analysis and urban airflow modelling [9]. Such innovation can inform planners on effective application of nature-based solutions (NBS) to tackle adverse environmental effects of air pollution in urban landscapes. Furthermore, machine learning and AI can enhance our understanding of native environmental patterns to optimise landscape interventions by monitoring ecosystems and biodiversity, managing afforestation and deforestation, minimising adverse environmental impacts in agriculture via disease detection, weed detection, soil sensing, and protecting peatlands to preserve these vital ecosystems [10]. Planners can identify the most suitable native species and optimal locations for ecological corridors and predict resilience scenarios using AI models trained on native plant species and urban performance measures. This adaptive quality raises the question of whether AI can enhance our ability to create urban spaces that are biodiverse, sustainable, performance-tested, and replicable in the real world.

1.1.5       Aim and research questions

This paper aimed to develop a strategy that utilises native ecosystems and digital innovation to revitalize post-industrial European cities. The questions guiding our paper are as follows:

a.      Can natural rehabilitation methods and technology develop a comprehensive strategy for reviving the landscape of degraded cities and urban areas that have undergone industrial decline, rapid urbanisation, or climatic stress?

b.      Could artificial intelligence be used to enhance our understanding of native environmental patterns to optimize landscape interventions, and how?

2.1       Literature review

The transformation of NBS into a fundamental component of sustainable urbanism reflects a broader shift from reactive environmental policies to proactive, systems-oriented design. Nature-based solutions refer to actions that protect and promote sustainable use of land and water ecosystems, tackle social, economic and environmental problems effectively, and simultaneously provide human well-being, ecosystem services and biodiversity benefits [2]. Russo et al. (2025) examined the contributions of both native and non-native plant species to biodiversity and ecosystem services. They emphasized the ecological value of native plants for supporting local wildlife, acknowledged the potential for non-native species to disrupt ecosystem dynamics, and advocated for the use of conservative strategies, which would integrate native species to ensure the preservation of ecosystem function in urban areas [5]. Hansen et al. (2015) emphasized the importance of integrating ecosystem services into urban planning [3]. However, critiques of NBS highlight concerns about implementations that prioritize aesthetic appeal over ecological integrity, a practice sometimes described as "greenwashing". Zarei and Shahab (2025) caution against bureaucratic or top-down NBS implementations that risk excluding local communities or failing to consider site-specific environmental histories. They support the collaborative development of green infrastructure adapted to local sites, particularly in European post-industrial cities with complex property ownership and zoning issues [8]. A good example of such collaborative planning is seen in Salzburg, Austria, where the city used public engagement to implement a policy (green space declaration 1985) which was later incorporated into the city’s development concept in 2007 [3].

Europe can benefit from policies that have been promoted by the European Union (EU), such as the biodiversity strategy for 2030, the green infrastructure strategy 2013, and common agricultural policy (CAP) reform 2023-27. Meanwhile, Stockholm in Sweden, is praised for being a forerunner for sustainable urban development, as it stands out for early adoption of the ecosystem services concept since the early 2000s [3]. Ecosystem services should integrate significant environmental issues in planning or political agendas to meet the needs of urban planners and managers. An analysis of plans in Polish cities showed that ecosystem services with a market orientation receive more attention in planning [3].

The opportunities for modelling and predicting the effects of green infrastructure are on the rise because of digital innovation. Cristina et al. (2020) discussed how Landsat data archives combined with advanced image processing and analysis methods are used to monitor the changes of green areas across 10,000 urban centres [4]. AI and machine learning was employed to analyse urban change, using built-up surface grids to model population distribution, and time series of high-resolution NDVI composites to extract and monitor vegetation from satellite data [4]. Digital twin technology is also used a lot in urban planning for environmental monitoring. Teutscher et al., (2025) present a digital twin framework for urban planning that integrates real-time meteorological data and computational fluid dynamics to enhance pollution predictions and support adaptive urban planning [9]. Although there has been an increase in initiatives using machine learning (ML) and artificial intelligence (AI) to address societal and global issues, there is still a need to determine how these technologies can be most effectively used to combat climate change [10].

3          Methodology

This study adopted a qualitative, comparative case study approach to analyse the integration of native ecosystems and digital tools in green strategies for post-industrial European cities. Our conceptual framework for understanding the dimensions of landscape transformation in formerly industrial areas draws upon theories of ecological urbanism, nature-based solutions, and digital landscape innovation. These perspectives provide a lens through which to explore how to leverage emerging technologies and native ecological processes to transform degraded urban landscapes.

We began our methodology by identifying five analytical themes through a review of current discourse in urban landscape architecture: ecological restoration, digital innovation, design approach, community and policy integration, and outcomes of ecological intervention. 

The next phase involved a keyword search on online databases and identified 250 articles published within the last forty years. These studies were passed through title screening and were subsequently narrowed down to 25 papers. To ensure relevance to the analytical themes, we then conducted abstract screening using our inclusion criteria to narrow the literature down to ten papers.

Next, we progressed to the third phase, which sought to reduce the literature to ten cases that qualify for the case study analysis phase; using our inclusion criteria, we selected studies that:

a.      Focused on cities with a post-industrial heritage.

b.      Documented existing or ongoing green transformation projects.

c.      Incorporated the use of native ecosystems; and

d.      Demonstrated application of digital tools and/or artificial intelligence (AI) in planning or environmental monitoring.

We initially utilized OpenAI's ChatGPT to grade each of the ten papers as either having fully met, partially met, or not met the screening criteria. Next, we reviewed the same papers to confirm their selection and identified three relevant case studies. involved. Although none of these three satisfied all four inclusion criteria, these satisfied at least three criteria. Finally, we conducted a thematic analysis of each study before interpreting and then narrating the findings.

4          Results and Discussions

4.1       Case 1 – Urban Green Resilience (Starczewski et al., 2023)

This paper examined the patterns of green space coverage in 32 Polish cities (including 12 post-industrial cities) between 2006 to 2018. The authors analysed the direction of transformation of green spaces during the 12-year timeframe using QGIS software, to ascertain whether there had been increased green spaces within urban areas in the cities. The authors sub-grouped the cities into low, medium and high intensity zones, depending on how much development the zones had undergone over the 12-year period.

Their findings revealed that the overall green space coverage in the post-industrial cities reduced by 0.1%, and low intensity zones were the only regions to realise a rise in green spaces. 2.6% of green spaces were converted to brown areas, with the overwhelming majority (approximately 50%) being converted to industrial and urban units. An overwhelming majority of green spaces (88%) were realised from agricultural land via secondary succession, with only a marginal (10.5%) proportion reclaimed from industrial and construction sites. In one of the post-industrial cities (Lódz´), the authorities sought to introduce policy to improve greening efforts by withdrawing from the sale of land for urban use. Another process suggested for greening was revitalization, which would involve restoring degraded urban areas while retaining their historic rhetoric value, with the benefit of aiding tourism and encouraging local engagement among residents.

The study showed that natural rehabilitation methods can be applied to revive the landscape of degraded cities, although there were concerns about costs when applying such changes in highly urbanized regions, and they were preferably applied to regions that had fewer concrete structures.

There was no mention of the use of artificial intelligence to study environmental patterns. However, the use of the Urban Atlas which utilises satellite imagery, as well as the use of GIS, proves that there is a place for technological advancements in reproducing the environments being studied. We could therefore hypothesize that there is room for the use of AI in studying and generating green patterns, which would generate data quickly and accurately based on available and verifiable databases.

4.2       Case 2 – Reviving urban greening through nature-based solutions (Dogan et al., 2023)

In this case study, the authors discuss the transformation of vacant post-industrial land in the Mirafiori Sud region of the city of Turin into urban living labs (ULLs). The paper highlighted policy changes aimed at creating green spaces in Mirafiori, including the Green Infrastructure Strategic Plan by the City of Turin which focused on ecosystem services, the FUSILLI project which sought to blend food and art with green transformation, and European-led projects (proGIreg and CONEXUS) which were focused on NBSs.  

The paper proves that comprehensive environmental rehabilitation strategies can be developed using natural methods, as the urban living labs projects are heavily based on NBS, with Orti Generali being a prime example. Orti Generali are achieving food production and social inclusion through establishment of green spaces and community gardens while simultaneously replacing degraded land. Also, in collaboration with Mirafiori Foundation, Orti Generali created long-term projects “the Grand Galà of Mirafiori” and “Being elderly in Mirafiori Sud”, projects which attracted many young and elderly people to take care of green areas. Additionally, the success of Orti Generali’s circular kiosk which utilises the products cultivated in the garden and creates a space for social cohesion, and a collaborative policy push, led to the city of Turin officially implementing the idea of the “circular pizza” which promotes sustainable cooking processes in the preparation of pizza.

This study did not directly refer to the use of artificial intelligence, but we could identify an opportunity for the use of artificial intelligence in the early phase of some projects. At the early stages of their project, proGIreg and FUSILLI proposed activities to convert local creativity into experiments and ended up engaging a wide selection of individuals to aid their experiment design prior to commencement of proper testing. The use of artificial intelligence in such experimental phases could potentially reduce time, personnel and financial costs in similar projects, and as such, may result in an increased capacity to run multiple studies simultaneously while reallocating expenses elsewhere.

4.3       Case 3: Tackling climate change with machine learning (Rolnick et al., 2022)

In our opinion, it is impossible to envisage a modern city without electricity. However, this study highlights that electricity systems are responsible for approximately 25% of global man-made greenhouse gas (GHG) emissions annually. This underscores the urgent need to transition to low-carbon energy sources such as solar, wind, hydro, and nuclear power. Machine learning (ML) enables this transition by forecasting electricity supply and demand using historical data, physical model outputs, images and even video data. It also balances electrical systems by determining optimal output from controllable generators and supports the efficient placement and maintenance of controllable sources like geothermal plants, nuclear facilities, and hydropower dams, using seismic and satellite data. Moreover, ML can reduce fossil fuel emissions by detecting methane leaks in natural gas pipelines and compressor stations via sensor networks or satellite data, prompting timely maintenance actions.

Similarly, transportation systems are vital to urban life but account for another 25% of global energy-related CO₂ emissions. Unlike electricity, transport emissions are harder to mitigate due to factors such as policy influence on end-users and the high energy density requirements of vehicular fuels. Nonetheless, emissions can be reduced through strategies such as minimizing travel demand, improving vehicle efficiency, adopting alternative fuels and electrification, and shifting to lower-carbon transport modes. ML-based disruptive technologies could reduce or replace transportation demand altogether. Additionally, ML contributes to vehicle energy efficiency through advanced consumption design, aerodynamic efficiency improvements, and autonomous vehicles (AV). It can also analyse travel behavior, aiding in the optimal planning of public transport routes and infrastructure.

Traditionally, building energy modelling and temperature forecasting relied on detailed physical models. However, ML methods like deep belief networks can improve accuracy with less computational cost. Satellite radar images can generate a global map of human settlements at meter-level resolution. Therefore, high-resolution satellite image segmentation can now precisely outline building footprints on a national scale. Meanwhile, ML can extract data on urban climate issues using web-scraping and text-mining. The integration of large, diverse data sources is essential for smart cities, with machine learning techniques such as data matching, fusion, and ensemble learning playing a critical role. When incorporated into urban planning, these smart city initiatives can significantly enhance sustainability and encourage low-carbon lifestyles. For example, ML and AI can coordinate heating and cooling networks, solar power, EV and bike charging stations, and optimize public lighting by adjusting intensity based on foot traffic patterns.

Precision agriculture has the potential to reduce soil carbon emissions while simultaneously enhancing crop yields, thereby contributing to reduced deforestation pressures. To support this, satellite imagery plays a critical role in estimating carbon sequestration and monitoring GHG emissions from terrestrial ecosystems. ML techniques can be employed to cost-effectively monitor forest and peatland health, predict fire risks, and support sustainable forestry practices. Given that peatlands store more carbon than any other terrestrial ecosystem makes their degradation a significant threat to climate stability. When peatlands dry, they release substantial amounts of carbon through decomposition and become highly susceptible to fire. Therefore, monitoring and protecting these ecosystems from drainage and drought is essential to preserving their carbon storage capacity and mitigating GHG emissions. Additionally, ML supports the automation of afforestation by identifying suitable planting sites, monitoring plant health, detecting weed presence, and analysing long-term ecological trends. Combined with remote sensing technologies, ML can also facilitate the automatic detection of tree cover and deforestation through aerial imagery analysis. While forest fires contribute to atmospheric CO₂ emissions, it is important to recognise that low-intensity fires play a natural role in maintaining forest ecosystems. However, excessive fire suppression can lead to the accumulation of combustible biomass, thereby increasing the risk of large, uncontrolled wildfires. Such events can result in the destruction of mature trees, topsoil erosion, substantial carbon emissions, biodiversity loss, and prolonged ecosystem recovery. By leveraging advanced tools to identify high-risk regions, firefighters can implement controlled burns and strategically clear vegetation to reduce fuel loads and prevent catastrophic fire outbreaks.

While this case study does not focus on cities, it combines evidence from multiple urban systems to demonstrate how AI and machine learning can support green transformations in post-industrial cities. The emphasis has been placed on demonstrating the capabilities of artificial intelligence to detect, analyse, and model environmental data, to inform the restoration of native ecosystems. The opportunities of urban decarbonisation, climate resilience, and the challenge of ecosystem degradation are acute in post-industrial cities, therefore, the study establishes a conceptual and methodological foundation that cities with post-industrial histories can adapt when implementing native ecosystem restoration and smart landscape planning. Future research will focus on specific case studies where these innovations have been directly integrated into green urban rehabilitation frameworks.

5          Conclusion

Our study has examined the potential for the integration of native ecosystem and digital tools to serve as mutually reinforcing strategies in the sustainable transformation of post-industrial European cities. We sought to ascertain if natural rehabilitation and technology can create robust strategy for reviving landscape, coping with rapid urbanisation and climate change, and how landscape interventions can be optimised by using artificial intelligence to study environmental patterns. The research questioned whether natural rehabilitation methods and technology could revive degraded urban landscapes, and if AI could enhance our understanding of native patterns for the landscape interventions.

The findings show that the use of native flora as part of nature-based solutions is essential in the rehabilitation of post-industrial cities, due to cost efficiency, sustainability and adaptability. This is evident in initiatives such as the Mirafiori urban living labs and urban greening policies in Polish cities, where community engagement and ecological restoration improved environmental quality and social cohesion. However, limitations exist in highly urbanised zones due to cost and land use constraints. While few existing projects explicitly applied AI in these interventions, the literature highlights significant potential for AI and machine learning in urban sustainability. AI can enhance the design, implementation, and management of urban greening projects in various ways, including predictive modelling for pollution and climate scenarios, ecosystem monitoring, afforestation planning, and carbon stock estimation. Its use can improve efficiency, precision, and scalability in ecological planning.

Future research could explore the development of AI models trained on local biodiversity and urban performance data to support context-sensitive interventions. Policymakers should promote collaboration between ecologists, urban planners, and data scientists to support inclusive, digital, and ecologically resilient urban development. Integrating digital landscape tools into mainstream urban policy could position European cities at the forefront of climate-resilient planning in the 21st century.

Conflict of interest: The authors declare no conflict of interest.

Contributions: Precious Ovat – Conceptualisation, Data gathering, Writing (original draft preparation), Writing (review and editing). 

Faithman Ovat – Data gathering, Writing (review and editing).

Funding: This research received no external funding.

REFERENCES

1.      Starczewski, T., Rogatka, K., Kukulska-Kozieł, A., Noszczyk, T., and Cegielska, K. Urban green resilience: Experience from post-industrial cities in Poland. Geoscience Frontiers, 2023, vol. 14, no. 4, pp. 1-11

2.      Dogan, E., Cuomo, F., and Battisti, L. Reviving Urban Greening in Post-Industrial Landscapes: The Case of Turin. Sustainability, 2023, vol. 15, no. 12760, pp. 1-16

3.      Hansen, R., Frantzeskaki, N., McPhearson, T., Rall, E., Kabisch, N., Kaczorowska, A., Kain, J.H., Artmann, M., and Pauleit, S. The uptake of the ecosystem services concept in planning discourses of European and American cities. Ecosystem Services, 2015, vol. 12, pp.228–246.

4.      Corbane, C., Martino, P., Panagiotis, P., Aneta, F.J., Michele, M., Sergio, F., Marcello, S., Daniele, E., Gustavo, N., and Thomas, K. The grey-green divide: multi-temporal analysis of greenness across 10,000 urban centres derived from the Global Human Settlement Layer (GHSL). International Journal of Digital Earth, 2018, vol. 13, no. 1, pp. 101-118

5.      Russo, A., Esperon-Rodriguez, M., St-Denis, A., and Tjoelker, M.G. Native vs. Non-Native Plants: Public Preferences, Ecosystem Services, and Conservation Strategies for Climate-Resilient Urban Green Spaces. Land, 2025, vol. 14, no. 954, pp. 1-22

6.      Wilson, M.V., Hibbs, D.E, and Alverson, E.R. Native plants, native ecosystems and native landscapes: an ecological definition of "native" will promote effective conservation and restoration. Kalmiopsis, 1991, vol. 1, pp. 13-17

7.      Aronson, M.F.J., Lepczyk, C.A., Evans, K.L., Goddard, M.A., Lerman, S.B., and MacIvor, J.S. Biodiversity in the city: key challenges for urban green space management. Frontiers in Ecology and the Environment, 2017, vol. 15, no. 4, pp.189–196.

8.      Zarei, M. and Shahab, S. Nature-Based Solutions in Urban Green Infrastructure: A Systematic Review of Success Factors and Implementation Challenges. Land, 2025, vol. 14, no. 818, pp 1-26.

9.      Teutscher, D., Bukreev, F., Kummerländer, A., Simonis, S., Bächler, P., Rezaee, A., Hermansdorfer, M., and Krause, M.J. 2025. A digital urban twin enabling interactive pollution predictions and enhanced planning. Building and environment, 2025, vol. 281, no. 113093

Rolnick, D., Donti, P.L., Kaack, L.H., Kochanski, K., Lacoste, A., Sankaran, K., Ross, A.S., Milojevic‑Dupont, N., Jaques, N., Waldman‑Brown, A., Luccioni, A., Maharaj, T., Sherwin, E.D., Mukkavilli, S.K., Kording, K.P., Gomes, C.P., Ng, A.Y., Hassabis, D., Platt, J.C., Creutzig, F., Chayes, J.,  and Bengio, Y. 2022. Tackling Climate Change with Machine Learning. ACM Computing Surveys, Vol. 55, No. 2, Article 42