Mapping Metropolitan Structures with Digital Models as a Supporting Tool in Spatial and Strategical Planning—The Case Study of the GZM Metropolis

Abstract

This study presents the results of comprehensive functional-spatial analyses conducted using cellular models in relation to the cities of the GZM Metropolis and its surroundings. The Abbreviation “GZM” stands for Górnośląsko-Zagłębiowska Metropolia, due to its location, which in English has been recognized as the GZM Metropolis. The GZM Metropolis, the largest metropolitan area in Poland, has a complex administrative and spatial structure that includes 41 very diverse municipalities, which poses a significant challenge in interpreting data and understanding its complexity. The research was conducted by a multi-person and interdisciplinary team using various tools, including geographic information systems (GIS) and statistical data. The spatial models built on the basis of the collected data were visualized using augmented reality tools to facilitate data interpretation. Special attention was paid to environmental aspects, especially blue-green infrastructure, which plays a key role in maintaining this heavily urbanized area. Furthermore, the authors developed urbanization scenario models for the GZM Metropolis based on their own approaches to cellular modeling and examined the integration of artificial intelligence techniques to further refine these forecasts.

1. Introduction

1.1. Introduction—Mapping Urban Spaces, Mapping Metropolitan Areas

In research on city comparison, Baty states that the self-similarity observed across many spatial levels implies that the processes that drive agglomeration and clustering in small cities are similar to those in large cities, or indeed, in cities of any size [1].

The dynamic development of metropolitan areas requires the use of advanced analytical tools that enable a precise understanding of urbanization processes. In the face of challenges related to spatial planning and the pursuit of sustainable development, cellular models based on kilometer grids are being used more and more, as well as other spatial modeling techniques such as augmented reality (AR) and virtual reality (VR). These tools allow for precise mapping of metropolitan structures.

In the publication “Models of the Structure of the GZM Metropolis” [2], researchers used cellular models that enabled a detailed analysis of the structure of the GZM Metropolis. The area was divided into administrative units using a 1 km² grid, which enabled detailed research of municipalities and districts, showing the variation in urban density or biologically active area. These models enable precise mapping of urbanization processes and analysis of indicators related to development, infrastructure, and population density. With these tools, one can notice functional differences between the central and outer areas of the metropolis, which significantly affect its functioning and future development. This approach also allows for the simulation of scenarios for the development of urban and metropolitan areas, taking into account variables related to urbanization [2].

Bielecka states that exploration of the Web of Science database showed that research in GIS spatial analysis modeling for land use change began in the early 1990s and has continued since then, with substantial growth in the 21st century [3]. Many researchers work with cellular data models. Yeh, Xia, Li (2021) [4] emphasize that, due to their simplicity, cellular models develop quickly, but this simplicity limits their ability to represent realistic urban phenomena at local scales, which leads to model modifications and complexity. However, cellular models have significant advantages when applied to areas on a larger scale. Additionally, they are flexible and can have various applications, but they can also cause difficulties for users due to the lack of a specifically defined main goal [4].

Cellular models, derived from the concept of cellular automata (CA), allow for a greater degree of objectivity in research results and a stronger connection to the real physical three-dimensional space [5]. This advantage particularly applies to larger areas that include complex spatial structures consisting of many diverse administrative units [6].

Cellular models, thanks to their ability to standardize and integrate diverse data, are a key tool for spatial–metropolitan analyses that allow for in-depth interpretation of complex administrative and functional structures [2]. A key problem in the use of cellular models (CA) in spatial analysis is the right size of the modular grid. Cells that are too small or too large can have a key impact on the results of the analyses. Nevertheless, the research decided to use the same mesh size for all analyses. Thanks to this, it was possible to achieve a certain standard between the different topics of analysis [7].

The Ruhr Area was mapped in the article “Combined Small- and Large-Scale Geo-Spatial Analysis of the Ruhr Area for an Environmental Justice Assessment” [8]. According to the researchers, this tool takes into account large-scale environmental data and can also help identify smaller areas, which, when combined, can provide a complete overview of the entire area—in this case, the Ruhr Valley [9]. Mapping is also useful for conducting other analyses that can affect sustainable development and review of the entire agglomeration in areas of different sizes–urban, but also rural [8].

Geolocation data and RS data in mapping were described in the article “Mapping Urban Areas Using a Combination of Remote Sensing and Geolocation Data” [10]. According to the authors, geolocation data sets are often difficult to obtain and are available mainly on a small scale, which significantly limits their use in large-scale research. RS data sets are used for large-scale mapping of urban areas. Combining these data can improve the accuracy and expand the scope of research to the limits of urban areas [10].

Taubenböck H. et al. state that measuring, analyzing, and hence understanding the complex processes of urbanization need to proceed beyond isolated case studies. They conducted a study using remotely sensed data (Landsat data) to show the measurement of spatial urban expansion over time for 12 world megacities. The constantly increasing availability and accessibility of modern remote sensing technologies provides new opportunities for a wide range of urban applications [11].

The geo-informatic Land Use/Land Cover (LULC) system, whose data is used for mapping, among other things, is beneficial because it reduces the resources required by traditional methods [12]. LULC also provides useful information on land transformations [13]. The model can be used to estimate unobserved cases and predict future changes in land use [14], and the results obtained provide precise information on spatial optimization on a small scale, which becomes significant [15]. This data set was also used in the urban planning model of the Tabriz metropolitan area, where the detailed results on land use can be used in the framework of reduce, reuse, and recycle policy [16].

M. Solanki, Laddha, Kangda, Noroozinejad Farsangi (2023) [17] indicate that the use of modern technologies such as AR or VR is also becoming an increasingly important element of analysis and planning of urban spaces. As indicated in works on designing and visualizing space, these technologies enable the intuitive presentation of complex information and the effective engagement of stakeholders in decision-making processes. These modern technologies are crucial in the context of creating digital twins of cities [18], which allow for virtual modeling and forecasting of metropolitan area development scenarios [17].

Tomasz Bradecki (2021) [19] claims that mapping public spaces is more effective in a 2D variant and a physical model, and analysis through 3D models in the form of single views is useless. The use of augmented reality in mapping with 3D models is viewed positively, thanks to the ability to easily identify a selected part of the city; such analysis provides better results in understanding the complexity of a given area and modern cities. The downside of augmented reality is the need to use additional apps and devices. Mapping data in the form of 3D models is useful in urban analysis, as well as in teaching and for people without specialized knowledge [19]. Sarah Maria Schroeter (2021) [20] emphasizes that mapping space is used as a tool in the design process, and its goal is to understand individual parts of the city, increasing awareness of spatiality.

Furthermore, in the future, the integration of spatial models with artificial intelligence (AI) tools will be key [21]. As indicated in the studies on the GZM Metropolis structures, the use of AI-based tools, such as MidJourney, allows for experimental prediction of urbanization scenarios. However, attempts to use artificial intelligence to create images of the future have proven effective for specific locations, but not for the entire municipality or metropolis [2].

1.2. GZM Metropolis

The GZM Metropolis is a special area in the European space [22]. The metropolis is the first association of municipalities in Poland to be granted special legal and organizational frameworks by law [23]. This association consists of 41 independent, self-governing municipalities, each with a specific scope of competence. It specifically includes issues of spatial order and the related planning authority, financial independence, and ensuring social infrastructure and access to it for residents. The creation of the association did not diminish the competences of any of the municipalities. Within the association, no differentiation or hierarchy of individual municipalities was introduced. As a result, the association is made up of very diverse municipalities: from small rural municipalities to large cities [24], but sharing the same competences. At the same time, the authorities of the GZM Metropolitan Area received their own, not very broad, competences in the field of spatial planning and strategic planning, and much broader competences in the field of public transport organization and regional promotion [25].

The GZM Metropolis covers an area that stands out among other European metropolises in terms of a complex, polycentric settlement network and a complicated functional-spatial structure. These phenomena are the result of development based mainly on hard coal mining and heavy industry. Urbanization factors and history have had a key impact on the spatial polarization of the entire area. Industrial functions were carried out both within the dense areas of the previous urbanization and in areas with low development intensity with a leading agricultural function. Industrial areas have been accompanied by settlements and worker’s colonies since the beginning [26]. Currently, areas with residential functions, both in single-family and multi-family buildings, are interspersed with areas with industrial, post-industrial, and infrastructural functions. The entire area is a mosaic of very diverse functions and spatial structures [27]. All of this is intersected by a relatively dense network of road and rail routes. In the last decade, there has been a noticeable trend of spreading development into areas that were previously unoccupied [28].

The current shape of the Metropolis is the result of both intensive processes of urbanization and industrialization taking place at different rates from the beginning of the 19th century to the 1980s, as well as rapid processes of deindustrialization initiated by political and economic changes in Poland. A typical feature of the current landscape of the Metropolis is its fragmentation and lack of continuity [29]. Lang (2003) [30] describes such areas as edgeless cities, explaining that the mixture of various land uses is a typical form of developing metropolis.

At the moment, the GZM Metropolis is still facing many economic, social, and environmental problems. The effects of rapid deindustrialization and the shift away from the heavy industry economy towards high-tech and service industries, which started in the late 1980s, continue to be felt. These processes have led to a decrease in the attractiveness of the region, unemployment, declines in GDP, environmental pollution, and depopulation. The answer to these problems and challenges are numerous transformation and revitalization projects led mainly by local governments. After the establishment of the GZM association, these activities are much better coordinated and focused on specific, selected strategically important areas. Such a state makes it difficult to implement policies with sustainable development goals. In the current GZM Development Strategy for 2022–2027, certain development goals have been adopted. They focus on five areas: climate change adaptation, mobility and accessibility, space and social cohesion, metropolitanity, and innovation, and cooperation and openness [31].

1.3. Challenges and Specifics of Spatial Analysis of the GZM Metropolis

In connection with receiving certain competences in the scope of normative spatial planning and strategic planning by the union authorities, challenges have been posed in the scope of analyses and diagnoses of the initial state [32]. These competencies were less related to normative spatial planning, as self-governing municipalities retained all their previous competencies in the area of creating planning documents, including local plans that have the character of regulatory and zoning plans. The competences of the Metropolis GZM authorities have been focused on strategic planning. Its primary goal is to promote sustainable development.

Due to the implementation of planning processes, serious problems were identified at the analysis stage. The difficulty in their preparation was mainly due to the complexity of the entire area, and above all, the incomparability and incompatibility of the input data [33]. Recognizing the entire Metropolis as an autonomous planning entity necessitated the creation of new data resources with spatial references. These resources were to contain data that had not been standardized before [34]. Such needs have forced the use of new research methods and tools.

Analyzing and then planning the development of such a spatial structure proved to be a huge challenge [35], especially when the priority of the sustainable development principle was adopted [36]. Collecting data and then analyzing it within the GZM Metropolis required a different approach than what was known and used so far in municipalities, on a local scale. There were also significant differences in the regional level analysis. The scale of metropolitan analyses required consideration of both certain features of local analyses and certain features typical of regional analyses. On the one hand, it was necessary to identify the exact locations of the phenomena being studied, and on the other hand, it was intentional to attempt some generalization of the collected data [37,38]. The aggregation of analyses performed at local scales did not yield satisfactory results due to the inconsistency and incompatibility of analytical methods used in different municipalities and the complexity of the functional-spatial structure of the area.

At the beginning of the research, it was assumed that the cellular automata model method, due to its unique features allowing for standardization of results, can be considered the most suitable for spatial analyses of metropolitan areas, such as the GZM Metropolis [39].

2. Materials and Methods

In connection with the decision by the Silesian University of Technology, a research university, to cooperate with the authorities of the GZM Metropolis in the preparation of analyses, a selection of methods for analyzing cellular models was made. They were to cover the entire metropolitan area and a very wide range of research subjects, from spatial analysis to social, economic, and infrastructure. They were also supposed to use GIS and CAD tools. Certain assumptions regarding the standards of presenting and imaging the results of analyses were adopted at the outset of the work. The decision was made to use modern imaging tools, including augmented reality and extended reality. This decision was made to make the results of the analysis more accessible to non-professionals.

For the purposes of this article, the following research questions were posed: Can models be useful in mapping metropolitan areas as opposed to classic 2D mapping? How can models be created to better understand spatial data? Can AI support be helpful in mapping metropolitan areas, and if so, in what way?

Due to the analytical work being conducted, research on the effectiveness and accuracy of the analytical method was conducted in parallel [40]. The following working hypotheses were initially adopted at the beginning of these research works:

• Assessment of the applicability of cellular models for analyzing complex functional-spatial structures at supra-local and metropolitan scales.
• Assessment of the potential of augmented reality (AR) and extended/immersive reality (XR) tools to improve the accessibility, comprehensibility, and dissemination of spatial analysis results among non-professional audiences.
• Assessment of the applicability of the applied analytical and visualization tools in strategic and scenario-based spatial planning aimed at supporting sustainable metropolitan development.

The following methods were used in the research, illustrated in Figure 1: a literature review (based on Polish and foreign scientific studies, information was gathered on the spatial analysis of the metropolis), in situ research (observations and measurements were made in selected areas of the metropolis, which allowed for verification of the data obtained from GIS analysis), spatial analysis using GIS (geographical data was visualized in GIS systems and spatial modeling techniques were used, as well as the generation of functional and administrative maps), statistical data (use of publicly available databases, including data from the Central Statistical Office (GUS) and specialized reports prepared by GZM Metropolis), and 3D modeling and visualization technologies, which were used for spatial research in the GZM Metropolis and allowed for a deeper analysis of urbanization processes. 3D models were created using SketchUp software, which is commonly used in spatial and architectural design, and were then adapted for Sketchfab and Augment platforms, which enable interactive visualization of these models in augmented reality (AR) and online visualization.

This article briefly presents a broad spectrum of research, conducted interdisciplinarily between the Faculty of Architecture of the Silesian University of Technology and the University of Economics in Katowice, as well as based on scientific studies based on the PBL (Project Based Learning) formula. The research was expanded to include mapping trials using artificial intelligence with the MidJourney program V6.1, which generated graphics based on commands and photographs taken.

This article presents a methodological framework for metropolitan-scale spatial analysis based on cellular models, spatial simulations, and advanced visualization techniques. The study adopts an applied and exploratory approach and focuses on assessing the usefulness of these tools for analyzing complex functional-spatial structures, communicating analytical results to non-professional audiences, and supporting strategic and scenario-based spatial planning aimed at sustainable development. The proposed approach is demonstrated using the case of the GZM Metropolis as a complex, polycentric metropolitan area.

3. Results

The research was conducted in several areas using various analytical tools. These include models of density and intensity of development, models of accessibility, models of functional and spatial layout, models of the city image according to Kevin Lynch’s theory, and spatial simulation models. All of the above served to synthesize data obtained from various sources such as spatial databases, demographic and field analyses, and the GIS (Geographic Information System) portal.

3.1. Cellular Models

Data cellular models (also known as raster or grid) are a way of presenting spatial data in which the area of the space is divided into a regular grid of cells (pixels). Each cell has an assigned value that describes a certain characteristic of a given space segment [41]. The advantages of using a raster model include the simplicity of the data structure, the ease of numerical analysis, and the usefulness in calculating spatial simulations [42]. However, it should be remembered that using this model involves generalizing data, and as a result, it often leads to a loss of geometric accuracy of individual shapes. Therefore, it is not used (without a clear need) for visualizing linear objects or presenting data on a small scale. Using a raster model for spatial statistical analysis, or reducing data to uniform analytical units, provides a better picture of the spatial distribution of the information presented than, for example, using administrative divisions. However, caution should also be exercised in this case: when determining the size of a given raster cell, the nature of the presented feature should be taken into account, because cells that are too small may enable the identification of persons to whom the presented data refer (thus violating the provisions of the GDPR).

The article used data cellular models to present the spatial distribution of the following basic urban indicators in the GZM area:

• Building density—the ratio of the total built-up area to the land area (Figure 2).
• Building coverage ratio—the percentage of building coverage in relation to the total area of the cell (Figure 3).
• The share of buildings in urbanized areas—that is, the percentage of the area of buildings to the area of urbanized areas in a given cell (Figure 4).
• Blue-green infrastructure—green and water areas present in the urban structure (Figure 5).

Also used were cellular models concerning the communicative accessibility of the GZM Metropolis—the length of roads (Figure 6), the number of cars (Figure 7), the 15 min metropolis (Figure 8), and the functional-spatial structure (Figure 9)—presenting the diversification of functions throughout the metropolis.

Using models, it was possible to analyze and demonstrate a clear variation in the intensity of building density in the urban areas of the GZM Metropolis. Analyses showed clear differences between central and peripheral areas. The center of the metropolis is dominated by high-intensity structures—dense development with a large share of building area in the cell—while the outskirts have lower intensity with a larger share of green areas and more dispersed development.

The conducted research confirmed the effectiveness of the use of the cellular model method in spatial analyses of complex and extensive spatial structures. However, the results of these studies revealed some important conditions and limitations associated with the use of such methods. A prerequisite is to properly select the size of the model grid cell. The size of the grid can affect the quality of the analysis results obtained. By definition, the grid should be built on as large a field as possible, but at the same time not generalize the results too much. The size of the grid should therefore be adapted to the expected results and to their applications. Large-field grids can be useful for analysis in strategic planning in building general models of functional-spatial structures covering large areas. Small-field meshes may be more useful in local and normative planning.

The study also revealed some limitations of the use of cellular models. They concern, in particular, the analysis of small point and linear objects. This method can only be used for area objects.

The choice of a kilometer grid was deemed suitable for conveying information, including population data, from irregular source zones [43]. For the GZM metropolis, the area was divided into fields of 1 km², which often encompass several administrative areas, to ensure that the values contained within are easily understandable [2].

3.1.1. Building Density Model

The spatial distribution of the intensity of development within the metropolis is illustrated in Figure 2. High-intensity areas are highlighted—the darkest cells, and medium- and low-intensity zones. This shows that the “intensity cores” occur in the central part of the metropolis and that the intensity gradually flattens out towards the periphery.

3.1.2. Building Area Participation Model

The plan map (Figure 3a) shows the percentage of built-up area in each cell relative to the total area. The 3D model (Figure 3b), sketchfab: https://skfb.ly/oGPss (accessed on 12 November 2025) gives a visual effect of “height”—cells with a larger share of development are stretched upwards, which facilitates intuitive comparison of intensity.

The analysis presented in Figure 4 is an expansion of the research on the relationship between built-up area and urbanized areas within the GZM Metropolis. This indicator plays a key role in assessing the compactness of urban structures and the degree of building saturation in areas where the urbanization process is already advanced. It allows for the identification of areas with high intensity of space use and those that have retained a larger share of undeveloped land within the existing urban fabric.

3.1.3. Model of Building Density in Urbanized Areas

The plan map (Figure 4a) shows the percentage of built-up area in each cell relative to the total area of urbanized land in each raster cell. This allows for precise identification of areas with the highest degree of urbanization. The highest values, exceeding 70% of the built-up area in urbanized areas, are concentrated in the central parts of the metropolis (especially within Katowice and Gliwice). These areas are characterized by dense residential and commercial development and very intensive land use, confirming the hypothesis of strong “intensity cores.”

The 3D model (Figure 4b), sketchfab: https://skfb.ly/oGP6o (accessed on 12 November 2025) gives a visual effect of “height” to individual cells. Cells with a larger built-up area ratio are stretched upwards, creating a three-dimensional urban sculpture effect that reflects the scale and structure of the built-up area.

3.1.4. Blue-Green Infrastructure Model

Under the blue-green infrastructure model (Figure 5), it was possible to identify cells with high natural potential—where the share of greenery and water is relatively high. This model shows that despite its industrial image, the metropolis has a significant number of areas with significant ecological values that can become the foundation for revitalization efforts in the future.

Map 2D (Figure 5a) shows the spatial distribution of the percentage of green space and water in each raster cell. The darkest shades of green indicate areas with the highest share of blue-green infrastructure, which are mainly concentrated on the outskirts of the metropolis, along watercourses, and within large forest complexes and recultivated post-industrial areas.

The three-dimensional model (Figure 5b), sketchfab: https://skfb.ly/oGPnY (accessed on November 2025), similar to the previous models, represents cells with a high proportion of blue-green infrastructure through their vertical extensions, creating a three-dimensional map of the “height” of ecological potential. This makes it possible to quickly identify key natural enclaves and ecological chains in the structure of the metropolis.

The model of the blue-green infrastructure within the GZM Metropolis shows that the whole area, despite its dominant industrial character, has significant areas with high ecological value. The identified areas with a high proportion of green spaces and water are a potential foundation for revitalization efforts aimed at improving the quality of life for residents, adapting to climate change, and protecting biodiversity.

3.1.5. Road Length Model

According to the authors, the analysis of road length (Figure 6a) provides information on their intensity, based on the sum of road length per square kilometer of area. The model (Figure 6b), sketchfab: https://skfb.ly/oFF87 (accessed on 12 November 2025) shows that the highest density of roads in the GZM Metropolis occurs in Katowice, Tychy, and Dąbrowa Górnicza—these cities have major transportation hubs and an extensive road network, which is a result of their large area. According to researchers, car transport is key to the functioning and development of metropolises, even on an international scale—the main European roads run through the GZM Metropolis area. The GZM Metropolis also includes numerous national routes—highways and expressways. The most notable is the Spine Road (Drogowa Trasa Średnicowa), which connects many major cities in the metropolis [2].

3.1.6. The Number of Cars Model

According to the authors, the analysis of the number of cars in the GZM Metropolis (Figure 7a) provides information on the intensity of the number of cars based on the indicator of the number of cars per 1000 inhabitants in individual GZM Metropolis municipalities in 2020. The 3D model (Figure 7b), sketchfab: https://skfb.ly/oK9Cr (accessed on 12 November 2025) shows that the highest intensity of the number of cars occurs in the central part of the metropolis, and the lowest intensity was recorded in the municipalities located on the outskirts of the GZM Metropolis [2].

3.1.7. The 15 Min Metropolis Model

The authors of the publication “Models of the structure of the GZM metropolis” [2] developed a 15 min GZM model at the metropolis scale (Figure 8a,b) and at the district scale (Figure 8c,d). As they note, well-functioning cities must have good public transportation and basic services within a 15 min radius. To conduct the analysis, 2D maps were used, showing the average number of stops (Figure 8a) and the average number of shops (Figure 8b) per square kilometer in the district. They were used to identify locations with the potential for 15 min cities. The synthesis of the above analyses (Figure 8c) and the results presented in the model (Figure 8d), sketchfab: https://skfb.ly/oGOZn (accessed on 12 November 2025) allowed for the identification of four such points, which are the centers of Katowice, Gliwice, Dąbrowa Górnicza, and Będzin [2].

3.1.8. Functional-Spatial Structure Model

An analysis of functional-spatial structure models, illustrated in Figure 9, is a significant element of research concerning the space of the GZM Metropolis. The goal was to create a detailed map of the functional diversity of metropolitan areas, identify the dominant functions of individual spatial units, and assess the availability of public services [2].

The study was conducted using a regular hexagonal grid of area approximately 1 km², which ensured a uniform data structure and enabled precise comparison of the analyzed areas. The use of a hexagonal grid also allowed for the avoidance of distortions resulting from traditional rectangular grids, which increased the accuracy of spatial representation. To better illustrate the results, 2D (Figure 9a) and 3D models (Figure 9b), sketchfab: https://skfb.ly/oHPvL (accessed on 12 November 2025) were used, which enabled visualization of structural differences within the metropolis. Additionally, an index and typological analysis was conducted to provide a detailed understanding of the urban and functional dynamics of the region. Using these tools, it was possible to capture both local and system-wide relationships that influence the spatial structure of the GZM Metropolis [2].

An analysis of the functional-spatial structure of the GZM Metropolis revealed significant differences in the distribution of functions in individual areas. In the central parts of the metropolis, especially in Katowice and Gliwice, residential and service functions dominate. However, peripheral cities, such as Sosnowiec, Świętochłowice, and Zabrze, show a greater presence of industrial functions. Identifying these differences was a key element of the study, enabling the identification of areas requiring special urban planning actions, such as the revitalization of industrial areas or improving access to public services. Thanks to a detailed functional analysis, it was possible to capture both local and systemic relationships affecting the spatial structure of GZM Metropolis [2].

3.1.9. Advantages of Cellular Models

Cellular models (raster) enabled numerical and visual analysis of the spatial structure of the GZM Metropolis, revealing a clear structural dichotomy between compact cores of high-density development in the central parts and areas of lower density on the periphery. An analysis of the blue-green infrastructure model confirmed the existence of numerous areas with high natural values that may serve as the basis for revitalization and adaptation activities in the future. Presenting results in the form of interactive 3D models with augmented reality (Figure 10) is a valuable tool that supports the design and spatial planning process and can be easily used for demonstration.

The results indicate that the application of cellular models at a supra-local and metropolitan scale enables a coherent and comparable analysis of complex functional–spatial structures across the GZM Metropolis.

3.2. Spatial Simulations

Spatial simulations are used to recreate or predict values in both space and time. An example of spatial simulation of incomplete measurement data in space is interpolation. This is a mathematical model (e.g., NN—Nearest Neighbor, TIN, IDW—Inverse Distance Weighted, Kriging) [44] that allows the creation of continuous maps from point data. Interpolation is widely used in environmental analyses (e.g., meteorology, hydrology, geology), but in the case of urban planning and social analysis, it is also used, for example, to continuously represent the distribution of population density [45,46]. Another example is the analysis of the rank of individual public transport stops in GZM Metropolis, determined by the average number of lines serving the stop and the average number of services serving the stop (Figure 11).

On the other hand, an example of space-time simulation is the various types of predictions used to assess the probability of future spatial development. Predictions can be based on predefined and indicated simulation rules, creating mathematical models (e.g., Markov chains) or artificial neural networks (machine learning models), which take into account many factors, including nonlinear relationships between factors (MLP—Multi-Layer Percpetron) [4,5,48,49,50].

For the purposes of this article, an assessment of the prediction made by the Terrset program was prepared using land cover/use data from the Corine Land Cover (CLC) program [51]. The simulation was performed based on data from 2012 and 2018; analyzing the changes that occurred during that time, the transition probability matrices were calculated, and then the simulation was performed for 2024 (Markov chains) (Figure 12).

From the perspective of contemporary urbanization challenges, a simulation of the placement of new residential buildings can be a particularly useful tool. Therefore, to assess the degree of fit of the forecast, the building prediction was compared to actual data using the State Geodetic and Cartographic Resource materials (BDOT10k: PTZB) (Figure 13). The Kappa Index (which takes values from −1 to 1) was 0.53, which can be interpreted as a moderate or weak level of agreement between the forecast and reality. The presented model does not exhibit high predictive precision. It is insufficient for precise local-scale planning decisions, but it can be used to analyze trends in urbanization directions or to compare different predictive scenarios.

The model performance is influenced by several factors, including significant differences in resolution (BDOT10k data are much more precise than the highly generalized CLC data), simplifications inherent in the Markov model, and the sensitivity of the Kappa Index to class prevalence, as the comparison concerns maps with a highly imbalanced class distribution. In land-use analyses where classes related to new development are marginal, and the spatial pattern is largely stable, the Kappa Index may underestimate the model’s ability to identify areas undergoing actual change. Therefore, the simulation results should be interpreted with caution.

The model is not intended to predict the exact locations of individual investments, but rather to identify areas with an increased probability of urbanization pressure at the metropolitan scale, making it suitable as a support tool for strategic and scenario-based spatial planning. Focusing on areas indicated by the calculations as most likely to be converted into new construction, local spatial planning policy can be adapted in advance.

A closer look at the areas where the prediction turned out to be false may lead to an analysis of the factors that may have an impact on the simulation. Predictive models can be further improved by developing a coherent spatial database on land use and by applying more advanced modeling approaches that incorporate user-defined locational factors. A good example of a broad approach to the analysis and forecast of residential construction growth can be found in the collective work edited by P. Gibas in 2017 [53].

The conducted research has shown high usefulness of cellular model methods, especially for strategic and scenario-based planning, which includes the identification of strengths, weaknesses, threats, and opportunities. The results of such analyses are highly suitable for analyses that are the basis for building models of functional and spatial structures.

4. Discussion

4.1. Attempts to Create Models Using Artificial Intelligence

Artificial intelligence is a tool for creating digital images based on input scripts, which was used in research on the MidJourney tool in the process of creating architectural concepts by Anna Jaruga-Rozdolska [54]. The authors of the publication “Models of the structure of the GZM Metropolis” [2] used artificial intelligence techniques to present scenarios—positive and negative visions of the future of public spaces in the GZM Metropolis. Bradecki, Bal, Mól, and Sanigórska (2024) [55] conducted experiments with generating an image of a city using the MidJourney platform. Researchers note that the results presented by artificial intelligence are ineffective; the images produced are not clear enough and are inconsistent in terms of scale, and therefore the results of the experiments at this stage are not fully reliable [55].

Mapping the GZM Metropolis

The authors attempted to create a map of the GZM Metropolis using an artificial intelligence tool through the “Midjourney” platform. The illustrations depict the image of the entire GZM Metropolis in the form of 2D plans and a 3D model, and then their positive and negative visions. Attempts were also made to visualize the development scenarios of the three main areas of Katowice based on the photos taken. The effects were achieved using the keywords “GZM Metropolis, Poland, mapping urban spaces, 2D map/3D model” and “Realistic positive/negative future scenario in 100 years from now for the development of a public space in Katowice, using mapping urban spaces”, taking into account the type: 2D map or 3D model, as well as the positive or negative vision.

Artificial intelligence presented the GZM Metropolis on a 2D plan (Figure 14a) and a 3D model (Figure 14b), highlighting the extensive network of national, regional, and local roads with purple lighting. The GZM Metropolis was identified as a city, not an area, which could suggest that it is a metropolis. Due to the skyscrapers, the presented place may refer to Katowice (the central hub of GZM Metropolis) and its modern high-rises. The similarity scale can be set to very low. The scenario is characterized by a predominance of built-up areas in the central core of the metropolis, and as it moves away from it, the structure gradually transitions into green spaces and undeveloped areas. The traffic pattern is clear—a highway runs through the middle of the metropolis as the main transportation axis.

Artificial intelligence presented some scenarios of the development of the GZM Metropolis on a 2D plan (Figure 15a) and a 3D model (Figure 15b), in a positive scenario, by highlighting the green areas in the center, high-rise buildings, and the main communication route, which runs through the GZM Metropolis. The areas were marked with a specific color, depending on their type, e.g., green for green areas and orange for high-density development areas. In the positive scenario, built-up areas predominate, but they are irregularly distributed, interspersed with grasslands, resulting in a less compact spatial structure. The metropolis is surrounded by highways, and its interior is served only by lower-category roads.

Artificial intelligence presented the development of the GZM Metropolis on a 2D plan (Figure 16a) and a 3D model (Figure 16b) in a negative scenario, taking into account significant unused areas, lack of distinction on the model of communication routes, and chaotic dispersion of green areas, mainly outside the center. Artificial intelligence did not distinguish individual areas, creating illustrations in a synthetic and unreadable way. The negative scenario assumes chaotic and dispersed land use, without a dominant type of development. Built-up and green areas are mixed, which affects the readability of the entire structure. The traffic pattern is ambiguous—highways run through and around the metropolis, and the layout of internal roads differs little from that of the main routes.

The advantage of AI is integrating data sets, which enables AI simulation models to provide comprehensive solutions for urban planning and increase the ability to predict the development of cities [56]. Currently, the results presented by AI are limited to visual representation as a depiction of alternative spatial visions. It lacks the further ability to independently distinguish types of land use or to indicate specific results using numerical data. To increase the analytical value, it is possible to integrate AI results with semantic analysis or pattern recognition. The elements of the image could be identified as individual types of land use and then compared between scenarios.

The use of AR and XR, supported with AI methods, allowed the illustration of possible scenarios for the development of the analyzed area. In this respect, the use of cellular models turned out to be impossible. It was necessary to use other tools that could illustrate the effects of the development of the studied area in a better, faster, and easier way in the future. It was also intentional to use such tools that would not operate in a hermetic and professional language. Images generated by AI were intended to be addressed to a non-professional audience, so they had to operate with images, visualization, and animation.

The use of AI imaging tools, similarly to cellular models, should be placed in the category of strategic planning, in which it is important to build alternative development scenarios. Unlike cellular models, the use of AI imaging tools may have much broader applications in the future. It should be noted, however, that at the moment this tool can only support the work of a spatial analyst. However, this tool is not the leading tool in the tested applications.

5. Conclusions

Research has shown that using cellular models in performing spatial analysis on a supra-local scale yields good and sufficient results. They allow for maintaining certain consistent spatial data standards. These tools are useful for spatial planners, decision-makers, and non-professionals. The studies confirmed the possibility of publishing the results of the analysis in non-professional media. These data are usually understandable and well-received.

The applied methods prove useful not only for spatial planners and decision-makers but also non-professional audiences, as the results can be communicated in a clear and accessible form through simplified visualizations. The AR and XR tools employed in this study exhibit a high degree of usability and intuitive interaction, supporting users’ comprehension and interpretation of complex spatial configurations. As a result, these tools can improve the communicability and interpretability of spatial analysis results for non-professional audiences.

Spatial simulations are a crucial complement to static analyses, enabling the recreation or prediction of values in both space and time. Research has demonstrated the utility of interpolation methods (e.g., NN, TIN, IDW, Kriging), applied to create continuous maps from point data in urban and social analyses, such as the continuous representation of population density or the analysis of the rank of public transport stops in the GZM Metropolis, as well as space–time simulations. In the context of the GZM Metropolis’s strategic planning, predicting future spatial development is of key importance, achievable through mathematical models like Markov chains, utilizing land cover/use data such as Corine Land Cover (CLC). The executed simulation of new residential building placement, despite relying on generalized CLC data, resulted in a Kappa Index value of 0.53 when compared against the more precise BDOT10k data, indicating a moderate-to-weak level of agreement. This result confirms the exploratory character of the simulation and highlights its limitations, particularly those related to data generalization, model simplifications, and the sensitivity of the Kappa Index to highly imbalanced class distributions. Consequently, the simulation should not be interpreted as a precise predictive tool, but rather as a means of identifying broader spatial tendencies and areas of increased urbanization pressure at the metropolitan scale.

Such analyses support early-stage, strategic, and scenario-based planning processes by enabling local and supra-local spatial policies to be adapted in advance, rather than serving as deterministic forecasts of individual investments. Overall, the study demonstrates that the applied analytical and visualization tools are suitable for strategic and scenario-based spatial planning and can effectively support decision-making processes aimed at sustainable metropolitan development. In the case of the GZM Metropolis, this approach may support the formulation of metropolitan-scale spatial strategies aimed at balancing urban development pressure with environmental protection and social accessibility. In particular, the identification of areas of increased urbanization pressure can inform decisions related to the prioritization of compact development zones and the coordination of supra-local spatial policies.

However, significantly better results can be achieved by enriching the presentation of these analyses with other means adapted to specific topics, scopes, and issues. Additional tools that help convey information to broader groups of non-professional audiences are extended reality and augmented reality. The integration of spatial simulation results with blue-green infrastructure planning and the 15 min city model allows for the identification of locations where compact urban development can be combined with ecological continuity and improved access to everyday services. Such a combined approach may support spatial strategies focused on limiting urban sprawl, strengthening public transport-oriented development, and enhancing environmental resilience within the GZM Metropolitan Area.

The scenario-based approach adopted in this study allows the results to be directly interpreted in relation to specific planning themes, such as blue-green infrastructure, public transport accessibility, and the 15 min city concept. Future research should focus on improving predictive accuracy through the development of more coherent and higher-resolution spatial databases, the incorporation of locally defined locational and socio-economic factors, and the application of more advanced, hybrid modeling approaches. At the same time, the limitations resulting from data generalization, model simplifications, and the exploratory nature of the applied simulations should be acknowledged when interpreting the results.

Involving a wider range of participants in planning processes is particularly important for both normative and strategic spatial planning. These two forms of planning in the Polish spatial planning system are particularly important at the supra-local and regional levels. The research presented in this article provides an example of how such tools and approaches can be applied to support normative and strategic spatial planning at the supra-local and regional levels within the Polish spatial planning system.