Path-Dependent Urban Expansion in Arid Cities: A Multi-Decadal Remote Sensing Analysis and ANN–CA–Markov Modelling of Saudi Arabian Cities (1984–2034)

Path-Dependent Urban Expansion in Arid Cities: A Multi-Decadal Remote Sensing Analysis and ANN–CA–Markov Modelling of Saudi Arabian Cities (1984–2034) | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Path-Dependent Urban Expansion in Arid Cities: A Multi-Decadal Remote Sensing Analysis and ANN–CA–Markov Modelling of Saudi Arabian Cities (1984–2034) Alireza Babapoorkamani, Liana Ricci, Tazyeen Alam This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8447461/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Urbanisation in arid environments evolves through distinctive spatial processes shaped by abundant developable land, strong environmental constraints, and infrastructure-led planning. Despite the rapid growth of Saudi Arabia’s major cities, the long-term interplay between land-cover change, demographic dynamics, and future expansion pathways remains insufficiently understood. This study reviews a consistent four-decade record of urban growth (1984–2024) for Riyadh, Jeddah, Makkah, and Madinah and examines it through an integrated framework combining multi-sensor remote sensing, demographic indicators, landscape fragmentation metrics, and ANN–CA–Markov modelling. Across all four cities, urban expansion follows a shared three-phase trajectory: an initial phase of fragmented and discontinuous growth (1984–1994), a prolonged period of corridor-driven consolidation aligned with major infrastructure investments (1994–2014), and a recent shift toward outward suburban diffusion (2014–2024). While this temporal sequence is highly synchronised, its spatial expression differs markedly. Fragmentation metrics (Patch Density and Edge Density) identify the mid-2010s as a peak of morphological discontinuity, most pronounced in the basin-confined cities of Makkah and Madinah. Jeddah, constrained by its coastline, retains a predominantly linear growth form, whereas Riyadh expands multi-directionally across an unconstrained plateau. Coupling demographic change with land consumption shows that population growth alone cannot explain observed expansion patterns: Riyadh maintains relatively stable land-use efficiency, while Jeddah and Makkah experience phases of disproportionately land-intensive development. The ANN–CA–Markov simulations reproduce observed spatial patterns with high agreement (Kappa 0.61–0.85) and project continued path-dependent expansion to 2034, with future growth largely reinforcing established corridors rather than generating new development fronts. By explicitly linking multi-decadal reconstruction, demographic efficiency, spatial fragmentation, and predictive modelling, this study advances a path-dependent interpretation of arid-city urbanisation and provides a transferable framework for understanding and anticipating urban growth in rapidly transforming desert environments. Urban Studies Urban expansion Arid cities Remote sensing ANN–CA–Markov Land consumption efficiency Saudi Arabia Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Introduction Urban expansion in fast-growing cities is a spatially uneven and temporally dynamic process shaped by demographic growth, infrastructure development, uneven implementation of plans and regulatory frameworks, and environmental constraints. According to the United Nations World Urbanization Prospects, Saudi Arabia has undergone one of the most rapid urban transitions in the Middle East, with population growth concentrated in a small number of major metropolitan areas (United Nations, 2023). In arid-region cities, urbanisation processes are particularly visible and consequential: limited vegetation cover, reflective soils, and the absence of buffered landscapes render land conversion both spatially legible and environmentally impactful (Fastner et al., 2023 ). Urban growth in desert environments therefore follows morphological logics distinct from those of temperate regions characterised by outward urban expansion, low-density development, and strong dependence on engineered infrastructure, water systems, and episodic reconstruction. Analysing these trajectories requires frameworks capable of integrating long-term land-cover change with demographic dynamics and spatial modelling (Zhang et al., 2025 ). Remote sensing–based studies have documented the expansion of major Saudi cities, including Riyadh, Jeddah, Makkah, and Madinah, consistently revealing outward growth, conversion of bare land to built-up areas, and strong links to population increase and national development strategies (Rahman, 2016 ; Alkhaldi et al., 2025 ). These four cities are therefore the focus of this study, as they combine demographic significance with sufficiently consistent long-term LULC records. Although Dammam was considered, its functional integration within a broader coastal–industrial urban system—extending beyond administrative boundaries to include Dhahran, Khobar, Qatif, Ras Tanura, and Al Jubail—would require a different spatial delineation and introduce additional challenges for multi-decadal satellite-based reconstruction and modelling (Al Atni, 2015 ; Alhajri, 2025 ). To ensure analytical consistency and comparability, Dammam is not included. Despite extensive work on Saudi urban expansion, three gaps remain. First, most studies rely on short time spans or widely spaced temporal snapshots, limiting insight into the intensity and sequencing of urbanisation across multiple decades. Second, demographic change—although a central driver of land consumption—is rarely integrated analytically with spatial expansion metrics. Third, predictive modelling approaches using Cellular Automata (CA), Markov chains, or hybrid ANN–CA frameworks often depend on inconsistent LULC inputs or omit density, fragmentation, and temporal path-dependence as modelling drivers (Resch et al., 2025 ). At the same time, shifts in national planning frameworks—such as rezoning initiatives, urban development corridors, and evolving Urban Growth Boundaries—have influenced where and how expansion occurred, shaping observed spatial trajectories even when not directly incorporated as modelling variables (Wang et al., 2024 ). Recent urban analytics research increasingly emphasises the need to coupling demographic processes with spatial form. Indicators such as Decadal Growth Rate (DGR), Land Consumption Rate (LCR), and Land Consumption Efficiency (LCE) quantify whether cities expand proportionally to population growth or consume land inefficiently (Cai et al., 2020 ). Complementary metrics, including Built-up Density (Db) and Expansion Percentage Change (EPC), reveal how growth is absorbed spatially. However, these indicators are seldom integrated with multi-decadal LULC reconstructions or embedded within spatially explicit modelling frameworks. In arid cities—where land conversion has pronounced environmental implications and population growth has been rapid over four decades—such integration is critical for understanding both drivers and consequences of expansion (Angel, 2023 ). This study addresses these gaps through an integrated framework analysing urban expansion in Riyadh, Jeddah, Makkah, and Madinah from 1984 to 2024. These cities represent heterogeneous arid urban systems shaped by different economic functions and physical constraints: Riyadh as a rapidly expanding capital on an arid plateau (Jarrar & Al-Homoud, 2024 ), Jeddah as a coastal metropolis influenced by port activity and post-disaster reconstruction (Tammar, 2017 ), Makkah as a pilgrimage-driven city undergoing large-scale redevelopment (Al-Saud & Goussous, 2023 ), and Madinah as a historically layered city with steady but spatially constrained growth (Rashid & Bindajam, 2015 ). Together, they provide a comparative setting for examining how demographic pressure, infrastructure-led development, and environmental constraints interact in arid contexts. Local planning interventions—particularly zoning updates and transport-led development— have influenced urban expansion and further contextualise observed patterns, alongside the role of unplanned settlements in cities such as Jeddah and Makkah (Aljoufie et al., 2013; Almazroui et al., 2017; El-Shorbagy, 2020 ; Alfakhrani et al., 2025 ). Methodologically, the study reconstructs a consistent 40-year LULC record using multisensory satellite imagery, integrates demographic indicators from census and gridded population datasets, and employs spatial metrics capturing expansion intensity, density, and fragmentation. These empirical patterns are operationalised within a hybrid ANN–CA–Markov framework to simulate near-term urban expansion, incorporating physical constraints and demographic pressures, and validated using multiple agreement metrics (Hu et al., 2020 ; Xu et al., 2022 ; Hou et al., 2023 ; Tharik et al., 2025 ; Alam & Banerjee, 2023 ). Overall, this study contributes by (i) producing a consistent four-decade LULC reconstruction for Saudi Arabia’s major cities, (ii) explicitly coupling demographic efficiency indicators with spatial expansion and fragmentation metrics, and (iii) embedding these relationships within a validated, spatially explicit modelling framework. By linking long-term reconstruction, demographic–spatial coupling, and predictive simulation, the study advances urban analytics approaches suited to rapidly transforming arid environments. 2. Study Area and Data 2.1 Study Areas Saudi Arabia is a predominantly arid country, with desert ecosystems covering more than 70% of its territory (Sayed & Masrahi, 2023). Since the early 1980s, oil-driven development has triggered rapid land use and land cover (LULC) transformation, placing the Kingdom among the fastest urbanising regions in the Middle East (Alqurashi & Kumar, 2014; Alqahtany, 2025). Multitemporal satellite imagery and remote sensing techniques therefore provide a critical means of monitoring long-term urban expansion, with post-classification comparison (PCC) widely used for detecting LULC change (Liu et al., 2023). The study focuses on four major Saudi cities—Riyadh, Jeddah, Makkah, and Madinah—selected for their demographic significance and contrasting geographical and functional contexts. Riyadh, the capital, is situated on the Najd Plateau; Jeddah is a major coastal port and commercial hub on the Red Sea; Makkah is a mountainous pilgrimage city in the Hijaz region; and Madinah occupies an elongated sedimentary basin constrained by volcanic fields and surrounding highlands. Together, these cities represent Saudi Arabia’s dominant administrative, economic, and pilgrimage-driven urban systems and account for most of the national urban growth. Their contrasting settings provide a robust comparative framework for analysing how environmental constraints and socio-economic functions shape long-term urban expansion trajectories. Figure 1 presents built-up area maps for the four cities between 1984 and 2024, illustrating distinct morphological characteristics, expansion directions, and density gradients. These patterns form the empirical basis for subsequent analysis of land consumption, growth efficiency, and urban form dynamics. To ensure consistency across temporal and spatial analyses, standardised metropolitan extents were defined for each city, encompassing both historical urban cores and the full perimeter of outward expansion over the 40-year period (Table 2 ). These extents were delineated beyond administrative boundaries to capture peri-urban transitions and regionally connected development fronts. All demographic indicators and urban expansion metrics (Db, LCR, LCE, PD, and ED) were computed within these standardised extents, ensuring full comparability across cities and years. 2.2 Data Sources The analysis integrates multi-temporal satellite imagery, census-based population data, and ancillary geospatial datasets to examine long-term urban expansion across the four study areas. Built-up dynamics were derived from Landsat imagery for 1984, 1994, 2004, 2014, and 2024, obtained from the USGS EarthExplorer platform. Landsat 5 TM, Landsat 7 ETM+, and Landsat 8/9 OLI scenes were selected based on availability, ensuring consistent spatial resolution (30 m) and dry-season acquisition. All imagery was radiometrically and atmospherically corrected, mosaicked where necessary, and clipped to standardised metropolitan extents. A summary of satellite datasets is provided in Table A2 , while full scene metadata are reported in Supplementary Tables S1–S4. Population statistics were obtained from the General Authority for Statistics (GASTAT) for census years 1992, 2004, 2010, and 2022 and compiled at the city level. Decadal population growth rates were calculated using exponential interpolation, and 2024 population estimates were derived from city-specific 2010–2022 trends. All demographic indicators were computed within the same standardised metropolitan extents to ensure consistency with spatial analyses. Built-up maps derived from the LULC classification were used to calculate Db, LCR, and LCE following standard formulations (Supplementary Material). Additional geospatial datasets were used to characterise terrain and accessibility. A 30 m SRTM DEM was processed to derive slope, capturing topographic constraints on urban expansion, particularly in Makkah and Madinah. Road networks extracted from OpenStreetMap were converted into distance-to-road layers to represent accessibility. Global Human Settlement Layer (GHSL) population grids were used to support interpretation of settlement patterns and cross-validate spatial density gradients. Together, these datasets provide a coherent foundation for the classification, spatial analysis, and modelling framework applied in this study. 3. Methodology This study implements a three-phase, integrated framework combining multi-decadal landcover reconstruction, demographic and spatial coupling analysis, and spatially explicit urban growth modelling. The workflow (Fig. 2 ) ensures end-to-end consistency through standardised metropolitan areas, unified classification protocols, and cross-validated datasets. 3.1. Image Pre-processing We obtained multi-temporal satellite imagery from Landsat TM, ETM+, and OLI (30 m), ASTER (15 m), and Sentinel-2 MSI (10 m) for the benchmark years 1984, 1994, 2004, 2014, and 2024. Standard radiometric and atmospheric correction procedures were implemented to ensure spectral consistency across sensors and seasons. Landsat images were corrected using LEDAPS, while Sentinel-2 scenes were processed with Sen2Cor to retrieve surface reflectance. All images were geometrically corrected to the WGS84/UTM coordinate system, and scenes were mosaicked where necessary to verify that the entire metropolitan area was covered. Cloud masking and quality filtering were applied using FMasks algorithms and QA bands to remove clouds, shadows, and other artefacts. Finally, all datasets were clipped to the study boundary, producing spatially harmonised, analysis-ready mosaics suitable for classification and temporal comparison. 3.2. Land Cover Classification A consistent land cover classification scheme comprising four categories—soil/bare land, built-up areas, vegetation, and water bodies—was developed to reflect the dominant landscape structure and urbanisation processes characteristic of arid environments. The selected classes align with the study’s focus on anthropogenic expansion and land transformation. Classification was done in ENVI using a supervised manual workflow. For each time point, multispectral imagery was visually examined to identify representative training samples for all classes. Soil and bare land were delineated based on their high reflectance in red and shortwave infrared bands; vegetation was identified by its strong near-infrared reflectance; built-up areas were recognised through their heterogeneous tonal and textural signatures; and water bodies were distinguished by their low reflectance in visible and near-infrared wavelengths. ENVI’s classification workflow was used to assign pixels to classes, followed by iterative manual refinement to correct misclassifications typically encountered in arid landscapes where spectral confusion between built-up and bare soil is common. Ancillary datasets—including SRTM-derived elevation and slope, OpenStreetMap road networks, and high-resolution basemaps—were applied selectively to guide interpretation and ensure classification consistency across time, while avoiding computational biases introduced by automated multi-layer classification models. ‎ 3.3 Accuracy Assessment‎ and Change Detection Analysis The classified land cover maps were validated using a stratified random sample of reference points derived from high-resolution imagery and temporally stable basemaps. Confusion matrices were generated for each classified year, and standard accuracy metrics—including overall accuracy, producer’s accuracy, user’s accuracy, and the Kappa coefficient—were calculated to assess thematic reliability across the multi-decadal period. The classification achieved stable accuracy levels suitable for subsequent spatial and temporal change analysis. A summary of overall accuracy and Kappa values for all cities and time points is reported in Table S5. Land cover change was quantified using PCC, which compares classified maps across successive time steps. Transition matrices were generated for each temporal interval to identify the magnitude and direction of class conversions, with particular attention to transitions from bare land to built-up land as the dominant urbanisation pathway. The spatial distribution of changes was visualised using thematic maps to identify growth fronts, infill processes, and peri-urban expansion patterns. Vegetation dynamics were further contextualised using NDVI-based inspection (Appendix B, Equation B1). 3.4 Urban Sprawl Analysis Urban sprawl was assessed using a combination of quantitative growth indicators and spatial pattern metrics. Built-up area was calculated for each time point to derive the Urban Expansion Rate (UER), enabling comparison of growth intensity across decades. Built-up Density (Db) was computed to quantify the proportion of urbanised land relative to the total metropolitan extent (Equation B2). To characterise spatial configuration and fragmentation of the urban fabric, landscape metrics, including Patch Density (PD) and Edge Density (ED), were calculated for each city and temporal stage. PD measures the number of built-up patches per unit area and serves as an indicator of dispersed development and leapfrogging growth, while lower values reflect more consolidated urban form (Equation B3). Because all analyses were constrained to standardised metropolitan extents (Table A1 ), PD values are directly comparable across cities and time periods. ED quantifies the total length of built-up boundaries per unit landscape area and captures the geometric complexity of urban form (Equation B4). Higher ED values indicate fragmented, irregular expansion, whereas lower values reflect more compact and spatially cohesive development. Considered together, PD and ED provide a complementary diagnostic of urban fragmentation, while UER captures the rate of short-term built-up expansion between successive periods. To quantify both the magnitude and efficiency of urban expansion, a set of complementary indicators was employed. Expansion Percentage Change (EPC) measures the decadal increase in built-up land (Eq. 2 ), while the Decadal Growth Rate (DGR) standardises this change into an annualised metric, enabling comparison across periods. The Land Consumption Rate (LCR) captures the pace of land conversion following SDG 11.3.1 conventions. Land Consumption Efficiency (LCE) was calculated as the ratio of LCR to population growth rate derived from census data, providing an indicator of whether spatial expansion outpaced demographic change. LCE values greater than 1 indicate land-extensive growth, whereas values below 1 reflect compact or population-aligned development. Together, these indicators provide a coherent framework for evaluating both the speed and efficiency of urban expansion across the four cities. 3.5 Urban Growth Modelling and Prediction Urban expansion was simulated using an integrated ANN–CA–Markov framework combining stochastic transition estimation, data-driven suitability modelling, and spatial allocation to reproduce observed urban growth patterns. Markov chains were used to estimate land-cover transition probabilities from observed changes during the 2004–2014 and 2014–2024 periods, providing the basis for projecting future conversions under existing development trajectories. Spatial transition potentials were generated using an Artificial Neural Network (ANN) trained on observed land-cover transitions (Equations B5–B9), incorporating key drivers of arid-region urbanisation, including slope, elevation, distance to roads, and population density, to capture non-linear growth relationships. Predicted transitions were allocated spatially using a Cellular Automata (CA) model, which enforces neighbourhood effects and spatial contiguity, reproducing the corridor-driven and path-dependent expansion patterns identified in the empirical analysis. Model performance was evaluated through back-casting validation, in which the model simulated the 2024 built-up configuration from earlier conditions. Comparison with observed land-cover maps shows good agreement across all four cities, as reflected in overall, histogram, and location Kappa statistics (Table D1 ). Following validation, the calibrated model was applied to generate 2034 urban expansion projections under a business-as-usual assumption. Full mathematical formulations, network architecture, and training parameters are provided in the Supplementary Material. 4. Results 4.1 Multi-decadal land cover transitions (1984–2024) The multi-temporal LULC series, derived using a semi-automatic classification approach (Congedo, 2021 ), reveals a consistent pattern of rapid urbanisation across all four cities between 1984 and 2024, characterised by systematic conversion of bare soil and vegetation into built-up land. Exact land-cover area values for built-up, soil, and vegetation classes across all cities and time periods are reported in Table S6 in Supplementary Section S.3, while Figure S1 provides a visual comparison of long-term trends. Despite contrasting geographical contexts—coastal (Jeddah), mountainous (Makkah and Madinah), and plateau-based (Riyadh) (Almalki et al., 2022 )—all cities exhibit strong outward expansion accompanied by a long-term decline in soil cover and only modest increases in vegetation. Across the study period, built-up land increased substantially in all cities, indicating a shared regional trajectory of land consumption driven by sustained demographic and economic growth (Table 1 ). Patch Density (PD) highlights distinct fragmentation trajectories shaped by local morphological constraints. Riyadh exhibits a gradual increase in PD, peaking in 2014, consistent with a transition from early discontinuous suburban growth toward progressive consolidation of peripheral development. Jeddah maintains comparatively low PD values throughout the period, reflecting its corridor-oriented, coastline-constrained expansion, where growth primarily extends existing urban fronts rather than forming isolated patches. In contrast, Makkah records the highest PD values, particularly in 1994 and 2014, indicating pronounced fragmentation associated with basin-confined topography and episodic pilgrimage-driven development. Madinah displays a more oscillatory pattern, with fragmentation peaking in 2014 during rapid peripheral expansion before declining by 2024 as dispersed patches merged into more contiguous urban fabric. Edge Density (ED) further elucidates the evolving spatial complexity of urban form (Table 1 ). Jeddah exhibits moderate but steadily increasing ED values, consistent with linear expansion along established coastal and transport corridors. Makkah shows a rise in ED until 2014 followed by a decline in 2024, signalling partial morphological consolidation within a constrained basin environment. Riyadh records a pronounced increase in ED over the study period, reflecting intensifying edge complexity associated with outward suburbanisation and fragmented peri-urban growth. Madinah registers the highest ED values overall, with a peak in 2014 corresponding to rapid expansion along basin corridors. Together, PD and ED provide a complementary diagnostic of urban fragmentation dynamics, distinguishing phases of dispersed, edge-intensive growth from subsequent periods of consolidation despite continued increases in total built-up area. Table 1 Built-up Density (Db), Patch Density (PD), and ED trend values for all cities (1984–2024) City Db PD ED 1984 1994 2004 2014 2024 1984 1994 2004 2014 2024 1984 1994 2004 2014 2024 Jeddah 0.0249 0.0278 0.0374 0.0499 0.0933 0.581 0.444 0.699 1.330 1.366 12.71 15.56 22.43 32.46 58.85 Madinah 0.0555 0.0788 0.0882 0.1655 0.2254 0.339 0.259 0.408 0.827 0.850 0.69 0.58 0.97 1.69 2.16 Makkah 0.0312 0.0907 0.1403 0.2023 0.3052 2.078 7.262 2.449 9.907 5.238 0.16 0.53 0.77 0.86 0.68 Riyadh 0.0377 0.0610 0.1038 0.1442 0.1930 1.629 0.893 2.280 9.503 1.394 22.37 24.32 31.61 63.87 48.23 4.2 Expansion intensity (EPC and DGR) Decadal expansion metrics indicate that, despite variation in absolute growth across the four cities, all exhibit a broadly consistent temporal trajectory in three phases (Tang et al., 2022 ; Zhang et al., 2025 ): a phase of accelerated urban expansion prior to 2004, a period of moderated growth between 2004 and 2014, and a renewed intensification during 2014–2024. This three-stage pattern is clearly reflected in the EPC and DGR values summarised in Table S7. The spatial expression of these expansion intensities is further illustrated in Figure S2 in the Supplementary Materials. Rather than isolating each city, these results allow cross-city comparison, showing that Makkah had the most extreme early expansion (1984–1994) and Jeddah dominated the final decade (2014–2024) while Riyadh and Madinah followed smoother multi-phase transitions. To complement EPC, DGR and LCR were calculated for each city (Table 3), revealing steady long-term land consumption in Riyadh and Jeddah, contrasted with extreme early-period conversion in Makkah followed by moderation. LCE values (Table 2 ) indicate strong inter-city contrasts: Jeddah shows disproportionately high land consumption relative to population growth during 2010–2022, whereas Riyadh maintains consistently balanced efficiency across all census intervals. Makkah exhibits marked early-period inefficiency linked to rapid redevelopment and topographic constraints, while Madinah shows fluctuating efficiency due to its basin morphology and limited buildable land. Table 2 Land Consumption Rate (LCR) per decade and Land Consumption Efficiency (LCE) for census intervals City LCR LCE 1984–1994 1994–2004 2004–2014 2014–2024 1992–2004 2004–2010 2010–2022 Riyadh 0.620 0.699 0.390 0.338 1.63 1.44 1.24 Jeddah 0.115 0.348 0.333 0.868 0.38 0.87 12.76 Makkah 1.906 0.546 0.442 0.509 6.88 1.66 1.15 Madinah 0.420 0.119 0.877 0.362 1.02 0.34 1.56 4.3 Spatial clustering and Spatiotemporal Growth Patterns KDE and the weighted temporal composite reveal a clear three-stage trajectory of spatial expansion across all four cities. The early period (1984–1994) was marked by fragmented, discontinuous growth; the mid-period (1994–2014) showed consolidation along major corridors and urban centres; and the recent decade (2014–2024) experienced outward diffusion into emerging suburban belts, particularly in low-slope, accessible areas (Liu et al., 2016 ). This transition from leapfrogging sprawl to more corridor-based development aligns with multi-stage urban growth patterns reported in the broader urbanisation literature (Li et al., 2013 ). Following this three-stage trajectory, Fig. 3 further illustrates the spatial configuration of these patterns by mapping the KDE-based clustering of newly developed built-up areas across the four cities. In addition to the KDE surfaces, the final composite panel visualises all decades of expansion through a weighted temporal composite, in which the RGB channels encode the timing and persistence of urban growth. Early development is represented in red, mid-period expansion in green, and later growth in blue, while mixed hues (such as yellow, cyan, and magenta) indicate areas that experienced more than one wave of development. White tones mark locations characterised by sustained, multi-decadal intensification. This composite therefore captures not only when expansion occurred but also how consistently growth has accumulated over time. Together, these spatial layers reveal persistent long-term hotspots—along Jeddah’s inland eastern corridor, within the southern and eastern basins of Makkah, across the western and northern valleys of Madinah, and along Riyadh’s northern expansion axes—highlighting the decisive role of infrastructure, land availability, and topographic constraints in shaping urban growth trajectories. 4.4 Population density trends Population density patterns derived from GHSL data (1985–2025) show a clear intensification across all four cities, closely mirroring the spatial distribution of built-up expansion (Figure E1 ). Details on GHSL datasets, preprocessing steps, and density visualisation methods are provided in Supplementary Section S.2. During the early decades (1985–2005), density increases were largely concentrated within historical cores and first urban rings—particularly in Riyadh and Jeddah—reflecting compact development and rapid demographic growth (UN-Habitat, 2020a; 2020b). From 2015 onwards, all cities exhibit a pronounced shift toward peripheral demographic diffusion, consistent with the outward expansion corridors identified through KDE and the weighted temporal composite analysis. Despite contrasting geographical settings, the four cities display highly synchronised demographic–spatial dynamics over the 40-year period. Topography strongly regulates density redistribution in Makkah and Madinah, where growth concentrates within buildable valleys and sedimentary basins while mountainous barriers channel expansion along transport corridors (UN-Habitat, 2020c; 2020d). In Jeddah, coastal constraints drive inland densification structured by major highway axes (UN-Habitat, 2020b), whereas Riyadh exhibits multi-directional demographic expansion enabled by extensive land availability and a metropolitan-scale road network (UN-Habitat, 2020a). GHSL-derived density trajectories confirm a strong coupling between population growth and land consumption across all cities, particularly during the 2004–2014 decade, when rapid built-up expansion coincided with marked demographic intensification. The emergence of new suburban belts during 2014–2024 underscores a shift toward outward, accessibility-driven urban growth shaped jointly by environmental constraints, infrastructure networks, and land availability. 4.5 Urban-growth modelling and prediction The weighted temporal composite (Built-up Temporal Intensity, BTI) indicates that the 2014–2024 decade exhibits the highest cumulative intensity of built-up expansion across all four cities (Alqurashi et al., 2016 ). Persistent multi-decadal hotspots are concentrated along Jeddah’s eastern corridor, Riyadh’s northern axes, and the basin-aligned valleys of Makkah and Madinah. These BTI-derived hotspots closely align with the high-probability transition zones used in the ANN–CA–Markov simulations, confirming that future growth trajectories are strongly shaped by long-term spatial inertia. The ANN–CA–Markov framework was applied to generate spatial predictions of built-up expansion for 2034. Transition probability matrices derived from the 2004–2014 and 2014–2024 periods show that bare soil to built-up conversion dominates urban growth, accounting for approximately 85–92% of all transitions across the four cities, consistent with findings from arid-region case studies (Almadini & Hassaballa, 2019 ). Suitability surfaces indicate that low-slope, highly accessible areas near major transport corridors exhibit the strongest transition potentials, reinforcing the role of terrain and infrastructure as key drivers of urban expansion (Vahid & Aly, 2025 ). Model validation through back-casting reveals clear performance differences linked to urban morphology and geomorphology (Table D1 ). Jeddah achieves the highest agreement (Kappa = 0.85), reflecting its relatively predictable corridor-based growth pattern, while Riyadh and Madinah show moderate-to-high accuracy (Kappa = 0.68 and 0.72). Makkah records the lowest agreement (Kappa = 0.61), consistent with modelling challenges in mountainous, basin-confined urban systems where spatial transitions are less predictable. The 2034 prediction maps (Fig. 4 ) reveal distinct but coherent expansion trajectories across the four cities. Jeddah expands predominantly inland toward the east–southeast, Riyadh exhibits strong northward and north-eastern growth along major arterials, Makkah’s expansion remains constrained within southern and eastern valleys, and Madinah shows peripheral diffusion along western and northern basin corridors. These simulated patterns reinforce the empirical KDE and BTI analyses and align with recent remote-sensing–based urban growth studies. 5. Discussion By integrating multi-decadal LULC reconstruction with demographic indicators, fragmentation metrics, and spatially explicit modelling, this study advances understanding of how long-term urban form in arid cities emerges through the interaction of population dynamics, infrastructure investment, and environmental constraints. Across all four cities, Patch Density (PD) and Edge Density (ED) reveal a pronounced fragmentation peak around the mid-2010s, indicating phases in which spatial expansion temporarily outpaced consolidation. This pattern is most pronounced in the basin-confined cities of Makkah and Madinah, while Jeddah maintains consistently lower fragmentation reflecting its linear, corridor-oriented growth constrained by the coastline. These trajectories suggest that, in arid environments, fragmentation represents a recurrent transitional state rather than a linear precursor to compact urban form, challenging classical compact–sprawl dichotomies. The broadly consistent sequencing of growth phases observed across geographically and functionally diverse cities suggests shared arid-region urbanisation tendencies, rather than a uniform developmental logic. Despite differences in magnitude and spatial expression, all four cities exhibit a comparable progression from early fragmented expansion to corridor-driven consolidation and, more recently, suburban diffusion (Li et al., 2013 ; Liu et al., 2016 ). This convergence reflects the sensitivity of desert landscapes to accessibility, limited vegetation cover, and open terrain, while differences in timing and intensity are mediated by local topography and infrastructure hierarchies. Regulatory interventions—such as redevelopment programmes, transport investments, and growth boundary adjustments—have further shaped these trajectories, even though they are not explicitly modelled here (Al-Ansi et al., 2023 ; UN-Habitat, 2020). The integration of demographic efficiency indicators with spatial metrics demonstrates that population growth alone does not explain land consumption dynamics (Angel et al., 2011 ). Periods of disproportionate land uptake—particularly in Jeddah and Makkah—highlight the role of infrastructure-led development, redevelopment cycles, and spatial constraints in decoupling demographic pressure from land conversion (Imam et al., 2016 ; Tang et al., 2024 ). Consistent with regional evidence, nearly all new urban development occurs through the conversion of bare desert land rather than vegetated surfaces (Almadini & Hassaballa, 2019 ), reinforcing the distinctive land-consumption logic of arid urban systems. Embedding these empirical patterns within an ANN–CA–Markov framework allows urban expansion to be interpreted as a path-dependent process governed by spatial inertia. The simulations indicate that near-term growth predominantly reinforces historically established development corridors rather than generating new spatial fronts. Higher predictive accuracy in unconstrained cities such as Jeddah, and lower accuracy in topographically complex cities such as Makkah, align with broader findings that modelling stability declines in geomorphologically constrained arid environments (Abdelkarim, 2025 ; Gaur & Singh, 2023 ). These results are consistent with machine-learning studies showing that historical accessibility structures remain dominant predictors of future urban growth in arid regions (Selmy et al., 2023 ). Taken together, the findings demonstrate that urban expansion in arid cities is shaped by a combination of environmental constraints, infrastructure-led accessibility, and historical spatial inertia, producing broadly generalisable growth processes with locally specific spatial expressions. This integrated analytical approach provides a transferable framework for interpreting long-term urban dynamics in desert environments, with implications for anticipating future land consumption patterns beyond population-centred planning paradigms. 6. Conclusion This study brings together four decades of evidence on urban expansion in Riyadh, Jeddah, Makkah, and Madinah by integrating long-term LULC reconstruction, demographic indicators, and spatially explicit modelling. Despite their very different geographical settings, all four cities exhibit a broadly similar temporal progression of growth, moving from fragmented early expansion to corridor-based consolidation and, in the most recent decade, to more diffuse suburban development. At the same time, the spatial expression of this growth remains highly uneven, shaped by infrastructure networks and persistent topographic constraints. The KDE analysis and the weighted temporal composite highlight long-lived development corridors and hotspots, pointing to a strong degree of spatial inertia in arid-region urbanisation. Linking demographic trends with land-use change further shows that population growth alone cannot fully account for observed patterns of land consumption. While Riyadh displays relatively stable expansion efficiency over time, Jeddah and Makkah experience periods in which land uptake far exceeds demographic growth. These divergences reflect the influence of coastal and basin-confined morphologies, redevelopment cycles, and accessibility-driven expansion, and they underline the limitations of population-centred interpretations of urban growth in environmentally constrained settings. The ANN–CA–Markov simulations reproduce observed expansion patterns with a high level of spatial agreement and suggest that future growth to 2034 is likely to follow established trajectories rather than generate entirely new development fronts. Predicted expansion remains concentrated along existing corridors—eastward in Jeddah, northward in Riyadh, and within topographically accessible basins in Makkah and Madinah—reinforcing the importance of historical infrastructure and terrain in shaping long-term urban form. Overall, the findings indicate that urban expansion in Saudi Arabia’s major cities is strongly path-dependent and not explained by demographic growth alone, with infrastructure investment and environmental context playing a central role. However, from a planning perspective, this implies a need to manage corridor-driven pressures, recognise the vulnerability of basin and valley landscapes, and guide future growth toward more spatially coherent and accessibility-efficient forms. Further research could extend this framework by incorporating scenario-based modelling, socio-economic drivers, and climate-related constraints, helping to support more anticipatory approaches to urban planning in rapidly transforming desert environments. In addition, investigating if and how urban development patterns have been influenced by specific sectoral and spatial policies or plans would help to identify future needs for effective urban plans and policies. References Rahman, M. (2016). Detection of Land Use/Land Cover Changes and Urban Sprawl in Al-Khobar, Saudi Arabia: An Analysis of Multi-Temporal Remote Sensing Data. ISPRS International Journal of Geo-Information , 5 (2), 15. https://doi.org/10.3390/ijgi5020015 Alkhaldi, G. F., Mosbah, E. B., & Emam, A. A. (2025). Assessment of Land Cover Changes and an Exploration of the Sustainability Key Factors at Al-Ahsa Oasis in Saudi Arabia. Sustainability , 17 (23), 10821. https://doi.org/10.3390/su172310821 United Nations, Department of Economic and Social Affairs, Population Division. (2023). World Urbanization Prospects 2022: Summary of Results. United Nations. https://population.un.org/wup/ Alam, T., & Banerjee, A. (2023). Characterizing land transformation and densification using urban sprawl metrics in the South Bengal region of India. Sustainable Cities and Society , 89 , 104295. https://doi.org/10.1016/j.scs.2022.104295 Alhajri, M. F. (2025). Urban challenges in Dammam metropolitan area: the oil capital of Saudi Arabia. Frontiers in Built Environment , 11 . https://doi.org/10.3389/fbuil.2025.1581017 Al Atni, B. S. (2015). The internationalisation of urban planning strategies: Environmental sustainable urban centres in the Kingdom of Saudi Arabia [Doctoral dissertation, The University of Manchester]. ProQuest Dissertations & Theses Global. https://www.proquest.com/docview/1827522994?pq-origsite=gscholar&fromopenview=true&sourcetype=Dissertations%20&%20Theses Xu, T., Zhou, D., & Li, Y. (2022). Integrating ANNs and Cellular Automata–Markov Chain to Simulate Urban Expansion with Annual Land Use Data. Land , 11 (7), 1074. https://doi.org/10.3390/land11071074 Al-Saud, L.E., & Goussous, J.S. (2023). Waterfront Urban Regeneration and Its Impact on Socio-cultural Activities (The Case of Jeddah Waterfront). In: Kostopoulou, S., Herrera-Franco, G., Wood, J., Al-Kodmany, K. (eds) Cities’ Vocabularies and the Sustainable Development of the Silkroads. SRSTDCH 2021. Advances in Science, Technology & Innovation . Springer, Cham. https://doi.org/10.1007/978-3-031-31027-0_3 Angel, S. (2023). Urban expansion: theory, evidence and practice. Buildings and Cities . https://storage.googleapis.com/jnl-up-j-bc-files/journals/1/articles/348/6447a3db8ae23.pdf Cai, G., Zhang, J., Du, M., Li, C., & Shu, P. (2020). Identification of urban land use efficiency by indicator-SDG 11.3.1. PLoS ONE 15(12): e0244318. https://doi.org/10.1371/journal.pone.0244318 Fastner, K. Djibrilla, S. Nguyen, TT. & Buerkert, A. (2023) Telecoupled urban demand from West. African cities causes social-ecological land use transformation in Saharan oases . PLoS ONE 18(9): e0289694. https://doi.org/10.1371/journal.pone.0289694 Hou, Y., Kuang, W., & Dou, Y. (2023). Observing the compact trend of urban expansion patterns in global 33 megacities during 2000–2020. Journal of Geographical Sciences . https://doi.org/10.1007/s11442-023-2180-0 Hu, Y., Wang, F., Guin, C., & Zhu, H. (2020). A Spatio-Temporal Kernel Density Estimation Framework for Predictive Crime Hotspot Mapping and Evaluation. Applied Geography . https://doi.org/10.48550/arxiv.2006.00272 Tharik, M., Arumugam, K., & Vijayaraghavalu, S. S. (2025). Spatiotemporal assessment and simulation of land use and land cover dynamics in coastal Tamil Nadu using CA–ANN modelling for sustainable development planning. Results in Engineering , 28 , 107771. https://doi.org/10.1016/j.rineng.2025.107771 Jarrar, O. & Al‐Homoud, M. (2024). Sustainable urban development in Riyadh: a projected model for walkability. International Journal of Urban Sustainable Development . https://doi.org/10.1080/19463138.2024.2421163 Yahya M., N., & Abdullahi, H. (2023). Urban Growth Studies for the Holy City of Makkah, Saudi Arabia, through Literature Synthesis. Journal of Hunan University Natural Sciences , 50 (5). https://doi.org/10.55463/issn.1674-2974.50.5.13 Rashid, M. & Bindajam, A. (2015). Space, movement and heritage planning of the historic cities in Islamic societies: Learning from the Old City of Jeddah, Saudi Arabia. Urban Design International. Springer NATURE. 20, 107–129. https://doi.org/10.1057/udi.2014.6 Resch, B., Kolokoussis, P., Hanny, D., Brovelli, M., & Boulos, M. N. (2025). The generative revolution: AI foundation models in geospatial health—applications, challenges and future research. International Journal of Health Geographics . https://doi.org/10.1186/s12942-025-00391-0 Tammar, A. (2017). Assessing Resilient Post-Disaster Recovery of a Flash-Flood-Prone Area: A Study of the City of Jeddah, Kingdom of Saudi Arabia. Zhang, Q., He, J., Ren, C., Chang, C., Tang, W., Ju, X., Saitiniyaz, A., Guo, X., Cui, L., & Liu, L. (2025). Sensitive shrubland and unique urbanization patterns: the key to understanding emerging towns growth in arid regions. Environmental Monitoring and Assessment . https://doi.org/10.1007/s10661-025-13693-4 Almadini, A. M., & Hassaballa, A. A. (2019). Depicting changes in land surface cover at Al-Hassa oasis of Saudi Arabia using remote sensing and GIS techniques. PLOS ONE , 14 (11), e0221115. https://doi.org/10.1371/journal.pone.0221115 McGarigal, K., & Marks, B. J. (1995). FRAGSTATS: spatial pattern analysis program for quantifying landscape structure. https://doi.org/10.2737/PNW-GTR-351 Faichia, C., Tong, Z., Zhang, J., Liu, X., Kazuva, E., Ullah, K., & Al-Shaibah, B. (2020). Using RS Data-Based CA–Markov Model for Dynamic Simulation of Historical and Future LUCC in Vientiane, Laos. Sustainability , 12 (20), 8410. https://doi.org/10.3390/su12208410 Hamad, R., Balzter, H., & Kolo, K. (2018). Predicting Land Use/Land Cover Changes Using a CA-Markov Model under Two Different Scenarios. Sustainability , 10 (10), 3421. https://doi.org/10.3390/su10103421 Li, C., Li, J., & Wu, J. (2013). Quantifying the speed, growth modes, and landscape pattern changes of urbanization: a hierarchical patch dynamics approach. Landscape Ecology , 28 (10), 1875–1888. https://doi.org/10.1007/s10980-013-9933-6 Tang, X., Sheng, L., & Zhou, Y. (2022). Exploring complex urban growth and land use efficiency in China’s developed regions: implications for territorial spatial planning. Frontiers of Earth Science , 16 (4), 1040–1051. https://doi.org/10.1007/s11707-022-0973-6 United Nations Human Settlements Programme (UN-Habitat), & Ministry of Municipal and Rural Affairs (MOMRA). (2019a). Jeddah City Profile . https://unhabitat.org/sites/default/files/2020/04/jeddah.pdf United Nations Human Settlements Programme (UN-Habitat), & Ministry of Municipal and Rural Affairs (MOMRA). (2019b). Madinah City Profile . https://unhabitat.org/sites/default/files/2020/04/madinah.pdf United Nations Human Settlements Programme (UN-Habitat), & Ministry of Municipal and Rural Affairs (MOMRA). (2019c). Makkah City Profile . https://unhabitat.org/sites/default/files/2020/03/makkah.pdf United Nations Human Settlements Programme (UN-Habitat), & Ministry of Municipal and Rural Affairs (MOMRA). (2019d). Riyadh City Profile . https://unhabitat.org/sites/default/files/2020/03/riyadh.pdf Vahid, R., & Aly, M. H. (2025). An Integrated CA–Markov Modeling Framework for Forecasting Land Use and Land Cover Dynamics in Arkansas, USA. Geomatics , 5 (4), 62. https://doi.org/10.3390/geomatics5040062 Zhang, Q., Liu, S., Niu, Y., Hu, Y., Li, L., Cai, E., Zhang, Y., & Zhao, M. (2025). Multi-Scenario Simulation of the Dynamic Relationship Between Land Use and Carbon Storage in the Urbanization Process: A Case Study of Zhengzhou, China. Land , 14 (6), 1227. https://doi.org/10.3390/land14061227 Abdelkarim, A. (2025). Monitoring and forecasting of land use/land cover (LULC) in Al-Hassa Oasis, Saudi Arabia based on the integration of the Cellular Automata (CA) and the Cellular Automata-Markov Model (CA-Markov). Geology, Ecology, and Landscapes , 9 (1), 13–44. https://doi.org/10.1080/24749508.2022.2163741 EL-SHORBAGY, A.-M. (2020). Unplanned Settlements in Saudi Arabia. The Case of Al-Sabeel District, Jeddah. Journal of Settlements and Spatial Planning, 11(2), 81–94. https://doi.org/10.24193/JSSP.2020.2.03 Alqurashi, A. F. (2021). Quantification of Urban Patterns and Processes through Space and Time Using Remote Sensing Data: A Comparative Study between Three Saudi Arabian Cities. Sustainability , 13 (22), 12615. https://doi.org/10.3390/su132212615 Alqurashi, A., Kumar, L., & Sinha, P. (2016). Urban Land Cover Change Modelling Using Time-Series Satellite Images: A Case Study of Urban Growth in Five Cities of Saudi Arabia. Remote Sensing , 8 (10), 838. https://doi.org/10.3390/rs8100838 Almalki, K. A., Al Mosallam, M. S., Aldaajani, T. Z., & Al-Namazi, A. A. (2022). Landforms characterization of Saudi Arabia: Towards a geomorphological map. International Journal of Applied Earth Observation and Geoinformation , 112 , 102945. https://doi.org/10.1016/j.jag.2022.102945 Dahy, B., Al-Memari, M., Al-Gergawi, A., & Burt, J. A. (2024). Remote sensing of 50 years of coastal urbanization and environmental change in the Arabian Gulf: a systematic review. Frontiers in Remote Sensing , 5 . https://doi.org/10.3389/frsen.2024.1422910 Congedo, L. (2021). Semi-Automatic Classification Plugin: A Python tool for the download and processing of remote sensing images in QGIS. Journal of Open Source Software , 6 (64), 3172. https://doi.org/10.21105/joss.03172 Gaur, S., & Singh, R. (2023). A Comprehensive Review on Land Use/Land Cover (LULC) Change Modeling for Urban Development: Current Status and Future Prospects. Sustainability , 15 (2), 903. https://doi.org/10.3390/su15020903 Imam, A., Alhaddad, B., & Roca, J. (2016). REMOTE SENSING EFFICIENCY FOR URBAN ANALYSIS OF MECCA AND SURROUNDS. ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences , XLI-B8 , 905–910. https://doi.org/10.5194/isprsarchives-XLI-B8-905-2016 Liu, Z., He, C., & Wu, J. (2016). General Spatiotemporal Patterns of Urbanization: An Examination of 16 World Cities. Sustainability , 8 (1), 41. https://doi.org/10.3390/su8010041 Selmy, S. A. H., Kucher, D. E., Mozgeris, G., Moursy, A. R. A., Jimenez-Ballesta, R., Kucher, O. D., Fadl, M. E., & Mustafa, A. A. (2023). Detecting, Analyzing, and Predicting Land Use/Land Cover (LULC) Changes in Arid Regions Using Landsat Images, CA-Markov Hybrid Model, and GIS Techniques. Remote Sensing , 15 (23), 5522. https://doi.org/10.3390/rs15235522 Tang, X., Liu, F., & Hu, X. (2024). Urban growth simulation and scenario projection for the arid regions using heuristic cellular automata. Scientific Reports , 14 (1), 21106. https://doi.org/10.1038/s41598-024-71709-4 Wang, W., Luan, W., Jing, H., Zhu, J., Zhang, K., Ma, Q., Zhang, S., & Liang, X. (2024). Quantitative Assessment of Urban Expansion Impact on Vegetation in the Lanzhou–Xining Urban Agglomeration. Applied Sciences , 14 (19), 8615. https://doi.org/10.3390/app14198615 General Authority for Statistics (GASTAT). (2023). Population Census Data of Saudi Arabia. https://www.stats.gov.sa/en/ Al-Ansi, N. A., Uddin, B., Alhrabi, A., & Wahid, J. (2023). Impacts of Urban Growth Policy on Loss of Identity in Expanding Saudi Cities: A Case Study of Buraydah. International Journal of Sustainable Development and Planning , 18 (10), 2975–2987. https://doi.org/10.18280/ijsdp.181001 Alfakhrani, A., Alzamil, W. S., & Hanafi, W. H. H. (2025). Evaluating the Procedural Framework of Informal Settlement Upgrading Makkah City, Saudi Arabia: Challenges and Lessons for Sustainable Urban Renewal. International Journal of Sustainable Development and Planning , 20 (7), 2817–2827. https://doi.org/10.18280/ijsdp.200707 Angel, S., Parent, J., Civco, D. L., Blei, A., & Potere, D. (2011). The dimensions of global urban expansion: Estimates and projections for all countries, 2000–2050. Progress in Planning , 75 (2), 53–107. https://doi.org/10.1016/j.progress.2011.04.001 Additional Declarations The authors declare no competing interests. Supplementary Files Appendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8447461","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":565432961,"identity":"24ceb8ba-23b8-4efd-b409-90c99a04ecad","order_by":0,"name":"Alireza Babapoorkamani","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0002-1878-8655","institution":"Department of Civil Engineering and Architecture, University of Pavia, Italy","correspondingAuthor":true,"prefix":"","firstName":"Alireza","middleName":"","lastName":"Babapoorkamani","suffix":""},{"id":565432962,"identity":"9a7d7de3-0b2a-4fa5-b550-4adba80d92b5","order_by":1,"name":"Liana Ricci","email":"","orcid":"https://orcid.org/0000-0003-0299-4136","institution":"UN-Habitat","correspondingAuthor":false,"prefix":"","firstName":"Liana","middleName":"","lastName":"Ricci","suffix":""},{"id":565432963,"identity":"73c1d6cc-2e6f-4ee2-96bf-0493b6454740","order_by":2,"name":"Tazyeen Alam","email":"","orcid":"https://orcid.org/0000-0002-6076-0947","institution":"Department of Civil and Environmental Engineering and Architecture (DICAAR), University of Cagliari, Italy","correspondingAuthor":false,"prefix":"","firstName":"Tazyeen","middleName":"","lastName":"Alam","suffix":""}],"badges":[],"createdAt":"2025-12-25 08:12:40","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8447461/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8447461/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":99493806,"identity":"3e509510-062a-4374-aef5-d60357bd164d","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2404729,"visible":true,"origin":"","legend":"","description":"","filename":"ManuscriptAnonymous.docx","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/d68c96a3b23c1eb6d4253700.docx"},{"id":99493811,"identity":"a2e313fd-74ec-4539-8bbc-f8ed1bc40c03","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs8447461.json","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/a8b7bd39fe3ef036725bd962.json"},{"id":99790062,"identity":"0ed518d6-3cea-48fb-bd6e-2fb8460a3492","added_by":"auto","created_at":"2026-01-08 12:54:07","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":149880,"visible":true,"origin":"","legend":"","description":"","filename":"rs84474610enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/a4ae75e0a24a357729a1463e.xml"},{"id":99790873,"identity":"ba86b8a3-c5ba-480f-b422-c456f8e455af","added_by":"auto","created_at":"2026-01-08 12:58:48","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":252018,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/a27752da77b7a1aa5d6f7152.jpeg"},{"id":99493818,"identity":"c03c0069-4872-4fbf-a0fa-e43457ceba9a","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":203035,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/575d36a2169be6364369e706.jpeg"},{"id":99493820,"identity":"f00e1d4e-c7a0-43fc-a1a1-6aed5b83619d","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":208249,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/16caaebb861919708e4bcae7.jpeg"},{"id":99790106,"identity":"afe7b9ea-dd32-4642-a3f0-fa73fb75f919","added_by":"auto","created_at":"2026-01-08 12:55:45","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":604806,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/000a25c0f3f7b91c3549723d.png"},{"id":99493822,"identity":"eb14017c-56b4-4f6b-a898-a09ddc61273d","added_by":"auto","created_at":"2026-01-05 05:50:45","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20296,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/587947045a352da186b8768a.png"},{"id":99493816,"identity":"fb485224-5c03-4b54-af2d-12f9634bed11","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1044247,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/b8423a2b257fe2df2ae83739.png"},{"id":99493826,"identity":"00219174-dd26-46c3-92f8-99f751a0f334","added_by":"auto","created_at":"2026-01-05 05:50:45","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":291902,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/ad26063810abbcc555d70ba4.png"},{"id":99493824,"identity":"2432e61a-0a31-4789-9050-61dd4954e6c5","added_by":"auto","created_at":"2026-01-05 05:50:45","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":167091,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/55002eb9ce8af6a365c0eb6d.png"},{"id":99790679,"identity":"59e72478-591a-4eee-9f00-2063b4547bd7","added_by":"auto","created_at":"2026-01-08 12:58:32","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":233119,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/3082e49cfa199016aef0d654.png"},{"id":99791203,"identity":"768b7b7e-d0c5-4adc-95a2-187803f1416f","added_by":"auto","created_at":"2026-01-08 12:59:17","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":104525,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/da7d3d71441120b9a8f5adf4.png"},{"id":99791168,"identity":"0b0f40f8-d622-4d1d-a22c-09b166336907","added_by":"auto","created_at":"2026-01-08 12:59:13","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7296,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/a81493e6806bfda112055da1.png"},{"id":99790369,"identity":"0d47c9ab-f903-4a76-820b-88905d0e4ccf","added_by":"auto","created_at":"2026-01-08 12:57:57","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":92702,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/fcd08d08bb0acbc8dd55674c.png"},{"id":99493819,"identity":"b5054599-8952-42de-a7d7-7eee89934cb0","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"xml","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":147407,"visible":true,"origin":"","legend":"","description":"","filename":"rs84474610structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/961fbc85c501e3a88b05454a.xml"},{"id":99791072,"identity":"5cbab3e4-0413-41a5-9a4b-f349e9a6e7ca","added_by":"auto","created_at":"2026-01-08 12:59:03","extension":"html","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":158531,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/dcabe219f5826049c9e303a5.html"},{"id":99790072,"identity":"810b86bd-7a20-4981-9a9c-4764943832a8","added_by":"auto","created_at":"2026-01-08 12:54:31","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":640411,"visible":true,"origin":"","legend":"\u003cp\u003eBuilt-up area maps for the four study cities across the five temporal stages (1984–2024): (a) Jeddah, (b) Makkah, (c) Madinah, and (d) Riyadh. The maps illustrate spatial patterns of urban expansion.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/9ba0669d857b7bcbd45443e0.png"},{"id":99791164,"identity":"7a2e7bec-e949-4f33-b5c5-d319fe2ab450","added_by":"auto","created_at":"2026-01-08 12:59:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":273307,"visible":true,"origin":"","legend":"\u003cp\u003eOverall Methodological Framework\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/c0a3a45a93a041e15d686e80.png"},{"id":99493814,"identity":"0faa0afc-5d1e-46ff-b452-bed70ade2060","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":812126,"visible":true,"origin":"","legend":"\u003cp\u003eKernel Density Estimation (KDE) maps of newly built-up areas (1984–2024) for the four study cities (top to bottom: Madinah, Riyadh, Makkah, Jeddah) and four decadal intervals (left to right: 1984–1994, 1994–2004, 2004–2014, 2014–2024). Darker tones indicate higher concentrations of new development. The composite panel summarises the timing and persistence of urban growth, with RGB colours representing early (red), mid-period (green), and late (blue) expansion.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/b97a3780ca1661933583a62a.png"},{"id":99493813,"identity":"214af1d2-c0e2-4f8e-a9c5-c611d618d025","added_by":"auto","created_at":"2026-01-05 05:50:44","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":537810,"visible":true,"origin":"","legend":"\u003cp\u003eDifference Built-up area in 2024 and 2034 prediction (a) Jeddah (b)Makkah (c)Madinah (d)Riyadh\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/bc4c1a0d0fd9cd857cf668b5.png"},{"id":100406243,"identity":"44bb3e8d-e7aa-4316-845a-5041a65f38a6","added_by":"auto","created_at":"2026-01-16 12:58:59","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2936862,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/527c210e-f163-4d64-bc7d-c7aa38af8159.pdf"},{"id":99791155,"identity":"325bea34-fd88-4479-9201-1c43a5268304","added_by":"auto","created_at":"2026-01-08 12:59:12","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1088655,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-8447461/v1/da88add85d5289f163404708.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003ePath-Dependent Urban Expansion in Arid Cities: A Multi-Decadal Remote Sensing Analysis and ANN–CA–Markov Modelling of Saudi Arabian Cities (1984–2034)\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eUrban expansion in fast-growing cities is a spatially uneven and temporally dynamic process shaped by demographic growth, infrastructure development, uneven implementation of plans and regulatory frameworks, and environmental constraints. According to the United Nations World Urbanization Prospects, Saudi Arabia has undergone one of the most rapid urban transitions in the Middle East, with population growth concentrated in a small number of major metropolitan areas (United Nations, 2023). In arid-region cities, urbanisation processes are particularly visible and consequential: limited vegetation cover, reflective soils, and the absence of buffered landscapes render land conversion both spatially legible and environmentally impactful (Fastner et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Urban growth in desert environments therefore follows morphological logics distinct from those of temperate regions characterised by outward urban expansion, low-density development, and strong dependence on engineered infrastructure, water systems, and episodic reconstruction. Analysing these trajectories requires frameworks capable of integrating long-term land-cover change with demographic dynamics and spatial modelling (Zhang et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRemote sensing\u0026ndash;based studies have documented the expansion of major Saudi cities, including Riyadh, Jeddah, Makkah, and Madinah, consistently revealing outward growth, conversion of bare land to built-up areas, and strong links to population increase and national development strategies (Rahman, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Alkhaldi et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). These four cities are therefore the focus of this study, as they combine demographic significance with sufficiently consistent long-term LULC records. Although Dammam was considered, its functional integration within a broader coastal\u0026ndash;industrial urban system\u0026mdash;extending beyond administrative boundaries to include Dhahran, Khobar, Qatif, Ras Tanura, and Al Jubail\u0026mdash;would require a different spatial delineation and introduce additional challenges for multi-decadal satellite-based reconstruction and modelling (Al Atni, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Alhajri, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). To ensure analytical consistency and comparability, Dammam is not included.\u003c/p\u003e \u003cp\u003eDespite extensive work on Saudi urban expansion, three gaps remain. First, most studies rely on short time spans or widely spaced temporal snapshots, limiting insight into the intensity and sequencing of urbanisation across multiple decades. Second, demographic change\u0026mdash;although a central driver of land consumption\u0026mdash;is rarely integrated analytically with spatial expansion metrics. Third, predictive modelling approaches using Cellular Automata (CA), Markov chains, or hybrid ANN\u0026ndash;CA frameworks often depend on inconsistent LULC inputs or omit density, fragmentation, and temporal path-dependence as modelling drivers (Resch et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). At the same time, shifts in national planning frameworks\u0026mdash;such as rezoning initiatives, urban development corridors, and evolving Urban Growth Boundaries\u0026mdash;have influenced where and how expansion occurred, shaping observed spatial trajectories even when not directly incorporated as modelling variables (Wang et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRecent urban analytics research increasingly emphasises the need to coupling demographic processes with spatial form. Indicators such as Decadal Growth Rate (DGR), Land Consumption Rate (LCR), and Land Consumption Efficiency (LCE) quantify whether cities expand proportionally to population growth or consume land inefficiently (Cai et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Complementary metrics, including Built-up Density (Db) and Expansion Percentage Change (EPC), reveal how growth is absorbed spatially. However, these indicators are seldom integrated with multi-decadal LULC reconstructions or embedded within spatially explicit modelling frameworks. In arid cities\u0026mdash;where land conversion has pronounced environmental implications and population growth has been rapid over four decades\u0026mdash;such integration is critical for understanding both drivers and consequences of expansion (Angel, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis study addresses these gaps through an integrated framework analysing urban expansion in Riyadh, Jeddah, Makkah, and Madinah from 1984 to 2024. These cities represent heterogeneous arid urban systems shaped by different economic functions and physical constraints: Riyadh as a rapidly expanding capital on an arid plateau (Jarrar \u0026amp; Al-Homoud, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), Jeddah as a coastal metropolis influenced by port activity and post-disaster reconstruction (Tammar, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), Makkah as a pilgrimage-driven city undergoing large-scale redevelopment (Al-Saud \u0026amp; Goussous, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), and Madinah as a historically layered city with steady but spatially constrained growth (Rashid \u0026amp; Bindajam, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Together, they provide a comparative setting for examining how demographic pressure, infrastructure-led development, and environmental constraints interact in arid contexts. Local planning interventions\u0026mdash;particularly zoning updates and transport-led development\u0026mdash; have influenced urban expansion and further contextualise observed patterns, alongside the role of unplanned settlements in cities such as Jeddah and Makkah (Aljoufie et al., 2013; Almazroui et al., 2017; El-Shorbagy, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Alfakhrani et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMethodologically, the study reconstructs a consistent 40-year LULC record using multisensory satellite imagery, integrates demographic indicators from census and gridded population datasets, and employs spatial metrics capturing expansion intensity, density, and fragmentation. These empirical patterns are operationalised within a hybrid ANN\u0026ndash;CA\u0026ndash;Markov framework to simulate near-term urban expansion, incorporating physical constraints and demographic pressures, and validated using multiple agreement metrics (Hu et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Xu et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Hou et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Tharik et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Alam \u0026amp; Banerjee, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eOverall, this study contributes by (i) producing a consistent four-decade LULC reconstruction for Saudi Arabia\u0026rsquo;s major cities, (ii) explicitly coupling demographic efficiency indicators with spatial expansion and fragmentation metrics, and (iii) embedding these relationships within a validated, spatially explicit modelling framework. By linking long-term reconstruction, demographic\u0026ndash;spatial coupling, and predictive simulation, the study advances urban analytics approaches suited to rapidly transforming arid environments.\u003c/p\u003e"},{"header":"2. Study Area and Data","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study Areas\u003c/h2\u003e \u003cp\u003eSaudi Arabia is a predominantly arid country, with desert ecosystems covering more than 70% of its territory (Sayed \u0026amp; Masrahi, 2023). Since the early 1980s, oil-driven development has triggered rapid land use and land cover (LULC) transformation, placing the Kingdom among the fastest urbanising regions in the Middle East (Alqurashi \u0026amp; Kumar, 2014; Alqahtany, 2025). Multitemporal satellite imagery and remote sensing techniques therefore provide a critical means of monitoring long-term urban expansion, with post-classification comparison (PCC) widely used for detecting LULC change (Liu et al., 2023).\u003c/p\u003e \u003cp\u003eThe study focuses on four major Saudi cities\u0026mdash;Riyadh, Jeddah, Makkah, and Madinah\u0026mdash;selected for their demographic significance and contrasting geographical and functional contexts. Riyadh, the capital, is situated on the Najd Plateau; Jeddah is a major coastal port and commercial hub on the Red Sea; Makkah is a mountainous pilgrimage city in the Hijaz region; and Madinah occupies an elongated sedimentary basin constrained by volcanic fields and surrounding highlands. Together, these cities represent Saudi Arabia\u0026rsquo;s dominant administrative, economic, and pilgrimage-driven urban systems and account for most of the national urban growth. Their contrasting settings provide a robust comparative framework for analysing how environmental constraints and socio-economic functions shape long-term urban expansion trajectories.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents built-up area maps for the four cities between 1984 and 2024, illustrating distinct morphological characteristics, expansion directions, and density gradients. These patterns form the empirical basis for subsequent analysis of land consumption, growth efficiency, and urban form dynamics.\u003c/p\u003e \u003cp\u003eTo ensure consistency across temporal and spatial analyses, standardised metropolitan extents were defined for each city, encompassing both historical urban cores and the full perimeter of outward expansion over the 40-year period (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). These extents were delineated beyond administrative boundaries to capture peri-urban transitions and regionally connected development fronts. All demographic indicators and urban expansion metrics (Db, LCR, LCE, PD, and ED) were computed within these standardised extents, ensuring full comparability across cities and years.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Data Sources\u003c/h2\u003e \u003cp\u003eThe analysis integrates multi-temporal satellite imagery, census-based population data, and ancillary geospatial datasets to examine long-term urban expansion across the four study areas. Built-up dynamics were derived from Landsat imagery for 1984, 1994, 2004, 2014, and 2024, obtained from the USGS EarthExplorer platform. Landsat 5 TM, Landsat 7 ETM+, and Landsat 8/9 OLI scenes were selected based on availability, ensuring consistent spatial resolution (30 m) and dry-season acquisition. All imagery was radiometrically and atmospherically corrected, mosaicked where necessary, and clipped to standardised metropolitan extents. A summary of satellite datasets is provided in Table \u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003eA2\u003c/span\u003e, while full scene metadata are reported in Supplementary Tables S1\u0026ndash;S4.\u003c/p\u003e \u003cp\u003ePopulation statistics were obtained from the General Authority for Statistics (GASTAT) for census years 1992, 2004, 2010, and 2022 and compiled at the city level. Decadal population growth rates were calculated using exponential interpolation, and 2024 population estimates were derived from city-specific 2010\u0026ndash;2022 trends. All demographic indicators were computed within the same standardised metropolitan extents to ensure consistency with spatial analyses. Built-up maps derived from the LULC classification were used to calculate Db, LCR, and LCE following standard formulations (Supplementary Material).\u003c/p\u003e \u003cp\u003eAdditional geospatial datasets were used to characterise terrain and accessibility. A 30 m SRTM DEM was processed to derive slope, capturing topographic constraints on urban expansion, particularly in Makkah and Madinah. Road networks extracted from OpenStreetMap were converted into distance-to-road layers to represent accessibility. Global Human Settlement Layer (GHSL) population grids were used to support interpretation of settlement patterns and cross-validate spatial density gradients. Together, these datasets provide a coherent foundation for the classification, spatial analysis, and modelling framework applied in this study.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Methodology","content":"\u003cp\u003eThis study implements a three-phase, integrated framework combining multi-decadal landcover reconstruction, demographic and spatial coupling analysis, and spatially explicit urban growth modelling. The workflow (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e) ensures end-to-end consistency through standardised metropolitan areas, unified classification protocols, and cross-validated datasets.\u003c/p\u003e\n\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\n \u003ch2\u003e3.1. Image Pre-processing\u003c/h2\u003e\n \u003cp\u003eWe obtained multi-temporal satellite imagery from Landsat TM, ETM+, and OLI (30 m), ASTER (15 m), and Sentinel-2 MSI (10 m) for the benchmark years 1984, 1994, 2004, 2014, and 2024. Standard radiometric and atmospheric correction procedures were implemented to ensure spectral consistency across sensors and seasons. Landsat images were corrected using LEDAPS, while Sentinel-2 scenes were processed with Sen2Cor to retrieve surface reflectance. All images were geometrically corrected to the WGS84/UTM coordinate system, and scenes were mosaicked where necessary to verify that the entire metropolitan area was covered. Cloud masking and quality filtering were applied using FMasks algorithms and QA bands to remove clouds, shadows, and other artefacts. Finally, all datasets were clipped to the study boundary, producing spatially harmonised, analysis-ready mosaics suitable for classification and temporal comparison.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\n \u003ch2\u003e3.2. Land Cover Classification\u003c/h2\u003e\n \u003cp\u003eA consistent land cover classification scheme comprising four categories\u0026mdash;soil/bare land, built-up areas, vegetation, and water bodies\u0026mdash;was developed to reflect the dominant landscape structure and urbanisation processes characteristic of arid environments. The selected classes align with the study\u0026rsquo;s focus on anthropogenic expansion and land transformation.\u003c/p\u003e\n \u003cp\u003eClassification was done in ENVI using a supervised manual workflow. For each time point, multispectral imagery was visually examined to identify representative training samples for all classes. Soil and bare land were delineated based on their high reflectance in red and shortwave infrared bands; vegetation was identified by its strong near-infrared reflectance; built-up areas were recognised through their heterogeneous tonal and textural signatures; and water bodies were distinguished by their low reflectance in visible and near-infrared wavelengths. ENVI\u0026rsquo;s classification workflow was used to assign pixels to classes, followed by iterative manual refinement to correct misclassifications typically encountered in arid landscapes where spectral confusion between built-up and bare soil is common. Ancillary datasets\u0026mdash;including SRTM-derived elevation and slope, OpenStreetMap road networks, and high-resolution basemaps\u0026mdash;were applied selectively to guide interpretation and ensure classification consistency across time, while avoiding computational biases introduced by automated multi-layer classification models.\u003c/p\u003e\n \u003cp\u003e\u0026lrm;\u003cstrong\u003e3.3 Accuracy Assessment\u0026lrm; and Change Detection Analysis\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eThe classified land cover maps were validated using a stratified random sample of reference points derived from high-resolution imagery and temporally stable basemaps. Confusion matrices were generated for each classified year, and standard accuracy metrics\u0026mdash;including overall accuracy, producer\u0026rsquo;s accuracy, user\u0026rsquo;s accuracy, and the Kappa coefficient\u0026mdash;were calculated to assess thematic reliability across the multi-decadal period. The classification achieved stable accuracy levels suitable for subsequent spatial and temporal change analysis. A summary of overall accuracy and Kappa values for all cities and time points is reported in Table S5.\u003c/p\u003e\n \u003cp\u003eLand cover change was quantified using PCC, which compares classified maps across successive time steps. Transition matrices were generated for each temporal interval to identify the magnitude and direction of class conversions, with particular attention to transitions from bare land to built-up land as the dominant urbanisation pathway. The spatial distribution of changes was visualised using thematic maps to identify growth fronts, infill processes, and peri-urban expansion patterns. Vegetation dynamics were further contextualised using NDVI-based inspection (Appendix B, Equation B1).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003e3.4 Urban Sprawl Analysis\u003c/h2\u003e\n \u003cp\u003eUrban sprawl was assessed using a combination of quantitative growth indicators and spatial pattern metrics. Built-up area was calculated for each time point to derive the Urban Expansion Rate (UER), enabling comparison of growth intensity across decades. Built-up Density (Db) was computed to quantify the proportion of urbanised land relative to the total metropolitan extent (Equation B2).\u003c/p\u003e\n \u003cp\u003eTo characterise spatial configuration and fragmentation of the urban fabric, landscape metrics, including Patch Density (PD) and Edge Density (ED), were calculated for each city and temporal stage. PD measures the number of built-up patches per unit area and serves as an indicator of dispersed development and leapfrogging growth, while lower values reflect more consolidated urban form (Equation B3). Because all analyses were constrained to standardised metropolitan extents (Table \u003cspan class=\"InternalRef\"\u003eA1\u003c/span\u003e), PD values are directly comparable across cities and time periods. ED quantifies the total length of built-up boundaries per unit landscape area and captures the geometric complexity of urban form (Equation B4). Higher ED values indicate fragmented, irregular expansion, whereas lower values reflect more compact and spatially cohesive development. Considered together, PD and ED provide a complementary diagnostic of urban fragmentation, while UER captures the rate of short-term built-up expansion between successive periods.\u003c/p\u003e\n \u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003eTo quantify both the magnitude and efficiency of urban expansion, a set of complementary indicators was employed. Expansion Percentage Change (EPC) measures the decadal increase in built-up land (Eq. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e), while the Decadal Growth Rate (DGR) standardises this change into an annualised metric, enabling comparison across periods. The Land Consumption Rate (LCR) captures the pace of land conversion following SDG 11.3.1 conventions. Land Consumption Efficiency (LCE) was calculated as the ratio of LCR to population growth rate derived from census data, providing an indicator of whether spatial expansion outpaced demographic change. LCE values greater than 1 indicate land-extensive growth, whereas values below 1 reflect compact or population-aligned development. Together, these indicators provide a coherent framework for evaluating both the speed and efficiency of urban expansion across the four cities.\u003c/p\u003e\n \u003cp\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\n \u003ch2\u003e3.5 Urban Growth Modelling and Prediction\u003c/h2\u003e\n \u003cp\u003eUrban expansion was simulated using an integrated ANN\u0026ndash;CA\u0026ndash;Markov framework combining stochastic transition estimation, data-driven suitability modelling, and spatial allocation to reproduce observed urban growth patterns. Markov chains were used to estimate land-cover transition probabilities from observed changes during the 2004\u0026ndash;2014 and 2014\u0026ndash;2024 periods, providing the basis for projecting future conversions under existing development trajectories. Spatial transition potentials were generated using an Artificial Neural Network (ANN) trained on observed land-cover transitions (Equations B5\u0026ndash;B9), incorporating key drivers of arid-region urbanisation, including slope, elevation, distance to roads, and population density, to capture non-linear growth relationships. Predicted transitions were allocated spatially using a Cellular Automata (CA) model, which enforces neighbourhood effects and spatial contiguity, reproducing the corridor-driven and path-dependent expansion patterns identified in the empirical analysis. Model performance was evaluated through back-casting validation, in which the model simulated the 2024 built-up configuration from earlier conditions. Comparison with observed land-cover maps shows good agreement across all four cities, as reflected in overall, histogram, and location Kappa statistics (Table \u003cspan class=\"InternalRef\"\u003eD1\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eFollowing validation, the calibrated model was applied to generate 2034 urban expansion projections under a business-as-usual assumption. Full mathematical formulations, network architecture, and training parameters are provided in the Supplementary Material.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"4. Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003e4.1 Multi-decadal land cover transitions (1984\u0026ndash;2024)\u003c/h2\u003e\n \u003cp\u003eThe multi-temporal LULC series, derived using a semi-automatic classification approach (Congedo, \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e), reveals a consistent pattern of rapid urbanisation across all four cities between 1984 and 2024, characterised by systematic conversion of bare soil and vegetation into built-up land. Exact land-cover area values for built-up, soil, and vegetation classes across all cities and time periods are reported in Table S6 in Supplementary Section S.3, while Figure S1 provides a visual comparison of long-term trends. Despite contrasting geographical contexts\u0026mdash;coastal (Jeddah), mountainous (Makkah and Madinah), and plateau-based (Riyadh) (Almalki et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e)\u0026mdash;all cities exhibit strong outward expansion accompanied by a long-term decline in soil cover and only modest increases in vegetation. Across the study period, built-up land increased substantially in all cities, indicating a shared regional trajectory of land consumption driven by sustained demographic and economic growth (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003ePatch Density (PD) highlights distinct fragmentation trajectories shaped by local morphological constraints. Riyadh exhibits a gradual increase in PD, peaking in 2014, consistent with a transition from early discontinuous suburban growth toward progressive consolidation of peripheral development. Jeddah maintains comparatively low PD values throughout the period, reflecting its corridor-oriented, coastline-constrained expansion, where growth primarily extends existing urban fronts rather than forming isolated patches. In contrast, Makkah records the highest PD values, particularly in 1994 and 2014, indicating pronounced fragmentation associated with basin-confined topography and episodic pilgrimage-driven development. Madinah displays a more oscillatory pattern, with fragmentation peaking in 2014 during rapid peripheral expansion before declining by 2024 as dispersed patches merged into more contiguous urban fabric.\u003c/p\u003e\n \u003cp\u003eEdge Density (ED) further elucidates the evolving spatial complexity of urban form (Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). Jeddah exhibits moderate but steadily increasing ED values, consistent with linear expansion along established coastal and transport corridors. Makkah shows a rise in ED until 2014 followed by a decline in 2024, signalling partial morphological consolidation within a constrained basin environment. Riyadh records a pronounced increase in ED over the study period, reflecting intensifying edge complexity associated with outward suburbanisation and fragmented peri-urban growth. Madinah registers the highest ED values overall, with a peak in 2014 corresponding to rapid expansion along basin corridors. Together, PD and ED provide a complementary diagnostic of urban fragmentation dynamics, distinguishing phases of dispersed, edge-intensive growth from subsequent periods of consolidation despite continued increases in total built-up area.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eBuilt-up Density (Db), Patch Density (PD), and ED trend values for all cities (1984\u0026ndash;2024)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"16\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eCity\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"5\"\u003e\n \u003cp\u003eDb\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"5\"\u003e\n \u003cp\u003ePD\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"5\"\u003e\n \u003cp\u003eED\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1984\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1994\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2004\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2014\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2024\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1984\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1994\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2004\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2014\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2024\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1984\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1994\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2004\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2014\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2024\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eJeddah\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0249\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0278\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0374\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0499\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0933\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.581\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.444\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.699\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.330\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.366\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e15.56\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e22.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e32.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e58.85\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMadinah\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0555\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0788\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0882\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1655\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.2254\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.339\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.259\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.408\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.827\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.850\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.16\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMakkah\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0312\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0907\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1403\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.2023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.3052\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.078\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.262\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.449\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e9.907\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5.238\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.68\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRiyadh\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0377\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.0610\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1442\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.1930\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.629\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.893\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.280\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e9.503\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.394\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e22.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e24.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e31.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e63.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e48.23\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003e4.2 Expansion intensity (EPC and DGR)\u003c/h2\u003e\n \u003cp\u003eDecadal expansion metrics indicate that, despite variation in absolute growth across the four cities, all exhibit a broadly consistent temporal trajectory in three phases (Tang et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e; Zhang et al., \u003cspan class=\"CitationRef\"\u003e2025\u003c/span\u003e): a phase of accelerated urban expansion prior to 2004, a period of moderated growth between 2004 and 2014, and a renewed intensification during 2014\u0026ndash;2024. This three-stage pattern is clearly reflected in the EPC and DGR values summarised in Table S7. The spatial expression of these expansion intensities is further illustrated in Figure S2 in the Supplementary Materials. Rather than isolating each city, these results allow cross-city comparison, showing that Makkah had the most extreme early expansion (1984\u0026ndash;1994) and Jeddah dominated the final decade (2014\u0026ndash;2024) while Riyadh and Madinah followed smoother multi-phase transitions.\u003c/p\u003e\n \u003cp\u003eTo complement EPC, DGR and LCR were calculated for each city (Table 3), revealing steady long-term land consumption in Riyadh and Jeddah, contrasted with extreme early-period conversion in Makkah followed by moderation. LCE values (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e) indicate strong inter-city contrasts: Jeddah shows disproportionately high land consumption relative to population growth during 2010\u0026ndash;2022, whereas Riyadh maintains consistently balanced efficiency across all census intervals. Makkah exhibits marked early-period inefficiency linked to rapid redevelopment and topographic constraints, while Madinah shows fluctuating efficiency due to its basin morphology and limited buildable land.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eLand Consumption Rate (LCR) per decade and Land Consumption Efficiency (LCE) for census intervals\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eCity\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"4\"\u003e\n \u003cp\u003eLCR\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003eLCE\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1984\u0026ndash;1994\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1994\u0026ndash;2004\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2004\u0026ndash;2014\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2014\u0026ndash;2024\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e1992\u0026ndash;2004\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2004\u0026ndash;2010\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e2010\u0026ndash;2022\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRiyadh\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.620\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.699\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.390\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.338\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.24\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eJeddah\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.115\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.348\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.333\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.868\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.76\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMakkah\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.906\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.546\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.442\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.509\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e6.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.15\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMadinah\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.420\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.119\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.877\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.362\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.56\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"left\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003e4.3 Spatial clustering and Spatiotemporal Growth Patterns\u003c/h2\u003e\n \u003cp\u003eKDE and the weighted temporal composite reveal a clear three-stage trajectory of spatial expansion across all four cities. The early period (1984\u0026ndash;1994) was marked by fragmented, discontinuous growth; the mid-period (1994\u0026ndash;2014) showed consolidation along major corridors and urban centres; and the recent decade (2014\u0026ndash;2024) experienced outward diffusion into emerging suburban belts, particularly in low-slope, accessible areas (Liu et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e). This transition from leapfrogging sprawl to more corridor-based development aligns with multi-stage urban growth patterns reported in the broader urbanisation literature (Li et al., \u003cspan class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eFollowing this three-stage trajectory, Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e further illustrates the spatial configuration of these patterns by mapping the KDE-based clustering of newly developed built-up areas across the four cities. In addition to the KDE surfaces, the final composite panel visualises all decades of expansion through a weighted temporal composite, in which the RGB channels encode the timing and persistence of urban growth. Early development is represented in red, mid-period expansion in green, and later growth in blue, while mixed hues (such as yellow, cyan, and magenta) indicate areas that experienced more than one wave of development. White tones mark locations characterised by sustained, multi-decadal intensification. This composite therefore captures not only when expansion occurred but also how consistently growth has accumulated over time. Together, these spatial layers reveal persistent long-term hotspots\u0026mdash;along Jeddah\u0026rsquo;s inland eastern corridor, within the southern and eastern basins of Makkah, across the western and northern valleys of Madinah, and along Riyadh\u0026rsquo;s northern expansion axes\u0026mdash;highlighting the decisive role of infrastructure, land availability, and topographic constraints in shaping urban growth trajectories.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003e4.4 Population density trends\u003c/h2\u003e\n \u003cp\u003ePopulation density patterns derived from GHSL data (1985\u0026ndash;2025) show a clear intensification across all four cities, closely mirroring the spatial distribution of built-up expansion (Figure \u003cspan class=\"InternalRef\"\u003eE1\u003c/span\u003e). Details on GHSL datasets, preprocessing steps, and density visualisation methods are provided in Supplementary Section S.2. During the early decades (1985\u0026ndash;2005), density increases were largely concentrated within historical cores and first urban rings\u0026mdash;particularly in Riyadh and Jeddah\u0026mdash;reflecting compact development and rapid demographic growth (UN-Habitat, 2020a; 2020b). From 2015 onwards, all cities exhibit a pronounced shift toward peripheral demographic diffusion, consistent with the outward expansion corridors identified through KDE and the weighted temporal composite analysis.\u003c/p\u003e\n \u003cp\u003eDespite contrasting geographical settings, the four cities display highly synchronised demographic\u0026ndash;spatial dynamics over the 40-year period. Topography strongly regulates density redistribution in Makkah and Madinah, where growth concentrates within buildable valleys and sedimentary basins while mountainous barriers channel expansion along transport corridors (UN-Habitat, 2020c; 2020d). In Jeddah, coastal constraints drive inland densification structured by major highway axes (UN-Habitat, 2020b), whereas Riyadh exhibits multi-directional demographic expansion enabled by extensive land availability and a metropolitan-scale road network (UN-Habitat, 2020a). GHSL-derived density trajectories confirm a strong coupling between population growth and land consumption across all cities, particularly during the 2004\u0026ndash;2014 decade, when rapid built-up expansion coincided with marked demographic intensification. The emergence of new suburban belts during 2014\u0026ndash;2024 underscores a shift toward outward, accessibility-driven urban growth shaped jointly by environmental constraints, infrastructure networks, and land availability.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003e4.5 Urban-growth modelling and prediction\u003c/h2\u003e\n \u003cp\u003eThe weighted temporal composite (Built-up Temporal Intensity, BTI) indicates that the 2014\u0026ndash;2024 decade exhibits the highest cumulative intensity of built-up expansion across all four cities (Alqurashi et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e). Persistent multi-decadal hotspots are concentrated along Jeddah\u0026rsquo;s eastern corridor, Riyadh\u0026rsquo;s northern axes, and the basin-aligned valleys of Makkah and Madinah. These BTI-derived hotspots closely align with the high-probability transition zones used in the ANN\u0026ndash;CA\u0026ndash;Markov simulations, confirming that future growth trajectories are strongly shaped by long-term spatial inertia.\u003c/p\u003e\n \u003cp\u003eThe ANN\u0026ndash;CA\u0026ndash;Markov framework was applied to generate spatial predictions of built-up expansion for 2034. Transition probability matrices derived from the 2004\u0026ndash;2014 and 2014\u0026ndash;2024 periods show that bare soil to built-up conversion dominates urban growth, accounting for approximately 85\u0026ndash;92% of all transitions across the four cities, consistent with findings from arid-region case studies (Almadini \u0026amp; Hassaballa, \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e). Suitability surfaces indicate that low-slope, highly accessible areas near major transport corridors exhibit the strongest transition potentials, reinforcing the role of terrain and infrastructure as key drivers of urban expansion (Vahid \u0026amp; Aly, \u003cspan class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eModel validation through back-casting reveals clear performance differences linked to urban morphology and geomorphology (Table \u003cspan class=\"InternalRef\"\u003eD1\u003c/span\u003e). Jeddah achieves the highest agreement (Kappa\u0026thinsp;=\u0026thinsp;0.85), reflecting its relatively predictable corridor-based growth pattern, while Riyadh and Madinah show moderate-to-high accuracy (Kappa\u0026thinsp;=\u0026thinsp;0.68 and 0.72). Makkah records the lowest agreement (Kappa\u0026thinsp;=\u0026thinsp;0.61), consistent with modelling challenges in mountainous, basin-confined urban systems where spatial transitions are less predictable. The 2034 prediction maps (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e) reveal distinct but coherent expansion trajectories across the four cities. Jeddah expands predominantly inland toward the east\u0026ndash;southeast, Riyadh exhibits strong northward and north-eastern growth along major arterials, Makkah\u0026rsquo;s expansion remains constrained within southern and eastern valleys, and Madinah shows peripheral diffusion along western and northern basin corridors. These simulated patterns reinforce the empirical KDE and BTI analyses and align with recent remote-sensing\u0026ndash;based urban growth studies.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"5. Discussion","content":"\u003cp\u003eBy integrating multi-decadal LULC reconstruction with demographic indicators, fragmentation metrics, and spatially explicit modelling, this study advances understanding of how long-term urban form in arid cities emerges through the interaction of population dynamics, infrastructure investment, and environmental constraints. Across all four cities, Patch Density (PD) and Edge Density (ED) reveal a pronounced fragmentation peak around the mid-2010s, indicating phases in which spatial expansion temporarily outpaced consolidation. This pattern is most pronounced in the basin-confined cities of Makkah and Madinah, while Jeddah maintains consistently lower fragmentation reflecting its linear, corridor-oriented growth constrained by the coastline. These trajectories suggest that, in arid environments, fragmentation represents a recurrent transitional state rather than a linear precursor to compact urban form, challenging classical compact\u0026ndash;sprawl dichotomies.\u003c/p\u003e \u003cp\u003eThe broadly consistent sequencing of growth phases observed across geographically and functionally diverse cities suggests shared arid-region urbanisation tendencies, rather than a uniform developmental logic. Despite differences in magnitude and spatial expression, all four cities exhibit a comparable progression from early fragmented expansion to corridor-driven consolidation and, more recently, suburban diffusion (Li et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Liu et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). This convergence reflects the sensitivity of desert landscapes to accessibility, limited vegetation cover, and open terrain, while differences in timing and intensity are mediated by local topography and infrastructure hierarchies. Regulatory interventions\u0026mdash;such as redevelopment programmes, transport investments, and growth boundary adjustments\u0026mdash;have further shaped these trajectories, even though they are not explicitly modelled here (Al-Ansi et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; UN-Habitat, 2020). The integration of demographic efficiency indicators with spatial metrics demonstrates that population growth alone does not explain land consumption dynamics (Angel et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Periods of disproportionate land uptake\u0026mdash;particularly in Jeddah and Makkah\u0026mdash;highlight the role of infrastructure-led development, redevelopment cycles, and spatial constraints in decoupling demographic pressure from land conversion (Imam et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Tang et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Consistent with regional evidence, nearly all new urban development occurs through the conversion of bare desert land rather than vegetated surfaces (Almadini \u0026amp; Hassaballa, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), reinforcing the distinctive land-consumption logic of arid urban systems.\u003c/p\u003e \u003cp\u003eEmbedding these empirical patterns within an ANN\u0026ndash;CA\u0026ndash;Markov framework allows urban expansion to be interpreted as a path-dependent process governed by spatial inertia. The simulations indicate that near-term growth predominantly reinforces historically established development corridors rather than generating new spatial fronts. Higher predictive accuracy in unconstrained cities such as Jeddah, and lower accuracy in topographically complex cities such as Makkah, align with broader findings that modelling stability declines in geomorphologically constrained arid environments (Abdelkarim, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Gaur \u0026amp; Singh, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These results are consistent with machine-learning studies showing that historical accessibility structures remain dominant predictors of future urban growth in arid regions (Selmy et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTaken together, the findings demonstrate that urban expansion in arid cities is shaped by a combination of environmental constraints, infrastructure-led accessibility, and historical spatial inertia, producing broadly generalisable growth processes with locally specific spatial expressions. This integrated analytical approach provides a transferable framework for interpreting long-term urban dynamics in desert environments, with implications for anticipating future land consumption patterns beyond population-centred planning paradigms.\u003c/p\u003e"},{"header":"6. Conclusion","content":"\u003cp\u003eThis study brings together four decades of evidence on urban expansion in Riyadh, Jeddah, Makkah, and Madinah by integrating long-term LULC reconstruction, demographic indicators, and spatially explicit modelling. Despite their very different geographical settings, all four cities exhibit a broadly similar temporal progression of growth, moving from fragmented early expansion to corridor-based consolidation and, in the most recent decade, to more diffuse suburban development. At the same time, the spatial expression of this growth remains highly uneven, shaped by infrastructure networks and persistent topographic constraints. The KDE analysis and the weighted temporal composite highlight long-lived development corridors and hotspots, pointing to a strong degree of spatial inertia in arid-region urbanisation. Linking demographic trends with land-use change further shows that population growth alone cannot fully account for observed patterns of land consumption. While Riyadh displays relatively stable expansion efficiency over time, Jeddah and Makkah experience periods in which land uptake far exceeds demographic growth. These divergences reflect the influence of coastal and basin-confined morphologies, redevelopment cycles, and accessibility-driven expansion, and they underline the limitations of population-centred interpretations of urban growth in environmentally constrained settings. The ANN\u0026ndash;CA\u0026ndash;Markov simulations reproduce observed expansion patterns with a high level of spatial agreement and suggest that future growth to 2034 is likely to follow established trajectories rather than generate entirely new development fronts. Predicted expansion remains concentrated along existing corridors\u0026mdash;eastward in Jeddah, northward in Riyadh, and within topographically accessible basins in Makkah and Madinah\u0026mdash;reinforcing the importance of historical infrastructure and terrain in shaping long-term urban form.\u003c/p\u003e \u003cp\u003eOverall, the findings indicate that urban expansion in Saudi Arabia\u0026rsquo;s major cities is strongly path-dependent and not explained by demographic growth alone, with infrastructure investment and environmental context playing a central role. However, from a planning perspective, this implies a need to manage corridor-driven pressures, recognise the vulnerability of basin and valley landscapes, and guide future growth toward more spatially coherent and accessibility-efficient forms. Further research could extend this framework by incorporating scenario-based modelling, socio-economic drivers, and climate-related constraints, helping to support more anticipatory approaches to urban planning in rapidly transforming desert environments. In addition, investigating if and how urban development patterns have been influenced by specific sectoral and spatial policies or plans would help to identify future needs for effective urban plans and policies.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eRahman, M. (2016). Detection of Land Use/Land Cover Changes and Urban Sprawl in Al-Khobar, Saudi Arabia: An Analysis of Multi-Temporal Remote Sensing Data. \u003cem\u003eISPRS International Journal of Geo-Information\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(2), 15. https://doi.org/10.3390/ijgi5020015 \u003c/li\u003e\n\u003cli\u003eAlkhaldi, G. F., Mosbah, E. B., \u0026amp; Emam, A. A. (2025). Assessment of Land Cover Changes and an Exploration of the Sustainability Key Factors at Al-Ahsa Oasis in Saudi Arabia. \u003cem\u003eSustainability\u003c/em\u003e, \u003cem\u003e17\u003c/em\u003e(23), 10821. https://doi.org/10.3390/su172310821 \u003c/li\u003e\n\u003cli\u003eUnited Nations, Department of Economic and Social Affairs, Population Division. (2023). World Urbanization Prospects 2022: Summary of Results. United Nations. https://population.un.org/wup/ \u003c/li\u003e\n\u003cli\u003eAlam, T., \u0026amp; Banerjee, A. (2023). Characterizing land transformation and densification using urban sprawl metrics in the South Bengal region of India. \u003cem\u003eSustainable Cities and Society\u003c/em\u003e, \u003cem\u003e89\u003c/em\u003e, 104295. https://doi.org/10.1016/j.scs.2022.104295 \u003c/li\u003e\n\u003cli\u003eAlhajri, M. F. (2025). Urban challenges in Dammam metropolitan area: the oil capital of Saudi Arabia. \u003cem\u003eFrontiers in Built Environment\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e. https://doi.org/10.3389/fbuil.2025.1581017\u003c/li\u003e\n\u003cli\u003eAl Atni, B. S. (2015). The internationalisation of urban planning strategies: Environmental sustainable urban centres in the Kingdom of Saudi Arabia [Doctoral dissertation, The University of Manchester]. ProQuest Dissertations \u0026amp; Theses Global. https://www.proquest.com/docview/1827522994?pq-origsite=gscholar\u0026amp;fromopenview=true\u0026amp;sourcetype=Dissertations%20\u0026amp;%20Theses \u003c/li\u003e\n\u003cli\u003eXu, T., Zhou, D., \u0026amp; Li, Y. (2022). Integrating ANNs and Cellular Automata\u0026ndash;Markov Chain to Simulate Urban Expansion with Annual Land Use Data. \u003cem\u003eLand\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(7), 1074. https://doi.org/10.3390/land11071074 \u003c/li\u003e\n\u003cli\u003eAl-Saud, L.E., \u0026amp; Goussous, J.S. (2023). Waterfront Urban Regeneration and Its Impact on Socio-cultural Activities (The Case of Jeddah Waterfront). In: Kostopoulou, S., Herrera-Franco, G., Wood, J., Al-Kodmany, K. (eds) Cities\u0026rsquo; Vocabularies and the Sustainable Development of the Silkroads. SRSTDCH 2021. \u003cem\u003eAdvances in Science, Technology \u0026amp; Innovation\u003c/em\u003e. Springer, Cham. https://doi.org/10.1007/978-3-031-31027-0_3\u003c/li\u003e\n\u003cli\u003eAngel, S. (2023). Urban expansion: theory, evidence and practice. \u003cem\u003eBuildings and Cities\u003c/em\u003e. https://storage.googleapis.com/jnl-up-j-bc-files/journals/1/articles/348/6447a3db8ae23.pdf\u003c/li\u003e\n\u003cli\u003eCai, G., Zhang, J., Du, M., Li, C., \u0026amp; Shu, P. (2020). Identification of urban land use efficiency by indicator-SDG 11.3.1. PLoS ONE 15(12): e0244318. https://doi.org/10.1371/journal.pone.0244318\u003c/li\u003e\n\u003cli\u003eFastner, K. Djibrilla, S. Nguyen, TT. \u0026amp; Buerkert, A. (2023) Telecoupled urban demand from West. \u003cem\u003eAfrican cities causes social-ecological land use transformation in Saharan oases\u003c/em\u003e. PLoS ONE 18(9): e0289694. https://doi.org/10.1371/journal.pone.0289694\u003c/li\u003e\n\u003cli\u003eHou, Y., Kuang, W., \u0026amp; Dou, Y. (2023). Observing the compact trend of urban expansion patterns in global 33 megacities during 2000\u0026ndash;2020. \u003cem\u003eJournal of Geographical Sciences\u003c/em\u003e. https://doi.org/10.1007/s11442-023-2180-0\u003c/li\u003e\n\u003cli\u003eHu, Y., Wang, F., Guin, C., \u0026amp; Zhu, H. (2020). A Spatio-Temporal Kernel Density Estimation Framework for Predictive Crime Hotspot Mapping and Evaluation. \u003cem\u003eApplied Geography\u003c/em\u003e. https://doi.org/10.48550/arxiv.2006.00272\u003c/li\u003e\n\u003cli\u003eTharik, M., Arumugam, K., \u0026amp; Vijayaraghavalu, S. S. (2025). Spatiotemporal assessment and simulation of land use and land cover dynamics in coastal Tamil Nadu using CA\u0026ndash;ANN modelling for sustainable development planning. \u003cem\u003eResults in Engineering\u003c/em\u003e, \u003cem\u003e28\u003c/em\u003e, 107771. https://doi.org/10.1016/j.rineng.2025.107771 \u003c/li\u003e\n\u003cli\u003eJarrar, O. \u0026amp; Al‐Homoud, M. (2024). Sustainable urban development in Riyadh: a projected model for walkability. \u003cem\u003eInternational Journal of Urban Sustainable Development\u003c/em\u003e. https://doi.org/10.1080/19463138.2024.2421163\u003c/li\u003e\n\u003cli\u003eYahya M., N., \u0026amp; Abdullahi, H. (2023). Urban Growth Studies for the Holy City of Makkah, Saudi Arabia, through Literature Synthesis. \u003cem\u003eJournal of Hunan University Natural Sciences\u003c/em\u003e, \u003cem\u003e50\u003c/em\u003e(5). https://doi.org/10.55463/issn.1674-2974.50.5.13 \u003c/li\u003e\n\u003cli\u003eRashid, M. \u0026amp; Bindajam, A. (2015). Space, movement and heritage planning of the historic cities in Islamic societies: Learning from the Old City of Jeddah, Saudi Arabia. \u003cem\u003eUrban Design International. \u003c/em\u003eSpringer NATURE. 20, 107\u0026ndash;129. https://doi.org/10.1057/udi.2014.6\u003c/li\u003e\n\u003cli\u003eResch, B., Kolokoussis, P., Hanny, D., Brovelli, M., \u0026amp; Boulos, M. N. (2025). The generative revolution: AI foundation models in geospatial health\u0026mdash;applications, challenges and future research. \u003cem\u003eInternational Journal of Health Geographics\u003c/em\u003e. https://doi.org/10.1186/s12942-025-00391-0\u003c/li\u003e\n\u003cli\u003eTammar, A. (2017). Assessing Resilient Post-Disaster Recovery of a Flash-Flood-Prone Area: A Study of the City of Jeddah, Kingdom of Saudi Arabia.\u003c/li\u003e\n\u003cli\u003eZhang, Q., He, J., Ren, C., Chang, C., Tang, W., Ju, X., Saitiniyaz, A., Guo, X., Cui, L., \u0026amp; Liu, L. (2025). Sensitive shrubland and unique urbanization patterns: the key to understanding emerging towns growth in arid regions. \u003cem\u003eEnvironmental Monitoring and Assessment\u003c/em\u003e. https://doi.org/10.1007/s10661-025-13693-4\u003c/li\u003e\n\u003cli\u003eAlmadini, A. M., \u0026amp; Hassaballa, A. A. (2019). Depicting changes in land surface cover at Al-Hassa oasis of Saudi Arabia using remote sensing and GIS techniques. \u003cem\u003ePLOS ONE\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(11), e0221115. https://doi.org/10.1371/journal.pone.0221115 \u003c/li\u003e\n\u003cli\u003eMcGarigal, K., \u0026amp; Marks, B. J. (1995). \u003cem\u003eFRAGSTATS: spatial pattern analysis program for quantifying landscape structure.\u003c/em\u003e https://doi.org/10.2737/PNW-GTR-351 \u003c/li\u003e\n\u003cli\u003eFaichia, C., Tong, Z., Zhang, J., Liu, X., Kazuva, E., Ullah, K., \u0026amp; Al-Shaibah, B. (2020). Using RS Data-Based CA\u0026ndash;Markov Model for Dynamic Simulation of Historical and Future LUCC in Vientiane, Laos. \u003cem\u003eSustainability\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(20), 8410. https://doi.org/10.3390/su12208410 \u003c/li\u003e\n\u003cli\u003eHamad, R., Balzter, H., \u0026amp; Kolo, K. (2018). Predicting Land Use/Land Cover Changes Using a CA-Markov Model under Two Different Scenarios. \u003cem\u003eSustainability\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(10), 3421. https://doi.org/10.3390/su10103421 \u003c/li\u003e\n\u003cli\u003eLi, C., Li, J., \u0026amp; Wu, J. (2013). Quantifying the speed, growth modes, and landscape pattern changes of urbanization: a hierarchical patch dynamics approach. \u003cem\u003eLandscape Ecology\u003c/em\u003e, \u003cem\u003e28\u003c/em\u003e(10), 1875\u0026ndash;1888. https://doi.org/10.1007/s10980-013-9933-6 \u003c/li\u003e\n\u003cli\u003eTang, X., Sheng, L., \u0026amp; Zhou, Y. (2022). Exploring complex urban growth and land use efficiency in China\u0026rsquo;s developed regions: implications for territorial spatial planning. \u003cem\u003eFrontiers of Earth Science\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(4), 1040\u0026ndash;1051. https://doi.org/10.1007/s11707-022-0973-6 \u003c/li\u003e\n\u003cli\u003eUnited Nations Human Settlements Programme (UN-Habitat), \u0026amp; Ministry of Municipal and Rural Affairs (MOMRA). (2019a). \u003cem\u003eJeddah City Profile\u003c/em\u003e. https://unhabitat.org/sites/default/files/2020/04/jeddah.pdf \u003c/li\u003e\n\u003cli\u003eUnited Nations Human Settlements Programme (UN-Habitat), \u0026amp; Ministry of Municipal and Rural Affairs (MOMRA). (2019b). \u003cem\u003eMadinah City Profile\u003c/em\u003e. https://unhabitat.org/sites/default/files/2020/04/madinah.pdf \u003c/li\u003e\n\u003cli\u003eUnited Nations Human Settlements Programme (UN-Habitat), \u0026amp; Ministry of Municipal and Rural Affairs (MOMRA). (2019c). \u003cem\u003eMakkah City Profile\u003c/em\u003e. https://unhabitat.org/sites/default/files/2020/03/makkah.pdf \u003c/li\u003e\n\u003cli\u003eUnited Nations Human Settlements Programme (UN-Habitat), \u0026amp; Ministry of Municipal and Rural Affairs (MOMRA). (2019d). \u003cem\u003eRiyadh City Profile\u003c/em\u003e. https://unhabitat.org/sites/default/files/2020/03/riyadh.pdf \u003c/li\u003e\n\u003cli\u003eVahid, R., \u0026amp; Aly, M. H. (2025). An Integrated CA\u0026ndash;Markov Modeling Framework for Forecasting Land Use and Land Cover Dynamics in Arkansas, USA. \u003cem\u003eGeomatics\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(4), 62. https://doi.org/10.3390/geomatics5040062 \u003c/li\u003e\n\u003cli\u003eZhang, Q., Liu, S., Niu, Y., Hu, Y., Li, L., Cai, E., Zhang, Y., \u0026amp; Zhao, M. (2025). Multi-Scenario Simulation of the Dynamic Relationship Between Land Use and Carbon Storage in the Urbanization Process: A Case Study of Zhengzhou, China. \u003cem\u003eLand\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(6), 1227. https://doi.org/10.3390/land14061227 \u003c/li\u003e\n\u003cli\u003eAbdelkarim, A. (2025). Monitoring and forecasting of land use/land cover (LULC) in Al-Hassa Oasis, Saudi Arabia based on the integration of the Cellular Automata (CA) and the Cellular Automata-Markov Model (CA-Markov). \u003cem\u003eGeology, Ecology, and Landscapes\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(1), 13\u0026ndash;44. https://doi.org/10.1080/24749508.2022.2163741 \u003c/li\u003e\n\u003cli\u003eEL-SHORBAGY, A.-M. (2020). Unplanned Settlements in Saudi Arabia. The Case of Al-Sabeel District, Jeddah. Journal of Settlements and Spatial Planning, 11(2), 81\u0026ndash;94. https://doi.org/10.24193/JSSP.2020.2.03 \u003c/li\u003e\n\u003cli\u003eAlqurashi, A. F. (2021). Quantification of Urban Patterns and Processes through Space and Time Using Remote Sensing Data: A Comparative Study between Three Saudi Arabian Cities. \u003cem\u003eSustainability\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(22), 12615. https://doi.org/10.3390/su132212615 \u003c/li\u003e\n\u003cli\u003eAlqurashi, A., Kumar, L., \u0026amp; Sinha, P. (2016). Urban Land Cover Change Modelling Using Time-Series Satellite Images: A Case Study of Urban Growth in Five Cities of Saudi Arabia. \u003cem\u003eRemote Sensing\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(10), 838. https://doi.org/10.3390/rs8100838 \u003c/li\u003e\n\u003cli\u003eAlmalki, K. A., Al Mosallam, M. S., Aldaajani, T. Z., \u0026amp; Al-Namazi, A. A. (2022). Landforms characterization of Saudi Arabia: Towards a geomorphological map. \u003cem\u003eInternational Journal of Applied Earth Observation and Geoinformation\u003c/em\u003e, \u003cem\u003e112\u003c/em\u003e, 102945. https://doi.org/10.1016/j.jag.2022.102945 \u003c/li\u003e\n\u003cli\u003eDahy, B., Al-Memari, M., Al-Gergawi, A., \u0026amp; Burt, J. A. (2024). Remote sensing of 50 years of coastal urbanization and environmental change in the Arabian Gulf: a systematic review. \u003cem\u003eFrontiers in Remote Sensing\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e. https://doi.org/10.3389/frsen.2024.1422910 \u003c/li\u003e\n\u003cli\u003eCongedo, L. (2021). Semi-Automatic Classification Plugin: A Python tool for the download and processing of remote sensing images in QGIS. \u003cem\u003eJournal of Open Source Software\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(64), 3172. https://doi.org/10.21105/joss.03172 \u003c/li\u003e\n\u003cli\u003eGaur, S., \u0026amp; Singh, R. (2023). A Comprehensive Review on Land Use/Land Cover (LULC) Change Modeling for Urban Development: Current Status and Future Prospects. \u003cem\u003eSustainability\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(2), 903. https://doi.org/10.3390/su15020903 \u003c/li\u003e\n\u003cli\u003eImam, A., Alhaddad, B., \u0026amp; Roca, J. (2016). REMOTE SENSING EFFICIENCY FOR URBAN ANALYSIS OF MECCA AND SURROUNDS. \u003cem\u003eISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences\u003c/em\u003e, \u003cem\u003eXLI-B8\u003c/em\u003e, 905\u0026ndash;910. https://doi.org/10.5194/isprsarchives-XLI-B8-905-2016 \u003c/li\u003e\n\u003cli\u003eLiu, Z., He, C., \u0026amp; Wu, J. (2016). General Spatiotemporal Patterns of Urbanization: An Examination of 16 World Cities. \u003cem\u003eSustainability\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(1), 41. https://doi.org/10.3390/su8010041 \u003c/li\u003e\n\u003cli\u003eSelmy, S. A. H., Kucher, D. E., Mozgeris, G., Moursy, A. R. A., Jimenez-Ballesta, R., Kucher, O. D., Fadl, M. E., \u0026amp; Mustafa, A. A. (2023). Detecting, Analyzing, and Predicting Land Use/Land Cover (LULC) Changes in Arid Regions Using Landsat Images, CA-Markov Hybrid Model, and GIS Techniques. \u003cem\u003eRemote Sensing\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(23), 5522. https://doi.org/10.3390/rs15235522 \u003c/li\u003e\n\u003cli\u003eTang, X., Liu, F., \u0026amp; Hu, X. (2024). Urban growth simulation and scenario projection for the arid regions using heuristic cellular automata. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(1), 21106. https://doi.org/10.1038/s41598-024-71709-4 \u003c/li\u003e\n\u003cli\u003eWang, W., Luan, W., Jing, H., Zhu, J., Zhang, K., Ma, Q., Zhang, S., \u0026amp; Liang, X. (2024). Quantitative Assessment of Urban Expansion Impact on Vegetation in the Lanzhou\u0026ndash;Xining Urban Agglomeration. \u003cem\u003eApplied Sciences\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(19), 8615. https://doi.org/10.3390/app14198615 \u003c/li\u003e\n\u003cli\u003eGeneral Authority for Statistics (GASTAT). (2023). Population Census Data of Saudi Arabia. https://www.stats.gov.sa/en/ \u003c/li\u003e\n\u003cli\u003eAl-Ansi, N. A., Uddin, B., Alhrabi, A., \u0026amp; Wahid, J. (2023). Impacts of Urban Growth Policy on Loss of Identity in Expanding Saudi Cities: A Case Study of Buraydah. \u003cem\u003eInternational Journal of Sustainable Development and Planning\u003c/em\u003e, \u003cem\u003e18\u003c/em\u003e(10), 2975\u0026ndash;2987. https://doi.org/10.18280/ijsdp.181001 \u003c/li\u003e\n\u003cli\u003eAlfakhrani, A., Alzamil, W. S., \u0026amp; Hanafi, W. H. H. (2025). Evaluating the Procedural Framework of Informal Settlement Upgrading Makkah City, Saudi Arabia: Challenges and Lessons for Sustainable Urban Renewal. \u003cem\u003eInternational Journal of Sustainable Development and Planning\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(7), 2817\u0026ndash;2827. https://doi.org/10.18280/ijsdp.200707 \u003c/li\u003e\n\u003cli\u003eAngel, S., Parent, J., Civco, D. L., Blei, A., \u0026amp; Potere, D. (2011). The dimensions of global urban expansion: Estimates and projections for all countries, 2000\u0026ndash;2050. \u003cem\u003eProgress in Planning\u003c/em\u003e, \u003cem\u003e75\u003c/em\u003e(2), 53\u0026ndash;107. https://doi.org/10.1016/j.progress.2011.04.001\u003cspan dir=\"RTL\"\u003e \u003c/span\u003e\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Urban expansion, Arid cities, Remote sensing, ANN–CA–Markov, Land consumption efficiency, Saudi Arabia","lastPublishedDoi":"10.21203/rs.3.rs-8447461/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8447461/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eUrbanisation in arid environments evolves through distinctive spatial processes shaped by abundant developable land, strong environmental constraints, and infrastructure-led planning. Despite the rapid growth of Saudi Arabia\u0026rsquo;s major cities, the long-term interplay between land-cover change, demographic dynamics, and future expansion pathways remains insufficiently understood. This study reviews a consistent four-decade record of urban growth (1984\u0026ndash;2024) for Riyadh, Jeddah, Makkah, and Madinah and examines it through an integrated framework combining multi-sensor remote sensing, demographic indicators, landscape fragmentation metrics, and ANN\u0026ndash;CA\u0026ndash;Markov modelling. Across all four cities, urban expansion follows a shared three-phase trajectory: an initial phase of fragmented and discontinuous growth (1984\u0026ndash;1994), a prolonged period of corridor-driven consolidation aligned with major infrastructure investments (1994\u0026ndash;2014), and a recent shift toward outward suburban diffusion (2014\u0026ndash;2024). While this temporal sequence is highly synchronised, its spatial expression differs markedly. Fragmentation metrics (Patch Density and Edge Density) identify the mid-2010s as a peak of morphological discontinuity, most pronounced in the basin-confined cities of Makkah and Madinah. Jeddah, constrained by its coastline, retains a predominantly linear growth form, whereas Riyadh expands multi-directionally across an unconstrained plateau. Coupling demographic change with land consumption shows that population growth alone cannot explain observed expansion patterns: Riyadh maintains relatively stable land-use efficiency, while Jeddah and Makkah experience phases of disproportionately land-intensive development. The ANN\u0026ndash;CA\u0026ndash;Markov simulations reproduce observed spatial patterns with high agreement (Kappa 0.61\u0026ndash;0.85) and project continued path-dependent expansion to 2034, with future growth largely reinforcing established corridors rather than generating new development fronts. By explicitly linking multi-decadal reconstruction, demographic efficiency, spatial fragmentation, and predictive modelling, this study advances a path-dependent interpretation of arid-city urbanisation and provides a transferable framework for understanding and anticipating urban growth in rapidly transforming desert environments.\u003c/p\u003e","manuscriptTitle":"Path-Dependent Urban Expansion in Arid Cities: A Multi-Decadal Remote Sensing Analysis and ANN–CA–Markov Modelling of Saudi Arabian Cities (1984–2034)","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-05 05:50:39","doi":"10.21203/rs.3.rs-8447461/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"baf2c8f5-538e-406c-8161-18af4dfa3b98","owner":[],"postedDate":"January 5th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":60199908,"name":"Urban Studies"}],"tags":[],"updatedAt":"2026-01-05T05:50:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-05 05:50:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8447461","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8447461","identity":"rs-8447461","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}