1. Introduction
1.1 Background and Problem Statement
Erbil, also known as Hawler, is one of the oldest cities that still exist today; in fact, the citadel of Erbil has been home to people for at least seven thousand years according to the UNESCO World Heritage Centre (2014). Erbil city, as well as the old part of the city, has changed almost beyond recognition since 2003, when regime change in Iraq occurred and Erbil expanded from an administrative centre to a large metropolitan area in which urban sprawl became the dominant growth pattern (Jawhar & Atakara, 2026). Erbil's city guidelines for development have changed dramatically since the inception of the master plan a few generations ago – even recently. As development continues, further modifications and extensions to those guidelines will follow. The city's growth is continuing to evolve, along with the addition of leapfrog (development from the city core to outer areas) and radial developments on the outskirts of the city (Abbas, 2019). The courtyard house (mal-i hewshidar), which has dominated in recent decades, has been supplanted at the building scale by two newer types – detached peripheral villas and, more recently, high-rise apartment buildings (Abbas, 2018, 2019; Jawhar, 2018). The creation of new districts in Erbil such as Empire World, Royal City, and English Village has brought a new design and lifestyle to a community whose previous domestic lifestyle was centred around the inward facing passively cooled home and now has also seen more changes in the heating and cooling process than before as well as greater changes in the environment from the small housing units to the larger districts. The present study quantifies both ends of this transformation, the macro-scale geospatial sprawl through Geographic Information System (GIS) analysis of multi-temporal Landsat imagery, and the micro-scale behavioural consequences through Partial Least Squares Structural Equation Modeling (PLS-SEM) analysis of household survey data, within a unified mixed-methods framework. Framed within urban economics, this transformation matters because it converts spatial growth into a recurring household energy burden and reshapes land and housing markets, raising the infrastructure and service costs of metropolitan expansion.
This double dislocation-macro-scale sprawl coupled with micro-scale typological inversion-is the empirical object of this study. Hot-arid and semi-arid cities across the Middle East and North Africa region are experiencing analogous transitions that now dominate the regional urbanisation literature. Tehran's four-stage demographic sequence from urbanisation through re-urbanisation (Talkhabi et al., 2022) illustrates how post-war rapid expansion couples with peripheral fragmentation; Dubai's aggressive built-up expansion against negligible green-area growth (Abulibdeh, 2021) shows how Gulf oil-economy urbanisation produces extreme typological concentration without attendant sustainability outcomes; and Dhaka's partial verification of the carbon benefits of transit-oriented development (TOD) for commuting trips (Ashik et al., 2022) illustrates the conditional nature of transit-oriented interventions in Global South megacities. Erbil's specificity lies elsewhere: it has undergone the sprawl-typology inversion without any functional mass-transit infrastructure, no urban rail, no bus rapid transit (BRT), only nascent fixed-route bus service. This infrastructural void means TOD discussion for Erbil is strictly forward-looking, and it shapes both the empirical scope of this study and the character of its policy recommendations, discussed in the upcoming Section 6. Three interlocking problems drive the research question. First, macro-scale sprawl has elongated commuting distances, entrenched car dependency, and consumed peri-urban agriculture, a process this study quantifies through GIS-based Land-Use/Land-Cover (LULC) analysis. Second, the shift to vertical living has imposed energy-intensive cooling regimes on a population whose practices were calibrated to horizontal typologies, a practice–infrastructure mismatch (Shove et al., 2012). Third, the cumulative environmental footprint remains unquantified at household scale: residential electricity demand has grown substantially over the past decade, yet no peer-reviewed study has attributed this growth to typological or behavioural drivers. The research question is therefore both spatial and causal: how has urban sprawl evolved geospatially over two decades, and which mechanisms link sprawl and typology to household energy demand and environmental awareness?
1.2 Knowledge Gaps, Contributions, and Objectives
The study was driven by three knowledge gaps: 1: a systematic search of the scientific literature on Iraqi cities' sustainability returned very few articles integrating the social dimensions in an empirical analysis using PLS-SEM and GIS-based sprawl analysis. 2: Although practice theory has been discussed in relation to variance-based SEM, especially in the fields of energy and urban studies (Abbas, 2019; Bin & Dowlatabadi, 2005; Fan et al., 2025; Glaeser, 2011; Shove et al., 2012; Strengers, 2013), there have been very few instances of its empirical application. 3: There is a lack of evidence linking macro geospatial change and micro-behavioural outcomes in hot-arid vertical living environments (Genovese et al., 2023). In hierarchical terms, Gap 1 is empirical (no integrated GIS–PLS-SEM evidence for Iraqi cities), Gap 2 is methodological (social-practice theory rarely operationalised in variance-based SEM), and Gap 3 is theoretical–contextual (no macro–micro linkage for hot-arid vertical living).
The study contributes on three registers. Methodologically, it integrates two complementary analytical traditions, GIS-based remote sensing and PLS-SEM household analysis, within a single mixed-methods framework, demonstrating how geospatial macro-scale sprawl indicators can be empirically linked to micro-scale lived experience. Theoretically, it integrates Shove–Pantzar–Watson practice theory (Shove et al., 2012) with Hair–Ringle–Sarstedt PLS-SEM (Hair et al., 2019) in a hot-arid MENA setting. Empirically, it provides one of the first sprawl-to-energy elasticity estimates for any city in the MENA region, grounded in both Landsat-derived geospatial evidence and household survey data.
The study pursues five objectives: (i) quantify Erbil's urban sprawl through multi-temporal GIS classification of Land-Use/Land-Cover for 2004, 2014, and 2024; (ii) operationalise USI as a formative construct via self-reported perceptual indicators; (iii) quantify the hypothesised USI → HRA → LS → EC → EDA pathway through PLS-SEM; (iv) test whether the effect of USI on EC is serially mediated through HRA and LS—the horizontal-to-vertical transition; and (v) derive policy recommendations integrating geospatial planning with behavioural change interventions. (Perceptual indicators are used deliberately because, under social-practice theory, it is residents’ perception of sprawl — not the geometric metric itself — that conditions household practice; the GIS strand supplies the corresponding objective measurement of the same phenomenon). Four research questions follow. RQ1: What is the magnitude and rate of Erbil's urban expansion 2004–2024 as measured through GIS LULC classification? RQ2: Does USI directly affect self-reported EC? RQ3: Does HRA mediate the USI–EC relationship, and in what direction? RQ4: To what extent does self-reported EC predict EDA? The gaps, objectives, and research questions align directly: Gap 1 motivates Objectives (i)–(ii) and RQ1; Gap 2 motivates Objectives (iii)–(iv) and RQ2–RQ3; Gap 3 motivates Objective (v) and RQ4.
2. Literature Review
The scholarship relevant to this study spans four overlapping domains. The sprawl-measurement and GIS tradition established compactness, density-gradient, mixed-use indicators, and Shannon's entropy as its operational toolkit (Ewing & Hamidi, 2015), with recent remote-sensing applications extending these measures to different cities in the region. Additional analyses have documented built-up expansion through Shannon's entropy and road-network indicators (Hamad, 2020). Further studies have tracked analogous dynamics in different cities in Iraq using integrated GIS, remote sensing, and Shannon's entropy (Hamad, 2020; Mohammed, 2013; Petrović & Lobanov, 2022). Research on Erbil as a rare case, and on peri-urban agricultural loss using GIS-based LULC classification, remains needed and is closely related to the present study (Al-Quraishi & Mustafa, 2024; Khzr et al., 2022; Rash et al., 2023). This body of research links sprawl metrics to metropolitan-scale carbon emissions, providing the empirical basis for hypothesising household-scale sprawl–energy pathways.
The residential-typology and high-rise literature develops along two lines. Foundational synthesis (Genovese et al., 2023) concentrates on the psychosocial consequences of vertical living: isolation, reduced neighbourhood attachment, and fear responses (Barr et al., 2011). This reframes high-rise residences as an arena for producing and contesting sustainable lifestyles, linking high-rise living to larger discussions around consumption and environmental identity. Research in multiple global cities has provided evidence of the typological inversion happening morphologically, but has not tested the behavioural impact of this shift; this study aims to fill that void (Fleischmann et al., 2021; Lee et al., 2017; Taubenböck et al., 2020; Taubenböck et al., 2018).
Work on household energy in hot-arid climates has largely concentrated on thermal-comfort engineering and envelope performance (Bin & Dowlatabadi, 2005). Practice-theoretic reframings (Shove et al., 2012), studies of continuity and change (Gram-Hanssen, 2011), and analyses of peak-demand dynamics (Strengers, 2013) recast energy as the aggregate by-product of routinised practices rather than isolated rational choice (Geels et al., 2015; Lasswell & Kaplan, 2017).
PLS-SEM methodological literature has undergone substantial refinement. The Hair et al. (2019) procedural guidelines provide the current reporting standard. Henseler, Ringle, and Sarstedt resolved discriminant-validity problems with the HTMT criterion (Henseler et al., 2015); Roemer, Schuberth, and Henseler further strengthened this with HTMT2 (Roemer et al., 2021). The inverse square-root method (Kock & Hadaya, 2018) updates sample-size determination; Kock's full-collinearity approach (Kock, 2015) provides a rigorous common-method-bias diagnostic; and Nitzl, Roldán, and Cepeda-Carrión's mediation protocol (Nitzl et al., 2016) completes the toolkit deployed in the upcoming Section 4.
3. Theoretical Framework and Hypotheses
3.1 Theoretical Foundations and Conceptual Model
Two theoretical streams converge in the conceptual model. The first is the sprawl-measurement and remote-sensing tradition advanced (Ewing & Hamidi, 2015; Khzr et al., 2022; Mohammed, 2013) as well as refined through Shannon's entropy and GIS-based applications (Alsharif et al., 2015; Shadman Roodposhti et al., 2016). This tradition treats sprawl as a multi-dimensional phenomenon captured by density, continuity, concentration, centrality, proximity, mixed use, and nuclearity, measurable through multi-temporal satellite imagery. Recent empirical analyses link sprawl metrics directly to residential energy demand: successive studies have established that every unit increase in objective sprawl indices is associated with measurable increases in per-capita residential electricity consumption (Akın & Erdoğan, 2020; Bhatta, 2010a; Kalogiannidis et al., 2025; Subasinghe et al., 2016). The second stream is the social-practice theory of energy consumption in which domestic energy use emerges as the by-product of routinised practices constituted by materials, competences, and meanings (Gram-Hanssen, 2011; Pohlmann, 2018; Rabiu & Jaeger-Erben, 2022; Strengers, 2010, 2013). TOD is deliberately excluded from the structural model because Erbil possesses no functional mass-transit infrastructure.
Integration of these streams rests on a causal logic seldom articulated explicitly in either literature. Macro-scale sprawl and the shift in residential typology are co-produced: land economics in decentralising cities shape vertical redevelopment in the contested core and detached villa construction on the periphery (Ahlfeldt & Pietrostefani, 2019; Glaeser, 2011). Each typology imposes its own material infrastructure, central AC, elevators, back-up generators, sealed glazing, and reconfigures the practices through which households meet everyday needs. Reconfigured practices, not typological materials alone, drive self-reported energy demand. The GIS analysis in the following sections quantifies the macro-scale process; the PLS-SEM analysis documents the micro-scale process; together they yield a unified mixed-methods account. This integrative logic is also read against the compact-city debate, which contends that densification can lower per-capita energy use; the present study treats that as a competing expectation rather than an assumption, and acknowledges that residential transformation in Erbil is additionally shaped by land prices, demographic pressure, and policy incentives beyond the constructs modelled here.
The resulting conceptual model is presented in Figure 1. The PLS-SEM model comprises four antecedent and mediating constructs (USI, HRA, LS, EC) plus one endogenous outcome (EDA). Formative versus reflective specification follows (Hair et al., 2019; Jarvis et al., 2003). Table 1 sets out each construct with its specification and self-reported indicators. USI and HRA are specified formatively because their indicators are defining characteristics that jointly constitute each construct rather than interchangeable reflections of it (Jarvis et al., 2003); formative validity is therefore assessed through indicator collinearity (VIF < 3.3) rather than through internal-consistency reliability alone.
Table 1: Latent constructs, specification, and self-reported indicators. Source: authors, based on the literature cited.
|
Construct
|
Specification
|
Self-Reported Indicators
|
Basis in Literature
|
|
Urban Sprawl Intensity (USI)
|
Formative
|
USI1: perceived distance from home to city centre; USI2: perceived daily commuting time; USI3: perceived peripheral-area dependence on private vehicle; USI4: self-reported neighbourhood density; USI5: perceived leapfrog fragmentation. All 5-pt Likert.
|
(Al-Sharif et al., 2014; Jiang et al., 2007; Lityński, 2021; Liu & Meng, 2020)
|
|
High-Rise Adoption (HRA)
|
Formative
|
HRA1: floor level; HRA2: years since moving from horizontal typology; HRA3: self-reported building-age category; HRA4: building-height category.
|
(Ala-Mantila et al., 2013; Feng et al., 2024; Lee & Braham, 2017)
|
|
Lifestyle Shift (LS)
|
Reflective
|
LS1: perceived change in reliance on mechanical cooling; LS2: change in daily car-trip frequency; LS3: loss of outdoor-courtyard activity; LS4: change in neighbourhood interaction; LS5: change in meal-preparation practices; LS6: change in natural light/ventilation use.
|
(Behnisch et al., 2022; Genovese et al., 2023; Smiraglia et al., 2021; Wu et al., 2021)
|
|
Energy Consumption (EC)
|
Reflective
|
EC1: perceived electricity-bill burden; EC2: self-reported daily AC operation hours; EC3: perceived thermal-discomfort compensation effort; EC4: perceived reliance on private generator; EC5: perceived household energy spending relative to income.
|
(Lyons et al., 2018; Navamuel et al., 2018; Popkin, 1999; Zhao & Zhang, 2018)
|
|
Environmental Degradation Awareness (EDA)
|
Reflective
|
EDA1: perceived local air-quality deterioration; EDA2: perceived UHI intensification; EDA3: perceived noise-pollution burden; EDA4: perceived loss of green/open space; EDA5: overall self-reported environmental-degradation concern.
|
(Frumkin, 2002; Ibimilua et al., 2020; Lyons et al., 2018; Ma et al., 2018)
|
Figure 1. The Sprawl Practice Energy Nexus as an Ontological Translation. Source: authors' analysis.
3.2 Hypothesis Development
Six hypotheses are proposed (Table 2). No TOD hypothesis appears because Erbil's lack of functional transit makes such a test empirically invalid. In brief, H1 follows from the commuting and cooling burdens of peripheral living; H2 from the land economics of decentralising cities; H3 from the practice-theoretic claim that material infrastructures reconfigure daily routines; H4 from routinised practices as the proximate driver of energy demand; H5 from the resulting serial logic; and H6 from experience-based awareness formation.
Table 2: Hypotheses (H1–H6) and literature basis. Source: authors, based on the literature cited.
|
ID
|
Hypothesis
|
Key Literature
|
|
H1
|
USI exerts a significant positive direct effect on self-reported EC among Erbil apartment households.
|
(Jawhar, 2018; Thanoon & Haykal, 2020)
|
|
H2
|
USI exerts a significant positive direct effect on HRA.
|
(Bhatta, 2010b; Travisi et al., 2010; Zhang, 2021; Zhao, 2010)
|
|
H3
|
HRA exerts a significant positive direct effect on LS.
|
(Brueckner & Largey, 2008; Ewing et al., 2003; Hamidi et al., 2018)
|
|
H4
|
LS exerts a significant positive direct effect on self-reported EC.
|
(Mouratidis, 2019; Santana et al., 2009)
|
|
H5
|
HRA and LS serially mediate the USI → EC relationship.
|
(Navamuel et al., 2018; Zhao & Zhang, 2018)
|
|
H6
|
Self-reported EC exerts a significant positive direct effect on EDA.
|
(Johnson, 2001; Rubiera-Morollón & Garrido-Yserte, 2020)
|
4. Materials and Methods
4.1 Research Design and Study Area
The study adopts a convergent parallel mixed-methods design (Creswell & Plano Clark, 2017) combining two methodologically distinct but conceptually integrated analytical strands: (i) a GIS analysis of multi-temporal LULC change to quantify the macro-scale sprawl phenomenon, and (ii) a PLS-SEM analysis of household survey data to quantify the micro-scale behavioural and environmental consequences. PLS-SEM is selected over covariance-based SEM (CB-SEM) per (Hair et al., 2019) because the model contains formative constructs (USI, HRA), the aim is prediction and theory development, Likert-based self-report data are expected to be non-normal, and feasible field-work sample sizes favour the distribution-free algorithm.
Erbil is selected for three reasons: (i) the sprawl–typology inversion has unfolded in an exceptionally compressed 2003–2024 window, yielding analytical leverage; (ii) the UNESCO-listed citadel core generates a policy-salient heritage–growth tension (UNESCO World Heritage Centre, 2014); and (iii) its semi-arid continental climate (Köppen BSh) shares morphological and thermal-load properties with Mashhad (Kalvová et al., 2003; Köppen, 2011; Naserikia et al., 2019) and Gulf cities (Elessawy, 2021; Fuccaro, 2001). Three sampling strata (or segments) were used for Erbil's residential areas: the inner core and citadel buffer, the intermediate mid-rise belt, and the outer residential periphery. The GIS component of the study involved analysing the spatial growth of Erbil during three different time periods (2004, 2014, 2024). There were six steps in the workflow.
4.2 GIS Data and Analytical Workflow
Step 1: Obtain the data. Obtain cloud-free Landsat satellite images for the three target years (2004, 2014, 2024) for the entire Erbil Metropolitan area (bounding box approximately 27 km E–W × 17 km N–S, with the UNESCO Erbil Citadel positioned approximately at the centre).
Step 2: Pre-process and Georeference the data. Project the Landsat scenes to WGS84 / UTM Zone 38N, perform atmospheric corrections, and trim the scenes to the study area and a surrounding buffer zone.
Step 3: Each Landsat scene was classified into three classes (Urban, Bareland, Green) in ArcGIS Pro using Supervised Maximum-Likelihood classification, adapted from the Anderson Level 1 scheme (Anderson et al., 1976). Training samples were collected from high-resolution reference imagery and used to train the Maximum-Likelihood Classifier; classification quality was subsequently assured through the multi-step qualitative validation protocol described below. Classification quality was assured through a four-step qualitative validation protocol rather than a formal confusion-matrix statistic, because independent high-resolution reference data for the archival epochs — particularly 2004 — are too scarce over Erbil to support a statistically defensible stratified-random validation sample. First, the deliberately coarse three-class scheme (Urban, Bareland, Green) maximises spectral separability: built-up, bare-soil, and vegetated surfaces occupy distinct, well-documented spectral regions in Landsat imagery, so the dominant source of LULC error — confusion among spectrally similar sub-classes — is designed out of the analysis. Second, each classified scene was visually cross-validated against the highest-resolution reference imagery available for its epoch, anchored on stable, dateable landmarks (the citadel core, Sami Abdulrahman Park, the airport, the major radial corridors) and on developments whose construction dates are public (e.g., Empire World, Royal City); misclassified clusters identified in this inspection were corrected through iterative re-training before the final maps were accepted. Third, cross-epoch logical-consistency screening confirmed that class transitions follow plausible urbanisation pathways (bareland → urban; green → urban or bareland) with no large-scale implausible reversals. Fourth, the resulting trajectory was triangulated against independent published analyses of Erbil and other Kurdistan-Region cities, which report the same direction and comparable magnitude of built-up expansion and entropy decline over the same period (Hamad, 2020; Khzr et al., 2022; Mohammed, 2013; Rash et al., 2023). Because the GIS strand serves as aggregate-level contextual corroboration for the household analysis rather than as the hypothesis-testing instrument, and because the headline findings — the near-tripling of the urban footprint and the monotonic entropy decline — are driven by the dominant urban/non-urban contrast, they are robust to the residual pixel-level error that a coarse three-class supervised classification can be expected to retain.
Step 4: Post-classification analysis in Python. Classified raster outputs were exported as RGB raster images and post-processed in Python using the NumPy and PIL libraries. A colour-threshold classification function (red: Urban; yellow: Bareland; green: Green) re-extracted the class membership of every pixel, enabling reproducible quantitative analysis independent of proprietary GIS software.
Step 5: Quantitative metrics. For each epoch the analysis computed: (a) class-wise area in square kilometres, scaled from pixel counts using the study-area total of approximately 132 km²; (b) class-wise percentage shares; (c) annualised conversion rates between consecutive epochs (km² / year and ha / year); and (d) Shannon's entropy of the LULC composition, H = -Σ p_i × ln(p_i), as a standard sprawl-diversity index in which lower values indicate dominance of a single class, typically the urban class in sprawl-affected cities (Akın & Erdoğan, 2020; Johnson, 2001; Kalogiannidis et al., 2025).
Step 6: Visualisation. Composite LULC maps (Figure 2), quantitative summary charts (Figure 3), and a transition-flow diagram (Figure 4) were rendered in Python's Matplotlib library to ensure publication-quality reproducibility.
4.3 Survey Design and PLS-SEM Procedure
Methodological declaration. All five PLS-SEM constructs (USI, HRA, LS, EC, EDA) are measured exclusively through self-reported survey instruments. No electricity meters, air-quality or temperature sensors, or utility-company records enter the PLS-SEM analysis. The GIS component (Section 4.2) provides independent macro-scale corroboration but does not feed directly into the PLS-SEM measurement model. This separation preserves the internal consistency of each analytical strand while permitting their joint interpretation in Section 5. EDA is operationalised as a perceived-awareness construct (a subjective attitude held by residents), not a directly measured environmental outcome. The two strands are integrated by triangulation at the interpretation level rather than by linking individual households to GIS pixels: the GIS results corroborate the macro-scale context, while the PLS-SEM results model the micro-scale behavioural pathway.
Survey items were adapted from the sources cited in Table 1. All items were rendered as 5-point Likert scales. The instrument was translated into Kurdish (Sorani) and Arabic using the back-translation method (Brislin, 1970), pilot-tested with 40 respondents (Cronbach's α, indicator loadings, and HTMT ratios guided item retention), and administered face-to-face on tablets by two trained native-language enumerators. The target population was apartment-dwelling households in Erbil residing on the third floor or above in buildings of six or more storeys, whose primary respondent had previously lived in a horizontal typology (courtyard or low-rise house / villa), an essential criterion for the LS retrospective-contrast design. Minimum N was estimated via the inverse square-root method (Kock & Hadaya, 2018), yielding N_min ≈ 155. The achieved sample of N = 460 valid responses substantially exceeds this minimum. Because the design is cross-sectional and relies on self-reported, perception-based measures, the relationships reported below are interpreted as predictive associations rather than established causal effects; retrospective items may also be subject to recall bias, mitigated by the salience of the horizontal-to-vertical move as a major life event. Recruitment followed the three sampling strata — the inner core and citadel buffer, the intermediate mid-rise belt, and the outer residential periphery — ensuring that the sample spans the full centre-to-periphery gradient over which sprawl intensity varies.
PLS-SEM analysis proceeds in three tiers using SmartPLS 4.0. Tier 1 (Preliminary Diagnostics) comprises univariate outlier screening, Harman's single-factor test, and full-collinearity VIF (< 3.3 required per (Kock, 2015)). Tier 2 (Measurement Model) uses the path-weighting scheme with 10,000 bootstrap resamples and the consistent PLSc algorithm; assessment follows (Hair et al., 2019) for indicator reliability (loadings ≥ 0.708), composite reliability (ρ_c ≥ 0.70), convergent validity (AVE ≥ 0.50), and discriminant validity (HTMT < 0.85; (Henseler et al., 2015)) with HTMT2 robustness check (Roemer et al., 2021). Tier 3 (Structural Model) reports R² and adjusted R² for each endogenous construct, Stone–Geisser Q² values via blindfolding (D = 7), and the specific indirect-effect test for H5 per (Nitzl et al., 2016).
5. Results and Analysis
5.1 Macro-Scale Urban Expansion and Landscape Change
The multi-temporal land-use/land-cover classification for Erbil in 2004, 2014, and 2024 is shown in Figure 2, created by performing a supervised classification of Landsat images using ArcGIS Pro, with post-processing completed using Python to prepare the data for display and analysis. Collectively, the three panels show a substantial amount of spatial transformation taking place at an accelerated rate. The initial landscape can best be described as predominately bareland with several areas of pure vegetation and, over the course of the past 20 years, these areas have been increasingly occupied by urban land cover, especially along the eastern, north-eastern, and southern axes, which extend outward from the citadel. The focus of this GIS mapping project is on the boundaries of the Erbil urban area (city) only, and not the entire Erbil Governorate. The differences in Erbil's urban growth over the 20-year period are depicted using different colour categories on the map. The details quantifying the urban growth of Erbil are presented in Table 3.
Table 3: LULC composition, share, and Shannon's entropy of Erbil, 2004–2024. Source: authors' calculation from supervised LULC classification.
|
Year
|
Urban (km²)
|
Urban %
|
Bareland (km²)
|
Bareland %
|
Green (km²)
|
Green %
|
Shannon H
|
|
2004
|
31.16
|
23.4
|
87.49
|
65.8
|
14.30
|
10.7
|
0.855
|
|
2014
|
63.99
|
48.9
|
62.11
|
47.5
|
4.72
|
3.6
|
0.823
|
|
2024
|
88.68
|
67.3
|
36.79
|
27.9
|
6.20
|
4.7
|
0.766
|
Three findings stand out. First, urban land cover almost tripled, increasing by 57.5 km² in absolute terms, a 184.6 % expansion. Second, this expansion came overwhelmingly at the expense of bareland (a loss of 50.7 km²) and, in the first decade, of vegetated cover (collapsing from 14.3 km² to 4.7 km²). Green space partially recovered by 2024 (to 6.2 km²), reflecting the introduction of formal urban parks such as Sami Abdulrahman Park, but remains far below its 2004 level. Third, Shannon's entropy declined monotonically from 0.855 to 0.766, a clear signal of progressive landscape homogenisation under urban dominance, consistent with entropy-based sprawl indicators developed for cities in the region (Hamad, 2020; Khzr et al., 2022; Mohammed, 2013).
Figure 2. Multi-temporal Land-Use / Land-Cover classification of Erbil, 2004 – 2024. Three panels depict the supervised LULC classification (Urban / Bareland / Green) for the reference epochs. Urban land cover expanded from 31.2 km² (23.4 %) in 2004 to 88.7 km² (67.3 %) in 2024, a 184.6 % increase. Source: authors' classification of Landsat imagery in ArcGIS Pro and Python (NumPy / PIL).
Figure 3 visualises the multi-dimensional sprawl trajectory. Panel (a) shows the stacked composition shift; panel (b) shows the crossover point around 2014 when urban surpassed bareland as the dominant class; panel (c) reports annualised conversion rates of 328 ha / year (2004–2014), 247 ha / year (2014–2024), and 288 ha / year (2004–2024 average); panel (d) confirms the monotonic decline in landscape entropy. The 2004–2014 decade exhibited the most rapid conversion, consistent with the immediate post-2003 demographic and investment boom; the 2014–2024 decade shows modest deceleration but continued substantial expansion.
Figure 4 confirms that the dominant transition is bareland to urban, with smaller but consequential losses from the already-thin green-cover stock. The conversion flow provides aggregate-level contextual corroboration for hypothesis H2 in the PLS-SEM analysis: urban sprawl intensity is closely connected to high-rise adoption because the macro-scale availability of converted bareland enables both peripheral villa development and core-area vertical redevelopment.
Figure 3. Geospatial quantification of urban sprawl in Erbil (2004 – 2024). (a) Land-cover composition by area in km²; (b) percentage-share trajectory of the three LULC classes; (c) annualised urban-sprawl rate in hectares per year; (d) Shannon's entropy of LULC diversity. Source: authors' analysis.
Figure 4. Land-cover conversion-flow diagram, Erbil 2004, 2014, and 2024. Each coloured ribbon represents the stock of a single LULC class across the three epochs; ribbon widths are proportional to area in km². The dominant transition is bareland to urban. Source: authors' analysis.
5.2 Sample Profile and Measurement Model
Fieldwork yielded 460 valid responses after listwise deletion, substantially exceeding N_min ≈ 155. All respondents met the eligibility requirements: 18 years or older, currently residing on the third floor or above in buildings of six or more storeys, and with prior experience of horizontal-typology dwelling. Table 4 summarises the sample's demographic, dwelling, and tenure characteristics.
Table 4: Sample demographic, dwelling, and tenure characteristics (N = 460). Source: authors' fieldwork. Mean household size = 5.6 persons; mean children under 18 = 2.3.
|
Variable
|
Category
|
Frequency (n)
|
Percentage (%)
|
|
Age
|
18–24
|
65
|
14.1
|
|
|
25–34
|
160
|
34.8
|
|
|
35–44
|
127
|
27.6
|
|
|
45–54
|
62
|
13.5
|
|
|
55–64
|
33
|
7.2
|
|
|
65+
|
13
|
2.8
|
|
Gender
|
Male
|
249
|
54.1
|
|
|
Female
|
200
|
43.5
|
|
|
Prefer not to say
|
11
|
2.4
|
|
Education
|
No formal
|
14
|
3.0
|
|
|
Primary
|
25
|
5.4
|
|
|
Secondary
|
87
|
18.9
|
|
|
Bachelor
|
190
|
41.3
|
|
|
Master
|
113
|
24.6
|
|
|
PhD
|
31
|
6.7
|
|
Monthly income (IQD)
|
Under 500 000
|
17
|
3.7
|
|
|
500 000 – 1 M
|
87
|
18.9
|
|
|
1 M – 2 M
|
135
|
29.3
|
|
|
2 M – 3 M
|
105
|
22.8
|
|
|
3 M – 5 M
|
65
|
14.1
|
|
|
Over 5 M
|
45
|
9.8
|
|
|
Prefer not to say
|
6
|
1.3
|
|
Apartment floor
|
3rd – 5th
|
129
|
28.0
|
|
|
6th – 10th
|
166
|
36.1
|
|
|
11th – 15th
|
89
|
19.3
|
|
|
16th – 20th
|
47
|
10.2
|
|
|
21st or higher
|
29
|
6.3
|
|
Building height
|
6–12 storeys
|
157
|
34.1
|
|
|
13–20 storeys
|
196
|
42.6
|
|
|
21+ storeys
|
107
|
23.3
|
|
Previous dwelling
|
Traditional courtyard house
|
214
|
46.5
|
|
|
Single-family villa
|
145
|
31.5
|
|
|
Townhouse / low-rise
|
74
|
16.1
|
|
|
Other
|
27
|
5.9
|
|
Tenure
|
Owner-occupier
|
225
|
48.9
|
|
|
Tenant
|
136
|
29.6
|
|
|
Family-owned (shared)
|
83
|
18.0
|
|
|
Other
|
16
|
3.5
|
The sample is well-suited to the lifestyle-transition focus of the study. Nearly half (46.5 %) had previously lived in a traditional courtyard dwelling, with a further 31.5 % from single-family villas, together establishing a substantial biographical contrast against current apartment life. The age distribution skews toward economically active middle-aged adults (62.4 % aged 25–44), and the gender split is approximately balanced (54.1 % male, 43.5 % female).
Table 5 presents the outer loadings of the 17 retained measurement indicators in the re-specified model (also shown in Figure 5); all exceed the recommended 0.708 threshold (minimum: EDA3 = 0.757). Eight original indicators were removed during re-specification — LS1 and LS6, which measured energy use rather than lifestyle and caused the EC–LS discriminant-validity overlap, together with the weak indicators HRA1, LS4, EC4, EC5, EDA1, and EDA2.
Figure 5. PLS-SEM measurement model with outer loadings and R² values for endogenous constructs (SmartPLS 4.0 output). Construct circles show R² values; arrows from constructs to indicator boxes show standardised outer loadings.
Table 5: Outer loadings of measurement indicators. Source: SmartPLS 4.0 output, N = 460.
|
Item
|
Loading
|
Item
|
Loading
|
Item
|
Loading
|
|
USI1
|
0.831
|
USI2
|
0.871
|
USI3
|
0.823
|
|
USI4
|
0.875
|
USI5
|
0.884
|
HRA2
|
0.872
|
|
HRA3
|
0.919
|
HRA4
|
0.850
|
LS2
|
0.858
|
|
LS3
|
0.887
|
LS5
|
0.881
|
EC1
|
0.883
|
|
EC2
|
0.894
|
EC3
|
0.850
|
EDA3
|
0.757
|
|
EDA4
|
0.936
|
EDA5
|
0.859
|
—
|
—
|
All five constructs satisfy the conventional thresholds for reliability and convergent validity (Table 6), with all five AVE values satisfying the 0.50 cut-off, consistent with (Hair et al., 2019). USI exhibits the strongest internal-consistency reliability (α = 0.910, ρ_c = 0.933), followed by HRA (α = 0.855, ρ_c = 0.912), while HRA shows the highest convergent validity (AVE = 0.775).
Table 6: Construct reliability and convergent validity. Source: SmartPLS 4.0 output. All constructs satisfy ρ_c > 0.70; AVE values range from 0.729 (EDA) to 0.775 (HRA), all satisfying the 0.50 threshold per Hair et al. (2019).
|
Construct
|
Cronbach's α
|
rho_a
|
ρ_c
|
AVE
|
|
USI
|
0.910
|
0.911
|
0.933
|
0.735
|
|
HRA
|
0.855
|
0.861
|
0.912
|
0.775
|
|
LS
|
0.849
|
0.859
|
0.908
|
0.766
|
|
EC
|
0.849
|
0.855
|
0.908
|
0.768
|
|
EDA
|
0.811
|
0.842
|
0.889
|
0.729
|
In the re-specified model, every construct pair satisfies the conservative HTMT threshold (maximum EC–LS = 0.825; Table 7), confirming that the five constructs are empirically distinct. The violations observed in the original specification (EC–LS = 0.976; EC–EDA = 0.938) stemmed from energy-loaded lifestyle items (LS1, LS6) and weak indicators; their removal resolves the overlap while preserving the construct set.
Table 7: Discriminant validity, Heterotrait–Monotrait (HTMT) ratio matrix. Source: SmartPLS 4.0 output. Threshold: HTMT < 0.85 (conservative).
|
|
EC
|
EDA
|
HRA
|
LS
|
USI
|
|
EC
|
—
|
|
|
|
|
|
EDA
|
0.605
|
—
|
|
|
|
|
HRA
|
0.769
|
0.633
|
—
|
|
|
|
LS
|
0.825
|
0.683
|
0.778
|
—
|
|
|
USI
|
0.788
|
0.591
|
0.799
|
0.745
|
—
|
5.3 Structural Model and Hypothesis Testing
The direct-path results are striking. H2 (USI → HRA) is highly significant (β = 0.707), confirming that perceived urban sprawl is strongly associated with the adoption of vertical living, consistent with the land-economics mechanism (Ahlfeldt & Pietrostefani, 2019; Glaeser, 2011). H3 (HRA → LS) is one of the strongest paths in the model (β = 0.673, t = 23.831), demonstrating that the typological transition fundamentally reconfigures daily practices. H4 (LS → EC) remains the strongest predictor of EC (β = 0.384, p < 0.001), confirming that Lifestyle Shift is the dominant household-level predictor of self-reported energy consumption. H1 (USI → EC) is also supported (β = 0.318, p < 0.001), and the small but significant direct path HRA → EC (β = 0.174, p = 0.001) indicates that high-rise adoption exerts a modest residual influence on energy consumption beyond the indirect route through Lifestyle Shift. H6 (EC → EDA) demonstrates that households' direct experience of energy use raises their awareness of environmental degradation, with a significant effect (β = 0.207, p = 0.001). The least significant direct pathway is USI → EDA (β = 0.176, p = 0.016), indicating that urban sprawl influences environmental-degradation awareness through a modest direct path alongside stronger behavioural mediators. R² values are HRA = 0.500, LS = 0.453, EC = 0.609, and EDA = 0.344, all in the moderate-to-substantial range (Hair et al., 2019). The model explains 60.9 % of the variance in self-reported energy consumption. Table 9 reports the bootstrap-based mediation analysis following (Nitzl et al., 2016). Figure 6 presents the bootstrapped structural model, with bootstrap p-values on each path. Table 8 reports the formal hypothesis tests.
Figure 6. PLS-SEM structural model with bootstrap p-values from 10,000 resamples (β reported in Table 8). R² for endogenous constructs: HRA = 0.500, LS = 0.453, EC = 0.609, EDA = 0.344.
Table 8: Direct-path hypothesis-testing results. Source: SmartPLS 4.0, 10,000 bootstrap resamples, two-tailed, 5 % significance.
|
Hypothesis / Path
|
β
|
SD
|
t
|
p
|
95 % CI
|
Decision
|
|
H2: USI → HRA
|
0.707
|
0.033
|
21.296
|
< 0.001
|
[0.638, 0.769]
|
Supported
|
|
H3: HRA → LS
|
0.673
|
0.028
|
23.831
|
< 0.001
|
[0.616, 0.728]
|
Supported
|
|
H4: LS → EC
|
0.384
|
0.047
|
8.261
|
< 0.001
|
[0.291, 0.474]
|
Supported
|
|
H1: USI → EC
|
0.318
|
0.057
|
5.629
|
< 0.001
|
[0.205, 0.425]
|
Supported
|
|
HRA → EC
|
0.174
|
0.051
|
3.442
|
0.001
|
[0.081, 0.277]
|
Supported
|
|
HRA → EDA
|
0.276
|
0.069
|
4.010
|
< 0.001
|
[0.130, 0.401]
|
Supported
|
|
H6: EC → EDA
|
0.207
|
0.064
|
3.259
|
0.001
|
[0.079, 0.332]
|
Supported
|
|
USI → EDA
|
0.176
|
0.073
|
2.408
|
0.016
|
[0.039, 0.325]
|
Supported
|
Table 9: Indirect-effect and mediation-analysis results. Source: SmartPLS 4.0 bootstrapping output, 10,000 resamples. H5 is the serial USI → HRA → LS → EC pathway.
|
Indirect Path
|
β
|
SD
|
t
|
p
|
95 % CI
|
|
USI → HRA → LS
|
0.476
|
0.035
|
13.659
|
< 0.001
|
[0.407, 0.543]
|
|
HRA → LS → EC
|
0.259
|
0.033
|
7.933
|
< 0.001
|
[0.194, 0.324]
|
|
USI → HRA → LS → EC
|
0.183
|
0.026
|
7.050
|
< 0.001
|
[0.133, 0.238]
|
|
LS → EC → EDA
|
0.080
|
0.029
|
2.755
|
0.006
|
[0.027, 0.141]
|
|
HRA → LS → EC → EDA
|
0.054
|
0.019
|
2.779
|
0.005
|
[0.019, 0.094]
|
|
USI → HRA → EDA
|
0.195
|
0.047
|
4.132
|
< 0.001
|
[0.095, 0.281]
|
|
USI → HRA → LS → EC → EDA
|
0.038
|
0.014
|
2.727
|
0.006
|
[0.013, 0.068]
|
|
USI → EC → EDA
|
0.066
|
0.022
|
2.935
|
0.003
|
[0.024, 0.111]
|
|
USI → HRA → EC
|
0.123
|
0.038
|
3.251
|
0.001
|
[0.055, 0.205]
|
|
HRA → EC → EDA
|
0.036
|
0.015
|
2.431
|
0.015
|
[0.012, 0.069]
|
|
USI → HRA → EC → EDA
|
0.025
|
0.011
|
2.353
|
0.019
|
[0.008, 0.050]
|
H5, the central mediation hypothesis, is strongly supported. The serial indirect effect USI → HRA → LS → EC yields β = 0.183, t = 7.050, p < 0.001, with 95 % CI [0.133, 0.238] firmly excluding zero. This is the largest of the indirect pathways from USI to EC and confirms that the effect of urban sprawl on household energy consumption operates substantially through the serial mechanism of high-rise adoption reshaping daily lifestyles. The four-step mediation USI → HRA → LS → EC → EDA is also significant (β = 0.038, p = 0.006), reaching all the way to environmental awareness. Table 10 reports Stone–Geisser Q² values computed via blindfolding with omission distance D = 7. All four endogenous constructs achieve Q² > 0.20, confirming moderate-to-substantial predictive relevance per (Hair et al., 2019).
Energy Consumption shows the strongest predictive relevance (Q² = 0.461), followed by High-Rise Adoption (0.384) and Lifestyle Shift (0.339). EDA achieves an acceptable Q² = 0.234. The USI Q² = 0 reflects its exogenous status in the structural model. Model fit for the re-specified model is acceptable (saturated-model SRMR = 0.063, below the 0.08 threshold; NFI = 0.839).
Table 10: Stone–Geisser Q² values for endogenous constructs. Source: SmartPLS 4.0 blindfolding output, omission distance D = 7.
|
Construct
|
SSO
|
SSE
|
Q² = 1 − SSE/SSO
|
Interpretation
|
|
EC
|
—
|
—
|
0.461
|
Substantial predictive relevance
|
|
LS
|
—
|
—
|
0.339
|
Moderate predictive relevance
|
|
HRA
|
—
|
—
|
0.384
|
Substantial predictive relevance
|
|
EDA
|
—
|
—
|
0.234
|
Moderate predictive relevance
|
|
USI
|
—
|
—
|
0.000
|
Exogenous (no Q² applicable)
|
The dual empirical apparatus is internally consistent and mutually reinforcing. The GIS analysis documents the macro-scale phenomenon, 184.6 % urban expansion over 2004–2024, monotonic decline in landscape entropy, and the bareland → urban transition flow, providing the geospatial substrate within which the PLS-SEM household sample is situated. The PLS-SEM analysis confirms that this macro-scale transformation has measurable micro-scale consequences: USI strongly predicts HRA, HRA predicts LS, LS strongly predicts EC, and EC predicts EDA, with the serial USI → HRA → LS → EC pathway being highly significant. The integrated framework therefore provides one of the first multi-method empirical sprawl-to-energy models for any city in the region.
6. Discussion and Conclusion
6.1 Principal Findings
Four principal findings emerge from the integrated GIS and PLS-SEM apparatus. First, the GIS analysis (Section 5.1) shows that Erbil's urban footprint expanded by 184.6 % between 2004 and 2024, accompanied by a 58 % decline in bareland and a 57 % decline in green space. Green space partially recovered by 2024, driven by the development of Sami Abdulrahman Park and related municipal interventions, though it remains far below its 2004 level. Shannon's entropy has decreased from 0.855 to 0.766 over this same period, which also provides an additional method of measuring urban sprawl using an entropy-based sprawl signature that is comparable with prior studies of Soran (Hamad, 2020) and Duhok (Mohammed, 2013).
Second, the PLS-SEM analysis confirms that this macro-scale geospatial transformation has measurable behavioural consequences: USI strongly predicts HRA (β = 0.707, p < 0.001), consistent with the land-economics mechanism (Glaeser, 2011) and the density-synthesis evidence (Ahlfeldt & Pietrostefani, 2019). For Erbil specifically, this means that the high-rise developments now dotting the city, Empire World, Royal City, English Village, and their successors, are not a spontaneous architectural preference but a structural consequence of sprawl-driven land economics, empirically corroborated by the documented conversion of bareland into urban land cover.
Third, HRA produces a measurable Lifestyle Shift (β = 0.673, p < 0.001), one of the strongest paths in the model, materialising the practice-theoretic expectation that material infrastructures reshape competences and meanings (Shove et al., 2012). The observed effect size is consistent with the foundational description (Genovese et al., 2023) of how high-rise environments restructure daily routines through material constraints (elevators, sealed envelopes, absent ground-level interfaces) that cascade into habit and norm reconfiguration. Fourth, LS predicts self-reported EC (β = 0.384, p < 0.001) and substantially mediates the HRA → EC relationship, providing the first PLS-SEM-based empirical confirmation of practice-theoretic mediation in a hot-arid MENA setting. The serial mediation USI → HRA → LS → EC (β = 0.183) is the most analytically consequential finding of the study and the one with the sharpest policy implications. EC subsequently predicts Environmental Degradation Awareness (β = 0.207), suggesting that households' direct experience of high bills and thermal discomfort correlates with heightened perception of broader environmental problems, a result consistent with the perception–behaviour literature (Ewing & Hamidi, 2015; Genovese et al., 2023; Subasinghe et al., 2016) and indicative of a latent constituency for demand-side environmental policy among Erbil's apartment-dwelling population (Abbas, 2018; UNESCO World Heritage Centre, 2014). We note that household-level factors such as income, education, household size, dwelling age, and tenure are plausible alternative explanations for the energy and awareness outcomes; controlling for these in a respondent-level dataset is recommended, and the directional reading here is therefore offered as a predictive rather than a definitive causal account.
Erbil presently lacks any functional TOD infrastructure: no urban rail, no BRT, only nascent fixed-route buses, making TOD methodologically inadmissible as an empirical PLS-SEM construct. It is nonetheless retained here as a theoretical policy benchmark against which Erbil's business-as-usual (BAU) trajectory can be evaluated.
International evidence for TOD's potential is well established. Ashik et al. (2022) found significant transport-CO₂ reductions for work and school trips in Dhaka's TOD neighbourhoods. Fan et al. (2025) identify 2,000–2,500 persons/km² density and 0.8–0.9 land-use-mix entropy as carbon-neutrality accelerators across 340 Chinese cities. Ibraeva et al. (2020) conclude that TOD's energy and emission benefits are conditional on integrated land-use–transit planning. Erbil's BAU trajectory replicates the Tehran signature documented by (Bokaie et al., 2016; Talkhabi et al., 2022), demographic inflation outpaced by physical expansion, with attendant urban-heat-island intensification. The sprawl → HRA → LS → EC chain demonstrated in Section 5 suggests that without TOD principles introduced before motorisation calcifies, aggregate residential energy will continue rising with sprawl intensity, regardless of building-envelope improvements.
6.2 Comparative Context and Contributions
Placing the Erbil findings in comparative context clarifies both their specificity and their transferability. Tehran offers the closest demographic and climatic analogue. The four-stage demographic analysis of the Tehran Metropolitan Region (Talkhabi et al., 2022) documents exactly the urban-growth–sprawl decoupling Erbil now exhibits, and the LST–LULC analysis (Bokaie et al., 2016) demonstrates that such decoupling produces measurable urban-heat-island consequences. The Erbil case extends this evidence base by modelling, via PLS-SEM, the behavioural-mediating pathway that Tehran studies have hypothesised but not empirically tested.
Dhaka, analysed by Ashik et al. (2022), provides the most rigorous empirical baseline for benchmarking Erbil's hypothetical future TOD performance. Its finding that TOD reduces transport-CO₂ for work and school trips but not for non-work trips transfers directly to Erbil, where commuting is expected to dominate aggregate vehicle-kilometres. The implication is that TOD's carbon benefits in Erbil, were the infrastructure introduced, would likely be front-loaded on commuting reduction rather than spread evenly across the trip portfolio, reinforcing the case for corridor-specific rather than network-wide implementation.
Gulf cities — particularly as analysed by Abulibdeh in Qatar and across eight arid and semi-arid GCC cities (Abulibdeh, 2021; Patel et al., 2024) — establish the upper bound of typological and thermal stress that Erbil could approach under unchecked BAU continuation.
The rapid pace of built-up growth in Dubai compared to the almost nonexistent growth of green areas is the model that Erbil now appears to be following as it develops outwards. The results from studying Erbil thus place it importantly in between: the behavioural patterns are like those of Tehran; the potential for future transit-oriented development would be like what has occurred in Dhaka; and thermal-stress risks are already similar to those in the Gulf. These comparative attributes yield policy implications specific to each location rather than generic ones.
This research has three main contributions, each of which extends beyond its empirical Erbil claim. The first (methodological) contribution is the successful operational integration of macro-scale GIS-based sprawl measurement with micro-scale PLS-SEM-based behavioural analysis within one convergent parallel mixed-methods research design (Creswell & Plano Clark, 2017). It generates two complementary data streams in which both bodies of evidence support and contextualise one another without the methodological problems of forcing them together into a single aggregate model; the second indicates that practice theory is (theoretically) quantitatively operable, not just evoked in variance-based SEM when the practice-shift construct is defined as a retrospectively explicit contrast – as evident in Table 9, where the statistically significant indirect effect through LS (H5) provides direct support for the assertion (Shove et al., 2012) that materials, competences, and meanings are the mechanism by which built-infrastructure change converts into energy-behaviour change. The previous literature provides this evidence as a theoretical assertion but rarely has it been quantitatively verified.
The third (empirical) contribution is that the study is among the first to present a sprawl-to-energy elasticity estimate, using both Landsat-derived geospatial data and survey data from households. The size of the measured effects allows empirical rather than anecdotal comparison with Tehran (Talkhabi et al., 2022), Gulf cities (Abulibdeh, 2021; Patel et al., 2024), and Dhaka (Ashik et al., 2022), and provides a reproducible benchmark for successor studies in Sulaymaniyah, Duhok, and comparable MENA settings.
6.3 Policy Recommendations, Limitations, and Conclusions
Five propositions follow from the findings, adapted to Erbil's current infrastructural realities:
(1) Prepare for TOD in the next master-plan cycle. The 5Ds, density, diversity, design, destination accessibility, distance to transit, should become binding design criteria for the three radial corridors (120 m Street, Kirkuk Road, Dohuk Road) before motorisation calcifies retrofit costs.
(2) Institute a Residential Typology Impact Assessment. High-rise proposals above ten storeys should submit a standardised lifestyle-and-energy forecast calibrated against this study's H1–H6 coefficients, with mandatory consideration of the documented USI → HRA → LS → EC pathway.
(3) Codify the Citadel buffer zone. Strict height and use-mix codes aligned with the UNESCO revitalisation agreement (Abbas, 2018, 2019) will pre-empt the fate of other MENA heritage cores.
(4) Mandate residential energy disclosure. U-values, AC-load projections, and post-occupancy lifestyle profiles at point of sale or lease will create demand-side pressure for efficient stock, addressing the LS → EC pathway directly.
(5) Build longitudinal data infrastructure combining GIS and survey data. The KRG Ministry of Municipalities should support successor studies that triangulate the Landsat-derived LULC time series with meter-level energy data and household behavioural surveys.
There are six limitations that should be acknowledged. First, because the PLS-SEM was designed as a cross-sectional analysis, strict counterfactual inferences cannot be made with these data; directional claims are limited by theoretical ordering; however, the longitudinal GIS analyses at the macro-scale partially compensate by establishing temporal precedence at the macro scale. Second, all PLS-SEM constructs rely on self-reporting; meter-level validation will be the focus of future research. Third, the original specification exhibited an EC–LS HTMT of 0.976; the re-specified model resolves this (EC–LS = 0.825), though future replications should continue to monitor the conceptual proximity of lifestyle and energy measures. Fourth, as discussed above, the effects of TOD are not available for testing through empirical means (Ashik et al., 2022; Fan et al., 2025). Fifth, findings are specific to Erbil's semi-arid climate and KRG governance context; cross-MENA generalisability requires replication. Sixth, the LULC classification is validated qualitatively — through visual cross-checks against epoch-matched reference imagery, logical-consistency screening of class transitions, and triangulation with independent published studies — rather than through a formal confusion-matrix statistic, because reference data sufficient for a stratified-random accuracy sample are unavailable for the archival epochs; the coarse three-class scheme and the corroborative (rather than inferential) role of the GIS strand bound the impact of residual classification error, and successor studies with contemporary ground-truthing are planned.
Four extensions are prioritised: (i) triangulation of self-reported EC with meter-level or utility-company data; (ii) multi-group SEM comparing apartment, villa, and courtyard residents to formalise typology differences, integrating their GIS-mapped neighbourhoods; (iii) a longitudinal panel tracking households across future TOD pilot interventions, combining annual GIS LULC updates with repeated household surveys; and (iv) extension to Sulaymaniyah, Duhok, and comparable MENA cities, replicating the integrated GIS and PLS-SEM methodology.
Erbil's post-2003 transformation is neither inevitable nor irreversible. The evidence assembled, 184.6 % expansion of urban land cover documented through multi-temporal GIS classification, and a fully specified PLS-SEM behavioural pathway from urban sprawl through high-rise adoption and lifestyle reconfiguration to household energy consumption and environmental awareness, indicates that the city's trajectory is shaped at three scales: macro-scale master-plan choices visible in the Landsat record, meso-scale typology approvals expressed in the new high-rise stock, and micro-scale household practices captured in the survey instrument. The PLS-SEM results reveal these three scales to be closely linked. A sustainable Erbil requires coherent action at all three scales. The citadel at the city's heart, now under renewed UNESCO protection, stands as a structural reminder that the city's future, like its history, is being authored today by choices still reversible. Theoretically, the study shows that the practice–infrastructure mismatch — vertical material infrastructures meeting practices calibrated to horizontal living — is the mechanism that converts macro-scale sprawl into micro-scale energy demand, extending social-practice theory into a quantitative, variance-based setting and linking it to the urban economics of residential transformation.
Acknowledgements
The authors sincerely thank Hourakhsh Ahmad Nia, Roksaneh Rahbarianyazd (Alanya University), Murat Yakar (Mersin University), and Islam Hamdi Elghonaimy (University of Bahrain) for their generous support and encouragement.
Funding
No specific funding was received for this research.
Conflicts of Interest
The authors declare that there is no conflict of interest.
Data Availability Statement
Data are available upon request.
Institutional Review Board Statement
This study involved an anonymous, voluntary household survey of adult respondents, conducted in accordance with the ethical principles of the Declaration of Helsinki. Informed consent was obtained from all participants before participation, no personally identifying information was collected, and respondents were free to withdraw at any time without penalty. Formal ethical review and approval were not required for this non-interventional perception survey in accordance with the applicable institutional and national regulations.
CRediT Author Statement
Shad Sherzad Jawhar: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Visualization, Writing – original draft. Sarko Hassan Sleman: Investigation, Resources, Data curation, Writing – review & editing. Cemil Atakara: Supervision, Validation, Project administration, Writing – review & editing. All authors have reviewed and approved the final version of the manuscript.
References
Abbas, A. (2018). The uniqueness of Erbil Citadel buffer zone as compared to the general theory of buffer zones. Proceedings of the 6th International Conference on Heritage and Sustainable Development. https://documentserver.uhasselt.be/handle/1942/28132
Abbas, A. (2019). A proposal for a methodology for the adaptive reuse of traditional buildings in the buffer zone of Erbil Citadel [Doctoral dissertation, Hasselt University]. https://documentserver.uhasselt.be/handle/1942/28252
Abulibdeh, A. (2021). Analysis of urban heat island characteristics and mitigation strategies for eight arid and semi-arid gulf region cities. Environmental Earth Sciences, 80(7), 259. https://doi.org/10.1007/s12665-021-09540-7
Ahlfeldt, G. M., & Pietrostefani, E. (2019). The economic effects of density: A synthesis. Journal of Urban Economics, 111, 93–107. https://doi.org/10.1016/j.jue.2019.04.006
Akın, A., & Erdoğan, M. A. (2020). Analysing temporal and spatial urban sprawl change of Bursa city using landscape metrics and remote sensing. Modeling Earth Systems and Environment, 6(3), 1331–1343. https://doi.org/10.1007/s40808-020-00766-1
Al-Quraishi, A. M. F., & Mustafa, Y. T. (Eds.). (2024). Natural resources deterioration in MENA region: Land degradation, soil erosion, and desertification. Springer. https://doi.org/10.1007/978-3-031-58315-5
Al-Sharif, A. A. A., Pradhan, B., Shafri, H. Z. M., & Mansor, S. (2014). Quantitative analysis of urban sprawl in Tripoli using Pearson’s chi-square statistics and urban expansion intensity index. IOP Conference Series: Earth and Environmental Science, 20, 012006. https://doi.org/10.1088/1755-1315/20/1/012006
Ala-Mantila, S., Heinonen, J., & Junnila, S. (2013). Greenhouse gas implications of urban sprawl in the Helsinki Metropolitan Area. Sustainability, 5(10), 4461–4478. https://doi.org/10.3390/su5104461
Alsharif, A. A. A., Pradhan, B., Mansor, S., & Shafri, H. Z. M. (2015). Urban expansion assessment by using remotely sensed data and the relative Shannon entropy model in GIS: A case study of Tripoli, Libya. Theoretical and Empirical Researches in Urban Management, 10(1), 55–71. https://um.ase.ro/no101/5.pdf
Anderson, J. R., Hardy, E. E., Roach, J. T., & Witmer, R. E. (1976). A land use and land cover classification system for use with remote sensor data (U.S. Geological Survey Professional Paper 964). U.S. Government Printing Office. https://doi.org/10.3133/pp964
Ashik, F. R., Rahman, M. H., & Kamruzzaman, M. (2022). Investigating the impacts of transit-oriented development on transport-related CO2 emissions. Transportation Research Part D: Transport and Environment, 105, 103227. https://doi.org/10.1016/j.trd.2022.103227
Barr, S., Shaw, G., & Coles, T. (2011). Sustainable lifestyles: Sites, practices, and policy. Environment and Planning A: Economy and Space, 43(12), 3011–3029. https://doi.org/10.1068/a43529
Behnisch, M., Krüger, T., & Jaeger, J. A. G. (2022). Rapid rise in urban sprawl: Global hotspots and trends since 1990. PLOS Sustainability and Transformation, 1(11), e0000034. https://doi.org/10.1371/journal.pstr.0000034
Bhatta, B. (2010a). Analysis of urban growth and sprawl from remote sensing data. Springer. https://doi.org/10.1007/978-3-642-05299-6
Bhatta, B. (2010b). Causes and consequences of urban growth and sprawl. In Analysis of urban growth and sprawl from remote sensing data (pp. 17–36). Springer. https://doi.org/10.1007/978-3-642-05299-6_2
Bin, S., & Dowlatabadi, H. (2005). Consumer lifestyle approach to US energy use and the related CO2 emissions. Energy Policy, 33(2), 197–208. https://doi.org/10.1016/s0301-4215(03)00210-6
Bokaie, M., Zarkesh, M. K., Arasteh, P. D., & Hosseini, A. (2016). Assessment of urban heat island based on the relationship between land surface temperature and land use/land cover in Tehran. Sustainable Cities and Society, 23, 94–104. https://doi.org/10.1016/j.scs.2016.03.009
Brislin, R. W. (1970). Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology, 1(3), 185–216. https://doi.org/10.1177/135910457000100301
Brueckner, J. K., & Largey, A. G. (2008). Social interaction and urban sprawl. Journal of Urban Economics, 64(1), 18–34. https://doi.org/10.1016/j.jue.2007.08.002
Creswell, J. W., & Plano Clark, V. L. (2017). Designing and conducting mixed methods research (3rd ed.). SAGE Publications.
Elessawy, F. M. (2021). The abnormal population growth and urban sprawl of an Arabian Gulf city: The case of Abu Dhabi City. Open Journal of Social Sciences, 9(2), 245–269. https://doi.org/10.4236/jss.2021.92017
Ewing, R., & Hamidi, S. (2015). Compactness versus sprawl: A review of recent evidence from the United States. Journal of Planning Literature, 30(4), 413–432. https://doi.org/10.1177/0885412215595439
Ewing, R., Schmid, T., Killingsworth, R., Zlot, A., & Raudenbush, S. (2003). Relationship between urban sprawl and physical activity, obesity, and morbidity. American Journal of Health Promotion, 18(1), 47–57. https://doi.org/10.4278/0890-1171-18.1.47
Fan, T., Ren, Y., & Chapman, A. (2025). Unveiling the carbon neutrality pathways of compact cities: A simulation-based scenario analysis from China. Humanities and Social Sciences Communications, 12(1), 1205. https://doi.org/10.1057/s41599-025-05545-w
Feng, W., Chen, J., Yang, Y., Gao, W., Zhao, Q., Xing, H., & Yu, S. (2024). The impact of building morphology on energy use intensity of high-rise residential clusters: A case study of Hangzhou, China. Buildings, 14(7), 2245. https://doi.org/10.3390/buildings14072245
Fleischmann, M., Romice, O., & Porta, S. (2021). Measuring urban form: Overcoming terminological inconsistencies for a quantitative and comprehensive morphologic analysis of cities. Environment and Planning B: Urban Analytics and City Science, 48(8), 2133–2150. https://doi.org/10.1177/2399808320910444
Frumkin, H. (2002). Urban sprawl and public health. Public Health Reports, 117(3), 201–217. https://doi.org/10.1093/phr/117.3.201
Fuccaro, N. (2001). Visions of the city: Urban studies on the Gulf. Middle East Studies Association Bulletin, 35(2), 175–187. https://doi.org/10.1017/s0026318400043339
Geels, F. W., McMeekin, A., Mylan, J., & Southerton, D. (2015). A critical appraisal of Sustainable Consumption and Production research: The reformist, revolutionary and reconfiguration positions. Global Environmental Change, 34, 1–12. https://doi.org/10.1016/j.gloenvcha.2015.04.013
Genovese, D., Candiloro, S., D’Anna, A., Dettori, M., Restivo, V., Amodio, E., & Casuccio, A. (2023). Urban sprawl and health: A review of the scientific literature. Environmental Research Letters, 18(8), 083004. https://doi.org/10.1088/1748-9326/ace986
Glaeser, E. (2011). Cities, productivity, and quality of life. Science, 333(6042), 592–594. https://doi.org/10.1126/science.1209264
Gram-Hanssen, K. (2011). Understanding change and continuity in residential energy consumption. Journal of Consumer Culture, 11(1), 61–78. https://doi.org/10.1177/1469540510391725
Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. European Business Review, 31(1), 2–24. https://doi.org/10.1108/EBR-11-2018-0203
Hamad, R. (2020). A remote sensing and GIS-based analysis of urban sprawl in Soran District, Iraqi Kurdistan. SN Applied Sciences, 2(1), 24. https://doi.org/10.1007/s42452-019-1806-4
Hamidi, S., Ewing, R., Tatalovich, Z., Grace, J. B., & Berrigan, D. (2018). Associations between urban sprawl and life expectancy in the United States. International Journal of Environmental Research and Public Health, 15(5), 861. https://doi.org/10.3390/ijerph15050861
Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science, 43(1), 115–135. https://doi.org/10.1007/s11747-014-0403-8
Ibimilua, A. F., Ibimilua, F. O., & Ogundare, B. A. (2020). Urban sprawl: Environmental consequence of rapid urban expansion. Malaysian Journal of Social Sciences and Humanities (MJSSH), 5(6), 110–118. https://doi.org/10.47405/mjssh.v5i6.411
Ibraeva, A., Correia, G. H. de A., Silva, C., & Antunes, A. P. (2020). Transit-oriented development: A review of research achievements and challenges. Transportation Research Part A: Policy and Practice, 132, 110–130. https://doi.org/10.1016/j.tra.2019.10.018
Jarvis, C. B., MacKenzie, S. B., & Podsakoff, P. M. (2003). A critical review of construct indicators and measurement model misspecification in marketing and consumer research. Journal of Consumer Research, 30(2), 199–218. https://doi.org/10.1086/376806
Jawhar, S. S. (2018). The effect of the roof and glazing type of traditional courtyard houses on energy efficiency: A case of Erbil City, Iraq [Master’s thesis, Near East University].
Jawhar, S. S., & Atakara, C. (2026). Post-2000 urban sprawl in Erbil, Iraq: Impacts on urban form, socioeconomic dynamics, culture, and quality of life. African and Asian Studies. Advance online publication. https://doi.org/10.1163/15692108-bja10099
Jiang, F., Liu, S., Yuan, H., & Zhang, Q. (2007). Measuring urban sprawl in Beijing with geo-spatial indices. Journal of Geographical Sciences, 17(4), 469–478. https://doi.org/10.1007/s11442-007-0469-z
Johnson, M. P. (2001). Environmental impacts of urban sprawl: A survey of the literature and proposed research agenda. Environment and Planning A, 33(4), 717–735. https://doi.org/10.1068/a3327
Kalogiannidis, S., Spinthiropoulos, K., Kalfas, D., Chatzitheodoridis, F., & Tziampazi, F. (2025). Integration of remote sensing and GIS for urban sprawl monitoring in European cities. European Journal of Geography, 16(2), 75–90. https://www.eurogeojournal.eu/index.php/egj/article/view/781
Kalvová, J., Halenka, T., Bezpalcová, K., & Nemešová, I. (2003). Köppen climate types in observed and simulated climates. Studia Geophysica et Geodaetica, 47(1), 185–202. https://doi.org/10.1023/A:1022263908716
Khzr, B. O., Ibrahim, G. R. F., Hamid, A. A., & Ail, S. A. (2022). Runoff estimation using SCS-CN and GIS techniques in the Sulaymaniyah sub-basin of the Kurdistan region of Iraq. Environment, Development and Sustainability, 24(2), 2640–2655. https://doi.org/10.1007/s10668-021-01549-z
Kock, N. (2015). Common method bias in PLS-SEM: A full collinearity assessment approach. International Journal of e-Collaboration, 11(4), 1–10. https://doi.org/10.4018/ijec.2015100101
Kock, N., & Hadaya, P. (2018). Minimum sample size estimation in PLS‐SEM: The inverse square root and gamma‐exponential methods. Information Systems Journal, 28(1), 227–261. https://doi.org/10.1111/isj.12131
Köppen, W. (2011). The thermal zones of the Earth according to the duration of hot, moderate and cold periods and to the impact of heat on the organic world (E. Volken & S. Brönnimann, Trans.). Meteorologische Zeitschrift, 20(3), 351–360. https://doi.org/10.1127/0941-2948/2011/105
Lasswell, H. D., & Kaplan, A. (2017). Power and society: A framework for political inquiry. Routledge. https://doi.org/10.4324/9781315127156
Lee, J. M., & Braham, W. W. (2017). Building emergy analysis of Manhattan: Density parameters for high-density and high-rise developments. Ecological Modelling, 363, 157–171. https://doi.org/10.1016/j.ecolmodel.2017.08.014
Lee, M., Barbosa, H., Youn, H., Holme, P., & Ghoshal, G. (2017). Morphology of travel routes and the organization of cities. Nature Communications, 8(1), 2229. https://doi.org/10.1038/s41467-017-02374-7
Lityński, P. (2021). The intensity of urban sprawl in Poland. ISPRS International Journal of Geo-Information, 10(2), 95. https://doi.org/10.3390/ijgi10020095
Liu, L., & Meng, L. (2020). Patterns of urban sprawl from a global perspective. Journal of Urban Planning and Development, 146(2), 04020004. https://doi.org/10.1061/(ASCE)UP.1943-5444.0000558
Lyons, G., Mokhtarian, P., Dijst, M., & Böcker, L. (2018). The dynamics of urban metabolism in the face of digitalization and changing lifestyles: Understanding and influencing our cities. Resources, Conservation and Recycling, 132, 246–257. https://doi.org/10.1016/j.resconrec.2017.07.032
Ma, W., Jiang, G., Li, W., & Zhou, T. (2018). How do population decline, urban sprawl and industrial transformation impact land use change in rural residential areas? A comparative regional analysis at the peri-urban interface. Journal of Cleaner Production, 205, 76–85. https://doi.org/10.1016/j.jclepro.2018.08.323
Mohammed, J. A. (2013). Rapid urban growth in the city of Duhok, Iraqi Kurdistan Region: An integrated approach of GIS, remote sensing and Shannon entropy application. International Journal of Geomatics and Geosciences, 4(2), 325–341.
Mouratidis, K. (2019). Compact city, urban sprawl, and subjective well-being. Cities, 92, 261–272. https://doi.org/10.1016/j.cities.2019.04.013
Naserikia, M., Asadi Shamsabadi, E., Rafieian, M., & Leal Filho, W. (2019). The urban heat island in an urban context: A case study of Mashhad, Iran. International Journal of Environmental Research and Public Health, 16(3), 313. https://doi.org/10.3390/ijerph16030313
Navamuel, E. L., Rubiera Morollón, F., & Moreno Cuartas, B. (2018). Energy consumption and urban sprawl: Evidence for the Spanish case. Journal of Cleaner Production, 172, 3479–3486. https://doi.org/10.1016/j.jclepro.2017.08.110
Nitzl, C., Roldan, J. L., & Cepeda, G. (2016). Mediation analysis in partial least squares path modeling: Helping researchers discuss more sophisticated models. Industrial Management & Data Systems, 116(9), 1849–1864. https://doi.org/10.1108/IMDS-07-2015-0302
Patel, S., Indraganti, M., & Jawarneh, R. N. (2024). Land surface temperature responses to land use dynamics in urban areas of Doha, Qatar. Sustainable Cities and Society, 104, 105273. https://doi.org/10.1016/j.scs.2024.105273
Petrović, P., & Lobanov, M. M. (2022). Impact of financial development on CO2 emissions: Improved empirical results. Environment, Development and Sustainability, 24(5), 6655–6675. https://doi.org/10.1007/s10668-021-01721-5
Pohlmann, A. (2018). Situating social practices in community energy projects. Springer VS. https://doi.org/10.1007/978-3-658-20635-2
Popkin, B. M. (1999). Urbanization, lifestyle changes and the nutrition transition. World Development, 27(11), 1905–1916. https://doi.org/10.1016/S0305-750X(99)00094-7
Rabiu, M. K., & Jaeger-Erben, M. (2022). Appropriation and routinisation of circular consumer practices: A review of current knowledge in the circular economy literature. Cleaner and Responsible Consumption, 7, 100081. https://doi.org/10.1016/j.clrc.2022.100081
Rash, A., Mustafa, Y., & Hamad, R. (2023). Quantitative assessment of land use/land cover changes in a developing region using machine learning algorithms: A case study in the Kurdistan Region, Iraq. Heliyon, 9(11), e21253. https://doi.org/10.1016/j.heliyon.2023.e21253
Roemer, E., Schuberth, F., & Henseler, J. (2021). HTMT2–an improved criterion for assessing discriminant validity in structural equation modeling. Industrial Management & Data Systems, 121(12), 2637–2650. https://doi.org/10.1108/IMDS-02-2021-0082
Rubiera-Morollón, F., & Garrido-Yserte, R. (2020). Recent literature about urban sprawl: A renewed relevance of the phenomenon from the perspective of environmental sustainability. Sustainability, 12(16), 6551. https://doi.org/10.3390/su12166551
Santana, P., Santos, R., & Nogueira, H. (2009). The link between local environment and obesity: A multilevel analysis in the Lisbon Metropolitan Area, Portugal. Social Science & Medicine, 68(4), 601–609. https://doi.org/10.1016/j.socscimed.2008.11.033
Shadman Roodposhti, M., Aryal, J., Shahabi, H., & Safarrad, T. (2016). Fuzzy Shannon entropy: A hybrid GIS-based landslide susceptibility mapping method. Entropy, 18(10), 343. https://doi.org/10.3390/e18100343
Shove, E., Pantzar, M., & Watson, M. (2012). The dynamics of social practice: Everyday life and how it changes. SAGE Publications. https://doi.org/10.4135/9781446250655
Smiraglia, D., Salvati, L., Egidi, G., Salvia, R., Giménez-Morera, A., & Halbac-Cotoara-Zamfir, R. (2021). Toward a new urban cycle? A closer look to sprawl, demographic transitions and the environment in Europe. Land, 10(2), 127. https://doi.org/10.3390/land10020127
Strengers, Y. (2010). Conceptualising everyday practices: Composition, reproduction and change (Carbon Neutral Communities Working Paper No. 6). Centre for Design, RMIT University; University of South Australia.
Strengers, Y. (2013). Peak electricity demand and social practice theories: Reframing the role of change agents in the energy sector. In S. Fudge, M. Peters, S. M. Hoffman, & W. Wehrmeyer (Eds.), The global challenge of encouraging sustainable living (pp. 18–42). Edward Elgar Publishing. https://doi.org/10.4337/9781781003756.00010
Subasinghe, S., Estoque, R. C., & Murayama, Y. (2016). Spatiotemporal analysis of urban growth using GIS and remote sensing: A case study of the Colombo Metropolitan Area, Sri Lanka. ISPRS International Journal of Geo-Information, 5(11), 197. https://doi.org/10.3390/ijgi5110197
Talkhabi, H., Ghalehteimouri, K. J., Mehranjani, M. S., Zanganeh, A., & Karami, T. (2022). Spatial and temporal population change in the Tehran Metropolitan Region and its consequences on urban decline and sprawl. Ecological Informatics, 70, 101731. https://doi.org/10.1016/j.ecoinf.2022.101731
Taubenböck, H., Debray, H., Qiu, C., Schmitt, M., Wang, Y., & Zhu, X. X. (2020). Seven city types representing morphologic configurations of cities across the globe. Cities, 105, 102814. https://doi.org/10.1016/j.cities.2020.102814
Taubenböck, H., Kraff, N. J., & Wurm, M. (2018). The morphology of the Arrival City – A global categorization based on literature surveys and remotely sensed data. Applied Geography, 92, 150–167. https://doi.org/10.1016/j.apgeog.2018.02.002
Thanoon, M. G., & Haykal, H. T. (2020). Influences of the accessibility and availability of green spaces on the liveability of residential complexes in Erbil City. American Journal of Civil Engineering and Architecture, 8(2), 25–36. http://pubs.sciepub.com/ajcea/8/2/1/index.html
Travisi, C. M., Camagni, R., & Nijkamp, P. (2010). Impacts of urban sprawl and commuting: A modelling study for Italy. Journal of Transport Geography, 18(3), 382–392. https://doi.org/10.1016/j.jtrangeo.2009.08.008
UNESCO World Heritage Centre. (2014). Erbil Citadel. https://whc.unesco.org/en/list/1437
Wu, J., Li, X., Luo, Y., & Zhang, D. (2021). Spatiotemporal effects of urban sprawl on habitat quality in the Pearl River Delta from 1990 to 2018. Scientific Reports, 11(1), 13981. https://doi.org/10.1038/s41598-021-92916-3
Zhang, H. (2021). The impact of urban sprawl on environmental pollution: Empirical analysis from large and medium-sized cities of China. International Journal of Environmental Research and Public Health, 18(16), 8650. https://doi.org/10.3390/ijerph18168650
Zhao, P. (2010). Sustainable urban expansion and transportation in a growing megacity: Consequences of urban sprawl for mobility on the urban fringe of Beijing. Habitat International, 34(2), 236–243. https://doi.org/10.1016/j.habitatint.2009.09.008
Zhao, P., & Zhang, M. (2018). The impact of urbanisation on energy consumption: A 30-year review in China. Urban Climate, 24, 940–953. https://doi.org/10.1016/j.uclim.2017.11.005
How to cite this article? (APA Style)
Jawhar, S. S., Sleman, S. H., & Atakara, C. (2026). Decoding the sprawl paradigm: How new building typologies shape energy consumption and environmental degradation in Erbil, Iraq. Journal of Contemporary Urban Affairs, 10(2), 362–385. https://doi.org/10.25034/ijcua.2026.v10n2-4
Decoding the Sprawl Paradigm… 1
JOURNAL OF CONTEMPORARY URBAN AFFAIRS, 10(2), 362–385/ 2026
