Quantifying the Thermal and Energy Impacts of Urban Morphology Using Multi-Source Data: A Multi-Scale Study in Coastal High-Density Contexts

Abstract

Urban thermal environments, characterized by the interplay between indoor and outdoor conditions, pose growing challenges in high-density coastal cities. This study proposes a multi-scale, integrative framework that couples a satellite-derived land surface temperature (LST) analysis with microscale building performance simulations to holistically evaluate the high-density urban thermal environment in subtropical climates. The results reveal that compact, high-density morphologies reduce outdoor heat stress (UTCI) through self-shading but lead to significantly higher cooling loads, energy use intensity (EUI), and poorer daylight autonomy (DA) due to restricted ventilation and limited sky exposure. In contrast, more open, vegetation-rich forms improve ventilation and reduce indoor energy demand, yet exhibit higher UTCI values in exposed areas and increased lighting energy use in poorly oriented spaces. This study also proposes actionable design strategies, including optimal building spacing (≥15 m), façade orientation (30–60° offset from west), SVF regulation (0.4–0.6), and the integration of vertical greenery to balance solar access, ventilation, and shading. These findings offer evidence-based guidance for embedding morphological performance metrics into planning policies and building design codes. This work advances the integration of outdoor and indoor performance evaluation and supports climate-adaptive urban form design through quantitative, policy-relevant insights.

1. Introduction

1.1. Background

With the increasing global emphasis on achieving carbon neutrality, addressing climate change has become a pressing global priority. Energy consumption, air pollution, and global warming are recognized as critical challenges [1,2,3]. The building sector accounts for over 40% of global energy consumption and approximately 36% of greenhouse gas emissions [4], making it a key focus area for many countries striving to achieve sustainable development goals. Moreover, merely reducing building energy consumption is no longer sufficient to meet modern residents’ high expectations for building performance. There is a growing awareness of the importance of thermal comfort in both indoor and outdoor spaces [5].

Hong Kong, one of the most densely populated cities in the world and a major international financial hub, provides a representative context for studying the impact of urban form on microclimate and energy. Characterized by compact development, a high population density, and limited land availability, nearly 7 million residents are concentrated in just 7% of the city’s land area. This extreme density creates significant challenges for daylighting, ventilation, and urban thermal management. Although ambient daylight levels often exceed 10,000 lux [6], dense building clusters hinder indoor daylight access. Low neighborhood permeability also hinders natural ventilation, reduces thermal comfort, and increases building energy use. Meanwhile, the utilization of distributed renewable energy systems, especially solar photovoltaics, has been greatly supported by the Hong Kong government to address both energy and climate challenges.

1.2. Literature Review

In recent years, researchers have increasingly explored the complex interactions between urban morphology, microclimate, and building energy performance. However, a majority of the existing studies tend to examine either indoor or outdoor thermal environments in isolation.

For outdoor comfort, studies have focused on UHI mitigation and urban shading. Ibrahim et al. [7] optimized urban courtyard designs using the Universal Thermal Climate Index (UTCI) to enhance outdoor thermal comfort. However, this study did not assess indoor thermal conditions or energy use. Evola et al. [8] proposed an integrated workflow combining UTCI with urban canyon energy demand modeling. While this study was innovative, it remained focused on outdoor comfort and did not address indoor thermal comfort or energy use. Nicholson et al. [9] validated the use of Ladybug Tools for optimizing outdoor thermal comfort through urban shading strategies. Although the study provided valuable insights, it failed to evaluate energy performance or the urban heat island effect. Zhang et al. [10] conducted a multi-scale UHI assessment to identify driving factors under varying climatic conditions. However, their research was also limited to outdoor thermal dynamics.

Conversely, indoor-oriented studies have investigated cooling loads or retrofit strategies with minimal reference to the microclimate. Zhang et al. [11] used machine learning models to analyze and predict building heating and cooling loads, with less consideration of the outdoor microclimate. Similarly, Tootkaboni et al. [12] investigated the impacts of building cooling strategies and building components on indoor energy demand and thermal comfort without thoroughly addressing the peripheral microclimate environment. Aloshan et al. [13] simulated the impact of school building facade renovation on energy efficiency and optimized renovation strategies for cooling demand reduction, while also failing to correlate with the local climate. Errebai et al. [14] quantified the effects of the UHI on the cooling energy demand in Montreal using meteorological simulations and station data. However, their analysis excluded outdoor thermal comfort metrics and failed to connect local UHI mitigation strategies with indoor cooling loads. Jia et al. [15] conducted a comprehensive study on indoor thermal and ventilation comfort in a mechanically ventilated university classroom in Hong Kong. By combining 5-min realtime monitoring and student perception surveys, they identified the relative importance of thermal sensation and air freshness (CO₂) on overall comfort. While the study focused on short-term indoor environmental dynamics, it did not account for outdoor thermal drivers or multi-scale urban morphology impacts.

Although a few efforts attempt to bridge this gap, they remain limited in spatial scale, temporal resolution, or methodological integration. For example, Mirzabeigi et al. [16] explored thermal adaptation strategies using optimization or adaptive models, but did not systematically address UHI effects or multi-scale interactions. Several studies (e.g., De Luca et al. [17], Sun et al. [18], Xu et al. [19], Pompei et al. [20], and Wang et al. [21]) provided valuable contributions in modeling or prediction but were restricted to either indoor or outdoor domains, lacking an integrated perspective.

Table 1. Summary of the relevant literature on urban heat, energy performance, and thermal comfort.

Author(s) and Year	Methodology	Indicators	Key Findings	Limitations
De Luca et al. (2021) [17]	CFD + EnergyPlus simulation of Nordic high-rise clusters	Indoor cooling load, outdoor UTCI	Layout variations significantly affected both indoor and outdoor comfort	Limited integration of radiative or seasonal thermal variations
Sun et al. (2023) [18]	Genetic algorithm inverse design of school courtyard forms	UTCI, solar radiation, height/coverage	An optimized building layout reduced UTCI by 1.2–1.6 °C	Focused on summer performance at individual building scale; limited generalizability
Xu et al. (2020) [19]	Adaptive HVAC setpoints using dynamic clothing insulation models	Indoor comfort temperature, HVAC energy	Adaptive control strategies reduced energy use by up to 65.5%	Focused only on indoor comfort; the outdoor thermal impact was not considered
Pompei et al. (2024) [20]	ML-based UTCI prediction using RF, GBDT, and SVR	Air temp, humidity, SVF, PET, UTCI	Built density and vegetation were key variables	Lacked integration with indoor metrics or energy demand outcomes
Wang et al. (2023) [21]	Landsat LST time series via GEE	LST, UHI pattern, NDVI/NDBI/NDWI	UHI expanded over 20 years; imperviousness and vegetation were key factors	Daytime only; lacked indoor metrics

summarizes selected key studies, including their methodology, indicators used, and major limitations.

1.3. Research Gaps

Although some studies have attempted to bridge the gap between outdoor microclimates and indoor thermal environments, significant limitations remain. The reviewed literature highlights critical gaps in the understanding and analysis of urban thermal environments. First, most studies focus predominantly on either outdoor thermal conditions or indoor energy performance, with limited efforts to integrate these two interconnected aspects. This isolation leaves a critical void in comprehensively addressing the interplay between urban microclimates and building-level thermal dynamics. Second, there is a notable absence of large-scale seasonal analyses leveraging advanced remote sensing tools, such as Google Earth Engine (GEE), to validate microclimatic impacts at the urban scale. Furthermore, while simulation tools like EnergyPlus and computational fluid dynamics (CFD) are widely applied in building and thermal studies, their integration with urban-scale UHI simulations remains insufficiently explored.

To address the intertwined challenges of urban heat and energy resilience, this study proposes a cross-scale, multi-indicator research framework that integrates satellite-based thermal observations with simulation-driven building performance modelling, thereby linking mesoscale and microscale analyses. Specifically, the framework combines land surface temperature (LST) data acquired from a satellite with a parametric simulation platform to enable data-driven weather adaptation and a comprehensive indoor–outdoor thermal analysis. By spatially differentiating microclimate impacts and quantifying their influences on building performance, the framework offers actionable insights for climate-adaptive urban design and urban heat island mitigation. It further provides a foundation for prioritizing adaptive interventions in vulnerable areas, ultimately supporting form–performance co-optimization in climate-resilient urban development.

The remainder of this paper is organized as follows: Section 2 outlines the workflow of the study, detailing the methodology and data applied in both mesoscale and microscale simulations. Section 3 presents the results, delivering a comprehensive analysis of the findings, and discusses the limitations. Finally, Section 4 summarizes the main findings and provides further suggestions based on the overall analysis.

2. Materials and Methods

2.1. Description of the Study Area

Hong Kong is situated in southeastern China, at approximately 22.3193° N latitude and 114.1694° E longitude. It is one of the most densely populated cities globally, with a population density of 25,900 people per square kilometer. According to the Köppen–Geiger climate classification, Hong Kong falls under the Cwa climate type (C = temperate, w = dry winter, a = hot summer), characterized by distinct seasonal variations [22]. Housing in Hong Kong comprises both public and private sectors, with approximately 45% of the population residing in public housing estates [23]. The existing public housing estate types are categorized into 15 standardized building layouts designed by the Hong Kong Housing Authority [24]. These include “Harmony,” “Trident,” “Slab,” “H-block,” “Y-shaped,” “Single-tower Type I,” and “Twin-tower” designs. Among public housing estates constructed after 2000, around 70% feature a cross-shaped layout, which is the dominant type of public housing estate today [25]. In 2024, Hong Kong hosted a total of 260 public housing estates, comprising more than 850,000 units. The distribution of existing, under-construction, planned, and proposed housing estates across districts is shown in Figure 1.

To enable an in-depth analysis, two representative urban blocks were selected for detailed simulation. As illustrated in Figure 2, Case 1 (Kowloon City) and Case 2 (Wong Tai Sin) were chosen based on spatial variations in land surface temperature (LST) across the study area.

2.2. Research Methodology

This study adopts a multi-scale methodological framework that integrates mesoscale remote sensing with microscale building performance simulation to evaluate the relationships between urban form, thermal comfort, and energy performance. As illustrated in Figure 3, the framework is structured into two stages:

• The mesoscale analysis extracts urban heat island patterns and morphological features using (Google LLC, Mountain View, CA, USA; https://code.earthengine.google.com/, accessed on 27 November 2024) and QGIS v3.30.2 (QGIS Development Team, Grüt, Zurich, Switzerland) based on seasonal Landsat 8 imagery.
• The microscale simulation evaluates indoor–outdoor environmental performance using a combination of Dragonfly v1.8.0 (Ladybug Tools LLC, New York, NY, USA), EnergyPlus v23.2.0 (U.S. Department of Energy, Washington, DC, USA), Ladybug v1.8.0, and Honeybee v1.8.0 (Ladybug Tools LLC, New York, NY, USA) tools within the Rhino–Grasshopper (Robert McNeel & Associates, Seattle, WA, USA) environment.

Landsat 8 data provide mesoscale insights into urban heat patterns driven by urban morphology and environmental conditions [27]. When integrated into indoor thermal environment simulations, they serve as realistic boundary conditions, ensuring that building-scale models accurately reflect local outdoor thermal influences. Seasonal variations in vegetation, urban morphology, and temperature across the 18 districts were analyzed [28]. Key urban morphological indicators, such as the building density, average building height, tree coverage, number of floors, and aspect ratio, were calculated for further analysis. In the microscale stage, an integrated simulation platform was developed using Grasshopper v1.8.0, and the workflow was shown in Appendix A. The simulation workflow was divided into interconnected modules: (1) Dragonfly v1.8.0 was used to downscale rural weather data by incorporating morphological inputs into the Urban Weather Generator (UWG) (Massachusetts Institute of Technology, Cambridge, MA, USA), thereby producing localized EPW files that reflect microclimatic UHI effects. (2) EnergyPlus v23.2.0 was applied to simulate indoor energy consumption and thermal behavior. (3) Ladybug v1.8.0 served as the interface for Radiance v5.4 (Lawrence Berkeley National Laboratory, Berkeley, CA, USA) and UTCI calculations, enabling spatial assessments of daylight autonomy (DA) and outdoor thermal comfort. The resulting simulation outputs included both indoor indicators (EUI, cooling load, and DA) and outdoor indicators (UTCI), offering an integrated diagnosis of climate sensitivity and performance across different urban typologies. The framework is illustrated in Figure 4. The simulation methods and parameters employed in this study were rigorously validated, with the results demonstrating strong agreement between the experimental measurements and simulation outputs [17], confirming the reliability and accuracy of the approach.

2.3. Data Source and Material

Landsat 8 Collection 2 Level-2 surface reflectance and surface temperature (SR/ST) images were obtained from the United States Geological Survey (USGS, Sioux Falls, SD, USA) [29,30] for the period between December 2022 and November 2023, covering one full climatic year. Based on Hong Kong’s subtropical climate, the data were grouped into four seasons: winter (December 2022–February 2023), spring (March–May 2023), summer (June–August 2023), and autumn (September–November 2023). Only cloud-free images (cloud cover <10%) were selected using the QA_PIXEL band, and a seasonal median composite was generated for each period. Urban and rural zones were classified using the official land planning raster from the Hong Kong Planning Department (Hong Kong SAR Government, Hong Kong, China).

Through a cluster-based and threshold-informed analysis of LST and morphology, a set of representative urban block typologies was extracted to capture the spectrum of thermal vulnerability conditions present within the city. These typologies were not chosen as isolated cases but as archetypal urban forms reflecting varying combinations of building geometry, vegetation cover, and density gradients. The building environment data for these areas were obtained from Hong Kong’s Common Spatial Data Infrastructure (CSDI, Development Bureau of the Government of the Hong Kong Special Administrative Region, Hong Kong, China) [31], including essential information such as building footprints, building heights, road networks, terrain, and vegetation coverage. All spatial data were processed in QGIS to generate high-resolution 3D building models for the two selected urban blocks. Twelve key urban morphological indicators [32] were calculated within a 200 m radius of each block, these indicators were used as input parameters for the Urban Weather Generator (UWG) model to generate urban EPW weather files, ensuring that microclimatic influences on the building thermal environment were accounted for. While this study focuses on Hong Kong, the methodology is broadly transferable to other urban contexts where comparable satellite imagery and spatial datasets are available, providing a scalable framework for identifying and analyzing thermally vulnerable urban typologies.

To ensure modeling consistency, detailed settings for the building envelope, HVAC systems, usage schedules, and urban environmental factors were unified with reference to similar studies. These parameters, including window performance, shading, material composition, internal loads, and infiltration rates, are summarized in

Table 2. Summary of model settings for the parameters.

Component	Indicators	Values
Window [33]	Window-to-wall ratio (WWR)	0.22
	Solar heat gain coefficient (SHGC)	0.32
	U-value	5.75 W/m2·K
	Transmittance	0.8
Material [34,35]	Exterior walls: reinforced concrete, gypsum, and light-colored mosaic tiles Roofs: reinforced concrete, gypsum, polystyrene insulation, and light-colored mosaic tiles Floors: concrete slabs Shading: 1 m overhangs with a tilt angle of 65°
Infiltration	0.0001 m3/s·m2
Schedule [36]	Occupancy, lighting, equipment, ventilation, cooling
Load [37]	Occupant density	0.088 people/m2
	Lighting power density	5 W/m2
	Equipment power density	8 W/m2
	Ventilation air exchange rates	0.65
HVAC [38]	Cooling temperature	25 °C
	Type	packaged terminal air conditioner (PTAC)
Urban microclimate	Vehicle heat emissions [27,39]	8 W/m2
	Vegetation albedo [34]	0.6

. The modeling assumptions reflect typical specifications for Hong Kong’s public housing and were used to simulate the cooling energy demand and indoor thermal performance across all seasons.

2.4. Evaluation Metrics

2.4.1. Land Surface Temperature

Land surface temperature (LST) (see Equation (1)) [40] refers to the radiative temperature of the Earth’s surface, derived from satellite or airborne remote sensing data. It is a critical indicator for analyzing urban heat island effects, especially in densely urbanized areas. LST is influenced by factors such as the surface type, vegetation cover, and urban morphology. Tools like GEE, QGIS v3.30.2, and Python v3.11 (Python Software Foundation, Wilmington, DE, USA) are useful for analyzing LST across large urban areas.

$$LST={\displaystyle \frac{BT}{1+\left(\lambda \times \frac{BT}{\rho }\right)\times \mathit{ln}\epsilon }}$$

(1)

where λ represents the wavelength of emitted radiance, defined as the peak response and the average of the limiting wavelengths, set as 11.5 μm; ρ is hc/σ, set as 1.438 × 10⁻² m·K; h is Planck’s constant (6.626 × 10⁻³⁴) J·s; σ is Boltzmann constant (1.38 × 10⁻²³) J/K; c is the velocity of light (2.998 × 10⁸ m/s); and ε is the emissivity.

2.4.2. Urban Heat Island Intensity

The urban heat island intensity (UHII) (see Equation (2)) [41] is commonly defined as the temperature difference between urban and rural areas. In this study, the urban heat island intensity is calculated using land surface temperature (LST) as follows:

$$UHII=T_{urban}-T_{rural}$$

(2)

where $T_{urban}$ represents the average land surface temperature of the urban area, and $T_{rural}$ refers to that of a reference rural area.

2.4.3. Energy Use Intensity

The energy use intensity (EUI) (see Equation (3)) [42] is a metric used to quantify the energy efficiency of a building or an urban area by measuring the total energy consumption relative to the floor area.

$$EUI={\displaystyle \frac{E_t}{A}}$$

(3)

where E_t represents the total annual energy consumption of the building or area (typically measured in kilowatt-hours, kWh); and A is the total floor area (measured in square meters, m²). EUI is expressed in units of kWh/m²/year.

2.4.4. Cooling Energy Consumption

Cooling energy consumption (Q_c) (see Equation (4)) [43] is a measure of the total energy required to maintain indoor thermal comfort by removing excess heat from a building or space using air conditioning or other cooling systems. It is influenced by factors such as the outdoor temperature, building insulation, window-to-wall ratio, internal heat gains, and HVAC system efficiency.

$$Q_c={\displaystyle \frac{Q_{c,t}}{\epsilon }}$$

(4)

where Q_c means the cooling energy consumption (kWh); Q_c,t means the total cooling load (kWh); and $\epsilon$ means the efficiency of the cooling system (dimensionless), and is typically represented by the Coefficient of Performance (COP) of the system.

2.4.5. Universal Thermal Climate Index

The Universal Thermal Climate Index (UTCI) (see Equation (5)) [42] is a comprehensive metric used to assess outdoor thermal comfort by considering the combined effects of air temperature, wind speed, humidity, and solar radiation on human physiology. UTCI is expressed as an equivalent temperature T_UTCI and is calculated using a complex mathematical model based on the human energy balance and thermoregulation.

$$T_{UTCI}=f (T_a,v,RH,S)$$

(5)

where $T_a$ is the air temperature (°C); $v$ is he wind speed at 10 m above ground (m/s); $RH$ is the relative humidity (%); and $S$ is the solar radiation (W/m²).

2.4.6. Daylight Autonomy

Daylight autonomy (DA) (see Equation (6)) [44] is a metric used to evaluate the percentage of time a given space receives sufficient natural daylight to meet a predefined illuminance threshold over a specific period. DA is commonly used in building performance assessments to analyze daylight availability and reduce reliance on artificial lighting.

$$DA={\displaystyle \frac{\sum _{i=1}^n{H}_i}{T}}\times 100$$

(6)

where H_i is the number of hours when the illuminance at a specific point meets or exceeds the threshold; T is the total number of hours in the analysis period; and n represents the number of calculation points or intervals.

3. Results and Discussion

3.1. Seasonal Variation Analysis

Figure 5a illustrates the spatial distribution of building footprints across Hong Kong, where high-density urban fabrics (black) are concentrated in Kowloon and northern Hong Kong Island. Figure 5b–f present the seasonal variation in land surface temperature (LST) after masking out water bodies. The white areas represent regions originally occupied by water, which have been excluded from the LST calculation to eliminate water bias.

Figure 5c–f illustrate the spatial distribution of land surface temperature (LST) across the study area during the four seasons: summer, winter, spring, and autumn. The highest average LST values were observed in summer, with urban cores like Kowloon City and Mong Kok exceeding 38 °C, while peripheral regions such as Tai Po recorded significantly lower values near 24 °C. In winter, the spatial disparity was less pronounced, with average LST values ranging from 12 °C to 18 °C. The annual average land surface temperature (LST) in urban areas is 28.77 °C, while that in rural areas is 25.56 °C. Therefore, the annual average UHII in Hong Kong is approximately 3.2 °C, which is consistent with values reported in previous scientific studies. For example, Siu et al. [44] found annual mean UHII values ranging from 1.5 °C to 2.8 °C based on long-term air temperature records, with peak daily values exceeding 3.8 °C. More recent satellite-based studies (e.g., Xu et al. [45] and Chakraborty et al. [46]) reported surface UHII values typically between 1 °C and 3 °C, but exceeding 3 °C in densely developed urban areas during summer.

Dense urban areas consistently exhibited higher LST values due to reduced vegetation and increased anthropogenic heat, whereas regions with vegetation cover exceeding 50% showed reductions in the LST of up to 5 °C compared to adjacent built-up areas, underscoring the cooling benefits of green infrastructure. In spring and autumn, intermediate LST patterns were observed, reflecting transitional thermal behaviors. Elevated LST values in summer are likely to correlate with higher cooling energy demands, particularly in high-density urban districts. The temporal and spatial variations in the LST highlight Hong Kong’s pronounced seasonal changes. These findings confirm the strong seasonality of the LST in Hong Kong, aligning with the observations of Yee and Kaplan [47].

3.2. Regional Variation Analysis

Figure 6a presents the average land surface temperatures (LSTs) for each district in autumn, spring, summer, winter, and the annual mean. A clear seasonal trend is observed in which summer yields the highest LST in all districts, while winter is coolest. Spatially, there is a pronounced urban–rural contrast. Densely built urban districts consistently register higher LSTs than outlying greener districts. For example, KLC—one of the most urbanized areas—reaches an average summer LST of about 39 °C, whereas a less urbanized, vegetation-rich district like ILD stays around 30 °C in summer. Spring and autumn LST patterns are intermediate, but still mirror this spatial variation (urban areas remain warmer than rural ones). In winter, absolute temperatures drop for all districts (e.g., KLC averages a high of 25 °C, versus a low of 22 °C in ILD), slightly narrowing the gap. Nevertheless, the hottest districts remain a few degrees warmer than the coolest, even in cooler seasons. This seasonal and spatial LST pattern underscores a strong link between urban morphology and surface heating: the city’s core heats up dramatically in summer and remains warm in winter relative to its greener periphery. Figure 6b shows the seasonal urban heat island intensity (UHII) for each district. A notable pattern is that the most urbanized districts exhibit consistently positive UHII in all seasons, whereas districts with extensive green or open areas often have near-zero or even negative values. For example, some outlying districts (with mountains, parks, or coastal features) show negative UHI intensities in summer (around −0.2 to −0.5 °C). However, the dense inner-city districts are about 0.5 to 0.9 °C hotter than the baseline on average, contributing to a daytime UHI effect (and potentially an even larger effect at night). Spatially, the urban core consistently shows the highest UHI intensity (on the order of 1 °C above others on average annually), whereas the outlying/new territories districts show a minimal or negative UHII. This confirms that Hong Kong’s urban heat island effect is significant in dense districts and persists through seasons, albeit with a varying magnitude, reinforcing that built-up areas remain hotter than their surroundings year-round.

The spatial analysis of the LST and UHII across Hong Kong’s 18 districts highlights substantial thermal variability closely associated with the urban morphological characteristics. High-density urban cores, such as Kowloon City, Yau Tsim Mong, Kwun Tong, and Sham Shui Po, consistently experience elevated LST values, which are particularly evident in summer (39 °C in Kowloon City and 38 °C in Yau Tsim Mong), primarily due to the high building density, narrow street canyons, extensive impervious surfaces, and minimal vegetation coverage. Conversely, districts with lower-density development and substantial greenery, including Tai Po, Sha Tin, Wong Tai Sin, and portions of the Island District, exhibit a significantly lower LST and minimal UHI effects. Wong Tai Sin, despite its central location, maintains moderate temperatures due to abundant urban greenery, while outlying districts benefit from enhanced natural ventilation and extensive vegetation cover. Intermediate districts like the Central and Western District and the Eastern District demonstrate mixed influences of the dense urban form mitigated by pockets of green space.

3.3. Typical Urban Block Created

Table 3. Statistical values of average urban morphology parameters of the 18 districts of Hong Kong.

District	BD	aBH (m)	mBCD (m)	BSC	SVF	FAR	PS (m2)	AR	CN	BO (°)	VCR	Alb
CW	0.44	48.23	63.68	0.344	0.45	7.15	22,826	1.8	0.77	269	0.447	0.055
EST	0.43	59.16	81.60	0.513	0.54	8.47	38,282	1.2	0.62	160	0.412	0.061
ILD	0.42	49.38	91.19	0.455	0.57	6.93	33,828	1.0	0.69	185	0.412	0.080
KC	0.32	71.11	91.53	0.418	0.56	7.62	71,212	1.5	0.64	234	0.382	0.065
KLC	0.52	58.59	63.72	0.575	0.41	10.11	44,042	1.5	0.67	154	0.436	0.045
KT	0.34	78.55	92.43	0.314	0.55	8.83	60,627	1.7	0.61	203	0.366	0.073
NTH	0.33	76.76	81.16	0.380	0.57	8.32	50,008	1.4	0.64	165	0.429	0.084
SK	0.49	75.47	78.09	0.300	0.51	12.31	49,441	1.5	0.63	123	0.439	0.064
SSP	0.37	65.22	81.31	0.455	0.51	8.13	54,336	1.3	0.66	147	0.418	0.084
ST	0.29	76.35	101.53	0.524	0.56	7.33	78,422	1.7	0.64	176	0.370	0.075
STH	0.42	58.96	110.74	0.369	0.48	8.32	64,493	1.3	0.65	148	0.422	0.085
TM	0.29	66.76	103.41	0.609	0.56	6.43	76,469	1.3	0.66	190	0.404	0.093
TP	0.31	73.27	115.49	0.276	0.56	7.66	86,405	1.2	0.61	142	0.403	0.078
TW	0.49	59.15	79.68	0.593	0.50	9.72	40,614	1.8	0.66	134	0.429	0.061
WC	0.74	43.38	65.26	0.418	0.59	10.65	27,180	2.3	0.44	159	0.403	0.078
WTS	0.34	63.05	97.12	0.576	0.52	7.07	70,558	1.5	0.64	175	0.415	0.079
YL	0.43	68.37	91.01	0.278	0.56	9.85	59,640	1.5	0.68	161	0.403	0.065
YTM	0.59	52.84	64.62	0.528	0.39	10.47	28,099	1.5	0.72	215	0.440	0.052

characterizes the average urban morphology of public housing estates across Hong Kong’s eighteen districts, revealing a clear gradient from low-rise, expansive developments to slender, high-rise towers. Estates in ST and TM display low building densities (BD ≈ 0.29) and moderate heights (aBH ≈ 66–76 m), with high site coverages (BSC ≈ 0.60) that produce relatively modest FARs (6.43–7.33). In contrast, SK’s public estates combine very small building footprints (BSC ≈ 0.30) with tall towers (aBH ≈ 75 m), yielding the highest FAR of 12.3. WC stands out with the highest density (BD ≈ 0.74) yet the lowest mean height (43 m) and the smallest plot areas (~27,180 m²), reflecting tightly packed mid-rise blocks. Floor-area ratios across all districts range from approximately 4.8 to 12.3, encapsulating both the low-density New Town model and the high-density “skinny-tower” typology. Sky view factors vary from 0.39 in YTM (dominated by street-canyon effects) to 0.59 in WC, while street connectivity peaks in YTM (CN = 0.72) and is weakest in the TP (CN ≈ 0.61). Aspect ratios (1.2–2.3), the vegetation coverage ratio (VCR ≈ 0.36–0.44), and vegetation albedo (Alb ≈ 0.045–0.093) show only limited spatial variation. These morphological distinctions among public housing estates form the basis for understanding their diverse thermal behaviors in subsequent analyses.

The integrated analysis of urban morphology indicators across Hong Kong’s 18 districts in summer reveals both distinct spatial variations (Figure 7 and Figure 8) and significant correlations (Figure 9) with the thermal environment. The Pearson correlation coefficient patterns identify FAR (r = 0.32) and BD (r = 0.21) as primary positive drivers of the elevated LST, reinforcing the conclusion that high-density urban development significantly exacerbates heat accumulation. Similarly, aBH (r = 0.12), AR (r = 0.11), and CN (r = 0.11) demonstrate moderate positive correlations, indicating that taller, more compact urban forms and highly connected street networks contribute further to increased urban temperatures. In contrast, variables such as Alb (r = −0.42), mBCD (r = −0.41), and SVF (r = −0.40) exhibit negative correlations, highlighting their roles in mitigating urban heat. Specifically, a higher surface albedo effectively reduces heat absorption, while increased building spacing (higher mBCD) and improved sky visibility (higher SVF) facilitate heat dissipation through enhanced nocturnal radiative cooling and improved urban ventilation.

District-specific morphological profiles further elucidate these quantitative findings. For instance, districts such as YTM and KLC, characterized by densely packed, high-rise urban typologies (high BD, high FAR, low SVF, and low vegetation coverage), correspond to a higher LST and pronounced UHII. Conversely, districts like TP and ST, with a comparatively lower BD, higher vegetation coverage (VCR), and improved sky openness (high SVF), quantitatively exhibit lower thermal stress. Notably, YL represents an exceptional case where rapid suburban urbanization, despite a moderate BD and increased VCR, results in unexpectedly high LST values, suggesting the influence of intermediate morphological conditions or mixed land-use patterns, meriting a deeper micro-scale analysis.

Combining these spatial and morphological insights, two representative typologies are proposed: a high-density, compact urban morphology exemplified by KLC or YTM, and characterized by high BD, high FAR, low SVF, and limited vegetation; and a moderate-density, vegetation-rich morphology represented by WTS or TP, featuring a lower BD, moderate building heights, high vegetation cover, and improved sky openness. These contrasting typologies provide valuable reference models for the detailed microscale analysis.

3.4. Urban Energy and Thermal Results

3.4.1. Simulation Results of Universal Thermal Climate Index

This section evaluates outdoor thermal comfort by analyzing variations in the UTCI for representative urban blocks during the hottest week, July 22 to 28, covering an 84 h monitoring period from 8:00 am to 7:00 pm daily. Spatial and temporal variations in the UTCI values were recorded by sensor clusters strategically positioned within each urban block. Figure 10 presents the spatial distribution of average UTCI values across the two selected urban typologies, reflecting the impacts of urban morphology, building layouts, and microclimatic factors. The observed UTCI values predominantly ranged between 31 °C and 34 °C, with red-colored zones indicating regions experiencing elevated heat stress nearing 34 °C.

Although the spatial extent and intensity of high UTCI values appear more pronounced in Case 2, a complementary statistical analysis of UTCI categories (Figure 11) reveals a more nuanced interpretation. While Case 2 exhibits broader heat exposure due to its lower building density (BD = 0.47) and higher sky view factor (SVF = 0.37), the percentage of time achieving UTCI category 4 (very strong heat stress) is actually longer for Case 1, indicating that compact layouts can also exacerbate thermal stress in specific microclimatic conditions. In contrast, Case 2 shows a higher percentage of time in UTCI category 0 (no thermal stress), suggesting that its greater vegetation coverage (tree coverage = 0.32) and openness improve comfort during non-peak hours. However, limited shading in large open spaces leads to elevated UTCI values during peak sunlight periods.

Meanwhile, Case 1’s higher-density layout provides localized shading that reduces direct radiation, lowering the UTCI in shaded areas. Yet, its relatively low vegetation coverage (0.20) and restricted airflow create hotspots of accumulated heat stress. This comparison demonstrates that while urban compactness offers some shading benefits, it does not universally guarantee thermal comfort, especially under high solar exposure. It should be noticed that this analysis primarily considers shading effects and excludes the cooling benefits of vegetation evapotranspiration.

3.4.2. Simulation Results of Energy Use Intensity

Table 4. Comparison of energy consumption by category.

Categories	Energy Use Intensity (kWh/m2)
	Case 1	Case 2
Cooling	67.96	59.80
Interior lighting	36.14	45.14
Electric equipment	68.44	56.27
Total	172.53	161.20

compares the energy use intensity (EUI) of public housing buildings between the two case study areas, emphasizing differences in the cooling load, indoor lighting, electric equipment energy consumption, and overall energy efficiency. The cooling load for simulated buildings in Case 1 is 67.96 kWh/m², which is significantly higher than the 59.8 kWh/m² in Case 2, representing a 12% increase. This can be attributed to the dense urban morphology in Case 1, which elevates outdoor temperatures and subsequently increases the cooling demand. This finding aligns with the GEE analysis, confirming that the urban heat island effect is more severe in Case 1. Then, indoor lighting energy consumption is higher in Case 2 (45.14 kWh/m²) compared to Case 1 (36.14 kWh/m²). This can be attributed to the greater average building height in Case 2 (56.38 m) compared to Case 1 (41.25 m). Taller buildings often have greater building depths, making it more challenging for natural light to penetrate into the building’s core. Additionally, areas such as corridors and elevator lobbies, typically located in the central parts of buildings and lacking access to natural daylight, further contribute to the increased reliance on artificial lighting in Case 2.

Conversely, electric equipment energy consumption is higher in Case 1 (68.44 kWh/m²) compared to Case 2 (56.27 kWh/m²). This discrepancy can be partially attributed to the smaller floor area in Case 1 buildings (972.74 m²) compared to Case 2 (1326.33 m²). A smaller floor area likely results in a higher equipment power density (W/m²), as a similar number of plug loads and specialized equipment are distributed across a more confined space. Additionally, the concentrated internal heat gains from equipment in Case 1 could exacerbate the cooling load, indirectly increasing overall energy consumption.

The total energy consumption in Case 1 (172.53 kWh/m²) exceeds that of Case 2 (156.49 kWh/m²), indicating that the urban morphology of Case 1 negatively impacts multiple aspects of energy performance, including the cooling load, ventilation, and lighting. In contrast, the urban form of Case 2 appears more conducive to reducing the overall energy demand. These findings suggest that Case 1 requires greater attention to improving the thermal insulation of building envelopes and enhancing equipment efficiency, while Case 2 should focus on optimizing daylighting design to reduce the reliance on artificial lighting.

3.4.3. Simulation Results of Detailed Cooling Energy Consumption

To ensure spatial and morphological consistency between outdoor thermal comfort and indoor energy demand assessments, this section examines the cooling energy consumption of the same representative buildings (Case 1 and Case 2) analyzed in the UTCI study. Specifically, cooling loads were assessed across three typical floors to explore how urban morphology and orientation affect indoor thermal performance.

Figure 12a presents the typical simplified floor plan of the selected public building used for simulation in both case areas. The central lobby connects to eight surrounding rooms (Public room_1 to Public room_8), each oriented in a different direction. The radar chart in Figure 12b uses labels D1 to D8, which correspond clockwise to Public room_1 through Public room_8, respectively. This mapping reflects the response of cooling loads to the room orientation under different urban and solar exposure conditions. The middle of rooms 6 and 7 faces due south, which is defined as the primary orientation, maximizing natural daylight and being most significantly influenced by solar radiation. Rooms 1 and 8 have exterior walls facing northwest and southwest, respectively, while rooms 2 and 3 face northwest and northeast. Rooms 4 and 5 face northeast and southeast, and rooms 6 and 7 face southeast and southwest. This arrangement ensures diverse orientations for the building’s external facades, balancing daylight access and solar exposure.

Figure 12b shows the distribution of the cooling load. For Case 1, the cooling load distribution shows pronounced directional variations, particularly in the southeast direction (rooms D5 and D6), where cooling loads are 25–35% higher than those in Case 2, peaking at 11,000 kWh. This can be attributed to Case 1’s high building density (0.64), which amplifies the urban heat island effect by increasing heat reflection and retention in the densely packed urban environment. Additionally, the low mean sky view factor (0.22) restricts heat dissipation, while limited natural ventilation exacerbates the indoor cooling demand. In contrast, Case 2’s cooling loads in the same direction remain below 8000 kWh, demonstrating the mitigating effects of its lower building density (0.47) and higher tree coverage (0.32), which enhance heat dissipation and shading, respectively. However, the lower shading effect due to the higher mean sky view factor (0.37) exposes some indoor spaces to direct solar radiation, particularly in open areas with reduced adjacent building coverage. As a result, Case 2 exhibits a more uniform cooling load distribution but with notable peaks in the northeast (rooms D4) and southwest (rooms D7 and D8) directions, where cooling loads exceed those in Case 1 by 15–20%. This can be attributed to its open urban morphology, which decreases shading and allows direct solar radiation to penetrate indoor spaces during the early morning hours. Similarly, the southwest direction experiences higher cooling loads due to increased heat gain from warm air flows in less obstructed urban spaces, compounded by an insufficient shading design. In contrast, in Case 1, the cooling load is relatively lower in the northwest (rooms D1 and D2) and northeast (rooms D3 and D4) directions. This is likely due to the compact urban morphology, which limits solar exposure and provides more effective shading.

Figure 13 illustrates the cooling load distribution across eight building orientations (D1–D8) and three vertical zones (bottom, middle, and top) for Cases 1 and 2. Overall, Case 2 exhibits greater vertical sensitivity, with top-floor units consistently recording higher cooling loads due to increased solar exposure and limited shading, which are particularly evident in directions D3, D4, and D7, where middle-floor loads exceed 9000 kWh. In contrast, Case 1 displays less variation by height but more pronounced differences by orientation. Directions D5 and D6, likely west- or southwest-facing, show the highest cooling loads across all floors, exceeding 11,000 kWh at the ground level, suggesting intense afternoon solar gains compounded by the dense urban form and reduced ventilation. Conversely, D3, D4, D7, and D8 in Case 1 maintain the lowest loads (4000–5000 kWh), benefiting from favorable shading and orientation.

Between cases, directional performance diverges. Case 1 consistently surpasses Case 2 in the cooling load for D5 and D6 across all floors, underscoring the influence of the compact urban morphology and UHI effects. However, in orientations like D3 and D4, Case 2 often records higher loads, with D4 reaching over 9000 kWh at the middle level. The most significant discrepancy occurs in D3 at the middle floor, where Case 2’s load is approximately 54.5% higher than that of Case 1. These results highlight the complex interplay between orientation, height, and urban context, emphasizing the need for differentiated design strategies tailored to façade exposure and vertical layout.

3.4.4. Simulation Results of Daylight Autonomy

This section examines variations in indoor daylight autonomy (DA) to evaluate the impacts of the building layout, orientation, and urban environment on indoor lighting conditions. Figure 14 describes the distribution of daylight autonomy (DA) across building orientations and floors for Case 1 and Case 2. Overall, DA values increase with floor height in both cases, reflecting reduced shading at higher elevations. Case 1, with a higher building density (0.64) and lower sky view factor (SVF = 0.22), experiences severe ground-level shading, particularly in directions D2 and D3, where DA falls below 5%. Moderate improvements are observed on upper floors, peaking around 19% at the top in southwestern orientations, though the northern directions remain constrained due to dense urban surroundings. In contrast, Case 2’s lower density (0.47) and higher SVF (0.37) allow for better daylight access and a more uniform distribution. Ground floor DA ranges from 5–10% across most orientations, while upper floors consistently exceed 12%, particularly in southern directions. However, localized obstructions still limit performance in directions such as D8.

Overall, Case 1 shows greater directional and vertical variability in DA due to its compact urban form, while Case 2 benefits from a more open configuration that enhances daylight penetration. These differences highlight the role of urban morphology in shaping indoor daylight conditions and inform passive design considerations for optimizing natural lighting.

Table 5. Relationships between urban form indicators and performance metrics.

Performance	Case 1	Case 2
UTCI (highest value)	33.7 °C (high density, lower SVF, and limited tree coverage)	>34 °C (higher sky view factor and insufficient shading, increased solar exposure, and localized heat accumulation)
EUI	172.53 kWh/m2/year (higher cooling and equipment use due to a dense morphology and higher FAR)	161.20 kWh/m2/year (lower density, lower FAR, and higher lighting demand)
Cooling load	4000–11,000 kWh (higher cooling loads due to a dense morphology and lower tree coverage)	6000–9000 kWh (lower cooling loads due to better ventilation and higher tree coverage)
DA	3–24% (limited by a high building density, shading, and average building height)	5–18% (lower density and shorter buildings, and constrained by a poor orientation)

synthesizes the relationship between urban form characteristics and key performance indicators across the two representative cases. Case 1, characterized by a high density, low sky view factor (SVF), and limited tree coverage, exhibits a reduced UTCI due to shading but experiences higher cooling loads (up to 11,000 kWh), elevated EUI (172.53 kWh/m²/year), and poor daylight autonomy (3–24%) driven by obstructed ventilation and limited daylight access. In contrast, Case 2’s lower density, higher SVF, and greater vegetation coverage enhance ventilation, reduce the cooling demand (6000–9000 kWh), and lower the EUI (161.20 kWh/m²/year). However, its increased sky openness leads to a higher UTCI (>34 °C) in exposed zones and DA remains suboptimal (5–18%) in certain orientations due to inadequate shading. These findings highlight the trade-offs between compactness and openness: while dense forms provide thermal buffering, they hinder airflow and daylight; open forms enhance ventilation but risk solar overheating. Balanced urban design optimizing SVF, integrating vegetation, and refining the orientation is essential to mitigate the UHI, improve comfort, and reduce energy use in high-density contexts.

Although this study provides a robust framework for analyzing indoor and outdoor energy and environmental performance, it has certain limitations. (1) This study does not comprehensively assess the evapotranspiration effects of vegetation, despite its well-documented ability to significantly mitigate the urban heat island effect and improve thermal comfort. (2) For the case block near the coastline, the potential impact of water bodies on the local microclimate, such as the cooling effect, was not considered in this study. A more detailed evaluation of vegetation evapotranspiration and water body influences will be incorporated into future research.

4. Conclusions

This study establishes a cross-scale analytical framework that bridges remote sensing-derived outdoor thermal patterns with simulation-based assessments of indoor energy performance and comfort in high-density urban environments. By integrating a Landsat 8-based LST analysis with detailed microscale simulations of UTCI, EUI, cooling load, and DA, this research provides new insights into how urban morphology governs both environmental exposure and building performance.

The results demonstrate that compact, high-rise typologies benefit from self-shading, reducing the UTCI in ground-level spaces, but simultaneously experience an elevated cooling energy demand (up to 11,000 kWh) and low daylight access (DA 3–24%) due to limited ventilation, a reduced sky view factor (SVF), and poor solar orientation. In contrast, open, mid-rise configurations with a higher SVF and vegetation coverage exhibit improved ventilation and a reduced EUI (161.20 kWh/m²/year), but also suffer from a higher UTCI (>34 °C) and increased artificial lighting demand in poorly oriented units. These findings reveal that no single urban form performs optimally across all indicators. Instead, climate-adaptive urban design must balance the building density, spacing, height, and orientation to mitigate heat stress while enhancing energy performance and daylight access. Specific recommendations include (1) maintaining a minimum building spacing of 15–20 m to facilitate ventilation and reduce heat accumulation; (2) orienting primary façades 30–60° from west to avoid peak solar gains; (3) integrating continuous tree belts and vertical greenery to support both shading and evapotranspiration cooling; and (4) regulating SVF between 0.4 and 0.6 to balance solar access and thermal comfort.

For effective policy implementation, these recommendations should be translated into urban planning codes and building guidelines, e.g., by embedding SVF and DA thresholds into daylighting and energy benchmarks, or incentivizing façade greening through development bonuses. Coordinated action between planning authorities, architects, and environmental regulators is essential to operationalize performance-driven morphological planning. The proposed framework not only quantifies the trade-offs between microclimate mitigation and indoor performance but also offers a pathway for integrating spatial and energy strategies in future urban policy.

As an initial reference, this study suggests three key directions for future research: (1) climate adaptability—validating morphological thresholds (e.g., SVF and spacing) across different climatic zones; (2) occupant behavior coupling—integrating dynamic user actions (e.g., shading and HVAC setpoint adjustments) into simulation workflows; and (3) policy integration—assessing the impacts and feasibility of strategies (e.g., façade greening incentives) through urban-scale modeling.