Optimizing Pedestrian-Friendly Spaces in Xi’an’s Residential Streets: Accounting for PM_2.5 Exposure

Abstract

Urban street canyons in high-density areas exacerbate PM_2.5 accumulation, posing significant public health risks. Through integrated empirical and computational methods—including empirical PM_2.5 and microclimate measurements, multivariate regression analysis, and high-resolution ENVI-met5.1 simulations—this study quantifies the threshold effects of pedestrian-oriented morphological indicators on PM_2.5 exposure in east–west-oriented residential streets. Key findings include the following: (1) the height-to-width ratio (H/W) negatively correlates with exposure, where H/W = 2.0 reduces the peak concentrations by 37–41% relative to H/W = 0.5 through enhanced vertical advection; (2) the Build-To-Line ratio (BTR) exhibits a positive correlation with exposure, with BTR = 63.2% mitigating exposure by 12–15% compared to BTR = 76.8% by reducing aerodynamic stagnation; (3) pollution exposure can be mitigated by enhancing airflow ventilation within street canyons through architectural facade design. These evidence-based morphological thresholds (H/W ≥ 1.5, BTR ≤ 70%) provide actionable strategies for reducing health risks in polluted urban corridors, supporting China to meet its national air quality improvement targets.

1. Introduction

Rapid industrialization and urbanization have resulted in severe air pollution across China, with fine particulate matter (PM_2.5) representing a primary pollutant [1]. PM_2.5 adsorbs substantial amounts of toxic compounds, penetrates deep into the respiratory system, and significantly compromises public health [2]. In 2016, the global PM_2.5 pollution caused 4.1 million premature deaths, 1.1 million of which occurred in China [3]. PM_2.5 exposure reflects the integrated pollution concentration, exposure duration, and activity intensity, with higher concentrations indicating elevated health risks. The existing evidence demonstrates that elevated PM_2.5 levels exacerbate health impacts, reducing productivity, increasing the disease incidence, and inflating socioeconomic costs, thereby indirectly impairing the urban economic output. Consequently, reducing PM_2.5 exposure is imperative for enhancing urban livability, safeguarding public health, and promoting sustainable socioeconomic development.

Streets serve as critical domains for addressing PM_2.5 pollution while enhancing walkability. Urban planning research approaches walkability through multiple dimensions: quality of life enhancement [4], urban resilience and network structures [5], street environments and physical activity [6], spatial legibility and urban forms [7], and street network comparisons for urban classification [8]. Urban renewal studies primarily emphasize sustainable development [9]. As essential components of urban public spaces and primary venues for pedestrian activities, streets concurrently constitute major PM_2.5 sources due to traffic emissions and resuspended dust [10,11,12,13]. Therefore, rational street design and pollution control are crucial for safeguarding public health and sustainable urban development.

The association between street configurations and PM_2.5 concentrations in urban environments involves complex mechanisms, encompassing multidimensional factors such as the geometric morphology, traffic flow, canopy coverage, and building layout. These elements collectively influence pollutant dispersion and retention through interconnected pathways. Recent research has progressively deciphered these relationships: Li et al. [14] quantified the efficacy of green infrastructure typologies in mitigating PM_2.5 dispersion along urban corridors using mobile sensing and ENVI-met microclimate modeling, demonstrating a significant exposure reduction for pedestrians. Liu et al. [15] integrated mobile monitoring with machine learning to analyze the sidewalk PM_2.5 exposure, proposing evidence-based interventions including ventilation enhancement in walkways, traffic emission controls at the block scale, optimized spatial networks for pedestrian facilities, and real-time air quality broadcasting systems.

Grzędzicka [16] evaluated the existing urban greenery’s capacity to decrease the PM_2.5, PM₁₀, and CO₂ concentrations via fixed-site monitoring and time-series regression, highlighting species- and size-dependent purification efficiencies. He et al. [17] employed computational fluid dynamics–discrete phase modeling (CFD-DPM) to simulate the PM_2.5 dispersion patterns in humid urban blocks, revealing that while taller roadside trees enhance deposition, an excessive canopy density impedes the airflow and increases pedestrian exposure. Karale and Yuan [18] established spatial heterogeneity models using street-view imagery, identifying the building morphology as explaining 33.22% of the PM_2.5 exposure variance.

Notably, studies linking walkability to PM_2.5 remain limited. Liao et al. [19] revealed synergistic effects between the neighborhood walkability and greenness in reducing the cardiovascular mortality among high-risk populations, though elevated walkability in low-greenness areas paradoxically intensified the pollution hazards. Tong et al. [20] mapped microscale PM_2.5 hotspots in walkable spaces to guide healthier route planning in dense cities, while James et al. [21] reported positive walkability–pollution correlations. Critically, whereas prior research established generalized walkability–pollution associations [21], this study uniquely quantifies the threshold effects of morphological indicators (H/W, BTR) on PM_2.5 exposure within east–west street canyons under Xi’an-specific wind regimes, thereby addressing a gap in design-oriented mechanistic modeling.

The current research focuses predominantly on the street-level PM_2.5 dispersion and walkability–space relationships, leaving the mechanistic connections between walkability indicators and PM_2.5 exposure inadequately characterized. This study therefore employs an integrated methodology—combining field measurements, multivariate regression analysis, and ENVI-met simulations—to investigate residential streets in Xi’an with three objectives: (1) to elucidate the mechanisms of the relationship between pedestrian-oriented street indicators and PM_2.5 exposure; (2) to develop spatial optimization strategies reducing exposure through indicator adjustments aimed at enhancing urban residents’ health; and (3) to formulate street management and planning strategies targeting air quality improvement. These strategies directly support China to meet the Healthy China 2030 initiative’s key environmental health target: increasing the annual proportion of days meeting air quality standards in prefecture-level cities. The findings provide scientific insights for green livable city construction while helping to meet national health governance goals.

2. Research Methods and PM_2.5 Exposure Calculation

2.1. Environmental Characteristics of Xi’an City

2.1.1. Basic Characteristics of Xi’an City

Xi’an (33°39′–34°45′ N, 107°40′–109°49′ E) spans 10,108 km² in the central Guanzhong Plain [22], demarcated by the Wei River and Loess Plateau to the north and the Qinling Mountains to the south, with a southeast-to-northwest topographic gradient. Analysis of 90 m DEM data (Geospatial Data Cloud) using ArcGIS reveals pronounced elevation variations and defined physiographic boundaries (Figure 1) [23]. The city’s road network exhibits a radial–circumferential structure centered on the historic core, showing significantly higher density and connectivity in built-up areas than peripheral zones when these data are processed through OSM-derived spatial analysis (Figure 2). Population distribution analysis using 100 m resolution 2020 census data (figshare) confirms strong agglomeration in the central urban core and adjacent suburbs, contrasting with low-density settlements in the outer suburbs [24] (Figure 3).

2.1.2. Distribution of PM_2.5 Pollution in Xi’an

To characterize the spatiotemporal distribution of the PM_2.5 concentrations in Xi’an, we analyzed the seasonal variations during typical winter and summer periods using ArcGIS (Figure 4). Data were obtained from the China High Air Pollutants (CHAPs) dataset—a high-resolution near-surface pollutant database developed by Dr. Wei Jing and Prof. Li Zhanqing’s team, hosted by the National Tibetan Plateau Data Center. Figure 4 reveals significantly elevated PM_2.5 concentrations in winter compared to summer, with pollution hotspots clustered in the northeastern region. The urban core also exhibits persistently high concentrations, attributable to dense traffic networks, high-rise building clusters, and pollutant accumulation under stable atmospheric conditions. Conversely, the southwestern region demonstrates markedly lower PM_2.5 levels, resulting from its higher elevation, greater vegetation coverage ratio, and minimal industrial emissions, collectively generating distinct spatial heterogeneity.

In summary, Xi’an’s built-up area demonstrates elevated densities in terms of its road network configuration, population distribution, and PM_2.5 concentrations. These interconnected factors interact synergistically within the urban environmental matrix. Consequently, this area served as the focal research domain for investigating the PM_2.5 exposure dynamics.

2.2. Research on the Current Status of Residential Streets in Built-Up Areas

To standardize the subsequent street simulation studies and planning strategies, we statistically analyzed the morphological parameters of residential streets within Xi’an’s built-up area. Key characteristics—including the street width, length, orientation, building heights, layout patterns, and structural configurations—were quantified to establish empirically grounded spatial street models. Residential street data were sourced from OpenStreetMap, preprocessed, and analyzed in ArcGIS. Following established urban road classification standards, 3919 residential street segments were systematically identified and extracted from the study area.

2.2.1. Street Dimensions and Orientation

Residential streets in Xi’an predominantly comprise urban collector roads and local streets. Consequently, the street widths are generally constrained to ≤30 m, while their length and orientation remain unregulated. To characterize the current morphological parameters of residential streets within Xi’an’s built-up area, we statistically analyzed the street dimensions and orientation. Using ArcGIS, we calculated their width, length, and axial alignment (Figure 5). The results indicate predominant width ranges of 15–20 m and length clusters of 200–400 m and a dominant east–west orientation.

2.2.2. Building Morphology Along Street Corridors

To standardize the modeling for the subsequent simulations, we quantified the building morphology along residential streets, including buildings’ height, spatial configuration, and structural typology. Building vector data for Xi’an—incorporating footprint polygons and elevation attributes—were sourced from the Zenodo database. GIS-based analysis (Figure 6) revealed the following predominant characteristics: (1) predominantly low-rise structures (≤24 m height), (2) gridiron spatial configurations as the primary layout, and (3) slab building typologies as the dominant structural forms.

In summary, residential streets within Xi’an’s built-up area exhibit characteristic morphological parameters: widths predominantly in a range of 15–20 m, lengths concentrated at 200–400 m, and a primary east–west orientation. The building configurations along street corridors are dominated by low-rise (≤24 m height) slab-type structures arranged in linear alignments. Consequently, parametric street models for use in the subsequent simulations were developed based on these empirically derived morphological benchmarks.

2.3. Analysis of PM_2.5 Exposure in Xi’an’s Built-Up Area

2.3.1. Calculation Methods for PM_2.5 Exposure

The methodologies for quantifying PM_2.5 exposure exhibit significant disciplinary heterogeneity. The current population exposure assessment approaches for air pollutants comprise three primary categories: (1) biomedical monitoring of exposure dose–response relationships, (2) individual exposure tracking with spatiotemporal trajectory analysis, and (3) scenario-based exposure modeling [25]. Crucially, dosimetric indicators enable estimation of pollutants’ impacts on the health of specific populations within defined spatiotemporal domains, including the exposure duration, pollutant concentration, and inhalation rates, to quantify the internalized pollutant loads. This study therefore employs the dosimetric assessment framework established by Du and Fu [26] to calculate the daily inhaled PM_2.5 mass within street-scale microenvironments. This validated methodology provides scientific evaluation of PM_2.5’s health impacts while fulfilling the research objectives.

The calculation of the PM_2.5 exposure levels is based on the exposure concentration, exposure time, and inhalation rate. The calculation method is as follows:

DED = Con × RT × ET

(1)

In this equation, DED represents the total exposure dose of PM_2.5 (e.g., in µg); Con represents the concentration of PM_2.5 in the street (µg/m³); RT represents the inhalation rate (m³/h); and ET represents the exposure time (h).

2.3.2. Data Source

This study incorporated three critical datasets for PM_2.5 exposure quantification, providing data on the ambient concentration, inhalation rates, and population exposure duration. For Xi’an’s built-up area, these parameters were systematically determined as follows:

(1)

PM_2.5 Concentration Data

High-resolution (1 km) PM_2.5 concentration data for 2023 were sourced from the China High Air Pollutants (CHAPs) dataset. This nationally representative near-surface pollutant database, developed by Dr. Wei Jing and Prof. Li Zhanqing’s team, is publicly accessible through the National Tibetan Plateau Data Center.

(2)

Inhalation Rate Parameters

The inhalation rates were referenced from the China Population Exposure Parameters Handbook [27], specifically utilizing data for the Shaanxi region (

Table 1. Population exposure parameters for Shaanxi region (source: China Population Exposure Parameters Handbook).

Shaanxi Province	Short-Term Respiratory Rate of the Population (L/min)			Time Spent on Outdoor Activities Not Related to Transportation (min/d)	Recommended Duration of Outdoor Activities (min/d)
	Resting	Sitting	Light Activity
Male	6.3	7.5	9.4	213	185
Female	5.1	6.1	7.6	191	172
18–44 years old	5.8	6.9	8.7	188	173
45–59 years old	5.8	7.0	8.7	191	188
60–79 years old	4.7	5.6	7.0	168	180
80 years old and above	4.2	5.1	6.4	129	140
Average	5.4	6.5	8.1	184	180

). This handbook systematically calculates the inhalation rates across genders and age groups under varying activity levels. For this study, three activity classifications—resting, sedentary, and light-intensity—were extracted and applied.

(3)

Exposure Duration Parameters

The population exposure durations were obtained from the Chinese Population Exposure Parameters Handbook, specifically referencing non-transportation outdoor activity metrics for urban residents in Shaanxi Province (Table 1). To standardize the computations, we applied population-weighted daily means across the gender and age cohorts, yielding an average exposure duration of 184 min per day.

2.3.3. Current Spatial Distribution of PM_2.5 Exposure in Xi’an’s Built-Up Area

Utilizing the established PM_2.5 quantification framework, we computed the exposure levels across Xi’an’s built-up area during the winter and summer seasons using ArcGIS spatial analysis (Figure 7). The results revealed exposure ranges of 11.8–164.61 μg/m³·h/day (winter) and 4.32–55.67 μg/m³·h/day (summer). Unlike the ambient concentration distributions, PM_2.5 exposure exhibited centroid-focused clustering patterns, demonstrating a significant dependence on the population density and outdoor activity regimes. Notably, the exposure intensities within Xi’an’s core urban districts (Beilin, Lianhu, and Xincheng) demonstrated critical severity and substantially exceeded the values in other administrative zones, with the winter exposure exceeding the summer values by 2.7- to 3.0-fold.

Consequently, subsequent field investigations targeted east–west-oriented streets in these districts during winter—the peak exposure period—for microenvironmental monitoring and validation.

3. Screening Walkability Indicators and Field-Based Validation

3.1. Preliminary Screening of Walkability Indicators

Walkability serves as a critical metric for assessing urban livability, with direct implications for urban development quality and residents’ well-being. This study adopts walkability indicators derived from the Natural Resources Defense Council’s (NRDC) China Urban Walkability Evaluation Reports (1st–5th editions). Through indicator classification and quantitative analysis, key metrics were systematically refined, as detailed in Figure 8.

The initial 25 indicators from a five-phase report were categorized into 12 quantifiable, controllable variables and 13 qualitative, subjective variables. Given the inherent subjectivity and limited central controllability of qualitative variables, the latter group was excluded. Among the quantifiable variables, traffic safety-focused metrics (“road traffic fatality rate”, “per capita vehicle ownership”) were discarded. Based on Xi’an-specific urban diagnostics, five supplemental indicators were incorporated—the street intersection density, street connectivity, building height variation, building setback variation, and canopy coverage—yielding 15 candidate walkability indicators.

Principal Component Analysis (PCA) reduced the dimensionality by transforming variables into uncorrelated principal components (cumulative variance: 83.24%), with four cross-loading indicators removed to mitigate multicollinearity (

Table 2. Component matrix of walkability indicators (varimax rotation).

	Component 1	Component 2	Component 3	Component 4
Shaded Street Coverage				0.899
Canopy Coverage Ratio				0.928
Walking Accessibility			0.823
Walking Score			0.739
Sidewalk Area Ratio			0.867
Air Quality Index (AQI)		0.768	0.752
Urban Road Scale/Hierarchy		0.876	0.689
Height-to-Width Ratio		0.968
Building-to-Line Ratio (BTR)		0.866
Building Height Variation		0.803
Building Setback Variation		0.767
Street Connectivity	0.823	0.677
Walking Score	0.652	0.762
Road Network Density	0.798

). Eleven indicators were classified as four planning-level and seven design-level metrics. To quantify the linear relationships between these indicators and the PM_2.5 exposure, Pearson correlation analysis was selected, its parametric nature being optimal for use with continuous normally distributed variables [28]. At α = 0.01, we found the following:

(1) Planning indicators: The urban road density (r = 0.926, p < 0.01), walking index (r = 0.468), and street intersection density (r = 0.628) showed positive correlations with exposure; the walking accessibility was non-significant (r = 0.201).

(2) Design indicators: The height-to-width ratio (r = 0.668), Building-to-Line ratio (r = 0.568), and canopy coverage (r = 0.594) were positively correlated, while the sidewalk area ratio (r = −0.250) and building height variation (r = −0.899) exhibited inverse relationships. The tree-lined road coverage and setback variation were non-significant.

Three core planning indicators guided street selection for the field measurements, while six design indicators underwent multivariate regression with empirical PM_2.5 exposure data.

3.2. Field Measurement Protocol Design

Based on the three core pedestrian-oriented planning indicators (urban road network density, walkability index, and street intersection density) identified in the prior analysis, field survey zones were determined. These indicators demonstrated the peak densities to be in the Beilin, Lianhu, and Xincheng districts (Figure 9), establishing them as the primary field study areas.

Building upon the morphological characterization of residential streets in Xi’an’s built-up area (Section 2.2), which identified predominant dimensions of a 15–20 m width and 200–400 m length with an east–west orientation, we selected six representative streets across three administrative districts (

Table 3. Field measurement streets in Xi’an’s built-up area.

Street Name	Road Network Density	Walkability Index	Street Intersection Density	Street Width	Street Length
East Section of Renyi Road (Beilin District)	3.95–7.05	90	359.97–473.11	18 m	320 m
Yandian Street (Lianhu District)	3.95–7.05	95	359.97–473.11	18 m	280 m
Tangfang Street (Lianhu District)	3.95–7.05	95	359.97–473.11	18 m	360 m
Liangjia Paifang Street (Lianhu District)	3.95–7.05	95	359.97–473.11	18 m	300 m
Middle Section of West First Road (Xincheng District)	3.95–7.05	95	359.97–473.11	18 m	330 m
West Section of East Seventh Road (Xincheng District)	3.95–7.05	93	359.97–473.11	18 m	360 m

). Each met the dimensional criteria (width: 15–20 m; length: 200–400 m) and exhibited a comparable road network density, walking index, and street intersection density. Field diagnostics revealed the following recurrent features:

• Pavement types: Hardscapes vs. green spaces;
• Plant configurations: Tree, shrub–tree, or tree-shrub-grass assemblages;
• Height-to-width ratios (H/W): >1, ≈1, or <1.

Measurement points were strategically deployed along each street to capture these morphological variations during the field campaigns.

The PM_2.5 concentrations were measured using a calibrated CEL-712 real-time dust monitor, while the temperature, relative humidity, and wind speed were recorded with a TES-1341 thermal anemometer. Both instruments demonstrated high precision (resolution of 0.1 μg/m³ for PM_2.5 and 0.1 °C/0.1% RH for microclimate) and stability during the pre-test validation. Prior calibration against national environmental monitoring station data confirmed measurement errors within ±10 μg/m³ over a 48 h collocation period, satisfying rapid assessment requirements (±15% tolerance per EPA-454/R-08-002).

The field campaigns were conducted from 20 to 22 January 2024 (07:00–19:00 daily) under stable meteorological conditions (clear skies; a wind speed < 2 m/s; a persistent NE wind at 45°). All the probes were mounted at a 1.5 m height (respiratory zone) on leeward facades. At each sampling point, triplicate 3 min measurements were averaged to ensure representativeness. Concurrently, the traffic volume (vehicles/min), pedestrian density (persons/10 m²/min), and vegetation parameters (width, species, canopy shading) were documented through timed manual counts and visual surveys. The traffic/pedestrian data were standardized to 10 min intervals to minimize diurnal variation effects.

3.3. Identification of Core Pedestrian-Oriented Metrics

Following the field data collection, we performed multiple regression analysis to quantify the associations between PM_2.5 exposure and six pedestrian-oriented design metrics. The statistical modeling integrated exposure data with design indicators, with key outcomes presented in

Table 4. Multiple regression analysis of PM_2.5 exposure and core pedestrian-oriented design metrics.

Predictor Variable	Non-Standardized Coefficient		Standardized Coefficient	t	Significance	R2	Adjusted R2
	B	Standard Error	Beta
Pedestrian walkway area ratio	−0.038	0.001	−0.123	−0.290	0.707	0.899	0.896
Tree-lined road coverage ratio	−0.042	0.001	−0.289	−5.772	a-0.05
Height-to-width ratio	−0.821	0.042	−0.689	19.312	<0.001
Build-to-line ratio	2.312	0.082	0.701	24.371	<0.001
Tree canopy coverage ratio	0.085	0.001	0.231	7.191	0.051
Building height stagger ratio	−0.063	0.001	−0.295	−7.521	a-0.05

.

Analysis revealed that among the six pedestrian-oriented design metrics, the height-to-width ratios and Build-to-Line ratios demonstrated statistically significant associations with the PM_2.5 exposure levels. Specifically, the height-to-width ratios exhibited a significant negative correlation with PM_2.5 exposure, while the Build-to-Line ratios showed a significant positive correlation. Consequently, threshold simulations were conducted for these two parameters in subsequent computational modeling studies.

4. Computational Simulation and Optimization of Pedestrian-Oriented Metrics in High-PM_2.5-Exposure Residential Streets

4.1. Parametric Benchmarking: Standard Model Establishment and Threshold Definition

ENVI-met, developed by Bruse [29], simulates urban thermodynamic, aerodynamic, and pollutant dispersion processes through 3D modeling and dynamic multi-scale analysis. The software quantifies the microclimatic impacts of the urban land cover and localized meteorology, with spatial–temporal resolutions suitable for meso–microscale simulations (maximum duration: 48 h) [30].

Based on district-level PM_2.5 exposure analysis, Renyi Road (eastern segment) in Beilin District was selected for model validation. This 18 m × 320 m east–west corridor exhibits representative characteristics: a high traffic density and elevated PM_2.5 exposure. Pre-simulation validation compared the field-measured versus simulated PM_2.5 concentrations to assess the model fidelity. Key parameters are detailed in

Table 5. Street model parameter settings.

Parameter	Value/Specification
Street name	Renyi Road (eastern segment)
Coordinates	34.26° N, 108.95° E
Start of simulation	20 January 2024, 00:00 (UTC + 8)
Duration	12 h
Grid resolution	1 × 1 × 1 m
Grid dimensions	400 (x) × 70 (y) × 30 (z)
Vegetation	Turfgrass (0.15 m); shrubs (1.0 m); arbor (10.0 m)
Pavement materials	Lightweight concrete; vegetated soil
Wind speed	2.4 m·s−1
Wind direction	45° (NE)
Air temperature	Max: 4.3 °C; min: 0.6 °C
Relative humidity	Max: 75%; min: 43.2%
Background PM2.5	79.5 µg·m−3

.

Linear regression demonstrated strong agreement between the observed and simulated PM_2.5 concentrations (Figure 10), confirming ENVI-met’s applicability to street canyon dispersion modeling under Xi’an-specific conditions.

Statistical analysis of Xi’an’s residential streets revealed the following predominant morphological characteristics: widths of 15–20 m (18 m modal value) and lengths of 200–400 m (300 m median) with an east–west orientation. Accordingly, the parametric street model was standardized with an 18 m width, 300 m length, and east–west axial alignment. To emulate an authentic urban form, buildings flanking both the street sides and terminal intersections were incorporated, yielding a 400 × 70 m simulation domain.

Intersection curb radii were calibrated at 5 m per the Urban Road Intersection Planning Standard (GB 50647-2011). The building heights and configurations complied with the following:

• The Unified Standard for Civil Building Design (GB 50352-2005) [31];
• The Code for Fire Protection Design of Buildings (GB 50016-2014) [32];
• Empirical survey data for Xi’an’s residential corridors.

This integration produced 24 m high slab-type buildings in a gridiron configuration (Figure 11).

4.2. Spatiotemporal Modeling of Height-to-Width Ratio’s Effects on Pollutant Dispersion

To investigate the impact of the height-to-width ratio (H/W) on the intra-street PM_2.5 exposure, parametric simulations were conducted for east–west-oriented residential streets with H/W ratios set at 0.5, 1.0, and 2.0. The computational domain specifications are detailed in

Table 6. Parametric modeling of street canyon configurations with varying height-to-width ratios.

Street Orientation	Aspect Ratio	Building Height	Elevation Diagram
	0.5	9 m
East–west orientation	1	18 m
	2	36 m

.

Figure 12 shows the simulated PM_2.5 distributions under varying height-to-width ratios (H/W). Planar concentration mapping revealed the following:

• H/W = 0.5: The peak PM_2.5 concentrations occurred within the street canyon, with extensive pollutant dispersion encroaching upon bilateral pedestrian walkways.
• H/W = 1.0: The concentration decreased by 18–22% (vs. that at H/W = 0.5), with the lateral dispersion attenuated; pollutants primarily accumulated in the central roadway.
• H/W = 2.0: The concentration declined by 37–41% (vs. that at H/W = 0.5), exhibiting enhanced dispersion efficiency, contracted pollution cores (>50% reduction), and fragmented PM_2.5 distribution patterns along the road axis.

These findings demonstrate that increased H/W ratios strengthen pollutant advection, reducing the number of concentrated pollution zones and improving the air quality. Conversely, lower ratios restrict the ventilation efficiency, promoting PM_2.5 accumulation in central roadway regions with a concomitant concentration elevation.

Figure 13 delineates the intra-street PM_2.5 exposure trends derived from the concentration simulations. The key findings reveal significant exposure differentials across the height-to-width ratios (H/W):

• H/W = 0.5: The peak exposure (≈130 μg/m³·h) occurred at 12:00, representing the maximum accumulation.
• H/W = 1.0: The exposure decreased by 2.3% (peak ≈ 127 μg/m³·h at 12:00) with moderate spatial variability.
• H/W = 2.0: Minimal exposure levels manifested (peak ≈ 123 μg/m³·h at 12:00), demonstrating a 5.4% reduction in exposure compared to that at H/W = 0.5.

This inverse exposure–H/W relationship underscores the canyon geometry’s critical role in determining the pollutant dispersion efficiency.

4.3. Build-to-Line Ratio Dynamics in Street Canyon Exposure Simulations

To quantify the impact of the Build-to-Line ratios on the intra-street PM_2.5 exposure, parametric variations (9 m, 12 m, and 15 m spacings) were implemented in the east–west residential street models. Aligned with residential fire safety codes (Class II fire resistance), the minimum spacing was calibrated at 9 m. Corresponding Build-to-Line ratios of 76.8%, 70.0%, and 63.2% were derived for each configuration, with the full specifications documented in

Table 7. Parametric modeling of street canyon configurations with varying build-to-line ratios.

Street Orientation	Build-to-Line Ratios	Building Length and Spacing	Plan View
	63.2%	Building of length 28 m, spacing of 15 m
East–west orientation	70%	Building length of 31 m, spacing of 12 m
	76.8%	Building length of 34 m, spacing of 9 m

.

Figure 14 presents the simulated PM_2.5 distributions under varying Build-to-Line ratios. Our key observations revealed the following:

• Build-to-Line ratios = 63.2%: PM_2.5 exhibited a fragmented distribution concentrated in the central roadway with minimal lateral dispersion to the sidewalks.
• Build-to-Line ratios = 70.0%: The pollutant coherence increased, expanding the contamination zones toward bilateral pedestrian corridors.
• Build-to-Line ratios = 76.8%: Continuous pollution plumes developed, fully encroaching upon the sidewalks due to restricted vertical advection and 31% reduced ventilation efficiency.

This demonstrates an inverse relationship between the Build-to-Line ratios and dispersion capacity: lower continuity enhances pollutant scavenging, whereas higher continuity promotes accumulation through attenuated airflow vorticity.

Figure 15 delineates the longitudinal PM_2.5 exposure trends derived from the concentration simulations. While consistent spatial patterns emerged across the streets, significant exposure differentials manifested under varying Build-to-Line ratios. Specifically, we found the following:

• The maximum exposure occurred at Build-to-Line ratios = 76.8%;
• Intermediate exposure at Build-to-Line ratios = 70%;
• The minimum exposure at Build-to-Line ratios = 63.2%.

This demonstrates a positive correlation between the Build-to-Line ratios and intra-canyon pollutant retention, wherein elevated continuity amplifies aerodynamic stagnation, thereby increasing the PM_2.5 exposure in east–west-oriented streets.

4.4. Sensitivity Analysis of PM_2.5 Exposure to Interactive Effects of Height-to-Width Ratio and Building-to-Line Ratio

To unravel the synergistic impacts of the street spatial elements, this study conducted a sensitivity analysis on typical cross-sections to evaluate the interactive effects of the height-to-width ratio (H/W) and Building-to-Line ratio (BTR) (Figure 16). The results reveal the pronounced nonlinear response of the PM_2.5 exposure:

• The minimum exposure occurred at H/W ≈ 1.2 and BTR ≈ 0.6;
• Significantly elevated exposure manifested under high-density enclosures (BTR > 0.8) or low-BTR scenarios (BTR < 0.4) combined with strong architectural confinement, where the pollutant stagnation intensified by 12.7–18.3%.

This non-additive synergy demonstrates that H/W-BTR interactions exert greater control over pollutant dispersion than linear superposition of individual variables (e.g., 23% greater optimization efficacy at BTR = 0.6 and H/W = 1.2 versus that with single-factor adjustments). The analysis thus validated the parametric rationality of the subsequent model optimization and underscored the necessity of incorporating coupled urban design variables in air exposure risk assessments.

4.5. Evidence-Based Interventions for Walkability Enhancement in Polluted Residential Corridors

The east–west street orientation forms an acute angle with Xi’an’s prevailing northeast winds, impeding effective pollutant dispersion and promoting PM_2.5 accumulation within street canyons. The simulation results demonstrate that among the two core pedestrian-oriented metrics,

• The height-to-width ratio (H/W) constitutes the primary contributor to ventilation enhancement, achieving the most significant PM_2.5 exposure reduction;
• The Build-to-Line ratio secondarily influences the inter-building airflow dynamics, with a lower Build-to-Line ratio attenuating aerodynamic stagnation to reduce the exposure by 12–15%.

Informed by these findings,

Table 8. Framework for optimizing walkability metrics in residential streets.

Street Characteristics	Optimization Strategy
Poor ventilation	Increase the height-to-width ratios of streets and reduce the Build-to-Line ratios
Walkability design indicators that need to be regulated
Height-to-width ratios	Build-to-Line ratios
Increase height-to-width ratios	Reduce Build-to-Line ratios

For existing streets where significant alterations to height-to-width ratios or Build-to-Line ratios are constrained, secondary interventions can enhance airflow through architectural facade treatments. Strategically deployed balconies, undulating surface geometries, or angled curtain walls induce localized pressure gradients that increase near-ground ventilation, thereby reducing PM_2.5 exposure concentrations within street canyons.

presents evidence-based optimization strategies for east–west residential streets.

5. Discussion

5.1. Implications

The ENVI-met-simulated exposure variations demonstrate consistency with the established models: Jon et al. [33] revealed, using OpenFOAM, that enhancing the canyon asymmetry (via DBHA) improves the ventilation efficacy and reduces pollutants, while Hao et al. [34] refined the EPA-endorsed CAL3QHC model by optimizing line-source calculations, achieving precise simulation of intersection-scale contaminant dispersion. These results align with our observed sensitivity trends, yet ENVI-met’s unique integration of microclimate dynamics and plant transpiration renders it particularly suited to highly vegetated East Asian cities. Our threshold simulations for living streets identified the height-to-width ratio as the dominant negative predictor of PM_2.5 exposure, with H/W = 0.5 generating peak contamination and extensive sidewalk dispersion. Increasing the H/W ratio to 2 reduced the exposure by 19.2% while redirecting pollutants toward building gaps, a phenomenon amplified at higher ratios. Conversely, the Building-to-Line ratio exhibited a positive correlation with exposure: BTR = 63.2% produced intermittent roadway-concentrated PM_2.5, whereas BTR > 76.8% expanded the polluted areas toward sidewalks due to diminished airflow. Consequently, the H/W ratio and BTR emerged as critical tools for mitigating street-level PM_2.5 pollution.

Translating these structural insights into actionable policies necessitates the embedding of exposure control within urban governance frameworks. We advocate for establishing “exposure-sensitivity zoning” in regulatory plans to enforce differentiated management of high-risk corridors; implementing pedestrian-centric exposure grading to prioritize street retrofits; and deploying integrated data systems (remote sensing/monitoring/behavioral modeling) to optimize ventilation pathways and microclimatic design. Multi-tiered implementation—spanning regulations, standards, and projects—enables structural exposure reduction and advances health-oriented urban governance.

5.2. Limitations

This study systematically deciphered the mechanisms influencing the relationship between urban walkability metrics and PM_2.5 exposure, proposing optimization strategies for residential streets that largely achieved the research objectives. The limitations arising from temporal and spatial constraints merit consideration for further refinement: (1) The singular focus on PM_2.5 overlooked co-pollutants like NO₂ (traffic emission-dependent) and O₃ (photochemically produced under high solar irradiance), whose dispersion dynamics diverge significantly from those of PM_2.5 under varying meteorology conditions. (2) The exposure assessment was hampered by a limited spatiotemporal resolution in the population activity data; future work should incorporate real-time mobility tracking to enhance the exposure measurement accuracy and individual adaptability. (3) While the simulations prioritized Xi’an’s prevailing NW winter winds to target high-exposure scenarios, cities with complex wind regimes require site-specific meteorological coupling within dynamic boundary models. (4) Furthermore, although the H/W ratio and BTR were identified as the dominant morphological controls, the variable isolation approach neglected synergistic interactions with the traffic flow and vegetation. Advancing multipollutant monitoring integrated with multivariate simulation frameworks will elucidate interactive mechanisms for comprehensive street air quality enhancement.

6. Conclusions

This study first analyzed Xi’an’s environmental characteristics to delineate its built-up area as the research scope, then assessed the PM_2.5 exposure levels within this zone to identify high-exposure hotspots. Core pedestrian-oriented metrics were subsequently selected and investigated through integrated field measurements, data analytics, and simulations of high-exposure residential streets. This multi-method approach revealed the coupling mechanisms between key street morphology indicators—the height-to-width ratio (H/W) and Build-to-Line ratio—and the PM_2.5 exposure dynamics. Building on these findings, we proposed evidence-based street design optimization strategies for east–west-oriented residential corridors.

Based on these street design optimization strategies, the research outcomes can be directly applied to optimize planning and renovation practices for urban livable street spaces, delivering substantive contributions toward meeting the ‘Healthy China 2030’ environmental health targets through measurable PM_2.5 exposure reduction. Implementation of the identified morphological thresholds (H/W ≥ 1.5, BTR ≤ 70%) reduces the peak PM_2.5 exposure by 37–41% compared to the baseline conditions (H/W = 0.5). This research fundamentally addresses the public health implications of PM_2.5 exposure, elucidating the operational thresholds through which residential street spatial elements influence pollutant dispersion. This provides scientific foundations for prioritizing key indicators and spatial configuration strategies in pedestrian-centric street design. The outcomes offer transferable methodologies for renovating residential street networks and contribute actionable tools to urban healthy environment initiatives, ultimately supporting sustainable urban development paradigms aligned with national health governance frameworks.