Air purification and cooling effects of green spaces: Mechanism,quantification,and interaction

Buffer analysis

GS not only lower land surface temperatures (LST) within their boundaries but also facilitate cooling in surrounding areas through heat exchange, thereby alleviating the UHI effect. Assessment of GS impacts on surrounding environmental temperatures is often conducted using remote sensing observations combined with multiple buffer zone analyses. This approach constructs temperature-distance curves based on temperature sequences at varying distances from the GS center or boundary. Typically, nonlinear relationships between temperature and distance are fitted using exponential or polynomial functions (Peng et al., 2024; Tan et al., 2021). In general, the cooling effect decreases with increasing distance from the GS, while ambient temperatures increase accordingly. Beyond a certain distance, the temperature stabilizes or even begins to decline. This distance threshold is defined as the turning point where ambient temperature stops increasing or where the slope of temperature-distance curve approaches zero—indicating the disappearance of the cooling effect (Peng et al., 2024). Based on this turning point, several quantitative metrics have been established to characterize the cooling effect of GS, including cooling distance (the distance from the GS boundary to the turning point), cooling range (the buffer zone between the GS boundary and the turning point), cooling intensity (the temperature difference between the turning point and the GS’s average or boundary temperature), and cooling gradient (the ratio of cooling intensity to cooling distance) (Chang and Li, 2014; Feyisa et al., 2014). Additionally, cumulative indices such as park cumulative cooling intensity and cooling gradient index provide a broader perspective on cooling effects (Peng et al., 2021). While this method effectively captures the influence of GS on surrounding temperatures, identifying turning points is time-consuming and labor-intensive, making it more suitable for small-scale studies such as urban parks and roadside greenbelts. Moreover, the quantified cooling effect is strongly influenced by surrounding land use types. A higher proportion of ecological land around the GS generally results in lower cooling intensity and gradient, whereas a higher proportion of built-up land leads to higher values (Cheng et al., 2015; Peng et al., 2024). Consequently, this approach may not always accurately reflect the intrinsic cooling capacity of GS.

Cooling efficiency assessment

The buffer zone method provides a detailed representation of the cooling effect of individual urban green spaces on their surroundings. However, it is insufficient for assessing the cooling effect of all green spaces at the city scale. Specifically, it does not address the key question of how much the citywide average temperature would decrease with a given increase in urban GS coverage. This limitation hinders its applicability in providing scientific support for urban GS planning. To quantify the impact of GS expansion on regional temperature reduction, the concept of cooling efficiency (CE) has been proposed. CE is defined as the negative ratio of air temperature or LST change to the fractional vegetation cover (FVC) change (-ΔLST/ΔFVC), representing the extent to which urban temperature decreases per 1% increase in vegetation coverage (Li et al., 2024c; Wang et al., 2019a, 2019b, 2024). However, this approach attributes all urban temperature variations solely to changes in FVC, neglecting other critical factors contributing to temperature heterogeneity. This may lead to an overestimation or underestimation of the CE (Cheng et al., 2022). To address this limitation, subsequent studies have developed multiple linear regression models with LST as the dependent variable and FVC, elevation, and nighttime light intensity as independent variables (Cheng et al., 2022, Yu et al., 2025a). The negative regression coefficient of FVC is then regarded as the adjusted cooling efficiency. By incorporating key factors affecting LST, such as topography and anthropogenic influences, these models have revealed that earlier estimates of CE tend to be overestimated (Yang et al., 2022). It is further indicated that CE is significantly influenced by both vegetation coverage levels and background climatic conditions (Cheng et al., 2022; Wang et al., 2019b, 2024). Consequently, CEs are only comparable when calculated under similar vegetation cover conditions and comparable background temperatures (Zhao et al., 2023).

Temperature difference-based quantification

GS’s cooling effect can be quantified by the average temperature difference between GS and other land use types. This method is relatively simple and easy to implement, making it suitable for large-scale analyses. For global-scale analysis, controlling for interference factors such as distance between the target and control groups, climatic background, and elevation is essential. A commonly used approach is the moving window analysis, which assumes that the climatic background is similar between target and control pixels. To further ensure comparability, control pixels are typically selected to have an elevation difference of less than 100 m from the target pixels (Li et al., 2015, 2023). Under these conditions, the observed temperature difference between target and control pixels can be attributed to the GS’s cooling effect. This method is particularly valuable for analyzing temperature effects associated with vegetation greening, degradation trends, and afforestation projects, where temporal trends are involved. It is also referred to as the “space-for-time” approach (Li et al., 2023; Peng et al., 2014; Zhu et al., 2023). Traditional time-series analyses struggle to distinguish the unidirectional impact of vegetation changes on local climate from satellite-derived vegetation indices and temperature observations due to their co-evolution (Gao et al., 2021; Li et al., 2018). In contrast, the space-for-time method enhances the ability to isolate the climatic effects of vegetation changes by controlling for spatial similarities between target and control pixels, thereby reducing the influence of complex factors such as climate variability (Li et al., 2023). At the urban scale, the temperature difference method is also widely used to assess the cooling effect of urban GS. For instance, comparative studies have demonstrated that GS exhibits significantly lower temperatures than built-up areas. Further analysis revealed that GS without trees were less effective in reducing LST, with their cooling effects being approximately two to four times weaker than those induced by urban trees (Schwaab et al., 2021). Despite efforts to minimize spatial heterogeneity between control and target groups, microclimatic factors such as slope, aspect, and urban-scale parameters like building density remain challenging to fully control. These factors can introduce biases in the assessment of cooling effect.

Numerical simulation

Advancements in observational accuracy, the accumulation of multi-source datasets, and improvements in computational technology have made it possible to simulate interactions among the atmosphere, ocean, and land surface using numerical methods. These developments have driven the evolution of comprehensive mathematical models, particularly Earth system models (ESMs) (Pan et al., 2024). In evaluating the cooling effect of forests through numerical simulations, the first step is to select an appropriate ESM and input multi-source datasets, including land use data, climate data, and forest characteristics. Various experimental scenarios, such as forest expansion or degradation, are then defined, and the model is run to simulate the impact of forest changes on temperature, thereby revealing the cooling potential of forests (Zhang et al., 2024). Due to computational constraints, global-scale numerical simulations are conducted at coarse resolutions. For urban-scale applications, high-resolution microclimate models, such as ENVI-met (Wu et al., 2025) and Computational Fluid Dynamics (CFD) (Ahn et al., 2024), have been developed. These models focus on interactions among urban elements such as GS, buildings, and roads, allowing for a detailed simulation of thermal processes at the city level. The flexibility of numerical simulation offers significant advantages in assessing the cooling effect of GS. By configuring different scenario parameters—such as GS area, spatial distribution, and vegetation types—researchers can systematically investigate the contribution of various influencing factors to cooling potential. This makes numerical simulation a valuable tool in urban planning, providing scientific evidence for decision-makers to optimize the spatial layout of GS to maximize cooling benefits. However, numerical simulations still have inherent limitations. Simplifications in the underlying physical processes, shortcomings in parameterization schemes, and uncertainties in input datasets may lead to biases in the simulation of vegetation-related surface energy distribution (Duveiller et al., 2018; Forzieri et al., 2020), affecting the accuracy of model outputs.

Weighing method

Weighing method is the most direct and technically mature approach for quantifying dust retention by vegetation. This method involves a series of complex procedures, including sampling, washing, filtration, drying, and weighing, to determine the amount of particulate matter retained per unit leaf area. By extrapolating from the total leaf area of individual plants or entire plant communities, the total dust retention capacity can be estimated (Chaudhary and Rathore, 2018; Zha et al., 2018). However, this method has limitations in determining the size distribution of retained particulates (e.g., PM₁₀, PM_2.5, and PM₁). Although sieving techniques or re-suspending particles for graded monitoring can be used to differentiate particle sizes, these approaches cannot fully guarantee that the original composition of dust particles remains unchanged or that their accumulation state on plant surfaces is accurately preserved (Wang and Wang, 2014).

Formula-based method

The dry deposition formula estimates the dust retention capacity of vegetation by incorporating pollutant concentration, dry deposition velocity, leaf area, and resuspension rate into an empirical equation. Among these parameters, dry deposition velocity is the most critical indicator for assessing the ability of vegetation to capture particulate matter. This velocity can be derived through wind tunnel experiments (Wu et al., 2021), artificial smoke chambers (Zhang et al., 2021), or mathematical simulations (Feng et al., 2022) under natural wind conditions, as well as through field observations. In general, larger and heavier particles exhibit higher dry deposition velocities, leading to more effective removal of these pollutants by vegetation (Zhang et al., 2020). The dry deposition formula primarily characterizes the deposition process of airborne particles on individual leaf surfaces as airflows pass through vegetation. However, obtaining more precise results at larger spatial scales requires extensive field observations and sophisticated experimental setups, which often entail significant time and financial costs.

Modeling method

The modeling method enables the assessment of vegetation’s capacity to capture particulate matter on larger spatial scales. The model serves as a simplification of real-world systems, integrating field observations with laboratory measurements to thoroughly analyze the mechanisms of vegetation-driven dust deposition. By incorporating extensive experimental data and empirical parameters, modeling provides a theoretical framework and data-driven support for large-scale assessments. The most representative and widely used models are iTree (Pace et al., 2021) and ENVI-met (Deng et al., 2019). These models require the input of site-specific and meteorological data for the study area. Site-specific data include tree species, diameter at breast height, tree height, canopy size, and health status, while meteorological data encompass precipitation, wind speed, and other climatic factors. By integrating all relevant data, these models can estimate and report the amount of particulate matter removed by trees. However, the modeling method shares similar limitations with the numerical simulation approach previously discussed for quantifying cooling effect, including process simplifications, model biases, and high computational demands.

Statistical method

With advancements in remote sensing and other geospatial technologies, macro-scale studies on the role of GS in reducing air pollution have gained increasing attention. Remote sensing data are used to track air quality and its temporal and spatial variations. This information, when combined with statistical methods such as correlation analysis, regression analysis, and machine learning, allows for an indirect assessment of the impact of GS on air quality improvement (Chen et al., 2019b; Venter et al., 2024). For example, a linear mixed model was used to evaluate the impact of GS on PM₁₀ and PM_2.5 concentrations in Zhengzhou across different spatial scales (Lei et al., 2021), which revealed that while GS did not significantly influence PM_2.5 concentration, they significantly reduced PM₁₀ concentration at nearly all scales. Additionally, increasing the contact area between GS patches and surrounding urban areas could further reduce PM₁₀ concentration. Furthermore, statistical analysis of five major cities indicated an inverse proportional relationship between PM_2.5 concentration and GS coverage (Chen et al., 2019b). Neighborhood-scale analysis also showed that the relationship between tree cover and PM₁₀ concentrations was nonlinear. As tree cover increased, PM₁₀ concentration decreased, but the rate of reduction slowed down, displaying a diminishing marginal effect (Yang et al., 2023).

The weighing method, formula-based method, and modeling method are essentially methods for quantifying the dust retention capacity of vegetation. Among these methods, the estimation accuracy decreases from high to low, the economic and time costs decrease from high to low, and their applicability spans from small to large spatial scales. All three methods rely on field measurements, which yield relatively accurate results. However, they are challenging to apply for continuous analysis over long time series and large spatial extents. Furthermore, these methods primarily focus on quantifying the dust retention capacity of vegetation and do not sufficiently address other GS air purification mechanisms, such as chemical absorption or the reduction of particulate matter sources. In contrast, statistical methods, by utilizing remote sensing data—characterized by large-scale coverage, long temporal durations, and high frequency—enable continuous quantification of the impact of GS on air quality over extensive areas. This method can also identify the combined effect of all purification pathways in GS on air pollutant concentrations, not just the dust retention ability of vegetation. However, statistical methods also have certain limitations. Attribution methods based on correlations face challenges in establishing causal relationships between GS and air quality improvement. They are susceptible to interference from factors such as meteorological conditions and changes in pollution sources, which may lead to biased results.

GS cooling reducing emission, transformation, and enhancing deposition

Temperature plays a critical role in the emission, transformation, transport, and deposition of air pollutants (Yu et al., 2025b). High temperatures often exacerbate air pollution (Chen et al., 2024), suggesting that GS cooling can contribute to air purification. This process primarily operates through three key pathways: (1) Emission reduction pathway. Lower temperatures help reduce pollutant emissions. Cooler environments decrease energy demand for air conditioning, thereby reducing anthropogenic emissions (Chen et al., 2024; Qin et al., 2025). Additionally, lower temperatures suppress the release of biogenic VOCs, which are precursors of ozone and secondary aerosols (Wu et al., 2019; Xu et al., 2023). (2) Transformation suppression pathway. Reduced temperatures slow down photochemical reactions, thereby inhibiting the formation of secondary aerosols and ozone, ultimately leading to lower pollutant concentrations (Lei et al., 2022). (3) Deposition enhancement pathway. Lower temperatures slow the movement of suspended particles, increasing their likelihood of deposition and reducing airborne pollutant concentrations (Yang et al., 2023). Thus, by lowering ambient temperatures, GS indirectly contribute to air purification by reducing pollutant emissions, suppressing their transformation, and enhancing their deposition. This process can be summarized as the “GS cooling - inhibit emission and transformation and enhance deposition - promote air purification” pathway.

Particularly in cases where the direct air pollution mitigation effects of GS remain limited (Pfannerstill et al., 2024; Venter et al., 2024), their indirect effects in improving air quality through cooling may be more significant. During summer, biogenic VOC emissions contribute up to 70% of secondary organic aerosol formation in China, far exceeding anthropogenic sources (Wu et al., 2019). As a major source of biogenic VOCs, GS exacerbate ozone and aerosol pollution to some extent (Kelly et al., 2018). Moreover, pollutant concentrations are largely influenced by dispersion conditions (Yu et al., 2022). Tall urban trees can obstruct ventilation, hinder pollutant dispersion, and thus exacerbate localized air pollution levels. A study in Shanxi has demonstrated that, compared to direct air purification mechanisms—such as dust capture and source reduction—GS are more effective in mitigating air pollution through their cooling effect, with this indirect pathway accounting for as much as 61.45% of the total air purification effect (Yu et al., 2025b).

GS air purification reducing downward longwave radiation

The decrease in pollutant concentrations caused by GS air purification alters longwave radiative forcing, thereby affecting the surface energy balance and ultimately influencing temperature. In general, a higher concentration of particulate matter enhances downward longwave radiation. Under heavy pollution conditions, this leads to increased longwave radiation reaching the urban surface, resulting in warming. This effect is particularly pronounced at night, as the absence of solar shortwave radiation eliminates counteracting influences, leading to a more significant warming effect induced by longwave radiation. A study conducted across 36 cities in China identified the biogeochemical effects of urban aerosols and haze pollution as key contributors to the nighttime UHI effect (Cao et al., 2016). In semi-arid cities, the UPI effect has been found to contribute up to 0.7 ± 0.3 K to nighttime warming, primarily due to the strong longwave radiative forcing of coarse aerosols, which increased downward longwave radiation and intensified nighttime heat retention (Cao et al., 2016). By reducing particulate matter concentrations through air purification, GS can reduce downward longwave radiation at night, leading to a cooling effect that indirectly enhances their overall temperature-regulating capacity. This process can be summarized as the pathway of GS air purification - reduction of downward longwave radiation - promote nighttime cooling.

Cooling-enhanced vegetation-induced diffusion and deposition

Unfavorable atmospheric diffusion conditions are a key factor contributing to air pollutant accumulation, leading to regional-scale severe air pollution episodes. Enhancing the rapid and effective diffusion of near-surface urban air pollutants is crucial for mitigating the adverse health effects of air pollution (Yang et al., 2019). Previous studies investigating the impact of GS on urban air pollutant dispersion have primarily focused on large-scale effects, emphasizing how increased surface roughness from vegetation hinders ventilation and exacerbates air pollution (Tsoka et al., 2020). However, at the local or community scale, GS contribute to cooling, lowering the ambient temperature and creating temperature gradients between green and non-green areas. This thermal contrast can enhance airflows, promoting pollutant diffusion and dilution. Based on this mechanism, there is a potential pathway of “GS cooling - convection - pollutant diffusion - promote air purification”.

GS can also promote dust deposition by lowering leaf surface temperatures through transpiration, thereby creating favorable conditions for atmospheric pollutant deposition. According to the kinetic theory of gases, particulate matter tends to migrate from high-temperature regions to low-temperature regions in the presence of a temperature gradient. This phenomenon, known as the thermophoretic effect, arises from differences in molecular kinetic energy and particle collisions caused by temperature variations. Notably, the thermophoretic effect is also a fundamental principle in the design of air purification devices (Yang et al., 2023). Compared to other land cover types, vegetation with lower leaf surface temperatures significantly enhances pollutant deposition efficiency. Additionally, green spaces, with large leaf area, provide extensive surface area for pollutant capture. The continuous renewal of plant leaves further sustains long-term deposition capacity, making the dust-retention capacity of vegetated areas substantially higher than that of other land surfaces. This means that enhancement of deposition efficiency becomes even more critical given the extensive deposition surface provided by vegetation. Specifically, GS can reduce leaf temperatures, accelerate dust deposition through thermophoretic movement, and enhance dust-retention efficiency, forming a cascading pathway of “GS cooling - thermophoresis - leaf dust deposition - promote air purification”.

During extreme heat events, GS exhibit a stronger cooling effect (Schwaab et al., 2021). Moreover, evidence suggests that the ability of vegetation to alleviate high temperatures plays a particularly important role in reducing PM₁₀ levels under extreme heat conditions. Specifically, the indirect cooling effect of GS was found to reduce PM₁₀ concentrations by 27.11% under average summer conditions and by 39.76% during extreme heat events (Yang et al., 2023). These findings indicate that as the cooling effect of GS intensifies, their air purification capacity is also enhanced, providing further evidence for the effectiveness of vegetation cooling-mediated indirect pollutant removal.

Air purification-enhanced vegetation transpiration and photosynthesis

GS improve air quality by reducing atmospheric pollutant concentrations, thereby decreasing the scattering, reflection, and absorption of solar shortwave radiation by pollutants. As a result, more solar radiation is available for direct absorption by vegetation, which in turn enhances physiological processes, increases transpiration efficiency, and strengthens photosynthetic activity. This amplification of vegetation function contributes to greater evaporative cooling and carbon sequestration. Based on this mechanism, two synergistic pathways can be identified, i.e. the pathway of “GS air purification - shortwave radiation enhancement - transpiration enhancement - promote cooling”, and the pathway of “GS air purification - shortwave radiation enhancement - photosynthesis enhancement - promote cooling”.

On heavily polluted days in major industrial cities, the amount of direct solar radiation reaching the land surface can be reduced by nearly 40% compared to unpolluted conditions. Haze alters both light intensity and quality, significantly suppressing vegetation transpiration and photosynthetic efficiency, thereby affecting plant growth and the ecological benefits of GS. Studies have shown that if strict air pollution control measures were implemented to reduce anthropogenic emissions to pre-industrial levels, forests could experience an additional cooling effect of up to 0.4°C due to increased solar radiation availability (Ge et al., 2023). This finding confirms the positive impact of low-pollution environments on vegetation-induced cooling through transpiration. Similarly, regarding photosynthesis, during weekends when industrial emissions and traffic activities are reduced, lower atmospheric particulate matter concentrations improve light conditions. As a result, 64% of observed regions in Europe exhibited a significant increase in chlorophyll fluorescence—a natural signal of plant photosynthesis (He et al., 2023). This further substantiates the role of cleaner air in enhancing vegetation photosynthetic activity. These findings collectively demonstrate that the air purification of GS can indirectly enhance transpiration and photosynthesis, thereby amplifying their cooling effect.