With the predominant industrialization and urbanization, the construction of urban infrastructure and land use changes have dramatically replaced natural ecosystem areas (He et al., 2021), resulting in significant environmental deterioration, economic losses, social challenges, and public health issues (Li K. et al., 2023; He, 2023; Wei et al., 2023). Addressing such challenges have been a consensus of the public, governments and organizations (Liu X. et al., 2023). Even though, cities are highly dependent on heavy fossil-fuel use. Authorities estimate that fossil fuel combustion accounts for 90% of total CO₂ emissions in 2023 (Friedlingstein et al., 2023), resulting in a rapid rise in atmospheric concentrations of greenhouse gases on Earth, ultimately causing global warming (Gao et al., 2021; Shen and Zhao, 2024). For example, a 1.1°C increase in global average temperature worldwide presents the largest increase in the last 1,000 years (Intergovernmental Panel On Climate Change, 2023). Furthermore, it is reported that the urbanization has contributed to one-quarter of the average annual temperature increase of Guangdong Province in the last 70 years (Zhong et al., 2023). The atmospheric studies in Guangdong, Hong Kong, and Macao in China have concluded that greenhouse gases are the most important determinants of future average and extreme temperatures (Zheng et al., 2022).

The global atmospheric CO₂ concentration in 2022 will be about 1.5 times that of the pre-industrial revolution, in addition to total global CO₂ emissions of about 40.9 billion tons in 2023, of which China will be the largest emitter, accounting for about 35% of the total emissions (Friedlingstein et al., 2023). Cities only cover 3% of the Earth’s surface, but they are now representing more than 55% of the global population. Moreover, the spatial distribution of carbon emissions in cities are much higher than rural counterparts, indicating that cities and their adjacent regions are among the highest contributors to carbon emissions (Figure 1). In China, the situation is more alerting since urban areas contribute as much as 90% of national carbon emissions (Li, 2010). With this background, the United Nations Framework Convention on Climate Change was signed by more than 150 countries in 1992, in order to collectively maintain atmospheric greenhouse gas concentrations at a stable level. The 2015 Paris Agreement requires all signatory nations to reduce their greenhouse gas emissions. In September 2020, China explicitly set a double carbon target for 2030 in order to achieve carbon neutrality by 2060.

Achieving carbon neutrality generally involves two primary approaches: carbon emission reduction and carbon sink augmentation. Carbon reduction is the ability of green spaces to reduce energy consumption in buildings, transportation, etc., by improving the microclimate of the space. The carbon sequestration is to increase the carbon sequestration capacity of greenfield plants, with a focus on strengthening the ecological construction and protection of urban green spaces. The terrestrial biosphere shows significant potential for carbon sequestration. The studies indicate that in China, the terrestrial biosphere annually sequestered an average of approximately 1.11 (±0.38) ×10⁹ tons of carbon dioxide between 2010 and 2016, roughly equivalent to 45% of the concurrent annual anthropogenic carbon dioxide emissions (Wang J. et al., 2020). Urban green spaces represent primary green ecological resources in densely populated and economically developed urbanized regions. These spaces also constitute the main natural carbon sinks within urban ecosystems and possess a unique ability for self-purification and self-regulation. It is demonstrated that urban green space contributes to improving urban environmental quality by sequestering carbon (Nowak et al., 2018). This is mainly manifested through vegetation employing photosynthesis to reduce atmospheric CO₂ while releasing O₂ (Russo et al., 2014), playing a crucial role in regulating the balance of atmospheric carbon and oxygen, and enhancing urban environmental quality.

Urban greenspace ecosystem comprises individual plant entities within the city, and its carbon sequestration capacity relies on the carbon sequestration abilities of vegetation. Urban green space vegetation serves as the main component of urban greenspace carbon sinks. Through photosynthesis and respiration within their leaves, plants accumulate a net carbon amount, sequestrating CO₂ within the vegetation, soil, and water bodies, thereby reducing atmospheric CO₂ concentrations (Dong and He, 2023). Moreover, urban green spaces effectively mitigate heat islands (Liu H. et al., 2023), decrease urban energy use, and thereby reduce urban carbon emission. Plants play a dual role of reducing CO₂ emissions and increasing CO₂ sequestration, making them the primary carbon sink in landscape architecture (Bao, 2011). Overall, ecosystem carbon sequestration, including urban greening carbon sinks, has become a focal task in the top-level design of China’s efforts toward achieving carbon neutrality. To address warming challenges, it is essential to include more carbon sinks during urban planning and design to sequestrate greenhouse gases such as CO₂ and CH₄, as well as to alleviate environmental deterioration such as urban flooding and heat islands. Quantitative analysis of the carbon sequestration capacity of plants is an important research topic since it not only comprehensively demonstrates their carbon sequestration and oxygen release abilities and influencing factors but also provides the basis for creating, designing, managing, and improving urban natural carbon sinks. Furthermore, the quantitative analysis aids in scientifically selecting tree species for urban green spaces in a low-carbon era, offering theoretical references for strong carbon sink plant configurations, ecological landscape construction, and the development of eco-friendly urban landscapes. Overall, the quantification of green space carbon sink potential is a crucial means for alleviating global greenhouse and heat island effects, serving as a key player in supporting sustainable urban development and mitigating climate change.

Sample plot measurement refers to an estimation method of setting up sample plots in typical areas with good forest growth and conducting continuous observation in the sample plots to obtain the changes in carbon stocks per unit time (Zhou et al., 2013). This method quantifies growth over time, considering the time when plants start growing until the time of their harvest. It is mainly applied to investigating the carbon sink potential of macro-scale green spaces. There has been a long history of sample plot measurement application, mainly in forestry and agricultural production, and it is a well suggested method by the IPCC for forest carbon sink assessment (Ouyang et al., 1999). This approach estimates the carbon sequestration of plants by harvesting and weighing all aboveground and belowground organic matter within sampling plots. The average values from these plots were used to estimate the biomass of the entire forest or individual plants, which were then converted into an average annual carbon sequestration rate for green vegetation.

In practical applications, it often involves the use of highly accurate measured data obtained from standard tree analysis to construct ecological indicators and biomass growth equations (Lee et al., 2014). For instance, the relationship between diameter at breast height (DBH), tree height and age indicators of trees was utilized to develop an allometric growth equation to estimate the carbon sequestration capacity of forests (Zhang et al., 2019). The biomass method acquires measured data through extensive field surveys to assess the biomass of plants at different time and measure their photosynthetic intensity. Overall, it is now the most commonly used method for calculating forest carbon storage because of its direct, explicit, and technically straightforward advantages. However, it has several limitations, such as destructive experimentation, inability for continuous observation, difficulty in accounting for root production and litterfall, and complex processing procedures.

The model estimation method, based on a plot inventory, is a convenient and precise approach suitable for quantifying carbon sequestration of urban green spaces in large areas or urban regions. By the end of the 21st century, leveraging big data systems, the United States Forest Service, systematically analyzed and developed several convenient plant carbon sequestration calculation systems. At current, the widely utilized systems include Citygreen, the NTBC (National Tree Benefit Calculator), Pathfinder, and the I-Tree Eco. These models receive extensive application in assessing the ecological benefits of urban green spaces. For instance, the Citygreen calculation system, in conjunction with on-site surveys, has been employed to estimate the total carbon sequestration of green spaces in cities like Shenzhen (Chen et al., 2009), Shanghai (Xu, 2010), and Shenyang (Liu et al., 2008). And quantification of carbon sequestration benefits of green space plants in Nanjing residential area using NTBC (Li Q. et al., 2023). The I-Tree Eco module has been used in cities in the United Kingdom (Monteiro et al., 2019), the United States (Ning et al., 2016), Thailand (Intasen et al., 2017), and Hungary (Kiss et al., 2015) to assess the ecological value of urban green spaces. Moreover, the I-Tree Eco can estimate the carbon sequestration of individual plants based on specific characteristics. Overall, these estimation systems estimate carbon sequestration on a large regional scale by considering vegetation characteristics, meteorological data, site features, greenspace area, and surrounding environments so that the results are comprehensive. However, it is important to note that the physiological, ecological, and climatic parameters of the associated models were calibrated based on the U.S. conditions. Consequently, when such methods are applied to regions with different ecological systems and climatic conditions, substantial errors may occur. Therefore, these models are primarily suitable for use in areas with geographical and climatic conditions similar to those in the United States, upon some essential validation.

The RS method, using satellite remote sensing to acquire various vegetation status parameters, offers the advantage of rapid, real-time, and large-scale data acquisition. It can well address the limitations of model-based estimations. The basic principle is to obtain various parameters of vegetation within the survey areas by sensing techniques on the basis of ground survey, and to estimate the impact of changes in land use and green space coverage on carbon stocks through spatial classification of vegetation and analysis of time series. Sensing-based assessments of net ecosystem productivity (NEP) are widely applied in carbon sink studies at regional and urban scales, such as some studies utilized QuickBird, RS and GIS to estimate the annual CO₂ uptake per tree in Los Angeles (Mcpherson et al., 2008), and combined LiDAR with QuickBird to estimate carbon stocks in urban trees (Schreyer et al., 2014). Meanwhile, there are studies on the use of remote sensing-based explicit forest carbon stock calculation models to realize high-resolution mapping of forest carbon stocks at large scales and dynamic monitoring of forest carbon sinks globally (Zhu et al., 2024), as well as the combination of ground-based measurements and LiDAR data to estimate the carbon stocks of urban trees (Gülin and Bosch, 2021). Research on the surface carbon storage of urban green spaces in Auckland also used LiDAR data combined with field surveys, suggesting that this approach not only provided more accurate data but also reduced the frequency and cost of actual measurements (Wang V. et al., 2020). This is because the method can offer detailed information on plant biomass in complex urban areas (Alonzo et al., 2014). However, due to the spatial heterogeneity and temporal dynamics of urban green spaces, the use of RS estimation methods for estimating carbon storage may encounter challenges like the inability to accurately determine plant quantity and difficulties in separating overlapping tree canopy boundaries (Richardson and Moskal, 2014). In addition, individual ecological differences among species can also lead to significant errors.

The micrometeorological method is based on microclimate monitoring and involves continuous dynamic monitoring of near-surface atmospheric flow conditions and atmospheric CO₂ concentrations. This approach indirectly estimates the carbon flux of vegetation (Liu et al., 2018). It focuses on the quantitative study of carbon dioxide and water exchange between the atmosphere and terrestrial ecosystems (Grimmond et al., 2002). In Phoenix, the micrometeorological method was adopted, revealing that CO₂ concentrations at midday were lower than before sunrise mainly due to the strong sunlight at midday which promotes photosynthesis in vegetation (Idso et al., 1998), while in Essen using mobile measurement, it was shown that CO₂ concentrations in the city were lower in summer months given vegetation photosynthesis and other effects (Henninger and Kuttler, 2010). The micrometeorological method can accurately estimate regional-scale carbon fluxes, with eddy covariance being the most commonly used method. For instance, Velasco et al. used the eddy covariance method to show that urban CO₂ concentrations are lower during the daytime during the plant growing season, mainly due to the high photosynthetic capacity of plants at this time of year (Velasco and Roth, 2010), similar to the conclusion reached by Rana et al. (2021). Although the micrometeorological method allows for continuous in-situ observations, the unstable near-surface atmosphere and replicated topographic conditions can introduce significant errors into the results (Wilson et al., 2002). For example, atmospheric inversion and data from the Shangri-La observation point in the complex terrain of the Hengduan Mountains resulted in higher estimated data (Wang J. et al., 2020). The micrometeorological method also relies on site-specific observations and cannot represent carbon flux values for the entire region. At the regional scale, there are challenges related to low spatial resolution, making it difficult to differentiate the carbon sink amounts for different ecosystems (Piao et al., 2022).

The assimilation method is a micro-scale research method. Since leaves are the most important organs for photosynthesis in trees, the assimilation method calculates plant carbon sequestration by determining the instantaneous net photosynthetic rate per unit of leaf area from the instantaneous CO₂ concentration of the leaves and the change in water content. Infrared Carbon Dioxide Gas Analyzers have been widely used from the 1950s to the present day, and the basic principle of their work is to calculate the net photosynthetic rate by determining the difference in CO₂ concentration entering leaf stomata. Measurements of photosynthesis rates of Wisconsin field species in the 1990s were an early application of the use of PPS to measure net photosynthetic rates of plants (Reich et al., 1995). By the end of 20th century, some studies reported the adoption of a photosynthesis meter to quantify plant photosynthetic intensity in a representative urban green space in Guangzhou and estimated the CO₂ uptake of the plants through the reaction equation of the photosynthesis process (Yang, 1996). Using infrared gas analyzers, measurement of the annual carbon dioxide uptake by urban plants in Korea was conducted, resulting in an empirical formula for future estimation. Furthermore, Han proposed a method for calculating plant carbon sequestration and oxygen release using the net assimilation rate formula (Han, 2005). The assimilation method experiment requires only a small amount of plant leaves, making it friendly to plants and widely applicable, and it has been employed in studies on common garden plants in locations such as West Bengal (Biswas et al., 2014), Rome (Gratani et al., 2016), and Fuzhou (Wang, 2010).

The assimilation method directly measures the carbon sequestration capacity of a single plant, and the results are accurate, allowing a direct comparison of carbon sequestration benefits among different plants. However, its outcomes are significantly influenced by temporal and spatial factors. This is attributed to the rhythmic cyclical variations in the photosynthetic rate of plant leaves, which systematically fluctuate with diurnal and seasonal changes. The primary rationale behind these fluctuations is the close relationship between photosynthetic rates and physiological and ecological environmental conditions such as atmospheric temperature, relative humidity, and solar radiation (Figure 3) (Evans and Santiago, 2011). For instance, the impact of vertical illumination results in peak solar radiation and daylight duration during the summer in the northern hemisphere, leading to a noticeable increase in carbon fixation by plants during this season compared to others. Although winter exhibits relatively weaker carbon fixation capacity throughout the year, the method still remains a high efficiency (Weissert et al., 2017). Moreover, air with higher relative humidity enhances stomatal conductivity in plant leaves, thereby elevating the rate of photosynthesis (Wang et al., 2019). These conditions necessitate labor-intensive hourly measurements of plant photosynthetic rates and impose significant requirements on weather conditions.

Whilst there have been many methods for urban green space carbon sink estimation, the advantages and disadvantages of each method are distinct (Table 1). Furthermore, there is no direct way to quantify carbon sinks in urban green spaces, and existing studies on urban green space carbon sink often employ methodologies developed for forest carbon sink (Gratani et al., 2016). Urban green spaces, primarily composed of artificial greenery, differ significantly from natural green spaces such as forests and grasslands. Urban green spaces also exhibit high heterogeneity, with plant species determined by early urban planning, incorporating both local and introduced species. This prominent contrast with naturally evolved green spaces implies that carbon sink in urban green spaces is more complex (Zhao et al., 2023). Therefore, existing methods for green space carbon sequestration may not be entirely applicable to urban settings. The sample plot measurement suitable for large-scale natural environments like forests, faces limitations in regions with diverse tree species distributions, rendering standard plots dissimilar to the entire region. Moreover, its destructive nature, involving irreversible vegetation removal, makes it unsuitable for urban green spaces designed and constructed through human planning. Explorations into non-destructive methods, such as the biomass method, for studying carbon sequestration in Bangkok park green spaces have been undertaken (Singkran, 2022), and the use of remote sensing to quantify and map forest structure to estimate forest above-ground biomass to quantify carbon stocks and fluxes in tropical forests (Mohd Zaki and Abd Latif, 2017), these studies explore additional possibilities for the sample plot inventory method.

Micrometeorological methods demonstrate high accuracy in regional-scale estimations but are sensitive to variations in urban spatial characteristics, leading to significant errors in different regions. Results from these methods reflect carbon sequestration at the regional scale and cannot discern individual plant carbon sequestration at a finer level. Model estimation incorporates green space information from similar regions into the model for estimation, potentially introducing errors due to ecological differences between regions. RS estimates carbon sequestration by calculating various green quantity indicators from satellite imagery, but the considerable variations in carbon sequestration benefits among different plant species result in two-dimensional indicators that may not fully reflect the actual carbon sequestration benefits of urban green spaces. Assimilation method is the most accurate way to quantify the benefits of plant carbon sequestration, but its results may include large errors owning to experimental weather fluctuations. Moreover, scattered distributions of tree species in the experiment require a long period of time, resulting in limitations of labor dependence. The assimilation amount method is only for single plant scale, so that extensive sample plant data should be carried out when facing the large-scale urban green spaces.

The relationship between net photosynthesis rate and PAR can be described by the net photosynthesis-light response curve which refers to the tendency of the photosynthetic rate of plant leaves to change with the change in the PAR intensity when other environmental conditions are kept unchanged (Figure 5). The light response curves are mainly derived by fitting the photosynthetic rate values exhibited by the plants at different levels of light intensity. In essence, the light response curves show a functional relationship between the photosynthetic rate and light intensity, so that it can be used as a model to predict photosynthetic rate.

Many models have been used to fit net photosynthetic light-response curves, such as exponential model (Eq. 1) (Bassman and Zwier, 1991), rectangular hyperbolic model (Eq. 2) (Baly, 1997), non-rectangular hyperbolic model (Eq. 3) (Gates, 1977; Hardwick, 1977), and mechanistic model (Eq. 4) (Ye, 2007). These models describe photosynthetic physiology and can reasonably analyze the whole process of the light response curve so that they are capable of obtaining accurate data and fitting curves with high reliability. However, there are no models suitable for all situations. For instance, the data fitted by the non-rectangular hyperbolic model and the exponential model are more in line with actual data, but the prediction results of the mechanistic model were more accurate and realistic (de Lobo et al., 2013). Accordingly, the mechanistic model was later modified and simplified as Eq. 5 (Ye et al., 2013). In practice, the fitting can be done by Ye’s photosynthetic computational model and the computational model built upon Microsoft Excel (de Lobo et al., 2013). P n = P ⁡ max ⁡ ( 1 − e − α I P m a x ) − R d (1) P n = α I P m a x α I + P m a x − R d (2) P n = α I + P ⁡ max ⁡ − α I + P m a x 2 − 4 θ α I P m a x 2 θ − R d (3) P n = α × 1 − β × I 1 + γ × I × I − I c o m p (4) P n = α × 1 − β × I 1 + γ × I × I − R d (5)

where Pn represents the net photosynthetic rate [μmol (CO₂)m^-2·s^-1], which is the difference between the overall photosynthetic rate and the dark respiration rate of the plant; Pmax is the maximum net photosynthetic rate [μmol (CO₂)m^-2·s^-1]; θ is the curvature of the curve; I is the photosynthetic photon flux density; α is the initial slope of the light response curve at I = 0, representing the initial quantum efficiency of the photosynthetic process [μmol (CO₂) (photon)⁻¹]; Rd is the dark respiration, the amount of carbon dioxide released by the plant in dark [μmol (CO₂) (photon)⁻¹]. I _comp is the light compensation point [μmol (photon)m^-2·s^-1], and β and γ are adjustment factor.

Solar radiation is the primary energy source of the Earth’s surface, where the portion of solar radiation that can be utilized by green plants for photosynthesis is defined as PAR, also known as light intensity. PAR serves as the energy source for plant biomass and constitutes a crucial parameter for assessing plant photosynthetic potential (Zhou and Xiang, 1996). The current measurement methods for PAR include the energy system (W·m²) for determining PAR illuminance (Q_PAR) and the quantum system (μmol·m^-2·s^-1) for determining PAR density (U_PAR). The PAR density is also known as the photosynthesis photon flux density (PPFD), presenting the number of photons incident on a unit surface per unit time (Sun et al., 2017). Under illumination by light sources with different spectral structures, the ratio between leaf photosynthetic rate and photosynthetically active quantum flux density exhibits minimal variation (McCree, 1972). Compared with the energy system, the quantum system for determining the photosynthetically active quantum flux density U_PAR is more reasonable. Therefore, quantum measurement systems are increasingly prevalent in fields such as agriculture and ecology.

However, U_PAR depends on wavelength and therefore cannot be derived directly from solar irradiance (Wang et al., 2021). Instruments specifically designed for directly measuring the light quantum flux U_PAR are not widespread, and most meteorological stations lack regular observation platforms for PAR. As a result, U_PAR is generally calculated in an indirect way, such as using UV-visible band fluxes to estimate (Sun et al., 2017), or using meteorological datasets to develop models (García-Rodríguez et al., 2021). On the meteorological conversion, the conventional unit for solar radiation in meteorological parameters is horizontal total radiation (W·m^-2), which can be quantitatively converted into light quantum flux (μmol·m^-2·s^-1) (Wang et al., 2005). More importantly, meteorological research indicates a stable proportion of PAR in total solar radiation. Therefore, the meteorological conversion relationship involves calculations of the photosynthetically active coefficient and quantum conversion coefficient, which are usually calculated by the following empirical formula (Zhou et al., 1984). Q P A R = η Q (6) U P A R = μ Q P A R (7) U P A R = μ η Q (8) where η is the photosynthetic efficiency coefficient, representing the proportion of photosynthetically active radiation energy in the total solar radiation; μ is the quantum yield indicating the number of quanta per unit photosynthetically active radiation energy, with units of μmol·J^-1; Q is the total horizontal solar radiation flux, measured in W·m^-2; Q _PAR is the photosynthetically active radiation intensity, measured in W·m^-2; U_PAR is the photon flux of photosynthetically active radiation, measured in μmol·m^-2·s^-1.

The current annual average value of η in Eq. 6 is between 0.409 and 0.477 worldwide (Akitsu et al., 2022). Observational studies indicate that η is influenced by both astronomical and meteorological factors. Long-term continuous synchronous observations of Q and PAR in locations such as Beijing, Yantai, and Zhengzhou revealed the stability of η values (Wang and Shui, 1988). A suitable calculation formula for plains was derived as Q_PAR = 0.42 Q, where η = 0.42. Subsequent research on cities like Chengdu, Kunming, and Guangzhou yielded similar conclusions (Wang and Shui, 1990), and other scholars found similar results that η is 0.39 ± 0.04 (Zhou et al., 1984). Consequently, the quantum conversion coefficient for China can be established as 0.42.

There is no systematic climatological research result about the value of the quantum conversion factor μ. McCree concluded μ = 4.57 μmol J⁻¹ (McCree, 1972), but the value of μ in Eq. 7 is inconsistent in different regions of China, and at present the intermediate value is 4.55 μmol J⁻¹ (Dong et al., 2011) so based on Eq. 8 that the conversion relationship between Q and PAR is U_PAR = μ×η = 4.55 × 0.42 Q = 1.911 Q.

Although no study has found a significant relationship between plant net photosynthetic rate and its size (Weissert et al., 2017), it is advisable to select healthy tree species with a planting age between 3 and 10 years for experiment. Since it is always challenging to measure the height of large deciduous trees using in-situ methods, auxiliary measurements are often carried out by tools such as cherry pickers and tower cranes (Chen et al., 2003; Zhao et al., 2008). However, this approach is applicable to only a few tree species with short measurement cycles. Given the diverse and scattered distribution of plant species in urban green spaces, as well as constraints posed by topography, the feasibility of this method is low. Alternatively, in vitro measurements are generally adopted. Whilst the photosynthetic rate of some plants may slightly decrease shortly after detachment, most tree species exhibit a high and stable photosynthetic rate within the first hour after detachment (Tang and Wang, 2011). As a result, the light response curves of plants measured within the first hour of leaf detachment can largely represent in situ measurements.

Currently, the main method for in vitro measurements is to restore the water supply to detached branches and leaves to alleviate water stress, thereby restoring their photosynthetic capacity to the in situ level (Qiang et al., 2017). Typically, plant branches are cut and immediately immersed in water, and the cut surface is maximized by making a second diagonal cut in the water, and subtracting excess foliage from the branches will reduce water loss (Xu, 2006). After measurements, the mechanism model of light response through photosynthesis such as leaf floating was adopted to fit formula (Ye et al., 2013), with the initial values of each parameter: α = 0.06, β = 0.002, γ = 0.01, and Rd = 1, and the limiting range of 0 < α < 0.1, 0.002 < β < 0.01, 0.01 < γ < 0.03, and 0 < Rd < 3. The measurement steps are as follows:

(1) Prune the middle branches of the plant canopy by pruning shears, and place the cut ends into water quickly. Afterwards, make another diagonal cut approximately 3 cm from the initial cut to increase the water absorption area of the branches, during which remove the majority of leaves or leaflets from the branches to minimize water loss from detached plant materials.

It is a consensus embraced by governments and organizations worldwide that urban green spaces are efficient natural carbon sinks and a crucial strategy for mitigating climate change. The loss of carbon sinks due to built-up area increase during urbanization should be compensated by urban green spaces (Nowak et al., 2013). In this context, the carbon offsetting capacity (COC), the ratio between carbon emissions from human activities and the carbon sink of green spaces, has been attracting attention (Chen, 2015). Accordingly, quantitative studies on urban green space carbon sinks not only elucidate their paramount significance for cities but also clarify their ecological value, providing a foundation for subsequent management endeavors. At present, research mainly focuses on the indirect quantification of carbon sinks, holding crucial implications for investigating carbon offset capabilities, as well as ecological and economic benefits. In comparison and more importantly, directly quantifying the individual carbon sequestration capacity of urban green space allows the identification of locally suitable high-efficiency carbon sequestration plants, which can guide future urban green space construction and maximize ecological benefits (Table 3). Despite urban green spaces contributing a fraction of carbon sink capacity compared to urban carbon emissions, appropriate design and management of vegetation, as the primary component of urban green spaces, can significantly enhance future carbon sink capacity. This enhancement must be based on an understanding of differences in carbon sequestration benefits among various plant species. In addition, the ecological benefits of plants, such as cooling and fire prevention, hold equal importance in climate change mitigation (Murray et al., 2018; Liu X. et al., 2023). In future research, integrating studies on plant carbon sequestration with quantifications of other ecological benefits may prove more advantageous for the sustainable development of cities.

Methodologies for quantifying carbon sequestration in urban green spaces exhibit a diverse trend, and existing methods tailored to different scales and types of green spaces are predominantly concentrated in fields such as forests, grasslands, and agricultural areas. Concurrent studies are exploring the integration of various research methods. For instance, employing the accumulation method principle, directly utilized ArcGIS to estimate the carbon sequestration in Russian forests (Malysheva et al., 2018) and used the IUEMS carbon sequestration and oxygen release model to study forest carbon sequestration in the Qinling Mountains based on net ecosystem productivity (Ma et al., 2022). The CASA model was adopted to estimate the carbon sinks of greenfield vegetation in some major cities in China based on satellite images, meteorological data and vegetation data (Xu et al., 2023), or extracting plant species and leaf area indices from large green spaces by satellite and RS correlation techniques to estimate carbon sink benefits (Johnson et al., 2023). In future, more fusion methods should be explored for simpler and more accurate studies.

To address the challenges of experimental requirements in the assimilation method, this study proposes a method by establishing a model for estimating photosynthetic rate based on light-response curves. This involves measuring the net photosynthetic rate of plants under different levels of light intensity and utilizing mechanistic model of light-response of photosynthesis to fit the light -response curve. The actual hourly PAR is incorporated to simulate the plant’s hourly net photosynthetic rate on corresponding dates. The carbon sequestration of plants can be finally quantified using the formula for calculating plant carbon sequestration. This standardized external environment setting allows for the carbon sequestration comparison among different plants under the same conditions, enabling more accurate selection of plants with high carbon sink. The study presents controlled adjustments to the light intensity environment, CO₂ concentration and other conditions in the leaf chamber of the photosynthesis measurement instrument to ensure a consistent experimental environment. Although light intensity and atmospheric CO₂ concentration have the most significant impact on plant photosynthetic rates (Xiao et al., 2019), other external factors can also influence photosynthetic rates. Therefore, in future experiments, there is a need to control the other conditions to make the data more accurate.

There have been many models for net photosynthesis-light response curve fitting, and studies have shown that there are uncertainties in the fitting models (Fang et al., 2015). Moreover, the calculation results of some data of these models may be out of the normal range, this study, to use the light response curve fitting model, follows the mechanism model of photosynthesis response to light. It is a modified model for the rectangular hyperbolic model, having some advantages compared with previous models, and more accurate and real data can be obtained when calculating α. However, the data such as P _max will be out of the normal physiological range under some circumstances (de Lobo et al., 2013), which should be overcome to increase the accuracy of the results in the future research. Finally, although the predicted net photosynthetic rate values based on the light response curve exhibit a strong correlation with the actually measured values, this does not mean that the results of the two experimental methods are consistent. The photosynthetic rate estimation method based on the light response curve is only applicable to preliminary research on quantifying carbon sequestration in a large number of urban green spaces. More precise carbon sequestration data requires actual hourly measurements.

Quantifying carbon sink in urban green spaces is crucial for guiding future urban green space planning and management efforts. Therefore, research should not only concentrate on the direct carbon sink capacity of plants but should also comprehensively explore aspects such as the reduction in carbon emissions, self-generated carbon emissions, and factors influencing carbon sink. This will provide clearer paths for the future construction of urban green space, and help to promote the limited urban green space to play the greatest carbon sink benefits. Regarding carbon emissions, urban green spaces, through shading and transpiration, can lower the overall temperature of the city, thereby reducing building energy consumption and indirectly decreasing urban carbon emissions (Dong and He, 2023). Studies indicate that trees in urban centers indirectly absorb more CO₂ than they do directly (McHale et al., 2007). Concerning self-generated carbon emissions, although urban green space vegetation can sequester carbon through photosynthesis, the carbon emissions generated throughout its entire lifecycle, including production, transportation, and management can offset a portion of the carbon sequestration (Park et al., 2021). Deducting the total carbon emissions from the overall carbon sequestration quantity is meaningful for understanding the net carbon sequestration benefits of plants.

On carbon sinks, their influencing factors can be categorized into plants themselves, and planning and design. Regarding plants, indicators such as LAI, size, tree age, tree diameter, and tree height have a significant impact (Othman et al., 2019; Shadman et al., 2022). Moreover, the amount of carbon sequestered by plants varies geographically. Native plants can also show better carbon sequestration efficacy given their local adaptability that means small carbon emissions from maintenance management (Wang et al., 2015). Additionally, past research has predominantly focused on trees and shrubs, neglecting the substantial carbon sink role played by extensive lawns in urban green spaces (Amoatey and Sulaiman, 2020), thus necessitating a quantitative study. On planning and design, plants show different carbon sink abilities with different planting designs. Quantifying the carbon sink capacity variation under different planning and design schemes is important to provide guidance for optimization. Planting density of plant communities had a significant positive relationship with their carbon sequestration (O’Donoghue and Shackleton, 2013), but excessively high density can have opposite effects (Mexia et al., 2018). The proportion of trees and shrubs, plant community hierarchy, biodiversity and vegetation spatial types also significantly affect the carbon sink capacity. Meanwhile, there is a spatial correlation between human activities and greenfield carbon sinks (Dong et al., 2023). On management, digital means can be used to realize the analysis and accurate management of urban green spaces. For example, a digital platform for high carbon sequestering plant communities developed by Python and other programming software in Xi’an provided a scientific guidance for decision-making in plant community design (Wang et al., 2023).

Overall, increasing the carbon sink of green space cannot be only considered in an individual aspect, but should be considered comprehensively from the whole life cycle of plant selection, community design, and maintenance management. However, only from the construction perspective to study the carbon sink of urban green space has a temporary nature. In future, there is a need to study the influencing factors to optimization carbon sink of urban green space so that it can play long-term and better carbon sink benefits. It is also possible to form evaluation standards and systems related to carbon sinks on the basis of carbon sink measurement and quantification, so as to provide a practical basis for evaluating various types of urban green areas.