Mental health issues have become more prevalent among urban residents in recent years due to rapid urbanization and escalating life stressors (Galea and Vlahov, 2005). As exposure to unfavorable environments tends to adversely affect people's mental health and lead to mental disorders, the influence of natural, built and social environments on people's mental health has drawn increasing attention in many fields, including health geography, epidemiology, and urban planning (Houlden et al., 2018; Silva et al., 2018). Among multiple environmental factors that may affect people's mental health, greenspace exposure is a central component of people's access to natural environments. Previous research has posited that the interconnection between individuals' mental well-being and their proximity to green spaces in residential areas is a salient and influential determinant (van den Bosch and Meyer-Lindenberg, 2019; Banay et al., 2019).

Further studies reveal some mechanisms of how greenspace exposure affects people's mental health. Exposure to greenspace has been found to facilitate recuperation from stress and mental exhaustion (Zhang et al., 2020), promote physical activity among individuals (Van den Berg et al., 2019), serve as a filter for health-endangering pollutants like particulate matter and noise (Kardan et al., 2015), and foster neighborhood social cohesion among residents (Maas et al., 2009). However, despite the theoretical plausibility of these pathways, the body of empirical research examining one or more of these factors remains limited and often yields conflicting results. For instance, some studies found that greenspace exposure has a positive effect on mental health (Roberts and Helbich, 2021; Zhang et al., 2021), while others observed no significant association between greenspace exposure and mental health (Yoo et al., 2021; Zhou et al., 2020).

Lack of attention on socioeconomic disparities, different methods of measuring greenspace exposure and neglect of individual mobility might lead to those inconsistent findings. As for socio-economic disparities, different social groups with unequal accessibility to greenspace can have different health outcomes (Li et al., 201; Apparicio et al., 2017; Liu et al., 2021). For example, residents with higher socioeconomic status have access to better economic/social/political resources for improving their neighborhood greenspace, thus enjoying more health benefits from greenspace compared to those with lower socioeconomic status (Li et al., 2016). In Hong Kong, although much of the greenspace is public greenspace and is provided by the government, they may still be distributed unequally across different neighborhoods with different socioeconomic conditions, as dwellings in residential areas with more greenspaces have higher housing prices or rent and disadvantaged social groups are less able to afford the properties there (Lu et al., 2018). Current research lacks discussion about how the greenspace-mental health relationship will vary among communities with different socioeconomic statuses.

When assessing greenspace exposure, remote sensing data is commonly employed to measure the extent of greenery (Markevych et al., 2017). One widely used metric is the normalized difference vegetation index (NDVI), which quantifies vegetation canopy (Pettorelli et al., 2005). Nevertheless, the efficacy of such aerial measurements in accurately capturing individuals' on-the-ground perception of vegetation remains questionable (Li et al., 2018). This may have led to some contradictory findings on the relationships between greenspace exposure and mental health (Houlden et al., 2017; Taylor et al., 2018; Yang et al., 2019). In detail, the NDVI and street-view greenness evaluate different aspects of greenspace exposure and thus may affect mental health through different mechanisms. The NDVI measures top-down greenspace, which contains the mediating effect of areal greenspace on improving air quality and reducing noise, as vegetation absorbs air pollutants and noise (Dzhambov et al., 2018; Mueller et al., 2020; Son et al., 2021). On the other hand, street-view greenness measures human-scale greenspace. It mainly measures eye-level street plants, which can precisely reflect people's exposure to the greenery around them. This kind of exposure may mainly affect mental health through relieving pressure, encouraging outdoor activities and improving social cohesion (Li et al., 2018; Wang et al., 2021a). In a preliminary study conducted in Beijing, it was found that streetscape greenery exhibited a negative association with depressive symptoms among older adults, whereas the same association was not observed for NDVI-based measurements (Helbich et al., 2019). These findings suggest that streetscape and remotely sensed greenspace exposure metrics likely capture distinct aspects of greenery and may differ in their underlying pathways and impacts on individuals' mental health.

Regarding the neglect of individual mobility, previous studies on the relationship between greenspace exposure and mental health have traditionally taken a residence-based approach to exposure assessment. With this approach, individual exposure is assessed based on the administrative unit (e.g., census tract) that contains a person's residential address or a circular or network buffer around this address (Klompmaker et al., 2019). However, an administrative unit centered on a person's residential address is an overly simplistic context that may not be in line with the true spatial context of the person's environmental exposure. In real life, people typically encounter multiple exposures while performing their daily routines outside their homes (e.g., work, education or leisure activities) (Kwan, 2012, 2018a; Chaix, 2018; Zhang et al., 2018). Discrepancies between residence-based and dynamic contexts can result in exposure misclassification, which in turn may lead to biased estimates (Kwan, 2018b; Liu et al. 2023). For example, people who live in neighborhoods with lots of greenspaces but travel to their workplaces by public transit and work on high floors may enjoy much less greenery than people who live in neighborhoods with few greenspaces (e.g., tenement areas) but walk to workplaces via sidewalks or paths with lots of greenspaces. To address this issue, a substantial body of research has integrated GPS with measures of greenspace exposure to assess dynamic greenspace exposure and explore its relationship with residents' mental health (Mears et al., 2021). Most of these studies on greenspace-mental relationships tend to focus on the overall status of mental health over recent periods (South et al., 2021; Vanaken and Danckaerts, 2018; Wang et al., 2021b).

Nevertheless, individuals' mental states and emotions tend to fluctuate temporarily, contingent on various events and activity-specific contexts (Kuppens et al., 2010; Zhang et al., 2018). Taking short-term psychological stress as an example, numerous experimental or quasi-experimental studies have demonstrated that human physiological stress indicators (such as electroencephalogram [EEG] patterns and salivary cortisol levels) exhibit significant changes after a short period (mostly in 30 min) of exposure to greenspace (Jiang et al., 2014; Jones et al., 2021; Yao et al., 2024; Yuan et al., 2023). Such fluctuations cannot be captured by assessments focusing on the overall long-term status of mental health. While short-term stress is not equivalent to one's overall mental health state, the accumulation of short-term stress is likely to lead to mental health issues like anxiety or depression (Cohen, 2000; Stratakis and Chrousos, 1995). To capture fluctuations in short-term stress in daily real-world environments, several studies have employed the geographically-explicit Ecological Momentary Assessment (GEMA) methods to investigate the relationship between dynamic green space exposure and momentary stress level (Fancello et al., 2023; Kondo et al., 2020; Mennis et al., 2018; Zhang et al., 2023). However, most of these studies have adopted top-down approaches to greenspace measurement and have not achieved robust and consistent results like their experiment-based counterparts. For instance, Kondo et al. (2020) did not observe the stress-relieving effect of greenspace exposure based on land use data. It is still unclear how greenspace exposure (especially eye-level exposure) will affect people's momentary mental health under the influence of stressful events in their daily lives.

To fill these three research gaps, this study examines the relationship between different greenspace exposure (assessed with static and dynamic approaches) and people's mental health using survey data from two communities in Hong Kong. This study contributes to current greenspace-mental health research in the following three aspects: (1) Instead of using only a residence-based or static approach, we also measure individual dynamic greenspace exposure, mental health status and momentary stress using GPS, activity diaries, questionnaires and ecological momentary assessment (EMA). EMA is a monitoring method that assesses people's emotional or mental states at the moment they occur in natural settings. This approach reduces the potential biases and inaccuracies associated with retrospective reporting (Shiffman et al., 2008). It has finer spatiotemporal resolutions and can better reflect the spatiotemporal relationship between greenspace and mental health. (2) As different measurements of greenspace exposure might have different mental health effects, we use both NDVI-based and street view-based measurements of greenspace exposure in this study. Recent advancements in deep learning techniques have facilitated the automated extraction of streetscape greenery features, including trees and green walls, from street-view images obtained through online street-view services (LeCun et al., 2015). (3) As the association between greenspace and mental health may differ between different socio-economic contexts, we choose two communities with different socio-economic characteristics to compare the relationships between greenspace and the mental health of their residents.