Changes in Green Space Use During a COVID-19 Lockdown Are Associated With Both Individual and Green Space Characteristics

Introduction

On March 11 2020, the World Health Organisation declared COVID-19 an international health emergency. Following the announcement, countries across the world took various measures to slow the spread of the virus that causes COVID-19, including stay at home orders, closures of schools and workplaces, and significant limits on travel. These restrictions inevitably disrupted people’s daily activity patterns, social interactions, and use of leisure facilities such as gyms, cafes, and places of worship. Such disruptions caused uncertainty and instability, taking a toll on individual health and community wellbeing (Godinić et al., 2020; Wang and Boros, 2021). The effects were exacerbated in urban areas where dense living arrangements, longer and more severe restrictions, higher infection risk, and unequal access to services and nature coalesce (Maher, 1994; Sharifi and Khavarian-Garmsir, 2020; Mouratidis, 2021; Spotswood et al., 2021). Outdoor green spaces were one of few recreational places that remained accessible during periods of lockdown.

The importance of urban green spaces for ecosystem services—the benefits humans derive from nature—is well established in the literature (Bolund and Hunhammar, 1999; Breuste et al., 2013; Haase et al., 2014). Green space use is positively associated with general physical health and mental health and wellbeing (Grilli et al., 2020). For instance, Olszewska-Guizzo et al. (2021) suggests that COVID-19 restriction periods may have contributed to a heightened risk of mental health disorders, such as depression and/or reduced cognitive functioning and that green spaces are a way to offset the neuropsychological effects of such periods. In view of this, multiple studies have highlighted increased frequency of green space use during periods of COVID-19 restrictions (Berdejo-Espinola et al., 2021; da Schio et al., 2021; Lu et al., 2021) and suggested this is due to their multifunctionality and their capacity to mitigate some of the negative effects of the COVID-19 pandemic on human health and wellbeing. Changes in frequency of urban green space use during COVID-19 lockdowns show that urban green spaces were discovered by some during the pandemic (new park users; Berdejo-Espinola et al., 2021) and rediscovered by other individuals (previous users, re-engaged and increased their use; Venter et al., 2020). Yet, there is also evidence that some regular users decreased their use of green spaces during the pandemic (Khalilnezhad et al., 2021). Travel restrictions imposed during the pandemic present a useful study system for understanding the temporal and spatial patterns of green space use. These conditions present an opportunity to gain deeper insights into features of green spaces associated with increased use during times of crisis and the self-perceived benefits of visiting green spaces during these times.

Green spaces serve a variety of functions for individuals in the community and perceptions of green spaces can be diverse. Individuals may visit green spaces to access specific features such as walking paths, playgrounds, or ecosystem services such as shade and clean air, offered by that particular space. Yet, not all urban green spaces offer the same types, qualities, and quantities of ecosystem services and not every individual has the same demands for such services. Studies show that biophysical features of green spaces, such as tree and grass cover, vegetation complexity, blue spaces, biodiversity, size, and shape play a role on people’s decision to visit green spaces. After conducting a survey of individuals in Portugal, Madureira et al. (2018) found that richness and diversity of plant species was a highly attractive feature of green spaces. However, Shanahan et al. (2017), found that people surveyed in Brisbane and the United Kingdom did not preferentially visit green spaces with higher plant species richness or greater tree cover despite evidence that these features offer the potential for improved nature-based experiences and greater wellbeing benefits. Further, Zhang X. et al. (2020) found that irregularly shaped green spaces were associated with pronounced increases in people using the space for walking. Rey Gozalo et al. (2019) discovered that the size of green spaces was positively correlated with the frequency of walking, exercising, and relaxing in Spain. Location and the facilities within green spaces also play a role in how attractive the green space will be to potential visitors. Anecdotal evidence suggests that frequent green space use for physical activity was greater when the time taken to reach a green space was shorter (Dallimer et al., 2014). Voigt et al. (2014) found that facilities in green spaces that promote relaxation and leisure activities were preferred by residents in Berlin, Germany.

Although previous quantitative research has explored patterns of green space use and features of green spaces associated with frequency of use (Dallimer et al., 2014; Madureira et al., 2018; Dade et al., 2020), we know relatively little about green space characteristics that encourage people to repeatedly visit the same green spaces, and we have limited empirical knowledge of the extent to which patterns of green space use and features associated with use change when routine activities abruptly shift; as was the case during the COVID-19 pandemic. In this study we examine green space characteristics and sociodemographic factors associated with changes in frequency of green space use during a period of COVID-19 restrictions in Brisbane, Australia, and explore changes in self-perceived benefits associated with visiting green spaces. We (i) quantify how the frequency of green space use changed during the COVID-19 lockdown, (ii) determine the extent to which these changes were associated with sociodemographic and green space characteristics, (iii) identify changes in self-perceived benefits of using green spaces during the restrictions period, and (iv) analyse the association between changes in self-perceived benefits of using green spaces and biophysical green space characteristics.

Materials and Methods

We conducted a survey (n = 1,002) in the Brisbane Local Government Area, Australia, to capture people’s urban green space visitation patterns, including change in frequency of, and self-perceived benefits associated with, using urban green spaces before and during a stay-at-home restriction (hereafter “lockdown”). We only considered survey participants that provided information about the specific green spaces they visited before and during the pandemic (n = 372). We developed statistical models to identify which sociodemographic characteristics and urban green space biophysical variables were associated with changes in frequency of green space use and/or self-perceived benefits of green space use during COVID-19 compared to before the pandemic.

Study Area and Data Collection

The Brisbane Local Government Area (hereafter “Brisbane”) has an estimated human population of 1.27 million residents, approximately 4.9% of Australia’s population with a population density of 947 individuals per km² (Brisbane City Council, 2021). Brisbane’s green space network comprises more than 2,100 urban parks, picnic grounds, pocket parks, riverside spaces, botanic gardens, nature reserves, and beaches. These green spaces are widespread across the city containing both native and non-remnant vegetation cover that provides habitat and connectivity to over 80 different vegetation communities and over 2,300 species of wildlife and native plants (Shanahan et al., 2017; Brisbane City Council, 2020a,b).

Relative to other jurisdictions, Brisbane experienced few COVID-19 infections and related lockdowns during the first year of the pandemic. In Brisbane, the most significant lockdown restrictions were introduced on 23 March 2020 and continued until 2 May 2020 when measures were partially relaxed. During lockdown residents were instructed to work from home where possible; practice social distancing and good hygiene and leave home only for essential trips. The restrictions involved the closure of schools and universities, indoor fitness and sports facilities, and all food, drink, and cultural venues. One of the few reasons people were expressly permitted to leave home was for recreation or physical activity in a public green space. However, these trips were limited to the immediate residential neighbourhood and could be conducted with no more than two people from different households (Australian Government Department of Health, 2020).

Our survey was administered immediately following this lockdown period (survey distributed in June 2020) using a market research company (Q&A).¹ Participants were invited to complete the survey according to four nested stratification criteria that ensured the sample reflected a range of Brisbane’s demographic groups, broad socioeconomic spread, and an even spatial distribution across the city. The stratification rules were as follows: (a) an equal number of males and females, (b) an equal number of participants above and below 45 years of age, (c) income quartiles reflecting those of the whole Brisbane population and (d) an even distribution of participants’ location of residence across Brisbane (North/South/East/West side of Brisbane). The survey was delivered online by a market research company in accordance with the University of Queensland Human Research Ethics Approval, approval number 2020001073. All participants were at least 18 years old and provided written consent to participate in the survey. The survey asked participants to report whether they had visited an urban green space (a) before and (b) during the COVID-19 lockdown. If they answered yes, they were asked to list up to seven of the green spaces they visited, specifying the name and the frequency of use during both periods (never, once every 2 weeks, once a week, 2–3 days a week, 4–5 days a week, and 6–7 days a week). For this study, we used the first-listed green space in each time period (hereafter “nominated green space”) since 60.05% of survey respondents listed only one green space, and to simplify the interpretation of the analysis. We then geolocated each nominated green space with reference to a Brisbane City Council spatial dataset of public green spaces (Brisbane City Council, 2020c). We also asked survey participants to provide either their exact address, the address to the nearest 10 houses or the location of the street corner closest to their home, depending on what they felt most comfortable revealing. We used this information to geolocate residences using either the exact address when provided by participants, or the mid-point of a street in other cases.

Dependent Variables

Change in Frequency of Green Space Use During COVID-19 Lockdown

This variable was computed as the difference between frequency of green space use during the COVID-19 lockdown [never (1), once every 2 weeks (2), once a week (3), 2–3 days a week (4), 4–5 days a week (5) and 6–7 days a week (6)] and frequency of green space use before lockdown. Given that the frequency of use variable is not continuous we coded change in frequency into three categories, where 0 is no change in frequency of green space use during lockdown compared to before the pandemic; 1 is increased use of green spaces during the pandemic compared to before lockdown; and 2 is decreased use of green spaces during lockdown compared to before the pandemic.

Change in Self-Perceived Benefits of Urban Green Spaces

Participants were asked to report on the extent to which ten common benefits associated using green spaces had increased or decreased in importance during the pandemic related lockdown using a 5-point Likert scale (1 = much more important, 5 = much less important). For the analysis, we constructed two new variables named “psychological benefits” and “nature interactions.” For “psychological benefits” we constructed mean scale scores taking each participant’s mean response across three reasons in the survey questionnaire: reduction of stress, reduction of anxiety, and reduction in depression (Cronbach’s alpha = 0.94). We followed the same process to construct a “nature interactions” variable using three reasons: connection to nature, appreciation of the environment, and provision of clean air (Cronbach’s alpha = 0.89). We use the individual Likert variables as outcome variables for physical health benefits, spiritual connection, social interactions, and family interactions.

Independent Variables

Biophysical Factors

Vegetation characteristics (tree cover, grass cover, and foliage high diversity -FHD-) were derived from LiDAR data and other high-resolution imagery at a resolution of 30 m (see Caynes et al., 2016 for more details). We calculated the proportion of grass and tree cover within each green space using the raster and fasterize R packages package in R (Ross, 2018; Hijmans, 2021). FHD is a measure of vegetation vertical complexity that accounts for how evenly vegetation is distributed among vertical strata (Caynes et al., 2016). Vegetation (or forest) strata can be composed by three classes of vertical layers. The lowest layer is an herbaceous/shrub story with grasses, herbs, and shrubs. The understory has trees above the shrub layer and below the canopy; and the overstory comprises the highest layer of vegetation in a forest, including canopy trees (Berger and Puettmann, 2000). Caynes et al. (2016) separated the FHD data into five discrete height intervals, including very low (≥ 0.15– < 1 m), low (≥ 1–< 2 m), medium (≥ 2– < 5 m), high (≥ 5– < 10 m) and very high vegetation (≥ 10 m). As such, FHD values are high where vegetation is more evenly distributed across the vertical strata and low where vegetation is less evenly distributed. We calculated mean FHD within each green space using the raster and fasterize packages in R (Ross, 2018; Hijmans, 2021). Data on blue spaces present within a green space was obtained from the Brisbane City Council (2018). This is an indicator variable where a 1 indicates the presence of blue spaces in a green space and a 0 indicates no blue space. The shape of a green space was characterised using the shape index defined by McGarigal et al. (2012), calculated by dividing the perimeter of a green space by the minimum perimeter of that green space if the perimeter was rearranged to a maximally compact shape (circle). The shape index is minimum for maximally compact green spaces and increases as green space increases in complexity. For this study, we calculated each urban green space size and shape using the landscapemetrics R package (Hesselbarth et al., 2019).

Distance

We measure the distance from each participant’s residence to the nominated urban green space visited before the COVID-19 lockdown by calculating the Euclidean distance between the two points using the gDistance function from the regos R package (Bivand and Rundel, 2020). In some cases, this will overestimate or underestimate the distance that needs to be travelled to reach the green space, but a detailed integrated transport network dataset was not available.

Facilities

Data on facilities present within each urban green space were obtained from Brisbane City Council (2018). During the COVID-19 lockdown, use of several green space facilities, including playgrounds, dog parks, water fountains, barbeque, and picnic areas were restricted; therefore, we excluded them for our analysis. We included only presence of a carpark and access by public transport. Presence of a carpark is a dichotomous variable where 0 indicates no carpark is present at the individual’s nominated green space and 1 indicates that the green space does have an attached carpark. Public transport is also a dichotomous variable where 0 indicates no public transport access at the individual’s nominated green space and 1 indicates public transport is available at the green space.

Change of Urban Green Space Visited During Lockdown

We operationalised this variable by comparing the first nominated green space visited before and during lockdown. We then computed it by creating a dichotomous variable where 0 indicates no change and 1 indicates that the individual changed the green space visited during COVID-19 compared to before the pandemic related lockdown.

Statistical Analysis

We first conducted a series of descriptive analyses to evaluate the differences between urban green spaces visited before and during lockdown. Tables 1, 2 summarize descriptive statistics for each of the measures included in the regression analyses (Supplementary Data Sheets 1, 2 respectively).

TABLE 1

Table 1. Characteristics of the 372 study participants that responded to items regarding green space visited and reasons for visiting green spaces during the COVID-19 lockdown.

TABLE 2

Table 2. Characteristics of urban green spaces visited before and during the COVID-19 lockdown (n = 300).

Second, we estimated a series of regression models to address our research questions. The first of these models is a multinomial logistic regression analysis to determine how the frequency of green space use changed during the COVID-19 lockdown, and the extent to which this was associated with sociodemographic and green space characteristics. The dependent variable in this model is change in frequency of use of green space during the COVID-19 lockdown compared to before, coded into three categories, where 0 = no change, 1 = increased use of green space during lockdown, and 2 = decreased use of green space during lockdown. The category “no change in frequency of green space use” is used as the reference category in the multinomial regression. Our selection of independent variables is informed by the results of previous research on people’s perceptions and use of green spaces (Ode Sang et al., 2016; Braçe et al., 2021), and also scholarship on the frequency of green space use during COVID-19 (Uchiyama and Kohsaka, 2020; da Schio et al., 2021; Lu et al., 2021). Studies consistently identify gender, age, and income as associated with their perception of greens spaces and frequency of use of green spaces not only in “non-pandemic” times, but also during periods of COVID-19 restrictions (Braçe et al., 2021). The influence of green space characteristics on changes in the frequency of green space use during the COVID-19 pandemic is yet to be examined in depth. However, a study by Lu et al. (2021) examining use of urban green space in Asian cities during the pandemic found that during COVID-19 restrictions, urban residents preferred large nature parks to smaller urban parks and also visited green spaces closer to the city centre. Drawing on this limited research and previous research examining green space characteristics that influence use during “non-pandemic” times (for example see Abkar et al., 2010; Dade et al., 2020) we selected the following green space characteristics as independent variables: shape of green space, size of green space, presence of blue spaces, tree and grass cover, FHD, presence of car park, and presence of public transport node. We conducted Spearman’s correlations and VIF statistics. All VIFs were below 3.6 suggesting multicollinearity would not be problematic, and Spearman’s correlations were below 0.43 except in the case of tree and grass cover which returned a value of –0.66 (Supplementary Table 1). Tree cover was therefore excluded from the model, while grass cover of the green space visited before lockdown, FHD of the green space visited before lockdown, the presence of blue spaces, size of the green space visited before lockdown, and shape of the green space visited before lockdown were retained to represent biophysical features of the green space. To ensure model parsimony [lowest Akaike Information Criterion (AIC) (Burnham and Anderson, 2002)] and the selection of the best fitting model we added variables in a stepwise (hierarchical) fashion. Refer to Supplementary Table 2 for details on the stepwise models not presented in the main text. Model 1 contains only the sociodemographic characteristics; Model 2 is our final model including independent variables that are both informed by the literature and ensure model parsimony. We used the nnet R package (Venables and Ripley, 2002) to run multinomial logit regressions. The analytic sample included survey participants with responses for items regarding urban green space engagement both before and during COVID-19 lockdowns (n = 300, Supplementary Data Sheet 2).

The second stage of our empirical analyses comprises a suite of six regression models (Models 3–8) that examine the association between changes in the importance of self-perceived benefits of green spaces during the pandemic and particular biophysical green space characteristics. Models 3–8 investigate changes in the importance of specific benefits associated with visiting green spaces including: nature interactions; psychological health benefits; physical activity; spiritual connection; and social and family interactions. Each model includes independent variables representing gender, age, income, size of green space visited during lockdown, shape of green space visited during lockdown, grass cover of green space visited during lockdown, presence of blue spaces in the green space visited during lockdown, and FHD of green space visited during lockdown. Variables were selected based on previous research showing that size, shape, FHD, and grass cover can influence reasons for visiting green spaces (Dade et al., 2020). Although Dade et al. (2020) also found that facilities influenced reasons for attending green spaces, we did not include facilities in these models given that all facilities at public green spaces were closed during the COVID-19 lockdown. Models 3 and 4 are generalised linear regression models (GLM) assuming a gamma distribution because both the “nature interactions” (Model 3) and “psychological health benefits” (Model 4) dependent variables include positive-only values, and the error distributions are right skewed. The dependent variables “physical activity” (Model 5), “spiritual connection” (Model 6), “social interactions” (Model 7), and “family interactions” (Model 8) are ordinal variables containing five categories; thus, we estimate ordered logistic regression models for these data. We used the MASS R package to run the ordered logistic regressions (Venables and Ripley, 2002). The analytic sample for Models 3–8 include all survey participants who responded to items regarding green space visited and reasons for visiting green spaces during COVID-19 lockdown (n = 372, Supplementary Data Sheet 1).

All statistical models were evaluated using goodness of fit tests. For multinomial logistic regression models, we used a Hosmer–Lemeshow Test from the generalhoslem R package (Jay, 2019), for generalised linear models we used a Likelihood Ratio Test from the lmtest R package (Zeileis and Hothorn, 2002), and for the ordered logistic regression we used a Lipsitz Test from the generalhoslem R package. We calculated models’ R squared using the rsq R package (Zhang, 2021). All models were assessed for and met normality of residuals. All data analyses were carried out with R Studio V1.2 (RStudio Team, 2020).

Results

Descriptive Analysis

The sample comprised 372 individuals of which 48.3% were females, and the average age was 43 (Table 1). Of the 372 individuals, 78.8% either increased or did not change their frequency of green space use, while 21.2% decreased their visits to green spaces. There was a great deal of flux in which green spaces were visited, with 33.3% of individuals visiting a different green space during the lockdown than before.

Although in aggregate there appears to be little change between the characteristics of green spaces used before and during the pandemic (Table 2), examining individual cases suggests several important differences in the characteristics of green spaces used during COVID-19 compared to before at the individual level. For example, before COVID-19 lockdown, a large section of the population visited green spaces two to 11 times farther away from their residences than the green space visited during lockdown. Conversely, during the COVID-19 lockdown, fewer individuals visited green spaces two to five times farther away from their residences than the green space visited before lockdown. Although a large section of the sample population (66%) visited green spaces of approximately the same size before and during the COVID-19 lockdown. During lockdown, a fraction of the sample population (9 and 9.6%) visited green spaces 10 or more times smaller and larger, (respectively), than the green space visited before lockdown. Urban green space shape patterns did not appreciably change during the lockdown (Table 2). There is a slight upward shift in grass cover from before [Interquartile Range (IQR) = 22.29–50.6] to during (IQR = 24.8–52.0) the lockdown. Although some individuals (13.6%) visited green spaces with less grass cover during lockdown (from 25 to 50% cover to 0%) many individuals (20.3%) visited green spaces with larger amounts of grass cover when compared to the green space visited before lockdown. Foliage height diversity followed a similar trend, with 16 and 18% of the sample population visiting green spaces with lower and higher FHD indexes (Table 2).

Change in Frequency of Green Space Use During COVID-19 Lockdown

A summary of the associations between dependent and independent variables of Models 1 and 2 are shown in Table 3. Detailed results of the multinomial logistic regressions are displayed in Table 4 and are presented as relative risk ratios (RRR). RRR are the exponentiated regression coefficients and can be interpreted as the relative likelihood of an event occurring between two groups (e.g., change in frequency of use of green spaces during lockdown) given exposure to the variable of interest (e.g., shape of a green space) (Simon, 2001). The category “no change in frequency of green space use” is used as the reference category in the multinomial regression. Hosmer-Lemeshow goodness of fit tests indicates a significant improvement in model fit between the base demographic model (Model 1 p < 0.0361) and our final model (Model 2 p < 0.6949). The test score for Model 2 is well above the 0.05 threshold indicating the model is correct. It is also approaching one; therefore, demonstrating a better fit.

TABLE 3

Table 3. Significant associations between dependent and independent variables of eight regression models including sociodemographic and green space characteristics of green spaces used before (Model 1–2) and during (Model 3–8) a COVID-19 lockdown.

TABLE 4

Table 4. Change in frequency of use of urban green spaces during the COVID-19 lockdown explained by sociodemographic and green space characteristics variables.

Results of Model 2 demonstrate that individuals were more likely to increase their use of green space during lockdown compared to before if they were female compared to male (RRR = 2.413, p < 0.01) and younger compared to older (RRR = 0.970, p < 0.01). Individuals who visited the same green space during lockdown compared to the green space visited before the lockdown tended to increase their frequency of use (RRR = 0.570, p < 0.01). Urban green space characteristics also played a role. Individuals were more likely to increase their frequency of green space use during the lockdown compared to before the lockdown if the green space they visited had a carpark (RRR = 1.217, p < 0.01), was accessible by public transport (RRR = 1.289, p < 0.01), contained a blue space (RRR = 1.307, p < 0.01), and/or had lower FHD –vegetation is less evenly distributed across vertical layers– compared to green spaces with higher FHD (RRR = 0.211, p < 0.01). Individuals were more likely to decrease their frequency of green space use during the lockdown compared to before if they were female (RRR = 1.179, p < 0.01), or changed their nominated green space during the lockdown compared to before (RRR = 1.437, p < 0.01), individuals were also more likely to reduce their frequency of green space use during the lockdown, compared to before, if the green space they visited did not have a carpark (RRR = 0.913, p < 0.01), was not accessible by public transport (RRR = 0.760, p < 0.01), and had higher FHD –vegetation is more evenly distributed across the vertical structure– compared to green spaces with lower FHD (RRR = 2.933, p < 0.01).

Changes in Self-Perceived Benefits of Urban Green Space Use

In Models 3–8 we examine changes in self-perceived benefits associated with urban green space use during the pandemic compared to before. A summary of the associations between dependent and independent variables of Models 3–8 are shown on Table 3. Model 3 shows that being male (b = –0.01, p < 0.05) and older (b ≤ 0.001, p < 0.001) was significantly associated with reporting an increase in the importance of urban green spaces for nature interactions during the lockdown. Characteristics of the green space visited during the lockdown did not influence change in the importance of nature interactions. Model 3 accounts for 11.2% of variance in the outcome variable and was significant (Likelihood ratio test, p < 0.001). Similarly, Model 4 indicates that being male (b = –0.001, p < 0.01) and older than 43 (b = 0.001, p < 0.001) were significantly associated with reporting an increase in the importance of urban green spaces for psychological benefits during the lockdown. Model 4 accounts for 16.3% of the variance in the outcome variable, which was significant (Likelihood ratio test p < 0.001). Models 5–8 are ordered logistic regression models. Lipsitz goodness of fit tests for all models returned a p-value above the 0.05 threshold indicating that all models satisfy the proportional odds assumption and are correctly specified. Results of Model 5 reveal that younger individuals are more likely than older people to report increases in the importance of green spaces for physical activity during the lockdown than before (RRR = 0.983, p < 0.01). Model 6 indicates that the odds of reporting increased importance of green spaces for spiritual connection is higher for females compared to males (RRR = 1.516, p < 0.1), younger individuals compared to older individuals (RRR = 0.974, p < 0.01), and higher income earners (RRR = 1.071, p < 0.05). Using a smaller green space with greater shape complexity is also associated with higher odds of reporting increases in the importance of green spaces for spiritual connection during the lockdown (RRR = 0.999, p < 0.1; RRR = 1.211, p < 0.1). Results of Model 7 show the odds of reporting an increase in the importance of green spaces for social interactions during the COVID-19 pandemic was significantly higher for female compared to males (RRR = 1.461, p < 0.1), younger individuals (RRR = 0.983, p < 0.01), and higher income earners (RRR = 1.068, p < 0.05). Model 8 reveals that the odds of reporting increases in the importance of green spaces for family interactions were higher for females compared to males (RRR = 1.411, p < 0.1), younger people compared to older individuals (RRR = 0.982, p < 0.01), higher income earners (RRR = 1.062, p < 0.05), and those who used green spaces with more complex shapes during the lockdown compared to those who visited more compact green spaces (RRR = 1.265, p < 0.05).

Discussion

In this study we discovered that changes in the frequency of green space use and the importance of green spaces for self-perceived benefits use during a COVID-19 lockdown were associated with a number of individual characteristics and green space characteristics. In sum, our findings suggest four key takeaway points.

Our first key finding is that frequency of green space use increased during COVID-19 for some individuals but decreased for others. This pattern of change in frequency of green space use was particularly evident among females in the sample and suggests that the impact of COVID-19 restrictions on changes in frequency of green space use among females may have been moderated by green space characteristics. Our results provide support for those reported by Braçe et al. (2021) who found gender differences in perceptions of green space characteristics, with females attributing greater importance to characteristics, such as pleasant views, playgrounds, lightning, and safety than their male counterparts.

Secondly, individuals who did not change the park visited during the lockdown period were more likely to increase their frequency of visits. Increasing frequency of use may be associated with individuals having more free time and using green spaces for a greater variety of activities during COVID-19 when other facilities such as gyms and cafes were unavailable. Also, being familiar with the available features and other people who use the green space has the potential to increase opportunities for social, physical, and psychological benefits. Increased frequency of use was also associated with specific features of green spaces visited pre-lockdown; in particular the presence of a blue space. Recent research suggests that more frequent visits to blue spaces are associated with positive wellbeing and lower rates of mental distress Abkar et al. (2010) and White et al. (2021) suggests that water in green spaces is one of the most important biophysical factors that contribute to individuals’ positive mood change. Thus, the presence of blue spaces might enhance the odds of individuals increasing their frequency of green space use especially during times of stress.

Individuals who went to a different green space during lockdown were more likely to report a decrease in visitation compared to before the lockdown. This could be explained by the routine activity theory (Cohen and Felson, 1979), which posits that changes in the structure of the patterns of daily activity could explain other events. During COVID-19 lockdowns, people’s daily life was forced to drastically change because important urban nodes or places where daily activities are usually carried out were closed. Therefore, routine daily life during lockdown was characterised by reduced mobility, prolonged stays at home, telework or job losses, disruptions in social relationships, and declines in physical activity levels; all with implications for mental wellbeing (Ogden, 2020; Salari et al., 2020; Biroli et al., 2021; Mckeown et al., 2021; Mouratidis, 2021). One potential explanation for why individuals visited a different green space during the lockdown is that green spaces took on some of the functions of other nodes during the lockdown (for example gyms, cafes). That is, individuals visited green spaces with different characteristics, to satisfy unmet demands resulting from the restrictions. However, it is also plausible that the new green space visited during lockdown did not fulfil individuals’ needs and may have caused them to decrease their frequency of use. For example, we found that green spaces with a higher proportion of canopy cover among forest strata (FHD), which tend to be located in the outskirts of the city, were associated with a remarkable decrease in visits during lockdown. In contrast, green spaces with fewer vegetation vertical layers (which are more characteristic of neighbourhood parks) were associated with increases in use. Our results may suggest that local green spaces and in closer proximity to people’s homes can act as nature-based solution by fulfiling needs that are not necessarily related to nature interactions during times of crisis.

Our third finding is that people travelled to green spaces two to 11 times closer to their home during the lockdown compared to before, suggesting that the mobility restrictions imposed to stop the spread of COVID-19 may have changed individuals’ daily activity patterns and further supporting a role for the routine activity theory in explaining the changes. Therefore, visiting urban green spaces with different characteristics may have served as a pathway to accommodate people’s new daily activity patterns and needs resulting from the COVID-19 restrictions. However, the opportunity to access the services provided by either the same or a different green space may also depend on the configuration of the urban landscape and individuals’ access to green spaces. Proximity was not the only indicator of accessibility that was associated with increased frequency of green space use during the lockdown. Individuals who visited green spaces with carparks and/or public transport nodes were also more likely to increase their frequency of green space use during lockdown. Although Brisbane residents have excellent accessibility to green spaces, relative to many other capital city residents (Berdejo-Espinola et al., 2021) this finding illustrates that ensuring equitable access across all sociodemographic groups and neighbourhood settings is pivotal for facilitating green space use to satisfy unmet needs during times of crisis.

Finally, our results show that the importance of self-perceived benefits associated with using green spaces changed markedly during the COVID-19 lockdown, and that these changes were associated with sociodemographic variables and to a lesser extent with green space characteristics. Despite the closure of facilities within green spaces, such a playgrounds, benches, and sporting features, on average, individuals reported increased importance of green spaces for their cultural ecosystem services. This reinforces the importance of biophysical features of green spaces in everyday urban life. We found that males and older individuals were more likely to express greater increases than females and younger people in the importance of green space for nature interactions and mental health benefits during the COVID-19 lockdown compared to before the pandemic. This might be related to changes in routine activity patterns for males, including teleworking, looking after children, and reallocation of household responsibilities that posed new stressors in their daily life (Douglas et al., 2020; Biroli et al., 2021). While the importance of urban green spaces for mental wellbeing is well established in the literature (Keniger et al., 2013; Zhang J. et al., 2020), these benefits have been articulated more strongly by females in times of crisis (Burnett et al., 2021; Gastelum-Strozzi et al., 2021; Lopez et al., 2021). In fact, Gastelum-Strozzi et al. (2021) have shown that males in Mexico City indicated feeling remarkably fewer effects of the pandemic on mental health than females. It may be that for males living in Brisbane, utilising green spaces for mental wellbeing has emerged as a condition of the COVID-19 pandemic after making other places for social interactions and physical health unavailable. In contrast, we found that females were more likely than males to report increased importance of green spaces for spiritual, social, and family interactions. This reflects research conducted by Burnett et al. (2021) in the United Kingdom, who found that females who reduced their use of green spaces during the pandemic were more likely to agree that they missed interacting with others in green spaces than males. Future research could explore the idiosyncrasies across sexes.

The shape of green spaces also emerged as an important feature associated with changes in benefits associated with green space use. Individuals using irregularly shaped green spaces were more likely to report increased importance of the green space for social interactions during the lockdown. Scholarship on social cognition suggests that social gazing—visual behavior between two or more individuals (Kleinke, 1986)—is one form to communicate and an important component of social interaction (Emery, 2000). Thus, spending time in less compact green spaces might help increase the number of everyday visual encounters where people can at least see others while socially distancing and stimulating a sense of social connection even when physical distancing is enforced. Further, individuals using irregularly shaped green spaces were more likely to report increased importance in green spaces for spiritual connection during the COVID-19 pandemic. Across Brisbane, the shape of urban green spaces tend to be relatively compact, with a mean shape index of 1.45 [compact shape = 1 according to McGarigal et al. (2012)], meaning that irregularly shaped green spaces are less common. Thus, the uniqueness of compact green spaces may facilitate the development of a sense of place attachment—the emotional bond between person and place (Scannell and Gifford, 2010). Stedman (2003) suggests that individuals do not become directly attached to the physical features of a green space, but rather to the meaning that those features represent. Thus, attachment to a green space could be related to the attachment to the experiences individuals can have in a green space, perhaps nature experiences (for example, birdwatching), or to the attachment to the social interactions that the green space affords them. Unique identifiers of the green space territory, such as a unique shape are important to facilitate opportunities for social connection, particularly during stressful times. In this context, future pandemic-resilient urban planning and policy might consider the size and shape of green spaces to ensure safe access to green spaces with appropriate social distancing.

While this study extended current knowledge on urban green spaces and their use during the COVID-19 pandemic it is not without limitations. We note that the sample is relatively small, and caution should be used when generalising the findings to other settings. Further, the study is conducted in Brisbane, Australia which had been relatively protected from COVID-19 compared to other regions during the early phase of the pandemic. Another limitation is that we asked individuals to report about the green spaces visited before and during a COVID-19 lockdown and did not account for the influence of household dynamics and how other members of the household may affect green space visitation. We encourage new research to explore this avenue. Lastly, our data are retrospective, given that participants reported their frequency of use and perceptions of the benefits of green spaces before and during the restrictions period 30–90 days after their use. This means that there may be some inaccuracy in people’s reports due to the memory recall errors, however, we expect recall bias to be limited due to the relatively short reference period (Ayhan and Isiksal, 2004).

Our findings suggest a number of areas for future research and considerations for policy and practice. First, they highlight the need to ensure accessibility to urban green spaces with appropriate ecosystem services for all sociodemographic groups and all geographic areas, before, during, and after disasters/collective crises as they are important multifunctional sites that can serve as a nature-based coping mechanism for communities and individuals (Berdejo-Espinola et al., 2021). Accessibility to green spaces must consider not only proximity to, and availability of green spaces, but also modes of access, such as car parking, and public, and active transport (the latter not considered in this study). Yet, making green spaces available is only the first step; ensuring that they are matched to the needs of the population is also vital. Our findings also suggest that features found to be associated with mental health and wellbeing benefits such as blue spaces may facilitate higher usage during periods of crises. Thus, the presence of blue spaces may have the potential to benefit communities. Further, the complexity of green space shape may be important for facilitating place attachment and supporting the social function of green spaces. Planning and co-designing crisis-resilient urban areas while considering the key individual and community factors might lead to more accessible, functional urban green spaces.

Data Availability Statement

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

Ethics Statement

The studies involving human participants were reviewed and approved by The University of Queensland Human Research Ethics Committee. The patients/participants provided their written informed consent to participate in this study.

Author Contributions

VB-E and RZ: conceptualisation, formal analysis, and writing original draft. VB-E, RZ, RF, and AS-C: methodology. VB-E, RZ, AS-C, JR, and RF: writing-review and editing. All authors contributed to the article and approved the submitted version.

Funding

VB-E received funding from the Australian RTP Scholarships.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo.2022.804443/full#supplementary-material

Footnotes

1. ^ https://qandapanel.com.au/

References

Abkar, M., Kamal, M., Mariapan, M., Maulan, S., and Sheybanic, M. (2010). Role of Urban Green Spaces in Mood Change. Austral. J. Basic Appl. Sci. 4, 5352–5361.

Google Scholar

Australian Government Department of Health (2020). Social distancing for coronavirus (COVID-19) [WWW Document]. Canberra: Australian Government Department of Health.

Google Scholar

Ayhan, Ö, and Isiksal, S. (2004). Memory Recall Errors in Retrospective Surveys: A Reverse Record Check Study. Qual. Quant. 38, 475–493. doi: 10.1007/s11135-005-2643-7

CrossRef Full Text | Google Scholar

Berdejo-Espinola, V., Suárez-Castro, A. F., Amano, T., Oh, R. R. Y., Fielding, K. S., and Fuller, R. A. (2021). Urban green space use during a time of stress: A case study during the COVID-19 pandemic in Brisbane, Australia. People Nat. 3:10218. doi: 10.1002/pan3.10218

PubMed Abstract | CrossRef Full Text | Google Scholar

Berger, A. L., and Puettmann, K. J. (2000). Overstory Composition and Stand Structure Influence Herbaceous Plant Diversity in the Mixed Aspen Forest of Northern Minnesota. Am. Midland Natural. 143, 111–125. doi: 10.1674/0003-0031(2000)143[0111:ocassi]2.0.co;2

CrossRef Full Text | Google Scholar

Biroli, P., Bosworth, S., Della Giusta, M., Di Girolamo, A., Jaworska, S., and Vollen, J. (2021). Family Life in Lockdown. Front. Psychol. 12:687570. doi: 10.3389/fpsyg.2021.687570

PubMed Abstract | CrossRef Full Text | Google Scholar

Bivand, R., and Rundel, C. (2020). rgeos: Interface to Geometry Engine - Open Source (‘GEOS’). R package version 0.5-3. Vienna: R Core Team.

Google Scholar

Bolund, P., and Hunhammar, S. (1999). Ecosystem services in urban areas. Ecol. Econom. 29, 293–301.

Google Scholar

Braçe, O., Garrido-Cumbrera, M., and Correa-Fernández, J. (2021). Gender differences in the perceptions of green spaces characteristics. Soc. Sci. Quart. 102, 2640–2648. doi: 10.1111/ssqu.13074

CrossRef Full Text | Google Scholar

Breuste, J., Schnellinger, J., Qureshi, S., and Faggi, A. (2013). Urban Ecosystem services on the local level: Urban green spaces as providers. Ekológia 32, 290–304.

Google Scholar

Brisbane City Council (2020a). Bushland reserves map. Brisbane, BNE: Brisbane City Council.

Google Scholar

Brisbane City Council (2020b). Biodiversity in Brisbane. Brisbane, BNE: Brisbane City Council.

Google Scholar

Brisbane City Council (2020c). Park — Locations - Data. Brisbane, BNE: Brisbane City Council.

Google Scholar

Brisbane City Council (2021). Brisbane Community Profiles. Brisbane, BNE: Brisbane City Council.

Google Scholar

Brisbane City Council (2018). Park Facilities and Assets locations - Park Facilities and Assets locations — CSV - Data. Brisbane: Brisbane City Council.

Google Scholar

Burnett, H., Olsen, J. R., Nichols, N., and Mitchell, R. (2021). Change in time spent visiting and experiences of green space following restrictions on movement during the COVID-19 pandemic: a nationally representative cross-sectional study of UK adults. BMJ Open 11:44067. doi: 10.1136/bmjopen-2020-044067

PubMed Abstract | CrossRef Full Text | Google Scholar

Burnham, K., and Anderson, D. (2002). Model Selection and Multimodel Inference - A Practical Information-Theoretic Approach. New York, NY: Springer.

Google Scholar

Caynes, R. J. C., Mitchell, M. G. E., Wu, D. S., Johansen, K., and Rhodes, J. R. (2016). Using high-resolution LiDAR data to quantify the three-dimensional structure of vegetation in urban green space. Urban Ecosyst. 19, 1749–1765. doi: 10.1007/s11252-016-0571-z

CrossRef Full Text | Google Scholar

Cohen, L. E., and Felson, M. (1979). Social Change and Crime Rate Trends: A Routine Activity Approach. Am. Sociol. Rev. 44, 588–608. doi: 10.2307/2094589

CrossRef Full Text | Google Scholar

da Schio, N., Philips, A., and Fransen, K. (2021). The impact of the COVID-19 pandemic on the use of and attitudes towards urban forests and green spaces: Exploring the instigators of change in Belgium. Urban For. Urban Green. 65:127305. doi: 10.1016/j.ufug.2021.127305

CrossRef Full Text | Google Scholar

Dade, M. C., Mitchell, M. G. E., Brown, G., and Rhodes, J. R. (2020). The effects of urban greenspace characteristics and socio-demographics vary among cultural ecosystem services. Urban For. Urban Green. 49:126641. doi: 10.1016/j.ufug.2020.126641

CrossRef Full Text | Google Scholar

Dallimer, M., Davies, Z. G., Irvine, K. N., Maltby, L., Warren, P. H., Gaston, K. J., et al. (2014). What Personal and Environmental Factors Determine Frequency of Urban Greenspace Use? Int. J. Environ. Res. Public Health 11, 7977–7992. doi: 10.3390/ijerph110807977

PubMed Abstract | CrossRef Full Text | Google Scholar

Douglas, M., Katikireddi, S. V., Taulbut, M., McKee, M., and McCartney, G. (2020). Mitigating the wider health effects of covid-19 pandemic response. BMJ 369:m1557. doi: 10.1136/bmj.m1557

PubMed Abstract | CrossRef Full Text | Google Scholar

Emery, N. J. (2000). The eyes have it: the neuroethology, function and evolution of social gaze. Neurosci. Biobehav. Rev. 24, 581–604. doi: 10.1016/s0149-7634(00)00025-7

PubMed Abstract | CrossRef Full Text | Google Scholar

Fagerland, M. W., and Hosmer, D. W. (2012). A Generalized Hosmer–Lemeshow Goodness-of-Fit Test for Multinomial Logistic Regression Models. Stata J. 12, 447–453. doi: 10.7759/cureus.10054

PubMed Abstract | CrossRef Full Text | Google Scholar

Gastelum-Strozzi, A., Infante-Castañeda, C., Figueroa-Perea, J. G., and Peláez-Ballestas, I. (2021). Heterogeneity of COVID-19 Risk Perception: A Socio-Mathematical Model. Int. J. Environ. Res. Public Health 18:11007. doi: 10.3390/ijerph182111007

PubMed Abstract | CrossRef Full Text | Google Scholar

Godinić, D., Obrenovic, B., and Hudaykulov, A. (2020). Effects of Economic Uncertainty on Mental Health in the COVID-19 Pandemic Context: Social Identity Disturbance, Job Uncertainty and Psychological Well-Being Model. Int. J. Innovat. Econom. Dev. 6, 61–74. doi: 10.18775/ijied.1849-7551-7020.2015.61.2005

CrossRef Full Text | Google Scholar

Grilli, G., Mohan, G., and Curtis, J. (2020). Public park attributes, park visits, and associated health status. Landsc. Urban Plann. 199:103814. doi: 10.1016/j.landurbplan.2020.103814

CrossRef Full Text | Google Scholar

Haase, D., Larondelle, N., Andersson, E., Artmann, M., Borgström, S., Breuste, J., et al. (2014). A Quantitative Review of Urban Ecosystem Service Assessments: Concepts, Models, and Implementation. AMBIO 43, 413–433. doi: 10.1007/s13280-014-0504-0

PubMed Abstract | CrossRef Full Text | Google Scholar

Hesselbarth, M., Sciaini, M., Wiegand, K., Nowosad, J., and Kimberly, A. (2019). landscapemetrics: an open-source R tool to calculate landscape metrics. Ecography 42:1657.

Google Scholar

Hijmans, R. (2021). raster: Geographic Data Analysis and Modeling. Vienna: R Core Team.

Google Scholar

Jay, M. (2019). generalhoslem: Goodness of Fit Tests for Logistic Regression Models. Vienna: R Core Team.

Google Scholar

Keniger, L. E., Gaston, K. J., Irvine, K. N., and Fuller, R. A. (2013). What are the Benefits of Interacting with Nature? Int. J. Environ. Res. Public Health 10, 913–935. doi: 10.3390/ijerph10030913

PubMed Abstract | CrossRef Full Text | Google Scholar

Khalilnezhad, M. R., Ugolini, F., and Massetti, L. (2021). Attitudes and Behaviors toward the Use of Public and Private Green Space during the COVID-19 Pandemic in Iran. Land 10:1085. doi: 10.3390/land10101085

CrossRef Full Text | Google Scholar

Kleinke, C. L. (1986). Gaze and eye contact: A research review. Psychol. Bull. 100, 78–100. doi: 10.1037/0033-2909.100.1.78

PubMed Abstract | CrossRef Full Text | Google Scholar

Lopez, B., Kennedy, C., Field, C., and McPhearson, T. (2021). Who benefits from urban green spaces during times of crisis? Perception and use of urban green spaces in New York City during the COVID-19 pandemic. Urban For. Urban Green. 65:127354. doi: 10.1016/j.ufug.2021.127354

PubMed Abstract | CrossRef Full Text | Google Scholar

Lu, Y., Zhao, J., Xueying, W., and Lo, S. M. (2021). Escaping to nature during a pandemic: A natural experiment in Asian cities during the COVID-19 pandemic with big social media data. Sci. Total Environ. 777:146092.

Google Scholar

Madureira, H., Nunes, F., Vidal Oliveira, J., and Madureira, T. (2018). Preferences for Urban Green Space Characteristics: A Comparative Study in Three Portuguese Cities. Environments 5:23.

Google Scholar

Maher, C. (1994). Residential Mobility, Locational Disadvantage And Spatial Inequality In Australian Cities. Urban Policy Res. 12, 185–191.

Google Scholar

McGarigal, K., Cushman, S., and Ene, E. (2012). “FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps,” in Computer software program produced by the authors at the University of Massachusetts, Amherst, (Amherst, MA: University of Massachusetts). doi: 10.1007/s10661-015-4727-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Mckeown, B., Poerio, G. L., Strawson, W. H., Martinon, L. M., Riby, L. M., Jefferies, E., et al. (2021). The impact of social isolation and changes in work patterns on ongoing thought during the first COVID-19 lockdown in the United Kingdom. PNAS 118:2102565118. doi: 10.1073/pnas.2102565118

PubMed Abstract | CrossRef Full Text | Google Scholar

Mouratidis, K. (2021). How COVID-19 reshaped quality of life in cities: A synthesis and implications for urban planning. Land Pol. 2021:105772. doi: 10.1016/j.landusepol.2021.105772

PubMed Abstract | CrossRef Full Text | Google Scholar

Ode Sang, Å, Knez, I., Gunnarsson, B., and Hedblom, M. (2016). The effects of naturalness, gender, and age on how urban green space is perceived and used. Urban For. Urban Green. 18, 268–276. doi: 10.1016/j.ufug.2016.06.008

CrossRef Full Text | Google Scholar

Ogden, R. S. (2020). The passage of time during the UK Covid-19 lockdown. PLoS One 15:e0235871. doi: 10.1371/journal.pone.0235871

PubMed Abstract | CrossRef Full Text | Google Scholar

Olszewska-Guizzo, A., Mukoyama, A., Naganawa, S., Dan, I., Husain, S. F., Ho, C. S., et al. (2021). Hemodynamic Response to Three Types of Urban Spaces before and after Lockdown during the COVID-19 Pandemic. Int. J. Environ. Res. Public Health 18:6118. doi: 10.3390/ijerph18116118

PubMed Abstract | CrossRef Full Text | Google Scholar

Rey Gozalo, G., Barrigón Morillas, J. M., and Montes González, D. (2019). Perceptions and use of urban green spaces on the basis of size. Urban For. Urban Green. 46:126470.

Google Scholar

Ross, N. (2018). fasterize: Fast Polygon to Raster Conversion. Vienna: R Core Team.

Google Scholar

RStudio Team (2020). RStudio: Integrated Development for R. Boston, MA: RStudio Team.

Google Scholar

Salari, N., Hosseinian-Far, A., Jalali, R., Vaisi-Raygani, A., Rasoulpoor, S., Mohammadi, M., et al. (2020). Prevalence of stress, anxiety, depression among the general population during the COVID-19 pandemic: a systematic review and meta-analysis. Globalizat. Health 16:57. doi: 10.1186/s12992-020-00589-w

PubMed Abstract | CrossRef Full Text | Google Scholar

Scannell, L., and Gifford, R. (2010). Defining place attachment: A tripartite organizing framework. J. Environ. Psychol. 30, 1–10.

Google Scholar

Shanahan, D. F., Cox, D. T. C., Fuller, R. A., Hancock, S., Lin, B. B., Anderson, K., et al. (2017). Variation in experiences of nature across gradients of tree cover in compact and sprawling cities. Landsc. Urban Plann. 157, 231–238. doi: 10.1016/j.landurbplan.2016.07.004

CrossRef Full Text | Google Scholar

Sharifi, A., and Khavarian-Garmsir, A. R. (2020). The COVID-19 pandemic: Impacts on cities and major lessons for urban planning, design, and management. Sci. Total Environ. 749:142391. doi: 10.1016/j.scitotenv.2020.142391

PubMed Abstract | CrossRef Full Text | Google Scholar

Simon, S. D. (2001). Understanding the odds ratio and the relative risk. J. Androl. 22, 533–536.

Google Scholar

Spotswood, E. N., Benjamin, M., Stoneburner, L., Wheeler, M. M., Beller, E. E., Balk, D., et al. (2021). Nature inequity and higher COVID-19 case rates in less-green neighbourhoods in the United States. Nat. Sustain. 2021, 1–7.

Google Scholar

Stedman, R. C. (2003). Is It Really Just a Social Construction?: The Contribution of the Physical Environment to Sense of Place. Soc. Nat. Resour. 16, 671–685. doi: 10.1080/08941920309189

CrossRef Full Text | Google Scholar

Uchiyama, Y., and Kohsaka, R. (2020). Access and Use of Green Areas during the COVID-19 Pandemic: Green Infrastructure Management in the “New Normal.”. Sustainability 12:9842. doi: 10.3390/su12239842

CrossRef Full Text | Google Scholar

Venables, W. N., and Ripley, B. D. (2002). Modern Applied Statistics with S. Fourth. New York, NY: Springer.

Google Scholar

Venter, Z., Barton, D., Gundersen, V., Figari, H., and Nowell, M. (2020). Urban nature in a time of crisis: recreational use of green space increases during the COVID-19 outbreak in Oslo, Norway. Environ. Res. Lett. 15:104075. doi: 10.1088/1748-9326/abb396

CrossRef Full Text | Google Scholar

Voigt, A., Kabisch, N., Wurster, D., Haase, D., and Breuste, J. (2014). Structural Diversity: A Multi-dimensional Approach to Assess Recreational Services in Urban Parks. AMBIO 43, 480–491. doi: 10.1007/s13280-014-0508-9

PubMed Abstract | CrossRef Full Text | Google Scholar

Wang, F., and Boros, S. (2021). Mental and physical health in general population during COVID-19: Systematic review and narrative synthesis. PJAMP 13, 91–99. doi: 10.29359/bjhpa.13.1.10

CrossRef Full Text | Google Scholar

White, M. P., Elliott, L. R., Grellier, J., Economou, T., Bell, S., Bratman, G. N., et al. (2021). Associations between green/blue spaces and mental health across 18 countries. Sci. Rep. Vol. 11:8903.

Google Scholar

Zeileis, A., and Hothorn, T. (2002). Diagnostic Checking in Regression Relationships. R News 2, 7–10.

Google Scholar

Zhang, D. (2021). rsq: R-Squared and Related Measures. Vienna: R Core Team.

Google Scholar

Zhang, J., Yu, Z., Zhao, B., Sun, R., and Vejre, H. (2020). Links between green space and public health: a bibliometric review of global research trends and future prospects from 1901 to 2019. Environ. Res. Lett. 15:063001.

Google Scholar

Zhang, X., Melbourne, S., Sarkar, C., Chiaradia, A., and Webster, C. (2020). Effects of green space on walking: Does size, shape and density matter? Urban Stud. 57:4209802090273.

Google Scholar