Effect of Forest Users’ Stress on Perceived Restorativeness, Forest Recreation Motivation, and Mental Well-Being during COVID-19 Pandemic

Even though the COVID-19 pandemic has discouraged travel and people’s movements, the number of visitors to forests near cities which are easily accessible by private vehicle is increasing in Korea. This study aims to investigate the relationship between stress, perceived restorativeness, forest recreation motivation, and the mental well-being of forest users. A survey of forest users was conducted at three recreational forests near Seoul in the summer of 2020. A total of 1196 forest users (613 males and 583 females) participated in the study. As a result of the data analysis, it was found that stress had a negative correlation with perceived restorativeness, forest recreation motivation, and mental well-being; perceived restorativeness had a positive correlation with mental well-being, and forest recreation motivation had a positive correlation with mental well-being. For the relationship between stress and mental well-being, the fitness index that was mediated by the perceived restorativeness and the forest recreation motivation found that the model was statistically suitable. Through this study, a research model was derived that, if the stress of forest users is reduced, direct or indirect effects on perceived restorativeness, forest recreation motivation, and mental well-being are increased. Further, a multi-group analysis found that the effect of perceived restorativeness and forest recreation motivation on the mental well-being of the male group was higher than the effect on the female group. Using this research model to find ways to promote health in forests can be utilized for forest management or forest healing.


Introduction
Currently, 54% of the world's population lives in urban areas. This figure has increased by more than 30% since 1950, and by 2050, 66% of the world's population is expected to be urban [1]. From a psychological point of view, the urban lifestyle increases the demand for cognitive activities such as working and maintaining social relationships [2]. Humans use their resources and energy to relieve stressful situations; however, persistently relying on this method can cause disease. Low levels of stress can be tolerated. Still, when the severity of the stress increases or persists, it destroys biological systems. This is described as the "fight-or-flight response" [3]. Differences in individual responses to stressful situations are likely to lead to changes in immune function and the spread of disease [4,5] to support health and well-being [6]. In particular, forests have been reported to have physiological and psychological recovery effects; for example, by being in contact with a forest environment for just 15 min. Furthermore, due to the COVID-19 pandemic, Hypothesis 2 (H2). Perceived restorativeness will have a positive effect on mental well-being.

Hypothesis 3 (H3). The forest recreation motivation will have a positive effect on mental wellbeing.
Hypothesis 4 (H4). Perceived restorativeness will mediate the effect of stress on mental well-being Hypothesis 5 (H5). Forest recreation motivation will mediate the effect of stress on mental wellbeing.

Study Subject and Research Method
This study was conducted by people who visited Mt. Gwanak Urban Forest, Mt. Bukhan National Park, and Mt. Yumyeong Natural Recreation Forest in the Seoul Metropolitan Area and the Gyeonggi-do province ( Figure 1). Since it was impossible to control the movement of forest users, the survey was conducted for adults who were 19 years of age or older, wearing a mask, and descending from the midpoint of each forest on weekends and weekdays between 1 May and 15 July 2020, from 9 am to noon and from 2 to 5 pm. The average temperature of the site was 20-25 °C (Refer to an article [22] for detailed information related to the location of the study).

Psychosocial Well-Being Index-SF
The PWI-SF that was developed by Chang [30] is an 18-item Likert-4 scale used to measure the physical and psychological state experienced or felt by a person in recent weeks. The PWI-SF gives each item a score of (0~3) and the sum of the scores for each item is distributed between 0-54 points: up to 8 points (health group); <9-26 points; (potential stress group); <27-54 points (high-risk group). A higher score indicates a higher stress level [31]. As a result of analyzing the reliability of social and psychological stress, Cronbach's α was 0.88.

Psychosocial Well-Being Index-SF
The PWI-SF that was developed by Chang [30] is an 18-item Likert-4 scale used to measure the physical and psychological state experienced or felt by a person in recent weeks. The PWI-SF gives each item a score of (0~3) and the sum of the scores for each item is distributed between 0-54 points: up to 8 points (health group); <9-26 points; (potential stress group); <27-54 points (high-risk group). A higher score indicates a higher stress level [31]. As a result of analyzing the reliability of social and psychological stress, Cronbach's α was 0.88.

Perceived Restorativeness Scale
The PRS that was developed by Hartig et al. [32] is a tool to measure the psychological recovery effect of the recovery environment. This scale is a Likert-7 scale consisting of 4 factors, including 16 items: two items of 'being away'; five items of 'fascination'; four items of 'coherence'; and five items of 'compatibility'. A scale that was adapted by Kim [33] was used in this study. As a result of reliability analysis, Cronbach's α was found to be: 'being away 0.85'; 'fascination 0.74'; 'coherence 0.60'; and 'compatibility 0.87'. The overall score was 0.87.

Forest Recreation Motivation Scale
The FRMS that was developed by Yeon [34] is a tool to measure users' motivation to manage recreation resources and their recreation experience. This scale is a Likert-5 scale consisting of 22 items and five factors: 'nature-friendly motive', six items; 'social-friendly motive', seven items; 'daily life change motive', four items; 'self-exploration motive', three items; and 'natural achievement motive', two items. As a result of analyzing the reliability with the FRMS, Cronbach's α was 'nature-friendly motive 0.75'; 'social-friendly motive 0.86'; 'motivation to pursue changes in daily life 0.65'; 'self-exploration motive 0.62'; and 'nature achievement motive 0.74'. The overall score was 0.91.

Mental Well-Being Scale (K-MHC-SF: Korean Version of the Mental Health Continuum Short Form)
According to the happiness theory, the mental well-being scale that was developed by Keyes [35,36] is a tool to measure a person's level of mental well-being. This scale is a Likert-6 scale consisting of 14 items and three factors: emotional well-being (questions 1-3); psychological well-being (questions 9-14); and social well-being (questions 4-8). As a result of analyzing trust using the mental well-being scale that was adapted by Lim et al. [37], Cronbach's α was 'emotional well-being 0.94'; 'well-being 0.87'; and 'Psychological wellbeing 0.93,'. The overall score was 0.95.

Research Model and Hypothesis
Structural equations are used to verify the stress, perceived restorativeness, motivation for forest recreation, and the mental well-being of forest users.
The research model is identified by setting the hypothesis as follows ( Figure 2). The PRS that was developed by Hartig et al. [32] is a tool to measure the psychological recovery effect of the recovery environment. This scale is a Likert-7 scale consisting of 4 factors, including 16 items: two items of 'being away'; five items of 'fascination'; four items of 'coherence'; and five items of 'compatibility'. A scale that was adapted by Kim [33] was used in this study. As a result of reliability analysis, Cronbach's α was found to be: 'being away 0.85′; 'fascination 0.74′; 'coherence 0.60′; and 'compatibility 0.87′. The overall score was 0.87.

Forest Recreation Motivation Scale
The FRMS that was developed by Yeon [34] is a tool to measure users' motivation to manage recreation resources and their recreation experience. This scale is a Likert-5 scale consisting of 22 items and five factors: 'nature-friendly motive', six items; 'social-friendly motive', seven items; 'daily life change motive', four items; 'self-exploration motive', three items; and 'natural achievement motive', two items. As a result of analyzing the reliability with the FRMS, Cronbach's α was 'nature-friendly motive 0.75′; 'social-friendly motive 0.86′; 'motivation to pursue changes in daily life 0.65′; 'self-exploration motive 0.62′; and 'nature achievement motive 0.74′. The overall score was 0.91.

Mental Well-Being Scale (K-MHC-SF: Korean Version of the Mental Health Continuum Short Form)
According to the happiness theory, the mental well-being scale that was developed by Keyes [35,36] is a tool to measure a person's level of mental well-being. This scale is a Likert-6 scale consisting of 14 items and three factors: emotional well-being (questions 1-3); psychological well-being (questions 9-14); and social well-being (questions 4-8). As a result of analyzing trust using the mental well-being scale that was adapted by Lim et al. [37], Cronbach's α was 'emotional well-being 0.94′; 'well-being 0.87′; and 'Psychological well-being 0.93,'. The overall score was 0.95.

Research Model and Hypothesis
Structural equations are used to verify the stress, perceived restorativeness, motivation for forest recreation, and the mental well-being of forest users.
The research model is identified by setting the hypothesis as follows ( Figure 2).

Hypothesis 1.
(H1-1). Stress will have a negative effect on perceived restorativeness. The natural environment is emerging as an external strategy for coping with stress. It is said that stress should be immediately alleviated and healed, and the forest is the best place for this [38]. Forest users experience stress reduction [21] and they can escape from daily life by enjoying a restorative environment [13]. Therefore, this study aims to verify the relationship between stress and the perceived restorativeness of forest users. (H1-2). Stress will negatively affect the forest recreation motivation. Outdoor recreation motivates people to escape from excessive work stress [39][40][41]. Leisure activities in the natural environment help people to cope with stress and are crucial for satisfying several needs. Therefore, in leisure activities in an outdoor setting, recovery from stress is the main recognized effect [42,43]. This study aims to verify the relationship between stress and the forest recreation motivation of forest users. (H1-3). Stress will negatively affect mental well-being. Stress is negatively correlated with well-being [44], and old-age stress threatens the happiness of the elderly [45]. Therefore, stress management helps people to cope with negative situations and positively affects their mental well-being [46]. Contact with the natural environment for 15 min promotes a reduction in cortisol concentration, which rises during periods of stress [16] and increases human immunity. A simple slow stroll helps to relieve stress and encourages mental stability [17], so we want to verify the relationship between stress and the mental well-being of forest users.

Hypothesis 2.
Perceived restorativeness will have a positive effect on mental well-being. Exposure to nature positively correlates with emotional and psychological well-being (PWB) [25,47]. When the ability to pay attention is restored by contact with the natural environment it has a positive emotional effect by alleviating the adverse effects of stress. In particular, the effect of emotional enhancement varies depending on the level of exposure to the recovery environment effect [23]. This study aims to verify the relationship between forest users' perceived restorativeness and mental well-being. Hypothesis 3. The forest recreation motivation will have a positive effect on mental well-being. An important motivation for people to engage in outdoor activities in their natural environment is to enjoy be close to nature [28]. The motivation of park visitors to restore their focus and experience comfort while spending time in nature encourages their interaction with nature [48]. Thus, the park environment can contribute to mental health and well-being by allowing visitors to participate in and benefit from recreational activities in all areas within the park [27]. This study aims to verify the relationship between forest recreation motivation and the mental well-being of forest users. Hypothesis 4. Perceived restorativeness will mediate the effect of stress on mental wellbeing. Nature has a stress-buffer effect on psychological well-being [49]. To solve stressful situations, people tend to accept their environment by facilitating positive perceptions about it or devising coping mechanisms to affect their mental well-being [50,51]. Humans improve their mood by alleviating the adverse effects of stress in a natural environment. The natural environment is closely related to people's perception of a therapeutic environment, as well as an improvement in mood [52]. This study aims to examine the relationship between stress and mental well-being, which is a mediating variable in people's perception of a recovery environment. Hypothesis 5. Forest recreation motivation will mediate the effect of stress on mental well-being. Motivation is the driving force to achieve or maintain a goal [53], and selfdetermination motivation leads to positive psychological function improvement. Therefore, people with self-determination motivation encourage their own performance, persistence, and well-being improvement [54]. Motives for engaging in forest recreation include having high expectations for health promotion, and its effects on mental and social health. The natural environment can contribute to mental health and well-being through participation in activities and obtaining benefits [26,27]. For forest visits, natural beauty is the essential motive for visiting [34], and fulfillment of this motive affects mental well-being. This study aims to verify the motivation for forest recreation, a factor in the relationship between stress and well-being.

Analysis Method
To test the hypothesis of this study, SPSS 26.0 and AMOS 18.0 were used, and the analysis method and procedure were as follows. A frequency analysis was performed to determine the general characteristics of forest users, and confirmatory factor analysis, reliability analysis, and correlation analysis were performed to examine the validity and reliability of the measurement variables that were used for the hypothesis testing. Mean, standard deviation, skewness, and kurtosis were calculated to check the general tendency and normality of the variables of this study, including stress, perceived restorativeness, forest recreation motivation, and mental well-being. The variable skewness ranged from −0.29~0.20, and the kurtosis range was from −0.40~0.50, which satisfied skewness < 3 and kurtosis < 10, indicating that all variables were normally distributed [55]. To test the hypothesis, the path coefficient of the research model was verified by applying the structural equation, and bootstrapping was performed to verify the mediating effect. A multi-group analysis was performed to confirm whether there was a difference in the path coefficient of the research model according to gender.

Demographic and Sociological Characteristics of Study Subjects
In this study, a total of 1196 samples were collected from Mt. Gwanak (n = 385); Mt. Bukhan (n = 493); and Mt. Yumyeong (n = 318), from a total of 613 males (51.3%) and 583 females (48.7%). The largest age groups of forest users were 202 men in their 60s (33.0% of men) and 215 women in their 50s (36.9% of women). The frequency of forest visits was high, with 287 men (46.8%) and 237 women (40.7%) visiting once or twice a week. As companions for forest visits, the family was found to be high with 218 males (35.6%) and 263 females (45.1%). Males (28.2%) visited alone more frequently than females (19.0%), and it was found that women (31.6%) visited the forest with friends more often than men (29.2%). The most common length of time spent in the forest was '1-3 h', with 281 males (45.8%) and 274 females (47.0%). When visiting the forest, as the main activity, climbing/walking was the highest with 523 men (60.8%) and 520 women (59.9%). As for the benefits of the forest, clean air was found to be the most important, with 402 males (32.9%) and 401 females (31.3%). Other general characteristics are presented in Table 1.

Descriptive Statistics and Correlation Analysis between Major Variables
As a result of Pearson's correlation analysis between variables, stress had the highest negative correlation with mental well-being (r = −0.46, p < 0.001), and forest recreation motivation had a positive correlation with mental well-being (r = 0.31, p < 0.001). The Pearson correlation coefficient (r) ranged from −1 to 1 and was confirmed to be significant in this study as −0.46 < r < 0.31, p < 0.001. In conclusion, because of Pearson's correlation analysis, stress was negatively correlated with perceived restorativeness, forest recreation motivation, and mental well-being. Perceived restorativeness was found to correlate with forest recreation motivation and mental well-being positively. Forest recreation motivation was positively associated with mental well-being (Table 2).

Differences in Stress Levels according to Demographic Characteristics
A chi-squared test was performed to verify the difference in stress levels according to the demographic characteristics of the subjects of this study. As a result, it was found that 7.9% belonged to the 'healthy group'; 82.5% to the 'potential stress group'; and 9.5% to the 'high-risk group' (Table 3). The difference in stress level according to the age (χ 2 = 20.50, p < 0.05) of forest users and the time spent visiting the forest (χ 2 = 16.82, p < 0.05) was found to be significant. Specifically, there was a significant difference in the high-risk group in their 40s. There was a significant difference in the 'high-risk group' when the length of stay in the forest was less than 30 min or between 30 min to 1 h, and a significant difference was found in the health group when the duration of stay in the forest was less than 5 h. On the other hand, there were differences in stress levels according to gender, frequency of forest visits, and companions who visited the forest, but they were not significant.  Note: 0-8 points were defined as a healthy group; 9-26 points, a potential stress group; and 27-54 points were defined as a high-risk group, * p < 0.05.

Confirmatory Factor Analysis
First, to verify the measurement model through confirmatory factor analysis, item parceling of the stress variables (18 items) consisting of a single factor was performed. Confirmatory factor analysis is an analysis estimate for individual items. As the number of items increases, the number of unknowns increases, and the estimation error increases. In a previous study [56], item grouping was performed, therefore, in this study, 18 items were created, and 3 item bundles of 6 items were created.
Next, the fit index of the confirmatory factor analysis model was confirmed. For the fit of the model, the absolute fit indices χ 2 and RMSEA [57], and the incremental fit indices IFI [58], TLI [59], and CFI [60] were used, and the acceptance criteria for the fitness index are presented, in Table 4. As shown in Table 4, the measurement model fit of this study was IFI = 0.943; TLI = 0.929; CFI = 0.943; and RMSEA = 0.072, all of which met the minimum standards, indicating that the model was statistically suitable. Table 4. Goodness-of-fit index of the measurement model. As a result of the confirmatory factor analysis, the measurement model's standardized regression coefficient (factor load) was 0.85 to 0.87 for stress; 0.40 to 0.84 for recovery environment perception; 0.57 to 0.86 for forest recreation; and 0.80 to 0.90 for mental wellbeing. Since the standardized regression coefficients of all the measured variables were above the minimum standard of 0.4 that was suggested by Kim [61], it can be seen that the measurement model composed of 15 measured variables and four latent variables is valid (Table 5).  Note. B = estimates; SE (standard error); β = standardized estimates; t = B/se; C.R (critical ration); *** p < 0.001.

Hypothesis Verification through Research Model Verification
As a result of the fit index of the model to the research model, as shown in Table 6, IFI = 0.935; TLI = 0.919; CFI = 0.935; and RMSEA = 0.076, indicating that the model was statistically suitable. According to the parameter estimates of the research model (Figure 3), stress is related to the perceived restorativeness (β = −0.46, p < 0.001); forest recreation motivation (β = −0.26, p < 0.001); and mental well-being (β = −0.44, p < 0.001), and each was found to be significantly explained by negative. In other words, it was confirmed that the higher the stress, the lower the perceived restorativeness, forest recreation motivation, and mental well-being. Therefore, it was found that Hypothesis 1-1, which stated that stress would have a negative effect on the perceived restorativeness; Hypothesis 1-2, which indicated that stress would have a negative impact on the forest recreation motivation; and Hypothesis 1-3, which stated that stress would have a negative effect on mental well-being were accepted. The perceived restorativeness was positively significant for mental well-being (β = 0.16, p < 0.001), and it was confirmed that the higher the perceived restorativeness, the higher the mental well-being. In addition, it was found that Hypothesis 2, which stated that perceived restorativeness would have a positive effect on mental well-being was correct. The forest recreation motivation was positively significant for mental well-being (β = 0.22, p < 0.001), and it was confirmed that the higher the forest recreation motivation, the higher the mental well-being. Therefore, Hypothesis 3, which stated that forest recreation motivation positively affects mental well-being was correct (Table 7). motivation on mental well-being (χ 2 = 6.72, p < 0.05). Specifically, it was confirmed that the influence of perceived restorativeness and forest recreation motivation on mental well-being was more significant in the male group than in the female group. On the other hand, it was found that the difference in the effect of stress on the perceived restorativeness, the forest recreation motivation, and the mental well-being according to gender was not significant.  Next, the indirect effect was estimated by performing bootstrapping. Bootstrapping randomly restored and extracted 2000 virtual samples from the original data. The 90% confidence intervals (CI) estimated using bootstrapping were significant at the significance level of 0.1 when 0 was not included. As shown in Table 8, the indirect pathway (B = −0.07, CI: −0.12 to −0.04) in which stress affects mental well-being through the perceived restorativeness, or the indirect pathway (B = −0.06, CI: −0.11 to −0.06) in which stress affects mental well-being through forest recreation motivation was found to be significant. Since stress directly affects mental well-being, perceived restorativeness and forest recreation motivation partially mediate the relationship between stress and mental well-being. Therefore, Hypothesis 4, which sets the relationship between stress and mental well-being as the perceived restorativeness, and Hypothesis 5, which states that the relationship between stress and mental well-being will mediate the relationship between stress and mental wellbeing, are correct. The effect of stress on mental well-being was calculated to investigate the effects of perceived restorativeness and forest recreation motivation in more detail. The direct effect is the magnitude of the direct effect of one latent factor on another potential factor without going through any latent factor; the indirect effect is the value multiplied by the intermediate path, and the total effect is the value of the direct effect plus the indirect effect. Because of the analysis, as for the indirect effect of stress on mental well-being, the perceived restorativeness (−0.07) had a negatively higher indirect effect than the forest recreation motivation (−0.06), and the total effect was high. This indicates that the indirect effect of perceived restorativeness on the effect of stress on mental well-being was increased.

Multi-Group Analysis
A multi-group analysis was conducted to verify whether the path analysis model of forest users' stress on mental well-being was suitable for other samples. In a multi-group analysis, the research model must satisfy the assumption of morphological identity and measurement identity in a sequence. The morphological identity was verified by classifying the data by gender and confirming whether the factor structure of the male (51.3%) and female (48.7%) groups was the same. Morphological identity means that if the model fit index is good in all groups then the shape identity assumption is satisfied. As the fitness index of each group is presented in Table 9, the male group had IFI = 0.953; TLI = 0.941; CFI = 0.952; and RMSEA = 0.068, and the female group had IFI = 0.931; TLI = 0.914; CFI = 0.931; and RMSEA = 0.077, confirming that the fitness index met the minimum standard and indicating that the shape identity assumption was satisfied. Next, the measurement identity was verified. Measurement identity shows whether the same tool was used for measurement. It is confirmed by demonstrating the χ 2 difference between the constrained model that applied the equalization constraint to the factor load of the measured variable between the gender groups, and the free model that did not use the equalization constraint. If there is no significant difference in the fit between the two models, it is judged that the measurement equality assumption is satisfied. As the χ 2 difference test results are presented in Table 10, the difference in the χ 2 fit between the models is 9.67, which is not significant in the difference in degrees of freedom 11, so the measurement equality is satisfied. Note. χ 2 (Chi-Square test of model fit); df (degrees of freedom); IFI (incremental fit index); TLI (Tucker Lewis index); CFI (comparative fit index); RMSEA (root mean square error of approximation); *** p < 0.001. Figure 4 and Table 11 show the results of the χ 2 test by applying the equalization constraint to each path coefficient, in order to verify whether there was a difference in the path coefficients according to gender since the assumptions of morphological identity and measurement identity were satisfied. As a result of the analysis, it was found that there was a significant difference in the effect of the perceived restorativeness on mental well-being (χ 2 = 6.95, p < 0.01) and the impact of forest recreation motivation on mental well-being (χ 2 = 6.72, p < 0.05). Specifically, it was confirmed that the influence of perceived restorativeness and forest recreation motivation on mental well-being was more significant in the male group than in the female group. On the other hand, it was found that the difference in the effect of stress on the perceived restorativeness, the forest recreation motivation, and the mental well-being according to gender was not significant.   Note. Arrows (→) indicate the direction of influence. * p < 0.05, ** p < 0.01, *** p < 0.001.

Discussion
This study revealed that stress was negatively correlated with perceived restorativeness, forest recreation motivation, and mental well-being. In addition, it was confirmed that the perceived restorativeness and the forest recreation motivation partially mediated the relationship between stress and mental well-being, and the direct and indirect effects of stress on mental well-being were confirmed. Moreover, it was confirmed that there was a difference in influence according to gender.

The Stress of Forest Visitors
The stress level of forest users, the subjects of this study, was found to be 8% in the health group, 82.5% in the potential stress group, and 9.5% in the high-risk group. Looking at previous studies, in the study of office workers by Chang [30], the high-risk group for stress was 22%; in the online survey for office workers by Kim [62], the high-risk group for stress was 25.5%; and in a study of Seoul citizens by the Seoul Institute [9], the highly stressed group during the COVID-19 period was found to be 14.9%. In this study, the proportion of the high-risk group for stress was lower than in previous studies. These results came from the fact that 50.2% of forest users in this study visited the forest at least once a week. It can be seen that exposure to nature, that is, the frequency of visits, contributed to personal psychological benefits and well-being [63].
There was also a significant difference in the length of stay in the forest by the stress level. In particular, the 'health group' showed the highest rate at more than 5 h, and the 'high-risk group' showed the highest rate at less than 30 min. Through this, it was found that there was a difference in the amount of time spent in the forest according to the stress level. In addition, exposure to the natural environment directly affects the stress hormone cortisol, and the natural environment reduces stress and enhances immunity [17,52,64,65]. Through many previous studies, it is possible to infer that stress is reduced in the natural environment. Since the subject of this study was a person descending from the forest and  Note. Arrows (→) indicate the direction of influence. * p < 0.05, ** p < 0.01, *** p < 0.001.

Discussion
This study revealed that stress was negatively correlated with perceived restorativeness, forest recreation motivation, and mental well-being. In addition, it was confirmed that the perceived restorativeness and the forest recreation motivation partially mediated the relationship between stress and mental well-being, and the direct and indirect effects of stress on mental well-being were confirmed. Moreover, it was confirmed that there was a difference in influence according to gender.

The Stress of Forest Visitors
The stress level of forest users, the subjects of this study, was found to be 8% in the health group, 82.5% in the potential stress group, and 9.5% in the high-risk group. Looking at previous studies, in the study of office workers by Chang [30], the high-risk group for stress was 22%; in the online survey for office workers by Kim [62], the high-risk group for stress was 25.5%; and in a study of Seoul citizens by the Seoul Institute [9], the highly stressed group during the COVID-19 period was found to be 14.9%. In this study, the proportion of the high-risk group for stress was lower than in previous studies. These results came from the fact that 50.2% of forest users in this study visited the forest at least once a week. It can be seen that exposure to nature, that is, the frequency of visits, contributed to personal psychological benefits and well-being [63].
There was also a significant difference in the length of stay in the forest by the stress level. In particular, the 'health group' showed the highest rate at more than 5 h, and the 'high-risk group' showed the highest rate at less than 30 min. Through this, it was found that there was a difference in the amount of time spent in the forest according to the stress level. In addition, exposure to the natural environment directly affects the stress hormone cortisol, and the natural environment reduces stress and enhances immunity [17,52,64,65]. Through many previous studies, it is possible to infer that stress is reduced in the natural environment. Since the subject of this study was a person descending from the forest and engaged in forest activities, their stress was reduced. In this study, 8.2% of men were in the high-risk group, and 11% of women were in the high-risk group, confirming that women had higher stress. Previous studies have shown that women are more stressed than men regarding mental health [66]. In addition, among forest users during COVID-19, the female group was more stressed than the male group. For example, women reported a higher prevalence of anxiety, depression, and stress due to the spread of COVID-19 [67], and female college students experienced more negative emotions [68].

Hypothesis Test
The success of Hypothesis 1-1 showed that stress negatively correlated with perceived restorativeness. This means that users who visited the forest experienced a reduction in stress. Stress resilience increases in people who visit a forest multiple times. The phytoncide that is released by trees reduces stress hormones and increases immunity by activating NK cells [13,15,69]. Therefore, in this study, it can be concluded that the stress of forest users and the perceived restorativeness are negatively correlated; in other words, stress is reduced and perceived restorativeness is increased.
The success of Hypothesis 1-2 showed that stress was negatively correlated with forest recreation motivation. Previous studies reported that recreational activities such as canoeing, backpacking, mountain climbing, caving, and hunting to relieve stress are highly motivating and bring satisfaction to participants [70]. Leisure acts as a buffer between stressors. Leisure activities that are not emotionally, cognitively, or physically burdensome and are enjoyed alone or with others can remove people from stressors or give them a temporary break. Light leisure activities in stressful situations provide health and wellness benefits. Leisure activities satisfy the motivation that is needed to cope with stressful situations and positively impact a person's life [71,72]. Therefore, in this study, since stress is negatively correlated with forest recreation motivation, it can be concluded that forest users' stress is reduced, and forest recreation motivation is increased.
The success of Hypothesis 1-3 showed that stress was negatively correlated with mental well-being. This was consistent with the research results showing that stress is negatively associated with well-being and psychological well-being. For example, it has been reported that stress in old age threatens the happiness of the elderly and is negatively correlated with well-being [45,73]. It was found that the more exposure to the natural environment, the higher the psychological well-being; while the higher the stress, the lower the psychological well-being [74]. Therefore, since there is a negative correlation between stress and mental well-being in this study, it can be concluded that the stress of forest users is reduced, and mental well-being is increased.
The success of Hypothesis 2 showed that perceived restorativeness had a positive correlation with mental well-being. Previous studies reported that exposure to nature has a positive relationship with emotional and psychological well-being (PWB) and that there are differences in the effect on the level of exposure to the recovery environment. [23,25]. Therefore, since there is a positive correlation between perceived restorativeness and mental well-being in this study, it can be concluded that the perceived restorativeness of forest users and mental well-being is increased.
Hypothesis 3: it was found that forest recreation motivation positively correlated with mental well-being. Visitors' participation in recreation contributed to their mental health and well-being, and motivation came in the form of use. This leads to user satisfaction with the user experience. Motivational satisfaction is related to well-being, and motivation and well-being are positively correlated [27,34,53,71].
Hypothesis 4: Perceived restorativeness mediates the relationship between stress and mental well-being. Stress directly affects mental well-being, and the indirect effect on mental well-being through perceived restorativeness is significant. Since the impact of the perceived restorativeness on mental well-being was smaller than the direct effect, it was found that perceived restorativeness partially mediated the relationship between stress and mental well-being.
Hypothesis 5 showed that stress directly affected mental well-being, and the indirect pathway that affected mental well-being through forest recreation motivation was significant. Moreover, because the effect of forest recreation motivation on mental well-being was smaller than the direct effect, it was found that the forest recreation motivation partially mediated the relationship between stress and mental well-being.
Based on the above evidence, it can be concluded that when forest users reduce their level of stress in the forest environment, their perceived restorativeness, forest recreation motivation, and mental well-being all increase.
This structural equation derived the effect decomposition By deducing that forest users' stress is reduced in the natural environment, if the stress decreases by 1, there is a direct effect that increases by 0.44 on mental well-being. It can be concluded that the indirect impact of the perceived restorativeness increases by 0.07 and increases by 0.51 in total. Additionally, suppose the stress of forest users decreases by 1. In that case, there is a direct effect that increases by 0.44 on mental well-being, and it can be concluded that the indirect effect of the motive for forest recreation increases by 0.06 and increases by 0.50 in total. As a result, a research model can be derived that stress reduction in forest users increases their perceived restorativeness, forest recreation motivation, and mental well-being.

Multi-Group Analysis
We conducted a multi-group analysis to verify whether there is a difference in path coefficients according to gender through a research model suggesting that forest users' stress reduction increases their perceived restorativeness, forest recreation motivation, and mental well-being. As a result of this study, there were significant differences in perceived restorativeness on mental well-being and the effect of forest recreation motivation on mental well-being according to gender. Specifically, it was confirmed that the influence of forest recreation motivation and perceived restorativeness on mental well-being was more significant in the male group than in the female group. In addition, men showed a higher direct effect of stress on perceived restorativeness, and women showed a higher direct impact of stress on mental well-being. In the study of subjects who experienced the forest, men had a heightened perception of the forest recovery environment [75]. There was a positive relationship between men who were engaged in physical activity and their psychological well-being [76]. A lifestyle that supports well-being also differs according to gender [77], based on these previous studies; thus, the following conclusions can be drawn. The effect of stress on mental well-being did not differ by gender. Still, the effect of perceived restorativeness and forest recreation motivation on mental well-being was higher in men than in women.

Limitations
First, this study is meaningful because it was a face-to-face study conducted on forest users when studies showed that public stress was high due to anxiety about COVID-19 transmission and the absence of a vaccine. However, there were limitations, as it was carried out while practicing social distancing in a limited environment. Second, there was a limit to the psychological research that was conducted with the survey during the COVID-19 period. Therefore, in future studies, pre-and post-test verification studies and physiological effects studies should be conducted. Third, differences were verified through a multi-group analysis of gender in the research model for stress, perceived restorativeness, forest recreation motivation, and mental well-being. However, it is expected that new studies will be conducted through analysis between various groups based on the research model of future research. In addition, since stress reduction in forest users can increase their perceived restorativeness, forest recreation motivation, and mental wellbeing, various studies on the forest environment and forest healing for stress reduction should be conducted. In conclusion, this study has significance in developing a research model suggesting that forest users' stress reduction leads to improvement in their perceived restorativeness, forest recreation motivation, and mental well-being. It is expected to serve as primary data for finding ways to promote forest healing in the future.

Implications
People avoid contact with others and travel short distances with their families due to COVID-19. It has become important for people to manage stress and lead a healthy life. This study found that perceived restorativeness and forest recreation motivation were mediators in reducing the stress of forest users and increasing their mental well-being. Moreover, there was a difference in the influence of men's perceived restorativeness and forest recreation motivation on mental well-being, compared to women. To increase the perceived restorativeness and the forest recreation motivation among forest users, it is necessary to manage forest recreation in a way that is suitable for all users and various types of use. In addition, to increase perceived restorativeness, a natural environment that considers the factors of Being Away, Fascination, Coherence, and Compatibility is required.

Conclusions
This study is significant because it confirmed that the forest is a recovery environment that reduces stress and increases the perceived restorativeness, forest recreation motivation, and mental well-being of those citizens who visited the forest during the COVID-19 pandemic. Therefore, forest recreation management that can increase forest perceived restorativeness and forest recreation motivation is an important consideration for health promotion in the natural environment. Institutional Review Board Statement: It has been confirmed that this study is subject to deliberation exemption by Article 16 of the Bioethics Act and Article 13 of the Enforcement Regulations. This study involves research on human subjects, but does not include subjects in a vulnerable environment, and belongs to research that uses information disclosed to the general public, or research that does not collect or record personally identifiable information. In addition, although this study faced the research subjects directly for the questionnaire, the IRB did not review it because the research subjects were anonymous and sensitive information was not collected or recorded, according to Article 23 of the Personal Information Protection Act (https://www.irb.or.kr/menu02/commonDeliberation.aspx (accessed on 30 April 2020)).