This study suggests that 2 hours forest walking significantly lengthened mean actual sleep time and mean immobile time estimated by actigraphy and that sleep depth and sleep quality were improved, as evaluated by SMHSQ. In addition, some participants had a nap(s) after forest walking. Forest walking may contribute to improvement of subsequent sleep for individuals with sleep complaints. Exercise and emotional improvement initiated by walking in forested areas may bring both increased sleeping hours and improved subjective sleep quality.
Regarding objective sleep parameters, a previous review suggested that exercise increased total sleep time (TST) but did not significantly affect sleep onset latency (SOL) ; our results come close to supporting this. Recent studies reported that sleep emerges locally and is regulated in a use-dependent (homeostatic) manner [40, 41]. The study showed that arm immobilization locally decreased slow wave activity in subsequent sleep; slow wave activity is thought to reflect sleep need . Exercise during the prior wake period therefore might induce sleep.
Regarding emotional effects, this study suggests that 2 hours' forest walking significantly improved anxiety as measured by STAI-S. Furthermore, self-rated sleep depth for individuals whose STAI-S scores decreased by more than the average was much improved versus in individuals who had STAI-S scores decreased by less than the average--even when adjusted by forest walking session, which was a relevant factor for sleep improvement. From these results, it appears that improvement of self-rated sleep depth may depend on not only exercise but also improvement of psychological factors. Since forest walking, which does not require specific techniques, is a widely available activity, walking in forested areas may be a practical method to improve sleep that is easily applicable in daily life.
Our study suggests that afternoon forest walks had a greater effect on actual sleep time and immobile minutes than those taken in the forenoon. The time of day when exercise is undertaken is one factor related to the subsequent sleep response [23, 39]. Youngstedt et al.  reported that SOL and wakefulness after sleep onset (WASO) were influenced by the time of day when exercise was completed, whereas TST was not influenced by the time of day. However, this study revealed that forest walking in the afternoon much improved actual sleep time and immobile minutes but not SOL compared with forest walking in the forenoon.
According to a two-process model [42, 43] sleep and waking are regulated by circadian rhythms (Process C) and homeostasis (Process S). To examine why the afternoon session causes improvements in actual sleep time and immobile minutes, the core body temperature should be measured to determine the change of circadian phase (Process C) . Since the measurement was not available in this study, we cannot conclude why the afternoon session caused improvements in actual sleep time and immobile minutes. However, one possible reason that the afternoon session brought improvements could be by homeostatic mechanism recovery after exercise (Process S). Exercise may amplify core body temperature. A steep decline of core body temperature before nocturnal sleep was reported to induce sleep . The afternoon session, with a shorter interval between the end of walking and the onset of nocturnal sleep compared with the forenoon session, might be profitable for sleep improvements. On the other hand, because both forenoon and afternoon sessions are conducted in the daytime, they might not affect the circadian phase (Process C), as shown by the sleep start times of the forenoon and afternoon session participants which did not change on the nights before and after forest walking. Furthermore, the possibility still remains that the higher percentage of participants taking a nap(s) after forest walking in a forenoon session compared with an afternoon session was related to an apparent extension of nocturnal sleep duration in the participants in the afternoon sessions.
Intensity of exercise is also a factor related to subsequent sleep . The intensity of forest walking in this study was considered light exercise on the Borg scale because the study was conducted in a forested area with few steep mountain paths. Since this study suggests that forest walking improved some sleep conditions, the intensity of exercise can be seen as appropriate.
This study has some limitations. First, there was no control group for walking in non-forested areas adjusted by exercise strength and light intensity. Therefore we cannot claim with certainty whether immediate improvements in the characteristics of sleep were brought on by walking only or by walking specifically carried out in forested areas. However, a previous study showed that STAI-S score was lower on a forest-walking day compared with another day with exercise in non-forested areas . Furthermore, the present study suggests that improvement of anxiety measured by STAI-S was associated with self-rated sleep depth, even after adjusting for the timing of forest walking (forenoon vs. afternoon). From these two results, walking in forested areas, where emotional effects would be expected, might produce much improvement of sleep compared with walking in non-forested areas. Second, we evaluated only the immediately discernable effects of sleep.
The conclusions of this study are that 2 hours forest walking lengthened actual sleeping hours and immobile minutes and that it improved self-rated depth of sleep and sleep quality for individuals who had sleep complaints. The self-rated depth of sleep depended on emotional improvements. Forest walking in the afternoon prolonged actual sleeping hours and immobile minutes. Further studies using a randomized controlled trial design need to be carried out to evaluate whether a series of forest walks improves slight insomnia and slight sleep complaints. Such a study should reveal what factors in forest walking are responsible for improving sleep complaints.