Substance and claimed effects

Time in nature — voluntary exposure to forested, vegetated, or otherwise green outdoor environments, typically in single visits of 20–120 minutes accumulated to a target of roughly two hours weekly. Claimed effects span stress physiology (lower salivary cortisol, lower sympathetic tone, lower blood pressure), affect and mood (reduced rumination, lower depression and anxiety symptoms, improved subjective wellbeing), sleep quality (indirect, via stress and circadian rhythm), attention restoration (improved performance on directed-attention tasks), and immune function (transient upregulation of natural killer (NK) cell activity). Long-term observational data also links residential greenness with reduced all-cause mortality, presumed mediated by physical activity, air-quality improvements, stress buffering, and social cohesion Hartig et al. 2014.

Evidence by addressing question

Mechanism

Four overlapping pathways carry the effects, none of which require shinrin-yoku to be "magic" Hartig et al. 2014:

• Stress Reduction Theory (SRT, Ulrich). Natural scenes containing water, depth cues, mid-range complexity, and non-threatening vegetation rapidly down-regulate sympathetic arousal — measurable as falls in cortisol, blood pressure, heart rate, and electrodermal activity within ~15 minutes of exposure. The original behavioural anchor is Ulrich 1984: cholecystectomy patients with a window view of trees had shorter postoperative stays (mean 7.96 vs 8.70 days), fewer "negative" nurse notes, and lower analgesic requirements than matched patients facing a brick wall.
• Attention Restoration Theory (ART, Kaplan). Directed attention is a finite resource depleted by urban/work environments full of "hard fascination" (traffic, deadlines). Natural scenes elicit "soft fascination" — bottom-up stimulus capture that doesn't demand top-down inhibition — letting the directed-attention system replenish. Berman, Jonides & Kaplan 2008 ran two within-subject experiments: a 50-minute Ann Arbor arboretum walk vs an urban-downtown walk, then a photographic-exposure variant. Backwards digit-span (a working-memory probe) improved after nature exposure but not after urban; effect size d ≈ 0.5.
• Phytoncides (plant volatile organic compounds). Coniferous forests release α-pinene, β-pinene, limonene, and other monoterpenes that humans inhale in measurable concentrations. Li et al. 2009 exposed 12 healthy adult men to hinoki-cypress phytoncides via a humidifier in a hotel room for three nights; NK-cell activity rose significantly, percentages of granulysin-, perforin-, and granzyme A/B-expressing lymphocytes rose, urinary adrenaline and noradrenaline fell. Li 2010 showed a 3-day forest trip increased NK activity ~50% with effect persisting ≥7 days; a 2-day trip persisted ≥30 days. Mechanism: stress-hormone fall + direct terpene action on NK precursors.
• Indirect mediators. Greenspace exposure increases walking minutes (physical activity), lowers PM_2.5 exposure relative to roadside urban air, supports social cohesion, and may improve sleep via dawn-light exposure. The Nurses' Health Study mediation analysis attributed roughly equal contributions to physical activity, air pollution, social engagement, and mental health James et al. 2016.

Evidence

Dose-response in population data. White et al. 2019 analysed Monitor of Engagement with the Natural Environment survey data from 19,806 adults in England. Compared to no nature contact, the odds of reporting good health peaked at ≥120 min/week (OR ≈ 1.59, 95% CI 1.39–1.81) and the odds of high wellbeing peaked at 200–300 min/week (OR ≈ 1.76). Below 60 min/week the effect was indistinguishable from zero — a genuine threshold. Identical pattern across men, women, older/younger, occupational groups, urban/rural, those with long-term illness, and those without. Distribution of the 120 minutes (one long visit vs several short) didn't matter.

Acute physiological effects. Park et al. 2010 ran cross-over field experiments in 24 Japanese forests (n=280, mean age 21.7). On test days subjects walked for 16 min and viewed the landscape for 14 min in a forest, then on a matched day did the same in a city. Forest sessions produced lower salivary cortisol (mean −12.4%), lower pulse (−7.0%), lower systolic BP (−1.4%), higher parasympathetic and lower sympathetic activity (HRV-derived). City sessions produced no such shifts.

Blood pressure meta-analysis. Ideno et al. 2017 pooled 20 trials (n=732). Systolic BP was −1.4 mmHg lower in forest vs non-forest conditions (95% CI not significant in primary analysis); for subjects with elevated baseline, SBP fell by ≈ −5.4 mmHg. Diastolic BP fell by −1.75 mmHg (95% CI −2.38 to −1.13, p<0.001, I² = 24%). Effect is modest but consistent.

Greenspace umbrella meta-analysis. Twohig-Bennett & Jones 2018 pooled 143 studies and found statistically significant decreases in diastolic BP, salivary cortisol, heart rate, type-2 diabetes incidence, all-cause mortality, and cardiovascular mortality with greenspace exposure. Quality grade was mixed and heterogeneity high.

Mortality. Rojas-Rueda et al. 2019 meta-analysed nine prospective cohorts covering ~8 million participants. Hazard ratio for all-cause mortality per 0.1 increase in NDVI (a satellite-derived greenness index, within 500 m of the residence) was 0.96 (95% CI 0.94–0.97) — a 4% reduction per modest greenness step. James et al. 2016 followed 108,630 women in the U.S. Nurses' Health Study over 2000–2008; the highest greenness quintile saw a 12% lower all-cause non-accidental mortality rate vs the lowest (HR 0.88, 95% CI 0.82–0.94).

Mental health. Bratman et al. 2015 randomised 38 healthy adults to a 90-minute walk in either a quiet Stanford-area natural setting or a traffic-heavy arterial road. The nature group showed reduced self-reported rumination (RRS subscale) and reduced subgenual prefrontal cortex perfusion (ASL fMRI) — the same brain region hyperactive in major depression. Yeon et al. 2021 meta-analysed 20 forest-therapy trials and reported Hedges's g ≈ −1.13 for depression and ≈ −1.72 for anxiety; effect sizes are likely inflated by clinical-population enrichment and Korean-cohort dominance, but direction is consistent. Astell-Burt & Feng 2019 followed 46,786 Australian adults >45 yr for ~6 years: ≥30% tree-canopy neighbourhoods showed lower psychological distress (OR 0.69, 95% CI 0.61–0.78) and better self-rated health than those with <10%; the effect did not appear with grass-only greenspace.

Attention restoration in clinical populations. Faber Taylor & Kuo 2009 ran a within-subjects RCT in 17 children (7–12 yr) clinically diagnosed with ADHD; each child completed three 20-minute guided walks (city park, downtown, residential) one week apart. Backwards-digit-span improved significantly more after the park walk than after the downtown walk (Cohen's d = 0.52) or the residential walk (d = 0.77). The authors note effect sizes are comparable to those reported for a standard dose of methylphenidate — a remarkable benchmark.

Protocol

Target dose for population-scale benefit is ≥120 minutes weekly, with the optimum landing in the 200–300 min/week band White et al. 2019. The 120 minutes can be one long visit or several short ones; distribution does not change the threshold. Acute physiological effects (cortisol, BP, parasympathetic shift) appear within ~15 minutes of forest exposure Park et al. 2010, so short bursts compound. NK-cell upregulation after a 2- to 3-day forest visit persists 7–30 days Li 2010, suggesting periodic immersive trips may be a useful complement to weekly maintenance.

The type of nature appears to matter, but less than expected. Forests trend strongest in physiological studies (phytoncides, sound profile); urban tree canopy ≥30% is the threshold for chronic mental-health effects Astell-Burt & Feng 2019; coastal/blue spaces and parkland likely substitute reasonably well for stress reduction but lack the phytoncide pathway. Solitary vs social, walking vs sitting, screens-on vs screens-off — none of these have been rigorously dose-graded; the field defaults to "screens away, attention on the surroundings."

Contraindications

Effectively none for the substance itself. Population-specific cautions are environmental, not behavioural: tick-borne disease in endemic regions (Lyme, encephalitis), poison-ivy/oak exposure, allergy season (pollen-load reactivity), heat-stroke risk in extreme conditions, snake/wildlife considerations in specific habitats, UV exposure without sun protection at altitude. The MA literature does not flag medication interactions.

Misconceptions

Common misframings:

• "Forest bathing is a meditative practice." Shinrin-yoku in the original Forestry Agency framing is exposure-based, not contemplative — walking and looking, not sitting in lotus. Meditation effects on stress add on top, but the nature exposure carries effects independent of mindfulness training Park et al. 2010.
• "You need a forest." Urban parks, residential tree canopy, and even window views show effects Ulrich 1984, Astell-Burt & Feng 2019. Forests are stronger but not gatekeepers.
• "Phytoncide effects = forest is anti-cancer." The NK-cell findings are biomarker-level, not endpoint-level. The leap from "NK activity rises in 12 men in a hotel" to "fewer cancers in the population" is unsupported by direct trial data; longevity associations with greenness exist but have multiple mediators.
• "Any time outdoors counts." Roadside exercise in heavy traffic produces opposite effects on stress markers; the surroundings matter, not just being out of the house Bratman et al. 2015.

Failure modes

Most failure modes are exposure dilution: short visits below the threshold, phone-on visits where the directed-attention drain from notifications continues, low-greenness routes (mostly impervious surfaces with token planters), or "nature" framed as a workout track where cardio-vascular load swamps stress reduction. Adverse-weather or insect-density conditions can flip the affect outcome negative — see the operative-conditions guard Hartig et al. 2014.

History

Shinrin-yoku was coined in 1982 by the Japanese Forestry Agency (Akiyama Tomohide) as a public-health framing for forest recreation; the term was the substance's first formalisation, and the Japanese government has since designated >60 official forest-therapy bases. The underlying intuition — that natural settings restore — predates the term by millennia (monastic gardens, English landscape parks, Olmsted's urban park theory, Thoreau, Muir). The modern scientific lineage runs through Ulrich's 1984 hospital-window paper, Kaplan & Kaplan's 1989 The Experience of Nature (attention restoration theory), and Qing Li's Tokyo-based phytoncide-and-NK-cell program from the mid-2000s onward.

Stakes

Population-attributable: White et al. 2019's threshold effect implies the modal adult (sub-60 min/week) is leaving wellbeing on the table that is statistically indistinguishable from zero contact. Rojas-Rueda et al. 2019's mortality hazard ratio (~4%/0.1 NDVI step) compounds across decades of low-greenness residence. James et al. 2016: a woman in the lowest-greenness quintile has a ~12% higher all-cause non-accidental mortality rate than one in the top quintile, mediated through physical activity, air pollution, mood, and social ties.

Payoff

Acute (within minutes): salivary cortisol −12%, pulse −7%, BP −1.4 mmHg systolic, parasympathetic shift Park et al. 2010. Within 90 minutes: measurable reduction in rumination and in subgenual PFC activity Bratman et al. 2015. After 20 minutes in a park: directed-attention performance comparable to a methylphenidate dose in children with ADHD Faber Taylor & Kuo 2009. After 2–3 days in a forest: NK-cell activity ~50% higher, persisting 7–30 days Li 2010. At population scale and over years: lower mortality, lower incident diabetes, lower depression incidence Twohig-Bennett & Jones 2018.

Out of scope

Adjacent entries this dossier does not cover but the article should point at: morning sunlight exposure (circadian-alignment substance, distinct mechanism), aerobic exercise (overlapping but separable), gardening as horticultural therapy (an active variant with its own literature), forest schools and pediatric nature-deficit syndrome (developmental scope), wilderness expeditions and adventure therapy (clinical-population scope), and indoor plant exposure (negligible phytoncide concentrations; mostly aesthetic).

Credibility range

Optimist case

The convergence is genuinely impressive. Acute physiological effects (cortisol, BP, HRV) replicate across Japanese, European, and North American trials with different protocols, different forest types, and different populations. The dose-response landed by White et al. is a real threshold, not a linear-with-noise — exactly what a recommend-able intervention looks like. Mechanism is plural and parsimonious: nothing exotic, just stress physiology + attention biology + immune cell trafficking + air-quality + walking. Effect sizes in attention restoration (d ≈ 0.5–0.8) and in ADHD-children attention (matching a stimulant) are not subtle. Mortality cohorts are large (millions) and methodologically robust. The intervention is free, available to most of the population, has trivially low side-effect profile, and has been embedded in Japanese public-health policy for four decades. If a pharma company filed an NDA with this evidence package, the indication would clear.

Skeptic case

Effect-size heterogeneity is large; the BP meta-analysis effect (−1.4 mmHg SBP) is modest and not clinically dramatic in normotensives. Most acute-physiology trials are small (n < 30), short, and from a handful of Japanese research groups with topical incentives — publication bias is plausible. Yeon et al. 2021's near-implausibly large effect sizes (g > 1) on depression and anxiety suggest Korean-cohort-and-clinical-population artefact more than generalisable signal. The "phytoncide → NK → cancer" chain is biomarker-only; no trial has shown nature exposure reduces incident cancer at a population endpoint. Greenspace-mortality associations are heavily confounded — wealthier areas have more trees, lower noise, better air, healthier residents — and mediation analyses don't fully neutralise this. NDVI-based exposure is a proxy for everything from "park within walking distance" to "lawn fertiliser run-off," and exposure misclassification is endemic. The 120-min threshold is from cross-sectional self-report, not RCT; reverse causation (healthier people visit nature more) is not ruled out. The field is also rife with motivated reasoning — environmental researchers wanting nature to win.

Author's call

Land in the middle-optimist range. The acute physiological effects are real and replicable; the cognitive (attention restoration) effects are real and useful; the depression/rumination effects are real but more modest than the headline forest-therapy meta-analyses suggest; the mortality association is real but probably not 12% large after full confounding adjustment. The action is low-burden, near-zero-cost, side-effect-free, and complements the catalogue's other low-friction levers (sunlight exposure, walking). Score evidence at 4 (one strong, consistent body of work across mechanism and outcome, multiple meta-analyses, no large negative trials, but modest individual trial sizes and heterogeneity bar 5); controversy at 1 (effect-size argument exists but no one credibly argues nature exposure is bad for stress markers).

Stakeholder and incentive map

• Pro: Japanese Forestry Agency (origin), Korean Forest Service (largest active research program), urban-planning and parks advocacy (Trust for Public Land, City Parks Alliance), environmental psychology academia (Hartig, Bratman, Berman, Kaplan lineage), the forest-therapy guide industry (ANFT and similar certification bodies). Strongest non-aligned voice is public-health journals' editorial line that promotes nature-prescriptions.
• Counter: Pharma (substitute for stimulants, antidepressants, antihypertensives — though no organised opposition). Skeptic-camp researchers concerned about confounding (Lancet Planetary Health editorial commentaries 2021). Conservative health-economics literature questioning ROI on green-infrastructure spending vs other public-health priorities.
• Commercial niche: Eco-tourism, retreat operators, "shinrin-yoku" certification — small relative to the underlying free-access public-goods nature of the intervention. The substance is largely uncapturable by markets, which is part of why the literature stays honest.

Population variability

The 120-min threshold appeared equally in men and women, across ages, across occupational groups, in urban and rural residents, and in those with and without long-term illness White et al. 2019 — a population-wide effect with no major modifier. Subgroups where effect may be larger:

• Baseline-stressed. Hypertensives respond more than normotensives in BP studies Ideno et al. 2017; ruminators show larger sgPFC and rumination shifts Bratman et al. 2015; ADHD children show clinically meaningful attention restoration Faber Taylor & Kuo 2009. The intervention is dose-equivalent across populations but the felt benefit scales with baseline impairment.
• Urban high-density residents. Marginal benefit per minute is higher because baseline exposure is lower; Astell-Burt's tree-canopy threshold suggests the deficit is real.
• Older adults. Tree canopy associated with reduced incident dementia in 11-year Australian follow-up (Astell-Burt et al. 2020, ancillary work); mortality protection mediated heavily by maintained physical activity.
• Mental-health populations. Yeon et al.'s implausible effect sizes still aim in the right direction; the clinical population likely benefits more than the healthy population on the affect axis.

Subgroups where the substance plausibly under-delivers: those with severe seasonal allergies (pollen-load may invert affect), tick-endemic-zone residents avoiding undergrowth, mobility-impaired adults for whom the "walking" pathway is closed but the "viewing" pathway remains via window/garden access.

Knowledge gaps

• No long-duration RCT with clinical endpoints (incident depression, MI, mortality). All long-term data is observational.
• Dose-response is from self-report cross-sectional data; an RCT randomising urban adults to ≥120 vs <60 min/week for 2+ years would be definitive but is not in the pipeline.
• Type-of-nature comparison studies are underpowered. Forest vs urban park vs blue space (coast/lake) vs grass-only — these have not been head-to-head dose-graded.
• Phytoncide → NK → clinical endpoint chain stops at NK biomarker. No cancer-incidence trial.
• Mechanism of attention restoration vs simple "screen and stressor removal" — current designs don't isolate the natural-scene effect from the away-from-work effect cleanly.
• Sleep effects are inferred (lower cortisol, less rumination, more physical activity, possibly more daylight) rather than directly measured by polysomnography after defined-dose nature exposure. A clear gap.
• Children: most modern attention studies are in healthy adult cohorts; pediatric data is mostly the ADHD population and observational schoolyard studies.

Coverage vs brief. The brief named stress markers, mood, sleep, blood pressure, attention restoration, and immune function. The article covers all six: stress markers and blood pressure under evidence, mood across evidence + payoff, attention restoration in its own ad-hoc section, immune function via the phytoncide / NK-cell strand inside mechanism and protocol. Sleep is the thinnest leg — discussed in payoff and scored a conservative 2 in meta — because direct PSG-graded data is sparse; flagged in the dossier's knowledge gaps.

Category choice. Filed under Light & Environment. The substance is environmental exposure rather than a movement, mental, or screening behaviour; the closest semantic neighbour in the existing taxonomy is the light-and-environment category that already houses sunlight exposure. If a future Outdoors or Environment category is split out, this entry moves with it.

Scoring calls worth flagging.

• beauty_direct & beauty_cumulative both 0. Tempting to score beauty_cumulative as a 1 via chronic cortisol reduction → skin → years, but the literature does not trace nature exposure to skin/hair endpoints. Spec encourages honest zeros; the article body doesn't cover beauty either, so the score and the article agree.
• longevity = 3, not 4. The mortality data is real (Rojas-Rueda meta, James et al. Nurses' Health) but heavily mediated and observational; greenness is a confound-rich exposure proxy. A 3 reflects the consistent direction without overclaiming the magnitude after confounding adjustment.
• mood = 4, not 5. Bratman PNAS rumination + sgPFC quieting is one of the strongest mechanistic findings, and the Yeon meta-analysis effect sizes are big — but those big effect sizes are likely inflated by Korean clinical-population enrichment, so calling this on a level with effective psychiatric intervention overshoots.
• focus = 3. Berman d ≈ 0.5 is meaningful; Faber Taylor & Kuo's methylphenidate-comparable result in ADHD children is the headline. Did not score 4 because outside ADHD, the effect on day-to-day deep work is moderate and the protocol burden (going outside) is what limits practical impact for office workers.
• evidence = 4. The body of work is substantial — four decades, multiple meta-analyses, mechanism replication — but the individual trials are small (n < 30 is common in the physiological work), the long-term endpoint data is observational, and there is no clinical-endpoint RCT. A 5 would require at least one of those.
• sleep = 2. The thinnest non-zero. There is a real chain (cortisol down, daylight up, walking up) but the dossier's knowledge gap on PSG-graded dose-response means this score is held conservatively.

Separate-entry candidates surfaced during writing.

• Gardening as horticultural therapy — an active variant with its own literature (RCTs in elderly populations, post-traumatic recovery), distinct enough in cadence and effort burden to warrant its own scoring.
• Forest school / pediatric nature exposure — the developmental literature is large enough to be its own thing; this entry only nods at it via the ADHD result.
• Wilderness expeditions / outdoor adventure therapy — clinical-population intervention, different dose, different population.
• Indoor plant exposure and biophilic design — almost certainly weaker but commonly asked about; would be a useful "honest zero-ish" entry.

Future-link candidates. When they exist in the catalogue, link to Morning Sunlight, Walking, Aerobic Exercise, Sleep Hygiene, Time Off Screens, Gardening. Currently flagged in the out-of-scope closing section but unlinked.

Hard call: cadence. Set to weekly because the threshold finding (White et al. 2019) is weekly-dosed and that's the actionable framing. Could equally have been daily (since most readers will hit 120 min via roughly 20 min most days), but weekly preserves the "cumulative dose" framing that the literature actually measured.

What was excluded. Wilderness-survival framings, neoprimitivist or grounding/earthing claims (no good evidence base for the latter), and the "biophilia hypothesis" as an evolutionary framing for why the substance works — interesting but speculative and not load-bearing for the action.