RESEARCH DOSSIER
Green and Blue Space Exposure and Population Health
Abstract. Green and blue space exposure links the physical organization of vegetated and aquatic environments with the distribution of opportunities for visual, recreational, social, and multisensory contact. This critical narrative monograph investigates how exposure, access, quality, experienced contact, and social circumstances relate to psychological and population-health outcomes. It compares historical restoration theories, short-term environmental experiments, cohort and register-linked longitudinal studies, community randomized greening trials, systematic reviews, and recent umbrella reviews, distinguishing proximal affective or cognitive responses from psychiatric, cardiometabolic, and mortality endpoints. The evidence supports intervention-specific psychological effects and multiple favorable population-level associations while identifying important residual confounding, exposure misclassification, heterogeneous effects, and uneven evidence for blue-space outcomes. The resulting synthesis frames natural environments as multidimensional social and ecological opportunities whose health consequences require transparent causal contrasts, reliable measurement, equitable access, and long-term evaluation. This work is a critical narrative synthesis of published studies and reports no original experiments, new systematic-review protocol, or pooled effect estimates.
Keywords: green space; blue space; population health; environmental psychology; nature exposure; green infrastructure; coastal and inland water; spatial epidemiology; mental health; restoration; attention restoration theory; stress recovery; environmental justice; urban greening; causal inference; longitudinal cohort; randomized environmental intervention; physical activity; mortality; Artificial Era
Introduction: Environmental Exposure as a Population Health Research Object
Green and blue space exposure denotes the relationship between populations and accessible vegetated or aquatic environments across residential, occupational, educational, recreational, and transit settings. Green environments include parks, street trees, woodland, grassland, community gardens, and vegetation integrated into built landscapes. Blue environments include rivers, lakes, canals, wetlands, coasts, and other outdoor water settings. Their health significance arises through an exposure process: a physical feature becomes an opportunity for visual, multisensory, behavioral, social, or environmental contact under particular conditions of access, safety, and time. Population health concerns the distribution of these contacts and their associations with mental, physical, functional, and social outcomes across groups and life stages. This subject therefore joins environmental psychology with epidemiology, urban planning, health geography, behavioral medicine, and public policy. Foundational syntheses by Hartig et al (2014) and Markevych et al (2017) establish the multi-pathway nature of the question.
The central research question is how characteristics of nearby or visited nature relate to health, which aspects of that relationship may be causal, and whose opportunities and outcomes are represented by available evidence. A rigorous answer distinguishes environmental availability from actual contact, subjective experience from remotely sensed vegetation, and psychological improvement from downstream disease or mortality. It also asks whether observed benefits arise from reduced environmental hazards, changes in behavior, restorative experiences, social interaction, or common causes such as neighborhood advantage. The framework of White et al (2020) is especially useful because aquatic environments offer recreational opportunities and sensory qualities alongside distinctive hazards and accessibility constraints.
This is a critical narrative synthesis of identifiable original studies, longitudinal investigations, experiments, systematic reviews, meta-analyses, and methodological papers. It neither recruits participants nor computes a new pooled effect. Claims about observational association, intervention effects, plausible mediation, and policy extrapolation are identified by their respective evidentiary bases. The object of the monograph is population-level green and blue space exposure and its health distribution; the microprocess of attention restoration, legal questions of environmental justice, and generalized climate-threat communication are treated as neighboring fields when essential to interpretation.
Conceptual Definitions and the Taxonomy of Natural Environments
A natural environment is not a uniform treatment. A wooded trail with uneven paths, a shaded public square, a fenced ornamental lawn, a polluted canal, and an accessible coastal promenade each produce different opportunities and constraints. Exposure descriptors should specify ecological content, spatial scale, physical condition, usable access, experienced contact, temporal duration, and the relevant outcome. A residential vegetation index measures chlorophyll-related surface reflectance and may capture inaccessible private landscaping; a walking diary measures reported behavior and may miss views through windows; a street-level image measures visible greenery from sampled viewpoints and may omit enclosed parks. These representations concern related constructs, yet they should be analyzed as distinct variables.
The classification of green spaces may use land cover, land use, vegetation type, tree canopy, biodiversity, management, ownership, perceived naturalness, or accessibility. Blue-space classification adds water-body type, coast versus inland location, water visibility, bank accessibility, water quality, flood vulnerability, and opportunities for contact with water. Hybrid locations—wooded riverbanks, urban wetlands, coastal parks—introduce collinearity between green and blue exposure. Gascon et al (2017) and White et al (2020) explain why comparing environments solely by a binary green/blue label can suppress important physical and experiential differences. The recent overview of quality measurement by J. Yang et al (2026) reports many noninterchangeable indicators and underscores the need to specify what the quality score actually represents.
Population health outcomes also require semantic precision. Self-rated general health, life satisfaction, depression symptoms, diagnosed psychiatric disorders, blood pressure, metabolic disease, birth outcomes, and mortality sit at different points in possible causal chains. An increase in momentary positive affect after a walk is a proximal psychological change. A difference in mortality among neighborhoods is a distal population association shaped by multiple exposures and social histories. These endpoints cannot be combined into a single statement of benefit without explaining their time scales, measurement properties, and evidentiary status.
Historical Formation of Nature and Health Research
The modern research tradition combines environmental psychology, environmental epidemiology, public health, and ecological planning. Early clinical-environment evidence included Ulrich (1984), a comparison of matched postoperative patients whose hospital windows overlooked either vegetation or a brick wall. The result helped stimulate research on views of nature, but its small retrospective design limits generalization and cannot resolve the causal contribution of room assignment or correlated care conditions. Later environmental psychology formalized theories of recovery from directed-attention demands and stress responses, including Kaplan, 199590001-2) and field-based psychophysiological comparisons by Hartig et al., 200300109-3).
In the 2000s and 2010s, researchers increasingly combined health surveys, geographic information systems, satellite imagery, residential histories, and administrative records. This changed the unit of analysis from a carefully selected restorative scene to neighborhood exposures repeated across large populations. The longitudinal mover analysis by Alcock et al (2014), the Barcelona schoolchildren investigation by Dadvand et al (2015), and the Danish register-linked cohort of Engemann et al (2019) illustrate distinct attempts to examine environmental conditions across time. Each made different assumptions about residential choice, environmental stability, and outcome measurement.
The field simultaneously expanded beyond vegetation. Coastal health associations examined by Wheeler et al (2012), long-term residential green and blue exposure reviews by Gascon et al (2015), and aquatic-environment synthesis by White et al (2020) introduced settings in which water visibility, access, recreation, weather, and perceived risk may differ from forest or park exposure. Large cohort reviews such as Rojas-Rueda et al., 201930215-3) and later umbrella reviews such as Wang et al (2025) transformed scattered findings into population-scale summaries. The resulting research history is an accumulation of methods and competing causal explanations, rather than a linear progression from a single experiment to settled policy.
Theories of Restoration and Psychological Functioning
Attention restoration theory proposes that environments offering compatible activity, a sense of being away, sufficient experiential extent, and relatively effortless fascination can support recovery from demanding use of directed attention. Kaplan, 199590001-2) formulated this account as a theory of conditions that may permit attentional replenishment, not as a claim that every natural scene automatically improves cognition. Berman et al (2008) subsequently contrasted nature and urban exposure using attention tasks after walking and image viewing. The experiments provide controlled, short-term evidence about selected cognitive outcomes, while the generalization to everyday school performance, lifelong cognition, or clinical disorders requires different research designs.
Stress recovery accounts focus on affective and physiological responses to perceived environmental safety and natural content. Ulrich (1984) influenced this line historically, and Hartig et al., 200300109-3) examined restoration in field settings using repeated affective, attention, and cardiovascular measures. These theories are partly complementary: a given environment may change mood and directed attention at different rates, and an attractive setting can be distracting or physically challenging depending on the visitor. An environment that facilitates recovery for one population may be inaccessible, unsafe, culturally unfamiliar, or sensory-overloading for another.
Studies of nature exposure and rumination have proposed more specific mental processes. In Bratman et al (2015), a controlled comparison of nature and urban walking examined self-reported rumination and a neural measure associated with self-referential processing. The resulting experimental contrast is informative about a short exposure under a particular protocol. It cannot independently establish a long-term prevention effect on depression incidence. A sound synthesis connects cognitive restoration to clinical and population outcomes through explicit mediating hypotheses and time-dependent designs, while resisting the transfer of short laboratory effects into unsupported lifetime effect estimates.
Harm Reduction, Capacity Restoration, and Capacity Building
A productive integrative account organizes possible mechanisms around three broad functions: reducing harms, restoring depleted capacities, and building capacities through repeated activity and social opportunity. Markevych et al (2017) articulated this conceptual organization after transdisciplinary review. Harm reduction includes possible moderation of heat, traffic-related pollution, and environmental noise, depending on vegetation design, street geometry, maintenance, and local meteorology. The presence of trees alone does not establish that exposure to every pollutant decreases; local canopy can alter dispersion, and pollen or allergen exposure can be clinically relevant for some groups.
Restoration involves changes in stress appraisal, mood, perceived control, physiological arousal, or attention during or after contact. Capacity building includes walking, play, organized recreation, casual encounters, social cohesion, and the development of routine access to public environments. In an urban neighborhood, a new shaded park may simultaneously create activity opportunities, influence thermal comfort, change visible disorder, and affect residents' evaluation of local institutions. These pathways can be concurrent and statistically correlated, so assigning a percentage of total benefit to any single mechanism requires well-designed mediation studies and valid temporal ordering.
Hartig et al (2014) describes multiple potential pathways linking nature contact to population health, and Kuo (2015) advances candidate mechanisms that connect environmental experience with human physiological function. Their conceptual contributions are valuable for hypothesis formation; mechanisms with limited direct empirical confirmation remain hypotheses. The central methodological implication is that population studies should measure changes in potential mediators alongside exposure and outcomes. Otherwise an observed association between greenness and mental health is compatible with several different causal structures.
Exposure, Contact, Use, and Dose
Environmental availability is the distribution of potential opportunities. Contact is actual proximity or sensory experience. Use is participation in a setting for an identifiable activity. Dose is a multidimensional description of exposure intensity, frequency, duration, and content. These quantities need separate analysis because a resident may live beside a large park without entering it, while another resident may experience substantial nature contact along a daily commute. Residential buffers therefore represent exposure opportunity under assumptions about mobility; they are not direct measures of experienced time in nature. Bardhan et al (2024) reviews how street-visible and geographically proximate vegetation can yield different mental-health associations.
A prominent empirical analysis by White et al (2019) examined reported nature contact during the previous seven days in a large English survey. Greater likelihood of good self-rated health and high well-being was observed among people reporting at least 120 minutes compared with those reporting no such contact. This is an observational threshold pattern based on self-reported exposure and health; it does not establish an individualized therapeutic prescription, a universal minimum dose, or a precise causal threshold. Differences in the ability to travel, baseline health, neighborhood opportunity, and recreation preferences remain relevant even after statistical adjustment.
Dose models should investigate nonlinearities, seasonality, repeated contact, recovery intervals, and the distinction between brief restorative experiences and cumulative environmental opportunity. A strong design would combine geospatial context with consent-based mobility traces, structured experience sampling, weather and pollution records, and repeated outcome measures. Such designs can discriminate frequent short contacts from rare intensive visits while acknowledging participant burden, differential smartphone access, and location privacy. Without these controls, the language of a natural 'dose' may imply a precision that most existing environmental observations have not attained.
Population Health as a Distributional and Causal Question
The public-health value of natural environments depends jointly on the average relationship between exposure and outcome and the distribution of environments across populations. A population with high mean tree cover may still contain neighborhoods where residents cannot safely reach a maintained park. Conversely, an urban district with relatively little vegetation may offer highly used waterfront facilities, shaded pedestrian routes, or nearby periurban landscapes. The proper unit of inference can be an individual, household, school, neighborhood, or municipality, and each introduces different forms of clustering and ecological bias.
A causal study asks what would happen to a defined population if a feasible environmental intervention changed a specified aspect of exposure, compared with a plausible alternative. That intervention might plant street trees, restore a polluted riverbank, add accessible paths, or maintain existing vegetation. An observational association between NDVI and mortality cannot directly quantify any one such intervention because NDVI aggregates distinct changes. Estimates also depend on the covariates included, the baseline landscape, follow-up duration, and whether residents move or public services change concurrently. The heterogeneous cohort evidence synthesized by Rojas-Rueda et al., 201930215-3) and the intervention-specific evidence of South et al (2018) are therefore distinct contributions.
Selection into neighborhoods, longitudinal mobility, health-related constraints on recreation, local policy placement, and environmental disinvestment all create pathways of confounding or reverse causation. Studies should distinguish total effects of environmental changes from controlled direct effects after adjusting mediators such as physical activity. Adjusting for a mediator can remove part of the pathway of interest, while failing to adjust for prior socioeconomic confounders can generate spurious correlations. Later chapters treat these questions through separate designs, outcome families, and implications for urban decision-making.
Spatial Exposure Measurement and Ecological Validity
Satellite-derived vegetation measures, including the normalized difference vegetation index, quantify vegetation-related reflectance within a defined spatial and seasonal sampling procedure. They enable broad comparisons and historical reconstruction, but they cannot by themselves determine public entry, tree shade at sidewalk level, vegetation diversity, perceived safety, or actual time spent in a setting. Land-cover datasets may better distinguish tree canopy from grass or water; cadastral and public-access information can distinguish publicly usable sites from private vegetation. Buffer choices around residences, schools, and workplaces must be justified in terms of the hypothesized exposure pathway and the population's mobility. The study summarized by Rojas-Rueda et al., 201930215-3) restricted its quantitative synthesis to longitudinal studies using NDVI, a methodological choice that supports comparability while narrowing the definition of green space.
Street-level photographs and computer-vision classifications capture the environments visible along sampled roads and pedestrian routes. They can improve exposure validity for visual restoration hypotheses, but image collection has its own sampling biases and temporal variation. Bardhan et al (2024) reviews associations between street-visible green and blue features and mental health, illustrating why satellite greenness, seen greenness, and physical accessibility are different constructs. A route lined with trees may be represented differently in overhead and horizontal imagery; each measure can be correct for a distinct research question.
Measurement error is consequential. In ecological studies, exposure is assigned from an address or administrative area, with uncertainty about where the person spent time. Errors can be correlated with deprivation, age, disability, homeworking, and travel. Strong designs disclose image dates, spatial resolution, seasonal sampling, buffer geometry, geocoding precision, missingness, and accuracy validation. The exposure metric should be described sufficiently for another team to reconstruct its meaning, even when precise participant locations remain private.
Measuring Blue Space as a Distinct Environmental Exposure
Aquatic environments require indicators that reflect both water presence and meaningful contact. Distance to coast or inland water can measure geographic proximity, whereas waterfront visibility, accessible shoreline length, recreational facilities, water cleanliness, and route safety reflect different dimensions of opportunity. The English coastal analysis of Wheeler et al (2012) identified an association between proximity to the coast and self-reported good health and proposed stress and activity pathways. Coastal proximity simultaneously tracks housing markets, local economies, and region-specific infrastructure, so identifying the water-specific contribution requires more than geographic association.
The systematic review by Gascon et al (2017) identified positive findings in parts of the quantitative blue-space literature while emphasizing variation across types of exposure, outcomes, and research designs. The model presented by White et al (2020) adds mechanisms relevant to aquatic landscapes, including recreation, social engagement, nature connection, and the interactions between environmental health and human activity. A polluted or dangerous waterfront can increase contact with harmful conditions even if visitors report that the view is attractive. Water access thus demands attention to safety alongside psychological experience.
A recent 2026 meta-analysis by B. Yang et al (2026) compared visual blue-space exposure with spatial proximity or coverage in studies of adult depression-related outcomes. It reported a very small pooled inverse association and judged certainty very low, with no statistically significant difference between metric classes. The numerical distinction among metrics should therefore be used to refine hypotheses and measurement rather than to declare one visual or spatial method superior. Future research needs repeated direct contact measures and intervention data that disentangle water-related effects from accompanying green infrastructure and neighborhood conditions.
Outcomes, Instruments, and Psychological Construct Validity
Mental-health studies use distinct kinds of endpoints: positive affect, perceived stress, distress scales, symptom inventories, clinical interviews, administrative diagnoses, psychotropic medication use, and service use. These endpoints reflect different processes, measurement windows, and thresholds. The randomized vacant-lot study of South et al (2018), for example, assessed reported components of distress rather than physician-confirmed depression remission. The Danish study of Engemann et al (2019) used register-based psychiatric outcomes across development rather than a short-term mood inventory. Conflating their outcomes would misrepresent both findings.
Cognitive and functional outcomes include directed-attention performance, working memory, executive control, classroom attention, school achievement, mobility participation, and ability to engage in everyday activities. Experimental attention tasks from Berman et al (2008) offer strong control over immediate conditions while sampling narrow task performance. School-linked longitudinal measures from Dadvand et al (2015) broaden ecological relevance but require adjustment for schooling and neighborhood conditions. Validity depends on relating the instrument to the proposed construct and specifying whether the scale or task generalizes across age, language, culture, and disability.
Physical health may be measured through blood pressure, sleep, physical activity, metabolic markers, incidence of diagnosed disease, reproductive outcomes, hospitalizations, or all-cause mortality. Biological markers can respond to momentary activity and measurement conditions, while chronic disease integrates years of exposures. A measurement model should therefore specify the causal time horizon and lag between exposure, mediator, and endpoint. Reviews such as Wang et al (2025) synthesize heterogeneous endpoint families, but the presence of statistically favorable effects across categories does not produce a common latent variable called 'nature health' with a single measurement unit.
Cross-Sectional Studies, Selection, and Spatial Confounding
Cross-sectional exposure–outcome comparisons are efficient for generating hypotheses and mapping inequities. They are especially common when remote sensing is available for a large number of residential areas. Their principal interpretive challenge is that exposure and outcome are often assessed simultaneously. A person with poor mental health may limit outdoor activity, choose or be constrained to a particular neighborhood, or report poorer environmental quality because of current mood. Neighborhood affluence can affect both measured greenness and access to healthcare. These mechanisms can create or amplify associations without an intervention effect of the environment itself.
Spatial confounding can arise through transportation access, pollution, street safety, urban density, municipal expenditure, residential segregation, local employment, and housing conditions. Standard adjustment for age, income, and educational attainment may not fully represent these structures. The field should use causal diagrams to distinguish pre-exposure confounders, intermediate pathways, colliders, and variables influenced by prior exposure. Adjusting for concurrent physical activity, for example, can obstruct an important route through which green space may affect health. Adjusting for neighborhood attributes measured after a greening intervention can obscure consequences of the intervention.
Geographic autocorrelation and multilevel clustering complicate conventional standard errors; individuals in nearby areas share context and sometimes exposure. The appearance of an extremely precise point estimate in a large cross-sectional dataset can mask serious model-based uncertainty. Gascon et al (2015) already identified heterogeneous exposure assessment and limited causal evidence in earlier long-term studies. Later synthesis by Wang et al (2025) documents more research but also reinforces the need to evaluate certainty separately for each endpoint, intervention, and population.
Longitudinal Evidence and Residential Mobility
Longitudinal designs add temporal order and can examine environmental changes before subsequent changes in health. Alcock et al (2014) studied individuals relocating to neighborhoods with different degrees of greenness and examined mental-health trajectories using repeated survey measurements. Its within-person dimension is informative because stable traits can be partly separated from changing environmental context. Yet moving is a selected event, often connected to employment, family circumstances, relationship changes, housing costs, and health. Changes coincident with relocation may still produce residual confounding.
Register-linked longitudinal cohorts can connect childhood environment with later diagnoses. Engemann et al (2019) examined cumulative childhood residential greenness and psychiatric outcomes in Denmark. The study reported higher subsequent psychiatric risk among those with low childhood greenness and explored adjustment for urbanization and familial characteristics. It provides evidence of temporally ordered association across an extended period. Exposure histories may still proxy social and physical features not completely captured by available records, and effect estimates from Denmark cannot automatically be transported to settings with different institutions, climate, urban form, or healthcare systems.
Longitudinal studies should model residential mobility as both a source of exposure variation and a selection process. Marginal structural models, person-fixed-effect designs, and event-time comparisons may address selected biases when their assumptions are satisfied. Pre-trends and negative-control outcomes can test aspects of a proposed causal story. Migration of healthier households toward more attractive locations, attrition from repeated surveys, and time-varying confounding require explicit handling. An increasing number of person-years or observations is valuable only insofar as measured exposure and causal identification remain credible.
Experimental Manipulations and Naturalistic Interventions
Experiments directly manipulate a specific exposure or intervention and can offer stronger causal identification for near-term outcomes. Randomized walking or viewing experiments, including Berman et al (2008) and Bratman et al (2015), compare environmental conditions while measuring attention, mood, or rumination. Their strengths include defined treatment periods, controlled contrasts, and observable task outcomes. Their limitations include short follow-up, convenience sampling, awareness of environmental assignment, and the inability to isolate every physical feature of an outdoor setting. Effects of a particular walk should be described as effects of that tested experience under its assigned comparator.
Community-based interventions examine environmental changes as public goods. South et al (2018) randomized vacant-lot clusters in Philadelphia to greening, trash cleanup, or control conditions and measured mental-health outcomes among nearby adults. The trial provides a rare population-facing experimental test of a concrete urban greening intervention. Its measured effects cannot be separated entirely into vegetation, removal of disorder, maintenance, visual improvement, and institutional signaling, because those changes occur together in the treatment package. Such practical complexity is a feature of a real intervention and should be documented rather than hidden.
Related research by Branas et al (2018) examined crime, violence, and fear around restored vacant land. It supports the importance of environmental intervention as a neighborhood-level process but answers different outcomes from the mental-health trial. In both cases, cluster randomization, participant follow-up, local implementation fidelity, and spatial interference matter. The mechanisms and effect sizes may change if similar interventions occur in a different urban context. The next generation of trials should specify whether the policy question concerns planting, maintenance, safe accessibility, social participation, or a combined environmental program.
Systematic Reviews, Meta-Analysis, and Certainty of Evidence
Evidence synthesis can increase precision and reveal recurring patterns across studies, provided that exposure definitions and outcomes are sufficiently aligned. Twohig-Bennett & Jones (2018) reviewed and meta-analyzed diverse green-space associations with health; Rojas-Rueda et al., 201930215-3) restricted its mortality meta-analysis to longitudinal studies with NDVI measures. The latter reported a pooled all-cause mortality hazard ratio of 0.96 per 0.1 higher residential NDVI within buffers of 500 meters or less, with a 95% confidence interval from 0.94 to 0.97 and high heterogeneity. This is an adjusted observational association across the included cohorts, not the estimated effect of planting a specified number of trees.
The 2025 umbrella review of Wang et al (2025) brought together meta-analyses across green and blue exposure and multiple health endpoints. Umbrella reviews can describe breadth and appraise the certainty of pooled literature, but they may inherit overlapping primary studies, varying review quality, endpoint duplication, and geographically selective original research. The later quality-focused umbrella review J. Yang et al (2026) documents the methodological diversity of what investigators call environmental 'quality'. An authoritative synthesis therefore reports what was pooled, the population and design that produced the estimate, and the degree of overlap and heterogeneity, rather than using the number of reviews as a proxy for causal certainty.
For blue space, Han & Zhang (2026) assessed certainty across six meta-analyses, while B. Yang et al (2026) reported a pooled standardized mean difference near −0.04 for depression-related outcomes with very low certainty. A statistically detectable effect can still be individually small and vulnerable to confounding. The proper comparison is between the study's outcome scale, plausible baseline risk, intervention feasibility, uncertainty, and expected distributional impact. A meaningful evidence hierarchy remains conditional on the specific causal question: a high-quality community randomized trial may be compelling for a concrete greening package, whereas an extensive cohort may be better suited to longer-term mortality patterns.
Depression, Distress, and Common Mental Health Outcomes
The association between natural environments and mental health is among the most intensively investigated topics in environmental epidemiology, yet the strength of evidence depends strongly on how exposure and outcome are operationalized. A cross-sectional association between green residential surroundings and lower depressive symptom scores is consistent with an environmental benefit, but also with residential selection, social advantage, and differences in perceived safety. Earlier reviews such as Gascon et al (2015) emphasized insufficiency of causal evidence, particularly for specific environmental quality and blue-space exposures. The later umbrella review by Wang et al (2025) reported favorable mental-health associations but did not eliminate the heterogeneity and inherited biases of constituent studies.
A particularly informative intervention is South et al (2018), in which 110 clusters of vacant lots were randomized in Philadelphia. Among 342 participants with follow-up data, the greening intervention was associated with significantly lower reports of feeling depressed and worthless relative to no intervention. The reported relative reduction for the depressed-feeling item was 41.5%, while the composite measure of poor mental health showed a 62.8% relative reduction whose confidence interval crossed the null and whose reported p value was 0.051. This distinction is scientifically decisive: the composite finding should not be described as conventionally statistically significant. Outcomes were self-reported and tied to one intervention package in one city, rather than diagnostic recovery or a nationwide treatment effect.
Experimental studies of short natural exposure identify proximal psychological changes, including reported rumination and affect, under relatively well-controlled comparisons Bratman et al (2015). Longitudinal mental-health studies Alcock et al (2014) add evidence about trajectories after residential changes but retain selection problems. Taken together, these lines support serious consideration of environmental interventions for population mental health while leaving open which environmental attributes produce durable reductions in diagnosable disorders. Mental-health endpoints need consistent follow-up windows, prospective registration, independent replication, and explicit handling of access barriers and adverse conditions.
Psychological Well-Being, Pleasure, and Nature Contact
Psychological well-being encompasses positive affect, life satisfaction, perceived functioning, meaning, and other constructs that are associated but not interchangeable. Nature-contact surveys often combine overall self-rated health and subjective well-being because both can be recorded in large household samples. White et al (2019) found that English respondents reporting at least 120 minutes of nature contact in the preceding week more often reported good health and high well-being than those reporting none, after adjustment for selected individual and neighborhood factors. The result is a population-level dose–response association, and its public familiarity should not obscure that self-selection and reverse direction remain possible.
Positive experiences can originate in several ways: quiet, aesthetic appreciation, moderate activity, an encounter with other people, perceived escape from immediate demands, or connection to personally meaningful places. The conceptual account of White et al (2020) stresses that blue spaces may provide particular forms of affordance and connection as well as physical and environmental risk. People may value a riverbank because of social memory, exercise, fishing, views, or ecological familiarity. Accordingly, self-reported satisfaction with a visit does not isolate the contribution of water to that experience.
A research program concerned with subjective well-being should measure both immediate affective change and sustained functioning. Ecological momentary assessment can track within-day effects but can be biased by who participates, when notifications arrive, and the settings participants select. Repeated pre/post measures without a matched control are vulnerable to regression to the mean and ordinary daily cycles. The strongest interpretations combine environmental documentation, valid psychological instruments, suitable comparators, and longer follow-up.
Anxiety, Stress, and Physiological Recovery
Stress-related hypotheses are grounded in the expectation that some environments reduce exposure to demands, support perceived safety, or permit attentional recovery. Hartig et al., 200300109-3) studied psychophysiological and psychological restoration in natural and urban field settings, drawing attention to blood pressure, emotions, and attention as potentially distinct endpoints. The design offers evidence for short-term restorative patterns in the sampled participants. It does not make physiological restoration a universal response to every green or blue environment, especially where weather, allergen exposure, physical danger, or social exclusion alter perceived and actual safety.
Affective responses and physiological measures can diverge. Heart rate varies with physical exertion; salivary cortisol depends on diurnal timing, anticipation, meals, medication, and collection procedures. An observed lower stress rating after a walk can reflect time away from work or ordinary relaxation independent of the landscape. Conversely, the same environment can increase exertion and heart rate while improving mood. Mechanism-focused studies should preregister primary outcomes, standardize measurement time, describe environmental and activity differences, and examine competing explanations.
The relation to population-level anxiety disorders remains less certain than associations with short-term distress and restorative experience. Broad reviews Hartig et al (2014) and Markevych et al (2017) present plausible routes from environmental contact to stress-related health, but evidence for disorder prevention requires prospective diagnostic endpoints. Research in high-risk groups should distinguish ordinary anxiety symptoms, clinically significant impairment, weather-related risks, and social conditions shaping access to safe outdoor environments. Translating this evidence into clinical guidance should be contingent on the relevant population and intervention evidence.
Childhood Development, School Context, and Life-Course Exposure
Childhood exposure is especially important because opportunities for play, mobility, social interaction, and cognitive engagement depend on caregivers, schools, and neighborhood organization. In Barcelona, Dadvand et al (2015) linked green-space measures at school, home, and commuting routes to repeated cognitive development measures in primary schoolchildren. The design highlighted school environmental opportunity as a potential determinant of learning-related outcomes. Nevertheless, school resources, residential socioeconomic patterning, air pollution, and unmeasured contextual features complicate causal attribution to vegetation alone.
Engemann et al (2019) used Danish registers and high-resolution measures of childhood green space to investigate psychiatric outcomes from adolescence into adulthood. It reported that children with the lowest level of surrounding greenness had up to 55% higher relative risk for several psychiatric outcomes compared with those with the highest, with statistical adjustment for multiple family and urbanization variables. The phrase 'up to' describes the upper end of findings across outcomes; it should not be recast as a universal 55% population risk increase. Such results are also association estimates, not the effects of a randomized redistribution of childhood greenspace.
The life-course perspective requires separating sensitive periods from accumulated exposure. A school-age child may experience vegetation through supervised activity and play, whereas a teenager's independent mobility and an older adult's physical accessibility generate different exposure pathways. Repeated residential addresses do not provide complete exposure histories, especially for time spent at school or in transit. Longitudinal studies should include changing mobility, environmental quality, urban redevelopment, family circumstances, and cross-generational inequalities. These distinctions also prevent a school greenness result from being generalized directly to adult dementia risk or all mental disorders.
Physical Activity, Mobility, and Everyday Function
Green and blue spaces can support physical activity by supplying pleasant, safe, and connected locations for walking, play, cycling, and outdoor recreation. However, their presence does not guarantee participation, and greater reported activity among residents of green neighborhoods may reflect selection into those settings. Safe entrances, lighting, accessible routes, public transportation, toilets, maintenance, and perceived legitimacy of use determine whether physically present environments are functionally available. The causal object is frequently a bundle of spatial opportunity, social invitation, and environmental comfort rather than a vegetation index.
Markevych et al (2017) places activity and social opportunity within capacity-building pathways, distinct from harm reduction and psychological restoration. In this model, a park may influence cardiometabolic outcomes indirectly by increasing regular movement, while a scenic waterfront could encourage walking through a different set of sensory and recreational opportunities. The review by White et al (2020) emphasizes both overlapping and distinct aquatic mechanisms. Evidence comparing exercise in green settings against exercise in other contexts must distinguish the effect of exercise itself from the incremental effect of environmental setting.
Mobility outcomes are influenced by disability, age, household responsibilities, gendered safety concerns, climate, and social norms. Studies that include only able-bodied recreational visitors can exaggerate policy transferability. Inactivity may also arise from fear of harassment, environmental hazards, or insufficient time rather than a lack of nominal green area. Intervention evaluations should measure actual visits, step counts or activity minutes when valid, access among previously excluded groups, and potential substitution from other activities. A net public-health claim requires considering changes in total behavior rather than visits counted in isolation.
Cardiovascular, Metabolic, and Mortality Outcomes
Distal physical-health outcomes provide a powerful test of public-health relevance because they integrate the consequences of many behavioral, environmental, and social processes over time. Twohig-Bennett & Jones (2018) reported associations across multiple physical-health outcomes in its systematic review and meta-analysis, while Rojas-Rueda et al., 201930215-3) focused specifically on longitudinal measures of green exposure and all-cause mortality. The latter included nine studies with approximately 8.3 million individuals from seven countries. Its pooled adjusted hazard ratio was 0.96 per 0.1 greater residential NDVI within a buffer of 500 meters or less; heterogeneity was very high, with reported I² of 95%. This combination of precision and heterogeneity cautions against universalizing a single effect.
The longitudinal nature of mortality cohorts helps establish exposure before death, but residential environmental quality may be correlated with housing, medical access, neighborhood deprivation, physical activity, thermal exposure, and pollution. Different cohorts may adjust different covariates, and a uniform NDVI increment may denote unlike environmental transformations in different climates. A hazard ratio for nearby greenness cannot specify whether planting street trees, expanding a public park, restoring a waterfront, or improving residential gardens will replicate the estimate. The policy question must be matched to an intervention with a plausible, validated pathway.
The broad umbrella review Wang et al (2025) reported favorable associations for multiple cardiovascular and metabolic endpoints while identifying outcomes with heterogeneous or unclear patterns. Claims should therefore be outcome-specific, and observational effect estimates should not be restated as confirmed clinical prevention effects. Population health assessment should combine effect uncertainty with local exposure distribution, baseline disease rates, competing interventions, and long-term maintenance costs.
Environmental Pathways: Heat, Air, Noise, and Other Hazards
Urban vegetation modifies the local physical environment through shade, evapotranspiration, surface permeability, and interactions with built form. These processes can alter the heat burden experienced during daily activity, though magnitude depends on canopy structure, water availability, pavement, wind, and local climate. Greenspace may also change acoustic conditions and exposure to some airborne pollutants. The integrated framework by Markevych et al (2017) treats these as harm-reduction pathways that operate alongside psychological experiences and social affordances, requiring separate environmental measurements to test the pathways.
Causal attribution requires recording conditions before and after greening, using appropriate spatial comparators, and avoiding the assumption that more vegetation uniformly means less pollution. Street trees can alter pollutant dispersion in constrained street canyons, and some planted species increase pollen exposure. Blue-space proximity may add weather-related and water-quality risks, including flooding, drowning hazards, contamination, or dangerous heat exposure on unshaded promenades. Public-health evaluation thus needs an assessment of benefits, risks, and the distribution of both.
A multicomponent intervention should document local temperature, shade, pollution, noise, maintenance, and use before asserting that each potential pathway operated. Health improvements after a greening program may result partly from reduced visual disorder or a stronger sense of public care, even when air-quality changes are modest. Conversely, improved physical comfort can influence activity indirectly. A strong mechanism analysis avoids equating correlation among several mediators with confirmed independent causal contributions.
Environmental Quality, Biodiversity, and Experienced Safety
Quantity, quality, and usability are independent dimensions of environmental opportunity. A neighborhood may have many hectares of vegetation but poor paths, contaminated soil, inaccessible entrances, or inadequate maintenance. Another may contain a smaller but shaded and well-used park that supports regular visits. Public-health research increasingly recognizes vegetation type, ecological condition, water quality, pedestrian access, perceived safety, quietness, visual comfort, and social welcome as distinct candidate effect modifiers. The quality-focused umbrella review by J. Yang et al (2026) identified multiple domains and metrics rather than a single agreed quantitative definition of green-and-blue-space quality.
Ecological features such as habitat diversity, tree structure, and seasonal change may affect experiences, but a measured biodiversity index does not automatically predict psychological benefit. Effects depend on which sensory and biological characteristics people encounter and how they interpret them. A naturalistic woodland can offer distinctive environmental experiences and simultaneously create mobility barriers for people with impaired balance or limited access to transport. Studies should therefore test physical quality and perceived quality together and evaluate whether the same features influence different groups in different ways.
Safety has at least three meanings: objective hazards, perceived vulnerability, and the stability of public management. An unlit entrance may reduce visits even when vegetation is abundant. Water features can attract visitors and introduce drowning or flood risks, while poorly maintained urban areas can reinforce avoidance and fear. The vacant-land interventions studied by South et al (2018) and Branas et al (2018) illustrate why environmental improvement may change both perceptual and behavioral conditions. Such bundled interventions should be compared with genuinely feasible alternatives rather than idealized contrasts with a perfectly managed natural setting.
Green and Blue Space as Social Infrastructure
Natural public spaces function as places where routines, informal encounters, organized exercise, civic events, family care, and intergenerational contact occur. Their social effects depend on who is present, how the space is governed, whether there are reasons to stay, and whether different groups can use it without conflict. The conceptual approaches of Hartig et al (2014) and Markevych et al (2017) identify social cohesion as a possible pathway to health. The pathway is empirical, not automatic: a newly renovated park can become a meeting place, remain underused, or generate tensions over access, commercial activity, noise, and competing priorities.
A useful design separates environmental change from changes in social use. Observational counts, structured time-use records, interviews, social-network measures, and repeated neighborhood surveys can assess changes in participation. Perceived collective efficacy, trust, neighborhood belonging, and frequency of social contact should be operationalized independently. When a greening project introduces visible maintenance and municipal investment, any gain in social trust may reflect institutional responsiveness as well as the physical landscape. A mediation model that fails to measure public management risks assigning that benefit inaccurately to vegetation itself.
Blue spaces can provide recognizable locations for community activity, recreation, and cultural identity. White et al (2020) discusses feedback between contact with aquatic environments, well-being, and environmental action. However, attractive waterfront redevelopment may replace affordable community facilities with commercial uses or limit traditional access. The health significance of a waterfront depends on the social organization of permission, affordability, transport, and safety as much as on the presence of water. Research should report which groups benefit and whether existing users lose access.
Environmental Inequality and Unequal Exposure Opportunities
Environmental exposures are distributed through urban history, housing markets, infrastructure investment, and institutional decisions. Residential greenness often correlates with economic advantage, yet the distribution varies by country and municipality. Comparisons should characterize both nominal access and actual usability. Publicly owned greenspace within a short straight-line distance can still be functionally inaccessible because of highways, steep grades, barriers, opening hours, discrimination, or fees. Access should therefore be measured in realistic travel time and with population-specific mobility constraints.
Environmental inequality can alter both baseline health and the likely effect of a new intervention. In South et al (2018), the intervention's reported reduction in the depressed-feeling item was particularly pronounced in a subset of neighborhoods below the poverty line. That subgroup result is informative but requires caution about multiplicity, context, and transferability. The coastal proximity findings of Wheeler et al (2012) also raised the possibility of socioeconomic differences in observed association. Neither result establishes that every low-income area will experience the same effect from identical landscape changes.
Fair distribution requires analysis of who can use a place, who bears maintenance and displacement costs, and whether the intervention changes local housing affordability. A greening program may improve physical conditions while increasing rents or stimulating speculative redevelopment. Such possibilities require longitudinal housing, migration, and social outcomes, not assumption alone. Community participation can improve the relevance of interventions by identifying the actual environmental barriers residents experience. Equity analysis belongs within causal evaluation because treatment uptake and resulting exposures are structured by social conditions.
Disability, Aging, Gender, and Cultural Variation
Population heterogeneity is fundamental to environmental-health inference. Adults with mobility limitations may benefit from level paths, accessible transit, shade, benches, and safe crossings even if their total neighborhood NDVI is modest. Older adults may depend on close and reliable access, while children require safe supervision, opportunities for play, and school-based exposure. A waterfront that is accessible to recreational runners may be inaccessible to a wheelchair user or hazardous for an unsupervised child. Measuring environmental opportunity across populations therefore requires more than age or disability as adjustment covariates.
Gendered patterns of safety, caregiving time, and use of public environments can shape exposure and its social meaning. Cultural familiarity, symbolic significance, religious and leisure practices, and attitudes toward water also affect the desirability and use of places. White et al (2020) treats environmental engagement as a socially situated process, and Hartig et al (2014) notes that public-health research must account for variability in opportunities and pathways. Transportable evidence asks whether an exposure-outcome relationship remains plausible under the relevant local practices and institutions.
Researchers should use accessible and linguistically valid instruments and recruit beyond healthy recreational volunteers. Quantitative analyses can predefine effect modification by age, ability, deprivation, and context, but subgroup comparisons need adequate sample sizes and multiplicity control. Qualitative work can reveal mechanisms absent from standard geographic datasets, such as fear of exclusion, absence of toilets, or the importance of culturally familiar activities. Mixed-methods analysis can clarify why a nominally large supply of nature produces little contact for some populations.
Urban Design, Interventions, and Public Policy
Planning interventions should be specified in terms of their delivered components: planting and shade, repair of damaged ground, access routes, physical activity facilities, shoreline rehabilitation, maintenance, lighting, or public programming. These components affect different causal pathways. For a district with severe heat exposure, tree canopy and accessible shade may be central; in a neighborhood with unsafe vacant land, maintenance and visible reduction of disorder may matter; around an existing river, safe paths and water-quality restoration may unlock contact that geography alone did not provide. The research models of Markevych et al (2017) and White et al (2020) encourage such mechanism-specific planning.
Population impact assessments need an explicit counterfactual. A city can allocate resources to new parks, improve the quality of existing green spaces, protect tree canopy, or connect isolated facilities. A comparison that measures only the added hectares may overlook a highly effective accessibility improvement. The randomized package evaluated by South et al (2018) establishes a useful example of intervention-specific inference, but the magnitude and duration of its effects cannot simply be inserted into unrelated cities or used to price cardiovascular mortality gains.
The policy objective should be measurable: improved nearby access, more safe visits, greater active mobility, lower environmental burden, increased psychological well-being, or reduced health inequality. Outcome assessment must be proportionate to plausible timescales. It is reasonable to measure perceived safety soon after restoring vacant land; demonstrating lower chronic disease incidence demands longer observation and robust clinical data. Lifecycle maintenance, climate resilience, and ecological sustainability should be considered alongside health outcomes, with benefit estimates updated as actual use and conditions change.
Blue Space: Restoration, Recreation, and Risk
Blue spaces have received less cumulative research attention than green spaces, and the quality and nature of evidence are uneven. Gascon et al (2017) reviewed quantitative studies of outdoor aquatic environments and health; White et al (2020) developed a narrative synthesis of possible pathways. Coastal proximity studies such as Wheeler et al (2012) contribute population-level associations, while later evidence reviews Han & Zhang (2026) and B. Yang et al (2026) draw attention to differences in outcome, measurement, and certainty. These strands collectively justify treating aquatic environments as a serious research object with distinctive characteristics.
In a 2026 synthesis, B. Yang et al (2026) pooled six studies of adult depression-related continuous outcomes within a wider eight-study review and reported a very small inverse association, standardized mean difference −0.04 with a 95% confidence interval from −0.06 to −0.02. The certainty was rated very low. The comparison between visual and spatial measures was exploratory and lacked a statistically reliable subgroup difference. This result warns against presenting waterfront visibility as a proven clinical treatment or ranking all blue-space types by apparent numerical effect sizes. The measured outcome and uncertainty matter as much as the direction of association.
Potential benefits include scenic contact, leisure, exercise, social use, perceived restoration, and connection to place. Countervailing conditions include poor water quality, flood exposure, hazardous access, excessive heat on unshaded banks, and inequitable development. Distinct aquatic environments should therefore be analyzed separately when data permit. A lake, an urban canal, and an ocean coast have different hydrology, sensory cues, social uses, and risks. Effective interventions may couple water-quality restoration with public access and green infrastructure, creating a need for designs that can examine the contribution and interaction of each component.
Parks, Street Trees, Waterways, and Multi-Component Landscapes
The umbrella category of green and blue space is useful for public communication but can obscure intervention-relevant distinctions. Street trees may create continuous shade and visual greenery along ordinary routes, whereas a large park may allow prolonged recreation. Urban wetlands may provide ecological services alongside limited recreational access; community gardens may combine soil contact, social activity, and food production; coastal promenades can support movement with little surrounding vegetation. The consequence for research is the need to define the functional exposure rather than assume equivalence from a shared natural label.
Spatially complex landscapes are especially challenging because green and blue exposures co-occur. A park around a lake can be coded as park area, waterside access, vegetation, canopy, biodiversity, and recreation infrastructure at once. Estimating independent coefficients for each feature in a regression model can be unstable when measures are strongly correlated and when omitted features jointly determine access. Factor analyses or latent constructs may summarize observations but do not establish that the latent variable is a causal agent. Mechanism-specific hypotheses and intervention contrasts remain essential.
J. Yang et al (2026) demonstrates the wide diversity of environmental quality constructs represented in recent reviews. Bardhan et al (2024) similarly illustrates how visual evidence can capture aspects of exposure missed by coarser spatial measures. For policy evaluation, the preferred dataset integrates geometry, actual routes, site conditions, maintenance, visual features, activity observations, and public feedback. Scientific precision increases when an apparently simple statement such as 'more nature improves health' is replaced by an explicit specification of what changed, where, for whom, and under which conditions.
Causal Identification and Competing Explanations
Causal claims require a well-defined environmental intervention, an appropriate comparison, a plausible assignment mechanism, and a transparent account of what was measured before and after exposure. Population surveys can identify patterned associations, whereas randomized environmental interventions and credible quasi-experimental changes can provide stronger evidence for particular actions. Each design nonetheless answers a specific question. The effect of randomized greening of previously blighted land South et al (2018) is not identical to the effect of living near a naturally vegetated neighborhood over decades, and neither is identical to the short-term effect of a prescribed walk Bratman et al (2015).
Competing explanations include healthy-resident selection, differences in wealth, unrecorded institutional quality, changing healthcare access, neighborhood safety, historical housing policies, and simultaneous transport improvements. These factors may act at different scales. Individual adjustment for income cannot fully account for correlated community services; neighborhood fixed effects cannot control every changing local influence. Randomization addresses confounding at assignment but can coexist with attrition, incomplete adherence, differential outcome reporting, interference, or contamination. The 2018 Philadelphia trial South et al (2018) is particularly instructive because its controlled design strengthens inference for the intervention package while leaving interpretation of individual pathways open.
A rigorous design first draws a causal diagram with pre-exposure determinants, hypothesized mediators, and outcomes. It then specifies the target causal contrast, whether the estimand concerns the effect of offering an intervention or the effect among those who actually use the modified landscape. Propensity methods, difference-in-differences, instrumental variables, interrupted time series, and individual fixed effects can support inference only under their respective substantive assumptions. The quality of an identification strategy is determined by whether those assumptions fit the local environmental process, not by the sophistication of the statistical estimator alone.
Heterogeneity, Publication Bias, and Replicability
Evidence on green and blue spaces is heterogeneous because researchers study distinct exposures, outcomes, climates, urban histories, and populations. In Rojas-Rueda et al., 201930215-3), the high between-study heterogeneity in mortality associations indicates that a pooled effect is a summary across nonidentical contexts. A small standard error around the pooled estimate does not remove uncertainty about transfer to a specific city. Review-level summaries Wang et al (2025) can further contain overlapping primary studies, so apparent agreement among reviews may arise from repeatedly analyzing some of the same cohorts.
Publication bias is especially relevant where many spatial buffers, psychological scales, and model specifications are available. Selective presentation of significant buffers can exaggerate confidence even when each reported regression is arithmetically correct. Preregistration of primary exposure, outcome, analysis window, and subgroup comparisons can reduce such flexibility. Multiverse analysis and specification curves may show the stability of conclusions across justified alternatives. When research measures many potentially correlated indicators, correction for multiplicity and transparent reporting of null results become central to trustworthy interpretation.
Independent replication should preserve a core causal question while testing variation across climate, vegetation type, water settings, and public institutions. A single urban greening trial with positive findings deserves replication in communities with different prior green cover and safety profiles. Short experimental cognitive improvements should be tested with additional tasks, longer follow-up, and diverse samples. Reviews must distinguish direct replication, conceptual replication, and transportability across policies. More studies are most useful when they close identifiable gaps in causal inference, measurement validity, and population representation.
Ethical Practice and Participatory Environmental Research
Research on environmental exposure can involve geolocation, residential records, clinical outcomes, and sensitive information about inequality. Publicly available environmental imagery does not remove the privacy implications of linking exact residences with health records. Consent-based mobility collection should minimize location precision where possible, control access to identifiable traces, describe data retention, and permit participation without continuous tracking when feasible. Open analysis code and reproducible exposure algorithms can coexist with protected individual-level geolocation data.
Participatory methods can identify barriers that maps overlook: exclusionary opening hours, unsafe crossings, poorly maintained surfaces, social conflict, inaccessible facilities, fear of harassment, and the loss of valued cultural uses. Community involvement can improve the specification of relevant outcomes and the design of interventions, yet it must not become a substitute for rigorous outcome measurement. Participants should be told whether the project tests an experimental intervention, conducts an observational evaluation, or invites deliberation about municipal priorities.
Ethical analysis also concerns opportunity costs and unintended effects. A new waterfront project may serve tourism while displacing customary users; urban greening may raise property values and strain housing affordability; additional vegetation may alter allergies, fire risk, or perceived surveillance. Researchers should measure adverse and distributional outcomes, especially among groups for whom new access is uncertain. Benefits are scientifically meaningful when their beneficiaries, mechanisms, and countervailing costs are specified rather than assumed.
Comparative Evidence Assessment Across Research Designs
The original experiment of Ulrich (1984) is historically influential and hypothesis-generating but limited by its retrospective comparison and small matched groups. Controlled field and laboratory studies Hartig et al., 200300109-3), Berman et al (2008), and Bratman et al (2015) provide more direct contrasts for immediate cognitive or affective outcomes. Large-scale repeated observational studies Alcock et al (2014), Dadvand et al (2015), and Engemann et al (2019) capture different temporal and developmental scales with residual selection and contextual confounding. The community randomized studies South et al (2018) and Branas et al (2018) offer more compelling causal evidence about concrete interventions and short- to medium-term neighborhood effects.
The mortality meta-analysis Rojas-Rueda et al., 201930215-3) supports an association between surrounding residential greenness and all-cause mortality among included cohorts, while its high heterogeneity and potential residual confounding limit causal extrapolation. The green-space syntheses Twohig-Bennett & Jones (2018) and Wang et al (2025) indicate a broad research base, with evidence quality varying by endpoint. For blue space, Gascon et al (2017), Han & Zhang (2026), and B. Yang et al (2026) reveal a smaller, heterogeneous literature in which some positive associations coexist with very low certainty for a narrowly defined depression outcome.
The central evidentiary distinction is between a directionally consistent association and an estimated causal response to a manipulable environmental change. The former can motivate policy evaluation; the latter is needed to credibly predict an intervention's health effect. Certainty cannot be assigned universally to 'nature exposure.' It must be assigned to a population, exposure definition, outcome, intervention contrast, follow-up period, and risk-of-bias assessment. This principle explains why a robust result concerning acute self-reported mood cannot certify long-term disorder prevention and why a mortality association cannot prescribe a specific greening package.
Research Design Priorities for the Next Generation
Future studies should combine precise exposure measurement with credible contrasts and outcome-specific longitudinal assessment. For green settings, satellite and street-level imagery, public-access layers, vegetation structure, and direct use measurement can be integrated to reveal whether benefits depend on visible canopy, safe routes, biodiversity, or program participation. For blue settings, water quality, shore access, water visibility, and recreational use should be measured independently wherever possible. The exposure is a structured environmental opportunity, and each data layer should correspond to a substantive hypothesis rather than be added merely because technology makes it available.
Large municipal interventions could use phased implementation, lottery allocation when ethically possible, preannounced evaluation plans, and well-matched comparison areas. Repeated surveys combined with administrative indicators would allow researchers to track near-term experience, medium-term activity, and longer-term clinical or functional outcomes. Natural experiments need evidence on parallel pre-intervention trends and concurrent developments. Where randomized assignment is possible, researchers should predefine primary endpoints and examine both intention-to-treat and exposure uptake without confusing the two estimands.
The international evidence base should expand across rapidly urbanizing regions, smaller communities, diverse climates, and populations underrepresented in leisure-oriented research. Attention is needed to people with disabilities, children, older adults, precariously housed residents, and those experiencing displacement. Multi-city replication should report local ecology, maintenance, governance, and access conditions so that transferability can be evaluated. Scientific progress depends on understanding the conditions under which a proposed pathway operates, not simply accumulating more statistically significant spatial associations.
Digital Measurement and the Artificial Era
Digital environmental research increasingly relies on satellite imagery, mobile sensors, computer vision, geographic information systems, and prediction models. These tools can map spatial variation at scales that would previously have required extensive manual observation. Bardhan et al (2024) illustrates how visible environmental features can be studied using street-view imagery, while J. Yang et al (2026) documents the growing complexity of environmental quality metrics. Computational methods should be treated as measurement instruments whose classification errors, temporal coverage, and population validity require independent auditing.
Artificial systems may also mediate environmental contact by recommending routes, ranking destinations, predicting shade or air quality, and influencing which spaces people encounter. A recommender that favors popular commercial waterfronts can change observed visitation patterns independently of ecological opportunity. A navigation system optimized for speed may underrepresent accessible green routes, while a personalized accessibility guide could improve practical access for some users. Such effects require empirical evaluation of actual behavior and distributional consequences; technical capability alone does not establish psychological or public-health benefit.
Within the horizon of Artificial Era · From Homo to Artificial, these technologies create a research question about the organization of human–environment experience through artificial mediation. The philosophical account of Artificial in Aisentica is a separate authorial framework. The empirical claims of this monograph remain grounded in human population studies, validated environmental measurements, and testable intervention hypotheses. Future research should compare human-only route selection, algorithmic recommendations, and assisted decisions with attention to autonomy, privacy, ecological sustainability, and equitable access.
Synthesis of Established Findings and Open Questions
Green and blue space exposure contributes to a research field spanning short psychological experiences and long population-health trajectories. Evidence is most directly interpretable where exposure and outcome are specifically defined. Controlled environmental comparisons support some short-term affective and cognitive effects; a randomized urban greening trial supports intervention effects on selected self-reported mental-health items; longitudinal studies connect residential greenness with later health and psychiatric outcomes; meta-analyses show favorable associations for several physical-health endpoints. Each line comes with a different causal target and level of uncertainty Berman et al (2008) South et al (2018) Engemann et al (2019) Rojas-Rueda et al., 201930215-3).
The research supports a practical emphasis on quality, access, maintenance, and locally appropriate interventions. It also supports preserving the distinction between the physical environment and experienced exposure. A park may be visible but inaccessible, nearby but unsafe, large but underused, or small yet frequently visited. The influence of blue space is particularly sensitive to measurement definitions and remains less firmly established for some clinical outcomes than public enthusiasm suggests Gascon et al (2017) B. Yang et al (2026). An integrated program should recognize environmental hazards, behavior, restoration, and social opportunity as interacting candidate pathways rather than competing universal explanations.
The largest unresolved problems concern causal identification across long time horizons, mediation by social and biological processes, the transfer of findings across institutions and populations, and the equitable distribution of access and benefits. Future systematic syntheses should clearly report constituent study overlap and evidence certainty. Future primary studies should document intervention implementation, exposure uptake, and unintended outcomes. These priorities establish the conditions under which environmental-health research can advance from association toward credible population-level guidance.
Conclusion
Green and blue space exposure is a multidimensional environmental and psychological process linking the material organization of places with opportunities for contact, perception, behavior, social participation, and health. Its population significance depends on the qualities and distribution of these opportunities and on the biological, psychological, and social conditions of those who encounter them. The available literature includes a substantial body of observational associations, targeted experiments, community randomized interventions, and systematic syntheses. Together they provide a reasoned scientific basis for investigating and improving access to safe, usable natural environments, with conclusions calibrated to the designs that support them.
The decisive scientific standard is correspondence between exposure, mechanism, outcome, and inference. Satellite vegetation indices, visible canopy, park visitation, accessible shoreline, and perceived environmental quality represent different aspects of the research object. Momentary mood, cognitive performance, psychological distress, psychiatric incidence, metabolic disease, and mortality are different outcomes with different time horizons. The most defensible findings preserve these distinctions and explicitly address selection, confounding, measurement error, heterogeneity, and equity.
A mature population-health approach treats natural environments as socially organized conditions that can be modified and evaluated. It tests which changes produce benefits, for whom, in which settings, at what cost, and with what unintended consequences. Across the transition From Homo to Artificial, new measurement and decision-support systems may sharpen environmental evaluation, provided that their empirical role remains open to independent assessment. The research agenda is therefore an evidence-based study of environments, experiences, and public consequences, sustained by reliable observation, transparent uncertainty, and intervention-specific inference.
References
- Alcock, I., White, M. P., Wheeler, B. W., Fleming, L. E., & Depledge, M. H. (2014). Longitudinal effects on mental health of moving to greener and less green urban areas Environmental Science & Technology 48(2), 1247–1255 https://doi.org/10.1021/es403688w
- Bardhan, M., Li, F., Browning, M. H. E. M., Dong, J., Zhang, K., Yuan, S., İnan, H. E., McAnirlin, O., Dagan, D. T., Maynard, A., Thurson, K., Zhang, F., Wang, R., & Helbich, M. (2024). From space to street: A systematic review of the associations between visible greenery and bluespace in street view imagery and mental health Environmental Research 263(Pt 3), 120213 https://doi.org/10.1016/j.envres.2024.120213
- Berman, M. G., Jonides, J., & Kaplan, S. (2008). The cognitive benefits of interacting with nature Psychological Science 19(12), 1207–1212 https://doi.org/10.1111/j.1467-9280.2008.02225.x
- Branas, C. C., South, E., Kondo, M. C., Hohl, B. C., Bourgois, P., Wiebe, D. J., & MacDonald, J. M. (2018). Citywide cluster randomized trial to restore blighted vacant land and its effects on violence, crime, and fear Proceedings of the National Academy of Sciences 115(12), 2946–2951 https://doi.org/10.1073/pnas.1718503115
- Bratman, G. N., Hamilton, J. P., Hahn, K. S., Daily, G. C., & Gross, J. J. (2015). Nature experience reduces rumination and subgenual prefrontal cortex activation Proceedings of the National Academy of Sciences 112(28), 8567–8572 https://doi.org/10.1073/pnas.1510459112
- Dadvand, P., Nieuwenhuijsen, M. J., Esnaola, M., Forns, J., Basagaña, X., Alvarez-Pedrerol, M., Rivas, I., López-Vicente, M., de Castro Pascual, M., Su, J., Jerrett, M., Querol, X., & Sunyer, J. (2015). Green spaces and cognitive development in primary schoolchildren Proceedings of the National Academy of Sciences 112(26), 7937–7942 https://doi.org/10.1073/pnas.1503402112
- Engemann, K., Pedersen, C. B., Arge, L., Tsirogiannis, C., Mortensen, P. B., & Svenning, J.-C. (2019). Residential green space in childhood is associated with lower risk of psychiatric disorders from adolescence into adulthood Proceedings of the National Academy of Sciences 116(11), 5188–5193 https://doi.org/10.1073/pnas.1807504116
- Gascon, M., Triguero-Mas, M., Martínez, D., Dadvand, P., Forns, J., Plasència, A., & Nieuwenhuijsen, M. J. (2015). Mental health benefits of long-term exposure to residential green and blue spaces: A systematic review International Journal of Environmental Research and Public Health 12(4), 4354–4379 https://doi.org/10.3390/ijerph120404354
- Gascon, M., Zijlema, W., Vert, C., White, M. P., & Nieuwenhuijsen, M. J. (2017). Outdoor blue spaces, human health and well-being: A systematic review of quantitative studies International Journal of Hygiene and Environmental Health 220(8), 1207–1221 https://doi.org/10.1016/j.ijheh.2017.08.004
- Han, K.-T., & Zhang, X.-L. (2026). Blue space and health/well-being: An overview of six meta-analyses and their evidence certainty (PRIOR item 1) Water Research 288(Pt A), 124530 https://doi.org/10.1016/j.watres.2025.124530
- Hartig, T., Evans, G. W., Jamner, L. D., Davis, D. S., & Gärling, T. (2003). Tracking restoration in natural and urban field settings Journal of Environmental Psychology 23(2), 109–123 https://doi.org/10.1016/S0272-4944(02)00109-3
- Hartig, T., Mitchell, R., de Vries, S., & Frumkin, H. (2014). Nature and health Annual Review of Public Health 35, 207–228 https://doi.org/10.1146/annurev-publhealth-032013-182443
- Kaplan, S. (1995). The restorative benefits of nature: Toward an integrative framework Journal of Environmental Psychology 15(3), 169–182 https://doi.org/10.1016/0272-4944(95)90001-2
- Kuo, M. (2015). How might contact with nature promote human health? Promising mechanisms and a possible central pathway Frontiers in Psychology 6, 1093 https://doi.org/10.3389/fpsyg.2015.01093
- Markevych, I., Schoierer, J., Hartig, T., Chudnovsky, A., Hystad, P., Dzhambov, A. M., de Vries, S., Triguero-Mas, M., Brauer, M., Nieuwenhuijsen, M. J., Lupp, G., Richardson, E. A., Astell-Burt, T., Dimitrova, D., Feng, X., Sadeh, M., Standl, M., Heinrich, J., & Fuertes, E. (2017). Exploring pathways linking greenspace to health: Theoretical and methodological guidance Environmental Research 158, 301–317 https://doi.org/10.1016/j.envres.2017.06.028
- Rojas-Rueda, D., Nieuwenhuijsen, M. J., Gascon, M., Perez-Leon, D., & Mudu, P. (2019). Green spaces and mortality: A systematic review and meta-analysis of cohort studies The Lancet Planetary Health 3(11), e469–e477 https://doi.org/10.1016/S2542-5196(19)30215-3
- South, E. C., Hohl, B. C., Kondo, M. C., MacDonald, J. M., & Branas, C. C. (2018). Effect of greening vacant land on mental health of community-dwelling adults: A cluster randomized trial JAMA Network Open 1(3), e180298 https://doi.org/10.1001/jamanetworkopen.2018.0298
- Twohig-Bennett, C., & Jones, A. (2018). The health benefits of the great outdoors: A systematic review and meta-analysis of greenspace exposure and health outcomes Environmental Research 166, 628–637 https://doi.org/10.1016/j.envres.2018.06.030
- Ulrich, R. S. (1984). View through a window may influence recovery from surgery Science 224(4647), 420–421 https://doi.org/10.1126/science.6143402
- Wang, X., Feng, B., & Wang, J. (2025). Green spaces, blue spaces and human health: An updated umbrella review of epidemiological meta-analyses Frontiers in Public Health 13, 1505292 https://doi.org/10.3389/fpubh.2025.1505292
- Wheeler, B. W., White, M., Stahl-Timmins, W., & Depledge, M. H. (2012). Does living by the coast improve health and wellbeing? Health & Place 18(5), 1198–1201 https://doi.org/10.1016/j.healthplace.2012.06.015
- White, M. P., Alcock, I., Grellier, J., Wheeler, B. W., Hartig, T., Warber, S. L., Bone, A., Depledge, M. H., & Fleming, L. E. (2019). Spending at least 120 minutes a week in nature is associated with good health and wellbeing Scientific Reports 9, 7730 https://doi.org/10.1038/s41598-019-44097-3
- White, M. P., Elliott, L. R., Gascon, M., Roberts, B., & Fleming, L. E. (2020). Blue space, health and well-being: A narrative overview and synthesis of potential benefits Environmental Research 191, 110169 https://doi.org/10.1016/j.envres.2020.110169
- Yang, J., Akafekwa, T., Williams, G., Elliott, J., Verma, A., & Greenhalgh, C. (2026). Quality of urban green and blue spaces with a specific focus on public health outcomes: An umbrella review Environmental Research 306(Pt 1), 125141 https://doi.org/10.1016/j.envres.2026.125141
- Yang, B., Zhou, Y., Liu, H., & Li, A. (2026). Comparing visual-exposure and spatial proximity/coverage metrics of urban blue space in relation to depression-related outcomes: A systematic review and meta-analysis Frontiers in Public Health 14, 1876283 https://doi.org/10.3389/fpubh.2026.1876283