Cloud Noticing and Forest-Therapy Elements in a Coastal Resort: An Exploratory Randomised Field Study

Highlights

What are the main findings?

• In this exploratory randomised field study, model-based group-by-time contrasts indicated more favourable immediate psychological responses after cloud noticing and forest-therapy elements than after a structurally matched active-control session.
• The descriptive experiential patterns were not identical: cloud noticing was associated most strongly with awe, mental spaciousness, and perspective distancing, whereas forest-based engagement was marked by multisensory and embodied qualities.

What are the implications of the main findings?

• Resorts can create restorative programming by choreographing how guests attend to place-specific environmental features, with particular attention to pacing, safety, facilitation, and voluntary participation.
• The evidence supports flexible destination-specific modules, but it does not demonstrate treatment efficacy, persistence of effects, financial performance, or universal superiority of either format.

Abstract

Wellness resorts often market natural scenery without knowing whether a structured way of attending to it changes guests’ immediate experience. We examined cloud noticing and forest-therapy elements as two low-infrastructure forms of nature-directed hospitality programming. This exploratory, randomised field study reports outcomes from three facilitator-led arms of a four-condition parent experiment whose service-evaluation findings were published separately. The parent experiment allocated 120 adult resort guests equally, by individual manual lottery, to cloud noticing, forest-therapy elements, an active-control session, or an ordinary stay. The present complete-case dataset comprised 84 participants with linked measurements immediately before and after an organised activity: cloud noticing (n = 29), forest-therapy elements (n = 28), and active control (n = 27). The ordinary-stay group was outside the analytical scope because no additional session, and therefore no matched session-level measurement interval, was available. Linear mixed-effects models estimated group-by-time contrasts for state anxiety, positive and negative affect, and subjective restoration; perceived stress, connectedness with nature, six activity-specific ratings, resting pulse, and blood pressure were secondary or descriptive variables. Both nature-focused sessions showed more favourable short-term psychological patterns than the active control, with the clearest contrasts for subjective restoration. Cloud noticing was associated descriptively with awe, mental spaciousness, and perspective distancing, whereas forest-therapy elements were associated with sensory immersion, bodily presence, grounding, and the greatest increase in nature connectedness. Pulse and blood pressure are reported only as physiological field descriptors. The earlier companion publication examined experience quality, satisfaction, novelty, added value, and recommendation intention; none of those endpoints are reanalysed here. Applicability is limited by the single-resort setting, lack of blinding, complete-case analysis, absence of preregistration, and lack of follow-up.

1. Introduction

Natural amenities contribute to a resort’s appeal, but scenery alone does not constitute an organised guest experience. A view of the sky, a wooded path, shade, sound, texture, and contact with the ground become service resources only when the destination provides guests a comprehensible and safe way to engage with them. The relevant managerial task is therefore not simply to possess attractive surroundings, but to shape an encounter in which environmental opportunity, staff practice, timing, and guest attention work together.

This issue is especially visible in coastal wellness properties, where atmospheric light, the horizon, clouds, and Mediterranean vegetation are commonly used in branding while remaining largely unstructured during the stay. A low-infrastructure innovation can occur when such features are converted into a clearly bounded activity with an intentional sequence, defined facilitation, safety provisions, and explicit limits. Tourism innovation may thus arise through a new configuration of existing resources and organisational knowledge rather than through physical expansion (Hjalager, 2010).

Environmental psychology and nature-health research provide reasons to expect immediate restorative responses to contact with forests, green areas, and blue environments, including changes in affect, stress, attention, and physiological activation (Ulrich, 1983; Ulrich et al., 1991; S. Kaplan, 1995; Hansen et al., 2017; Britton et al., 2020; White et al., 2020). Evidence concerning forest bathing and forest therapy is encouraging but methodologically diverse, with substantial variation in protocols, samples, and endpoints (Farrow & Washburn, 2019; Stier-Jarmer et al., 2021; Queirolo et al., 2024; Kaleta et al., 2025). For hospitality research, the additional question is how these possibilities can be translated into a credible and manageable visitor offer (Ohe et al., 2017; Farkić et al., 2021). A separate investigation of a three-day guided forest-bathing programme has likewise shown that the average immediate change and differences in guests’ starting orientations are separate issues for programme design (Racz & Armano, 2026a).

Deliberate attention offers a plausible bridge between environmental exposure and experienced restoration. Research on noticing nature indicates that actively registering natural details and one’s reactions can add value beyond the duration of exposure (Passmore & Holder, 2017; Passmore et al., 2025). In a coastal setting, attention can be directed outward and upward toward clouds, light, visual depth, motion, and the sea horizon, or nearby toward scent, sound, temperature, texture, posture, and footing. These modes use the same destination at different perceptual distances and may therefore produce related but distinguishable immediate experiences.

The current manuscript is transparently linked to the randomised parent study reported by Racz and Armano (2026b), but it answers a different question. The earlier article used the final assessment of all four conditions to evaluate a 22-item service-performance module covering organisation and quality of the experience, satisfaction, perceived innovativeness, added value, and recommendation intention. Here, the analytical sample is restricted to participants with paired observations surrounding one of the three organised sessions, and the outcomes are psychological, experiential, and basic physiological responses. No service-performance endpoint, statistical model, or principal result from the earlier publication is repeated.

Accordingly, the object of inquiry is an immediately experienced restorative episode within a hospitality setting, not a clinical intervention and not a second evaluation of service quality. The present exploratory analysis asks whether two different ways of directing attention to destination nature show more favourable response patterns than an organised control activity, whether their activity-specific ratings differ descriptively, and what those patterns imply for responsible resort programming. All inferences are confined to the activity period.

2. Theoretical Background and Theory-Guided Expectations

2.1. Restorative Experience in Wellness Tourism

Attention Restoration Theory describes how fascination, psychological distance from routine, environmental scope, and person–environment fit may replenish directed attention (R. Kaplan & Kaplan, 1989; S. Kaplan, 1995). Stress Reduction Theory instead emphasises relatively rapid affective and bodily responses to environmental cues compatible with safety (Ulrich, 1983; Ulrich et al., 1991). Within tourism, neither process occurs in isolation from the service encounter. Guests respond to the environment as it is framed by instructions, pacing, group conditions, facilitator conduct, perceived choice, and congruence with the destination.

A scenic location may therefore remain merely pleasant if the visitor is hurried, distracted, or uncertain about what to do. Conversely, a highly scripted activity may feel artificial even in an attractive setting. We conceptualise a restorative experience as an emergent property of three interacting elements: what the place affords, how the programme organises participation, and how the guest engages with the opportunity.

2.2. Nature-Directed Attention, Soft Fascination, and Experiential Distance

Cloud noticing and forest-therapy elements organised attention at different spatial and sensory scales. The theoretical distinction concerns the attentional task, not a meteorological typology. Cloud noticing favoured a wide visual field, distant focus, visible depth, slow movement, changing edges, light, and openness. The activity did not ask participants to identify cloud genera, select preferred cloud forms, or assign predetermined symbolic meanings. Evidence from landscape-perception research suggests that the proportion of visible sky and the organisation of sky–horizon information can influence preference and attentional allocation, while experimental work on upward canopy views indicates that a moderate presence of sky can support attention restoration (Lin et al., 2023; Tasser et al., 2023). These findings justify treating an open and dynamically changing sky as an attentional affordance, but they do not justify claiming that a particular cloud type caused the observed responses.

Attention Restoration Theory provides the first mechanism linking this task to restoration. Soft fascination is generated when environmental change holds attention without demanding effortful problem solving, thereby leaving room for reflection and psychological distance from routine (R. Kaplan & Kaplan, 1989; S. Kaplan, 1995). Slowly evolving clouds can meet this description because they are continuously changing yet do not require rapid response, factual classification, or goal-directed performance. The horizon and visual depth additionally enlarge the perceived scope of the scene, strengthening the compatibility between a wide environmental field and the subjective sense of mental room.

Awe theory provides a second, complementary pathway. Awe is elicited when perceived vastness challenges the current frame of reference and requires cognitive accommodation (Keltner & Haidt, 2003; Yaden et al., 2019). Experimental work links awe to expanded time perception and reduced preoccupation with the immediate self (Rudd et al., 2012; Piff et al., 2015). In the present context, mental spaciousness and perspective distancing are therefore interpreted as provisional, activity-specific descriptors of widened attentional scope and temporary decentring from immediate concerns. They are not treated as validated mechanisms, and the study did not manipulate sky extent, cloud cover, motion, brightness, or form separately.

2.3. The Role of Active Control

An organised comparison session was required because simply receiving additional programme time can change self-reporting. Contact with a facilitator, temporary relief from routine, group participation, quietness, and a clearly marked beginning and end may all be restorative. The active control reproduced these general conditions but omitted directed engagement with clouds, vegetation, multisensory forest cues, mindfulness, and reflective nature-based prompts. It therefore provided a more demanding benchmark than an untreated comparison.

2.4. Theory-Guided Expectations

Because neither the hypotheses nor the analysis plan was preregistered, the following statements are theory-guided expectations for an exploratory randomised field study rather than confirmatory hypotheses.

H1. 

Improvements in state anxiety, affect, and subjective restoration were expected to be larger in the cloud-noticing and forest-therapy-elements groups than in the active control.

H2. 

After the session, cloud noticing was expected to show the highest descriptive ratings for awe, mental spaciousness, and perspective distancing.

H3. 

After the session, forest-therapy elements were expected to show the highest descriptive ratings for sensory immersion, bodily presence, and grounding.

H4 

(Ancillary and descriptive). Resting pulse and blood pressure were expected to decline numerically more in the two nature-focused groups than in the active control; these variables were treated only as physiological field descriptors and not as evidence of biological restoration or clinical benefit.

3. Materials and Methods

3.1. Analytical Scope and Relationship to the Earlier Companion Publication

The source project was a prospective field experiment with four conditions: cloud noticing, forest-therapy elements, the active control, and standard stay-implemented during normal operations at a coastal wellness resort. Recruitment, consent, allocation, programme delivery, and data collection belonged to one approved parent protocol. The present exploratory companion paper uses only the portion of that dataset needed to examine immediate changes around the organised sessions. Appendix A,

Table A1. Direct comparison of the published companion article and the present manuscript.

Element	Published Service-Evaluation Companion	Present Exploratory Companion Analysis
Shared parent protocol	Same prospective four-condition resort experiment; N = 120 randomised 1:1:1:1.	Same parent protocol, allocation, interventions, and ethics approval; no new recruitment or intervention.
Analytical sample	114 final assessments from all four conditions.	84 linked pre-post observations from the three organised session arms: 29 cloud noticing, 28 forest-therapy elements, 27 active control.
Comparator logic	Standard stay included because every arm could complete a final service evaluation.	Standard stay excluded because it had no study-organised session and no equivalent session-level pre-post interval; active control is the structural comparator.
Timing	Final post-stay/service assessment.	Immediately before and after the organised two-hour session.
Outcome family	Experience quality, satisfaction, perceived innovativeness/novelty, added value, and recommendation intention.	State anxiety, affect, subjective restoration, perceived stress, nature connectedness, descriptive activity-specific items, and physiological field descriptors.
Analytical focus	Service-performance comparisons and predictors of recommendation intention.	Linear mixed-effects group-by-time contrasts and descriptive item-level patterns.
Unavoidable factual overlap	Setting, allocation, participant flow, intervention identity, ethics approval, and basic sample descriptors.	The same protocol facts are transparently cross-referenced because they cannot differ across reports.
Non-overlap	No psychological change models or paired session-level tables used.	No service-performance endpoint, regression model, principal table, hypothesis, or conclusion is reanalysed.
Distinct contribution	How guests evaluated the organisation and value of the service.	What changed immediately around the organised activities and whether their descriptive experiential patterns differed.

Note. The two reports share one parent protocol but address non-overlapping outcome families and analytical purposes.

, directly compares the shared protocol, analytical samples, outcome families, overlapping factual elements, and non-overlapping analyses in the earlier and present publications.

The published companion paper evaluated the final, performance-oriented service module in the full four-condition sample (Racz & Armano, 2026b). Its dependent variables were experience quality, satisfaction, perceived innovativeness, added value, and recommendation intention. By contrast, this article uses paired pre-session and post-session psychological and physiological observations together with separate experiential ratings. The two reports, therefore, share a parent protocol but not an outcome family, primary tables, hypotheses, or central analyses.

Standard-stay participants were not exposed to a study-organised session and did not have an equivalent before-and-after programme interval. Their data were appropriate for the earlier service-value comparison but not for the present question. The active-control arm, which matched the duration, group organisation, and facilitator contact of the added activities, is the relevant comparator for estimating whether explicitly nature-directed content contributed beyond participation in any structured session.

3.2. Setting, Participants, Random Allocation, and Analytic Sample

Fieldwork took place before the summer peak at an anonymised Northern Adriatic health-and-wellness property. Within the same resort, one location provided a wooded or park-forest environment, and another, situated at its edge, opened toward the sky and sea horizon. These spatial characteristics allowed the two nature-focused formats to be delivered without naming the commercial property or destination.

Adults were eligible when they could safely undertake low-intensity outdoor participation, understood the study procedures, agreed to complete questionnaires, and provided consent. Acute health problems incompatible with the activity, inability to follow elementary instructions, or refusal of consent led to exclusion. The recruited guests were a convenience sample from a wellness-oriented resort and should not be regarded as representative of tourists more generally.

Individual allocation occurred after consent and baseline measurement. A card was prepared for every participant using a temporary operational code. Co-occupants of the same room received separate A and B suffixes and entered the draw independently, so the room was never the allocation unit. The draw materials were stored apart from research responses, and the room-based codes were removed before analysis.

Cards were repeatedly mixed and selected until each of the four conditions contained 30 participants. This procedure produced equal allocation (1:1:1:1; N = 120) and did not allow choice based on preference, arrival time, prior nature or wellness experience, or facilitator expectations. Because the activities were visibly different, masking participants and guides after assignment was not feasible.

Recruitment began with 176 contacted guests. Of these, 142 showed interest, and 128 were screened; eight did not enter the study, leaving 120 randomised participants. Before the final assessment, one person left the cloud-noticing arm, two left the forest-therapy elements, and three left the active control, whereas the 30 standard-stay participants completed the final questionnaire. The paired analysis reported here, therefore, contains 29, 28, and 27 participants in the three organised arms, respectively (total n = 84) (Figure 1). Withdrawals did not require justification, and reasons were not collected (

Table 1. Participant flow in the parent study and present companion analysis.

Study Phase	Number of Participants
Contacted hotel guests	176
Expressed initial interest	142
Entered screening	128
Excluded before inclusion	8
Included and randomised	120
Initially allocated to each of four conditions	30 per condition
Completed programme—cloud noticing	29
Completed programme—forest-therapy elements	28
Completed programme—active control	27
Completed final evaluation—standard stay	30
Total final evaluation sample in parent study	114
Present matched pre-post companion sample	84
Serious adverse events	0

Note. Table 1 shows the complete parent-study flow; only participants from the three facilitated sessions contributed to the paired analyses in this article. Outcomes from the standard-stay arm are available in the earlier service-evaluation report.

).

3.3. Intervention Conditions and Standardisation

Each organised condition lasted approximately two hours and was delivered to a group by a facilitator. To keep the formats comparable, the two nature-focused sessions used the same broad progression: arrival and safety orientation, attentional settling, guided exploration, an extended period with little facilitator speech, a short invitation to articulate or reflect on what had been noticed, and closure. The content of attention—not the duration or overall structure—was the main planned difference.

Delivery occurred outdoors in stable, clear weather. Before starting, all participants received the same guidance concerning voluntary involvement, comfortable positioning, movement within agreed boundaries, and the freedom to remain seated or stand. They were also told that there was no preferred response and that speaking during the final exchange was optional. Facilitators used prepared prompts, refrained from interpreting participants’ experiences therapeutically, and made no promises of health improvement.

3.3.1. Cloud-Noticing Condition

The cloud-noticing session took place where the wooded resort grounds opened toward the sea-facing horizon. The site offered an unrestricted sky view without placing participants on the shoreline. Facilitator invitations prioritised clouds and atmospheric change; nearby trunks, bark, and vegetation were not used as the central objects of attention.

The sequence allocated approximately 10 min to orientation and eye-safety advice, 15 min to awareness of air, temperature, ambient sound, posture, and openness, and 20 min to guided observation of cloud shape, edges, colour, illumination, shadow, layering, direction, and apparent speed. Participants then observed mainly in silence for about 45 min, followed by a 15 min invitation to reflect on one atmospheric feature that had held their attention and a 15 min voluntary closing exchange. They were instructed never to look at the sun and to alter their position if sustained upward viewing became uncomfortable.

Cloud noticing concerned how clouds emerged, moved, changed, divided, and disappeared. No meteorological labelling, cloud typology, or predetermined metaphor was requested. The session translated the general principle of noticing nature into a brief cloud-focused resort activity; it was not a mindfulness course and did not reproduce the original two-week intervention protocol.

3.3.2. Forest-Therapy-Elements Condition

Forest-therapy elements were delivered inside the resort’s wooded grounds. The format assembled a limited set of practices commonly used to slow and deepen engagement with a forest environment, adapted to a short group programme for hotel guests. It was not represented as clinical care, a personalised prescription, certification-based treatment, or a complete standard method.

Timing paralleled the cloud session. About 10 min were devoted to orientation, route limits, safety, and choice; approximately 15 min to slower movement and awareness of light, colour, sounds, scents, surfaces, shade, temperature, posture, and footing; and about 20 min to exploring whichever natural element attracted attention without naming or judging it. A 45 min period of quiet movement, standing, or stillness followed, with limited speech from the facilitator. The final 30 min comprised reflection on a memorable sensation or encounter and an optional group integration.

The programme was intentionally responsive to this site, the participating adults, and the practical conditions of group delivery. Its conceptual basis was flexible, nature-oriented facilitation rather than adherence to an external licensed protocol.

3.3.3. Active-Control Condition

Active-control participants received a two-hour, facilitator-led session scheduled in the same way as the experimental activities. It combined orientation, neutral informational content, low-intensity seated or quiet participation, and a common closing procedure. The control was designed to hold constant the experience of additional programme time, staff attention, group organisation, and a period of relative quiet.

Nature-focused ingredients were deliberately absent. The facilitator did not ask participants to observe the sky, clouds, horizon, vegetation, natural acoustics, scent, shade, texture, or contact with the ground, and there was no guided sensory exploration, slow forest sequence, mindfulness exercise, reflective nature invitation, or nature-based integration. Quiet periods remained neutral components of the schedule rather than contemplative instructions.

3.4. Measures Used in This Companion Analysis

One multi-module questionnaire package served the approved parent project. Separate sections covered consent, background and wellness characteristics, repeated psychological and nature-related variables, immediate experiential ratings, physiological field descriptors, contextual comments, and the service-performance module. The last of these-22 items concerning quality and perceived service value belongs to the earlier publication and was excluded in full from the analyses presented here (Racz & Armano, 2026b).

Four repeated psychological variables were designated as principal outcomes for this report. State anxiety was measured with the state form of the State-Trait Anxiety Inventory (Spielberger, 1983), and affect with the positive and negative subscales of PANAS (Watson et al., 1988). Immediate restoration was represented by a brief contextual module grounded in perceived-restorativeness research (Hartig et al., 1997). Because this abbreviated module was developed for the field setting, it is not presented as equivalent to the complete Perceived Restorativeness Scale.

Perceived stress and connectedness with nature were secondary repeated outcomes, informed, respectively, by the Perceived Stress Scale (Cohen et al., 1983) and the Connectedness to Nature Scale (Mayer & Frantz, 2004). Six post-session single-item indicators-awe, mental spaciousness, perspective distancing, sensory immersion, bodily presence, and grounding-were included to describe theory-relevant qualities of the activities. A common 1–5 response format ranged from 1 (not at all) to 5 (very strongly). Appendix A,

Table A2. Common prompt, full item wording, and response anchors for activity-specific indicators.

Indicator	Item Wording Following the Common Prompt	Interpretive Status
Awe	The activity gave me a sense of awe or wonder.	Single descriptive item
Mental spaciousness	The activity gave me a sense of mental spaciousness or room in my thoughts.	Single descriptive item
Perspective distancing	The activity helped me gain some distance from my immediate worries or concerns.	Single descriptive item
Sensory immersion	I felt absorbed in the activity through my senses.	Single descriptive item
Bodily presence	The activity made me feel present in and aware of my body.	Single descriptive item
Grounding	The activity made me feel grounded or anchored in the present place.	Single descriptive item

Note. Common prompt: ‘Immediately after the activity, to what extent did this experience evoke each of the following?’ Response anchors were 1 = not at all, 2 = slightly, 3 = moderately, 4 = strongly, and 5 = very strongly. The items are reported individually; no composite score or internal-consistency coefficient was calculated.

, reports the common stem and complete English item wording used for reporting. Single-item indicators were selected to limit respondent burden in the immediate field assessment and because the constructs were intended as distinct descriptive qualities rather than one latent scale. Internal-consistency coefficients are therefore neither available nor conceptually appropriate. These items are not diagnostic instruments, validated scales, or transportable measures, and they are interpreted only at the item level.

Additional items documented whether instructions were understandable, participation was comfortable, duration felt suitable, the activity seemed useful, participants would consider repeating it, and the measurements felt burdensome. Step count, temperature, perceived disturbance, visibility of the sky, and sun exposure were recorded to contextualise delivery. Written comments were retained for background interpretation only and were not analysed as a qualitative dataset.

3.5. Physiological Measurements and Data Collection

Pulse and blood pressure were obtained with a validated automatic upper-arm device after participants had rested in a seated position for at least five minutes. Pre- and post-session readings followed the same field procedure in comparable locations. They were retained only as physiological field descriptors. Posture, time of day, ambient conditions, expectation, quiet sitting, device error, and regression to the mean could all contribute to change; the measurements were not clinical endpoints, mechanistic biomarkers, or evidence of biological restoration.

Anonymous codes connected each participant’s baseline and final questionnaire. The first assessment preceded the assigned session, and the second followed its completion. Neither direct identifiers nor the temporary room-related allocation codes were retained in the analytical file. Only cases with linkable observations at both time points in an organised programme arm entered the present dataset.

3.6. Statistical Analysis

The analysis was restricted to complete paired cases from the cloud-noticing, forest-therapy-elements, and active-control arms; missing outcomes were not replaced. Means and standard deviations describe continuous measures, while counts and percentages describe categorical variables. Linear mixed-effects models were used as the principal framework for the four primary repeated psychological outcomes. Each model included fixed effects for group, time, and the group-by-time interaction, with a random participant intercept for the two linked observations. The principal estimands were the interaction contrasts: cloud noticing versus active control, forest-therapy elements versus active control, and cloud noticing versus forest-therapy elements. Model contrasts are reported as unstandardised B estimates with standard errors, 95% t-based confidence intervals, t statistics with 81 residual degrees of freedom, two-sided p values, and Hedges’ g.

Hedges’ g was calculated for each between-group difference in change using the pooled change-score standard deviation and the small-sample J correction; its 95% confidence interval used the conventional sampling-variance approximation. The direction of g follows the favourable response: reductions in anxiety and negative affect were sign-reversed in Figure 2 so that positive values consistently favour the first-named nature-focused condition. Residual quantile-quantile and residual-versus-fitted plots were the prespecified diagnostic checks for marked non-normality and heteroscedasticity; given the modest sample, these checks were interpreted descriptively rather than as proof of assumptions. As a robustness check, the direction and magnitude of model-based contrasts were compared with the original unadjusted difference-in-change estimates; conclusions were unchanged because both approaches used the same complete two-wave observations.

The three planned contrasts were corrected within each primary outcome using Holm’s sequential procedure; raw and Holm-adjusted p values are reported together. Secondary repeated outcomes, activity-specific ratings, acceptability, movement, environmental variables, and physiological field descriptors remained estimation-focused and descriptive because they were not preregistered and included several study-specific single items. The available recruitment window and equal four-arm allocation determined the parent sample; no prospective power calculation was performed. With 84 paired completers, small effects cannot be excluded. Complete-case analysis is valid only conditional on observed completers, and missing-completely-at-random cannot be established because withdrawal reasons were not collected. Neither the study nor the theory-guided expectations were registered in advance; interpretation, therefore, prioritises effect sizes, interval width, and consistency across related outcomes.

3.7. Ethical Considerations

The FoRRestMind Institute Ethics Committee approved the complete parent protocol (004-2026; 15 March 2026), and the work followed the Declaration of Helsinki. The ethics review and approval decision were conducted separately from the research team. Authors affiliated with the Institute did not participate in the Committee’s deliberation, voting, or approval decision for this protocol; their institutional roles were disclosed, and recusal constituted the conflict-management procedure. Before enrolment, guests received information about the purpose and procedures, handling of data, freedom to skip questions, and the right to stop without penalty or explanation. Written consent preceded participation. The reviewed materials comprised the protocol, participant information and consent documents, questionnaires, anonymisation procedures, and outdoor-safety arrangements. Analysis used anonymised records, and no serious adverse events occurred.

4. Results

4.1. Participant Characteristics

Paired observations were available for 84 participants: 29 after cloud noticing, 28 after forest-therapy elements, and 27 after active control. Loss to follow-up was small but confined to the organised conditions.

Table 2. Baseline demographic and tourism-profile characteristics in the three completed programme arms.

Characteristic	Cloud Noticing (n = 29)	Forest-Therapy Elements (n = 28)	Active Control (n = 27)
Age, M (SD)	47.9 (10.7)	49.2 (11.4)	48.7 (10.8)
Women, n (%)	18 (62.1)	17 (60.7)	17 (63.0)
Men, n (%)	11 (37.9)	11 (39.3)	10 (37.0)
Higher education, n (%)	20 (69.0)	19 (67.9)	17 (63.0)
Previous experience with wellness practices, n (%)	12 (41.4)	11 (39.3)	10 (37.0)
Previous experience with guided nature activities, n (%)	9 (31.0)	10 (35.7)	8 (29.6)
Time in nature at least once per week, n (%)	16 (55.2)	15 (53.6)	14 (51.9)
Main reason for stay: rest/wellness, n (%)	23 (79.3)	22 (78.6)	21 (77.8)
Expected usefulness of programme, M (SD)	4.1 (0.6)	4.0 (0.7)	3.8 (0.7)

Note. Entries describe completers in the three facilitated arms. These factual characteristics necessarily correspond to the same participants reported in the parent service-evaluation article; no service-evaluation outcome is included.

reproduces only the basic descriptors required to characterise these completers and to maintain consistency with the previously disclosed parent sample. All psychological values used in the current analysis, including baseline levels, appear with their post-session counterparts in

Table 3. Changes in primary psychological outcomes from baseline to the end of the programme.

Outcome	Group	Baseline M (SD)	Final M (SD)	Mean Change	Standardised Change d
STAI-S	Cloud noticing (n = 29)	42.1 (8.7)	34.8 (7.9)	−7.3	0.88
STAI-S	Forest-therapy elements (n = 28)	41.8 (8.4)	35.6 (8.1)	−6.2	0.75
STAI-S	Active control (n = 27)	42.5 (8.9)	39.4 (8.6)	−3.1	0.35
PANAS positive affect	Cloud noticing (n = 29)	28.4 (6.2)	34.2 (6.5)	+5.8	0.91
PANAS positive affect	Forest-therapy elements (n = 28)	27.9 (6.5)	32.8 (6.1)	+4.9	0.78
PANAS positive affect	Active control (n = 27)	28.1 (6.3)	30.0 (6.4)	+1.9	0.30
PANAS negative affect	Cloud noticing (n = 29)	20.6 (5.8)	15.4 (5.1)	−5.2	0.95
PANAS negative affect	Forest-therapy elements (n = 28)	21.1 (5.6)	16.3 (5.2)	−4.8	0.89
PANAS negative affect	Active control (n = 27)	20.8 (5.7)	18.9 (5.5)	−1.9	0.34
Subjective restoration	Cloud noticing (n = 29)	3.1 (0.8)	4.3 (0.6)	+1.2	1.67
Subjective restoration	Forest-therapy elements (n = 28)	3.0 (0.7)	4.4 (0.5)	+1.4	2.30
Subjective restoration	Active control (n = 27)	3.0 (0.8)	3.5 (0.7)	+0.5	0.67

Note. Subjective restoration used a 1–5 response range. Within-group standardised values are descriptive summaries, and positive values denote movement in the favourable direction.

.

4.2. Primary Psychological Outcomes

Linear mixed-effects models indicated omnibus group-by-time effects for subjective restoration, F(2.81) = 14.92, p < 0.001; positive affect, F(2.81) = 4.49, p = 0.014; negative affect, F(2.81) = 4.15, p = 0.019; and state anxiety, F(2.81) = 3.12, p = 0.049. Descriptively, state-anxiety scores fell by 7.3 points after cloud noticing, 6.2 after forest-therapy elements, and 3.1 after the active control. Positive affect rose by 5.8, 4.9, and 1.9 points, respectively; negative affect declined by 5.2, 4.8, and 1.9 points. The largest restoration gain followed the forest-focused session, closely followed by cloud noticing (Table 3).

Model-based contrasts were clearest for subjective restoration (

Table 4. Linear mixed-effects model contrasts for primary psychological outcomes.

Outcome	Planned Contrast	B (SE), 95% CI	t(81); Raw p	Holm p	Hedges’ g
STAI-S	Cloud noticing vs. active control	−4.20 (1.74), −7.65 to −0.75	−2.42; 0.018	0.053	0.50
STAI-S	Forest-therapy elements vs. active control	−3.10 (1.66), −6.41 to 0.21	−1.86; 0.067	0.133	0.38
STAI-S	Cloud noticing vs. forest-therapy elements	−1.10 (1.64), −4.36 to 2.16	−0.67; 0.505	0.505	0.13
PANAS positive affect	Cloud noticing vs. active control	3.90 (1.34), 1.24 to 6.56	2.92; 0.005	0.014	0.61
PANAS positive affect	Forest-therapy elements vs. active control	3.00 (1.30), 0.41 to 5.59	2.30; 0.024	0.048	0.47
PANAS positive affect	Cloud noticing vs. forest-therapy elements	0.90 (1.25), −1.58 to 3.38	0.72; 0.474	0.474	0.14
PANAS negative affect	Cloud noticing vs. active control	−3.30 (1.17), −5.63 to −0.97	−2.82; 0.006	0.018	0.59
PANAS negative affect	Forest-therapy elements vs. active control	−2.90 (1.15), −5.19 to −0.61	−2.52; 0.014	0.027	0.52
PANAS negative affect	Cloud noticing vs. forest-therapy elements	−0.40 (1.11), −2.60 to 1.80	−0.36; 0.720	0.720	0.07
Subjective restoration	Cloud noticing vs. active control	0.70 (0.17), 0.37 to 1.03	4.24; <0.001	<0.001	0.94
Subjective restoration	Forest-therapy elements vs. active control	0.90 (0.18), 0.55 to 1.25	5.11; <0.001	<0.001	1.20
Subjective restoration	Cloud noticing vs. forest-therapy elements	−0.20 (0.14), −0.48 to 0.08	−1.43; 0.157	0.157	0.28

Note. B is the group-by-time interaction contrast. Holm p values correct the three planned contrasts within each outcome. For anxiety and negative affect, negative B values favour the first-named group; Hedges’ g is shown as an absolute favourable-direction magnitude. Confidence intervals are unadjusted 95% intervals. CI = confidence interval; SE = standard error.

and Figure 2). After Holm correction within each outcome, both nature-focused conditions retained contrasts with the active control for positive affect, negative affect, and restoration. The cloud-noticing contrast for state anxiety narrowly crossed the adjusted 0.05 threshold (raw p = 0.018; Holm p = 0.053), and the forest-therapy contrast for anxiety remained uncertain. Direct contrasts between the two nature-focused formats were small, and unadjusted and adjusted intervals were compatible with no difference. The stronger inference is therefore a shared advantage over active control on several psychological outcomes, not overall superiority of one nature-focused format.

4.3. Descriptive Experiential Patterns

The six activity-specific ratings provided the clearest descriptive distinction between the two nature-focused sessions (

Table 5. Post-programme experiential indicators by condition.

Experiential Indicator	Cloud Noticing M (SD)	Forest-Therapy Elements M (SD)	Active Control M (SD)
Awe	4.4 (0.6)	3.7 (0.7)	2.9 (0.8)
Mental spaciousness	4.5 (0.5)	3.8 (0.7)	3.0 (0.8)
Perspective distancing	4.2 (0.7)	3.6 (0.8)	2.8 (0.8)
Sensory immersion	3.8 (0.7)	4.5 (0.5)	3.1 (0.8)
Bodily presence	3.7 (0.8)	4.4 (0.6)	3.0 (0.8)
Grounding	3.6 (0.8)	4.5 (0.5)	3.1 (0.7)

Note. Ratings were obtained after the session on five-point scales. They are concise descriptors created for this study and should not be interpreted as diagnostic or psychometrically validated scales.

). Cloud noticing had the highest means for awe, mental spaciousness, and perspective distancing, whereas forest-therapy elements had the highest means for sensory immersion, bodily presence, and grounding. Active control had lower means on every indicator. Because these are separate study-specific single items, the findings are descriptive and do not establish validated experiential patterns, latent mechanisms, or stable signatures.

The pattern is compatible with atmospheric and distal attention during cloud noticing and proximal, embodied attention during forest-focused engagement. It should not be interpreted as evidence that specific cloud types produced specific psychological effects: cloud morphology was neither classified nor manipulated. The programme indicators require replication with validated multi-item measures and observed guest choices.

4.4. Perceived Stress, Nature Connectedness, and Physiological Field Descriptors

Stress scores decreased in every arm, with larger numerical reductions after cloud noticing and forest-therapy elements. Nature connectedness changed little in the active control, increased after cloud noticing, and increased most after the forest-focused session. These secondary outcomes are presented as estimates without confirmatory multiplicity claims (

Table 6. Changes in secondary self-reported outcomes.

Outcome	Group	Baseline M (SD)	Final M (SD)	Mean Change	Standardised Change d
PSS	Cloud noticing (n = 29)	19.8 (5.4)	16.7 (5.2)	−3.1	0.58
PSS	Forest-therapy elements (n = 28)	20.2 (5.7)	17.1 (5.4)	−3.1	0.56
PSS	Active control (n = 27)	19.5 (5.1)	18.2 (5.0)	−1.3	0.26
Nature connectedness	Cloud noticing (n = 29)	3.4 (0.7)	3.8 (0.7)	+0.4	0.57
Nature connectedness	Forest-therapy elements (n = 28)	3.3 (0.7)	3.9 (0.6)	+0.6	0.92
Nature connectedness	Active control (n = 27)	3.4 (0.6)	3.5 (0.7)	+0.1	0.15

Note. PSS = Perceived Stress Scale. Nature connectedness was scored from 1 to 5. Standardised values describe within-group movement in a favourable direction.

).

Mean pulse and blood-pressure values decreased from baseline in each condition. The two nature-focused arms showed somewhat larger numerical reductions than the active control, but no clinical or biological interpretation is warranted. These values are reported solely as physiological field descriptors because posture, timing, weather, expectation, quiet sitting, measurement error, and regression to the mean could explain part or all of the observed change (

Table 7. Changes in physiological field descriptors.

Outcome	Group	Baseline M (SD)	Final M (SD)	Mean Change	Standardised Change d
Resting heart rate	Cloud noticing (n = 29)	76.2 (9.1)	71.4 (8.5)	−4.8	0.55
Resting heart rate	Forest-therapy elements (n = 28)	75.8 (8.7)	71.5 (8.2)	−4.3	0.51
Resting heart rate	Active control (n = 27)	76.5 (8.9)	74.6 (8.6)	−1.9	0.22
Systolic blood pressure	Cloud noticing (n = 29)	128.4 (12.8)	122.8 (11.9)	−5.6	0.45
Systolic blood pressure	Forest-therapy elements (n = 28)	127.9 (13.1)	122.8 (12.4)	−5.1	0.40
Systolic blood pressure	Active control (n = 27)	128.1 (12.6)	126.0 (12.2)	−2.1	0.17
Diastolic blood pressure	Cloud noticing (n = 29)	81.6 (8.4)	78.2 (7.9)	−3.4	0.42
Diastolic blood pressure	Forest-therapy elements (n = 28)	81.9 (8.1)	78.7 (7.8)	−3.2	0.40
Diastolic blood pressure	Active control (n = 27)	82.0 (8.3)	80.7 (8.1)	−1.3	0.16

Note. Pulse is reported in beats per minute and blood pressure in mmHg. Values are descriptive field observations only. They are not clinical endpoints, mechanistic biomarkers, or evidence of biological restoration.

).

4.5. Environmental Conditions and Acceptability

Participants moved relatively little in all three sessions. Recorded temperatures and perceived disturbance were similar across arms; the cloud-noticing site provided the greatest sky visibility, and the forest-focused activity involved slightly more slow movement. These observations assist interpretation of implementation but cannot eliminate every environmental or behavioural alternative explanation (

Table 8. Movement and environmental conditions during the programme.

Indicator	Cloud Noticing M (SD)	Forest-Therapy Elements M (SD)	Active Control M (SD)
Steps during programme	312 (95)	428 (120)	295 (88)
Air temperature, °C	22.4 (1.2)	22.1 (1.3)	22.3 (1.1)
Subjective noise/disturbance, 1–5	1.6 (0.6)	1.7 (0.7)	1.8 (0.7)
Sky visibility, 1–5	4.6 (0.5)	3.2 (0.8)	3.4 (0.7)
Sun exposure, 1–5	2.8 (0.7)	2.5 (0.6)	2.6 (0.7)

Note. Step counts refer to low-intensity movement during the approximately two-hour session. Environmental observations are implementation descriptors rather than independently manipulated exposures.

).

Immediate acceptability was positive throughout (

Table 9. Immediate acceptability of the three programme conditions.

Indicator	Cloud Noticing M (SD)	Forest-Therapy Elements M (SD)	Active Control M (SD)
Clarity of instructions	4.6 (0.5)	4.7 (0.5)	4.4 (0.6)
Comfort during participation	4.5 (0.6)	4.7 (0.5)	3.8 (0.8)
Appropriateness of duration	4.3 (0.7)	4.4 (0.6)	4.0 (0.7)
Subjective usefulness	4.4 (0.6)	4.5 (0.6)	3.4 (0.8)
Willingness to participate again	4.5 (0.6)	4.6 (0.5)	3.5 (0.9)
Measurement burden	2.0 (0.8)	2.1 (0.7)	2.2 (0.8)

Note. Items used five-point response scales; higher scores on measurement burden indicate a less favourable experience.

). Compared with the active control, both nature-focused sessions received higher mean ratings for usefulness and willingness to take part again. Comfort was highest after forest-therapy elements, while cloud noticing retained the clearest cloud-related item-level pattern. The perceived measurement burden was low and comparable. These responses are immediate intentions and appraisals, not evidence of purchasing, repeat attendance, or market demand.

5. Discussion

5.1. Principal Findings

This exploratory companion report focused on paired immediate responses in the three facilitator-led conditions of the randomised parent experiment. Model-based contrasts indicated more favourable changes after cloud noticing and forest-therapy elements than after the active control for several primary psychological outcomes, with the strongest evidence for subjective restoration. Holm-adjusted results remained favourable for positive affect, negative affect, and restoration, whereas the adjusted cloud-noticing contrast for state anxiety was borderline. Direct comparisons between the two nature-focused formats remained small and uncertain.

The formats differed descriptively in experiential character. Cloud noticing was associated with awe, mental spaciousness, and temporary distance from immediate concerns; attention within the wooded area was associated with sensory absorption, bodily awareness, grounding, and the largest increase in nature connectedness. These item-level patterns are compatible with different attentional orientations, but they do not establish distinct psychometric profiles or causal mechanisms.

Some improvement also occurred in active control. This is not a nuisance finding: a scheduled pause, staff attention, group containment, and removal from ordinary routine may themselves support short-term relief. Larger changes in the nature-focused arms are compatible with an added contribution from the content and direction of attention, but the design cannot isolate each element of expectation, facilitation, social context, or exposure.

5.2. Implications for Tourism Experience and Resort Service Design

From a management perspective, the study treats destination nature as material for service design rather than as a therapeutic claim. Clouds, trees, sounds, and surfaces do not need to be labelled intrinsically curative. A resort can instead offer a time-limited, voluntary, and safely facilitated encounter with an intelligible sequence and clear boundaries between wellness programming and healthcare. This reading complements the separate three-day forest-bathing study, in which programme response and guests’ initial orientations were examined as distinct considerations for hospitality delivery (Racz & Armano, 2026a).

Place fit is essential. A property with an unobstructed sky and horizon can credibly support cloud noticing, while accessible wooded ground can support slower, multisensory exploration. The theoretical rationale concerns openness, visual depth, low-demand dynamic change, and the direction of attention-not a typology of cloud forms. Other destinations-urban hotels, islands, mountains, gardens, terraces, or waterfronts-would need to build experiences from their own affordances. The generalisable proposition is a design process: identify authentic environmental resources and organise a suitable form of guest attention around them.

The contrasting item-level patterns should not be converted into fixed customer types. Individual guests may value either format, both, or neither, and their choice may depend on interest, mobility, available time, weather, and accessibility. A cautious service model would describe each module, allow elective or sequential participation, and then evaluate actual selection, completion, willingness to pay, repeat uptake, and later memory or recommendation.

Facilitation remains an operational competency rather than a script-reading task. Guides must protect physical and psychological safety, regulate pace and silence, preserve voluntary sharing, adapt to changing weather and group needs, and avoid implying medical benefit. Protocols, training, withdrawal procedures, and contemporaneous documentation of delivery conditions are therefore central elements of quality assurance.

5.3. Relationship to the Published Companion Evidence

The two articles derived from the parent experiment have deliberately non-overlapping analytical purposes. The prior report used 114 final assessments from all four conditions to study perceived service performance and value: experience quality, satisfaction, novelty, added value, and recommendation intention (Racz & Armano, 2026b). This report uses 84 paired observations from the three facilitated conditions to estimate changes in psychological, experiential, and basic physiological variables. It does not reuse the earlier dependent variables, principal tables, hypothesis tests, or conclusions.

This distinction matters scientifically, as well as procedurally. Guests may consider an activity well organised and worth recommending without exhibiting measurable short-term restoration; equally, a transient psychological response does not establish that a programme is attractive, feasible, or commercially sustainable. Taken together, the publications describe separate dimensions of the same service encounter, while neither supports claims about durable health, observed behaviour, costs, profitability, or destination competitiveness.

Shared factual elements—setting, recruitment flow, allocation method, programme identity, ethical approval, and basic participant characteristics—are disclosed consistently and cross-referenced because they belong to one parent protocol. Repetition has otherwise been minimised, and the current manuscript is self-contained only to the extent necessary for readers to understand and reproduce the present analysis.

5.4. Limitations and Future Research

Interpretation is constrained by several features. The research involved one resort and a self-selected, wellness-oriented guest population, so results may differ across destinations, cultures, seasons, accommodation categories, age distributions, and levels of prior interest in nature. Neither participants nor facilitators were masked, leaving expectancy and demand characteristics as plausible influences. Although the expected usefulness was recorded, it was not experimentally controlled.

The standard-stay arm could not contribute to the session-level paired analysis because it had neither an organised study activity nor a matching programme interval. Active control is therefore informative about nature-specific content within an added session, but the study does not estimate ordinary change during the stay. Attrition was low overall, yet occurred unevenly across the three programme arms, and withdrawal reasons were unavailable.

Several restoration and experiential items were deliberately brief adaptations for the resort context. The six activity-specific indicators were single items, so internal-consistency reliability, factor structure, invariance, and temporal stability cannot be inferred. Their wording and anchors are now reported in Appendix A, but psychometric validation is still required before use as scales in other destinations. Physiological measurements were equally limited: only resting pulse and blood pressure were collected under pragmatic field conditions, without blinding or a clinical protocol. They are therefore physiological field descriptors, not biomarkers of restoration.

The sample size reflected a finite recruitment opportunity rather than prospective power estimation. Although linear mixed-effects models provide an appropriate group-by-time analysis for the repeated observations, the complete-case sample was not adequate for small-effect exclusion, mediation, extensive subgroup modelling, or a richly adjusted covariate model. Missing values were not imputed, withdrawal reasons were unavailable, and the missing-completely at-random assumption cannot be verified. The absence of prospective registration requires an estimation-focused and exploratory reading. Finally, all measurements were immediate; the study cannot address durability, clinical importance, purchase behaviour, return visits, price tolerance, or financial return.

Future studies should preregister multi-property protocols, use established experiential instruments, include delayed assessments and seasonal replication, and measure actual tourism decisions alongside implementation costs. Within-person crossover work could test whether guests’ experience sky- and forest-focused activities as complements. Further research should also examine accessibility, cultural meaning, group size, facilitator expertise, weather, and landscape morphology as determinants of fidelity and response.

6. Conclusions

In this exploratory paired analysis of three facilitated arms from a four-condition resort experiment, cloud noticing and forest-therapy elements showed more favourable immediate psychological patterns than an organised active control. Neither nature-focused format was generally superior. Descriptively, cloud noticing emphasised awe, mental spaciousness, and perspective distancing, while the forest-based session emphasised sensory immersion, bodily presence, grounding, and connection with nature. These single-item patterns are provisional and do not constitute validated experiential patterns.

For hospitality practice, the findings show how existing environmental affordances can be converted into structured restorative programming through careful attention to sequence, safety, choice, pace, and facilitation. The innovation lies in the design of the encounter, not in declaring natural features to be treatments. Replication across properties, seasons, and behavioural endpoints is necessary before drawing conclusions about lasting wellbeing or commercial and destination-level outcomes.