Investigating the impact of indoor wood element combinations on human subjective thermal perception in cold region using virtual reality technology

Wooden materials can influence the human body’s thermal perception of indoor environments through visual stimuli, particularly in cold regions. Optimizing the allocation of indoor wood elements has the potential to conserve resources, reduce emissions, and enhance quality of life. This study investigated how different combinations of indoor wood elements affect human thermal perception in cold regions by using virtual reality to simulate various living room environments. The participants experienced a living room environment with different wood element configurations and reported their subjective thermal perceptions. The results revealed that the coverage and natural color of wood significantly affect thermal perception, with natural color having the most substantial influence. In contrast to conventional beliefs, increasing wood coverage does not consistently increase warmth. Notably, yellow-white wood tones evoke a greater sense of warmth than other warm tones do, challenging established views on warm tone applications. These findings offer new perspectives on the use of wood in indoor environmental design.

In cold regions, long heating seasons demand efficient thermal management, making materials such as wood a priority for their insulating properties, energy-saving potential, and effective influence on thermal perception [1]. Early efforts to understand indoor thermal perception involved adjusting key physical parameters influencing the thermal environment, including the air temperature, mean radiant temperature, relative humidity and air velocity [2,3,4]. Recently, however, the scope has broadened to include nonthermal factors—elements that affect a person’s perception of temperature but are not directly related to the physical measurement of heat transfer. These factors include demographic characteristics, emotional state, and perceived interactions [5]. These nonthermal factors have been shown to affect people’s thermal perceptions, contributing positively to energy conservation and emission reduction [6, 7]. Among these factors, visual elements, such as color and natural landscapes have been identified as particularly influential, as vision plays a key role in human perceptions of warmth and cold, according to theoretical models such as ecological psychology and psychophysics [8, 9].

Visual perception plays a significant role in shaping indoor environmental sensation and is a key factor influencing overall indoor thermal experience [10, 11]. Even small changes in interior design, such as the choice of color or the introduction of natural textures, can result in energy savings by broadening the acceptable range of indoor temperatures, thereby reducing overall building energy consumption [12, 13]. The mechanism by which visual effects influence thermal perception is complex and involves cognitive and sensory processes. Among various visual elements, wood stands out as a particularly intriguing material. Its natural textures, warm tones, and inherent esthetic appeal not only enhance the visual quality of a space but also, as a commonly used building material and interior design element, contribute to energy efficiency and psychological comfort. These qualities make it valuable to explore how wood influences indoor thermal perception, bridging theoretical exploration and real-world applications [14].

Wood affects visual thermal perception

The visual elements of wood can influence human perception of the thermal environment, which is associated with the ability of the material to evoke human biophilia [15]. "Biophilia" is the human instinct used to approach the natural world [16]. Wood scenes evoke a subjective sense of nature and life, enhancing people’s satisfaction with the environment [7]. Previous studies have examined the relationship between the proportion of wood in interiors and human responses, revealing various psychological and physiological experiences such as comfort and warmth induced by different levels of wood coverage [17, 18]. Subsequent findings revealed an inverted U-shaped pattern, where the feeling of "comfort" initially increased with more wood, peaked, and then gradually declined [19, 20]. This pattern aligns with later observations, indicating that environments with moderate wood coverage are considered the most comfortable, but it is still unclear whether this variation is also reflected in the perception of warmth [18]. Sungeun’s recent work established a relationship between wood coverage and thermal perception, and concluded that as wood coverage increases, people’s perceptions of the indoor thermal value increase [21]. This linear relationship differs from the aforementioned U-shaped pattern observed for comfort perception, suggesting a need to further clarify the exact relationship between wood coverage and thermal perception.

Natural color refers to the original color of the wood itself, which, unlike artificial color, is understood to influence the perception of the indoor thermal environment. Natural color is generally believed to be related to the hue–heat hypothesis. Benndorf, R. noted that natural wood color is commonly perceived as organic and warm [22]. Subsequent investigations revealed that inside warmth is related to the inherent color of the wood, as the natural appearance of the original material conveys a sense of warmth. The natural red hue of some wood types, in particular, is commonly considered the warmest color, further reinforcing this effect [19, 23]. Wood with natural colors is considered warmer, not only because the natural colors tend to fall within the warm spectrum, ranging from yellow to red but also because they convey the naturalness of wood [24, 25]. However, considering that many natural color classifications of wood exist, most of which are warm colors, further refining the perceptual differences influenced by various primary colors is important.

In addition to wood visual elements such as coverage and color, wood species, texture, and application location are considered important attributes that affect the visual perception and psychological impression of an indoor wood space, and their impacts on thermal perception need to be further explored [20, 26].

Subjective evaluation of visual thermal perception

Although physiological monitoring, which measures indices such as heart rate, skin temperature, and sweat rate, is considered an accurate way to capture human thermal perception through objective data [27, 28], these physiological changes often require sustained temperature stimulation in a dynamically changing visual environment, as they depend mainly on the body’s response to actual physical stimuli [29]. Specifically, temperature changes directly impact the body’s thermoregulatory mechanisms, leading to physiological responses such as an increased heart rate, vasoconstriction or vasodilation of the skin’s blood vessels, and increased sweating. However, visual stimuli typically cannot directly alter the body’s thermal balance, resulting in only subtle changes in physiological indices [30, 31]. Unfortunately, subjective evaluation is extremely sensitive to weak changes. The multidimensional setting of subjective evaluation can not only reveal how people perceive the thermal environment but also determine their acceptance of the environment and temperature preference. Personal attributes, such as clothing or health status, and environmental factors, such as humidity or air velocity, are thoroughly considered [32]. Finally, the multidimensional questionnaire attributes also help reveal potential inconsistencies between evaluation dimensions. These contradictions are observed and further explored through subjective assessment [33].

Virtual reality and thermal perception

In visual perception, constructing complex or changing scenes can be challenging. Virtual reality (VR) technology offers an innovative solution by enabling the rapid, repeatable, and cost-effective creation of scenarios that are challenging to replicate in the real world. By offering precise control over experimental variables, VR enables the replication and observation of complex scenarios that are nearly impossible to measure in physical environments [34].

Furthermore, VR influences indoor environmental perception and adaptation [35, 36], creating new opportunities for exploring indoor thermal perception. VR allows for the manipulation of individual factors independently, which helps in identifying how specific design elements—such as lighting, textures, and spatial arrangements—affect thermal perception [37, 38]. Previous studies have demonstrated that the human body can respond to changes in indoor environments as perceived in VR, replicating the physiological and psychological responses found in real-world conditions [36].

However, despite these promising findings, significant questions remain about the applicability of VR in thermal perception studies. One major concern is whether traditional subjective evaluation methods used for real environments are appropriate for assessing thermal perception changes in VR settings. Additionally, the use of head-mounted devices in VR may cause discomfort for participants, potentially influencing their perception and overall experience [39]. Therefore, the validity and reliability of VR-based thermal perception models need to be further explored and refined, focusing on developing a robust VR evaluation model that accurately captures the complexity of human thermal perception in virtual settings.

This study is based on validating the applicability of the subjective evaluation system for indoor human thermal environment perception in VR settings. Taking Harbin, China, as a typical example, it further examines the effects of different design combinations of indoor wooden wall surfaces on human subjective thermal perception within the temperature range constrained by winter heating. This study aims to refine indoor wooden design elements that affect human thermal perception and determine the influence of various collocation methods and the degree of allocation of these elements in the context of typical cold-region indoor environments.

Research area

Harbin, the capital of Heilongjiang Province in China (125°42′–130°10′E, 44°04′–46°40′N; altitude: 180–200 m), lies within a severe cold climate zone. The area experiences extended periods of 4–6 months each year, during which the average temperature remains below 0 °C, with extremely low temperatures dropping below − 30 °C. Despite the widespread implementation of winter heating systems, issues such as inadequate indoor heating or extreme heating conditions continue to pose significant challenges in maintaining stable and comfortable indoor thermal environments. This makes Harbin a relevant case for studying the indoor thermal conditions in cold regions.

The experimental research framework is illustrated in Fig. 1.

Calibration of thermal perception assessment in a virtual environment

To effectively study subjective thermal perception via VR equipment, it is essential to validate whether the selected VR devices can evoke a sufficient sense of presence [40], ensuring that participants’ evaluations of the thermal environment in the virtual setting are reliable.

As shown in Table 1, a preliminary experiment was conducted. The process, equipment selection, and other related requirements were the same as those in the following formal experiments. The study scenarios include both real and virtual environments featuring identical wooden and nonwooden wall environments. The primary objectives of this step were twofold: (1) to verify whether thermal perception trends in response to scene changes are consistent between VR and real environments and (2) to validate the appropriateness of using the same subjective thermal perception evaluation method in both settings.

Selection of physical variables and levels

Focus group discussions aid in identifying the wooden visual elements that affect the thermal perception of indoor spaces. The literature indicates that organizing focus group discussions with 6–12 people is effective [41]. The selection of focus group participants and process design were based on Jing’s experimental organization [20]. The focus group protocol is provided in Appendix 1. The moderator led the discussion of the prepared topics. Using the focus group results, we first identified the most important wood wall elements that affect people’s thermal perceptions: wood coverage, natural wood color, and wood layout. The effect of wood coverage is noticeable only when the degree of coverage considerably changes, as such changes must create a substantial visual difference to be consciously perceived. Generally, the proportion of this change is between 15 and 25% [18]. Natural wood color is preferred over artificial color matching, and an out-market survey shows that the current market mainly offers decorative wood-based colors: yellow-white (balsam fir, spruce), reddish-brown (red cedar, Pinus koraiensis), and brown (black walnut, elm). On the basis of a summary of 367 living room examples, which were collected through field surveys and online databases representing a regional context (Harbin and surrounding areas in Northeast China), we considered design controls when defining our experimental variables to mitigate the impact of excessive factor levels on the experimental results and ultimately to determine the levels of influencing factors, as shown in Table 2.

Setting of the experimental scene

The sample living room used in this study measures 5.5 m in width, 6.5 m in depth, and 2.8 m in height and was selected as a representative living room in Northeast China. The window area was carefully controlled to maintain a window-to-floor area ratio below 1:7, ensuring that the influence of window size or natural lighting on the experimental results was minimized [42]. Only essential furniture, such as sofas, tables, chairs, and lighting fixtures, was arranged in the living room. An indoor-only simple furniture layout can mitigate the influence of nonwooden design elements on the perceptual results.

The standard scene was modeled via Google SketchUp software. We avoided creating an overly dark interior by adjusting the sun angle to allow sunlight to illuminate the room, and no other light sources were placed in the room to avoid the impact of the interior color temperature on the experimental results, among other factors, and to regulate environmental parameters such as space brightness and contrast. Finally, the test scene enabled roaming animation for subsequent VR scene experiments.

Because there were too many test models composed of different variables, we applied the orthogonal method to effectively reduce the sample size of the study and the number of experimental bases [20]. In this study, 24 visual scenes (4A ×ばつ 3B ×ばつ 2C) were designed for our comprehensive experiments. For the orthogonal experiments, only 12 scenes were needed to obtain representative spatial sample data, and the experimental results are shown in Table 3.

The control experiments included a white nonwooden wall, meaning that the wood coverage was 0% (test 1), a yellow-white wooden wall (test 2) and a yellow-white solid paint wall (test 3) with the same coverage of 100%, as shown in Table 4. Tests 2 and 3 were compared to determine whether the thermal perception effects of yellow-white wood stem from the material’s color properties as a visual attribute or simply from the influence of color itself.

Questionnaire

The questionnaire used in this study addressed the relationship between thermal perception and wood visual elements. The verbal anchors were chosen according to ISO 10551 [43], and the questionnaire contained a fixed scale to indicate the thermal states (Table 5). The thermal sensation and preference were evaluated with a 7-point scale ending with the choices "Cold" and "Hot" and "much cooler" and "much warmer", respectively. A 4-point scale was used to evaluate acceptance, and a 5-point scale was used to evaluate comfort [44].

Experimental environment and equipment

The experiment took place within the environmental simulation cabin to ensure that the required environmental conditions were met. The cabin is 4.2 m ×ばつ 4.0 m ×ばつ 3.0 m, and the laboratory and its internal layout are shown in Fig. 2.

In terms of comfort in the range of internal temperatures, the acceptable operative temperature range for 80% of the residents in Harbin in winter is 18.0–25.0 °C [4], where indoor environments, particularly living rooms, are typically heated with central heating and can be classified within Categories II and III according to ISO 17772-1 standards [45]. Considering the characteristics of winter heating in Harbin, this study aims to explore the thermal perception of participants under extreme heating conditions, thereby providing a more comprehensive understanding of indoor comfort across different temperature conditions. Additionally, the inclusion of 27 °C as a setting is justified for examining potential variations in thermal perception under extreme heating scenarios [46]. To ensure precise temperature control, we selected four uniform temperature gradients: 18 °C, 21 °C, 24 °C, and 27 °C. Table 6 shows the instruments used to collect the indoor environmental variables of the environmental simulation cabin. These variables included the indoor air temperature, relative humidity, mean radiant temperature, and air velocity. The instruments were positioned at the front, center, and rear of the cabin, with sensors for radiant temperature and air velocity specifically placed in the center. These sensor placements were determined on the basis of the ASHRAE 55–2020 [47] and ISO 7726:2002 [48] guidelines, ensuring that the sensors were situated 0.6 m above the floor, at least 1.0 m away from nearby surfaces, and sufficiently far from any heating or cooling sources. All the parameters were recorded at 1-min intervals from the start to the end of each subject group’s session.

Considering that this study primarily investigates the impact of temperature variation on thermal perception, other variables need to be controlled within the requirements specified by ISO 17772-1 [45]. Humidity was maintained at a relatively comfortable level of 40% (the standards require the indoor humidity to be controlled between 30 and 50%). Because the interior space created by the experimental booth is enclosed, the air velocity at the central measuring point in the laboratory was controlled to be less than 0.05 m/s. Additionally, the indoor climate was essentially unaffected by the heat generated by the lighting system, and the black globe temperature and illumination remained stable.

An HTC VIVE Pro 2 VR headset was used, featuring 5 K resolution (4896 ×ばつ 2448 binocular), a 120° field of view, and a refresh rate up to 120 Hz, ensuring high-definition scene presentation and an immersive participant experience.

Participants

Thirty participants (15 men and 15 women) were selected for the experiment. The participants were college students without visual impairments, including color blindness or eye diseases. All the participants had lived in Harbin for more than 2 years to ensure consistency in their perceptions of the indoor thermal environment under cold region conditions. In terms of demography, a relatively homogeneous sample was chosen to reduce variability due to regional and climatic unfamiliarity, while acknowledging that a more diverse sample could enhance the generalizability of the findings [20]. The participants’ average body mass index (BMI) was 22.36 kg/m², which falls within the normal range (18.5–24.9 kg/m²) recommended for healthy individuals by the World Health Organization, and this BMI is suitable for studying psychological and physiological perceptions.

To avoid the influence of changes in clothing thermal resistance on the experimental outcomes, the same clothes were worn at all four temperatures to enable participants to perceive temperature changes from cool to warm. Throughout the experiment, the participants were instructed to wear sweaters or pullovers, pants, and closed-toe shoes to ensure similar thermal resistance, in accordance with ISO 9920 standards [7]. The participants sat in a chair and performed light activities, with a metabolic equivalent of task (MET) of 1.2 [49]. Table 7 lists the demographic information. The experiment took place between December 15 and December 28, 2021, during the winter season.

Procedure

During the observation period for each group of visual environmental scenes in the multiscenario experiment, long-term use of VR glasses could induce subjective discomfort in participants, so the shortest possible time was given for thermal adaptation to the VR environment. We eliminated the impact of discomfort on the experimental results by setting the observation time for each scene to 3 min [50]. Additionally, to minimize visual distraction before the subsequent image was presented, a blank page was shown between the two sets of pages for 30 s [51]. The experiment was conducted over multiple days, with each participant exposed to only one fixed temperature condition per day to ensure full thermal adaptation and avoid cumulative fatigue effects. The sequence of temperature conditions was randomized across participants to avoid the influence of experiential awareness caused by progressively increasing or decreasing temperatures.

As shown in Fig. 3, after arriving at the rest area of the experimental cabin, the participants rested for approximately 20 min and provided basic personal data to participate in the experiment. In the experimental room, the participants were first asked to listen to light music for 30 min to relax, creating a real living room environment for their sensory immersion [52]. The participants were subsequently instructed to wear the VR glasses and maintain a seated posture in the test area to adapt to the spatial visual environment under the various conditions.

After observing a group of scenes, participants were prompted by an audio signal to complete the questionnaire via verbal responses. To minimize disruptions caused by removing the VR headset, a high-sensitivity microphone was placed in the experimental cabin, enabling participants to respond verbally without additional movement. The investigators in the control room recorded the answers in real time, ensuring an uninterrupted immersive experience. The participants then entered a resting period to establish a neutral physical and mental state and to avoid negative effects of image recall on the results of the subsequent experiments. After evaluating half of the VR scene, the participants were instructed to remove the VR glasses and listen to a light piece of music to alleviate the visual fatigue and discomfort caused by long-term wearing of VR glasses and to reinforce their spatial experience in the living room environment. After the participants completed the entire experiment, the study team numbered and archived the experimental data. In this study, we randomized the order of the scenes for each subject to eliminate any order effects. The participants completed the test after completing all the scene experiments.

Statistical analysis

In thermal perception research, subjective evaluations under different environmental conditions are typically analyzed through regression with the operative temperature. This method allows for the determination of specific temperature values and ranges for various thermal indices within each distinct visual environment. Statistical analysis was conducted using SPSS version 23 software (IBM Corp., Armonk, NY). Studies have questioned traditional thermal sensation vote (TSV) statistical methods based on linear regression, as these methods treat TSV as a continuous variable, neglecting its ordinal discrete nature [53]. In the ASHRAE 7-point scale, the discrete categories of TSV (− 3 to + 3) are not equidistant but are ordered, meaning that they inherently possess ordinal properties rather than interval scale properties. Therefore, using linear regression to model TSV, which assumes equidistant spacing, can lead to unreasonable inferences and results. To address this issue, recent research has recommended the use of ordinal regression models (such as ordinal probit regression and ordinal logistic regression), which can effectively handle the ordinal nature of TSV and avoid the biases introduced by assuming continuous variables [43, 54]. Thus, employing ordinal regression models for TSV analysis not only better aligns with the actual data structure, but also enhances the accuracy and reliability of statistical inferences.

An ordinal probit model was applied to model the relationship between the operative temperature and TSV. This approach enables probabilistic estimation of categorical response distributions and estimates the cutoff points between TSV levels, offering a more rigorous and interpretable treatment of ordinal data [55].

Logistic regression was used to analyze thermal preferences (TPs), determining the overall tendency of preference rather than relying solely on individual scale points. For percentage calculations, responses were categorized as "Preferring Warmer" for positive values (+ 1, + 2, + 3) and "Preferring Cooler" for negative values (− 1, − 2, − 3), while neutral responses (0) were excluded from percentage computations to ensure a clearer representation of preference trends. Polynomial regression was applied to model the relationship between the percentage of dissatisfied individuals (PD) and temperature, defining the acceptable thermal range (PD ≤ 20%) [56]. One-way Analysis of Variance (ANOVA) with post hoc Tukey’s test (p < 0.05) was used to evaluate significant differences across test scenes, such as coverage levels and wood colors.

VR validation

For the preliminary experimental results, both real and VR environments showed similar perception trends across the four evaluation metrics in response to changes in indoor visual environments. Taking TSV as an example, the neutral temperature in real environments with wooden wall surfaces was 22.4 °C, which was lower than the 23.4 °C observed in nonwooden environments. In virtual environments, this effect was more pronounced, with the neutral temperature for wooden wall surfaces decreasing to 22.1 °C compared with 23.7 °C for nonwooden environments. This finding indicates that the effect of wooden walls on lowering the neutral temperature was more significant in VR environments, with a change range approximately ± 0.5 °C greater than that in real environments, which is within the experimental margin of error [36]. These findings confirm that VR technology effectively replicates the thermal perception trends observed in real environments across different visual settings.

Because subsequent studies primarily aim to assess the impact of different wooden visual environments on thermal perception and do not require precise measurements of absolute temperatures under various visual conditions, the application of VR in this study is deemed feasible.

Analysis of factor significance

Thermal sensation votes were used to investigate the influence of coverage rate, natural wood color, and layout style on thermal perception. Using thermal sensation votes as an example, Table 8 presents the statistical significance (p < 0.05) of each factor and its interactions on thermal sensation evaluation. These results provide insights into the relative importance of each factor in shaping thermal sensation ratings.

These factors indicate that at each temperature level, the coverage rate and natural wood color significantly affect thermal sensation, with natural wood color having the most significant effect. Layout mode had no significant effect on thermal sensation, and the interaction effect between the three factors and temperature was not significant. This study also investigated the interaction between coverage and natural wood color. Because the effect of this interaction is still not significant, in the following subsections, the variables of cover and natural wood color are examined separately but are no longer included in the layout analysis.

Impact of wood coverage on thermal perception

Thermal sensation analysis

An ordinal probit model was employed to estimate the probability distributions of TSV responses at various temperatures under different wood coverage scenarios. All the models successfully converged, with the temperature coefficients being consistently positive and statistically significant at conventional levels (p < 0.05).

As illustrated in Fig. 4a, this curve represents the predicted probability of TSV values within [− 1, 0, + 1] in the probit model across different operative temperatures under varying coverage conditions. The temperature range corresponding to a comfort probability ≥ 0.8 exhibits an expanding-then-contracting trend with increasing coverage, reaching its peak at 60% coverage. Considering the definition of the comfort temperature range as TSV between ± 0.5 according to ASHRAE, we further conducted subsequent research by applying a weighted average method based on predicted probabilities. This approach preserves the ordinal nature of the original TSV scale while producing a continuous outcome.

Neutral temperature

To calculate the neutral temperature under different coverage scenarios, we determine the expected thermal sensation vote (eTSV) at each temperature by taking the weighted average of the predicted probabilities for all TSV categories. The expression is provided in Eq. (1):

where \(j\) represents the discrete TSV category level and \(P\left( {{\text{TSV}} = j|T} \right)\) represents the predicted probability of receiving a vote of \(j\) at temperature \(T\).

In the formulation, the ordinal probit model assumes a linear effect of temperature on an underlying latent continuous variable. Although both the probit and linear regression models yield a continuous function of temperature, the probit model is constructed on the basis of the estimated probabilities of each discrete TSV category. The eTSV is then computed as the weighted average of these category probabilities. This approach preserves the ordinal and discrete nature of the original TSV scale while producing a continuous outcome, thereby avoiding the equal interval assumption implicit in treating ordinal ratings as continuous variables in linear regression.

As shown in Figure 4b, the neutral temperature, defined as the operative temperature at eTSV = 0, decreases with increasing wooden wall coverage up to 60 %, allowing participants to feel more comfortable in lower-temperature environments. When coverage was 60 %, the neutral temperature reached its lowest point, and participants demonstrated the greatest tolerance to cold. Beyond 60 % coverage, the neutral temperature increased, reducing participants’ comfort in colder environments.

Comfortable temperature threshold

The eTSV values are obtained within the range of [−0.5, +0.5], as shown in Figure 4b. The lower limit of the comfortable temperature in the environment with varying degrees of wood coverage on the walls also tended to first decrease and then increase, indicating that participants’ tolerance to the cold environment first increased but then decreased. The lowest comfortable temperature was reached at 60 % coverage, at which the participants were best able to withstand a cold environment.

The upper limit of comfortable temperatures continued to decrease, indicating that participants were better able to tolerate the hot environment when the wood coverage was low. This effect weakened with increasing coverage.

Comfortable temperature range

As the coverage increased, the range of comfortable temperatures gradually increased, and the adaptability of the participants to the thermal environment improved. When the coverage reached 60 %, the range of comfortable temperatures in the environment reached its maximum, indicating that participants could tolerate the widest range of temperatures at this coverage level. As the coverage continued to increase, this range decreased. This trend mirrors the changes observed in the neutral temperature.

Thermal preference analysis

Preferred temperature

Fig. 5 shows the curves for cold and heat preferences under different coverage conditions. The black dotted curve represents the combined probability of "Preferring Warmer" and "Preferring Cooler" at the same temperature; the lowest point of this curve indicates the temperature at which the highest proportion of participants preferred no change in the indoor thermal environment, which is considered the participants’ preferred temperature.

The results show that the temperature preferred by participants in a wooden visual environment first decreases and then increases with changes in the amount of wood coverage. When the coverage was less than 60%, the preferred temperature decreased as the coverage increased, suggesting that higher wood coverage led to a preference for lower indoor temperatures. However, when coverage exceeded 60%, the preferred temperature increased with increasing coverage, and participants preferred a warmer indoor temperature. This conclusion confirms the previous conclusions regarding the statistics used to evaluate the perception of warmth.

Thermal acceptance analysis

Acceptable thermal threshold

The temperature corresponding to the percentage of dissatisfied participants was set as the acceptable thermal temperature range, as observed from the range of curves falling within the gray-shaded region in Figure 6a. The lower limit of the acceptable thermal temperature for the environment initially decreased as the wood coverage increased, enabling participants to gradually tolerate a lower indoor temperature. When the coverage exceeds 60 %, the acceptable thermal temperature increases with increasing coverage, and the ability of participants to adapt to cold decreases. Overall, at 60 % wood wall coverage, the lower threshold of acceptable thermal temperature is reached.

For the upper limit, the curve did not fall within the range of PD ≤ 20 % within the operating temperature range when the coverage was greater than 60 %, indicating that excessive wood coverage reduces human tolerance to high temperatures. When wood coverage is low, such as 20 %, people have relatively high thermally acceptable temperatures, which is consistent with previous ratings of thermal sensation. When coverage is low, people can tolerate relatively high indoor ambient temperatures.

Acceptable thermal interval

When the coverage rate was less than or equal to 60 %, increasing the coverage rate gradually increased the comfort temperature interval, and the adaptability of the participants in the indoor thermal environment increased. When the coverage rate reached 60 %, the decrease in the temperature interval accepted by the participants was less significant. This conclusion aligns with the results of thermal sensation and thermal preference discussed earlier in this paper.

Thermal comfort analysis

As shown in Fig. 7a, in the low-temperature environment (18 °C and 21 °C), as the coverage increased, the mean thermal comfort votes (mTCVs) first decreased but then increased, with the lowest value being 60%. Under these conditions, as shown in Fig. 7b, the number of samples selected for "0: comfort" first increased in the low-temperature environment and then decreased, peaking when the coverage level was 60%. This comprehensively proves that as coverage increases, people’s evaluations of comfort in the thermal environment first increase but then decrease. This value peaks at 60%, where thermal comfort is at its highest value.

In the high-temperature environment (27 °C), the mean thermal comfort rating initially decreased but then increased, whereas the percentage of comfort votes (PCVs) initially increased but then decreased; both values peaked at 20%, indicating that people can better adapt to a warmer indoor environment when the wall is covered with less wood. The statistical results regarding the relationship between wood coverage and thermal comfort ratings in the wood-based visual environment further confirmed these conclusions.

Impact of natural wood color on thermal perception

Thermal sensation analysis

On the basis of further analysis of the comfort probability results obtained from the ordinal probit model under different color conditions, Fig. 8a illustrates the predicted probability of TSV falling within the acceptable range [− 1, 0, + 1] under each condition. Compared with the nonwood scenario, all wood-colored environments present a broader temperature range in which the comfort probability exceeds 0.8, indicating an expansion of the comfortable temperature range.

Neutral temperature

Figure 8b shows that among the three natural wood colors tested, yellow-white wood had the lowest neutral temperature and the strongest cold adaptability, followed by reddish-brown and brown wood, which had greater cold adaptability than did the nonwooden wall design environment.

Comfortable temperature threshold

The lower limit for comfortable temperature for the various natural wood colors was lowest for yellow-white, followed by reddish-brown, and highest for brown, indicating that in low-temperature environments, individuals are more comfortable in rooms with yellow-white wood on the walls, followed by those with reddish-brown and brown wood on the walls.

For the upper limit, Figure 8b shows the following order from highest to lowest: brown, nonwooden, reddish-brown, and yellow-white. The respondents noted that it was more difficult to reach higher interior temperatures when the original interior wall wood was reddish-brown or yellow-white. However, a brown wood collage on the wall allowed them to adapt to the high-temperature environment better than the nonwooden collage and provided them with a sense of comfort.

Comfortable temperature range

As shown in Figure 8b, in the rooms containing wooden walls, changing the natural wood color did not significantly alter the comfortable temperature range; this requires further differentiation in subsequent studies.

Thermal preference analysis

Preferred temperature

As shown in Figure 9, among the three wood colors tested, yellow-white wood had the lowest preferred temperature, followed by reddish-brown and brown wood, and the preferred temperature was significantly lower than that in the nonwooden environment. This conclusion confirms previous conclusions regarding the sensation of warmth.

Thermal acceptance analysis

Acceptable thermal threshold

The range of curves falling within the gray-shaded region is shown in Figure 10a. The lowest acceptable temperature limit of the living room with wood decoration was lower than that of the plain wall environment without wood material, regardless of the natural wood color, and the living room with wood decoration provided greater adaptability to cold. In addition, the lower limit of the acceptable temperature in a room with wooden walls varies according to the natural wood color, with yellow-white being the lowest, followed by reddish-brown and brown being the highest. Therefore, people in rooms with yellow-white wood can tolerate lower indoor temperatures, followed by people in living rooms with reddish-brown or brown wooden walls.

Regarding the upper threshold of the acceptable temperature, Figure 10b indicates that brown has the highest threshold, followed by nonwooden, reddish-brown, and yellow-white. Consequently, people can better tolerate higher indoor temperatures when a brown wood collage is used for interior walls. In contrast, people find it more difficult to accept higher indoor temperatures in rooms with yellow-white and reddish-brown wooden walls than in rooms with nonwooden walls.

Acceptable thermal interval

In each wooden visual environment, the interval widths were arranged in the following order: yellow-white had the widest interval, followed by reddish-brown, with brown having the narrowest interval. Accordingly, the participants had the strongest adaptability to ambient temperature in the yellow-white wooden visual environment, followed by the reddish-brown and brown environments. This conclusion refines the findings regarding comfortable temperature ranges as a function of the natural wood color of the walls and confirms previous conclusions regarding thermal preferences.

Thermal comfort analysis

As shown in Fig. 11a, the mean thermal comfort values in the low-temperature environment (18 °C and 21 °C) were generally lower for wood walls than for nonwooden walls, and the yellow-white wood wall environment had the lowest mean thermal comfort vote value. Additionally, no significant differences were detected between the reddish-brown and brown wood walls. In Fig. 11b, for the percentage of comfort ratings in the thermal environment, the percentage of ratings was significantly higher for each type of wood wall than for the nonwooden wall; the percentage was the highest for the yellow-white environment and did not significantly differ between the reddish-brown and brown environments. Thus, yellow-white wood walls can provide better thermal comfort at low temperatures, followed by reddish-brown or brown wood walls.

In the high-temperature environment (27 °C), the brown wood wall had the lowest mean thermal comfort score, while the percentage of comfort ratings was the highest. This confirms the conclusions of this study regarding the relationship between natural wood color and the thermal perceptions of wood in the visual environment.

The results show that wood coverage has a nonlinear effect on human thermal perception. At 60% coverage, people perceive the environment to be warmest; the overall trend shows an initial increase in perceived warmth, followed by a decrease, as people’s preferences for coverage change in a nonlinear manner.

However, while previous studies suggest that changes in wood coverage have a linear effect on human thermal perception, the U-shaped patterns observed in our investigation likely result from multiple contributing factors. First, human preference for indoor wood tends to increase with increasing coverage, after which it decreases [20]. Additionally, extensive wood coverage may visually induce a sense of heaviness and solidity, often creating the illusion of a cooler temperature [57]. Furthermore, a large expanse of wood might evoke psychological associations with natural environments, such as forests and trees, leading to a perceived drop in temperature [58].

The natural wood color of walls influences human perceptions of warmth, which may stem from the inherent visual attributes of wood or simply from the effect of color itself. To investigate this further, an ordinal probit model was applied to analyze the thermal sensation ratings for two wall types with the same yellow-white color—one with 100% wood coverage and one with solid paint (as shown in Table 4, Tests 2 and 3)—across four temperature levels. This analysis revealed that natural wood surfaces exhibited a lower rate of change in thermal sensation compared with solid-colored surfaces. This comparison suggests a broader eTSV range across temperatures for natural wood, indicating enhanced adaptability to temperature variations. Consequently, participants perceive natural wood surfaces as more suitable for colder conditions than solid paint surfaces. It has been further demonstrated that in environments featuring yellow-white wood, participants reported better thermal evaluation at lower ambient temperatures than in environments with reddish-brown wood, which challenges the hue–heat hypothesis that red tones are associated with greater warmth. This discrepancy could be explained by the psychological expectation that yellow-white wood creates, potentially evoking a stronger sense of "naturalness". The lighter, more natural appearance of yellow-white wood might foster a deeper connection to nature, thereby enhancing thermal perception under colder conditions. This hypothesis warrants further discussion and investigation.

Furthermore, different evaluation measures may reveal variations in assessment outcomes. In the thermal acceptability analysis, the 80% wood coverage environment tolerated lower ambient temperatures than the 60% coverage environment, which contrasts with the findings from the thermal sensation and thermal preference analyses. With respect to color selection, the thermally neutral temperature for brown wood was similar to that of nonwooden environments, but its thermally preferred temperature was notably lower.

This difference in perception may be attributed to the use of a statistical questionnaire as a tool for measuring an individual’s sense of environmental temperature. Thermal sensation scores are mainly devised to measure one’s instant sensations, which reflect the degree of direct impact of the present environment on the physical perceived temperature. Concurrently, thermal acceptability scores are used to assess an individual’s subjective satisfaction with the temperature in the environment, reflecting a person’s level of tolerance and acceptance. Thermal preference scores focus primarily on an individual’s future expectations, where individuals could be physiologically warm but psychologically perceive a temperature beyond their acceptance level. Therefore, quantifying perceived temperature from different angles may yield slightly varying findings, which highlights the complexities and challenges in understanding how humans perceive and process environmental temperature information.

The analysis of various virtual scenes indicates that wooden visual environments significantly influence indoor thermal perceptions. Among these factors, natural wood color has a more pronounced impact on thermal perception, whereas the layout contributes only marginally. As the coverage increased, participants’ acceptance of the cold environment first increased but then decreased, with the highest value being 60%. With low wood coverage, the respondents were better able to accept the warmer indoor environment. In terms of natural wood color, individuals in yellow-white wooden wall environments presented greater tolerance to lower temperatures than did those in reddish-brown and brown wood environments. In warmer settings, brown wood was still preferred over nonwood environments because it conveyed a cooler temperature impression, even though brown wood is generally classified within the warm color spectrum according to the hue–heat hypothesis.

This study also has certain limitations. First, although the visual effects of wood on thermal perception indoors in a virtual environment were simulated, real tactile and olfactory experiences related to wood in a virtual environment are difficult to simulate accurately; that is, distinguishing the visual differences between real wood and wooden textured stickers is difficult, which may have affected the results. Second, in TP analysis, responses from the original 7-point scale were classified into broader categories ("Preferring Warmer" or "Preferring Cooler") for logistic regression, which may have led to a loss of granularity in capturing nuanced thermal preferences. Finally, the study relied on subjective evaluations without incorporating physiological measurements, and the relatively homogeneous sample, which was limited to participants from a single regional context, may have constrained the generalizability of the findings. Future studies should address these limitations to obtain a more comprehensive understanding.

The datasets used in the current study are available from the corresponding author upon reasonable request.

VR:

Virtual reality

RH:

Relative humidity

BMI:

Body mass index

MET:

Metabolic equivalent of task

TSV:

Thermal sensation vote

eTSV:

Expected thermal sensation vote

PD:

Percentage of dissatisfied individuals

mTCV:

Mean thermal comfort vote

PCV:

Percentage of comfort vote

TP:

Thermal preference

ANOVA:

Analysis of variance

We thank Professor Hong Jin and Professor Chao Shen for their valuable advice on the experimental design and guidance on data analysis.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Authors and Affiliations

1. School of Architecture, Harbin Institute of Technology Shenzhen, Shenzhen, 518055, China

   Qishen Guo

2. Key Research Base of Humanities and Social Sciences of Guangdong Province, Center for Digital Technology of Space Governance, Harbin Institute of Technology Shenzhen, Shenzhen, 518055, China

   Qishen Guo

3. School of Architecture and Design, Harbin Institute of Technology; Key Laboratory of Cold Region Urban and Rural Human Settlement Environment Science and Technology, Ministry of Industry and Information Technology, Harbin, 150006, China

   Hongpeng Xu & Jianmei Wu

4. Faculty of Architecture and Urban Planning, Chongqing University, Chongqing, 400044, China

   Zirui Fang

5. Key Laboratory of New Technology for Construction of Cities in Mountain Area, Ministry of Education, Chongqing University, Chongqing, 400044, China

   Zirui Fang

Contributions

QG wrote and edited the original draft, validated the findings, conducted the investigations, performed the formal analysis, and curated the data. HX provided supervision, contributed to methodology development, and conceptualized the study. JW participated in writing, reviewing, and editing, as well as providing supervision for the study. ZF contributed to the data analysis and visualization. All the authors read and approved the final manuscript.

Corresponding author

Correspondence to Hongpeng Xu.

Ethics approval and consent to participate

This study was approved by the Research Ethics Committee at the School of Architecture, Harbin Institute of Technology. All participants provided informed consent prior to their involvement in the study.

Consent for publication

Not applicable.

Competing interests

The author(s) declare that they have no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix 1

Focus group discussion topics and key findings.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions