Exploring the Impact of Visual and Aural Elements in Urban Parks on Human Behavior and Emotional Responses

Abstract

As cities progress into high-quality developments, the demand for urban parks that enhance residents’ well-being and sustainability is increasing. Traditional visual-centric design methods no longer suffice. Given that vision and hearing are the primary sensory pathways through which people perceive their environment, exploring their relationship with landscape experiences offers a novel perspective for optimizing the audiovisual perception quality of urban parks. This study explores the relationship between visual and auditory elements and landscape experiences to optimize urban parks’ sensory quality. Using visual perception, soundscape perception, sound source perception, and behavioral vitality, this study evaluates the audiovisual perception quality of a representative wetland park in Chengdu’s ring ecological zone. By quantifying relationships between audiovisual characteristics, behavioral vitality, and emotional feedback, several emotional assessment models were constructed. The results show that lawns, pavements, and sound pressure levels significantly impact vitality. A sound pressure level of 77 dB has been identified as a critical threshold in emotional perception models. Consequently, distinct emotional prediction models can be employed to enhance landscape design across various sound pressure level zones. This research provides scientific evidence and flexible strategies for designing urban open spaces that improve landscape experiences based on multisensory perception.

1. Introduction

The WHO’s “Healthy Cities” initiative emphasizes the importance of urban environments that promote health and well-being [1]. In response, agencies in various countries, such as the U.S. Environmental Protection Agency, National Health Service of the U.K., and the National Health Commission of the People’s Republic of China, have published guidelines for designing open spaces that are accessible, safe, and conducive to physical activity and mental relaxation [2,3,4]. Green and open spaces help reduce stress and anxiety, enhance well-being, and improve the quality of life, directly impacting mental health positively. Parks, as typical natural landscapes in cities, have been widely proven to regulate urban residents’ emotions—such as inducing pleasure, alleviating anxiety, and boosting self-confidence—and promote behaviors such as recreation and social interaction. However, due to the numerous intrinsic factors influencing emotional feedback and behavior, many studies have insufficiently addressed the complex interactions between environmental characteristics and people’s emotional and behavioral responses.

1.1. The Health Benefits of Parks

Urban parks and green spaces are essential components of sustainable urban environments, offering numerous health benefits to residents. Research consistently demonstrates that exposure to natural environments in urban areas can significantly improve mental and physical health. For instance, studies have shown that green spaces help reduce stress and anxiety, enhance mood, and promote overall well-being [5,6]. Parks provide settings for physical activities such as walking, jogging, and cycling, which are crucial for maintaining physical health and preventing chronic diseases [7,8].

Moreover, the presence of natural elements like trees, water features, and diverse plantings can create restorative environments that facilitate recovery from mental fatigue and stress [9,10]. These environments are particularly effective in urban settings where residents are often exposed to high levels of stress and limited access to natural scenery [11]. The role of parks in fostering social interactions and community engagement is also significant. Well-designed green spaces can act as social hubs where people gather, interact, and build social ties, contributing to social cohesion and community well-being [12,13].

Furthermore, parks contribute to environmental health by improving air quality, reducing urban heat island effects, and providing habitats for urban wildlife, thus enhancing urban biodiversity [14,15]. These ecological benefits, in turn, support human health by creating healthier living environments. Therefore, the integration of parks and green spaces in urban planning is critical for promoting sustainable and healthy urban living [16,17].

1.2. The Effects of Visual or Auditory Elements on Human Behavior and Emotions

Visual elements in urban parks and green spaces significantly influence human behavior and emotional well-being. The presence of natural scenery, such as trees, flowers, and water features, has been shown to enhance mood and reduce stress levels. Research indicates that visual exposure to green environments can lower heart rates and blood pressure, thereby promoting relaxation and recovery from mental fatigue [9,18]. Additionally, the visual presence of water features in urban parks has been linked to increased tranquility and reduced anxiety, further highlighting the importance of incorporating diverse natural elements in urban design [19]. Moreover, the diversity and complexity of visual stimuli in natural environments can enhance cognitive function and creativity, as exposure to varied natural scenes stimulates different areas of the brain [20]. These environments provide a sense of escape from the urban hustle and bustle, contributing to restorative experiences that are essential for mental health [6].

Auditory elements in urban parks and green spaces also play a crucial role in shaping human behavior and emotional responses. Natural sounds, such as birdsong, rustling leaves, and flowing water, have been found to significantly enhance mental well-being and reduce stress. These sounds create a calming and restorative auditory environment that can lower cortisol levels and promote relaxation [21,22]. Research has demonstrated that auditory elements influence the overall perception of environmental quality and satisfaction with urban spaces. For instance, areas with pleasant natural sounds are more likely to be perceived as restorative and are preferred for recreational activities [23,24]. The soundscape can also affect the choice of activities within a park, encouraging behaviors such as leisurely walking, meditation, and social interaction in areas dominated by natural sounds [25]. Conversely, the presence of noise pollution, such as traffic sounds, can detract from these benefits, increasing stress levels and reducing the restorative potential of green spaces [26].

The psychological benefits of visual and auditory elements in urban parks underscore the need for the careful design and maintenance of urban parks. Incorporating a variety of plants, ensuring seasonal changes in scenery, and maintaining cleanliness are essential for maximizing the positive effects on human behavior and emotions [27,28]. Strategies such as creating water features, planting sound-absorbing vegetation, and designing quiet zones away from urban noise sources can enhance the acoustic quality of parks [29]. By creating visually appealing environments and prioritizing natural soundscapes, urban planners can promote healthier and more fulfilling urban lifestyles.

1.3. The Interactive Effects of Visual and Auditory Elements on Human Behavior and Emotions

The interaction between visual and auditory elements in urban green spaces profoundly impacts human behavior and emotions, creating multisensory experiences that contribute to overall well-being. While research has shown that both visual and auditory stimuli individually enhance mental health and promote positive behaviors, understanding their combined effects remains an emerging field. Studies have demonstrated that environments combining visually appealing landscapes with pleasant natural sounds are perceived as more restorative and beneficial than those with a single sensory stimulus [30,31,32]. This synergistic effect underscores the importance of designing urban spaces that cater to multiple senses to maximize health benefits.

Despite these insights, significant gaps remain in the research. Most existing studies rely heavily on qualitative assessments and correlational analyses, which, although valuable, do not provide a comprehensive understanding of the underlying mechanisms. There is a notable lack of quantitative studies that explore the mathematical relationships between visual and auditory elements and their combined impact on human behavior and emotions [23,33,34]. Without such data, it is challenging to develop predictive models that can inform the design of more effective restorative environments. Furthermore, many studies do not consider the potential moderating factors such as individual differences in sensory processing, cultural backgrounds, and situational contexts that might influence the perceived benefits of multisensory environments [35,36,37]. Addressing these gaps requires more rigorous experimental designs and advanced analytical techniques.

Considering the research review, this study aims to investigate the integrated effects of visual and aural stimuli from urban nature on human behavior and emotional responses, especially in the context of traffic noise influence. Drawing upon sustainable landscape design, this research will focus on understanding how the visual landscape and auditory environment of urban green spaces influence emotional responses such as satisfaction, pleasantness, and calmness, as well as behaviors observed within these environments.

Using Bailuwan Wetland Park in Chengdu as a case study, this research will utilize spatial analysis techniques, including space syntax and QGis, to quantify visual landscape attributes. A soundwalk and survey will assess perceived aural characteristics, while onsite observations will document prevalent activities. The data will be used to develop a model that explores how visual and auditory stimuli interact to affect behavior and emotions. By integrating findings from both visual and aural perspectives, this study seeks to contribute to the advancement of sustainable landscape design practices that foster positive emotional experiences and encourage beneficial behaviors within urban environments.

2. Materials and Methods

2.1. Research Site

This research was conducted at the Bailuwan Wetland Park, a typical urban green space surrounding the Chengdu Ring Expressway, as depicted in Figure 1. Located in the southeast corner of Chengdu, Sichuan Province, the park is adjacent to an expressway, resulting in significant traffic noise issues and concentrated citizen complaints. The total area of the park is 2 square kilometers, with approximately one-third covered by water. As a national urban wetland park, it primarily serves recreational, sightseeing, and ecological conservation functions. The park features diverse landscape elements, including water bodies, lawns, woodlands, and paved pavements. The current severe noise problems and the public’s pursuit of a high-quality living environment pose higher demands for the sustainable development of the area.

Considering the park’s construction status, open areas, functional diversity, and environmental sound pressure levels, ten measurement points were selected within the research site to evaluate landscape elements and soundscape satisfaction. The selected points meet the following criteria: (1) Each point has unique visual and acoustic characteristics, distinguishing them from each other. (2) The distance between measurement points is greater than 300 m to eliminate interference from the soundscape of adjacent points. (3) Measurement points cover most visual elements within the research site, offering open views and rich visual elements, suitable for acoustic measurement and visual factor calculations. (4) Each location should have an activity space of no less than 50 m × 50 m, allowing researchers to move freely to gain a more comprehensive acoustic landscape experience, while also not causing excessive differences in the sound environment [38,39].

Nevertheless, we admit there might be some potential bias regarding site selection. For instance, by focusing on areas with open views and rich visual elements, this study might underrepresent more enclosed or vegetated areas that contribute differently to the overall soundscape. The requirement for a 300 m distance between points could result in the exclusion of smaller but significant park areas, leading to a potential underrepresentation of their soundscape characteristics.

2.2. Measurements

This study constructs an indicator system based on visual and auditory environmental characteristics to explore the impact mechanisms of the Bailuwan Wetland Park landscape on human activities and emotional perception. Using QGIS spatial syntax calculations and field surveys, audiovisual environment data for the ten points were collected. Questionnaire surveys quantified behavioral vitality intensity and emotional perception indicators to explore the relationship between audiovisual environments and human activities and emotions. Finally, multiple regression analysis was used to form a predictive model for optimizing audiovisual environments in ecological parks around urban expressways (Figure 2).

2.2.1. Visual Landscape Composition

Considering the objective existence and subjective perception of visual landscape factors [40], this study selected the proportions of different landscape elements within people’s visual fields to represent visual characteristics. Researchers first conducted a site survey of the research area, using a panoramic camera, Gopro, to capture panoramic images of the visual landscape. Through semantic segmentation model (Qgis 3.36.2), the visual landscape elements in the site were ultimately summarized and extracted into four types: woodland (Wo), lawn (L), water bodies (Wa), and pavements (P). Subsequently, geographic information system (QGIS) and spatial syntax methods were used to calculate the composition of the landscape perceived visually by people throughout the entire research area.

2.2.2. Auditory Environmental Characteristics

This study used the equivalent continuous A-weighted sound pressure level (L_Aeq) and sound source perception indicators to describe the park’s sound environment. L_Aeq is a standardized measure for environmental sound pressure levels, expressing decibel values and variations in sound intensity, and has been shown to significantly impact soundscape perception [41]. A multi-channel signal analyzer (AWVA6290L+) recorded and measured the sound environment at each measurement point. Additionally, based on traffic flow data from surrounding expressways and field measurements, a noise map of the site was calculated using the noise prediction model (NPL) [see http://resource.npl.co.uk/acoustics/techguides/crtn/ (accessed on 18 March 2024)]. Sound sources were used to describe which sounds dominate within the study site.

Sound source perception consists of two parts: the type of sound source and the intensity of perception. This study employed a combination of soundwalks and questionnaires to conduct research on sound source perception. A soundwalk is a standardized method widely used in soundscape research to measure auditory perception [42]. This method primarily involves organizing participants to take walks along predetermined routes, during which they focus on experiencing the surrounding soundscape environment and record and evaluate the sounds they hear and their relationship with the surrounding landscape environment at designated research points [43,44]. The sound environment characteristic evaluation part of the questionnaire is divided into two sections. Section 1 is a semi-open questionnaire used to record the types of sound sources at the research points. Section 2 uses a five-point Likert scale (1–5, where 1 represents “not perceived at all” and 5 represents “very strongly perceived”) to evaluate the perceived intensity of different types of sound sources. In the pre-experiment, researchers mainly recorded the categories of sound sources to distinguish the differences in the acoustic environment at different research points and to formulate subsequent experimental methods. In the formal experiment, participants were invited to refine the specific types of sound sources. Based on two experiments, the sound sources in Bailuwan Wetland Park can be categorized into four types: traffic sounds (S1), including car horns, rail bicycle rings, motor vehicles, highways, heavy vehicles, and airplanes; construction sounds (S2), including alarms, radios, and loudspeakers; human sounds (S3), including exercising, cleaning, whistling, walking, parent–child activities, talking, laughing, and singing; and natural sounds (S4), including insects, tree murmur, wind, twittering of birds, ripple, fish diving, and dogs barking.

2.2.3. Behavioral Vitality Characteristics

This study used behavioral vitality (V) to quantify the behavioral vitality characteristics of the site. Behavioral vitality intensity refers to the diversity and intensity of activities within different areas, reflecting the impact of the landscape environment on human behavior. These data were collected through field surveys and expert assessments, recording activities lasting over 3 min within the measurement points and rating their diversity using a five-point Likert scale (1–5, where 1 indicates “single activity” and 5 indicates “highly diverse/intense activities”).

2.2.4. Emotional Perception Evaluation

Based on the classic Russell complex emotion model [45], the Perceived Affective Quality (PAQ) model, and international soundscape standards [46], previous experiments typically selected pleasure, calmness, and annoyance to describe emotional perceptions in the environment. Satisfaction was also selected to evaluate the overall psychological impact of the coupled audiovisual environment in urban green spaces around expressways. However, considering that calmness and annoyance are antonymous in the Chinese context, this study retains the positive emotional indicators of pleasure (E1), calmness (E2), and satisfaction (E3) as the emotional evaluation indicators for this research. Data were collected through a soundwalk and questionnaires, using a five-point Likert scale (1–5, where 1 indicates “very disagreed” and 5 indicates “very agreed”).

2.3. Data Collection

The data collection for this study involved both desk research and field surveys. Initially, desk research was conducted to obtain data on visual environmental characteristics using geographic remote sensing systems. Google satellite images and the Qgis 3.36.2 software platform were employed to acquire remote sensing images of the study site. The proportions and distributions of four types of research elements were measured. Boundaries for each element were delineated using QGIS Maptiler, and the visible areas of landscape elements throughout the entire park were analyzed with Depthmap X, a spatial syntax software, considering buildings and woodland as visual barriers. A grid of 50 m × 50 m, corresponding to the experimental plot space, was applied. Vector circles were generated in QGIS based on the actual positions and ranges of each measurement point, from which the proportions of visual landscape elements were extracted (Figure 3).

Data on auditory environmental characteristics, behavioral vitality characteristics, and psychological perception indicators were primarily collected through field surveys. The surveys were conducted on clear days with no obstructions, rain, snow, or lightning, ensuring good air quality and visibility, and wind speeds below 5 m/s. Observations were made between 8:00 AM and 17:00 PM. To eliminate interference from other factors, 30 participants aged 20–30, with a gender ratio of approximately 1:1, were recruited locally. All participants had good vision and hearing to control for perceptual biases due to individual differences. This study received ethical approval, and informed consent was obtained from all participants. Prior to the survey, participants received necessary training and briefings.

The selection of participants aged 20–30 was driven by this study’s focus on controlling for potential perceptual biases related to sensory acuity and cognitive processing, which can vary significantly with age. Young adults generally possess stable and consistent sensory capabilities, which minimizes variability due to age-related sensory decline. This demographic also tends to have a higher level of familiarity with digital interfaces and technology, which was utilized in the data collection process, ensuring that all participants could engage with the survey tools effectively. Additionally, by focusing on a specific age group, this study aims to produce more homogeneous and reliable data, thereby reducing confounding variables and enhancing the internal validity of the findings. Future studies could expand the age range to explore how these factors evolve across the lifespan.

Once the survey began, 30 participants were divided into five groups. Each group, led by a researcher, followed different paths through the 10 measurement points for soundwalk experiments. Participants freely walked and stayed within the soundwalk areas, observing environmental characteristics and listening to ambient sounds for 5 min [47]. Afterward, they completed and submitted online questionnaires via mobile devices. A total of 250 questionnaires were distributed, and 244 valid responses were obtained, meeting the effect size requirements for visual and soundscape evaluation studies [48].

Simultaneously, researchers measured and evaluated the auditory environment and behavioral vitality characteristics at each point. After calibrating the multi-channel signal analyzer, two amplifiers and microphones equipped with wind shields were placed approximately 1.6 m above ground at the center of each measurement point to record environmental sound pressure levels for 3 min. The equivalent continuous A-weighted sound pressure level (Laeq) was obtained using the multi-channel signal analyzer. Additionally, during the park’s opening hours, professional observers recorded the types of functional activities lasting more than 3 min within the measurement points every two hours (10:00, 12:00, 14:00, and 16:00) and rated their diversity.

2.4. Data Analysis and Visualization

There are two main methods to explore the relationship between emotions and audiovisual environmental characteristics. The first method involves correlation analysis to investigate the factors influencing positive and negative emotions, but it cannot show the interactions between these factors. The second method involves constructing relational equations to predict the impact of the overall environment on human activities and emotional feedback based on the quality of audiovisual environments. Based on this, this study employs modeling and map visualization techniques to process the data, providing an intuitive representation of the park’s behavioral vitality intensity and emotional responses.

The data analysis in this study can be divided into three main steps. First, we constructed a relational model between audiovisual environmental characteristics and behavioral vitality characteristics, with behavioral vitality characteristics as the dependent variable and objective elements of the audiovisual environment as independent variables, to predict the intensity of functional activities throughout the park. Next, using emotional perception elements (pleasure, calmness, annoyance, and satisfaction) as dependent variables, we analyzed the impact of the proportions of various visual elements, perceived intensities of different sound sources, and behavioral vitality levels, constructing a multiple linear regression model to describe the relationship between landscape environment and emotional perception. This model aims to discover the driving factors that affect emotional changes, providing support for the optimization of the audiovisual environment in ecological parks along urban expressways. Finally, we utilized QGIS to visualize the models, predicting the intensity of functional activities and emotional feedback throughout the park, thereby guiding the sustainable development and health of the park.

3. Results

3.1. Manipulation Checks

The collected questionnaire data were subjected to Shapiro–Wilk and Kolmogorov–Smirnov tests to assess normality, with the results presented in

Table 1. Significant (p) values of normality test.

	Significant (p-Value)
	Shapiro–Wilk Test	Kolmogorov–Smirnov Test
Traffic sounds (S1)	0.290	0.776
Construction sounds (S2)	0.417	0.601
Human sounds (S3)	0.240	0.854
Natural sounds (S4)	0.285	0.762
Behavioral vitality (v)	0.354	0.625
Pleasure (E1)	0.220	0.860
Calmness (E2)	0.178	0.900
Satisfaction (E3)	0.242	0.890

. A significance level of α = 0.05, p > 0.05, indicates that the data meet the normal distribution criteria and are suitable for subsequent analysis. Reliability tests were conducted on the collected questionnaires, and the results indicated that the Cronbach’s α for the satisfaction evaluation exceeded 0.6, demonstrating the overall validity of the collected data [49]. An independent-samples t-test was performed based on gender grouping. The Levene’s test for equality of variances yielded values greater than 0.05, indicating the homogeneity of variances. The t-test for equality of means also showed significance values greater than 0.05, indicating that gender differences did not affect the evaluation results.

3.2. Descriptive Analysis

The visual environmental characteristics, auditory environmental characteristics, behavioral vitality characteristics, and psychological perception results for the ten measurement points at the study site are shown in

Table 2. Objective environmental and psychological perception data for measurement points.

Point Location		Wo	Wa	L	P	Others	LAeq	S1	S2	S3	S4	V	E1	E2	E3
NO.1	Mean	34%	1%	52%	13%	0%	75.9	3.77	1.68	2.41	4.68	3	3.14	2.73	3.3
	Std. Dev	-	-	-	-		-	0.73	0.76	0.78	0.47	-	0.69	0.86	0.63
NO.2	Mean	17%	2%	65%	16%	0%	75.9	4.86	1.18	1.29	4.25	3	2.18	1.57	2.43
	Std. Dev	-	-	-	-		-	0.35	0.38	0.52	0.69	-	0.89	0.73	0.98
NO.3	Mean	22%	6%	56%	16%	0%	74.6	4.45	1.17	1.76	4.55	3	2.41	1.93	2.69
	Std. Dev	-	-	-	-		-	0.56	0.38	0.82	0.56	-	0.77	1.14	0.83
NO.4	Mean	29%	14%	45%	12%	0%	71.4	4.42	1.26	2.05	4.42	3	3.05	2.42	3.42
	Std. Dev	-	-	-	-		-	0.67	0.44	0.69	0.59	-	0.69	0.88	0.59
NO.5	Mean	6%	7%	77%	10%	0%	75.1	4.45	1.23	2.55	4.32	0	2.55	1.73	2.82
	Std. Dev	-	-	-	-		-	0.66	0.42	0.58	0.63	-	0.58	0.69	0.65
NO.6	Mean	3%	18%	66%	13%	0%	76.8	4.92	1.71	1.54	4.17	2	1.96	1.42	2.21
	Std. Dev	-	-	-	-		-	0.28	0.93	0.58	0.75	-	0.73	0.64	0.82
NO.7	Mean	18%	17%	54%	11%	0%	79.2	4.41	1.17	1.97	4.29	3	3.07	2.55	3.35
	Std. Dev	-	-	-	-		-	0.67	0.38	0.72	0.61	-	0.94	1.19	0.9
NO.8	Mean	30%	14%	49%	7%	0%	78.5	4.65	1.22	2.04	4.04	3	3.09	2.39	3.21
	Std. Dev	-	-	-	-		-	0.56	0.41	0.81	0.86	-	0.65	1.01	0.63
NO.9	Mean	57%	0%	27%	16%	0%	66.3	3.67	1.95	2.05	4.90	2	3.86	3.52	4.64
	Std. Dev	-	-	-	-		-	0.64	1.13	0.90	0.29	-	0.77	1.05	0.71
NO.10	Mean	25%	23%	34%	13%	5%	69.3	3.00	1.54	4.23	3.83	2	3.42	2.46	3.62
	Std. Dev	-	-	-	-		-	1.00	0.93	0.80	1.06	-	0.84	1.01	0.78
Mean		24%	11%	52%	13%	1%	74.3	4.26	1.41	2.19	4.35	2.4	2.87	2.27	3.17

Wo = woodland, Wa = water bodies, L = lawn, P = pavements, L_Aeq = A-weighted sound pressure level, S1 = traffic sounds, S2 = construction sounds, S3 = human sounds, S4 = natural sounds, V = behavioral vitality, E1 = pleasure, E2 = calmness, and E3 = satisfaction.

and Figure 4.

The overall satisfaction level in Bailuwan Wetland Park is relatively high, with an average of 3.12. The highest satisfaction level is at point 9 (4.14/5), while the lowest is at point 6 (2.21/5). There are four points with satisfaction levels below 3.00: point 2 (2.43/5), point 3 (2.69/5), point 5 (2.82/5), and point 6 (2.21/5). The satisfaction levels for the remaining points range between 3.00 and 4.00.

3.2.1. Descriptive Analysis of Visual Environmental Characteristics

Using a geographic information system (QGis) and space syntax, the proportions of lawn, woodland, water, and pavement elements within the site were calculated, as shown in Figure 5. It was found that lawns are distributed throughout the site, with higher visibility near the viewing platform on the east side of the park. Water bodies are also widely distributed, mainly concentrated in the northern part of the site near point 6, Tinglu Island. Woodland in the western area is scattered, while in the eastern area, it is mainly concentrated in the southern part. A few pavements intersect the site, with higher density and higher pavement grades in the north.

The proportions of visual elements for the 10 measurement points were compared, as shown in Table 2 and Figure 5. The proportion of lawn (mean = 51.54%) was the highest, followed by woodland (mean = 24.02%), pavement (mean = 13.68%), and water (mean = 10.19%) as the least among visual elements. Six measurement points had lawn proportions greater than 50%: point 5 (78%), point 6 (66%), point 2 (65%), point 3 (57%), point 7 (54%), and point 1 (52%). Point 8 (49%) and point 4 (45%) were slightly below 50%, while point 9 and point 10 had the lowest proportions, at 27% and 33%, respectively.

Regarding woodland elements, most points had proportions between 20 and 35%, including point 1 (34%), point 8 (30%), point 4 (28%), point 10 (25%), and point 3 (22%), with only point 9 (57%) exceeding 50%. The points with the least visible woodland elements were point 6 (3%) and point 5 (6%), followed by point 2 (17%) and point 7 (18%).

Water is an important component of the wetland park landscape, and except for point 9, all points had visible water elements. Although water elements were not a design necessity, they played a significant role in psychological perception. The highest visibility for water elements was at point 10, located near Baxian Bridge and adjacent to the water. Points 6, 7, 8, and 4 had water proportions of between 10 and 20%, while the remaining points had less than 10% visibility. There were no large hardscape areas within the site, and the proportion of pavements at each point was similar, fluctuating around 10%. The highest pavement visibility was at points 2, 3, and 9, each with 16%, indicating a dense pavement network. The lowest was point 8, with a pavement grade of only 7%.

3.2.2. Analysis of Auditory Environmental Characteristics

The average sound pressure level across the 10 measurement points in the site was 74.3 dBA. Eight out of these ten points had sound pressure levels exceeding 70 dBA, surpassing the noise standards stipulated by China’s “Environmental Quality Standards for Noise” for areas adjacent to expressways. The highest sound pressure level was recorded at point 7 (79.2 dBA), followed by point 8 (78.5 dBA) and point 6 (76.8 dBA). Points 2 (75.5 dBA), 1 (75.5 dBA), 5 (75.5 dBA), and 3 (75.5 dBA) all had sound pressure levels around 75 dBA. Only points 10 and 9, which were farther from the expressway, had decibel values meeting the standards, slightly below 70 dBA, as shown in Table 2 and Figure 6.

Based on the traffic flow on the northern side of the site and the measured sound pressure levels at the 10 points, a noise map of the site was generated using the noise prediction model (NPL) [5], as illustrated in Figure 6. The simulation results indicated that the expressway was the primary noise source within the site, with the highest environmental sound pressure levels observed in the northern part, gradually decreasing with distance.

The perception of traffic noise was strong at all 10 measurement points, with perception levels above 3.00. Points 9 (3.67/5) and 10 (3.00/5) exhibited slightly weaker perceptions, whereas point 6 (4.92/5) had the strongest. The perception of mechanical noise was the weakest across all points, with levels below 2.00. As a wetland park, natural sounds were strongly perceived at all points, particularly at points near woodland, such as point 9, which experienced abundant bird calls. The strongest perception of human sounds was at point 10 (4.23/5), located by the water, where traffic noise was minimal and human conversations and laughter were clearly audible (Figure 7).

3.2.3. Analysis of Behavioral Vitality Characteristics

Based on the distribution of landscape elements within the site, Bailuwan Wetland Park can be divided into four zones, each characterized by its predominant landscape element: the Ecological Conservation Area where forest land is most abundant, the Traffic Activity Area with the highest proportion of roads, the Recreational Area dominated by lawn elements, and the Wetland Protection Area which is richest in water body elements. The area mainly consists of nature reserve areas and wetlands, with an overall lower level of activity, and the average value is 2.4. Open space points 1, 2, 7, 8, 3, and 4 are located in the northwest region of the site, with a behavioral vitality rating of only 3. Points 9 and 10 follow with an activity rating of 2. Point 5 shows minimal human activity, with an activity rating of 0, as depicted in Figure 8.

Correlation analysis between site activity intensity and audiovisual environmental characteristics revealed significant associations with woodland (p < 0.001 < 0.05), lawn (p < 0.001 < 0.05), pavement (p < 0.001 < 0.05), and sound pressure level (p < 0.001 < 0.05). At the same time, the Shapiro–Wilk test and the Kolmogorov–Smirnov test results for behavioral vitality both have p-values greater than 0.05, indicating that the data conform to a normal distribution. Based on this finding, a multiple regression model was attempted with site activity intensity as the dependent variable and woodland, lawn, and sound pressure levels as independent variables. During model fitting, it was found that woodland elements exhibited significant collinearity with site activity intensity (VIF > 10) and were subsequently removed, resulting in a final predictive model for site activity intensity with an R² = 0.597 (Equation (1)):

V = −16.211–−0.58L + 0.9P + 0.275L_Aeq

(1)

The model was validated, as shown in

Table 3. Feasibility assessment of site functional activity fit models.

Model Fit (R2)	Durbin–Watson Test	Attribute	Estimate B	Standard Error	t-Value	p-Value	VIF
0.597	1.622	lawn	−0.058	0.004	−16.043	<0.001	1.885
		pavement	0.090	0.011	8.115	<0.001	1.898
		lAeq	0.275	0.016	17.127	<0.001	2.785

. All independent variables were significant at <0.005, the D-W test result was 1.622 ≈ 2, and VIF values were all less than 10, indicating high model reliability. The model suggests that site activity intensity is negatively correlated with the proportion of lawn in the field of view and positively correlated with pavement visibility and the sound pressure level. In other words, areas with higher greenery exhibit lower activity intensity, whereas areas with better pavement accessibility and livelier sound environments show higher activity intensity.

Mapping the model using QGIS generated the predicted map of site activity intensity (Figure 9). It reveals that areas with high activity intensity are concentrated in the northern side of the site and along the riverbanks, while the utilization of water surfaces is minimal.

3.3. Analysis of Factors Influencing Sound Source Perception and Emotional Perception

3.3.1. Analysis of Factors Affecting the Intensity of Sound Source Perception

A correlation analysis was conducted between the perceived intensity of four types of sound sources and visual landscape elements, as well as functional activity levels (see

Table 4. Correlation analysis of the perceived intensity of sound sources with audiovisual environmental characteristics and behavioral vitality features.

		Wo	Wa	L	P	V
S1	Correlation (Pearson’s r)	−0.347 **	−0.093	0.471 **	0.161 *	0.063
	Significant (p)	<0.001	0.183	<0.001	0.021	0.366
S2	Correlation (Pearson’s r)	0.183 **	−0.023	−0.195 **	0.194 **	−0.062
	Significant (p)	0.008	0.747	0.005	0.005	0.376
S3	Correlation (Pearson’s r)	0.124	0.340 **	−0.367 **	−0.119	0.235 **
	Significant (p)	0.074	<0.001	<0.001	0.088	0.001
S4	Correlation (Pearson’s r)	0.204 **	−0.291 **	0.038	−0.240 **	−0.037
	Significant (p)	0.003	<0.001	0.588	<0.001	0.591

Wo = woodland, Wa = water bodies, L = lawn, P = pavements, S1 = traffic sounds, S2 = construction sounds, S3 = human sounds, S4 = natural sounds, and V = behavioral vitality. * Significant at 5% level. ** Significant at 1% level.

). The results indicate that the perceived intensity of traffic sounds is significantly positively correlated with the visibility of lawns and pavements and negatively correlated with the visibility of woodland. Pavements with high visibility, often the primary source of traffic sounds, tend to have better accessibility, resulting in stronger perceived traffic noise. Conversely, areas with high lawn visibility are typically more open and lack barriers that could impede noise propagation, leading to a stronger perception of traffic sounds.

The perceptibility of construction sounds shows a positive correlation with the visibility ratio of woodland and pavements and a negative correlation with the visibility ratio of lawns. This may be due to construction devices in parks often being located along roads with dense pedestrian traffic or concealed within wooded areas. The perceptibility of human sounds is positively correlated with the visibility of water bodies and the level of behavioral vitality and inversely correlated with the visibility of lawns. In these settings, lawns are generally not favored, and people tend to congregate near water bodies and in areas with higher levels of functional activity.

The perceptibility of natural sounds is primarily related to the visibility of woodland and pavements; increased woodland and reduced pavement visibility are associated with more pronounced natural sounds. Interestingly, the visibility of water bodies, a common source of natural sounds, shows an inverse relationship with the perception of natural sounds in this study. This anomaly may be attributed to the masking effect created by the excessive presence of human sounds around water bodies [50].

3.3.2. Analysis of Factors Influencing Emotional Perception

The analysis indicates that within a certain range of sound pressure levels, human psychological perception of the soundscape is related to the intensity of the sound pressure level. However, when the sound pressure level exceeds this range, the relationship between the two changes. The relationship between the sound pressure level and people’s psychological perception of satisfaction at 10 points in the study site is illustrated in Figure 10. Correlation tests revealed that when the sound pressure level is below 77 dB, the correlation coefficient is R = −0.915, with a p-value of less than 0.01, indicating a significant negative correlation between the two. Conversely, when the sound pressure level exceeds 77 dB, the correlation coefficient changes to R = 0.985, with a p-value of 0.111, which is greater than 0.05, indicating no significant correlation between the psychological perception of satisfaction and decibel values. The perceived satisfaction initially increases and then gradually stabilizes.

Subsequently, a correlation analysis was performed on the other two types of emotions, audiovisual landscape characteristics, and behavioral vitality features. The analysis confirmed that the perception of pleasant emotions (E1) is influenced by visual landscape elements, such as woodlands, lawns, and pavements, and auditory landscape elements, such as traffic noise, human sounds, and natural sounds, as well as the level of behavioral vitality. Specifically, higher visibility of trees and roads, lower visibility of lawns, an increase in human and natural sounds, and more active populations contribute to a higher degree of pleasant emotions.

The perception of calm emotions (E2) is influenced by the visibility of woodlands, lawns, and pavements, the level of behavioral vitality, the perceived intensity of traffic noise, and the A-weighted sound pressure level. A denser woodland, higher road accessibility, fewer lawns, lower sound pressure levels, and reduced functional activity are associated with a greater likelihood of experiencing calm emotions. Details are provided in

Table 5. Correlation analysis of emotional perception with audiovisual environmental characteristics and behavioral vitality features.

		Wo	Wa	L	P	S1	S2	S3	S4	LAeq	V
E1	Correlation (Pearson’s r)	0.472 **	0.029	−0.494 **	0.128 *	−0.466 **	0.036	0.276 **	0.130 *	−0.261 **	0.120 *
	Significant (p)	<0.001	0.658	<0.001	0.046	<0.001	0.572	<0.001	0.043	<0.001	0.045
E2	Correlation (Pearson’s r)	0.467 **	−0.070	−0.432 **	0.197 **	−0.314 **	0.055	0.055	0.110	−0.075	−0.496 **
	Significant (p)	<0.001	0.275	<0.001	0.002	<0.001	0.390	0.141	0.088	0.244	<0.001

Wo = woodland, Wa = water bodies, L = lawn, P = pavements, S1 = traffic sounds, S2 = construction sounds, S3 = human sounds, S4 = natural sounds, V = behavioral vitality, and L_Aeq = A-weighted sound pressure level. * Significant at 5% level. ** Significant at 1% level.

.

Additionally, a correlation analysis of satisfaction revealed that satisfaction is positively correlated with calmness and pleasure, with a p-value of less than 0.001. In summary, the perception of calm emotions is related to decibel levels; when the decibel level is below 77 dB, it can simultaneously affect the psychological perception of satisfaction. However, the variation in pleasant emotions is not related to sound pressure levels. When the decibel level exceeds 77 dB, satisfaction can be enhanced by increasing the sense of pleasure, thereby improving overall perceived satisfaction.

3.4. Emotional Perception Model Prediction

To directly benefit landscape design and perception prediction, we further explored the emotional model equations. After excluding variables with a Variance Inflation Factor (VIF) greater than 10, such as the proportion of lawn elements, and variables with a less significant impact on the dependent variable (e.g., visual area and human sounds), we constructed a model using the perception intensity of pleasant emotions as the dependent variable. The independent variables included lAeq, woodlands, water bodies, traffic noise, and natural sounds, resulting in an R² of 0.500 (Equation (2)).

E1 = 2.915–−0.362S₁ + 0.193S₄ + 0.27Wo + 0.015Wa–−0.032V

(2)

All independent variables in this model were statistically significant (p < 0.05), with a Durbin–Watson test value of 1.547, approximately 2, and VIF values all less than 10, indicating high model reliability, as shown in

Table 6. Feasibility test of the emotional perception fitting models.

Emotion	Model Fit (R2)	Durbin–Watson Test	Attribute	Estimate B	Standard Error	t-Value	p-Value	VIF
E1	0.500	1.547	Woodland	0.025	0.402	0.433	<0.001	1.493
			Water bodies	0.015	0.004	0.145	0.014	1.376
			Activity intensity	−0.362	0.006	−0.039	0.049	1.216
			Traffic sound	−0.362	0.047	−0.388	<0.001	1.139
			Natural sound	0.193	0.054	0.182	<0.001	1.074
E2	0.542	1.645	Woodland	0.051	0.005	0.787	<0.001	2.538
			Pavement	−0.040	0.015	−0.188	0.008	2.021
			LAeq	−0.027	0.022	−0.108	0.023	3.261
			Activity intensity	−0.121	0.065	−0.121	<0.001	1.739

. The model suggests that, in addition to the visibility of woodlands and water bodies and the perceived intensity of natural sounds, higher levels of functional activity and a stronger perception of traffic noise negatively impact the perception of calm emotions.

For the calm emotion perception model, the independent variables are woodlands, pavements, behavioral vitality level, and traffic noise, with R² = 0.542 (Equation (3)). The lawn element was excluded due to a VIF > 10.

E2 = 3.771–−0.027L_Aeq + 0.051Wo–−0.040P–−0.121V

(3)

All independent variables in this model were statistically significant (p < 0.05), with a Durbin–Watson test value of 1.645, approximately 2, and VIF values all less than 10, indicating high model reliability, as shown in Table 6. The model indicates that, in addition to the visibility of woodlands, a higher visibility of pavements, increased levels of behavioral vitality, and lAeq negatively affect the perception of calm emotions.

Mapping the two previous models (Figure 11) reveals that the distribution tendencies of the two types of emotions are consistent, exhibiting a trend of being higher in the south and lower in the north. This pattern is in direct contrast to the distribution of negative emotions. People tend to experience more positive emotions around open water surfaces and large areas of woodland.

4. Discussion

This study, using Bailuwan Wetland Park as a case study, integrates GIS analysis and soundwalk field surveys. From both visual and auditory dimensions, it considers the proportion of visual elements such as lawn, woodland, water bodies, and pavements, as well as the sound pressure level and perceived intensity of four types of sound sources—traffic noise, construction noise, human sounds, and natural sounds—as independent variables. Behavioral vitality levels and emotional perceptions are used as dependent variables to innovatively construct a mid-scale ecological park psychological perception prediction model. This research highlights how visible elements and human activity influence the perceived sound source perception intensity, further contributing to the understanding of soundscape research.

4.1. Discussion on the Influence of Sound Source Perception Intensity

The analysis of factors influencing sound source perception intensity reveals significant correlations between visual landscape elements, functional activity levels, and the perceived intensity of different sound sources. This interplay between visual and auditory elements provides valuable insights for urban park soundscape design, particularly in enhancing acoustic environments to align with user expectations and needs [51,52].

The strong correlation between traffic noise perception and the visibility of grassland and roads highlights the role of visual openness in amplifying undesirable sounds. Grassland areas, being open, offer minimal noise obstruction, leading to a higher perception of traffic noise. In contrast, the negative correlation with tree visibility underscores the effectiveness of vegetative buffers in mitigating noise, aligning with studies on green infrastructure in urban soundscapes [53,54]. Incorporating denser tree cover in high-visibility road areas could reduce the perceived impact of traffic noise.

The relationship between mechanical noise and the visibility of trees and roads suggests that mechanical sources, often near pathways or within wooded areas, are more perceptible. The negative correlation with grassland visibility indicates that open grasslands may be less affected by mechanical noise, possibly due to spatial separation from these sources. This supports research on strategic spatial planning to alter noise perception in urban environments [51,55]. Placing mechanical equipment in less visually prominent locations can reduce their auditory impact.

The correlation between human sound perception and the visibility of water elements and activity levels highlights how people interact with their surroundings. Water bodies, typically seen as tranquil, paradoxically contribute to higher human sound perception due to increased human activity. This masking effect, where natural sounds are overshadowed by human noise, echoes findings on the dual role of water features in soundscape quality [56,57]. Designing buffer zones around water bodies could preserve their natural soundscapes by introducing seating away from the water to minimize noise pollution.

Lastly, the inverse relationship between water visibility and natural sound perception, despite water being a source of natural sounds, underscores the complexity of soundscape design. Prominent water features often attract human activity, diminishing natural sound perception. This can be mitigated by creating quiet zones or enhancing the prominence of natural sounds through landscape design, such as using sound-reflective surfaces or amplifying natural sounds in quieter park areas [43,57].

4.2. Emotional Responses and Audiovisual Landscape Elements

From the three emotional perception models, it was found that the visibility of woodlands and pavements, as well as the perceived intensity of traffic and natural sounds, are significant factors affecting human emotional perception.

Woodlands and natural sounds play a crucial role in fostering positive emotional responses. The presence of trees and greenery has been associated with reduced stress levels and improved mental well-being [9,58]. The natural sounds commonly found in these environments, such as birdsong, rustling leaves, wind, and water, have a significant therapeutic effect on humans [59,60]. These sounds contribute to a sense of tranquility and relaxation, counteracting the stress and anxiety often caused by urban noise. In this study, Bailuwan Wetland Park, located near an expressway, suffers from severe traffic noise pollution. The contrast with the surrounding urban landscape enhances positive feedback when people see woodland landscapes and hear natural sounds. This phenomenon aligns with Attention Restoration Theory, which suggests that natural environments provide restorative experiences that help individuals recover from mental fatigue [9].

The models also reveal that the visibility of water features is significantly correlated with overall satisfaction, although this correlation is weaker in other emotional models. Water is a unique landscape element within the wetland park, present throughout the entire area. During the soundwalk experiments, its presence was consistently perceived. Water features may have a limited immediate visual impact on rapidly changing emotions such as joy, but they subtly influence overall satisfaction through various sensory pathways. The color of the water, its dynamic movements, and the sounds it produces (such as flowing or splashing water) stimulate both visual and auditory perceptions. These sensory experiences contribute to a more profound and enduring effect on people’s overall satisfaction with different locations within the park [54,55]. Studies have shown that water features can enhance the aesthetic quality of a landscape, promote relaxation, and improve the overall mood of visitors [56]. The soothing sounds of water can mask unpleasant noises, further enhancing the auditory environment and providing a calming backdrop for park visitors [57].

4.3. Guidance of Landscape Design Based on Sound Pressure Level and Emotional Perception

The analysis identified 77 dB as a pivotal threshold in the emotional response to park environments impacted by traffic noise. Below this threshold, overall satisfaction with the park environment declines as sound pressure levels increase, whereas above 77 dB, no significant relationship was found between satisfaction and sound pressure levels. Therefore, it is recommended that noise mapping should use 77 dB as a key delineation for zoning and targeted environmental improvements.

For park areas where sound pressure levels are below 77 dB, we recommend using a model based on promoting calm emotions to guide design improvements. This model involves controlling sound pressure levels primarily through physical noise reduction measures, such as the installation of sound barriers and protective vegetation. Additionally, perceptual quality can be enhanced by increasing the visibility of woodland areas and reducing the visibility of hardscape elements at the visual level. Furthermore, the model reveals a negative correlation between site activity levels and the perceived calmness of the soundscape. Therefore, the site activity level model can provide further guidance on enhancing the perceived soundscape calmness, such as by appropriately increasing the visibility of grassy areas. These findings offer more refined and feasible strategies for park environment optimization, moving beyond the sole reliance on sound pressure control for improving park quality.

In contrast, for areas where sound pressure levels exceed 77 dB or where an effective reduction in sound pressure is not feasible, a different approach is necessary. In such zones, the focus should shift from sound pressure control to enhancing the perceived pleasantness of the soundscape. This can be achieved by reducing the perception of traffic noise, amplifying the perception of natural sounds, increasing the visual prominence of greenery, and reducing the visibility of paved surfaces. In addition, reducing the activity level can improve people’s perceived pleasantness of a soundscape to a certain extent, which reflects the necessity of controlling the intensity of human activities in wetland parks. These strategies indicate that, even in acoustically challenging environments, it is possible to improve the perceptual quality of the soundscape. This dialectical approach highlights the need for landscape design strategies that are responsive to specific acoustic conditions, offering a new perspective on the enhancement of urban park environments.

4.4. Limitations

This study faces several limitations in experimental design and data analysis. Firstly, the experiment was conducted in July, during a period of elevated environmental temperatures, which may have influenced subjective perception evaluations. Future research could address this by conducting studies across different seasons to examine how varying temperatures affect perception. Secondly, the limited number of testing points for measuring soundscape perception within the study site may introduce minor biases into the models. Expanding the number of testing points within and across different parks could improve the accuracy and generalizability of the findings. Thirdly, the study site’s landscape is predominantly composed of natural elements, with a lack of hardscape features and built structures. Future research could explore the visual and acoustic effects of diverse landscape compositions, including built environments, to provide a more comprehensive understanding of soundscape perception in different settings.

Moreover, all participants in this study were young adults aged 20–30. This age group may have a higher tolerance for environmental conditions such as temperature and noise, which could potentially influence the generalizability of the findings to other age groups. Future research should consider a broader demographic to assess how age-related differences might impact soundscape perception and environmental tolerance. The inclusion of a more diverse participant pool could enhance the applicability of the results across different population segments.

5. Conclusions

This study, centered on human subjects, employed mathematical modeling to depict the impact mechanisms of environmental audiovisual characteristics on human subjective emotional perception and activities. By integrating this understanding with site-specific environmental information, the study visualized the perceptual quality issues related to audiovisual senses in the studied park. Based on these findings, it proposed leveraging audiovisual interactions to drive sustainable development in parks, offering scientific foundations and flexible strategies to optimize the audiovisual perception quality of urban parks with diverse functionalities. This approach provides a new pathway for the sustainable development of urban green spaces near expressways, aimed at promoting residents’ health and well-being. The predictive models developed in this research provide direct design guidance for shaping high-quality urban parks without the need for professional interpretation, thus reducing potential deviations caused by technical and professional language barriers in design objectives. Additionally, they expand the research and practical pathways from single-element perception to the coupling of multisensory elements in landscape perception studies, contributing empirically to the design of urban open spaces that are human-centered and based on multisensory perception. The feasibility of the research methods employed has been validated, suggesting their broad applicability to other cities and types of parks, thereby laying the groundwork for accumulating universally applicable experiences in optimizing urban park audiovisual quality.