1. Introduction
The creation of sustainable landscapes is fundamental to advancing the landscape architecture industry and achieving ecological, social, and economic resilience. Landscape performance (LP) research, grounded in principles of spatial design, employs scientifically robust methods to identify and quantify the multidimensional outcomes of landscape interventions. This provides a critical foundation for optimizing spatial configurations, which refer to the arrangement and organization of objects or elements within a given space, influencing how that space is used and perceived [1]. Such configurations are essential for enhancing environmental quality, and fostering sustainable practices. A central challenge in contemporary research lies in developing systematic, objective frameworks for evaluating complex landscape systems [2,3].
The landscape environment is a complex, open system where ecological processes, spatial forms, and human perceptions interact in interdependent layers. The ecological environment forms its foundation, functioning as the biophysical substrate that regulates material cycles and energy flows, with a core focus on preserving ecological integrity. Built environments arise from the interplay between natural systems and human interventions, manifesting as spatial forms that meet human demands for functionality and experience. Humans, in turn, interact with the landscape through these built environments, where spatial configurations influence perception and behavior, while human feedback further shapes and optimizes the landscape [4]. Therefore, the characteristics of natural processes, physical spaces, and human perceptions collectively reflect the state of the landscape environment, and their coordination is crucial for sustainable development.
To evaluate LP comprehensively, previous work has developed diverse indicator systems tailored to specific goals, employing various measurement tools to quantify performance metrics [5]. These tools can be categorized into two types: subjective evaluation data and objective measurement data. Subjective data are typically collected via questionnaires, interviews, and behavioral records, reflecting people’s perceptions, experiences, and satisfaction with the landscape. In contrast, objective data are used to analyze the physical and ecological characteristics of the landscape environment, such as biodiversity monitoring, carbon emission calculations, land-use analysis, and spatial data processing [6,7]. Together, these data types provide multidimensional insights for LP evaluation.
However, the complexity of the landscape environment presents challenges for performance evaluation. Trade-offs and synergies often exist among indicators. For instance, as Luo et al. [8] demonstrated, built environments often face inherent tensions between environmental, economic, and social sustainability objectives. Conventional analytical approaches, which rely on univariate or bivariate methods, fail to capture the intricate interdependencies among multiple drivers of LP. To address these limitations, performance evaluation must extend beyond efficient measurement of landscape components to rigorously test the validity of variable selection and elucidate underlying mechanisms. Traditional descriptive statistics, while useful, lack the depth to fully unravel the complexities of LP.
Structural equation modeling (SEM) offers a transformative approach for analyzing multifactorial relationships within LP research. This multivariate method enables researchers to model causal pathways among variables, identify key drivers of LP, and disentangle direct and indirect impact mechanisms. By doing so, SEM clarifies how design strategies shape high-performance landscapes, bridging the gap between theoretical frameworks and empirical validation. Emerging applications of SEM in LP research demonstrate its potential to revitalize analytical and evaluative practices. However, as the field grows, a systematic synthesis of current progress, challenges, and future directions is urgently needed to consolidate knowledge and guide innovation.
This review synthesizes advancements in SEM applications for LP research, with three objectives: (1) to analyze global trends in SEM adoption across LP studies; (2) to evaluate predominant research dimensions, methodologies, and data types; and (3) to propose expanded applications of SEM for addressing evolving challenges in landscape sustainability. By mapping the state of the art and identifying critical gaps, this work aims to strengthen theoretical foundations and methodological rigor in LP assessment, ultimately supporting the design of resilient, high-impact landscapes.
1.1. Landscape Performance
Landscape performance (LP) refers to the extent to which landscape architectural design practices achieve their intended goals while adhering to sustainable development principles [9]. LP research emerged in the 1990s and gained momentum with the growing emphasis on sustainability, particularly the ecological performance of completed projects. The Sustainable SITES Initiative was established to provide authoritative standards for advancing sustainable landscape design. In the 21st century, as social–ecological systems grew increasingly complex, the Landscape Architecture Foundation (LAF) launched the Landscape Performance Series (LPS) research program in 2010 to enhance the quality and sustainability of urban landscape practices. By 2014, LP had become a standard topic within the Council of Educators in Landscape Architecture (CELA).
LP research is structured around environmental, social, and economic goals (Figure 1), offering a comprehensive, precise, and scientifically grounded framework for refining performance evaluation indicators and methods [10]. The LPS, an online platform, comprises four key sections: case studies, benefit toolkit, fast fact library, and scholarly works on LP. As of 2024, the case library features over 200 exemplary built projects, serving as a valuable resource for designers, agencies, and advocates seeking quantitative metrics and methodologies. Existing research primarily explores how sustainable landscape design strategies generate multifaceted impacts, particularly in ecosystem services and human health and well-being [11,12]. However, most studies focus on selecting evaluation indicators or merely highlighting the sustainability features of completed projects. These analyses often rely on descriptive summaries of performance outcomes, lacking deeper insights into the specific design strategies or elements driving particular benefits.
The advent of digital technology has transformed LP research from linear, single-dimensional benefit assessments to multi-dimensional data correlation studies. This shift has seen research methodologies evolve from graphical descriptions to model-based interpretations, with data processing advancing from basic factor superposition to dynamic linkage analysis. Initially rooted in traditional landscape evaluation, LP research has broadened its scope due to the expansion of research categories and advancements in evaluation techniques. This progression underscores the collaborative innovation in needs, perspectives, and methodologies within the landscape architecture research knowledge system. The core objective of this research extends beyond assessing landscape quality; it aims to establish causal relationships between design strategies and their outcomes, thereby offering targeted insights to enhance design efficacy. Consequently, contemporary LP research demands sophisticated algorithms and robust modeling frameworks supported by advanced methodological platforms.
1.2. Structural Equation Modeling
SEM can be traced back to Spearman’s (1904) introduction of latent variables into data analysis [13] and Wright’s (1934) development of path analysis in genetics [14]. Building on these foundations, Jöreskog (1978) formulated the preliminary concept of SEM [15]. Since then, SEM has evolved through interdisciplinary research integrating multiple disciplinary systems, such as psychology, sociology, behavioral science, and ecology, earning recognition as one of the three major statistical advancements in recent decades [16]. Leveraging diverse modeling methods, such as factor analysis, path analysis, and hierarchical models, SEM enables researchers to analyze complex relationships among variables, including multiple pathways and feedback loops. This capability allows for the modeling, testing, and refinement of intricate theoretical frameworks, offering valuable insights into potential relationships within observational datasets.
At its core, SEM evaluates the degree of fit between the covariance matrix of a theoretical model and the covariance matrix derived from empirical data. Researchers begin by constructing a theoretical model based on prior knowledge, and then input observational data into a path diagram for parameter estimation and goodness-of-fit analysis. The model is iteratively refined based on fit evaluation results until an optimal fit is achieved [17,18] (Figure 2). Compared to classical multivariate analysis methods such as factor analysis and principal component analysis, SEM can represent complex attributes that cannot be directly measured but are inferred from observed variables [19]. Additionally, SEM quantifies direct, indirect, and total associations among multiple variables [20], making it a powerful tool for exploring the intricate interactions within landscape environments.
As illustrated in Figure 3, SEM encompasses several analytical methods, including correlation mechanism analysis (CMA), mediation effect analysis (MEA), and multi-group analysis (MGA). Correlation mechanism analysis, also known as Confirmatory Factor Analysis (CFA), assesses the structural validity of measurement models by determining whether observed indicators effectively reflect one or more latent variables. This analysis focuses on factor correlations rather than causal relationships. Mediation effect analysis examines how a mediating variable transmits the influence between a cause and its outcome. SEM’s path analysis capabilities address limitations of traditional regression methods, such as large standard errors, by simultaneously estimating all model parameters [21,22]. Multi-group analysis evaluates the consistency of factor structures across different groups and identifies significant differences in path parameters. This approach tests whether a hypothesized model remains stable across diverse samples, enhancing the generalizability of theoretical frameworks.
2. Methods
2.1. Article Selection
This study adopts a systematic literature review approach to search and extract relevant existing works of literature [23,24]. By employing the terms “landscape performance” and “structural equation model” in a search, we collected studies between 2010 and 2024 in the Web of Science Core Collection and Scopus databases. We also considered their alternative terms coming from synonyms, related terms, and broader or narrower terms, where “landscape space” can be extended to include “urban space”, “built environment”, and “ecological space”; “performance” can be replaced by “quality”, “efficiency”, “benefit”, and “effect”. The Boolean search term “OR” was used to incorporate congeneric keywords mentioned above in the search string. After removing the duplicates, we examined the titles and abstracts of the publications and removed irrelevant studies, reviews, and theoretical studies. We obtained the full texts for the remaining publications and read the papers. Records that met the predetermined screening criteria were included in the literature analysis list. After the screening, a total of 245 valid publications were identified that applied SEM techniques in the field of LP research. Figure 4 illustrates the systematic research framework, together with the results in the main steps.
2.2. CiteSpace Analysis
Following the acquisition of the research literature, CiteSpace [25] was used for bibliometric analysis and visualizing a wide range of connectivity (clusters) between different papers, thereby uncovering the underlying patterns and thematic structures within the paper database. As a Java-based program, CiteSpace demonstrates exceptional capabilities in co-citation and co-occurrence analysis. This analysis identifies clusters of co-cited references and creates networks of co-occurring keywords. CiteSpace generates color-coded bibliographic data maps that offer profound visual representations of latent connections and research dynamics [26]. As a result, the evolutionary trajectory of landscape performance research can be revealed, which elucidates critical keywords within publications, and identifies prevailing research themes and conceptual frameworks within the discipline.
2.3. Thematic Analysis
After analyzing the general trends in the application of SEM to LP research, we extracted the following information from those publications, categorizing and coding each record according to its attributes, and all information was organized into Microsoft Excel 2021: (1) general information of studies; (2) LP research dimensions; (3) SEM analytical methods; and (4) data types. Finally, the thematic analysis results were visualized by Origin 2025 [27].
Specifically, “general information” includes basic attributes of the publications, such as authors, publication year, journal, title, keywords, and abstract. The geographic locations and research scales of each study were also documented.
Based on the research objectives and scope, the LP research dimensions are categorized into three themes, ecological environment performance, built environment performance, and cognitive environment performance, reflecting the complexity of landscape systems. The “ecological environment performance” dimension was ascribed to articles that focus on natural processes, ecosystem services, and environmental sustainability. The “built environment performance” dimension corresponded to articles that evaluate the physical composition of spaces and their interactions with human activities, emphasizing land-use patterns, spatial configurations, and the functionality of infrastructure in influencing landscape benefits. Lastly, the “cognitive environment performance” was matched to articles that address human perceptions, satisfaction, mental health, and socio-cultural values.
In terms of SEM application methods, we identified three main types: correlation mechanism analysis, mediation effect analysis, and multi-group analysis. In several cases, the reviewed articles employed multiple methods and, therefore, were accordingly ascribed to “multi-type analyses”. These distinctions highlight the methodological diversity in LP research and provide insights into the strengths and limitations of each approach.
The classification of data types primarily distinguishes between subjective evaluation data and objective measurement data. Depending on research objectives, these categories are further divided into techniques such as surveys, behavioral observations, field sampling, and open-source datasets. Several studies used a combination of both subjective and objective data; therefore, we classified them into “subjective and objective combination”. This classification framework effectively reveals how different data sources complement each other in evaluating LP, offering a more comprehensive perspective for multidimensional research.
3. Results
SEM was first applied to LP research in 2011 [28], aligning with the launch of the Landscape Performance Series (LPS) program. This seminal study examined linkages between land use and commuting behavior, setting a precedent for SEM’s utility in LP analysis. Since then, annual publications have grown steadily, peaking in 2022 (Figure 5). By November 2024, 245 studies were identified, predominantly published in Urban Forestry & Urban Greening (n = 27) and Landscape and Urban Planning (n = 25). Table 1 presents the number and proportion of articles grouped by peer-reviewed journal based on the number of articles published in each journal within the dataset.
Geographically, China and the USA emerged as primary research hubs, reflecting their leadership in sustainability-focused landscape research. Collaborative networks spanned Europe, with notable partnerships between the Netherlands and the USA, and England and Australia (Figure 6a,b). Over 59.5% of studies focused on city/regional scales, while global and block-level analyses were less common. Asia, North America, and Europe dominated research coverage, with Oceania underrepresented.
3.1. Preliminary Bibliometric Analysis Visualized Using CiteSpace
The chronological graph presented by CiteSpace 6.3.R1 illustrates the research priorities in the historic garden field during the 2010–2024 period. The focus of research on LP ranges from “biodiversity”, “conservation”, and “sustainability” (Figure 7a,b) to “urban form”, “behaviour”, and “community”, reflecting a paradigm shift toward balancing ecological preservation with urban development imperatives while foregrounding urban environments’ public health implications. Moreover, according to the clustering analysis of the word cloud (Figure 7c), investigations into the LP have primarily focused on three topics, namely, travel behavior (Cluster #0), neighborhood park (Cluster #1) and environmental cognition (Cluster #2), which have received considerable attention in recent years.
The betweenness centrality (pink circles), shown in Figure 7a,b, measures the importance of nodes in the network and highlights key nodes that act as bridges connecting different research clusters. The thickness of the pink circles reflects the magnitude of the centrality value, with thicker circles denoting greater influence within the network. Keywords such as “structural equation model”, “conservation”, and “travel behaviour” exhibit high betweenness centrality, indicating their pivotal roles in fostering interdisciplinary connections, integrating ecological, behavioral, and urban research, thereby advancing the field of LP studies.
3.2. Research Dimensions
Cognitive environment performance studies dominated LP research (64.8%), followed by built environment (24.3%) and ecological environment (10.9%) analyses (Figure 8). Early cognitive studies (2011–2014) emphasized behavioral perception feedback, shifting toward preference evaluation and physiological health assessment post 2015. Built environment research gained traction after 2017, evolving from spatial resource allocation efficiency to spatial ontology composition mechanisms by 2023. In contrast, ecological performance studies remained limited, likely due to prior dominance by ecological sciences. Recent urbanization pressures, however, have spurred landscape architecture interest in topics like urban park carbon dynamics.
3.3. Analytical Methods
Mediation effect analysis (MEA) was the most prevalent SEM technique (45%), followed by correlation mechanism analysis (CMA, 43%). Both methods are often combined with multi-group analysis (MGA) and multitype measurements to enhance robustness (Figure 9). The popularity of SEM methods has evolved over time. CMA dominated early studies, but MEA surged post 2015, reflecting demand for causal pathway exploration. Since 2020, MGA adoption increased, driven by needs for cross-sample validation and model precision.
Cross-dimensional analysis (Figure 10) revealed the flexibility of SEM in addressing diverse research needs by adapting the analytical approach to specific objectives. MEA is frequently applied to investigate the indirect effects between cognitive and ecological environment performance, whereas CMA supports built environment measurement frameworks. MGA models are frequently utilized as extensions of these methods, enabling validation of theoretical models across varying landscape spatial types and demographic characteristics. The application of SEM across multiple dimensions facilitates a comprehensive examination of complex relationships among variables. This integrative approach provides valuable insights into the composite functions and correlation mechanisms of landscape environments, advancing the understanding of their multifaceted dynamics.
3.4. Data Types
The development of data sources follows a similar pattern of development as research dimensions (Figure 11). Evaluation data based on cognitive analysis have consistently been the dominant source, accounting for 56% of the total. In the early stages, data collection was constrained by the reliance on standard social science methods, such as surveys, interviews, and observational records, to obtain first-hand evaluation data. Objective environmental data, on the other hand, were primarily derived from publicly available government audit reports.
Since 2017, the rapid development and widespread adoption of digital analytics have enabled more precise measurements of built environment performance, resulting in the emergence of increasingly segmented data hotspots. Furthermore, among the studies reviewed, 80 articles utilized data from multiple sources, with 53 combining subjective and objective data—an approach that has gained significant popularity in recent years.
Notably, despite widespread concerns about the subjectivity of evaluation data and its potential to generate inaccurate results [29], the use of this method in LP research has not declined, nor has it been swiftly replaced by objective environmental data. Instead, evaluation data remain the preferred choice for most researchers. This preference underscores the unique advantages of human subjective perception and behavioral data in practical applications. These data types, supported by ongoing technological advancements, demonstrate a capacity for continuous refinement. Together, subjective and objective data exhibit a mutually reinforcing relationship, contributing to the evolution of research methodologies in the field.
4. Discussion
4.1. Research Dimensions of SEM-Based LP Research
Through a comprehensive review and synthesis of 245 studies, this paper organizes the dimensions of LP research, the application of SEM methods, and the data types involved (Table S1). LP has long been a central theme in research and practice, evolving through three distinct stages: ecological environment performance, built environment performance, and cognitive environment performance.
-
(1). Ecological environment performance
Ecological environment performance research examines interactions among organisms or events within landscape ecosystems and their mechanisms of interaction with the broader environment [30]. These mechanisms include direct and indirect relationships, feedback effects, community changes, and responses to ecosystem processes. SEM’s strength lies in its ability to model these multidimensional networks, enabling simultaneous analysis of influencing factors, the identification of direct and indirect effects, and the differentiation between multiple causal pathways. Moreover, SEM allows for the estimation and comparison of pathway strengths, offering deeper insights into causal relationships [31].
In ecological performance studies, SEM is often employed to explore indirect relationships among observed variables. Typically, these models rely on observed variables measured through grid-based sampling plots—a traditional method in ecological research. However, the adoption of latent and composite variables (where response variables are represented by observed variables) remains relatively limited [32]. For example, Yin et al. [33] demonstrated that SEM’s capacity to disentangle green space structure indirectly influences particulate matter distribution by altering micrometeorological conditions, offering insights beyond the numerical relationships typically provided by traditional correlation analyses [34].
-
(2). Built Environment Performance
The built environment, defined as the intentionally designed urban spatial framework supporting human activities [35], plays a vital role in enhancing urban vitality, health, safety, and sustainability [36]. Built environment performance depends largely on spatial composition, where distinct configurations lead to divergent outcomes. This “composition-performance” paradigm highlights the necessity of analyzing spatial mechanisms and assessing how configurations optimize performance.
Most studies employ parametric methods to quantify spatial components, categorize influencing factors, and build hierarchical index systems via SEM. This approach enables researchers to identify the pathways, directions, and strengths of factors affecting performance [37]. However, indicator systems vary across studies due to the built environment's complexity and diverse research objectives. Researchers typically derive variables and causal pathways from literature reviews.
Despite progress in understanding the built environment, most studies remain exploratory and lack fully developed theoretical models. Consequently, whether these models can be generalized to other contexts remains uncertain. To bridge this gap, longitudinal evaluations and large-scale datasets are needed to establish systematic indicator systems [38,39], facilitating broader model validation and application.
-
(3). Cognitive Environment Performance
Cognitive environment performance represents the most studied dimension, closely tied to advances in habitat and behavioral sciences. Unlike ecological and built environment performance, which focus on the physical attributes, cognitive research emphasizes landscape perception as a core subject. It explores the relationships among psychological needs, motivations, and behaviors [40,41], typically through three research areas: cognitive preference evaluation, physiological health assessment, and behavioral perception feedback.
Landscape perception encompasses sensory-driven environmental information processing and subsequent emotional interpretation [42]. This process operates at three levels: cognitive (decoding spatial forms), behavioral (fulfilling functional needs through environmental interaction), and emotional (assigning meaning to spaces to cultivate place attachment) [43]. These levels collectively mediate human–environment interactions [44]. However, perceptual dynamics are inherently abstract and latent, necessitating measurable variables to operationalize these constructs. SEM excels in modeling such latent variables, explaining its dominance in cognitive performance research due to methodological robustness and reliability.
Research in this domain extensively integrates psychological and behavioral theories, including Attention Restoration Theory [45], Theory of Planned Behavior [46], and Self-Determination Theory [47]. Landscape scholars have adapted these models by incorporating factors specific to their research objectives, revealing how spatial and user traits shape perceptions [48]. A critical gap persists, however, as most models focus on perception-to-outcome pathways while neglecting how physical features directly trigger perceptions [49]. This oversight hinders design applications. Bridging this gap requires integrating objective landscape indicators with perceptual models to establish actionable design principles for improving landscape design quality [50].
-
(4). Summary
SEM has been extensively utilized in various dimensions of LP research owing to its adaptability to diverse research contexts. Through path diagrams, it effectively illustrates direct and indirect relationships between observed and latent variables, facilitating a comprehensive understanding of complex interactions.
The scale of LP research has evolved from localized contexts, including parks, campuses, and communities, to encompass city, regional, and national levels. Small-scale studies examine specific factors and mechanisms affecting performance within localized areas [51]. Nevertheless, these studies frequently encounter geographic constraints as their findings typically apply only to similar environments and may not be generalizable to different spatial contexts. To address this limitation, developing LP models with cross-environmental generalization capabilities is essential.
In contrast, large-scale studies at municipal or national levels typically emphasize theoretical model generalizability. These studies, however, often neglect detailed physical landscape characteristics and specific respondent locations [52]. Consequently, while planners and policymakers utilize SEM findings at these scales to identify priority factors and develop macro-level strategies, the findings lack practical insights at spatial or physical levels. Moreover, environmental variables frequently demonstrate significant spatial heterogeneity, exhibiting varying effects across different geographic contexts [53].
To overcome these challenges, future research should incorporate Geographic Information Systems (GISs) to integrate objective environmental indicators into LP studies [54]. The creation of hotspot maps combining cognitive and environmental feature data enables researchers to verify and analyze both direct and integrated variable effects. Furthermore, local spatial autocorrelation models can help identify spatial heterogeneity in variable effects [55], facilitating the recognition of specific geographic patterns and providing targeted, actionable recommendations for LP enhancement.
4.2. Analytical Methods for SEM-Based LP Research
Based on the application of SEM in developing theoretical models for LP research, this section explores SEM analytical methods to address the multiple requirements of such studies. SEM enables systematic analysis of LP-influencing factor relationships, mediation effects, and cross-group comparisons.
-
(1). Correlation mechanism analysis
The landscape environment functions as a complex system in which LP is shaped by interactions among various subsystems, including spatial patterns, ecological communities, socio-economic factors, and resource distribution. Correlation mechanism analysis (CMA) aims to decipher these interactions and their effects on LP. Researchers typically develop LP models incorporating both latent and observed variables based on prior knowledge. By leveraging SEM, researchers could quantify variable relationships and test model fit against empirical data. This methodology has been widely adopted in LP research, particularly in studies of built environment performance [56]. Moreover, Crawford et al. [57] found hierarchical distinctions among influencing factors, where some are governed by higher-order latent traits. Accordingly, hierarchical second-order structural models can be constructed to identify the core dimensions of influencing factors, integrating a well-balanced set of elements within the measurement framework [58]. This approach enhances the understanding of the relationships between the observed indicators and the latent variables, resulting in a more robust model with improved explanatory and predictive capabilities.
-
(2). Mediation effect analysis
Mediation effect analysis (MEA) explores how independent variables indirectly influence dependent variables through mediators. Uncovering these mechanisms advances the theoretical understanding of variable interplay, particularly in bridging subjective perceptions and objective environments. For instance, the Stimulus–Organism–Response (SOR) theory explains how the objective environment influences behavior. It posits that when individuals encounter stimuli from the built environment, they form subjective perceptions that shape their judgments and decisions [59]. Contemporary studies integrate objective metrics and subjective perceptions to demonstrate how environmental attributes mediate psychological states, which, in turn, drive decision-making and actions [60,61]. MEA requires theory-grounded hypotheses about mediation pathways, avoiding overreliance on statistical significance. Rigorous mediation studies should prioritize theoretical coherence and contextual relevance to ensure robust conclusions.
-
(3). Multi-group analysis
Multi-group analysis (MGA) assesses the validity of theoretical models across different samples or contexts [62]. While models may fit specific datasets well, their generalizability to broader populations remains uncertain. LP studies face variability in individual perceptions such as gender, age, or socio-economic status. Similarly, landscape design objectives and LP mechanisms often differ across regions and spatial typologies. To address this, researchers classify samples into subgroups using variables like spatial typology or demographic traits. Then, MGA was conducted to examine the heterogeneity among subgroups. Guo [49], for instance, demonstrated that age and gender mediate how audiovisual interactions affect perceived restorative quality in urban parks. Expanding beyond demographics, Li et al. [63] applied MGA to compare landscape typologies (e.g., activity-oriented vs. leisure-oriented parks), uncovering how spatial features and soundscapes jointly shape restorative perceptions. By uncovering both shared patterns and unique differences in LP, MGA contributes to developing theoretical models that are broadly applicable and adaptable to diverse contexts [64].
-
(4). Multitype measurements
Recent advances enable hybrid SEM approaches integrating machine learning algorithms such as artificial neural networks (ANNs) [65,66], random forests (RFs) [67], support vector machines (SVMs) [68], and Markov Chain Monte Carlo (MCMC) [69]. These integrations enhance causal inference by cross-validating outcomes across methods, improving model precision and reliability [70]. Other studies combine SEM with Importance–Performance Analysis (IPA), leveraging SEM path coefficients as weights to rank indicator importance [71]. Indicators are plotted on a 2D grid (importance vs. performance), where quadrant positions (e.g., high importance/low performance) highlight priorities for landscape management [72]. However, performance scores often derive from subjective satisfaction metrics rather than environmental indicators. Notably, certain environmental indicators have been shown to exhibit threshold effects related to sustainable development goals [73,74]. Future research could merge SEM, threshold analysis, and IPA to set dynamic benchmarks for critical indicators [75], offering evidence-based guidance for policy development and environmental management.
4.3. Data Selection for SEM-Based LP Research
Most studies in LP research predominantly rely on subjective data due to the emphasis on cognitive environment performance. Although subjective survey data offer advantages in terms of simplicity and rapid large-scale data collection, these have limitations regarding scientific rigor and accuracy. Demographic characteristics, especially age, can significantly influence LP mechanisms. MGA within SEM serves as an effective tool for investigating these individual differences. Through the integration of detailed social and demographic variables, researchers can develop targeted landscape design strategies for diverse population groups.
However, survey data may provide limited insights when investigating specific causal relationships and localized parameters underlying the benefits of landscape environments. Palardy et al. [76] recommended conducting qualitative interviews with residents to explore deeper “why” questions. Similarly, Wang and Liu [77] integrated qualitative interviews with extensive field observations of urban open spaces to aid in interpreting quantitative SEM results. This approach, while effective, is more suitable for meso-scale studies, such as those focusing on city districts or specific communities, where qualitative methods like field surveys and interviews can yield detailed insights. However, when applied to large-scale research involving broader spatial extents, such as national or multi-regional studies, the data collection process can become labor-intensive due to the increased complexity and scope.
To address cognitive data biases, scholars increasingly pair surveys with physiological data to enhance the accuracy and reliability of research findings. Tools such as eye-tracking devices, wearable electrocardiograms (ECGs), and skin conductance (EDA) sensors have been employed in LP studies [78]. Additionally, there is an emerging trend toward integrating quantitative and qualitative data for a more comprehensive understanding [79]. For example, Liu et al. [80] employed NDVI to quantify residents’ exposure to greenery, thereby enhancing the comparability of research metrics. Similarly, Wang et al. [81] utilized VR technology to control external variables and standardize landscape environment presentation.
Regarding objective measurements, numerous studies highlight the importance of improving data accuracy in their discussions and future research recommendations. Advanced technologies, including mobile signals, social media analytics [82], high-resolution remote sensing [83], night-time light [84] and street view imagery [85], have enabled multi-source data collection, providing comprehensive and real-time measurements [86]. However, pursuing data accuracy alone is insufficient. Researchers should select appropriate measurement indicators and data sources aligned with their specific research objectives and environmental contexts. Excessive emphasis on high-precision data may lead to a restricted understanding of the research subject, overlooking the inherent complexity and multidimensional characteristics of LP. Future studies should emphasize data diversity and complementarity, striving to balance investigative scope and depth while maintaining accuracy.
Furthermore, contemporary research predominantly relies on cross-sectional data, leaving temporal aspects of landscape performance largely unexplored. Cross-sectional data, collected at specific time points, represents multiple observational units during a single period. Given that landscapes are dynamic systems, LP inherently evolves over time. Landscape architecture projects typically require longer post-construction periods than architectural projects to fully manifest their performance characteristics [87]. However, existing research largely depends on single time-point or short-duration data collection, inadequately capturing landscape environments’ dynamic nature and seasonal variations affecting performance. Longitudinal studies tracking participants across seasons and monitoring environmental conditions through climate variations can provide deeper insights into landscape performance mechanisms and the dynamic interactions between landscape patterns, environmental factors, and socio-economic conditions [55].
4.4. Trends and Recommendations
To date, SEM has gained prominence in LP research for its capacity to model latent variables, quantify direct/indirect effects, and account for measurement errors. Nevertheless, despite its strengths, several limitations and challenges need to be addressed to improve its accuracy and reliability. A significant drawback is that SEM represents correlation rather than proven “causality”. Cross-sectional data collected by researchers are typically collected at the same point in time, resulting in limitations when it comes to inferring causal relationships due to the lack of a temporal sequence. In such cases, the causal relationships between variables are often based on people’s perceptions or on existing research and experience.
Moreover, the influence of scale effects on the results should be considered as small sample sizes or overparameterization at local scales can distort models. While large, high-quality datasets improve robustness, their acquisition remains resource-intensive. Additionally, SEM outputs for a specific study area are often presented as numerical matrices, and spatial analysis is thus required to highlight key areas for targeted improvement. To address these limitations and challenges, we propose three recommendations:
From a research perspective, it is essential to integrate studies on the landscape environment and cognitive environment performance. High-quality landscape design can deliver tangible environmental benefits and observable social benefits. However, as noted earlier, many theoretical models were originally developed for behavioral and psychological studies. The specific mechanisms through which environmental characteristics influence human behavior and mental states remain insufficiently explored. Additionally, landscapes are dynamic systems that evolve over time, making it challenging to establish standardized evaluation benchmarks. To address these issues, longitudinal studies tracking landscape lifecycle phases are recommended to identify construction patterns and evolutionary trends
From a technological perspective, creating an open-source database dedicated to LP research is a critical step. Such a database should facilitate the sharing and categorization of spatial information from research cases, including, but not limited to, three-dimensional geometric data, non-geometric data and performance evaluation data. In addition, SEM-based performance measurement models and their associated datasets should be incorporated into the database. This would enhance research transparency and reproducibility while providing theoretical and data support for researchers in related fields. Furthermore, integrating SEM with GIS offers promising opportunities. Developing spatial analysis algorithms based on SEM could enable the precise spatial identification of key factors affecting LP, thereby improving spatial resolution in the analysis of LP.
From an implementation mechanism perspective, how to integrate landscape space with local regulations and policies so that the research results can be adopted and applied in practice is the next problem that must be faced. Future work could focus on testing the results of SEM by conducting case studies on indicators that exhibit either high correlation or low performance within the SEM outcomes. Currently, discussions on applying SEM results to design practices are limited as much of the research remains in the conceptual validation and modeling phase. The lack of such studies will reduce the true effectiveness and evidence-based value of assessment results in the LP metrics. In terms of evaluating the benefits after implementation, continuous tracking and testing of real cases will validate model efficacy and inform iterative refinements.
5. Conclusions
LP research is crucial for scientifically guiding the design and construction of high-quality, sustainable landscape environments. SEM supports the exploration of underlying mechanisms in landscape environments. This review systematically analyzed 245 studies from two academic databases, revealing that LP research has evolved through distinct stages, each reflecting shifting priorities and methodologies. SEM is widely adopted for addressing latent variables and elucidating complex relationships. However, challenges hinder its practical application in improving landscape quality. These include insufficient attention to spatial location information and temporal evolution characteristics, which limit the applicability of findings in design practices. Additionally, subjectivity in cognitive data and the need for greater precision in objective data constrain SEM-based measurements’ accuracy.
Based on the above conclusions, the following recommendations are made for future work in this area.
-
Integrate landscape and cognitive environment performance: conduct long-term analyses to identify construction models and evolutionary trends over the entire lifecycle.
-
Establish an open-source database: share spatial information and theoretical models, integrating SEM with geographic information technologies to enhance spatial analysis precision.
-
Apply research findings in design practices: validate results through case studies and continuously track the performance of real landscape renewal projects.
There are limitations to this study. Due to the broad scope of landscape performance, there is low homogeneity in the literature on relevant assessment metrics, which made it challenging to conduct a rigorous literature screening and cross-sectional comparisons using quantitative review methods such as meta-analysis or evidence gap mapping. Therefore, we adopted a qualitative review method, in which we purposively searched and analyzed the literature according to the research question, which may lead to a certain degree of subjectivity in the research results. In future work, a more precise research scope needs to be defined with a view to providing a more comprehensive and objective theoretical basis for landscape performance assessment.
With advancements in statistical analysis and data science, SEM will continue to evolve, shifting LP research from superficial descriptions to deeper mechanistic insights. Methodologies are expected to progress from mathematical statistics to spatial information statistical analysis, fostering refined measurement models and comprehensive performance quantifications. We anticipate that the synergistic advancement of methodological research and practical applications will lead to more case studies integrating SEM into LP analysis and performance-enhancing design. This will expand SEM’s application scope and advance LP’s theoretical development, enhancing the theory’s practical operability.
Conceptualization, X.H. and Z.L.; methodology, X.H.; software, X.H.; validation, H.C., M.Y. and Y.S.; formal analysis, X.H.; investigation, X.H.; resources, X.H.; data curation, H.C., M.Y. and Y.S; writing—original draft preparation, X.H.; writing—review and editing, Z.L.; visualization, X.H., H.C., M.Y. and Y.S.; supervision, Z.L.; funding acquisition, Z.L. All authors have read and agreed to the published version of the manuscript.
No new data were created or analyzed in this study.
The authors declare no conflicts of interest.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Enlarge this image.Figure 1. Indicator system established by the Landscape Performance Series research program.
Enlarge this image.Figure 3. Basic analysis method of SEM (where measured variable is represented by a rectangle, latent variable is represented by an ellipse, and the arrow indicates the direction of influence).
Enlarge this image.Figure 5. Number of publications on the application of SEM in landscape performance research from 1 January 2010 to 30 November 2024.
Enlarge this image.Figure 6. (a) Network map of countries; (b) geographic areas covered by 245 publications.
Enlarge this image.Figure 7. (a) Chronological distribution of keyword co-occurrence frequencies; (b) connectivity of keyword co-occurrence frequency; (c) keyword cluster ranking processed in Citespace 6.3.R1.
Enlarge this image.Figure 9. Proportion of SEM analysis methods applied in the literature retrieved.
Number and proportion of articles in the top ten journals (ranked by the number of articles published).
| Journal | n (%) |
|---|---|
| Urban Forestry and Urban Greening | 27 (11) |
| Landscape and Urban Planning | 25 (11) |
| Transportation Research Part A: Policy and Practice | 12 (5) |
| Sustainable Cities and Society | 11 (4) |
| Landscape Ecology | 11 (4) |
| Transportation Research Part D: Transport and Environment | 10 (4) |
| Journal of Transport Geography | 9 (4) |
| International Journal of Environmental Research and Public Health | 9 (4) |
| Cities | 9 (4) |
| Travel Behaviour and Society | 8 (3) |
Supplementary Materials
The following supporting information can be downloaded at:
References
1. Zhang, Y.; Jin, Y.; Steemers, K.; Cao, K. 3.20—The Study of the Effects of Built Form on Pedestrian Activities: A GIS-Based Integrated Approach. Comprehensive Geographic Information Systems; Huang, B. Elsevier: Oxford, UK, 2018; pp. 330-344. ISBN 978-0-12-804793-4
2. Barnes, M. Landscape Performance in Practice: Insights and Recommendations. Landsc. Rev.; 2024; 20, pp. 18-26. [DOI: https://dx.doi.org/10.34900/lr.v20i2.1244]
3. Yang, B.; Li, S.; Binder, C. A Research Frontier in Landscape Architecture: Landscape Performance and Assessment of Social Benefits. Landsc. Res.; 2016; 41, pp. 314-329. [DOI: https://dx.doi.org/10.1080/01426397.2015.1077944]
4. Marsh, W.M. Landscape Planning: Environmental Applications; Wiley: New York, NY, USA, 2005.
5. Shen, Z.; Peng, X.; Du, C.; Kim, M. Quantifying Sustainability and Landscape Performance: A Smart Devices Assisted Alternative Framework. Sustainability; 2023; 15, 13239. [DOI: https://dx.doi.org/10.3390/su151713239]
6. Lin, H.; Wang, J.-B.; Zhang, X.; Hu, F.; Liu, J.; Hong, X.-C. Historical Sensing: The Spatial Pattern of Soundscape Occurrences Recorded in Poems between the Tang and the Qing Dynasties amid Urbanization. Humanit. Soc. Sci. Commun.; 2024; 11, 730. [DOI: https://dx.doi.org/10.1057/s41599-024-03251-7]
7. Feng, L.; Wang, J.; Liu, B.; Hu, F.; Hong, X.; Wang, W. Does Urban Green Space Pattern Affect Green Space Noise Reduction?. Forests; 2024; 15, 1719. [DOI: https://dx.doi.org/10.3390/f15101719]
8. Luo, Y.; Li, M.-H. A Study of Landscape Performance: Do Social, Economic and Environmental Benefits Always Complement Each Other?. Landsc. Archit. Front.; 2014; 2, pp. 42-56.
9. Ndubisi, F.; Whitlow, H.; Deutsch, B. Landscape Performance: Past, Present, and Future. Landsc. Archit.; 2015; 1, pp. 40-51. [DOI: https://dx.doi.org/10.14085/j.fjyl.2015.01.0040.12]
10. Yang, Y.; Lin, G. The development, connotations, and interests of research on landscape performance evaluation for evidence-based design. Landsc. Archit. Front.; 2020; 8, pp. 74-83. [DOI: https://dx.doi.org/10.15302/J-LAF-1-020023]
11. Chen, Z. Assessing the Impact of High-density High-heterogeneity Urban District Landscape on Psychological Health and Optimizing via Evidence-based Design. Landsc. Archit.; 2018; 25, pp. 106-111. [DOI: https://dx.doi.org/10.14085/j.fjyl.2018.01.0106.06]
12. Wang, J.; Zhu, A.; Wang, M. Correlating Physical Forms of Riparian Zones in Urban Parks with Effective Eco-services Provision of Water Self-purification: A Case Study of Shanghai Mengqing Park. Landsc. Archit.; 2016; 8, pp. 118-127. [DOI: https://dx.doi.org/10.14085/j.fjyl.2016.08.0118.10]
13. Spearman, C. The Proof and Measurement of Association between Two Things. Am. J. Psychol.; 1904; 15, pp. 72-101. [DOI: https://dx.doi.org/10.2307/1412159]
14. Wright, S. The Method of Path Coefficients. Ann. Math. Stat.; 1934; 5, pp. 161-215. [DOI: https://dx.doi.org/10.1214/aoms/1177732676]
15. Mueller, R.O. Basic Principles of Structural Equation Modeling; Springer Texts in Statistics; Springer: New York, NY, USA, 1996; ISBN 978-1-4612-8455-0
16. Hershberger, S.L. The Growth of Structural Equation Modeling: 1994–2001. Struct. Equ. Model.; 2003; 10, pp. 35-46. [DOI: https://dx.doi.org/10.1207/S15328007SEM1001_2]
17. Peng, J.; Yan, S.; Strijker, D.; Wu, Q.; Chen, W.; Ma, Z. The Influence of Place Identity on Perceptions of Landscape Change: Exploring Evidence from Rural Land Consolidation Projects in Eastern China. Land Use Policy; 2020; 99, 104891. [DOI: https://dx.doi.org/10.1016/j.landusepol.2020.104891]
18. Sharafatmandrad, M.; Khosravi Mashizi, A. Visual Value of Rangeland Landscapes: A Study Based on Structural Equation Modeling. Ecol. Eng.; 2020; 146, 105742. [DOI: https://dx.doi.org/10.1016/j.ecoleng.2020.105742]
19. Spasojevic, M.J.; Grace, J.B.; Harrison, S.; Damschen, E.I. Functional Diversity Supports the Physiological Tolerance Hypothesis for Plant Species Richness along Climatic Gradients. J. Ecol.; 2014; 102, pp. 447-455. [DOI: https://dx.doi.org/10.1111/1365-2745.12204]
20. Grace, J.B.; Anderson, T.M.; Seabloom, E.W.; Borer, E.T.; Adler, P.B.; Harpole, W.S.; Hautier, Y.; Hillebrand, H.; Lind, E.M.; Pärtel, M. et al. Integrative Modelling Reveals Mechanisms Linking Productivity and Plant Species Richness. Nature; 2016; 529, pp. 390-393. [DOI: https://dx.doi.org/10.1038/nature16524]
21. Iacobucci, D.; Saldanha, N.; Deng, X. A Meditation on Mediation: Evidence That Structural Equations Models Perform Better Than Regressions. J. Consum. Psychol.; 2007; 17, pp. 139-153. [DOI: https://dx.doi.org/10.1016/S1057-7408(07)70020-7]
22. Zhao, X.; Lynch, J.G., Jr.; Chen, Q. Reconsidering Baron and Kenny: Myths and Truths about Mediation Analysis. J. Consum. Res.; 2010; 37, pp. 197-206. [DOI: https://dx.doi.org/10.1086/651257]
23. Moher, D.; Liberati, A.; Tetzlaff, J.; Altman, D.G. The PRISMA Group. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med.; 2009; 6, e1000097. [DOI: https://dx.doi.org/10.1371/journal.pmed.1000097]
24. Tranfield, D.; Denyer, D.; Smart, P. Towards a Methodology for Developing Evidence-Informed Management Knowledge by Means of Systematic Review. Br. J. Manag.; 2003; 14, pp. 207-222. [DOI: https://dx.doi.org/10.1111/1467-8551.00375]
25. Chen, C. CiteSpace II: Detecting and Visualizing Emerging Trends and Transient Patterns in Scientific Literature. J. Am. Soc. Inf. Sci. Technol.; 2006; 57, pp. 359-377. [DOI: https://dx.doi.org/10.1002/asi.20317]
26. Waltman, L.; van Eck, N.J.; Noyons, E.C.M. A Unified Approach to Mapping and Clustering of Bibliometric Networks. J. Informetr.; 2010; 4, pp. 629-635. [DOI: https://dx.doi.org/10.1016/j.joi.2010.07.002]
27. Origin. 2025; Available online: https://www.originlab.com/getstarted (accessed on 25 February 2025).
28. Van Acker, V.; Witlox, F. Commuting Trips within Tours: How Is Commuting Related to Land Use?. Transportation; 2011; 38, pp. 465-486. [DOI: https://dx.doi.org/10.1007/s11116-010-9309-6]
29. Zhou, Y.; Yang, L.; Yu, J.; Guo, S. Do Seasons Matter? Exploring the Dynamic Link between Blue-Green Space and Mental Restoration. Urban For. Urban Green.; 2022; 73, 127612. [DOI: https://dx.doi.org/10.1016/j.ufug.2022.127612]
30. Eisenhauer, N.; Bowker, M.A.; Grace, J.B.; Powell, J.R. From Patterns to Causal Understanding: Structural Equation Modeling (SEM) in Soil Ecology. Pedobiologia; 2015; 58, pp. 65-72. [DOI: https://dx.doi.org/10.1016/j.pedobi.2015.03.002]
31. Grace, J.B. Structural Equation Modeling and Natural Systems; Cambridge University Press: Cambridge, UK, 2006; ISBN 978-0-521-54653-9
32. Zhu, Y.; Liu, C.; Peng, C.; Zhou, X.; Xie, B.; Li, T.; Li, P.; Zou, Z.; Tang, J.; Liu, Z. Applications of Structural Equation Modeling in Plant Functional Trait Research. Environ. Rev.; 2024; 32, pp. 525-538. [DOI: https://dx.doi.org/10.1139/er-2023-0128]
33. Yin, Z.; Zhang, Y.; Zhang, R.; Chen, G.; Cong, Y.; Ma, K. Structure of an Urban Green Space Indirectly Affects the Distribution of Airborne Particulate Matter: A Study Based on Structural Equation Modelling. Urban For. Urban Green.; 2022; 72, 127581. [DOI: https://dx.doi.org/10.1016/j.ufug.2022.127581]
34. Chen, X.; Pei, T.; Zhou, Z.; Teng, M.; He, L.; Luo, M.; Liu, X. Efficiency Differences of Roadside Greenbelts with Three Configurations in Removing Coarse Particles (PM10): A Street Scale Investigation in Wuhan, China. Urban For. Urban Green.; 2015; 14, pp. 354-360. [DOI: https://dx.doi.org/10.1016/j.ufug.2015.02.013]
35. Cervero, R.; Kockelman, K. Travel Demand and the 3Ds: Density, Diversity, and Design. Transp. Res. Part Transp. Environ.; 1997; 2, pp. 199-219. [DOI: https://dx.doi.org/10.1016/S1361-9209(97)00009-6]
36. Gehl, J. Cities for People; Island Press: Washington, DC, USA, 2013; ISBN 978-1-59726-984-1
37. Nicolas, C.; Kim, J.; Chi, S. Quantifying the Dynamic Effects of Smart City Development Enablers Using Structural Equation Modeling. Sustain. Cities Soc.; 2020; 53, 101916. [DOI: https://dx.doi.org/10.1016/j.scs.2019.101916]
38. Kline, R.B. Principles and Practice of Structural Equation Modeling, 4th ed; Guilford Press: New York, NY, USA, 2016; ISBN 978-1-4625-2334-4
39. Li, Z.; Han, X.; Lin, X.; Lu, X. Quantitative Analysis of Landscape Efficacy Based on Structural Equation Modelling: Empirical Evidence from New Chinese Style Commercial Streets. Alex. Eng. J.; 2021; 60, pp. 261-271. [DOI: https://dx.doi.org/10.1016/j.aej.2020.08.005]
40. Bechtel, R.B.; Marans, R.W.; Michelson, W. Methods in Environmental and Behavioral Research; Van Nostrand Reinhold Co.: New York, NY, USA, 1987; ISBN 978-0-442-21157-8
41. Moore, G.T. New Directions for Environment-Behavior Research in Architecture. Center for Architecture and Urban Planning Research, University of Wisconsin: Madison, WI, USA, 1981.
42. Purcell, A.T. Landscape Perception, Preference, and Schema Discrepancy. Environ. Plan. B Plan. Des.; 1987; 14, pp. 67-92. [DOI: https://dx.doi.org/10.1068/b140067]
43. Hashemnezhad, H.; Heidari, A.A.; Mohammad Hoseini, P. Sense of Place” and “place Attachment. Int. J. Archit. Urban Dev.; 2013; 3, pp. 5-12.
44. Fiske, S.T.; Cuddy, A.J.C.; Glick, P. Universal Dimensions of Social Cognition: Warmth and Competence. Trends Cogn. Sci.; 2007; 11, pp. 77-83. [DOI: https://dx.doi.org/10.1016/j.tics.2006.11.005]
45. Kaplan, S. The Restorative Benefits of Nature: Toward an Integrative Framework. J. Environ. Psychol.; 1995; 15, pp. 169-182. [DOI: https://dx.doi.org/10.1016/0272-4944(95)90001-2]
46. Ajzen, I. From Intentions to Actions: A Theory of Planned Behavior. Action Control: From Cognition to Behavior; Kuhl, J.; Beckmann, J. Springer: Berlin/Heidelberg, Germany, 1985; pp. 11-39. ISBN 978-3-642-69746-3
47. Li, W.; Feng, T.; Timmermans, H.J.P.; Li, Z.; Zhang, M.; Li, B. Analysis of Citizens’ Motivation and Participation Intention in Urban Planning. Cities; 2020; 106, 102921. [DOI: https://dx.doi.org/10.1016/j.cities.2020.102921]
48. Shahangian, S.A.; Rajabi, M.; Zobeidi, T.; Tabesh, M.; Yazdanpanah, M.; Hajibabaei, M.; Ghazizadeh, M.J.; Sitzenfrei, R. Perceptions of How Occupants Adopt Water Conservation Behaviors under Psychosocial Processes: A Complementary Dual-Stage SEM-ANN Perspective. Sustain. Cities Soc.; 2024; 106, 105354. [DOI: https://dx.doi.org/10.1016/j.scs.2024.105354]
49. Guo, X.; Liu, J.; Albert, C.; Hong, X.-C. Audio-Visual Interaction and Visitor Characteristics Affect Perceived Soundscape Restorativeness: Case Study in Five Parks in China. Urban For. Urban Green.; 2022; 77, 127738. [DOI: https://dx.doi.org/10.1016/j.ufug.2022.127738]
50. van Wee, B.; De Vos, J.; Maat, K. Impacts of the Built Environment and Travel Behaviour on Attitudes: Theories Underpinning the Reverse Causality Hypothesis. J. Transp. Geogr.; 2019; 80, 102540. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2019.102540]
51. Hui, H.; Liu, B.; Sun, Y.; Zhang, S. Influential Factors and Mechanism of Sense of Place in Historic Districts. Proceedings of the Proceedings of 5th International Conference on Civil Engineering and Architecture; Hanoi, Vietnam, 16–18 December 2022; Kang, T. Springer Nature: Singapore, 2024; pp. 935-945.
52. Mullenbach, L.E.; Larson, L.R.; Floyd, M.F.; Marquet, O.; Huang, J.-H.; Alberico, C.; Scott Ogletree, S.; Aaron Hipp, J. Cultivating Social Capital in Diverse, Low-Income Neighborhoods: The Value of Parks for Parents with Young Children. Landsc. Urban Plan.; 2022; 219, 104313. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2021.104313]
53. Peng, J.; Qi, J.; Cui, C.; Yan, J.; Dai, Q.; Yang, H. Research on the Impact of the Built Environment on the Characteristics of Metropolis Rail Transit School Commuting—Take Wuhan as an Example. Int. J. Environ. Res. Public. Health; 2021; 18, 9885. [DOI: https://dx.doi.org/10.3390/ijerph18189885]
54. Chang, P.-J.; Tsou, C.-W.; Li, Y.-S. Urban-Greenway Factors’ Influence on Older Adults’ Psychological Well-Being: A Case Study of Taichung, Taiwan. Urban For. Urban Green.; 2020; 49, 126606. [DOI: https://dx.doi.org/10.1016/j.ufug.2020.126606]
55. Huang, Y.; Aguilar, F.; Yang, J.; Qin, Y.; Wen, Y. Predicting Citizens’ Participatory Behavior in Urban Green Space Governance: Application of the Extended Theory of Planned Behavior. Urban For. Urban Green.; 2021; 61, 127110. [DOI: https://dx.doi.org/10.1016/j.ufug.2021.127110]
56. Xue, F.; Yao, E.; Jin, F. Exploring Residential Relocation Behavior for Families with Workers and Students: A Study from Beijing, China. J. Transp. Geogr.; 2020; 89, 102893. [DOI: https://dx.doi.org/10.1016/j.jtrangeo.2020.102893]
57. Crawford, D.W.; Jackson, E.L.; Godbey, G. A Hierarchical Model of Leisure Constraints. Leis. Sci.; 1991; 13, pp. 309-320. [DOI: https://dx.doi.org/10.1080/01490409109513147]
58. Li, Z.; Lin, X.; Han, X.; Lu, X.; Zhao, H. Landscape Efficiency Assessment of Urban Subway Station Entrance Based on Structural Equation Model: Case Study of Main Urban Area of Nanjing. Buildings; 2022; 12, 294. [DOI: https://dx.doi.org/10.3390/buildings12030294]
59. Ul Islam, J.; Rahman, Z. The Impact of Online Brand Community Characteristics on Customer Engagement: An Application of Stimulus-Organism-Response Paradigm. Telemat. Inform.; 2017; 34, pp. 96-109. [DOI: https://dx.doi.org/10.1016/j.tele.2017.01.004]
60. Koohsari, M.J.; Badland, H.; Sugiyama, T.; Mavoa, S.; Christian, H.; Giles-Corti, B. Mismatch between Perceived and Objectively Measured Land Use Mix and Street Connectivity: Associations with Neighborhood Walking. J. Urban Health; 2015; 92, pp. 242-252. [DOI: https://dx.doi.org/10.1007/s11524-014-9928-x]
61. Ma, L.; Dill, J.; Mohr, C. The Objective versus the Perceived Environment: What Matters for Bicycling?. Transportation; 2014; 41, pp. 1135-1152. [DOI: https://dx.doi.org/10.1007/s11116-014-9520-y]
62. Kaplan, D. Structural Equation Modeling: Foundations and Extensions; 2nd ed. SAGE Publications, Inc.: Thousand Oaks, CA, USA, 2009; ISBN 978-1-4522-2657-6
63. Li, W.; Liu, Y. Predicting the Impact of Integrated Audio-Visual Environments on Perceived Restorative Benefits across Different Park Types: A Field Study Based on Seven Parks in Hangzhou, China. Urban For. Urban Green.; 2024; 101, 128517. [DOI: https://dx.doi.org/10.1016/j.ufug.2024.128517]
64. Li, Z. Returning to land: A case study of landscape architecture based on site type. Landsc. Archit.; 2015; 4, pp. 117-118.
65. Davies, F.; Goode, M.; Mazanec, J.; Moutinho, L. LISREL and Neural Network Modelling: Two Comparison Studies. J. Retail. Consum. Serv.; 1999; 6, pp. 249-262. [DOI: https://dx.doi.org/10.1016/S0969-6989(98)00009-5]
66. Zhou, Y.; Zhang, L.; Chiaradia, A.J.F. An Adaptation of Reference Class Forecasting for the Assessment of Large-Scale Urban Planning Vision, a SEM-ANN Approach to the Case of Hong Kong Lantau Tomorrow. Land Use Policy; 2021; 109, 105701. [DOI: https://dx.doi.org/10.1016/j.landusepol.2021.105701]
67. Jiang, X.; Ji, L.; Chen, Y.; Zhou, C.; Ge, C.; Zhang, X. How to Improve the Well-Being of Youths: An Exploratory Study of the Relationships among Coping Style, Emotion Regulation, and Subjective Well-Being Using the Random Forest Classification and Structural Equation Modeling. Front. Psychol.; 2021; 12, 637712. [DOI: https://dx.doi.org/10.3389/fpsyg.2021.637712]
68. Hong, W.; Wang, Y. Evaluation of Highway Construction Foreman’s Competency Based on Support Vector Machine. J. Phys. Conf. Ser.; 2019; 1168, 032106. [DOI: https://dx.doi.org/10.1088/1742-6596/1168/3/032106]
69. Rudincová, K.Ž.; Ženka, J.; Bosák, V.; Slach, O.; Istrate, A.; Nováček, A. Perception and Preferred Reuse of Vegetated Vacant Lots along Main Streets in a Shrinking City. Urban For. Urban Green.; 2023; 85, 127956. [DOI: https://dx.doi.org/10.1016/j.ufug.2023.127956]
70. Hu, X.; Dong, C.; Zhang, Y. Dynamic Evolution of the Ecological Footprint of Arable Land in the Yellow and Huaihai Main Grain Producing Area Based on Structural Equation Modeling and Analysis of Driving Factors. Ecol. Inform.; 2024; 82, 102720. [DOI: https://dx.doi.org/10.1016/j.ecoinf.2024.102720]
71. Martilla, J.A.; James, J.C. Importance-Performance Analysis. J. Mark.; 1977; 41, pp. 77-79. [DOI: https://dx.doi.org/10.1177/002224297704100112]
72. Li, H.; Lv, L.; Zuo, J.; Bartsch, K.; Wang, L.; Xia, Q. Determinants of Public Satisfaction with an Urban Water Environment Treatment PPP Project in Xuchang, China. Sustain. Cities Soc.; 2020; 60, 102244. [DOI: https://dx.doi.org/10.1016/j.scs.2020.102244]
73. Cawson, J.G.; Burton, J.E.; Pickering, B.J.; Penman, T.D. Moisture Thresholds for Ignition Vary between Types of Eucalypt Forests across an Aridity Gradient. Landsc. Ecol.; 2024; 39, 70. [DOI: https://dx.doi.org/10.1007/s10980-024-01864-6]
74. Wu, Y.; Liu, Y.; Zeng, H. Ecosystem Service Supply–Demand Ratio Zoning and Thresholds of the Key Influencing Factors in the Pearl River Delta, China. Landsc. Ecol.; 2024; 39, 162. [DOI: https://dx.doi.org/10.1007/s10980-024-01964-3]
75. Li, D.; Cao, W.; Dou, Y.; Wu, S.; Liu, J.; Li, S. Non-Linear Effects of Natural and Anthropogenic Drivers on Ecosystem Services: Integrating Thresholds into Conservation Planning. J. Environ. Manag.; 2022; 321, 116047. [DOI: https://dx.doi.org/10.1016/j.jenvman.2022.116047] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36104875]
76. Palardy, N.P.; Boley, B.B.; Johnson Gaither, C. Residents and Urban Greenways: Modeling Support for the Atlanta BeltLine. Landsc. Urban Plan.; 2018; 169, pp. 250-259. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2017.09.006]
77. Wang, X.; Liu, Z. Neighborhood Environments and Inclusive Cities: An Empirical Study of Local Residents’ Attitudes toward Migrant Social Integration in Beijing, China. Landsc. Urban Plan.; 2022; 226, 104495. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2022.104495]
78. Chang, P.-J.; Ho, L.-C.; Suppakittpaisarn, P. Investigating the Interplay between Senior-Friendly Park Features, Perceived Greenness, Restorativeness, and Well-Being in Older Adults. Urban For. Urban Green.; 2024; 94, 128273. [DOI: https://dx.doi.org/10.1016/j.ufug.2024.128273]
79. Lin, H.; Hong, X.-C.; Wen, C.; Hu, F. The Historical Sensing of Urban Forest Based on the Indicators of CES and Landscape Categories: A Case of Kushan Scenic Area in CHINA. Ecol. Indic.; 2024; 166, 112440. [DOI: https://dx.doi.org/10.1016/j.ecolind.2024.112440]
80. Liu, Y.; Wang, R.; Grekousis, G.; Liu, Y.; Yuan, Y.; Li, Z. Neighbourhood Greenness and Mental Wellbeing in Guangzhou, China: What Are the Pathways?. Landsc. Urban Plan.; 2019; 190, 103602. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2019.103602]
81. Wang, Y.; Lu, S.; Xu, M.; Zhang, Y.; Xu, F. What Influences Stress Reduction in Urban Forests: Environment Types or Personality Traits?. Urban For. Urban Green.; 2024; 92, 128187. [DOI: https://dx.doi.org/10.1016/j.ufug.2023.128187]
82. Donahue, M.L.; Keeler, B.L.; Wood, S.A.; Fisher, D.M.; Hamstead, Z.A.; McPhearson, T. Using Social Media to Understand Drivers of Urban Park Visitation in the Twin Cities, MN. Landsc. Urban Plan.; 2018; 175, pp. 1-10. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2018.02.006]
83. Givoni, B. Impact of Planted Areas on Urban Environmental Quality: A Review. Atmos. Environ. Part B Urban Atmos.; 1991; 25, pp. 289-299. [DOI: https://dx.doi.org/10.1016/0957-1272(91)90001-U]
84. Ren, W.; Lu, P.; Wang, Y.; Hong, X.-C. Understanding the Association between Urban Noise and Nighttime Light in China. Sci. Rep.; 2024; 14, 31472. [DOI: https://dx.doi.org/10.1038/s41598-024-83125-9] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39732997]
85. Rossetti, T.; Lobel, H.; Rocco, V.; Hurtubia, R. Explaining Subjective Perceptions of Public Spaces as a Function of the Built Environment: A Massive Data Approach. Landsc. Urban Plan.; 2019; 181, pp. 169-178. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2018.09.020]
86. Liu, Z.; Ding, M.; He, C.; Li, J.; Wu, J. The Impairment of Environmental Sustainability Due to Rapid Urbanization in the Dryland Region of Northern China. Landsc. Urban Plan.; 2019; 187, pp. 165-180. [DOI: https://dx.doi.org/10.1016/j.landurbplan.2018.10.020]
87. Dai, D.X.; Li, M.H. Research Development of Landscape Performance Assessment in America. Landsc. Archit.; 2015; 1, pp. 25-31.
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Li, Zhe
; Chen, Haini; Yu, Mengyao; Shi, Yi