Direct and indirect impacts of urbanization on ecosystem health based on PLS-SEM in Xiangyang, China

doi:10.21203/rs.3.rs-3593308/v1

Download PDF

Research Article

Direct and indirect impacts of urbanization on ecosystem health based on PLS-SEM in Xiangyang, China

https://doi.org/10.21203/rs.3.rs-3593308/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

This study proposes a framework for investigating the direct and indirect impacts of urbanization on ecosystem health by introducing partial least squares structural equation model (PLS-SEM), the method is then applied to Xiangyang, Hubei Province, China. The validity and reliability evaluation show that PLS-SEM model is reasonable. The results showed that the level of ecosystem health in Xiangyang decreased significantly in 2010–2020. Only spatial urbanization (SU) had a direct impact on ecosystem health (-0.251/-0.262), showing a negative correlation. Population urbanization (PU) had an impact on economic urbanization (EU) (0.687/0.662), and economic urbanization (EU) had an impact on spatial urbanization (SU) (0.634/0.702). Population urbanization (PU) and economic urbanization (EU) have indirect effects on ecosystem health index (EHI). This study provides a quantitative method to determine the causes of the decline in ecosystem health, which is essential for more effective measures to maintain ecosystem health. The two objectives of this study are: (1) To establish a framework for analyzing the impacts of urbanization on ecosystem health; (2) To quantify the direct and indirect impacts and interactions of urbanization on ecosystem health.

Ecosystem health (EH)

urbanization

PLS-SEM

Xiangyang

An ecosystem is an ecological support network for the maintenance of the Earth's environment. However, these networks have been severely affected by human activities initially aimed at improving welfare and economic development. (Fang et al., 2019). Especially in the fast-growing cities of the developing world (Chi et al., 2021; Fang et al., 2021; Li et al., 2021a; Liu et al., 2007). Urbanization plays a crucial role in driving ecosystem change that affects overall ecosystem health (Li et al., 2021b; Aguilera et al., 2020; Lapointe et al., 2020; Liu et al., 2019; McDonoughet al., 2020). Rapid urbanization severely disrupts ecosystems health and threatens their sustainability (Li et al., 2021b., 2021c). Promoting sustainable development and urbanization while mitigating associated disturbances and damage to ecosystems has become an urgent issue. An effective approach is to elucidate the spatial relationship between urbanization level (UL) and ecosystem health index (EHI). However, the mechanism by which UL affects EHI is unclear. Therefore, we sought to fill these gaps by examining EHI using a variety of models and multi-source data, using an integrated spatial approach identification mechanism.

In recent decades, EHI assessment has developed rapidly, and some progress has been made in the development of theories, concepts, and assessment methods to help understand the factors that influence EHI (Li et al., 2021b, 2021c; Peng et al., 2015). Although there is no consensus on the exact definition of “ecosystem health,” the theory and understanding of EHI has been enriched and expanded to some extent (Liu et al., 2015). Currently, there is no standard method for evaluating EHI. However, using the vitality, organization, and resilience of ecosystems to establish an EHI assessment index system (Costanza et al., 1992; Peng et al., 2015), is the most representative and widely used method and has been widely validated by various studies (Chen et al., 2022; Ning et al., 2016; OU et al., 2018) .

Previous studies have examined in detail the interactions and coupling between urbanization level (UL) and ecosystems (Li et al., 2021b, 2021c; Hsiao et al., 2007). For example, Li et al. (2021B) investigated the factors that connect UL and EHI and measured their degree of coupling using a coupling coordination model. Wu et al. (2021) explored the relationship between UL and EHI using a piecewise linear regression model. Despite numerous studies, the exact impact of UL on EHI remains unclear. In addition, the laws governing cities and ecosystems vary from administrative level to administrative level. In addition, previous studies have shown significant differences in the effects of UL on ecosystems at different stages (Wang et al., 2018). China's cities (such as urban districts, county-level cities, and counties) vary greatly in terms of ecological background, resource availability and allocation, intensity of human activities, UL, and policies (Li et al., 2015; Zeng et al., 2017). Zeng et al. (2017) found that the impact of administrative levels on land use intensity has been present and increasing over the past few decades. Li et al. (2015) found that urban construction sites with higher administrative levels, such as those designated by national plans, expanded more rapidly. Ouyang (2020) found that the rate of urban land expansion varies significantly at different stages of urban development.

Grossman and Krueger (1995) suggested that the U-shaped Simon Kuznets curve could provide a theoretical basis for assessing the impact of different stages of urbanization on EHI. The Simon Kuznets curve was originally applied to economics, but it has also been environmental science to explain the relationship between economic growth and environmental impacts. Therefore, whether the relationship between UL and Ehi can be defined by the Simon Kuznets curve requires further investigation (Stern et al., 1996).

Previous studies of EHI have focused mainly on evaluation, with relatively few studies on its driving mechanisms (Pan et al., 2021; Hsiao et al., 2007). The present research mainly uses regression analysis, principal component analysis, correlation analysis and geological detector to determine the influence of natural factors (such as climate, topography) and socio-economic factors (such as construction land area ratio, land use intensity, per capita GDP on EHI (He et al., 2019a; Li., 2021b, 2021c; Wu., 2021; Xiao., 2020; Xie. 2021). Previous studies have shown that urban development indicators, such as GDP per capita, the proportion of urban population and the proportion of development land, are negatively correlated with EHI (Xiao et al., 2020; OU et al., 2018). Although the impact of UL on EHI depends on the stage of urbanization and the level of administration, few studies have explored the different impact mechanisms related to the internal size of cities. In addition, the relationship between EHI and UL exhibits significant spatial heterogeneity due to geographical complexity and process interdependence. It has been recognized that traditional statistical analysis methods can not quantify the spatial interactions between drivers and their effects on EHI, and that EHI within regions may be influenced by neighboring units and other factors (Chen et al., 2019; Xiao et al., 2020). The driving mechanism of EHI can be better studied by considering their spatial dependence and heterogeneity (Chen et al., 2019). Therefore, multi-scale spatial models are needed to analyze and explain the mechanism of the interaction between UL and EHI. Despite progress in studying the interaction between UL and EHI and the factors driving this relationship, some gaps remain. The sheer scale and heterogeneity of such relationships make it difficult to provide scientific and accurate guidelines for the sustainable development of cities and to make policy recommendations to ensure the protection and healthy functioning of urban ecosystems. There is a need for a scientific assessment of ecosystem health degradation from an urbanization perspective, rather than just based on a particular aspect of degradation.

In contrast to the above methods, structural equation modeling (SEM) is a quantitative research method based on statistical analysis techniques for dealing with multifactorial causality (Hayes et al., 2017). Based on path analysis, factor analysis, regression analysis and analysis of variance, the relationship between multiple factors and results can be determined by establishing the relationship between empirical data and theoretical analysis. SEM is used to estimate latent variables and create a complex variable prediction model (cepeda-carrion., 2019). In addition to the results of traditional multivariable statistical analyses, this approach provides a better understanding of direct and indirect interactions between factors (ren., 2021; Schweiger., 2016). Considering that the degradation of ecosystem health is the result of urbanization, the direct and indirect impacts of urbanization on ecosystem health and their interactions were studied by SEM model.

China is now entering a phase of accelerated urbanization, and clarifying the drivers and differences of the impact of urbanization (UL) on the ecosystem health index (EHI) will improve control strategies at the local and regional levels, ensure ecosystem protection and healthy operation. Xiangyang's rapid urbanisation over the past 20 years has seriously disrupted the structure, function and health of ecosystems. Studying the direct and indirect effects of urbanization level (UL) on ecosystem health index (EHI) is important for maintaining ecosystem health and promoting effective ecosystem management and control mechanisms. Therefore, we have chosen Xiangyang as our study area to analyse changes in ecosystem health and to explore the direct and indirect impacts of urbanization on ecosystem health and their interrelationships. The two objectives of this study are: (1) to establish a framework for analyzing the impacts of urbanization on ecosystem health; (2) to quantify the direct and indirect impacts and interactions of urbanization on ecosystem health. Our findings aim to propose a macro-management strategy to ensure sustainable urbanization and ecosystem protection, taking into account different control measures in different cities.

2.1. Study area and date

Xiangyang (32°04' N、112°05' E) is lies in the north-west of the Chinese province of Hubei in the middle reaches of the Han River. It has jurisdiction over three counties, three towns and three urban districts. Xiangyang has a wide variety of geological and geomorphological types, with the central hinterland being a river alluvial plain. The climate in this area has obvious seasonal characteristics. Xiangyang is located in the Han River basin, Han River is the largest tributary of the Yangtze River, the Xiangyang River basin area of about 87.7%. With the further development of urbanization and industrialization in Xiangyang, various urbanization problems in the process of urban expansion manifest themselves as the imbalance of urban ecosystem health. Xiangyang's total area is 19,727.68 square kilometers (Fig. 1), accounting for 10.6 percent of the total area of China's Hubei Province. At present, the total resident population is 5.26 million.

The main data used in this paper are Xiangyang land use (LULC) data for 2010 and 2020. In this study, land use data were extracted (at 30-meter resolution) from Landsat (7,8) ETM/OLI remote sensing images in 2010 and 2020. The images were downloaded from the U.S. Geological Survey (USGS) (http://landsat.USGS.gov/index.php). Based on the Chinese Academy of Sciences Resource Environmental Science and the data center platform, the overall accuracy reached about 85%, finally, the data of land use in Xiangyang were obtained, including six major categories and 19 minor categories (Fig. 1). The related parameters of population urbanization, economic urbanization and spatial urbanization are shown in Table 1. The same projection coordinate system (CGS2000) and spatial resolution (100m × 100m) were used for all data.

Table 1

Summary of data
Data type	Data name	Data sources
Population urbanization subsystem	Urban population density raster dataset in 2010 and 2020(+)	Resources and Environmental Sciences of the Chinese Academy of Sciences (http://www. resdc.cn)
Economic urbanization subsystem	Night-time density raster dataset in 2010 and 2020(+)	National Tibetan Plateau Scientific Data Center (TPDC) (http://data.tpdc.ac.cn).
	POI density raster dataset in 2010 and 2020(+)	Autonavi Map website (https://ditu.amap.com/).
Spatial urbanization subsystem	Distance to main traffic network raster dataset in 2010 and 2020(—)	Resources and Environmental Sciences of the Chinese Academy of Sciences (http://www. resdc.cn)
	Built-up area density raster dataset in 2010 and 2020(+)	LULC

2.2. Framework design

The study proposes a research framework to explore the direct and indirect impacts of urbanization on ecosystem health. Urbanization is a complex process that includes population, economic, spatial, and social urbanization subsystems (Chen and Chi., 2022; GE et al., 2021; Xing et al., 2021; Zhu et al., 2020). This study provides a qualitative analysis of urbanization from the perspectives of population, economy and space (Peng et al., 2017a). A partial least squares structural equation model (PLS-SEM) was introduced to investigate the complex factors and their interactions affecting ecosystem health. The evolution of ecosystem health and its influencing factors in Xiangyang over the past 10 years were discussed. The PLS-SEM model was constructed with the data of ecosystem health index (EHI) distribution, population urbanization, economic urbanization and spatial urbanization as variables, to set out the research assumptions:

1. Urbanization is divided into three dimensions: population urbanization, economic urbanization and spatial urbanization.
2. Urbanization has a direct impact on ecosystem health.
3. Population urbanization has an impact on spatial urbanization through economic urbanization.
4. Population urbanization and economic urbanization have indirect effects on ecosystem health through spatial urbanization.

Then the index of variable loading is established, and the relationship between exogenous variable (population urbanization, economic urbanization and spatial urbanization) and endogenous variable (ecosystem health index) is assumed as path coefficient. The direct and indirect impacts of urbanization on ecosystem health were evaluated by path coefficient and variable loading.

The research area grid used in this paper is calculated by using grid method in Arc GIS software (the grid size is 1 × 1km, the research area produces 21176 grid units). In this study, based on the size of the grid chosen by previous researchers for the equivalent study area, and the results of the tests, a 1 × 1km window was used to calculate and display the data of ecosystem health index (EHI), population urbanization, economic urbanization and spatial urbanization. The window size is appropriate because it represents a suitable distance to reflect the unique spatial characteristics of Xiangyang.

2.3. Ecosystem health index assessment

We used an ecosystem health index (EHI) assessment framework based on ecosystem vitality, ecosystem organization, and ecosystem resilience. This method has been widely validated and has been applied several times in similar studies (Ning et al., 2016; OU et al., 2018). Based on the vitality-organization-resilience approach proposed by Costanza et al. (1992), we established a framework for EHI assessment:

\(EH{I_{it}}=\sqrt[3]{{{V_{it}} \times {O_{it}} \times {R_{it}}}} {\mathbf{(}}1{\mathbf{)}}\)

Among these, V_it, O_it, and R_it refer to three traditional ecosystem indicators; ecosystem vitality, organization, and resilience (Li et al., 2021c; Pan et al., 2020). EHI_it represents the regional ecosystem health index in the first unit of the t-year. We graded the EHI results into five categories: good health (0.75-1), relatively good health (0.65–0.75), general health (0.55–0.65), relatively weak health (0.45–0.55), and weak health (0-0.45)(Ou et al., 2018) .

(1) Ecosystem vitality (EV) is usually defined as ecosystem metabolism or net primary productivity. The vegetation coverage (NDVI) was selected to reflect the vitality of the ecosystem according to the technical standard for ecological assessment (HJ192-2015), by calculating the proportion of different land use types (Meng et al., 2018; OU et al., 2018).

(2) Ecosystem organization (EO) refers to the stability and complexity of regional ecosystem structure. Ecosystem organization (EO) indicators can be assessed by different computational methods and can be analyzed from the perspective of landscape connectivity and spatial heterogeneity (Turner et al., 1989). To assess EO, landscape heterogeneity and landscape connectivity indicators were used by reference scholars (Xie et al., 2021; Das et al., 2021). Using three landscape measurement landscape heterogeneity metrics, modified simpson's diversity index (MSIDI), shannon's evenness index (SHEI) and shannon's diversity index (SHDI). Interspersion and juxtaposition index (IJI), contagion index (CONTAG), and landscape division index (DIVISION) are used to calculate Landscape connectivity. For the weights of LH and LC, the results of relevant scholars are synthesized and obtained by combining the characteristics of the study area (Xie et al., 2021; Das et al., 2021; Peng et al., 2017a, 2017B, 2015). Landscape heterogeneity (LH) was 0.5 and Landscape connectivity (LC) was 0.5.

Table 2

Weight of each parameters.
EO	weight	Sub-index layer	weight
LH	0.5	MSIDI	0.25
		SHDI	0.15
		SHEI	0.10
LC	0.5	IJI	0.25
		CONTAG	0.15
		DIVISION	0.10

EO formula:

\(EO=(0.5 \times LH)+(0.5 \times LC) {\mathbf{(}}2{\mathbf{)}}\)

\(\begin{gathered} EO=(0.25 \times L{H_1}{\text{+}}0.15 \times L{H_2}{\text{+}}0.1 \times L{H_3})+ \hfill \\ (0.25 \times L{C_1}{\text{+}}0.15 \times L{C_2}{\text{+}}0.1 \times L{C_3}) {\mathbf{(}}3{\mathbf{)}} \hfill \\ \end{gathered}\)

EO stands for ecosystem organization, while LH and LC stand for landscape heterogeneity and landscape connectivity. LH is divided into three categories: LH₁(MSIDI), LH₂(SHDI) and LH₃(SHEI). In addition, LC is divided into three parts: LC₁(IJI), LC₂(CONTAG), and LC₃(DIVISION) .

(3) Ecosystem resilience (ER) refers to the ability of an ecosystem to maintain its structure and patterns. Ecosystems are healthy only after long-term resistance and recovery from outside disturbance (Lopez et al., 2013; Colding et al., 2007). Ecosystem resilience assesses the resilience and resistance of ecosystem to external forces. Based on previous studies (Peng et al., 2017a; Peng et al., 2017b; Xiao et al., 2019; Xie et al., 2021; Das et al., 2021), the ecosystem resilience formula and its coefficients are as follows:

\(ER=0.6 \times re{\text{s}}il+0.4 \times resist {\mathbf{(}}4{\mathbf{)}}\)

ER includes “Resil” and “Resist”, “Resil” refers to the ecosystem resilience, and “Resist” refers to ecosystem resistance.

Because of Xiangyang's urban sprawl and human interference, the study focused on resilience. Therefore, give a weight to resilience (0.6) and a weight to resistance (0.4). The coefficients for each LULC type are shown in Table 3.

Table 3

ER coefficient used for LULC types.
Land use type	Cultivated	Forest	Grass	Water	Urban	Unused
Resilience Coefficient	0.4	0.8	0.8	0.7	0.2	0.3
Resistance Coefficient	0.6	1.0	0.6	0.8	0.3	0.5

2.4. The PLS-SEM model specification

PLS-SEM was used to estimate the causal network between latent variables and manifest variables. A manifest variable is a quantity that can be directly observed or measured, also known as an observed variable. Latent variables are those that can not be directly observed, but can be constructed from one or more manifest variables by some theory or hypothesis. PLS-SEM usually includes measurement model and structure model. Measurement models (or external models) are explanatory models consisting of latent and manifest variables (Ullman and Bentler, 2012). This can be expressed in the following formula:

\(X={\Lambda _x}\xi +\delta \quad \quad \quad {\mathbf{(}}5{\mathbf{)}}\)

\(Y={\Lambda _y}\eta +\varepsilon \quad \quad \quad {\mathbf{(}}6{\mathbf{)}}\)

Equations (5) and (6) are exogenous and endogenous indicators respectively. Λ refers to the relationship between the latent variable and manifest variable; δ and ε refer to the measurement error. Structural (or internal) models are path maps that reflect the relationship between latent variable effects (Hoyle, 1995; Kline, 2015) :

\(\eta =\beta \eta +\Gamma \xi +\xi {\mathbf{(}}7{\mathbf{)}}\)

ηis an endogenetic latent variable, ξis an exogenous latent variable, and ᵦ represents the influence of exogenous latent variable on endogenetic latent variable. Γ denotes the effect of some endogenous latent variables on other endogenous latent variables. PLS-SEM was used to assess the main factors affecting ecosystem health index (EHI) distribution. PLS-SEM relaxed the multivariate normal distribution assumptions during parameter estimation relative to covariance-based structural equation models (CB-SEM)(Hair et al., 2011; Shen et al., 2016) .

If CB-SEM model is used in the study of ecosystem health and urbanization data, the hypothesis can not be guaranteed because of the normal distribution of the data. However, PLS-SEM relaxes the relevant limitations, is suitable for exploratory research, and can effectively assess interactions between variables (Fan et al., 2016; Tenhaus et al., 2005).

The degradation of ecosystem health is the result of urbanization, which has direct and indirect effects on ecosystem health. Urbanization is a complex process that includes demographic, economic, spatial, and social urbanization subsystems (Chen and Chi, 2022; GE et al., 2021; Xing et al., 2021; Zhu et al., 2020). This study takes Xiangyang as an example to conduct a qualitative analysis of urbanization from a Population, economic and spatial perspective (Peng et al., 2017a). Based on PLS-SEM model, research hypotheses are proposed: 1. Urbanization is divided into three dimensions: population urbanization, economic urbanization and spatial urbanization. 2. Urbanization has a direct impact on ecosystem health. 3. Population urbanization and economic urbanization affect ecosystem health through spatial urbanization. An integrated raster data set with a size of 1 km × 1 km cells was used in the PLS-SEM Model (Table 1). The 2010 and 2020 ecosystem health and urbanization data were divided into grid layers, with the average pixel value of each grid serving as a sample set.

3.1. EHI distribution in 2010 and 2020

From 2010 to 2020, Xiangyang's ecosystem health index (EHI) changed significantly (Fig. 3). The ecosystem health index (EHI) showed an overall trend of degradation, mainly occurring around cities and along main traffic routes, especially to the east of Xiangcheng and Fancheng (Fig. 3: a). It indicates that urbanization is the main cause of degradation of ecosystem health index (EHI) .

From 2010 to 2020, ecosystem vitality (EV) in Xiangyang changed somewhat (Fig. 3), mainly around cities, especially to the east of Xiangcheng and Fancheng (Fig. 3: b), the eastward expansion of the city makes the green patches in the east of the city disappear gradually.

From 2010 to 2020, ecosystem organization (EO) in Xiangyang changed considerably (Fig. 3), mainly around cities and along main traffic routes (Fig. 3: c, d, e, f), urbanization and the construction of traffic networks have significantly degraded the organization of ecosystem health.

From 2010 to 2020, ecosystem resilience (ER) in Xiangyang changed somewhat (Fig. 3), mainly around cities, especially to the east of Xiangcheng and Fancheng (Fig. 3:g), the eastward expansion of cities has changed the resilience of ecosystem health.

Urban expansion and traffic networks have significant impacts on the ecosystem health index (EHI), and the interaction between urbanization and EHI needs to be explored.

3.2. Multicollinearity test

Correlations between independent variables are common, and multicollinearity tests are needed to prevent “serious multicollinearity” in independent variables (> 10). After SmartPLS 4 software examined 12 datasets in 2010 and 2020, the variance expansion factor (VIF) was used to detect multicollinearity of 6 variables, the results showed VIF values for variables ranging from 1 to 1.43(Table 1), indicating no multicollinearity problems between variables (Thompson et al., 2017) .

3.3. Reliability, Convergence Validity and Differential Validity Analysis result

An assessment of the reliability and validity of the data allows an effective assessment of the reliability of PLS-SEM models (Chen et al., 2015; Hair et al., 2019). The reliability, convergence validity and differential validity of the data were analyzed according to the hypothesis model, and the rationality of the data and the model was tested.

Table 4

Reliability and convergence validity table
Dim	Item						Item Reliability	Composite Reliability	Convergence validity
Dim	Item	Esti	VIF	STDEV.	Est./STDE.	P-Value	R-square	CR	AVE
EH(2010)	EH	1.000	1.000	0.000	/	/	/	/	/
PU(2010)	pop	1.000	1.000	0.000	/	/	/	/	/
EU(2010)	nig	0.890	1.331	0.007	133.714	***	0.472	0.680	0.748
	poi	0.839	1.331	0.011	78.116	***	/	/	/
SU(2010)	bui	0.955	1.023	0.003	336.277	***	0.362	0.476	0.551
	tra	-0.436	1.023	0.009	48.279	***	/	/	/
EH(2020)	EH	1.000	1.000	0.000	/	/	/	/	/
PU(2020)	pop	1.000	1.000	0.000	/	/	/	/	/
EU(2020)	nig	0.885	1.430	0.006	153.367	***	0.439	0.709	0.774
	poi	0.875	1.430	0.005	170.642	***	/	/	/
SU(2020)	bui	0.952	1.046	0.002	521.919	***	0.443	0.591	0.578
	tra	-0.500	1.046	0.006	81.472	***	/	/	/
Note:***=P < 0.001

As shown in Table 4, all items had standardized estimates greater than 0.43, which indicates that all factor loads meet structural validity requirements (Moeinaddini et al., 2020). Est/STDE, all above 1.96, indicating significant. All p-values are less than 0.001, which means that all items are significant. Item Reliability is the R-square of the standardized estimate, which is more than 0.36 and acceptable, and 0.5 or more represents good item reliability. Item reliability is the ability of dimensionality to explain item. Composite reliability (CR) refers to the internal consistency of our dimensionality and item (J öreskog et al., 1969), which is greater than 0.47. We say that the composite reliability (CR) is acceptable, and all the composite reliability (CR) is greater than 0.47. AVE refers to the average ability of a dimension to explain an item. It is recommended that the dimension be greater than 0.5(Nasution et al., 2020), this means that my dimensions have good explanatory power, which is generally acceptable at 0.36, and that the convergence validity of all my dimensions is at the recommended level.

Table 5

Reliability, convergence validity and discriminant validity analysis table
DIM	ITEM	Item Reliability	Composite Reliability	Convergence validity	Discriminate Validity
DIM	ITEM	STD.LOADING	CR	AVE	EU	PU	SU
EU(2010)	2	0.839–0.890	0.680	0.748	0.865
PU(2010)	1	/	/	/	0.687	1.000
SU(2010)	2	0.436–0.955	0.476	0.551	0.601	0.387	0.743
EU(2020)	2	0.875–0.885	0.709	0.774	0.880
PU(2020)	1	/	/	/	0.662	1.000
SU(2020)	2	0.500-0.952	0.591	0.578	0.663	0.407	0.760
Note: The diagonal bold font is the AVE square root value, and the lower triangle is the Pearson correlation of the dimension.

As shown in Table 5, each of our dimensions has its own item, and all of its standard factor loads are greater than 0.43, our composite reliability (CR) is greater than 0.47, and our convergence validity (AVE) is greater than 0.5, (Clacs Fornell, 1981), we put the square root of AVE diagonally, and the lower triangle is the Pearson correlation of dimensions, indicates that our dimensions have discriminant validity.

Furthermore, the model was tested by the fit degree index (SRMR), which was consistent with the recommended value (< 0.08) in 2010 and 2020. The model was established and had statistical significance.

3.4. Research model hypothesis analysis result

The results in Table 4,5 show that the model is valid. Table 6 shows the hypothetical results of the model, and Fig. 4 shows the model diagram drawn from the results of Table 6. Figure 4 shows that only spatial urbanization (SU) has a direct impact on ecosystem health index (EHI) (2010(-0.251)/2020(-0.262)), a negative correlation. Population urbanization (PU) has an impact on economic urbanization (EU) (2010(0.687)/2020(0.662)), and economic urbanization (EU) has an impact on spatial urbanization (SU) (2010(0.634)/2020(0.702)), population urbanization and economic urbanization have indirect effects on ecosystem health.

Table 6

Research model hypothesis analysis
	DV	IV	Estimate	STDEV.	Est./STDE.	P-Value	Hypothesis
2010	EH	PU	-0.099	0.009	10.578	0.000	Not Support
		EU	0.094	0.012	8.086	0.000	Not Support
		SU	-0.251	0.007	37.293	0.000	Support
	SU	PU	-0.048	0.023	2.118	0.000	Not Support
		EU	0.634	0.023	27.655	0.034	Support
	EU	PU	0.687	0.027	25.261	0.000	Support
2020	EH	PU	-0.097	0.007	13.620	0.000	Not Support
		EU	0.093	0.009	9.862	0.000	Not Support
		SU	-0.262	0.007	36.159	0.000	Support
	SU	PU	-0.058	0.026	2.237	0.025	Not Support
		EU	0.702	0.020	35.196	0.000	Support
	EU	PU	0.662	0.019	34.180	0.000	Support

3.5. Effects of latent variables and manifest variables on EHI distribution

Figure 4 shows the PLS-SEM model of the relationship between the 2010–2020 ecosystem health index (EHI) distribution and the latent and manifest variables of urbanization influencing factors. The results showed that only spatial urbanization (SU) had a direct impact on ecosystem health index (EHI) (2010(-0.251)/2020(-0.262)), which is a negative correlation. Population urbanization (PU) has an impact on economic urbanization (EU)(2010(0.687)/2020(0.662)), and economic urbanization (EU) has an impact on spatial urbanization (SU)(2010(0.634)/2020(0.702)), population urbanization and economic urbanization have indirect effects on ecosystem health index (EHI). Spatial urbanization (SU) is a manifest variable of urbanization on ecosystem health index, which has a direct impact on ecosystem health index. Population urbanization (PU) and economic urbanization (EU) are the latent variables of urbanization on ecosystem health index, which have indirect effects on ecosystem health index. The latent variable has an effect on ecosystem health index through manifest variable, that is, population urbanization and economic urbanization have an indirect effect on ecosystem health index through spatial urbanization.

As shown in Fig. 4, this relationship changes over time, and overall, spatial urbanization has a direct impact on the ecosystem health index, population urbanization and economic urbanization have indirect effects on ecosystem health index through spatial urbanization. This general trend has not changed, but the extent to which spatial urbanization has a direct impact on the ecosystem health index is increasing (2010(-0.251)/2020(-0.262)). The impact of population urbanization (PU) on economic urbanization (EU) was weakening (2010(0.687)/2020(0.662)). The impact of economic urbanization (EU) on spatial urbanization (SU) was enhanced (2010(0.634)/2020(0.702)).

4.1. Direct impact of factors on EH distribution

Our results suggest that spatial urbanization has a direct impact on ecosystem health index, with a negative correlation, that is, with the expansion of spatial urbanization (urban land and main traffic network), this has led to a decline in the ecosystem health index, and the trend has become more pronounced over time. Human activities are therefore a major factor in the decline of the ecosystem health index. This finding is consistent with the findings of some scholars (Xiao et al.,2020).

4.2. Indirect impact of latent variables on EHI distribution

The complex interaction between urbanization level (UL) and ecosystem health index (EHI) is difficult to understand. In this study, PLS-SEM model was used to explore the interaction among population urbanization, economic urbanization and spatial urbanization, and the impact on ecosystem health index (EHI). It is generally considered that population urbanization, economic urbanization and spatial urbanization have a direct impact on ecosystem health index. However, this study demonstrates that only spatial urbanization (SU) has a direct impact on ecosystem health, population urbanization (PU) and economic urbanization (EU) have indirect effects on ecosystem health index through spatial urbanization (SU). And the effect is linear (Fig. 4), that is, PU - EU - SU - EH. It is the first time that this viewpoint is put forward in the research field of the relationship between urbanization and ecosystem health.

The 2010 and 2020 models show (Fig. 4) that population urbanization has a direct impact on economic urbanization, but the magnitude of the impact is weakening year by year, indicating a gradual weakening of the relationship between population concentration and economic activity; It may be related to demographics such as population ageing, employment and unemployment rates, but this assumption needs to be tested in detail. Economic urbanization has a direct impact on spatial urbanization, and the impact is increasing year by year, indicating that the relationship between economic activities and urban construction and main traffic network construction is gradually strengthened. It may have something to do with urbanization, main traffic network and other infrastructure construction, but this hypothesis needs to be tested.

The results show that population urbanization (PU) and economic urbanization (EU) have indirect effects on the distribution of ecosystem health index (EHI). This is because changes in population urbanization (PU) and economic urbanization (EU) have altered spatial urbanization (SU). Thus, we can maintain the ecosystem health index (EHI) by improving the structure of population urbanization (PU) and economic urbanization (EU). Therefore, this phenomenon must be considered in future research. We need to look for breakthroughs in how urbanization affects the ecosystem health index (EHI), and further analyze the principles and mechanisms of the process.

4.3. Limitations and implications

In contrast to existing research on the relationship between urbanization and ecosystem health, we have innovatively proposed a research framework, the relationship and interaction between urbanization and ecosystem health were discussed from the perspective of urbanization. Makes up for the lack of quantitative analysis in existing research. This paper focuses on the direct and indirect impacts of urbanization on ecosystem health, which provides a new perspective for the study of ecosystem health index (EHI). By constructing a balance system between urbanization and ecosystem health, the impact of human activities on ecosystem health can be controlled within a reasonable range, thus achieving sustainable development of environment and society.

The reasons for changes in the ecosystem health index (EHI) tend to vary from region to region. Therefore, the framework can be extended to different research areas to provide targeted guidance for the protection of ecosystem health (EH). However, factors affecting the distribution of the ecosystem health index (EHI), such as social urbanization, should be taken into account more fully.

1. A framework for analyzing the impact of urbanization on ecosystem health index was constructed.

2. Between 2010 and 2020, Xiangyang's ecosystem health declined significantly, with only spatial urbanization (SU) having a direct impact on ecosystem health index (EHI), a negative correlation.

3. Population urbanization (PU) has an impact on economic urbanization (EU), economic urbanization (EU) has an impact on spatial urbanization (SU), and population urbanization and economic urbanization have an indirect impact on ecosystem health index (EHI).

Overall, the study presents an innovative approach to the direct and indirect impact of urbanization on changes in ecosystem health index (EHI). These findings have important implications for more effective measures to maintain the relationship between urbanization and ecosystem health under different conditions.

Aguilera, M.A., Tapia, J., Gallardo, C., Nú˜nez, P., Varas-Belemmi, K., 2020. Loss of coastal ecosystem spatial connectivity and services by urbanization: natural-to-urban integration for bay management. J. Environ. Manag. 276, 111297.
Chi, Y., Liu, D., Xing, W., Wang, J., 2021. Island ecosystem health in the context of human activities with different types and intensities. J. Clean. Prod. 281, 125334.
Costanza, R., Norton, B., Haskell, B., 1992. Ecosystem Health: New Goals for Environmental Management. Island Press, Washington.
Chen, W., Chi, G., 2022. Urbanization and ecosystem services: the multi-scale spatial spillover effects and spatial variations. Land Use Pol. 114, 105964.
Chen, W., Chi, G., Li, J., 2019. The spatial association of ecosystem services with land use and land cover change at the county level in China, 1995–2015. Sci. Total Environ.669, 459–470.
Cepeda-Carrion, G., Cegarra-Navarro, J.-G., Cillo, V., 2019. Tips to use partial least squares structural equation modelling (PLS-SEM) in knowledge management. J. Knowl. Manag. 23, 67–89.
Chen, W., Chi, G., 2022. Urbanization and ecosystem services: the multi-scale spatial spillover effects and spatial variations. Land Use Pol. 114, 105964.
Chen, X., Liu, X., Hu, D., 2015. Assessment of sustainable development: a case study of Wuhan as a pilot city in China. Ecol. Indic. 50, 206–214.
Colding, J. 2007. ‘Ecological land-use complementation’for building resilience in urban ecosystems. Landscape Urban Plann. 81 (1-2), 46-55.
Das, M., Das, A., Pereira, P., & Mandal, A. 2021. Exploring the spatio-temporal dynamics of ecosystem health: A study on a rapidly urbanizing metropolitan area of Lower Gangetic Plain, India. Ecological Indicators, 125, 107584.
Fan, Y., Chen, J., Shirkey, G., John, R.,Wu, S.R., Park, H., Shao, C., 2016. Applications of structural equation modeling (SEM) in ecological studies: an updated review. Ecol. Process.5, 19.
Fang, C., Cui, X.G., Liang, L.W., 2019. Theoretical analysis of urbanization and ecoenvironment coupling coil and coupler control. Acta Geograph. Sin. 74 (12),2529–2546.
Fang, L., Wang, L., Chen, W., Sun, J., Cao, Q., Wang, S., Wang, L., 2021. Identifying the impacts of natural and human factors on ecosystem service in the Yangtze and Yellow River Basins. J. Clean. Prod. 314, 127995.
Grossman, G.M., Krueger, A.B., 1995. Economic growth and the environment. Q. J. Econ.110 (2), 353–377.
Ge, F., Chen, W., Zeng, Y., Li, J., 2021. The nexus between urbanization and traffic accessibility in the middle reaches of the Yangtze River Urban Agglomerations, China. Int. J. Environ. Res. Publ. Health 18 (7), 3828.
Hsiao, C., 2007. Panel data analysis-advantages and challenges. Test 16 (1), 1–22.
He, J., Pan, Z., Liu, D., Guo, X., 2019a. Exploring the regional differences of ecosystem health and its driving factors in China. Sci. Total Environ. 673, 553–564.
Hair, J.F., Ringle, C.M., Sarstedt, M., 2011. PLS-SEM: indeed a silver bullet. J. Mark. Theory Pract. 19, 139–152.
Hair, J.F., Risher, J.J., Sarstedt, M., Ringle, C.M., 2019. When to use and how to report the results of PLS-SEM. Eur. Bus. Rev. 31, 2–24.
Hayes, A.F., Montoya, A.K., Rockwood, N.J., 2017. The analysis of mechanisms and their contingencies: PROCESS versus structural equation modeling. Australas. Mark.J. AMJ 25, 76–81.
Hoyle, R.H., 1995. Structural Equation Modeling: Concepts, Issues, and Applications. SAGE Publications. https://books.google.com.sg/books?id=CoM5DQAAQBAJ.
Jöreskog, K.G., 1969. Statistical Analysis of Sets of Congeneric Tests. Educational Testing Service. https://books.google.ch/books?id=xK58oAEACAAJ.
Kline, R.B., 2015. Principles and Practice of Structural Equation Modeling. Fourth edition. Guilford Publications. https://books.google.com.sg/books?id=Q61ECgAAQBAJ.
Li, H., Wei, Y.D., Liao, F.H., Huang, Z., 2015. Administrative hierarchy and urban land expansion in transitional China. Appl. Geogr. 56, 177–186.
Li, W., Liu, C., Su, W., Ma, X., Zhou, H., Wang, W., Zhu, G., 2021a. Spatiotemporal evaluation of alpine pastoral ecosystem health by using the Basic-Pressure-State-Response Framework: a case study of the Gannan region, northwest China. Ecol. Indicat. 129, 108000.
Li, W., Wang, Y., Xie, S., Cheng, X., 2021b. Coupling coordination analysis and spatiotemporal heterogeneity between urbanization and ecosystem health in Chongqing municipality, China. Sci. Total Environ. 791, 148311.
Li, W., Xie, S., Wang, Y., Huang, J., Cheng, X., 2021c. Effects of urban expansion on ecosystem health in Southwest China from a multi-perspective analysis. J. Clean. Prod. 294, 126341.
Liu, J., Dietz, T., Carpenter, S.R., Alberti, M., Folke, C., Moran, E., Pell, A.N., Deadman, P., Kratz, T., Lubchenco, J., Ostrom, E., Ouyang, Z., Provencher, W.,Redman, C.L., Schneider, S.H., Taylor, W.W., 2007. Complexity of coupled human and natural systems. Science 317 (5844), 1513–1516.
Lapointe, M., Gurney, G.G., Cumming, G.S., 2020. Urbanization alters ecosystem service preferences in a small island developing state. Ecosyst. Serv. 43, 101109.
Liu, X., Pei, S., Wang, S., Wen, Y., Li, X., Wu, J., Chen, J., Feng, K., Liu, J., Hubacek, K.,Davis, S., Yu, L., Liu, Z., Wu, C., Cai, Y., Yuan, W., 2019. Global urban expansion offsets climate-driven increases in terrestrial net primary productivity. Nat. Commun. 10 (1), 1–8.
Liu, Y.X., Peng, J., Wang, A., Xie, P., Han, Y.N., 2015. New research progress and trends in ecosystem health. Acta Geograph. Sin. 35 (18), 5920–5930.
L´opez, D.R., Brizuela, M.A., Willems, P., Aguiar, M.R., Siffredi, G., Bran, D. 2013.Linking ecosystem resistance, resilience, and stability in steppes of North Patagonia.Ecol. Ind. 24, 1–11.
Meng, L., Huang, J., Dong, J., 2018. Assessment of rural ecosystem health and type classification in Jiangsu Province, China. Sci. Total Environ. 615, 1218–1228.
McDonough, L.K., Santos, I.R., Andersen, M.S., O’Carroll, D.M., Rutlidge, H.,Meredith, K., Oudone, P., Bridgeman, J., Gooddy, D.C., Sorensen, J.P.R., Lapworth, D.J., MacDonald, A.M., Ward, J., Baker, A., 2020. Changes in global groundwater organic carbon driven by climate change and urbanization. Nat. Commun. 11 (1), 1–10.
Moeinaddini, M., Asadi-Shekari, Z., Aghaabbasi,M., Saadi, I., Zaly Shah,M., Cools,M., 2020.Proposing a newscore tomeasure personal happiness by identifying the contributing factors. Measurement 151, 107115..
Ning, L.X., Ma, L., Zhou, Y.K., Bai, X.L., 2016. Spatiotemporal variations of ecosystem health of the coastal zone in Jiangsu Province based on the PSR model. China Environ. Sci. 36 (2), 534–543.
Nasution, M.I., Fahmi, M., Jufrizen, Muslih, Prayogi, M.A., 2020. The quality of small and medium enterprises performance using the structural equation model-part Least Square (SEM-PLS). J. Phys. Conf. Ser. 1477, 052052..
Ou, W., Zhang, L., Tao, Y., Guo, J., 2018. A land-cover-based approach to assessing the spatio-temporal dynamics of ecosystem health in the Yangtze River Delta region.China Popul. Resour. Environ. 28, 84–92.
Ouyang, X., Zhu, X., 2020. Spatio-temporal characteristics of urban land expansion in Chinese urban agglomerations. Acta Geograph. Sin. 75 (3), 571–588.
Peng, J., Liu, Y., Wu, J., Lv, H., Hu, X., 2015. Linking ecosystem services and landscape patterns to assess urban ecosystem health: a case study in Shenzhen City, China. Landsc. Urban Plann. 143, 56–68.
Peng, J., Tian, L., Liu, Y., Zhao, M., Hu, Y., Wu, J., 2017a. Ecosystem services response to urbanization in metropolitan areas: thresholds identification. Sci. Total Environ.607–608, 706–714.
Peng, J., Liu, Y., Li, T., Wu, J. 2017b. Regional ecosystem health response to rural land use change: a case study in Lijiang City, China. Ecol. Ind. 72, 399–410.
Pan, Z., He, J., Liu, D., Wang, J., 2020. Predicting the joint effects of future climate and land use change on ecosystem health in the Middle Reaches of the Yangtze River Economic Belt, China. Appl. Geogr. 124, 102293.
Pan, Z., He, J., Liu, D., Wang, J., Guo, X., 2021. Ecosystem health assessment based on ecological integrity and ecosystem services demand in the Middle Reaches of the Yangtze River Economic Belt, China. Sci. Total Environ. 774, 144837.
Ren, L., Li, J., Li, C., Dang, P., 2021. Can ecotourismcontribute to ecosystem? Evidence from local residents' ecological behaviors. Sci. Total Environ. 757, 143814.
Stern, D.I., Common, M.S., Barbier, E.B., 1996. Economic growth and environmental degradation: the Environmental Kuznets Curve and sustainable development. World Dev. 24 (7), 1151–1160.
Schweiger, E.W., Grace, J.B., Cooper, D., Bobowski, B., Britten, M., 2016. Using structural equationmodeling to link human activities towetland ecological integrity. Ecosphere 7, e01548.
Shen,W., Xiao,W., Wang, X., 2016. Passenger satisfaction evaluation model for urban rail transit: a structural equation modeling based on partial least squares. Transp. Policy 46, 20–31.
Turner, M.G. 1989. Landscape ecology: the effect of pattern on process. Annu. Rev. Ecol. Syst. 20 (1), 171–197.
Tenenhaus, M., Vinzi, V.E., Chatelin, Y.-M., Lauro, C., 2005. PLS path modeling. Comput. Stat. Data Anal. 48, 159–205.
Thompson, C.G., Kim, R.S., Aloe, A.M., Becker, B.J., 2017. Extracting the variance inflation factor and other multicollinearity diagnostics from typical regression results. Basic Appl. Soc. Psychol. 39, 81–90.
Ullman, J.B., Bentler, P.M., 2012. Structural Equation Modeling. Handbook of Psychology, Second edition https://doi.org/10.1002/9781118133880.hop202023.
Wu, J., Cheng, D., Xu, Y., Huang, Q., Feng, Z., 2021. Spatial-temporal change of ecosystem health across China: urbanization impact perspective. J. Clean. Prod. 326,129393.
Wang, J. Y., Xi, F. M., Liu, Z., Bing, L. F., Alsaedi, A, Hayat, T., Guan, D. B. (2018). The spatiotemporal features of greenhouse gases emissions from biomass burning in China from 2000 to 2012. Journal of Cleaner Production, 181, 801–808.
Xiao, Y., Xiong, Q., Pan, K. 2019. What is left for our next generation? Integrating ecosystem services into regional policy planning in the Three Gorges Reservoir Area of China. Sustainability 11 (1), 3.
Xiao, Y., Guo, L., Sang, W., 2020. Impact of fast urbanization on ecosystem health in mountainous regions of Southwest China. Int. J. Environ. Res. Publ. Health 17 (3),826.
Xie, X., Fang, B., Xu, H., He, S., Li, X., 2021. Study on the coordinated relationship between urban land use efficiency and ecosystem health in China. Land Use Pol. 102,105235.
Xing, L., Zhu, Y., Wang, J., 2021. Spatial spillover effects of urbanization on ecosystem services value in Chinese cities. Ecol. Indicat. 121, 107028.
Zhu, C., Zhang, X., Zhou, M., He, S., Gan, M., Yang, L., Wang, K., 2020. Impacts of urbanization and landscape pattern on habitat quality using OLS and GWR models in Hangzhou, China. Ecol. Indicat. 117, 106654.
Zeng, C., Zhang, A., Liu, L., Liu, Y., 2017. Administrative restructuring and land-use intensity-A spatial explicit perspective. Land Use Pol. 67, 190–199.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Direct and indirect impacts of urbanization on ecosystem health based on PLS-SEM in Xiangyang, China

Status:

Version 1

Abstract

Figures

1. Introduction

2. Methods

2.1. Study area and date

2.2. Framework design

2.3. Ecosystem health index assessment

2.4. The PLS-SEM model specification

3. Results

3.1. EHI distribution in 2010 and 2020

3.2. Multicollinearity test

3.3. Reliability, Convergence Validity and Differential Validity Analysis result

3.4. Research model hypothesis analysis result

3.5. Effects of latent variables and manifest variables on EHI distribution

4. Discussion and Conclusion

4.1. Direct impact of factors on EH distribution

4.2. Indirect impact of latent variables on EHI distribution

4.3. Limitations and implications

5. Conclusion

References

Additional Declarations

Status:

Version 1