1. Introduction
Since China launched the “Reform and Opening-up” policy in 1978, the country has experienced—and continues to experience—rapid urban development, with the proportion of urban population (also often referred to as urbanization rate) increasing from 17.92% to 54.8% in 2015. This growth, however, has brought various environmental issues in China [
1,
2,
3,
4]. Therefore, a new-type of urbanization was proposed in 2014 to improve the quality and sustainability of urban development [
5,
6]. Given limited agglomeration effects for single cities as well as imbalanced spatial distribution of resources and environmental conditions, Chinese governments emphasize the significance and urgency of taking urban agglomeration (UA) as a main form of new-type urbanization [
1,
7]. Urbanization in UA areas will have a profound impact on economic growth in China. However, it is also posing practical or potential threats to the eco-environment [
8,
9,
10,
11]. On the one hand, UA areas in China with 20% of the land area concentrate 60% of the total population and 80% of the total economic output. On the other hand, these UA areas contribute over 67% of industrial emissions [
7]. Urbanization planning compiled and implemented by Chinese governments is the mandatory policy to guide healthy regional development. Consequently, a better understanding of the relationship between urbanization (U) and eco-environment (E) and integrating it into plans compiled for UA are required to achieve the sustainable development of the whole country.
Previous literature engaged in this work focused mostly on coordination between U and E. After the theory of complex social-economic-natural ecosystem [
12] was proposed, many scholars regarded U and E as complex subsystems and measured the level of coordinated development of the coupled U-E system using various methods. For example, the Coupling Degree Model [
13], the Coupling Coordination Degree Model [
14], and the Dynamic Coupling Coordination Degree Model [
10] were used to explore the coordination level in the coupled system. Grey Correlation Analysis [
15] was also employed to quantify the covariation between the two variables. These quantitative analyses between the U and E subsystems can reveal the coordination degree and provide urban planners with guidance to develop reasonable policies for the balanced development. However, the level of coordinated development calculated from the above researches is a single value, which cannot disclose the interacting mechanism between U and E.
To explore the interaction in the coupled U-E system, many studies have been conducted based on a series of methods from multiple perspectives. The relevant literature on the interaction mainly falls into three categories. The first category of studies employs empirical models to reveal the curvilinear relationship between the comprehensive levels of the two subsystems. After the Environmental Kuznets Curve (EKC) [
16,
17] was proposed to represent the relationship between economy and environment, the logarithmic curve [
18] was used to describe the relationship between urbanization and economic development. Then, Huang et al. [
19] deduced a double-exponential function between U and E based on the two curve-based models above. A growing body of empirical studies subsequently verified this relationship in many countries and cities [
20,
21]. However, these studies could not detail the internal driving mechanism, thereby being incapable of offering effective guidance for urban planning. The second category is to measure non-linear relationships between the two subsystems by identification of driving mechanisms. Taking Jiangsu province in China as an example, Song et al. [
22] used a System Dynamics Model to demonstrate diverse causal feedbacks from human activities to environmental changes. Similar studies were conducted in Shenzhen city by Burak et al. [
23], and in Jiangxi province by Liu et al. [
24]. The Environmental Computable General Equilibrium (CGE) model [
25] was also employed to investigate the impact of urban development policies on the economy and environment. Although these studies concentrated on the functional path within the coupling system and provided a better understanding of interaction mechanism between factors of two subsystems, they could not uncover the spatial heterogeneity of the effect caused by the same factors in different regions. The third category explores mathematical relationships between the two subsystems utilizing econometric models, including correlation analysis, the Ordinary Least Squares (OLS) model, stepwise regression and panel data model [
26,
27]. However, these studies just mainly focused on the relationship between U and a single ecological factor, such as resource production [
26], carbon emissions [
27], land use [
28], water resource [
29], and air quality [
30]. Moreover, there were few studies that investigate spatial heterogeneity using spatial regression models, such as Spatial Lag Model, Spatial Error Model, and Geographically Weighted Regression (GWR) [
2].
The studies above have enriched our understanding of the relationship between U and E. However, most of the literature [
2,
3,
4,
8,
9,
10,
17,
21,
28] concentrated on unidirectional causality from urban development to various eco-environmental responses. For example, Fang et al. [
2] found that urbanization in China had a significant negative impact on air quality index. The research by Robert J.R. Elliott [
4] showed that urbanization in China boosted energy intensity. Research on the influence of urbanization on SO
2 emissions were also conducted in Shandong Province, China [
20]. Instead, only a few scholars paid attention to the constraining effect of the environmental change on urbanization [
24,
25]. There was even less literature [
15,
22] focusing on the bidirectional relationship between U and E. As pointed out in the literature, such as Dinda [
31] and Fang [
2], this bidirectional influence mechanism between the eco-environment and urbanization has been ignored by most studies. In fact, urbanization and eco-environment are mutually influential and coupled. Environmental degradation and pollution emission also have influences on urbanization process. As for the study area, those studies mainly focused on the spatial scale of a single administrative unit at provincial or municipal level, instead of UAs, which span the boundary of administrative regions. Recently, UAs, paid great attention by Chinese governments, have been regarded as a main form of new-type urbanization in China. Therefore, it is necessary and urgent to study the relationship between U and E in UAs, especially those in the infancy of development. The deficiency in the existing studies motivates the need of an analysis framework to quantify the two-way relationships between U and E in UAs systematically. This framework will help us uncover how urbanization process poses effects on eco-environmental changes and vice versa.
Therefore, in this study, we established such a framework to: (1) identify how U exerted coercing effects on E; (2) how E posed constraining effects on U; and (3) whether this interaction varied temporally and spatially in a UA area in China. UAMRYR, one of the five national UAs, was taken as the study region to examine the feasibility of the proposed framework and the study period covers the year of 2000 to 2015. The results of this study can provide scientific reference for policy-makers to compile effective development plans to promote regional sustainability.
5. Conclusions
The coordinated development between urbanized development and eco-environment protection is an important issue to be solved in China’s new-type urbanization. Identifying and quantifying the impact of U on the E system as well as the reversed relationship is key to solving this issue. Based on the traditional research, which mainly focused on the impact of U on E, we proposed a quantitative analysis framework that takes into account the bidirectional interacting mechanism for investigating the influence from U to E and the constraint from E to U. After the comprehensive evaluation of U level and E quality as well as the identification of the key interacting factors in these two subsystems, this framework detects the coercion from U level and its key factors to E as well as the constraint from the quality of E and its key factors to U based on global regression (OLS here). Further, the framework employs local regression (GWR here) to examine the regional differences of the four types of relationships.
Using this proposed framework, we revealed the interaction between U and E and its spatial differences in UAMRYR during 2000–2015. From the overall relationship between the two subsystems, there was a significant coercion and restraint effect in the U-E system from 2005 to 2015. In the direction of U to E, the coercing effect increased, indicating that the impact of urbanization process on the E system got increasingly exacerbated. In terms of geographical variation in this interaction, it showed a spatial pattern of “low in southwest and high in northeast” in 2010. Comparatively speaking, this pattern became more obvious in 2015, indicating that under the different urbanization stages and eco-environmental background conditions, this influence mechanism presents different trends of development. On the other hand, in the direction of E to U, the constraining effect decreased moderately and showed a spatially stationary pattern, which demonstrated that the intensity of the constraint of E (agricultural land and water environment) on U was relatively small. This is mainly due to the good conditions of the two natural factors in the study area. In terms of influential factors, there were four key human factors (percentage of urban population, proportion of built districts in the total land area, total retail sales of consumer goods per capita and percentage of the tertiary industry in GDP) and two key natural factors (total volume of water per capita and discharge of industrial waste water) in the U-E coupling system. In the coercing direction of U-factors on E, the three factors (percentage of urban population, proportion of built districts in the total land area and total retail sales of consumer goods per capita) exerted a significant impact on E subsystem in 2005–2015. The most serious coercive factor that caused the eco-environmental deterioration was always proportion of built districts in the total land area, although the negative effect it posed became weaker during the study period. Next came percentage of urban population, which had a fluctuating effect on E. As for total retail sales of consumer goods per capita, it effected E positively in 2005, while afterwards led to increasingly negative coercion. Spatially, the impact from percentage of urban population gradually spread from PLEZ to WMA along the northwest-southeast axis of the UAMRYR region. The coercion caused by a proportion of built districts in the total land area on E showed a high level along the northwest-southeast axis but weak outside the axis in 2005. The positive effect of total retail sales of consumer goods per capita on E shifted from the east to the west of the region. In the constraining direction of E-factors on U, although the two key natural factors have a significant and spatially stable constrained effect on U from 2005–2015, they were weak in interpreting the variability of urbanization level. It can be concluded that eco-environmental change in the study area had not yet become a dominant factor constraining the development of urbanization during the study period.
The above findings of the case study will provide an important reference for the compilation of UAMRYR’s new-type urbanization plan. Population urbanization and spatial urbanization should coordinate with economic urbanization and social urbanization, so as to mitigate the negative impact of spontaneous urbanization on the ecological environment. Especially in the region of WMA and PLEZ located in the north and east of the study region, these government control efforts should be strengthened. The government should focus on immigrants’ education and employment, and realize sustainable urbanization with harmonious development of population quantity and quality. The coercion intensity of population urbanization on the ecological environment in the region of RCZTUA was the least and constantly weakened. Thus, this development mode of population urbanization should be adopted by the government of WMA. While emphasizing the expansion of urban built-up areas, more measures should be taken to focus on investing infrastructure as well as adjusting and upgrading industrial structure so as to reduce the coercion of economic development and urban expansion on the eco-environmental system. With respect to the development and construction of built-up areas, the setting of strict development threshold will be the focus of planning.
The framework provides support for investigations on the mutual interaction between U and E subsystems. Our case study demonstrated the usability and applicability of the framework. This analytical framework does not only apply to a specific region such as the study area or specific variables such as U and E in this article. Instead, this framework is useful for establishing a comprehensive understanding of the relationship between any variables as well as its spatial patterns in other regions within the context of urban sustainability.