Introduction

Social–ecological systems (SES) are complex adaptive systems (CAS) consisting of humans, ecosystems and their interactions1. Continuously adapting and evolving elements (human or non-human) and interactions produce macro- or system-level outcomes, such as landscape patterns, and can lead to SES changes, such as regime shifts or transformations2. Although social–ecological interaction has been a long-term focus in the field of sustainability, the nonlinear dynamics of SES, such as threshold and collapse, have not attracted attention in the urban field until recently. Characteristics of social–ecological interaction can provide support in decision-making for sustainable urban management. Urbanization is a multi-scale, dynamic, and complex evolution accompanied by fundamental changes in the structure and function of SES. In terms of theoretical research, there are some theoretical frameworks that help people understand the nonlinear characteristics of the evolution of urban SES by proposing and explaining conceptual models related to regime shift, among which are the adaptive cycle theory3,4, basin of attraction5 and press-pulse dynamics framework6. In terms of empirical research, existing studies mainly focus on dynamic analysis and quantitative evaluation of agricultural and fishery SES, such as agricultural transformation7 and fishing community sustainability8. In recent years, few scholars have attempted to quantitatively analyze the dynamics and turning points of urban SES, mainly including natural resource development9 and urban sustainability10. Although the dynamics of urban SES have begun to gain attention, there are still many pressing questions to be addressed regarding the social–ecological interaction of cities, such as the differences in the social–ecological coupling trajectories of different cities, the driving mechanisms of urban social–ecological evolution, and the impacts of social–ecological coupling on sustainable management. In summary, a specialized framework and empirical research are still lacking to identify the trajectory and dynamics of the social–ecological coupling of urbanization.

The synthesis of the adaptive cycle and the green to red loop provides a new method for studying urban social–ecological coupling dynamics. Adaptive cycle theory believes that a dynamic system will go through four phases (exploitation, conservation, release, and reorganization) and be in a cycle of change11. This model identifies low resilience traps and potential for adaptation12 and has been used to explain and analyze specific phases in the cyclic behavior of complex ecosystems13. In ecology, indicators such as standard deviation, skewness, kurtosis, critical deceleration, and flicker serve as quantitative early warning signals for regime shifts between these phases14,15,16,17. However, SESs are complex adaptive systems shaped by the interaction of social and ecological factors. As a result, the emergent system boundaries are not merely additive combinations of social and ecological boundaries18,19. The diverse conceptualizations of SES, along with varied research questions and methods, hinder the comparison of research results across regions, impeding knowledge generalization and theory building20,21. In addition, a more focused framework is needed to clarify the mechanisms of agricultural transition and urbanization processes, rather than relying solely on a few characteristic indicators and qualitative empirical knowledge. The green to red loop conceptual framework, proposed by Cumming et al.22, has the potential to address these shortcomings23. This framework breaks down the processes of agricultural transition and urbanization into four states: green loop, red loop, green trap, and red trap. Each state is defined by positive or negative feedback loops between humans, local ecosystem functions, and the external environment22. The green loop represents a negative feedback mechanism between the human population and the availability of local ecosystem functions they depend on, maintaining a dynamic balance between human population size and natural resources22,24. The red loop, on the other hand, is a positive feedback mechanism involving the urban population, local ecosystem functions, and imported natural resources. Its sustainability depends on whether human consumption remained within the local environmental carrying capacity22,24. A green trap occurs when there is inadequate food production, while a red trap occurs when people over-consume and fail or refuse to adapt to ecological declines22. This deductive prototype is regarded as a valuable template for identifying the status of SES, facilitating comparison, summary, and the transfer of insights across regions23. The integration of the adaptive cycle and the green to red loop offers a new perspective: it quantitatively identifies general patterns in complex SES through feature indicators and enhances understanding of the driving mechanisms behind agricultural transition and urbanization through feedback loops. This approach aids in understanding the complex feedback and regime shifts in urban SES in the context of environmental change.

This study proposes a new analytical framework that combines the adaptive cycle and green to red loop theory in order to elucidate the long-term evolution and mechanism of urban SES and conducts a case study. First, an analytical framework for the trajectory of urban social–ecological coupling is constructed, and its theoretical basis and operational logic are elaborated in detail. Second, the social–ecological coupling trajectory of China’s Pearl River Delta (PRD) is quantitatively identified in the period between 1980s and 2019, and the impacts of its rapid prosperity on sustainable urban management are assessed. The PRD is chosen as a case study area for the reason that it is one of the fastest developing regions in China. During its rapid prosperity, both its social and ecological structure have undergone significant changes, making it a world representative region for the evolution of urban SES. Finally, the driving mechanism of SES evolution and unsustainable risk dynamic changes are analyzed from multiple perspectives, and certain exploratory findings beneficial to the SES regime shift management are discussed. This paper provides an analytical framework for further research into social–ecological coupling. In addition, it also expands the present empirical knowledge of SES evolution and its driving mechanisms and is significant for promoting regional sustainable management.

Methods

Study area

The PRD is an urban agglomeration situated in the southern part of China. It is surrounded by mountains and hills in the east, west, and north, and faces the sea in the south, with a geographical range of 21°17′–23°55′N, 111°59′–115°25′E (Fig. 1). The main landform of the PRD’s central area is a plain with a dense river network and surrounding hills. The precipitation is greatest in summer, lesser in winter, with typhoons most commonly occurring in summer and autumn25,26. Its administrative area envelops 7 cities including Guangzhou, Shenzhen, Zhuhai, Foshan, Jiangmen, Dongguan, and Zhongshan, as well as parts of Huizhou (its urban area and counties of Huiyang, Huidong, and Boluo) and Zhaoqing (its urban area and counties of Gaoyao and Sihui). The PRD is adjacent to Hong Kong and Macao, one of the first regions in China to implement the reform and opening-up policy27. After the “reform and opening up” in the 1980s, the PRD rapidly transformed into one of the most economically developed regions in China. It is thus a representative region for studying the SES evolution process and mechanism within the context of urbanization.

Fig. 1
figure 1

Geographical location of the Pearl River Delta created using ArcGIS 10.2. Geographical information obtained from the National Catalogue Service For Geographic Information website (https://www.webmap.cn/).

Full size image

Materials

The research data encompass remote sensing data, land use data, meteorological data, basic geographic data, and socioeconomic data. Table 1 provides details on the data sources, time periods, and resolutions. To ensure consistency in spatial data accuracy, all data were resampled at a resolution of 100 m.

Table 1 Details of the dataset and data source information used in this study.
Full size table

Social–ecological coupling trajectory framework

Driven by multiple factors such as population, economy, land use and climate, there are complex linear or non-linear, spatial or non-spatial relationships between social and ecological elements in cities29,30,31. These complex causal feedbacks further lead to self-organization and emergence, making the identification of multi-stability and evolutionary dynamics in urban SES, especially in rapidly developing SES, challenging. Faced with open and complex giant systems like cities, we need a new focus and methods to understand the unique dynamic process of urban SES evolution. A new framework is needed, which should be specially designed based on the general laws of urban development, fully taking into account the diverse social–ecological elements and complex social–ecological interactions in cities, in order to better understand the dynamic laws and mechanisms of urban SES evolution (Fig. 2).

Fig. 2
figure 2

Understanding the process of rapid prosperity from the perspective of the green to red loop.

Full size image

In this study, we construct a new analytical framework, i.e., the social–ecological coupling trajectory framework, to help understand the dynamics of long-term evolution and mechanisms of urban SES. This analytical framework is inspired by two theories of adaptive cycle and green to red loop, including four steps: (1) assessment of the social–ecological evolution dynamics of SES, (2) identification of the state and trajectory of SES, (3) analysis of evolutionary mechanism of SES; and (4) analysis of the impacts of rapid prosperity (Fig. 3).

Fig. 3
figure 3

Analytical framework for the coupling dynamics of urban SES.

Full size image

The first step is the analysis of the characteristics of urban social and ecological elements, and the evaluation of the dynamics of social–ecological coupling. In the process of urbanization, social and ecological elements are incorporated into the urban operating system, merging to form relatively stable SES structures, thus achieving specific urban system functions. As the adaptive cycle progresses, social and ecological elements will undergo periodic coupling and decoupling11. In the phases of exploitation and conservation, the elements are coupled in a coordinated and stable manner in order to ensure the formation of new structures and their stabilization. In the phases of release and reorganization, the partial or complete decoupling of elements leads to the decay of the original structure, and the released elements would be redistributed into new structures of the next cycle. These rhythmic coupling and decoupling enable the quantification of urban social–ecological dynamics (Fig. 4).

Fig. 4
figure 4

The relationship between the adaptive cycle and the social–ecological coupling dynamics.

Full size image

In this study, a comprehensive indicator system is used to integrate multiple statistical data to evaluate the social function of SES, and the Remote Sensing Ecological Index (RSEI) is used to integrate multiple remote sensing information to evaluate the ecological function of SES. On this basis, the coupling coordination index is used to combine social and ecological functions in order to evaluate the dynamics of the social–ecological evolution of SES, thereby achieving the scale correlation of elements, functions, and systems. The details of the calculation are explained in the section of “SES coupling coordination degree”.

The second step is to identify the tipping point and trajectory of urban SES and divide them into different modes of evolution. In the field of ecology, indicators such as standard deviation, skewness, kurtosis, critical deceleration, and flicker can quantitatively describe the regime shift of an ecosystem15. In the field of social ecology, identifying the tipping points in the evolution of SES is crucial for formulating sustainable urban development plans, but relevant frameworks and quantitative methodologies are lacking. In this study, we propose to use the partial and complete decoupling of social and ecological subsystems as an indicator of regime shift. As mentioned in the first step, the social–ecological coupling of urban SES undergoes rhythmic changes along with an adaptive cycle. In theory, decoupling in the release phase would lead to a deterioration of the coupling coordination relationship. We believe this is a breakthrough point for quantitatively capturing the regime shift signals of urban SES. Therefore, the decrease in the growth rate of the coupling coordination degree is considered to be a signal for a regime shift, which divides the social–ecological coupling process into multiple phases in order to clarify the evolution trajectory of urban SES (Fig. 4). Typical modes of urban SES evolution can be obtained by summarizing the evolution trajectory characteristics of different cities. Details of this identification process are explained in the section of “SES coupling coordination degree”.

The third step is to clarify the causal feedback of trajectory and analyze the mechanisms of SES evolution. Taking decoupling (the release phase in adaptive cycle) as a tipping point, the social–ecological evolution trajectory can be divided into multiple green to red loop states (Fig. 5). It is important to note that urbanization does not always adhere strictly to the four stages outlined. In reality, cities may experience intersecting, overlapping, or skipping of stages depending on the country, region, climate, economy, and politics. Since the complexity of real-world problems must be simplified when mapped into theoretical frameworks, our model divides urban development into four standard stages according to the green to red loop. This approach may have limitations, and specific patterns not covered will be discussed in the case analysis.

Fig. 5
figure 5

The matching between adaptive cycle and green to red loop.

Full size image

Based on the green to red loop theory, the evolution mechanisms of urban SES can be presented as causal feedback loops (Fig. 6). Due to the limitations of method feasibility and data availability, the variables and feedback loops included in this step are only used as a theoretical basis for understanding the research results and not as part of the quantitative research. The feedback loops are designed based on the definition of sustainable urban development, i.e., improving the quality of life (QoL) without leaving an ecological burden, and by taking into account QoL and ES deficit as key variables32,33.

Fig. 6
figure 6

Dominant feedback loops of social–ecological coupling in urbanization.

Full size image

QoL includes components of well-being that are related to general factors such as health, welfare, freedom of choice, and basic liberties, as well as more specific ones such as the availability of food, clothing, drinkable water, educational facilities, health care, and income34. Many studies have attempted to provide an appropriate representation of all those dimensions using social, economic, and environmental indicators35,36. Given the research focus and data availability, the QoL in this study mainly includes transportation facilities, education, health care, technological services, and consumption levels of residents. The QoL affects the choice of variables in the section of “SES coupling coordination degree”, but is not calculated as specific indicator. ES includes providing food, clear water and energy, improving quality of water and air, providing support for habitat and biodiversity as well as landscapes for leisure and entertainment. ES deficit is the difference between population demand for ES and local ES supply capacity. The greater the deficit of ES, the greater the burden that SES causes on the external environment. The basic principle of feedback loop is that negative feedback promotes mutual inhibition and coordination to form a stable state, while positive feedback enhances the system’s response to small changes.

Specifically, the social–ecological interdependencies in the green trap and green loop can be represented as a dominant loop composed of blue blocks in Fig. 6. In the green trap, people are trapped in poverty due to the inability of poor or overexploited ecosystems to provide sufficient resources, and their quality of life cannot be guaranteed, which results in unsustainable SES development. Systems in the green trap must be reorganized in order to improve the economic quality of life. In the green loop, people earn income by developing local ecosystems, such as deforestation and mining of minerals, to ensure economic quality of life. At the same time, the local ecosystem is able to provide enough ESs for people because it is not excessively disturbed, i.e., there is no ES deficit, or the ES deficit is very small. Therefore, in the green loop, people's quality of life is guaranteed and there is little load on external environment, and SES is in a relatively stable state. The red loop and the red trap are post-urbanization phases, which can be represented as a dominant loop in Fig. 6, including the blue and red blocks. In the red loop, people promote urbanization through the development of natural and land resources in order to pursue prosperity. The increase in construction land creates two positive feedbacks related to population and wealth (① and ② in Fig. 6) by supporting the development of secondary and tertiary industries, promoting the prosperity of urban society and economy, and encouraging the continuous expansion of construction land. Increasing construction land is encroaching on ecological areas, and increasing population and industrial activities are disrupting ecological processes. These impacts result in a reduction in the local supply of ES. When the local ecosystem is unable to meet people’s demand for ES, an ES deficit occurs. Although people can obtain ESs from the external environment as a supplement, ES deficit will inevitably have a negative impact on people’s ecological quality of life and social equity. Majewska et al.37 point out that the lack of well-landscaped green areas (squares, outdoor gyms, small parks) near one’s own home is becoming a serious problem in cities. In the red loop, the economic quality of people's lives has improved, which is a considerable burden on the external environment, while SES is still in a relatively stable state. When human activities interfere with an ecosystem beyond its resilience, even a small change can trigger regime shifts, driving SES into the red trap. Such regime shifts often result in negative impacts, such as degradation of ecosystems and comprehensive collapse of social, economic, and ecological systems38. Therefore, this negative regime shift should be avoided.

The fourth step is to analyze the challenges that rapid prosperity brings to urban sustainable management. Rapid prosperity triggers rapid changes in the social–ecological structure, affecting the stability and sustainability of social–ecological functions and posing challenges for the sustainable urban management25,39. This study assesses changes in ES supply and ecological environmental management performance in each phase of the social–ecological coupling in order to analyze the cumulative effects of SES evolution on sustainable urban management. The details of the calculation are explained in the sections of “selection and quantification of ecosystem service” and “ecological management efficiency”.

This framework analyzes the evolution process of urban SES from four aspects: social–ecological coupling, multi-stability, causal feedback, and sustainable management. It proposes a method for social–ecological study, combining the adaptive cycle and green to red loop theory in order to identify important time periods and areas of sustainable urban development. Through a comprehensive analysis of the dynamics and impacts of the social–ecological coupling, it also provides suggestions and insights for the sustainable urban management.

SES coupling coordination degree

The coupling coordination degree represents a comprehensive index derived from capacity coupling and the physical capacity coupling coefficient model, which is used to describe the SES evolution process30,40,41. This study uses the coupling coordination degree to describe the social–ecological coupling level of SES. The formula is given as Eq. (1):

$$ {\text{D}} = \sqrt {C \times T} $$
(1)
$$ {\text{C}} = 2 \times \sqrt {\frac{{{\text{RSEI}}_{i} \times M_{i} }}{{({\text{RSEI}}_{i} + M_{i} )^{2} }}} $$
(2)
$$ {\text{T}} = \frac{{{\text{RSEI}}_{i} + M_{i} }}{2} $$
(3)

where D represents the SES coupling coordination degree, expressed as a function of the coupling index (C) and the coordination index (T). The value range of D is 0–1, with a higher value indicating a higher level of coupling. C and T denote the functions of ecological state and social–economic state, respectively. The socio-economic state is represented by M, which is calculated using index evaluation based on the entropy weight method (Table 2). The selection of variables referred to existing research30,40,41 and relevant aspects of the feedback loops mentioned earlier. Detailed weight parameters are provided in Supplementary Table S1. The ecological state is represented by the RSEI, a comprehensive ecological index. The RSEI combines multiple indices: Normalized Difference Vegetation Index (NDVI), Wetness Index (WET), Dryness Index (NDSI), and Land Surface Temperature (LST). These indices were calculated using Landsat 5 and Landsat 8 data and were integrated into the RSEI through Principal Components Analysis (PCA) after normalization. For more details on RSEI calculation, refer to the existing researches42,43,44.

Table 2 Evaluation index system of social economic state in the PRD.
Full size table

The social–ecological coupling dynamics of PRD are measured and fitted based on the D index. An upward trend in D values indicates that SES stabilizes the system structure through a higher level of social–ecological coupling, while a non-upward trend in D values (stopping growth or decreasing) indicates that SES is undergoing social–ecological decoupling, either partially or completely. The decoupling leads to the decay of the system structure and is used as a signal for regime shift of SES in this study. After comparing multiple fitting results, the BiDoseResp equation provides the best fitting effect, with an R-Squared value exceeding 0.98. The standard equation for this model is provided as Eq. (4). In this formula, y corresponds to the D value, x corresponds to the time (year), and all other symbols are regression parameters. The parameters of logx01 and logx02 are used to identify the tipping points of SES.

$$ y = A_{1} + \frac{{(A_{2} - A_{1} ) \times p}}{{1 + 10^{{(\log x_{01} - x) \times h_{1} }} }} + \frac{{(A_{2} - A_{1} ) \times (1 - p)}}{{1 + 10^{{(\log x_{02} - x) \times h_{2} }} }} $$
(4)

Selection and quantification of ecosystem service

Ecosystem service represents the link between the social economic system and the ecosystem, directly reflecting the coupling of humans and nature. Since the PRD is a delta urban agglomeration area, this study chose water yield (WY), soil retention (SR), carbon storage (CS), and grain production (GP) as the evaluation objects of ecosystem services. WY, SR, and CS were calculated by the InVEST software (version 3.11) mainly based on land use data, while GP was calculated based on the statistical data of grain production for each city. Per capita data were used to reflect the direct supply capacity of all ESs. Since the PRD population increased by 3.1 times during the study period, while its ESs are limited, using per capita data may better reflect the ecological quality of life that people can obtain. Among them, the research scale of water yield (WY), soil retention (SR), and carbon storage (CS) is 100 m, while the research scale of population and grain production (GP) is urban scale. Furthermore, by standardizing the ES value to be 0–1, the evaluation results of different ecosystem services become comparable.

Ecological management efficiency

The input–output ratio represents the ecological management performance through the ratio of ecological management input and output. This ratio is also a common method in the field of sustainability evaluation45,46. In this study, input denotes people’s efforts on ecological protection and management, while output represents the ecological quality. A smaller input–output ratio indicates a greater output under the same input, reflecting stronger SES sustainability. This study calculates the input by the index evaluation method. The evaluation index system is comprised of six indicators in four aspects: wastewater treatment, solid waste treatment, ecological construction, and legal supervision (Table 3). The indicator weights are obtained by the entropy weight method. See Supplementary Table S2 for the specific weight values. Lastly, the output is represented by RSEI.

Table 3 Evaluation index system of ecological management input in the PRD.
Full size table

Results

Social–ecological coupling trajectory

The coordination degree fitting results indicate that the speed of SES coupling in the PRD was stable from 1986 to 1989. It then increased from 1990 to 1992 and declined from 1993 to 2000. Finally, it began to accelerate again in 2001 and then decelerated between 2007 and 2019 (Fig. 7a). Further analysis suggests that the SES changed from green loop to red loop, with an evolution trend to red trap (Fig. 7b).

Fig. 7
figure 7

Evolution process of social–ecological synergy in the PRD region from 1986 to 2019.

Full size image

At the city scale, the degree of coordination among cities in the PRD shows an increasing trend (Fig. 8). Analyzing geographical spatial scale, Guangzhou, located in the central region, has consistently maintained a leading level of coordination, with Shenzhen in the southeastern region closely following. Between 2016 and 2019, Foshan, Dongguan, and Huizhou experienced an increase in their coordination levels, reaching a secondary leading position. These cities established strong spatial and inter-city connections with Guangzhou and Shenzhen, respectively. Overall, by 2019, the coordination degree of cities in the eastern part of the PRD was higher compared to the western part. Based on the fitting results on the city scale, the SES evolution trajectories may be divided into two types: step mode and skip mode (Fig. 9). In the step mode, the SES goes through green trap, green loop, and red loop states in turn. The observed deceleration of coupling coordination may be an early warning signal of regime shift. In the skip mode, SES directly switches from the green trap to the red loop, bypassing the green loop state, thus showing a relatively stable change rate of coupling coordination throughout the research period. Cities including Guangzhou, Zhuhai, Foshan, Huizhou, Zhongshan, Jiangmen and Zhaoqing belong to the step mode, while Shenzhen and Dongguan are part of the skip mode.

Fig. 8
figure 8

Distribution of coupling coordination degree in the PRD from 1986 to 2019. Note: Dongguan (DG), Foshan (FS), Guangzhou (GZ), Huizhou (HZ), Jiangmen (JM), Shenzhen (SZ), Zhaoqing (ZQ), Zhongshan (ZS), and Zhuhai (ZH).

Full size image
Fig. 9
figure 9

SES evolution mode of cities in the PRD from 1986 to 2019.

Full size image

All cities in the step mode experience three states: green trap, green loop, and red loop. However, their deceleration behaviors before the regime shift are different. The SES evolution trajectory of Guangzhou, Foshan, and Huizhou (Group A) has two deceleration periods, with the system being in the green loop state for a short period (Fig. 9a). On the other hand, even though the SES evolution trajectories of Zhongshan, Jiangmen, and Zhaoqing (Group B) also have two deceleration periods, their decelerations are not as obvious as those of group A (Fig. 9b). Thus, when compared to group A, the evolution speed of group B is slower, and its green loop period is longer. The evolution trajectory of Zhuhai (Group C) differs significantly from that of other cities in the step mode (Fig. 9c). For Zhuhai, no green trap to green loop hysteresis was observed. Moreover, its deceleration from green loop to red loop was the most obvious among the cities. Compared to the step mode, the SES evolution of Shenzhen and Dongguan (Group D), which are under the skip mode, is more profound. In addition, its risk of falling into the red trap is also notably higher (Fig. 9d). See Supplementary Fig. S1 for the fitting curve of each group.

Increasing sustainability challenges

For ecosystem service supply capacity per unit area across various cities (Fig. 10), Shenzhen and Dongguan saw increases of 18% and 14% respectively, while other cities generally remained stable. In the northern part of the PRD, Zhaoqing, Foshan, and Guangzhou experienced increases of 50% or more. In contrast, Zhongshan, Zhuhai, and Jiangmen in the western region saw increases ranging from 26 to 43%, while the growth in Dongguan, Huizhou, and Shenzhen in the east did not exceed 9%. CS in the PRD decreased from 1980 to 2000 but remained stable from 2000 to 2019. Dongguan, Foshan, Shenzhen, Zhongshan, and Zhuhai all experienced decreases of more than 11% in CS, while other cities remained stable. GP showed a significant downward trend, with 6 out of 9 cities experiencing declines of over 82%, and Zhongshan, Foshan, and Guangzhou seeing reductions greater than 95%. These results highlight significant spatial differences in ecological quality of life within the PRD and suggest an increasing risk that urban SES may impose on the external environment. This trend reflects the trade-offs between ESs and urban development during the research period.

Fig. 10
figure 10

Supply of ecosystem services in the PRD from 1980 to 2019. Note: Dongguan (DG), Foshan (FS), Guangzhou (GZ), Huizhou (HZ), Jiangmen (JM), Shenzhen (SZ), Zhaoqing (ZQ), Zhongshan (ZS), and Zhuhai (ZH).

Full size image

Ecological management efficiency indicates that the challenges facing sustainable development in the PRD have rapidly increased since 1995. In terms of the ecological management effort, there was small input with a downward trend (0.18 to 0.10) between 1986 and 1995. It then continuously increased from 0.10 to 0.72 between 1995 and 2019 (blue line in Fig. 11). In terms of ecological quality, RSEI fluctuated between 0.78 and 0.84 from 1986 to 2003. However, since 2003, it continuously decreased from 0.84 to 0.74 (green line in Fig. 11). Ecological management efficiency in the PRD thus exhibits a staged change from high efficiency to low efficiency. With a continuous decrease in the input–output ratio from 0.23 to 0.12 between 1986 and 1995, high ecological quality was maintained with little effort during this period. Furthermore, since 1995, the cost of ecological quality maintenance continued to increase, while the input–output ratio increased to 0.97 in 2019 (orange line in Fig. 11). The increasing input–output ratio reflects a continuous decline of ecological recovery, indicating that the sustainable development of the PRD is faced with serious challenges. The growing ecological management efforts implemented since 1995 have failed to reverse the declining trend of ecological quality. Their failure demonstrates that socio-economic interference may have notable negative consequences to regional sustainable development through cumulative effects.

Fig. 11
figure 11

Dynamic changes of ecological management performance in the PRD region in the period 1986–2019.

Full size image

Discussion

Impact of policy reform on SES evolution

Driven by rapid social and economic prosperity, SES in the PRD has shifted from green loop to red loop, with a trend to red trap. These observations are consistent with the idea that SES evolves in phases47. Further analysis has shown that policy reform has a decisive role in SES phase development.

The PRD represents a pilot area for “reform and opening up”, which recognizes economic prosperity as its primary goal. Within this reform, a series of new systems related to contemporary enterprises, finance and taxation, and foreign trade promoted the PRD so that it would rapidly complete the transformation from a planned economy to a socialist market economy. Similarly, the medical reform and housing marketization encouraged the improvement of medical efficiency and the growth of commercial housing48. The preferential tax policy implemented during the reform has attracted a large amount of commercial investment. Furthermore, adequate employment opportunities, advanced education and medical care, and sound urban infrastructure have attracted people to relocate to the PRD, resulting in a 3.1-fold increase in the region’s permanent population between 1986 and 2019. The rapidly growing population and high intensity of human activities drove the PRD’s transition from green loop to red loop. However, once the socio-economic disturbance exceeds the carrying capacity of the local ecosystem, ecological resilience may sharply decline and thus cause irreversible negative consequences38,49. This is visible as the ecological efficiency takes on a downward trend and urbanization increases during the SES evolution39.

During the research period, there were two decelerations of the coupling coordination degree in the PRD. Both may be related to the intensity of human activity that can approach the threshold of ecological carrying capacity. Moreover, the recovery of coupling coordination may be affected by people's awareness of ecological protection, active intervention in ecological management, and adjustment of socio-economic structure. At present, SES in the PRD is experiencing a second slowdown, and reducing its risk of falling into red trap is a serious issue. One possibility is that the relationship between environmental pressure and per capita income is an inverted U-shaped curve50. In other words, the environmental pressure increases with the per capita income during the SES coupling stage and decreases if SES enters the decoupling stage. Faced with unprecedented challenges in balancing ecological security and economic prosperity, sustainable urban development should encourage the decoupling of economic prosperity and environmental pressure. China's agenda of ecological civilization is introducing a series of policies conducive to the goal of decoupling, which may in turn trigger a positive tipping point for SES in the PRD.

Cumulative effect of rapid prosperity

During the research period, the ESs supply and ecological management efficiency continued to decline as socio-economic prosperity grew. The characteristics of ecological management efficiency partially reflected the fact that the PRD’s rapid prosperity had a negative cumulative effect on its sustainability. As the intensity of human disturbance was still within the threshold of ecosystem resilience before 1995, the self-repair ability of the ecosystem made low input and received high returns. However, when the disturbance intensity exceeds the threshold, ecological and environmental problems can threaten the livability of cities. After the importance of ecological security was realized, the investment of funds and personnel into ecological management increased and the legislature concerned with ecological protection improved. However, even though increasing investments into ecological management have achieved certain results, inefficient ecological management has not been successfully reversed because the ecosystem resilience has been damaged. This indicates that with a trade-off between prosperity and ecological security, long-term high-intensity human activities damage the quality of the environment and have a far-reaching negative impact on regional sustainable development.

According to the entropy theory, the influence of a variable system can be divided into fast and slow variables. The cumulative effect of slow variables is an important cause of the resilient degradation and collapse of the ecosystem, which in turn determines the critical threshold of the structure and system change of SES51,52. Thus, identifying key slow variables and strengthening the positive feedback or weakening the negative feedback of the current state to trigger a positive tipping point may be a solution to slowing down or even reversing the SES constant approach to the red trap53,54,55. In the context of the Sustainable Development Goals, the cumulative ecological effects of rapid prosperity provide an important reference for developing countries seeking rapid prosperity to scientifically formulate their own development plans.

Inspiration from SES specific behaviors

In general, the evolution trajectories show that the coupling coordination rate of SES in the PRD tends to decline before entering a new state. This observation demonstrates that “slowing down” is an early warning sign of regime change not only for the ecosystem16 but also for SES in urban areas. When an ecosystem is close to its tipping point, there may be a critical deceleration or growth rate, flicker, or other warning signals, which have been observed through biological physical chemical monitoring and control experiments16,17,56. Similar to regime shift, “non-marginal change” is a notion found in socio-economic literature to contrast gradual marginal change. In socio-economic systems, regime shift is most commonly related to the macroeconomic cycle, which is characterized by “creative destruction” during an economic downturn14. However, in SES, regime shift involves more complex internal correlations and external influences, making its recognition much more challenging57,58. The deceleration signal found in this study provides a basis for detecting early warning signs of SES regime shift in urban areas.

In the PRD, the SES evolution trajectories are affected by multiple factors and show two different modes, i.e., step and skip. The expansion of construction land is the main factor driving the change of landscapes, ES supply, and livelihood in urbanization areas59. The positive feedback formed by construction land, population, and social wealth is the primary driving force for the coordinated evolution of the socio-economic subsystem and ecological subsystem in the PRD. Under the step mode, the negative impact of construction land expansion on the ecosystem is gradual rather than thorough. Moreover, the relationship between humans and nature evolves in a coordinated way during this relatively slow change. However, under the skip mode, expansion rate of construction land exceeds a certain threshold, the positive feedback effect of the social subsystem is strengthened, the local ecosystem is degraded, and the relationship between humans and nature is changed into a red loop. In this process, obtaining remote ecosystem services rather than limiting the expansion of local construction land further strengthens the feedback mechanism of the skip mode.

Limitations

One of the limitations of this study is that ES per capita used to represent the support of local ecosystems for urban ecological quality of life may be uncertain. Insufficient ES per capita means that people have to take external environmental ES as a supplement to ensure ecological quality of life. The potential for urban SES to cause a burden on its external environment is increasing. However, the urban system is a system that continuously relies on external supply of resources. The living needs of a urban large population are difficult to meet with the local ecosystem and need to be supplemented from the external environment. Although ES per capita can reflect the increasing population pressure facing an ecosystem, it cannot provide more detailed information about the supply and demand process of ES. On the other hand, local ecosystems can serve not only the local population, but also the population of other regions60. Therefore, a more detailed analysis of urban ES supply and demand balance requires a specialized investigation of ES sources, flows, and sinks.

Second, the assumption about the expansion of the city’s construction land may be flawed. In this study, large-scale construction land encroachment on local ecological land is considered detrimental to the sustainability of urban SES. This assumption is theoretically reasonable based on the definition of sustainable urban development, but real-world situations require a more detailed discussion. In the practice of urban land management, the selection of highly degraded ecological areas around cities for construction and restoring ecosystems in remote areas is a strategy that can be considered. This is a complex cross-regional issue that requires specialized and in-depth research into the effects, impacts, stakeholders, and ecological compensation based on cross-regional land replacement.

In addition, the theoretical framework proposed in this study has limitations when addressing the complexity of urban symbiosis. The PRD, as a significant special economic zone in China, promotes integrated urban development through coordinated social, economic, and ecological planning by the government61. In this context, the urban elements and functions of the nine cities within the PRD are closely interconnected and can theoretically be viewed as a larger-scale SES research subject. The results from the social–ecological coupling analysis (in the section of “social–ecological coupling trajectory”) also support the notion that the overall development pattern of the PRD aligns with the assumptions of the theoretical framework. However, while fostering interaction and functional sharing among urban elements in cities, the integrated development of urban agglomerations—such as large-scale population movements between cities62 and spatial functional divisions within urban agglomerations63—tends to blur the boundaries of individual cities. This blurring can lead to errors in research that relies on administrative regions as system boundaries. In addition, due to inter-city relationships, the transmission of urban ecological risk and their interactions with system elasticity exhibit nonlinear network characteristics64. Factors such as shared water resources, air pollution diffusion, ecological protection and restoration, and geographic information sharing further complicate SES research in urban agglomerations. Therefore, a more comprehensive approach is needed, incorporating rese theories and methodologies related to urban agglomeration symbiosis, multi-center structures, and multi-scale cascading.

Although the results of this study have direct theoretical support and indirect verification from similar studies47, their further evaluation could make them more convincing. In further study, workshops should be organized to compare and evaluate the rationality and accuracy of similar models. In addition, due to the limitations of method feasibility and data availability, the social–ecological coupling mechanism of urban SES in this study is only qualitatively explained and not further quantitatively analyzed. With the rapid development of big data, cloud computing, and digital twin, quantitative simulation of urban social–ecological coupling mechanism, as an important foundation for building smart cities, will become an important trend in future urban governance.

Conclusion

This study proposed an analytical framework to uncover the dynamics and mechanism of urban social–ecological coupling and simulated the social–ecological evolution of China's rapidly prosperous PRD as a case study. The results revealed the evolution mode and driving mechanism of SES in areas of rapid urbanization with the hope of providing solutions to regional sustainable management. The main conclusions of this research include: (1) The social–ecological relationship in the PRD evolves in phases. Its four ESs have continuously declined to 40% or less of its original value in the 1980s. Because of the disproportionate decline of the ecological management efficiency, the PRD faces sustainability challenges; (2) The “reform and opening up” is the decisive factor promoting the evolution of SES to red loop. Moreover, the cumulative effect of sustained high-intensity socio-economic interference is the reason for the disproportionate increase in sustainability challenges; and (3) The deceleration of SES presents a possibility of predicting the regime shift of complex systems. Furthermore, the skip behavior may provide a basis for identifying the regime shift thresholds. This paper expands the knowledge of SES sustainable management from the perspective of social–ecological coordinated development. In addition, it provides reference for understanding the characteristics of global rapid urbanization, especially for developing countries in order to formulate reasonable urbanization development plans.