Evolution of County Socio–Ecological Systems in Nature Reserves in Western China Over the Past 30 Years

The evolution process of socio–ecological systems from 1990 to 2020 in the Tianzhu Tibetan Autonomous County, located in the National Key Ecological Function Area of western China, was analyzed quantitatively based on resilience theory and methodology combined with catastrophe theory and adaptive cycle theory, through field investigation, questionnaires and interviews, social data collection and remote sensing data analysis. The results show that the coordinated development of socio–ecological systems has made great progress in recent decades in the study area, and the coordinated development of social systems and ecosystems has a high degree of coupling and a strong connection. Changes in ecosystem resilience regularly surpassed changes in the social system, indicating the significant impact and success of ecological protection policies and projects in recent decades. In future, improvements in the social sub-system will be the key to developing the socio–ecological system in the study area. Enhancing social sub-system resiliency, implementing transformational development and green industry development, and transforming and realizing ecological product values are important topics for further investigation in the study area. Substantial changes in policy, production, population and climate change are needed to promote the evolution of socio–ecological systems. Stable national policies are crucial for improving people's livelihoods and providing ecological protections.


INTRODUCTION
Many ecosystems may undergo regime shifts (Rocha et al., 2015Biggs et al., 2018;Guerrero et al., 2018), which are large, sudden, and persistent changes in the function and structure of an ecosystem (Scheffer and Carpenter, 2003;Scheffer and Jeppesen, 2007;Scheffer et al., 2009). Identifying the stages of socio-ecological system evolution and the drivers of long-term regime shifts is crucial for the success of future systems management (Reyers et al., 2018). Identifying regime shifts in the context of intertwined socio-ecological dynamics remains a scientific challenge and an emerging frontier in socio-ecosystem research (Reyers et al., 2018;Wu et al., 2020). Sustainability science has made great progress in constructing theoretical foundations for socio-ecological systems (Holling, 2001;Chapin et al., 2009;Pacheco-Romero, 2021) and devising conceptual frameworks to operationalize these theories (Ostrom, 2009). Sustainable development should strive to maintain desirable institutions or prevent the creation of damaging institutions by managing the drivers that influence key feedback processes (Reyers et al., 2018).
Based on a literature analysis, current socioecological research mainly focused on three areas. The first area is the study of adaptive governance of socio-ecological systems. Transformational governance and collaborative governance are the two most common approaches. The most significant features of transformational adaptation are restructuring, path-shifting, innovativeness, multiscale attributes, systemwide attributes, and persistence (Fedele et al., 2019). Such governance involves significant shifts in fundamental properties, functions, or interactions within the social, ecological, or socioecological system (Fazey et al., 2018), alters the current trajectory of the system by pushing the system toward an alternative direction (Dakos et al., 2015), and changes the system to a new state that has not previously existed in the region (Biggs et al., 2010;Feola, 2015). In terms of collaborative adaptation and governance, ecosystems constitute complex entities that span spatial and temporal scales and often do not align well with various jurisdictional and political demarcations (Bodin, 2017). Thus, the ability to fit the scale and extent of ecosystems with appropriate structures of governance is affected by institutional differentiation (Bodin, 2017). The second area is research on regime shifts in socio-ecological systems. Typically, the key dynamics of a socio-ecological system across system thresholds are determined by interactions between multiple variables (Norgaard et al., 2009;Wei et al., 2018;Wu et al., 2020); however, the most common methods for identifying the evolutionary stages of socio-ecological systems include analyzing a single change point of social or ecological components or exploring its driving force. The disadvantage of these methods is that their interactions are often overlooked; a single indicator change point may not accurately capture relevant information regarding key regime shifts. The third area comprises the driving force of socio-ecological evolution. Research on the driving force of system and regime shifts is also an emerging focus in socio-ecological systems research. Usually, multiple driving forces are involved in systemic change (Rocha et al., 2015). However, agricultural production and climate change are often intertwined and become the main drivers of regime shifts Reyers et al., 2018).
In the study of the evolution of socio-ecological systems, analyses are usually based on the change point of a single social or ecological component; they include the continuous t-test and f-test for the difference between two or more systems or early warning signal prediction methods (Scheffer et al., 2009. In addition, Mann-Kendall analysis and regime shift point analysis are employed. However, the key dynamics in socioecological systems are usually determined by the interaction between multiple variables. Therefore, the interaction between society and ecology is often ignored when the change point method is used for analysis. As a result, some key regime shifts may not be informed accurately. This study considers the interaction process between the components of socio-ecological systems and applies the catastrophe theory and adaptive cycle theory to analyze the evolution of socio-ecological systems. It considers this multi-objective drive under socio-ecological interactions from a comprehensive and systematic perspective. The results can provide a lens for regional socio-ecological system management, thereby providing a scientific basis and decision-making basis for ecological management and regional sustainable development and related regions. A nature reserve is a complex social and ecological system (Cumming and Allen, 2017), and for nature reserve management, the concept of a socioecological system is very useful (Cumming et al., 2015;Ban et al., 2013) because it can be defined clearly. Resilience, an important component of sustainability, can be viewed as the likelihood that a system can retain its identity. This paper takes the evolution of the county socio-ecological system in the key national ecological function areas in western China as a case and analyzes the characteristics of system evolution over the past 31 years. The structure of this paper includes six chapters: the first chapter is the introduction, the second introduces the research area, the third introduces the research methods and theories, the fourth describes the results of analyses, the fifth is the discussion, and the last chapter draws the conclusions. The main highlights of the research are: (1) social systems and ecosystems have a high degree of coupling and a strong connection in Tianzhu County; (2) by solving the contradictions of the sub-systems, the socioecological systems completed the evolution of each stage; and (3) through the evolution of adaptive cycles, the socio-ecological system moves toward sustainability.

STUDY AREA
The Tianzhu Tibetan Autonomous County (hereafter referred to as Tianzhu County) is located at the eastern end of the Hexi Corridor in Northwest China ( Fig. 1) at the intersection of the Qinghai-Tibet Plateau, Loess Plateau and Inner Mongolia Plateau. The terrain is high in the northwest and low in the southeast, and the altitude range is 2,040-4,874 m. The Tianzhu County was taken as a typical case area in this study based mainly on the following considerations. First, it is located in the National Key Ecological Function Area of western China, which is a water conservation area. The Qilian Mountain National Nature Reserve in the Tianzhu County covers an area of 4,390 km 2 , accounting for 66.9% of the county. The county is also the source of the Shiyang River, an important inland river in northwest China located in the middle and upper reaches of the Yellow River Basin. Second, it is a key ecological security barrier area and an ecological function area in China, and its ecological status is extremely important. The Tianzhu County is dominated by grassland and forest, among which water conservation forest accounts for 96.4% of its total forest area and 40% of the water conservation forest area in the Qilian Mountains. Annual runoff reaches 1.024 billion m 3 , which plays an irreplaceable, strategic role in maintaining national ecological security. Third, the Tianzhu County is also the only Tibetan Plateau autonomous county in the Hexi Corridor in Northwest China. The promotion of social development and improvement of living standards is as important as the protection of ecology.
The Tianzhu County is dominated by the planting industry and animal husbandry. In 2020, the Tianzhu County had a permanent resident population of 150,600 people, with GDP of 4.89 billion yuan and an urbanization rate of 46.37%. Its economy is underdeveloped, and the contradiction between agriculture, forestry, grass and animal husbandry is prominent. Over the years, the immoderate utilization of resources in Tianzhu has caused degradation of the ecosystem; however, since approximately 2017, a large number of concentrated ecological restoration and remediation works have been carried out (Fig. 2).

Methodology
(1) Most natural and anthropogenic changes follow a common process in which a constant change in factors triggers abrupt changes (Scheffer et al., 2001). The catastrophe theory addresses the phenomena and laws of transitions from one stable state to another (Thorn, 1969;Zeeman, 1976). It describes the evolution of forms and is particularly suitable for the analysis of sudden dynamics caused by gradually changing forces (Li, 2018). It can detect the transformation of human and environmental stochasticity in resilience shifts between different equilibria, thus revealing the adaptability of a resilient system. The outcome of applying the catastrophe theory represents the transformation in a coupled system over time rather than a simple summation of the values of all indicators (Li, 2018).
Abrupt changes in ecosystems refer to the transformation process of ecosystem structure and function in response to gradual changes in resources and environmental elements (such as grazing intensity) or sudden changes (such as sudden extreme climate events or natural disasters) during ecosystem evolution. Catastrophe theory usually contains models with different equilibrium surfaces, such as folds, cusps, swallowtails and butterflies. The dimension of the control variables in the mutant system determines the computational model, which means that the number of indicators in the sub-system determines the model choice. The catastrophe phenomenon is illustrated by a potential function ( Table 1). The normalization is via a basic calculation formula for comprehensive analysis and judgment in the catastrophe theory, and it can be obtained by decomposing the bifurcation equation. The normalization formula can be used to quantify the recursive operation of a system and further determine the value of the mutation membership function that represents the state of the system.
Since the catastrophe theory follows a hierarchical process, evolution values are calculated from indicator to sub-system and then to socio-ecological system. For example, the social sub-system includes four groups of indicators: policy, economy, production, and population. First, the resilience values of the four groups are calculated. Because the policy group has 1 indicator (1 dimension), the fold model is used for policy resilience values; the economy and production groups have four indicators (four dimensions), and so both groups adopt the butterfly model. The population group uses a cusp model. Afterward, the butterfly model is used to evaluate the resilience values of the social sub-system from the values of the four groups of indicators. The calculation of the ecological sub-system uses the same process. Finally, the resilience values of the socio-ecological system are derived through the cusp model (two-dimensional) by integrating the overall resilience value of the society and the ecosystem. The description of catastrophe theory is summarized in the normalization formulas in Table 1.
(2) The adaptive cycle theory describes how systems change. The adaptive cycle has four stages Gunderson & Holling, 2002): rapid growth and exploitation (c), protection (j), collapse or release (creative reconfiguration) (x), and renewal or recombination (a). In the development phase (c), the system is in a state of rapid growth, which for the socio-ecological system indi- cates that its resilience and coordination are constantly improving. During the protection phase (j), socio-ecological resilience slows down. In the release phase (x), the coordination between the components in the system is weakened, and the regulation and control ability of the system is decreased. In the recombination stage (a), the uncertainty generated in the release stage helps to readjust the state of the system, and the system enters a new stage. In this study, the evolutionary stages were identified by using evolutionary progression, the adaptive cycle theory and the implementation time of important policies in the study area.
(3) Transformation of socio-ecological evolution progression. The levels of socio-ecological evolutionary resilience derived from the catastrophe calculations are very close to each other (Poston and Ian, 1978). At the present stage, there is no unified standard for the classification of socio-ecological evolutionary resilience. After referring to a number of similar studies, the study divides the resilience of the system into five levels (Huang, 2009;Sun, 2011;Li, 2018Li, , 2022. Therefore, through K-means clustering analysis in SPSS software, the comprehensive values in different systems were divided into five different levels to generate five levels of interval values in the clustering results. Because the catastrophe theory was used for hierarchical calculation, the index and the cluster values of sub-systems and socio-ecological factors were analyzed separately (Li, 2018). Each indicator was then classified according to Table 2, which shows different levels of coordination and resilience from low (level 1) to high (level 5).

Index Framework and Data Sources
According to the socio-ecological system framework of Ostrom (2009), the core sub-systems were first divided into social and ecological subsystems based on the social, economic, and political settings in the study area. In terms of the social subsystem, the key indicators that can reflect comprehensively the social-and ecological-related policies, economic development, agricultural and industrial production, and population of the study area in the past decades were selected based on the availability of data. In the ecosystem sub-system, the key indicators of resources, habitat and system services were chosen for analysis according to the characteristics of the natural ecological environment, the positioning of the study area in the national key ecological function area and nature reserve, the dependence on natural resources, and the impact of climate change.
For example, habitat quality (H1) and ecosystem service value (H3) reflect the quality of natural resources provided by the ecosystem for the survival and continuation of organisms in the system and are important indicators for measuring the level of ecological quality, while water conservation is a service function of the study area in its role as an important ecological functional area. Therefore, total water conservation (H4) was chosen as one of the evaluation indicators. Specific indicators are selected as shown in Table 3. The degree of influence of each evaluation index for the socio-ecological system is different. Therefore, according to the principle of entropy, the complexity degree of each index and various different states were measured, and the entropy weight method was introduced to weight the index objectively. Because all the indicators had different units, they were standardized to eliminate the influence of different dimensions and directions. After normalization by Eqs. 1 and 2, all values were adjusted to a nominally universal scale.
Negative indicators : where x ij is the value of index j in the ith year, max{x j } is the maximum value of index j, and min{x j } is the minimum value of index j. All selected indicators are grouped separately, taking into account both positive and negative effects. The higher the value of the index is, the better the development of the social ecosystem is, as normalized by Eq. 1. For a negative indicator, a smaller value indicates a better socio-ecological state, which is normalized by Eq. 2. Then, the entropy weight of each index was calculated by the entropy weight method (Table 3). Policy intensity indicates the strength and influence of the policy. In this study, the policy issuing agency and the type of policy were scored. Policy issuing agencies were assigned scores in decreasing order to reflect their strength according to their governmental level: national, provincial, municipal, and district/county levels were assigned strengths of 4, 3, 2, and 1, respectively. For documents of different policy types, according to the principle decreasing effectiveness, scores for laws, administrative regulations, planning documents, and other normative documents were assigned as 4, 3, 2, and 1, respectively. The formula for calculating policy intensity is: where EP is the total score of policy intensity, EI i is the score of the policy issuing agency of Article i policy, and ET i is the score of the policy type of Article i policy. The policy intensity value of a region in a certain year was the sum of the scores of the policy issuing agency and the policy types of n policies in the region in that year (Liu, 2021). Through the collection of ecology, resources, and other related policies in the study area, 623 national, provincial, municipal, and county policies were collected. Through statistical analysis, the policy intensity of the study area since 1990 was scored comprehensively.

RESULTS AND ANALYSIS
Based on the catastrophe theory and adaptive cycle theory, the time changes in the evolution process of each group, social sub-system and ecological subsystem, and social-ecological system were evaluated. The results are shown in Figures 3, 4, and 5.

Evolution of Social Sub-Systems
From 1990 to 2020, social sub-system resilience rose from level 3 to level 4. It then experienced a substantial decline in 1995 before improving again, gradually rising from low resilience (level 1) to high resilience (level 5). After 2018, it experienced a decline from level 5 to level 4. The evolution of the social sub-system is the result of policy, economy, production and population factors.
(1) Policy changes fluctuated greatly during the study period, and policy intensity peaked in 2007 (policy intensity was 197) and 2017 (policy intensity was 478). Since the 1990s, China's grasslands, forests and water resources have been managed by laws and systems. At the national macro policy level, starting with the Grassland Law enacted and implemented in 1985, problems such as inadequate management and grassland degradation due to poor management have been addressed by laws and systems. At the same time, grassland ownership and use rights were stipulated, which have played a very important role in promoting the use, management, protection and construction of grassland. Due to the prominent concentration of water resource problems in the study area, water-related management policies were strongly implemented, with the intensity of these policies peaking in 2007. Under these policies, the area of irrigated land decreased, the number of pumping wells declined, and groundwater extraction stabilized. Simultaneously, an extremely strict water resource management system led to continuous improvement in the basin water resource management system, resulting in an improved water resource situation in the study area. Since December 2016, to address ecological environmental issues in the Qilian Mountains Nature Reserve, ecological protection, restoration and renovation of these mountains have been promoted. Comprehensive management has been carried out at the provincial, municipal and county levels, and measures such as the delineation of ecological protection red lines, the shrinking list of industries with access to key ecological functional areas, and the refinement of management measures to balance grass and livestock have been completed. During the study period, the intensity of these policies peaked in 2017.
(2) Production. The planting and animal husbandry structures were constantly adjusting in the study area. Before the 1990s, Tianzhu mainly cultivated traditional crops such as rapeseed, highland barley and potato. Continuous adjustment of the industrial structure led the proportion of planting area of grain, cash crops and forage material to change from 65:16:19 in 1990 to 22:40:38 in 2016 (Fig. 6), which is more suitable for the region. Further adjustments to the agricultural industrial structure have led to improvements in agriculture   and animal husbandry and their corresponding facilities (Table 4); moreover, improvements have benefited water-saving efforts and intensive, green and large-scale farming practices.
(3) Population and economy. The total population and the number of agricultural populations declined during the study period. The total population decreased from 209,400 in 1990 to 152,000 in 2020, while the agricultural population decreased from 186,700 to 136,200. Due to natural birth rate decreases, ecological migration and labor force transfers to outside areas, the population showed a relatively large decline after 2012 and, after a brief period of stability, a large decrease again in 2017. In key ecological function areas, decreases in the total population and agricultural population greatly reduce human activity impacts; these changes benefit ecological protection and habitat quality improvement efforts. However, for the social system, population decreases impact worker numbers, which may affect economic development and social transformation to some extent. The economy rose in the study area, with stepwise improvements in 1992, 1996, 2005, and 2011. Per capita disposable income in rural areas increased from 466 yuan in 1990 to 8,844 yuan in 2020, an increase of nearly 19 times. The per capita disposable income in urban areas increased from 1,225 yuan in 1990 to 27,736 yuan in 2020, an increase of approximately 22.6 times. The GDP increased from 100.7 million to 488.481 million during the study period.

Evolution of Ecological Sub-Systems
From 1990 to 2020, ecosystem resilience in the study area improved gradually through constant adjustment, from level 3 in 1990 to level 5 in 2020. The evolution of ecological sub-systems is the result of resources, habitat, climate change and other factors.
(1) Resources. The period 1990-1998 was one of continuous growth in the absolute population of Tianzhu. With this increase in population, to meet the demand for food, the planting area was expanded continuously, and grassland resources were lost. The grassland area was reduced continuously, the grassland carbon storage decreased slowly, and the grassland degradation became quite serious (Table 5). This situation paralleled the overall development trend and characteristics of grassland resources in China during this time. Studies have shown that the grassland conditions in China were at their worst in the 1990s (Sun et al., 2014;Robinson et al., 2017). According to the monitoring data released by Tianzhu, throughout the 1990s, the yield of natural grassland was measured in Xixiulong and Songshan. The results showed that the grass yield in a normal year was 3193.5 kg/ha, of which winter and spring yields were 3396 kg/ha. However, the average grass yield from 1989 to 2000 was 2715 kg/ha, of which winter and spring yields were 3129 kg/ha, 15% lower than that in normal years and 32.2% lower than that in the 1950s (Tianzhu County Annals Compilation Committee, 1994).
Since 2000, China has implemented the ''Farmland for Forest'' program, which aims to restore grassland conditions. In addition, the grassland and livestock balance plan, grazing prohibition and rotational grazing plan were implemented. However, because livestock production is the main source of livelihood for most households, these programs hinder grassland restoration to some extent. At the same time, the growth of the market economy promotes the growth of animal husbandry, which in turn further affects grassland resources.
The area of forest resources increased throughout the study period, and the growth of forest areas was the fundamental reason for the increase in carbon storage in Tianzhu. The increases in forest carbon storage and forest area were closely related to the implementation of various policies and ecological engineering. Forest carbon stocks increased from 8 million tons in 1990 to 10.92 million tons in 2020. The area of forestland remained basi- Grassland carbon storage decreased from 16.08 million tons in 1990 to 15.73 million tons in 2020 due to the long-term trend of grassland degradation during the study period.
(2) Habitat quality and ecosystem services. From 1990 to 2020, the habitat quality in the Tianzhu County showed a trend of fluctuating growth, while the habitat degradation index showed a trend of fluctuating decrease, and the overall habitat quality in the Tianzhu County showed an improving trend. The mean value of habitat quality increased from 0.7207 in 1990 to 0.7356 in 2020. The value of ecosystem services increased from 89.26 million yuan in 1990 to 95.41 million yuan in 2020 and that of water conservation increased from 1.008 billion m 3 in 1990 to 1.703 billion m 3 in 2020 (Table 6).

Identification of Evolutionary Stages
The overall evolution of the socio-ecological system is represented as a wave of upgrading changes that represents the adaptive transformation of the coupled system. Evolution characteristics are based on the system components of the sub-systems and the socio-ecological system. According to the characteristics of stages (c, j, x, a) in the adaptive cycle theory, the implementation or adjustment of major ecological and social policies, and the characteristics of social and ecological sub-systems, four stages of socio-ecosystem evolution were identified.
The first adaptive cycle (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998) was defined by rapid development in socio-ecological system resilience, as the system entered the growth and development phase (c) in 1991 and then entered a rapid development phase (c) in 1995, followed by a conservation phase (j). The socio-ecosystem resilience increased rapidly from level 1 to level 4. Production and economic development generally presented an upward trend, and the intensity of policies fluctuated within a certain range. In terms of ecology, resources and habitat quality improved rapidly, and climate change experienced great fluctuations during this period.
The second adaptive cycle (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007) was characterized by a massive reconstruction of the socio-ecological system, as resilience increased from level 3 to level 5. In 1998, system resilience decreased from level 4 to level 3 after a release phase (x) and then, after a period of renewal and reorganization from 1999 to 2002, again entered the development (c) and protection (j) stages (a), and the social-ecosystem resilience reached level 5.
The third adaptive cycle (2005-2018) comprised smaller levels of reconstruction and development of the socio-ecological system, as resilience grew from level 3 to level 4 from 2005 to 2015. This period experienced a small cycle of renewal-developmentprotection and then rose to the highest resiliency level of 5 in 2014. Regarding the social sub-system, the economy and production were relatively stable. After 2010, the resilience of these indicators increased by one level, but that of population showed a large and continuous decline. The change in policy intensity was the greatest during this stage, though after peaking in 2007, it showed a moderate and declining trend. In 2017, policy intensity reached its highest point during the study period. In terms of the ecosystem, resource amounts were relatively stable, but habitat quality changed intensely, and the climate changed steadily. The fourth adaptive cycle process (after 2018) was defined by a rapid decrease in socio-ecological system resilience from level 5 to level 4; after experiencing a large release stage (x), the system entered its current recombination and reconstruction stage (a). In 2018, due to the comprehensive improvement in the ecology of the Qilian Mountains, the social and economic development of Tianzhu was affected to a certain extent. The regional GDP and industrial and agricultural added value fell significantly, and population resilience further decreased. The amount of resources was stable, the habitat improved after a small decline, and climate change was relatively stable. After the adaptation stage of reconstruction and renewal, the evolution of the socio-ecological system should continue to develop in a positive direction.

Analysis of Main Contradictions in the Evolution Stage
In the first adaptive cycle, production was more unstable, and imperfect laws and systems led to problems in resource management. The cultivated land area was constantly expanding to meet the food demand. Grassland area reduction and degradation were obvious due to imperfect systems and laws. The systematic management of water resources also began to appear, but the contradiction between the structural supply for and demand of water resources was prominent due to undeveloped infrastructure and the water conservancy project.
During the second adaptive cycle, the habitat quality degraded substantially. The absolute increase in population size slowed, but grassland degradation further intensified. The main reason for this decline was that livestock production was the main source of livelihood for most families. At the same time, the development of the market economy promoted the continuous growth of the livestock industry. The excessive demand for grassland resulted in deep-seated ecological problems, such as a continuous reduction in downstream water supplies. Grassland degradation at this stage became an important factor restricting the sustainable development of animal husbandry and the key factor leading to the degradation of habitat quality in Tianzhu County.
In the third adaptive cycle, water resource challenges were concentrated and prominent, and grassland degradation still existed. The rapid development of agriculture and animal husbandry had brought about serious environmental problems, such as continuous increases in water consumption, overgrazing, and water and soil erosion. Therefore, adjusting the production structure became the key to continuously improving socio-ecological system resilience, but the problem of grassland degradation still existed due to the increasing pursuit of economic benefits.
In the fourth adaptive cycle, the harmonious development between society and ecology became the most critical issue. Farmers and herdsmen began to receive training on the skills required to transform industries, and the planning and development of ecological industries and ecotourism were carried out in the county. In the short term, top-down centralized ecological governance policies had a substantial impact on social and economic development and promote the transformation of social and economic development. During this stage, coordinated development between society and ecology became a long-term problem that continued during subsequent development.

DISCUSSION
(1) There was a high coupling between the evolution process of the social system and the ecosystem in the study area, and this coupling has continued to increase. During 1993During -1994, the gap between the resilience of the social system and that of the ecosystem was the largest. In subsequent development, this coupling continuously strengthened, and the gap between the two sub-systems narrowed continuously. In 2000In -2002In , 2010In -2011, the resilience of the social system and that of the ecological system developed synchronously.
(2) Population change, policy intensity and its changes, production structure adjustment and climate change played important roles in the evolution process of socio-ecological systems. The population, policy, climate change and production fluctuated greatly during the study period, and their changes were more critical for the overall impact of the socio-ecological system. The key transition points of population resilience occurred in 2012 and, after a brief period of stability, from 2013 to 2017. Population resilience rapidly decreased from level 5 to level 1 by 2020. According to the characteristics of population production, it may be difficult for population resilience to rise again, which may have a positive impact on ecological protection and habitat improvement. However, there may be adverse effects on the social system. The intensity of the policy varied at different times in the study period, directly affecting the resilience of the social system. Compared with the steady growth of the economy, the resilience of production declined during 1994-1995 and 1999-2000, which had a greater impact on fluctuations in the social system. In the ecosystem, the fluctuation range and frequency of climate change were significantly higher than those of resources and habitats, and the impact of climate change on the ecosystem was more obvious.
(3) The social and ecological sub-systems combined to create a complex system in which a balance was achieved between their internal and external contradictions, allowing for the adaptive cycle evo-lution process to be completed at each stage and finally develop in a sustainable direction. While the study area served as an important ecosystem service function, its economic development speed also improved, indicating that ecological protection and economic development had been accomplished to a certain extent. However, the resilience of the social sub-system was always lower than that of the ecosystem, which is closely related to the positioning of the study area as a national key ecological function area and nature reserve; this discrepancy indicates that in future, how to improve the social subsystem should be examined.
(4) The dynamic evolution of the adaptive cycle enabled environmental management institutions to understand the basic state of changing systems; for socio-ecological systems to transform, the fore-loop period of the cycle in the adaptive phase (development and protection), regulatory policy and its effects are essential. In the back-loop period (release and restructuring phase), as the system changes, the critical task comprises maintaining the balance of the system and keeping it within safe operational boundaries (Li, 2018). In future studies, we will study the internal relationship between different key drivers considering resilience transition and further analyze the process and mechanism of ecosystem evolution.

CONCLUSIONS
(1) Based on the resilience theory and methodology, combined with the catastrophe theory and adaptive cycle theory, the study analyzed quantitatively the evolution process and adaptive cycle process of the social sub-system, ecological sub-system and socio-ecological systems in the study area from 1990 to 2020. The results showed that the coordinated development of the socio-ecosystem in the study area has made great progress in recent decades. The coordinated development of the social system and ecosystem had a high degree of coupling and strong connection, although the change in the resilience of the ecosystem was always stronger than that of the social system. This result shows that ecological protection policies and ecological projects have achieved great success in the Tianzhu County over the past decades. On the basis of continued ecological protection, improving the resilience of the social system will become a major task in future.
(2) As a result of policy promotions, the forest area and carbon storage in the study area increased continuously, but the grassland area and quality showed a trend of continuous degradation due to the development of animal husbandry. Therefore, we must strengthen the ecological restoration and management of grasslands in future, emphasize research on the balance of grass and livestock, and protect grassland ecology while supporting the healthy development of animal husbandry.
(3) Enhancing the resilience of the social system, carrying out transformation development and green industry development, and transforming and realizing the value of ecological products are important topics to be discussed in depth regarding the further development of the research area. The simultaneous decrease in the total population and agricultural population will have a certain negative impact on future production and the economy. Agricultural production and climate change are intertwined, and their impacts may increase in future. As a national key ecological functional area and nature reserve, stable national policies play a vital role in social and ecological development. Therefore, in future development, policies should be formulated to promote both social development and ecological protection, which will support the coordinated development of socio-ecological systems.

ACKNOWLEDGMENTS
This study is financially supported by the National Key R&D Program of China (2022YFC3800705).

Conflict of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

OPEN ACCESS
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecom mons.org/licenses/by/4.0/.