Fairness analysis and compensation strategy in the Triangle of Central China driven by water-carbon-ecological footprints

This study proposes water–carbon–ecological footprints to form footprint family indicators for identifying the ecological compensation and regional development equilibrium in the Triangle of Central China (TOCC). The occupation of natural capital stock and flow consumption can be illustrated using a three-dimensional ecological footprint model, and Gini coefficient is integrated into the evaluation framework for fairness measurement from various aspects. Quantificational ecological compensation standards can be given with considering indicators associated with ecological resource conversion efficiency and willingness to pay. Results reveal that ecological and carbon footprints in the TOCC demonstrate rising trends from 2000 to 2015, while its water footprint presents a fluctuating trend. The majority of average Gini coefficients exceed the threshold value of 0.4 under different footprints, thereby indicating poor overall fairness of regional development. Water footprint in Jingmen, Xiangtan, and Yichun show relatively higher compensation expenses, while Yichang, Zhuzhou, and Fuzhou exhibit higher received compensation values compared with other cities. Carbon footprint in Wuhan, Loudi, and Xinyu indicate high compensation expenses due to their overuse of biological resources. Maximum amounts of compensation expense appear in Nanchang and Wuhan from the perspective of ecological footprint. This study can provide a theoretical reference for sustainable development in the TOCC by performing a comparative analysis with Beijing–Tianjin–Hebei urban agglomeration and developed countries.


Introduction
Sustainable development has become a common concern (Chen et al. 2017;Shulla et al. 2020). However, unreasonable utilization of natural resources has led to contradictions between natural capital supply-demand in China (Kaczan et al. 2013;Wang et al. 2019). Some environmental issues associated with climate anomalies, resources exhaustion, and water security have already played serious impacts on China's natural ecosystems and human living environments (Araujo et al. 2019;Chen et al. 2020), which forced the society to reexamine the traditional concept of development which focuses on economic growth. It is thus necessary to evaluate regional ecological sustainability and to determine compensation strategy for achieving regional fair development.
Footprint method is a frontier method in sustainable development assessment, which originated from the term of ecological footprint and was formally introduced to the scientific community in the 1990s (Fang et al. 2014). The widely used ecological footprint is considered as one of the most important achievements in the field of sustainable development. However, the single footprint index cannot meet the needs of sustainable development evaluation (Kinzig et al. 2011), especially with increasingly obvious characteristics of the interaction, complexity and integrity of the earth ecosystem. With response it, the concept of footprint family (FF), such as water, carbon, energy, phosphorus, and nitrogen footprints, has received much attention to integrate ecological conservation and socio-economic development (Song et al. 2018;Dabkienė et al. 2020;Su et al. 2020). Combination of ecological, water and carbon footprints are regarded as a more common model within FF approaches, which comprehensively not only considers the issues of land carrying capacity, water resources carrying capacity, and carbon emissions/sinks, but assesses the impacts of human consumption of biological resources, water resources and greenhouse gas emissions on the earth's environmental system. This combination has extensive, complementary and comprehensive advantages. Numerous effects regarding FF have been made for evaluation of the regional resource-environmental system, such as to quantitatively evaluate the spatial-temporal variations in resource and environment pressure (Chen et al. 2021a,b), to comprehensively understand environmental issues, policy formulation, to assess the trade-offs between different environmental concerns (Vanham et al. 2019), and to jointly evaluate the resource reallocation . Ecological, water, and carbon footprints embody the basic connotation of resource conservation and environmental friendliness required by the construction of ecological civilization, which can be seen as a supplement to the sustainability debate (Galli et al. 2012).
Certain differences in economic development and resource utilization in various regions result in unfairness in the use of ecological resources and environmental protection between regions to some degree. Ecological compensation is an important way to promote regional coordinated development in response to such differences (Wang et al. 2020;Yang et al. 2020a). The regional ecological resources can serve itself and others due to the positive externality of ecological services. If a region can provide more ecological services than it consumes (i.e., ecological carrying capacity greater than ecological footprint), it can receive ecological compensation. It should be specially mentioned that ecological compensation is an equivalent term to payment for ecosystem services (PES) (Pan et al. 2017;Fan and Chen 2019;Yu and Xu 2012;Fu et al. 2021), where PES is a conservation scheme that uses economic means to incentivize service providers to ensure the sustainable supply of ecosystem services. For example, Yu et al. (2020) focused on the trends of terminology, location, types of PES. They indicated that there were various terminologies, and interestingly "ecological compensation" was mostly used especially in Chinese research community. They also adopt ecological compensation as an alternative term for PES studies because of its inclusiveness and representation of empirical practices. Actually, there are efforts undertaken for ecological compensation (Zhao et al. 2021;Fan and Chen 2019;Jiang et al. 2019;Yang et al. 2020b;Koh et al. 2017;Sonter et al. 2020;Vaissière and Meinard 2021;Reid et al. 2015). For example, Fan and Chen (2019) established a comprehensive evaluation framework for identifying the spatial characteristics of land uses and ecological compensation from 2000 to 2015 in Sichuan Province, China. Koh et al. (2017) took the ecological compensation as a policy instrument in Sweden, results showed that ecological compensation guaranteed the sustainable and fair benefits of ecosystem services. Vaissière et al (2021) used ecological compensation policy to response the negative impacts on biodiversity. Results revealed that ecological compensation is key to stop the erosion of biodiversity in French. It can be seen that ecological compensation played an increasingly important role in biodiversity protection, ecological environment restoration, and regional coordinated development. But there is lack of institutional standardization and compensation standards. It is thus necessary to propose a novel method for determining compensation standard (He et al. 2018), At present, the main methods to determine the ecological compensation standard are the opportunity cost method, willingness to pay method, ecosystem service value method, and ecological footprint method (Yang et al. 2020a). However, there are some difficulties in the process of implementing ecological compensation, especially in China with a top-down administrative management system (Wang and Wall 2007). A comprehensive ecological compensation framework based on FF analysis is a practical response to this difficulty. FF aims to track natural resource occupancy, greenhouse gas emissions, and water resource consumption generated by human activities from the biosphere, atmosphere, and hydrosphere (Galli et al. 2013). Compared with traditional compensation methods, the FF method can effectively avoid the influence of subjective factors and objectively determine the subject and object of ecological compensation and their specific amounts.
Despite the abovementioned efforts, many questions have not yet been resolved. For example, the traditional ecological footprint model is difficult to reflect the importance of maintaining the stock of natural capital to maintain the stability of ecosystems. In the ecological footprint accounting, it is necessary to develop an improved method to distinguish the difference between the stock and flow of natural capital. The analysis of the consumption of natural capital stock can include a more detailed description of the ecological situation. Aiming at the sustainable utilization of natural capital in urban agglomeration, the research has gradually changed to threedimensional ecological footprint model, which is based on footprint size with spatial attribute and footprint depth with temporal attribute. Moreover, considerable differences exist in the spatial distribution of China's biological resource consumption within different land-use types, and the resulting ecological environment problems remain unresolved (Xue et al. 2014). Judging the fairness between FF and the regional economic development with consideration for the asymmetry of economic growth is necessary (White 2007). The Gini coefficient is a widely used quantitative indicator of balances based on the Lorentz curve that can evaluate the regional fair development (Figueras and Duro 2015;Shu and Xiong 2018;Teng et al. 2011;Chen et al. 2017;Dai et al. 2018;Yang and Fan 2019). For example, Chen et al. (2016) used Gini and deviation indexes for evaluating differences in the per capita consumption of fossil energy across 30 Chinese provinces from 1997 to 2013. Results revealed that the Gini rate of Chinese inter-provincial fossil energy consumption was less than 0.3. Shu and Xiong (2018) constructed a non-grouped Gini index for identifying regionally balanced development of economy and environment in China.
The objective of this study targets to develop a comprehensive ecological compensation model driven by water-carbonecological footprints. Gini coefficient will be used for measuring fairness of the regional development. Some critical questions will be answered, such as footprints deficit or surplus, regional development equilibrium, and compensation strategies. A real-world case study in the Triangle of Central China (TOCC) will be performed to validate the applicability of the developed model. The TOCC is a super large national level urban agglomeration, which occupies an important position in China's regional development pattern. However, serious environmental pollution has hindered its sustainable development. Findings will provide a robust decision-making reference for the sustainable development in the TOCC.

Problem's statement
The TOCC (108°21′E~118°28′E and 20°09′N~33°20′N) consists of Hubei, Hunan, and Jiangxi Provinces, which has an area of 32.61×10 4 km 2 with a center of Wuhan. The TOCC mostly includes the four city groups, i.e., Poyang Lake City Group (PLCG), Chang-Zhu-Tan City Group (CZTCG), Wuhan Metropolitan Area (WMA), and Xiang-Jing-Yi City Group (XJYCG) (Figure 1). It should be specially mentioned that WMA and XJYCG are integrated as one urban agglomeration named as WMA& XJYCG. Table 1 gives the specific city division. The TOCC is an important part of the Yangtze River Economic Belt, China, and has four biodiversity protection ecological function zones with important ecological service functions. Figure 2 shows the spatial distribution of per unit of GDP and population in the TOCC across Yangtze River Economic Belt. In 2017, its total population reached 125 million with a GDP of 7.90 trillion RMB ¥ (ranking fifth in China's urban agglomeration). The TOCC has created 9.6% of the total economic output on basis of 3.4% and 9.0% of China's land area and population, respectively. High energy consumption and serious pollution caused by the large number of industries in the TOCC results in high water pollution emissions. Insufficient capital investment and utilization efficiency have led to low pollution control efficiency due to the lack of water pollution control measures. Average annual control efficiency of the TOCC is 0.53, which is constantly lower than the overall level of the Yangtze River Economic Belt that results in the deterioration of water quality and eutrophication of lakes and reservoirs. Long-term water pollution problems have led to changes in the structure of biological communities, with the total number of fish reduced by 38.1%, and decreased biodiversity. Generally, the pressure of ecological environment and problems of ecological security in the TOCC during economic development are more prominent.

Footprint family indicators
Water footprint model The general water footprint can be divided into the direct and indirect water footprints with a unit of m 3 , but it cannot reveal the relationship between the consumption of natural resources and the ecosystems' carrying capacity. In this study, water footprint refers to the area of water resources needed for human life, production and natural environment. It can be divided into domestic, production and ecological water footprints according to the water-use characteristics (Chen et al. 2021a).
where TWF represents water footprint (hm 2 ); TWC denotes water carrying capacity (hm 2 ); M denotes the population; WF denotes per capita water footprint (hm 2 /cap); WC denotes per capita water carrying capacity (hm 2 /cap); when WD>0, it represents a water deficit, otherwise, a water surplus; b w represents the global equilibrium factor of water resources; h w represents the global average water production (m 3 /hm 2 ); B represents per capita water consumption (m 3 /cap); c represents water resources production factor that is a ratio of the average output of regional water resources to the average output of global water resources; D represents regional water resources availability; Previous studies have shown that 60% of water resources should be reserved for the maintenance of ecological environment, thus a = 0.6. Carbon footprint model: the coefficient method is used for determining the amounts of carbon absorption (CA) and carbon emissions (CE), as follows (Chen et al. 2021a): where AR i denotes the area of ith land type (hm 2 ); s i denotes the carbon absorption coefficient of ith land type (tC/hm 2 ); o j denotes the amount of energy consumption (t); g j denotes standard coal conversion coefficient; q j denotes the emission coefficient of ith energy (tC/t); NCE denotes the amount of net carbon emissions (tC). In addition, the carbon carrying  capacity (CC) and carbon footprint (CF) can be calculated as follows (Chen et al. 2021a): where NSP f , NSP g , and NSP a denote the carbon sequestration capacity of forest, grassland, and cultivated land (tC/ hm 2 ), respectively; R f , R g and R a denote the proportions of carbon sequestration in terms of forest, grassland, and cultivated land, respectively; N is the total population; when CD>0, it denotes a carbon deficit, otherwise, a carbon surplus.
Ecological footprint model: the traditional ecological footprint (EF) model is presented as Eqs. (10)-(13), where EC denotes the total regional ecological capacity that refers to the total area of bio productive land for providing human survival and development (hm 2 ); i refers to a certain natural capital category; n refers to the total number of natural capital categories; j refers to different land-use type; this study uses ArcGIS with spatial resolution of 1km×1km to extract land classification that is divided into six categories (i.e., cultivated land, grassland, forest land, water land, construction land, and fossil fuel land); r j refers to equilibrium factor; y j refers to yield factor; p j and c j refers to average production capacity and consumption level of a certain natural capital category, respectively; a i refers to land area converted by a certain natural capital category (hm 2 ); A j denotes the actual land use area (hm 2 ); N denotes the population; ef e and ec e denote the ecological footprint and carrying capacity (hm 2 /cap), respectively; when ED > 0, it represents an ecological deficit, otherwise, an ecological surplus.
The traditional ecological footprint model focuses on the measurement of flow capital, while ignoring stock capital that plays a critical role in regional ecosystem balance. In response to such concern, the three-dimensional ecological footprint (EF 3D ) model introduces two indexes (footprint depth and size) to represent the extent to which humans consume natural capital stock and occupy natural capital flows. The formula is as follows (Chen et al. 2021a): where EF depth is ecological footprint depth that refers to the multiples of land area theoretically required to maintain the existing level of resource consumption in the region. EF depth reflects the consumption of natural capital stock that exceeds the ecological carrying capacity. EF size is ecological footprint size that refers to the area occupying biologically productive land within the region's carrying capacity. It reflects the level of natural capital flow occupied by human beings (Fang 2015).
where ED is the total ecological deficit; EC is the total ecological carrying capacity. When EF depth equals to 1, it means that flow capital can just meet the demand for resource consumption; when EF depth is greater than 1, it means that flow capital cannot meet consumption demand, and stock capital must be consumed. Notedly, the three-dimensional footprint in this study is partially simplified based on some assumptions: (a) ignoring the supply and demand of natural capital of different types of biological production land, and (b) without consideration of the transfer effect of cross-regional trade on natural capital.
For details of regional fairness analysis and ecological compensation model, please refer to the Appendix.

Data sources
In this study, the social-economic data are mostly obtained from Hunan, Jiangxi and Hubei Provincial Statistical Yearbook from 2000 to 2015. The data associated with biological resource consumption and energy consumption are collected for calculating different footprints. Among of them, biological resources are divided into agricultural, animal, forest and aquatic products. Energy resources are divided into industrial consumption of energy and electricity (Table 2). For example, Table 3 presents some average socio-economic data in different urban agglomerations. The annual average precipitation ranges from 1,134 to 2,239 mm, and the altitude is between 20 and 3,105 m. Cultivated land, forest land, and water have a major position of the regional land uses (Figure 3).

Results analysis
Water-carbon-ecological footprints analysis The general ecological and carbon footprints in urban agglomerations keep rising during the periods from 2000 to 2015 ( Figure 4). The annual growth rates of carbon footprint in the WMA&XJYCG and PLCG reach 6.42% and 9.86%, respectively; that of ecological footprint is 5.0%. In comparison, water footprint in urban agglomerations shows a fluctuating trend. Its average values range from 0.606 hm 2 /cap (Yichang, W5) to 1.435 hm 2 /cap (Jingmen, W6), from 0.632 hm 2 /cap (Hengyang, C7) to 1.120 hm 2 /cap (Xiangtan, C3), and from 0.628 hm 2 /cap (Jingdezhen, P3) to 1.283 hm 2 /cap (Yingtan, P4) in the WMA&XJYCG, CZTCG, and PLCG, respectively ( Figure 5). Most of water footprint is contributed by domestic and production water usage. The corresponding water carrying capacity ranges from 0.446 hm 2 /cap (Xiaogan, W8) to 4.949 hm 2 /cap (Xianning, W10), from 2.286 hm 2 /cap (Xiangtan, C3) to 5.530 hm 2 /cap (Zhuzhou, C2), and from 4.308 hm 2 /cap (Jingdezhen, P3) to 13.151 hm 2 /cap (Fuzhou, P9). As shown in Figure 6, high water carrying capacity mostly appears in the southeast of TOCC due to its abundant rainfall, especially in Fuzhou and Yingtan of the PLCG. By contrary, low water carrying capacity exists in the northwest of TOCC due to its highly intensity of population and economics, especially in Wuhan and Xiaogan of the WMA&XJYCG. Figure 7 illustrates the changes of carbon footprint. The general carbon performance is negative because of carbon deficit greater than zero, especially in WMA&XJYCG and PLCG, with average values of 0.4248 and 0.4624 hm 2 /cap, respectively. Specifically, average carbon footprints in WMA&XJYCG, CZTCG, and PLCG amount to 0.4411, 0.4338, and 0.4919 hm 2 /cap, respectively. Ezhou, Loudi, and Xinyu demonstrate high carbon footprints with averages of 1.181, 1.330, and 1.431 hm 2 /cap, respectively, mainly due to large consumption of raw coal, followed by power generation in these areas. For example, the carbon emission in Wuhan is as high as 3.69×10 8 tC from 2000 to 2015, of which   coal-fired based carbon emission is 1.36×10 8 tC, accounting for 36.86% of the total. Average carbon carrying capacities in WMA&XJYCG, CZTCG, and PLCG reach 0.0163, 0.0189, and 0.0337 hm 2 /cap, respectively. Approximately 94.2% of the total carbon absorption is contributed by forest land, followed by water land (4.02%). Ji'an and Yichang demonstrate a high contribution to the total carbon carrying capacity due to their wide areas of forest land and crops, with values of 0.2736×10 6 and 0.2684×10 6 hm 2 , respectively. High carbon carrying capacities can be observed in the southeast TOCC, while low carbon carrying capacities are mainly distributed in the northeast and southwest areas (Figure 8). Rapid economic development is normally accompanied by huge carbon footprints (e.g., Wuhan). On the contrary, the urbanization process and coal-consumption level in Xiantao and Tianmen is relatively slow, coal consumption is slow, which results in their low-level carbon footprints.
The average ecological footprint in the WMA&XJYCG increases from 2.435 in 2000 to 5.465 hm 2 /cap in 2015 with an annual growth rate of 5.55%; that in the CZTCG raises from 2.231 in 2000 to 4.126 hm 2 /cap in 2015 with an annual growth rate of 4.52%; that in the PLCG increases from 1.416 in 2000 to 3.245 hm 2 /cap in 2015 with an annual growth rate of 5.34% ( Figure 9). The per capita ecological footprint is generally contributed by grassland and cultivated land, which account for approximately 82.58% of the total, followed by fossil fuel land with a share of 15.71% of the total. By comparison, ecological carrying capacities reach 0.3301, 0.3009, and 0.4229 hm 2 /cap in WMA&XJYCG, CZTCG, and PLCG, with ecological deficits of 3.6098, 3.2546, and 2.1334 hm 2 /cap, respectively. Except for water and construction lands, ecological carrying capacities in other land types remarkably decrease from 2000 to 2015. Compared with 2000, ecological carrying capacities of cultivated and forest lands decrease by 0.31 hm 2 /cap in 2015. As shown in Figure 10, the ecological footprint in the TOCC increases, especially in Ezhou, due to its high-intensity population and energy consumption. Overall, the kernel distribution maps of water-carbon-ecological footprints in different urban agglomerations present unimodal patterns, implying that the water-carbon-ecological performances in TOCC are not reach the extent of becoming divided (Figure 11). The ecological, water, and carbon footprints in the WMA&XJYCG mostly lie in the intervals of [2,4]

Regional fairness analysis
The Gini coefficient can reflect fairness of water, carbon and ecological footprints in spatial distribution with respect to different influencing factors (i.e., water resources, population, GDP). The following criteria are used to classify the matching degree: G< 0.2 for absolute match; 0.2 ≤G < 0.3 for comparative match; 0.3 ≤ G < 0.4 for relative match; 0.4 ≤ G < 0.5 for general mismatch; 0.5 ≤ G < 0.6 for comparative mismatch, and G ≥ 0.6 for serious mismatch (Wu et al. 2017). From the perspective of water footprint (Figure 12a and Figure 13b), its population Gini coefficient has a slight change over the  periods from 2000 to 2015, with an annual average of 0.373 basically below the threshold value. Its average GDP Gini coefficient reaches 0.450 at a general mismatch state, suggesting a poor match relation between water resource footprint and GDP growth. The average value (0.442) of water resources Gini coefficient presents a general mismatch state, especially in 2012 with a value of 0.553 that significantly exceeds the threshold value and is at comparative mismatch state. The above variations mostly arise from significant changes in precipitation yet slight changes in water footprint in these years, resulting in the mismatch between water resources and water footprint. In general, the relationship among population distribution, economic development, water resources, and water footprint are not harmonious due to their comprehensive Gini coefficient reaching 0.421. In terms of carbon footprint (Figure 12b and Figure   (c) (d) Figure 10. Spatial-temporal dynamic variations of ecological footprint in the TOCC. Figure 11. The kernel distribution map of water-carbon-ecological footprints in different urban agglomerations. 0.516), respectively. The average value of comprehensive Gini coefficient is 0.456, which is in the state of general mismatch. It is worth noting that both GDP and water resources Gini coefficients are above the warning line, while population Gini coefficient is higher than the threshold value only after 2011.

Compensation strategies driven by family footprint
Results indicate significant differences in ecological service consumption values (Figure 14). The CZTCG demonstrates high service consumption values of water resources in Changsha and Changde that account for approximately 37.82% of the total. Nearly 42.98% of the total service consumption value of water resources in the PLCG come from Yichun and Nanchang. In terms of carbon concern, its high service consumption values in the WMA&XJYCG exist in Compensation strategies in different urban agglomerations from the perspective of water-carbon-ecological footprints are illustrated in Figure 15. From a water-footprint aspect, Wuhan and Jingmen in the WMA&XJYCG pay high compensation expenses with averages of −11,093.30×10 4 and −20,339.23×10 4 RMB ¥, respectively, while Yichang receives a high amount of compensation with a value of 20,339.23×10 4 RMB ¥. Ezhou and Tianmen demonstrate high compensation expenses in terms of per capita level of −58.78 and −23.49 RMB ¥, respectively, while Xianning

Model discussion
Evaluation of water-carbon-ecological footprint family in urban agglomeration systems is far more complex than usual ones. This mainly results from the requirement of considering multiple economic, social, and environmental factors, as well as the complicated interactions among various cities within an urban agglomeration. There are two major reasons driving extended studies regarding footprint family for urban agglomerations. Firstly, the traditional ecological model cannot effectively reflect ecological sustainability. A three-dimensional ecological footprint model is thus proposed with essential connotations of EF depth and EF size . Although the three-dimensional ecological footprint can provide further comprehensive management strategies for decision makers, it usually enters into trouble when other environmental concerns are considered. The footprint induced by economic development and human activities in the TOCC is not limited by only one. This study thus integrates water, carbon, and ecological footprints into a comprehensive compensation evaluation framework for illustrating the change track of environmental impact in the TOCC. Secondly, most of the previous studies can hardly set reasonable standards and measure regional fairness caused by the differences in biological resources consumption. With these concerns, this study proposes a footprint family-based compensation approach, which can effectively avoid the influence of subjective factors. The Gini coefficient is then merged into the footprint family evaluation framework and is used to identify the fairness among different cities and urban agglomerations.
In fact, more footprints such as nitrogen, phosphorus, energy might be considered into this study, because they can reflect the potential resource consumption or environmental deterioration during the development of urban agglomeration. Future research needs to use a highly integrated model to more accurately assess the regional environmental pressure. Additionally, the developed model did not consider multiple  uncertainties normally expressed as interval, fuzzy, and stochastic parameters into the general framework. The introduction of these uncertainties probably leads to significant differences in water-carbon-ecological footprints and compensation amounts. The last one is to apply reliability-resilience-vulnerability (RRV) indexes to the footprint family problems (Lu et al. 2019), which is beneficial to comprehensively evaluate regional environment safety and to establish management policies for joint long-term control of RRV at the regional scale (Asefa et al. 2014).

Comparisons with the previous studies
A comprehensive comparison among different urban agglomerations is illustrated by consulting relevant literature and practices. Corresponding indicators of the WMA&XJYCG, CZTCG, PLCG, and Beijing-Tianjin-Hebei urban agglomeration (BTHUA) are transformed into the interval [0,1] on the basis of the per capita three-dimensional ecological footprint, footprint depth, and footprint size of TOCC ( Figure 16). Averages of the footprint depth and per capita threedimensional ecological footprint of the WMA&XJYCG and CZTCG are larger than those of the TOCC. In contrast, BTHUA's footprint depth is smaller than that of TOCC. High consumption of natural capital stock can be observed in the TOCC, which is mainly due to the growth of ecological footprint caused by the agriculture/animal husbandry development and fossil energy consumption. The minimum and maximum per capita ecological footprint can be found in the BTHUA and WMA&XJYCG, respectively. These situations show the TOCC with a poorer sustainable development situation compared with the BTHUA. Moreover, the other comparisons with some countries are presented in Figure 17. The per capita ecological footprint of 10 developed countries during the study period basically shows a downward trend. The per capita ecological footprint of the United States decreases from 10.25 in 2000 to 8.17 hm 2 /cap in 2015. During the periods from 2000 to 2010, the per capita ecological footprint of the WMA&XJYCG, PLCG, and CZTCG is smaller than that of the United States and Canada. However, the per capita ecological footprint in the TOCC is gradually larger than that of the United States and Canada by 2010 and exhibits an increasing trend. The per capita ecological footprint of the CZTCG exceeds the average value of the developed countries (6.53 hm 2 / cap) in 2004, while the per capita ecological footprint of the WMA&XJYCG and PLCG exceeds the average value of the developed countries (5.79 hm 2 /cap) in 2009. These findings indicate that the area of bioproductive land required for per capita resource consumption in the TOCC is gradually increasing and the regional sustainability presents a deviation state.

Policies implication
In terms of water resources utilization, it should focus on agricultural water use and strengthen construction of    agricultural water-saving project. To improve the carbon deficit in the TOCC, it needs to adjust the industrial structure and energy structure, and eliminate high-pollution and high energy-consumption industries. It is also necessary to improve forest coverage and to strengthen forest protection. In the case of widespread ecological deficit, it is required to find a way to promote sustainable development, such as improving the liquidity of natural capital (increase footprint size) and reducing the consumption of stock capital (decrease footprint depth). The former is determined by local natural resource endowment and is not easy to change. The latter is closely related to the deficit of energy consumption footprint, which can decrease the total consumption of fossil energy and optimize the structure of energy consumption to reduce footprint depth and the depletion of natural capital stock. Moreover, the cities that pay for ecological compensation are mostly concentrated in the central of urban agglomerations with rapid economic development and high GDP. The number of ecological resources consumed by economic development exceeds the regional ecological carrying capacity, and the ecological resources of other regions need to be occupied and compensated. Due to geographic location and terrain constraints, the compensated area has slowed economic development and strong ecological carrying capacity, which can provide ecological services to other cities and obtain corresponding ecological compensation. Future developments must adhere to green requirements that highlight three aspects. First, the TOCC should follow the harmony between humans and nature, cherish superior natural environment, and act in strict accordance with ecological laws. Second, the TOCC should strengthen the relation between development and protection and protect the ecological environment as an important premise in the rapid economic development. Finally, it should clarify the relation between ecological civilization construction and development mode and aggressively promote low-carbon energy technology and comprehensively improve total factor productivity.

Conclusions
Some conclusions can be summarized as follows: (a) the TOCC demonstrates a surplus in the water footprint but its ecological and carbon footprints continue to rise (annual growth rates of 6.2% and 5.0%) with evident footprint deficits. Most of ecological and carbon footprints are from grassland and cultivated land. (b) Average Gini coefficients of population, GDP, and water resources reach 0.373, 0.450, and 0.442, respectively, from the perspective of water footprint. Carbon footprint presents an increasing trend in Gini coefficients of population, GDP, and water resources from 2000 to 2015; notably, Gini coefficients of population and water resources exceed the threshold value. Slightly varying Gini coefficients of population, GDP, and water resources with averages of 0.395, 0.457, and 0.516, respectively, from the perspective of ecological footprint, which should be paid more attention because most coefficients are higher than 0.40. (c) Cities that pay ecological compensation are typically concentrated in the center of each urban agglomeration with high GDP.

Willingness to pay indicator
where W R denotes R city's the indicator of willingness to pay; l R denotes R city's development stage coefficient; p denotes a province's per capita GDP; p R represents R city's per capita GDP; ln R denotes R city's per capita income; ln denotes a province's per capita income; A R is the urban per capita disposable income of R city; m R is the urban population of R city; B R is rural per capita net income of R city; n R is the rural population of R city.
Ecological service supply coefficient where β R is the supply coefficient of ecological services of R city; EC R is water/carbon/ecological carrying capacity of R city; V R is the amount of money arising from supply of ecological services of R city; M is the total amount of ecological services in each province, which is represented by regional investment in pollution control in this study.

Ecological service consumption coefficient
where Rec R is the comprehensive correction coefficient of R city; U R is average ecological footprint of ten thousand-yuan GDP of R city; W R is average willingness to pay in R city; U is average ecological footprint of ten thousand-yuan GDP in a province; W is average willingness to pay in a province; α R is ecological service consumption coefficient of R city; F R is ecological service consumption values of R city; EF R is ecological footprint of R city; M is the total amount of ecological services in each province.

Amount of ecological compensation
where X R is the amount of received ecological compensation of R city. A positive value of X means that the city's ecological compensation amount is a net inflow; otherwise, it is a net outflow.