How does the green efficiency of urban land use evolve in the urban agglomeration of China's middle Yangtze river?

The urban agglomeration in the middle reaches of the Yangtze river (MYRUA) is the second largest national urban agglomeration in China, with an excellent ecological foundation. As the construction of urban agglomerations accelerated towards both extension and connotation, land prices of various cities rose, and the contradiction in the demand and utilization of land resources became acute. Therefore, exploring the spatiotemporal evolution and driving mechanisms of land green utilization efficiency within MYRUA had positive significance. This article focused on 28 prefecture‐level cities in the MYRUA and measured their urban land green use efficiency (ULGUE) from 2006 to 2018 using the super slack‐based measure model with undesirable output (Super‐SBM‐U). Then, the improved Fixed Malmquist–Luenberger index was developed to improve the accuracy of dynamic efficiency analysis. Further, the internal and external driving factors were explored using the index decomposition and the spatial error model with fixed effects. The results showed that the land use mode of the urban agglomeration in the middle reaches of the Yangtze River was relatively green, but efficiency gaps still existed. Although the green total factor productivity of the whole cluster has improved by 3.30% a year on average, the excessive pursuit of technological innovation has led to the stagnation of urban land management in various cities, and there are numerous inherent contradictions in urban development. But at the same time, feasible economic development directions, industrial upgrading, high‐quality employees, land marketization, and positive environmental behaviors are conducive to the green and sustainable development of ULGUE. This article depicted the internal land use characteristics of the MYRUA from the perspective of spatiotemporal evolution and efficiency, providing a theoretical reference for the green development and coordinated management of urban agglomeration land.


Recommendations for Resource Managers
• Natural features and socioeconomic heterogeneity lead to a significant discrepancy in the green efficiency of urban land use.Thus, they should be regarded as the scientific basis of policy design for an entire urban agglomeration.• Improving regional connections is vital for enhancing the exemplary and radiation effects of core cities on green land use.
• While the importance of technology has been stressed, the management of related systems should keep pace to provide sustainable forces for the development of urban agglomerations.

| INTRODUCTION
In the process of continuous urbanization, the Chinese government has drawn "three redlines," that is, an ecological redline, a permanent basic farmland redline and an urban development boundary on urban spatial planning, further restricting the supply of urban construction land (Jiang, 2021;Zhang et al., 2019).China used to rely on large-scale land input to stimulate economic growth, which caused a series of problems, such as cultivated land destruction, idle development projects, and environmental pollution, which restricted the high-quality development of cities (Liu, Jin, et al., 2020;Yang et al., 2020).In 2019, the "Proposal on Establishing the Spatial Planning System and Implementation Supervision of Land Use" issued by the State Council of China has emphasized that the concept of green development must be integrated into the whole procedure to improve the efficiency of urban land use (ULUE) and make it become a priority.It is imperative to reinforce the "Five Adjustment Initiatives" by controlling the total amount of construction land, optimizing the new development projects, animating the stock, improving the circulation efficiency, and enhancing the quality of construction land to tackle the challenges posed and promote the transition of land use mode.(Shen et al., 2021;Yu et al., 2019).
The land is the primary carrier of social production, people's lives, and ecological space, where ULUE reflects the flow speed and value realization of total factors (Xu et al., 2019).Both parametric and non-parametric methods have been proposed to evaluate the efficiency of this type of land use.As a parameterized method for evaluating land use efficiency, some scholars established an evaluation index system and then adopted a multilevel comprehensive evaluation model for quantification, while others constructed specific production functions, such as the Cobb-Douglas production and Stochastic frontier functions (Liu, Xiao, et al., 2020).Data envelopment analysis (DEA), as a nonparametric method, did not require setting production functions in advance and adopted the optimal principle to configure the objective weight of input and output for each evaluation object.As such, it became popular in terms of efficiency measurement (Dong & Wu, 2020;Yu et al., 2019).First, scholars mainly applied traditional DEA models, such as CCR, BCC and directional distance function (DDF), to their analyses (Liu et al., 2018;Zhou et al., 2021).Combined with green development, scholars have performed in-depth studies and paid more attention to the negative output brought about by production (An et al., 2021), ultimately producing ULGUE.Meanwhile, diverse improved models have been proposed.Hailu and Veeman (2001) minimized the undesirable output as input variables.Chen et al. (2020) extended the slack-based measure (SBM) model with waste and emissions of industrial production, while Yang et al. (2021) considered carbon emissions as an undesirable index.Furthermore, to reveal the dynamics of efficiency in different periods, scholars have successively introduced the Malmquist index (MI) and Malmquist-Luenberger index (MLI) (Jiang, 2021;Li & Chen, 2021).Some scholars have also proposed a slacks-based global DEA for Chinese urban green economic growth and a global Malmquist index (Tian & Feng, 2021).These indexes were mainly calculated by DDF and SBM models, demonstrating the change of GTFP and its source through index decomposition (Sun et al., 2020;Xie et al., 2019).With the help of the DEA models and MLI, the spatiotemporal evolution of efficiency could be demonstrated intuitively (Han & Zhang, 2020;Li & Chen, 2021;Ren et al., 2019).The exploratory data analysis (Zhao et al., 2015), the Tobit model (Xie et al., 2019) and the spatial regression model (He et al., 2020;Tian & Feng, 2023) were introduced to analyze the factors that contribute to it, where some scholars found that ULGUEs were spatially related at different scales, such as the national, the provincial and urban agglomeration (Yang et al., 2022).Chen et al. (2022) adopted the geographically weighted regression model to explore the spatial effect between economic development, land transaction, popular agglomeration, the construction of transport facilities and ULGUE at the national level.Taking marketization, industrial structure and fiscal expenditure as the basic factors, Song et al. (2022) constructed the panel regression analysis model to analyze the relationship between economic transition and ULGUE in the different reaches of the Yellow River Basin.Under carbon emission constraints, some researchers used the Tobit model for analysis and found that the disposal income, level of urbanization and technological development have an impact on land use efficiency.Tian and Feng (2023) applied the global Malmquist index and spatial econometric model for the effect and its mechanisms of green technological innovation (GTN) on resource curse Green, and found that technological innovation (GTN) would help improve resource utilization efficiency and environmentally-friendly products.A preliminary review of the aforementioned studies showed that the areas with a higher level of ULGUE tended to converge in the space, and urbanization, industrial structure, economic level, and marketization were considered as the basic influencing factors of ULGUE (Ge et al., 2021).
The available research still had some shortcomings.First, the output dimension was limited to the socioeconomic perspective and environmental constraints while ignoring residents' well-being, which neglected the people-oriented concept.Second, most research focuses on urban agglomerations at the macro level, taking it as a whole.The discussion of inner cities and subgroups is slightly inadequate (Li et al., 2021).Third, the MLI was widely used to analyze the change of GTFP from the internal perspective but lacked consistency in time series (An et al., 2021;Long et al., 2020) and the consideration of external factors.However, the role of geographic spatial effect and environmental governance in discussing the promotion path of ULGUE is significant because of the regional heterogeneity and its influence on a policy design oriented toward coordinated development (Ge et al., 2021).Therefore, further effort was needed to evaluate the ULGUE and comprehensively reveal its dynamics.
The significance of MYRUA in promoting China's high-quality development has again been brought into the spotlight.China's "14th 5 Year Plan" and the "Outline of Long-term Objectives in 2035" suggest that the government should promote the coordinated development of MYRUA through the construction of Wuhan and the Changsha-Zhuzhou-Xiangtan metropolitan areas, ultimately building it into a national growth pole.Based on this, we first apply the Super-SBM-U model considering economic prosperity, people's well-being, and environmental sensitivity to evaluate ULGUE in the MYRUA from 2006 to 2018.Second, we propose an improved MLI to analyze the spatiotemporal evolution of ULGUE and explore its internal driving factors through index decomposition.The results examine the management and technical conditions of study areas and provide a better understanding of ULGUE.Additionally, we construct the spatial regression model to reveal the external driving factors from the perspective of economy, industrial structure, transportation, quality of employees, marketization, and governments' awareness.This contributes to formulating targeted policies and guiding synergetic and sustainable development for urban agglomerations.

| Study area
The MYRUA covers 326,100 km 2 and ranks second among approved national urban agglomerations in China.Located in the centre of China, it connects not only the East and the West but also the South and the North, becoming the hub connecting three other national city groups: the Yangtze River Delta Urban Agglomerations, Pearl River Delta Urban Agglomerations, and Chengdu-Chongqing Urban Agglomerations.This region has advanced natural conditions with the longest river, that is, the Yangtze River, running through it.
Since 2006, the MYRUA with Wuhan, Changsha, and Nanchang as the core has gradually taken shape, but the leading role of Wuhan and Changsha is more evident than that of Nanchang (Zheng & Qingyun, 2021).In 2015, China's Development and Reform Commission officially issued the "Urban Agglomeration Development Plan" for the MYRUA (hereafter referred to as "Plan") to guide the region's integration.It defines 28 prefecture-level cities across the Hunan, Hubei, and Jiangxi provinces as the scope of the MYRUA, covering six members of the Wuhan Metropolitan area (WHMA), the whole Changsha-Zhuzhou-Xiangtan City Group (CZT), and the entire Poyang Lake City Group (PLCC), as shown in Figure 1.

| The green efficiency of urban land use (ULGUE)
With the concept of green development gradually penetrating land use research, the field has reached a consensus that ULGUE plays a vital role in realizing the sustainable development of cities (Chen et al., 2016;Tang et al., 2021).There are many similar ideas in previous research, such as "intensive," "ecological," and "low carbon," though they have an obvious bias on the economy and environment (Fu et al., 2020).Coordinating economic development with ecological protection and human well-being has been further considered in ULGUE.Specifically, in this research, ULGUE refers to a gradual reduction in ecological impact and resource intensity over entire land production cycles, realizing the goals of economic prosperity, social well-being, and environmental friendliness.

| The green total factor productivity of urban land use system
The concept of total factor productivity (TFP) was first put forward by Timberger.The TFP measures the efficiency of transforming total input into output in socioeconomic activities while simultaneously taking all input factors (including labor, capital, land, and so on) into consideration, which is a comprehensive embodiment of regional development quality and management.The improvement of TFP refers to the output growth brought by all other factors except the effects of capital, labor, and other input factors.It is usually attributed to the progress of science and technology and the improvement of technical efficiency.
Urban land use is generally a long-term and continuous activity.In this process, the level of production technology is constantly developing.The GTFP described in this paper is a concept of growth rate.Based on the perspective of sustainable development, the present study brings the adverse environmental effects produced by industrial production into the TFP research framework to evaluate the socioeconomic and ecological conditions of the whole process of land use within the jurisdiction of a city.The analysis combines ML and GTFP indices and requires decision-making unit (DMU) data to be implemented as panel data.The result can analyze the change in productivity and explore the role of technical efficiency and the technical progress in the change of ULGUE through index decomposition.
The efficiency change index (ECI) refers to the change of GTFP brought by the optimal allocation of resources, which is the embodiment of regional management efficiency.
The technology change index (TCI) measures the effect of technological progress on the change of GTFP, which reflects the level of regional scientific and technological development.

| Urban land use in urban agglomerations
Physically, a city is a typical open system with considerable inflows and outflows of resources and energy.Around a central city, several cities of different sizes and endowments are linked together through traffic connection, talent gathering, capital circulation, industrial division, and innovation transmission, gradually forming a functional community, i.e., an urban agglomeration.In the process of social, economic, and ecological integration, each member plays their respective role.With the expansion of urban space demand, land use tends to be an interconnected process instead of a self-sufficient one (Zhao et al., 2020).On the one hand, removing market barriers will bring about a new wave of land demand and trigger a large-scale concentration of capital, labor, and other factors, which may lead to the factor input exceeding the regional carrying boundary, thus inhibiting the improvement of ULGUE.On the other hand, under the market competition, due to the rising cost of new land, some inefficient development projects will be gradually replaced by high-value-added activities, and the land use subjects will also take initiatives to upgrade technology to promote the transformation of urban land use to be green and efficient.Therefore, whether the critical points of the urban agglomeration of a joint and single city regulation are the same has become a key and complex issue of urban agglomeration land management.This issue is particularly crucial at the starting point of construction for urban agglomerations, which will primarily affect the direction and process of the future development of urban agglomerations.

| Traditional data analysis models: CCR and BBC
The first model of DEA is CCR, proposed by Charnes, which considers constant return to scale (CRS); that is, outputs increase in the same proportion as inputs.The CCR model can be expressed as mathematical programming in (1): where θ is the efficiency score; x and y are the inputs and outputs; s xi − and s yj + represent their slack variables; and λ r r epresents the weight coefficient.Then the BCC model relaxes the CRS assumption to variable return to scale (VRS) by adding a constraint condition in Equation ( 2), denoting that the outputs of the BCC model will not increase by the same proportion as the inputs.Both of these models are the basis of other DEA models. (2)

| Super-SBM model with undesirable output
Traditional DEA models include the proportional adjustment of inputs and outputs, but fail to consider the relaxation improvement, causing the results to be biased.It is impossible to sort the efficient units by the level of efficiency because of the same value.Therefore, Tone (2002) proposed a non-radial and non-angular Super-SBM model that overcomes the shortage of no feasible solutions when the DMU is outside the production possibility set (Tang et al., 2021;Zhang et al., 2021).However, ULGUE involves both favorable benefits and some adverse effects, so the Super-SBM-U model is extended.The specific model is as follows.
Suppose there are n DMUs, and each DMU has m inputs, denoted as , and q undesirable outputs, denoted as b k q ( = 1,2, , ) k ⋯ .The unit to be measured is recorded as DMU o .Its efficiency can be evaluated with model (3):

HONG and MAO
Natural Resource Modeling where λ r represents the weight coefficient; and s xi − , s yj + , and s bk − represent the slack variables of inputs, expected output, and undesirable output, respectively.

| The fixed MLI
The MLI based on the DDF model is calculated by Here, FMLI > 1 indicates that GTFP has been improved between periods t and t + 1; indicates the improvement of management; and TCI > 1 indicates advances in technology.

| The spatial error model with fixed effects
Through the results of Lagrange multiplier tests, the SEM with fixed effects is constructed to explore the internal mechanism of the spatio-temporal evolution of ULGUE in MYRUA.The specific model is as follows: where Y refers to the FMLI for each city i and the year t; X is the explanatory variable; βis the regression coefficient of the independent variable; μ i indicates fixed effect; ρ denotes the coefficient of the spatial error; W is the spatial adjacency weight matrix; and ε it and v it are random error terms, which are assumed to be normalized at zero mean value and constant variance (Elahi et al., 2021(Elahi et al., , 2022)).

| Variable selection
According to the statistical data of the three provinces in 2020, the proportion of the secondary and tertiary industries reached 95% in the three major central cities of MYRUA, while the proportion in other cities also exceeded 80%.Therefore, we assume that the secondary and tertiary industries are the economic foundations of a city (Yao & Zhang, 2021), and the indicator system is constructed in Table 1.Following the Cobb-Douglas production function, labor and capital are basic production inputs.Land is the necessary element for human activities in urban development, which must be included in the measurement of ULGUE (Fu et al., 2020;Tang et al., 2021), so the inputs include three basic elements.The green utilization efficiency of urban land emphasizes the coupling of the three subsystems of "economy society ecological environment" (Liang et al., 2019), so the expected output is divided into economic output, social output, and environmental output (Hu et al.,2018).output reflects the social benefits that people gain in the process of urbanization, namely, the improvement of the quality of life and well-being level of urban residents.This article uses social well-being as a representation, and specific indicators include total retail sales of consumer goods and average salary of employees on the job.Environmental output reflects the positive output of urban land use and the degree of environmentally friendly development.This article uses environmental friendliness as a representation, and the specific indicator is Public green space area.Based on previous research, the undesirable output mainly collects the pollution discharged by industrial production (Xie et al., 2019;Yu et al., 2020).

| Selection of driving factors for the spatio-temporal pattern of ULGUE
Based on relevant research, this paper selects economic development (ECO), industrial structure (IND), transport accessibility (TRANS), human capital quality (HUMAN), the level of land marketization (MARKET),' and the behavior of local government (GOV) as the factors influencing the dynamics of ULGUE.
(1) ECO: Conceptualized by the regional GDP per capita.Relative studies have shown that, on the one hand, economic development could accelerate land transfer and promotes the improvement of ULGUE (Chen et al., 2021).On the other hand, with the continuous development of the economy, the contradiction between the supply and demand of land resources becomes increasingly serious, which will negatively impact ULGUE (Zhao et al., 2020).Therefore, the influence of economic development on ULGUE has yet to be verified.(2) IND: Measured by the ratio of the added value of the tertiary and secondary industries.The tertiary industry is more environmentally friendly than the secondary industry (Song & Chen, 2021).Therefore, the higher ratio contributes to the efficient allocation of land resources.
T A B L E 1 Index system of ULGUE.(3) TRANS: Characterized by the area of roads per capita.A well-developed transport network can accelerate the marketization of resources and environmental factors, prompt energyintensive enterprises to take the initiative to adopt green transformation measures, and provide endogenous impetus to optimize the cross-regional allocation of resources, thereby improving ULGUE (Zhao et al., 2021).( 4) HUMAN: Quantified by the number of college students per 10,000 population.Human capital is the core of innovation, and helps to develop more green technologies to enhance ULGUE (Chen et al., 2021).( 5) MARKET: Expressed as the proportion of land auctioned to the total area of land sold in the same year.Lu et al. (2022) indicate that the degree of marketization of land elements significantly contributes to the improvement of construction land and promotes the overall efficiency of ULGUE.(6) GOV: Denoted by the utilization rate of industrial solid waste.The disposal of industrial waste reflects the government's awareness of environmental protection and the implementation of environmental protection policies, which profoundly impact the transformation of polluting enterprises and efficiency gain (Xu et al., 2021).

| Data source
The raw data are mainly obtained from China's City Statistical Yearbook, China's Urban Construction Statistical Yearbook, and China's Regional Economic Statistical Yearbook from 2007 to 2019.The economic data of 28 prefecture-level cities are all exponentially reduced based on 2005.According to the depreciation rate of 9.6%, the fixed asset investment is converted into stock data through the perpetual inventory method (Young, 2003).

| Overall ULGUE of the MYRUA
The Super-SBM-U model is used to calculate the ULGUE from 2006 to 2018.The average value of the MYRUA is 0.984, indicating that the potential impact of urban land use on socioeconomic activities, people's living, and the ecological environment has been highlighted in the region.The efficiency of each city is presented in Figure 2. Changsha maintains the leading role throughout the study period, but the unit with the lowest efficiency changed from Ji'an to Yichang.However, Xiangtan has the lowest average efficiency and is two times lower than Changsha because of the continuous extensive land use mode.

| Cross-section differences of ULGUE
Based on the average ULGUE from 2006 to 2018, all cities are split into two groups.Half are inefficient, whose values are below 1, and the other half are efficient.Then, by applying the natural breaks method, the results are divided into four intervals for spatial comparison: low efficiency (<0.786), medium efficiency (0.786-1.000), and high efficiency (1.001-1.146),and ultrahigh efficiency (>1.146).Specifically, the distribution of efficiency hierarchies and differences among three core cities and internal metropolitan areas are discussed.As shown in Figure 3, there are five low-efficiency units.Among them, Xiangtan, Jiujiang, and Yichun are adjacent to the core cities, where Yichang and Jingmen have been planned as the sub-cities.The efficient units are located on the axis between the core cities and external urban agglomerations, including the Wuhan-Yangtze River Delta, Wuhan-Chengdu Chongqing, Changsha-Pearl River Delta, and Nanchang-Pearl River Delta, revealing the advancement of cross-regional connection for green land use.
Three core cities belonged to different hierarchies.Changsha is in the first class of ULGUE, while Wuhan and Nanchang are in the second class.The efficiency of cities on the Wuhan-Changsha axis is higher than those on the Wuhan-Nanchang axis, and Changsha-Nanchang axis.The intercity connection between Wuhan and Changsha seems stronger than between Nanchang.
In terms of three metropolitan areas, the ULGUE of the WHMA, CZT, and PLCC during 2006-2018 is 1.033, 1.071, and 0.968, respectively (Table 2).The WHMA and CZT are higher than the average level of the whole cluster, while the efficiency PLCC is relatively low.Moreover, each region has significant efficiency gaps (Figure 3).Both the maximum and the minimum ULGUEs are in the CZT.The most efficient unit in the PLCC is Yingtan, whose value is 1.73 times higher than that of Jiujiang.The difference in the WHMA is relatively lower.The city with the highest efficiency is Huanggang.As the worst city, Xianning is at the medium-efficiency level, indicating that the regional development is more coordinated in WHMA than in the other two areas.

| Spatio-temporal characteristics of ULGUE and its driving forces
Taking production data 2006 as a reference, the FMLI from 2007 to 2018 is calculated to demonstrate the variation of GTFP in adjacent years.It is then decomposed into ECI and TCI to explore the internal driving factors.Moreover, the spatial regression model has been introduced to reveal the external forces.

| The improved FMLI and its decomposition of the MYRUA
It can be seen from Figure 4 that the GTFP of urban land in the MYRUA first gradually increases, then slightly fluctuates during 2011-2015, and then drastically increases in recent years.From 2006 to 2018, the average annual growth rate (AAGR) of the whole cluster is 3.30%.Among all members (Table 3), the greatest progress is made by Yiyang, of which the AAGR is 6.37%, though the advancement of GTFP is hardly observed in Yichun.Index decomposition reveals that the TCI of the MYRUA increased by 59.55% during the study period, and its AAGR is 3.97%.In contrast, its ECI is almost unchanged and even decreases in some periods.Except for Ji'an, the TCI of each prefecture-level city is greater than ECI, signifying that technical innovation is the main driving force for the favorable change of ULGUE in the MYRUA.Additionally, nine cities make progress both in TCI and in ECI.None are located in Hubei province, and Changsha is the only core city.
The improved fixed Malmquist-Luenberger index (FMLI) and its decomposition of the urban agglomeration in the middle reaches of the Yangtze river (MYRUA).

| External driving forces of the spatio-temporal evolutionary patterns of ULGUE
Before running the regression model, Moran's I index is calculated in ArcGIS to verify whether spatial dependence is relevant to land use efficiency.Each year, the z-score values of Moran's I index are all significantly positive pass the 15% significance test.Moreover, the LM test, Wald test, LR test, and Hausman test for the robustness of the spatial panel data all reject the original hypothesis, denoting that the fixed effect is superior to the random effect.Thus, this paper adopts a spatial error model with fixed effects.The log-likelihood value of it is 322.538, and the R2 is 0.418, and the results are shown in Table 4.
(1) ECO is significantly positive, indicating that economic development helps to promote technological innovation and optimize the allocation of resources, improving ULGUE.(2) IND is significant at the 5% level, indicating that the higher the proportion of tertiary industries, the faster the improvement of ULGUE.As the industrial structure is upgraded, the proportion of energy-consuming industries will gradually reduce.Still, the high-valueadded industries will expand, thus saving energy per unit area and improving output efficiency, achieving the goal of environmental protection and economic development.
(3) TRANS is positive but does not pass the significance test, indicating that the promotion effect of the transportation network on ULGUE in MYRUA is less evident.Therefore, the city cluster should take advantage of the transportation network to strengthen regional linkages and promote the efficiency spillover from predominance areas.(4) HUMAN is significantly positive, indicating that the improvement of human capital quality has a positive effect on the progress of ULGUE.In other words, if the labor could be transformed into innovative inputs, it will promote land use transformation.( 5) MARKET is significant at the 10% level, indicating that the level of land marketization one of the driving factors.
Based on land value, the market will allocate land resources to more efficient units, thus improving overall efficiency.Therefore, the marketization of land factors can not only control the demand for land resources, but also improve the efficiency of land utilization, achieving green and sustainable development.(6) GOV is positive and significant, indicating that raising the local government's environmental awareness and regulation is conducive to green and high-quality land use, balancing regional development and ecological protection.

| DISCUSSION
From the spatial distribution of green and efficient cities in the urban agglomeration, it can be determined that they were located on the axis between the core cities and external urban agglomerations, confirming that cross-regional connections are conducive to the promotion of ULGUE.In globalization and urbanization, the urban agglomeration has gradually become the basic unit of resources allocation and regional interaction (Gaitani et al., 2014;He et al., 2017).
Though each entity has a specialization, cross-regional connections can promote the flow of resources, the expansion of markets, and the acquirement of advanced technology and management, which contributes to increasing the potential for realizing the green use of urban land.
Regarding the disparities among core cities, the results show that their ULGUE belongs to different hierarchies, and their intercity connections vary in strength, which is similar to Lu et al.'s (2020) research results.Unlike the traditional economic core in the MYRUA, Changsha has displaced Wuhan and become the leader in the green use of urban land.Nanchang exhibits the lowest efficiency, so it should promote the construction of regional transportation and communication facilities to achieve closer contact with others.To take full advantage of core cities, it is imperative to ascertain the center and functional zones in the region.The multi-core pattern will be consolidated and exert a greater positive effect by facilitating connections between core cities.The metropolitan area is at a critical stage and provides robust support for urban agglomeration development.The comparison of these subgroups reveals significant differences among them.The ULGUE of the WHMA and CZT is higher than the average level of the whole cluster, while the PLCC is lagging.Though the GTFP of urban land use has been improved, their main driving forces differ.Therefore, these differences must be considered when observing the coordinated development of MYRUA.The WHMA has a sound industrial foundation and exhibits distinct advantages in terms of resource aggregation, which is the core of technological innovation.The CZT has not only uses traditional production factors but also actively optimizes the layout of advanced manufacturing, high-end service, and high-tech industries, which dominate management skills.Though the PLCC owns plentiful ecological environment resources, the inadequate transportation network, infrastructure construction, and industrial structure have hindered its optimal land allocation.These results further validate the research conclusions of Lu et al. (2023) and Wang et al. (2022).The level of economic development, transportation infrastructure, fiscal expenditure, and other factors have significant spatial spillover effects on the green use of urban land.However, according to the first law of geography, as the distance between cities increases, the spatial spillover effect gradually decreases, forming a spatial pattern of "cities clustered and divided" (Lu et al., 2020).
With the help of index decomposition, the results show that technological progress is the primary source of favorable changes in ULGUE in the MYRUA, which coincides with relevant research (Liu, Xiao, et al., 2020).However, most cities face issues between poor management and unprecedented technological investment.The exposed contradiction reminds us that the exclusive emphasis on technology has its limits, and reforming the management system is significant for urban land's green and sustainable development.On the one hand, technology is a critical factor for resource conservation, cost reduction, and quality improvement, contributing to the advancement of ULGUE.On the other hand, the application of new technologies reqiures adequate market demand and a high-quality workforce, which calls for a scientific management system.
The spatial correlation of ULGUE is evident in the city cluster, which is interwoven by multiple external factors.In terms of factors outside the land use system, economic level, industrial upgrading, high-quality employees, degree of land marketization, and environmental governance are conducive to the convergence of ULGUE.With the high-speed development of the economy, the transition from quantity to quality has been promoted.Additional labor will be engaged in the research and development of energy-saving technologies and forwardlooking green technologies.The tertiary industry, represented by the service industry, will stand out under market selection (Chen et al., 2016;Liang et al., 2019).This will reduce the input of land and produce more green production factors, providing new growth impetus for the green development of urban land and promoting the coordinated improvement of GTFP.Under the concept of green and sustainable development, the government will intervene in land production activities through a series of measures such as increasing financial investment in green technology and infrastructure, establishing a market for pollution emission rights, and carrying our joint control of pollution to improve ULGUE at the macro level (Ge et al., 2021).More importantly, with the in-depth implementation of the regional coordinated development strategy, the technology will be easier to obtain by backward cities from developed regions at low cost through human capital flow and management experience exchange, thus speeding up the process of reducing the regional differences of ULUE.

| CONCLUSIONS
Taking the MYRUA as an example, this paper first constructs a comprehensive evaluation index system that considers economic, social, and ecological benefits and applies the Super-SBM-U model to analyze spatial differences of ULGUE from 2006 to 2018.Then, the improved FMLI is established to investigate the evolution characteristics of GTFP in time and space.Compared with existing research (An et al., 2021;Li et al., 2021;Long et al., 2020), this study avoids the problem of having no feasible solutions and strengthens the consistency of temporal comparison for panel efficiency.Furthermore, we explore the internal and external driving forces of efficiency improvement considering the geographical spatial effect.The major conclusions can be drawn as follows: (1) From 2006 to 2018, the average value of ULGUE in MYRUA was 0.984, and there were significant efficiency gaps among member cities and internal city groups.Three core cities remain ahead of others in ULGUE but belong to different hierarchies.Changsha plays a leading role in promoting the CZT to get ahead of the WHMA and PLCC.High-efficiency cities are located on the axis between core cities and external urban agglomerations, including the Wuhan-Yangtze River Delta, Wuhan-Chengdu Chongqing, Changsha-Pearl River Delta, and Nanchang-Pearl River Delta.
(2) During the study period, the GTFP of urban land has gradually increased with an AAGR of 3.30%.Here, technological innovation is the primary internal motivation, while the low efficiency of management has hindered the further development of ULGUE.
(3) The evolution of ULGUE in MYRUA has spatial correlation characteristics.In terms of factors outside the land use system, economic development, industrial upgrading, highquality employees, land marketization, and government's environmental behavior are conducive to the improvement of ULGUE, while the role of transportation has not been fully exploited yet.
Based on the results and the reality of the MYRUA, we propose the following policy suggestions for the green use of urban land and the integrated development of urban agglomerations.
First, improving regional connections is vital to enhance the radiation effects of central cities on ULGUE.Core cities should serve as a bridge to establish cross-regional connections with other urban agglomerations, which contributes to producing advanced technology and tapping into new markets.It needs to improve internal transportation and support the strategy of opening up to boost the spatial circulation of resources and promote the cooperation of cities inside the urban agglomeration.For areas with lower ULGUE, it is necessary to promote the talent introduction policy and facilitate the industrial transformation, which can relieve the stress of population aggregation and overwhelming land demands on core cities, thus contributing to the convergence of development in the whole city cluster.
Second, Combining the efficient market and the active government could alleviate the conflict between unprecedented technological investment and relatively backward management to achieve the green and coordinated use of urban land.Specifically, the government should pay more attention to control the expansion of construction land, through strengthening environmental regulations, advocating for the application of green technology, and establishing a modern management system to provide a favorable market environment for land transactions.Project developers should take over the responsibility of innovative investment and labor training to tap into the potential of construction land.Only when management skills keep pace with technological progress, can innovation release the full potential to improve the sustainability of urban land use.
Third, each region's ULGUE and individual characteristics should be regarded as the scientific basis of policy design for the urban agglomeration.The government should be aware of ULGUE to establish a comprehensive assessment system considering economic development, people's well-being, and green ecology.Various land development policies should be proposed based on natural endowments, outstanding issues, and evaluation to realize the industrial division's spatial layout and mutual complementary function.This could be supported by several functional areas, such as Wuhan could intensify R&D in electronic information, new energy vehicles and biomedicine, Changsha could focus on engineering equipment, aerospace, and new materials, jointly supporting the development of regional integration.
The space-time dynamics of ULGUE and the driving forces analysis can reveal development differences and the potential mechanism of urban land use.However, there are still many limitations in this paper.First, we regard construction land as a single input without division of functional purposes, making it challenging to implement improvement measures for specific types of land.Second, regarding indicators related to people's living, we only selected two objective data for description: i.e., the total retail sales of social consumer goods and the average salary of urban employees.In the future, subjective evaluations of urban residents, such as residential satisfaction and life happiness, are worthy of consideration to enrich the evaluation index system.Third, the discussion of the relationship between internal and external influencing factors is also worth to deeply study.Therefore, the subsequent research in this article will quantitatively explore the structural characteristics of the ULGUE connection network within the MYRUA based on this study.
Following Liang et al. (2019),Hu et al. (2018),Chen et al. (2021), andTang et al. (2021) in this article, the economic output reflects the economic level of regional urban land use, and is characterized by the economic prosperity indicator, including added value of the secondary and tertiary industries.Social HONG and MAO Natural Resource Modeling | 9 of 22

F
I G U R E 2 The urban land green use efficiency (ULGUE) of each prefecture-level city during 2006-2018.

F
I G U R E 3 Spatial differences of efficiency of urban land green use (ULGUE) in the urban agglomeration in the middle reaches of the Yangtze river (MYRUA).
To demonstrate the efficiency variation of panel data, the Super-SBM-U model is applied to ensure feasible solutions.Furthermore, data in 2006 is taken as a reference, providing a consistent standard for time series comparison.The index is constructed as follows.First, model (3) is used to obtain the contemporaneous efficiencies input and output vectors of DMU o in periods t and t + 1, respectively.Like the MLI, the FMLI can be decomposed into the ECI and TCI.
+1 , it is the efficiency of input and output data in period t + 1 which refers to the production frontier in period t.Therefore, the MLI of different periods has its own production frontiers.Furthermore, if the data exceeds the production possibility set in period t, there are no feasible solutions. .
The improved FMLI and its decomposition of prefecture-level cities.
T A B L E 3Abbreviations: ECI, efficiency change index; FMLI, fixed Malmquist-Luenberger index; TCI, technology change index.
Results of the estimated model.