Land-Use Transformation and Landscape Ecological Risk Assessment in the Three Gorges Reservoir Region Based on the “Production–Living–Ecological Space” Perspective

Abstract

Rapid urbanization and land-use change cause risk in regional ecological security. It is very significance to explore the evolutionary trend of land-use change and landscape ecological risk (LER) in an ecologically fragile area, especially in terms of maintaining sustainable development in a regional ecological environment. We selected the Three Gorges Reservoir Region (TGRR) as the study area based on land-use and land-cover data for 2000, 2010, and 2020. The land-use classification system used here was constructed using the perspective of the production–living–ecological space (PLES). The GIS spatial-analysis technique and FRAGSTATS 4 software were used. We used the method of the land-use transfer matrix, the landscape ecological risk assessment model, the ecological contribution rate of land-use transfer, and spatial autocorrelation analysis. We performed quantitative analysis of the spatio-temporal pattern of PLES and its LER in the TGRR over the past 20 years. The results show that: (1) The area of human living space (HLS) has expanded significantly—by 1469.37 km² (+326.66%), while the area of agricultural production space (APS) has been compressed by both the urban/rural living space (URLS) and the forestland ecological space (FES), particularly during the last 10 years; (2) The overall LER results were medium, but LER is increasing; (3) The LER in the northern area of the Yangtze River is higher than in the south. The Wanzhou district and the downstream areas had a lower LER; and (4) The transformation from agricultural production space to forestland ecological space and urban/rural living space has had a higher contribution rate to the LER compared to other events. These results can be used as a reference for land planning, sustainable development, and ecological civilization construction in ecologically fragile areas.

1. Introduction

Since period of reform and opening up began, China has witnessed rapid economic and social development. However, many cities still rely mainly on occupying large amounts of farmland, forests, pastureland, and bodies of water to improve their urbanization [1]. To some extent, this rapid urbanization has intensified the human pursuit of materials and the demand for natural resources. This has led to the degradation of natural ecosystem services and an increasing ecological risk [2,3]. “Land-use transformation” refers to changes in regional land-use forms due to social and economic development. It is mainly reflected in changes in the function of the production–living–ecological space (PLES). Therefore, the ecological risks caused by the transformation of land use is essentially caused by the imbalance of the PLES. “Ecological risk” refers to the negative ecological impact of environmental changes and human social activities on the natural ecology [4,5,6]. Landscape ecological risk (LER) assessment is an important part, a supplement, and an expansion of ecological risk assessment. It is a method for studying ecological risk from the perspective of the spatial heterogeneity, landscape patterns, and ecological process interactions concerned in landscape ecology [7,8]. As an essential branch of environmental risk assessment, it mainly emphasizes the comprehensive analysis of the various possible large-scale disasters in the regional ecological environment, which has a guiding significance for the improvement and development of the quality of the regional ecological environment [9,10,11]. This method supports the study of ecological risk from the perspective of spatial heterogeneity and landscape pattern–ecological process interactions, which are central players in landscape ecology. With the development of global climate-change and ecological-risk research, conducting LER assessments using the perspective of land-cover change has become the mainstream approach [12,13]. Based on the leading function of land use, combining the spatial evolution of PLES with the transformation of land use is a new entry point for solving the coordinated development problem of the PLES and promoting regional sustainable development. The report of the 18th National Congress of the Communist Party of China (CPC) and the 14th 5-Year Plan of the CPC Central Committee on National Economic and Social Development both put forward the goal to develop the nation’s PLES. Clarifying the spatial evolution process of the PLES and the landscape ecological risks against the background of rapid social and economic development for optimizing the territorial spatial layout and the sustainable development of the Three Gorges Reservoir Region (TGRR) is of great significance. This paper focuses on possible strategies for connecting the PLES concept with the leading functions of land use, combined with the long-term evaluation of the ecological risk faced by the landscape so as to explore the relationship between the spatial-type conversion of Production–Living–Ecological space and the landscape ecological risk.

In recent years, substantial progress has been made regarding research on LER. Scholars have attempted to use multiple indicators to build a comprehensive LER model so as to evaluate environmental hazards and values, as well as the risks of the landscape on different scales. There has been a considerable focus on the exploration of LER assessment methods and models, mainly on risk source–receptor relationships. For example, Ayre et al.’s (2012) Bayesian network model was based on an ecological risk assessment framework to evaluate the forest landscape of north-eastern Oregon at three levels: landscape disturbance, habitat, and ecological resources [14]. Hayes (2004) conducted a regional ecological risk assessment of the offshore marine environment in northwest Washington using relative risk models [15]. Forbes (2013) improved predictive systems models (PSMs) to address the complexity and relevance of ecological risk assessment [16]. Paukert et al. constructed a landscape-scale ecological risk assessment for the lower Colorado River basin on four watershed scales in terms of land use, waterway development and diversions, and human development [17]. Researchers in China have been further exploring the factors affecting LER from the perspective of landscape ecology, gradually elucidating the dynamic changes in LER and its spatial and temporal patterns. The diversity metrics, uniformity, dominance, separation, and fragmentation indices of the landscape were introduced by Fu (1996) and Chen (1996) to lay the foundation for the analysis of the spatial patterns of the landscape [18,19]. Zeng et al. (1999) constructed an ecological risk model using the weighted areas of landscape components and ecological risk intensity using an expert scoring method to explain the spatial characteristics and the intrinsic formation mechanism of ecological risk [20]. In a further approach for interpreting LER indices, spatial statistical methods were introduced to analyse the spatial correlation of LER indices [21,22], and geographic detectors were introduced to quantitatively analyse changes in LER and their driving forces [23]. However, from the perspective of risk sources or risk receptors, most of the studies have focused on a single land-use type, and few studies have been conducted to analyse LER from a PLE perspective [24,25]. Moreover, the research has been limited to an administrative scale by the lack of grid-unit scale.

The TGRR is not only a national strategic freshwater resource pool, but it is also an important ecological barrier area in the Yangtze River Basin and a model area of “ecological priority and green development” in the Yangtze River Economic Belt, covering 26 districts and counties in Chongqing and Hubei. Based on three phases of land-use data in the TGRR, this paper connects land-use types with the PLES, analyses the evolution of land-use types using the land-transition matrix, and explored the spatial and temporal patterns of ecological risk in the TGRR using the LER assessment model. This study also investigates the influence of the contribution ratio of land-use-function transformation on LER from the PLES perspective. The goal is to provide a theoretical basis and reference for both the efficient use of land resources and the formation of comprehensive risk-prevention decisions in the TGRR.

2. Materials and Methods

2.1. Study Area

The TGRR is located in southwestern China (105°49′–111°39′ E, 28°31′–31°43′ N), crossing the mountain valleys in central Hubei and the ridge and valley belt in eastern Sichuan, separated by the Daba Mountains in the north and the Sichuan–Hubei Plateau in the south. It includes 22 districts and counties under the jurisdiction of Chongqing and 4 districts and counties in Hubei Province, and it has a total area of 57,543 km² (Figure 1). The area has a subtropical humid climate and is mainly mountainous and hilly, with large topographic undulations. The forest coverage-rate is about 48.2%. The soil types are mainly purple soil, lime soil, yellow soil, and rice soil. With the increase in the water-storage level during the construction of the Three Gorges Project, the soil erosion in the study area reached a serious level. A large number of forest ecosystems were destroyed, and a large number of sudden geological disasters were caused. It is an important part of the Economic Zone in the Upper Reaches of the Yangtze River and an ecological environmental barrier in the Middle and Lower Reaches of the Yangtze River. The rapid deterioration of the ecological environment has transformed the Three Gorges Reservoir area into one of the main ecologically fragile areas in the Upper Reaches of the Yangtze River.

2.2. Data

The land-cover data were derived from GlobeLand30’s global surface-cover database (http://www.globallandcover.com/ (accessed on 28 May 2022)) and include data for 3 periods: 2000, 2010, and 2020. GlobeLand30 was created and donated to the United Nations by the Chinese government. It is the first high-resolution land-cover mapping product in the world, and it has a long time-span and high accuracy [13,26,27]. The spatial resolution of the data is 30 metres, and the land-cover types in the study area include arable land, woodland, grassland, shrubland, wetland, bodies of water, and artificial surfaces: a total of 7 types. To avoid potential confounding variables when comparing data from different years, it was important to consider the changes in the Yangtze River’s water level that were caused by the storage of water in the TGRR. In this paper, the Yangtze River’s main stem for the three sampled years was extracted using ArcGIS 10.4, employing the “union” function to form the largest surface. The data from each year was “identified” and “dissolved” using this surface so that the pre-processed data were more comparable and robust. Combining the theory of PLES with the results from previous research [28,29,30], each land-use category was linked with the leading function of the PLES. We calculated the average area of each land-use category using the weighting method, and we assigned an ecological environmental quality index to each PLES land-type (

Table 1. Association of the PLES classification system with the land-use categories.

PLES Land Use Dominates the Functional Classification		Secondary Classification of the Land-Use Classification System	Ecological Environmental Quality Index
Class I	Class II
Production space	Agricultural production space (APS)	Arable land	0.293
Living space	Urban/rural living space (URLS)	Artificial surface	0.010
Ecological space	Forest ecological space (FES)	Woodland, shrubland	0.883
	Pasture ecological space (PES)	Grassland	0.798
	Water ecological space (WES)	Body of water, wetland	0.521

).

2.3. Methods

2.3.1. Division of the LER Assessment Unit

The grid-based GIS method was used. The grid was used as the research unit, and we used the “create fishnet” function in the ArcGIS 10.4 data management module. The TGRR was divided into several grids and converted into grid data for subsequent studies. Considering the sensitivity of the risk plot size to the landscape parameters and the calculation workload, based on existing studies [31,32,33], we found it appropriate to use 2–5 times the average patch area for the risk plot. The average patch area in the study area was about 7.8 km², so a square grid of 6 km × 6 km was used. The study area was divided into 1846 LER patches, based on which the LER was calculated for each risk patch.

2.3.2. The PLES Land-Use Transfer Matrix

The PLES land-use transfer matrix can reflect the structural change characteristics of the PLES land-type in the study area over time [34]. The formula is as follows:

$$S_{ij}=\left[\begin{array}{ccc}{S}_{11}& S_{12} \dots & S_{1n}\\ S_{21}& S_{22} \dots & S_{2n}\\ S_{n1}& S_{n2} \dots & S_{nn}\end{array}\right]$$

(1)

where S refers to the area of the PLES; n refers to the number of land-types in the PLES; and i and j refer to the PLES land-types at the beginning and end of the study, respectively.

2.3.3. The LER Assessment Model

The LER assessment is the basis for judging the adverse effects of human activities and natural disasters on the functions and structures of regional ecosystems. The landscape pattern is the result of many ecological influences, including external interference on different scales [35]. According to previous research results [36,37], we combined the landscape disturbance index (E_i) and the landscape vulnerability index (V_i), which affect the stability of the regional ecosystem. The LER in the study area was calculated in order to analyse the evolutionary trend of the regional LER due to the PLES land-use change [38]. The formulas are as follow:

$$LE{R}_i=E_i·V_i$$

(2)

where $LE{R}_i$ refers to the integrated LER of the i-th grid. This is a regional ecological risk based on spatial patterns, and it can measure the degree of the destruction of natural ecosystems.

$$E_i=aC{n}_i+bS{n}_i+cD{n}_i$$

(3)

where $E_i$ refers to the landscape disturbance index, which indicates the exogenous causes of regional ecological risk. It measures the magnitude of disturbances originating from natural and anthropogenic factors at the landscape level. The values a, b, and c indicate the weights of the landscape fragmentation index (C_i), the landscape isolation index (S_i), and the landscape dominance index (D_i), respectively. Thus, a + b + c = 1. According to numerous studies [24,25,39], C_i is considered to be the most important value, followed by S_i and D_i. Thus, a, b, and c are usually assigned uniform values in different regions, and their values are 0.5, 0.3, and 0.2, respectively. $C{n}_i$, $S{n}_i$, and $D{n}_i$ are the normalised indicators of C_i, S_i, and D_i, respectively.

$$V_i={\sum }_{j=1}^n{V}_j\frac{A_{ij}}{A_i}$$

(4)

where $V_i$ refers to the landscape vulnerability index, which indicates the extent of human development. Its ecological significance is that the V_i grows gradually larger with an increase in human development intensity, and then the LER becomes larger. This paper refers to existing studies and combines the scientific nature of the mathematical formulas to normalise and assign vulnerability coefficients to each landscape-type. The vulnerability coefficients ($V_j$) of artificial surfaces, arable land, woodlands, shrublands, grasslands, bodies of water, and wetlands are 1, 0.5, 0.4, 0.3, 0.2, 0.1, and 0.1, respectively [40]. $A_i$ is the sum of the areas of all landscape-types in the i-th cell. $A_{ij}$ is the total area (km²) of the j-th landscape-type in the i-th cell [39].

$$C_i={\sum }_{j=1}^n{C}_{ij},\hspace{1em}{C}_{ij}=N_{ij}/A_{ij}$$

(5)

where $C_i$ refers to the landscape fragmentation index, which indicates the degree of landscape fragmentation at a given time. Higher values represent greater degrees of the fragmentation of the landscape and greater human interference with the landscape. $C_{ij}$ is the landscape fragmentation metric of the j-th land-use landscape-type in the i-th cell. $N_{ij}$ is the number of land-use patches of the j-th landscape-type in the i-th cell. N is the total number of landscape-types in this grid [39].

$$S_i={\sum }_{j=1}^n{S}_{ij},\hspace{1em}{S}_{ij}=\frac{A_i}{2A_{ij}}\sqrt{\frac{N_{ij}}{A_i}}$$

(6)

where $S_i$ refers to the landscape isolation index, which indicates the degree of isolation of the different plaques in a landscape-type. The larger the value, the more complex the corresponding spatial distribution of the landscape. $S_{ij}$ is the landscape isolation of the j-th landscape-type in the i-th cell of the grid [39].

$$D_i=\ln n+{\sum }_{j=1}^N\frac{A_{ij}}{A_i}\ln \frac{A_{ij}}{A_i}$$

(7)

where $D_i$ refers to landscape dominance index, which indicates the dominance of one or more landscape-types in the landscape structure [39].

2.3.4. Spatial Autocorrelation Analysis

In this paper, GeoDa software was used to perform the spatial autocorrelation analysis of the LER. This was used to test whether the attribute values of certain spatial variables were significantly correlated with those of neighbouring areas. The analysis can be split into a global and a local spatial autocorrelation. The global spatial autocorrelation is a quantitative indicator of spatial correlation with other spatial units for one spatial unit [41]. Local spatial autocorrelation more strongly shows the spatial aggregation of LER, which can graphically depict the spatial aggregation of the ecological risk. These can be classified as high–high, high–low, low–low, low–high, or as an insignificant aggregation [42].

There are many global autocorrelation statistics, the most widely used of which is Moran’s I, with the following equation. The formula is below:

$$\mathrm{Moran}’\mathrm{s} I=\frac{n}{{\sum }_{i=1}^n{\left(x_i-\overline{x}\right)}^2}·\frac{{\sum }_{i=1}^n{\sum }_{j=1}^n{W}_{ij}\left(x_i-\overline{x}\right)/\left(x_j-\overline{x}\right)}{{\sum }_{i=1}^n{\sum }_{i=1}^n{W}_{ij}}$$

(8)

where n is the number of spatial units involved in the analysis; $x_i$ and $x_j$ are the observed values of the i-th and j-th spatial units; $\overline{x}$ is the average value of $x_i$; $W_{ij}$ is the spatial weight matrix between $i$ and $j$, which represents the proximity of each spatial unit. The value range of $\mathrm{Moran}’\mathrm{s} I$ is [–1, 1]. Subject to passing the significance level test, an $I$ value closer to 1 indicates a more substantial positive spatial correlation of the variable, while a value close to −1 indicates a more substantial negative spatial correlation. When $I$ = 0, it indicates a random distribution in space.

2.3.5. Ecological Contribution Rate of PLES Land-Use Transformation (LEI)

LEI refers to the change in regional ecological environmental quality caused by the transformation of a certain PLES land-use-type. Quantifying the ecological and environmental effects caused by land-use-type transformation from both positive and negative aspects. The land-use conversion types that affect the changes of regional ecological environmental quality can be separated. It is convenient to explore the leading factors that cause the regional ecological environmental changes. The formula is below [43]:

$$\mathrm{LEI}=\left({\mathrm{LE}}_{1 }-{\mathrm{LE}}_0\right)\mathrm{LA}/\mathrm{TA}$$

(9)

where LEI refers to the ecological contribution rate of a certain land-type change. ${\mathrm{LE}}_{0 }$ and ${\mathrm{LE}}_1$ refer to the ecological environmental quality index of a certain land-type at the beginning and the end of the change. LA refers to the area of a certain land-type change. TA refers to the total area of TGRR.

3. Results

3.1. Analysis of the Spatial Evolution of the PLES

This paper analyses the spatial changes in the PLES and transition matrix from 2000 to 2020. Regarding the spatial changes of the PLES (

Table 2. Spatial evolution of the PLES in the TGRR from 2000 to 2020 (km²).

PLES		2000	2010	2020	Area of Change 2000–2010	Ratio %	Area of Change 2010–2020	Ratio %
Production space	APS	23,340.97	23,407.20	22,238.48	66.23	0.28	−1168.72	−4.99
Living space	URLS	449.81	662.15	1919.19	212.34	47.21	1257.04	189.84
Ecological space	FES	27,473.96	28,762.86	28,613.91	1288.90	4.69	−148.95	−0.52
	PES	5125.37	3591.03	3551.19	−1534.34	−29.94	−39.84	−1.11
	WES	1152.72	1119.58	1220.06	−33.14	−2.87	100.48	8.97
	Subtotal	33,752.05	33,473.48	33,385.16	−278.57	−0.83	−88.32	−0.26

and Figure 2), the TGRR was dominated by ecological space and agricultural production space, and the ecological space was larger. In terms of specific changes, the urban/rural living space increased the most significantly, from 449.81 km² in 2000 to 1919.19 km² in 2020 (+326.66%). This is explained by the rapid promotion of urbanisation after Chongqing became a municipality directly governed by the Central Government, especially after the expansion of the city and rapid industrial development. The agricultural production space did not change significantly from 2000 to 2010, but it decreased by 1168.72 km² between 2010 and 2020 (−4.99%). This indicates that the protection of agricultural production space in the TGRR was relatively well-controlled. With the expansion of urban/rural living space, reclamation and other measures to restore agricultural production space also progressed simultaneously. Ecological space was slowly decreasing, and the rate of decrease in the first 10-year period was higher than that in the second 10-year period. The rates of decrease in pasture ecological space and water ecological space were larger.

As observed in the land-use transition matrix (

Table 3. Spatial transition matrix of the PLES in the TGRR from 2000 to 2010 (km²).

2000	2010					Sum
	APS	URLS	FES	PES	WES
APS	21,093.78	232.55	1591.56	369.13	53.95	23,340.97
URLS	70.26	366.64	5.83	5.23	1.86	449.82
FES	1605.25	6.84	25,068.58	774.77	18.52	27,473.96
PES	575.17	51.85	2074.72	2409.64	13.98	5125.36
WES	62.74	4.28	22.17	32.26	1031.27	1152.72
Sum	23,407.20	662.16	28,762.86	3591.03	1119.58	57,542.83

and

Table 4. Spatial transition matrix of the PLES in the TGRR from 2010 to 2020 (km²).

2010	2020					Sum
	APS	URLS	FES	PES	WES
APS	20,644.58	1128.96	1213.84	352.45	67.37	23,407.20
URLS	18.02	628.51	5.61	8.50	1.51	662.15
FES	1231.20	97.40	26,582.62	811.53	40.11	28,762.86
PES	325.96	59.25	799.13	2372.49	34.21	3591.04
WES	18.72	5.06	12.71	6.23	1076.87	1119.59
Sum	22,238.48	1919.18	28,613.91	3551.20	1220.07	57,542.84

and Figure 3), urban/rural living space was mainly transferred from agricultural production space from 2000 to 2010, with an area of 232.55 km². This was followed by the transfer of pasture ecological space, with an area of 5.83 km². Urban/rural living space was still mainly transferred from agricultural production space from 2010 to 2020, with an area of 1128.96 km². Compared with the previous 10 years, the extent of encroachment on agricultural production space by urban/rural living space increased considerably from 2010 to 2020, followed by the transfer of forestland ecological space and pasture ecological space to urban/rural living space. The quantitative changes and shifts in the PLES showed that the urban/rural living space expanded significantly. In contrast, the agricultural production space was compressed, but it was also supplemented to a certain extent.

3.2. Analysis of Spatial and Temporal Changes in LER

The LER of each risk zone was calculated using the LER model. Based on a comprehensive 3-year LER and using the natural breakpoint method, the LER level of the TGRR was classified into 5 categories: lowest (0–0.09), low (0.09–0.18), medium (0.15–0.26), high (0.26–0.38), and highest (0.38–1). The area share of each landscape ecological risk class and the mean value were determined (

Table 5. LER-level area scale and mean value in the TGRR from 2000 to 2020.

LER Grade	Lowest	Low	Medium	High	Highest	Mean Value of LER
2000	20.68%	27.71%	23.24%	21.52%	6.84%	0.191
2010	20.52%	28.49%	20.58%	21.96%	8.45%	0.195
2020	19.80%	28.61%	20.03%	22.65%	8.90%	0.200

). The LER of the TGRR from 2000 to 2020 continued to increase, and the overall LER was at the medium level, with mean values of 0.191, 0.195, and 0.2, respectively. Specifically, the proportion of the highest LER gradually increased from 6.84% in 2000 to 8.9% in 2020, while the proportion of high LER was also gradually increasing, and the LER area with the lowest risk showed a decreasing trend.

The spatial distribution of the LER level in the TGRR from 2000 to 2020 was obtained by ordinary Kriging interpolation using ArcGIS geostatistical analysis (Figure 4). The highest and high LER areas were clearly defined and mainly concentrated in Wanzhou District and its upstream area. The risk in the northern area of the Yangtze River was obviously higher than that in the southern area. The highest LER areas were mainly distributed in the junction of Shapingba District, Jiangbei District, Beibei District, and Yubei District, as well as in the Changshou District, Fuling District, Fengdu County, and Zhongxian County. The low and lowest LER areas were mainly concentrated in Wanzhou District and downstream.

3.3. Spatial Autocorrelation Analysis of Ecological Risk in the Landscape

In this paper, GeoDa software was used to analyse the values of Moran’s I. The global autocorrelation statistics of the LER for the 3 periods were 0.5529, 0.6269, and 0.5944, all of which were greater than 0. This indicates that the LER indices of all districts and counties in the TGRR showed some autocorrelation in space within these 3 years. The spatial clustering was analysed and discussed using local autocorrelation LISA plots (Figure 5). The areas with high LER in the TGRR were clearly concentrated in the northern part of the main city of Chongqing, as well as in the Changshou District, Fuling District, and Fengdu County, extending to Zhongxian County. It is noteworthy that there is some high–high aggregation at the junction of Kaizhou District and Yunyang District.

3.4. The Impact of the Land Transformation of PLES on the Ecological Risk of the Landscape

The land-use transformation of the PLES affects the LER and security of the TGRR. Regional changes in LER are often accompanied by both improvement and deterioration. To a considerable extent, these may counteract each other within a certain area so that the overall situation remains relatively stable. Combined with the actual situation of the study area and information in the literature [30], the ecological contribution rate of land-use transformation can be used to analyse the impact of a certain change in spatial usage on regional ecological quality. Map algebra and the land-use transition matrix were used to explore the dominant factors affecting regional ecological changes leading to both the improvement and deterioration of ecological quality (

Table 6. Contribution of PLES transfer to effects on ecological environmental quality in the TGRR from 2000 to 2020.

Transformation of the PLES (Improvement)	Metrics Change	Weight of Contribution (%)	Transformation of the PLES (Degradation)	Metrics Change	Weight of Contribution (%)
APS-FES	0.01607	66.61	FES-APS	−0.01652	49.75
APS-PES	0.00382	15.84	APS-URLS	−0.00646	19.46
PES-FES	0.00322	13.34	PES-APS	−0.00524	15.79
APS-WES	0.00039	1.62	PES-URLS	−0.00153	4.62
URLS-APS	0.00024	0.99	FES-URLS	−0.00144	4.35
URLS-PES	0.00012	0.50	FES-PES	−0.00130	3.93
WES-FES	0.00010	0.42	FES-WES	−0.00023	0.69
URLS-FES	0.00008	0.34	WES-APS	−0.00022	0.66
WES-PES	0.00007	0.29	PES-WES	−0.00014	0.42
URLS-WES	0.00002	0.06	WES-URLS	−0.00011	0.33

).

The conversion of agricultural production space into forestland ecological space and pasture ecological space, and the conversion of pasture ecological space into forestland ecological space in the TGRR from 2000 to 2020 were the dominant factors affecting ecological improvement, amounting to a total contribution of 95.79%. In contrast, the conversion of forestland ecological space into agricultural production space, the conversion of agricultural production space into urban/rural living space, and the conversion of pasture ecological space into the agricultural production space were the dominant factors leading to ecological degradation, with a total contribution of 85%. It is noteworthy that the expansion of urban/rural living space was mainly derived from the shrinking of agricultural production space. The transformation of agricultural production space to urban/rural living space indirectly led to changes in landscape fragmentation, landscape isolation, and landscape dominance indices, which directly affected the landscape vulnerability metrics. This increased the original landscape vulnerability metrics, thus enhancing the regional LER.

4. Discussion

Based on existing studies [28,29,30,34], through PLES land-use change and landscape pattern interaction research, the connection between these can be better analysed at the macro level. This paper presents a land-use classification system from the perspective of the PLES. We used the land-use transfer matrix to analyse the spatial evolution of the PLES, taking a 6 km × 6 km grid as the evaluation unit. We referred to the existing research [24,25,31,32,33,36,37,38,39,40], built a LER assessment model, and used spatial autocorrelation analysis to analyse the spatio-temporal distribution characteristics of LER in the study area over the last 20 years. The ecological environment of the TGRR has been improved by protection policies in recent years, but in general, the ecological environment is very fragile. Therefore, it is necessary to conduct a LER assessment for this area. Based on our LER analysis of the TGRR, we propose the following: First, for medium and high-grade ecological risk areas, importance should be attached to the impact of urban construction and the increase of industry along the river on the ecological risks on both sides of the Yangtze River. The anthropogenic disturbance of highly vulnerable landscape-types, such as bodies of water, should be reduced, and wetland protections should be strengthened to reduce the risk of soil erosion in the TGRR. Second, for low-grade ecological risk areas, environmental governance should be strengthened, and an ecological buffer zone should be established. The possibility of forestland and grassland degradation should be reduced, and desertification should be prevented. As future protection, the integrity and heterogeneity of the landscape of the TGRR should be respected, along with the self-restoration of nature.

Changes in the PLES influence changes in the landscape pattern of the TGRR, which in turn affects the ecological risk of the region. In this paper, we have tried to use the ecological contribution rate of land-use transformation [43] and make a preliminary exploration of the strength of the PLES land-use types on the improvement and deterioration of the ecological environment. The results indicate that the changes in regional LER are associated with the transformation of the PLE land-use function type. Therefore, the control of regional land-use transformation, the rational planning of land use, and ecological environmental protection should be strengthened. The system of the paid occupation of ecological land and the system of ecological space-use control should be implemented. The ecological environmental governance and protection level of the TGRR should be comprehensively improved. This paper provides a reference for land planning, sustainable development, and ecologically civilized construction in ecologically fragile areas with an adherence to the principle of “ecological priority” against the background of the integration of “dual carbon” goals and “multi-plans”. The concept of “dual carbon” will be fully integrated into the territorial space planning and play overall leading role in planning low-carbon space governance. We recommend the building of a scientific and reasonable low-carbon territorial spatial pattern and limiting territorial spatial development under the framework of “conditional planning”. We also recommend comprehensively optimizing human–land relations and promoting the low-carbon operation of the “economic–social–ecological” composite system [44].

It should be pointed out that the methods and ideas in this study need to be improved:

(1)

The paper relies too much on the present situation of land-use functions. However, the changes in factors, such as precipitation, topography, and soil organic-matter content, as well as changes in economic, social, policy, and demographic factors, affect changes in the PLES. The in-depth analysis of various factors in the human and land system is very promising to attempts to reveal land-use-change and landscape-pattern problems more fully. Therefore, in future studies, it is necessary to further evaluate and predict the evolution of PLES from multiple natural, social, and economic perspectives.

(2)

The calculation methods and ecological significance of the parameters in the LER model were further elucidated in this paper by analysing a large quantity of research. However, LER assessment is a complex process that requires considering multiple uncertainties. These factors determine the comprehensive evaluation results. In the LER assessment of the TGRR, the method and process need to be further improved.

(3)

The method by which to bring the ecological environment index of different PLES land-use-types more in line with objective reality requires further analysis.

5. Conclusions

This paper based is on land-use and land-cover data on the TGRR from 2000, 2010, and 2020. The land-use classification system was constructed using the perspective of the PLES. The GIS spatial analysis technique and FRAGSTATS 4 software were used with a basis in the methods of the land-use transfer matrix, LER assessment model, ecological contribution rate of land-use transfer, and spatial autocorrelation analysis. The spatio-temporal pattern of PLES and its LER in the TGRR in the last 20 years is discussed quantitatively. The results indicate that:

(1)

From 2000 to 2020, the spatial area of the TGRR changed significantly. The area of living space increased by 1469.37 km² (+326.66%), while the production space area shrank by 1102.49 km² (−4.72%). Moreover, the ecological space decreased by 366.88 km² (−1.09%). In terms of PLES land-use function-type transfer, agricultural production space was mainly compressed by urban/rural living space and forestland ecological space, especially over the last ten years. The extent to which urban/rural living space displaced agricultural production space has increased significantly.

(2)

Based on the spatial and temporal changes in the LER, the ecological risk in the TGRR showed special changes in stages. Globally, the TGRR was at a medium LER from 2000 to 2020. However, the mean value of the LER showed an increasing trend. Locally, the boundaries of the highest and high LER areas were clear, concentrated in the upper reaches of the area. The risk in the northern area of the Yangtze River was significantly higher than that in the southern area.

(3)

The spatial autocorrelation analysis revealed that the high–high aggregation areas were mainly located in the northern part of the main city of Chongqing, as well as in Changshou District, Fuling District, and Fengdu County. Furthermore, the high–high aggregation areas and low–low aggregation areas were highly consistent with the distribution ranges of high-risk areas and low-risk areas.