Evaluation of ESV Change under Urban Expansion Based on Ecological Sensitivity: A Case Study of Three Gorges Reservoir Area in China

: In recent years, ecosystem service values (ESV) have attracted much attention. However, studies that use ecological sensitivity methods as a basis for predicting future urban expansion and thus analyzing spatial-temporal change of ESV are scarce in the region. In this study, we used the CA-Markov model to predict the 2030 urban expansion under ecological sensitivity in the Three Gorges Reservoir area based on multi-source data, estimations of ESV from 2000 to 2018 and predictions of ESV losses from 2018 to 2030. Research results: (i) In the concept of green development, the ecological sensitive zone has been identiﬁed in Three Gorges Reservoir area; it accounts for about 35.86% of the study area. (ii) It is predicted that the 2030 urban land will reach 211,412.51 ha by overlaying the ecological sensitive zone. (iii) The total ESV of Three Gorges Reservoir area showed an increasing trend from 2000 to 2018 with growth values of about USD 3644.26 million, but the ESVs of 16 districts were decreasing, with Dadukou and Jiangbei having the highest reductions. (iv) New urban land increases by 80,026.02 ha from 2018 to 2030. The overall ESV losses are about USD 268.75 million. Jiulongpo, Banan and Shapingba had the highest ESV losses.


Introduction
An ecosystem is a unity of the biota and the abiotic surroundings in a certain space of nature [1]. It is also an ecological functional unit formed by the continuous exchange of energy and materials between organisms and their abiotic environment [2][3][4]. Ecosystems can provide a range of services for human production and livelihoods [5][6][7]. Ecosystem services are the materials that humans need to obtain from the Earth's ecosystems and natural environment [8,9]. Ecosystem services are diverse, and the services are unique and irreplaceable for different ecosystem [10][11][12]. For example, forests are the most essential ecosystems for human life, providing wood, regulating air, promoting soil formation and supporting social benefits such as spiritual, landscape and educational values [13]. The evaluation of ESV can measure the status of a regional ecosystem.
The evaluation of ESV is a method based on ecology, economics and sociology [14][15][16]. It can quantitatively evaluate ecosystem services from the perspective of monetary value and better reflect the change in ESV and provide decision makers with knowledge to make policy adjustments [16,17]. In the early days, its evaluation methods made some progress in emphasizing the impact of socio-environmental change on ecosystem services (supply, regulation, support and entertainment services) [18][19][20]. Research scholars began to use east. The central and western parts of the reservoir area are dominated by platforms and hills. The eastern part is close to the Daba Mountain and has many rolling hills [39]. There are abundant vegetation types in the Three Gorges Reservoir area, and the overall spatial distribution of vegetation coverage shows the characteristics of high in the east and low in the west [40]. The east is mostly broad-leaved forests, bushes and grasslands; the central and western area are farming area with many cultivated plants and crops [41]. The Three Gorges Reservoir area is an important ecological barrier in the Yangtze River Basin, China's strategic water resources reserve. Therefore, the prediction of ESV change in the Three Gorges Reservoir area is important for China's sustainable development under the future urban expansion.

Materials
The two main types of data used in this study are spatial data and statistical yearbook data. Spatial Data. (1) Administrative boundary vector data of Three Gorges Reservoir area (SHP format). (2) Soil dataset provided by Harmonized World Soil Database (HWSD) and Cold Arid Regions, Available online: http://www.westdc.westgis.ac.cn (accessed on 17 April 2019), which contains the spatial coordinates and properties of the soil (GRID format).
(3) Digital elevation model (DEM), which was downloaded from the Geospatial Data Cloud Available online: http://www.gscloud.cn/ (accessed on 26 April 2020). In the above data, the DEM data, with a resolution of 30 m, can be extracted into slope and elevation. (4) The highway, the primary road and railroad data procured from OpenStreetMap Available online: http://www.openstreetmap.org (accessed on 13 July 2020). The population density data was provided by Landscan Available online: https://landscan.ornl.gov/ (accessed on 18 December 2020). (5) The 1000 m Normalized Difference Vegetation Index (NDVI) data and 30 m land use and land cover change (LUCC) data were obtained from the Data Centre for Resources and Environmental Sciences, Chinese Academy of Sciences (RESDC) Available online: http://www.resdc.cn (accessed on 26 April 2020). NDVI and LUCC data were obtained by remote sensing image processing. NDVI data was based on inversion of SPOT/VEGETATION and MODIS satellite remote sensing and LUCC data was generated by manual visual interpretation based on Landsat 8 remote sensing images. (6) China Nature Reserve, the earthquake and landslide vector data acquired from RESDC. China Nature Reserve are surface vector data, the rest are point vector data.
Statistical data: The statistical data include average annual rainfall (2018), average annual temperature (2018), number of days with wind and sand wind speed ≥10 m/s (2018), annual evaporation (2018), groundwater mineralization and groundwater burial depth. Meteorological data were mainly obtained from 22 meteorological stations around the Three Gorges Reservoir area by the Chinese Meteorological Science Data Sharing Service Available online: http://data.cma.cn/site/index.html (accessed on 17 January 2020). According to each weather station, the inverse distance interpolation (IDW) method of ArcGIS was used to convert meteorological statistics into a raster image of meteorological data in the study area. Groundwater mineralization and groundwater depth of burial data were obtained by querying the statistical yearbooks of different districts and counties in the Three Gorges Reservoir area.

Methods
In this paper, we use the overlay analysis function of GIS, Markov chain and cellular automata (CA) to simulate the future urban expansion. From the scenario of ecological conservation, predicting future urban expansion, estimating ESV during 2018-2030 and forecasting ESV losses at 2030 in the Three Gorges Reservoir area were performed. The research framework of this study is presented in Figure 2.

Ecological Comprehensive Sensitivity and Zone Identification
In this research, the comprehensive ecological sensitivity was calculated from three sensitivities of soil erosion, land desertification and soil salinization. The calculation formula is shown below: where EES denotes the comprehensive ecological sensitivity, and SS i , Z i and S i are the weight values of soil erosion, land desertification and soil salinization, respectively. The coefficient of variation method was used to calculate ω a , ω b and ω c as 49.63%, 32.00% and 18.37%, respectively [42]. R i , SG i , LS i and C i were selected for the evaluation of SS i . The ecological sensitivity factors were classified into five classes named insensitive, mildly sensitive, moderately sensitive, highly sensitive and extremely sensitive, which were categorized into classes 1, 2, 3, 4 and 5, respectively ( Table 1). The four indicators are calculated as follows: where SS i denotes the soil erosion sensitivity; R i is the rainfall erosivity factor; SG i is the soil type factor; LS i is the relative height factor; and C i is the normalized difference vegetation index factor. The ecological sensitivity factors were divided into five classes named insensitive, mildly sensitive, moderately sensitive, highly sensitive and extremely sensitive, which were categorized into classes 1, 2, 3, 4 and 5 (Table 1). The evaluation of the Z i required the I i , W i and SL i . Its classification and assignment methods were the same as those of SS i ( Table 2). The formula of Z i was as follows: where Z i denotes the land desertification sensitivity; I i is the dryness index factor; W i is the number of days on which wind-blown sand speeds are 6m/s; SL i is the slope factor. The evaluation of the S i required the E i , GS i , GD i and LUCC i . Its classification and assignment methods were the same as those of SS i ( Table 3). The formula of S i is as follows: where S i is the salinization sensitivity; E i is the evaporation index factor; GS i is the degree of mineralization of ground water factor; GD i is the groundwater depth factor; LUCC i is the land use and land cover change factor. The assessment results of ecological sensitivity were divided into 5 degrees, namely, insensitive, mildly sensitive, moderately sensitive, highly sensitive and extremely sensitive. The eleven factors required to calculate the spatial distributions of SS i , Z i and S i were presented in Figure 3. Based on the results of the SS i , Z i , S i and EES (Figure 4), using the natural breakpoint method in ArcGIS, the high sensitivity zone in the EES is proposed to determine the location of ecological sensitive zone.

CA-Markov Model
The CA-Markov model is composed of CA model, Markov chain and multi-criteria evaluation (MCE) [29]. The CA model is a discrete, finite state composition of the meta-cell model. It can simulate complex dynamic systems with spatial-temporal characteristics according to certain local rules [43]. Markov chains create the transfer matrix and probability between land use types for multiple time periods in the past through spatial comparison analysis, which are the basic data for predicting future land use patterns. MCE refers to the selection of expansion factors to construct a land use transition suitability image collection. CA-Markov model can effectively predict future land use dynamics [44]. The prediction process equation of the CA-Markov model is shown below [45]: where C(t j ) and C(t j+1 ) are the states of the cell at time t j and t j+1 , respectively; F is the transition rule; N is the domain of the cellular. In this study, elevation, slope, earthquake, landslide, highway, main road, railroad and population density were selected as the driving factors affecting urban expansion. Among them, elevation, slope, earthquake and landslide are negative indicators, and the rest are positive indicators. The suitability evaluation maps for earthquakes, landslides, highways, main roads and railroads were calculated by the kernel density tool, and finally, the normalized driver maps were obtained ( Figure 5). The Kappa coefficient was applied to check the accuracy of the CA-Markov model simulation results [46]. The formula for calculating the Kappa coefficient is shown below: where n is the total number of cell sizes in the raster; a 1 is the number of cell sizes in the real raster of the urban land; a 2 is the number of cell sizes in the non-urban land; b 1 is the number of cell sizes in the simulated raster of the urban land; b 2 is the number of cell sizes in the non-urban land; s is the number of cell sizes in the real raster and the simulated raster that correspond to each other.

Evaluation of ESV
Ecosystems provide the ecological products and services that humans need [47]. In order to estimate the value of ecosystems, basic value transfer, expert modified value transfer and spatial explicit function modeling were used [5,14,47]. Costanza et al. proposed to divide ecosystem services into 17 services and established a global value equivalence factor system to calculate the global ESV. This method has been widely accepted and used, but it cannot be directly applied to evaluate ESV in China [48]. This paper will be based on the recent research results of Xie et al. in Table 4, classify 6 types of ecosystems into the 4 following types of ecosystem services: provisioning services, regulating services, habitat services and entertainment services. A VPUA-based work has published that in China, and unit equivalent coefficients were estimated as USD 503.2 of ESV per ha [19]. For each ESV category, the calculation formula of ESV k is as follows: where ESV k is total value of ecosystem services; ESV i is the ESV for a particular primary service class (i); EC i,j is the equivalent coefficients of the secondary service class (j) in a particular primary service class (i); n is the four kinds of primary service class that include provisioning services, regulating services, habitat services and entertainment services; m is the total number of the secondary service class; Area j is the area of class (j) in a land type of a hectare and Area u is a hectare. Table 4 shows the equivalent coefficients table for ESV per unit area in China.

The Evaluation of Soil Erosion Sensitivity
Soil erosion is the serious damage to natural resources of water, soil and land productivity [49].The soil erosion sensitivity result was calculated by formula (2), as shown in Table 5 and Figure 4a. It shows a lower sensitivity at both ends and a higher sensitivity in the middle. The percentages of mildly sensitive zones and moderately sensitive zones are 30.00% and 22.99%, respectively. Next, the highly sensitive and insensitive zones accounted for 19.88% and 15.85%, respectively. Finally, the least percentage of the extremely sensitive zone is 11.28%, and it is mainly concentrated in Yunyang, Zigui, Fengjie, Kaizhou and Wanzhou.

The Evaluation of Land Desertification Sensitivity
Land desertification is generally defined as the loss of surface soil due to soil erosion, a reduction of agricultural land use value and ecological degradation [50]. In the spatial distribution of land desertification sensitivity (Figure 4b), extremely and highly sensitive zones occur mainly in Fengjie County, Wuxi County, Yunyang County and the upper part of Xingshan County. From the amount of land desertification sensitivity, as shown in Table 5, the moderately sensitive zone and the mildly sensitive zone accounted for the highest percentages of 24.06% and 22.96%, respectively. The extremely sensitive zone, highly sensitive zone and insensitive zone accounted for 18.40%, 17.80% and 16.79%, respectively.

The Evaluation of Soil Salinization Sensitivity
Soil salinization is a process in which salts from the soil substrate or groundwater rise to the surface, causing salts to accumulate in the surface soil after the water evaporates [51]. The spatial heterogeneity of the distribution of soil salinity sensitivity in the study area is shown in Figure 4c. The characteristics of the spatial distribution showed a gradual decrease from northeast to southwest. The higher sensitivity zones are mainly in Fengjie County, Wuxi County and Xingshan County. In the evaluation table of soil salinity sensitivity (Table 5), the highest percentages of the moderately and highly sensitive zones are 28.19% and 28.02%, respectively. The mildly sensitive and insensitive zones accounted for 25.10% and 9.75%, respectively. The least percentage was observed in the extremely sensitive zone at 8.94%.

The Comprehensive Evaluation and Zone Identification of Ecological Sensitivity
The comprehensive sensitivity of the ecological environment was calculated by formula (1), as shown in Table 5 and Figure 4d. In general, the highly sensitive zone is mainly concentrated in Wuxi, Yunyang and Fengjie. It is close to the Daba Mountain Nature Reserve and belongs to high forest cover areas with high terrain elevation [52]. The moderately sensitive zone is mainly concentrated in Zhong, Fengdu, Badong and Zigui.
The following zone of low sensitivity is mainly concentrated in the nine central urban areas of Chongqing, Fuling and Jiangjin. As shown in the comprehensive sensitivity evaluation table of the ecological environment (Table 5), the study area mainly showed moderate sensitivity (about 29.99%). The proportions of extremely sensitive, highly sensitive, low sensitive and insensitive zone are 10.91%, 24.95%, 26.47% and 7.68%, respectively.
In the identification of ecological sensitive zone, we use the spatial analyst tool in ArcGIS to extract the extremely and highly sensitive zone in the comprehensive ecological sensitivity, the vector data of Chinese nature reserves were converted into raster grid data, and then both were mosaicked. Finally, we obtained the spatial distribution map of the ecological key zone and the ecological core zone in the Three Gorges Reservoir area ( Figure 6).  (Figure 7c), using the Kappa coefficient formula to verify the accuracy, the Kappa coefficients for the three land types and the whole were calculated in Figure 8. The Kappa coefficients for nonurban, urban and water bodies are 86.73%, 83.14% and 92.19%, respectively, while the accuracy of the overall pattern was 86.74%, which indicated that the overall simulation accuracy was better [53], it can provide a basis for further simulation.

The Simulation of Urban Expansion Based on Ecological Sensitivity
This study will use the trained sample to predict the state of urban expansion in the Three Gorges Reservoir area in 2030, the urban expansion simulation map of the study area in 2030 (Figure 7e) and the area of urban land is about 226,594.43 ha. The focus is on the protection of ecological core zone and ecological key zone in the context of "Ecological Priority and Green Development" so that the urban expansion can be effectively adjusted. Using spatial overlay and analysis tools, the adjusted land use is shown in Figure 7f. Excluding the part of urban land that overlaps with the ecological core zone and ecological key zone, the area of this part of land is about 15,181.92 ha, and finally, the area of urban land is adjusted to 211,412.51 ha in 2030.

The Analysis of the Urban Expansion Simulation
The results of the 2030 urban land expansion were simulated based on ecological sensitivity in the Three Gorges Reservoir area, and we identified the development direction and area change of urban expansion in different districts in the study area. A map of the urban expansion change in the study area from 2018-2030 is shown in Figure 9, the faster growth in urban land is mainly occurring around the main urban areas of Chongqing. The statistical analysis yields a statistical map of the urban expansion area as shown in Figure 10, the northwestern part of Banan shows the most growth in urban land, about 9903.69 ha; followed by the northwestern part of Yubei, with an increase of 9400.41 ha; the growth area of Jiulongpo, Shapingba and Beibei are about 7448.76 ha, 7308.09 ha and 6133.68 ha, respectively; Yuzhong growth area is only 100.80 ha in urban land, and it is the only city-wide urban functional core area in Chongqing and past large-scale urban expansion has resulted in less available land for development and construction. Next, the smallest areas of urban land growth are in Xingshan, Zigui and Wushan. These areas are constrained by natural conditions, economic development and policy regulation, making them unsuitable for future urban expansion.

ESV Change from 2000 to 2018
The ESV was calculated by Formulas (9) and (10) and land use data as shown in Table 6, Figures 11 and 12.    Table 6, it can be found that the total ESV of the study area is USD 38. The results indicate that although the Three Gorges Reservoir area has been affected by urban expansion to some extent, the overall value has maintained an increasing trend, mainly due to the increasing area of broadleaved forests, which provide a large amount of ESV.  As shown in Figure 12, the four types of value in the study area showed an increasing trend from 2000 to 2018. We found that regulating services accounted for the highest proportion of overall ESV, about 79.88% (2000), 80.40% (2010) and 80.29% (2018), respectively. The proportion of each type of ecosystem service function had the following order of magnitude: regulating services > habitat services > provisioning services > cultural and amenity services.

Prediction of ESV Losses from 2018 to 2030
The CA-Markov model predicts the urban expansion pattern of the study area in 2030, and the results show that there will be a certain degree of outward urban expansion, mainly in the vicinity of the main city of Chongqing. The urban expansion will lead to a decrease in the value of ecosystem services. Our model suggests that the newer urban area is 80,026.02 ha from 2018 to 2030. In the ten land use types (Figure 13a), The most transferred land to urban areas is dry land and paddy field, about 33,903.54 ha and 28,381.23 ha, respectively, while the least transferred area is wetland, lake and river, about 6.48 ha and 61.38 ha, respectively. Table 7    The increasing urbanization will encroach on other ecological lands and directly lead to the losses of ESV by the equivalent coefficients. We found that the total ESV losses is USD 268.81 million in the Three Gorges Reservoir area (Figure 13b). The highest values of dry land and broadleaved forest were lost with USD 68.41 million (25.45%) and USD 58.17 million (21.64%), respectively. The variation of the overall ESV losses was explored from the district scales (Figure 14a). Jiulongpo, Banan, Shapingba and Yubei have higher losses (more than USD 25 million). This result indicates that these areas are at risk of ecological degradation. Yuzhong, Xingshan and Wuxi have the lowest losses, mainly due to the highest level of urbanization in Yuzhong, resulting in limited land for urban development. In addition, Xingshan and Wuxi are as important ecological reserves in China, which limit the large-scale expansion of urban land use. On average, Dadukou, Jiulongpo, Shapingba and Nanan have the highest unit ESV losses (more than 400 USD/ha) (Figure 14b).

The Rationality of Delimiting Ecological Sensitive Areas
First, because of the limitation of collecting data and the need to delimit ecological sensitive areas, according to the actual situation of the Three Gorges Reservoir area, this study selects relevant indicators from 2018 to identify the ecological sensitive areas in the study area in terms of three sensitivities (i.e., soil erosion sensitivity, land desertification sensitivity and soil salinity sensitivity) so as to identify the scope and number of ecological sensitive areas and provide a basis for achieving green development in the region [54]. As of now, the indicators and data will be further updated, which is a hot topic for future research.
Second, the grid size used to identify ecological sensitive areas is 30 m × 30 m in this paper. However, some of the raw data (NDVI, soil type, groundwater mineralization and groundwater depth) are of higher resolution, but still can accurately identify the extent of ecological sensitive zone. We believe that higher resolution data will be used in the future to improve the accuracy of delimiting ecological sensitive zones.
Third, in the comprehensive ecological sensitivity evaluation, the areas with high sensitivity or above account for 35.86% of the study area, with an area of 20,639.71 km 2 , mainly in Wushan, Fengjie and Yunyang. This part is a forest reserve with mainly mountainous terrain and high elevation, and its ecosystem types are diversified and spatially heterogeneous, which are vulnerable to some natural disasters and human factors. The harm to the ecological sensitive zone should be reduced. Meanwhile, a corresponding ecological compensation mechanism is established [55], the ecological advantages will be transformed into economic advantages, the ecological and economic benefits of the study area will be improved, and the green development concept of "making all-out efforts to protect it, and forbidding large-scale development" will be fully implemented.

The Accuracy of Simulated Urban Expansion
This study combines the ecological sensitivity, CA and Markov models in the context of green development to simulate and predict the future urban spatial patterns in the Three Gorges Reservoir area. The urban construction land area is predicted to increase by 80,026.02ha between 2018 and 2030, while the main expansion areas are in the northwestern part of Banan, the northwestern part of Yubei, Jiulongpo, Shapingba and Beibei. The analysis revealed that the future urban expansion in the study area is mainly occupying paddy fields and drylands, occupying an area of 33,903.54ha and 28,381.23ha, respectively.
The driving factors affecting urban expansion were mainly selected as elevation, slope, earthquake, landslide, highway, main road, railroad and population density. These factors performed relatively well in simulating urban expansion in 2018, with an overall kappa coefficient of 86.74% compared to the actual 2018 urban sites, but urban expansion is a complex dynamic system [56]. It is not enough to consider only the above eight driving factors. In the future, more socioeconomic and ecological data should be obtained to provide a basis for improving the accuracy of urban expansion simulation. At the same time, different prediction scenarios [57][58][59] and different grids sizes [33,60] play different roles in the simulation of urban expansion.

Estimation of ESV Dynamics and Prediction of Future ESV Losses as an Effective Tool for Ecological Safety Management
The estimation of ESV can provide a basis for the region to achieve sustainable development [61]. The ESV change in the study area were calculated by using the land use data from 2000-2018 and the ESV value equivalent table. The results show that the ESV in the Three Gorges Reservoir area increased from USD 38,320.49 million in 2000 to USD 41,964.75 million in 2018, and the overall ESV change showed an increasing trend. This may be due to the continuous increase in the area of forest land (especially broadleaf forest) due to the implementation of the reforestation project, which also brings an increase in the value of different ecosystem service functions [62,63]. However, it is noteworthy that between 2000 and 2018, there are 16 districts and counties with decreasing ESV per unit. The most significant decreases (more than 5000 USD/ha) were observed in Dadukou and Jiangbei. With increasing urbanization, we see that ESV will show a decreasing trend and ecosystems will face increasing pressure in the region. In the four different ecosystem services, regulating services provide the highest value, followed by habitat services, and finally provisioning services and entertainment services.
As urbanization continues, it may cause degradation of the ecosystem [64,65]. This study uses the CA-Markov model to predict the trend of urban expansion in 2030, and the study finds that the growth of urban land in the study area is mainly through the occupation of dry land and paddy fields. From 2018 to 2030, the overall ESV of the Three Gorges Reservoir area will be reduced by about USD 268.81 million. At the district and county scales, the ESV of Jiulongpo, Banan and Shapingba will be reduced most significantly, and the ecological pressure in the area will be further increased. In starting from the perspective of the four ESV, we found that regulating services values decreased the most. Therefore, it is necessary to combine the ecological sensitivity method and ESV evaluation results to achieve a more reasonable and targeted implementation of ecological spatial management and steadily realize the green development strategy.
This study provides insights into ecological conservation under sustainable urbanization by predicting change in urban land expansion and calculating ESV, using the Three Gorges Reservoir area as an example, which can be of guidance to other countries. Especially, developing countries are facing different problems caused by rapid urbanization, such as India [66], Sri Lanka [53] and Egypt [67]. Each country faces different ecological problems. We should consider geohazard sensitivity and select suitable indicators for the region where geohazards occur frequently. In ESV evaluation, there are differences in different countries' ecosystem value, and the equivalent coefficients table should be changed appropriately according to the regional characteristics. Although there are differences between the spatial scales of different countries, it is necessary to identify the ecological sensitive zones and to predict the future development direction of urban land and the dynamic change of ESV through scientific and rational identification. This quantitative evaluation provides a reference basis for ecological spatial control and effective macro-control.

Conclusions and Outlook
In the context of economic globalization, the urbanization process of human beings has become irreversible. Handling the relationship between environmental protection and economic development has become particularly critical. In this study, we simulate the future urban expansion and ESV change in the study area based on ecological sensitivity. Under the guidance of green development concept, the socioeconomic maximization is achieved, and the relationship between ecological protection and urban expansion is coordinated. This study shows that:

1.
In the comprehensive ecological sensitivity assessment, we found that the ecological sensitive zone is about 20,639.71 km 2 in the Three Gorges Reservoir area, accounting for 35.86% of the total study area. This part of the area is in Wushan, Fengjie and Yunyang.

2.
The results of the study show that the overall ESV in the Three Gorges Reservoir area showed an increasing trend from 2000 to 2018. The growth was about USD 3644.26 million. From the perspective of ESV change in districts and counties, we found that 16 districts and counties were decreasing in unit ESV, Dadukou and Jiangbei decreased most significantly. In four ecosystem services, regulating services provided the highest ESV.

3.
In the context of ecological priority and green development, the 2030 urban land was predicted and simulated. In 2018-2030, about 80,026.02 ha of new construction land will be added to the Three Gorges Reservoir area, and the overall ESV will lose USD 268.75 million. The largest losses are in Jiulongpo, Banan and Shapingba.
Although, this method can better delimit ecologically sensitive zones and estimate the dynamic changes of ESV in the Three Gorges Reservoir area, there is still much room for improvement. We will use high-resolution land use data and remote sensing inversion data (FVC, WET, NDBSI, LST and NPP) in the future work. This allows us to better identify zones of ecological sensitivity. Meanwhile, we should also make different simulations of urban expansion for different scenarios, such as developmental orientation or ecological orientation. It can provide better prediction and estimation of ESV changes for future urban expansion in the Three Gorges Reservoir area.