Spatial-Temporal Evolution of Ecosystem Service Value in Guilin, China from 2000 to 2020: A Dual-Scale Perspective

Abstract

Assessing land use-based changes in ecosystem service values (ESVs) is a beneficial approach for land resource planning and ecologically sustainable development. Located in the south of China, Guilin is one of the first Sustainable Development Goals (SDGs) Innovation Demonstration Zones set up by China. It is a typical ecotourism city with an important ecological and economic status. In recent years, the time series, model fit, and spatial scale of ESV assessment in Guilin have needed to be improved in the context of rapid urbanization and natural change. In this study, an improved ESV assessment methodology was utilized to incorporate the effects of biomass, soil conservation, and precipitation and to adjust the equivalence factors based on the ratio of geographic and environmental parameters to the national average to make them heterogeneous in time and space in improving the practical fit of the assessment results. The study analyzed the evolution of land use and its contribution to ESVs in Guilin from 2000 to 2020. County and 3 km × 3 km grid scales were combined to reveal both broad and detailed spatial and temporal characteristics of ESVs in Guilin. The results show that the expansion of building land in Guilin is notable, and the amount of land use transfer continues to increase. ESVs fluctuated in a lateral S-shape, with significant differences in ESV effectiveness between counties, consistently high ESVs near waterbodies and ecological reserves, and low ESVs near commercial and industrial land and cultivated land. Despite the recovery trend in ESVs in the later years, there is still a gap between 2020 and 2000. To a certain extent, it helps Guilin optimize land allocation from different perspectives and promote ecological improvement and resource planning optimization.

1. Introduction

In 2015, the United Nations adopted the 2030 Agenda for Sustainable Development, which outlines 17 Sustainable Development Goals (SDGs). These goals set standards for balanced and sustainable development across economic, social, and environmental dimensions over the next 15 years [1]. Ecosystem services provide the necessary support and security for human survival and development and are considered an essential foundation for realizing the SDGs [2,3,4]. Research on ecosystem services, as an emerging interdisciplinary field, lies at the intersection of ecology, economics, and sociology and has become a focus of contemporary research. An ecosystem service value (ESV) is defined as the value of direct or indirect benefits provided by ecosystems to human society and is a key indicator of regional ecological quality [5]. As a carrier of human agricultural production and the impacts of various activities, changes in land use patterns affect ecosystem services and are a significant driver of changes in ESVs [6,7,8]. Therefore, land use-oriented ESV research provides essential support for regional ecological conservation and development decisions and has attracted worldwide attention [9,10].

There was a growing recognition of the significance of ecosystem service values. The monetization of ecosystem services was first proposed by Costanza et al. in 1997, using a unit area value equivalent approach to assess ESVs on a global scale indirectly [11]. Since then, research related to ESVs has rapidly expanded worldwide. However, China is a developing country, and its economic and ecological conditions differ significantly from the rest of the world. Based on Costanza’s classification and assessment methodology system, Xie et al. [12] synthesized the results of China’s regional research, the state of understanding of ecological services by the Chinese people and decision-makers, the knowledge and experience of experts, and statistical data, and proposed an ecosystem service assessment system suitable for China’s actual situation after improvement [13,14], which has been widely applied.

Relevant research covered different administrative and geographical scales. In terms of spatial scales, relevant studies covered medium and large scales, such as national [15,16], regional [4,17], provincial, municipal [18,19], and counties [20,21], as well as geographical scales, such as watersheds [22,23,24], typical regions [25,26], and grids [27,28]. Yet, different scales have different characteristics, and a single scale often does not adequately reflect spatial information.

The same ecosystem service may have different values in different regions or groups, and generalized equivalence factors may not sufficiently account for such differences [29]. Some scholars attempted to correct the equivalent factor using the same correction factor for the same time frame in the study area. Cai et al. [30] quantitatively assessed the value of ecosystem services in the Baiyangdian region in the context of ecological water diversion projects (EWDPs) using an optimized appraisal system of Net Primary Production (NPP) data and found that EWDPs increased the watershed area and thus the value of regulation services. Li et al. [31] improved the assessment model by considering the effects of NPP, precipitation, and population level to predict the land use pattern in the Sichuan Plateau region in 2030. The use of the same correction coefficient, based on the overall situation in the study area, increases to some extent the practical applicability of the results of the regional assessments. However, the equivalence coefficients could be corrected even further regarding temporal and spatial precision to make them more comparable and realistic across regions. Therefore, adjusting the equivalence factors according to the specific characteristics of the local environment is significant.

Guilin, a well-known eco-tourism city in Guangxi Province in southern China, is one of the first Sustainable Development Goals Innovation Demonstration Zones set up by China, with the theme of “Sustainable Use of Landscape Resources”. The area is rich in ecological landscape resources, such as the Lijiang River, which is listed on UNESCO’s World Natural Heritage list, and Mao’er Mountain, which is listed as a World Biosphere Reserve. Ecology is the keystone of Guilin’s tourism development, which drives the city’s economic growth. However, Guilin is located in an underdeveloped region of China with a weak economy. In recent years, against the background of rapid urbanization and natural changes, Guilin has faced severe challenges, such as high land use intensity and impaired ecosystem service functions. Therefore, it is of great practical significance to assess the characteristics of spatial and temporal changes in the ESVs in Guilin and to analyze its development strategy in the context of regional analysis.

The current study of ESVs in Guilin is relatively rough regarding time series, methodological precision, and scale. Zhao et al. [32] used object-oriented land use classification and ecosystem service value assessment methodology to study the land use change and the growth trend of ecosystem service values in Guilin in 2002, 2012, and 2017 and found that Guilin’s woodland ecosystem contributes the most, providing an ecosystem service function that far exceeds the production function. Li et al. [33] explored the effects of anthropogenic drivers on the value of ecosystem services in the Lijiang River Basin in Guilin from 1995 to 2015. Related studies have explored the ecosystem service value of Guilin. Still, the time series are short, applying uncorrected Chinese regional coefficients and lacking the assessment of county-level administrative units, which reduces the scientific validity and accuracy of our current understanding of Guilin’s ESVs.

This paper analyzed the land composition of Guilin and investigated the evolutionary characteristics of regional land use by using the land movement attitude and land transfer matrix. Using an assessment model that was more in line with the characteristics of the regional environment, a dual-scale perspective was used to assess ESVs in the study area over 20 years. The county scale explored the county’s broader, macroscopic spatial and temporal evolution and identified county ecological efficacy. The 3 km × 3 km grid scale revealed detailed, microscopic trends in the spatial distribution of ESVs, revealing the dynamics of ecological highlights and areas of vulnerability, and applying the Moran’ I index, LISA explored global and local aggregation.

Based on the above study framework, we mainly performed the following: (1) we extended the time series of existing regional ESV studies; (2) we adopted an improved ESV assessment methodology that incorporated the effects of biomass, soil retention, and precipitation changes, and adjusted the equivalence factors according to the ratio of actual regional environmental parameters to the national average parameters, making them heterogeneous in time and space, and improving the comparability and practical applicability of the results; (3) we supplemented the spatial-scale study, analyzing the spatial and temporal evolution of the dual-scale perspective of county and grid ESVs, which overall provides a dual-scale perspective that can analyze the effectiveness of ecological governance in each county at the macro level, identify high- and low-efficiency regions, and formulate and adjust macroscopic policies, as well as accurately identify ecological priority regions at the micro level to guide targeted governance measures.

Providing data reference for SDG 15 (Protection and Restoration of Terrestrial Ecosystems) in Guilin, which is of great significance to the green and high-quality development of the region and the enhancement of people’s well-being. In part, it provides information support for SDG 17 (partnerships for realizing the SDGs), which is a reference for equally underdeveloped and ecologically resource-rich tourist-oriented cities.

2. Materials and Methods

2.1. Study Area

Figure 1 presents the general schematic map of the study area. Guilin is located in southern China and northeastern Guangxi, with a longitude of 109°36′ to 111°29′ East and a latitude of 24°15′ to 26°23′ North, covering a total area of about 2690 square kilometers. The city consists of the Guilin Municipal District in the central area and 13 surrounding counties. The central area has a flat topography with more plains, is densely populated, and is surrounded by mountains. The overall elevation in the region is low, high in the west, north, and southeast, and lower in the center. Guilin has a subtropical monsoon climate, with a mild climate, long summers and short winters, and abundant rainfall. The population increased from 4.46 million in 2000 to 4.95 million in 2023, and the Gross Domestic Product (GDP) increased from CNY 30.3 billion to CNY 252.3 billion in 2019.

Guilin is rich in ecological landscape resources, such as the Lijiang River, which is listed on UNESCO’s World Natural Heritage List, and Mao’er Mountain, which is listed on the World Biosphere Reserve. It is also an ancient capital with a ten-thousand-year history, and many well-known Chinese literary works and paintings have been associated with it. The unique karst landscape, high-quality soil and water environment, and extended human history give Guilin great potential for sustainable development with significant ecological and economic values [34]. However, Guilin is located in an underdeveloped, multi-ethnic region of China with relatively weak economic strength, and at the same time, faces serious challenges such as high land use intensity and impaired ecosystem service functions in the process of urbanization and industrialization.

2.2. Data Sources

The land use datasets of the study area were extracted from the Land-Use and Land Cover of China (CNLUCC) database [35] in 2000, 2005, 2010, 2015, and 2020, with a spatial resolution of 30 m.

Table 1. Data types and sources.

Data Type	Data Name	Spatial Resolution	Source	Data Provision
Land use	CNLUCC	30 m	Landsat Series, CBERS-1	Resource and Environment Science and Data Center (http://www.resdc.cn (accessed on 18 October 2023))
NPP	MOD17A3H database	500 m	Terra	The database in GEE (https://doi.org/10.5067/MODIS/MOD13A2.061 (accessed on 30 October 2023))
NDVI	MOD13A2	1000 m	Terra	The database in GEE (https://doi.org/10.5067/MODIS/MOD17A3HGF.061 (accessed on 30 October 2023))
Precipitation	CHIRPS Daily	0.05°	multiple satellites and meteorological stations	The database in GEE (https://chc.ucsb.edu/data/chirps (accessed on 30 October 2023))

shows the data information.

The dataset is based on multi-source Landsat images such as TM, ETM, OLI, and China–Brazil Earth Resources Satellite 1 (CBERS-1) data. After being processed and validated through various correction methods, it provides a land use dataset for the China region with an accuracy of over 90%. This dataset has been widely used in investigations and research in the fields of land resources, ecology, and hydrology in China. With reference to the first-level classification of China’s national standard classification of land use, land use types were reclassified into six categories: cultivated land, forestland, grassland, waterbody, building land, and unused land. Specific classification information can be found in Appendix A,

Table A1. Land use structure and proportion in Guilin.

Reclassified Land Types	Subclasses
Cultivated Land	11 Paddy field, 12 Dry land
Forestland	21 Forest, 22 Forest, 23 Sparse woods, 24 Other forestland
Grassland	31 High-covered grassland, 32 Medium-covered grassland, 33 Low-covered grassland
Waterbody	41 Rivers and canals, 42 Lakes, 43 Reservoir ponds, 46 Beach
Building land	51 Urban residential areas, 52 Rural residential areas, 53 Other building land
Unused Land	63 Saline land, 64 Marshland, 65 Bare land, 66 Bare rock, 67 Other unused land

Note: The number is CNLUCC code.

. The data of grain output, planting area, and grain unit price in Guilin used in this study were derived from the Guilin Economic and Social Statistical Yearbook (https://tjj.guilin.gov.cn/ (accessed on 6 November 2023)) of each year and the China Statistical Yearbook (https://www.stats.gov.cn/sj/ndsj/ (accessed on 6 November 2023)).

Google Earth Engine is an online cloud platform provided by Google for visualizing, analyzing, and processing Earth science data on a global scale. It combines a large number of remotely sensed data and geospatial datasets with planetary-level analytics capabilities. In this study, the GEE platform was used to obtain data on net primary productivity (NPP), normalized difference vegetation index (NDVI), and precipitation for the study area and the country as a whole (Table 1), and these data were used to calculate three spatial and temporal correction coefficients—EF_NPP, EF_FVC, EF_P for each year.

2.3. Methodology

The overall technical flow of the study is shown in Figure 2.

2.4. Single Land Use Dynamic Degree

Single land use dynamic degree can be used to describe the degree of change in the land use structure in a particular study area, which is essentially a study of the change in the area of a particular land use type over a certain period of time [36]. The formula is as follows:

$$D_s={\displaystyle \frac{S_2-S_1}{S_{1}T}}\times 100\mathrm{\%}$$

(1)

S₁ is the area of land use type at the beginning of the study; S₂ is the area at the end of the study; T is the duration of the study cycle in years. Of these, D_s is the dynamic degrees of single land types. If D_s > 0, the area of the land use type is increasing. If D_s < 0, the area of the land use type is decreasing. The greater the absolute value of D_s, the greater the range of variation in land area.

2.5. Land Use Transfer Matrix

The land use transfer matrix can characterize the evolution of the land use structure before and after a certain period of time [37]. It clearly depicts the direction of shifts between different land use types and the specific size of the area. In practice, it can be obtained from the intersection of land use maps. The specific mathematical expression is as follows:

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

(2)

where S_ij is the area (km²) of land use type i converted to another land use type j from the beginning to the end of the study period. i (i = 1, 2, …, n) and j (j = 1, 2, …, n) are the land use types at the beginning and end of the study, respectively. n is the number of land use types.

2.6. Spatiotemporal Correction Coefficient

Ecosystems undergo continuous changes in their internal structure and external morphology across different periods and spatial scales. Therefore, their ecological service functions and values change continuously over time and space. Previous researches indicate that biomass significantly influences ecosystem functions such as nutrient cycling, atmospheric regulation, raw material and food production, and biodiversity [38]. Soil conservation is closely related to precipitation, topographic gradient, soil properties, and vegetation cover [39]. Precipitation changes are related to hydrologic regulation and water supplies [40]. This study set Net Primary Productivity (NPP), Fractional Vegetation Cover (FVC), and precipitation as spatiotemporal correction coefficients, adjusting the equivalence factor based on previous work. Adjustment of the equivalence factors based on the ratio of the actual regional environmental values to the national average values and adjustment of the equivalence factors for the table of equivalence factors for ecosystem service values (ESVs) of Guilin before spatiotemporal correction (

Table 2. EF_NPP, EF_FVC, and EF_P correction coefficients in Guilin from 2000 to 2020.

County	EFNPP					EFFVC					EFP
	2000	2005	2010	2015	2020	2000	2005	2010	2015	2020	2000	2005	2010	2015	2020
Gongcheng	2.05	1.88	1.92	1.80	1.89	0.74	0.68	0.72	0.67	0.72	2.88	2.86	2.95	3.88	3.01
Guanyang	1.91	1.64	1.71	1.71	1.67	0.72	0.67	0.70	0.66	0.66	3.00	2.70	2.90	3.65	2.95
Municipal district	1.86	1.70	1.71	1.73	1.74	0.65	0.58	0.61	0.60	0.62	3.42	2.94	3.02	4.35	3.51
Lingchuan	2.12	1.86	1.89	1.81	1.82	0.75	0.68	0.71	0.66	0.67	3.40	2.93	3.08	4.21	3.45
Lingui	1.95	1.71	1.79	1.70	1.70	0.73	0.65	0.67	0.63	066	3.51	2.93	3.01	4.32	3.68
Lipu	2.17	1.96	2.07	1.89	2.04	0.74	0.67	0.70	0.64	0.70	2.74	2.72	2.82	3.65	2.74
Longsheng	2.14	1.85	1.98	1.87	1.74	0.79	0.71	0.75	0.73	0.68	3.14	2.57	2.73	3.66	3.10
Pingle	2.17	2.00	2.06	1.95	2.06	0.72	0.66	0.69	0.66	0.74	2.57	2.77	2.84	3.48	2.70
Quanzhou	1.94	1.64	1.72	1.81	1.70	0.71	0.65	0.69	0.67	0.68	2.79	2.34	2.60	3.23	2.71
Xing’an	2.11	1.78	1.83	1.80	1.78	0.74	0.67	0.70	0.66	0.64	3.22	2.78	2.98	3.95	3.18
Yangshuo	2.14	1.96	2.08	1.95	2.04	0.73	0.66	0.70	0.63	0.70	3.12	2.86	2.85	3.91	3.06
Yongfu	2.10	1.78	1.93	1.81	1.86	0.77	0.67	0.72	0.65	0.70	3.67	3.11	3.07	4.43	3.85
Ziyuan	2.07	1.73	1.84	1.77	1.65	0.75	0.67	0.74	0.68	0.67	2.92	2.39	2.68	3.33	2.67

) in Section 2.6 was carried out. To capture the dynamic heterogeneity of ecosystem service functions in time and space and to enhance comparability among the interior of the region, and taking into account the resolution of the data sources (Table 1) and the number of calculations, the spatial and temporal correction coefficients for each county were calculated in this study. The formulas are as follows:

• NPP Spatiotemporal Correction Coefficient:

$${EF}_{NPP}={\displaystyle \frac{{NPP}_i}{{NPP}_C}}$$

(3)

where EF_NPP represents the correction factor for net primary productivity. NPP_i is the mean value of NPP for the year in the study site, and NPP_c is the mean value of NPP for the year in China. The EF_NPP was used to adjust nine categories of ecosystem services: food production, raw materials, gas regulation, climate regulation, environment purification, nutrient cycling, biodiversity conservation, and cultural services.

2.

FVC Spatiotemporal Correction Coefficient:

$${EF}_{FVC}={\displaystyle \frac{{FVC}_i}{{FVC}_c}}$$

(4)

$$FVC={\displaystyle \frac{NDVI-{NDVI}_{min}}{{NDVI}_{max}-{NDVI}_{min}}}$$

(5)

where EF_FVC represents the correction factor for fractional vegetation cover. FVC_i is the mean value of FVC for the year in the study site, and FVC_c is the mean value of FVC for the year in China. NDVI is the normalized difference vegetation index. EF_FVC was used to amend the soil conservation service function.

3.

Precipitation Spatiotemporal Correction Coefficient:

$${EF}_P={\displaystyle \frac{P_i}{P_c}}$$

(6)

where EF_P represents the correction factor for precipitation. P_i and P_c are the average annual precipitation for the study site and for the country as a whole. The EF_P was used to correct hydrological regulation and water supply service functions.

The three calculated correction coefficients are shown in Table 2.

2.7. Calculation of Ecosystem Service Value

Xie et al. [14] proposed a unit-area equivalent factor method suitable for China’s scenario based on the indirect valuation of ecosystem services based on the market theory proposed by Costanza et al. [41], taking into account China’s socio-economic conditions and ecosystems based on the results of China’s regional research, Chinese people’s and decision-makers understanding of ecological services, experts’ knowledge and experience, and statistical data. The equivalence factor method for China’s national scenario is proposed. The method facilitates data acquisition, enables easy statistical analysis and cross-comparison, and finds widespread application in China. This study employed an improved method of equivalence factor per unit area to assess the ESVs in Guilin. Correction of the equivalence coefficient for the ESVs in Guilin is based on the ratio of the actual regional environmental value to the national average value. Each image element was modeled based on the land use type to which it belongs, using equivalent coefficients calibrated within the region based on actual time and environmental characteristics, improving the internal comparability and practical application of the results.

The value of a standard equivalence factor is defined as the market economic value of the average yield of natural grain per 1 hectare per year in the study area. The standard equivalence factor value is calculated by multiplying the natural grain yield per unit area by the cultivation area. Meanwhile, consider the combined ratio of planting costs to benefits. The formula is as follows:

$$D={\displaystyle \frac{1}{7}}\times P\times Q$$

(7)

where D is the standard equivalence factor, and 1/7 is the ratio of food benefits to costs estimated by Xie. P is the natural grain yield per unit area, obtained by dividing the yield by the area of cultivated land. Q is the price of the grain.

Checking the grain production and cultivated land area in the Guilin Economic and Social Statistical Yearbook for each of the five years from 2000 to 2020, the average value of grain production per unit area in Guilin was calculated to be 5233.55 kg/hm² over the 20 years. Due to the annual food prices being affected by social demand, policy regulation, and many other factors, in order to increase the comparability of the data and reduce the error, the average price of food in 2020 was taken to be 2.52 CNY/kg (Data from China Statistical Yearbook 2021). Substituting into the equation yields a value of 1884.08 CNY/hm² for an equivalence factor in Guilin.

The land use types in the equivalent coefficients table for ESVs differed from the reclassification criteria of this study using the six categories. Therefore, this paper adjusted the table according to the characteristics of land use types and the actual situation in Guilin. Checking the yearbook, the ratio of paddy field to dry land in Guilin is about 7:3; the ratio of forest land mainly composed of coniferous forest, broad-leaved forest, and shrub is about 4:4:2; grassland took the average value of scrub and meadow; there is no glacier snow in the study area, so the waterbody corresponded to the water system; and the unused land equivalence factor corresponded to the value coefficient of the desert. In addition, only natural ecosystem service value was considered in this paper, so the building land ecosystem service equivalence factor was assigned as 0 [42,43].

The formula for calculating ESV is as follows:

$$ESV=D\times {\sum }_f{\sum }_k\left(V_{kf}^{EF}\times A_k\right)$$

(8)

where ESV is the total ecosystem service value of the area (CNY). A_k is the area of land use category k in the area (hectare); $V_{kf}^{EF}$ is the equivalence factor for the ecosystem service function of land in category f of land use type k corrected by the spatial and temporal correction coefficient EF; D is the value of a standard equivalence factor.

After organizing the above rules, the Guilin equivalence factor value of 1884.08 CNY/hm² was multiplied by the merged

Chinese equivalent coefficients table for the ESV. The equivalent coefficients table for ESV per unit area in Guilin before spatiotemporal correction was calculated (

Table 3. The equivalent coefficients table for ecosystem service value (ESV) in Guilin before spatiotemporal correction (CNY·hm⁻²·a⁻¹).

Primary Ecosystem Service Types	Secondary Ecosystem Service Types	Cultivated Land	Forestland	Grassland	Waterbody	Unused Land
Provisioning	Food production	2274.08	455.95	565.22	1507.26	18.84
	Raw materials	344.79	1051.32	838.42	433.34	56.52
	Water supply	−3457.29	542.62	461.60	15,619.02	37.68
Regulating	Gas regulation	1842.63	3447.87	2929.74	1450.74	207.25
	Climate regulation	955.23	10,313.45	7752.99	4314.54	188.41
	Environment purification	280.73	3059.75	2562.34	10,456.64	584.06
	Hydrological regulation	3739.90	7351.68	5680.50	192,628.34	395.66
Supporting	Soil protection	595.37	4197.73	3570.33	1752.19	244.93
	Nutrient cycling	318.41	320.30	273.19	131.89	18.84
	Biodiversity conservation	350.44	3824.68	3250.03	4804.40	226.09
Culture	Entertainment culture	152.61	1676.83	1431.90	3560.91	94.20

).

2.8. Coefficient Sensitivity Assessment

The coefficient of sensitivity (CS) was used to assess the uncertainty of the ESV results [44,45,46]. The formula is as follows:

$$CS=\left|{\displaystyle \frac{\left({ESV}_j-{ESV}_i\right)/{ESV}_i}{\left(V_{jk}-V_{ik}\right)/V_{ik}}}\right|$$

(9)

where CS is the coefficient of sensitivity. The ecological service value equivalent coefficients for each land type were calculated by increasing the coefficients by 50%, where ESV_i and ESV_j are the ESV before and after adjustment, respectively. V_ik and V_jk are the equivalence coefficients for the land use type k before and after adjustment. If CS > 1, it indicates that the ESV is elastic for the equivalence coefficients, and the assessment results are low in credibility. If CS < 1, the results are credible.

2.9. Exploratory Spatial Data Analysis

2.9.1. Global Spatial Autocorrelation

The global spatial autocorrelation measure of ESV in Guilin under the 3 km × 3 km grid was calculated using Moran’s I index with the following formula:

$$Moran’sI={\displaystyle \frac{n{\sum }_{i=1}^n{\sum }_{i=1}^n{w}_{ij}\left(x_i-x\right)\left(x_j-x\right)}{{\sum }_{i=1}^n\left(x_i-x\right){\sum }_{i=1}^n{\sum }_{j=1}^n{w}_{ij}}}$$

(10)

where n is the number of grids; x_i and x_j are the attribute values of grids i and j; x is the attribute mean value; and w_ij is the spatial weight matrix. The inverse-distance weights matrix was used in this study. If Moran’s I > 0, it indicates a spatial positive correlation of ESV. If Moran’s I < 0, it indicates the presence of spatial negative correlation.

2.9.2. Local Indicators of Spatial Association

Local indicators of spatial association (LISA) reveal the local clustering characteristics of spatial unit attributes by analyzing the degree of difference between the spatial unit and the surroundings and the significance [47], which is calculated as follows:

$$I_I=z_i{\sum }_j{w}_{ij}{z}_j$$

(11)

$$z_i={\displaystyle \frac{n{\left(x_i-x\right)}^2}{{\sum }_i\left(x_i-x\right)}}$$

(12)

$$z_j=\left(x_j-x\right)$$

(13)

where z_i and z_j are the normalization of the observations on grids i and j, respectively.

3. Results

3.1. Land Use Change in Guilin

3.1.1. Land Use Status and Dynamics

The land use composition of Guilin is shown in

Table 4. Land use composition in Guilin from 2000 to 2020.

Year	Cultivated Land		Forestland		Grassland		Waterbody		Building Land		Unused Land
	km2	%	km2	%	km2	%	km2	%	km2	%	km2	%
2000	5361.14	19.49	17700.98	64.35	3735.86	13.58	260.08	0.95	448.77	1.63	2.12	0.01
2005	5357.92	19.48	17696.74	64.33	3729.02	13.56	261.54	0.95	461.70	1.68	2.13	0.01
2010	5342.28	19.42	17709.77	64.38	3708.59	13.48	272.35	0.99	473.15	1.72	3.11	0.01
2015	5319.10	19.34	17680.77	64.27	3708.23	13.48	278.15	1.01	517.91	1.88	4.87	0.02
2020	5255.74	19.11	17636.68	64.11	3687.86	13.41	294.04	1.07	629.87	2.29	3.80	0.01

, and the spatial distribution is shown in Figure 3. Between 2000 and 2020, the spatial distribution of different land use types in Guilin exhibited distinct regional patterns and demonstrated significant trends in their evolution.

Forestland was the most widely distributed (nearly 17,600 km², 64% of the area), mainly in the form of patches and surfaces in the northwestern, southwestern, and eastern zones of Guilin, and was widely distributed in areas such as Yongfu County, Ziyuan County, and the northern and southern parts of Lingchuan County. Cultivated land (nearly 5300 km², 20%) was distributed in a narrow north–south belt, mainly concentrated in the northeast, southeast, and central zones. Grassland (nearly 3700 km², 23%) distribution was relatively sporadic, with overall spatial integration low. The spatial distribution of building land and waterbodies shows a clear expansion trend, with waterbody reaching 294 km² in 2020 and building land being particularly prominent, reaching 630 km² in 2020.

From Figure 4, we can learn about the attitude of land use dynamics in Guilin from 2000 to 2020. Trends in the area of land types in Guilin have varied over the 20-year period.

The dynamic degree of a single land type of cultivated land was negative during the period 2000–2020, and the area continued to decrease with increasing absolute values, reaching a maximum in 2015–2020 (−0.24%). Forest land had a negative dynamic degree except for 2005–2010, when it declined slightly overall. Grassland dynamic degree was negative throughout the study period, with a continuous decrease in area and small absolute values, all less than 0.11%. The dynamic degree of building land increases year by year, and the total dynamic degree reaches 40.38%. The dynamic degree of waterbody is more than 1%. The unused land fluctuates during the study period, first increasing and then decreasing. Trends in the area of land types in Guilin have varied over the last 20 years.

3.1.2. Land Use Transfer Distribution and Quantity

Figure 5 illustrates the distribution of land use transitions from 2000 to 2020 across different research periods (a) and the quantities of those transitions (b). Between 2000 and 2020, forestland was mainly converted into cultivated land and grassland, concentrated in the northeastern and southern regions of Guilin. Cultivated land was primarily transformed into forestland and building land, focusing on urban centers and northeastern and southern regions. Grassland was primarily converted to forest and cultivated land. In contrast, the inflow of waterbody and building land exceeded the outflow. The waterbody primarily originated from the conversion of cultivated land and forestland. Building land was mainly converted from cultivated land, with a notable increase in the net transfer area, especially in Guilin’s urban core and commercial zones, showing expansion towards the city center.

During the study period, land sales of all types rose, and the building land area expanded, notably in the northwestern part of the municipal area and the eastern part of Lincoln County. Waterbodies increased slightly. Although the flow trend to ecological land can be seen, the encroachment of building land shows some problems.

3.2. Spatial and Temporal Characteristics of ESV

3.2.1. Characterization of Temporal Changes in ESV

Based on the land use results, we used the improved equivalence factor method to calculate the ESV situation in Guilin during the period of 2000–2020, and the results are shown in Figure 6: (a) is the total ESV, (b) is the ESV of each land use type, and (c) is the ESV of each ecosystem’s secondary service function. The sensitivity assessment results show that all CSs are <1, and the results are credible; the detailed results are shown in Appendix A,

Table A2. The coefficient of the sensitivity (CS) of different land use types from 2000 to 2020.

Year	Cultivated Land	Forestland	Grassland	Waterbody	Building Land	Unused Land
2000	0.0465	0.753	0.129	0.0722	0	0.00000515
2005	0.0465	0.753	0.128	0.0726	0	0.00000518
2010	0.0462	0.751	0.127	0.0754	0	0.00000754
2015	0.0460	0.750	0.127	0.0770	0	0.00001180
2020	0.0454	0.747	0.126	0.0813	0	0.00000922

.

In terms of total volume, the total ESV volume from 2000 to 2020 shows a horizontal S-shape. The total ESV volume in Guilin was about CNY 176.24 billion in 2000 and then dropped to CNY 154.72 billion in 2005, reaching its lowest value; it continued to grow from 2005 to 2015, reaching its highest value (CNY 178.72 billion) in 2015, while it retraced its steps in 2020, reaching a value of CNY 164.51 billion in 2020. And by 2020, ESV declined to CNY 164.51 billion. Overall, the value of environmental services in Guilin did not show a single upward or downward trend over the 20 years, but rather significant fluctuations.

Regarding land use types, forest land has the highest value, accounting for about 70% of the total value of all land use types. However, there is a slight overall downward trend in this contribution. It showed that Guilin’s forestland plays a great role in the ecosystem and generates high economic value. The next most important types were grassland and waterbodies, both of which accounted for similar proportions, about one-fifth of the total value, and provided a high economic value. The share of value contributed by grassland decreased minimally from year to year, while the share of ESVs provided by waterbodies increased overall. Forest land, grassland, and waterbodies were crucial for enhancing the total ESV of Guilin and sustaining its ecological service value. ESVs provided by unused land were the lowest among land use types. The highest ESVs across various land use types were all recorded between 2010 and 2015.

In terms of ecosystem service functions, hydrological regulation services ranked first among all service values, increasing from 36.85% in 2000 to 43.25% in 2020. It demonstrated consistent growth across the four periods, highlighting its importance to the ecosystem [48]. Secondly, climate regulation, biodiversity conservation, and gas regulation services provided relatively high values, with some fluctuations observed. These services experienced a slight increase from 2015 to 2020, but overall, they exhibited a declining trend. The water supply service recorded negative values and reached its lowest value of CNY −626 million in 2015. During the period from 2000 to 2005, all ESVs exhibited a declining trend. From 2005 to 2010, all functions showed an upward trend, except for a slight decline in water supply ESV. Between 2010 and 2015, all functions, except for water supply and hydrological regulation services, showed a declining trend. From 2015 to 2020, all functions, except for soil conservation services, exhibited a slight decline.

3.2.2. Spatial and Temporal Evolution of ESV at the County Scale

The total ESV is closely related to the size of the study area in order to eliminate the influence of the area factor in the calculation of ESVs and more intuitively reflect the level of ESVs of each county during the study period, in addition to calculating the total ESV, we also calculated the per-unit ESV per 1 km² in each county, which is expressed as ESVa. The total ESV and per-unit ESV (ESVa) for each county in Guilin during the period from 2000 to 2020 are shown in Figure 7a,b.

In terms of total volume, the distribution of total ESV rankings by county was generally stable, with the highest being Allstate County at CNY 22.83 billion and the lowest being the Municipal District at CNY 2.61 billion. The highest ESV growth rates over the 20-year period were in the Municipal District (22.46%) and Lingchuan County (13.56%) during the 2010–2015 period. The lowest growth rates were in Ziyuan County (−16.67%) and Yongfu County (−15.13%) during the 2000–2005 period. Regions with significant fluctuations in the rate of change were mainly Municipal District, Quanzhou, and Lingchuan, while those with moderate fluctuations were Pingle, Lipu, and Gongcheng Yao Autonomous Region.

Boundaries were established at significant numerical breakpoints using the natural breakpoint method, which grouped similar values to maximize inter-group differences. The data were categorized into five value zones: high, higher, medium, lower, and low. It effectively represented the spatial variation characteristics of ESVa across different counties [49], as shown in Figure 8.

As can be seen in Figure 7b, ESVa in the counties showed an overall horizontal s-shape, with more marked fluctuations than the total ESV. Lingchuan County was a long-term leader in ESVa, peaking at 7.82 million CNY/km² in 2015. However, it was volatile, showing that its high potential coexists with instability. Municipal Districts, on the contrary, continued to be the lowest, reaching a minimum value of 4.5 million CNY/km² in 2005. On the contrary, Pingle County and Gongcheng Yao Autonomous Region had small fluctuations in ESVa.

ESVa in Guilin exhibited higher values in the north and west and lower values in the south and east. High-value areas, concentrated in regions like Lingchuan, Xing’an, and the Multinational Autonomous County of Longsheng, had extensive forest cover (about 70%) and favorable ecological conditions, such as mountain ranges and well-developed water systems. By 2015, high-value areas expanded to include Yangshuo and Gongcheng, indicating more balanced ecological development. In contrast, low-value zones were found in the urban district and the low-lying areas of Jinzhou and Guanyang, where urbanization, intensive land use, and limited forest cover reduced ecosystem service value.

3.2.3. Spatial and Temporal Evolution of ESV at the Grid Scale

In terms of the grid scale, in reference to relevant literature and considering the size and spatial characteristics of the study area, as well as the applicability of the results, the 3 km × 3 km grid was constructed [50,51,52]. On this basis, the ESV of each grid was calculated. Meanwhile, the natural breakpoint method was combined with the distribution of ESV, and the ESV was divided into five classes: high, higher, medium, lower, and low, to obtain the spatial distribution pattern of ESV in Guilin in 20 years, as shown in Figure 9.

In terms of the overall pattern, the ESV in Guilin exhibited a spatial distribution with higher values in the north and west and lower values in the south and east. The city also demonstrated a phenomenon of urban “hollowing,” with values increasing towards the periphery. The areas with relatively higher ESV were concentrated in dense woodlands and waterbodies. The ESV values in the ecological reserve sites, such as the areas along the Lijiang River and its tributaries, Mao’er Mountain in Xing’an County, Qingshitan Reservoir in Lingchuan County, and the Yulong River in Yangshuo County, were all consistently high. The areas with low ESV values were mainly located around the city center, Quanzhou County, and Gongcheng Yao Autonomous County, with denser industrial and residential areas.

ESV in Guilin declined significantly from 2000 to 2005, reaching its lowest point, with high-value areas contracting, particularly around Lingchuan and Yongfu, and ecological stability in the northwest dropping. From 2005 to 2015, ESV gradually increased, especially along the Lijiang River, peaking in 2015. However, from 2015 to 2020, ESV generally declined again, with a shift towards the southwest, highlighting the negative ecological impacts of land conversion for construction.

3.2.4. Spatial Autocorrelation Analysis of ESV in Guilin

The global Moran’s I index values for Guilin from 2000 to 2020 can be found in Appendix A,

Table A3. Moran’s I index of the spatial distribution of ESV in Guilin.

Parameter	2000	2005	2010	2015	2020
Moran’s I	0.477	0.469	0.457	0.438	0.453
p	0.001	0.001	0.001	0.001	0.001
z	38.3481	37.8110	36.9244	35.3083	35.7871

. It was observed that Moran’s I was >0.43 for all 20 years, 0.477, 0.469, 0.457, 0.438, and 0.453, respectively, indicating that there was spatial clustering of ESVs during the study period in Guilin. Z were 38.3481, 37.8110, 36.9244, 35.3083, and 35.7871, respectively, which far exceeded the critical values, indicating that the spatial aggregation of the data was very significant. All five p-values were <0.001, much less than 0.005, further confirming the significance of the spatial aggregation phenomenon, which indicated that the ESV in the study area had very significant spatial aggregation, with separated aggregation in the high- and low-value areas. Overall, spatial aggregation in Guilin was high but showed a continuous downward trend, except for a slight rebound in 2015–2020.

The global Moran’s I only revealed the type of agglomeration in the study area as a whole, and the autocorrelation between ESV domains in Guilin could be analyzed by using the local indicators of spatial association (LISA). Figure 10 presents the LISA aggregation map of the ESV in Guilin.

During the 20-year period, the spatial aggregation of ESV basically remained stable, with high significance for the high-high and low-low aggregation types and non-significance for the high-low and low-high types. Overall, there was a “hollowing out” near the city center, with a “low-high-low” trend in distribution. The high-high aggregation category areas were mainly located in Lingchuan, Xing’an, Yongfu, and Pingle County along the Lijiang River. The aggregation of provincial nature reserves, such as the Mao’er Mountain Nature Reserve located in the north of Xing’an and the Qingshitan Water Source Forest Nature Reserve in the northwestern part of Lingchuan, was increased, indicating that the ecological protection of the nature reserves was effective. The low-low aggregation type was mainly concentrated in the municipal district and the northern and southern regions, and the distribution was gradually extensive, which was in line with the overall planning of relieving the old city in urban land planning. However, the overall low-low area became more, indicating that the destruction and fragmentation of ecological factors were obvious in the process of urban construction and tourism development.

4. Discussion

4.1. Spatial and Temporal Characteristics of Land Use

As a tourist city with a karst landform, Guilin has a developed tourism industry, relatively more mountain terrain, and a low utilization rate of cultivated land [53]. Through the comprehensive analysis of the structure, dynamic change degree, and transfer matrix of land use, our results revealed the trend of land use change in Guilin.

The study revealed the current impact of building land expansion on the availability of cultivated land, which is consistent with the research of Lan [54] and others. Over the past 20 years, there have been significant changes in the areas of forestland, cultivated land, and grassland, with a notable decrease in total area due to their mutual transformation. The primary reason for the reduction in these land areas is the extensive expansion of building land [55]. The study highlights ongoing urban sprawl in Guilin, driven by construction activities, with pressures on land resources and significant spatial transformation [56]. The study area showed an “L-shaped” pattern centered around the city center and extending along both sides of the Lijiang River. In the municipal district, Lingui County, and Lipu County, development shifted from dispersion to concentration, forming strong agglomeration effects. Waterbody areas increased, possibly due to Guilin’s aesthetic landscape value and seasonal rice paddies. The study also revealed an increase in land transfers over 20 years, with cultivated land, grassland, and forestland transfers, alongside conversions of some building land to ecological uses, reflecting the influence of sustainable policies. Based on the study results, the government should take the following measures: To address the increase in building land and the reduction of cultivated land, forestland, and grassland, land use planning should be optimized, urban renewal should be promoted, building land efficiency should be improved, and urban green spaces should be expanded. Regarding the increase in water areas, water management should be strengthened to balance landscape needs with ecological functions. Additionally, the protection of cultivated land should be reinforced, land use conversion should be strictly controlled, and ecological restoration policies should be promoted to ensure the rational use of land resources and long-term ecological stability.

4.2. The Contribution of Land Use to ESV in Guilin

In the land use structure of Guilin, there were significant differences in the contribution of different types of land to ESV. Figure 11a illustrates the contribution of each land use type to the total area, and Figure 11b illustrates the contribution of each land use type to the total ESV.

As one of the valuable ecological resources in Guilin, the contribution of a waterbody to the ESV was as high as about 10%, with the total area at 1% share. And it is increasing year by year, reflecting the fact that waterbody possesses very high ecological value [57]. The Lijiang River and its surrounding waters are not only the core hub for water resource regulation and environmental purification but also the habitat of many species whose ecological value is inestimable. In addition, these waterbodies are unique tourist resources in Guilin and have a significant impact on enhancing citizens’ quality of life and promoting local economic development. In contrast, although cultivated land accounted for a large proportion of land use in Guilin, its contribution to ESV was comparatively limited. It reflects the limited support of cultivated land for ecological functions such as soil and water conservation, biodiversity maintenance, and climate regulation [58]. This is closely related to its wide distribution and highly intensive management approach [59]. To increase Guilin’s ESV, it is important to take advantage of the core strengths of the waterbody, in particular, to optimize the pattern of use of cultivated land and reduce the overexploitation of building land. In the future, Guilin needs to explore more eco-friendly farming models in land management. Forestland and grassland remained essentially the same in terms of total area and contribution to ESV. Forestland and grassland were the two most dominant land use types in Guilin. As crucial providers of ecosystem services, they not only play a significant role in mitigating climate change, stabilizing soil, and controlling soil erosion but also offer vital habitats that support biodiversity. Their contributions are essential for maintaining species diversity and improving carbon sequestration capacity [60]. To enhance the ESV in Guilin, careful planning of ecological management and land use is essential. This includes optimizing farmland structure, promoting ecological agriculture, reducing chemical fertilizers and pesticides, and improving the ecological functions of farmland. Additionally, it is important to strengthen the protection of forests and grasslands, implement effective ecological restoration efforts, and focus on the health of aquatic ecosystems through rigorous water quality monitoring and protection. The comprehensive measures aim to balance economic development and tourism with ecological protection, thereby advancing the development of ecological civilization in Guilin.

4.3. Analysis of ESV Effectiveness by County in Guilin

The method of equivalence factor per unit area adopted in this study has the advantages of requiring less data, being intuitive and easy to operate, and is applicable to the assessment of regional ESV at different scales [61]. Compared to other studies that use the same correction factor for the whole region on the same time scale [62,63], the introduction of three types of spatiotemporal correction factors for ESV with temporal and spatial pixel corrections made the result more practically applicable and comparable. Compared with studies in the same study area [32,33], a more accurate reflection of the actual situation within each county was obtained, which contributes to a better understanding of the spatial heterogeneity within counties and compares the differences between counties.

In our study, we analyze differences between total ESV and per-unit ESVa to identify county ecological effectiveness. The ranking comparison chart is shown in Figure 12. Most counties have near-unanimous aggregate ESV and ESVa rankings. It indicates that most counties are relatively stable in terms of ecosystem service performance levels. Yongfu and Lingchuan led in both total ESV and per-unit ESVa, reflecting their high supply, high performance, and superior ecological environment. On the contrary, Municipal District and Guanyang had low total ESV and per-unit ESVa, indicating that their ecological structure functions are limited and they are subject to strong anthropogenic disturbances due to the high proportion of constructed land, unused land, and cultivated land, affecting the ESV.

Lingchuan and Yangshuo ranked low in terms of total ESV but high in terms of ESVa, showing that their small land areas provide efficient ecological services. These counties are rich in ecological resources, located in the Lijiang River Basin, with a number of ecological zones such as Qingshitan and the Yulong River, and a dense distribution of high-value land use types such as water areas and forest land, which provided a high density of ecological service functions. Meanwhile, both of them actively responded to higher-level policies and promoted ecological restoration projects with remarkable results. Although the total ESV in these counties is low, the high ranking of ESVa suggests the full utilization of the potential for ecosystem service benefits in their areas and also demonstrates the spatial heterogeneity and advantageous distribution of ecosystem service provision. Guilin should continue to carry forward the governance policies of the region and form a model of governance that can be drawn on and promoted, leading the city to maintain and develop ecology efficiently.

On the contrary, Lingui and Quanzhou ranked significantly better in terms of total ESV than ESVa, especially Quanzhou, which topped the list for 20 consecutive years in terms of total ESV but ranked only twelfth in terms of ESVa, suggesting that it has a large amount of total ecosystem services but is concentrated in a few high-density areas. This characteristic suggests an uneven distribution of ESV within these counties, which requires the county’s attention and adjustment in land use planning and ecological protection policies to ensure a balanced supply of ESV and maximize socio-economic benefits. This characteristic suggests that ESVs are unevenly distributed and ecological utilization efficiency is low in these countries. The governments of these counties need to pay attention to and adjust their land use planning and environmental protection policies and learn from high-efficiency ecological governance models to ensure a balanced supply of ESVs and maximize socio-economic benefits.

In summary, by comparing and analyzing ESV at the county scale, this study may help Guilin to more efficiently and comprehensively formulate ecological protection strategies based on each county’s own situation and to promote the comprehensive enhancement of ecosystem services and sustainable development.

4.4. Spatial Patterns of ESV in Guilin from a Grid Perspective

From the grid ESV and spatial clustering in Guilin over 20 years, we can observe the dynamics of the inverted ecological highlights and weaknesses areas more finely to aggregate the situation, which is crucial for the maintenance and development of ecologically focused areas (e.g., the World Natural Heritage site of the Lijiang River, representative scenic area of the Mao’er Mountain Ecological Reserve, and Qingshitan Reservoir, etc.).

The ecosystem value level was low in areas with dense building land, such as the northern part of the Municipal District, the southeastern part of Lingui County, etc., where the area is densely populated with residential and commercial areas and where there are frequent human activities, and the vicinity of industrial and economic development zones such as the Quanzhou Industrial Concentration Zone located in the northeast, and the Pingle Industrial Concentration Zone located in the southeast, can show a low level of ecological value, which is consistent with the conclusions of the study [64]. The government should enhance the ESC of ecologically low-value key areas through green infrastructure construction, economic incentive policies, and ecological compensation mechanisms to realize the coordinated development of regional ecology and economy. Along the Lijiang River Basin, ecological scenic spots such as Qingshitan Reservoir in Lingchuan and Mao’er Mountain in Xing’an demonstrate sustained and significant high ESV levels. These areas should continue to maintain the ecological environment, implement strict protection of the core areas, and prohibit activities such as construction and occupation, deforestation and reclamation, development of commercial forests, and other activities that force the ecological environment to suffer direct or indirect damage, as well as prohibit tourism development projects such as construction facilities that permanently damage the original ecology. Establishing an ecosystem health monitoring system is suitable for vigorously developing “ecology +” tourism integration and other development modes, realizing ecological economic gains [65], and ensuring the long-term stability of environmental functions and rational use.

Spatial aggregation was evident in Guilin. In terms of localized spatial autocorrelation, the high-high aggregation areas are concentrated near ecological zones, and the persistently high ESV in these areas not only verifies the effectiveness of the policy and ecological management but also provides successful examples of ecological conservation in other areas. However, some of the high aggregation types around the ecological zone were reduced. Tourism development continues to infringe on the periphery of the ecoregion, and development projects that destroy the original ecology should be prohibited in Guilin’s landscape as the primary source of ESVs in Guilin. Meanwhile, the low-low aggregation area was mainly located in the urban building land and cultivated land concentration area, revealing the dual impact of urban development on ESV, suggesting that cities need to pay more attention to the protection and rational utilization of ecological space in future development.

In the future, to address the dual challenges of economic development and ecological protection, scientific management measures should be based on the division of ecological functional zones. Tourism infrastructure should be avoided in sensitive ecological areas to protect ecosystem integrity and stability. Strategies should be adjusted based on ESV changes to promote balanced development and conservation.

4.5. Research Limitations and Future Research Directions

In this study, only five periods of land use data from 2000 to 2020 were used in Guilin, and year-by-year land use data will be used in future studies to reflect the trend of ESVs more comprehensively. Moreover, the accuracy of land use is 30 m, and in the future, we will use data with higher accuracy to reduce the bias caused by reclassification and model calculation. In addition, the resolution of the datasets used for the spatiotemporal correction factors NPP, FVC, and precipitation varied. Because of the inconsistency in accuracy, we were only able to reconcile it with county accuracy, and future studies will use uniform, higher-resolution data to correct ESV results that are more consistent with real-world environmental characteristics. We will further explore methods of uncertainty analysis to enhance the reliability of the results.

The high share of water in ESV due to water resources, hydrological regulation, and rainfall correction parameters was noted in the methodology. However, despite a 30-year flood in 2015 with record precipitation, which coincided with a peak in total ESV, flooding caused negative impacts such as damage to hydrological systems and soil erosion. This indicates that ESV is not directly proportional to precipitation or waterbody quantity. Future work will focus on developing correction parameters for flood inundation to improve the model’s practical application. Due to these shortcomings, subsequent in-depth research is needed to provide technical support and sustainable development decisions for land optimization in Guilin.

5. Conclusions

In this study, three types of spatiotemporal correction coefficients were introduced to improve the ESV of Guilin from 2000 to 2020 with a dual-scale perspective of the county and a 3 km × 3 km grid, which provided a basis for decision-making on the sustainable and green development of the region.

The main conclusions of this paper are as follows:

(1)

The improved ESV assessment methodology considered the spatiotemporal heterogeneous effects of biomass, soil conservation, and precipitation, adjusted equivalence factors on a cell-by-cell basis according to the specific characteristics of regional environments, revealing ESV results with higher accuracy and practical applicability;

(2)

The dual-scale perspective, which obtained a wide range of trends and detailed spatial characterization results, contributed to the exploration of ecosystem conditions and optimization of land allocation from different perspectives. The county scale identified high-value and low-value as well as high-efficiency and low-efficiency areas. The grid scale offered a more detailed view of the spatial pattern characterized by “high in the west and north, low in the east and south”, along with the phenomenon of “hollowing out” aggregation. High-value areas were predominantly located in counties such as Lingchuan and Yangshuo and near ecological zones like the Lijiang River and Mao’er Mountain;

(3)

From 2000 to 2020, Guilin showed the highest but decreasing trend in the proportion of forest land among land use types; the expansion of building land was noticeable, and the flow of each land use increased yearly. Forestland types and the hydrologic regulation service function contributed the most to Guilin’s ESV, with the waterbody being extremely dominant. There are significant differences in ESVs between counties, with consistently high values near waterbodies and ecological reserves and low values near industrial and commercial land and cultivated land. The total ESV showed S-shaped fluctuation, with the lowest value in 2005 (CNY 154.716 billion) and the highest in 2015 (CNY 178.719 billion). Although there was a trend of ecological recovery in the later period, there was still a gap between 2020 and 2000. Guilin was burdened with the responsibility of balancing ecological protection and development.