Impacts of Land Use Changes on Landscape Patterns and Ecosystem Service Values in Counties (Villages) in Ethnic Regions of China: A Case Study of Jianghua Yao Autonomous County, Hunan Province

Abstract

This study, using Jianghua Yao Autonomous County in Hunan Province as a case, systematically analyzes the response of ecosystem service value (ESV) to land use and landscape pattern changes by employing landscape-level indices and landscape-type-level indices. The findings provide a reliable basis for scientifically formulating land use planning and ecological protection policies in ethnic regions, thereby promoting regional ecological security and sustainable development. This study reveals that (1) the land use structure in the county underwent significant changes between 2000 and 2020, with grassland and shrubland areas decreasing substantially by 71.66% and 78.41%, respectively, while urban and arable land areas increased significantly by 228.30% and 15.84%, respectively. Particularly under the scenario of prioritizing economic development, these changes led to increased landscape fragmentation and a decline in ecosystem service value (from CNY 296.571 billion in 2020 to CNY 287.959 billion in 2030). (2) In contrast, the scenarios of ecological protection and sustainable development significantly enhanced the region’s ecosystem service value by increasing forest and water area, effectively maintaining the stability of the landscape pattern. These findings provide important evidence for the formulation of scientifically sound land use planning and ecological protection policies, contributing to the dual goals of economic development, tourism growth and ecological protection in Jianghua Yao Autonomous County and similar regions.

1. Introduction

County-level ethnic regions in China refer to administrative areas at the county level where a particular ethnic minority constitutes the majority or a significant proportion of the population. These regions typically enjoy certain autonomous rights to safeguard the cultural, economic, and social development of ethnic minorities. County-level ethnic regions include autonomous counties and ethnic townships, which are widely distributed across the country, particularly concentrated in the southwest, northwest, and northeast regions. These areas possess a degree of administrative autonomy and receive preferential support from central fiscal policies to promote balanced regional development and maintain ethnic unity. The “Guiding Opinions of the State Ethnic Affairs Commission on Further Strengthening and Regulating the Protection and Development of Ethnic Minority Characteristic Villages and Towns in the New Era” states that “The construction of ethnic minority characteristic villages and towns should be integrated into the implementation plan of the rural revitalization strategy, achieving seamless connection and integrated promotion, and orderly advancing the protection and development of ethnic minority characteristic villages and towns” [1]. In recent years, with the comprehensive advancement of the rural revitalization strategy, the land utilization rate in ethnic minority villages has increased and land use methods have diversified. Changes in landscape patterns, which reflect the spatial and temporal distribution of land use, inevitably affect the quality and evolution of ecosystems, thereby impacting local ecological security. Therefore, exploring the spatiotemporal response of landscape pattern changes and ecosystem service value in county-level ethnic regions has increasingly become a crucial topic in the construction of ecological civilization.

As a core component of landscape ecology research, landscape patterns quantitatively describe the spatial structure characteristics of landscapes [2], revealing the mechanisms by which natural changes and human social development impact the ecological environment [3]. To optimize ecosystems and enhance the quality of ecosystem services, scholars both domestically and internationally have conducted in-depth studies on the evolution of landscape patterns. Killen et al. [4] and Alphan et al. [5] established various landscape indices to study and analyze the spatiotemporal changes in landscape patterns in the Carpathian region and Mediterranean towns. Shiliang Su et al. [6] proposed that changes in landscape patterns affect the evolution process, components, structure, and biodiversity of ecosystems, thereby influencing ecosystem services. Jiawei Gu et al. [7], using the Fengxi Provincial Nature Reserve in Guangdong Province as an example, explored the evolution patterns of landscape structures and discussed the driving forces behind these changes.

In 1997, Daily and Costanza successively introduced the concept of ecosystem services (ESV) [8] and the evaluation method for ecosystem service value (ESV) [9], laying the foundation for subsequent research. The main components of ecosystem services include regulating services, supporting services [10], cultural services, and provisioning services, which primarily manifest in the natural environmental conditions and utilities that sustain human survival and development within ecosystems. Holmes et al. [11] conducted comprehensive and in-depth studies on the functions, accounting methods, and applications of ESV. Based on the accounting methods proposed by Costanza et al. [9], Gaodi Xie et al. [12,13] developed new per-unit area value equivalents tailored to China’s national conditions, providing a foundational contribution to ESV research in China. Jie Yang et al. [14], using land use data from the Aksu River Basin from 2000 to 2020, explored the spatiotemporal evolution characteristics and driving factors of ESV. Baiting Zhang et al. [15] calculated the ESV of the Qilian Mountain region and investigated the dynamic changes in ESV. Gaodi Xie et al. [16] further refined the classification of ESV and revised the equivalent factor table, laying the groundwork for precise ESV assessment in China. The current mainstream methods for ESV calculation include the equivalent factor method, the functional value method, and the model calculation method, among which the equivalent factor method is widely used because of its standardized accounting, result consistency, and global applicability in calculating ecosystem service value [17].

In summary, current research on landscape patterns and ecosystem service value (ESV) is burgeoning, yet it predominantly focuses on provincial [18] and municipal scales [19], with primary subjects being watersheds [20] and wetlands [21]. There is a notable lack of in-depth studies targeting villages administered at the county level, particularly in ethnic regions where ecological systems require further exploration. Therefore, this study takes Jianghua Yao Autonomous County in Hunan Province as a case, utilizing remote sensing data from 2000, 2010, and 2020. By selecting landscape-level indices and landscape-type-level indices, we systematically analyze the spatiotemporal response of landscape pattern changes and ESV in Jianghua Yao Autonomous County from 2000 to 2020 and explore the changes in landscape patterns and ESV under different future scenarios. This research aims to effectively understand the ecological environment changes in ethnic minority villages, scientifically allocate and utilize land resources, improve regional ecological quality, and achieve the goal of sustainable development for both humans and nature.

2. Study Area and Data Sources

2.1. Overview of the Study Area

Jianghua Yao Autonomous County is located between 110°25′ and 112°10′ E longitude and 24°38′ and 25°15′ N latitude. It stretches 77.92 km from north to south and 72.5 km from east to west, covering a total area of 3248 square kilometers. The county comprises 16 townships and one state-owned forest farm, with a total population of 540,000, of which 70% are Yao people, making it the county with the largest Yao population in China. It is known as the “Capital of the Yao People in China”. Jianghua Yao Autonomous County features a mid-subtropical monsoon humid climate, diverse landforms, and significant regional and vertical differences, resulting in various microclimate zones. The county has a well-developed water system and abundant mineral resources. The area is rich in historical and cultural heritage, with several villages, including Shuidong Village in Chetianhe Town, Baojing Village in Daxu Town, and Jingtouwan Village in Dashiqiao Township, listed in the National Directory of Traditional Villages of China (Figure 1).

2.2. Data Sources and Preprocessing

This study comprehensively analyzes the landscape patterns and ecosystem service values of Jianghua Yao Autonomous County, utilizing multi-source data including remote sensing imagery, meteorological data, and demographic and economic data. The specific data sources and preprocessing steps are as follows:

• Land Use Data: Land cover data from 1985 to 2019 were sourced from open-access data published by Yang Jie et al. at Wuhan University [22], with a spatial resolution of 30 m. These data were used for LUCC monitoring and studies on ecosystem responses.
• Meteorological Data: These data were primarily used for regional land use multi-scenario simulation and prediction, including temperature and precipitation data from the National Earth System Science Data Center (http://www.geodata.cn, accessed on 15 March 2024) for the years 2010 and 2020, with a spatial resolution of 1 km. Precipitation data were measured in 0.1 mm units and temperature data in 0.1 °C units. These were processed into annual averages using ArcGIS.
• Population and GDP Distribution Data: These data were mainly used for regional land use multi-scenario simulation and prediction. Population distribution data were sourced from Popworld (https://www.worldpop.org/, accessed on 15 March 2024) and GDP data from the Resource and Environment Data Center of the Chinese Academy of Sciences (https://www.resdc.cn/, accessed on 16 March 2024). Data for the years 2010 and 2020 were selected, with a spatial resolution of 1 km. Missing values were filled using the raster calculator.
• DEM Data: DEM data were available for the years 2009 and 2020. The 2009 DEM data were sourced from the Geospatial Data Cloud (https://www.gscloud.cn/search, accessed on 5 April 2024) and the 2020 DEM data from NASA (https://lpdaac.usgs.gov/news/release-nasadem-data-products/, accessed on 5 April 2024), both with a spatial resolution of 30 m. These were resampled to 1 km, and slope and aspect data were extracted using ArcGIS.
• Other Data: Distance to various levels of roads and distance to rivers of grade three and above were vector data sourced from the Geospatial Remote Sensing Ecological Network (http://www.gisrs.cn/, accessed on 17 March 2024). These were converted to raster data using the Euclidean distance method in ArcGIS.

All data were based on ArcGIS 10.8.1 software and were projected and transformed to the WGS-1984-UTM-zone-49N coordinate system. Subsequently, all datasets underwent spatial resampling and standardization to eliminate differences among the data sources and ensure data consistency.

3. Analysis Methods

This study systematically investigates the land use changes and their impacts on ecosystem service values in Jianghua Yao Autonomous County, Hunan Province, across different years and future scenarios using the following three main methods: remote sensing image analysis, landscape pattern index calculation, and ecosystem service value assessment.

3.1. Land Use Scenario Prediction

Given the significant impact of land use and cover change (LUCC) on regional ecosystem service values, this paper first analyzes the characteristics of land use changes in Jianghua Yao Autonomous County from 2000 to 2020 using a land use transition matrix, laying the foundation for subsequent LUCC contribution analysis. The land use transition matrix, derived from system analysis methods, quantitatively describes the conversion relationships among different land use types using the Markov model. This model generates a two-dimensional matrix based on land cover changes in the same area at different times. Through in-depth analysis of this matrix, spatial location and area information of land type conversions between different periods can be obtained [23].

This paper employs the FLUS (Future Land Use Simulation) model to simulate and predict land use, exploring the response characteristics of regional ecosystem service values under different LUCC scenarios. The FLUS model is a comprehensive method for simulating future land use changes, integrating the impacts of human activities and the natural environment. By combining top-down system dynamics (SD) models with bottom-up cellular automata (CA) models, the FLUS model simulates various LUCC scenarios [16,24].

The FLUS model has been widely applied to address geographic process simulations and complex spatial optimization problems, such as large-scale land use changes, urban expansion, and nature reserve delineation. It handles land use distribution at the spatial level and predicts the quantitative changes in each land use type through the Markov model, thus exploring dynamic land use information in both spatial and quantitative dimensions [25].

Based on the historical development characteristics of land use in the study area and regional sustainable development needs, this paper set up the following three scenarios for LUCC prediction research:

(1) The natural development scenario predicts the quantity of each land type in 2030 based on the transition probabilities among land use types from 2010 to 2020, using the 2020 land use status as the baseline. (2) The economic development scenario prioritizes regional economic development needs by increasing the transition probability of arable land to impervious surfaces and forest and grassland to impervious surfaces. (3) The ecological protection scenario prioritizes the protection of land types with high ecosystem service functions, such as arable land, forest, and grassland, by increasing the transition probability of arable land to forest, reducing the transition probability of arable land to impervious surfaces, and reducing the transition probability of grassland and water bodies to unused land.

To determine the total land use change amounts for each scenario, the FLUS model was used to set the “Future Pixel Number” through the following steps:

(1)

Historical Data Analysis and Modeling: We analyzed land use data from 2000, 2010, and 2020 to construct a land use transition matrix using the Markov model. This matrix quantified the conversion relationships among different land use types and provided transition probabilities across years. These probabilities allowed us to estimate the future changes in land use types for the year 2030 under a natural development scenario.

(2)

Setting Future Pixel Numbers: For the natural development scenario, the FLUS model was used to predict the future pixel numbers for each land use type in 2030 based on historical spatial trends and transition probabilities. For the economic development priority and ecological protection priority scenarios, the future pixel numbers were adjusted according to different policy directions. Specifically, in the economic development priority scenario, the future pixel numbers for arable land and built-up areas were increased, while those for forests and grasslands were decreased. Conversely, in the ecological protection priority scenario, future pixel numbers for forests and water bodies were increased, with a corresponding reduction in arable land and built-up areas.

(3)

Scenario Simulation and Adjustment: After setting the initial future pixel numbers, we iteratively adjusted the simulation results to ensure that the total land use amounts were reasonable for each scenario. For example, in the economic development priority scenario, we prioritized the demand for arable land and built-up areas while maintaining minimal protection areas for forests and grasslands. In the ecological protection priority scenario, although arable land expansion was minimized, we ensured that the minimum required arable land for regional food security was preserved.

(4)

Relation Between Future Pixel Numbers and Land Use Change Amounts: The total land use change amounts presented were directly derived from the FLUS model’s simulation results of future pixel numbers. These pixel numbers reflect spatial changes in land area for each land use type. For instance, reductions in grasslands and increases in arable land were achieved through adjustments in future pixel numbers, considering both spatial distribution and policy directions for each scenario. This approach ensured that the land use change amounts in each scenario were scientifically justified and accurately represented the projected changes.

To ensure the accuracy and reliability of the FLUS model simulation results, we conducted a model validation process comparing simulated land use data with actual data from 2020.

Overall Accuracy: The ratio of correctly simulated grid numbers to the total grid numbers was calculated as 3,286,205/3,565,971 = 0.921, indicating a high level of accuracy in the simulation results.

Kappa Verification: The Kappa coefficient is a common metric used to evaluate the accuracy of model predictions, especially when dealing with classification tasks like land use simulation. The Kappa coefficient is calculated using the following formula:

$$kappa=(P_0-P_c)/(P_p-P_c)$$

where $P_0$ represents the proportion of correct simulation, $P\mathrm{c}$ represents the expected proportion of correct simulation under random conditions, and $P_p$ represents the maximum proportion of correct simulation under ideal conditions. In our study, the calculated Kappa coefficient was:

$$kappa={\displaystyle \frac{(0.921-\frac{1}{6})}{(1-\frac{1}{6})}}=0.906$$

This result demonstrates the high consistency and reliability of the model simulation. A Kappa value greater than 0.75 is typically considered to indicate good agreement between the simulation and the actual land use data, further confirming the reliability of the FLUS model predictions.

3.2. Calculation of Landscape Pattern Indices

Landscape pattern indices are used to quantitatively describe the spatial structure characteristics of landscapes and reveal the impact of land use changes on landscape patterns. This study employed Fragstats 4.2 software to calculate various landscape pattern indices, including the number of patches (NP), patch density (PD), the largest patch index (LPI), the landscape shape index (LSI), and the cohesion index (COHESION). These indices were used to measure the degree of landscape fragmentation, shape complexity, and connectivity among patches (

Table 1. The selected landscape indices and their definitions.

Landscape Index	Definition
Number of patches, NP	The number of patches reflecting different land use types typically indicates an increase in landscape fragmentation, as an increase in the number of patches usually signifies heightened landscape fragmentation.
Patch density, PD	The number of patches per unit area reflects the spatial distribution density of the landscape.
Largest patch index, LPI	The proportion of the area of the largest patch within the landscape measures the degree of landscape concentration.
Landscape shape index, LSI	The complexity of the shape of landscape patches is described, with higher values indicating more complex shapes.
Cohesion index, COHESION	The connectivity of landscape patches is reflected by this value; a higher value indicates stronger connectivity between patches.

).

3.3. Ecosystem Services Valuation

This study employs the ecosystem services valuation (ESV) method. Based on the unit area ecosystem service value equivalency table for different land use types from the existing literature, combined with land use classification data, the total value of ecosystem services for various land use types in different years and scenarios was calculated. The primary ecosystem services evaluated included provisioning services, regulating services, supporting services, and cultural services. By comparing the changes in ESV across different years and future scenarios, the impact of land use changes on regional ecosystem service functions was analyzed in depth.

Xie Gaodi and other scholars [12,13,26,27] optimized and adjusted the evaluation method proposed by Costanza et al. [9] based on expert knowledge, formulating an ecosystem service value equivalency factor table more suitable for China’s national conditions. They categorized ecosystem service functions into four major categories and nine secondary categories. To better reflect regional differences and the comparability of results, this study further referenced the regional correction coefficients for ecosystem service values proposed by Xie Gaodi et al. [28,29], and developed a localized ecosystem service value evaluation coefficient table for Hunan Province (

Table 2. Unit area service value coefficients for various ecosystem types in Hunan Province (CNY·hm−²·a−¹).

Ecosystem Services	Individual Service Function	Farmland	Forestland	Grassland	Water Bodies
Provisioning services	Food production	3852.20	1271.23	1656.45	2041.67
	Raw material production	1502.36	11,479.55	1386.79	1348.27
Regulating services	Gas regulation	2773.58	16,641.50	5778.30	1964.62
	Climate regulation	3736.63	15,678.45	6009.43	7935.53
	Water conservation	2966.19	15,755.50	5855.34	72,305.78
Supporting services	Waste treatment	5354.56	6625.78	5084.90	57,205.16
	Soil formation and retention	5662.73	15,485.84	8628.93	1579.40
	Biodiversity protection	3929.24	17,373.42	7203.61	13,213.04
Cultural services	Recreation and leisure	654.87	8012.57	3351.41	17,103.77
Total		30,432.38	108,323.85	44,955.17	174,697.25

), to evaluate and calculate the ecosystem service values of the region in different periods. The formula is as follows:

$$E_n=1/7{\sum }_{i=1}^m{\displaystyle \frac{o_i{p}_i{q}_i}{M}},$$

(1)

where $E_n$ represents the economic value of the food production capacity of the farmland ecosystem per unit area in Yao Autonomous County (yuan/hm²); i denotes the crop category; $o_i$ is the area of the i-th type of grain crop (hm²); $p_i$ is the yield per unit area of the i-th type of grain crop (kg/hm²); $q_i$ is the average price of the i-th type of grain crop (yuan/kg); and M represents the total area of grain crops.

The formula for calculating ESV is as follows:

$$ESV=\sum (A_k\times {VC}_k)$$

(2)

$${ESV}_f=\sum (A_k\times {VC}_{fk})$$

(3)

where ESV represents the ecosystem service value (CNY); $A_k$ is the area of land type (k) in the study area (hm²); ${VC}_k$ is the unit area value coefficient of that land type (CNY·hm−²·a−¹); ${ESV}_f$ and ${VC}_k$ are the value of individual ecosystem services (CNY) and the value coefficient (CNY·hm−²·a−¹), respectively.

4. Results and Analysis

4.1. Characteristics of Landscape Pattern Evolution

4.1.1. Characteristics of Landscape Area Change

By analyzing the land use data of Jianghua Yao Autonomous County from 2000 to 2020, the area transfer situation of various land types in different periods was obtained (Figure 2). During the period from 2000 to 2010, the main source of new farmland was forestland, with an area of 80,160 km². At the same time, part of the farmland area was transferred to forestland and construction land, with transfer areas of 38,509 km² and 6585 km², respectively. Forestland mainly transferred into farmland, with an area of 38,509 km², and transferred out to farmland and shrubland, with areas of 80,160 km² and 1435 km², respectively. During the period from 2010 to 2020, the main source of new farmland remained forestland, with an area of 103,940 km², while its main transfer directions were still forestland and construction land, with transfer areas of 57,162 km² and 13,314 km², respectively. The main source of new forestland was farmland, with an area of 57,162 km², and it mainly transferred out to farmland and water bodies, with areas of 103,940 km² and 1486 km², respectively. Construction land expanded significantly, mainly transferring from farmland and forestland, with transfer areas of 13,314 km² and 1365 km², respectively.

Overall, over time, the expansion of farmland and construction land was very significant, reflecting the increased demand for land resources due to socio-economic development, which also led to a reduction in natural land types such as forests and grasslands. This change reveals the profound impact of human activities on the natural environment, especially in terms of land use.

To further understand the future trend in this change, we predicted future scenarios to assess the potential long-term impacts of these changes (Figure 3). By analyzing the changes in landscape type areas in Jianghua Yao Autonomous County in 2000, 2010, 2020, and 2030 under natural development scenarios (

Table 3. Areas of various landscape types in 2000, 2010, and 2020 and under the 2030 natural development scenario.

Year and Proportion	Landscape Type Area (hm2)
	Grassland	Urban Areas	Farmland	Shrubland	Forest	Water Bodies
2000	306.45	1399.32	55,726.92	881.19	261,197.91	1425.6
Proportion %	0.10%	0.44%	17.36%	0.27%	81.39%	0.44%
2010	286.56	2081.52	58,980.06	309.69	257,939.19	1340.37
Proportion %	0.09%	0.65%	18.38%	0.10%	80.37%	0.42%
2020	105.75	3357.9	62,049.78	231.66	253,532.25	1660.05
Proportion %	0.03%	1.05%	19.33%	0.07%	79.00%	0.52%
2030 natural development scenario	86.85	4593.96	64,554.84	190.26	249,851.43	1660.05
Proportion %	0.03%	1.43%	20.11%	0.06%	77.85%	0.52%
Change rate %	−71.66%	228.30%	15.84%	−78.41%	−4.34%	16.45%
Average proportion %	0.06%	0.89%	18.80%	0.13%	79.65%	0.47%

), it was found that its matrix landscape was forestland, accounting for 79.65%; followed by farmland, accounting for 18.80%; and urban areas, water bodies, shrubland, and grassland had smaller areas, accounting for only 0.89%, 0.47%, 0.13%, and 0.06%, respectively. The land use types in Jianghua Yao Autonomous County changed significantly at these time points, where the urban land area continuously increased, with an increase of 228.30%, and the water body area increased by 16.45%. Shrubland and grassland areas decreased significantly, by 78.41% and 71.66%, respectively.

4.1.2. Characteristics of Landscape Pattern Changes

(1)

Landscape Level Index Analysis

Selected landscape indices, including NP (number of patches), FRAC_AM (Average Shape Index), PLADJ (Proximity Index), SHDI (Shannon Diversity Index), SHEI (Evenness Index), and AI (aggregation index), were used to analyze the landscape types of Jianghua Yao Autonomous County in 2000, 2010, and 2020 and under different future scenarios (2030 economic development priority scenario, 2030 ecological protection scenario, 2030 natural development scenario). The changes in landscape-level indices for Jianghua Yao Autonomous County are presented in

Table 4. Table of changes in landscape-level indices in 2000, 2010, and 2020 and under the future scenarios.

Year	NP	FRAC_AM	PLADJ	SHDI	SHEI	AI
2000	12,631	1.2589	96.1315	0.5422	0.3026	96.2055
2010	11,669	1.2667	96.1643	0.5555	0.3100	96.2385
2020	12,026	1.2721	95.7207	0.5868	0.3275	95.7955
2030 economic development priority scenario	18,449	1.3206	94.3247	0.642	0.3583	94.4000
2030 ecological protection scenario	11,137	1.2753	95.7119	0.5788	0.3230	95.7867
2030 natural development scenario	14,988	1.3038	94.9943	0.6121	0.3416	95.0693

.

From the changes in the number of patches (NP), it can be seen that between 2000 and 2010, the number of landscape patches in the study area gradually decreased, indicating a reduction in landscape fragmentation. However, from 2010 to 2020, the number of patches began to increase again, showing a resurgence in landscape fragmentation. Looking ahead to 2030, the changes in NP vary significantly under different scenarios. In the economic development priority scenario, NP increases significantly, indicating more severe landscape fragmentation; in the ecological protection scenario, NP decreases, indicating a trend towards landscape consolidation; while in the natural development scenario, the degree of landscape fragmentation falls between the two.

Based on the trends in FRAC_AM (Average Shape Index), SHDI (Shannon Diversity Index), and SHEI (Evenness Index), it can be seen that between 2000 and 2020, the shapes of landscape patches in the study area became increasingly complex, landscape heterogeneity increased significantly, and the overall landscape pattern became more complex. Looking ahead to 2030, the changes in these indices differ under the different scenarios. In the economic development priority scenario, FRAC_AM increases to 1.3206, indicating more complex landscape shapes; in the ecological protection scenario, FRAC_AM is 1.2753, with little change; while in the natural development scenario, FRAC_AM is 1.3038, with shape complexity falling between the two. Meanwhile, in the economic development priority scenario for 2030, SHDI increases to 0.642, indicating increased landscape diversity; in the ecological protection scenario, SHDI is 0.5788, with little change; while in the natural development scenario, SHDI is 0.6121, with diversity falling between the two.

Regarding SHEI, in the economic development priority scenario for 2030, SHEI increases to 0.3583, indicating more even landscape distribution; in the ecological protection scenario, SHEI is 0.323, with little change; while in the natural development scenario, SHEI is 0.3416, with evenness falling between the two. Based on the changes in PLADJ (Proximity Index) and AI (aggregation index), it can be seen that between 2000 and 2010, the aggregation of landscape patches in the study area fluctuated slightly; from 2010 to 2020, the connectivity among landscape patches decreased, leading to a reduction in aggregation. Looking ahead to 2030, in the economic development priority scenario, PLADJ decreases to 94.3247, indicating reduced connectivity among landscape patches; in the ecological protection scenario, PLADJ is 95.7119, maintaining high connectivity; while in the natural development scenario, PLADJ is 94.9943, with connectivity falling between the two. Similarly, in the economic development priority scenario for 2030, AI decreases to 94.4, indicating reduced aggregation of landscape patches; in the ecological protection scenario, AI is 95.7867, maintaining high aggregation; while in the natural development scenario, AI is 95.0693, with aggregation falling between the two.

In summary, the landscape pattern indices of Jianghua Yao Autonomous County changed significantly across different years and future scenarios. In the economic development priority scenario, landscape fragmentation intensifies, the number of patches increases, and shape complexity and diversity rise; however, connectivity and aggregation decrease, indicating significant damage to the landscape due to rapid urbanization and agricultural expansion. Conversely, in the ecological protection scenario, the landscape tends towards consolidation, the number of patches decreases, and connectivity and aggregation are higher, demonstrating the effectiveness of ecological protection measures in maintaining landscape integrity and stability. In the natural development scenario, landscape pattern changes fall between the two, with some degree of fragmentation and some retention of consolidation.

(2)

Landscape Type Level Index Analysis

Selected landscape type level indices, including NP (number of patches), PD (patch density), LPI (largest patch index), LSI (landscape shape index), and COHESION (landscape connectivity index), were used to analyze the landscape types of Jianghua Yao Autonomous County in 2000, 2010, and 2020 and under the 2030 natural development scenario. The changes in the landscape-type-level indices for Jianghua Yao Autonomous County are presented in

Table 5. Table of changes in landscape-level indices across different landscape types in 2000, 2010, and 2020, and under the future scenarios.

Index	Year	Grassland	Urban Areas	Farmland	Shrubland	Forestland	Water Bodies
NP	2000	888	2088	4422	1176	3388	669
	2010	665	2268	4422	540	3084	690
	2020	398	2505	4671	352	3274	826
	2030 natural development scenario	348	2236	5497	321	5760	826
PD	2000	0.2767	0.6506	1.3778	0.3664	1.0557	0.2085
	2010	0.2072	0.7067	1.3778	0.1683	0.9609	0.215
	2020	0.124	0.7805	1.4554	0.1097	1.0201	0.2574
	2030 natural development scenario	0.1084	0.6967	1.7128	0.1	1.7947	0.2574
LPI	2000	0.0091	0.0775	13.0452	0.0076	70.0895	0.0692
	2010	0.0088	0.1177	14.1049	0.0067	71.1536	0.0686
	2020	0.0013	0.2657	14.6248	0.005	68.4098	0.0786
	2030 natural development scenario	0.0012	0.4169	14.9926	0.005	69.7612	0.0786
LSI	2000	31.6239	47.936	77.5222	39.9444	35.8148	32.6587
	2010	28.6372	50.6951	77.992	25	34.6858	33.1347
	2020	21.7826	55.0362	85.3576	21.5392	38.1984	37.4522
	2030 natural development scenario	20.1111	58.3739	98.6724	20.8043	45.0351	37.4522
COHESION	2000	68.0901	85.7866	99.8203	78.5607	99.8851	94.1009
	2010	74.3591	88.1808	99.8298	76.8146	99.9049	93.5513
	2020	55.7652	94.1187	99.8302	75.8122	99.8978	92.775
	2030 natural development scenario	53.8298	94.8722	99.8247	75.133	99.9253	92.775

.

Based on the changes in grasslands, the COHESION index initially increased and then decreased, indicating that the connectivity among grassland patches initially strengthened but subsequently weakened. Meanwhile, NP, PD, LPI, and LSI continuously decreased and were projected to continue decreasing under the natural development scenario by 2030. This suggests a reduction in the fragmentation of grasslands, with a more sparse spatial distribution, an increase in the area of some patches, and a trend towards more regular patch shapes.

In urban areas, the PD and LPI indices initially increased and then decreased, indicating that the number and distribution density of urban patches first increased and then decreased. This reflects a process where the concentration and core areas of urban regions expanded and then contracted. Conversely, the NP, LSI, and COHESION indices continuously increased and were predicted to keep increasing under the natural development scenario by 2030. This reflects an accelerated urbanization process, with increased complexity in the shapes of urban patches, enhanced connectivity among patches, and an overall strengthening of the urban structure.

For croplands, the NP and PD indices remained stable from 2000 to 2010, but they showed an increasing trend from 2010 to 2020 and were projected to continue increasing under the natural development scenario by 2030. This indicates an increase in the fragmentation of cropland patches and a denser spatial distribution of croplands. The LPI, LSI, and COHESION indices also continuously increased and were predicted to keep rising under the natural development scenario by 2030. This suggests a significant increase in the area of certain cropland patches, increased complexity in patch shapes, and enhanced connectivity and overall structure of cropland patches. For shrubs, forests, and water bodies, the trends in the NP, PD, LPI, LSI, and COHESION indices exhibited similar patterns. Overall, the number and density of shrub and forest patches decreased, while the number and density of water body patches remained relatively stable.

4.2. Evolution of Ecosystem Service Value

4.2.1. Changes in Ecosystem Service Value

Based on the data of various landscape types and their ecosystem service values (ESVs) in Yao Autonomous County for the years 2000, 2010, and 2020 and under the future scenarios (

Table 6. The areas of various landscape types and their ecosystem service values (ESVs) in the years 2000, 2010, and 2020 and under the future scenarios.

Year	Farmland (hm2)	Forestland (hm2)	Shrubland (hm2)	Grassland (hm2)	Water Bodies (hm2)	Total ESV (CNY Billion)
2000	55,726.92	261,197.91	881.19	306.45	1425.6	3029.23
2010	58,980.06	257,939.19	309.69	286.56	1340.37	2999.68
2020	62,049.78	253,532.25	231.66	105.75	1660.05	2965.71
2030 nature growth scenario	64,554.84	249,851.43	190.26	86.85	1660.05	2933.19
2030 economic development priority scenario	70,244.91	243,540.54	188.37	69.93	1518.57	2879.59
2030 sustainable development scenario	57,578.22	256,988.79	155.61	85.95	1660.05	2989.11

), the following trends were observed:

From 2000 to 2020, the total ESV continuously declined, dropping from CNY 3029.23 × 10⁸ in 2000 to CNY 2965.71 × 10⁸ in 2020. This decrease was mainly attributed to the continuous reduction in the areas of forests, shrubs, and grasslands, which weakened the ecosystem service values of these land types, thereby leading to a decline in the total ESV. Meanwhile, the continuous expansion of cropland and the initial reduction followed by an increase in water area indicated that the expansion of agricultural activities contributed relatively less to the support and regulation services of the ecosystem, thus driving the decline in ESV.

Under the future scenarios, forest areas continue to decrease in the nature growth and economic development priority scenarios, while they recover somewhat under the sustainable development scenario. The reduction in forest areas has a significant negative impact on ESV, as forests play a crucial role in climate regulation, soil and water conservation, and biodiversity maintenance. Water areas remain relatively stable across future scenarios and continue to play important roles in providing water resources, regulating climate, and enhancing the environment.

The expansion of urban land exerts a crowding-out effect on other land use types, especially cropland and forest, thereby affecting ESV. Although the reduction in shrub and grassland areas has a relatively minor impact on ESV, the loss of their ecological functions still warrants attention. In the economic development priority scenario, rapid urbanization may lead to a further decline in ESV, whereas in the sustainable development scenario, measures such as controlling urban expansion and protecting forests can enhance ESV.

4.2.2. Ecosystem Service Functions and Values

Overall, the total ecosystem service value (ESV) of Jianghua Yao Autonomous County gradually declined from 2000 to 2020. Under the different scenarios for 2030, the extent of change in total ESV varies significantly. Specifically, under the 2030 nature growth and sustainable development scenarios, the total ESV shows minor fluctuations, while under the 2030 economic development priority scenario, the total ESV decreases to CNY 287.959 billion. This indicates that different land use scenarios have significantly different impacts on ESV.

In terms of individual ESVs, their importance ranks as follows: regulation services, support services, provisioning services, and cultural services (

Table 7. Ecosystem service values for 2000, 2010, and 2020 and the future scenarios (CNY billion).

Years	Provisioning Services	Regulation Services	Support Services	Cultural Services	Total ESV
2000	363.73	1534.01	915.72	215.77	3029.23
2010	361.11	1518.47	907.07	213.03	2999.68
2020	357.16	1502.79	895.60	210.16	2965.71
2030 nature growth scenario	353.79	1486.23	885.81	207.35	2933.19
2030 economic development priority scenario	348.74	1458.13	870.29	202.42	2879.59
2030 sustainable development scenario	359.15	1514.85	902.51	212.60	2989.11

). All ESV indicators show a declining trend, with regulation services having the highest value, making them the primary ecosystem service function in Jianghua Yao Autonomous County. Regulation services include climate regulation, hydrological regulation, and air purification, with the reduction in forest areas being the main reason for the decline in regulation service value. In contrast, cultural services (including tourism, recreation, and aesthetic value) show the smallest decrease. The expansion of urbanization has a relatively minor impact on cultural services, but changes in the ecological environment still have some negative effects on cultural services.

Under the 2030 nature growth and sustainable development scenarios, all ESVs show a declining trend. However, under the 2030 sustainable development scenario, there is potential for growth in all ESVs. This suggests that appropriate management and conservation measures can partially restore and enhance ecosystem service values.

From the above analysis, it is evident that the ecosystem service value (ESV) of Jianghua Yao Autonomous County underwent significant changes across different years and the future scenarios. The expansion of cropland and urban areas, while driving economic development, had a notable negative impact on ESV, particularly on regulation and support services. The reduction in forest areas significantly lowered the value of regulation and support services, whereas the relative stability of water areas played an important role in maintaining ESV. These changes indicate that land use changes have profound impacts on ecosystem service functions, especially in terms of regulation and support services.

4.3. Correlation Analysis between Landscape Pattern Indices and ESV

Further analysis showed that different landscape pattern indices had varying impacts on the ecosystem service values (ESVs). Therefore, we conducted a Pearson correlation analysis among the aforementioned landscape-level, landscape-type-level indices, and ESV using the correlation coefficient R as the evaluation metric (

Table 8. Correlations between landscape pattern indices and total ESV.

Landscape Pattern Index	Provisioning Services		Regulation Services		Support Services		Cultural Services		Total ESV
	R	p	R	p	R	p	R	p	R	p
NP	−0.672	0.328	−0.653	0.347	−0.655	0.345	−0.645	0.355	−0.655	0.345
FRAC_AM	−0.922	0.078	−0.923	0.077	−0.917	0.083	−0.918	0.082	−0.921	0.079
PLADJ	0.934	0.066	0.920	0.080	0.925	0.075	0.917	0.083	0.923	0.077
SHDI	−0.997 *	0.003	−0.990 *	0.010	−0.995 *	0.005	−0.990 *	0.010	−0.993 *	0.007
SHEI	−0.997 *	0.003	−0.990 *	0.010	−0.995 *	0.005	−0.990 *	0.010	−0.993 *	0.007
AI	0.934	0.066	0.920	0.080	0.925	0.075	0.916	0.084	0.923	0.077

Note: * means significant at the 10% level.

).

The correlation analysis results indicate that there are significant correlations between different landscape pattern indices and ecosystem service values (ESVs), closely related to the ecological functions of landscape characteristics. Specifically, the number of patches (NP) is negatively correlated with all service categories and total ESV (correlation coefficient R between −0.653 and −0.672, p-value between 0.328 and 0.355), suggesting that a high number of patches may lead to ecosystem fragmentation, affecting service provision, although this correlation is not statistically significant. The mean patch shape index (FRAC_AM) is significantly negatively correlated with ESV (R between −0.917 and −0.923, p-value between 0.077 and 0.083), indicating that complex patch shapes may increase edge effects, thereby reducing service efficiency. Patch adjacency (PLADJ) is positively correlated with ESV (R between 0.917 and 0.934, p-value between 0.066 and 0.083), reflecting that good connectivity among patches positively impacts service provision. The Shannon Diversity Index (SHDI) and Shannon Evenness Index (SHEI) are highly significantly negatively correlated with service categories and total ESV (R between −0.990 and −0.997, p-value between 0.003 and 0.010), suggesting that excessive landscape diversity may reduce ecosystem functions because of increased resource competition. Finally, the aggregation index (AI) is positively correlated with ESV (R between 0.916 and 0.934, p-value between 0.066 and 0.084), indicating that higher aggregation can effectively reduce the negative effects of fragmentation, enhancing ecosystem coherence and service provision capacity. These results highlight the importance of optimizing landscape structure to enhance ecosystem services and reveal the specific impact mechanisms of different landscape indices on ecosystem service functions.

5. Strategies for Landscape Pattern Optimization Based on ESV Enhancement

By analyzing the correlation mechanisms between landscape pattern indices and ecosystem service values (ESVs), the following landscape pattern optimization strategies are proposed to enhance ESVs:

• Overall Strategy: To achieve an overall increase in ESV, it is essential to balance ecological protection with economic development needs. Land use planning should adopt various measures to improve land use efficiency while protecting and restoring forests and water bodies and controlling the expansion rate and scale of urban and cropland areas [3]. Special attention should be given to protecting grasslands and shrubs to prevent encroachment by other land use types, thereby maintaining their ecological functions. Additionally, rational urban land layout planning, optimizing agricultural land structure, and promoting ecological agricultural technologies will help reduce environmental impacts and enhance the overall service functions of the regional ecosystem [30].
• Provisioning and Regulation Services: Expanding water areas through activities such as water sports or hot spring tourism can fully utilize the ESV of water bodies. Constructing ecological corridors [31], such as river corridors, green belts, and parks, can connect isolated ecological patches to form a complete ecological network, thereby enhancing the ecosystem’s regulatory capacity.
• Support Services: Protecting existing forests, restoring degraded ecosystems, and increasing green and shrub cover are crucial. Optimizing land use structure to improve landscape connectivity and stability will enhance the regional ecosystem service value. Additionally, developing agricultural experiences or rural tourism products unique to the Yao Autonomous County will help maintain landscape and natural environment diversity, supporting the core resources of tourism.
• Cultural Services: Attractive landscapes are the foundation for the development of tourism in the study area. Constructing aesthetically valuable hydrological and forest landscapes and integrating fragmented landscapes will enhance the overall beauty of Yao Autonomous County, thereby promoting tourism development.

These strategies aim to enhance ecosystem services by optimizing landscape structure and function, laying a solid foundation for regional sustainable development.

6. Discussion and Conclusions

This study analyzed the changes in land use types and landscape patterns in Jianghua Yao Autonomous County, Hunan Province, in 2000, 2010, and 2020 and under the natural development scenario for 2030, revealing the dynamic changes in land use and landscape patterns and their impacts on ecosystem service value (ESV). This study found significant reductions in grassland and shrub areas, with grassland area decreasing from 306.45 hectares in 2000 to 86.85 hectares in 2030, representing a reduction of 71.66%, and shrub area decreasing from 881.19 hectares in 2000 to 190.26 hectares in 2030, representing a reduction of 78.41%. Meanwhile, urban and cropland areas increased significantly, with the urban area expanding from 1399.32 hectares in 2000 to 4593.96 hectares in 2030, representing an increase of 228.30%, and the cropland area increasing from 55,726.92 hectares in 2000 to 64,554.84 hectares in 2030, representing an increase of 15.84%. These changes reflect the region’s rapid urbanization and agricultural expansion trends.

The landscape pattern indices of various land use types also showed significant changes. The number of patches (NP) and patch density (PD) generally indicated increased fragmentation of cropland and urban areas, while grassland and forest areas tended to become more contiguous. Changes in the largest patch index (LPI) and landscape shape index (LSI) indicated increased concentration and shape complexity of cropland and urban patches, while grassland and forest patches tended to have more regular shapes. The increase in the landscape cohesion index (COHESION) indicated enhanced connectivity between urban and cropland patches but decreased connectivity for grassland and shrub patches. These changes in landscape patterns significantly impacted the regional ecosystem service value, with provisioning service value increasing, but regulation and support service values declining. Particularly, under the economic development priority scenario, increased landscape fragmentation challenged the overall stability of the ecosystem.

This study shows that while rapid urbanization and agricultural expansion drove economic growth and food production to some extent, they negatively impacted regional ecosystem service functions. Especially under the economic development priority scenario, increased landscape fragmentation significantly reduced ESV, with declines in regulation and support service values, adversely affecting climate regulation, soil and water conservation, and biodiversity maintenance. In contrast, the ecological protection and sustainable development scenarios, through effective ecological protection measures such as expanding forest and water areas and reducing damage to grasslands and shrubs, significantly enhanced the regional ecosystem service value and maintained landscape pattern stability.

To balance ecological protection and economic development, future land use planning should comprehensively consider the needs of both. Various measures should be taken to improve land use efficiency, focusing on protecting and restoring forests and water bodies, controlling the expansion rate and scale of urban and cropland areas, and, particularly, strengthening the protection of grasslands and shrubs to prevent encroachment by other land use types, thereby maintaining their ecological functions. Additionally, rational urban land layout planning, optimizing agricultural land structure, and promoting ecological agricultural technologies will help reduce environmental impacts and enhance the overall service functions of the regional ecosystem.

One of the limitations of this study is the lack of consideration for Restricted Areas, such as traditional villages, in the FLUS model. This omission was primarily due to the unavailability of reliable data to accurately identify and locate these traditional villages within the study area. As a result, we were unable to incorporate these areas as restricted zones in the current modeling process. However, recognizing the importance of these areas for land use planning and ecosystem conservation, we acknowledge that future studies should focus on obtaining more comprehensive data to address this gap. Incorporating Restricted Areas into the FLUS model will improve the accuracy and reliability of land use scenario predictions.

Another limitation of this study is the focus solely on Jianghua Yao Autonomous County without comparing it to neighboring regions. While this focus allowed us to explore the unique ethnic, socio-economic, and ecological characteristics of this region in detail, we acknowledge that broader regional comparisons could provide further insights into how landscape pattern changes in Jianghua interact with, and are influenced by, surrounding areas. Future research could expand our scope to include adjacent regions to better understand the broader spatiotemporal dynamics at play while still considering the distinctive ethnic factors of Jianghua Yao Autonomous County. This would enhance the contextual understanding of land use and ecosystem service changes in a larger geographic context.