Identification of Land Use Conflict Based on Multi-Scenario Simulation—Taking the Central Yunnan Urban Agglomeration as an Example

Abstract

Land use conflict is an inevitable and objective phenomenon during regional development, with significant impacts on both regional economic growth and ecological security. Scientifically assessing the spatiotemporal evolution of these conflicts is essential to optimize land use structures and promote sustainable resource utilization. This study employs multi-period land use/land cover remote sensing data from China to develop a model for the measurement of land use conflict from the perspective of the landscape ecological risk. By applying the optimal landscape scale method to determine the most appropriate analysis scale, this research investigates the spatiotemporal evolution characteristics of land use conflicts in the Central Yunnan Urban Agglomeration from 2000 to 2020. Furthermore, by integrating the Patch-Generating Land Use Simulation (PLUS) model with the Multi-Objective Programming (MOP) algorithm, this study simulates the spatial patterns of land use conflict in 2030 under four scenarios: Natural Development (ID), Economic Development (ED), Ecological Conservation (PD), and Sustainable Development (SD). The findings reveal that, from 2000 to 2020, the proportion of areas with strong and moderately strong conflict levels in the Central Yunnan Urban Agglomeration increased by 2.19%, while the proportion of areas with weak and moderately weak conflict levels decreased by 1.45%, underscoring the growing severity of land use conflict. The predictions for 2030 suggest that the spatial pattern of conflict under various scenarios will largely reflect the trends observed in 2020. Under the ID scenario, areas with weak and moderately weak conflict levels constitute 57.5% of the region; this increases by 0.85% under the SD scenario. Conversely, areas experiencing strong and moderately strong conflict levels, which stand at 33.02% under the ID scenario, decrease by 1.04% under the SD scenario. These projections indicate that the SD scenario, which aims to balance ecological conservation with economic development, effectively mitigates land use conflict, making it the most viable strategy for future regional development.

1. Introduction

As industrialization and urbanization in China continue to accelerate, the demand for land resources has surged. This increase has deepened the inherent tension between the limited availability of land and the seemingly limitless human needs, exacerbating conflicts over land use [1]. The uncontrolled expansion of construction areas causes them to increasingly infringe upon farmland, forests, and other types of land, intensifying these conflicts. Such expansions adversely affect the structure and functionality of regional ecosystems and undermine the sustainable development of both economic and environmental systems, thus posing a serious threat to the long-term sustainability of urban regions. Consequently, identifying and addressing regional land use conflicts and implementing effective strategies to mitigate these human–land disputes are imperative in ensuring future sustainable development [2].

“Land use conflict” refers to the discrepancies and discord among various stakeholders regarding the methods and extent of land utilization, as well as conflicts between different land use practices and environmental concerns. If not adequately addressed, these conflicts could pose significant threats to sustainable urban development and land resource management. As such, achieving a balance between economic growth, social progress, and environmental protection has become a critical challenge for society [3,4]. Presently, land use conflict has become a prominent focus in international research. Due to the complexity of socioeconomic systems, land use conflicts display varied characteristics, including competitive pressures among different land use types, ecological risks stemming from disordered landscape patterns, and functional trade-offs among different land categories. In the early stages, technical limitations meant that the identification of land use conflicts primarily depended on expert assessments and participatory surveys [5], as well as methodologies such as game theory [6] and the Pressure–State–Response (PSR) conceptual model [7]. With advancements in technology, the introduction of 3S technologies has facilitated the use of multi-objective evaluation methods [8,9] and landscape ecological indices [10], providing a scientific basis for the characterization of the spatial features of land use conflicts and the development of differentiated governance strategies. Although previous studies have frequently utilized game theory to devise conflict resolution strategies, land use conflict is not solely a social issue—it is also a geographical phenomenon. Qualitative evaluation methods and governance strategies alone are inadequate in uncovering the spatial disparities and connections underlying these conflicts [1]. As research has evolved and technology has progressed, scholars have acknowledged that land use conflicts cannot be completely resolved or eliminated through a single governance strategy; rather, they can be mitigated and managed through spatial planning and other approaches [11]. The management of land use conflict involves not only addressing their impacts but also monitoring the dynamic processes and evolutionary trends that contribute to these conflicts, employing an interdisciplinary approach to achieve comprehensive governance [12,13]. From a temporal perspective, land use conflicts can be categorized into past, present, and future conflicts. While past and present conflicts have traditionally dominated the academic focus, recent research has extensively explored the simulation of future land use under various scenarios. Models such as system dynamics (SD) [14,15], ANN-CA [16,17], and CLUE-S [18] have been employed to simulate future land use developments. However, traditional land use forecasting models, often reliant on statistical or computational methods, struggle to fully capture the complexity and dynamism of land use changes. In contrast, the PLUS model significantly enhances the simulation accuracy [19]. The PLUS model integrates natural, social, and economic factors, enabling dynamic spatiotemporal simulations of various land use types at the patch level. Given that each scenario has distinct developmental sub-goals, the MOP algorithm is paired with the PLUS model to ensure that these sub-goals are met as effectively as possible. Multiple development scenarios are established in this study, with specific objective functions and constraints for each. This methodology allows for the simulation and prediction of future land use changes under differing policy priorities [20,21,22].

Currently, the swift development and urban expansion in Western China have accentuated land use conflicts, particularly due to the region’s complex terrain and challenging environmental conditions. Effectively identifying these conflicts and delineating a reasonable trajectory for territorial spatial development are critical prerequisites for the harmonization of regional socioeconomic ambitions with the goals of a sustainable ecological civilization. The Central Yunnan Urban Agglomeration, situated at the nexus of China’s Belt and Road Initiative and the Yangtze River Economic Belt, plays a pivotal role in China’s “Two Horizontal and Three Vertical” urbanization strategy. This rapid urbanization has led to significant transformations in the land use landscape, both urban and rural, highlighting the growing tension between economic and social development and ecological land use. This study focuses on the Central Yunnan Urban Agglomeration as a case study to develop a land use conflict measurement model based on landscape patterns. By applying optimal landscape scale analysis methods, it systematically explores the spatiotemporal evolution of the land use conflicts in the region from 2000 to 2020. Additionally, this study integrates the MOP-PLUS model to simulate land use conflict patterns under various development scenarios projected for 2030. This analysis aims to guide regional territorial spatial development, mitigate human–environment conflict, and promote the sustainable use of land resources alongside socioeconomic development.

2. Materials and Methods

2.1. Study Area

The Central Yunnan Urban Agglomeration is located in the eastern–central part of Yunnan Province (Figure 1), positioned between longitudes 100°43′73″ E and 104°82′40″ E and latitudes 22°99′73″ N and 27°03′19″ N. This urban cluster includes Kunming City, Qujing City, Yuxi City, Chuxiong Prefecture, and the northern part of the Honghe Hani and Yi Autonomous Prefecture, encompassing a total of 49 counties and cities. The total area spans 111,410 km², representing 29% of Yunnan Province’s total area. As of the end of 2020, the resident population of this area was approximately 21.95 million, accounting for 44.1% of the province’s total population. The regional gross domestic product (GDP) was CNY 1.51 trillion, constituting 61.47% of the province’s total GDP. The Central Yunnan Urban Agglomeration is located on the Yunnan–Guizhou Plateau and is characterized by a typical karst plateau topography. The terrain is predominantly mountainous and basin-like, with a gently undulating landscape that includes two-thirds of the province’s plains. This area boasts the richest dam resources in Yunnan Province, providing favorable conditions for rapid development and making it the most industrialized and urbanized region in Yunnan and the most developed area in the province. However, as construction land increasingly encroaches upon arable land and natural ecosystems, the conflict between urban space and ecological space has intensified. Consequently, the exploration of suitable development models for the future of the Central Yunnan Urban Agglomeration has become an urgent regional issue.

2.2. Data Sources

This study utilizes the LUCC 30 m land use data for the years 2000, 2010, and 2020, sourced from the Resource and Environment Science and Data Center of the Chinese Academy of Sciences. The data, with a resolution of 30 m, exhibited minor inconsistencies, such as fragmented voids and abnormal values. To rectify these issues, the focal statistics tool in ArcGIS 10.8 was employed, filling the majority of the values within each neighborhood to align with the research needs. The land use categories were reclassified into six primary types, namely cultivated land, forests, grassland, water, built-up land, and unutilized land, using the reclassification tool in ArcGIS. The specific factors influencing land use changes are listed in

Table 1. Data information.

Data Type	Data Name	Data Source
Human Factors	Population	GPW v4 Dataset (https://earthdata.nasa.gov/)
	GDP	China Gridded GDP Distribution Dataset
	Distance to Major Roads	OpenStreetMap
	Distance to Major Rivers
	Distance to Railways
	Distance to Train Stations	https://lbsyun.baidu.com/
	Distance to Government Offices
Natural Factors	DEM	Geospatial Data Cloud (http://www.gscloud.cn/)
	Slope
	NDVI	National Ecoscience Data Center
	Temperature	WorldClim v2.1 Climate Data (https://www.worldclim.org/)
	Precipitation
	Soil Type	HWSD v1.21 Soil Dataset (https://iiasa.ac.at/)
Limiting Factors	Nature Reserves
	Water Source Protection Areas

. Further standardization of the rows, columns, and spatial resolutions of both the land use data and driving factors was conducted, with the data reprojected to the WGS_1984_UTM_Zone_48N coordinate system using raster projection.

2.3. Research Methods

The research framework of this paper is divided into three parts:preprocessing of land use, human, natural, and limiting data; establishing population, GDP, DEM, and other factors as driving forces, using land use data from the Central Yunnan Urban Agglomeration for the years 2000, 2010, and 2020, and employing the MOP algorithm and PLUS model to predict the land use under four scenarios for 2030; calculating the land use conflict index and the spatial distribution of land use conflicts based on land use data from 2000 to 2020 and the 2030 multi-scenario simulation results through the land use conflict measurement model. As mentioned earlier, the research framework of this study is shown in Figure 2.

2.3.1. Land Use Conflict Measurement Model

This study develops a land use conflict measurement model grounded in the ecological risk assessment framework, which identifies risk sources, risk receptors, and risk effects from a landscape ecological risk evaluation perspective [23]. The model identifies spatial external pressure factors indicative of the intensity of spatial resource development and utilization as risk sources. It assesses the spatial ecological risk exposure, which represents the potential likelihood of risk occurrence under pressure from risk sources, using indicators related to the vulnerability of risk receptors. Risk effects are defined by the responses of different risk receptors when impacted by risk sources, with spatial unit stability factors chosen as indicators. As spatial units experience greater external pressure, the ecological risk exposure increases, and the spatial stability decreases, thereby elevating the likelihood of spatial ecological risks. This exacerbation in regional landscape ecosystem disturbance suggests intensified spatial conflicts.

The model synthesizes this concept into the equation “Land Use Conflict = External Pressure + Spatial Exposure − Spatial Stability”. It incorporates landscape ecology metrics such as the class area (CA), patch density (PD), and area-weighted mean patch fractal dimension (AWMPFD) to quantify the conflict levels [24]. The calculation methodologies are elaborated in

Table 2. Calculation methods for land use conflict indices.

Name	Calculation Formula	Significance
External Pressure	$P={\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n\left[\frac{2\ln (0.25P_{ij})}{\ln a_{ij}}(\frac{a_{ij}}{A})\right]}}$	Utilizes the AWMPFD to reveal the degree of disturbance and the influence of surrounding landscapes on the current landscape.
Vulnerability	$V={\displaystyle \sum _{i=1}^n{V}_i}\times {\displaystyle \frac{a_i}{A}}(n=6)$	Reflects the responsiveness of the land use system to external pressures. Vi represents the vulnerability index for each landscape type, and ai indicates the area of each type within a unit. Based on previous research [25,26], the values are set as follows: cultivated land—3, forest land—1, grassland—2, water—4, built-up land—6, and unutilized land—5.
Stability	$PD={\displaystyle \frac{n_i}{A}}S=1-{\displaystyle \frac{PD-P{D}_{\min }}{P{D}_{\max }-P{D}_{\min }}}$	Utilizes patch density to negatively reflect landscape stability, where ni is the number of patches of type i within a spatial unit, A is the total area of the spatial unit, and PD represents the patch density.
Land Use Conflict Index	$SCCI=P+V-S$	Reflects the degree of land use conflict within the evaluation unit; a higher value indicates stronger conflict. The classification of the land use conflict index is based on the cumulative frequency curve distribution and the inverted “U”-shaped evolutionary pattern of spatial conflict.

.

2.3.2. Optimal Landscape Scale Setting

The landscape pattern index is subject to a significant scale effect. In this study, the moving window method, implemented through the Fragstats 4.2 software, is used to calculate metrics based on a specified grid size. An excessively large or small grid size can considerably impact the results of landscape index calculations. Therefore, selecting an appropriate landscape grain size and analysis extent is crucial in enhancing the accuracy of these calculations. This study employs an optimal landscape scale selection method to determine the appropriate landscape grain size and analysis extent for the study area.

2.3.3. Area Information Loss Evaluation Model

The area information loss evaluation model is an effective tool for the quantitative assessment of the accuracy of scale conversion [27]. This model is utilized in this study to evaluate the total area loss within the research area, which aids in determining the appropriate grain size for landscape pattern analysis. The primary method involves using the area of each land cover type before scale conversion as the baseline. Subsequently, the areas of each type at the target scale post-conversion are compared to the baseline areas to determine the absolute or relative values of area loss at different scales. Smaller loss values, whether absolute or relative, indicate the more effective performance of the scale conversion method. This model enables the assessment of the area loss not only for specific types in spatial data but also for the overall area loss across the region. The formula for this model is as follows:

$$L_{\mathrm{i}}=(A_i-A_{bi})/A_{bi}\times 100$$

(1)

$$S_i=\sqrt{{\displaystyle \sum _{i=1}^n{L}_i^2/n}}$$

(2)

In the formula, L_i represents the relative value of the lost area; A_i denotes the area of landscape class i after grain size conversion; A_bi indicates the area of landscape class i before grain size conversion; n is the number of land use types; and S_i signifies the change in land area for the region. The larger the value of S_i, the higher the loss of precision in the area of the land type.

2.3.4. Optimal Selection of Landscape Grain Size and Extent

The landscape pattern indices employed in this study, facilitated by Fragstats 4.2, are categorized into three hierarchical levels: the patch level, class level, and landscape level. The patch and class levels focus on the analysis of individual patches and different types of patches, respectively. In contrast, the landscape level indices encompass an assessment of the overall characteristics within the study area. This research adopts landscape pattern indices that reflect three aspects of the landscape structure at the landscape level: fragmentation, aggregation, and heterogeneity.

(1)

The selection of the landscape scale involves nine indices: the class area (CA), division (DIVISION), largest patch index (LPI), splitting index (SPLIT), effective mesh size (MESH), Shannon’s diversity index (SHDI), Shannon’s evenness index (SHEI), aggregation index (AI), and area-weighted mean patch fractal dimension (AWMPFD) [28]. The analysis begins at a scale of 30 m and is extended to 300 m, with intervals of 30 m, to evaluate the response effects of these indices. By examining the trends in the scale response curves, appropriate scales for scale analysis are determined. Concurrently, the area information loss evaluation model is utilized to quantitatively assess the total area loss at various granularities, ultimately establishing the optimal scale size for landscape pattern analysis within the study area.

(2)

Drawing on methodologies from existing studies on optimal landscape amplitude analysis [29], and considering the geographical and ecological characteristics of the Central Yunnan Urban Agglomeration, a sampling belt oriented from southwest to northeast is established (Figure 3). This belt, which is 3 km wide and 22 km long, encompasses significant urban and natural landscapes, including Yuxi, the Dianchi Lake basin, and the main urban areas of Kunming and Qujing, covering a diverse range of land uses. Along the centerline of the belt, 23 sampling points are designated from southwest to northeast, numbered 1 through 23. Landscape index values are gathered using the moving window method across different amplitudes. The optimal analysis amplitude is determined by analyzing the trends in the landscape index value curves across these points. Using the 2020 land use data as a reference, four landscape indices are selected for detailed analysis: the patch density (PD), Shannon’s diversity index (SHDI), division (DIVISION), and largest patch index (LPI) [30,31]. Amplitude sizes are set at integer multiples of 30 m, up to a maximum of 3600 m, to calculate the landscape indices using moving windows of varying side lengths.

2.3.5. PLUS Model

The PLUS model integrates a land expansion analysis strategy with a cellular automata (CA) model, utilizing multiple types of random patch seeds. This integration not only facilitates a comprehensive analysis of the driving mechanisms behind land use changes but also enables precise predictions and simulations of the landscape dynamics under various scenarios. Compared to other models, such as CLUE-S and FLUS, the PLUS model enables significant advancements in the depth of transformation rule extraction and in simulating landscape dynamics. It is particularly adept in simulating the transformation processes of various land use types at the patch scale and can be effectively coupled with the MOP algorithm [32]. This capability provides robust support for planning decisions and could promote sustainable development within the study area.

The present study utilizes land use data from the Central Yunnan Urban Agglomeration for the years 2010 and 2015 to simulate the land use conditions for 2020. The simulation results are subsequently compared with actual land use data from 2020 to validate the accuracy of the PLUS model. The land use demand for 2020 is predicted using a Markov chain approach to determine the number of required land use pixels. The intensity of land use expansion is measured by the change in the patch area, following the method proposed by Wang B. S. [33]; see Equation (3). This approach is employed to calculate the domain weight of the land use conditions (

Table 3. Setting of domain weights.

Land Use Type	Cultivated Land	Forest	Grassland	Water	Built-Up Land	Unutilized Land
Domain Weights	0.44	0.16	0.07	0.33	0.93	0.06

). The Kappa coefficient for accuracy testing is 0.908, indicating that the PLUS model exhibits high simulation accuracy and is suitable for future land use simulations.

$$X_{\mathrm{i}}={\displaystyle \frac{T_{Ai}-T{A}_{\min }}{T{A}_{\max }-T{A}_{\min }}}$$

(3)

In the formula, X_i represents the domain weight parameter for land use type i; T_Ai denotes the expansion area of land use type i; TA_min indicates the minimum expansion area among all land use types; and TA_max represents the maximum expansion area among all land use types.

2.3.6. Multi-Scenario Simulation Settings

(1)

Natural Development Scenario (ID). This scenario adheres to a path of natural development, informed by land use changes from 2000 to 2020, and incorporates both natural and anthropogenic factors. It designates areas for natural protection and water source conservation as restricted zones. The Markov chain model is employed to forecast the demand for land use pixels in 2030 across various land use types.

(2)

Ecological Protection Scenario (PD). Amid the rapid expansion of the Central Yunnan Urban Agglomeration, the increasing threats to the ecological environment from human activities have become more pronounced. In response, the PD scenario prioritizes ecological development by valuing ecosystem services. This approach facilitates the prediction and planning of a land use structure aimed at optimizing the ecological benefits and maximizing the ecological efficacy. The calculation formula for the ecosystem service value is as follows:

$$f_1(x)={\displaystyle \sum _{i=1}^{6}es{v}_i\times x_i}$$

(4)

In the formula, esv_i denotes the ecosystem service value per unit area for land use type i (unit: CNY ten thousand per hectare); f₁(x) reflects the total benefit of the ecosystem service value. The methodology for the calculation of the ecosystem service value is derived from the research conducted by Xie G. D. [34]. Additionally, based on the findings of Costanza et al. [35] on global ecosystem services, the ecosystem service value coefficient for construction land is set to zero. The formula can thus be modified to

$$\max f_1(x)=0.5x_1+4.67x_2+1.55x_3+16.12x_4+0x_5+0.08x_6$$

(5)

(3)

Economic Development Scenario (ED). As the most significant area for economic growth in Yunnan Province, the Central Yunnan Urban Agglomeration should consider a “GDP-only” perspective in its strategic planning. This approach aims to highlight the benefits and challenges associated with potential land use conflicts resulting from unregulated development. The scenario is designed to prioritize economic benefits and involves establishing a multi-objective constrained objective function, detailed as follows:

$$f_2(x)={\displaystyle \sum _{i=1}^{6}e{c}_i\times x_i}$$

(6)

In the formula, ec_i represents the economic benefit per unit area for land use type i (unit: CNY ten thousand per hectare); x_i denotes the area of land use type i, where x₁ through x₆ correspond to cultivated land, forests, grassland, water, built-up land, and unutilized land, respectively; and f₂(x) denotes the total economic benefit. Drawing from the research by Wang Y et al. [36] and data from the “Yunnan Statistical Yearbook”, economic values for arable land, forest land, grassland, and water areas are calculated using output values from agriculture, forestry, animal husbandry, and fisheries. The economic value of construction land is estimated from the outputs of secondary and tertiary industries, while unused land is considered economically inactive. Based on economic data from 2000 to 2020, the economic benefits for each land type in 2030 are projected using the gray prediction model GM(1,1). The revised formula is expressed as follows:

$$\max f_2(x)=2.23x_1+0.22x_2+1.8x_3+0.06x_4+24.85x_5+0x_6$$

(7)

(4)

Sustainable Development Scenario (SD). The Central Yunnan Urban Agglomeration is a pivotal area for new-type urbanization in Yunnan Province. Accelerating its development represents a crucial strategy for the promotion of regionally coordinated growth. The “Central Yunnan Urban Agglomeration Development Plan” highlights the vision of creating an ecologically livable urban cluster on the plateau and establishing an ecological barrier to safeguard the regional ecological security. Accordingly, this scenario integrates ecological protection with economic development, aiming to optimize both aspects simultaneously. In this scenario, natural conservation and water source protection areas are incorporated into the simulated restricted zones. The goal is to maximize economic development while also achieving the maximum ecological benefits, ensuring the sustainable development of land use. The multi-objective optimization function can be expressed as follows:

$$MAX{\left\{f\right.}_1(x),\left.f_2(x)\right\}$$

(8)

(5)

Constraint Setting. Building on established research methodologies, relevant policy directives, and the actual conditions of the study area, the constraints for the objective functions are as specified in

Table 4. Multi-objective planning constraints.

Constraint Type	Constraint Conditions/(CNY Ten Thousand, hm2)	Remarks
Total Land Use Area	$S={\displaystyle \sum _{i=1}^6{x}_i=\mathrm{11,141,026.47}}$	The total land use area in the study area remains unchanged.
Cultivated Land Area	$\mathrm{2,221,300}\le x_1\le \mathrm{2,254,262.22}$	The cultivated land area has been decreasing and converted into other uses over the past few decades, which is irreversible. Thus, the cultivated land area in 2020 is set as the upper limit, with the lower limit established based on the “National Land Use Overall Planning Outline (2006–2020) Adjustment Plan”.
Forest Land Area	$\mathrm{5,468,022}\le x_2\le \mathrm{6,032,279.68}$	The area in the Central Yunnan Urban Agglomeration in 2020 serves as the lower limit. The upper limit is set at 1.1 times the Markov-chain-predicted area for 2020, based on existing research [37].
Grassland Area	$\mathrm{2,948,925.15}\le x_3\le \mathrm{3,638,433.17}$	The minimum area is set based on the inertial development scenario [38], while the maximum area is set at 120% of the Markov-chain-predicted area for 2020, according to previous studies.
Water Area	$\mathrm{146,503.62}\le x_4\le \mathrm{158,870.79}$	The area of water bodies in the Central Yunnan Urban Agglomeration has been increasing annually. The area in 2020 is set as the lower limit, with the upper limit based on the Markov chain prediction for the natural development scenario in 2030.
Built-Up Area	$\begin{array}{l}\mathrm{266,209.65}\le x_5\le \mathrm{399,314.48} (economic priority)\\ \mathrm{266,209.65}\le x_5\le \mathrm{319,451.58} (ecological priority)\\ \mathrm{266,209.65}\le x_5\le \mathrm{346,072.55} (sustainable development)\end{array}$	The built-up area in 2020 is set as the lower limit based on existing research [37]. The maximum values for different scenarios are set as 1.5 times (economic priority), 1.2 times (ecological priority), and 1.3 times (sustainable development) the original area.
Unutilized Land Area	$\mathrm{15,526.8}\le x_6\le \mathrm{16,037.01}$	The area of unutilized land is set with the upper limit based on the natural development scenario and the lower limit based on the area in 2000.

. These constraints, along with the objective functions, are optimized using the MOP algorithm and subsequently processed in LINGO 18.0 to determine the optimal land use allocations under the PD and ED scenarios.

3. Results

3.1. Analysis of Optimal Landscape Scale

3.1.1. The Best Granularity Selection

Based on a 30 m resolution, landscape indices were computed at various scales ranging from 30 m to 300 m. The trends of these indices across the different scales were graphically represented (Figure 4). As the scale increased, significant variations were observed in all landscape indices. The fluctuations were relatively stable within the 30–120 m range, whereas they became more pronounced from 120 m to 300 m. Consequently, the optimal scale for analysis was determined to be within the range of 30 m to 120 m. The area information loss evaluation model was applied to assess the landscape area loss at these scales (Figure 5). The findings indicated that setting the scale at either 60 m or 120 m minimized the precision loss of the landscape area within the region. However, the precision loss rose sharply beyond the 120 m scale. Therefore, 120 m was established as the most suitable scale for detailed landscape analysis.

3.1.2. Optimal Amplitude Selection

At a base resolution of 30 m, integer multiples of this measurement were used to define the amplitude scales for the sampling strip, with the selected scales including 600 m, 1200 m, 2400 m, 3000 m, and 3600 m. The moving window method facilitated the selection of 23 sampling points along the strip’s centerline, from which the landscape index values were derived at varying amplitudes (Figure 6). An analysis of the Shannon’s diversity index revealed that the amplitude fluctuations were the most significant at 600 m and 3600 m, while being more subdued at 2400 m and 3000 m. The patch density index showed considerable fluctuations at 600 m and 1200 m, but these fluctuations diminished and stabilized at higher amplitudes, particularly at 2400 m and 3000 m. The proportion of the landscape area occupied by the largest patch displayed relatively uniform fluctuations across all amplitudes, with the smallest fluctuations at 3000 m. The division index exhibited considerable overall fluctuations, with similar patterns observed at 2400 m, 3000 m, and 3600 m, although the fluctuations at 3000 m were notably smoother. Given the variability in the landscape indices at different amplitudes and the size of the study region, an amplitude of 3 km × 3 km was selected as the most appropriate scale to conduct landscape analysis. This selected scale will underpin further calculations of landscape indices that are pertinent to land use conflicts within the study area.

3.2. Spatiotemporal Evolution of Land Use Conflict

Employing the land use conflict measurement model, this analysis computed the spatiotemporal patterns of land use conflict within the Central Yunnan Urban Agglomeration from 2000 to 2020. As depicted in Figure 7, the areas of high conflict are predominantly concentrated in the relatively low-lying central regions, specifically within the urban centers of Kunming, Yuxi, Chuxiong, and Qujing. The Central Yunnan Urban Agglomeration is identified as a pivotal area for economic development in Yunnan Province, characterized by strong dependency on limited land resources, which has precipitated intense conflict within these urban areas. During the accelerated phase of urbanization in the Central Yunnan Urban Agglomeration, the expansion of built-up land frequently encroached cultivated land and grassland territories, thereby converting them into zones of heightened land use conflict. In recent years, the vicinities of Dianchi Lake and Fuxian Lake, known for their fertile soils, have seen an escalation in conflicts between agricultural pursuits and river ecological protection, exacerbated by phosphate mining and urban development. As the second-largest city in Yunnan and a significant economic nucleus, Qujing has increasingly prioritized the development of both light and heavy industries, further intensifying the conflict between industrial expansion and ecological conservation in the region. The data from

Table 5. Proportions of land use conflict levels from 2000 to 2020 (%).

	Stable and Controllable	Weak Conflict	Moderately Weak Conflict	Moderately Strong Conflict	Strong Conflict
2000	10.33	25.75	34.02	21.06	8.84
2010	9.83	25.93	33.53	21.49	9.21
2020	8.23	24.92	33.40	22.22	9.87

clearly show that, between 2000 and 2020, the proportion of areas classified as having stable and controllable conflict levels decreased by 2.1%, and the proportions with weak and moderately weak conflict levels declined by 0.83% and 0.62%, respectively. In contrast, the proportion of areas with moderately strong conflict levels increased by 1.16%, and the proportion of those with strong conflict levels rose by 1.03%. These trends indicate a significant rise in land use conflicts over the past two decades within the Central Yunnan Urban Agglomeration, underscoring the urgent need for the formulation and implementation of effective development strategies to mitigate these conflicts.

Overall, the land use conflict index for the Central Yunnan Urban Agglomeration demonstrates a rising trajectory, with only a few regions remaining classified as stable and controllable, as illustrated in Figure 8. Notably, major cities such as Kunming, Qujing, and Yuxi have recorded conflict increase values exceeding 0.2, while the smaller surrounding cities have seen increases ranging from 0.05 to 0.2. Conversely, the conflict levels associated with agricultural land, forests, and grassland located farther from urban centers show a relatively stable trend. This pattern suggests that land use conflicts, as manifestations of the inherent contradictions in human–land relationships, are persistent, long-term, and stable. Therefore, the resolution of these conflicts is a complex endeavor that cannot be achieved swiftly; it demands sustained efforts and strategic interventions.

3.3. Multi-Scenario Simulation of Land Use Conflict

3.3.1. Overall Analysis of Land Use Conflict

According to the PLUS model simulation of the land use in the Central Yunnan Urban Agglomeration for 2030, landscape indices under various scenarios were computed with the Fragstats 4.2 software. This analysis resulted in the spatial distribution of land use conflict indices for these scenarios. The indices were classified into five categories by employing the natural breaks method: stable and controllable, weak conflict, moderately weak conflict, moderately strong conflict, and strong conflict (Figure 9). The projected land use conflict distribution for 2030 shows the continuation of the patterns observed from 2000 to 2020 (Figure 7), with more pronounced conflicts in the central region of the Central Yunnan Urban Agglomeration, notably in Kunming, Qujing, Yuxi, and Chuxiong. This trend underscores the inherent challenges posed by rapid urbanization. In the Central Yunnan Urban Agglomeration, the conflict levels are higher in the relatively flat, urban basin areas of the central region compared to the northern mountainous areas. This indicates that urban development is significantly influenced by the topographical diversity of this plateau region.

To conduct a more detailed analysis of the effectiveness of land use conflict mitigation under different development scenarios, the area proportions of land use conflicts at various levels from 2000 to 2020, as well as projections for 2030 under different development scenarios in the Central Yunnan Urban Agglomeration, were calculated and compared. The dynamic trends in the proportions of land use conflicts categorized as strong, moderately strong, weak, and moderately weak across these scenarios were illustrated (Figure 10).

From 2000 to 2020, there was a consistent, year-by-year decline in the proportions of weak and moderately weak land use conflict areas within the Central Yunnan Urban Agglomeration. The simulations for 2030 under three scenarios indicate that the proportions of weak and moderately weak conflicts will continue to be lower than those observed in 2020, adhering to the historical trends. Specifically, in the SD scenario, the proportion was the highest at 58.35%, closely followed by the ED scenario at 58.2% and the PD scenario at 57.84%. The ID scenario recorded the lowest proportion at 57.5%. This suggests that continuing the same historical development patterns may perpetuate the increase in land use conflict, emphasizing the critical need for coordinated and balanced conflict management strategies. The frequency of strong and moderately strong conflict levels shows an initial decline followed by an increase from 2000 to 2020. This pattern indicates that although the Central Yunnan Urban Agglomeration experienced some relief from land use conflict between 2000 and 2010, the subsequent decade saw a significant rise in conflict levels amid rapid development. By 2020, the proportion of areas with strong and moderately strong conflicts had reached 32.09%, underscoring the urgent need for effective land use conflict management. In the 2030 simulation results, across the four scenarios, the ID scenario exhibited the highest proportion of strong and moderately strong conflicts at 33.02%, followed by the PD scenario at 32.6% and the ED scenario at 32.3%, with the SD scenario showing the lowest at 31.98%, which represents a decrease of 1.04% compared to the ID scenario. These outcomes corroborate the findings related to weak and moderately weak conflicts, reiterating that the development strategies in the Central Yunnan Urban Agglomeration should not merely follow a trajectory of natural growth. Unrestricted large-scale urban expansion could further exacerbate regional conflicts.

In summary, the trajectory of the land use conflict levels in the Central Yunnan Urban Agglomeration indicates that the degree of conflict is expected to escalate between 2020 and 2030. The land use strategies implemented under the PD and ED scenarios can effectively mitigate these conflicts; however, each scenario primarily focuses on either ecological protection or economic development, thereby neglecting a balanced approach. This oversight could significantly hinder the sustainable development of central cities and may not be effective across the entire study area. From a broader regional development perspective, the SD scenario, which integrates considerations for both economic growth and ecological protection, somewhat alleviates land use conflicts and establishes a more robust and sustainable foundation for regional development. Consequently, it represents a more suitable strategic path for the area’s future development.

3.3.2. Local Analysis of Land Use Conflict

(1)

The main urban area of Kunming, the capital of Yunnan Province (Figure 11), has been selected as the focal point for this analysis. As the primary core urban district, Kunming experiences a high level of land use conflict. Notably, the ID scenario uncovers a wider range of conflict zones. As detailed in

Table 6. The transfer of local land area under each scenario from 2020 to 2030 (hm²).

Area 1	1–2	1–3	1–4	1–5	2–3	2–4	2–5	3–4	3–5	4–5
2020–2030 ID	−211.86	−81.27	11.88	445.59	−258.21	0	789.66	1.62	338.58	0
2020–2030 ED	0	−273.96	0	534.51	0	0	0	0	523.17	0
2020–2030 PD	241.47	0	17.19	0	−377.55	0	0	−0.81	0	0
2020–2030 SD	−362.34	−169.2	−64.8	589.59	−465.03	−0.27	1631.7	1.35	445.77	18.45
Area 2
2020–2030 ID	−233.73	−166.32	648.54	5445.81	−302.31	0.09	655.47	22.05	327.42	0
2020–2030 ED	0	−397.26	0	4487.76	0	0	0	0	468.72	0
2020–2030 PD	201.15	0	738.09	0	−558.81	0	0	11.34	0	0
2020–2030 SD	−360.45	−333.09	−359.1	6737.49	−584.46	0	1458.09	25.2	434.7	228.15
Area 3
2020–2030 ID	−815.31	−317.97	27.36	923.04	−16.56	0	20.16	2.34	29.79	0
2020–2030 ED	0	−732.42	0	904.77	0	0	0	0	42.03	0
2020–2030 PD	135.9	0	39.33	0	−135.9	0	0	1.35	0	0
2020–2030 SD	−1780.65	−635.49	−171.27	1336.41	−40.68	0	76.05	0.81	36.45	−0.09

Note: The numbers 1 to 5 represent arable land, forest land, grassland, water, and construction land, respectively. Positive values indicate a transfer relationship from 2020 to 2030, while negative values indicate the opposite.

, built-up land encroaches on 445.59 hectares of arable land, 789.66 hectares of forest land, and 338.58 hectares of grassland, thereby intensifying the competition among different types of land resources. Given Kunming’s role as the administrative, financial, and principal development hub of Yunnan Province, its land resources are under considerable strain. Without suitable policy interventions for future development, these severe conflicts are likely to worsen. Thus, when contemplating the future development trajectory of the main urban area, it is crucial to consider the trends in urban economic growth. When comparing the ED and SD scenarios, the extent of land use transition is relatively limited in the ED scenario, which primarily involves converting 534.51 hectares of cultivated land and 523.17 hectares of grassland into built-up land. Conversely, the SD scenario demonstrates a more balanced approach by converting cultivated land, forest land, and grassland into built-up land while also facilitating the transfer of land for other uses. This indicates a balanced development strategy that equally prioritizes urban growth and ecological protection. The SD scenario not only supports future urban economic development but also highlights the importance of ecological conservation.

(2)

The cities of Gejiu and Mengzi in Honghe Prefecture have been selected for the ecological protection analysis. As outlined in the “Yunnan Province Biodiversity Conservation Strategy and Action Plan for Gejiu City Implementation (2018–2030)”, this region is identified as a priority area on the southern edge of Yunnan’s tropical rainforest, highlighting its significance for biodiversity conservation. National policies and effective interventions have successfully limited human activities, allowing the ecosystem to remain relatively stable, which, in turn, results in a lower degree of land use conflict. In the analysis of the different simulated scenarios, the ID scenario follows the historical linear trends and shows the substantial conversion of cultivated land, forest land, and grassland into built-up land, totaling 5445.81 hectares, 655.47 hectares, and 327.42 hectares, respectively. This significant shift contributes to a higher level of conflict in this scenario compared to the others. The PD scenario shows lower levels and extents of land use conflict, with arable land transitioning to forest land and water bodies, covering areas of 201.15 hectares and 738.09 hectares, respectively, and grassland being converted to forest land across 558.81 hectares. This scenario prioritizes ecological protection without conflicts between built-up land and other land uses. Conversely, in the SD scenario, there is the notable conversion of cultivated land to built-up land over an area of 6737.49 hectares, while forest land, grassland, and water bodies are repurposed to cultivated land and forest land transitions to grassland over areas of 1052.64 hectares and 584.46 hectares, respectively. As a significant development area in Honghe Prefecture, it is essential to balance economic development with ecological protection, in alignment with the strategies outlined in relevant policies.

(3)

The agricultural regions within Qujing City have been selected to represent cultivated land for this analysis. There are significant variations in the degree of land use conflict across the different scenarios. In the ID scenario, the level of land use conflict in the arable regions of Qujing City peaks due to rapid economic and social development, leading to built-up land encroaching upon surrounding cultivated and other types of land, thus intensifying the conflict. In contrast, the SD scenario minimizes both the conflict level and the extent of the conflict area. This is achieved by restricting the conversion between cultivated land, forest land, and built-up land, effectively preventing the encroachment of construction land on arable and forest areas to the extent observed in the ID scenario. This approach not only promotes economic development but also prioritizes ecological protection. Therefore, under the sustainable development scenario, the degree of land use conflict is more favorable compared to the other scenarios.

4. Discussion

4.1. Land Use Changes in the Central Yunnan Urban Agglomeration

This study of the land use changes in the Central Yunnan Urban Agglomeration from 2000 to 2020 indicates that the primary transformation has been the expansion of built-up land, encroaching on the surrounding ecological lands, including cultivated land, grassland, and forest land. Taking the urban core of Kunming as an example, built-up land has occupied 445.59 hectares of cultivated land, 789.66 hectares of forest land, and 338.58 hectares of grassland. This expansion has exacerbated the land use conflicts in the surrounding areas, intensifying the competition for land resources. This phenomenon is driven by several factors. Rapid urbanization has led to the continuous expansion of the industrial scale, fostering rapid economic growth. Consequently, the population in the surrounding rural areas has begun migrating into the urban centers. Amid this shift, the demand for urban infrastructure and housing has surged, facilitating extensive urban expansion. This trend aligns with the findings from previous research [39]. The simulation results for four scenarios in 2030 indicate that the areas with high levels of land use conflict in the Central Yunnan Urban Agglomeration continue to be concentrated around Central Kunming, Qujing, Yuxi, and Chuxiong. Kunming, as Yunnan’s largest city and the primary hub of the agglomeration, along with Qujing and Yuxi, both significant urban centers, have experienced the rapid expansion of built-up land. This expansion has led to substantial and dramatic changes in land use, with the increasing conversion of other land types to built-up areas. In contrast, cities like Chuxiong, with less developed economies, exhibit slower rates of built-up land expansion, resulting in fewer land use conflicts. This outcome aligns with the findings of Deng et al. [40].

4.2. Multi-Scenario Simulations of Land Use Changes

As demonstrated in this study, the integration of the MOP algorithm with the PLUS model in multi-scenario simulations effectively facilitates the rational allocation of land resource quantities and the optimization of spatial configurations. The establishment of multi-objective planning constraints transforms the challenge into a multi-level relationship. By imposing constraints derived from higher-level conditions, the problem is decomposed into several interconnected sub-problems, each contributing a solution that, when aggregated, yields the optimal resolution for the primary issue. This approach significantly surpasses traditional land use change simulation models in accurately simulating spatial patterns of land use that align with actual land configurations. By merging the top-down quantity predictions of various land use types from the MOP algorithm with the bottom-up spatial optimization capabilities of the PLUS model, this integration simultaneously optimizes both the quantitative structure and the spatial layout of land use. This synergy enables comprehensive coordination and optimization, spanning from a holistic perspective to a localized focus.

Employing the coupled MOP-PLUS model, four distinct scenarios were simulated for the study area, focusing on projections for 2030. Predicting and simulating future land use changes provides invaluable insights, enabling decision-makers to devise rational land allocation strategies that effectively mitigate conflicts arising from urbanization, economic development, and ecological protection. The simulations revealed notable variations in the land use changes across different development scenarios. In the ED scenario, the expansion of built-up land was the most pronounced, with major cities like Kunming and Qujing converting substantial areas of cultivated land and grassland into construction zones for infrastructure development. Conversely, the PD scenario maintained a constant area of built-up land, predominantly featuring transformations between cultivated land and ecological land. The SD scenario demonstrated the more judicious expansion of built-up land, where conversions from cultivated, forest land, and grassland to built-up land were balanced by reciprocal transitions from other land types into these categories, fostering a dynamic equilibrium. This model underscores the importance of balancing economic growth with ecological preservation. According to the analysis and the policies proposed, this model is particularly applicable to cities in regions with complex topographies, such as Guangxi, Guizhou, and Sichuan, corroborating the findings of Shu et al. [41]. To realize the objectives of the SD scenario, policymakers must enforce stringent environmental governance measures in areas experiencing high levels of developmental conflict. It is crucial to maintain ecological protection boundaries, enhance land use management, and reconcile urban expansion with ecological conservation. Additionally, integrating the construction of green infrastructure into urban planning and establishing a robust environmental monitoring and evaluation system are essential steps toward sustainable urban development.

4.3. Research Limitations

This study quantitatively analyzed the future trends of land use change in the Central Yunnan Urban Agglomeration using the PLUS model, which incorporated 13 influencing factors for predictive simulations. However, it did not delve into the underlying mechanisms or assess the impact level of each factor on the predictions. This omission may result in a lack of in-depth understanding of the fundamental causes of land use change. Future research should focus on exploring the driving effects of the influencing factors on predictive simulations. For instance, the methods used by He et al. [42] to examine the interactions among influencing factors could provide insights for subsequent studies.

In the scenario simulations, numerous potential uncertainty factors exist. Land use change is a complex and dynamic process. Although this study considered the influences of the natural environment, human development, and local policy constraints, it did not adequately address factors such as the geological conditions, biodiversity, and climate change. Future research should aim to include these elements in the model to improve the accuracy of the simulations. This approach will ensure that the simulations are more scientifically sound and rational, thereby supporting the region’s pursuit of sustainable development.

5. Conclusions

(1)

Through this research, it has been established that a grain size of 120 m and a range of 3000 m are the optimal scales for the study area. This study assessed grain sizes from 30 to 300 m and ranges from 600 to 3600 m for the calculation of landscape indices. The resulting trend curves, which depict variations in the landscape index according to different grain sizes and ranges, demonstrate that adjusting these parameters induces varying levels of fluctuation in the index values. Notably, within the grain size interval of 30–120 m, the fluctuations are relatively stable. The area information loss evaluation model was utilized to ascertain that a grain size of 120 m provides the most suitable analysis scale for this area. Furthermore, trend graphs of the landscape index variations across different ranges indicate that a range of 3000 m offers the most appropriate analysis scope for this research area.

(2)

The issue of historical land use conflict in the Central Yunnan Urban Agglomeration is notably severe. An analysis of the land use conflict levels and the spatiotemporal distribution of these conflicts from 2000 to 2020 reveals that high-conflict areas are predominantly located in the relatively flat basin regions of Central Yunnan, specifically within the cities of Kunming, Qujing, Yuxi, and Chuxiong. These locales have endured intense land use conflict driven by rapid economic development, characterized by the relentless expansion of built-up land at the expense of cultivated land, forests, and grassland. Over the past two decades, the proportion of land use that is classified as stable and controllable has decreased by 2.1%, while the proportions of areas experiencing weak and moderately weak conflict have declined by 0.83%. In contrast, the proportions experiencing strong and relatively strong conflict have risen by 1.16%, signaling an increase in the severity of land use conflict issues within the region.

(3)

Future sustainable development scenarios may emerge as suitable development models for the Central Yunnan Urban Agglomeration. Analyzing the land use layout under four distinct scenarios projected for 2030 reveals that tailored policies are essential for different developmental contexts. For example, in the PD scenario, the ecological protection area in Honghe Prefecture shows promising advancements that align with the biodiversity conservation objectives outlined in relevant policies. In contrast, Kunming is positioned for sustained rapid economic growth under the ED scenario. This highlights the need for context-specific strategies that address both ecological integrity and economic development in the Central Yunnan Urban Agglomeration.