Spatial Simulation of Land-Use Development of Feixi County, China, Based on Optimized Productive–Living–Ecological Functions

Abstract

Rural revitalization places higher demands on the productive–living–ecological (P-L-E) spaces of towns and cities. It is necessary, therefore, to identify, evaluate, and optimize P-L-E spaces to better guide spatial planning. Existing studies typically evaluate a single space, lacking a comprehensive consideration of whole-area integration. This study, therefore, developed a coupled spatial/developmental suitability evaluation system for Feixi County, Anhui Province, China, combining spatial quality evaluation, a coupled coordination model, and future land-use simulation (FLUS) model. The spatial quality of Feixi County in 2010, 2015, and 2020 was obtained by applying the evaluation system to the spatial development pattern. The results were analyzed and verified using the landscape pattern index and development suitability evaluation. The results showed the following: (1) The coupling coordination degree of the region increased from 0.131 to 0.372, changing from low to moderate coordination. (2) Based on the FLUS model to better capture the uncertainty and stochastic basis of the development in the study area. The kappa coefficient and Figure of Merit (FoM) index of the land-use simulation accuracy verification index were 0.7647 and 0.0508, respectively, and the logistic regression ROC values were above 0.75, thus meeting accuracy requirements. This demonstrated that the simulation model—based on a factor library of the evaluation of resource and environmental carrying capacity and suitability for development and construction—could better reflect future land-use changes. (3) The simulation showed that under the baseline development scenario, the area’s spatial layout is too concentrated in terms of construction land, ignoring P-L-E coordination. Under the ecological optimization scenario, high-quality ecological space is ensured, but other types of spaces are lacking. Under the comprehensive guidance scenario, lagging ecological space is optimized and P-L-E spatial development is enhanced through aggregation, clustering, concentration and integration. This way, the spatial quantity structure and distribution form can meet P-L-E spatial development needs in Feixi County. In this study, on the basis of scientific assessment of the current P-L-E space, the FLUS model was applied to carry out a scenario simulation according to different objectives. Moreover, based on the construction of the coupling system of human–nature system, the driving factors were improved to enhance the prediction accuracy of the FLUS model. This study’s findings can help improve the scientificity, flexibility and management efficiency of Feixi County’s P-L-E spatial layout, thereby supporting its sustainable development.

1. Introduction

Since its reform and opening up, China has achieved remarkable results in the development and utilization of land space and the urbanization process has advanced rapidly. In the process of county territorial space development, while social and economic development has continued rapidly, unreasonable development has also led to a series of problems [1]. For example, the rapid expansion of urban construction land has encroached on a large amount of ecological space, causing land sanding, grassland degradation, shrinking wetlands and other problems, and ecological security has been affected [2,3,4]. The inefficient use of land space and the increasingly prominent conflict between supply and demand are concentrated in the inefficient use of production and living space caused by the uncontrolled expansion of urban construction land [1]. In the face of a series of problems caused by the unreasonable use of land resources, how to rationalize the layout and optimize the development of the productive–living–ecological (P-L-E) spaces in the county has become one of the key solutions [3]. In 2012, the Chinese government established the “Production-Life-Ecology Space” strategy in its territorial planning to achieve the goal of “intensive and efficient production space, livable and moderate living space, and beautiful ecological space”, and thus achieve the harmony between humans and nature [5].

Research on the optimization of P-L-E spaces mainly concerns ecosystem function, land cover and utilization and landscape value, which correspond, respectively, to ecological frameworks, socioeconomic orientation and landscape ecology [1,2,3,4,6,7]. Previous P-L-E research has focused on the development and application of spatial planning [8]. The European Union initiated its research on P-L-E spaces using a multifunctional agricultural classification system [9,10]. Then, a classification system for ecosystem and landscape functions was established based on multidisciplinary research [11,12,13]. Meanwhile, inspired by the concept of sustainable development, methods related to land use were introduced in the research [14,15,16,17]. In China, research on P-L-E spaces has mainly focused on index system development, main function identification and spatial optimization and control [6]. Here, spatial pattern optimization has mainly focused on policy formulation and management methods or the evaluation of a single space, such as farmland.

In general, the research on P-L-E space has mainly focused on the identification and classification of P-L-E space, spatial utilization quality evaluation and spatial simulation and optimization. Among them, P-L-E spaces can be divided into two types from the perspective of functional composite, one is the single-functional production, living and ecological space system, and the other is the composite P-L-E space system based on the multifunctionality of land. The commonly used spatial identification methods of P-L-E functions can be divided into the quantitative measurement method and subsumption classification method [18]. The former one identifies the evaluation system of production, living and ecological functions by constructing a quantitative one, but it is difficult to carry out multi-subject integration and multi-scale integrated expression [19,20,21]; the latter one essentially consolidates and classifies land-use data, bridging urban land-use classification, while compensating to a certain extent for the shortcomings of land-use classification in terms of insufficient consideration of ecological functions [22].

Spatial utilization quality evaluation can be generally divided into two categories. One starts from the basis of spatial composition and reveals the relationship between spatial functions and sustainable development [23], and the other directly takes the spatial type as the object and constructs the system from the perspectives of resource and environmental carrying capacity as well as development suitability evaluation [13,24]. Territorial spaces are formed by the interaction of ecosystems and human societies [25]. They are characterized by multiple constituents and nested spatiotemporal scales. This requires spatial governance schemes that consider development and utilization in tandem with the restoration and protection of nature [26]. In the context of rapid urbanization, ignoring P-L-E coupling optimization is not conducive to the optimization of the territorial spatial patterns [27,28]. In this regard, Guo et al. developed and optimized an ecological security pattern from the perspective of P-L-E spatial coupling [29]. Zhang [8] and Chen et al. [30], meanwhile, studied the spatiotemporal evolution of the degree of coupled coordination. Further, Zhang et al. evaluated the quality of spatial utilization and coupled coordination degree [18]. In general, the evaluation of the coupling coordination degree in the coupled human–nature system of the national space can measure the development coordination level of spatial socio-ecological processes in the study area from a functional perspective, so this paper starts from the perspective of landscape ecology and focuses on the process coupling including the coupling of natural and human processes and the spatiotemporal coupling of geographical processes [28,30].

In terms of spatial simulation and optimization, cellular automata have been widely used in the spatial evaluation and optimization of land [19]. Commonly used cellular automata include the geographical simulation and optimization system (GeoSOS), the future land-use simulation (FLUS) model and the conversion of land-use and its effects at small region extent (CLUE-S). Through application to land-use simulations [20,21,22] and urban expansion simulation and urban growth boundary delineation [31,32,33], it has been shown that, at this stage, among the subsequent improvements of multiple cellular automata, the FLUS model integrates a top-down system dynamics model and a bottom-up cellular automata, and establishes adaptive inertia and competition mechanisms to deal with the complex competition and interactions between different land-use types [31], which can better capture the uncertainty and stochasticity of development in the study area. Among the current studies using the FLUS model, Jiang [5], Penny [34] and Ma et al. [35] used the FLUS model to study land use and spatial patterns considering natural, economic and cultural factors. Xu [36], Mansour [37] and Ye et al. [38] mainly conducted simulations and monitoring in the urban context using the FLUS model. Wang [39] and Yang et al. [40] used the FLUS model to simulate ecohydrological services and ecological security patterns, respectively. Chen [41] and Sun et al. [42] considered population and arable land factors when using the FLUS model. Huo [43], Lin [44] and Wang et al. [45] focused on the coupling relationship in the application of FLUS model studies.

There are three main problems in current P-L-E research. First, there is the duplication of spatial boundary delineation due to the composite function of land use. Second, spatial identification at different geographic scales tends to be inconsistent [46]. Third, there is a lack of comprehensive integration research in the whole domain, which is mainly manifested by optimizing technical ideas and application practices [29,47]. Regarding the FLUS model, the problem is that the current model has been mainly studied in the context of pattern recognition and image processing and have not involved multidisciplinary applications when considering the main drivers of land-use/land-cover (LULC) changes, resulting in an underutilized model [48].

To fill the above research gaps, the present study aimed to integrate resource and environmental carrying capacity and the suitability of spatial development to support sustainable regional development. A coupled system reflecting the human–nature relationship was created, the driving factors were improved and three types of scenario simulations were carried out based on a factor library of the evaluation of resources and environmental carrying capacity and suitability for development and construction using landscape pattern index (LPI) calculation. The P-L-E spatial optimization of Feixi County, China, was taken as an example. The findings can guide the flexible and scientific management of P-L-E spaces and offer practical support for high-quality sustainable development.

2. Materials and Methods

2.1. Study Area

Feixi County (31°30′–32°00′ N, 116°40′–117°21′ E) is located in the middle of Anhui Province, southwest of Hefei (Figure 1). It is located in the urban radiation area and urban–rural interlacing area of Hefei, with a complex composition of P-L-E spatial types. The county serves various functions, such as a comprehensive service center in the southwest of the urban area, an ecological civilization demonstration area and a core area of new industrialization. It has a total of eight towns and four townships (Figure 1), with a total area of 1698 square kilometers, accounting for 14.8% of the total area of Hefei. It has a high northwest and low southeast topography.

On 1 January 2021, the population of Feixi County was 967,508, and was concentrated in Shangpai, Taohua, Zipeng, Guanting and Huagang in the northeast of Feixi County; the rest of the population was scattered in other towns in Feixi. At present, Feixi County is rich in labor resources, with 69.45% of the county’s population being aged 15–59. Compared with the Sixth National Census in 2010, the county population has increased by 222,500 in 10 years. The residential population is now 115,800 more than the registered population, achieving a net inflow of population and reversing population outflow, which shows the location advantages of Feixi (Figure 2).

Meanwhile, the pace of urbanization in Feixi County has accelerated, with 63.30% of the county’s population living in cities and towns, up 28.7% from 2010.With rapid urbanization and industrial development, Feixi County faces problems such as the heavy use of land resources, a serious contradiction between people and land. The Feixi County Master Plan (2015–2030) stipulates that by 2030, Feixi County should achieve a rational urban–rural spatial layout, functional division of labor and city–industry linkage. It is necessary, therefore, to evaluate spatial functioning and coordination based on the division of P-L-E areas and determine the most suitable spatial layout for Feixi County.

2.2. Data Sources

The data included land-use data, coupling coordination analysis series data, urbanization data, economic development data, population gathering data, digital elevation model (DEM) data, transportation advantages data, water resources data, land resources data, atmospheric resources data, soil erosion data, biological diversity data, ecological development data, vegetation resources data and food resources data. The data are detailed in

Table 1. Data sources within this study.

Category	Data Name	Data Format	Year	Data Source
Land database	Land-use data	30 m × 30 m raster	2010, 2015 and 2020	US Geological Survey (https://glovis.usgs.gov/ (accessed on 18 October 2021))
Socio-economic Factors	Coupling coordination analysis series data	Text data	2010, 2015 and 2020	China Rural Statistical Yearbook, Hefei City Statistical Yearbook and Anhui Province Ecological Environment Status Bulletin
	Urbanization	30 m × 30 m raster data	2015 and 2018	Resource and Environment Science and Data Center (http://www.resdc.cn/ (accessed on 18 October 2021))
	Economic development	1 km × 1 km per capita GDP	2015
	Population gathering	30 m × 30 m raster data	2020	Gaode Map POI data (https://lbs.amap.com/ (accessed on 18 October 2021))
Topography	Digital elevation model data (DEM)	30 m × 30 m raster	2015	National Basic Science Data Center (http://www.nbsdc.cn/ (accessed on 18 October 2021))
Locational Factors	Transportation advantages	30 m × 30 m raster	2020	Gaode Map POI data (https://lbs.amap.com/ (accessed on 18 October 2021))
Natural Environmental Factors	Water resources	1 km × 1 km raster data	2015	Resource and Environment Science and Data Center (http://www.resdc.cn/ (accessed on 18 October 2021)) and China National Environmental Monitoring Center (http://www.cnemc.cn (accessed on 18 October 2021))
	Land resources	Calculate the spatial distribution of chalky soil, clay and sand content	1995
	Atmospheric resources	1 km × 1 km raster data	2010 and 2018
	Soil erosion	1 km × 1 km erosion intensity	2015
	Biological diversity	Specific reference to spatial distribution data of ecosystem service values	2015
	Ecological development	Specific reference to spatial distribution data of ecosystem service values	2015
	Vegetation resources	NDVI data	2018	Big Earth Data Science Engineering Data Sharing Service System (https://data.casearth.cn/ (accessed on 18 October 2021))
	Food resources	NPP data	2015

. Using ArcGIS 10.2, the spatial data were resampled into 30 m × 30 m raster cells, and the CGCS2000_3_Degree_GK_Zone_38 coordinate format was uniformly adopted.

2.3. Data Processing

2.3.1. Identification and Classification of P-L-E Functional Areas in Feixi County

Taking the urban development boundary, permanent basic farmland red line and ecological protection red line as the primary principles [23], the ecosystem function system and landscape function system were integrated [24] to establish a more comprehensive and integrated P-L-E spatial classification system for Feixi County based on multifunctional land.

The subsumption classification method was used for functional space identification [6]. According to the multifunctionality of land use, the idea of zoning before classification was adopted [7,49,50,51,52] based on existing P-L-E space types in Feixi County, excluding spatial types corresponding to land-use types that did not appear. These were divided into three types (P-L-E) of spaces and two types of composite spaces corresponding to land-use types, where the composite spaces were classified based on their dominant functions. The five specific types of spaces are shown in

Table 2. Reclassification of land-use types of P-L-E spaces.

Space Classification		Land-Use Type	Further Subdivision of Land Use	Description
Tier 1	Tier 2
Single functional space	Ecological space	Woodland	Wooded land, shrubland, other wooded land	With climate regulation, atmospheric regulation, soil and water regulation and other ecosystem regulation functions play an important role in biodiversity
		Grass	Natural pasture, other grasslands	Provide biological products, atmospheric regulation, climate regulation, water conservation, soil and water conservation, ecological landscape and recreation
		Water and water facilities land	River water surface, lake water surface, inland mudflats	Plays an important role in regulating regional temperatures and stabilizing local climate
		Other land	Vacant land, saline land, swampy land, sandy land, bare land	Important ecological function of the landscape
	Production space	Cropland	Paddy field, watered land, dry land	Production function for providing food and other biological products
		Other land	Facility agricultural land, field	Has a certain production function
		Commercial land	Wholesale and retail land, accommodation and catering land, business and financial land, other commercial land	Important production functions such as providing business services
		Industrial and mining storage land	Industrial land, mining land, storage land	To provide people with industrial production and material storage places
		Transportation land	Railroad land, road land, street land, rural roads, pipeline transport land	Surface lines, yards, etc., that provide important transportation access for people
		Water and water facilities land	Hydraulic construction land	Provides an important watershed production function
	Living Space	Residential land	Urban residential land, rural residential land	Provides an important place for people to live and rest
		Special land	Religious land, funeral land	Provides for people’s special needs
Composite functional space	Eco–production space	Gardens	Orchards, tea plantations, other gardens	Ecosystem supply service functions that provide food, fruits and other biological products, as well as ecosystem support functions such as climate regulation, atmospheric regulation, nutrient cycling and soil and water regulation
		Water and water facilities land	Ditches, reservoir water, pond water	Has important ecological service functions and is important land for water conservation; also has certain production functions
	Production-living space	Public administration and public service land	Land for institutions and organizations; press and publication; science and education; culture, sports, and recreation; public facilities; scenic facilities; parks; green areas	Provide public services and products such as medical care, education, culture, sports, while having certain production functions

.

2.3.2. Construction of the Evaluation Index System for the Three Spatial Functions

Based on the analysis of the three spatial functions (P-L-E) and previous studies [8,23,24,53,54,55,56,57,58,59], and considering indicator stability and data availability, evaluation indicators were selected considering spatial types in Feixi County. Considering the complexity and repeatability of attribute data, when constructing and using the spatial utilization quality evaluation index system, the composite functional space involved in Table 2 was merged into a single functional space based on its dominant function.

As Feixi County is located in a rural area, the production space function indicators were divided into agricultural and nonagricultural production function indicators [49]. Four indicators were selected: per capita grain production, per capita arable land area, per capita output value of secondary and tertiary industries and per capita output value of agriculture, forestry, animal husbandry and fishery services [23,24,53]. People are the main service object of living spaces. To reflect the income, expenditure and welfare of residents, this was divided into social security and livelihood security [23], and five indicators were selected: number of hospital beds per 10,000 rural residents, per capita housing area in rural areas, rural employment structure and per capita annual income of rural residents and per capita living consumption expenditure of rural residents [53,54,55]. Ecological space is the basis of the P-L-E space [6], and its function is formed by the combination of ecosystem and ecological processes to maintain natural conditions for human survival and utility [24]. As the land in Feixi County is mainly agricultural land, four indicators were selected from the two perspectives of ecological environment formation and human activity interference: forest coverage rate, wetland area ratio, per capita agricultural fertilizer application and per capita water resources of rural residents [23,24,53,54,55,56,57,58]. Meanwhile, the entropy weight method and coefficient of variation method were used to calculate the weights of each index and establish the functional evaluation index system (

Table 3. P-L-E space utilization quality evaluation index system and weights.

P-L-E Space	First Grade Indexes	Second Grade Index	Interpretation of Indexes and Calculation Methods	Entropy Method Weights	Coefficient of Variation Method Weights	Combined Weights
Production space	Agricultural production function	Food production per capita	Total food production/total population, tons per 10,000 people	0.0017	0.0152	0.0084
		Arable land per capita	Total arable land area/total rural population, acre/person	0.0057	0.0280	0.0168
	Nonagricultural production function	Per capita output value of secondary and tertiary industries	Output value of secondary and tertiary industries/total rural population, million yuan/person	0.1742	0.1501	0.1621
		Per capita output value of agriculture, forestry, animal husbandry and fishery services	Total output value of agriculture, forestry, animal husbandry, and fishery services/rural employed population, million yuan/person	0.0468	0.0785	0.0626
		Structure of rural employment	Rural nonfarm employment/rural employment, percent	0.0039	0.0234	0.0136
Living Space	Life Security	Average annual income of rural residents	Total annual income of rural residents/total rural population, million yuan	0.0833	0.1066	0.0950
		Per capita living consumption expenditure of rural residents	Total living consumption expenditure/total rural population, million yuan	0.2022	0.1643	0.1833
		Rural housing area per capita	Total rural housing area/total rural population, m2/person	0.3275	0.2535	0.2905
	Social Security	Number of hospital beds per 10,000 rural residents	Total number of beds in hospitals and health centers/total rural population, number of beds/10,000 people	0.0101	0.0277	0.0189
		Percentage of rural residents receiving education	Total number of educated people in rural areas (kindergarten, primary, secondary)/total rural population, %	0.0990	0.0102	0.0546
Ecological Space	Natural conditions	Forest coverage	Total area of forest land/total area of evaluation unit, %	0.0186	0.0511	0.0349
		Wetland area ratio	Wetland area/total area of study area, %	0.0096	0.0365	0.0231
	Ecological status	Amount of agricultural fertilizer applied by rural residents per capita	Agricultural fertilizer application/total rural population, tons per 10,000 people	0.0172	0.0498	0.0335
		Water resources per rural resident	Total water resources/total rural population, tons per 10,000 people	0.0002	0.0051	0.0027

).

2.3.3. Construction of Driving Factors

The factor database was improved by referring to the evaluation system of resource and environmental carrying capacity and development suitability, as well as the classification method of Zhao [60]. The evaluation factors of resource and environmental carrying capacity and of the suitability of territorial space development were constructed considering the current situation and data availability, for a total of 37 evaluation indicators in 13 categories in two databases (

Table 4. Driving factors of P-L-E spatial land-use simulation.

Database	Categories	Evaluation Indicators	Data Description	Year	Data Source
Resource and environmental carrying capacity evaluation factors	Water resources	Average annual precipitation	1 km × 1 km raster data	2015	a
		Water supply	Specific reference to spatial distribution data of ecosystem service values [61]
		Hydrological regulation
		Distance from river system	30 m × 30 m raster data
	Land resources	Soil texture	Calculate the spatial distribution of chalky soil, clay and sand content	1995	a
	Vegetation resources	Annual normalized vegetation index	NDVI data	2018	b
	Food resources	Net primary productivity	NPP data	2015
		Food production capacity	Specific reference to spatial distribution data of ecosystem service values [61]	2015	a
		Raw material production capacity
	Atmospheric resources	Atmospheric pollution index	PM2.5 spatial interpolation of raster data	2018	c
		Average annual temperature	1 km × 1 km raster data	2010	a
	Soil erosion	Soil erosion	1 km × 1 km erosion intensity	2015	a
	Urbanization	Distance from each town center	30 m × 30 m raster data	2015	a
		Distance to mineral resources		2018
	Biological diversity	Biodiversity	Specific reference to spatial distribution data of ecosystem service values [61]	2015	a
Appropriateness evaluation factors for urban spatial development	Terrain conditions	Elevations	Calculated from DEM data	2015	d
		Slope
		Slope direction
	Transportation advantages	Distance to train station	30 m × 30 m raster data	2020	e
		Distance to highway
		Distance to national highway
		Distance to provincial road
		Distance to county road
		Distance to main road
		Distance to railroad
		Distance to waterway
	Economic development	GDP	1 km × 1 km per capita GDP	2015	a
	Population gathering	Population density	1 km × 1 km population distribution	2015	a
		Distance to park square	30 m × 30 m raster data	2020	e
		Distance to industrial park
		Distance to scenic spot
		Distance to important companies
		Distance to institution
		Distance to hospital
		Distance to financial services
		Distance to shopping center
	Ecological development	Environmental purification capacity	Specific reference to spatial distribution data of ecosystem service value [61]	2015	a

Note: data sources: a: Resource and Environment Science and Data Center (http://www.resdc.cn/ (accessed on 18 October 2021)); b: Big Earth Data Science Engineering Data Sharing Service System (https://data.casearth.cn/ (accessed on 18 October 2021)); c: China National Environmental Monitoring Center (http://www.cnemc.cn (accessed on 18 October 2021)); d: National Basic Science Data Center (http://www.nbsdc.cn/ (accessed on 18 October 2021)); e: Gaode Map POI data (https://lbs.amap.com/ (accessed on 18 October 2021)).

).

At the same time, in order to clearly reflect the positive or negative effects of each driving factor on the distribution probability of various types of spatial utilization, a logistic regression model was used to calculate the regression coefficients of each driving factor. In this study, 37 driving factors and five spatial raster data were processed in ASCII form and imported into Statistical Product and Service Solutions (SPSS, Version 25.0, IBM Corp, Armonk, NY, USA) software for analysis. The corresponding results were obtained and the model was verified by ROC curve.

2.4. Methods

2.4.1. Framework Design

The spatial simulation of the land-use development framework combining the coordination model, FLUS model and evaluation method of P-L-E space functions in Feixi County is shown in Figure 3.

First, the P-L-E space categories of Feixi County were identified and reclassified according to the natural and social data sets (Table 2). Second, a coupling-based spatial quality evaluation system was constructed to obtain the spatial utilization quality from 2010 to 2020 after importing the corresponding data. Finally, the FLUS model with improved natural-social drivers was combined with the landscape pattern index and development suitability evaluation to carry out a multiscenario land-use development simulation in 2030 and obtain the most suitable scenario.

2.4.2. Evaluation of Utilization Quality and Coupling Coordination Degree of P-L-E Space

(1)

Evaluation of the quality of the P-L-E spaces

The quality of P-L-E space utilization was calculated using Equation (1) according to the following weights:

$${\mathrm{Q}}_P,Q_L\mathrm{or}{Q}_E={\displaystyle {\sum }_{i=1}^{\mathrm{n}}{\alpha }_i}\times {x^{\prime }}_{ij},$$

(1)

where ${\mathrm{Q}}_P$, $Q_L$ and $Q_E$ are the evaluation results of the utilization quality of the three spatial subsystems, respectively; ${\alpha }_i$ is the index weight and $x_{ij}^{\prime }$ is the value of the operational indexes.

After obtaining the quality evaluation values for P-L-E spaces, the comprehensive utilization quality of P-L-E spaces was determined. Since P-L-E spaces comprise an interrelated system, and principal component analysis was used to calculate the comprehensive utilization quality of space Q. The formula is as follows:

$$Q=\tau \times Q_P+\varphi \times Q_L+\omega \times Q_E,$$

(2)

where $\tau ,\varphi ,\omega$ are the evaluation weights of the P-L-E space.

(2)

Coupling coordination degree model

The coupling model in physics was adopted to explore the degree of interaction and influence of the three types of spaces and to classify the coupling coordination degree, as shown in Equations (3) and (4):

$$C=3\times {\left[\frac{P_{\mathrm{i}}\times L_{\mathrm{i}}\times E_{\mathrm{i}}}{(P_{\mathrm{i}}+L_{\mathrm{i}}+E_{\mathrm{i}})}\right]}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$3$}\right.},$$

(3)

where $C_{}$ indicates the degree of coupling of P-L-E spaces. $C_{}$ ∈ [0, 1]; the closer $C_{}$ is to 1, the more coupled the system is. Conversely, the closer it is to 0, the more conflicting it is. $P_{\mathrm{i}}$, $L_{\mathrm{i}}$, $E_{\mathrm{i}}$ indicate the measurement index of each subsystem function.

$$D=\sqrt{(C_{}\times T)}{;}_{}T=\alpha P_{\mathrm{i}}+\beta L_{\mathrm{i}}+\gamma E_{\mathrm{i}},$$

(4)

where $D$ is the coupling coordination degree of the P-L-E space and $D$ ∈ (0, 1].

Based on previous research [53,54,55], expert opinions and the characteristics of Feixi County, the coupling coordination degree was divided into five types (

Table 5. Classification of coupling coordination degree.

Coupling Coordination Level	Range
Low-level coupling coordination	Di ∈ (0, 0.3]
Lower-level coupling coordination	Di ∈ (0.3, 0.35]
Medium coupling coordination	Di ∈ (0.35, 0.5]
Higher-level coupling coordination	Di ∈ (0.5, 0.8]
High-level coupling coordination	Di ∈ (0.8, 1]

). α, β and γ were the coefficients to be determined for the P-L-E spatial quality promotion rate. They were assigned one-third each because they are equally important for economic and social development.

2.4.3. Development Suitability Evaluation

LPI reflects the spatial heterogeneity of landscape patches and is an important part of landscape ecology. In this study, the suitability of urban space development was evaluated from the perspective of landscape ecology, focusing on P-L-E spaces, the division of restricted lines, the carrying capacity of resources and the environment and the development suitability system. At the same time, based on the LPI evaluation of the landscape patch index, a library of the driving factors of P-L-E spatial land-use simulation was constructed (

Table 6. Suitability index system for development and construction based on P-L-E spaces.

Suitability Level	Land Type	Corresponding Space Types
Suitable	Wholesale and retail land; accommodation and catering land; business and financial land; other commercial land; industrial land; mining land; storage land; railroad land; highway land; street land; rural roads; pipeline transport land; waterworks land; urban residential land; rural residential land; religious land; funeral land; institutional land; press and publication land; scientific and educational land; cultural, sports and entertainment land; public facilities land; scenic spots and facilities land	Production space, living space, production–living space
Generally suitable	Dry land; orchards; tea gardens; other garden land; ditches; reservoir water; pond water; parks and green spaces; facilities; agricultural land	Production space, eco–production space
Slightly suitable	Natural pasture; other grasslands; inland mudflats; fields	Ecological space, production space
Unsuitable	Paddy land; watered land; wooded land; shrub land; other wooded land; river water; lake water; inland mudflat; vacant land; saline land; marsh land; sandy land; bare land	Ecological space, production space

), and the FLUS driving factors were generated (Figure 4 and Figure 5). The simulation results, together with the development suitability results, can provide basic support for optimizing the spatial development pattern of the land [60,62].

Based on the background constraints and the P-L-E spatial classification system, a development suitability evaluation system for Feixi County was constructed in terms of spatial types, relating land-use types to the four suitability levels [63] (Table 6). A development suitability evaluation of the study area was then carried out. The results support the evaluation of multiscenario simulation results.

2.4.4. Multiscenario FLUS Simulation

The cellular automata model is a discrete-time dynamic system defined in the spatial distribution of grid format, which can deduce spatiotemporal changes in complex systems. The FLUS model is based on system dynamics and the cellular automata model, and it integrates the adaptive inertia competition mechanism of artificial neural networks (ANNs) and roulette. It imports data for space utilization types and various driving factors in the selected research area and obtains the development probability for each type of space. Then, it combines the neighborhood weight, conversion cost matrix and adaptive inertia coefficient to obtain the overall conversion probability of cells in the study area; establishes the cyclic iteration of spatial demand and status; and simulates results on that basis. This method can be used to simulate land-use changes under various driving forces (e.g., nature and society) to determine the rational allocation of the number and spatial distribution of the pixels of various spatial types in the study area. Considering the overall quality of the three (P-L-E) spatial areas, the coupling coordination degree and the LPI results, the parameters were adjusted and the driving factors and constraints were introduced into the FLUS model to ensure that the simulation was carried out in the predetermined direction. On that basis, following national development policies, simulations were carried out for the baseline scenario, ecological priority scenario and comprehensive optimization scenario.

Combining the P-L-E land-use simulation driving factors with the FLUS model requirements, vector data processing (e.g., kernel density analysis, normalization and logistic regression) was used to create maps of 37 driving factors affecting the spatial development of Feixi County (Figure 4 and Figure 5). Restricting factors were set from the perspective of ecological sensitivity and ecological space protection (Figure 6), reflecting the characteristics of resource and environmental endowment, risks and problems in the study area.

Since the future demand for each type of space varies, a Markov chain was used to predict the amount of each type of space and amount of change that would satisfy future development needs (

Table 7. Changes in input pixels.

	Number of Pixels for the Corresponding Year				2015–2020 Amount of Pixel Increase	2020–2025 Amount of Pixel Increase	2025–2030 Amount of Pixel Increase
	20 15	20 20	20 25	20 30
Production space	319,790	310,649	304,449	300,194	−9141	−6200	−4255
Ecological space	7451	9554	11,149	12,368	2103	1595	1219
Productive–living space	1831	971	956	969	−860	−15	13
Living space	21,952	28,893	32,677	34,846	6941	3784	2169
Eco–production space	27,346	28,303	29,138	29,992	957	835	854

). It is also used to simulate the change in space use by predicting that the situation of space-use type at time (t + 1) is only related to the situation of that type of space use at time t [64,65,66].

The formula is as follows:

$$S_{(t+1)}=P_{ab}\times S_{(t)},$$

(5)

where $S_t$ and $S_{(t+1)}$ are the matrix of the spatial utilization type situation at t + 1, and $P_{ab}$ denotes the transfer matrix of transformation from type a to type b. Spatial utilization data for 2015 and 2020 were used as the basis for parameter adjustment and testing the accuracy of the model to curb possible bias in the Markov chain when performing long-term projection and to forecast the land-use condition of 2030 under multiscenario simulation.

The back-propagation ANN algorithm (BP-ANN) in the FLUS model was used to calculate the evolution probability of each type of land use with the metacell spatial succession condition as the core. By fitting various types of spatial constraints and drivers, the metacellular transition probabilities were calculated in the output layer. The 3*3 Moore neighborhood model was also used to calculate the neighborhood influence factor, using the following equation:

$${\mathsf{\Omega }}_{p,k}^t=\frac{{\displaystyle {\sum }_{3\times 3}con\left(c_p^{t-1}=k\right)}}{3\times 3-1}\times w_k,$$

(6)

where $con\left(c_p^{t-1}=k\right)$ denotes the spatial tuple size occupied by land type k at the last (t − 1) iteration, $w_k$ denotes the neighborhood influence factor parameter of each type of space and ${\mathsf{\Omega }}_{p,k}^t$ is the neighborhood influence factor parameter of tuple space p at time t. The defined spatial types were ranked according to their expansion capacity, based on the change in patch area, as living space > ecological–production space > production–living space > production space > ecological space. The defined spatial types were ranked according to expansion capacity as living space > ecological–production space > production–living space > production space > ecological space; they were assigned according to the change in patch area (

Table 8. Neighborhood factor weights.

Land-Use Types	Living Space	Eco–Production Space	Productive–Living Space	Production Space	Ecological Space
$\mathrm{Neighborhood}\mathrm{factor}\mathrm{parameters}{w}_k$	1	0.74	0.64	0.52	0.45

).

The difference between the expected demand and the actual space allocated in the study area was calculated using the adaptive inertia coefficient. On that basis, site-specific development trends that contradicted the expected demand were automatically adjusted and thus corrected in the next iteration [34,35,67]. The formula is as follows:

$$I_k^t=\left\{\begin{array}{c}{\mathrm{I}}_{\mathrm{k}}^{\mathrm{t}-1},\left|{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-1}\right|\le \left|{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-2}\right|\\ {\mathrm{I}}_{\mathrm{k}}^{\mathrm{t}-1}\times \frac{{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-2}}{{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-1}},{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-1}<{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-2}<0\\ {\mathrm{I}}_{\mathrm{k}}^{\mathrm{t}-1}\times \frac{{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-1}}{{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-2}},0<{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-2}<{\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-1}\end{array}\right.,$$

(7)

where $I_k^t$ denotes the inertia coefficient corresponding to space type k during iterative run t, and ${\mathrm{D}}_{\mathrm{k}}^{\mathrm{t}-1}$ denotes the difference between the scale that satisfies the space utilization requirement and the allocated scale at (t − 1).

This study developed three simulation scenarios based on different spatial planning policies and constraints: (1) the 2030 baseline development scenario, (2) the 2030 ecological priority scenario and (3) the 2030 comprehensive guidance scenario. In the first scenario, no constraint variables were set based on human interference, such as policies to establish ecological protection zones and thus influence spatial planning. The land-use model was based on natural spatial drivers. Meanwhile, to avoid generating unconstrained forms of township land-use scenarios, the conversion coefficients between each spatial type were set with reference to quantitative indicators. The quantitative constraint indicators of Feixi County planning were transformed into the cost matrix of the adaptive inertia competition mechanism. The 2030 ecological priority scenario was based on the 2015 data for P-L-E spatial evaluation, and ecological space protection was taken as a variable in the constraint layer, focusing on the consideration of protecting ecological space and the carrying capacity of resources and the environment to ensure that the relevant spatial types are not reduced, and ecosystem restoration is carried out. In addition, the Feixi County Master Plan (2015–2030) and the evaluation content were introduced to account for the integrated development of five spatial types for the setting of the 2030 comprehensive guidance scenario (7d), which coordinated the conversion relationship of the five land types.

Conversion cost was used to indicate the ease of converting existing spatial types into demand types, and the probability of land-type transfer could be increased by a certain proportion according to the focus type [40,68]. Simulations were conducted under the baseline scenario, ecological priority scenario and comprehensive guidance scenario. The corresponding conversion cost matrix (

Table 9. Conversion cost matrix.

	Baseline Developments					Ecological Priorities					Comprehensive Guidance
	a	b	c	d	e	a	b	c	d	e	a	b	c	d	e
a	1	1	1	1	1	1	1	1	1	1	1	0	1	1	1
b	1	1	0	0	0	0	1	0	0	0	1	1	0	0	0
c	1	0	1	1	1	0	1	1	0	1	1	1	1	1	1
d	1	0	0	1	0	0	0	0	1	0	0	0	0	1	0
e	1	0	0	0	1	0	0	0	0	1	0	0	1	0	1

Note: a, b, c, d and e represent, respectively, production space, ecological space, production–living space, living space and ecological–production space; 0 means no transformation, and 1 means transformation is allowed.

) was obtained based on the changes in the three spatial types from 2015 to 2020 (Table 7). For the latter two scenarios, the ecological benefits of each type of space were ranked, with the ecological space as the best and the living space as the worst; the conversion principle was to not allow high-priority space to be converted into low-priority space. In the process of model implementation, it was shown that the conversion of ecological space into other space types was controlled, and the conversion of other space types into ecological space was appropriately increased.

The overall conversion rate of the space cell occupied by each spatial type was estimated as follows:

$$C{P}_{p,k}^t=P_{p,k}^t\times {\mathsf{\Omega }}_{p,k}^t\times I_k^t\times s{c}_{c\to k},$$

(8)

where $C{P}_{p,k}^t$ is the overall probability that the medium cell space p is converted between the original spatial type c and spatial type k at time t, $P_{p,k}$ denotes the likelihood that the cell space is converted between p and k, ${\Omega }_{p,k}^t$ denotes the neighborhood influence factor parameter, $I_k^t$ denotes the adaptive inertia coefficient and $s{c}_{c\to k}$ denotes the consumption of conversion between spatial types c and k.

Finally, the space utilization conversion type of the cell space was determined using the roulette wheel selection mechanism.

2.4.5. Evaluation of Landscape Patch Indexes

The five dimensions of area, density, shape, proximity and clustering were screened for landscape pattern indexes. Six commonly used indexes from patch area, maximum patch, number of patches, average perimeter area ratio, Euclidean nearest neighbor distance and separation were determined to measure the landscape characteristics of the study area and analyze the advantages and disadvantages under different scenario simulations in terms of landscape ecology [60].

Table 10. Meanings of the landscape patch index and the calculation formulas.

Landscape Patch Index	Meaning of the Index	Formula
Class area (CA)	Area sum of a certain patch type is the basis for calculating other indicators	$CA={\displaystyle \sum _{j=1}^n{a}_{ij}}\left(\frac{1}{10000}\right)$ $a_{ij}$$\mathrm{is}\mathrm{the}\mathrm{area}\mathrm{of}\mathrm{plaque}ij$
Largest patch index (LPI)	Proportion of the largest patch in a certain patch type to the whole landscape area; the change in its value can reflect the landscape dominance and the direction and strength of human activity in land use	$LPI=\frac{\max (a_{ij})}{A}\times 100$$,\mathrm{where}{a}_{ij}$$\mathrm{is}\mathrm{the}\mathrm{area}\mathrm{of}\mathrm{patch}ij$$,\mathrm{and}A$ is the total landscape area
Number of patches (NP)	Total number of all patches in a given patch type or landscape area, capable of describing landscape heterogeneity and landscape fragmentation	$NP=n_i$, $\mathrm{where}{n}_i$$\mathrm{is}\mathrm{the}\mathrm{number}\mathrm{of}\mathrm{plaques}i$
Perimeter–area ratio distribution (PARA_MN)	Average value of the perimeter–area ratio of each patch can measure the shape complexity of each type of patch	$PARA\_MN=\frac{p_{ij}}{a_{ij}}$, $\mathrm{where}{p}_{ij}$$\mathrm{is}\mathrm{the}\mathrm{perimeter}\mathrm{of}\mathrm{plaque}ij$$,\mathrm{and}{a}_{ij}$$\mathrm{is}\mathrm{the}\mathrm{area}\mathrm{of}\mathrm{plaque}ij$
Euclidean nearest-neighbor distance distribution (ENN_MN)	Distance to the nearest neighboring plaque of the same type enables the quantification of plaque isolation	$ENN\_MN=h_{ij}$, $\mathrm{where}{h}_{ij}$$\mathrm{is}\mathrm{the}\mathrm{linear}\mathrm{distance}\mathrm{from}\mathrm{patch}ij$ to the closest neighboring patch of the same kind
Splitting index (SPLIT)	Degree to which the landscape is separated can reflect the spatial structure of the landscape	$SPLIT=\frac{A^2}{{\displaystyle \sum _{j=1}^n{a}_{ij}^2}}$, $\mathrm{where}{a}_{ij}$$\mathrm{is}\mathrm{the}\mathrm{area}\mathrm{of}\mathrm{patch}ij$$,\mathrm{and}A$ is the total landscape area

shows the formulas and specific meanings.

LPI, NP, PARA_MN, ENN_MN and SPLIT were selected to calculate the landscape patch index composite values with the following equations:

$$A_i=0.2\times \left(\frac{LP{I}_i}{LP{I}_{\max }}+\frac{{NP}_{\min }}{N{P}_i}+\frac{PAR{A}_{M{N}_{\min }}}{PAR{A}_{M{N}_i}}+\frac{EN{N}_{M{N}_{\min }}}{EN{N}_{M{N}_i}}+\frac{SPLI{T}_{\min }}{SPLI{T}_i}\right),$$

(9)

where $A_i$ is the integrated value of the landscape patch index of the first-class scenario; $LP{I}_i$, $N{P}_i$, $PAR{A}_{M{N}_i}$, $EN{N}_{M{N}_i}$, and $SPLI{T}_i$ are, respectively, the value of the maximum patch index, patch number index, average perimeter area ratio index, Euclidean nearest-neighbor distance index, and separation index of the first-class scenario; $LP{I}_{\max }$ is the maximum value of the maximum patch index in each scenario; and $N{P}_{\min }$, $PAR{A}_{MN\min }$, $EN{N}_{M{N}_{\min }}$, and $SPLI{T}_{\min }$ are the minimum values of each index in each scenario.

3. Results

3.1. Evaluation Results of the Spatial Function and Coupling Coordination of P-L-E

A spatial evaluation system and coupled coordination model were constructed to analyze data for three spatial attributes (P-L-E) in Feixi County from 2010 to 2020, as well as the overall and subsystem utilization levels (

Table 11. Results for the spatial utilization quality of rural P-L-E in Feixi County.

	2010	2015	2020
Production space	0.033	0.197	0.406
Living space	0.002	0.198	0.541
Ecological space	0.025	0.012	0.029

and

Table 12. Results for the comprehensive utilization level of rural P-L-E space in Feixi County.

Year	2010	2015	2020
Feixi County	0.030	0.267	0.649

). The degree of coupled coordination was obtained for Feixi County (

Table 13. Results for coupling coordination degree.

Item	Coupling Degree C-Value	Coordination Index T-Value	Coupling Coordination D-Value	Coordination Level	Degree of Coupling Coordination
2020	0.652	0.212	0.372	3	Medium coupling coordination
2015	0.242	0.247	0.245	1	Low-level coupling coordination
2010	0.122	0.018	0.131	1	Low-level coupling coordination

).

Figure 7 shows the utilization quality of each of the P-L-E spatial types in Feixi County from 2010 to 2020, while Figure 8 shows the comprehensive utilization quality during this period. From 2010 to 2020, the spatial utilization quality of Feixi County increased steadily from 0.033 and 0.002 in 2010 to 0.197 and 0.198 in 2015, reaching 0.406 and 0.541 in 2020 for production and living spaces, respectively. The ecological space saw a trend of decline and then increase, from 0.025 in 2010 to 0.541 in 2015. In general, the comprehensive utilization level of P-L-E spaces in Feixi County was optimized from 0.030 in 2010 to 0.649 in 2020, showing a trend of gradual optimization.

The coupling coordination level in Feixi County was divided by the coupling coordination degree level (Table 5). We can see that the overall coordination level improved greatly, from low-level coupling coordination in 2010 to medium-level coupling coordination in 2020. During this period, the coupling coordination level increased from 0.131 to 0.245 in 2010–2015 and from 0.245 to 0.372 in 2015–2020, and the optimization rate increased (Table 13).

3.2. Spatial Distribution of the Suitability of Spatial Development of P-L-E

The evaluation results indicated that 82.64% of the total area of Feixi County was unsuitable for the development of P-L-E spaces, mainly important ecological space and permanent basic farmland; 9.48% was slightly suitable for development, mainly waters and basic farmland for general ecological regulation; 2.25% was generally suitable for development, mainly ecological–production space; and 5.63% was relatively suitable for development. The main area was the built-up area. The results showed that P-L-E spaces in Feixi County were mainly unsuitable for development (Figure 9).

3.3. Simulation Results of Spatial Distribution Scenarios

3.3.1. Qualitative Analysis for Driving Factors

Table 14. Logistic regression results for land-use driving factors.

Regression Coefficients	Production Space	Ecological Space	Productive–Living Space	Living Space	Eco–Production Space
Average annual precipitation	−0.001170	0.005513	-	−0.002057	0.001344
Water supply	0.000372	0.000393	-	-	0.000430
Hydrological regulation	-	0.005778	-	−0.000894	0.001649
Distance from river system	−1.763042	-	−2.190576	−1.415009	2.098721
Soil texture	0.000029	−0.000115	-	−0.000037	0.000019
Annual normalized vegetation index	0.569112	1.533602	0.689810	−0.430065	0.917689
Net primary productivity	−0.008672	0.003770	0.003542	0.001631	0.010750
Food production capacity	1.184822	−1.102773	−0.508086	0.176642	−0.206951
Raw material production capacity	3.704482	-	-	−5.913095	2.933071
Atmospheric pollution index	−3.572851	−3.798831	4.208653	2.351258	1.837250
Average annual temperature	0.902957	−4.204323	-	1.612922	−1.077030
Soil erosion	-	-	-	−0.006732	0.008976
Distance from each town center	-	10.039127	−10.456873	4.959463	−4.924353
Distance to mineral resources	1.221799	−4.180269	-	-	-
Biodiversity	−32.080199	51.261999	−20.012987	−9.472880	28.411155
Elevations	−0.419370	1.937141	0.355437	0.053486	−0.148695
Slope	-	0.454744	0.143120	-	−0.258234
Slope direction	-	-	-	-	−0.048880
Distance to train station	1.469992	−9.783845	2.375191	3.174095	−7.044560
Distance to highway	0.030546	−0.315732	−0.074175	0.051262	0.059441
Distance to national highway	3.151996	21.074044	−5.182357	−3.056930	−2.528056
Distance to provincial road	−1.214557	6.130374	-	−3.576987	3.428839
Distance to county road	5.762157	-	−6.812342	−9.871236	7.467694
Distance to main road	0.015106	−0.171118	−0.213137	−0.256120	0.059041
Distance to railroad	−0.070780	−0.158040	−0.152683	−0.090970	-
Distance to waterway	−3.340447	14.739082	-	2.503530	9.431854
GDP	−0.000062	−0.000114	-	0.000028	−0.000101
Population density	−0.000656	−0.001129	0.000157	0.000965	−0.000308
Distance to park square	-	−21.686782	6.168784	-	-
Distance to industrial park	6.416128	−18.046922	−2.132867	0.794588	-
Distance to scenic spot	−1.091852	−6.664396	-	3.345872	−3.753178
Distance to important companies	15.897312	13.451898	-	−13.167416	1.815647
Distance to institution	5.031040	−12.684662	−13.984502	−2.867755	8.391972
Distance to hospital	6.570663	−4.241231	-	19.957793	-
Distance to financial services	−10.140912	27.443215	10.804411	3.866193	4.440700
Distance to shopping center	7.033288	−7.178214	−5.664033	−5.706204	-
Environmental purification capacity	−4.075612	18.388761	2.966647	9.688627	-
Constant	0.777194	−6.399624	1.972114	2.088712	−1.243135
ROC	0.856200	0.984890	0.750651	0.837967	0.925830

shows the results of the logistic regression coefficients for each driver relative to the distribution of each spatial type, the results of regression constants and the ROC values for each group.

The regression results show that raw material production capacity, distance to national highway, distance to county road, distance to industrial park, distance to important companies, distance to institution, distance to hospital and distance to shopping center had a stronger positive influence on production space, while biodiversity had a significant negative relationship with it. Distance to financial services played a major positive role in the development of productive–living space, while the distance from each town center and biodiversity played a negative role. For the development of living space, environmental purification capacity and the distance to a hospital played the main positive role. For the expansion of eco–production space, biodiversity was a strong positive driver.

The ROC curve can be used to evaluate the reliability of the fitted results of the logistic regression model, and the high accuracy of the results can be judged when the ROC index is higher than 0.7. All the ROC values in this study were higher than 0.75, indicating that the results in this study for the degree of association between the drivers and the distribution of spatial types are highly accurate (Table 14).

3.3.2. Simulation Results of the Spatial Distribution Scenarios of Three Statuses

The kappa coefficient and Figure of Merit (FoM) index of the overall model simulation accuracy verification index were used to evaluate the feasibility of the FLUS model. The patterns of the ANN training samples include uniform sampling and random sampling; this study used uniform sampling (10/1000). The kappa coefficients of the baseline, ecological priority and comprehensive guidance scenarios were 0.6828, 0.7056 and 0.7647, respectively, and the FoM indexes were 0.0524, 0.0519 and 0.0508, respectively, which were within the feasible accuracy range (

Table 15. Accuracy assessment for FLUS model.

	Baseline	Ecological Priority	Comprehensive Guidance
Kappa	0.6828	0.7056	0.7647
FoM	0.0524	0.0519	0.0508

).

Figure 10 represents the change in the quantitative structure of the five P-L-E spatial types in Feixi County from 2015 to 2030, where the years 2015 to 2020 are based on the classification results, and the years 2020 to 2030 are the prediction results of the Markov chain.

Overall, in terms of quantitative structure, the P-L-E spatial areas in Feixi County showed different development trends under the three simulation scenarios from 2015 to 2030 (

Table 16. Multiscenario simulations of the number structure of P-L-E space sites in Feixi County.

Simulation Scenarios	Production Space		Ecological Space		Production–Living Space		Living Space		Eco–Production Space
	Number of Pixels/pcs	Percentage/%	Number of Pixels/pcs	Percentage/%	Number of Pixels/pcs	Percentage/%	Number of Pixels/pcs	Percentage/%	Number of Pixels/pcs	Percentage/%
Status scenario in 2015	310,649	82.10%	9554	2.53%	971	0.26%	28,893	7.64%	28,303	7.48%
Baseline scenario in 2030	303,101	80.11%	12,370	3.27%	1066	0.28%	34,846	9.21%	26,987	7.13%
Ecological priority scenario in 2030	301,956	79.80%	12,368	3.27%	1853	0.49%	34,847	9.21%	27,346	7.23%
Comprehensive guidance scenario in 2030	308,168	81.45%	7007	1.85%	1070	0.28%	34,847	9.21%	27,278	7.21%

). Ecological space increased from 2.53% to 3.27% under both the baseline development scenario and the ecological priority scenario but decreased to 1.85% under the comprehensive guidance scenario. Production space and eco–production space decreased from 82.10% and 7.48% in 2015 to 80.11% and 79.80%, and 81.45% and 7.13%, and 7.23% and 7.21% in 2030 under the three modeling scenarios, respectively, with the former decreasing more under the ecological priority scenario. By contrast, production–living space and living space increased from 0.26% and 7.64% to 0.28%, 0.49%, 0.28% and 9.21%, respectively, during the simulation period, with the former increasing significantly under the ecological priority scenario and the latter increasing significantly under all three scenarios. In terms of distribution (Figure 11), production space was spread over the whole area; ecological space was mainly distributed in the green wedge of Zipeng Mountain; living space and production and living space were concentrated in Shangpai, Taohua, Guanting, Zipeng, Shanan and Huagang; and ecological–production space was mainly scattered in the eastern part adjacent to Chaohu Lake.

In the baseline development scenario, driven entirely by natural factors, in terms of spatial quantity structure (Table 16), each type of space accounted for 80.11%, 3.27%, 0.28%, 9.21% and 7.13% compared to the scenario in which ecological space and living space increased significantly at the same time. Meanwhile, the other three types of spaces did not change significantly, indicating that under the situation of no human intervention, ecological space and living space did not change significantly. In terms of spatial distribution, new land development was concentrated in the vicinity of the existing transportation network and built-up areas. The town of Shangpai expanded to the west and south, showing a trend of “pancake spreading”, while new dense urban construction areas appeared in the northern Guanting area (Figure 11b).

The ecological priority scenario emphasizes the protection of ecological space and the carrying capacity of resources and the environment. In terms of spatial quantity structure (Table 16), the proportions of production space, ecological space, production–living space, living space and eco–production space are 79.80%, 3.27%, 0.49%, 9.21% and 7.23%, respectively. In this scenario, the production space is reduced by 2.30% while the eco–production space is reduced by 0.25%, and the ecological space, production-living space and living space are increased by 0.74%, 0.24% and 1.57%, respectively. In contrast, the growth trend of living space is more significant; however, in the ecological priority scenario, the transformation of production space is more inclined to living space. This reflects the irreversibility of urbanization, under which it is difficult to mitigate the harm caused to ecological spaces. The spatial distribution of each spatial development trend was similar to the baseline scenario, but under the control of the ecological protection red line, the expansion of living space was more moderate and the directionality was clearer (Figure 11c).

Under the comprehensive guidance scenario, accounting for the comprehensive development and multiple needs of the five types of spaces. In terms of spatial quantity structure (Table 16), the proportions of production space, ecological space, production-living space, living space and eco–production space are 81.45%, 1.85%, 0.28%, 9.21% and 7.21%, respectively. Compared with other scenarios, the transformation of the five types of spaces in the comprehensive guidance scenario is more balanced. In terms of spatial distribution, the expansion approach took the old city as the core and expanded to the west and south. The wedge-shaped embedded green space formed by Zipeng Mountain on the west side stabilized the ecological background, together with the connection area of Chaohu Lake on the east side (Figure 11d).

3.3.3. Comprehensive Evaluation Results of Landscape Patches

The results of the landscape patch index (

Table 17. Comprehensive evaluation of landscape patches in Feixi County under three scenarios.

Simulation Scenarios	CA/ha	LPI/%	NP/pc	PARA_MN	ENN_MN/m	SPLIT	Integrated Value A
Baseline	353508.7500	51.8460	8144.0000	555.1848	247.9873	2.3783	0.9653
Ecological priority	353508.7500	51.8460	8592.0000	559.1992	259.3545	2.4146	0.9435
Comprehensive guidance	353508.7500	51.8460	6731.0000	565.6931	260.3648	2.4136	0.9839

) showed that the three simulated scenarios had the same proportion of largest patches (LPI, 51.8460), indicating there was no difference in the dominance of the core patches. In comparison, the integrated guided scenario had better patch integrity, as reflected by the smaller number of patches (NP, 6731) and relatively smaller separation (SPLIT, 2.4136) (i.e., lower patch fragmentation). The baseline development scenario had more advantages in patch proximity and patch shape, as reflected by the smaller average perimeter-to-area ratio index (PARA_MN, 555.1848), the Euclidean nearest-neighbor index (ENN_MN, 5.548) and the distance index (ENN_MN, 247.9873). The core patches were more complete, the patches were more closely connected and the shape was simpler and more regular. At the same time, the composite value of the comprehensive guidance scenario (0.9839) was also higher than in the other scenarios, reflecting the leading role of dominant patches in the comprehensive guidance scenario, which has greater value for achieving regional coordination and integration.

4. Discussion

In this study, Feixi County was chosen as the study area, on the one hand, because Feixi County is dominated by cropland and each spatial type is clear, which is convenient to further advance the refinement of the driving factors based on the existing spatial simulation study mainly on cropland from the perspective of P-L-E space [41,69]. On the other hand, few studies have been conducted at the county level [70]. Since Feixi County is located in the core area of the Hefei metropolitan area and Wanjiang demonstration zone, it is an important part of Hefei city to build a “competitive national central city” in the future, so it has important evaluation and simulation value [43].

4.1. Spatial Function and Coupled Evaluation

Policy is a factor that influences changes in ecological issues [71]. From 2010 to 2015, Feixi County focused on economic development and neglected ecological protection, emphasizing the development of production and living space. This resulted in a serious stagnation of ecological space development. This is consistent with the earlier trend of China and other developing countries to develop their economies to some extent at the cost of environmental resources [72,73,74]. In 2015, in order to change the deteriorating local ecological environment, Feixi County adjusted its development focus to maintain economic development while focusing on ecosystem protection, ecological space restoration and ecological environmental management.

In general, under the guidance of the Feixi County government’s ideas of deepening land-use patterns, accelerating urban–rural integration and optimizing spatial layout, the results showed P-L-E spaces reaching an optimized state from 2010 to 2030. However, overall quality was still at a low level, and the degree of coupling and coordination needs further improvement.

4.2. P-L-E Space Pattern and Driving Forces

The transformation of the LULC is influenced by various natural, social and economic aspects [45,75], and there are current LULC simulations that focus on exploring the influence of a single factor on LULC change, such as population [41,76] or vegetation [45]. However, most of the studies that applied the FLUS model for simulation lack the integrated application of factors involving multidisciplinary drivers, such as the lack of consideration of other key factors such as hydrology in their use [48]. Therefore, in this study, based on the integration of natural, economic and social factors, a corresponding and complete factor system has been constructed for the actual situation and requirements of Feixi County.

The results obtained from the study show that between 2015 and 2020, there was a significant decrease in production space and production–living space, as well as a significant increase in the area of ecological space and living space, respectively, in Feixi County. However, according to the results of the simulation analysis based on improved driving factors, the amount of production space and production–living space was stable after 2020, which indicates that the adjustment of the development strategy of the government of Feixi County to ecological improvement began to show effects during this period [71], which is consistent with the trends reflected in Table 7, Figure 7 and the results obtained in Section 4.1.

With the accelerated industrialization and urbanization, Feixi County has been developing rapidly in social and economic fields, but at the same time, the deterioration of the ecological environment has become an urgent problem, which is directly reflected in the low quality of ecological space utilization (Figure 7). Since 2012, the Chinese government has clearly pointed out the importance of ecological civilization construction, and Feixi County has responded to the national call to strengthen the top-level design of ecological civilization, promulgating the “Comprehensive Creation of Clear Water, Green Shore and Industrial Excellence Beautiful Yangtze River (Feixi) Economic Belt Implementation Plan”, and then fully implementing the main functional area strategy. At the same time, special funds have been established to increase financial investment and adhere to the principle of “prevention and treatment at the same time” to rectify outstanding environmental problems. The direct performance reduces the area of production–living space, restores the area of ecological space and improves the carrying capacity of resources and the environment. Meanwhile, according to the results of logistic regression coefficients of driving factors, the main positive influence on ecological space was biodiversity, that is, the higher the degree of biodiversity, the higher the probability of ecological space. In this study, the corresponding policy measures factors and natural factors were also considered in the simulation, which are reflected in the constructed database of resource and environmental carrying capacity evaluation factors (Table 4).

In this study, the production space was dominated by farmland, and considering the previous studies on the relationship between population density and farmland [41,45], it is also clear from the logistic regression results that the population density also showed a negative correlation for the production space; that is, the more centralized the population, the lower the probability of the production space. According to the Hefei Municipal Bureau of Statistics, the urbanized population is increasing, the expansion of towns is accelerating, the area of production space is shrinking and the area of living space is increasing, which is consistent with the results in Table 7 and Figure 7.

The main reason for the absence of significant changes in the area of eco–production spaces compared to other spatial types may be that this type of space consists mainly of water bodies, which did not change significantly during the study time frame.

4.3. Development Suitability and Land-Use Simulation

The P-L-E spatial pattern plays an important role in sustainable land use and development [69,77]. Spatial optimization is a current research hotspot, and it is necessary to determine the appropriate spatial distribution of existed urban spaces in order to alleviate the contradictions in the urban development and planning process and to coordinate ecological, economic and social development [78,79].

The results for development suitability showed that 82.64% of unsuitable spaces were scattered throughout Feixi County, mainly permanent basic agricultural land and important ecological spaces, such as Zipeng Mountain and the area connected to Chaohu Lake (Figure 9). Resource and environmental carrying capacity and development potential should be reasonably evaluated, and such spaces should be better controlled to give full play to ecological benefits. The slightly suitable development space, generally suitable development space and relatively suitable development space are relatively small, and the former two are mainly water areas and basic farmland that play a general ecological regulating role while the latter is mainly the built-up area, concentrated in Shangpai, Taohua, Guanting and Zipeng. This built-up area can be suitable as the main development area in the future, coordinating the urban–rural relationship, undertaking urban functions and promoting overall development. This is consistent with the results of a previous study on suitable future construction areas in Hefei City [80]. According to the distribution of the P-L-E spaces in Feixi County, the unsuitable space is mainly production space and ecological space, the slightly suitable space is mainly production–living space and eco–production space, the generally suitable space is mainly production–living space and the suitable space is mainly living space.

In terms of spatial quantity structure, the three simulation scenarios based on the natural background of Feixi County were all dominated by production space (Table 16). Further, production space, ecological space, production–living space, living space and ecological–production space all showed approximate proportional relationships. Among them, ecological space and living space significantly increased under the baseline development scenario. The other three types of spaces did not change significantly. There is a significant decrease in production space under the ecological priority scenario, insignificant changes in eco–production space and growth in ecological space, production-living space and living space. The transformation of the five spatial types was therefore more balanced under the comprehensive guidance scenario.

In terms of spatial distribution, the three simulation scenarios differed in terms of corresponding development suitability results. Under the baseline development scenario, ecological space and living space in Zipeng Mountain and Shangpai in north-central Feixi County expanded significantly, and new, dense urban construction areas appeared in the area of Guanting. However, in order to better integrate the relationship between human and nature, urban development needs to respect the development goals and ecological principles proposed in the plan, and the distribution and expansion direction of living space needs to be reasonably planned, and the ecological red line and red line policy for cropland needs to be strictly implemented [81,82,83]. Under the ecological priority scenario, ecological space mainly expanded in the Zipeng Mountain area, which has been effectively protected, while the changes in the other spatial types were relatively small. However, this scenario does not avoid the occupation of production space including basic cropland in the ecological restoration process, and the contradiction between cropland protection and ecological protection has not been weighed, which is not conducive to the sustainable development of P-L-E space [84]. Under the comprehensive guidance scenario, the core of Shangpai expanded to the west and south. On the premise of not encroaching on ecological space, the expansion of living space was mostly mild, and the changes in ecological–production and production–living spaces were relatively small. A multispatial planning structure of the county, with Shangpai as the focus of development, is formed accordingly. Specifically, ecological space takes Zipeng Mountain as the core, and the ecological nodes are linked to form an ecological network. The production space in the north is integrated with the living space in the south for linkage and rational development. The ecological–production space and the production–living space are interpenetrated to meet multifunctional needs [6,85,86].

Meanwhile, the landscape patch index results showed that the number of patches under the comprehensive guidance scenario was smaller, and the separation degree was relatively small—that is, the patch integrity is better. Its combined value was also higher than in other scenarios, reflecting the greater value of achieving regional coordination and integration.

4.4. Recommendations and Future Strategies

Based on the evaluation of the quality of the utilization of the P-L-E spatial areas, the evaluation of the suitability of development and the simulation of the three spatial areas to obtain the above research results, we found that the following: (1) The 2030 baseline development scenario showed unrestricted spatial development and an overconcentration of urban construction land. This does not meet the future development needs of Feixi County and ignores the spatial coupling and coordination of P-L-E spaces. (2) The 2030 ecological priority scenario focuses on restoring biodiversity while ensuring the quantitative index requirements for the requirements of high-quality ecological space. Meanwhile, urban space shows a desirable expansion trend but lacks comprehensive service to other spatial types, which does not meet the planning needs of Feixi County. Additional consideration of other aspects such as basic cropland policy needs to be added [75]. (3) In the rapid urbanization process, population growth leads to an increased demand for land resources [70]. In the 2030 comprehensive guidance scenario, the expansion strategy for city spaces is combined with the development process, and the integrated development needs of the three spatial types, as well as the need for human–nature coupling, are considered. The control of population concentration areas and the encouragement of clustering to non-town core areas are in line with the logistic regression results. Compared with the other scenarios, the simulation results of the comprehensive guidance scenario were more intensive and efficient, showing a directional nature. Its spatial quantity structure and distribution form can achieve the development concept of the three spatial areas in Feixi County by 2030.

In light of the above discussion, the following recommendations are made for the optimization and adjustment of the future use of P-L-E spaces in Feixi County: (1) Focus on restoring biodiversity and increasing the area of ecological space; (2) guide the migration of people to non-urban core areas; and (3) ensure the implementation of ecological policies and strengthen special funding to protect the natural quality of the study area.

5. Conclusions

This study integrates resource and environmental carrying capacity and development suitability, and regarding the requirements of the P-L-E functions, chooses the P-L-E space of Feixi County as the research object, providing an integrated research system combining spatial quality evaluation, a coupled coordination model, FLUS model, landscape pattern index and development suitability evaluation. The system helps to fill the gap in the current P-L-E spatial research in the urban development process and overcome the lack of comprehensive ideas and practical applications of overall optimization. At the same time, the system integrates natural, social and economic factors, improves the drivers used in the current FLUS model simulation and helps to improve the predictive capability of the FLUS model.

This study demonstrates the quality of spatial utilization and the degree of coupling coordination in Feixi County from 2010 to 2020 and reveals the influence of the spatial distribution of the P-L-E space on spatial development from 2015 to 2030. Among them, 2015 is the inflection point of ecological spatial utilization quality in the study area, and the prominent ecological problems in Feixi County have improved after the adjustment of the development focus. From the perspective of spatial type change, the main reasons influencing the development of the P-L-E space in Feixi County are the natural environment, human activities and development policies. Human activities built on the basis of the natural environment have a negative impact on the development of the P-L-E space types, highlighted by the decrease in the area of production space in the population concentration areas. In contrast, policies focusing on ecological space development lead to an improvement trend in the ecological space. The simulation results in 2030 show that the area of production space decreases in all three types of scenarios. Under the baseline development scenario, ecological space and living space expand significantly. Under the ecological priority scenario, ecological space has been effectively protected, while the changes in the other spatial types were relatively small. Under the comprehensive guidance scenario, on the premise of not encroaching on ecological space, the expansion of living space was mostly mild, and the changes in eco–production and production–living spaces were relatively small. Together with the development suitability and the landscape pattern index, it is concluded that the spatial quantity structure and spatial distribution of the P-L-E space under the comprehensive guidance scenario are more consistent with the vision of the development of the P-L-E space in Feixi County in 2030.

In conclusion, the spatial utilization quality results, scenario simulation results and the revealed relationships between the driving factors and the development of the P-L-E space types in this study are conducive to improving the scientificity, flexibility and management efficiency of Feixi County’s P-L-E spatial layout, thereby supporting its sustainable development.

However, due to the complexity of the research object, the current study still has certain shortcomings, mainly in the current land spatial classification system and spatial identification still needs to be further defined and studied. Further systematic and identification research is needed in the future, so as to gain wider recognition, which can better guide the practice of P-L-E spatial optimization.