Unveiling scale effects in human settlement environment suitability through a novel multi-factor weighting approach

Abstract

Rural human settlement environment suitability is a key indicator for guiding land-use planning and promoting sustainable rural revitalization. This study develops a refined framework for evaluating the natural suitability of rural human settlement environments in central Hunan, China, using GIS-based spatial analysis. We propose an enhanced factor-weighting method that integrates normalized information entropy (NIE) and the coefficient of variation (CV) into a Pearson correlation coefficient (PCC)-driven approach, thereby improving the objectivity and sensitivity of weight determination. Comparative experiments demonstrate that this hybrid method captures the relative influence of natural factors more accurately than traditional PCC-based techniques. Results reveal a pronounced scale effect: as the spatial analysis scale decreases, disparities in factor weights become more evident. The spatial distribution of suitability exhibits a distinct east–high, west–low gradient, reflecting strong spatial heterogeneity across the region. Among all factors, terrain relief emerges as the dominant driver of suitability variations. Based on the calculated suitability index, the study delineates five functional zones and formulates corresponding development strategies tailored to local conditions. These findings offer a scientific basis for optimizing rural spatial planning and promoting sustainable development in central Hunan and similar regions. Beyond the case region, the framework provides a transferable approach for evaluating natural suitability in other mountainous rural areas. It also offers practical guidance for policymakers seeking to balance ecological protection with sustainable land-use development.

Introduction

Habitat suitability evaluation is a key topic in urban–rural planning, ecological environmental protection, and sustainable development research^1,2. In the mid-20th century, C.A. Doxiadis introduced the concept of Ekistics, the science of human settlements³—which laid the theoretical foundation for research on human settlements. In China, Academician Wu Liangyong proposed the theory of human settlement science in 2001⁴, dividing human settlements into five systems: natural, human, social, residential, and infrastructural, thereby providing a comprehensive theoretical framework for subsequent studies.

Currently, research on human settlement suitability primarily focuses on the analysis of natural environmental factors, the construction of comprehensive evaluation models, and the expansion of research scales. In terms of natural factors, scholars have investigated the effects of climate^5–7, topography^{8– 9}, hydrology^10,11, and vegetation¹² on human settlements^13–17. With continuous methodological advancements, recent studies have developed multi-index evaluation systems, often integrating GIS technology with methods such as principal component analysis (PCA)¹⁸, entropy weighting within the Pressure–State–Response (PSR) framework¹⁹, geographically weighted regression (GWR)²⁰, and minimum cumulative resistance (MCR) modeling²¹, thereby improving the scientific rigor and reliability of such evaluations. Additionally, the spatial scale of research has progressively shifted: earlier studies focused on national^22–24 or provincial levels^25,26, while more recent work has turned toward finer scales such as urban agglomerations^12,27, watersheds^28–30, and townships^31,32 to enhance the regional precision of suitability assessments.

Most previous studies have been conducted at larger spatial scales—either national or provincial^5,8,23. A limited number of medium-scale studies focusing on sub-provincial regions have often treated rural and urban areas as a single unit^13,33,34. In practice, however, the factors influencing urban settlement suitability are highly complex, with economic and political variables often playing dominant roles, whereas rural settlement suitability is more directly influenced by natural environmental factors^35,36. Furthermore, many studies that specifically address rural areas^31,32,37 tend to apply models developed for broader spatial contexts without scale-specific optimization, leaving room for methodological refinement.

However, existing HEI quantification methods still have several limitations. Many rely on subjective expert scoring or single-indicator weighting, which may reduce objectivity and sensitivity. In addition, some commonly used weighting methods are not well adapted to medium-scale rural contexts and lack scale-specific optimization. These issues highlight the need for a more objective and scale-adaptive weighting approach.

Accordingly, this study focuses on a medium spatial scale, examining the natural suitability of rural human settlement environments within a sub-regional area. It also aims to optimize the evaluation model to more accurately quantify the influence of natural constituent factors on human settlement suitability.

The central Hunan region is one of China’s major grain-producing areas. It represents a moderate spatial scale and features well-matched rainfall and temperature conditions, providing a favorable natural foundation for agriculture. However, the terrain is complex—dominated by hills and mountains—and land use is highly diverse^38,39. Urbanization remains relatively low, with a large proportion of the population engaged in agriculture^40,41. The region also experiences lagging economic development and limited infrastructure coverage⁴². Under these conditions, assessing the natural suitability of the human settlement environment in central Hunan holds strong practical significance for promoting sustainable agricultural development and improving residents’ quality of life.

At present, most research on central Hunan focuses on village construction styles and vernacular architecture, rather than on the natural suitability of the environment⁴³. While such studies have contributed to improving village environments and residential spaces, they have not systematically examined the natural environmental suitability of rural areas in central Hunan.

Meanwhile, some scholars have evaluated human settlement suitability across Hunan Province as a whole. For example, Li Bohua and colleagues examined macro-scale characteristics of human settlements in terms of economic coupling⁴⁴ and regional development levels⁴⁵. Zeng Yi et al., guided by national land spatial planning needs, classified the natural suitability of land space in Hunan into ecological, agricultural, and construction categories based on natural conditions and management objectives⁴⁶. While these studies provide comprehensive macro-level analyses, they do not offer in-depth investigations specific to central Hunan. Therefore, there remains significant room to expand research on the natural suitability of rural human settlements in this region.

This study aims to improve the existing evaluation model and reveal the spatial characteristics of rural settlement suitability in central Hunan, thereby providing a scientific basis for optimizing the rural human settlement environment, enhancing agricultural productivity, and guiding sustainable regional development.

Study area and data sources

Study area

The central Hunan region is situated in the middle of Hunan Province and primarily includes Loudi City, Shaoyang City, and Anhua County in Yiyang City, covering 18 county-level administrative units. It spans from 110°30′ to 112°30′ E and from 26°00′ to 28°00′ N, with a total area of approximately 34,100 km². The total population is around 10.87 million, yet the average urbanization rate is only 51.54%, significantly lower than both the provincial average (61.16%) and the national average (66.16%). The region has a large agricultural population (about 5.27 million), and its per capita GDP is approximately 46,000 CNY—roughly half of the national average.

Geologically, the region lies within the Xiangzhong–Xiangdongnan fold zone, with an average elevation of about 500 m. It is bordered by mountains to the west, south, and north. The Xuefeng Mountains in the west give the region a terrain pattern of higher elevation in the west and lower in the east (Fig. 1). The area is drained by the Zi River and Yuan River basins, forming a dense and complex hydrographic network. Fluvial erosion in the eastern region has shaped a variety of landforms dominated by hills. A central mountain ridge further divides this eastern hilly zone into the Loudi Basin and the Shaoyang Basin.

Fig. 1.

Elevation in central Hunan.

The underlying geology consists mainly of discontinuous red beds and limestone, with weak volcanic activity, rendering the area low in seismic risk. Favorable geochemical conditions have facilitated the formation of mineral deposits, and the region is rich in coal, iron, copper, antimony, and other mineral resources.

The map was generated using ArcGIS 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com).

Central Hunan has a subtropical monsoon climate, with a mean annual temperature of approximately 17 °C and average annual precipitation around 1,500 mm. The temporal alignment of rainfall and heat is favorable for rice cultivation and also supports the growth of tea, oil-tea camellia, and other cash crops. However, due to the concentration of rainfall in spring and summer, dry-season precipitation amounts to less than half of that during the wet season. Periodic floods and droughts pose recurrent threats to agricultural production.

Additionally, the region experiences high humidity levels alongside temperature extremes: in both summer and winter, relative humidity often exceeds 90%. This combination leads to hot and humid summers and cold, damp winters, increasing discomfort during periods of extreme weather.

Data sources

The basic data used in this study and their sources are as follows:

Administrative boundaries: County-level administrative boundary data for central Hunan were obtained from the National Geoinformation Resource Catalog Service (https://www.webmap.cn).

Socioeconomic data: County-level data on total population and economic indicators were collected from the Hunan Provincial Bureau of Statistics (https://tjj.hunan.gov.cn).

Population distribution: Gridded population data with a spatial resolution of 100 m, based on China’s Seventh National Census, were obtained from datasets shared by Professor Chen Yuehong’s research team on Figshare (https://figshare.com/s/d9dd5f9bb1a7f4fd3734).

Elevation data: Digital elevation model (DEM) data with a 30-meter resolution were acquired from the Geospatial Data Cloud (https://www.gscloud.cn).

Land use: Land use vector data (30-meter resolution) were provided by the International Research Center of Big Data for Sustainable Development Goals (CASEarth) (https://data.casearth.cn/thematic/glc_fcs30/314).

Hydrology and climate: River network data, monthly average temperature, and monthly average relative humidity data were sourced from the National Earth System Science Data Center (https://www.geodata.cn).

Vegetation index: Normalized Difference Vegetation Index (NDVI) data were accessed via NASA EarthData (https://www.earthdata.nasa.gov).

Although these datasets represent the latest publicly accessible sources, their temporal resolution does not fully capture rapid socio-economic changes occurring in some rural areas. This may introduce a degree of uncertainty into the analysis. However, as the study focuses primarily on natural environmental suitability—where the underlying variables (elevation, climate, hydrology, land cover) change relatively slowly—the impacts of data timeliness on the main findings are expected to be limited. Small gaps or missing pixels in the raster datasets were pre-processed using nearest-neighbor filling to maintain spatial continuity. These missing values were extremely limited in spatial extent, and therefore their influence on the final HEI results is negligible.

Research model and improvements to the calculation method

Research model

Feng Zhiming et al.²² developed a Human Settlement Environment Suitability Evaluation Model—known as the Human Settlement Environment Index (HEI)—which has been widely applied and serves as an important reference in related research. This study adopts the HEI model, integrating multiple natural factors to construct a comprehensive index for evaluating human settlement suitability (Fig. 2). The selected natural factors include the Relief Degree of Land Surface (RDLS), Temperature-Humidity Index (THI), Water Resource Index (WRI), and Land Cover Index (LCI).

Fig. 2.

Construction of the human settlement environment suitability index (HEI) in central Hunan.

Relief degree of land surface (RDLS)

1

In this equation, RDLS represents the Relief Degree of Land Surface. ALT is the average elevation (m) within a defined neighborhood around a grid cell. max(H) and min(H) denote the maximum and minimum elevations (m) within that neighborhood. P(A) is the area of flat land (slope ≤ 5°) within the neighborhood (km²), and A is the total area of the neighborhood (km²). The RDLS was calculated using a circular spatial neighborhood with a radius of 1.75 km.

Temperature-humidity index (THI)

2

In this equation, THI is the Temperature-Humidity Index; t is the monthly mean air temperature (°C); and f is the monthly mean relative humidity (%).

Water resource index (WRI)

3

In this equation, WRI is the Water Resource Index; P is the normalized mean annual precipitation; and Wa is the normalized surface water area. α and β are the weights for precipitation and water area, respectively. Following the study by Hao Huimei et al.¹⁵, we set α = 0.8 and β = 0.2.

Land cover index (LCI)

4

In this equation, LCI is the Land Cover Index; NDVI is the Normalized Difference Vegetation Index; and LT_i is the land use type coefficient for land cover type i (where i = 1, 2, …, 25). The land use type coefficients are provided in Table 1¹³.

Table 1.

Land use type coefficients.

Land use type	Paddy	Dry land	Forest	Shrub	Sparse forest	Other forest
Coefficient	1.0	0.7	0.6	0.6	0.4	0.4

Land use type	Grass (H)	Grass (M)	Grass (L)	Urban	Rural res.	Other built-up
Coefficient	0.6	0.6	0.6	0.8	0.6	0.4

Land use type	River	Lake	Res. & Pond	Shoal	Swamp	Bare land
Coefficient	0.6	0.6	0.6	0.4	0.5	0.2

*“Res.&Pond” is the abbreviation for “Reservoirs and Ponds”.

The data for each constituent factor were normalized and classified using the Natural Breaks (Jenks) method. Their spatial distribution across central Hunan is presented (Fig. 3).

Fig. 3.

Spatial distribution of natural suitability factors. (a) Terrain relief (RDLS); (b) Temperature-humidity index (THI); (c) Water resource index (WRI); (d) Land cover index (LCI).

The map was generated using ArcGIS 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com).

Human settlement environment suitability index (HEI)

The calculation equation for the Human Settlement Environment Suitability Index (HEI) is as follows (Eq. 5):

5

HEI denotes the Human Settlement Environment Suitability Index. P_RDLS, P_THI, P_WRI, and P_LCI represent the raster values of the four constituent factors—RDLS, THI, WRI, and LCI, respectively. W_RDLS, W_THI, W_WRI, and W_LCI are the corresponding weights assigned to these four factors.

This model is widely recognized for its scientific validity and practical applicability. However, the current approach to calculating the weights of constituent factors leaves room for improvement. Existing weighting methods mainly fall into two categories: the Analytic Hierarchy Process (AHP)^13,47,48 and correlation-based methods^22,49.

The AHP method relies on expert judgment to construct a hierarchical structure model for determining the weight distribution across different levels. However, it is highly dependent on expert scoring and is vulnerable to individual cognitive biases and subjective tendencies, which may compromise the model’s objectivity.

In contrast, correlation-based methods calculate the Pearson Correlation Coefficient (PCC) between each constituent factor and population density, and then convert these correlation coefficients into weights through range normalization.

Compared with AHP, the correlation-based method is relatively more objective. However, in previous implementations, the Pearson Correlation Coefficient (PCC) was used as the sole criterion for determining factor weights. This approach overlooks the distributional characteristics of the factor data themselves, which may lead to biased weight allocation. For example, in a region where the water resource index is consistently high and both precipitation and surface water distribution are relatively uniform, common sense suggests that the hydrological factor should not be assigned a high weight in the evaluation of natural suitability. Nevertheless, if the water resource index shows a strong correlation with population density, it may still be given an unreasonably high weight, thereby compromising the rationality of the weighting process.

To address this issue, this study retains the fundamental structure of the HEI model while optimizing and improving the weight calculation method. The objective is to ensure methodological objectivity while also aligning with reasonable, common-sense judgments about the significance of each factor.

For consistency across all factor calculations, the final HEI evaluation was conducted on a unified raster grid with a spatial resolution of approximately 1.2 km × 1.2 km. The output raster contains 250 columns and 279 rows (69,750 grid cells in total), covering the full bounding extent of the study area, while only the cells corresponding to the actual rural areas of central Hunan were used for the suitability analysis.

This figure is a schematic diagram illustrating the methodological workflow.

Improved method for calculating factor weights

The core idea of the improved weighting method proposed in this study is to retain the Pearson Correlation Coefficient (PCC) framework while introducing Normalized Information Entropy (NIE) and the Coefficient of Variation (CV) to correct for biases caused by differences in the statistical distributions of factor data (Fig. 4). Multiple weighting methods are compared and validated to identify the optimal approach.

Fig. 4.

Improved method for calculating factor weights.

Normalized information entropy and coefficient of variation

Information entropy (IE), originally proposed by C.E. Shannon in information theory, is a metric used to quantify the information content or uncertainty within a dataset. In this study, IE is applied to evaluate the informational contribution of each factor. To ensure comparability across different factors, the entropy values are normalized, resulting in Normalized Information Entropy (NIE). The formula for calculating NIE is shown in Eq. (6):

6

In this equation, NIE denotes the normalized information entropy; n is the number of possible states in the data; and p_i is the probability of the i-th state. Since a smaller NIE indicates a higher information contribution, we use (1 − NIE) to represent the contribution of information entropy to the weighting of each factor.

The Coefficient of Variation (CV) is a statistical measure of relative dispersion, reflecting the degree of variability in relation to the mean. It eliminates the influence of measurement units, enabling direct comparison of variability across different factors and ensuring that high-variability factors are not undervalued. The CV is calculated using Eq. (7):

7

In this equation, CV is the coefficient of variation; σ is the standard deviation; and µ is the mean of the data.

Validation of the improved method

To validate the feasibility of the improved weighting method, we calculated the Human Settlement Environment Index (HEI) for rural areas in central Hunan using several different weighting schemes. We then assessed the correlation between each resulting HEI and population density. A higher correlation indicates that the corresponding weighting method better captures the relationship between natural suitability and actual settlement patterns.

The various weighting methods compared in this study are summarized in Table 2.

Table 2.

Weight calculation methods compared in this study.

Method	Weighting method	Description
1	PCC[22]	PCC-based weights from a prior study
2	PCC	Weights calculated using only PCC
3	1-NIE	Weights calculated using only (1 − NIE)
4	CV	Weights calculated using only CV
5	PCC×(1-NIE)	Weights calculated using PCC × (1 − NIE)
6	(1-NIE)×CV	Weights calculated using (1 − NIE) × CV
7	PCC×CV	Weights calculated using PCC × CV
8	PCC×(1-NIE)×CV	Weights calculated using PCC × (1 − NIE) × CV
9	AHP[13]	AHP-based weights from a prior study

The weights assigned to each constituent factor using the different methods are presented in Table 3. Because the weights are obtained by correlating each method-derived factor score (PCC, 1 − NIE, CV, or their combinations) with population density, both positive and negative values may appear, indicating positive or negative associations respectively.

Table 3.

Weights of each factor under different calculation methods.

Method	1	2	3	4	5	6	7	8	9
WRDLS	− 0.28	− 0.48	0.24	0.43	− 0.44	− 0.72	0.45	− 0.72	− 0.29
WTHI	0.28	0.29	0.36	0.08	0.39	0.08	0.12	0.11	0.12
WWRI	− 0.19	− 0.14	0.23	0.27	− 0.12	− 0.13	0.27	− 0.12	0.43
WLCI	− 0.25	− 0.10	0.17	0.22	− 0.06	− 0.07	0.15	− 0.05	0.07

Using these weights, the HEI for rural human settlements in central Hunan was calculated under each method (Fig. 5).

Fig. 5.

HEI results derived from different weighting methods. (a) Method 1; (b) Method 2; (c) Method 3; (d) Method 4; (e) Method 5; (f) Method 6; (g) Method 7; (h) Method 8; (i) Method 9.

The map was generated using ArcGIS 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com).

We then calculated the correlation between each HEI and population density. Using Method 2 (PCC only) as the baseline, we computed the improvement rate in correlation for each method. The correlation coefficients and corresponding improvement rates are presented in Table 4.

Table 4.

HEI–population density correlation and improvement by Method.

Method	1	2*	3	4	5	6	7	8	9
Correlation	0.53	0.64	0.15	0.12	0.72	0.70	0.22	0.77	0.45
Improvement rate	− 17%	0%	− 77%	− 81%	13%	9%	− 66%	20%	− 30%

*Improvement Rate is calculated relative to Method 2.

As shown in Table 4, the HEI values calculated using Method 1 and Method 9—both of which apply weights directly cited from previous studies—exhibit lower correlations with population density than Method 2 (PCC only), as indicated by negative improvement rates. Method 2²² provides the baseline HEI using simple correlation-based weighting, but its performance is exceeded by Methods 5, 6, and 8.

Methods 3 and 4, relying solely on NIE or CV, yield relatively low correlation coefficients. Their combination in Method 7 also performs poorly. However, combining PCC with NIE (Method 5) or CV (Method 6) improves the correlation with population density. Notably, Method 8, which integrates all three—PCC, NIE, and CV—achieves the highest correlation coefficient of 0.77, representing a 20% improvement over the PCC-only baseline (Method 2).

These results demonstrate that Method 8 (PCC×(1 − NIE)×CV) is a highly feasible approach for calculating factor weights. It produces a more objective and accurate Human Settlement Environment Suitability Index (HEI) for the study area. Although population density is a practical and widely used proxy indicator for validating natural suitability in prior HEI research, it does not capture socio-economic dimensions such as economic development, education, and healthcare accessibility. Because this study focuses exclusively on natural environmental suitability, population distribution was selected as the primary validation indicator. Nevertheless, we acknowledge that multidimensional socio-economic validation would strengthen the comprehensiveness of the assessment, and this is identified as an important direction for future research. In addition, the present study focuses on developing an objective weighting approach, and therefore does not incorporate expert judgment or field-based validation. Future research should integrate expert assessments or field investigations to further support the rationality of the weighting results.

Effect of spatial scale on factor weights

A comparison between the factor weight distributions obtained in this study and those reported in previous literature reveals that the spatial scale of the study area significantly influences the allocation of weights among factors.

Table 5 lists the factor weights reported in multiple studies conducted at different spatial scales. These studies were selected because they employ similar indicator systems with consistent definitions and sign conventions, ensuring that the weight values are directly comparable. To facilitate comparison, we normalized the weights using range normalization and calculated the standard deviation of the weights in each study. A larger standard deviation indicates greater disparity in the allocation of weights among factors.

Table 5.

Comparison of factor weight distributions at different spatial scales.

Study area	Central China region22	Shaanxi province1	Henan province50	Ganjiang river basin51	Fenhe river basin29	Central Hunan rural (This study)
Total Area (104 km2)	56.47	20.58	16.70	8.35	3.95	3.41
WRDLS	− 0.28	0.14	0.26	0.15	0.29	− 0.72
WTHI	0.28	0.14	0.08	0.07	0.09	0.11
WWRI	− 0.19	0.10	0.19	0.06	0.07	− 0.12
WLCI	− 0.25	0.13	0.27	0.21	0.14	− 0.05
Std. Dev. of Weights	0.0367	0.0321	0.0948	0.1253	0.1460	0.2727

As shown in Table 5, studies focusing on smaller areas—such as the Fenhe River Basin²⁹, the Ganjiang River Basin⁵⁰, and the rural central Hunan region in this study—exhibit relatively high standard deviations in factor weights. For example, in rural central Hunan, the study area is only 34,100 km², the smallest among the listed cases. The factor weights are: RDLS = − 0.72, THI = 0.11, WRI = − 0.12, and LCI = − 0.05, resulting in a standard deviation of 0.2727, the highest among all studies. Among these factors, RDLS exhibits the largest absolute weight value, indicating its substantially greater influence relative to the other constituents and confirming its role as the dominant factor.

By contrast, studies conducted at broader spatial scales—such as the Central China region²² or entire provinces like Shaanxi²¹ and Henan⁵¹—show more balanced distributions of factor weights. The corresponding standard deviations are much lower, such as 0.0367 for Central China (RDLS = − 0.28, THI = 0.28, WRI = − 0.19, LCI = − 0.25) and 0.0948 for Henan (RDLS = 0.26, THI = 0.08, WRI = 0.19, LCI = 0.27).

In summary, the smaller the spatial scale of the study area, the more pronounced the differences in weight allocation among constituent factors. Certain factors are more likely to dominate, with weights significantly higher than others. In the case of rural central Hunan, for instance, terrain relief (RDLS) varies substantially across local areas, whereas temperature-humidity, surface water, and vegetation show relatively homogeneous spatial distributions. This leads to a much higher weight for RDLS compared to the other factors. This phenomenon highlights the importance of considering spatial scale effects when conducting natural suitability evaluations to ensure the scientific validity of the assessment.

While this study compares weight distributions across different regions to reveal scale effects, it does not include a multi-scale analysis within the same region due to the substantial data reconstruction required. Future research should conduct hierarchical multi-scale evaluations to more comprehensively examine how spatial resolution influences factor weight allocation.

Analysis of HEI in central Hunan

After calculating the HEI for the central Hunan region using Method 8, as developed in this study, the index values were standardized to a range of 0 to 100. Overall, the spatial distribution of suitability shows higher values in the central and eastern parts of the region and lower values in the west. These spatial patterns largely reflect the underlying environmental controls of the region. In particular, the strong variation in terrain relief, the distribution of river basins, and the east–west gradients in temperature and humidity jointly determine the observed suitability differences across central Hunan.

At the county level, Shuangfeng County (Loudi City) and Shaodong City (Shaoyang City) exhibit the highest HEI values, suggesting that their geographic and ecological conditions are most conducive to human habitation. In contrast, the northern part of Anhua County, the western areas of Xinhua, Longhui, and Dongkou Counties, and the southern Chengbu Miao Autonomous County show the lowest suitability levels. These areas are characterized by complex terrain and relatively low agricultural potential.

To better capture the differences in human settlement suitability across the region, the HEI values were classified into five levels using the Jenks Natural Breaks method, following the approach of Feng Zhiming et al.²². The five suitability levels, from lowest to highest, are defined as follows:

Unsuitable Zone (HEI 0–44.31, hereafter U.Z.);

Marginally Suitable Zone (HEI 44.32–58.43, M.S.Z.);

Generally Suitable Zone (HEI 58.44–71.37, G.S.Z.);

Moderately Suitable Zone (HEI 71.38–83.53, M.D.S.Z.);

Highly Suitable Zone (HEI 83.54–100, H.S.Z.).

The classification results are presented in Table 6.

Table 6.

HEI ranges and factor characteristics for each suitability zone.

Category		U.Z.	M.S.Z.	G.S.Z.	M.D.S.Z.	H.S.Z.
HEI	Range	0.00-44.31	44.32–58.43	58.44–71.37	71.38–83.53	83.54–100
	Mean	36.71	52.09	64.99	77.77	89.27
RDLS	Range	39–667	29–674	17–602	21–584	16–481
	Mean	306	264	214	146	86
THI	Range	46.74–58.73	47.18–58.94	48.18–58.93	48.27–58.98	53.42–59.01
	Mean	53.48	54.48	55.52	56.65	57.13
WRI	Range	0.04–0.90	0.02–0.95	0.02–0.93	0.01–0.90	0.01–0.77
	Mean	0.41	0.40	0.37	0.30	0.27
LCI	Range	0.13–0.77	0.16–0.77	0.13–0.74	0.04–0.75	0.06–0.73
	Mean	0.40	0.39	0.38	0.37	0.35

The map was generated using ArcGIS 10.8 (Esri Inc., Redlands, CA, USA; https://www.esri.com).

According to the classification results, most areas in central Hunan fall within the Generally Suitable Zone or higher. The suitability exhibits considerable spatial variation across the region (Fig. 6a). Summary statistics for each zone are provided in Table 7.

Fig. 6.

Suitable zone ranges (based on weighting Method 8). (a) Summary of suitable zone ranges; (b) Unsuitable zone; (c) Marginally suitable zone; (d) Generally suitable zone; (e) Moderately suitable zone; (f) Highly suitable zone.

Table 7.

Basic statistics for each natural suitability zone in central Hunan.

Category		U.Z.	M.S.Z.	G.S.Z.	M.D.S.Z.	H.S.Z.
Land	Area (104 km2)	0.24	0.55	0.74	0.82	1.06
	Proportion (%)	7.13%	16.15%	21.62%	23.99%	31.12%
Pop.	Total (104 ppl)	4.19	15.86	46.05	123.90	344.56
	Proportion (%)	0.78%	2.97%	8.61%	23.18%	64.46%
	Density (ppl/km2)	17.24	28.80	62.48	151.45	324.73
Temp.	Range (°C)	8.2–17.2	8.8–17.2	9.4–17.2	9.5–17.3	13-17.3
	Mean (°C)	13.3	14.0	14.8	15.6	16.0
RH	Range (%)	71–83	68–83	67–82	64–81	64–80
	Mean (%)	78.07	77.57	76.53	74.54	72.55
Prec.	Range (mm)	1214–1647	1214–1649	1212–1649	1212–1649	1212–1649
	Mean (mm)	1431	1425	1403	1365	1344
Elev.	Range (m)	35-1916	24-1929	23-1912	30-1896	42–910
	Mean (m)	828	694	567	405	282
GDP	PC (104 CNY)	2.76	2.83	2.97	3.10	3.54

Unsuitable zone

The Unsuitable Zone is primarily distributed along the northern and southern fringes, as well as in the western mountainous valleys of central Hunan. It encompasses Anhua County, Xinning County, Suining County, Dongkou County, Xinshao County, and Chengbu Miao Autonomous County (Fig. 6b), accounting for 8.12% of the region’s total area. This zone is characterized by high mountains, deep valleys, steep slopes, and karst depressions. The terrain is highly rugged, with large elevation differences and severely fragmented topography in many areas, which greatly limits agricultural development and infrastructure construction.

Due to long-term tectonic activity and erosion, the region features widespread karst landforms, rockfalls, and landslides. Some areas include sheer cliffs and deeply incised gorges. These geomorphic conditions result in extremely low land utilization and very poor human settlement suitability.

Hydrologically, the terrain also contributes to complexity and instability. Despite abundant precipitation, steep slopes and low groundwater retention lead to rapid surface runoff and poor soil water storage. In the rainy season, intense hillside runoff and severe river erosion frequently cause flash floods, debris flows, and landslides, posing risks to agriculture and human safety. In the dry season, insufficient water retention results in seasonal droughts in some areas, further compounding agricultural uncertainty.

Given these constraints, the zone is essentially unsuitable for large-scale agriculture. It is best suited for forest ecosystem conservation and limited forestry use, serving mainly ecological functions such as water source protection and soil retention. Some areas may support wildlife conservation and eco-tourism, but development intensity must be strictly regulated to ensure ecological stability and sustainability.

This zone contains only 0.78% of the total population of central Hunan, with a population density of 17.24 persons/km², far below that of other zones. Settlements are highly dispersed, with some villages limited to river valleys or low-slope foothill areas. Living conditions are difficult, and the limited agricultural productivity cannot support large populations.

Given the extremely low levels of agricultural productivity and settlement carrying capacity, this zone should prioritize ecological restoration, forest conservation, and the development of a carbon sink economy as its primary directions for future development (Table 8).

Table 8.

Development strategies for each natural suitability zone in central Hunan.

Zone	Development strategy
U.Z.	Prioritize ecological restoration, forest conservation, and the development of a carbon sink economy, given the zone’s low agricultural productivity and limited carrying capacity
M.S.Z.	Emphasize the sustainable use of forest resources and integrate ecological agriculture with biodiversity conservation to promote a sustainable development model tailored to the region’s characteristics
G.S.Z	Promote characteristic high-efficiency agriculture, strengthen soil and water conservation and ecological restoration projects, and improve living conditions to enhance overall human settlement suitability
M.D.S.Z.	Further optimize land use, develop high-efficiency cash crops and integrated farming systems, and implement soil and water conservation measures to improve sustainability. Strengthen small-scale water conservancy infrastructure to increase agricultural stability and disaster resilience
H.S.Z.	Focus on optimizing the agricultural structure, advancing modern agricultural technologies, improving infrastructure, and enhancing environmental governance to further improve the living environment and promote sustainable regional economic development

Marginally suitable zone

The Marginally Suitable Zone is primarily located along mountainous areas in central Hunan characterized by relatively steep topographic slopes. It spans portions of Anhua County, Suining County, and Chengbu Miao Autonomous County (Fig. 6c), covering 16.15% of the total area. The zone is dominated by steep mountainous terrain and mid- to high-elevation landforms, with complex and highly undulating topography. Long-term crustal activity and erosion have produced fragmented landforms such as deeply incised canyons, cliffs, and steep slopes, which significantly constrain the development and utilization of land resources.

Due to its rugged terrain, the zone exhibits unstable hydrological conditions. Surface runoff is rapid, and groundwater recharge is limited, resulting in poor soil moisture retention. During the wet season, concentrated precipitation leads to enhanced runoff, accelerated channel erosion, and pronounced soil loss. In the dry season, insufficient water storage impedes agricultural irrigation and vegetation growth. Although annual precipitation is relatively abundant, its effective use is limited; in many areas, rainwater quickly drains downslope, leading to alternating wet and dry cycles that increase agricultural uncertainty. The steep terrain also heightens the risk of landslides and debris flows, further weakening environmental stability and reducing arable land availability.

Given these constraints, large-scale agricultural development is largely infeasible. Only forestry and ecological agriculture are appropriate. In some areas, converting farmland to forest and closing hillsides for afforestation can help improve land use efficiency while reducing soil erosion and ecological degradation.

The zone has a low population density of 28.80 persons/km², accounting for only 2.97% of the region’s total population, underscoring its limited habitability. Due to steep slopes, infertile soils, and fragile ecological conditions, settlements are small in scale and highly dispersed, often confined to valleys or riverbanks. Infrastructure development is difficult, and the provision of public services is limited. With low agricultural productivity, out-migration of labor is common, and population aging is relatively pronounced.

Despite these limitations, the Marginally Suitable Zone holds important value for ecological protection and sustainable development. Future efforts should focus on enhancing sustainable forestry practices, promoting eco-friendly agriculture, and conserving biodiversity, thereby fostering a development model tailored to the region’s ecological characteristics (Table 8).

Generally suitable zone

The Generally Suitable Zone is located in the transitional belt between mountainous and hilly terrain in the western and southern parts of central Hunan, along the fringes of the Xuefeng and Nanling mountain ranges. It covers 21.62% of the region and is primarily distributed across Anhua County, Xinhua County, Suining County, and Chengbu Miao Autonomous County, with scattered areas in central Xinshao County, Longhui County, and parts of eastern Shuangfeng County and Shaodong City (Fig. 6d). The terrain is diverse, consisting of low mountains, hills, and foothill slopes with notable relief and some steep sections.

The mountainous landscape results in uneven distribution of water resources. High subsurface permeability and rapid surface runoff in certain areas limit water retention, leading to seasonal shortages that affect both agricultural irrigation and domestic water supply. Nevertheless, the zone receives relatively abundant precipitation, and small stream networks often develop along foothills. With appropriate management, these water resources can be effectively utilized.

Steep slopes hinder large-scale mechanized farming, restricting the development of broad-acre crops. However, the zone offers significant potential for forestry, specialty cash crops, and ecological agriculture. Due to terrain constraints, levels of agricultural mechanization remain low, labor demands are high, and production costs are relatively elevated.

This zone is home to 8.61% of the region’s population—considerably less than that of the moderately and highly suitable zones—indicating limited residential appeal. Rugged topography and prominent slopes pose challenges for infrastructure construction and constrain the development of large rural settlements. Although suitable for forestry and specialty agriculture, high labor intensity and modest economic returns limit its long-term capacity to support large populations. Additionally, complex terrain hampers transportation and the construction of agricultural infrastructure. Residents in some areas may face barriers to accessing education, healthcare, and commercial services, further reducing settlement desirability.

Overall, the Generally Suitable Zone should prioritize the development of high-efficiency specialty agriculture, enhancement of soil and water conservation, and implementation of ecological restoration projects to improve agricultural sustainability. Improving living conditions and access to public services is also essential to enhance the overall habitability of this zone (Table 8).

Moderately suitable zone

The Moderately Suitable Zone is widely distributed across the central transitional hilly areas of central Hunan. It is primarily concentrated in Lengshuijiang City, Lianyuan City, Xinhua County, Longhui County, and Xinning County (Fig. 6e), accounting for 23.99% of the total area. The zone is characterized by low hills and gentle slopes with moderate terrain relief. Between hills, there are small alluvial valleys and gently sloping farmlands, some of which still possess potential for agricultural development. However, undulating topography affects overall land availability, and steeper sections are unsuitable for large-scale mechanized farming. As a result, agriculture in this zone primarily relies on traditional intensive cultivation.

The area features well-developed small watershed networks and abundant surface runoff, offering generally favorable conditions for irrigation. However, in certain locations, heavy rainfall combined with significant elevation differences may lead to drainage challenges, localized waterlogging, and increased risks of flash floods and slope erosion. The climate is warm and humid, with adequate rainfall and sufficient sunshine, providing favorable conditions for cultivating a variety of cash crops. Nevertheless, in some areas, high soil permeability and limited water retention on slopes necessitate artificial land improvements—such as terracing or irrigation measures—to optimize agricultural production. The zone is particularly suitable for forestry, fruit orchards, tea cultivation, and other specialty crops. Gentler slopes can be converted into terraced fields for staple crops such as rice, corn, and tubers. However, low levels of mechanization and high labor intensity result in relatively high production costs.

This zone accommodates 23.18% of the region’s total population, with a population density of 151.45 persons/km², second only to the Highly Suitable Zone. The economy remains predominantly agricultural, and settlements are mostly rural villages located in hilly valleys and near water sources, typically arranged in a dispersed pattern. Although certain areas are rich in agricultural resources, terrain constraints and underdeveloped transportation and agricultural infrastructure hinder the zone’s overall development potential.

To further optimize land use in the Moderately Suitable Zone, emphasis should be placed on developing high-efficiency cash crops and integrated (three-dimensional) farming systems. These should be supported by soil and water conservation practices to promote agricultural sustainability. In addition, strengthening small-scale water conservancy infrastructure is essential for improving agricultural stability and disaster resilience⁵² (Table 8).

Highly suitable zone

The Highly Suitable Zone is located in the central and eastern basin areas of central Hunan. It encompasses most of Shuangfeng County, Shaodong City, Shaoyang County, Dongkou County, and Wugang City (Fig. 6f), covering 31.12% of the total area. The zone features relatively flat terrain and the lowest Relief Degree of Land Surface (RDLS) values in the region. It is primarily composed of valley basins, river plains, and alluvial plains with fertile soils and stable landforms. Small river plains influenced by the Zi River watershed are distributed between hills, providing reliable water sources that enhance the sustainability of both agriculture and human settlements.

Climatically, this zone lies within a subtropical humid monsoon region and exhibits the highest temperature-humidity index in central Hunan. It benefits from warm temperatures, moderate humidity, and abundant rainfall—an ideal combination for cultivating a wide range of crops. The dominant soil types are red earth, yellow earth, and alluvial soils, which retain moisture well, are well-aerated, and rich in organic matter, creating a stable and productive soil environment. The zone is well-suited for growing staple crops such as rice, wheat, and rapeseed, and also supports fruit orchards, vegetable production, and facility agriculture (e.g., greenhouses). In some areas, efficient mixed farming systems—such as grain-economic crop rotations—have been developed, forming highly productive agricultural models. Well-developed irrigation infrastructure further enhances agricultural adaptability, making this zone a vital base for grain and cash crop production in central Hunan.

Approximately 66.77% of the region’s population resides in this zone, far exceeding that of any other zone. With a population density of 324.73 persons/km², this area demonstrates a strong capacity to support human settlements due to its flat terrain, stable agricultural productivity, and well-developed infrastructure. Efficient transportation systems and a relatively mature economic structure have contributed to rapid economic growth, attracting population concentration and supporting well-established public services. As a result, the Highly Suitable Zone is one of the most economically dynamic regions in central Hunan.

Thanks to its favorable topography, reliable water resources, agricultural suitability, and high population carrying capacity, the Highly Suitable Zone holds the greatest development potential in central Hunan. However, rapid development also introduces challenges related to land use efficiency, environmental protection, and sustainable management. Future efforts in this zone should focus on optimizing agricultural structure, promoting modern agricultural technologies, improving infrastructure, and enhancing environmental governance to further improve the quality of the human settlement environment and ensure sustainable economic development (Table 8).

Recent studies on ecological quality assessment also emphasize that spatial patterns of suitability are shaped by multiple natural drivers—such as topography, climate, and vegetation—providing complementary evidence for the mechanisms identified in this study⁵³.

Conclusion

Optimization of the weighting method for natural suitability factors

This study adopts an objective, quantitative approach to determine the weights of natural suitability factors for human settlements, thereby avoiding the subjectivity associated with expert scoring and enhancing the scientific validity and comparability of the weighting process. By integrating entropy weighting, correlation analysis, and the coefficient of variation, we optimized the traditional correlation-based weighting method.

The incorporation of entropy weighting enables a more balanced distribution of information among factors, effectively highlighting those with greater influence on suitability outcomes. Meanwhile, the coefficient of variation adjusts for data variability, ensuring that factors with high dispersion are not underestimated. Together, these enhancements improve the overall stability and accuracy of the suitability assessment.

Through quantitative evaluation and comparison of multiple methods, the superiority of the proposed approach was verified. Unlike previous studies that relied on a single weighting technique, the method developed in this study ensures that weight allocation remains both objective and aligned with the intrinsic characteristics of the data.

Impact of Spatial scale on factor weight allocation

The results of this study reveal a clear scale effect in the distribution of factor weights. At smaller spatial scales, differences in weights among factors are more pronounced, with certain variables emerging as dominant. In contrast, at national or provincial scales, factor weights tend to be more evenly distributed. These findings highlight the importance of carefully considering spatial scale when conducting natural suitability evaluations to ensure the scientific rigor and accuracy of the assessment.

Characteristics and development strategies of rural human settlement environment suitability in central Hunan

The rural human settlement environment suitability in central Hunan exhibits pronounced spatial disparities in both suitability values and factor weight distributions. Spatially, overall suitability follows a pattern of being higher in the east and lower in the west.

In terms of factor contributions, RDLS is the most influential variable, significantly outweighing the contributions of other natural environmental factors. In contrast, THI, WRI, and LCI exhibit relatively uniform spatial distributions and exert weaker influences, insufficient to substantially alter the overall suitability pattern.

Based on the characteristics of natural resources, environmental carrying capacity, and agricultural development potential across different areas, targeted development strategies were formulated for each suitability zone. These strategies aim to achieve the coordinated goals of rational land development and ecological protection, thereby promoting sustainable rural development throughout central Hunan. It is important to clarify that these recommendations apply primarily to rural land and ecological development contexts shaped by natural environmental constraints and should not be extended to urban construction or socio-economically driven development scenarios.

Moreover, although this study is focused on central Hunan, the proposed weighting method—integrating PCC, normalized information entropy, and the coefficient of variation—is adaptable to other rural regions with diverse natural conditions. Its data-driven nature and modular design make it suitable for extension to areas with similar human–environment interactions, especially in mountainous or topographically heterogeneous landscapes. Future studies can further explore its applicability across broader geographic contexts.

Future research prospects

Several directions can be pursued in future research to deepen and extend the findings of this study:

Multi-scale factor weighting: Further explore how factor weight distributions vary across different spatial scales, particularly between small-scale and large-scale studies, in order to improve the precision and adaptability of suitability evaluations.

Influence of settlement typologies: Investigate how different rural settlement patterns affect the weighting of natural suitability factors, with the goal of identifying optimal development pathways for diverse settlement forms.

Impacts of climate change and extreme weather: Examine the influence of climate change and high-impact weather events on rural human settlement environments, and assess the risks they pose to agricultural production, ecological stability, and living conditions.

Intelligent modeling and cross-regional validation: Incorporate big data and artificial intelligence technologies to enhance the intelligence, automation, and responsiveness of suitability evaluation models. Conduct cross-regional comparative studies to assess the generalizability and practical feasibility of the proposed methods, providing more precise guidance for rural planning and sustainable development.

Multidimensional and dynamic model enhancement: Integrate socio-economic indicators, climate model projections, and long-term monitoring mechanisms to provide more comprehensive and dynamic validation of natural suitability. Extend the framework to other regions and apply spatial–temporal cross-validation to improve methodological robustness, generalizability, and reliability under varying geographical and climatic conditions.

Refinement and sustainability assessment of development strategies: Further refine development strategies in accordance with local socio-economic contexts and specific regional characteristics. Incorporate sustainability assessment tools—such as life-cycle analysis—to evaluate the environmental, economic, and social impacts of these strategies and enhance their practical applicability in rural planning.

Author contributions

Conceptualization, L.X.; methodology, L.X.; software, L.X.; validation, J.X. and X.L.; formal analysis, L.X.; investigation, L.X.; resources, X.L. and L.Z.; data curation, Y.L. and S.C.; writing—original draft preparation, L.X.; writing—review and editing, J.X.; visualization, Y.L. and S.C.; supervision, X.L.; project administration, L.X.; funding acquisition, L.X. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Hunan Provincial Natural Science Foundation of China (No. 2026JJ90040), the Scientific Research Project of Hunan Provincial Department of Education of China (No. 24C0383), and the Industry-Education Integration Project of the Ministry of Education of China (No. 230801960192831).

Data availability

The original contributions presented in this study are included in thearticle. Further inquiries can be directed to the corresponding authors.

Declarations

Competing interests

The authors declare no competing interests.