Evaluating the suitability of underground space development based on social and economic factors

ABSTRACT

Ma’anshan City, located in the Yangtze River Delta region, is a demonstration city for handling the industrial transfer of the Nanjing, Hefei, and Wanjiang Metropolitan Areas in China. Developing and utilizing underground space can effectively alleviate the land scarcity issue in urban development. The suitability evaluation of underground space development in Ma’anshan City offers valuable information support for the economic and efficient use of land resources. Drawing on POI, night light radiance, and population data, this study selects six indices, including POI kernel density, road node connection rate, night light radiance in the most recent year, change amplitude of night light radiance over the years, population density in the most recent year, and change amplitude of population density over the years. The study conducts a comparative analysis of the suitability of underground space development in Ma’anshan City using the Analytic Hierarchy Process (AHP) and the Combination Weight Method. Comparing the results of AHP and Combination Weight Method reveals that the latter can identify a more extensive range of highly suitable development zones than the former. This study provides an effective method for evaluating the suitability of underground space development in urban planning, which contributes to promoting sustainable urban development.

KEYWORDS:

• Underground space development
• suitability assessment
• analytic hierarchy process (AHP)
• combination weight method
• sustainable development

1 Introduction

Sustainable development refers to a mode of development that not only fulfills the needs of present generations but also preserves the capacity of future generations to meet their own needs (Cobbinah & Mensah Darkwah, 2016; Yan et al., 2018). Confronting severe challenges such as population explosion, resource depletion, and environmental degradation, sustainable development has become a consensus embraced by humanity (Yu Duo & Jiao, 2011). As a valuable spatial resource, underground space represents one of the most effective approaches to expanding urban areas and achieving sustainable urban development. The continuous growth of the population necessitates cities to provide increased space to accommodate people’s living, working, recreational, and transportation needs (Egidi et al., 2020). Concurrently, ecological protection requirements are causing urban expandable construction land to become increasingly scarce, leading cities to face progressively severe ‘urban diseases’ such as traffic congestion, environmental pollution, and frequent disasters (Chaosu et al., 2023; Li et al., 2021). In this context, the scientific and rational evaluation of urban underground space resources holds significant importance for urban planning and development (Zhu et al., 2016).

Currently, numerous scholars have conducted extensive research on the suitability evaluation of urban underground space development. For instance, Tan et al. (2021) proposed an evaluation method that combines the limited interval cloud model with genetic algorithms. They used Wuhan Yangtze New City as the evaluation area and established an evaluation system comprising 12 indices, including five geological factors. Liu et al. (2013) employed the analytic hierarchy process to determine the weights of 15 factors affecting underground space development, such as topography, stratum structure, stratum stability, and groundwater depth. They then applied grey mathematical theory to calculate the correlation between the evaluation index and the appropriate level, ultimately obtaining a suitability zoning map. Duan et al. (2021) argued that geology is fundamental to underground space development, analyzed Kunming’s geological conditions using the analytic hierarchy process, and proposed an evaluation system tailored for Kunming. From the perspective of research content and index selection, current suitability evaluations predominantly focus on geological conditions, while comprehensive research on socio-economic factors is relatively scarce. In reality, geological conditions, existing facilities, and various protection requirements determine the difficulty of underground space development, while socio-economic factors dictate the potential value of development. These factors collectively influence the potential level of underground space development and utilization. Therefore, assessing the suitability of urban underground space development based on social and economic factors holds significant research importance.

Humans and the Earth are interconnected and mutually influential. Socioeconomic factors refer to the economy and culture shaped by human activities, as well as various social phenomena associated with them, such as residential areas, transportation networks, administrative boundaries, population, history, and culture (Wilkerson et al., 2018). The status of these factors can profoundly reflect the developmental level and social civilization of the study area. In this paper, the selected factors include Points of Interest (POI), night light radiance, and population. POI data comprises the name, category, longitude, latitude, address, and other basic attributes of geographical entities. This data can describe the relevant characteristics of urban functional zoning and activity space distribution, reflecting people’s travel needs and urban land use status to some extent, and providing a reference for underground space development (Zhang et al., 2018). For instance, Chen-Xiao and Peng (2023) proposed a framework for evaluating urban underground space development levels based on POI data. They established evaluation methods for urban underground space performance and development demand and applied them to calculate the urban underground space supply-demand ratio for 217 streets in Shanghai. Night light remote sensing imagery has a high correlation with human activities, and also possesses the advantages of time-space continuity, independence, and objectivity. It can serve as a valuable socioeconomic proxy variable (Rehman et al., 2021). Su et al. (Preprint) studied the relationship between average nighttime light radiance and GDP in Shanghai using long time-series multi-source image data, finding a significant positive correlation between the two. High night light radiance indicates vigorous human social activities and high value for urban underground space development. As remote sensing technology advances, spatially continuous data sources such as remote sensing images are widely used in regional development research. Population distribution is a mapping generated by people in urban geographical locations to meet their activity needs. High population density drives the development of urban underground space (Wang & Liu, 2019). In studying the historical evolution of urban underground space, Bobylev (2009) revealed a positive correlation between population density and the total amount of urban underground space by analyzing the data relationship between population density and urban underground infrastructure capacity in Stockholm, Tokyo, and Paris. In summary, POI, night light radiance, and population are all crucial socioeconomic factors affecting the suitability evaluation of urban underground space development. Regarding multi-factor overlay, this paper adopts the combination weight method because the Analytic Hierarchy Process (AHP) is more subjective when determining weights, while the subjective and objective comprehensive weight method can reduce the subjectivity of weight setting, thereby enhancing the scientific validity of the evaluation.

2 Study area and data sources

2.1 Overview of the study area

Ma’anshan, a city situated in the Yangtze River Delta, serves as a demonstration city responsible for the industrial transfer of the Nanjing Metropolitan Circle, Hefei Metropolitan Circle, and Wanjiang Metropolitan Circle in China. The city encompasses a total area of 4,049 square kilometers, and administers three districts and three counties (Figure 1). According to China’s seventh population census, as of 1 November 2020, Ma’anshan City has a permanent population of 215,993. When compared to first-tier cities in China such as Beijing and Shanghai, the development of underground space in Ma’anshan City remains underdeveloped. The rational development and utilization of underground space can transform the city’s planar extension development into three-dimensional development, relocating some urban functions from the surface to the underground. This process provides more space and area for urban development and effectively addresses the issue of land scarcity in urban development. Currently, Ma’anshan has developed and constructed several underground space projects, including Wanda Plaza, Red Star Macalline, and Eight Hundred Companions. As the sole demonstration city for land resource conservation and intensive use in Anhui Province, Ma’anshan needs to innovate in terms of land resource conservation and intensive use. The development of this work is closely related to the assessment of social and economic factors. By taking Ma’anshan City as the research subject, this paper evaluates the suitability of urban underground space development from the perspective of social and economic factors using GIS software. The goal is to provide valuable insights for the future planning of urban underground space in Ma’anshan and other similar cities.

Figure 1. Location map of Ma’anshan City.

2.2 Data sources

The primary research data for this study consist of POI, night light radiance, and population data for Ma’anshan City.

For POI data, this paper utilizes the Gaode Map APP as its data source. Firstly, Python is employed to write a data crawling code to obtain POI within the research area. Subsequently, the basic POI data imported into ArcGIS undergoes initial processing via coordinate and spatial correction, eliminating useless POI information points outside the research area while retaining the effective POI information points. Lastly, the acquired POI information points are examined, and their address information is corrected and completed. Additionally, this paper also downloads Ma’anshan’s road vector data from OpenStreetMap and extracts road intersections as supplementary POI.

For night light data, this paper utilizes the VIIRS monthly composite night light remote sensing images provided by the Colorado School of Mines (https://eogdata.mines.edu/download_dnb_composites.html). Compared to the previous generation DMSP/OLS images, VIIRS images offer higher spatial resolution and a broader radiation range. Additionally, they more accurately reflect the spatial information of human social and economic activities on the surface after onboard radiation calibration. VIIRS monthly composite data products include ‘VCM’ and ‘VCMSL’ types. This paper selects the ‘VCM’ type of VIIRS monthly composite data for research.

For population data, this paper utilizes data from WorldPop (www.worldpop.org), a website that provides high-resolution geospatial data from 2000 to 2020 with a focus on low-income countries. The data is used to construct population distribution and demographic data, with a population raster resolution of 100 meters selected for this study.

3 Evaluation method and index system

3.1 Evaluation model

The evaluation of the development suitability of underground space in Ma’anshan is accomplished through the following steps: (1) establishing an index system to evaluate development suitability, including POI, night light radiance, and population; (2) using the ‘AHP method’ and ‘AHP_Entropy weight method’ to determine the weight of each index; (3) standardizing each index; (4) defining the spatial distribution and change characteristics of POI, night light radiance, and population; (5) weighted superposition of single factor evaluation results; (6) visualizing the distribution of development suitability for Ma’anshan’s underground space.

In the selection of indicators for the suitability assessment of underground space development, this study fully considers the availability of indicator data and takes into account three aspects: Points of Interest (POI), nighttime light radiation, and population. Specifically, the POI factor includes the density of POI distribution and the current transportation convenience index. This means that the study considers the distribution of points of interest and the influence of transportation convenience on underground space development. The density of POI distribution can reflect the concentration of urban vitality and demand, while the transportation convenience index can measure the accessibility and interconnectedness of underground space.

The nighttime light radiation factor includes the nighttime light radiation value for the most recent year (2021) and the nighttime light radiation growth rate from 2013 to 2021 (nine years). Nighttime light radiation is an indicator of urban development and activity levels, and it has important reference value for the development potential of underground space. The nighttime light radiation value and growth rate can reveal the level of urban prosperity and future development trends.

The population factor includes the population density for the most recent year (2020) and the population growth rate from 2013 to 2020 (eight years). Population is one of the core factors of urban development and has a significant impact on the demand and feasibility of underground space utilization. The population density and growth rate can reflect the population size, growth trends, and potential demand for underground space in the city.

These factors are key aspects that influence the feasibility and effectiveness of underground space development. By considering them comprehensively, the study establishes a comprehensive evaluation framework aimed at providing valuable guidance for assessing the suitability of underground space in the specific context of Ma’anshan City (Figure 2).

Figure 2. Evaluation index system for assessing the suitability of underground space.

development based on social and economic factors in Ma’anshan City

The suitability of underground space development in this paper is a comprehensive function of POI suitability, night light radiance suitability, and population suitability. The suitability index of underground space development can be calculated using Formula (4).

POI suitability index:

(1) $A\left(x\right)={\sum }_{j=1}^i\left[w_j\times A_{ji}\left(x\right)\right]$(1)

Night light radiance suitability index:

(2) $B\left(x\right)={\sum }_{j=1}^i\left[w_j\times B_{ji}\left(x\right)\right]$(2)

Population suitability index:

(3) $C\left(x\right)={\sum }_{j=1}^i\left[w_j\times C_{ji}\left(x\right)\right]$(3)

Suitability index:

(4) $S\left(x\right)=w_{A}A\left(x\right)+w_{B}B\left(x\right)+w_{C}C\left(x\right)$(4)

Where: A_ji(x), B_ji(x) and C_ji(x) represent the standardized values of each index; A(x), B(x), C(x) and S(x) represent the POI suitability, night light radiance suitability, population suitability, and underground space development suitability indices in the suitability assessment model; w_A, w_B, w_C and w_j are the weights assigned to each evaluation index.

3.2 Processing of index data

3.2.1 POI suitability index

The kernel density analysis method estimates the spatial distribution of POI data concentration by using a regular moving quadrat based on the distance attenuation law. As shown in Figure 3, several areas of high POI density can be identified, including Hanshan County, He County, Dangtu County, and the junction of Huashan District and Yushan District. A high density of POI distribution indicates a relatively complete urban infrastructure in that location, while a low density implies the opposite. The POI in Figure 3 does not include road intersections.

Figure 3. Kernel density of points of interest (POI).

The connectivity of roads can serve as an index of urban design. Compared to road networks with many dead ends and long sections, a well-connected network can link various urban infrastructures and facilitate the connection between ground and underground spaces. Road intersections are significant POIs. To evaluate the degree of traffic convenience, this study calculates the road node connection rate using the formula:

(5) $\mathrm{N}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{n}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}=\frac{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}}{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}+\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}\mathrm{e}\mathrm{n}\mathrm{d}\mathrm{s}}$(5)

The method involves dividing Ma’anshan City into 1000 × 1000 meter grids, calculating the road node connection rate for each grid, and assigning the value to the centroid of each grid. The distribution of the road node connection rate for the entire city of Ma’anshan is obtained using Kriging interpolation, as illustrated in Figure 4.

Figure 4. Distribution of road node connection rate.

3.2.2 Night light radiance suitability index

The nighttime light radiance in Ma’anshan for 2021 is depicted in Figure 5. According to Figure 5, Huashan District and Yushan District exhibit the highest levels of night light radiance, while other areas display elevated radiance only in limited zones, with most regions exhibiting low radiance. A higher radiance of nighttime lighting correlates with more vibrant human activity. To facilitate the growth of underdeveloped areas within the city, it is worth considering the development of underground spaces in locations where nighttime lighting is less prominent.

Figure 5. Nighttime light radiance in Ma’anshan, 2021.

From 2013 to 2021, the nighttime light radiance varies each year. The traditional method for characterizing the change in light radiance amplitude is by examining the change rate, as demonstrated in Formula (6):

(6) $NLCR=\frac{SO{L}_{2021}-SO{L}_{2013}}{SO{L}_{2013}}$(6)

In this context, SOL₂₀₁₃ denotes the total amount of nighttime light for a region in 2013, while SOL₂₀₂₁ represents the total amount of nighttime light for the same region in 2021. When NLCR (Nighttime Light Change Rate) is greater than 0, it indicates an increase in the total nighttime light amount within the region from 2013 to 2021. Conversely, when NLCR is less than 0, it suggests a decrease in the total nighttime light amount from 2013 to 2021, potentially signifying an economic recession in the region.

However, the computation of the nighttime light change rate only takes into account the information from the starting and ending years of the research period, neglecting the data within the time frame. As a result, the Mann Kendall (MK) trend test method is employed to examine the changing trend of the total nighttime light amount in Ma’anshan City from 2013 to 2021. The MK trend test can effectively differentiate whether a specific natural process is experiencing natural fluctuations or exhibiting a discernible change trend. In the MK test, the statistic is calculated as follows:

(7) $\left\{\begin{array}{c}\begin{array}{c}Z=\left(S-1\right)/\sqrt{n\left(n-1\right)\left(2n+5\right)/18},S>0\\ Z=0,S=0\\ Z=\left(S+1\right)/\sqrt{n\left(n-1\right)\left(2n+5\right)/18},S>0\end{array}\\ \hspace{0.17em}\end{array}\right.$(7)

Where S is:

(8) $S=\sum _{i=2}^n\sum _{j=1}^{i-1}sign\left(X_i-X_j\right)$(8)

Where sign represents a symbolic function. When it comes to the confidence level, if the condition is met ($\left|Z\right|\ge Z_{1-\alpha /2}$), there is a significant trend in the total nighttime light amount time series. A positive value of Z signifies an upward trend, while a negative value indicates a downward trend.

Figure 6 illustrates the change trend of nighttime light radiance in Ma’anshan City from 2013 to 2021. The values of −2, −1, 0, 1, and 2 represent significant decrease, slight decrease, basically unchanged, slight increase, and significant increase, respectively. In Ma’anshan City, Huashan District and Yushan District have a large proportion of areas with significant increase, indicating that the economic development in these two areas is fast, and they have become important economic regions in Ma’anshan City. The regions with significant rise and slight rise in He County and Hanshan County appear in a striped distribution, implying that there is a certain degree of correlation in their development, forming multiple emerging economic centers. However, the aggregation of areas with significant increase in Bowang District and Dangtu County is not as obvious as in other regions.

Figure 6. Magnitude of change in night light radiance from 2013 to 2021.

(95% confidence interval)

3.2.3 Population suitability index

The population density in 2020 is depicted in Figure 7. The figure illustrates that the population is primarily concentrated in Huashan District and Yushan District, with a few high-value accumulation areas in other regions. For sparsely populated areas, the government may consider talent diversion as a means to promote local economic development.

Figure 7. Population density in 2020.

The population factor is also analyzed by the Mann Kendall (MK) trend test, which yields values of −2, −1, 0, 1, and 2 indicating significant decrease, slight decrease, no change, slight increase, and significant increase, respectively, as shown in Figure 8. The figure shows that only a few regions have experienced slight or no changes in population density, while most regions have had significant changes, suggesting strong population mobility in Ma’anshan from 2013 to 2020. The comparison between Figures 8 and 6 indicates that the areas with significant population density increase also exhibit an upward trend in nighttime lights.

Figure 8. Amplitude of change in population density from 2013 to 2020 (95% confidence).

3.2.4 Standardization of data

Due to the complexity and diversity of the evaluation indices, their meanings and dimensions are different, and there is no direct comparability between the indices. In order to comprehensively calculate the data of each index, it is necessary to conduct dimensionless standardization processing on the data and standardize the values of each index to between 0 and 1. The standardized values can reflect the impact of the indices on the suitability of underground space development. When planning underground space, on the one hand, the infrastructure of the underground space, such as subway stations, should be as close as possible to major commercial complexes and large residential areas (i.e. densely populated areas); on the other hand, some underground space infrastructure should also be set up in the periphery of the city to guide the future development of urban underground space. In order to comprehensively consider the development of the urban center and periphery, A1, B1, and C1 are defined as negative indices, while A2, B2, and C2 are defined as positive indices. Therefore, in areas with currently low POI distribution density, low night light radiance, and low population density, underground space development can be considered to stimulate the development of underdeveloped areas. However, we do not prioritize areas with currently high POI distribution density, high night light radiance, and high population density because the development in these areas is relatively saturated. Areas with convenient transportation and significant growth in night light radiance and population density over the years indicate that their development is relatively stable and attractive, thus making their underground space highly valuable for development. The standardized calculation formula is as follows:

(9) $Forpositiveindices,Y_{ij}=\left(X_i-X_{min}\right)/\left(X_{max}-X_{min}\right)$(9)

(10) $Fornegativeindices,Y_{ij}=\left(X_{max}-X_i\right)/\left(X_{max}-X_{min}\right)$(10)

3.3 Weight determination of indices

3.3.1 Analytic hierarchy process (AHP) method

The Analytic Hierarchy Process (AHP) is a decision-making method that involves multi-level weight analysis. It was first proposed by American operations research scientist A L Saaty in the 1970s and is a system analysis method that combines qualitative and quantitative analysis. AHP can be used to determine the subjective weight of each evaluation index (Sun et al., 2013). The steps for AHP to determine the weight of evaluation indices are as follows: firstly, establish a hierarchical structure; secondly, construct a pairwise judgment matrix and use the scale method of 1 ~ 9 and its reciprocal to quantify the relative importance of pairwise elements; thirdly, calculate the subjective weight w’i of the index; fourthly, check the consistency. When CR < 0.1, it indicates that the consistency of the judgment matrix is reasonable. When CR ≥ 0.1, it indicates that the judgment matrix is unreasonable, and the consistency check needs to be performed again.

In the Analytic Hierarchy Process (AHP) method, we invited 10 experts to assess the importance coefficients of the criteria to determine their relative weights. These experts were selected based on their professional knowledge and experience in urban planning and underground space development. To collect their opinions, we designed a structured questionnaire that included the various criteria. The experts were asked to compare the criteria pairwise based on their perceived importance, thus forming a judgment matrix. Then, we averaged the judgment matrices from the 10 experts to obtain a representative judgment matrix representing the collective expert opinion. According to the general consensus among the experts, they unanimously agreed that Points of Interest (POI) are more important than Night Light Radiance, while Population has the lowest importance. Subsequently, we analyzed the collected data using the AHP method to determine the AHP weights of the criteria.

3.3.2 Entropy weight method

The entropy weight method is an objective weighting method that has gained wide usage in recent years. The smaller the entropy value of an index, the greater the variability of its values, the more information it provides, and the greater its weight (Feng et al., 2012). If there are m indices and n pixels in the study area, the index matrix is represented by R=(r_ij)_m×n. The entropy of the i-th index is defined as:

(11) $H_i=-\frac{1}{lnn}\sum _{j=1}^n{f}_{ij}ln{f}_{ij}$(11)

Where: H_i is the entropy of the i-th index; n is the number of units; When f_ij = 0, define f_ijlnf_ij = 0; f_ij is defined as:

(12) $f_{ij}=-\frac{Z_{ij}}{\sum _{j=1}^n{Z}_{ij}}$(12)

Where, Z_ij refers to the standardized index value of the j-th evaluation object under the i-th index.

The entropy weight of the i-th index is defined as:

(13) $w_i^{"}=\frac{1-H_i}{\sum _{i=1}^m\left(1-H_i\right)}$(13)

Where: w″_i is the entropy weight of the i-th index, 0 ≤w″_i ≤1; H_i is the entropy of the i-th index;$\sum _{i=1}^m{H}_i=1$; m is the number of evaluation indices.

3.3.3 Combination weight method

The combination weight method utilizes the AHP method to obtain subjective weight and the entropy weight method to obtain objective weight. In this study, the linear combination method was applied to obtain the combination weight for evaluating the suitability of underground space development, and the results are presented in Table 1. By combining the subjective weight w_i′ and the objective weight w_i”, we can obtain the combined weight w_i. To eliminate the interference of large fluctuations in data, α and β are used to represent the difference between w_i′ and w_i”. Furthermore, the concept of distance function is introduced.

Table 1. Weight of indices for evaluating the suitability of underground space development.

Target layer	Index layer	Sub index layer	Weight
			AHP method	Entropy weight method	Combination weight method
Suitability Index for Underground Space Development	POI (0.57)	POI kernel density（-）	0.250	0.501	0.344
		Road node connection rate（+）	0.750	0.499	0.656
	Night light radiance (0.30)	Night light radiance in the most recent year (2021)（-）	0.333	0.528	0.411
		Change amplitude of night light radiance over the years（+）	0.667	0.472	0.589
	Population (0.13)	Population density in the most recent year (2020)（-）	0.500	0.728	0.588
		Change amplitude of population density over the years（+）	0.500	0.272	0.412

The formula for calculating the combination weight is as follows:

(14) $w_i=\alpha w_i^{\prime }+\beta w_i^{{ }^{\prime \prime }}$(14)

Where: α and β are the weight distribution coefficients, $\alpha +\beta =1$

The expression for the distance function between subjective weight and objective weight is:

(15) $d\left(w_i^{\prime },w_i^{{ }^{\prime \prime }}\right)=[\frac{1}{2}\sum _{i=1}^n{\left(w_i^{\prime }-w_i^{{ }^{\prime \prime }}\right)}^2{]}^{\frac{1}{2}}$(15)

The difference between α and β is the difference between distribution coefficients:

(16) $D=\left|\alpha -\beta \right|$(16)

According to the above, the equation set is constructed as follows:

(17) $\left\{\begin{array}{c}d{\left(w_i^{\prime }-w_i^{{ }^{\prime \prime }}\right)}^2={\left(\alpha -\beta \right)}^2\\ \alpha +\beta =1\end{array}\right.$(17)

The distribution coefficients α and β for each weight can be calculated by solving Equations (17)(17) $\left\{\begin{array}{c}d{\left(w_i^{\prime }-w_i^{{ }^{\prime \prime }}\right)}^2={\left(\alpha -\beta \right)}^2\\ \alpha +\beta =1\end{array}\right.$(17) . Once the distribution coefficients are obtained, the combined weight can be calculated using the formula and the results are shown in Table 1.

4 Result

The spatial distribution results and weights of each index were used as inputs to the suitability evaluation model for underground space development to obtain the suitability index. Using the natural break grading method, the spatial distribution of the suitability index obtained by the AHP method was divided into five grades (very low, slightly low, moderate, slightly high, very high) using GIS. Figure 9 shows the distribution of suitability levels for underground space development obtained by the ‘AHP method’ (Figure 9a) and the ‘Combination weight method’ (Figure 9b). Both methods used the same threshold.

Figure 9. Suitability levels for underground space development.

The suitability of underground space development in Ma’anshan City was assessed using the combination weight method. The results showed that 23.95% of the city’s total area was highly suitable for underground space development, with the most suitable areas located in Huashan District, Yushan District, and the middle and eastern parts of Bowang District, as well as the central and southern parts of Dangtu County and the eastern part of He County. In addition, 30.84% of the total area had slightly high suitability, mainly concentrated in the western Bowang District, northern Dangtu County, eastern He County, and some parts of Hanshan County. These areas had relatively developed ground traffic, increasing night light radiance, and population density, indicating their high potential for underground space development. Moderate and slightly low suitability areas accounted for 27.28% and 16.01% of the total area, respectively, while the low suitability area accounted for only 1.92%, mainly distributed in He County and Hanshan County (Figure 9). Among the six administrative districts, Huashan District had the highest suitability for underground space development, with 46.72% of the total area falling into the very high suitability category and 22.98% in the slightly high suitability category. Yushan District and Dangtu County ranked second and third with very high suitability areas accounting for 44.59% and 39%, respectively. On the other hand, Hanshan County had the lowest suitability, with only 8.77% of very high suitability areas (Table 2).

Table 2. Proportions of suitability levels for underground space development in each district.

Suitability level	Proportion of Yushan District (%)		Proportion of Huashan District (%)		Proportion of Bowang District (%)
	AHP method	Combination weight method	AHP method	Combination weight method	AHP method	Combination weight method
Very low	2.41	1.7	9.5	0.28	11.18	2.2
Slightly low	18.02	19.04	22.75	9.46	25.71	18.17
Moderate	19.31	20.44	14.46	20.55	27.38	26.88
Slightly high	25.86	14.24	10.75	22.98	24.3	34.78
Very high	34.4	44.59	42.54	46.72	11.43	17.97
	Proportion of Dangtu County (%)		Proportion of Hanshan County (%)		Proportion of He County (%)
Suitability level	AHP method	Combination weight method	AHP method	Combination weight method	AHP method	Combination weight method
Very low	3.95	0.87	11.21	2.76	10.64	2.6
Slightly low	14.94	9.12	33.3	20.71	25.29	17.32
Moderate	21.37	17.48	33.08	38.15	27.2	27.95
Slightly high	34.48	33.52	17.51	30.11	22.97	31.59
Very high	25.26	39	4.9	8.77	13.91	20.52

The spatial distribution of suitability levels for underground space development in Ma’anshan City differs noticeably between the ‘AHP method’ and the ‘Combination weight method’, as depicted in Figure 9. According to the distribution, the very high suitability area obtained by the ‘Combination weight method’ is relatively large (7.69%), with a similarly large slightly high suitability area (6.75%). In contrast, the slightly low suitability area is significantly smaller (8.43%), as is the very low suitability area (6.88%), and there is no significant difference between the moderate suitability areas (0.87%). While the distribution of very high suitability areas identified by the ‘Combination weight method’ is generally consistent with that of the AHP method, the former assesses a larger area of very high suitability, identifying more plots suitable for underground space development.

The suitability zoning map of underground space development obtained in this study highlights Huashan District and Yushan District as the areas most worthy of attention. Huashan District, which is adjacent to Nanjing in the east and north, is the main urban area of Ma’anshan City, while Yushan District, located in the west, is the seat of the Ma’anshan Municipal Government and borders the Yangtze River. These two districts have strategic importance due to their location. The study proves the suitability of underground space development based on data results and highlights the importance of social and economic benefits for the development of underground space. The economic benefits of developing high-value areas outweigh the development costs. Based on social and economic factors, the suitability of underground space development in the six administrative regions of Ma’anshan City, ranked from high to low, is: Huashan District, Yushan District, Dangtu County, He County, Bowang District, and Hanshan County.

5 Discussion

To further explore the spatial cluster characteristics of the suitability index for underground space development in Ma’anshan City, this study used the 1000 × 1000 m grid divided in Section 2.2.1 and calculated the average suitability index of each grid. The ‘Cluster and Outlier Analysis’ tool was then used to determine the cluster type of each grid (see Figure 10). In Ma’anshan City, the suitability index did not show significantly higher or lower values compared to the surrounding units. Grids with low-low cluster were mainly located at the edge of the study area, while those with high-high cluster were mainly located in the central area of Huashan District, Yushan District, Dangtu County, and the east of He County. High-value clustering and low-value clustering belonged to a local spatial positive correlation region. Overall, a large number of grids in Ma’anshan City belonged to the local spatial positive correlation area for the suitability index. If large-scale underground space development is to be carried out, the high clustering areas, especially the high clustering area connecting Huashan District, Yushan District, Dangtu County, and He County, should be prioritized. These areas exhibit significant high clustering in the suitability index, indicating a greater potential and feasibility. The results of the global spatial autocorrelation analysis of the suitability index are shown in Figure 11. The Moran’s I index was greater than 0, with a z-value of 74.295229. Through hypothesis testing at a significance level of 0.0%, it was found that the suitability index of underground space development in Ma’anshan City has a spatial positive correlation on the whole. This suggests that there is a certain degree of similarity and correlation in the suitability index among neighboring areas

Figure 10. Local cluster analysis of the suitability index.

Figure 11. Global cluster analysis based on the suitability index.

Although this study has made important findings, there are still some limitations that need to be addressed. Firstly, the study did not fully consider the impact of different depths on underground space development, which involves different social needs and cost factors (Peng et al., 2023). Therefore, a more in-depth analysis of these factors is necessary.

Secondly, the interrelationship between underground space development and socio-economic conditions has not been clearly identified. The study did not delve into the relationship between underground space development patterns and socio-economic conditions. Future research can explore different combinations of underground space development patterns and socio-economic conditions, taking into account both costs and benefits to achieve optimal outcomes.

Lastly, the selected indicators of socio-economic conditions in this study may not be comprehensive enough. For example, the impact of carbon neutrality policies on underground space development was not considered. Underground space development can support carbon neutrality policies (Lyu et al., 2022), such as constructing underground carbon storage facilities to reduce greenhouse gas emissions or developing renewable energy facilities like geothermal and underground energy storage systems to promote clean energy utilization. Additionally, emphasizing walking, cycling, and public transportation in underground space planning and design to reduce car usage and carbon emissions is also aligned with carbon neutrality goals. Therefore, future research can explore a more diverse set of socio-economic indicators to comprehensively assess the suitability of underground space development.

In summary, this study provides an innovative perspective for assessing the suitability of urban underground space, contributing to optimizing the functional layout of underground space. However, further research is needed to understand the interaction between underground space development and socio-economic conditions in more depth, enhancing our understanding of the potential of underground space development and providing valuable insights for future urban planning and underground space development.

6 Conclusion

The significance of underground space in new urbanization is rapidly increasing. However, to ensure sustainable urban development and improve living standards, a rational evaluation system for underground space development suitability is crucial. The study’s conclusions are as follows:

1. Evaluation using the ‘Combination weight method’ shows that in Ma’anshan, the very high suitability area and slightly high suitability area cover 23.95% and 30.84% of the total area, respectively. The most suitable district for underground space development is Huashan District, with the very high suitability area covering 46.72% and the slightly high suitability area covering 22.98% of the total area.

2. The ‘Combination weight method’ used in this study effectively identified a wider range of highly suitable development zones, particularly in the central urban areas. This method, combining subjective and objective factors, is more practical than the Analytic Hierarchy Process (AHP) for assessing underground space development suitability. Therefore, the ‘Combination weight method’ can be considered for evaluating underground space development suitability in cities where underground space utilization is in its early stages.

Disclosure statement

No potential conflict of interest was reported by the author(s).