Suitability Evaluation of Underground Space Development by Considering Socio-Economic Factors—An Empirical Study from Longgang Region of China

Abstract

Underground space is considered a critical urban resource that can significantly promote sustainable development under rational planning. This study, taking the Longgang region in SE China as an example, comprehensively considers the dual influences of geological environmental factors and socio-economic factors, incorporating socio-economic factors as key cost indicators. Following the principle of “stratification and classification”, a detailed assessment framework was developed to evaluate the suitability of underground space for development across various depths, and a systematic analysis of development suitability was conducted. Specifically, we employed a fuzzy comprehensive evaluation method to assess the suitability of underground space. During this process, an analytic hierarchy process (AHP) was used to determine the weights of geological environmental indicators, and the group judgment matrix approach was applied to assign weights to the socio-economic factors. The results indicated that, for shallow spaces, areas with high resource potential accounted for approximately 10.0% of the region, while areas with relatively high resource potential accounted for 28.5%. For medium-depth spaces, areas with high resource potential comprised 19.9% of the region, and areas with relatively high resource potential accounted for about 35.1%. These findings suggest that the study area demonstrated promising prospects for the development and utilization of underground space. The proposed approaches ensured that the evaluation results were both scientific and reasonable. By integrating the impact of socio-economic factors into suitability evaluation, the outputs provided more scientifically grounded guidance for urban planning.

1. Introduction

With the rapid urbanization in China, cities are facing a series of challenges, including high population density, traffic congestion, and environmental degradation, which have gradually hindered the pace of urban development [1,2]. The rational development and utilization of underground space have become critical approaches and an inevitable trend to promote sustainable urban development, which can improve resource efficiency [3,4,5,6]. Nonetheless, underground space is a finite and non-renewable resource [7,8]. Inappropriate development and utilization not only lead to resource wastage but may also trigger a series of geological hazards [9,10], which severely threaten public life and property safety [11,12,13,14,15]. Evaluating the potential for urban underground space development can optimize the use of urban land resources, effectively alleviate population density and traffic congestion, and contribute to the improvement of the overall urban environment [16,17,18]. To achieve these goals, it is important to conduct evaluations of underground space suitability for large cities, particularly for tunnel construction projects [19,20].

Many researchers have investigated the suitability assessment of urban underground space development from diverse perspectives. In terms of methodologies, many studies have employed the analytic hierarchy process (AHP) for suitability evaluation [21,22], with optimal algorithms to enhance the effectiveness of results. To enable intelligent and proactive quantitative planning of underground space development, some scholars have integrated Artificial Intervention Genetic Algorithms (AIGAs) to establish intelligent planning models, guiding the development and utilization of urban underground space [23]. Moreover, different evaluation indicator systems have been developed for underground space resources, based on GIS and multiple mathematical tools (e.g., the least adverse grading method) [24]. The integration of the AHP with the fuzzy comprehensive evaluation (FCE) method has also been explored [25], leading to the construction of multi-level fuzzy comprehensive evaluation models for assessing the suitability of underground space [26,27]. However, traditional fuzzy AHP methods heavily rely on expert opinions and personal experience, resulting in high subjectivity. To address this, the optimal transfer matrix has been employed to refine the traditional AHP method, ensuring more objective and reliable weight determination for evaluation indicators [28,29]. Considering the impact of temporal dimensions, time-weighted methods have been introduced into underground space resource evaluations. Combining the entropy weight method and the time dimension weighting method, a hybrid weight allocation model of urban underground space resources was proposed, called the entropy weight and time weighting model (E-TW) [30]. Xu et al. [31] employed a Bayesian network model as a decision support tool to assess the suitability of underground space development in Changsha. Similarly, Dou et al. [32] proposed an innovative 3D geological suitability evaluation framework for underground space resources, successfully integrating both 2D and 3D geological data to enhance the accuracy of the evaluation. Moreover, the literature also reported studies coupling GIS and the AHP for urban sustainability assessment in other regional settings, especially in arid regions. For example, Alogayell et al. [33] assessed land suitability for urbanization in the NEOM region of Saudi Arabia by using the GIS and AHP applications. Specially, the assessment results can help in sustainable urbanization site selection, as well as optimal land use planning [34,35], which highlights the reliability of such methods.

A critical step in the suitability evaluation process is the comprehensive selection of evaluation indicators. The literature has often employed combinations of methods such as FCE, AHP, AIGA, and Bayesian network models to form an evaluation indicator system [36]. From the perspective of research content and indicator selection, current studies primarily focus on geological conditions. However, in addition to geological and environmental factors, socio-economic factors are also important in underground space suitability evaluation [37]. For example, some researchers have explored the relationships between urban underground space development and factors such as per capita GDP, population density, and real estate prices. Their findings highlighted that per capita GDP and population density positively predicted the development of urban underground space [38]. Yan et al. [39] analyzed the suitability of underground space development in Ma’anshan City using the AHP and combined weighting methods based on point-of-interest data, nighttime light radiation rates, and population data. Bobylev [40] investigated the link between population density and the capacity of urban underground infrastructure in Stockholm, Tokyo, and Paris, finding a positive correlation between higher population density and increased underground space capacity. Wang et al. [41] employed structural equation modeling to analyze factors influencing the development potential of urban underground space. Their findings highlighted five key factors: geological conditions, land prices, economic development levels, benefits of underground space development, and alignment with urban planning. In addition, studies on multi-criteria decision modeling for infrastructure development in desert environments also validated the positive roles of integration of socio-economic and environmental considerations in land suitability analysis [33,34].

Hence, it can be found that the development and utilization of underground space in urban areas have been widely explored, and various evaluation indicator systems have been proposed. However, previous studies generally face the following limitations: (i) Many studies focused on geological or environmental factors, without considering the impact of socio-economic factors and their interactions. (ii) Although some studies incorporated multi-dimensional evaluation systems, they lacked region-specific analyses for particular cities and failed to adequately integrate local urban development plans and practical needs. With regard to this issue, the present study aims to propose a new integrated assessment framework for underground space development in big cities. We comprehensively considered geological and socio-economic factors and the future urban development plan, which presented evident differences in comparison to other studies. Specially, we included several socio-economic factors that rarely exist in the literature, including the influence of spatial location on the efficiency of underground space development, the positive predictive effect of population density and economic level on the demand for underground space, the decisive impact of land resource value on the economic feasibility of development, and the role of land development and utilization factors in assessing the feasibility and potential of development. Longgang City in SE China was taken as a case study, and the suitability for the development and utilization of shallow (0 to −15 m) and middle-depth (−15 to −30 m) underground space were assessed, respectively.

2. Study Area

Longgang City is located at the southern part of Zhejiang Province of SE China (Figure 1). The geographical coordinates are about 27°28′ to 27°35′ north latitude and 120°27′ to 120°40′ east longitude. It is bordered by the East China Sea to the east and is adjacent to Cangnan County in the south and Pingyang County in the north. The city covers an area of approximately 150 km², including a small mountainous area of about 15 km² and a relatively large plain area of about 135 km².

The topography of the region is higher in the southwest and lower in the northeast, with the main geomorphic units including plains, hills, tidal flats, and alluvial landforms. The northeastern part of the area is near the coast, characterized by low-lying terrain predominantly composed of coastal alluvial plains, with elevations typically ranging from 0 to 20 m, representing a typical coastal plain region. In contrast, the southwestern part consists of an alternating pattern of hills and mountains, with higher elevations mostly within the range of 100 to 300 m, forming a region of low hills and mountains. The geological structure of the study area is relatively complex, with multiple fault zones. Prominent northeast-trending faults include the Wenzhou–Zhenhai Fault and the Taishun–Huangyan Fault. The stratigraphy of the study area primarily consists of pyroclastic rocks, rhyolites, and quaternary sedimentary layers, indicating relatively stable geological conditions. The study area is rich in groundwater resources, which can be classified into two main types: pore water in loose rock formations and fissure water in bedrock. In the plain areas along river and coastal zones, groundwater exhibits saline characteristics, which impose mild corrosiveness on concrete and reinforced concrete. This corrosiveness is particularly pronounced under alternating wet and dry conditions.

According to remote sensing interpretation and statistics, the total area of the existing underground space in Longgang City is ~15.4 km², which are mainly located in the old downtown and the eastern part of the city (Figure 2). Currently, the utilization of underground space in Longgang City is relatively singular, primarily focused on public facilities such as parking lots, water and power supply facilities, communication pipelines, and cables. The layout of these facilities has contributed to the intensive use of urban space to some extent. However, the developed underground space area in Longgang City accounts for only one-tenth of the total area, indicating significant potential for further underground space development. Nonetheless, there are several issues and deficiencies in the current utilization of underground space in Longgang City. First, the forms of underground space utilization are relatively limited, with fewer areas developed for residential, commercial, or transportation purposes. Second, the distribution of underground spaces is relatively fragmented and lacks effective connectivity and integration, which hinders the formation of a systematic underground space network. This, in turn, limits the overall efficiency of underground space utilization.

3. Data and Methods

3.1. General Overview of the Modeling Process

There are two categories of factors that were mainly considered in our analysis, namely the geological environmental and socio-economic factors. Based on these factors, two evaluations were first conducted by using the fuzzy comprehensive evaluation (FCE) method, namely geological environmental suitability and socio-economic value evaluations. The AHP method was utilized to obtain the weights of geological environmental factors, whereas the group judgment matrix was applied to determine the weight distribution of socio-economic factors. The details for the foundations of these methods will be described in the following sections. There are several reasons to set the FCE and AHP as our major modeling methods. The FCE method is suitable for addressing uncertainty in multi-criteria decision making and can effectively account for the fuzziness in geological, environmental, and socio-economic factors, making it particularly well suited for the evaluation of complex systems like underground space. The AHP method, on the other hand, establishes a hierarchical structure to determine the importance of each factor, helping to clarify the weights and ensuring the scientific and rationality of the evaluation. While TOPSIS and PROMETHEE are common multi-criteria decision-making methods, they have certain limitations in handling fuzziness and uncertainty. Therefore, the FCE and AHP methods are more appropriate for the requirements of this study.

Next, based on the results of geological environmental suitability and socio-economic value evaluations, a graded assessment of the underground space resource potential in Longgang City is conducted. The calculation formula for the potential value of underground space is as follows:

$$P_e=S_e\times E_e$$

(1)

where P_e represents the underground space resource potential value, S_e denotes the underground space suitability evaluation value, and E_e refers to the underground space socio-economic value evaluation value.

3.2. Fuzzy Comprehensive Evaluation Method

3.2.1. Model Principles

The FCE method is an integrated evaluation approach based on fuzzy mathematics [42]. It employs the theory of membership degrees to transform qualitative evaluations into quantitative analyses, enabling a comprehensive assessment of objects or systems constrained by multiple factors [43,44]. This method effectively addresses uncertainties and ambiguities in decision-making processes [45] and has been widely applied in the fields of urban development and urban underground space suitability evaluation [46,47].

Based on the criteria for evaluating the suitability of underground space development and utilization, the stratified weighted average method is adopted to perform two-layer score weighting. During the geological environmental factor analysis, the evaluation model is calculated using the following formula:

$$S_e={{\displaystyle {\sum }_{i=1}^{n}w}}_i{r}_i$$

(2)

where S_e is the suitability evaluation value of underground space; r_i is the evaluation value of the theme layer (geological environmental factors); and w_i is the theme layer weight. When it comes to the socio-economic factors, the evaluation model can be denoted as follows:

$$E_e={{\displaystyle {\sum }_{m=1}^{n}w}}_m{r}_m$$

(3)

where E_e is the social and economic value of underground space development; r_m is the evaluation value of the theme layer (socio-economic factors); and w_m is the weight of the theme layer. Hence, it can be found that the methods above have similar structures, but different categories of factors (namely r_i and r_m) will be used.

Taking the geological environmental factors (r_i) as an example, the calculation formula of the main layer evaluation model can be denoted as follows:

$$r_i={\displaystyle {\sum }_{j=1}^n{u}_{ij}{w}_{ij}}$$

(4)

where u_ij represents the j-th indicator under the i-th thematic layer, and w_ij represents the weight of the j-th indicator in the i-th thematic layer.

3.2.2. FCE Modeling Process

In our case, the main steps of the FCE modeling process are as follows [48,49]:

(i)

The relationship between the influencing factors of underground space development and the evaluation objects was analyzed. Based on the classification criteria of the evaluation objects, the factor set (H) was determined. Given that the criteria for underground space development and utilization are multi-level standards, the characteristics of each level were described and quantified to determine the evaluation set (V).

(ii)

The fuzzy relationship for each indicator was determined during this step. It should be noted that the indicator system included both qualitative and quantitative indicators. The former were mainly descriptive factors such as lithology and groundwater corrosivity, and their membership degree was determined according to expert experience and knowledge. The quantitative indicators were in numerical form such as land subsidence and soft soil thickness. According to the statistical characteristics of the data of each factor, we applied three different mathematical functions to determine their membership degree, namely the descending half-trapezoid membership function, the ascending half-trapezoid membership function, and the triangular distribution membership function, represented by μ (x). In these equations, x represents the value of the indicator. For the descending half-trapezoid membership function, the thresholds (a and b) were set as mean and mean plus standard deviation. For the ascending half-trapezoid membership function, the thresholds were mean minus standard deviation and mean values. For the triangular distribution membership function, the three thresholds (a, b, and c) were mean minus standard deviation, mean, mean plus standard deviation, respectively. The detailed equations for the three functions are as follows:

The first kind of descending semi-trapezoidal membership function can be calculated as follows:

$$\mu \left(x\right)=\left\{\begin{array}{cc}1& x<a\\ {\displaystyle \frac{b-x}{b-a}}& a\le x\le b\\ 0& x>b\end{array}\right.$$

(5)

The second kind of ascending semi-trapezoidal membership function is denoted as follows:

$$\mu \left(x\right)=\left\{\begin{array}{cc}1& x<a\\ {\displaystyle \frac{x-a}{b-a}}& a\le x\le b\\ 0& x>b\end{array}\right.$$

(6)

The third kind of triangular distribution membership function is calculated as follows:

$$\mu \left(x\right)=\left\{\begin{array}{cc}0& \begin{array}{ccc}x\le a& or& x\ge c\end{array}\\ {\displaystyle \frac{x-a}{b-a}}& a<x<b\\ {\displaystyle \frac{c-x}{c-b}}& b\le x<c\end{array}\right.$$

(7)

Then, the correlation matrix (R) among indicators was determined by using the Pearson correlations as follows:

$$R=\left[\begin{array}{cccc}{R}_{11}& R_{12}& \cdots & R_{1j}\\ R_{21}& R_{22}& \cdots & R_{2j}\\ ⋮& ⋮& ⋮& ⋮\\ R_{i1}& R_{i2}& \cdots & R_{ij}\end{array}\right]$$

(8)

where R_ij represents the Pearson correlation coefficient between the ith and the jth evaluation indexes, which can denoted as follows:

$$R_{ij}={\displaystyle \frac{{\displaystyle {\sum }_{k=1}^n(x_{ik}-\overline{x_i})(x_{jk}-\overline{x_j})}}{\sqrt{{\displaystyle {\sum }_{k=1}^n{(x_{ik}-\overline{x_i})}^2{(x_{jk}-\overline{x_j})}^2}}}}$$

(9)

where x_ik is the normalization value for the kth evaluation block of the ith evaluation index; x_i is the average value of the ith evaluation index. Next, the eigenvector corresponding to the maximum eigenvalue of the above matrix (8) should be found. The squares of the components of this eigenvector, after their normalization to unit length, can be used as the weights of features for subsequent use.

(iii)

The AHP and group judgment matrix methods were used to determine the weights of geological environmental and socio-economic factors, respectively. The details of these two methods will be presented in the following section. After determining the fuzzy evaluation matrix R and fuzzy weight W, the comprehensive evaluation model expression is B = W * R. B is a fuzzy comprehensive evaluation set, denoted by B = (b₁, b₂, …, b_n). b_n is the membership degree of the evaluation grade V_n to the grade fuzzy subset B obtained by the comprehensive evaluation.

3.3. Evaluation System Construction

3.3.1. Geological Environmental Factors

Underground space development is a multi-level, multi-criteria comprehensive problem. Its development is influenced by both geological environmental factors and socio-economic factors. In our analysis, four categories of geological environmental factors were taken into account, namely topography and landform, engineering geological conditions, hydrogeological conditions, and geological safety risks. The topography and landform are primarily evaluated based on surface elevation (Figure 3). The engineering geological conditions use indicators such as lithological composition, soil compressibility, soil bearing capacity, and fill soil (Figure 4). Hydrogeological conditions include groundwater table depth, groundwater abundance, and soil permeability (Figure 5). Geological safety risks are assessed based on the indicators such as shallow gas, cumulative ground settlement, and the ground settlement rate (Figure 6).

Based on the general requirements of local authorities, expert knowledge, and engineering geological surveys (e.g., in-site monitoring data), an evaluation indicator system suitable for the study area is developed. Each evaluation indicator is classified into four levels for underground space development: highly suitable (I), moderately suitable (II), generally suitable (III), and poorly suitable (IV). It should be noted that the sustainability evaluation is conducted for different vertical layers. Specifically, we considered two different layers, namely the shallow underground space (0~−15 m) (

Table 1. Suitability evaluation index system for the development and utilization of shallow underground space (0~−15 m).

Evaluation Factors	Evaluation Indices	Evaluation Index Classification Criteria
		I	II	II	IV
Terrain landform	Ground elevation	>6 m	4~6 m	2~4 m	<2 m
Engineering geology	Lithology	Bedrock	Cohesive soil	Gravel soil, sand, silt	Special soil
	Soil compression coefficient	<0.1 Mpa−1	0.1~0.5 MPa−1	0.5~1 Mpa−1	>1 MPa−1
	Soil bearing capacity	>120 KPa	100~120 KPa	80~100 KPa	<80 KPa
	Thickness of fill soil	No distribution	0~1.5 m	1.5~3.0 m	>3.0 m
	Soft soil thickness	No distribution	0~30 m	30~40 m	>40 m
Hydrogeology	Depth of the groundwater table	>0.8 m	0.6~0.8 m	0.4~0.6 m	<0.4 m
	Groundwater abundance	<5.0 m3/d	5.0~10.0 m3/d	10.0~15.0 m3/d	>15.0 m3/d
	Soil permeability	<0.01 m/d	0.01~0.5 m/d	0.5~1.0 m/d	>1.0 m/d
Geological safety risks	Shallow gas	Main distribution	Secondary distribution area	Potential distribution	Probability small area
	Cumulative land subsidence	<50 mm	50~100 mm	100~200 mm	>200 mm
	Land settlement rate	<10 mm/a	10~20 mm/a	20~30 mm/a	>30 mm/a

Note: Grade I assignment takes 10, Grade II takes 7, Grade III takes 4, and Grade IV takes 1.

) and the middle layer underground space (−15~−30 m) (

Table 2. Suitability evaluation index system for development and utilization of underground space middle layer. (−15~−30 m).

Evaluation Factors	Evaluation Indices	Evaluation Index Classification Criteria
		I	II	III	IV
Terrain landform	Ground elevation	>6 m	4~6 m	2~4 m	<2 m
Engineering geology	Lithology	Bedrock	Cohesive soil	Gravel soil, sand, silt	Special soil
	Soil compression coefficient	<0.1 MPa−1	0.1~0.5 MPa−1	0.5~1 MPa−1	>1 MPa−1
	Soil bearing capacity	>350 KPa	300~350 KPa	250~300 KPa	<250 KPa
	Thickness of fill soil	No distribution	0–1.5 m	1.5–3.0 m	>3.0 m
	Soft soil thickness	No distribution	0~30 m	30~40 m	>40 m
	Thickness of anti-burst soil layer	>40 m	30~40 m	20~30 m	<20 m
Hydrogeology	Depth of the groundwater table	>0.8 m	0.6~0.8 m	0.4~0.6 m	<0.4 m
	Groundwater abundance	<5.0 m3/d	5.0~10.0 m3/d	10.0~15.0 m3/d	>15.0 m3/d
	Soil permeability	<0.01 m/d	0.01~0.5 m/d	0.5~1.0 m/d	>1.0 m/d
Geological safety risks	Shallow gas	Main distribution	Secondary distribution area	Potential distribution	Probability small area
	Cumulative land subsidence	<50 mm	50~100 mm	100~200 mm	>200 mm
	Land subsidence rate	<10 mm/a	10~20 mm/a	20~30 mm/a	>30 mm/a

Note: Grade I assignment takes 10, Grade II takes 7, Grade III takes 4, and Grade IV takes 1.

).

3.3.2. Socio-Economic Factors

Based on data related to the economic development status, population distribution, land use patterns, and social demands of the study area, a comprehensive analysis of the socio-economic value of underground space development was conducted. It was concluded that the socio-economic value assessment of underground space development in the study area is primarily influenced by four key factors: spatial distribution, economic development, land use patterns, and infrastructure accessibility. Population conditions used population density as the evaluation indicator. Land resource value was evaluated using base land prices for commercial, residential, public service, and industrial land. Land development and utilization were assessed using the proportion of construction land and land development intensity as indicators. The detailed information of each evaluation indicator is shown in Figure 7, Figure 8, Figure 9 and Figure 10.

Based on the engineering geological survey results and monitoring data of Longgang City, each evaluation indicator was classified into four levels. These levels are high socio-economic value (I), relatively high socio-economic value (II), moderate socio-economic value (III), and low socio-economic value (IV). A socio-economic value evaluation indicator system for underground space development in Longgang City was established accordingly (

Table 3. Evaluation index system of social and economic value of underground space development.

Evaluation Factors	Evaluation Indices	Evaluation Index Classification Criteria
		I	II	III	IV
Spatial location	Spatial location	Urban core	Business centers and other urban administrative centers	Urban residential, office areas, and other public areas of the city	Protected areas such as urban style, history, and culture, as well as other prohibited and restricted areas
Population status	Population density	0.20 × 104/km2	0.15~0.20 × 104/km2	0.1~0.15 × 104/km2	<0.1 × 104/km2
Land resource value	Base land price of commercial land	Class I commercial	Class II commercial	Class III commercial	Class IV and V commercial
	Base land price of residential land	Class I residential	Class II residential	Class III residential	Class IV and V residential
	Base land price for public service land	Class I public service	Class II public service	Class III public service	Class IV and V public service
	Base land price for industrial land	Class I industrial	Class II industrial	Class III industrial	Class IV and V industrial
Land development and utilization	Proportion of construction land	>60%	30%~60%	<30%
	Land development intensity	Class IV and V density areas	Class III density areas	Class II density areas	Class I density areas

).

3.4. Determination of Indicator Weight

Determining the weights of evaluation indicators is a critical step in the suitability evaluation process. The validity of the indicator weights is closely linked to the reliability of the final evaluation outcomes, which underscores the importance of selecting an appropriate weighting method. In this study, the indicator weights for geological environment evaluation were assigned using the AHP, while the indicator weights for socio-economic value evaluation were determined using the group judgment matrix method. The key step in the AHP is constructing the judgment matrix, as the quality of the matrix directly affects the calculated weight rankings. Typically, the judgment matrix is derived through expert consultation. To mitigate the influence of individual preferences, the number of experts involved in the decision-making process should not be fewer than two, thereby adopting a group judgment matrix.

3.4.1. Analytic Hierarchy Process

The AHP is a decision-making technique that breaks down decision-related elements into various levels, including objectives, criteria, and options. It combines qualitative and quantitative analysis, forming a systematic hierarchical structure. The primary approach involves decomposing a complex issue into various components, considering the problem’s nature and the overall goals to be accomplished. These components are then grouped hierarchically based on their interrelationships and membership relationships. By making simple comparisons and calculations between the factors, the weights of different alternatives can be derived, allowing for the determination of their relative rankings and providing a basis for selecting the optimal alternative [50,51,52].

In the AHP method, determining the indicator weights requires constructing a judgment matrix for pairwise comparisons of elements at each hierarchical level and assigning quantified values to each indicator. The eigenvector associated with the maximum eigenvalue of the matrix represents the relative importance ranking of the evaluation factors. After normalization, the resulting eigenvector (W) gives the weight distribution of each evaluation indicator. During the construction of the judgment matrix, due to the diversity of underground space resources, the matrix may deviate significantly from consistency, leading to unreasonable weight distribution. Therefore, consistency checking is required, and the consistency ratio (CR) should be calculated as follows [53,54]:

$$CI={\displaystyle \frac{{\lambda }_{\max }-n}{n-1}},CR={\displaystyle \frac{CI}{RI}}$$

(10)

In this context, CI represents the consistency index of the matrix, λ_max refers to the maximum eigenvalue of the matrix, and n denotes the order of the matrix. RI is the random consistency index, which can be obtained from a table based on the order of the matrix (

Table 4. Matrix element judgment scale.

Scale	Implication
1	The two factors are equally important.
3	One factor is slightly more important than the other.
5	One factor is moderately more important than the other.
7	One factor is strongly more important than the other.
9	One factor is exceptionally more important than the other.
2, 4, 6, 8	Median of the above two adjacent judgments.
Reciprocal	The judgment of factor i compared with j is aij; then, the judgment of factor j compared with i is aji = 1/aij

and

Table 5. Random consistency index RI (Saaty).

n	1	2	3	4	5	6	7	8	9
RI	0	0	0.58	0.90	1.12	1.24	1.32	1.41	1.45

). When the CR value is less than 0.1, the comparison matrix is considered to exhibit satisfactory consistency.

In this study, the main modeling process of the AHP is as follows:

(i)

Establish the hierarchical structure model for weight calculation: The evaluation model is structured into three levels: the goal layer, the criterion layer, and the indicator layer (as shown in Figure 11), forming a complete site suitability evaluation process. The target layer represents the final result of the site suitability evaluation. The criterion layer classifies indicators based on their relevance, dividing the evaluation indicators into four categories: hydrogeological conditions, engineering geological conditions, geological safety risk conditions, and topographic and geomorphological conditions. The indicator layer consists of specific evaluation factors.

(ii)

Target–criterion layer comparison matrix and consistency test: According to the comparison matrix element discrimination scale, the target–criterion layer judgment matrix is established (

Table 6. Target–criterion layer judgment matrix.

Suitability A	Terrain Landform B1	Engineering Geology B2	Hydrogeology B3	Geological Safety Risk B4
Terrain landform B1	1	1/3	2	4
Engineering geology B2	3	1	3	5
Hydrogeology B3	1/2	1/3	1	2
Geological safety risks B4	1/4	1/5	1/2	1

). The maximum eigenvalue of the comparison matrix is obtained and substituted into the matrix. The normalized eigenvector corresponding to the maximum eigenvalue can be obtained as W = (0.0846,0.3891,0.2632,0.2631) ^T. Therefore, the consistency of the A-B (target–criterion layer) discriminant matrix meets the requirements.

(iii)

Criteria–index layer comparison matrix and consistency test: In the same way, a comparison matrix for the criterion–index layer can be established to calculate the maximum eigenvalue of the matrix and the corresponding normalized eigenvector. At the same time, a consistency test is required. The results show that the consistency test results of each comparison matrix meet the preset standards.

(iv)

Hierarchical total ranking and consistency test: After completing the hierarchical analysis of the target–criterion layer and the criterion–index layer, the weights of the two layers need to be weighted and integrated to perform hierarchical total ranking. The goal of this process is to calculate the relative importance of the weight of each index relative to the target layer and determine the overall weight of the index layer factors. The results are shown in

Table 7. Comprehensive weight of spatial indicators for the shallow layer.

Primary Indicators	Weights	Secondary Indicators	Weights	Comprehensive Weight
Terrain landforms	0.0846	Ground elevation	1	0.0846
Engineering geology	0.3890	Lithology	0.2589	0.1007
		Soil compression index	0.1651	0.0642
		Soil bearing capacity	0.1722	0.0670
		Thickness of fill soil	0.2883	0.1122
		Soft soil thickness	0.1155	0.0449
Hydrogeology	0.2632	Depth of groundwater table	0.3672	0.0966
		Groundwater abundance	0.2446	0.0644
		Soil permeability	0.3882	0.1022
Geological safety risk	0.2631	Shallow gas	0.4286	0.1128
		Cumulative ground settlement	0.2037	0.0536
		Ground settlement rate	0.3677	0.0967

and

Table 8. Comprehensive weight of spatial indicators for the middle layer.

Primary Indicators	Weights	Secondary Indicators	Weights	Comprehensive Weight
Terrain landforms	0.0846	Ground elevation	1	0.0846
Geotechnical conditions	0.3929	Lithology	0.2214	0.0870
		Soil compression index	0.2101	0.0825
		Soil bearing capacity	0.2472	0.0971
		Thickness of fill soil	0.1001	0.0393
		Soft soil thickness	0.1111	0.0437
		Thickness of anti-burst soil layer	0.1101	0.0433
Hydrogeology	0.2631	Depth of groundwater table	0.1532	0.0403
		Groundwater abundance	0.3895	0.1025
		Soil permeability	0.4573	0.1203
Geological safety risk	0.2630	Shallow gas	0.4367	0.1149
		Cumulative ground settlement	0.1989	0.0523
		Ground settlement rate	0.3644	0.0958

. In order to ensure the credibility of the analysis results, the total ranking of the hierarchy was tested for consistency, and the results showed that the consistency reached a satisfactory level.

3.4.2. Group Judgment Matrix Method

The process for calculating the weights of socio-economic indicators using the optimal transitive matrix in the group judgment matrix method is as follows:

(i)

Expert Selection: The selection of experts plays a crucial role in the evaluation results in the group judgment matrix method. To ensure the scientific and objective nature of the evaluation, the experts were chosen based on the following criteria: first, experts should have at least five years of research or practical experience in fields such as underground space development, urban planning, or geotechnical engineering; second, experts should possess a certain academic background or industry influence, having published multiple papers or participated in large-scale projects in the relevant fields.

(ii)

Weight Assignment: In the weight allocation process, experts assign scores to each evaluation factor based on the practical situation of underground space development and the importance of each indicator. Quantified scoring standards are used to ensure objectivity and standardization in the scoring process. All scoring results are input into the judgment matrix, and the final weights are calculated using the group judgment matrix method. The scoring criteria are based on the relative importance of each indicator in the context of underground space development.

(iii)

Handling Expert Discrepancies: To address potential biases and differences in expert ratings, this study adopts a discussion mechanism for expert scores. During the initial scoring phase, experts independently submit their scores. In cases of significant discrepancies, a collective discussion is organized, where experts share their reasoning and scoring basis, and consensus is reached through collaborative negotiation. If large differences still persist after the discussion, extreme value removal or weighted averaging methods will be applied to ensure the objectivity and consistency of the scoring results.

(iv)

Consistency Check: To ensure that the scores in the group judgment matrix method are consistent and reasonable, a consistency check is conducted. After calculating the weights, the consistency ratio (CR) is tested. If the CR value exceeds 0.1, the scoring matrix will be further revised. If the CR value is less than 0.1, the scoring matrix is considered to be sufficiently consistent, and the final weight values are deemed reliable. The core principle behind this approach is explored below.

The number of participants in the consultation decision-making process is denoted by m, and the judgment matrices provided by each participant are represented as A₁, A₂, …, A_m.

Let B = (b_ij)_n×n be an antisymmetric matrix. If b_ij = b_ik + b_kj, i, j, k ∈ {1, 2, …, n}, then B is called a transitive matrix.

Let B_l = [b_ij^(l)]² be an antisymmetric matrix, 1 ∈ {1, 2, …, n}; if there exists a transfer matrix C = (c_ij)_n×n, ${{\displaystyle {\sum }_{i=1}^n{\displaystyle {\sum }_{j=1}^n{\displaystyle {\sum }_{l=1}^n\left[c_{ij}-{b_{ij}}^{(l)}\right]}}}}^2$ as a minimum, then C is called the optimal transfer matrix of B₁, B₂, …, B_m.

Expert feedback scores are collected using the expert scoring method, and a judgment matrix is constructed based on these scores. Subsequently, the group judgment matrix is optimized and computed to determine the weights of each indicator. To validate the robustness of the weight allocation scheme, this study employed a sensitivity analysis method. In this process, slight variations were made to some of the weights, and their impact on the suitability evaluation results for underground space development was analyzed. The results indicated that changes in the weights within a reasonable range did not significantly affect the evaluation outcomes, thereby ensuring the stability and scientific reliability of the evaluation system. The results are shown in

Table 9. Comprehensive weight of socio-economic value evaluation indicators.

Primary Indicators	Weights	Secondary Indicators	Weights	Comprehensive Weight
Spatial location	0.1832	Spatial location	1	0.1832
Population status	0.1598	Population density	1	0.1598
Land resource value	0.3312	Base land price of commercial land	0.2771	0.0918
		Base land price of residential land	0.2653	0.0879
		Base land price for public service land	0.2690	0.0891
		Base land price for industrial land	0.1886	0.0625
Land development and utilization	0.3258	Proportion of construction land	0.4673	0.1522
		Land development intensity	0.5327	0.1736

.

4. Results

4.1. Geological Environmental Suitability Evaluation

The Longgang Plain covers a total area of approximately 121 km². Based on the urban development plan of Longgang City and the “Land Use Planning Map” specified in the Longgang City Spatial Master Plan (2021–2035), the city’s construction land was subdivided into 2071 evaluation units, with a total area of approximately 30.3 km². The results of the weighted calculations for the underground space suitability evaluation were assigned to each evaluation unit. According to the geological environment suitability evaluation standards for underground space (

Table 10. Evaluation criteria for geological environmental suitability of underground space.

Suitability Classification	Good Suitability	Fairly Good Suitability	Moderate Suitability	Poor Suitability
	Se ≥ 6.0	5.0 ≤ Se < 6.0	4.0 ≤ Se < 5.0	Se < 4.0

), the underground space suitability of Longgang City was categorized into four levels, ultimately forming suitability evaluation results for the development and utilization of underground space at varying depths. The evaluation results are shown in Figure 12 and Figure 13.

For the shallow underground space, areas with high suitability are primarily located in the southern plain region, near Ruqiaotou and the Dongmenyang community, with a total area of approximately 0.44 km², accounting for about 2.5% of the city. Areas with moderate suitability are mainly distributed in the central plain region near the Qihe, Liudian, and Huangzhong communities, as well as the southern region around Ruqiaotou–Linjia Courtyard, covering approximately 2.26 km², or 12.9%, of the area. Areas with general suitability are widely distributed in the central and eastern parts of the plain, covering about 12.35 km², which constitutes approximately 70.4%. Areas with low suitability are distributed in patches in the northern and eastern plain regions, covering around 2.49 km², or about 14.2% of the area.

For medium-depth underground space, areas with high suitability are primarily located in the central plain region, east of Huangzhong community, and in the southern areas near Ruqiaotou and Dongmenyang communities, covering approximately 0.64 km², which accounts for about 3.6% of the evaluated area. Areas with moderate suitability show a significant increase compared to the shallow layer, mainly concentrated in the central plain region, covering approximately 6.97 km², or 39.7%, of the area. Areas with general suitability are widely distributed in the central and certain eastern parts of the plain, spanning around 9.62 km², which constitutes about 54.9% of the area. Areas with low suitability are sporadically distributed in the central plain region, covering approximately 0.32 km², or about 1.8%, of the area.

4.2. Socio-Economic Value Evaluation

For the socio-economic value evaluation of underground space development, and in alignment with the urban development plan of Longgang City, this evaluation primarily relies on the “Land Use Planning Map” from the Cangnan County Longgang Town Urban Master Plan (2011–2030), which outlines the planned development zones (as shown in Figure 6). Based on the site conditions, appropriate adjustments were made. The space was partitioned, and a total of 2071 evaluation units were delineated, covering an area of 30.3 km². The weighted results of the socio-economic value evaluation were assigned to each evaluation unit. According to the underground space socio-economic value evaluation criteria, the socio-economic value of underground space in Longgang City was classified into four levels (

Table 11. Socio-economic value evaluation criteria for underground space.

Suitability Classification	High Socio-Economic Value	Fairly High Socio-Economic Value	Moderate Socio-Economic Value	Low Socio-Economic Value
	Ee ≥ 6.0	5.0 ≤ Ee < 6.0	4.0 ≤ Ee < 5.0	Ee < 4.0

). The final outcome is the socio-economic value evaluation for the development and utilization of underground space in Longgang City.

The areas with high economic value are primarily located in the northern part of the plain area, including the old town and the northeastern part of Longgang New City, covering an area of approximately 10.0 km², accounting for about 30.9% of the evaluation area. The areas with fairly high economic value are mainly located in the northern part of the plain area, including the old town and the northeastern Longgang New City, covering an area of approximately 5.7 km², accounting for about 18.8%. The areas with moderate economic value are mainly distributed in the central and eastern parts of the plain area, covering an area of approximately 3.0 km², accounting for about 9.9%. The areas with low economic value are distributed in patches in the western and southeastern parts of the plain area, covering an area of approximately 12.3 km², accounting for about 40.5% (as shown in Figure 14).

4.3. Evaluation of the Development and Utilization Potential of Underground Space

Based on the calculation results from the underground space resource potential formula, and after excluding the areas with already developed underground space, the underground space resource potential in Longgang City is classified into four levels: high potential, fairly high potential, moderate potential, and low potential (

Table 12. Underground space resource potential evaluation criteria.

Suitability Classification	High Resource Potential	Fairly High Resource Potential	Moderate Resource Potential	Low Resource Potential
	Pe ≥ 30.0	23.0 ≤ Pe < 30.0	16.0 ≤ Pe < 23.0	Pe < 16.0

). According to the underground space resource potential evaluation values (Pe), Longgang City’s underground space resource potential is divided into four categories based on the magnitude of the potential evaluation value.

Based on the results of the underground space resource potential zoning evaluation, Longgang City is divided into four categories of zones for underground space resource potential, categorized into shallow and middle layers (as shown in Figure 15, Figure 16 and Figure 17).

For the shallow underground space, the areas with high resource potential are sparsely distributed in the central and northern parts of the plain area, around the Taian, Chaoxiwu, and Tuchang communities, covering an area of approximately 1.76 km², accounting for about 10.0% of the evaluation area. The areas with fairly high resource potential are primarily located in the central and northern parts of the plain area, around the Xianyuan, Longpu, Lijiayang, Xiala, Xiangang, and Liudian communities and further to the north. Larger areas are also found in the central and northern parts of Longgang New City. This area covers approximately 5.0 km², accounting for about 28.5%. The areas with moderate resource potential are distributed in contiguous patches surrounding the areas with fairly high potential, mainly located to the north of the Xiqiao East River–Longhe–Longping–Liudian community line in the central and northern plain area, as well as around the Ruqiaotou, Jianhouyang, and Zhaochang communities and the central and northern parts of Longgang New City. The total distribution area is approximately 5.6 km², accounting for about 32.1%. The areas with low resource potential are distributed in the western and southeastern parts of the plain area, covering an area of approximately 5.2 km², accounting for about 29.4%.

For middle-layer underground space, the areas with high resource potential are sparsely distributed in the central and northern parts of the plain area, around the Taian, Chaoxiwu, Dongcheng, and Tuchang communities, with smaller distributions in the central and northern parts of Longgang New City. The total area of these zones is approximately 3.5 km², accounting for about 19.9% of the evaluation area. The areas with fairly high resource potential are distributed in a broader area surrounding the high-potential zones, primarily in the central and northern parts of the plain area, around the Xianyuan, Xiqiao, Longhe, and Liudian communities and further to the north, with significant areas also found in the central and northern parts of Longgang New City. This zone covers approximately 6.2 km², accounting for about 35.1%. The areas with moderate resource potential are interspersed between the higher potential zones or distributed on their periphery, with a total distribution area of approximately 3.2 km², accounting for about 18.2%. The areas with low resource potential are distributed in patches in the western and southeastern parts of the plain area, covering an area of approximately 4.7 km², accounting for about 26.8%.

5. Discussion

This chapter expands on the results we found and attempts to clarify the limitations during the analysis and provide a potential direction for future research. Then, some practical implementation insights are recommended for policymakers and urban planners.

The present study systematically investigated the impact of socio-economic conditions on the development of underground space based on the fuzzy comprehensive evaluation model. We considered the dual roles of geological environmental factors and socio-economic factors to establish a suitability evaluation system for underground space development and utilization. The proposed framework was considered comprehensive and scientific, especially compared with previous works. However, it should be noted that urban development is dynamic, particularly when it comes to socio-economic conditions and land use planning [55]. In our analysis, most of the geological environmental factors are temporally static (e.g., lithology, soil properties) or do not change rapidly (e.g., cumulative ground settlement and settlement rate). In contrast, the socio-economic factors mostly vary temporally, such as population density and base land price [56]. This leads us to conclude that the assessment results are relevant with time and depend more on the socio-economic factors. However, we note that it is difficult to quantify the temporal relevance of the results, and few efforts have been made regarding this point in the literature. Nevertheless, it is still recommended to better incorporate the update cycle to ensure the timeliness of the evaluation. For example, stakeholders can perform suitability evaluations every 3 to 5 years to effectively reflect the changes in urban development, population changes, and alterations in environmental conditions.

While our analysis highlighted the economic benefits of socio-economic factors on underground space development, certain limitations still exist, among which the selection of influencing factors should be stated first. For example, it has been found that public transport accessibility to urban areas and economic indicators (such as per capita GDP, employment density) have a relationship to the capacity of the urban underground space [57,58]. These factors were not considered in this study due to the lack of data availability. However, they were not completely absent from our analysis. For example, the indicator population density was included to reflect population status, which was similar with employment density. The base land price and land development intensity can be considered proxies of the GDP and economic level. In fact, some of these factors are spatially related to each other; thus, we believed that eliminating some factors avoided the issue of multicollinearity [59,60]. Moreover, the detailed effects of these factors are also associated with urban spatial structures, which may be rather complicated. For example, Zhang and Buyuklieva found that the urban structure of Shanghai City had a spatial cluster, and each factor had different roles in each single cluster. Hence, the question of how to incorporate these factors into future analysis is still an operational challenge since they are highly dynamic in terms of time and space.

In addition, some environmental factors that are not related to geological conditions and the financial feasibility of using underground spaces were not taken into account. This is mainly because we focused on technical and socio-economic assessments, while environmental and financial evaluations require more specific data and complex models for support. Regarding the latter topic, it has been reported that the cost for tunnel construction in China ranges from USD 10 to USD 15 million per km. This cost is a fraction of the cost experienced in other countries, mainly in the developed world, for example, USD 52 million in the US and USD 45 million in New Zealand [61]. This is mainly attributed among others to the low labor cost. In addition, the urban underground space in China is developed by nationalized businesses instead of private companies. Hence, once a project is determined as helpful for society, financial feasibility is not commonly the critical factor considered by local authorities. This is also supported by the literature review, since very few studies have been conducted regarding the financial feasibility of using underground space in China [62]. Hence, it can be expected that future studies will strengthen the comprehensive assessment of environmental impacts. Financial indicators such as cost–benefit analysis, project payback period prediction, risk assessment, and financing schemes will also be introduced to evaluate the financial feasibility of underground space development. To further clarify the reliability of the selected factors, we compared them with those of other, similar studies [63] and found that the number of factors in this study (12 geological environmental factors and 8 socio-economic factors) were similar or even larger than them. Hence, we conclude that the number of the present factors are sufficient and fit with our research objectives.

Another limitation regarding factor selection is the lack of opinions of the population living in the area. It is essential to thoroughly understand user perception of the underground space when constructing a human-friendly underground space [49]. However, it may be difficult to conduct such studies. First, personal opinions about the development of an underground space could be rather subjective. Second, investigations into the population’s opinions at the city level is time-consuming. Most of the similar studies conducted surveys based on questionnaires [64]. However, it is difficult to collect sufficient and efficient answers. In China, the population in big cities could be large (>4 × 10⁷ in our study area), and even tens of thousands of opinions are insufficient. A future study could further explore the question of how to enhance public participation through methods such as surveys and focus groups, in order to better understand residents’ views, needs, and potential concerns regarding underground space development. This would improve the feasibility and sustainability of the projects.

In this study, the internal variabilities of evaluation indicators and their impact on assessment results were not adequately addressed. In practical suitability evaluations, the internal variability of individual indicators may significantly influence the overall evaluation results. Even though other factors perform well, excessive internal variability within a single indicator (e.g., spatial heterogeneity of geological conditions or local fluctuations in traffic flow) may lead to overall evaluation results that deviate from actual conditions, particularly in the context of micro-scale planning and development. Future research could apply multiple techniques for analyzing and adjusting for internal variability across different indicators. For example, a variable weight comprehensive model could be adopted to dynamically adjust weights for indicators with significant internal variability, which could help reduce biases introduced by subjective weight assignments [63].

Moreover, the interactive mechanisms between underground space development and socio-economic conditions were not explicitly revealed. As Chen et al. [1] mentioned, socio-economic conditions are not only key factors influencing the suitability of development in an underground space but are also significantly impacted by the development itself, creating bidirectional interaction. However, this study did not explore this two-way relationship in depth, thus limiting the practical guidance provided by the results. Future studies should place greater emphasis on the coupling mechanisms between underground space development models and socio-economic conditions, clarifying the pathways and modes of their interactions.

Based on the research findings, the following implementation recommendations can be proposed for urban planners: (i) The evaluation results provide a clear understanding of the development potential in different areas. High-potential regions, such as the northern and north-central parts of Longgang New City, including communities like Taian and Chaoxiwu, should be prioritized for underground space development. These areas exhibit favorable geological and socio-economic conditions, which will help enhance development efficiency and economic returns. In contrast, areas with lower potential should consider postponing development, or, alternatively, the development plan could be adjusted to align with other urban growth needs. Such regions may face significant challenges in underground space development, and premature development could bring higher risks and costs. (ii) Underground space development should take into account both ecological protection and social demands. Strict environmental assessment and monitoring mechanisms must be established to prevent irreversible damage to natural resources and ecosystems during development. Particularly, ecological restoration and green development pathways should be integrated into the development process, promoting sustainable resource utilization. Moreover, special attention should be paid to the housing and living conditions of low-income populations, ensuring that underground space development does not lead to social inequities. (iii) Long-term monitoring and feedback systems should be established. As underground space development progresses, the original potential evaluation values may change. Therefore, a dynamic tracking system should be established to continuously monitor the evolution of underground space development. Urban planners should be prepared to make timely adjustments to plans and policies to address emerging issues and challenges in the future.

Last but not least, although this study uses Longgang City as a case study, the proposed assessment method for underground space development is also applicable to other cities and regions, particularly those facing rapid urbanization and increasing demand for underground space. By appropriately adjusting the evaluation indicator system and weight settings, this method can be widely applied to the assessment of underground space development potential in different settings.

6. Conclusions

The suitability evaluation of underground space development can optimize the use of land resources and promote urban sustainability. However, previous works often ignore the role of social and economic factors. In this study, we developed a general framework, aiming to establish a comprehensive factor system. The Longgang City in SE China was selected as the study area. Following the principles of “layering and classification”, the geological environmental conditions, geological safety risks, and socio-economic factors were all incorporated into the analysis. The AHP method was utilized to calculate the weight of each geological environmental factor. The results showed that the importance of each indicator varied, but shallow gas always had a relatively high weight. The suitability evaluation results of underground space development were presented according to the geological layers. We found that middle underground space had more suitable areas for potential development. Regarding the socio-environmental factors, their weights were determined by using the group judgment matrix method, the results of which highlighted the importance of spatial location and land development intensity. Moreover, regions with high economic value accounted for 30.9% of the area, those with relatively high economic value represented 18.8%, those with moderate economic value represented 9.9%, and regions with low economic value accounted for 40.5%.

Finally, by combining the results of geological environmental suitability and socio-economic value evaluation, a classification of underground space resource potential in Longgang City was conducted. For the shallow space, the percentages for different levels were 10.0% (high potential), 28.5% (relatively high potential), 32.1% (moderate potential) and 29.4% (low potential), respectively. In contrast, the middle space showed higher potential on the underground space resources, with the percentages reaching 19.9% (high potential), 35.1% (relatively high potential), 18.2% (moderate potential), and 26.8% (low potential), respectively.

Although this study focused on a specific site, we still highlighted its broader applicability beyond the case study area. The proposed method and framework can be easily replicated in other urban areas for assessing underground space development. Moreover, refining factor selection or testing the model in other regions should be further investigated in future research.