Study on the Early Warning Mechanism for Industrial Land Redevelopment in High-Tech Zones: A Multi-Dimensional Evaluation Based on Enterprise Life Cycle, Park Compatibility, and Land Use Efficiency

Abstract

In the era of new productive forces, the efficient utilization of industrial land in high-tech zones is critical for fostering technological innovation, intelligent manufacturing, and green development. However, constrained by limited land reserves, inefficient stock utilization, and sluggish industrial upgrading, high-tech zones must establish a scientific early warning mechanism for industrial land redevelopment. This study constructs a four-tier early warning system (normal, alert, warning, and response) based on three key dimensions: enterprise life cycle, enterprise–park compatibility, and industrial land use efficiency. Using the Jinan High-Tech Zone as a case study, this study conducts an empirical analysis of 360 industrial land parcels from 2020 to 2022, employing DEA, fixed effects models, GIS visualization, and MCDA methods. The results indicate a strong correlation between enterprise life cycle and land use efficiency, with significant spatial differentiation in enterprise–park compatibility. Efficient land use is concentrated in areas with well-defined functions and high industrial agglomeration. This study identifies 360 land use scenarios, with 12% classified as normal, 28% requiring monitoring, 52% requiring optimization, and 8% necessitating redevelopment. Based on these findings, a “warning–monitoring–regulation” closed-loop management model is proposed, providing decision-making support for dynamic land optimization and sustainable development in high-tech zones.

1. Introduction

Against the backdrop of China’s structural shift in economic development and the deepening of high-quality growth strategies, high-tech zones have long served as strategic engines of regional economic transformation, playing a critical role in advancing technological innovation, intelligent manufacturing, and green development [1]. However, the current industrial land use model in these zones faces mounting challenges [2]. On the one hand, land supply is severely constrained by national land quota controls and urban development boundaries, making traditional expansion-driven development paths increasingly unsustainable. On the other hand, substantial inefficiencies exist in the use of existing industrial land, including functional mismatches and spatial fragmentation, which have limited the capacity to support the growth of emerging and high-value-added industries [3]. Moreover, many of the zones’ early-stage industrial structures—centered on investment attraction and conventional manufacturing—have entered maturity or decline, with diminishing land productivity and enterprise dynamism. Meanwhile, emerging industries increasingly demand technologically advanced and environmentally sustainable spatial carriers, intensifying the structural mismatch between supply and demand [4,5,6,7].

At the national level, policy directives such as the National New-type Urbanization Plan (2014–2020) and the 13th Five-Year Land and Resources Plan have prioritized revitalizing underutilized land, restructuring land use patterns, and improving efficiency. As China’s first pilot zone for growth driver transformation, Shandong Province underscores the importance of reallocating inefficient industrial land toward digital, intelligent, and high-end sectors to advance structural upgrading [8,9]. Within this policy framework, developing a scientific and adaptive redevelopment mechanism is critical to overcoming land constraints and enabling sustainable industrial transformation in high-tech zones.

Despite growing scholarly attention to industrial land governance, several critical limitations remain in the existing research. First, most studies adopt a single-dimensional evaluation approach [10,11,12], lacking an integrated framework to identify land inefficiencies from the joint perspectives of enterprise dynamics, park compatibility, and land productivity. Second, prevailing models are often static, overlooking the spatiotemporal evolution of enterprise development and mismatches between industrial demand and spatial configuration. Third, there is an absence of policy-linked frameworks that connect land evaluation results to differentiated regulatory strategies, limiting their practical value in supporting redevelopment decisions [13,14,15,16].

To address the challenges of low industrial land use efficiency and structural imbalance in high-tech zones, this study develops a multi-dimensional early warning mechanism for industrial land redevelopment, aiming to establish a closed-loop governance system integrating identification, evaluation, and intervention. This mechanism comprehensively considers three core dimensions—enterprise life cycle, enterprise–park compatibility, and industrial land use efficiency—and systematically assesses industrial land risks from the micro, meso, and macro levels. Methodologically, this study integrates Data Envelopment Analysis (DEA), Fixed Effects Models (FEMs), Multi-Criteria Decision Analysis (MCDA), Analytic Hierarchy Process (AHP), and Geographic Information Systems (GISs) to support the quantification, visualization, and dynamic updating of the early warning model. This mechanism facilitates the transition from static assessment to dynamic governance of industrial land and enhances the precision and adaptability of land resource allocation in high-tech zones.

2. Literature Review

2.1. Development of High-Tech Zones and Industrial Land Redevelopment

As key drivers of regional economic growth and technological innovation, high-tech zones play a vital role in shaping land use patterns and the quality of industrial development. The existing literature mainly focuses on the development trajectory of high-tech zones in China [17,18,19], their operational models, and associated land use challenges, highlighting prominent issues such as the disjunction between industry and urban life [20], extensive land use practices, and insufficient innovation capacity [21]. In the field of industrial land redevelopment, researchers have concentrated on land intensification, stock land redevelopment models, and policy tools. Recently, optimizing land use structure and improving efficiency have become prominent strategies for facilitating urban renewal and industrial upgrading [22,23].

However, current studies tend to emphasize traditional industrial parks or general urban renewal cases, with limited exploration of the dynamic adjustment of industrial land within high-tech zones [24]. Systematic evaluations of industrial land redevelopment in these zones remain scarce, particularly in terms of fine-grained analyses of existing land stock. Moreover, the spatial-temporal evolution of industrial land use efficiency in high-tech zones and strategies for precise optimization have yet to be fully developed within a comprehensive theoretical framework.

2.2. Enterprise Life Cycle, Park Compatibility, and Industrial Land Use Efficiency

The enterprise life cycle theory has long served as a foundational perspective for understanding firm development dynamics. While recent studies have begun to explore how different life cycle stages affect land utilization, its integration into the field of industrial land research remains relatively limited [25]. Empirical evidence suggests that firms at different stages—start-up, growth, maturity, and decline—exhibit significantly heterogeneous land demand and usage behaviors. For example, early-stage firms tend to favor flexible and short-term land arrangements, mature firms prioritize stability and long-term occupancy, while declining firms are often associated with reduced land use efficiency. However, limited attention has been paid to the spatial distribution of these life cycle stages and their interactions with park-level spatial structures—factors that are essential for understanding dynamic shifts in land performance.

Another underexplored dimension is enterprise–park compatibility, which refers to the degree to which an enterprise aligns with the spatial, infrastructural, and policy advantages of its host park. The existing literature has primarily focused on compatibility between firms and employees, customers, organizational culture, or strategic objectives [26], while overlooking the land governance implications of spatial mismatch. Enterprises with low compatibility often face inefficient resource allocation, structural misalignment, and declining productivity—challenges that ultimately impair overall park performance. As such, improving enterprise–park compatibility has become an increasingly critical concern in industrial land governance.

Industrial land use efficiency serves as a key indicator for assessing the rationality of land resource allocation. Existing research has established multi-dimensional evaluation frameworks incorporating factors such as land intensification, economic output, and spatial performance, using diverse analytical techniques including Data Envelopment Analysis (DEA) [27,28], entropy-based weighting methods, TOPSIS [29], and matter-element models. While these approaches emphasize parallel evaluation of land use processes and outcomes, they largely remain static in nature, failing to capture the evolving dynamics of enterprise behavior and thereby constraining the precision of policy interventions.

Synthesizing existing studies, although enterprise life cycle, park compatibility, and land use efficiency have each received considerable attention, an integrated framework combining these three dimensions to support dynamic early warning and classification-based governance of industrial land in high-tech zones remains absent. Addressing this research gap, this study proposes a multi-dimensional early warning mechanism that integrates enterprise development dynamics, spatial matching efficiency, and comprehensive land performance, thereby providing refined and highly adaptive evaluation support for the redevelopment of industrial land in high-tech zones.

2.3. Application of Early Warning Mechanisms in Industrial Land Management

Early warning systems have been widely adopted in land management domains such as arable land protection, pollution control, real estate market regulation, and land reserve risk management, with positive outcomes [30]. These systems rely on indicator-based frameworks to monitor and assess land use risks in real time, providing scientific support for policy interventions. However, their application in industrial land management remains in its infancy, lacking an integrated research framework. Specifically, the following apply: (1) most existing warning systems focus on single dimensions, such as ecological safety or economic performance, and fail to capture the multifaceted dynamics of industrial land use; (2) current systems often rely on static assessments and do not fully consider the interplay between enterprise life cycles, park compatibility, and land use efficiency; and (3) there is a shortage of studies targeting the unique land characteristics of high-tech zones and developing tailored early warning frameworks to support industrial upgrading and land redevelopment.

In conclusion, although enterprise life cycle, park compatibility, and land use efficiency have been individually addressed in the existing literature, few studies have conceptually integrated them into a unified early warning framework. This lack of a comprehensive, multi-dimensional mechanism—linking time-based enterprise dynamics, spatial matching efficiency, and land use performance—remains a key research gap in supporting differentiated and adaptive industrial land redevelopment strategies in high-tech zones.

2.4. Definition of Core Concepts

A scientifically grounded and logically coherent early warning mechanism for industrial land redevelopment requires clear conceptual definitions of its key evaluation dimensions. This study identifies three such dimensions—enterprise life cycle, enterprise–park compatibility, and industrial land use efficiency—which collectively form the foundation of the analytical framework and ensure consistency in indicator selection and methodological rigor.

Enterprise life cycle refers to the sequential stages a firm undergoes during its development—start-up, growth, maturity, and decline—each characterized by distinct resource needs, operational efficiency, and growth dynamics [25]. In this study, a dual-indicator matrix combining cross-sectional operational efficiency and time-series growth performance is employed to classify firms into four life cycle stages: high growth and low efficiency (start-up), high growth and high efficiency (growth), low growth and high efficiency (maturity), and low growth and low efficiency (decline). This framework enables a dynamic reflection of firm-level land performance and development potential.

Enterprise–park compatibility is an indicator measuring the degree to which a firm’s development objectives align with the spatial functions and strategic direction of the host park [26]. Specifically, in this study, compatibility refers to the extent to which an enterprise matches the high-tech zone’s development priorities of “high technology content, high value-added output, and industry–city integration”. This dimension serves as a critical tool for identifying spatial mismatches and guiding the optimization of industrial agglomeration.

Industrial land use efficiency denotes the comprehensive performance of a land parcel in generating economic, social, and ecological benefits under a given level of input [28]. Building on traditional land efficiency assessments, this study incorporates a land intensification dimension to construct a multidimensional performance evaluation system. This enables the precise identification of underperforming land parcels and supports the prioritization of targeted redevelopment strategies.

3. Materials and Methods

3.1. Materials

3.1.1. Study Area

The Jinan High-Tech Zone (JHTZ) is one of China’s first national high-tech zones and a core area of the National Pilot Zone for New and Old Kinetic Energy Conversion, leveraging policy advantages to drive high-quality economic growth and industrial transformation. Over the past two decades, its central urban area has integrated seamlessly with Qilu Software Park, forming a model of industry–city integration. Currently, JHTZ hosts nearly 50,000 registered enterprises, including 990 high-tech firms and 102 listed or publicly traded companies. The industrial structure has evolved from traditional software sectors to emerging industries such as artificial intelligence, big data, and integrated circuits, with a well-developed industrial development framework. However, the zone’s land use composition reveals significant inefficiencies—construction land accounts for 82.42%, while industrial land represents only 2.20%, highlighting the urgent need for industrial land redevelopment due to diverse yet inefficient land utilization. As a representative case of industrial transformation and stock land redevelopment, JHTZ’s central urban area presents a significant research value.

This study focuses on the downtown area of the Jinan High-Tech Zone, aiming to construct a practice-oriented early warning mechanism for industrial land redevelopment. As a typical sample of national high-tech zones, this area is facing structural problems such as land tension, functional separation and lack of industrial renewal, which are common in global high-tech zones, and is also in the transition stage of “integration of industry and city”, whose diversified enterprise life cycle, complex spatial structure, and positive policy response provide an ideal scenario for empirical testing of the early warning mechanism. The core value of choosing this region lies in the following: (1) both the common features of global innovation parks and the national strategic orientation, and its development trajectory, reflect the evolution path of China’s high-tech zones; (2) as the core area of old and new kinetic energy conversion, it undertakes the task of exploring the transformation and upgrading of industries, which has the experimental nature and complexity of the policy; (3) it forms the typical scenarios of emerging industry cultivation and redevelopment of old industrial land in the transformation of the land in stock; and (4) it possesses rich data resources and dynamic verification conditions, providing an observable basis for mechanism construction.

By integrating international mature methods and local practices, this study effectively avoids region-specific bias, and its theoretical system and methodology have universal guiding value for many types of high-tech zones.

3.1.2. Data Sources

The data used in this study primarily encompass two aspects: industrial land use and enterprise development in the Jinan High-Tech Zone. (See Figure 1).

On the one hand, industrial land use data were obtained from the 2020–2022 Industrial Land Utilization Survey conducted in the Jinan High-Tech Zone, urban inefficient construction land redevelopment projects, Jinan’s master plan and regulatory zoning data, as well as supporting datasets from the “Per-Mu Benefit” evaluation reform and urban inefficient land redevelopment projects in the high-tech zone. Additionally, data from the Third National Land Survey were incorporated.

On the other hand, enterprise development and operational growth data were mainly sourced from statistical reports of Jinan High-Tech Zone’s Statistics Bureau, Tax Bureau, Human Resources and Social Security Bureau, and various park administrative offices. Supplementary sources included The Annals of Jinan High-Tech Industrial Development Zone, corporate annual reports published in the National Enterprise Credit Information Publicity System, and financial reports from publicly listed enterprises such as Inspur Electronic Information Industry Co., Ltd. (Jinan, China), Shandong Gold Mining Co., Ltd. (Jinan, China), and Joyoung Co., Ltd. (Jinan, China). For small and medium-sized enterprises (SMEs), specific indicators were acquired through third-party survey agencies commissioned by the administrative committee of the Jinan High-Tech Zone. These multi-source data ensure the reliability and accuracy of this study.

This study selects the central urban area of the Jinan High-Tech Zone as the empirical research region. To ensure data richness and accuracy, a wide range of heterogeneous sources was integrated, including current land use records, cadastral parcel vector data, and enterprise registration and operational statistics. On this basis, a sample framework was constructed by linking spatial parcel data with corresponding enterprise-level information, allowing for an in-depth analysis of land use patterns and firm dynamics. The presence of 990 registered high-tech enterprises in the study area is not only indicative of substantial industrial agglomeration but also reflects the region’s high technological intensity and spatial vitality. These characteristics collectively delineate the broader industrial structure and the integrated development of industry and urban space within the high-tech zone.

However, not all of these enterprises were included as direct objects of analysis. Instead, this study adopts the land parcel as the core research unit and matches corresponding enterprise data accordingly. Specifically, 360 land parcels with clearly defined industrial use attributes, identifiable spatial boundaries, and relatively complete enterprise data were selected as the main analytical sample. Since each parcel corresponds to a single enterprise, a total of 360 enterprises were ultimately included in the analysis. The decision to use parcels rather than individual enterprises as the primary unit of analysis was made for several key reasons:

(1)

Some enterprises, although qualified as high-tech firms, are still in incubation stages, lacking complete business data or stable output, thus making them unsuitable for evaluating lifecycle stages or land use performance;

(2)

Many parcels exhibit “multi-tenant” or “shared-office” patterns, where multiple micro or small enterprises occupy the same building or floor. This leads to ambiguous spatial attribution and blurred performance boundaries, making it difficult to establish a valid parcel–enterprise mapping;

(3)

Certain firms operate in non-industrial sectors (e.g., education or training) that fall outside the scope of industrial land performance evaluation, thereby affecting the internal consistency and generalizability of the model.

It is worth noting that although certain early-stage firms were excluded due to data limitations, a number of representative “start-up phase” enterprises were retained among the 360 valid parcels. Their inclusion in the lifecycle classification ensures that the model captures real structural heterogeneity.

In summary, this sampling strategy balances data controllability and spatial rationality, supports the logical closure required for early warning model construction, and provides a structured, scalable framework that can be adapted and extended to other regions in future applications.

3.2. Methods

To achieve dynamic early warning and classification-based governance for industrial land redevelopment in high-tech zones, this study establishes an integrated evaluation system across temporal, spatial, and performance dimensions based on three core aspects: enterprise life cycle, enterprise–park compatibility, and industrial land use efficiency. This system supports a closed-loop logic of “identification–classification–governance”.

In the design of the evaluation framework, the enterprise life cycle dimension captures the dynamic evolution of enterprise growth and corresponding changes in land demand; the enterprise–park compatibility dimension measures the degree of spatial matching and functional synergy; and the industrial land use efficiency dimension assesses the comprehensive benefits of land resource allocation. Logically interlinked, these three dimensions form a continuous chain of dynamic evolution, spatial adaptation, and performance optimization, providing multidimensional support for the early warning mechanism.

In terms of technical methodology, this study follows a structured sequence: First, Data Envelopment Analysis (DEA) is employed to evaluate enterprise operational efficiency, combined with the Fixed Effects Model (FEM) to assess enterprise growth performance, thereby constructing a dual-indicator matrix of “efficiency–growth” for life cycle stage identification. Second, DEA is further applied to measure both enterprise–park compatibility and multi-dimensional industrial land use efficiency, ensuring quantitative assessments of spatial synergy and resource performance. Subsequently, Multi-Criteria Decision Analysis (MCDA) is used to integrate scores across dimensions, with the Analytic Hierarchy Process (AHP) employed to determine indicator weights, and the Weighted Sum Model (WSM) applied to calculate comprehensive early warning scores for each land parcel. Finally, a quantile-based segmentation method is introduced to define early warning thresholds, classifying land parcels into four risk levels—normal, alert, warning, and response—to support tiered governance and targeted interventions.

This methodological system achieves an organic integration of data aggregation, indicator linkage, and dynamic warning, ensuring the scientific rigor, systematic structure, and practical operability of industrial land management. (See Figure 2).

3.2.1. Data Envelopment Analysis (DEA)

Data Envelopment Analysis (DEA) is a non-parametric linear programming method used to evaluate the relative efficiency of decision-making units (DMUs) under multiple input–output conditions [31]. This study employs the BCC (Banker–Charnes–Cooper) model, which distinguishes between technical efficiency and scale efficiency, making it particularly suitable for the diversified industrial structure of high-tech zones. By constructing a production possibility frontier, the model assesses each DMU’s deviation from the optimal frontier and provides guidance for optimization [32,33,34].

In the DEA model, the efficiency score θ for each DMU is calculated using the following formula:

$$\mathrm{m}\mathrm{a}\mathrm{x}\mathsf{\theta }={\displaystyle \frac{\sum _{\mathrm{r}=1}^{\mathrm{s}}{\mathrm{u}}_{\mathrm{r}}{\mathrm{y}}_{\mathrm{r}\mathrm{j}}}{\sum _{\mathrm{i}=1}^{\mathrm{m}}{\mathrm{v}}_{\mathrm{i}}{\mathrm{x}}_{\mathrm{i}\mathrm{j}}}}$$

(1)

where

max θ indicates that the objective function seeks to maximize the efficiency score θ, which represents the relative efficiency of each decision-making unit (DMU);

s is the number of output indicators;

m is the number of input indicators;

y_rj is the value of output r for DMU j;

x_ij is the value of input i for DMU j;

u_r is the weight assigned to output r;

v_i is the weight assigned to input i.

A DMU is considered efficient when θ = 1, indicating that it lies on the production frontier, while a score of θ < 1 reflects inefficiency relative to the best-performing units. The BCC model’s ability to distinguish between pure technical inefficiency and scale inefficiency further enhances its applicability in evaluating heterogeneous industrial land use contexts.

In this study, DEA serves as a core analytical method and is applied in three primary areas: (1) evaluating enterprise operational efficiency; (2) assessing enterprise–park compatibility; and (3) evaluating industrial land use efficiency. All of these applications adopt the input-oriented BCC DEA model, which aims to enhance resource utilization efficiency and minimize inputs under a fixed output condition, especially when resources cannot be easily adjusted. The input-oriented BCC model is particularly suitable for this context, as it effectively measures the rational allocation of resources and optimizes the efficiency of output per unit. In this study, the input-oriented BCC model is employed to assess enterprise operational efficiency, enterprise–park compatibility, and industrial land use efficiency, incorporating input indicators such as land area and infrastructure, which are rigid and cannot be readily adjusted. Therefore, this study utilizes this approach to explore how to more effectively utilize these rigid resources, thereby improving overall efficiency and land utilization levels.

3.2.2. Fixed Effects Model

The Fixed Effects Model (FEM) is a fundamental approach in panel data analysis, primarily used to control for unobserved, time-invariant individual characteristics, thereby improving the accuracy and robustness of model estimations [35,36]. It is particularly suitable for analyzing how variables change within entities over time, avoiding biases caused by ignoring individual heterogeneity. The model is expressed as follows:

$${\mathrm{G}\mathrm{r}\mathrm{o}\mathrm{w}\mathrm{t}\mathrm{h}}_{\mathrm{i}\mathrm{t}}=\mathsf{\alpha }+{\mathsf{\beta }}_1{\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{u}}_{\mathrm{i}}+{\mathsf{ϵ}}_{\mathrm{i}\mathrm{t}}$$

(2)

where

Growth_it is the dependent variable, representing the main business revenue growth rate of enterprise i in year t, used as a measure of enterprise growth performance;

$\mathsf{\alpha }$ is the constant term;

X_it is a vector of key explanatory variables, including total profit, industrial added value, fixed asset investment, and employment per unit output, reflecting enterprise operational efficiency and performance characteristics;

Z_it is a set of control variables accounting for other firm-level factors that may influence growth, such as firm age and industry classification;

β₁ and β₂ are coefficients to be estimated, representing the marginal effects of the corresponding variables;

${\mathrm{u}}_{\mathrm{i}}$ captures the individual fixed effects, controlling for unobserved, time-invariant heterogeneity across firms;

ϵ_it is the idiosyncratic error term, reflecting unobserved time-varying shocks at the firm level.

In this study, panel data from 2020 to 2022 at the enterprise level in the Jinan High-Tech Zone are utilized, and the FEM is employed to dynamically assess enterprise growth performance. By controlling for fixed differences in industry type, firm size, and other structural characteristics, the model effectively captures the impact of time-varying factors such as policy shifts and market dynamics, thereby enabling a dynamic evaluation of enterprise growth trajectories.

3.2.3. Multi-Criteria Decision Analysis (MCDA)

Multi-Criteria Decision Analysis (MCDA) is a systematic decision-support approach designed to address complex problems involving multiple, often conflicting objectives. Its core principle involves defining decision goals, selecting multi-dimensional evaluation criteria with assigned weights, and applying decision models to generate composite scores for alternative options, thus facilitating the integration and optimization of diverse information [37,38]. In this study, the Analytic Hierarchy Process (AHP) is used to determine the weights of evaluation indicators, and the Weighted Sum Model (WSM) is applied to calculate the composite scores.

Step 1: Define Evaluation Objectives and Criteria System

Clarify the evaluation objectives and establish a structured system of criteria to guide the assessment process.

Step 2: Standardization of DEA Evaluation Results

First, raw data are standardized to ensure comparability across different efficiency dimensions, as DEA-derived outputs may vary in units and scales.

Second, linear normalization is applied using the following formula:

$${\mathrm{x}}_{\mathrm{i}\mathrm{j}}^{\prime }={\displaystyle \frac{{\mathrm{x}}_{\mathrm{i}\mathrm{j}}-\mathrm{m}\mathrm{i}\mathrm{n}({\mathrm{x}}_{\mathrm{j}})}{\mathrm{m}\mathrm{a}\mathrm{x}({\mathrm{x}}_{\mathrm{j}})-\mathrm{m}\mathrm{i}\mathrm{n}({\mathrm{x}}_{\mathrm{j}})}}$$

(3)

where x′_ij is the normalized value of indicator j for unit i obtained through min-max scaling.

This transformation rescales all indicators to a [0, 1] range, facilitating consistent integration.

Step 3: Determining Indicator Weights Using the Analytic Hierarchy Process (AHP)

This study employs the expert scoring method to determine the weights of the indicators. Specifically, five experts with backgrounds in land management were invited to form an expert panel, ensuring the authority and scientific rigor of the evaluation. The experts established scoring standards based on the definitions and importance of each indicator, and then performed pairwise comparisons to assign relative importance scores ranging from 1 to 5 for each indicator. After scoring, the AHP relative matrix was constructed and consistency testing was conducted.

First, based on their experience and the evaluation framework, the experts conducted pairwise comparisons and assigned a relative importance score ranging from 1 to 5 for each indicator, resulting in the matrix element a_ij, where a_ij represents the relative importance of indicator i compared to indicator j.

Second, the consistency index (CI) is calculated as follows:

$$\mathrm{C}\mathrm{I}={\displaystyle \frac{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}-\mathrm{n}}{\mathrm{n}-1}}$$

(4)

where

$\mathsf{\lambda }$_max is the maximum eigenvalue of the matrix;

n is the order of the matrix.

Then, the consistency ratio (CR) is used to verify the consistency of the matrix, with the formula:

$$\mathrm{C}\mathrm{R}={\displaystyle \frac{\mathrm{C}\mathrm{I}}{\mathrm{R}\mathrm{I}}}$$

(5)

If CR < 0.1, the matrix is considered to have good consistency, and the final weights of the dimensions are determined.

Step 4: Calculation of the Composite Score

The composite early warning score for each unit is calculated using the Weighted Sum Model (WSM), which aggregates the standardized indicator values according to their respective weights:

$${\mathrm{S}}_{\mathrm{i}}={\sum }_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{w}}_{\mathrm{j}}·{\mathrm{x}}_{\mathrm{i}\mathrm{j}}^{\prime }$$

(6)

where

S_i is the composite score for unit i;

w_j is the weight assigned to indicator j;

x′_ij is the normalized value of indicator j for unit i.

This calculation integrates multiple evaluation dimensions into a single index, enabling comprehensive performance assessment and risk classification.

Step 5: Classification and Ranking of Results

Industrial land parcels are ranked in descending order based on their composite scores (S_i) derived from the Weighted Sum Model (WSM); the evaluation results are then further categorized using a quantile-based method to accommodate different research needs.

Step 6: GIS Visualization and Decision Support

Each parcel’s composite score is mapped onto the GIS system. Combined with the spatial layout of the high-tech zone, an efficiency map is generated to visually present the spatial distribution of composite evaluation results across land parcels.

In this study, Multi-Criteria Decision Analysis (MCDA) is applied in two key areas. First, it is used to evaluate the comprehensive performance of industrial land by integrating multiple efficiency dimensions, including intensive efficiency, economic benefits, social benefits, and ecological benefits. Second, MCDA is employed within the early warning mechanism to synthesize the three core evaluation dimensions—enterprise life cycle, enterprise–park compatibility, and multi-dimensional industrial land use efficiency—into a unified composite score for risk classification. This integration supports quantitative, tiered governance strategies for industrial land redevelopment.

3.2.4. Dual-Indicator Matrix

The dual-indicator matrix is a structured analytical tool designed to classify and evaluate objects based on two interrelated performance dimensions. It provides a visual and diagnostic framework by placing one indicator on the horizontal axis and the other on the vertical axis, thereby dividing the space into four quadrants. This classification allows for the differentiation of heterogeneous units and supports strategic decision-making across multiple evaluation contexts.

In this study, the dual-indicator matrix is specifically applied to assess enterprise life cycle stages by integrating two dimensions: operational efficiency (based on cross-sectional data) and growth performance (based on time-series analysis). This “efficiency–growth” coordinate framework enables the classification and structural differentiation of enterprise development trajectories.

4. Results

In response to the practical demand for industrial land renewal and redevelopment in high-tech zones, this study constructs an early warning mechanism tailored to industrial land redevelopment and conducts a systematic empirical evaluation of land use in the central urban area. By quantitatively analyzing enterprise development dynamics, spatial matching levels, and comprehensive land performance, different types of land parcels and their associated utilization risks are identified. These findings provide a decision-making basis for formulating differentiated and tiered governance strategies. The following presents the main empirical analysis results of this study.

4.1. Evaluation of Enterprise Life Cycle in High-Tech Zones Based on DEA and Dual-Indicator Matrix

4.1.1. Analysis of Enterprise Operational Efficiency

(1)

Enterprise Operational Efficiency Indicator System

Enterprise operational efficiency is a key indicator that reflects the level of resource allocation and business performance, playing a vital role in evaluating the efficiency of industrial land use. In this study, the Data Envelopment Analysis (DEA) method is employed, with land, capital, and labor as the main input variables, and output benefits and tax contributions as the main output variables, to establish an operational efficiency evaluation system. Based on the relative positioning of DEA scores within the sample, enterprises are categorized into high-efficiency and low-efficiency groups [39,40]. The detailed evaluation indicator system is presented in

Table 1. High-tech zone enterprise operational efficiency evaluation system.

Indicator Type	Indicator Name	Indicator Definition
Input Indicators	Land Supply Area [41]	Total land area occupied by the enterprise (m2)
	Fixed Asset Investment	Total capital invested in fixed asset acquisition (CNY)
	Number of Employees	Total number of employees in the enterprise or park (persons)
	Floor Area Ratio (FAR) [42]	Ratio of total building area to land area
	Building Density [43]	Ratio of built-up area to total land area
	Energy Consumption per Output	Energy consumed per unit of output (kWh per 10,000 CNY)
	Pollution Control Investment [44]	Total funds invested in environmental protection (CNY)
Output Indicators	Total Industrial Output Value	Total value of industrial production activities (CNY)
	Input–Output Ratio	Output generated per unit of input
	Land Profit Output Ratio	Profit per unit land area (CNY/m2)
	Tax Contribution per Unit Area	Tax revenue per unit land or building area (CNY/m2)
	Energy Consumption Reduction Rate [45]	Reduction percentage in energy use per unit output (%)
	Revenue per Unit Area	Revenue generated per unit land or building area (CNY/m2)

.

(2)

Evaluation Results and Analysis of Enterprise Operational Efficiency

The overall enterprise operational efficiency in the central urban area is relatively high, with 74.44% of enterprises classified as highly efficient. This performance is largely attributed to well-developed integrated services, commercial infrastructure, and an effective industry–city integration strategy, which have facilitated the growth of high-value-added industries such as intelligent manufacturing, big data, and financial technology.

However, challenges persist due to the high proportion of traditional software enterprises, lagging technological upgrades, and insufficient innovation in business models. These factors contribute to the existence of low-efficiency enterprises, which require enhanced resource optimization and targeted technological transformation support to improve their operational efficiency and align with the high-tech zone’s industrial upgrading goals.

4.1.2. Enterprise Growth Analysis

(1)

Enterprise Growth Analysis Index System

Enterprise growth reflects a company’s long-term development potential and market competitiveness. In this study, the Fixed Effects Model (FEM) is employed to evaluate enterprise growth, using the growth rate of main business revenue as the dependent variable. Key explanatory variables include total profit, industrial added value, fixed asset investment, and employment per unit output [46]. Based on the regression results, residuals are calculated to obtain growth deviation values. Enterprises with high growth are characterized by significantly positive residuals and high positive growth deviations, indicating performance beyond expected levels. The details of the enterprise growth evaluation indicator system are presented in

Table 2. Enterprise growth analysis index system in high-tech zones.

Index Type	Indicator Name	Indicator Definition
Dependent Variable	Main Business Revenue Growth Rate [47]	The percentage increase in a company’s main business revenue compared to the previous period (%)
Independent Variables	Profit	The difference between a company’s total revenue and total expenses over a given period (CNY)
	Industrial Added Value	The value added during the production process (CNY)
	Fixed Asset Investment [48]	The total investment in fixed assets made by a company or region over a given period (CNY)
	Employment per Unit Output	The number of employees required per unit of output (persons per CNY 10,000)

.

(2)

Evaluation of Enterprise Growth Performance

The evaluation of enterprise growth reveals that only 25.56% of enterprises exhibit high growth. The underlying causes can be attributed to two primary factors.

First, the ongoing industrial transition from traditional service sectors and software industries to emerging fields such as high-end software, big data, and artificial intelligence has yet to fully take effect. The prevalence of traditional industries and insufficient technological innovation constrain enterprise growth momentum.

Second, capital and technological constraints hinder the breakthrough potential of certain enterprises. Additionally, market volatility and the concentrated allocation of resources limit the expansion of small and medium-sized enterprises (SMEs), further restricting the overall growth potential within the industrial park.

4.1.3. Enterprise Life Cycle Assessment and Analysis Based on the Dual-Indicator Matrix

(1)

Dual-Indicator Vector Matrix

This study evaluates the enterprise life cycle based on a growth–efficiency matrix, where the horizontal axis represents market expansion capability and the vertical axis indicates resource utilization efficiency. Enterprises are categorized into four distinct stages: start-up stage (high growth, low efficiency); growth stage (high growth, high efficiency); maturity stage (low growth, high efficiency); decline stage (low growth, low efficiency). These classifications provide a structured framework for assessing the dynamic evolution of enterprises within the high-tech zone. Figure 3 illustrates the enterprise classification based on this matrix.

(2)

Enterprise Life Cycle Assessment Results

The enterprise life cycle assessment of the central district in the Jinan High-Tech Zone reveals distinct distribution patterns across different development stages. Start-up-phase enterprises account for 15%, characterized by both high growth potential and operational efficiency, predominantly concentrated in the software, artificial intelligence, and digital economy sectors. Growth-phase enterprises comprise 8%, exhibiting high growth but relatively low efficiency, often constrained by capital limitations and managerial challenges. The maturity phase encompasses 66% of enterprises, reflecting high operational efficiency but slowing growth, with some firms facing increasing pressure for technological upgrading and structural transformation. Declining enterprises represent 11%, suffering from low efficiency and stagnating growth, primarily due to industry obsolescence and resource misallocation, making them potential candidates for restructuring or exit strategies.

The spatial distribution of enterprises at different life cycle stages in the central urban area is visualized using GIS, as shown below Figure 4.

GIS spatial analysis reveals that the enterprise life cycle structure in the central urban area of the Jinan High-Tech Zone exhibits significant differentiation across development stages. Start-up- and growth-stage enterprises are predominantly concentrated in the core area of the high-tech zone, benefiting from the agglomeration of policy incentives, capital, and talent. This indicates that the core area possesses strong innovation capacity and growth-driving potential. In contrast, maturity- and decline-stage enterprises are mostly located in the earlier-developed zones, highlighting a lag in industrial renewal, resource reallocation, and technological upgrading in these regions.

From a mechanistic perspective, this spatial mismatch may stem from historical variations in investment promotion strategies, the rigidity of land allocation mechanisms, and the absence of a dynamic admission and exit system aligned with enterprise life cycle stages. These institutional factors contribute to the misalignment between spatial enterprise distribution and development needs, thereby constraining efficient land use and the sustainable evolution of the industrial park.

4.2. Evaluation and Analysis of Enterprise–Park Compatibility

Enterprise–park compatibility measures the alignment between enterprise resource demands, development strategies, and the industrial positioning of the park, directly influencing resource allocation efficiency and regional economic performance. Highly compatible enterprises can fully leverage park resources, drive industrial upgrading, and enhance land use efficiency, while low compatibility may lead to resource mismatches, hindering industrial development. Thus, this indicator is critical for optimizing industrial land use planning in high-tech zones.

Focusing on the core development objective of “industry–city integration,” this study constructs an enterprise–park compatibility evaluation framework based on the DEA model, incorporating three key dimensions: industrial advancement, urbanization degree, and integration outcomes.

A comprehensive assessment was conducted based on data from 360 industrial land parcels in the central urban area of the Jinan High-Tech Zone.

According to the relative distribution of DEA scores within the sample, enterprises are categorized into high, medium, and low compatibility groups. The detailed evaluation indicator system is presented in the

Table 3. Evaluation index system for enterprise–park compatibility in high-tech zones.

Index Type	Index Name	Index Description
Input Indicators	Number of Patents Granted per 10,000 People [49]	Number of patents granted per 10,000 people (units/10,000 people)
	Employee Settlement Rate	Percentage of enterprise employees settling in the local area (%)
	Pension Insurance Coverage Rate	Percentage of enterprise employees covered by pension insurance (%)
	Infrastructure Support Level	Degree of infrastructure development and service provision in the park
	R&D Investment as a Percentage of Revenue [49]	Proportion of enterprise revenue allocated to R&D (%)
	Enterprise Consumption in the Tertiary Sector	Proportion of enterprise spending on tertiary sector services (%)
	Employee Housing Subsidy Ratio	Percentage of employees receiving housing subsidies from the enterprise (%)
Output Indicators	High-Tech Industry Output as a Share of Total Industrial Output [50]	Proportion of high-tech industry output in the park’s total industrial output (%)
	Proportion of Employees with a Master’s Degree or Higher [51]	Percentage of employees with a master’s degree or higher (%)
	Per Capita Disposable Income of Urban Employees	Average disposable income of urban employees (CNY/year)
	Percentage of Employees with a Commute of Less Than 30 min	Proportion of employees whose commute time is under 30 min (%)

below.

The evaluation results indicate that high-compatibility enterprises (30%) are primarily concentrated in the park’s R&D bases and talent cultivation hubs, including leading high-tech enterprises such as Inspur Group and Jinan High-Tech Wanda Plaza Real Estate Co., Ltd. These enterprises play a crucial role in driving innovation and enhancing urban functionality. Moderate-compatibility enterprises (58%) mainly include certain manufacturing and service-oriented enterprises. Low-compatibility enterprises (12%) are predominantly traditional industries, single-service enterprises, and some real estate firms, which face challenges in talent attraction, technological upgrading, and industrial transformation. As a result, they struggle to support the development of high-end industries in the park, leading to industrial mismatches and inefficient resource utilization.

The enterprise–park compatibility assessment results for the central urban area are visualized using GIS, revealing the spatial distribution patterns as shown in Figure 5.

Through GIS-based visualization, it is evident that enterprises with low compatibility are predominantly concentrated in early-developed areas and traditional manufacturing clusters. These spatial patterns are largely a legacy of earlier park development policies that prioritized “land-based investment attraction” and adopted low-entry thresholds. Under such conditions, land was allocated without a dynamic evaluation mechanism for industrial suitability, resulting in a misalignment between some enterprises and the current strategic industrial orientation of the park. Furthermore, low compatibility also indicates a lack of coordination mechanisms between enterprise development and park objectives. To address this, the high-tech zone should establish a compatibility-oriented land access and exit mechanism, improve investment selection and performance evaluation systems, and promote structural optimization of the park’s industrial ecosystem through spatial reconfiguration and strategic resource reallocation.

4.3. Evaluation of Industrial Land Use Efficiency and Identification of Inefficient Land in High-Tech Zones

4.3.1. Definition of Industrial Land Use Efficiency and Inefficient Land

Industrial land use efficiency is comprehensively evaluated across four dimensions: intensiveness, economic benefits, social benefits, and ecological benefits. It reflects the capacity of land resources to maximize output while meeting economic, social, and ecological demands. Intensiveness efficiency assesses the effective utilization of land resources by optimizing spatial layout and resource allocation to enhance land productivity. Economic efficiency measures the economic contribution of land per unit area, including output value, tax revenue, and employment, indicating its role in supporting regional economic growth. Ecological efficiency evaluates the environmental impact of land use, covering energy consumption, pollution control, and green development. Social efficiency gauges the contribution of land use to social well-being, including employment promotion, livelihood improvement, and social equity.

Inefficient industrial land refers to land that fails to fully realize its potential. Its characteristics include low output efficiency, resulting in insufficient economic and social contributions; resource underutilization, where land remains undeveloped or has low utilization rates for extended periods; and spatial misallocation, characterized by fragmented enterprise distribution and a lack of industrial clustering and synergy.

4.3.2. Evaluation Method for Industrial Land Use Efficiency in High-Tech Zones

This study constructs a comprehensive evaluation system for industrial land use efficiency based on the DEA-BCC model and Multi-Criteria Decision Analysis (MCDA), systematically assessing performance across four dimensions: intensive efficiency, economic benefits, social benefits, and ecological benefits. The detailed evaluation index system is presented in the

Table 4. Evaluation index system for industrial land use efficiency in high-tech zones.

Index Type	Intensive Efficiency	Economic Benefits	Social Benefits	Ecological Benefits
Input Indicators	Land supply area	Total investment	Employment per unit of output value [52]	Energy consumption per unit of output value [53]
	Fixed asset investment	Number of employees	Per capita built-up area	Pollution treatment investment per unit of land
	Floor area ratio	Investment intensity per unit area [54]	Proportion of public facility land in built-up areas	—
	Building density	—	—	—
Output Indicators	Land development rate [54]	Business revenue per unit area [55]	Per capita retail sales of consumer goods	Sewage treatment rate
	Input–output ratio	Tax contribution per unit area	Pension insurance coverage rate [52]	Green coverage rate
	Land profit output rate	Net value added	Medical insurance coverage rate	Energy consumption reduction per unit [56]
	Industrial gross output per unit area	Profit	Unemployment insurance coverage rate	Industrial wastewater discharge per unit area [56]
	—	Growth rate of main business revenue	—	—
	—	Sales revenue of products or services	—	—

below.

The weights of the four dimensions—land use intensity, economic efficiency, social benefits, and ecological benefits—are determined using the Analytic Hierarchy Process (AHP), reflecting their relative importance in the overall evaluation of industrial land use efficiency. First, a pairwise comparison matrix is constructed based on expert judgment and normalized accordingly. Then, the average value of each row in the normalized matrix is calculated to derive the initial weight vector. Results show that intensive efficiency (47.09%) and economic benefits (28.40%) account for the highest proportions, indicating that improving spatial utilization and economic output capacity is key to optimizing land use efficiency. Although social benefits (17.15%) and ecological benefits (7.36%) have relatively lower weights, they nonetheless highlight the importance of social welfare and sustainable development. A consistency check was further performed, and the resulting consistency ratio (CR = 0.019 < 0.1) confirms a high level of logical consistency within the matrix. Based on the relative distribution of composite scores, the parcels are classified into three categories—high efficiency, medium efficiency, and low efficiency—to support subsequent analysis and policy recommendations.

4.3.3. Evaluation Results of Industrial Land Use Efficiency in High-Tech Zones

The evaluation results reveal a clear stratification of industrial land use efficiency within the central urban area.

Low-efficiency land includes 221 parcels, accounting for 61.39% of the total. These are primarily located in traditional manufacturing zones and are characterized by lagging industrial upgrading and low resource utilization efficiency. Among them, some enterprises are in the process of being phased out, while others are in the early stages of transformation but have not yet achieved performance improvement, resulting in generally inefficient land use.

Medium-efficiency land comprises 92 parcels, representing 25.56%. These plots are associated with enterprises that exhibit potential for optimization. With targeted policy interventions and improved resource allocation, these areas are expected to see enhanced land utilization performance.

High-efficiency land consists of 47 parcels, or 13.06% of the total. These are mainly concentrated in high-tech industrial clusters, benefiting from well-established industrial chains and efficient land resource allocation. In addition to early-transitioned high-tech enterprises, some technologically advanced or market-dominant traditional enterprises are also included, demonstrating strong development momentum and land productivity.

GIS analysis reveals that low-efficiency industrial land is widely distributed across the park. (See Figure 6). Overall, land use efficiency in the central area is relatively low, influenced by the following key factors: First, industrial structure aging. Established in 1995, some enterprises have entered maturity or decline phases, with limited R&D investment, making it difficult to align land planning with the needs of emerging industries. Second, a mismatch between dominant industries and high-efficiency land use demands. The area is still dominated by traditional software industries and small-to-medium enterprises, with insufficient integration of high-value-added industries, hindering optimal land resource allocation. Third, rapid industrial technology evolution vs. lagging infrastructure upgrades. The slow pace of infrastructure modernization has led to short-term inefficiencies in land utilization, restricting industrial transformation and high-tech sector expansion.

4.4. Construction and Application of the Early Warning Mechanism for Industrial Land Redevelopment in High-Tech Zones

This study develops an early warning mechanism for industrial land redevelopment in high-tech zones based on Multi-Criteria Decision Analysis (MCDA). It integrates three core dimensions: enterprise life cycle, enterprise-park compatibility, and industrial land use efficiency. By employing the Analytic Hierarchy Process (AHP) for weight assignment and the Weighted Sum Model (WSM) for comprehensive scoring, the model enables precise identification of inefficient land use and informs targeted optimization strategies. Based on the comprehensive score, industrial land is classified into four levels—normal, alert, warning, and response—to facilitate dynamic land use adjustments and enhance the efficiency of industrial land allocation.

4.4.1. Indicator System and Weight Allocation for the Early Warning Mechanism of Industrial Land Redevelopment in High-Tech Zones

(1)

Construction of the Indicator System

The early warning mechanism for industrial land redevelopment in high-tech zones is designed with a focus on systematic, multidimensional, and targeted assessment. A four-tier indicator system is established. First-tier indicator: comprehensive early warning objective. Second-tier indicators: three core dimensions—enterprise life cycle, enterprise–park compatibility, and industrial land use efficiency. Third-tier indicators: sub-categories including enterprise operational efficiency, growth potential, park–city integration objectives, and industrial land efficiency in terms of intensive use, economic contribution, social impact, and ecological sustainability. Fourth-tier indicators: quantifiable metrics, ensuring a structured and actionable evaluation framework. (See

Table 5. Hierarchical structure of the comprehensive evaluation indicator system for the early warning mechanism of industrial land redevelopment in high-tech zones.

First-Tier Indicator	Second-Tier Indicator	Third-Tier Indicator	Fourth-Tier Indicator
Comprehensive Early Warning Indicator System for Industrial Land Redevelopment in High-Tech Zones	Enterprise Life Cycle	Enterprise Operational Efficiency	Input–Output Ratio, Fixed Asset Investment, Unit Area Business Revenue, etc.
		Enterprise Growth Potential	Main Business Growth Rate, Industrial Added Value, Profit, etc.
	Enterprise–Park Compatibility	Park–City Integration	Employee Settlement Ratio, High-Tech Industry Output as a Percentage of Total Industrial Output, Per Capita Disposable Income of Urban Employees
	Industrial Land Use Efficiency	Intensive Utilization	Land Supply Area, Floor Area Ratio, Unit Area Profit Output, etc.
		Economic Benefits	Unit Area Investment Intensity, Unit Area Tax Contribution, Net Value Added, etc.
		Social Benefits	Unit Output Employment Rate, Pension Insurance Coverage, Per Capita Retail Sales, etc.
		Ecological Benefits	Green Coverage Rate, Sewage Treatment Rate, Unit Output Energy Consumption Reduction Rate, etc.

).

(2)

Weight Assignment in the Early Warning Mechanism for Industrial Land Redevelopment Using the AHP Method

To scientifically determine the relative weights of the three secondary indicators—industrial land use efficiency, enterprise life cycle, and enterprise–park compatibility—in the comprehensive early warning mechanism, this study applies the Analytic Hierarchy Process (AHP). AHP quantifies the relative importance of indicators through the construction of pairwise comparison matrices and ensures the rationality of judgments via consistency testing.

Given that this study is situated within the context of industrial land redevelopment under the new–old kinetic energy transition, the weight assignment follows the following logic: First, industrial land use efficiency directly reflects land performance and the urgency for redevelopment, serving as the foundational basis of the early warning system and is thus assigned the highest weight. Second, the enterprise life cycle represents a firm’s long-term utilization potential and development trajectory for industrial land, warranting intermediate importance. Third, enterprise–park compatibility measures the alignment between firm development direction and park strategy. Although it carries relatively lower importance, it remains essential for guiding structural optimization and spatial matching.

Based on expert scoring, the pairwise comparison matrix passed the consistency test (CR = 0.03 < 0.1), and the resulting weights are as follows: enterprise life cycle, 26.05%, enterprise–park compatibility, 10.62%, and industrial land use efficiency, 63.33%.

The final comprehensive score calculation formula is as follows:

S = 0.261M + 0.106E + 0.633C

where

M represents the enterprise life cycle score;

E represents the enterprise–park compatibility score;

C represents the industrial land use efficiency score.

Accurately reflecting the risk levels associated with industrial land redevelopment requires a clear translation of qualitative evaluation results into quantitative scores. Therefore, this study adopts a grade-based assignment method to convert the categorical outcomes of three core dimensions—enterprise life cycle, enterprise–park compatibility, and industrial land use efficiency—into numerical values within the early warning mechanism. Specifically, the enterprise life cycle is divided into four stages—start-up, growth, maturity, and decline—assigned scores from 4 to 1 (start-up = 4, growth = 3, maturity = 2, decline = 1). The other two dimensions are categorized into three levels—high, medium, and low—scored from 3 to 1 (high = 3, medium = 2, low = 1). This method is not meant to reflect the absolute strength of each indicator, but rather their relative representation of redevelopment risk, where lower scores denote higher risk.

Based on the assigned values and weights derived using the Analytic Hierarchy Process (AHP), a composite score is calculated for each land parcel. This enables dynamic monitoring, tiered risk classification, and precision governance, thereby providing a robust quantitative foundation for informed decision-making in the redevelopment of industrial land within high-tech zones.

4.4.2. Design and Implementation of the Comprehensive Early Warning Mechanism

The early warning mechanism for industrial land redevelopment in high-tech zones is constructed based on three core dimensions: industrial land use efficiency, enterprise life cycle, and enterprise–park compatibility. A composite score (S) is calculated by aggregating weighted and normalized indicators using the Weighted Sum Model (WSM), with indicator weights determined through the Analytic Hierarchy Process (AHP). The resulting scores range from 1.000 to 3.261, serving as the basis for classifying industrial land parcels into four early warning levels to guide differentiated management and redevelopment strategies.

To enhance the objectivity and reproducibility of the classification, this study applies a quantile-based segmentation method to determine threshold values. Specifically, the 25th, 50th, and 75th percentiles of the score distribution are 1.628, 1.998, and 2.552, respectively. For interpretability and alignment with conventional risk levels in urban planning practice, the thresholds were adjusted to 1.60, 2.00, and 2.55, closely approximating the empirical quantiles while enhancing practical decision-making utility.

Accordingly, industrial land parcels are categorized into four early warning levels:

Normal (S ≥ 2.55): reflects efficient land use and stable enterprise development; no intervention required.

Alert (2.00 ≤ S < 2.55): indicates early signs of inefficiency or decreasing compatibility; enhanced monitoring is recommended.

Warning (1.60 ≤ S < 2.00): signals significant inefficiency; optimization and policy adjustment are needed.

Response (S <1.60): denotes critical inefficiency or severe enterprise decline; urgent intervention and redevelopment are necessary.

This classification framework integrates statistical rigor with real-world applicability, providing a robust, data-driven tool for dynamic land use monitoring, risk stratification, and precision management in high-tech zone redevelopment. (See

Table 6. Policy responses for early warning levels in industrial land redevelopment.

Early Warning Level	Score Range (S)	Implications	Policy Recommendations
Normal	S ≥ 2.55	Optimal land use and enterprise stability with high compatibility between firms and the industrial park.	Maintain current policies and support innovation.
Alert	2.00 ≤ S < 2.55	Potential efficiency decline or mismatched industries, requiring close monitoring.	Strengthen monitoring and optimize resource allocation.
Warning	1.60 ≤ S < 2.00	Significant inefficiencies, such as low land productivity or declining enterprises.	Adjust policies, promote industrial upgrades, and enhance land use strategies.
Response	S < 1.60	Severe inefficiencies or high-risk enterprises, leading to resource wastage.	Implement urgent interventions, phase out inefficient enterprises, and initiate land redevelopment.

).

The spatial distribution of early warning levels for industrial land use in the central urban area indicates an overall low efficiency, with a significant proportion of land classified as inefficient. Specifically, only 12% (43 parcels) fall into the normal category, primarily located in areas with well-developed industrial structures and strong innovation capacity. In contrast, 28% (100 parcels) are classified as alert-level, signifying that while these plots hold development potential, optimization measures are required. The warning-level category accounts for 52% (189 parcels), highlighting substantial inefficiencies attributed to industrial aging and infrastructure deficiencies. The response-level category comprises 8% (28 parcels), indicating severe inefficiencies that necessitate urgent redevelopment interventions to restore productivity.

Several structural and spatial factors contribute to this distribution pattern. First, long industrial development cycles have resulted in aging industrial structures that no longer align with contemporary industry demands, leading to land use inefficiencies. Second, land scarcity combined with outdated planning has caused a mismatch between high-intensity development patterns and the evolving needs of industrial upgrading. Third, many enterprises have entered the mature or declining phase of their lifecycle, exhibiting reduced efficiency and limited growth potential, thereby impacting land productivity. Lastly, inadequate policy support and suboptimal resource allocation have hindered the introduction of high-end industries and restricted the effective redevelopment of underutilized land.

The spatial visualization of industrial land efficiency through GIS analysis provides further insight into these trends. Normal-level plots are primarily concentrated in functionally optimized areas with strong industrial agglomeration effects, where innovation-driven industries contribute to efficient land use. Alert-level plots are more geographically dispersed, suggesting the need for targeted interventions to enhance land productivity. Warning- and response-level plots, on the other hand, are predominantly located in older industrial zones characterized by low land utilization rates, particularly in early-developed areas that struggle to accommodate high-value industries. These findings emphasize the necessity of tailored land use policies that integrate dynamic monitoring and spatially differentiated redevelopment strategies to enhance land efficiency and support sustainable industrial transformation.

The distribution of early warning levels for industrial land in the central urban area has been visualized using GIS, revealing the spatial patterns of land use efficiency. (See Figure 7).

5. Discussion

Building on the early warning results, an in-depth analysis reveals that the suboptimal land use efficiency observed in more than half of the parcels is not merely a reflection of enterprise-level operational performance, but symptomatic of deeper institutional and systemic barriers. First, the initial policy orientation of “land-driven investment attraction” led to the influx of numerous traditional manufacturing enterprises, many of which lack innovation capacity and sustainable growth potential. This has resulted in a significant mismatch between current land use structures and the strategic shift toward high-end industries, thereby intensifying redevelopment pressures. Second, the prevailing land allocation mechanism lacks dynamic evaluation and tiered access control based on enterprise life cycle stages or industrial compatibility, leading to severe resource misallocation and spatial rigidity. Third, the absence of an integrated coordination mechanism linking industrial upgrading, spatial restructuring, and functional enhancement impedes the synchronized evolution of enterprise growth and land use optimization.

To translate early warning identification outcomes into actionable governance strategies, this study proposes a tiered policy response framework alongside corresponding institutional support mechanisms. For parcels classified as normal, it is recommended to strengthen resource security and support the sustained growth of high-performing enterprises through tax incentives, R&D subsidies, and infrastructure enhancements. For alert-level parcels, real-time monitoring and preventive interventions, such as technical consulting, financial support, and contingency planning, should be implemented. Warning-level parcels require active measures including land remediation, enterprise transformation support, and the reconstruction of public service functions to improve overall productivity. For response-level parcels, compulsory enterprise withdrawal, functional rezoning, and land redevelopment should be pursued, accompanied by compensation mechanisms and reemployment support to ensure smooth and orderly transitions.

At the institutional level, this study further proposes promoting differentiated rental schemes, exploring flexible land tenure mechanisms based on performance evaluation, and establishing a dual-circulation system that integrates internal park coordination with external market flows. Additionally, standardizing exit procedures and streamlining land use reclassification processes are essential to enhance land turnover efficiency. To mitigate social risks associated with enterprise withdrawal, this study recommends implementing vocational training, employment support, and multi-stakeholder coordination mechanisms.

From a broader perspective, this study advocates for a government-guided, market-driven model to accelerate the redevelopment of inefficient land, prioritizing the introduction of high-value, innovation-driven industries aligned with the park’s strategic goals. It further recommends the establishment of a dedicated industrial guidance fund and the construction of a GIS-based dynamic monitoring system to enhance data-informed decision making, improve policy transparency, and support the intelligent and sustainable governance of industrial land [57,58].

Looking ahead, the early warning model for industrial land redevelopment proposed in this study could be further refined along four key directions: (1) incorporating real-time monitoring data and dynamic policy feedback mechanisms to enhance model sensitivity and adaptability; (2) deepening sector-specific analyses to explore how different industry characteristics influence land use efficiency; (3) conducting cross-regional comparative studies to validate the generalizability and scalability of the early warning system; (4) leveraging artificial intelligence and machine learning technologies to develop intelligent early warning tools with stronger predictive capabilities; and (5) applying spatial autocorrelation analysis methods, such as Moran’s I and Local Indicators of Spatial Association (LISA), to uncover spatial clustering patterns of land use efficiency and early warning scores, thereby enhancing the spatial interpretability and precision of governance strategies.

Although this study employs a three-year panel dataset, the overall framework is designed to support rolling updates, approximating dynamic governance, and lays the foundation for future integration with real-time urban management platforms.

It is worth emphasizing that the effectiveness of post-evaluation interventions—such as land withdrawal, functional rezoning, and redevelopment incentives—is significantly shaped by China’s state-owned land system, which provides greater administrative flexibility and policy coordination. Therefore, while the proposed early warning framework possesses strong applicability and transferability, specific policy instruments should be localized to align with different regional land governance regimes.

6. Conclusions

This study makes innovative contributions in three key aspects: theory, methodology, and application. In terms of theoretical innovation, this study breaks through the traditional reliance on single-dimensional efficiency or static characteristics in industrial land management research. It is the first to construct a three-dimensional early warning mechanism that integrates enterprise life cycle, enterprise–park compatibility, and industrial land use efficiency, forming an integrated analytical framework covering “micro-level enterprise dynamics–meso-level spatial compatibility–macro-level performance evaluation”. This provides a new theoretical perspective for optimizing industrial land use in high-tech zones under the context of economic transformation. In terms of methodological innovation, this study comprehensively applies multiple quantitative tools, including Data Envelopment Analysis (DEA), Fixed Effects Model (FEM), Multi-Criteria Decision Analysis (MCDA), Analytic Hierarchy Process (AHP), and GIS spatial analysis, to establish a quantifiable, visualizable, and dynamically updatable early warning and classification governance model, significantly enhancing the accuracy of land use monitoring and the adaptability of policy interventions. In terms of application innovation, this study proposes a systematic pathway of “dynamic monitoring–risk classification–tiered governance”. Based on the composite early warning scores, land parcels are classified into four categories, normal, alert, warning, and response, and a GIS-based visualization system is employed to support spatial decision-making, providing a practical and scalable governance solution for industrial land redevelopment.

The empirical results reveal that the overall industrial land use efficiency in the central urban area remains relatively low, with over 60% of parcels classified under the warning and response levels, primarily concentrated in early-developed areas. The phenomena of inefficient land utilization, industrial aging, and insufficient enterprise growth momentum highlight the structural contradictions encountered during the industrial evolution and spatial renewal of high-tech zones. To address these issues, differentiated strategies—such as optimizing industrial structure, facilitating enterprise transformation and upgrading, and revitalizing underutilized land resources—are required to improve land use efficiency and promote high-quality regional development.

The early warning mechanism developed in this study not only provides a theoretical foundation and practical tool for optimizing the utilization of stock land in the central urban area but also offers a transferable methodological framework for the governance of industrial land in various types of innovation parks. Looking ahead, future research could introduce real-time dynamic monitoring and intelligent algorithms to promote adaptive evolution of the early warning model, deepen cross-regional comparisons and sector-specific analyses, and contribute to building a more resilient and forward-looking industrial spatial governance system in high-tech zones.