Dilute debris flows in hilly regions pose persistent threats to human life and property, as evidenced by catastrophic events in the Laomaoshan area of Dalian City, China, yet their underlying initiation mechanisms remain insufficiently understood. This study systematically investigates the mechanisms through an integrated approach combining physical model experiments, PFC3D numerical simulations, and theoretical mechanical analysis. Results indicate that the initiation is a progressive, gradation-controlled process driven by mesoscopic structural degradation. We propose a novel four-stage hydro-mechanical framework to elucidate this process: (1) Rainfall Infiltration, characterized by rapid matric suction dissipation and the initial weakening of strong contact force chains; (2) Initial Slip, driven by porosity-induced dilation and enhanced localized hydraulic forces at the slope toe; (3) Fissure Development, where fluid-solid velocity disparities exert substantial hydrodynamic drag on the soil skeleton, inducing retrogressive sliding; (4) Overall Mobilization, culminating in the complete collapse of the internal force network and rapid channelized fluidization. Furthermore, mesostructured soil dictates vulnerability, with discontinuously graded soils exhibiting higher susceptibility to sudden, liquefaction-like failure upon saturation. By linking mesoscopic internal yielding with macroscopic hydrodynamic entrainment, this study establishes a physically grounded framework. This coupled weakening-erosion mechanism distinguishes dilute debris flows from traditional viscous debris flows, offering a robust theoretical basis for improved prediction and early warning in analogous geomorphic environments.

Karst groundwater systems are characterized by highly heterogeneous flow networks, resulting in complex responses of groundwater levels to precipitation variability. The Jinan Spring Basin, one of the most representative karst spring systems in northern China, has experienced substantial changes in spring discharge due to variations in precipitation, groundwater exploitation, and hydrogeological conditions. However, the temporal scales at which precipitation signals control groundwater-level fluctuations and the mechanisms governing their transmission within the karst aquifer remain poorly understood. In this study, daily precipitation data from 30 meteorological stations and groundwater-level records at Baotu Spring during 2016–2018 were analyzed using global wavelet spectrum (GWS) and wavelet transform coherence (WTC) approaches. The dominant precipitation cycles, scale-dependent precipitation–groundwater relationships, and phase-derived groundwater response lags were quantified to reveal the hydrological response characteristics of the Jinan karst system. Three prominent precipitation periods were identified at 17.37, 29.22, and 330.57 days. Groundwater responses presented clear temporal-scale dependence, with short-period signals showing rapid but localized responses, while intermediate and long-period signals demonstrated stronger and more persistent coherence. The percentage of significant coherence area (PASC) increased from 23.50% at the 0–17.37 day scale to 64.28% at the 29.22–330.57 day scale, indicating that accumulated precipitation rather than individual rainfall events exerts the dominant control on groundwater-level variations. The spatial distribution of response lags revealed that rapid responses (17.37 days) mainly occurred in the southern recharge areas, reflecting preferential recharge through well-developed karst conduits. Intermediate responses (29.22 days) showed a progressive increase in lag time from south to north, indicating the influence of regional groundwater flow and aquifer storage. Long-period responses (330.57 days) were locally enhanced near major faults, suggesting structural controls on deeper groundwater circulation. This study reveals that precipitation signals in the Jinan Spring Basin are transmitted through multiple groundwater circulation pathways with distinct temporal characteristics. The identified multi-scale response patterns provide new insights into the internal structure and hydrological functioning of karst aquifers and offer scientific support for sustainable management of spring water resources.

This study employs field monitoring, hydrochemical analysis, lab experiments, and tracer tests to characterize multilevel groundwater flow systems and circulation patterns in karst aquifers, establishing a classification framework for spring protection. The main findings are as follows: (1) The Jinan area comprises two first-order karst water systems, specifically the Jinan Monocline Karst Water System and the Laiwu Basin Karst Water System. The Jinan Monocline system is further subdivided into five second-order intermediate flow systems. (2) The multilevel karst water circulation patterns are primarily classified into three major categories. The unconfined local flow system includes three circulation patterns. The confined intermediate flow system comprises three patterns, with the window-type intermediate system further divided into three subtypes. (3) The local flow system performs water conservation functions, serving as an indirect recharge source for the four major spring groups. The intermediate flow system constitutes the primary direct recharge source, mandating rigorous protection of the Cambrian-Ordovician karst aquifer in its recharge area. The confined regional system with deep circulation must adhere to the 'heat extraction without water withdrawal' development principle, preventing geothermal development activities from impacting the sustainable discharge of the spring clusters.

ABSTRACT Workflow of a spatio-temporal graph convolutional network predicting karst groundwater levels, with a Random Forest ranking driving factors. Accurate groundwater level prediction in karst aquifer systems remains challenging due to their complex hydrogeological characteristics, including rapid recharge responses, preferential flow paths, and strong spatial heterogeneity. This study develops a spatio-temporal graph convolutional network (STGCN) framework specifically designed for the karst spring system in Jinan, China. The framework treats the monitoring network as an interconnected graph, with well-to-well relationships encoded through a hydraulically-informed adjacency matrix. Multiple environmental and anthropogenic driving factors – precipitation, ecological water replenishment, sea-level pressure, pumping, air temperature, and wind speed – are integrated to comprehensively capture karst groundwater dynamics. Three years of daily monitoring data (2021–2023) were utilized for model development and validation. The proposed model achieved a root mean squared error of 0.2077 m, reducing prediction errors by approximately 40% compared to conventional approaches. Feature importance analysis revealed ecological water replenishment (46.1%) and precipitation (37.4%) as dominant drivers of groundwater variations. The model successfully captured both long-term seasonal trends and rapid short-term responses to rainfall events characteristic of karst systems. This framework provides an interpretable and accurate tool for groundwater management in complex karst terrains, supporting sustainable water resource decisions in urbanized karst regions.

Groundwater is a critical freshwater resource in Weihai, a coastal city in North China, yet its hydrochemical evolution and associated pollution risks under the combined influence of seawater intrusion and human activities remain insufficiently understood. This study established an integrated framework to characterize groundwater chemistry, quantify source contributions, evaluate water quality, and assess non-carcinogenic health risks. A total of 72 shallow groundwater samples collected during the dry season were analyzed using Piper and Gibbs diagrams, ion-ratio analysis, principal component analysis coupled with the absolute principal component score-multiple linear regression model, the entropy-weighted water quality index, and human health risk assessment. Groundwater was classified into four major hydrochemical facies: HCO₃-Ca·Mg, SO₄-Ca·Mg, Cl-Na, and Cl·SO₄-Ca·Mg. Multiple lines of evidence indicated that marine-related salinization is a major control on groundwater chemistry, particularly in southern Wendeng and coastal Rongcheng. Source apportionment identified five principal contributors: carbonate dissolution (27.7

To clarify the distribution pattern of the fault-karst groundwater flow system in the Baotu Spring Basin, Shandong Province, China, and to reveal the controlling mechanisms of groundwater flow and the response characteristics of solute transport, this study focused on the Cambrian-Ordovician carbonate rock distribution area within the basin. Karst water, pore water, and surface water samples were collected, and a comprehensive research approach was employed, integrating hydrochemical analysis, statistical analysis, and numerical simulation based on FEFLOW. The results indicate that calcium ions (Ca²⁺) are the dominant cations in the water bodies of the study area, with bicarbonate ions (HCO₃⁻) and sulfate ions (SO₄²⁻) as the primary anions. The average total dissolved solids (TDS) concentrations of karst water, pore water, and surface water are 570.73 mg/L, 746.96 mg/L, and 507.75 mg/L, respectively. The low coefficient of variation in ion concentrations reflects the overall stability of water quality across the study area. Hydrochemical characteristics are predominantly governed by rock weathering, with negligible influence from evaporation-crystallization processes, while atmospheric precipitation exerts a dilution effect. Calcite and dolomite are in a supersaturated state, whereas gypsum and halite are prone to dissolution. Verification results of the groundwater flow numerical model demonstrate that the selected hydrogeological parameters and source-sink terms are reasonable, enabling accurate characterization of the overall groundwater flow field, which flows from southeast to northwest in the study area. Fault zones significantly complicate groundwater flow paths: the Qianfoshan Fault induces flow-line deflection, while the Chaomidian Fault triggers flow-line convergence due to its high permeability. Using nitrate ions (NO₃⁻) as a tracer, this study confirms that fault permeability coefficient and time exert a significant regulatory effect on solute transport and pollution plume distribution. Under the dominance of fault-controlled permeability mechanisms, the pollution plume’s influence range expands dynamically over time as long as a hydraulic head difference persists. This study systematically clarifies the key evolutionary processes and core controlling factors of the fault-karst groundwater flow system in the Baotu Spring Basin, further enriches the theoretical system of hydrogeological research in karst areas of northern China, and provides reliable theoretical support and practical basis for the refined management of regional water resources, ecological protection, and sustainable development.

Karst aquifers exhibit complex groundwater-precipitation interactions shaped by climatic variability, internal structural heterogeneity and human disturbance. However, the mechanism by which climate-driven recharge signals are transformed by karst flow architecture and simultaneously modulated by anthropogenic pressure across multiple timescales remains unclear. Herein, we investigated the coupled influence of climate variability and human disturbances on groundwater dynamics in a typical heterogeneous, fracture-dominated karst aquifer in northern China. Three process-oriented insights emerged: (1) Human disturbance (particularly spring irrigation) had a prominent influence on seasonal groundwater decline and exceeded climatic forcing during critical periods. (2) Groundwater levels (GWLs) exhibited rapid (similar to 10 days), seasonal (similar to 40-80 days) and quasi-annual (similar to 320 days) variability in response to precipitation (P), reflecting the interaction between fast preferential pathways and slow storage components. (3) Strong spatial heterogeneity was evident; urbanised discharge zones showed markedly reduced recharge transmission owing to surface sealing, whereas permeable recharge areas remained highly responsive to P inputs. Flow partitioning indicated a pronounced dual-flow structure, with recharge zones exhibiting mixed behaviour and quick flow contributions of 31.2%-36.2%, whereas discharge zones were strongly dominated by slow flow (similar to 87.4%), highlighting the buffering role of aquifer storage in smoothing climatic signals. Overall, these findings advance the process-based understanding of how karst aquifers filter, store and transmit P signals under coupled climate-human forcing. These insights are transferable to other karst regions and support the diagnosis of the dominant factors controlling groundwater dynamics under global environmental change.

Facility agriculture is a modern intensive cultivation method that is widely seen as the future of global agriculture. However, large-scale emissions of concentrated pollutants during production pose serious threats to groundwater quality. Identifying the sources of pollutants and assessing source-specific risks are critical for developing effective risk mitigation strategies. In this study, a combination of methodologies including Self-Organizing Maps (SOM), K-means clustering, factor analysis, and ion ratio analysis were utilized to investigate pollution risks in a typical facility agriculture area in Shouguang City, Shandong Province, China. The groundwater quality in the study area is poor and slightly alkaline, with NO3− being the main pollutant. The chemical composition of groundwater in the aquifer is influenced by both human activities (41.89

Groundwater is a crucial water source and strategic resource, essential for sustaining both urban and rural livelihoods, supporting economic and social development, and maintaining ecological balance. This study investigates the hydrochemical properties and controlling factors of groundwater in the Baiquan basin (BQB) by analyzing water quality data collected during both dry and wet periods. Additionally, the suitability of groundwater for drinking and agricultural irrigation was evaluated. The findings reveal that groundwater in BQB is generally weakly alkaline and primarily consists of hard-fresh water. Although there are seasonal variations in the main ion concentrations, HCO3− and Ca2+ are the predominant anions and cations, respectively. Consequently, the hydrochemical type is mainly HCO3-Ca⋅Mg type, with a secondary classification of SO4⋅Cl-Ca ⋅ Mg. The hydrochemical composition is primarily influenced by the dissolution of carbonate and silicate minerals, as well as cation exchange processes. Additionally, it is affected by anthropogenic inputs, particularly from the use of agricultural fertilizers. The water quality assessment results indicated that all water samples are classified as either good or moderate, with a significant majority falling into the good category. Additionally, the northern section of the BQB exhibited lower entropy weight water quality index (EWQI) values during the dry season in comparison to the wet season. For irrigated agriculture, groundwater in the BQB serves as a high-quality water source for irrigation throughout both the dry and rainy seasons. Furthermore, non-carcinogenic risks are notably concentrated in the north-western and south-eastern regions of the study area. Health risks associated with nitrates in groundwater are elevated during the rainy season. Notably, non-carcinogenic risks for infants were significantly high across both seasons and substantially exceeded those for children and adults. These results provide valuable scientific insights for the management and development of groundwater resources in the BQB.

The formation of heterogeneous pore structures is primarily influenced by the depositional environment and diagenetic processes. The inhomogeneity of carbonate media leads to the development of microfractures and dissolution pore structures, which significantly contribute to the permeability of the aquifer. In this study, various theories and methods, including geological, hydrodynamic, fractal geometry, threshold segmentation, three-dimensional reconstruction, and pore network extraction methods are integrated. Focusing on the karst Baotu Spring area in northern China, representative carbonate rock samples from the Jinan Spring area were selected. Using computed tomography (CT) scanning technology, these core samples were digitally reconstructed, allowing for an in-depth analysis of the spatial structural characteristics of the pores, fractures, and dissolution pores. Additionally, on the basis of the Navier-Stokes equations, the fluid flow process within the pore structure of carbonate rocks was simulated, and the effective permeability was determined. In this study, the impact of pore characteristic parameters on permeability were explored. The results reveal that the total porosities of the carbonate rocks range from 0.89% to 10.35%, with effective porosities varying between 0.6% and 6.12%. The surface porosities of the samples exhibit significant variability, with the pore structure showing strong inhomogeneity and high permeability. Correlation analyses of key parameterssuch as the porosity, fractal dimension, equivalent diameter, and aspect ratioindicate that the pore shape, structural complexity, and size are crucial factors affecting permeability. Porosity is positively correlated with permeability, with variations in permeability being smaller at lower porosities. The fractal dimension has a nonlinear relationship with porosity: at low porosities, the pore structure is simpler and poorly connected, whereas at higher porosities, the pore count increases, the structure becomes more complex, and the connectivity improves, thereby increasing permeability. Comprehensive analysis suggests that the geometric complexity and structural inhomogeneity of pores are key factors controlling the permeability of carbonate rocks.

【Objective】 Accurate prediction of water quality is of great importance for protecting water ecological environment and improving water resource management, but difficult due to the combined influence of various factors including human activities, rainfall, temperature and hydrodynamic conditions which are complex and uncertain. In this paper we compare three models, the SARIMA model, Holt-Winters model and LSTM neural network model, for predicting water quality of reservoir. 【Method】 Electric conductivity of water was used as a proxy for water quality. Data measured monthly from 2012—2018 from Mishan Reservoir was used for model training, and the data measured in 2019 was used to test the models. 【Result】 The SARIMA and Holt-Winters models are comparatively simple, but because they only consider time series of water quality data, their accuracy is low. In contrast, the LSTM neural network model considers factors that affect water quality and implicitly represents the nonlinearity of the factors in their impact on water quality, it is more accurate than other two models. 【Conclusion】 In general, the LSTM neural network model is reliably, giving rise to large error only when there were sudden changes in the conductivity. Overall, it is reliable and accurate for predicting change in water quality induced by variations in the environment.

The coastal area is a major area of socio-economic development and the most active zone for human activities. With the rapid development of the economy and the strengthening of urban construction, the groundwater environmental problems in coastal areas are increasingly prominent. It is significant to investigate the groundwater chemical characteristics, hydrochemical types, and the factors that influence groundwater chemistry for water resources protection and groundwater resources development. In this paper, 48 groundwater samples and 5 surface water samples from the study area were analyzed with statistical analysis, Piper diagram, Gibbs diagram, mineral saturation index method, and the ionic scale factor, and explored the factors that influence groundwater chemistry. The modified Nemerow index method was also applied to evaluate the groundwater. The results show that the groundwater in the study area is neutral to weakly alkaline (average pH = 7.0~8.0). The groundwater chemical types are mainly Cl·SO4-Na and SO4·Cl-Ca·Mg. Hydrochemistry is mainly influenced by rock weathering and evaporative concentration. TDS was strongly correlated with TDS, Na+, Mg2+, K+, Ca2+, Cl−, SO42−, and the saturation index showed a gradual increase along the groundwater drainage flow path, it indicates that the main groundwater ions originate from the dissolution of halite, sulfate, and carbonate. Combining GIS technology and the kriging spatial interpolation method, we obtained the current situation map of groundwater quality in Laoshan District. The groundwater quality is mainly III water, and the overall water quality is good; IV and V water is mainly distributed in the middle and lower reaches of rivers, distributed in a belt pattern along the coastline. It is mainly influenced by both the human factor and seawater intrusion. It is significant for the utilization of groundwater resources and the management of seawater intrusion in the Laoshan District. In addition, the proposed research ideas and methods provide a reference for the study of groundwater genesis in other coastal areas in the world.

In order to protect the Qilitang geothermal hot spring in Weihai City, this paper discusses the elements of the Qilitang geothermal field, such as water source, heat source, geothermal field, heat accumulation model, and water and heat migration channel, through the methods of geothermal geological survey, geophysical exploration and geochemical exploration, and establishes the genetic conceptual model of the geothermal field and reveals its genetic mechanism. The research shows that: ① The chemical types of geothermal water are mainly SO4·HCO3−Ca·Na and HCO3−Na·Ca type water, and the water quality dynamics are relatively stable. ② The main source of hot water recharge is meteoric water, which circulates to about 2276 m underground along the Hengkou−Yanggezhuang deep fracture and is heated to about 114.39 ℃. At the intersection of the faults, springs emerge along the fracture zone. ③ The infiltrated groundwater continuously absorbs heat from surrounding rocks during its deep migration along the fault tectonic belt. The genetic type of hot spring in geothermal field is deep circulation−convection type. ④ Jiaodong hot spring geothermal fields such as Qilitang are controlled by faults. The area of geothermal anomaly is small and the scale of geothermal field is small. Although Jiaodong hot spring geothermal has good market prospects for development, exploitation must be controlled to avoid resource exhaustion and temperature drop of hot water caused by excessive exploitation. The research results have certain guiding significance for the development and utilization of geothermal resources in Weihai area.

准确掌握地下水的环境质量状况是合理确定地下水资源开发策略和有效进行地下水资源保护的重要前提.通过随机森林(random forest)法构建弥河-潍河流域地下水质量评价模型,结果表明:(1)随机森林法在进行地下水水质分类时具有分类精度高、泛化能力强等特点,且在进行超参数优化后,其分类精度会进一步提高,证明将随机森林法应用于地下水质量评价是可行的,并且其综合性能要优于逻辑回归模型;(2)研究区地下水水样均为Ⅳ类和Ⅴ类水,说明水质状况整体较差;(3)通过分类指标重要性评价可以看出,研究区地下水水质的主要影响指标为硝酸盐、总硬度和溶解性总固体,而此类指标的主要来源是蔬菜种植化肥的不合理使用及河流污染入渗,因此要进一步加强对蔬菜种植污染排放及河流水质的监测和控制.

This study enhances the understanding of the evolution of water transport in karst water flow systems. This paper explores the karst water flow system from a hydrogeochemical perspective. It has been discovered that in the recharge area, the hydrochemical effect of karst water is primarily influenced by dissolution and filtration. In the fault zone and concentrated discharge area, it is primarily influenced by mixing. In the geothermal area, the dominant factors are dissolution, precipitation, and dolomitization. From the southern to the northern area, the hydrochemical type of karst water gradually changes from HCO3-Ca type to SO4·Cl-Ca·Na type. Atmospheric precipitation is the primary recharge source of karst water, with a recharge elevation range of 181–1,495 m. According to geological drilling statistics, the groundwater transport depth transits from the shallow transport in the south to the deep transport in the north, with a depth range of 3,000–3,500 m, and using isotope data to obtain the groundwater retention time range is 6–27,000 years. The groundwater transport mode is divided into three levels: open-shallow transport karst cold water phreatic flow, semi-open-medium-deep transport karst cold water confined flow, and weak open-deep transport geothermal water confined water flow.

以济南市马武寨山为试验地点,采用11个指标,通过主成分分析确定影响研究区水土环境的3个主要成分为土壤生态因子、微地形因子、土壤养分因子,利用3个主成分对不同植被修复模式下露天采石矿废弃地及自然山体的水土环境进行研究,分析生态修复效果及存在的问题.研究表明:(1)土层厚度、枯枝落叶层厚度、NDVI和植被类型是土壤生态因子的主要影响因素;微地形因子中坡度是决定土壤侵蚀强度的主要因素;土壤侵蚀造成了一定程度的养分流失.(2)露天采石矿废弃地生态修复中,可通过种植多种乔木、增加土层厚度及土壤中黏粒含量等措施改善区域土壤生态环境,降低坡度减弱土壤侵蚀强度.(3)植被人工修复区域的生态环境得到了改善,植被与土壤的良性循环演替已初步形成.

Due to the influences of geological structures, many confined ascending springs occur in North China's karst area. The electrical conductivity (EC) of karst spring flow has been a fundamental variable in characterizing karst systems. However, deeply exploring the hidden nonlinear dynamic characteristics is challenging. To avoid overreliance on the fitting polynomial order in the detrending process by classic multifractal detrended fluctuation analysis (MFDFA), the intrinsic time-scale decomposition (ITD) method was applied to identify the external trend term of the data by decomposing the original data into different frequency modes, and the ITD-MFFA method was proposed to reveal the formation mechanism of spring EC complex characteristics in the Jinan spring area. The results showed that the EC sequences of karst springs in North China are characterized by multifractal behavior and antipersistence and display multiyear complexity with an overall decreasing trend. Different recharge sources, formation conditions, and seasonal precipitation might be the primary factors driving the spring's EC complexity. Compared with the traditional MFDFA method, the ITD-MFFA method improves the anti-interference ability and stability. The timely determination of the dynamic complexity of spring EC data and monitoring the future trends of the spring water quality have guiding significance for protecting karst springs.

Due to the multi-scale nature and complexity of the flow field, it has always been difficult for hydrogeologists to accurately grasp the water cycle characteristics of the karst water flow system. The ISM model and IsoSource model were employed to calculate the spring age and the mixing ratio of spring recharge sources of the four large spring groups in Jinan. The karst hierarchical groundwater flow system was identified based on multiple factors. We found that the groundwater ages of Baotu Spring and Heihu Spring are about 6a-10a, and that of Wulongtan Spring and Pearl Spring are about 30a-35a. The main recharge sources of spring water of the four large spring groups were recognized as artificial recharge water, Ordovician karst water, and Cambrian Zhangxia Formation karst water. There were distinct differences in the recharge proportion of the four large spring groups. Three groundwater flow systems were identified according to multiple factors, which were local, intermediate, and regional. Furthermore, different levels of groundwater flow systems were recharged to the four large spring groups in different proportions. The research results provided new insights into the analysis of the hierarchical cycle and evolution of karst water.

The objective of this study was to predict the dynamic change in the spring water level more precisely, to provide timely solutions for karst spring protection. Using the Jinan spring region as a case study, this study established a numerical model of a karst groundwater system, and optimized the mining layout. The calculated maximum extraction volume following the optimized exploitation layout was 0.69 m3/s, in order to ensure the continuous flow of spring water in the median water year. A coupled karst groundwater numerical model with dual structure was developed using the MODFLOW-Conduit Flow Process (CFP), which simulates and then precisely predicts changes in the water level of the karst springs. Here, the plane extension direction of the karst conduit was determined by a tracer test and correlation analysis of the spring water levels and groundwater levels of the observation wells. Meanwhile, the vertical location of the karst conduit was determined by layered monitoring of the groundwater temperature and conductivity. Based on this, a coupling model of seepage and conduit flow was created to simulate the dynamic change in the spring water level, and the dual-media coupling model improved the simulation accuracy of the spring water level. The current study confirmed that, compared to the porous media seepage model, the dual-media coupling model can simulate the groundwater level dynamic change more accurately in a heterogeneous karst aquifer in northern China. The coupling model was used to analyze the effect of supplementation and optimize mining, to ensure that spring water continues to flow during the dry season while supplying the mining demand.

Engineering construction changes the groundwater flow field and endanger the safety of buildings. Taking the large-scale underground complex along Jinan Jingshi Road as an example, the impact of engineering construction on the groundwater seepage field is obtained by numerical simulation, and the groundwater flow field repair models are established. These results show that the underground spatial structure will block the normal movement of groundwater, and will reduce the bearing capacity of the foundation after the groundwater level is raised. After adding diversion measures to the underground space, the water level at the upstream surface will decrease with time, and the closer it is to the natural state, the slower the water level attenuation rate is. Due to the difference in the stratum structure, there is a great difference in the time when the backwater level basically drops. The number of diversion wells required to repair the flow field under different geological conditions has a negative correlation with the structural parameters of diversion geometry. The established equation for predicting the number of diversion wells fully reflects the complexity and variability of geological conditions in karst areas. In addition to the structure of the fluid conducting geometry, the hydraulic gradient and permeability coefficient of the surrounding rock also affect the water conducting capacity of diversion measures. The permeability coefficient controls the speed of the water conducting rate, while the hydraulic gradient controls the occurrence of water conducting behavior. The implementation of diversion measures can reduce the impact of engineering construction on the groundwater environment and ensure that the impact of underground engineering construction on the water environment is controllable.