Introduction

The Tanlu fault zone in eastern China has been an active seismic zone since it formed in the Mesoproterozoic Era (Fig. 1). The fault zone stretches more than 2400 km in the territory of China and cuts through the eastern part of different tectonic units (Zhu et al. 1995; Gilder et al. 1999; Li et al. 2011). The scale of the fault is grand, and the structure is complex. The area is not only a main fault zone and a source zone for deep magma among a series of NNE-trending giant fracture systems in the East Asian continent (Chang 1996; Chen and Nábelek 1988; Shi and Han 2005; Chen et al. 2006; Lv et al. 2011) but is also the most important concentration area of gold deposits (Deng 1992; Lv et al. 2006; Wan 1993; Yang and Lv 1996; Wang et al. 2002) and other mineral resources (Fig. 1) (Zhai 1997; Xie et al. 2003) in China. There are three important metallogenic belts in the northwest of Jiaodong: the Sanshan Island ore control tectonic belt, Longkou-Laizhou ore control tectonic belt and Zhaoyuan-Pingdu ore control tectonic belt (He et al. 2015; Li et al. 2015).

Fig. 1
figure 1

Tanlu fault zone

Full size image

Jiaodong refers to the eastern part of the Jiaolai River in the Shandong Province of China, including Yantai, Weihai and the eastern part of Qingdao and is a major part of the Shandong Peninsula. These three metallogenic belts are all important branches of the Tanlu fault zone. The Jiaojia Gold Mine area is located at the southern end of the Longkou-Laizhou ore control tectonic belt (Zhang et al. 2012; Zhang 2012). Its area is less than 0.1% of the land area of China, but it has approximately 20% of the gold reserves and output of China, so its economic status is very important. There are many gold mines distributed in the Jiaojia Gold Mine area, such as the Xujiatong Gold Mine, Shangzhuang Gold Mine, Xincheng Gold Mine, Hedong Gold Mine, Longbu Gold Mine, Qiansunjia Gold Mine, Wangershan Gold Mine, Jiaojia Gold Mine and Sizhuang Gold Mine, in which the Xincheng Gold Mine, Wangershan Gold Mine, Jiaojia Gold Mine, and Sizhuang Gold Mine are the most important mines.

In recent years, with the increase in the mining depth, hydrogeological conditions have become more complex, and the water inflow of mines has increased annually, and mine water disasters often occur, resulting in economic losses amounting to tens of millions of dollars (Shi et al. 2012; Li et al. 2017; Shi et al. 2017; Qiu et al. 2017a, b; Krzysztof et al. 2016; Zhang et al. 2014). In a detailed geological survey report of the Jiaojia Gold Mine area, the method of prediction of mine inflow was the traditional large-well method and analogy method (Zhu 2012). However, production practice shows that the error between the predicted value and the actual value of the two methods is large. In this paper, based on borehole data, hydrogeological data, and mine water inflow data, the GMS is used to simulate the groundwater in the Jiaojia Gold Mine area to predict the water inflow and its dynamic change in the main gold mines and to provide a decision basis for the layout and safe mining of the drainage system of the mines. (Alaciani et al. 2018; Mirmahdi et al. 2019; James and Reed 2017; Yin et al. 2016; Soni and Manwatkar 2015; Niedbalska et al. 2014).

Study area

Most of the Jiaojia gold mine area contains exposed bedrock areas, and few are covered by Quaternary strata. According to the occurrence conditions of groundwater and hydraulic characteristics, the aquifers in the study area are divided into three major types: a Quaternary pore aquifer, bedrock weathered fractured aquifer, and bedrock structure fracture aquifer. The Quaternary pore aquifer is mainly distributed along the coast and on both sides of the rivers, and is mainly composed of sandy clay, various kinds of sands and gravels with local thin silty clay layers. Its water-bearing properties range from moderate to rich. The bedrock weathered fractured aquifer is located below the Quaternary pore aquifer and the southern exposed bedrock areas, is distributed across the entire study area and is composed of metagabbro, phyllic cataclastic rock, and monzonitic granite with poor to moderate water-bearing properties. The bedrock structure fracture aquifer is located underneath the bedrock weathered fractured aquifer, is distributed throughout the entire study area and is composed of mylonite, breccia, cataclastic rock, lamprophyre, and diorite porphyrite with poor to moderate water-bearing properties. At the bottom of the bedrock structure fracture aquifer is the water-resisting rock mass of monzonitic granite with a thickness of more than 1000 m.

In Jiaojia gold mine area, the north-east compressive torsional structures are developed (Fig. 2), and the Jiaojia fault zone is the main fault controlling the formation of gold mineralization, which is more than 20 km long and 80–200 m wide. The area developed a continuous main fracture surface marked by gray-black fault gouge and a complete set of fault-rock zones. Ore bodies mainly occur in tectonic altered rocks with a high alteration degree under the main fracture surface. The main gold mines in the Jiaojia Gold Mine area distributed along the fault zone are the Jiaojia Gold Mine, Wangershan Gold Mine, Sizhuang Gold Mine, and Xincheng Gold Mine, which have been mined to − 630, − 630, − 400, and − 680 m, respectively. The mine inflow is up to approximately 5800–9077, approximately 12,000–14,000, approximately 4800–5200, and approximately 1000–1200 m3/d, respectively.

Fig. 2
figure 2

Hydrogeological characteristics within the range of numerical simulation

Full size image

Method

In this study, the finite difference MODFLOW module simulation program based on GMS 7.1 software was selected to simulate the groundwater system in the study area. GMS 7.1 is an environmental simulation software approved by US Geological Survey (USGS) and Environmental Protection Agency (EPA). It is currently the most comprehensive groundwater modeling software and has been proved to be the most effective groundwater modeling system. MODFLOW is a grid-based finite difference method for 3D numerical simulation of groundwater flow in porous media. The water balance equation of each grid in the simulated area is established using space division time discretization. All grid equations are combined into a set of large linear equations, and the iterative solution of the equations can result in the head value of each grid, etc.

The mathematical model of groundwater is a set of mathematical relations describing the quantity, space, and time of the actual groundwater flow. The model has the ability to duplicate and represent the actual movement state of groundwater flow and is usually composed of partial differential equations and their definite conditions. According to the hydrogeological conditions of the simulated area, by studying the characteristics of groundwater recharge, discharge, and dynamic change, a three-dimensional unsteady groundwater flow system with heterogeneous isotropy and double-layer structure was established, which is described by the following three-dimensional fluid mathematical model.

$$ \Big\{{\displaystyle \begin{array}{l}\frac{\partial }{\mathrm{\partial x}}\left({K}_{xx}\frac{\partial h}{\partial x}\right)+\frac{\partial }{\mathrm{\partial y}}\left({K}_{yy}\frac{\partial h}{\partial y}\right)+\frac{\partial }{\mathrm{\partial z}}\left({K}_{zz}\frac{\partial h}{\partial z}\right)+\varepsilon +p=0\kern0.5em x,y,z\in \varOmega \\ {}h\left(x,y,z\right)=h\kern1.50em \begin{array}{cccc}& & & \end{array}\begin{array}{ccc}& & \end{array}\begin{array}{ccc}& & \begin{array}{cc}& x,y,z\in \varOmega \end{array}\end{array}\\ {}{\left.{K}_n\frac{\partial h}{\partial x}\right|}_{\varGamma }=q\begin{array}{cccc}& & & \end{array}\begin{array}{cccc}& & & \end{array}\begin{array}{ccc}& & \end{array}\begin{array}{ccc}& & \begin{array}{cc}& x,y,z\in \varOmega \end{array}\end{array}\end{array}} $$
(1)

where Ω is the seepage area; Kxx, Kyy, and Kzz are the permeability coefficients in the x, y, and z directions, respectively (m/d); Kn is the permeability coefficient in the normal direction of the boundary surface (m/d); ε is the source or sink term of the aquifer (1/d); p is the intensity of precipitation infiltration and water table evaporation (m/d); h is the groundwater level distribution of the aquifer (m); Γ is the lateral flow boundary of the seepage area; and n is the normal direction of the boundary surface.

The specific simulation steps are as follows:

  1. (1)

    Determination of the simulation scope and boundary conditions of the model.

  2. (2)

    Generalization of the vertical structure of the aquifer, and hydraulic properties of the fault zones.

  3. (3)

    Division of space, discretization of time and determination of the initial values.

  4. (4)

    Partition of the hydrogeological parameters.

  5. (5)

    Disposal of sources and sinks and selection of software packages.

  6. (6)

    Model identification and verification.

  7. (7)

    Prediction of mine water inflow based on the model.

Establishment of the conceptual model

Boundary conditions

The numerical simulation boundary is a range drawn up by referring to the hydrogeological boundary, which has little or no hydraulic connection with the outside world. The establishment of simulated boundary is conducive to the input of source and sink items and the study of flow law of groundwater. The northern boundary of the numerical simulation is the coastline, which is set as a constant water head boundary. The southern boundary is a natural watershed stretching from Mount Mi to Jinhua Mountain, which is set as a flux boundary. The east boundary and the west boundary are the Zhuliu River and Zhuqiao River, respectively, which are set as constant water head boundaries (Fig. 2).

Generalization of the fault zones

In the central part of the simulation area, northeast to southwest faults are developed, primarily including the Jiaojia fault zone and Wangershan fault zone (Fig. 3). The central part of the fracture, composed of fault gouge, mylonite, breccia, and cataclastic rock, belongs to the tight zone and water-resisting zone (Fig. 4). The barrier package included in the MODFLOW module of the map module in GMS can be used to simulate the barrier boundary conditions of thin layered, vertical, and low permeability objects in the underground system to impede the horizontal flow of groundwater. Therefore, the barrier package is set as the fault boundary due to the large dip of the Jiaojia fault zone and Wangershan fault zone.

Fig. 3
figure 3

Profile of line A-A'

Full size image
Fig. 4
figure 4

Schematic diagram of the fractures controlling the hydrogeological zoning

Full size image

Jiaojia fault zone

The length of Jiaojia fault zone is more than 20 km, the width is 100–200 m, the average strike is 40°, the tendency is northwest, the dip angle is 20–30° with local steeper and up to 70°.

The Jiaojia fault zone is a wide crush zone composed of cataclastic rocks, granitic cataclastic rocks, and there is gray-black and gray-white fault gouge with 10–20 cm thickness, continuous stability, local occurrence of small folds along the fault zone, which shows the properties of the compression torsion fracture. On the other hand, there are different thickness and discontinuous mylonite, breccia also along the fault zone. The cataclastic rocks and granitic cataclastic rocks are bended and symmetrical distributed in the hanging wall and footwall of the main fault, but the gold ores usually occur in the footwall of the main fault, where tectonic rock of the bonanza usually produces alteration and mineralization because of the hydrothermal liquid mineralization and finally forms the mineralization alteration tectonic zone.

Wangershan fault zone

The length of Wangershan fault zone is 13 km, the width is 20–120 m, the average strike is 40°, the tendency is northwest and the dip angle is from 45 to 60°. The characteristics of the tectonic rock in Wangershan fault zone are consistent with that in Jiaojia fault zone, only with slightly different in developing degree and scale. Wangershan fault zone is also a compression torsion fault and belongs to a secondary fault structure derived from Jiaojia fault zone.

Generalization of the vertical structure of the aquifer

Due to the similarity of the hydraulic characteristics between the weathered fractured aquifer and the bedrock structure fracture aquifer, the aquifer can be generalized as a bedrock fractured aquifer. In the vertical generalization, the Quaternary pore aquifer and bedrock fractured aquifer constitute the generalized aquifers for numerical simulation. There is a hydraulic connection between them. The top boundary of the model is the Earth’s surface, and the bottom boundary is monzonitic granite, which belongs to the water-resisting rock mass. The thickness of each aquifer is obtained according to the elevation of the ground (Fig. 5) and the bottom elevation of each aquifer (Figs. 6 and 7).

Fig. 5
figure 5

Contour map of the ground elevation

Full size image
Fig. 6
figure 6

Bottom elevation of the Quaternary pore aquifer

Full size image
Fig. 7
figure 7

Bottom elevation of the bedrock fractured aquifer

Full size image

The ground elevation of the Jiaojia Gold Mine area is high in the southeast and low in the northwest, with elevations ranging from 0 to 180 m (Fig. 5). The bottom elevation of the Quaternary pore aquifer is basically consistent with the ground trend, and the elevation range is approximately − 20–85 m (Fig. 6). The bottom elevation of the bedrock fracture aquifer increases successively from the faults to either end, and the elevation range is approximately − 1400–500 m (Fig. 7).

Establishment of the numerical model

Space division

We utilize a comprehensive hydrogeological map with a scale of 1:50000 in the Jiaojia Gold Mine area as the numerical simulation area map, and the simulated area is 112 km2 in size, with a north-south length of 18 km, and an east-west width of 14 km. According to the size of the study area and the requirement of calculation accuracy, square elements are used to segment the region, that is, the study area is divided into 60 lines and 60 columns along the plane. There are 642 effective grids in the Quaternary pore aquifer in the first layer (Fig. 8) and 1641 effective grids in the bedrock fractured aquifer in the second layer (Fig. 9), for a total of 2283 effective grids.

Fig. 8
figure 8

Plane subdivision diagram of the Quaternary pore aquifer

Full size image
Fig. 9
figure 9

Plane subdivision diagram of the bedrock fractured aquifer

Full size image

The front and side view of the model can be obtained by importing the elevation of the top and bottom into 2D Scatter Data in the form of a txt-file and then import the file into MODFLOW. The structure of and relationship between the layers, the general location and elevation relationship of the gold mines are thus visualized. The actual distance of a horizontal grid in the figure is 279 m (Fig. 10).

Fig. 10
figure 10

Aquifer profile in plane subdivision diagram

Full size image

From the above figures, compared with the model area, the Quaternary pore aquifer is very thin and does not cover the whole area. Most of the model area consists of exposed bedrock areas, which is consistent with the actual situation. The thickness of the bedrock fractured aquifer increases successively from the faults to either ends, and the nearest fault is the smallest. The simulated bottom of the gold mine is located at approximately − 1000 m (Fig. 10).

Time discretization and setting of the initial values

The simulation time is from January 1, 2017, to December 30, 2018, with a time step of 30 days, and there are 24 stress periods. Based on the groundwater head of the well and the starting head of the long-time observation well, the initial flow field of the two aquifers was determined by interpolation in 2D Scatter Data of the GMS as shown in Figs. 11 and 12.

Fig. 11
figure 11

Initial water level contour map of the Quaternary pore aquifer

Full size image
Fig. 12
figure 12

Initial water level contour map of the bedrock structure fracture aquifer

Full size image

Partition of the hydrogeological parameters

The parameter value assignments and partitions are implemented through the Layer-Property Flow Package; it can define the horizontal and vertical hydraulic conductivity for each layer. MODFLOW then computes the cell by cell conductances using the K values and the layer geometry. And the hydrogeological parameter partition maps of the Quaternary pore aquifer and bedrock fractured aquifer are shown in Figs. 13 and 14. The selection of partition parameters of the Quaternary pore aquifer is mainly based on hydrogeological characteristics, previous pumping tests of the aquifer and experience. Area I of the water-bearing properties of the Quaternary pore aquifer is rich while area II is moderate. Partitioning of the bedrock structure fracture aquifer is mainly based on the lithology and the distribution of fault zones. Area I of the water-bearing properties of the bedrock fractured aquifer contains intrusive rocks, area II is the concentrated distribution area of the fault zone, and area III consists of metamorphic rocks.

Fig. 13
figure 13

Hydrogeological parameter partitioning of the Quaternary pore aquifer

Full size image
Fig. 14
figure 14

Hydrogeological parameter partitioning of the bedrock fractured aquifer

Full size image

Disposal of sources and sinks

There are 5 sources and sinks, including the precipitation infiltration recharge from January 1, 2017, to December 30, 2018, which amounts to 1107.68 × 104 m3, lateral recharge of the Quaternary pore aquifer, which is 0.256 × 104 m3/d, lateral recharge of the bedrock fractured aquifer, which is 0.691 × 104 m3/d, river seepage recharge that is 0 m3/d and river runoff that has been 0 m3/d in recent years; the evaporation capacity has a maximum impact depth of 4 m. The sources and sinks are disposed of through the recharge package, river package, and evaporation package. The mine drainage data are derived from the drainage of 4 major gold mines, including the Jiaojia Gold Mine, Xincheng Gold Mine, Wangershan Gold Mine, and Sizhuang Gold Mine from January 1, 2017, to December 30, 2018, and disposed of through the well package and drain package. Table 1 summarizes the monthly mine drainage of the 4 gold mines in the last 2 years. In addition, two main fault zones, the Wangershan fault zone and the Jiaojia fault zone, are generalized by the barrier package in the MODFLOW module (Fig. 15).

Table 1 Mine water inflow data of the Jiaojia Gold Mine area
Full size table
Fig. 15
figure 15

Schematic diagram of the source and sink items

Full size image

Model identification and verification

The preconditioned conjugate gradient method (PCG) package is one of the solver packages for MODFLOW supported by GMS, is a numerical solver to calculate the water head values that vary over time. After calculation, the temporal and spatial distribution characteristics of the groundwater table can be obtained for the hydrogeological conceptual model under the given hydrogeological parameters and equilibrium terms. To make the model more in line with the hydrogeological conditions of the study area, the hydrogeological parameters, boundary values and equilibrium terms are identified by fitting the long duration curve of the hydrological observation borehole water head. The minimization of the error is achieved by repeatedly adjusting the hydrogeological parameters and equilibrium terms. The dotted line represents the 95% confidence interval. In this paper, the water head fitting curves of 3 hydrological observation boreholes are depicted in Figs. 16, 17, and 18, and the dots in the figures are the predicted values. The inverted triangles are the actual values of the hydrological observation borehole water head, and the dotted line represents the 95% confidence interval. Mean absolute error analysis of the fit is carried out (Table 2).

Fig. 16
figure 16

Water level duration curve fitting graph of the JG03 observation borehole

Full size image
Fig. 17
figure 17

Water level duration curve fitting graph of the JG08 observation borehole

Full size image
Fig. 18
figure 18

Water level duration curve fitting graph of the JG09 observation borehole

Full size image
Table 2 Error statistics of the observation borehole curve fitting
Full size table

According to the fitting curves of the long-time observation borehole and the fitting curve of the flow field, we determine that the mean absolute errors of most of the fitted points are less than 1 m; thus, the simulation model meets the requirement of model accuracy. The hydrogeological parameters of each parameter zone are determined through parameter identification and are shown in Table 3.

Table 3 Statistical hydrogeological parameters
Full size table

Prediction accuracy of the water inflow

To verify the feasibility of the model, we compare the actual water inflow with the predicted water inflow at the mining level. The comparison between the predicted water inflow results of the numerical simulation and the actual water inflow in each gold mine is shown in Table 4.

Table 4 Comparison between the actual water inflow and predicted water inflow
Full size table

As seen from the above table, the relative error of the simulated water inflow is small, which agrees with the hydrogeological conditions of the study area and reflects the dynamic characteristics of the groundwater system. This model can be used to predict the mine water inflow.

Results and discussion

Groundwater flow field

The figures below are the results of groundwater head calculations for the aquifers at the end of verification, that is, 2 years later (Figs. 19 and 20). The Quaternary pore aquifer and bedrock fracture aquifer are the main water intake layers for local residents and production. Due to the long-term and continuous drainage of the mine areas, the groundwater head in the simulated area of the Quaternary pore aquifer and bedrock fracture aquifer continuously decreased, particularly in the Jiaojia Gold Mine, Wangershan Gold Mine, and Sizhuang Gold Mine, and the falling funnel of the water level appeared clearly. The groundwater head of the bedrock fracture aquifer in the Xincheng Gold Mine declined from 15 to 5 m, a decrease of 10 m, and no falling funnel appeared. The groundwater head in the Wangershan Gold Mine declined from 50 m to − 27 m, i.e., dropped by 77 m, and had a falling funnel with a range of 1939 m2. The groundwater head in the Jiaojia Gold Mine declined from − 15 to − 57 m, a decrease of 42 m, and had a falling funnel with a range of 3255 m2. The groundwater head in the Sizhuang Gold Mine declined from 10 to − 27 m, a drop of 37 m, and had a falling funnel with a range of 609 m2. The general direction of groundwater flow was from southeast to northwest, which could impede the horizontal flow of groundwater, specifically in the southern part of the Jiaojia fault zone.

Fig. 19
figure 19

Flow field fitting map of the Quaternary pore aquifer at the end of simulation

Full size image
Fig. 20
figure 20

Flow field fitting map of the bedrock fractured aquifer at the end of simulation

Full size image

The study of the variation in the groundwater flow field could result in improved evaluation of the quantity of groundwater resources and provide a theoretical basis for enhanced protection of groundwater resources. The study has certain guiding significance for urban layout arrangement and planning, and provides theoretical support for the prevention and control of water inrush in gold mining.

Prediction of the mine water inflow

The prediction of mine water inflow was divided into two cases. On the one hand, the amount of mine water inflow varied with increasing mining depth; on the other hand, the amount of mine water inflow varied with increasing mining age. Accurate prediction of the mine water inflow is of great significance to the prevention and control of mine water inrush, the minimization of mining risks, and the maximization of mining interests.

Prediction of water inflow at different mining levels

The bottom elevation of drainage was set to the mining level, which is required to predict the water inflow, i.e., to simulate the amount of water inflow with increasing mining depth. The water inflow at different mining levels in the Jiaojia Gold Mine, Xincheng Gold Mine, Wangershan Gold Mine, and Sizhuang Gold Mine was obtained using the groundwater numerical simulation model (Table 5).

Table 5 Predictions of the water inflow at different mining levels
Full size table

From the above, we can see that the water inflow increased with increasing mining depth. The mine water inflow of Wangershan Gold Mine exhibited the largest increase, followed by that of the Jiaojia Gold Mine.

Prediction of the dynamic change in the mine water inflow

Suppose that the goldmines had a constant level of production every 5 years. Using the established groundwater numerical simulation model, the water inflow in the 4 mines over the next 5 years was predicted. Figures 21, 22, 23, and 24 are parts of the forecast result charts. The seasonal variation in the water inflow in each mining area was clear. On the whole, there was an upward trend in the summer and a downward trend in winter, and at the same level, the mine water inflow had a stable trend with the increase in mining years.

Fig. 21
figure 21

Prediction chart of the water inflow at − 650 m in the Jiaojia Gold Mine

Full size image
Fig. 22
figure 22

Prediction chart of the water inflow at − 650 m in the Wangershan Gold Mine

Full size image
Fig. 23
figure 23

Prediction chart of the water inflow at − 700 m in the Xincheng Gold Mine

Full size image
Fig. 24
figure 24

Prediction chart of the water inflow at − 450 m in the Sizhuang Gold Mine

Full size image

The seasonal variation in the water inflow in the Sizhuang Gold Mine was the least apparent, as it was close to the Zhuqiao River. The underground runoff was relatively constant, the water supply head was relatively stable, and a certain water-conducting structure existed. The water and groundwater of the Zhuqiao River could be recharged to the mine along the water-conducting structure.

Risk of water inrush in Jiaojia gold mine area

Based on the above prediction of water inflow, groundwater flow field, and the complexity of tectonic development based on fault fractal value of Jiaojia gold mine area (Fig. 25), information fusion technology (Qiu et al. 2017b) is applied to judge the risk of water inrush. In this paper, the risk index of mine water inrush in Jiaojia gold mining area is given to provides practical basis for safe mining of gold mine. According to the geometric method (Qiu et al. 2016), the water inrush risk index (W) is divided into four grades (Fig. 26), which are no water inrush (0.16 < W < 0.28), less water inrush (0.28 < W < 0.4), small amount of water inrush (0.4 < W < 0.52), water inrush (0.52 < W < 0.64), and easy water inrush (0.64 < W < 0.76).

Fig. 25
figure 25

Fault fractal value contour map

Full size image
Fig. 26
figure 26

Risk index of water inrush

Full size image

While:

  1. (1)

    No water inrush (0.16 < W < 0.28): It is mainly distributed in the northern, western, and eastern boundary parts of the simulation area, and distributes in strips and lies.

  2. (2)

    Less water inrush (0.28 < W < 0.4): It mainly distributes in the north, west, and east of the simulation area, and distributes in strips and lies in the inside of the no water inrush area.

  3. (3)

    Small amount of water inrush (0.4 < W < 0.52): It mainly distributes in the southeastern and central transitional areas of the simulation area with large area.

  4. (4)

    Water inrush (0.52 < W < 0.64): Mainly distributed in the middle and southeast of the simulation area, circular distribution.

  5. (5)

    Easy water inrush (0.64 < W < 0.76): Mainly distributed in the middle and southeast of the simulation area, circular distribution.

Conclusion

Groundwater numerical simulation model in Jiaojia gold mine area was established from 2017 to 2018 based on GMS model used to predicted the mine water inflow of main gold mines and groundwater flow field, and the main conclusions are as follows:

  1. (1)

    Numerical simulations of unsteady flow conditions for the period of 2 years show that the groundwater head in the simulated area continuously declines, and a groundwater funnel is formed in the main gold mine area; the groundwater level declined by 10–77 m with a range of approximately 609–3255 m2.

  2. (2)

    The general direction of groundwater flow in the study area is from southeast to northwest towards the sea, and the faults can impede the horizontal flow of groundwater, particularly in the southern part of the Jiaojia fault zone.

  3. (3)

    The mine water inflow in the study area increases with increasing mining depth. Among the main gold mines, the mine water inflow exhibits the largest increase in the Wangershan Gold Mine, followed by that of the Jiaojia Gold Mine.

  4. (4)

    According to the numerical simulation results, it is expected that the mine water inflow in the study area will increase with increasing mining depth, and the water inflow at the same level will gradually become stable with increasing mining time.

  5. (5)

    The seasonal variation in the water inflow in each mining area is clear. Overall, there is an upward trend in the summer and a downward trend in the winter, due to the development of faults in Jiaojia gold mine area, ground precipitation can seep into the mine through cracks, and the water inflow varies seasonally. The seasonal variation in the water inflow in the Sizhuang Gold Mine is the least notable because the water and groundwater of the Zhuqiao River can be recharged to the mine along a water-conducting structure, which makes the seasonal variation of its water intake small.

  6. (6)

    The easy water inrush area in Jiaojia gold mine area is mainly distributed in the middle and southeast of the simulation area, circular distribution. The water inrush area is mainly distributed in the middle and southeast of the simulation area, circular distribution.