Introduction

As the West Development Strategy proposed in China, the importance of mines is gradually being strengthened there (Li 2012). Thereupon, the status of mine in Ordos has also been enhanced (Li 2016). In the process of mining, mined areas are being expanded with the formation of mine water (Wei et al. 2017). Groundwater, as a kind of renewable resources, could be used for living and irrigation of the human (Tesfagiorgis et al. 2011). China has encountered different environmental problems for complex surroundings of mines (Wu et al. 2011). Especially in the Central-Western provinces of China (Wang et al. 2017), where the ecological environment is too fragile to serve much water, there is a high potential to cause the groundwater levels declined, streams dried up and other serious ecological effects (Dong et al. 1999). Li et al. (2000) preliminarily discussed possibility of coal mining under water-containing (CMWC) between different engineering geological areas in North Shaanxi, China. Fan (2005) summarized and commented the CMWC, coming up with some advices about water protection in Yushenfu mining area. Wang et al. (2010) studied the mentioned topic in ecologically fragile areas, and Miao et al. (2010) analyzed hydrogeological features in mining areas of Eastern Shenmu and Ordos.

More and more technologies were applied in geology and hydrogeology as the depth and breadth of research in this field increasing. Visual MODFLOW, based on Darcy’s law and foundation of Dupuit, were widely welcomed in groundwater study (McDonald and Harbaugh 1988). Early, it was used in assessment of effects for different land use ways on soil salinization and groundwater variety (Pavelic et al. 1997). Zhou et al. (2000) simulated the distribution of groundwater sources and forecasted the groundwater distribution in Suqian, constituting a reasonable exploitation project. Maha Abd Algafar et al. (2011) simulated hydrogeological condition in Khartoum city as well as concluding appropriate water levels to maintain. Shao et al (2014) joined the inversion methods into the study of the groundwater in North China Plain, improving the estimation of the groundwater withdrawal greatly. Combined the existing and new methods for modeling catchment-scale recharge (RCH) and evapotranspiration (ET), remote sensing information for RCH and ET in calibration was potential for development in the future (Doble and Crosbie 2016).

To analyze the feasibility of mining, most areas have been prospected with bore holes data. However, these results could pay hardly attention to a visual preview of mining on outer surroundings, which may lead an inaccurate judgement, whilst the achievement that resulted by mentioned fragments were still unavailable in a large part. Combined with local climate in this paper, influences of different precipitations, the most important factor, and continuous leakage followed by vertical hydraulic conductivity change on groundwater variety of groundwater resources of 231 working area, Zhuanlongwan Mining Area (ZMA) will be studied.

Data and methods

Study area

Zhuanlongwan mining area (ZMA) in Ordos, lies in the southernmost district of the Inner Mongolia Autonomous Region, which is between 39°30′ and 39°37′ north latitudes and 110°00′ and 110°07′ east longitudes (Fig. 1).

Fig. 1
figure 1

Location of ZMA

Full size image

The region is in an arid semi-arid temperate plateau continental climate. The mean annual precipitation of the area is 359.6 mm, concentrated from July to September in years, and the mean annual evaporation is 2485.2 mm. Two streams, i.e., the East Ulan Mulun River and Gongnieergai ditch, flow across this region. Atmospheric precipitation is directly evaporated mainly, or intercepted by the vegetation and coming to the underground, and the maintenance of surface runoff is largely benefitted from it.

Geology and hydrogeology

The district is a typical erosive plateau and is characterized by rugged terrain. The topography of CMZ varies from 1174 to 1414 m, higher in the north and lower in the south (Fig. 2).

Fig. 2
figure 2

Topographic elevation map for the region of ZMA

Full size image

Main sedimentary layers of the district are: Middle Jurassic Zhiluo group (sandstone), Lower Cretaceous Zhidan group (sandstone), Quaternary Pleistocene Sara Wusu group (alluvium) and Holocene (aeolian sandstone). Most of the district is mainly covered by Quaternary aeolian sand, bedrocks is only exposed on ravines, banks and some hillsides (Fig. 3). Most area is mainly covered by Quaternary aeolian sand, the other bedrocks is only exposed on ravines, banks and some hillsides.

Fig. 3
figure 3

Geological map of ZMA in horizontal directions

Full size image

The main studied mining area, i.e., 231 working area of II-3 coal seam, related in the middle Jurassic Zhiluo formation, whose stratum is nearly horizontal and geological condition is simple. Average thicknesses of each stratum in Region of Jungar Juu-Xinmiao, where ZMA located, is listed as next: aeolian 20 m, Sara Wusu 13 m, Zhidan 52 m, Zhiluo 24 m (Only listed the section on Coal seam II-3) (Fig. 4). While the particular drilling holes showed Sara Wusu distributes little with poor thickness, which would be tricky in modeling work in next part.

Fig. 4
figure 4

Stratigraphical column of main formations

Full size image

The sandstone water above the coal seam roof, i.e., upper strata layers covered on the coal seam, is the direct water filling source. In the present mining condition, however, there may still be water seepages in thin bedrocks, where the water conducted fracture zone directly affects the Quaternary aquifer, which greatly increases the risk of water inrush in the mining area. In this model, aeolian and Sara Wusu are the main layers to study, both composing an entire phreatic aquifer in ZMA.

Residual water resisting layer (RWRL)

In the progress of simulation, hydrogeological model could hardly clarify the pure litho-related problems, while fortunately, which needn’t do that either. After attempts and comparison, the whole strata system was divided in two layers theoretically, the upper one, the phreatic, was to study in the paper and the other, rest lower two strata, led to the problem to study in the paper. And the theory of residual water resisting layer (RWRL) was also used for simplification. In exchange, the bed rock, including Zhiluo and Zhidan, was treated as an independent object drawn by the theory of RWRL.

To preview the environmental effects of mining work, the necessity of discussing how mining affects the environment becomes obvious (Liu et al. 2018a, b). While mines are dug out under, bedrocks above the coal seam start to bear more stresses and cracks (Fig. 5). The larger mined area is produced, the deeper fractured zone will be in upper bedrocks. These zones are called diversion fractured zone (DFZ). Some fractures may touch phreatic aquifer, in other words, the aquifuge has been destroyed completely. As a result, mine water from aquifers begins to drop down through these cracks. Only if the water is kept pumping, the work of mining could continue. On the other hand, continuously pumping water may means something harmful to surrounding environment. Consequently, the thickness of residual water resisting layer (RWRL) is important to calculate and preview (Liu et al. 2017, 2018a, b).

Fig. 5
figure 5

Relations between RWRL and other zones in mining area

Full size image

According to factors of the mining coal thickness, the hard-rock lithology ratio, the working face length and the mining depth, adopting multiple regression analysis (Eq. 2) (Hu et al. 2012), the thicknesses of fractured zone in all bore hole dispositions were then calculated out.

$${H_f}=3.47M+28.36b+1.89\ln l+0.31{e^{\left( {5.346\frac{{426.243}}{s}} \right)}}+6.04,$$
(1)

where H f is the thicknesses of fractured zone, M is the mining coal thickness, b is hard-rock lithology ratio, l is the working face length, s is the mining depth. In this paper, the II-3 coal seam were full mined in design, i.e., l is the thickness of the seam.

With drilling data in ZMA, distances between the lower bound of phreatic aquifer and the coal seam roof, thicknesses of DFZ could be easily collected. Thicknesses of RWRL, i.e., difference between the mentioned statistics, distributed in each drill hole were then calculated. By Kriging interpolation in Surfer, the distribution of 231 working area, RWRL thickness when the II-3 coal seam were fully mined was successful previewed (Fig. 6).

Fig. 6
figure 6

Distribution of RWRL thickness

Full size image

The more thickness RWRL is, the safer where there will be. In Fig. 6, the khaki area, i.e., overlap of selection rounded by white line and 231 working area, is the part calculated as negative value, which means the part has been probably broken.

Results

The circulation of the groundwater system was simulated by the MODFLOW groundwater model (Harbaugh et al. 2000). To describe the actual shape of the region in the model, active and inactive cells were made a definition, with a boundary defined as a no flow condition (Whitaker 1986).

The groundwater system was generalized as a single-layered phreatic aquifer, thus the following equation of two-dimensional groundwater flow is applied in steady state (Bedient et al. 1994):

$$\frac{\partial }{{\partial x}}\left( {T\frac{{\partial H}}{{\partial x}}} \right)+\frac{\partial }{{\partial y}}\left( {T\frac{{\partial H}}{{\partial y}}} \right)+R - D=0,$$
(2)

where H is the groundwater head or elevation (m), R is the groundwater recharge (m/d), D is the groundwater discharge (m/d), x and y are horizontal dimensions (m) and T is the transmissivity (m2/d), varies spatially depending on the geological conditions.

Equation 2 permits to set up a groundwater model without specifying a vertical dimension of the ground layers.

Design of model

By Visual MODFLOW finite-difference numerical simulation, the whole area was subdivided into on average, which produced 120 rows and 120 columns by square mesh. As mentioned before, one Sara Wusu distributed is not contiguous to another, which could hardly make a successful layer up in modeling. In light of that, the layer was combined with that one upper, i.e., Holocene, and adopted the latter parameters, which could draw some geometrical difference but almost no actual impact.

In the model, each area with its own parameters was assigned into a single-layer inhomogeneous isotropic phreatic aquifer system. On the basis of the hydrogeological parameters obtained from the pumping test, the district could be divided into four respective zones (Fig. 7).

Fig. 7
figure 7

Location of wells to establish the calibration points for hydraulic heads and cross section of vertical layers in steady state

Full size image

Evaluation of parameters

Each geological formation, assigned by corresponding parameters (Table 1), was considered as a separate unit. Corresponding to Fig. 2, the greatest values were valued in alluvium and aeolian sand, whose structures are loose and susceptible to storage the groundwater. The other two formations, i.e., Zhidan and Zhiluo, have smaller hydraulic conductivities for their basement structures, which the former is more fractured.

Table 1 Optimized parameters values for units
Full size table

As a part of the arid and semi-arid climate area, the constancy of local stream flows extremely rely on the game between atmospheric precipitation and evaporation, which accordingly should also be considered for their hydrogeological effects. In the model, relevant parameters are referred from the local hydrological data and optimized. As the limited float between wet and dry months, the groundwater levels were considered steady each year. On the other hand, the extinction depth was defined as 5 m by native experience. Data of recharge and evaporation boundaries are shown in Table 2.

Table 2 Atmospheric recharge and evaporation of the model
Full size table

Calibration

The accurate calibration between the observed and the calculated water levels in steady state finds a successful simulation of the groundwater flow in ZMA. By trial and error for calibration, the root mean squared (RMS) for the correlation is finally considered as 1.01 m, with correlation coefficient 0.89 and residual mean 0.91 m (Fig. 8).

Fig. 8
figure 8

The steady-state calibration results of ZMA groundwater flow mode

Full size image

A stable gradient was arrayed from the northern district to the southern outlet of Ulan Mulun River in ZMA (Fig. 9). With a higher value of 1381.48 m in north to 1160.08 m lower in southern district of the alluvial zone, an overall drop of 221.4 m has been observed before. Recharge mainly comes from the lateral supply of northern highland. The phreatic results have been matched with the equipotential map, in order to obtain the final conclusion.

Fig. 9
figure 9

Simulated groundwater heads of ZMA in steady state

Full size image

In calibrated state, injection and dewatering tests were run with different wells, variety of head in simulated and observed wells had a permitted error (Fig. 10), which further indicated the reasonableness of the model.

Fig. 10
figure 10

Test of water injection and pumping tests

Full size image

Sensitivity analysis

Some sensitivity analyses were done to prove parameters which affected the final results (Panagopoulos 2012). The main object of corresponding study is to find out the influence of model parameters and other relevant factors on the aquifer system, picking the most sensible parameter(s) up from all, which would be a remarkable object in the next term. On the basis of successful calibration, sensitivity analysis was performed for horizontal and vertical hydraulic conductivities, specific storage, recharge of rainfall and river boundaries (El-Zehairy et al. 2018). According to Rejani et al. (2008), project of 50% increase and decrease of the inputs was adopted in assessment of sensibility (Ting et al. 2010).

The sensitivity of mentioned parameters was analyzed by a 50% plus or subtraction in horizontal and vertical hydraulic conductivities, storage coefficient, interflow, rainfall recharge and river boundaries as designed before (Gallegos et al. 2013). Consequently, results of the analysis were sorted out as Fig. 11. As expected, river boundary, i.e., stream seepage, is the most sensitive factor for study area, playing an important role in the model. Sensitivity to increasement of precipitation recharge is surprisingly relevant, which confirmed the importance of precipitation to local ecology from another point of view (Ajami et al. 2011). Obviously, the phreatic aquifer of ZMA shows higher sensitivity to changes in recharge and stream seepage than to changes in vertical or horizontal hydraulic conductivity. For recharge condition, well heads were affected more when parameter decreased than that when increased. Vertical hydraulic conductivity is the most insignificant factor to sensibility working and specific storage drew no function on it.

Fig. 11
figure 11

Sensitivity analysis at sites Q1, J2, Q2, 1, S1, S2, S4

Full size image

Discussion

Once the simulation could be well refined and capable of consistently combining the measured statistics and model, it will be applied to get insight into some characteristics of hypothesis in actual practice. According to the preview result of Fig. 5, with drilling data showing the average elevation of phreatic layer above 231 working place as 1204 m, the groundwater was defined to drop down through the way where fractured zone meets phreatic layer.

In the model, this situation was simulated as a progress of leakage with drain package. Recharge would be imagined as a maximum and a minimum, which represented the groundwater variety under two limit positions of precipitation with the normal. Afterwards in normal recharge, constant flows though the broken zone would lead to reduction of vertical permeability coefficients (Wang et al. 2014). Conditions of drain would also change for that. In summary, it could be divided into following situations.

Estimated lower and upper limits for the present situation Estimates were found by acting models with the proper combination of minimum and maximum for precipitation (1303.2 and 537.1 mm/year), while the local evapotranspiration did not have a synchronous change with the former values owing to years stability (Fernández et al. 2016). The predicted changes would be limiting feasible as long as the geometry of broken area is confined to be constant. Precipitation rates sharply affected upstream area, while distribution of groundwater in northern highland would never consistent with topography (Fig. 12). The broken area encountered a marked drop of groundwater levels for floor leakages there. On the contrary, the downstream especially where in Ulan Mulun River may be faced with a less hydrological infection, in which the normal hydraulic gradient of Ulan Mulun had a not bad abnormality. The maximum amount of outflow would be 36% greater and the minimum of that would be 51% less than the average estimated value.

Fig. 12
figure 12

Simulation of drainage in different precipitations: a position of the groundwater levels for average situation; b, c the situations for minimum and maximum of recharge

Full size image

Conclusions

A consistent framework, with types of geological, hydrological, mine geological and hydro-geological data, is rendered helpful by groundwater models. This study discussed how a numerical model could be performed to integrate the present data of different channels in a centralized way, which provided a framework to preview a result. The results indicated that MODFLOW model has good performance for simulation of groundwater flow field in ZMA.

By analyzing the sensitivity, hydrological factors presented a greater influence to water drainage in the calibrated and validated model, and the accuracy of completed system was well proved as well. In the study, by combining the MODFLOW model and mine geological calculation, possible results of climate change was obtained, and corresponding changes in water head and amount was quantified to reflect the effect of mining scheme. The loss of groundwater caused by mining work according to the current mining plan would make surface environment of ZMA born several more terrible water-apply pressures especially in the dry seasons. According to the results, an optimized mining plan needs to be designed and re-combined with the studied model but would not be studied here. This result will be required for improving and refining with further practices in the future. Nevertheless, in most commercial mine activities, companies pay more attention to benefits and speed of projects. This study organized a relatively complete scheme evaluation with limited materials, in order to reduce the loss of property and destruction of the environment, thus could be useful to support benefit and safety for the mine in the area.