Impact of Coal Mining on the Moisture Movement in a Vadose Zone in Open-Pit Mine Areas

: Long-term dewatering of groundwater is a necessary operation for mining safety in open-pit coal mines, as extensive dewatering might cause ecological problems due to dramatic changes in moisture movement in the soil, especially in ecologically fragile areas. In order to evaluate the impact of the coal mining operation on moisture movement in the vadose zone and vegetation, this paper presents a quantitative methodology and takes the Baorixile open-pit coal mine as a study example. A long-term in situ experiment (from 2004 to 2018), laboratory analysis, and numerical modelling were conducted to analyze the mechanisms and relationship among the dropping groundwater level, the vadose-zone moisture, and the ecological responses in the grassland area. The experiment data and modelling results suggest that groundwater level dropping during open-pit mining operation has limited inﬂuence on the vadose zone, exhibiting a variation of capillary water zone within a depth of 3 m while the vadose zone and soil water zone were at least 16 m deep. The critical evaporation depth of ground water is 8 m. The long-term inﬂuence radius of groundwater dewatering is about 2.72 km during the Baorixile mining operation, and the groundwater level change mainly inﬂuences the lower part of the intermediate vadose zone and the capillary water zone below 16 m, with little inﬂuence on the moisture contents in the soil water zone where the roots of shallow vegetation grow. The results from this study provide useful insight for sustainable development of coal mining in ecologically fragile areas.


Introduction
Hulunbuir Grassland, the most concentrated and representative area of temperate grassland in China, includes multiple types of grassland ecosystems. It is not only an important ecological barrier in North and Northeast China [1], but also a main base for the integrated development of coal and electricity in China. Statistics suggest that Hulunbuir Grassland degrades at a rate of 2% annually and most of the rivers in Hulunbuir Grassland area have suffered from flow decline and even flow cutoff in recent years [2,3]. Eco-environmental problems have become a major risk, endangering the sustainable development of the mining area and even the ecological security of the entire region [4,5].
Environmental problems related to water in open-pit mines are increasingly piquing researcher's interests. In recent years, large efforts had been devoted to studying the consequence of water resources [6][7][8][9][10][11][12] and contamination migration [13,14] during openpit coal mining. Furthermore, ecological issues due to dewatering in open-pit coal mining are also becoming attractive to researchers. As suggested by Chu et al. [15], the total ecological overburden of open-pit coal mines is 4.31 to 11.36 times that of underground coal mines. Therefore, great progress in studies has been made in terms of ecological areas. Field investigation, laboratory tests, and numerical simulations were implemented to investigate the vadose zone moisture movement under the condition of a drop in groundwater level during mining and clarify the vertical distribution characteristics of vadose zone soil moisture, which are of great significance for correctly evaluating the influence of coal mining on regional groundwater resources and surface ecology, thereby guiding the management on the ecological policies and design of the proper way to rehabilitate.

General Characterization of the Study Region
The Baorixile open-pit mining area is located in Hailar District and Chenbaerhu Banner of Hulunbuir city, Inner Mongolia Autonomous Region (see Figure 1). The area has a continental sub-frigid zone climate, with an annual average temperature of −2.6 • C within the range of −48-37.7 • C, an annual average precipitation of 315.0 mm, and an annual average evaporation of 1344.8 mm. The Bao #1 mine has a production capacity of 20.0 Mt/year and has been in production for 13 years, since 2005. The mining area has a monoclinic structure, with a north-northwest dip direction and a 5-15 • dip angle. Within the area, from bottom to top, the Lower Cretaceous Series of the Longjiang Formation, Damoguaihe Formation, and Yimin Formation, as well as the Quaternary System's Holocene series, are the mainly developed strata. The coal-bearing stratum is located in the Damoguaihe Formation of the Lower Cretaceous, where 20 minable coal seams in four minable coal seam groups have developed. areas. Field investigation, laboratory tests, and numerical simulations were implemented to investigate the vadose zone moisture movement under the condition of a drop in groundwater level during mining and clarify the vertical distribution characteristics of vadose zone soil moisture, which are of great significance for correctly evaluating the influence of coal mining on regional groundwater resources and surface ecology, thereby guiding the management on the ecological policies and design of the proper way to rehabilitate.

General Characterization of the Study Region
The Baorixile open-pit mining area is located in Hailar District and Chenbaerhu Banner of Hulunbuir city, Inner Mongolia Autonomous Region (see Figure 1). The area has a continental sub-frigid zone climate, with an annual average temperature of −2.6 °C within the range of −48-37.7 °C, an annual average precipitation of 315.0 mm, and an annual average evaporation of 1344.8 mm. The Bao #1 mine has a production capacity of 20.0 Mt/year and has been in production for 13 years, since 2005. The mining area has a monoclinic structure, with a north-northwest dip direction and a 5-15° dip angle. Within the area, from bottom to top, the Lower Cretaceous Series of the Longjiang Formation, Damoguaihe Formation, and Yimin Formation, as well as the Quaternary System's Holocene series, are the mainly developed strata. The coal-bearing stratum is located in the Damoguaihe Formation of the Lower Cretaceous, where 20 minable coal seams in four minable coal seam groups have developed. Within the mining area, the Quaternary porous water-bearing formation is a relatively stable aquifer affected by mining. The formation is of gravel stratum with a thickness of 17-30 m, which gradually becomes thinner from south to north and from west to east and wedges out to the east of the area. The Quaternary strata in the eastern part of the mining area is not water-bearing. The Cretaceous strata are directly covered by the Quaternary strata, the bottom of which is generally a layer of moraine pebbles, whereas the top of the coal-bearing strata is composed of mudstone, siltstone, and fine sandstone.

Evolution of Groundwater Level during the Mining Period
According to a report of a field investigation and water quality test by Inner Mongolia Coal Geological Exploration (Group) Co. Ltd. (www.nmmtdz.com.cn), from 2004 to 2016, the Baorixile open-pit mine recorded an annual average production capacity of 20 million tons/year, an annual average stripping capacity of 7.5 million m 3 /year, and an annual average water discharge of 3 million m 3 /year. In the early stage of dewatering, the water level elevation was 597.35 m, and the water discharged was 14,000 m 3 /d. In 2004, the groundwater level elevation of the first mining area was 570 m, the dewatering capacity of which was 4900 m 3 /d. The dewatering capacity of the stope was 8640 m 3 /d in 2010, Within the mining area, the Quaternary porous water-bearing formation is a relatively stable aquifer affected by mining. The formation is of gravel stratum with a thickness of 17-30 m, which gradually becomes thinner from south to north and from west to east and wedges out to the east of the area. The Quaternary strata in the eastern part of the mining area is not water-bearing. The Cretaceous strata are directly covered by the Quaternary strata, the bottom of which is generally a layer of moraine pebbles, whereas the top of the coal-bearing strata is composed of mudstone, siltstone, and fine sandstone.

Evolution of Groundwater Level during the Mining Period
According to a report of a field investigation and water quality test by Inner Mongolia Coal Geological Exploration (Group) Co. Ltd. (www.nmmtdz.com.cn, accessed on 20 February 2021), from 2004 to 2016, the Baorixile open-pit mine recorded an annual average production capacity of 20 million tons/year, an annual average stripping capacity of 7.5 million m 3 /year, and an annual average water discharge of 3 million m 3 /year. In the early stage of dewatering, the water level elevation was 597.35 m, and the water discharged was 14,000 m 3 /d. In 2004, the groundwater level elevation of the first mining area was 570 m, the dewatering capacity of which was 4900 m 3 /d. The dewatering capacity of the stope was 8640 m 3 /d in 2010, whereas the dewatering volume was 2941,100 m 3 in 2012. During the same period, affected by the dewatering and drainage of the pit, the groundwater level was negatively correlated with the amount of mining.

Study Methodology
Field investigation, laboratory tests, and numerical simulation were involved in studying the moisture movement in vadose zones. The numerical simulations were conducted to quantitatively describe the vertical movement of the moisture in vadose zones during groundwater level fluctuation.

Site Selection of the Field Investigation
To elucidate the soil type of the vadose zone in the Baorixile mining area, 16 test points were positioned along the north-south and east-west directions based on the groundwater runoff direction (from northeast to southwest), with the cross-section intersection centered in the mining area. The test points were selected according to the degree of water resource utilization along the investigated directions. Then, in situ testing was conducted on the lithological and structural features of the vadose zone in the coalfield. For the locations of the test points, refer to Figure 2. whereas the dewatering volume was 2941,100 m 3 in 2012. During the same period, affected by the dewatering and drainage of the pit, the groundwater level was negatively correlated with the amount of mining.

Study Methodology
Field investigation, laboratory tests, and numerical simulation were involved in studying the moisture movement in vadose zones. The numerical simulations were conducted to quantitatively describe the vertical movement of the moisture in vadose zones during groundwater level fluctuation.

Site Selection of the Field Investigation
To elucidate the soil type of the vadose zone in the Baorixile mining area, 16 test points were positioned along the north-south and east-west directions based on the groundwater runoff direction (from northeast to southwest), with the cross-section intersection centered in the mining area. The test points were selected according to the degree of water resource utilization along the investigated directions. Then, in situ testing was conducted on the lithological and structural features of the vadose zone in the coalfield.

The Lithologic Structure of the Vadose Zone
The on-site excavation and measurement (see Figure 2a,b) results show that the vadose zone is mainly composed of the upper humus layer and the lower deep sand soil layer, and there is a clear boundary between the two layers. The soil samples were analyzed for grain composition and then classified according to the international soil texture classification standard [47]. According to the data measured, the thickness of the humus layer is 30-60 cm, with an average thickness of approximately 40 cm. Comprehensive analysis suggested that the soil present at depths from 0 cm to 40 cm is a silt loam, and that below 40 cm is loamy sand. The on-site excavation and measurement (see Figure 2a,b) results show that the vadose zone is mainly composed of the upper humus layer and the lower deep sand soil layer, and there is a clear boundary between the two layers. The soil samples were analyzed for grain composition and then classified according to the international soil texture classification standard [47]. According to the data measured, the thickness of the humus layer is 30-60 cm, with an average thickness of approximately 40 cm. Comprehensive analysis suggested that the soil present at depths from 0 cm to 40 cm is a silt loam, and that below 40 cm is loamy sand.  Figure 2c) was constructed in the pasture near the dump site of the mining area. The test field included the selected typical profile, the humus soil thickness of which is 40 cm. Weather, soil moisture, and groundwater level data were obtained from 11 September, 2017, to 9 March, 2018 (180 days in total) (see Figure 2d). Meteorological elements were collected by an automatic weather station installed at the dump site. During the monitoring period, the total precipitation was 1532.1 mm, the daily average net radiation was 4.68 MJ/m 2 , the daily average temperature was −11 • C, the daily average relative air humidity was 68.7%, and the daily average wind velocity was 121.866 km/d.
The soil water content of the vadose zone was measured with EM − 50 m sensors. The soil water temperature and humidity sensor probes (type No. 5TM) were buried at 20 cm, 50 cm, 100 cm, and 300 cm below the ground. The data were collected every half an hour, and the water content changes obtained are shown in Figure 3.
Sustainability 2021, 13, x FOR PEER REVIEW 5 of 17 soil moisture, and groundwater level data were obtained from 11 September, 2017, to 9 March, 2018 (180 days in total) (see Figure 2d). Meteorological elements were collected by an automatic weather station installed at the dump site. During the monitoring period, the total precipitation was 1532.1 mm, the daily average net radiation was 4.68 MJ/m 2 , the daily average temperature was −11 °C, the daily average relative air humidity was 68.7%, and the daily average wind velocity was 121.866 km/d. The soil water content of the vadose zone was measured with EM − 50 m sensors. The soil water temperature and humidity sensor probes (type No. 5TM) were buried at 20 cm, 50 cm, 100 cm, and 300 cm below the ground. The data were collected every half an hour, and the water content changes obtained are shown in Figure 3. The moisture content of shallow soil at a depth of 20 cm varied greatly with time due to the low amount of precipitation and the dominant influence of evaporation during the monitoring period. The soil moisture content at a depth of 50 cm was the greatest and remained stable because of the decrease in soil texture and the effect of the local plants. The soil moisture content decreased with increasing depth from 100 cm to 300 cm, and the deeper the soil was, the more stable the water content. Generally, the soil moisture content increased with increasing depth. The groundwater level monitoring data were collected from the mine's groundwater level observation wells at a frequency of once a day. As it is plotted in Figure 3, from the establishment of the in situ test field to 3 October, 2017, the groundwater depth fluctuated around 20 m. Due to large-scale drainage activity at the Baorixile mine in early October, the groundwater level in the study area decreased. On 5 October, the groundwater level was 21 m, with a drop of 1 m. The groundwater level continued to decrease in December 2017 and fell to 21.5 m on 20 January, 2018; it rose again to 21 m in early March 2018.

Laboratory Tests
To quantitatively analyze the moisture movement in the vadose zone in the study area, measurement of the basic parameters of the vadose zone soil is required, including the soil bulk density, saturated hydraulic conductivity, and soil porosity. Equipment utilized in the laboratory tests is shown in Figure 4, including grading sieves, a variable head seepage device, and a nanoVoxel-3502E scanning platform for soil classification, hydraulic conductivity, and porosity, respectively. The moisture content of shallow soil at a depth of 20 cm varied greatly with time due to the low amount of precipitation and the dominant influence of evaporation during the monitoring period. The soil moisture content at a depth of 50 cm was the greatest and remained stable because of the decrease in soil texture and the effect of the local plants. The soil moisture content decreased with increasing depth from 100 cm to 300 cm, and the deeper the soil was, the more stable the water content. Generally, the soil moisture content increased with increasing depth. The groundwater level monitoring data were collected from the mine's groundwater level observation wells at a frequency of once a day. As it is plotted in Figure 3, from the establishment of the in situ test field to 3 October, 2017, the groundwater depth fluctuated around 20 m. Due to large-scale drainage activity at the Baorixile mine in early October, the groundwater level in the study area decreased. On 5 October, the groundwater level was 21 m, with a drop of 1 m. The groundwater level continued to decrease in December 2017 and fell to 21.5 m on 20 January, 2018; it rose again to 21 m in early March 2018.

Laboratory Tests
To quantitatively analyze the moisture movement in the vadose zone in the study area, measurement of the basic parameters of the vadose zone soil is required, including the soil bulk density, saturated hydraulic conductivity, and soil porosity. Equipment utilized in the laboratory tests is shown in Figure 4, including grading sieves, a variable head seepage device, and a nanoVoxel-3502E scanning platform for soil classification, hydraulic conductivity, and porosity, respectively.

Dry Soil Densities
The cutting-ring method was adopted to measure the soil bulk density of stratified sections [48]. The drying method was adopted to measure the stratified soil samples [49]. For each soil type after classification by grading sieves (see Figure 4a), four groups were measured and averaged.

Dry Soil Densities
The cutting-ring method was adopted to measure the soil bulk density of stratified sections [48]. The drying method was adopted to measure the stratified soil samples [49]. For each soil type after classification by grading sieves (see Figure 4a), four groups were measured and averaged.

Saturated Hydraulic Conductivity
The variable head seepage method (see Figure 4b) was adopted [50] to measure the saturated hydraulic conductivity k of the silt loam and loamy sand, and the average permeability coefficient 20 k of the two types of soil was calculated as follows:

Porosity of Soil
To grasp the microstructure of the soil along a typical profile more comprehensively and precisely, a nanoVoxel-3502E scanning platform (see Figure 4c) and digital core analysis technology were used on the soil samples, 3D extraction of the soil pores was carried out with the threshold segmentation method, as plotted in Figure 5, and the equivalent pore diameters in the silt loam and loamy sand were determined.

Saturated Hydraulic Conductivity
The variable head seepage method (see Figure 4b) was adopted [50] to measure the saturated hydraulic conductivity k of the silt loam and loamy sand, and the average permeability coefficient k 20 of the two types of soil was calculated as follows: where a is the sectional area of the variable head tube (cm 2 ), L is the height of the sample (cm), A is the sectional area of the sample (cm 2 ), t 1 and t 2 are the starting time and the ending time of the head reading (s), h 1 and h 2 are the starting water head and the ending water head (cm), η T is the hydrodynamic viscosity coefficient (kPa·s) at T • C, and η 20 is the dynamic viscosity coefficient (kPa·s) of water at 20 • C. η T /η 20 , the dynamic viscosity coefficient ratio, is derived from related literature [51].

Porosity of Soil
To grasp the microstructure of the soil along a typical profile more comprehensively and precisely, a nanoVoxel-3502E scanning platform (see Figure 4c) and digital core analysis technology were used on the soil samples, 3D extraction of the soil pores was carried out with the threshold segmentation method, as plotted in Figure 5, and the equivalent pore diameters in the silt loam and loamy sand were determined.

Dry Soil Densities
The cutting-ring method was adopted to measure the soil bulk density of stratified sections [48]. The drying method was adopted to measure the stratified soil samples [49]. For each soil type after classification by grading sieves (see Figure 4a), four groups were measured and averaged.

Saturated Hydraulic Conductivity
The variable head seepage method (see Figure 4b) was adopted [50] to measure the saturated hydraulic conductivity k of the silt loam and loamy sand, and the average permeability coefficient 20 k of the two types of soil was calculated as follows: where a is the sectional area of the variable head tube (cm 2 ), L is the height of the sample (cm), A is the sectional area of the sample (cm 2 ), 1 t and 2 t are the starting time and

Porosity of Soil
To grasp the microstructure of the soil along a typical profile more comprehensively and precisely, a nanoVoxel-3502E scanning platform (see Figure 4c) and digital core analysis technology were used on the soil samples, 3D extraction of the soil pores was carried out with the threshold segmentation method, as plotted in Figure 5, and the equivalent pore diameters in the silt loam and loamy sand were determined.

Results of Laboratory Tests
The soil samples were analyzed for grain composition and then classified according to the international soil texture classification standard [47]. The soil present at depths from 0 cm to 0.4 m was a silt loam, which was the same as the field observation, and that below 0.4 m was loamy sand. The measured average dry soil density of silt loam was about 1.36 g/cm 3 and that of loamy sand was about 1.39 g/cm 3 . According to the results obtained, the saturated moisture content of the silt loam was 0.4753 cm 3 , and that of the loamy sand was 0.4214 cm 3 . Based on the previously measured parameters of the soil profile, the Marqualdt-Levenberg parameter optimization algorithm in Hydrus was used for the inversion [52] and to obtain the empirical parameter of the vadose zone's soil moisture characteristics. The inversion data were sourced from the moisture content data measured from 11 September, 2017, to 9 March, 2018 (180 days in total), based on which θ s (cm 3 /cm 3 ), the residual moisture content θ r of the two layers, and the empirical parameter α(cm −1 ) were obtained. Specifically, Equation (3) was applied to determine the water retention curve [53]: which can be rewritten as Thus, Equation (4) can be simplified as In Equation (4), the subscript i means the i th number of measured data of θ i and h i . According to Equation (5), two steps are involved in the calculation [54]: Step 1: the initial value of m is assumed to range from 0 to 1. Then calculate α with Equation (3).
Step 2: where ε is the accuracy error defined as 0.0001, return to step 1, otherwise complete the calculation.
All the calculations were implemented with Matlab software. The final characteristic parameters of the soil moisture are shown in Table 1. The test results of saturated transmissibility and average permeability of the study soils are listed in Table 2. The method for calculating the saturated water transmissibility and average permeability from laboratory test data is referenced in the literature [51,52]. The average permeability of silt loam was about 9.623 cm/d, and that of loamy sand was about 363.3 cm/d. According to the test results of the water retention curves, hydraulic conductivity tests, saturated soil water content tests and numerical revision of Hydrus 1D, the parameters were summarized as in Table 1. The soil water retention curves obtained are shown in Figure 6a and the fitted curves by parameters in Table 1 are shown in Figure 6b. According to the test results of the water retention curves, hydraulic conductivity tests, saturated soil water content tests and numerical revision of Hydrus 1D, the parameters were summarized as in Table 1. The soil water retention curves obtained are shown in Figure 6a and the fitted curves by parameters in Table 1 are shown in Figure 6b. The silt loam mainly contained fine pores with an equivalent diameter of 100 μm and had a porosity of 41.87%. The loamy sand contained pores with equivalent diameters of 100-500 μm and had a porosity of 49.36%.

Numerical Simulation
This paper involves numerical modelling of the moisture variation and evaporation discharge in the vadose zone induced by different groundwater levels. The current computational methods on these aspects are various, such as MODFLOW, STOMP, OpenGe-oSys, TOUGH, and HYDRAUS. [55][56][57][58][59]. Among them, HYDRAUS is a widely used software to calculate the vadose zone based on Richard's flow equation [60][61][62]. This paper applied HYDRUS software to compute the generalized model.
During coal mining, the influence radius of dewatering and drainage in the studied stope is approximately 2.72 km. According to the data monitored on site, the groundwater level near the boundary reaches 20-21.5 m, ranging within 1.5 m. These fluctuations are obvious and due to seasonal changes. With a thickness of approximately 20 m, the vadose zone in the mining area is deep.
To quantitatively study the influence of groundwater level changes on the moisture and movement in the vadose zone, the model was solved using the Hydrus-1D finite element calculation model, since the studied area was about 2.72 km, which was large The silt loam mainly contained fine pores with an equivalent diameter of 100 µm and had a porosity of 41.87%. The loamy sand contained pores with equivalent diameters of 100-500 µm and had a porosity of 49.36%.

Numerical Simulation
This paper involves numerical modelling of the moisture variation and evaporation discharge in the vadose zone induced by different groundwater levels. The current computational methods on these aspects are various, such as MODFLOW, STOMP, OpenGeoSys, TOUGH, and HYDRAUS [55][56][57][58][59]. Among them, HYDRAUS is a widely used software to calculate the vadose zone based on Richard's flow equation [60][61][62]. This paper applied HYDRUS software to compute the generalized model.
During coal mining, the influence radius of dewatering and drainage in the studied stope is approximately 2.72 km. According to the data monitored on site, the groundwater level near the boundary reaches 20-21.5 m, ranging within 1.5 m. These fluctuations are obvious and due to seasonal changes. With a thickness of approximately 20 m, the vadose zone in the mining area is deep.
To quantitatively study the influence of groundwater level changes on the moisture and movement in the vadose zone, the model was solved using the Hydrus-1D finite element calculation model, since the studied area was about 2.72 km, which was large enough to ignore the transverse water recharge, and as demonstrated by Miao et al. [26], moisture movement in a vadose zone is mainly triggered by evaporation. Thus, 1-D modelling is close to 3-D modelling if the main concern is on the vadose zone.

Generalized and Numerical Model Setup
According to the observation of on-site excavation and classification by grading sieves, the lithologic structure of the typical vadose zone profile was generalized into two layers within a total thickness of 20 m. The upper layer was a silt loam layer with a thickness of 0.4 m, the lower layer was a loamy sand layer with a thickness of 19.6 m, and each soil medium was homogeneous. Surface runoff was not considered into this study on the vadose zone in the study area during rainfall. Due to the low rainfall and strong evaporation in this grassland area, the moisture in the vadose zone is mainly exchanged vertically, and its lateral flow could be neglected. The generalized model is shown in Figure 7a. sieves, the lithologic structure of the typical vadose zone profile was generalized into two layers within a total thickness of 20 m. The upper layer was a silt loam layer with a thickness of 0.4 m, the lower layer was a loamy sand layer with a thickness of 19.6 m, and each soil medium was homogeneous. Surface runoff was not considered into this study on the vadose zone in the study area during rainfall. Due to the low rainfall and strong evaporation in this grassland area, the moisture in the vadose zone is mainly exchanged vertically, and its lateral flow could be neglected. The generalized model is shown in Figure 7a.  Figure 7b) to monitor the changes in soil moisture content.

Governing Equations
The Van-Genuchten model [63] was used to calculate the parameters of the soil moisture in unsaturated media, and the unsaturated permeability coefficient was predicted in the form of a water retention curve. The calculation method of the water retention curves was the same as that of references [50,64,65]. Specifically, it was assumed that the groundwater level at the bottom of the soil profile remained unchanged during a short period and that the water evaporation intensity of the soil surface was 0. After the moisture  Figure 7b) to monitor the changes in soil moisture content.

Governing Equations
The Van-Genuchten model [63] was used to calculate the parameters of the soil moisture in unsaturated media, and the unsaturated permeability coefficient was predicted in the form of a water retention curve. The calculation method of the water retention curves was the same as that of references [50,64,65]. Specifically, it was assumed that the groundwater level at the bottom of the soil profile remained unchanged during a short period and that the water evaporation intensity of the soil surface was 0. After the moisture distribution in the soil reached a steady state, the moisture content data monitored at the depths of 20 cm, 50 cm, 100 cm, and 300 cm along the profile were fitted.

Definite Conditions
The initial conditions were set according to the field-measured moisture content. The upper boundary of model was set as the atmospheric boundary under natural conditions, which were mainly influenced by rainfall, evaporation, and crop transpiration, and irrigation was ignored in this problem. The amount of precipitation was directly measured by the automatic weather station in the area, and the crop transpiration was calculated with the help of the measured meteorological data and the Penman formula. The lower boundary was described with the pressure head indicated by the groundwater level. The fluctuation in the water level at an observation hole was used to represent the change in groundwater level and to select the deep drainage boundary. The bottom flux of the model was selected according to the fluctuation of the groundwater level.

Moisture Movement in a Vadose Zone during the Mining Period
In soil science, the vadose zone is divided into three belts from the surface to the water table: the soil water belt (capillary suspended water zone), the intermediate vadose belt, and the capillary water belt [66]. To compare the moisture movement and the range of capillary water rise under different groundwater extents, the variation in soil moisture content was chosen to reflect the vadose zone's moisture movement. The numerical simulation scheme was designed as follows: (1) Based on years of weather data and groundwater level variation characteristics, the moisture movement in the typical profile of the vadose zone was simulated during the coal mining period (2004-2016). The soil moisture movement in the intermediate vadose belt, close to the capillary water belt, during the continuous lowering of the groundwater level was analyzed under the mining conditions. (2) Based on the in situ test data, the moisture movement in the typical profile of the vadose zone during the monitoring period (from 11 September, 2017, to 9 March, 2018) was simulated. The influence of the groundwater level drop on the soil moisture contents in different vadose zone belts was analyzed, the maximum range of capillary rise was determined, and the critical depth to water, which can form a hydraulic relationship with the soil water belt, was simulated and predicted.

Groundwater Level and Vadose Zone Moisture Change during the Mining Period
Combined with the background of coal mining, the long-term moisture movement in the vadose zone under the conditions of periodic rainfall and a deep burial depth was simulated for the typical soil profile to obtain the influence of the long-term drop in the groundwater level on the moisture movement in different areas of the vadose zone.
The changes in soil moisture content at different burial depths during the mining period are shown in Figure 8.  The soil moisture contents of the profiles at the nodes (as shown in Table 3) were numerically calculated. The four-day meteorological data were used as the upper boundary input data, and the monitored value of the topsoil moisture content was used as the initial moisture content. The lower boundary conditions were set according to the depth changes in the groundwater level. The zonality of the vadose zone was determined, and the influence of the lowering of the groundwater level on the soil moisture content was analyzed according to the vertical distribution characteristics of the soil moisture. The simulation results are shown in Figure 9. Based on the changes in soil moisture content with depth, the ranges of the vadose zone's three belts (the soil moisture belt, the intermediate vadose belt, and the capillary water belt) were determined: (1) As the soil moisture content started to decrease significantly at a depth of 4 m below the surface, the area from the surface to a depth of 4 m could be classified as the soil moisture belt. In this belt, the variation in soil moisture content decreased with the increasing depth, and moisture distribution along the depth in this belt was different, as in Figure 9. Specifically, from November to March, as the temperature decreased, as suggested by Shen et al. [67], the temperature of the shallow vadose zone decreased until the soil moisture was frozen, during which time the soil moisture content was significantly reduced and more stable. In March, the moisture content of the soil moisture belt increased significantly, mainly due to the melting of frozen soil and surface snow. The difference in the moisture within a depth of 4 m convinced us that it is easily influenced by atmospheric change. (2) The area between the burial depths of 4-18.8   Based on the calculated results, the correlation curve between the vertical groundwater extent and the recharge amount is plotted in Figure 10. When the depth of the groundwater level was within 8 m, the recharge dropped from 421 mm/a to 100 mm/a and then remained unchanged, indicating that the recharge did not change as the groundwater level extended deeper than 8 m. Therefore, the critical evaporation depth of the typical profile of the vadose zone was 8 m. As the depth of groundwater level ranged from about 17 m to 21 m as displayed in Figure 8, according Figure 10 Based on the calculated results, the correlation curve between the vertical groundwater extent and the recharge amount is plotted in Figure 10. When the depth of the groundwater level was within 8 m, the recharge dropped from 421 mm/a to 100 mm/a and then remained unchanged, indicating that the recharge did not change as the groundwater level extended deeper than 8 m. Therefore, the critical evaporation depth of the typical profile of the vadose zone was 8 m. As the depth of groundwater level ranged from about 17 m to 21 m as displayed in Figure 8, according Figure 10, the groundwater level has little influence on the evaporation recharge. In other words, the fluctuation of groundwater level induced by dewatering and drainage in open-pit coal mines has limited impact on groundwater evaporation. remained unchanged, indicating that the recharge did not change as the groundwater level extended deeper than 8 m. Therefore, the critical evaporation depth of the typical profile of the vadose zone was 8 m. As the depth of groundwater level ranged from about 17 m to 21 m as displayed in Figure 8, according Figure 10, the groundwater level has little influence on the evaporation recharge. In other words, the fluctuation of groundwater level induced by dewatering and drainage in open-pit coal mines has limited impact on groundwater evaporation.

Discussion
An investigation into the soil moisture content in Figures 8 and 9 convinced us that coal mining influenced the groundwater level initially but kept stable thereafter. This tendency can be also seen in the intermediate vadose belt. Dewatering and drainage in the coal mine caused approximately 0.4 m to 0.8 m change in depth of the capillary water belt. The critical evaporation depth of the study area was about 8 m. Based on these analyzed results, the impact of dewatering in coal mines on plants can be analyzed as follows.
Hu et al. [68] denoted that 10% of the effective diameter of soil is able to represent the diameter of the continuous capillary in soil. According to Zhang [67], the capillary rise height h c in soil can be defined as 0.73/D ≤ h c (cm) ≤ 2/D (6) where D is the effective diameter of fine soil particles. As measured by Zhang [69], the average value of D in the Baorixile coal mine area is about 0.001 mm; thus, the variable height in the studied area ranged from 0.73 m to 2 m. In the studied area, there were three kinds of plants, namely, grasses, shrubs, and trees. The depth of their roots buried were approximately 0.5 m, 1.5 m, and 6 m, respectively. Therefore, the critical depth that could be influenced by soil water content of grasses ranged from 1.23 m to 2.5 m, of shrubs ranged from 2.23 m to 3.5 m, and of trees ranged from 6.73 m to 8 m.
Obviously, the dewatering of coal mines has limited influence on the grasses and shrubs because the soil water content started to change below the depth of 4 m. Trees in the studied area would be influenced slightly due to the moisture in intermedial vadose zone caused by variation of underground water level, but would recover soon after the dewatering stops.

Conclusions
Through the investigation of the hydrogeology and lithological structure of the vadose zone in the Baorixile mining area, according to the analysis on the data of in situ tests and numerical modelling results, the following conclusions were drawn: Trees would be influenced initially but recovered to normal soon after the dewatering stops. As a whole, the vegetation was mainly affected by the changes in atmospheric precipitation at the upper boundary.