Hydrotechnical and ecological principles of water resources management for a mined-out mine field

The article considers the issue of predicting and managing water resources of flooded mines based on hydrotechnical and ecological criteria. Reproduction of a natural-technogenic situation within mine fields with many unknowns regarding the condition of a rock massif and influence factors became possible through a step-by-step construction and identification of a numerical geofiltration model. Balance components of formation for a technogenic underground water mode, capacity parameters and zonal water permeability of a rock massif are estimated as a result of methodical adaptation and creation of a hydrodynamic model of two adjacent mines with a complex hydrogeological structure. A duration and speed for groundwater level restoration at different time periods, an area of underflooding and flooding on the surface is established according to predictive calculations. The inefficiency of submersible pumps use is proven when capital workings lose their drainage functions and a massif has rheological dynamics. Variants for long-term pumping of mine waters with subsequent discharge into a hydrographic network do not bring the territory into a self-draining mode. Therefore, the mode of closing mines with full restoration of underground water level in a monitoring mode with a combination of drainage reconstruction of the surface is the most acceptable.


Introduction
Along with the global factors of climate change and scarcity of fresh water resources, Ukraine has its own regional features of natural and technogenic character.According to water availability indices, Ukraine is a European country with a water supply deficit, where local water runoff per one citizen averages 1000 m 3 , while, for example, in Canada this figure is 94300 m 3 , in the USA -7400 m 3 , in Germany -1900 m 3 [1][2][3][4].At the same time, water availability in individual regions differs by almost a factor of 60: from 0.14 km 3 /year in the Kherson region to 7.92 km 3 /year in Zakarpattia [5].The Southeastern Ukraine has the least available water resources, though the largest water consumers are concentrated there, also the depth and scale of technogenic load on the hydrosphere is critical.For example, under the influence of drainage in mines of the Eastern Group in Western Donbas of Ukraine, the reserves of fresh water used for drinking and economic purposes have been depleted.The size of the cone of depression in the aquifer of overlying Buchak sediments is 11×37 km with a depth of maximum water lowering of 38.7 m.The Quaternary aquifer is drained on an area of about 9 km 2 , which caused damage to water supply for drinking and economic use in Mykolaivka, Petrivka, and IOP Publishing doi:10.1088/1755-1315/1348/1/012069 2 Katerynivka villages of Petropavlivskyi district.A rise in underground water level in the Mezhyhirskyi aquifer led to a complete shutdown and operation termination at the Samara water intake.According to data [6], as of 2018, all operational wells of Pershotravenskyi, Svitlohirskyi and Verbskyi water intakes have been conserved.
The problem of water supply became critical during the period of military aggression, when 80 % of drinking water supply is provided by surface sources, which are the most vulnerable during hostilities.The search for solutions regarding alternative sources of water supply has become a priority [7][8][9].
Under these conditions, the underground water resources, which are restored when mines are closed and flooded, cannot be neglected.But here it is necessary to have tools and to adapt the methodology of predicting hydrodynamic changes in groundwater mode for developing hydrotechnical and ecological water resource management schemes.Predictions of mine flooding in Europe are critically reviewed and compared with actual monitoring data in the paper [10].
Reliability of predictive calculations and, accordingly, controllability of processes depends on conception about transformation on filtration and capacity characteristics of mined-out areas during 40-60 years of a coal deposit operation.The vast majority of research is devoted to determination of filtration and capacity parameters at a stage of active shifts and formation of zones of waterconducting cracks.In addition, according to various data, filtration and capacity parameters vary by orders of magnitude, which means they are not universal and depend on specific geological, hydrogeological and technological conditions.Therefore, the goal of the article is to predict formation of a hydrogeological mode of underground waters within the adjacent mine fields of Yuvileina and Stepova mines in accordance with their closure schedule, considering possible options for water regulation and environmental protection measures on the surface.

Methods
The method of numerical modelling is the most reliable method of predicting a hydrodynamic mode for mine fields in the conditions of complex geological structure of mine fields and many unknown parameters of a disturbed rock massif.Advantages of numerical modelling are in the assessment of an underground water mode, complicated by almost any complex of natural and technogenic factorsfiltration heterogeneity in plan and cross-section, complex conditions at hydrodynamic boundaries, non-stationarity and nonlinearity of filtration processes, the introduction of internal sources-runoffs at different times, changes in permeability and capacity in time and many others.
Creation of a mathematical model of geofiltration for a mine field is preceded by a conceptual model in which a set of simplifications and assumptions about a complex real system is justified.At the same time, an adequate representation of an object should be ensured, including the main functions and avoiding secondary details.
Numerical methods for solving differential equations of geofiltration are implemented in practice with the help of appropriate specialized software [10].
Currently, the development level of computer technologies and equipment is so high that the issues of technical nature no longer arise as limiting or restraining factors of modelling.The issues of correct and complete representation of hydrogeodynamic objects in the model are of the greatest importance.That is why the PMWIN program (Processing Modflow for Windows, USA) is selected for simulation of filtration processes in a technogenically disturbed environment of mine fields.
Numerical modelling involves a grid breakdown of a filtration area and is based on a finitedifference method of solving differential equations, when partial derivatives are replaced by finite increments, and a differential equation is reduced to a system of algebraic equations.The process of non-stationary filtration of groundwater for a layered system of aquifers is described by a system of differential equations of the type [11] IOP Publishing doi:10.1088/1755-1315/1348/1/012069 where μ* is coefficient of specific storage, T = km is the water conductivity (k is filtration coefficient, m/h, m is aquifer capacity, m, H is groundwater head mark, m, t is time, h, qn is cross-flow between horizons, m/h.Unambiguity of solution for ( 1) is achieved by assigning boundary conditions, for which the spatial position and shape (point, contour, and plane), hydrodynamic conditions and its quantitative characteristics are established.
Boundary conditions of the first kind (H = const) reflect positions of hydro-and piezoisogypsums, as well as surface watercourses and reservoirs that are hydraulically connected with underground waters and are hydrodynamically perfect.In a case of imperfection, a parameter of additional filtration resistance of a reservoir bottom or a river bed is used, and then the boundary condition of the third kind Q =ƒ (Н) takes place, where Q is the cross-flow rate.
Mine workings are approximated by hydrodynamic condition of the third kind with an absolute mark of groundwater level of a productive stratum at the bottom of a productive stratum being developed, or the horizon of mining operations.
Boundary conditions of the second kind reflect a dependency of flow rate (Q) on coordinates and time.For conditions of mine fields, the boundary is watertight at Q = 0 and characterizes the zones of tectonic faults, which are the technical boundaries of a mine field or its blocks.In addition, water intake or absorption wells and sources are also reflected by the value of flow rate (Q = const).
When constructing a geofiltration model of a mine field, a certain algorithm of actions is performed, which involves considering the hydrodynamic openness of a deposit and a zone of waterconducting cracks when schematizing the conditions, as well as a mandatory method of identifying the model by reproducing the process of developing the deposit to establish the patterns of water inflows formation and changes in permeability parameters of the rock massif in time.
The effectiveness of using mathematical modelling method to assess a hydrodynamic mode of mine fields both at a stage of operation and mine closure is confirmed by approbation of the results at a number of facilities in the Central and Western regions of Donbas [11][12][13].

Natural and technogenic features of the research object
The leading factor in formation of natural and technogenic hydrodynamic environment of mine fields is a geological-structural factor.Accordingly, coal seams are split into types as "open" and "closed", which determines the amount of mine water inflows.Mine fields along the seam dip are broken into blocks by tectonic faults.
Stepova and Yuvileina mines are located in the eastern part of Western Donbas in Ukraine.They belong to the "semi-open type", where coal seams of both the "closed" and "open" types are developed."Closed" coal seams do not have a hydraulic relationship with water-saturated overlying deposits, and "open" coal seams do have this relationship within parts of a mine field (figure 1 [12][13]).Inclined parts of mine fields are hydrodynamically open (block 1, upper), and are separated by tectonic faults from the "closed" type blocks 2 and 3 (middle and lower).Confirmation of the main role of the geological-structural factor in a formation of mine field flooding is the actual data on the amount of water inflows into mine workings within mine fields (table 1), where the maximum water inflows are recorded in block 1.
Involvement of additional groundwater resources from overlying strata into drainage of Stepova and Yuvileina mines is confirmed by low mine water mineralization of 2.5-6.0 g/dm 3 (Stepova mine) and 2.0...3.7 g/dm 3 (Yuvileina mine) compared to mineralization of water in the nearby Dniprovska mine (up to 19 g/dm 3 ), which is hydrodynamically "closed".Within the mine fields of Western Donbas, there is a full cycle of observations over underground water mode, both natural and disturbed modes included (under the conditions of construction and mine operation).The results of mode observations for different periods of time are control data for establishing a compliance of the model with actual conditions within mine fields.

Model development and verification
The hydrodynamic model covers a territory of two mine fields -Stepova and Yuvileina, which are hydrodynamically connected through the overlying aquifer complex, and a coal seam is separated by a tectonic fault.Model dimensions, considering the existing influence of mining operations, are defined in plan using rectangular coordinates of 0.0...18800.0m in a latitudinal direction and 0.0...10800.0m in a meridional direction, the total area of the model is 126 km 2 .The research area is approximated using a grid of 9454 calculation blocks with a size of 200200 m, which made it possible to display a configuration of mine field areas and contours of natural hydrodynamic boundaries with sufficient accuracy.
In a vertical cross-section, the model of mine fields is represented by a 10-layer stratum (figure 2), consisting of covering Meso-Cenozoic deposits (calculation layers 1...3) and coal deposits (calculation layers 4...10), where the latter ones correspond to layers c6 and c6 1 and according to a tectonic scheme, they are represented by three blocks and a layer of rocks dividing them.
Methodology of geofiltration modelling involves solving epignostic or inverse problems in natural hydrodynamic conditions and conditions disturbed by mining operations in order to identify a model and a studied object.As a result of their solution, according to the position of groundwater levels in plan and cross-section, as well as the amount of water inflows into mine workings, boundary conditions of the model, filtration and capacity parameters of a rock massif are specified.The final assessment of identification results is performed based on data of a studied object functioning, i.e. the functional compliance is established, in which processes are reproduced on the model, the actual data on the course of which were previously recorded in natural conditions.To do this, drainage contours are set according to the plans of mining operations in periods where characteristic changes in the amount of water inflows are recorded, as well as data from mode observations.In order to solve the inverse non-stationary problems, 6 periods in operation of Stepova and Yuvileina mines are selected: 1963-1978, 1978-1986, 1986-1994, 1994-2002, 2002-2010, 2010-2020 (figure 3), for which the most characteristic trends in change of water inflows are recorded.The actual data on the amount of water inflows into mine workings, as well as the data of mode observations on a level of groundwater in the aquifer complex of Meso-Cenezoic sediments are used as controls.When reproducing a hydrodynamic situation on the model for time periods from 1963 to 2020, the deviation of water inflows on the model from the actual is 10-15 % under the conditions of zonal distribution of filtration parameters of Carboniferous stratum.Reliability of the results is also confirmed by coincidence of control data in observation wells and a position of groundwater level on the model (figure 4).According to calibration results, a deviation of estimated water level marks on the model from actually established water level marks in selected observation wells, equipped to cover the aquifer complex (a total of 13 wells), are determined by a standard error that does not exceed 0.5 m at a correlation coefficient value of 0.977.Based on the results of solving inverse identification problems, the following is established.1.The features of hydrodynamic mode of overlying sediments, composed of sandy-clay varieties of rocks, include a formation of a pronounced hydraulic depression on a basset contour of productive coal seams under overlying deposits.The latter is limited to the maximum in the southern part of a mine field of Stepova mine within the boundaries of calculation block 1 with absolute marks of a groundwater level in the center of 52.0 m (figure 5).Depth to a groundwater level of the first aquifer from the surface in the area of "Petropavlivski Lymany" landscape reserve is 0-1.5 m.
2. Formation of a hydrodynamic mode within mine fields occurs mainly due to groundwater feeding of overlying stratum (61.0%), to a much lesser extentcapacity reserves (19.2 %), inflow from adjacent territories (9.9%) and infiltration feeding (9.8%).The discharge of underground water is almost completely dominated by a technogenic componentthe mine drainage of Stepova and Yuvileina mines (97.3%), and there is only a small amount of discharge to a river network, reservoirs and watercourses (2.6%).
3. Destruction and decompaction of rocks in a zone of stopes, as well as presence of operating capital workings are considered by introducing increased indicators of permeability and gravity capacity (10 %) into the calculation.The specified percentages are determined according to a generally accepted practical modelling method by means of repeated simulation calibration of the model and correspond to an acceptable convergence of the object and model identification results.This methodical approach considers the total influence of massif fault by development periods and presence of residual volumes of mine workings.
Correctness of a determined parameter of gravitational capacity at a level of 10 % is confirmed: -considering the zone thickness of water-conducting cracks (up to 30 m), which exceeds the volume of mine workings; -compliance with an average value for highly fractured sedimentary rocks (shale, limestone, and sandstone).

Predictive calculations regarding the impact of closure and flooding of mines
A hydrogeological prediction of closure consequences for Yuvileina and Stepova mines is carried out according to four options.
In a case of complete liquidation and flooding of mines without water regulation (option 1) restoration (more than 83 % of the natural) level of underground water in the shaft of Yuvileina mine takes place within 5 yearsuntil 2028, in the shaft of Stepova minewithin the first three years until 2027 (figure 6).Further restoration of hydrodynamic mode in a coal massif occurs almost synchronously with a restoration of a level in the overlying sediments.
When groundwater levels are restored, considering the subsidence of undermined areas, the areas of flooding (with a depth to water of 0 m) and potential flooding of day surface (depth of groundwater level up to 1.5 m) are formed, covering the southern part of a floodplain of Samara River (figure 7 ).It should be noted that the selected area of predicted flooding and underflooding is drained under the action of mine drainage (figure 5), which is also evidenced by the results of a survey of local residents of Mykolaivka village, whose wells have no water at depths of 5.0...7.0 meters.That is, a change in a natural hydrogeological background of the southern part of a Samara River floodplain, including "Petropavlivski Lymany" reserve, took place precisely during the operation of the mine drainage.And when the mine is flooded, the level of underground water in the floodplain, including the area of the reserve, are restored to the natural pre-operational level.
However, in order to prevent flooding of the highway and part of Mykolaivka village, a possibility of long-term maintenance of a stationary drainage in block 1 of Stepova mine at the horizon of 145 m is considered according to option 2.
According to the calculated data, the maintenance of a stationary drainage under this option ensures 8 a stable groundwater level both in a coal massif and a water-bearing complex of Meso-Cenozoic sediments, which allows maintaining the existing level mode within a floodplain of Samara River.It should be noted that the operation of a stationary water drainage is effective under the conditions of preserving draining effect of capital mine workings at the horizon of 145 m.
In addition, maintenance of a stable level mode in this case is accompanied by an increase in water inflow in block 1 of Stepova mine by an amount from 7.0 to 12.0 % over the course of 20 years, which is associated with a general restoration of hydrodynamic mode within the mine fields of flooded mines (table 2).When using submersible pumps in the shaft of Stepova mine according to option 3, no significant effect of their operation is observed, which is associated with a leveling of a linear drain depression when it is flooded.This means that when flooded, the capital and preparatory workings lose their drainage functions.From comparing the volumes of a disturbed massif (30 extraction capacities of a coal seam) and a volume of capital and preparatory workings, it is determined that the hydraulic channels in a filtration massif make up less than 0.7 %.Considering the laminar flooding mode, this practically does not affect the predictive indicators.
During the continuation of operation in Yuvileina mine in conditions of closure and flooding of Stepova mine (option 4), water level recovery is similar to option 1, but the recovery dynamics in a water-bearing complex of overlying sediments on the mine field of Stepova mine is slowed.
Summarizing the above, the environmental friendliness of option 1 should be noted for restoring the natural conditions of "Petropavlivsʹki Lymany" landscape reserve and bringing the territory of mine fields closer to self-draining conditions.The complete restoration of groundwater level of the first aquifer from the surface within 10 years should take place in a monitoring mode with justification and arrangement of systemic water regulation measures (for example, clearing and deepening the bottoms of Kosminna and Velyka Sukha gullies within the Mykolaivka village).

Conclusion and recommendations
Model verification with an area of 126 km 2 is achieved by a simulated determination of gravity capacity with a correlation coefficient of 0.97 between the actual and model groundwater levels and a convergence (85-90 %) of water inflows, which made it possible to establish a significant relationship between rheological, filtration and capacity properties of a rock massif.
The created computer model of geofiltration for adjacent mine fields should be used as a permanent tool when making engineering decisions for hydro-eco-safe regulation of water resources in a minedout mine field or when synchronizing the cessation of mining operations and coal production.Given the established changes in drainage properties of a rock coal massif over time, one should focus on water regulation in a stratum over coal by reconstructing its drainage functions with achieving a selfdraining and recovery mode.

Figure 1 .
Figure 1.Hydrodynamic scheme of "semi-open type" coal seams.The inflow of water into mine workings of Stepova mine during operation varied from 215 m 3 /h to 1239 m 3 /h, at Yuvileina minefrom dozens to 1053 m 3 /h.

Figure 4 .
Figure 4. Calibration results of geofiltration model as of 2020level mode, inverse problem, non-stationary filtration.

Figure 5 .
Figure 5.A fragment of an area of mine fields in a zone of influence of mine drainage and formation of a depression cone.

Figure 6 .
Figure 6.Dynamics of restoring groundwater level when mines are flooded through option 1.

Figure 7 .
Figure 7. Predicted depth of groundwater level for a calculation period of 10 years according to the option 1.

Table 1 .
Formation of water inflows within mine field blocks in 2020.

Table 2 .
Predictive water inflows (option 2) when maintaining water drainage at the horizon of 145 m.