Hydrogeological characterization of low permeability medium for conceptualization of a mine site in Eastern Turkey

To obtain the hydrogeological properties of lithological units, a total of 11 small diameter (PQ which is 3.34 inches in diameter) drill holes and a total of 10 large diameter wells (17.5 inches in diameter) were drilled in the designated critical cross-sections (Fig. 1). The small diameter wells have been cored to specify the geotechnical properties of lithological units. Also, during the drilling process of small diameter wells, lugeon tests were conducted at each well to obtain permeability profile on the critical cross-sections. Following the drilling, vibrating wire piezometers (VWP) were installed at different depths to measure the pore water pressure at different zones of lithological units. It is seen that in Fig. 1, some wells do not coincide with corresponding critical sections (e.g. Pe1, Pb5). The reason for this is that the conditions of the mining area can change quickly and sometimes drilling could not be done at the planned points. In such cases, the wells are drilled as close as possible to the planned points.

In addition to the small diameter drill holes, seven out of 10 large diameter wells were drilled into carbonate rocks and three wells (groundwater sampling wells) into metasediments and diorite to characterize the groundwater conditions. The main purpose of drilling seven large diameter wells was to determine the hydraulic head values at both ends of cross-sections in the carbonate rocks surrounding the mine site to define the hydraulic boundary conditions. The aim of drilling three groundwater sampling wells is tapping three different permeable zones in the metasediments and diorite. Water samples were collected from these zones to determine the origin and recharge areas of the groundwater in different depths by using stable isotope analysis. The aim of stable isotope analysis is to determine the water source and recharge areas which is very important for judging whether there is a hydraulic connection between the metasediments and diorite unit.

Hydraulic head distribution along critical sections were monitored in both large-diameter wells and small-diameter drill holes. A contact gage was used for monitoring on a weekly basis in all wells from September 2017 to July 2018, followed by monitoring on a monthly basis.

Carbonate rocks, metasediments and diorite are the three main lithology exposed in the mine site. Hydrogeologically, metasediments and diorite are impermeable units compared to carbonate rocks. For regional scale, this classification, based on the lugeon test results, was assumed sufficient to characterize the regional groundwater flow system. Although these units have low permeability, the total porosity and specific retention are high. This results in high pore water pressure due to the lack of significant groundwater movement.

The 11 small diameter holes were used to obtain geotechnical data from cores; a representative geotechnical log of small diameter drill hole is given in Fig. 2. In logs, the lithological description of metasediment and diorite is made in detail to include faults, clay zones, skarn, gossan, quartzite etc. Similarly, texture was defined as massive, brecciated, foliated, sheared, fault etc. The grain size was classified as very fine/aphanitic to very coarse. Porosity of lithological units has been classified as non-porous to highly porous or with vuggy porosity. Permeability values obtained from lugeon tests were converted to hydraulic conductivity value in m/s and plotted next to RQD values.

The Lugeon test is one of the in-situ tests applied to determine hydraulic conductivity values especially in fractured rock mass (Tesema and Ekmekci 2019). The test is the most frequently used method to obtain the variation of the hydraulic conductivity in vertical profile, especially in heterogeneous media. Lugeon tests have been conducted in 11 small diameter drill holes. Test intervals were set up to 2 m to obtain a high-resolution characterization. The design and the application of the test were completed according to Houlsby (1991). Five pressure stages were applied at each test interval for at least 10 minutes and the water intake has been recorded regularly. In case of significant difference in water intake records, the test prolonged another 10 minutes to ensure the quasi steady-state conditions. The readings were evaluated in Lugeon Unit (Lu) for each of the five pressure stages. The lugeon values were then converted to hydraulic conductivity values (m/s) according to Houlsby (1991) and Roeper et al. (1992).

Determination of aquifer characteristics (e.g. Transmissivity, Storativity, Hydraulic Conductivity) is the main purpose of characterization studies. Pumping test is one of the widely applied in-situ tests to determine aquifer characteristics (Butler Jr 1990; Fetter 2001; Kruseman et al. 1970; Mishra et al. 2013; Walton 1970). Pumping tests have been planned to obtain hydraulic characteristics of the carbonate rock aquifer and the metasediments-diorite unit. Tests were carried out in six wells. These are Dwa-3, Dwb-2, Dwf-1, GSW-1, GSW-2 and GSW-3 wells. The test could not be conducted in other wells drilled in the carbonate rock unit because they were dry.

A total of 36 vibrating Wire Piezometers (VWP) were installed at different depths in small diameter drill holes to obtain the pore water pressure distribution at the study site. The VWPs were calibrated before installation and installed in fully grouted boreholes. The grouting material is prepared according to ASTM D5092 which consists of mix of water-cement and bentonite. Contreras et al. (2008) and Vaughan (1969) discussed that the pore water pressure measurement error is insignificant even if there is too much difference between grout permeability and ground permeability. The readings in kilopascal (kPa) were converted for barometric pressure and finally, hydraulic head values in meters above sea level (masl) were calculated using pressure head and elevation head values for each VWPs.

Stable isotopes provide useful tools to determine the possible recharge area and to explain the process affecting the groundwater flow system (Fackrell et al. 2020; Gat 1996; Toulier et al. 2019). A total of 78 samples were collected from wells, rainwater, surface water and springs. The sampling was not limited to the mine site, but it was performed on a regional scale. The analysis was done in the Stable Isotope Laboratory of Hacettepe University by Liquid Laser Analyzer of Los Gatos Inc.^TM.

Logging the large diameter wells in carbonate rocks that are fractured and partially karstified has been problematic due to loss of drilling mud. The circulation has been lost and therefore no cuttings have come out of the hole, and it is important to state that the complete loss of the drilling mud starts at a certain depth which indicates the presence of a karstic cavity/conduit. The estimates made over the elevation zones where complete loss of drilling mud starts may indicate karstic level.

According to the overall assessment, elevation of karst is around 1150 masl in the northern part of the mine, while it is around 1250 masl in southern part. In eastern and western parts, karstic zone elevation is estimated around 1050 masl and 1200 masl respectively. However, it is not clear that those zones indicate a productive karst aquifer or not.

The geotechnical logging was conducted in 11 small diameter drill holes to characterize the metasediments and diorite units. In the logs, detailed lithological description of metasediments and diorite is given with other information such as RQD, grain size, porosity, and hydraulic conductivity values.

The primary (intergranular) porosity of the diorite is very low, whereas it may be higher than 5 % in the massive metasediments. The altered metasediments and diorite have similar and higher intergranular porosity. The intergranular pores are saturated with water and due to their low hydraulic conductivity, the pore water is not easily drained by gravity forces. Fracture porosity has also developed in the metasediments and diorite. Due to high porosity and low specific yield, nearly all of the pore water is retained by electrostatic and capillary forces in the material. It resists flow because of the low hydraulic conductivity. Permeability values obtained from lugeon tests plotted next to RQD values. In general, RQD values are inversely related to permeability values. High RQD values are generally matching up with massive and foliated textures, on the other hand, brecciated and fault texture types are correlated with low RQD values. Grain size and porosity does not show variation in general so that in almost all piezometer wells, preponderant types are fine and very fine for grain size, non-porous and vuggy for porosity. It is important to note that several zones in the lithological logs have been described as faults. Faults are one of the main structural elements that control groundwater flow. Therefore, their existence, continuity in terms of depth and length, their width, the filling material, and its permeability need to be defined for a representative hydrogeological characterization. In the geotechnical logging practice, the number of fractures per geotechnical length has also been defined. Overall data shows that metasediment and diorite units have similar number of fractures, and these units were characterized by a mean fracture frequency (FF) of 15.24 fractures per run length in meters. Descriptive statistics of fractures for the drill holes are given in Table 1. In the table, all piezometer wells except Pb3 showed similar characteristics. The coefficient of variation of individual drill holes is generally very high. This suggests that the lithological unit is far from being homogeneous. The fracture frequency was correlated with the permeability values obtained from lugeon tests. Although in general, the two parameters are found to be compatible, they are not in some zones, where the fractures are filled with thick residue clayey material.

In the study area, 882 lugeon tests were performed in 11 piezometer wells. The test range is set to 2 m to achieve a high-resolution characterization. The tests are constructed and applied as described in the previous parts of the paper. Lugeon (Lu) values were converted to hydraulic conductivity (m/s) values by assuming a constant radius of influence. The Lugeon values were evaluated in critical sections to characterize the medium. Histograms of the results have shown significant vertical changes of permeability in the wells. Example histograms showing the results of a piezometer well Pa4 in the study area are given in Fig. 3. Values greater than 50 Lu were carefully evaluated and marked in red on histograms. Overall, the permeability decreases with depth. According to the results, medium characterized the first 30-50 m depth from the surface has high permeability, medium-low permeability between about 50-120 m and low-very low in sections deeper than 120 m. In general, high permeability values were obtained in shear zones at shallow depths. On the other hand, low-very low permeability values were calculated at depths greater than 100 m.

The evaluation of Lugeon test results according to lithologies has shown that metasediments and diorite units have similar hydrogeological characteristics and can be accepted as a single hydrostratigraphic unit. The statistics of the Lugeon test results in Lu and in m/s for each well are given in Table 2 and Table 3 respectively. The coefficient of variation in small diameter drill holes are very high, values changing between %33-%518, indicating a great heterogeneity. The Lugeon test results were drawn on the critical cross-sections as hydraulic conductivities (m/s) for the hydrogeological characterization. Cross-section A is given as an example in Fig. 4. Two large diameter wells were also drilled in the carbonate rocks at the both end of cross-section. On these wells, representative groundwater measurements were put as blue marks. At various depths, there are some permeable zones that can be observed. However, it is not clear that the permeable zones belong to the same extensive zone.

In order to obtain hydraulic characteristics of carbonate rocks and the metasediments-diorite units, pumping tests were conducted at six large diameter wells. A list of the tests performed in the large diameter wells are summarized in Table 4. As shown in the table, constant discharge has continued only couple of hours and a long recovery period has followed. The hydraulic conductivity values changing between E-08 to E-10 m/s. Hydraulic conductivity values converted from transmissivity values indicate a low permeability medium. The results obtained from pumping tests are lower than the lugeon tests. It is an expected result because lugeon test represents more local characterization. The pumping tests are representative for larger scales. A fracture in the close vicinity of the packer test zone may give high permeability for the short test period. Pumping tests are more representative for the area of influence of the well. As can be seen in Table 4, model types were fitted to the graphs representing the unconfined and/or leaky aquifers. This was also confirmed by geological and hydrogeological field observations.

Also, hydraulic head measurements from VWPs prove that these sections are also unconfined in general. Similar results from different models for the same aquifer type have given similar results which confirms the interpretation above. The metasediments-diorite unit may be under unconfined conditions at the uppermost section where the material is altered and intensely fractured. As revealed by the packer tests, a permeable zone is underlined by a less permeable zone. Another permeable zone beneath this low permeability zone may form a leaky or confined aquifer. However, the hydraulic heads measured at the VW piezometers suggest that these sections are also unconfined in general.

A total of 36 VWPs have been installed to characterize the pore water pressure distribution along the critical cross-sections. The vertical change in pore water pressure in metasediments-diorite unit has been recorded at the VWPs installed at different depths in the drill holes located along the critical cross-sections. The pressure heads measured at each VWP installed in drill holes were illustrated on cross-sections to give an idea on the actual pore water distribution with respect to the current situation and the ultimate pit bottom topography. A representative cross-section is given in Fig. 5. As seen in the figure, downward hydraulic gradient exists in Pa1, and Pa4, while in Pa2, the gradient is upward between VWP-3 (third piezometer of Pa2) and VWP-4 (fourth piezometer of Pa2) only. Similarly, in Pa3, again an upward gradient is observed between VWP-2 (second piezometer of Pa3) and VWP-1 (first piezometer of Pa3). The figure also suggests that the gradient is not uniform in the vertical profile confirming the heterogeneity of the medium. At the south-west part of the section in Fig. 5, hydraulic head distribution forms groundwater divide and it can be seen better in the conceptual model. In general, a downward hydraulic gradient was observed in almost all piezometers, but the gradient is not uniform.

The results of the stable isotope analysis showed that the deuterium and oxygen -18 (O-18) is characterized by a deuterium (D) excess of 14. The local meteorological water line (LMWL) was obtained from the equation of D=8(O-18) +14. The graph of oxygen-18 vs deuterium is given in Fig. 6. The plots of rainwater were shown as green dots, and the groundwater were shown as blue dots. The rainwater plots over a wide range. Rain may be quite rich with respect to heavy isotopes which indicates the summer rain and while the winter rain is much depleted and forms one of the end members. As can be seen in the figure, the groundwater plots between winter rain and the warmer period rain suggesting that groundwater is mainly recharged by spring rains.

The elevation–stable isotope regression equations were derived for both oxygen-18 and deuterium (Fig. 7). From both equations, different elevation estimations were obtained (Table 5). The results show that the metasediments and diorite are recharged from between 1230 masl and 1400 masl on average. Even with the highest area considered, the recharge area falls within the slopes of the carbonate rocks surrounding the mine site.

The groundwater system of the mine site consists of carbonate rocks surrounding the metasediments and diorite units in the center. The in-situ tests and measurements showed that the metasediments and diorite units act like single hydrostratigraphic unit compared to carbonate rocks. A general expectation of groundwater system of the mine site is that the carbonate rocks supply inflow toward center of the mine. However, ore drillings on the southern side of the mine have shown that a diorite intrusion acts like an impermeable barrier against groundwater flow (Fig. 8). According to the cross-section given in Fig. 8, the southern carbonate rock mass is divided into two parts: the larger part forming the main carbonate rock aquifer. The small part between the diorite intrusion and the metasediments cropping out at the mine site has small area for recharge. In this part one should not expect thick saturated zone. Most probably the thin saturated zone is being discharged toward the west to contribute to the springs at the flood plain of the Karasu River. This model also explains the dryness of the wells drilled in carbonate rocks in this part of the area as discussed in the previous sections.

Based on the available and recently obtained data, a hydrogeological conceptual model was constructed on a cross-section, which can be regionalized to represent the mine site (Fig. 9). The metasediments and diorite have high total porosity and low specific yield indicating that the pore water is retained by electrostatic forces in the medium and it resists flow due to low hydraulic conductivity. The vertical change in the hydraulic conductivity values suggests that the medium is heterogeneous.

The medium consists of different water bearing levels. The shallow layers are in under unconfined conditions and deeper layers are in leaky conditions. A hydraulic divide is also noted around the Pa4 piezometer which causes groundwater flow to take place to SSW and NNE directions. This interpretation was concluded as a result of piezometer readings.