Thirty hydrofracturing tests on the premises of the Friedrich Heinrich mine in the city of Kamp-Lintfort were carried out at a depth of 586 m, around 1000 m south-west from the shaft IV, in two boreholes. The first borehole is vertical with a length of 40 m drilled in compact claystone–siltstone series. The second is a 37 m long horizontal borehole drilled towards the NW direction in compact claystone–siltstone series. The tests were located in Middle Witten formations within the Westphalian A stage of the Carboniferous period.

On the premises of the Niederberg mine in the city of Neukirchen-Vluyn, 27 hydrofracturing tests were carried out at a depth of 630 m in three boreholes from which two were horizontal and one vertical. The horizontal wells were drilled with azimuths of 120 and 97∘ N in both sandstones and slates, whereas the vertical borehole penetrated mainly sandstones.

On the premises of the Lohberg coal mine in the vicinity of the city of Dinslaken, 84 hydrofracturing tests were performed in 10 boreholes. Four test locations were located on the fifth level (i.e. at a depth of 1315 m), near shaft II, with a maximum horizontal distance between test locations of around 2500 m. All test locations except one were located south of the Bruckhauser fault. Drilled boreholes had a length between approximately 33 and 60 m and penetrated either compact sandstone or claystone series with interlayering coal series. The horizontal boreholes were drilled in both NE and E directions. All boreholes were drilled in Upper Bochum formations within the Westphalian A stage of the Carboniferous period.

In three boreholes in the General Blumenthal coal mine in the vicinity of the city of Recklinghausen, 34 hydrofracturing tests were carried out. The drilling operations took place on the ninth level of the coal mine at depth of 975 m directly below the Dickebank coal seam. The horizontal boreholes were drilled along both NW and NNE directions. All boreholes, each 40 m length, were drilled in the sandstone series of the Upper Bochum formations of the Westphalian A stage of the Carboniferous period.

Seventy-six hydrofracturing tests were performed at two depths levels, i.e. 750 and 998 m in 10 boreholes within the Haus Aden coal mine in the vicinity of the city of Bergkamen. The boreholes drilled at 750 m depth, two of which were vertical and two horizontal, had a length of 38 to 42 m and penetrated mainly compact medium-to-fine grained sandstones, claystones, and few coal seams. The horizontal boreholes at those depths were drilled in the NE directions. At a depth level of 998 m, six boreholes (two of which were horizontal and drilled in the NE direction) with lengths ranging between 40 and 76 m were drilled. Tests at depth of 998 m were carried out in the Middle Bochum formations, whereas tests at depth of 750 m were carried out within the Upper Bochum formations. The formations penetrated by the boreholes in the Haus Aden mine belong to the Westphalian A stage of the Carboniferous period.

In the Westphalia coal mine in the vicinity of the city of Hamm (Westph.), 149 hydrofracturing tests were carried out in 13 boreholes at two depths levels, i.e. 1030 and 1250 m, where the maximum horizontal distance between tests amounted to approximately 2100 m. At depth of 1300 m, four boreholes, two of which were horizontal (drilled in both NE and NW directions), with a length between 40 and 50 m were drilled. The penetrated formations included mainly compact or fractured sandstone and, occasionally, coal layers. At depth of 1250 m, nine boreholes with lengths between 39 and 43 m were drilled into compact sandstone layers and some coal layers. All tests were carried out in the Middle Bochum formations of the Westphalian A stage of the Carboniferous period.

As a part of a coal bed methane exploration program, two deep vertical boreholes (i.e. Natrap-1 and Rieth-1) were drilled around 20 km south from the city of Münster, north of the Ruhr region. In both boreholes, located approximately 17 km away from each other, hydrofracturing and permeability tests were carried out in cased and open hole borehole sections. In the Natrap-1 well, located approximately 1.5 km north of the city of Hoetmar, cased hole tests were carried out in three perforated sections between 1380 and 1949 m depth, within eleven coal seams, and open hole tests between 1418 and 1935 m depth. In total, 17 hydrofracturing tests were carried out in both Westphalian A and B stages of the Carboniferous geological period in the Natrap-1 borehole. In the Rieth-1 borehole, located around 4 km south-west of the city of Drensteinfurt, hydrofracturing tests were carried out in 10 perforated cased borehole sections at depths between 1082 and 1582 m and open borehole sections between depths of 1694 and 1705 m. In total, 12 hydrofracturing tests were performed in coal seams and coal-bearing formations in the Westphalian A stage of the Carboniferous geological in the Natrap-1 borehole. An estimation of hydrofracturing tensile strength and fracture breakdown pressure was only possible for open hole tests, whereas no tests of the orientation of SHmax were carried out in the deep boreholes, primarily due to the difficulty of estimating fracture orientation in perforated borehole sections.

To interpret stress magnitudes based on the pressure curves recorded during hydrofracturing tests conducted in anisotropic and fractured Carboniferous rock mass (Fig. ), inversion techniques following the classical hydraulic fracturing theory were applied. For its application, it has been assumed that (i) a borehole is aligned with a principal stress axis. Such an assumption will be valid mainly for vertical boreholes, located away from mine workings or in deep boreholes; (ii) rock mass is homogeneous and isotropic. Neglecting coal seams, the compact sandstone and siltstone layers fulfill this assumption, i.e. fracture propagation remains not affected by the properties of the rock mass; (iii) fracturing fluid does not penetrate the rock prior to the fracture initiation. As the permeability of the rock mass of the Carboniferous rocks is extremely low, this assumption remains valid; (iv) once the fracture is initiated, it propagates in the direction perpendicular to the orientation of the Shmin. As this assumption remains true for vertical boreholes, it may not be true for the horizontal boreholes in cases when a borehole is not aligned with a principal stress axis. In few cases, drilling direction of horizontal boreholes was selected based on results from a vertical borehole accounting for mine geometry restrictions. For other cases, the orientation of horizontal principal stresses was unknown prior to testing and drilling direction was limited by the mine geometry. In a case of a borehole not being aligned with a principal stress axis, fracture orientation was excluded from the further analysis.

A schematic example of a hydrofracturing test from one of the coal mines in the Ruhr region with estimates of the shut-in pressure during fracture closure, Psi, fracture reopening pressure, Pr, the rock mass (hydrofracturing) tensile strength, Pco, and the breakdown pressure at fracture initiation, Pc, is presented in Fig. a. Assuming negligible pore pressure, P0, in impermeable and compact Carboniferous rock mass within the Ruhr region, for a case of vertical boreholes, where a vertical fracture was induced, Shmin is assumed to be equal to Psi: 1Shmin=Psi.

For the case of vertical boreholes and vertical fracture, SHmax was computed as follows: 2SHmax=3Shmin-Pr-P0.

Utilising the bulk density, ρb, of the rock mass of 2500 kgm-3, based on ρb of the Ruhr sandstone , and the true vertical depth (TVD) of a test location, z, Sv was computed using the following: 3Sv=∫0zρb(z)gdz.

Although ρb of the Carboniferous rocks in the greater Ruhr region could be lower (in the case of coal seams) or higher (in the case of siltstones or shales) than 2500 kgm-3, this ρb value is considered as a good first-order average approximation for the study region.

In a case of a horizontal borehole aligned with SHmax orientation and a vertical fracture induced, estimation of Shmin was calculated according to Eq. (). If a horizontal borehole was drilled along the SHmax orientation, instead of estimating SHmax magnitude as in the case of a vertical well, Sv magnitude was estimated instead, 4Sv=3Shmin-Pr-P0.

For the case of horizontal boreholes aligned with Shmin orientation and a radial fracture, Shmin estimation follows Eq. (). The same goes for the estimation of Shmin from horizontal boreholes aligned with Shmin orientation and a vertical fracture. Additionally, in this particular case, Sv was calculated with Eq. (). For a case of horizontal boreholes aligned with Shmin orientation and a horizontal fracture, Sv was assumed to be equal to Psi, 5Sv=Psi, whereas SHmax, for this particular case, was computed as follows: 6SHmax=3Sv-Pr-P0.

The results from vertical boreholes can be considered to yield better quality than horizontal boreholes, due to the uncertainty connected to their alignment with a principal stress axis.

In deep boreholes, the shut-in pressure was determined using a three-step analysis of the pressure plots which included: (i) pressure vs. flow rate plot, where the moment at which flow is stopped was used to estimate an upper bound on Psi; (ii) the so-called Muskat pressure plot for estimating the lower bound on Psi, assuming that the linear part of the plot characterises radial flow (i.e. stimulated fracture is closed); and (iii) within the two limits, Psi value marks the transition from a rapid linear pressure drop to a diffusion dominated pressure decrease, where the transition can be determined by a tangent (i.e. inflection point) method (Fig. b). In some hydrofracturing tests, determination of Psi could be only carried out through pressure versus flow rate plots. Considering the system stiffness, Pr was constrained based on the deviation of the linear pressure versus injected volume plot (Fig. c) which indicates fracture opening. Pc was determined as the maximum pressure registered during the fracture initiation phase of a hydrofracturing test (Fig. d). As the analysed deep boreholes were drilled vertically, Eqs. () and () were used to estimate Shmin and SHmax, respectively. Equation () and an assumption of ρb of the rock mass of 2500 kgm-3 was used to estimate Sv from the two deep boreholes.

The detailed results of Psi, Pr, Pc, and Pco averaged within each borehole and their uncertainties are presented in Table . Both Psi and Pr show relatively small uncertainties amounting to values between 2 and 4 MPa and present no significant variations, except increasing magnitudes with depth. Characteristic pressure peaks for fracture reopening pressure during subsequent slow pumping rates were observed. Pressures only slightly decreased during shut in, which is an indication of extremely low rock permeability. Pc, on the other hand, shows significant variations due to local rock strength variations and relatively high uncertainties of about 5 MPa on average. Pc values are high, approaching, at times, technical limits of the testing tool. The tensile strength of the rock mass, computed from the in situ stress tests, can be considered as high, with an average tensile strength of 6.5 MPa for the study region. An increasing trend of Pco with depth was observed.

The results of the determined in situ stress magnitudes are presented in Table . Based on the collected data, it can be observed that Shmin is significantly lower than the vertical stress (with an average ratio of 0.6), whereas SHmax is on average around 1.9 times higher than the Shmin, proving high differential horizontal stresses at depth in the studied area. The vertical stress derived from the horizontal boreholes agrees with the vertical stress computed for the overburden bulk density of 2500 kgm-3, with a few tests slightly exceeding this value. The SHmax magnitude is higher than the vertical stress with an average ratio of 1.2. However, there is an always-present uncertainty related especially to the estimation of SHmax based on the classical hydraulic fracturing approach used in this study, values of SHmax from both coal mines and deep boreholes are comparable, and therefore should be treated as reliable. Generally, the studied region represents strike-slip stress regime, where Shmin<Sv<SHmax (Fig. a). Based on the average stress values from each test location presented in Table , gradients of Shmin, SHmax, and Sv valid for depths between 0.6 and 1.7 km with their coefficient of determination, R2, are as follows: 7Shmin=0.0134z+1.2893;(R2=0.74),8SHmax=0.0248z+3.9588;(R2=0.52),9Sv=0.0234z+1.2675;(R2=0.92).

It can be observed that at depths between 1000 and 1300 m, the horizontal stresses especially are slightly lower than for tests carried out at shallower or greater depths, pointing towards a more extensional stress regime. Based on Fig. , which presents a normalised stress polygon with averaged results from each location (differentiating depth levels) computed within this study (Table ), and assuming negligible pore pressure, one can see that the majority of stress tests fall predominantly into a strike-slip stress regime and only a few present normal stress regime. Looking at Fig. b, which presents a so-called mean stress ratio, k (i.e. a ratio of average horizontal stress, Sh, and Sv), it could be concluded that stress does not change significantly at depth and that an average k value of 0.86 represents the studied area most accurately. For comparison, stress ratio computed based on the approach by for Young modulus, Esh, of 36 ± 11 GPa, which was based on 29 static measurements on fine-grained sandstone core samples extracted from the H13 borehole from the Westphalia coal mine at 1250 m depth , is presented in Fig. b.

After data evaluation, the quality ranking scheme developed by was applied for the derived Shmin magnitudes. The test results collected within this study and ones from already published studies are summarised in Table . In total, 429 hydrofracturing tests were carried out during the measurement campaign. Based on these tests, 429 unique Shmin data records were derived, from which 367 received the highest, i.e. A-quality, 19 data records received C-quality (due to the tests being carried out in cased borehole sections), and 43 data records where it was not possible to derive the Shmin magnitude received E-quality (due to either no pressure build-up in the tested interval or an estimation of Sv magnitude, instead of Shmin, in the case of several horizontal boreholes).

Furthermore, based on 429 hydrofracturing tests, 188 data records of SHmax magnitudes and 341 data records of Sv magnitudes were derived. The magnitudes of Sv are a combination of stress magnitude values derived from hydrofracturing tests for horizontal boreholes and estimations of Sv magnitudes based on the bulk density of the rock mass of 2500 kgm-3 for vertical boreholes (Table ). The stress data record quality assessment developed by , as of today, refers to Shmin magnitudes only, and therefore no quality was assigned to SHmax or Sv magnitudes.

Following the WSM data assessment guidelines and the WSM quality ranking scheme , an estimation of the mean orientation of the induced fractures, assumed to be equal to the mean SHmax orientation, was carried out in each borehole. For the estimation of the mean SHmax orientation and its standard deviation from each borehole needed for the quality assignment , the statistics of bi-modal data were utilised.

From 429 hydrofracturing tests, in 254 cases the orientation of the induced fractures was measured. From this data set, 38 SHmax orientation data records were derived. These acquired SHmax orientations increase the number of data records in the greater Ruhr region from 49, which are already available in the WSM database release from 2016 , to 87. In some hydrofracturing tests, several fractures were observed at the same test location. Once these fracture orientations differed from each other by more than 25∘, the data record was not taken into account for further analysis. This procedure was performed to exclude data records with high uncertainties of fracture orientations being an actual indicator of the SHmax orientation. The resulting SHmax orientation, including its standard deviation using the statistic of bi-modal data, is provided in the second last column of Table . Figure a shows the final stress map of the greater Ruhr region with the distribution of SHmax orientations, whereas Fig. b and c present rose plots of SHmax orientations from this and already published studies. Since the boreholes in the six investigated coal mines areas are close to each other, derived SHmax orientations are represented with six polar plots presenting mean values (solid black line) and its standard deviation (dashed black line) from each mine.

As presented in Table , the uncertainty of six SHmax orientations is considered as high, as it is either equal to or higher than 40∘. These data records were assigned the lowest i.e. E-quality according to the quality assessment by . Due to the short length of the tested intervals (i.e. less than 40 m), the rest of 32 SHmax orientation data records from hydrofracturing tests were deemed to be of D-quality according to the WSM quality assessment .

To showcase how a data set presented in this study can be utilised, slip and dilation tendencies were calculated for the major geological discontinuities discovered throughout the 700-year-long coal mining activities in the Ruhr region . The average SHmax orientation of 161∘ obtained from this study and Eqs. (), (), and () were used to constrain the mean in situ stress state at a depth of 1200 m, representing the coal mining depth. The assumption was made that the pore pressure starts from the surface level and that the pore fluid has a constant density of 1000 kgm-3. Another assumption, caused by the lack of data on fault geometry, assumes that all evaluated discontinuities are vertical. This may not be always true in reality, i.e. fault's dip may significantly differ (e.g. Fig. c) from the one assumed in this study. It is therefore advised to treat results of the slip and dilation tendency in a more relative manner rather than as absolute values. To address the full range of uncertainties on slip and dilation tendency values of major faults in the greater Ruhr region, including in situ stress tensor and fault geometry, a probabilistic assessment should be carried out in the future studies e.g.. Based on the abovementioned assumptions and stress gradients developed in this study, Pp of 11.8 MPa, Shmin of 17.4 MPa, SHmax of 33.7 MPa, and Sv of 29.3 MPa were used. Based on the methodology presented in , normal effective and shear stresses were computed for each discontinuity segment and simultaneously slip and dilation tendencies were calculated. Results of these computations are presented in Figs.  and . As it can be seen in Fig. , where the colour of a discontinuity segment represents different slip tendency values, the N–S and NW–SE-striking discontinuities (with azimuths of 5 and 137∘ N, respectively) have the highest slip tendency values. This indicates that these structures are the most prone to be reactivated (either a- or co-seismically) by pressure and/or temperature changes created during, e.g. geothermal fluid production or fluid injection. On the other hand, discontinuities that are considered “locked” in the prevailing stress state are the ones striking in the ENE–WSW and NNW–SSE directions (with azimuths of 71 and 161∘ N, respectively). Figure  presents the results of dilation tendencies for the discontinuities within the Ruhr region, where the colour of a discontinuity segment represents different dilation tendency values. It can be observed that the NNW–SSE-striking structures (with an azimuth of 161∘ N), i.e. ones striking along SHmax orientation, will be the ones that are the most likely to stay open (if not filled by, e.g. secondary fluid mineralisation) in the prevailing stress regime. These geological discontinuities will therefore be the most permeable ones within the region, and could be considered as potential targets for, e.g. establishing a geothermal systems. On the contrary, the hydraulically “dead” discontinuities will be the ones striking ENE–WSW (with an azimuth of 71∘ N). These impermeable structures could lead to, e.g. reservoir compartmentalisation.