Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Contents lists available at ScienceDirect

Journal of Rock Mechanics and

Geotechnical Engineering
journal homepage: www.jrmge.cn

Full Length Article

A complementary approach to quantify the basic GSI chart considering

scale effect on rock structure
Harun Sonmez a, *, Gulseren Dagdelenler a, Yilmaz Ozcelik b, Murat Ercanoglu a
a
Applied Geology Division, Geological Engineering Department, Faculty of Engineering, Hacettepe University, Beytepe, Ankara, 06800, Türkiye
b
Mining Engineering Department, Faculty of Engineering, Hacettepe University, Beytepe, Ankara, 06800, Türkiye

a r t i c l e i n f o a b s t r a c t

Article history: Geological strength index (GSI) has been widely used as an input parameter in predicting the strength
Received 30 September 2023 and deformation properties of rock masses. This study derived a series of equations to satisfy the original
Received in revised form GSI lines on the basic GSI chart. Two axes ranging from 0 to 100 were employed for surface conditions of
22 January 2024
the discontinuities and the structure of rock mass, which are independent of the input parameters. The
Accepted 31 January 2024
Available online 5 August 2024
derived equations can analyze GSI values ranging from 0 to 100 within ±5% error. The engineering di-
mensions (EDs) such as the slope height, tunnel width, and foundation width were used together with
representative elementary volume (REV) in jointed rock mass to define scale factor (sf) from 0.2 to 1 in
Keywords:
Basic GSI chart
evaluating the rock mass structure including joint pattern. The transformation of GSI into a scale-
Engineering dimension (ED) dependent parameter based on engineering scale addresses a crucial requirement in various engineer-
Hoek and Brown failure criterion ing applications. The improvements proposed in this study were applied to a real slope which was close
Quantification of GSI to the time of failure. The results of stability assessments show that the new proposals have sufficient
Quantitative GSI chart capability to define rock mass quality considering EDs.
Scale effect © 2025 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Published by Elsevier B.V. This
is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/
4.0/).

1. Introduction of rock masses. GSI chart has then been updated by a series of
studies performed by Prof. Hoek and other co-authors (Hoek and
Determination of the mechanical properties of the rock masses Brown, 1997; Hoek and Marinos, 2000; Marinos and Hoek, 2000,
is of great importance in engineering applications. However, it is 2001). The GSI chart proposed by Hoek and Marinos (2000) was
almost impossible to directly measure the strength and deforma- known as the basic GSI chart. As stated by Hoek et al. (2013), the
tion properties of rock masses in field including joint patterns. To basic GSI chart was constructed on the observations performed by a
overcome this difficulty, empirical tools have been a highly qualified and experienced geologist or engineering geologist.
attractive research topic among the rock mechanics community, Therefore, the use of basic GSI charts requires engineering judg-
particularly since the 1950s. As one of the valuable studies on ment and experience to quantify rock mass to be used for deter-
empirical tools, the Hoek-Brown criterion has been widely used in mination of its strength and deformation behavior. On the other
determination of the overall strength of rock masses for different hand, as mentioned by Hoek et al. (2013), the mapping of rock
types of rocks, which has gained widespread utilization in engi- masses or drilling cores is carried out by professionals who are less
neering applications (Hoek and Brown, 2019). Since the initial form familiar with these qualitative descriptions in some rock engi-
of GSI was introduced to the Hoek-Brown criterion in the early neering projects. While we respect the practitioners’ preference for
1990s, it has been widely used as a popular input parameter not GSI with sufficient experience, scientifically it is necessary to have a
only in the Hoek-Brown criterion but also in some other empirical quantification guide for selection of the GIS based on engineering
equations for prediction of the strength and deformation properties scale and importance.
Although the first attempt at the quantification of GSI was
performed by Sonmez and Ulusay (1999), the number of studies has
* Corresponding author. progressively increased since the basic GSI chart was first intro-
E-mail address: [email protected] (H. Sonmez). duced to the rock mechanics literature by Hoek and Marinos
Peer review under responsibility of Institute of Rock and Soil Mechanics, Chi- (2000). There is no doubt that each study in the literature has
nese Academy of Sciences.

https://doi.org/10.1016/j.jrmge.2024.01.029
1674-7755/© 2025 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

contributed to the improvement of the quantification of GSI. Un- GSI lines without considering the possible scale effect on the
fortunately, due to the use of different assumptions carried out in structure of rock mass.
almost every study, different GSI values can be obtained from each
quantification effort of GSI. Furthermore, another significant
outcome obtained from all these valuable contributions to the 3. Quantification of GSI lines of the basic GSI chart
quantification of the basic GSI chart is a necessity for clarification
on the scale effect, considering rock mass structure depending on The basic GSI chart has two axes, i.e. rock mass structure
the engineering dimensions (EDs). including five categories and discontinuity surface condition
As is well known, the scale effects on the use of Hoek-Brown ranging from “very poor” to “very good”. It is employed in devel-
failure criterion have been continuously emphasized since the first opment of the generalized formulations of GSI lines. In the quan-
introduction of Hoek-Brown failure criterion. A novel contribution of tified form of GSI by Hoek et al. (2013), while only RQD is used for
the equations proposed in this study for determination of the value defining the jointing degree of rock mass, the joint condition
of GSI is the ability to consider ED based on the scale effect by Hoek (JCond) is quantified by JCond76 and JCond89 obtained from the
and co-authors’ original recommendations. In this study, a series of ratings of 1976 and 1989 versions of RMR (Bieniawski, 1976, 1989).
equations for the calculation of a GSI value on the original GSI lines The value of GSI can be determined by summing the value obtained
are derived based on two axes, i.e. rock mass structure and surface from the horizontal axis (Scale A, which depends on joint condition
properties of discontinuities. The results shows that the GSI value as 2JCond76 and 1.5JCond89) and the value obtained from the ver-
obtained from the proposed equations can satisfy an error range tical axis (Scale B, which depends on RQD, linked to rock mass
within ±5% when compared with the original GSI lines. structure as RQD/2). The quantification approach of GSI used by
Hoek et al. (2013) is quite simple. However, as stated by Hoek et al.
(2013), redrawing the GSI lines parallel to each other was a minor
2. Brief of the geological strength index (GSI) modification to the basic GSI chart proposed by Hoek and Marinos
(2000). Although this modification was presented as a minor
The first version of Hoek and Brown's empirical failure criterion modification, it may cause the determination of slightly different
was introduced in 1980 (Hoek and Brown, 1980), including two GSI values when compared to the basic GSI chart. From a mathe-
constants, i.e. “mb” and “s”, which control the applicability of the matical perspective, defining fully parallel GSI lines is, of course, a
criterion from intact rock to heavily jointed rock masses. Rock mass simpler approach. However, the difference between the original GSI
rating (RMR) values obtained from Bieniawski's classifications of lines (GSIbasic) of the basic GSI chart and the fully parallel GSI lines
RMR76 and RMR89 (Bieniawski, 1976, 1989) were used to calculate (GSI2013) may be considerable for particularly weak to very weak
the “mb” and “s” constants. Later, the parameter GSI was introduced rock masses. The GSI lines on both the basic GSI chart and the
as the criterion to overcome the insufficiency of RMR, particularly GSI2013 chart were digitized to create two GSI matrices with a res-
for characterizing weak rock masses (Hoek, 1994). The latest olution of 160  201. For a simple and effective comparison, the
version of the GSI chart was published by Hoek and Marinos difference between fully parallel GSI lines (GSI2013) and original GSI
(2000). Marinos and Hoek (2001) introduced heterogeneous and lines (GSIbasic) is investigated by considering the difference in both
tectonically deformed sedimentary rocks to a GSI chart. The percentage and quantity (see Figs. 1 and 2).
extended GSI chart for heterogeneous and tectonically deformed Although Hoek et al. (2013) described the transformation of the
sedimentary rocks was published by Marinos (2017) and Marinos original GSI lines into fully parallel lines as a minor modification,
and Carter (2018). this change can lead to an increase in GSI values by up to 25%,
The necessity of the use of quantitative inputs in the GSI chart particularly when GSIbasic < 50 (see Figs. 1 and 2). On the other
was first proposed by Sonmez and Ulusay (1999). The quantitative hand, considering that the original GSI lines are drawn based on
two scales, i.e. structure rating (SR) and surface condition rating experiences from case studies, it is reasonable to preserve the
(SCR), were introduced to the basic GSI chart to minimize the ne- original GSI lines as much as possible in the quantification studies.
cessity of experience and/or judgment, which opens to misleading In this study, the original GSI lines have been preserved as much as
assessments. In the study of Sonmez and Ulusay (2002), the first possible in the quantification of the basic GSI chart.
version of the quantitative GSI chart was re-arranged by including The innovative approach by the sum of “Scale A00 and “Scale B00
the upper rock mass class named “Intact or Massive”. In addition, an used by Hoek et al. (2013) was followed in this study to derive a
equation between SR and Jv (see Eq. (1)) for easy calculation is series of equations of the original GSI lines on the basic GSI chart. As
introduced. Dinc et al. (2011) improved the determination of SR a result, the steps were followed.
value by using the volumetric joint count (Jv) and rock quality
designation (RQD), average discontinuity spacing (S) and block (1) Structure rating (SR) between 0 and 100 proposed by Sonmez
volume (Vb). For this, the researchers took into account the study by and Ulusay (1999) is preserved as it is.
Palmstrom (2005): (2) A standardized range between zero and 100 is applied to the
“surface condition rating (SCR)” proposed by Sonmez and
SR ¼  17:5 ln Jv þ 79:8 (1)
Ulusay (1999) and named “standardized surface condition
After the development of a quantitative GSI chart by Sonmez rating (SCRs).
and Ulusay (1999, 2002), the quantification of the GSI chart con- (3) The digitized GSI values of GSIbasic are employed for the
tinues as an attractive research topic among the rock mechanics quantification of the basic GSI chart.
community (e.g. Cai et al., 2004; Russo, 2009; Hoek et al., 2013;
Duran, 2016; Morelli, 2017; Schlotfeldt and Carter, 2018). The recent For every 10 points incremental of SR from 0 to 100, the digitized
study by Xia et al. (2022), as a reflection of necessity of quantifying GSI values were used in the new quantification. The generalized
GSI, was published as the latest effort on this topic. This study formulations of the original GSI lines were derived using the
considered RQD and SCR for defining the degree of jointing of rock interpolation technique provided in Fig. 3. “Rating from SCALE A (in
mass and surface condition rating of discontinuities. On the other the horizontal axis)” and “Rating from SCALE B (in the vertical
hand, the researcher considers Hoek et al. (2013)'s slightly modified axis)” were denoted as RSCALE-A and RSCALE-B, respectively:

155
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 1. The comparisons between fully parallel GSI lines (GSI2013) and original GSI lines (GSIbasic) were investigated on a basic GSI chart by considering both (a) the difference in
percentage, and (b) the difference in quantity (adapted from Hoek and Marinos, 2000; Hoek et al., 2013).

Fig. 2. The comparisons between fully parallel GSI lines (GSI2013) and original GSI lines (GSIbasic) were graphically investigated by considering both (a) the difference in percentage,
and (b) the difference in quantity.

RSCALEB ¼ 0:0016ðSRÞ2 þ 0:1972ðSRÞ þ 6:4295 (2b)

RSCALEA ¼ ½0:0018ðSRÞ þ 0:40225SCRS (2a)
GSI ¼ RSCALEA þ RSCALEB (2c)

156
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 3. Derivation of the generalized quantified GSI formulations (the values in the GSI matrix belong to the basic GSI chart): (a) 201 digitized GSIbasic points on GSI chart, (b)
relations between SCRs and GSI0 for different SR values from 0 to 80, (c) relation between SR and GSISCR ¼ 0, and (d) relation between SR and A constant).

The sensitivity of the generalized equations of the original GSI percentage error graphs were used to evaluate the sensitivity of the
lines was analyzed using 32,160 GSIbasic values (¼160  201 based generalized equations for the original GSI lines (see Fig. 4). As
on SR and SCRs) and the correlation coefficients (r) have been shown in Fig. 4, Eq. (2) could effectively determine the GSI values
determined based on their relations. The cross-correlation and within a range of 5% when compared with the original GSI lines.

157
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 4. Sensitivity analysis of the generalized equations of original GSI lines by using (a) cross-correlation and (b) percentage error graphs (blue point represents data point).

However, it should be noted that plotting a GSI value on the basic (2018) employed a scale rating up to 55 as a VFC-rating. VFC is
GSI chart is also part of engineering judgment for better under- almost the half of SR value as a result of this scale difference;
standing the mechanical behavior of rock mass. therefore, the relation of VFC-rating (VFC-rating ¼ 8.457ln(VFC)þ
48.851) is almost equal to the half of Eq. (1) proposed by Sonmez
3.1. Determination of structure rating (SR) based on scale effect and Ulusay (2002). The third parameter used by Schlotfeldt and
Carter (2018) was the approximate block volume which was used
Block size distribution in a unit volume of rock mass is a function almost the same as that in Cai et al. (2004), except for the close part
of the jointing degree of rock mass. As is known, engineering ap- of the scale of the VFC-rating to 55. While P32 as the two-
plications are constructed in/on rock mass composed of large dimensional (2D) measurement was considered as the fourth
blocks which tend to be less deformable and well interlocked. In parameter, RQD/Jn (used Jn ¼ 9 for three joint sets) is considered as
general, a rock mass composed of small blocks is much deformable the fifth parameter used by Schlotfeldt and Carter (2018).
and the overall strength is expected to be weaker than those of In this study, the volumetric joint count (Jv z VFC), the average
intact rock masses. In addition, an increase in the number of joint S, the average Vb, and the RQD were used as input parameters for
sets causes an increase in the roundness of the rock blocks, and as determining the scale-based SR.
such the structure of the rock mass becomes poorly interlocked.
Therefore, the shape and dimension of rock blocks are important 3.1.1. Jv (zVFC)
properties of rock mass in terms of deformability and strength. Sonmez and Ulusay (2002) proposed a relationship by using a
Dimension and shape of rock blocks are closely related to the range of Jv (¼1 joint/m3, 3 joints/m3, 10 joints/m3 and 30 joints/m3)
number of joint sets and average spacing of joint sets. This indicates corresponding to the description of block rating intervals of SR
that the volumetric joint count (Jv) is a useful parameter that in- (¼80, 60, 40 and 20) from “intact or massive” to “disintegrated”
cludes both the number of joint sets and the average spacing of rock mass structures in the basic GSI chart. In this study, theoretical
joint sets. By considering this, Sonmez and Ulusay (1999, 2002) upper and lower boundary data pairs of Jv and SR as (0.3 joints/m3
defined the rock mass structure using Jv. On the other hand, Dinc and100) and (100 joints/m3,and 0) were also included to data
et al. (2011) determined the structure rating (SR) by considering points for the lowest and the heights values of SR as 0 and 100,
the volumetric joint count (Jv) and the relations among Jv, RQD, respectively. The regression analysis between SR and Jv was
average joint spacing (S) and block volume (Vb) based on the work repeated by using all data point including boundary data pairs
of Palmstrom (2005). The use of average joint spacing and the (Fig. 5a). The new relation of SR ¼ 17.3ln(Jv)þ17.422 has a negli-
average volume of rock blocks were also reported by Cai et al. gible difference from the relation of SR ¼ 17.5ln(Jv)þ79.8 which
(2004) to define the structure of rock mass in the basic GSI chart. was introduced to the literature by Sonmez and Ulusay (2002). On
However, Hoek et al. (2013) used only RQD to define the structure of the other hand, Schlotfeldt and Carter (2018) considered the in-
rock mass. Schlotfeldt and Carter (2018) rearranged the scale of tervals of VFC-rating (0.5 fractures/m3, 3 fractures/m3, 10 fractures/
RQD defined by Hoek et al. (2013), after a brief discussion on the m3, 30 fractures/m3, 100 fractures/m3 and 300 fractures/m3) cor-
limitations of the use of RQD alone to define the structure of rock responding to the description of block rating intervals of VFC (¼55,
mass. In addition, volumetric fracture count (VFC) which is similar 40, 30, 20, 10 and 0) for the same rock mass structures. For a
to Jv was selected as the second parameter to define the structure of possible comparison of the two studies on the same graph, the VFC-
rock mass by Schlotfeldt and Carter (2018). The boundary values of rating scale was multiplied by 2 to approximate the SR values. It was
VFC are almost the same as Jv used by Sonmez and Ulusay (2002). revealed that the rating difference between Sonmez and Ulusay's
However, Schlotfeldt and Carter (2018) moved one row up the (2002) and Schlotfeldt and Carter's (2018) assessments was about
boundary values of VFC in the structure of rock mass defined by 20 due to moving up one row (see Fig. 5a). However, when the
Hoek and Marinos (2000). While Sonmez and Ulusay (2002) values of VFC multiplied by 0.333, both approaches overlap (see
defined a scale rating up to 100 as a SR, Schlotfeldt and Carter Fig. 5b).

158
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 5. (a) Evolution of SR (z2VFC-rating) with Jv data pairs used by Sonmez and Ulusay (2002) together with the theoretical boundary data pairs and VFC data pairs by considering
Schlotfeldt and Carter (2018), and (b) evolution of SR (z2VFC-rating) based on Jv and 0.333VFC.

3.1.2. Average discontinuity spacing (S) block in a rock mass can practically be calculated using the relation
The average spacing of discontinuities has been widely used in Vb ¼ S3. This assumption was also employed by Cai et al. (2004) for
rock mechanics for defining the jointing degree of rock mass the use of Vb in their quantified GSI chart proposal. Therefore, the
pffiffiffiffiffiffi
particularly for almost isotropic joined rock masses. It is an easily relation of S ¼ 3 Vb can be used in determining SR as an isotropic
obtained and useful parameter for defining the jointing degree of rock mass.
isotropic rock masses composed of almost equidimensional rock
blocks. The use of average S in determining SR by using the rela- 3.1.4. Rock quality designation (RQD)
tionship between Jv and S defined by Palmstrom (2005) as “Jv ¼ 3.3/ RQD has been an important geotechnical parameter in rock
S” for common and equidimensional block shape was introduced by engineering because it is one of the standard measured parameters
Dinc et al. (2011). On the other hand, Cai et al. (2004) also paired the in geotechnical drilling studies as introduced to rock mechanics
boundary values of average spacing of discontinuities for the literature by Deere (1964). The use of RQD in determining SR should
blockiness intervals of 0.01 m, 0.03 m, 0.1 m, 0.3 m, 1 m, and 3 m to be performed carefully. Since RQD is a directional parameter, it is
the corresponding SR intervals of 0, 20, 40, 60, 80 and 100, important to ensure that the rock mass being considered has a
respectively. In addition, the boundary values of the average satisfactory isotropic structure. In addition, spacing of joints may be
spacing of discontinuities of 0.06 m, 0.2 m, 0.6 m, 2 m, and 6 m, as the other limitation for the use of RQD to describe the structure of a
recommended by International Society for Rock Mechanics and rock mass. Palmstrom (2005) mentioned the limitations of the RQD
Rock Engineering ISRM (1981) and used by Pettifer and Fookes method as “… …. The application of RQD in rock engineering calcu-
(1994) and in RMR89, were also considered in this study. lations may lead to inaccuracy or errors”. Consequently, while RQD is
A comparison of three approaches was given on the same graph a practical parameter for core logging, it is not sufficient on its own
of S versus SR (see Fig. 6a). The three recommendations overlap to provide an adequate description of a rock mass (Bieniawski,
when the interval of S proposed by ISRM (1981) and Cai et al. (2004) 1989; Milne et al., 1998).
were multiplied by 0.606 and 0.303, respectively (see Fig. 6b). Some different approaches are available in the literature for
determining SR from RQD. For example, in the work of Dinc et al.
3.1.3. Rock block volume (Vb) (2011), the proposed empirical relations by Palmstrom (2005) to
Block volume distribution of the jointed rock mass is controlled predict Jv from RQD can be used in determining SR based on the
by the number, orientation, and true spacing of joint sets. For value of Jv. In addition, when the borehole data are not available, the
almost isotropic jointed rock masses, the average volume of rock theoretical relation between discontinuity frequency (l) and RQD
159
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 6. (a) Evolution of SR based on intervals suggestions of spacing of discontinuities (S) by considering Palmstrom (2005), ISRM (1981) and Cai et al. (2004) and (b) the overlapped
relations based on the values of Jv by using multipliers.

derived by Priest and Hudson (1976) may also be used in deter- When a multiplier is applied to different input recommendations,
mining RQD. For the same purpose, Priest (1993) presented a linear the logarithmic relationships could overlap. Based on this infer-
relation between RQD and l by considering the linear part of the ence, the multiplier was named as a scale factor (sf) to generalize
exponential relation proposed by Priest and Hudson (1976). On the the logarithmic type equation for the determination of engineering
other hand, Hoek et al. (2013) preferred the use of RQD/2 in the scale-dependent SR (see Eq. (3)). Furthermore, K in Eq. (3) has the
quantification of rock mass structure. This approach can be rear- flexibility to use different input parameters such as Jv, l, Save and Vb
ranged to satisfy SR ¼ RQD by considering RQD ¼ 0% for SRmin ¼ 0 for defining the degree of jointing of a rock mass (see Eq. (4)).
and RQD ¼ 100% for SRmax ¼ 100. Although a well-known spacing of It may not be possible to highlight one of the input parameters
joints has generally negative exponential statistical distribution, for determining the jointing degree of a rock mass. However, when
Hoek et al. (2013) assigned the boundary values of RQD/2 by using a the empirical characteristic of GSI-based design tools is employed,
range of 10% for the rock mass classes defined in the basic GSI chart. the use of Jv for determination of the SR value may be recom-
Due to this simplification, the higher values of SR are obtained from mended as a volumetric quantity depending on its capability to
Hoek et al. (2013), compared to Dinc et al. (2011), when the values describe rock mass structure. On the other hand, RQD can also be
of RQD are higher than 17%. This output is consistent with the used carefully to determine the value of K. The use of more than one
caution underlined for a recommendation of the use of the quan- relation to define rock mass structure class may be helpful to
tified GSI chart up to a slope of height of 20 m and a tunnel of width improve the quantitative assessment of GSI.
of 10 m by Hoek et al. (2013). However, at this point, it is better to
 
remember that the measurement of RQD for the rating of dis-
SRsf ¼  17:5 ln sf K þ 80 (3)
integrated and widely jointed rock masses is almost impossible.
Therefore, when the scale of interest depending on the ED in-
creases, the rating method by Dinc et al. (2011) seems to produce
K ¼ Jv (4a)
more acceptable results than that proposed by Hoek et al. (2013).
or
3.2. Modified SR equation with scale effect of ED
3:3
As aforementioned, a characteristic logarithmic type of equation K¼ (4b)
Save
is valid in determining SR value using different input parameters
which are widely used to define the jointing degree of rock mass.
160
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

K ¼ 3:3l (4c)

3:3
K ¼p ffiffiffiffiffiffi (4d)
3
Vb

where SRsf is the scaled structure rating, K is the jointing input

parameter, sf is the scale factor in the range of 0e1, Jv is the volu-
metric joint count (joints/m3), Save is the average joint spacing (m),
l is the average joint frequency (joints/m), Vb is the average rock
block volume (m3).
The studies on the quantification of GSI have been conducted
since 1999. However, the second of two highlighted shortcomings
in the quantified GSI charts underlined by Hoek et al. (2013) is
“Secondly, the use of RQD by the authors or some variation of the
volumetric joint count Jv or the block volume Vb, by the other authors,
limits the definition of rock structure to the dimension of the blocks.
This takes no account of the ratio of block size to the size of the tunnel
or slope which has a significant influence on the application of the GSI
chart for characterizing the rock mass.“.
As emphasized by Hoek et al. (2013) for the shortcoming of GSI
quantifications, quantitatively defining rock mass structure still
needs improvement in terms of scale effect for engineering appli-
cations. The SR needs to be decreased from the smaller size to the
larger size even for the same rock mass. This is mainly due to the
increase of joint number in the considered size, which is consistent
with the limitations on the use of GSI depending on scale empha- Fig. 7. (a) Schematically presentation of rock mass volume and strength relationships
sized by Hoek et al. (2013). By considering the scale effect on GSI in depending on scale effect, and (b) of scale factor in accordance with Hoek and co-
terms of ED, Hoek et al. (2013) limited the use of their quantified authors’ recommendation.
GSI chart for tunnels of about 10 m span and slopes of height less
than 20 m. They suggested reducing the GSI value for larger caverns
or slopes. This caution was given in the left upper part of the intact rock pieces will translate, rotate or crush in response to stresses
imposed” by Hoek (1983). As a commonly accepted conclusion, the
quantified GSI chart proposed by Hoek et al. (2013). After all these
explanations, it can be concluded as a scale dependency of the GSI ratio of average block dimension to the size of engineering appli-
cation as at least 1/10 seems to be meaningful and acceptable for
chart based on EDs. The reduction of GSI by Hoek and co-authors’
recommendations can be satisfied by moving vertically downward the minimum required volume (representative elementary volume,
REV) for the rock mass behavior in rock engineering practice. When
in the quantified GSI chart proposed by Hoek et al. (2013). However,
the answer to the question of how much it should be moved the limitations on the use of GSI depending on the scale empha-
sized by Hoek and co-authors are taken into account, the minimum
downward is still not available now. In this study, in order to give a
quantitative answer to the question, SR was transformed into a number of rock blocks equal to 10 in REV may be evaluated as the
lowermost rock mass volume in which the rock mass failure be-
scale-dependent parameter to sufficiently define rock mass struc-
ture with EDs, such as slope height and tunnel width. comes dominant. REV is only the lowermost limit of rock mass
volume. However, reducing the overall strength of a rock mass is
Numerous studies on the scale effect on the mechanical
behavior of rock mass are available in the literature. Duran (2016) continued due to the scale effect based on EDs for larger rock mass
volumes over REV.
summarized available methods on the minimum volume of rock
mass required to the applicability of the Hoek-Brown failure cri- When the scale factor denoted by sf in Eq. (3) is considered as
the ratio 1 to 10, SR is determined almost equal to 100 by using
terion since it was introduced to the literature in 1980. In these
studies, a minimum limit ratio of an engineering scale to block generalized formulation of SR given in Eq. (3). Therefore, SR ¼ 100 is
the upper value for sparsely jointed (intact or massive) rock mass
dimension between 6 and 20 was provided by different researchers
for applicability of the criterion (Duran, 2016). Schlotfeldt and class, and the ratio of 1e10 between sfi and Save given by Eq. (5) was
used for the upper limit of the scale factor (sf).
Carter (2018) indicated that the ratio of engineering scale to block
dimension should be at least ten or more for applicability of their V-
GSI chart. Cunha (1993) and Farahmand et al. (2018) emphasized sfi 1
the importance of dependency of an engineering property to the ¼ (5)
S 10
volume of interest in a rock medium. A simple schematic presen-
tation of rock mass volume and strength relationships depending where S is the average joint spacing (m) and sfi is the value of sf
on scale effect is illustrated in Fig. 7a. When a very small size value to satisfy SR z 100, depending on the value of S. The sche-
relative to the average block dimension is considered, the strength matic illustration of the calculation of sfi and its physical meaning is
of the rock medium has a wide range for intact rock and joint explained in Fig. 7b.
(Fig. 7a). However, the rock mass including equidimensional rock The graphical relations between scaled structure rating (SRsf)
blocks surrounded by three or more discontinuity sets, when the and scale factor (sf) varying from sfi to 1 is presented in Fig. 8a, for
rock mass volume increases, the anisotropic strength behavior of rock mass of size varying from very small to very large block size
rock mass disappears (Fig. 7a). In this case, the mechanical behavior based on the average S. The SR axis including sf ¼ 0.303 and 0.606
of rock mass is expected in accordance with the definition “the based on Cai et al. (2004) and ISRM (1981) is also shown. In
161
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 8. (a) The graphical relationship between structure rating (SR) and scale factor (sf) varying from 1 to sfi together with plots based on the works of Sonmez and Ulusay (2002),
ISRM (1981), Cai et al. (2004), and Hoek et al. (2013), and (b) the relationship between sf and ED.

addition, the RQD values based on Hoek et al. (2013) are also plotted behave anisotropic for the studied ED. For this case, “anisotropic -
on the same graph (see Fig. 8a). As can be seen from Fig. 8a, SRsf do NOT use GSI” reported by Hoek et al. (2013) should be consid-
close to 100 is obtained for the ED which is 10 times the average ered. On the other hand, for a rock mass depending on the increase
block size (or average spacing of discontinuity). When the ED and of ED, sfK can be higher than 0.33 which will be satisfactory to be
average spacing of discontinuities are equal to 50 m and 5 m or used in GSI.
more, SRsf is obtained as almost equal to 100. On the other hand, for
the same ED of about 50 m, when the average spacing of discon- 3.3. Standardized surface condition rating (SCRs)
tinuities is equal to 0.75, SRsf ¼ 66 (see Fig. 8a). By considering the
combined effect of scale factor (sf) and ED in Fig. 8a, sf (in m) is As aforementioned, both SR and SCRs are independent from the
defined by diving ED with 100 m as illustrated (see Fig. 8b). ED can input parameters with the values varying from 0 to 100. As re-
be selected as either slope height (H) or two times the tunnel span ported by Hoek (1983), the intact rock pieces in the rock mass will
width (B) in accordance with the description used by Hoek et al. translate, rotate, or crush in response to the imposed stresses. The
(2013). Practically, SRsf determined for an engineering dimension shear strength of discontinuities is also an important component
faced in a project design stage can be used for relatively small cases used in determination of the GSI. Five classes of joint surface quality
in the same rock mass, because this value will be conservative for presented on the basic GSI chart by Hoek and Marinos (2000) were
smaller EDs in relevant design projects. In addition, the ED can be defined arithmetically as equal rates to 20 points in the SCRs. By
considered as 2e3 times the foundation width (Bf) depending on using ratings for joint conditions given in RMR89 and the ratio of
the shape of the foundation from square to strip. Although the joint roughness number (Jr) to joint alteration number (Ja) given in
continuous relationship was defined between sf and ED, the cate- 1974 version of Q value (Q74) proposed by Barton et al. (1974), SCRs
gorical values of sf can be used for practical purposes by considering can be determined. Although Hoek et al. (2013) considered all five
the five classes of engineering dimensions from very small (VL) to joint condition parameters including aperture and persistency of
extremely large (EL) as defined in Fig. 8b. RMR for quantification of GSI, only three of them (i.e. roughness,
At this stage, it is necessary to discuss an important aspect re- weathering, and infilling) are directly related to the shear strength
ported by Hoek and co-authors. When the value of sfK in Eq. (3) is of joint surfaces as preferred by Sonmez and Ulusay (1999, 2002). In
less than ~0.33, SRsf is higher than the upper limit of SRsf ¼ 100. This this study, three alternatives were presented for determination of
condition indicates that the ED is not sufficiently large for the rock the SCRs based on RMR89 and Q74. In the first approach to determine
medium to behave as rock mass, while the rock mass tends to SCRs, the total maximum rating of 18 proposed by Bieniawski
162
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

(1989) was kept the same as those used by Sonmez and Ulusay (see Fig. 11).
(1999), but it was proportionally standardized by using The properties of the geomaterials observed in this site were
SCRs_max ¼ 100. In the second approach, a total maximum rating of investigated by Ozcelik and Sonmez (2020). P2 stratified tuff rock
30 given in RMR89 was considered and similarly, it was propor- mass unit includes two joint sets in addition to the stratification
tionally standardized by using SCRs_max ¼ 100. The properties of surface. The average spacing of discontinuities varies from ~20 cm
persistency and aperture may be related to the rock mass structure to ~1 m with an average value of about 30 cm. The average uniaxial
rather than surface conditions. Therefore, the use of SCRs by compressive strength and unit weight of the P2 rock mass unit
considering ratings of three parameters such as roughness, infilling, were 3.1 MPa and 13.2 kN/m3, respectively. The mi parameter used
and weathering is highly recommended in this study. In addition, in the Hoek-Brown failure criterion was selected as 8 for fine-
the ratio of joint roughness number (Jr) to joint alteration number grained tuff rock material. By considering the effects of both me-
(Ja) by Barton et al. (1974), was also introduced as the third chanical excavation and blasting production methods used in
approach to determine SCRs. The alternatives for determination of operation of the open pit, the disturbance factor (Df) of an average
the SCRs according to the parameters of RMR89 and the parameters of 0.85 was introduced following the guideline proposed by Hoek
of the Q74 system are given in Fig. 9. On the other hand, SCRs are also et al. (2002). However, Hoek and Brown (2019) recommended a
open to introducing ratings of new methods for a value between transitional disturbance relationship from the excavation surface to
0 and 100. the depth proposed by Rose et al. (2018). In fact, it is obvious that
In this study, GSI is transformed into a scale-dependent the importance of the transitional relationship for the disturbance
parameter based on the scale-dependent SR considering EDs, effect increases with the depth of slopes. Rose et al. (2018) assumed
which is highly needed in engineering applications. To avoid that D ¼ 1 from the surface of excavation to the depth of the fully
possible confusion, it has been suggested to use GSIED such as disturbed limit (FDL). However, the use of D ¼ 1 without consid-
GSIED¼60 m (means engineering dimension is about 60 m slope ering the method of excavation seems to be conservative. In this
height or 30-m tunnel span width) or GSIED¼20 m (means ED is study, the transitional disturbance relationship was rearranged by
about 20-m slope height or 10-m tunnel span width) to indicate considering the disturbance factor (Df) selected for the excavation
this distinction in the notation. On the other hand, the GSI deter- surface (see Table 1). On the other hand, when Df ¼ 1, the transition
mined for ED equal to 100 m or higher becomes independent from relationship proposed by Rose et al. (2018) is satisfied in the pro-
the ED. Hence, when the ED is sufficiently large, the use of GSI∞ is posal given in Table 1. Therefore, the use of Df which can be selected
recommended to define the scale-independent lowest value of in accordance with the recommendations based on the excavation
GSIED. The generalized form of the scaled GSIED chart is given in method by Hoek et al. (2002) was preferred for the same purpose.
Fig. 10. As can be calculated by the recent equation of Hoek-Brown Therefore, the FDL terminology preferred by Rose et al. (2018) was
failure criterion, the strength equation will be equal to the strength also renamed as disturbed limit (DfL).
of intact rock when GSI ¼ 100. Because the values of Hoek-Brown A simple judgment criterion suggested by Hoek and Karzulovic
constants’ “mb” and “s” are calculated as “mi” and 1, respectively (2000) for the possible thickness of the disturbed zone may change
when GSI ¼ 100. Therefore, GSIED ¼ 100 is represented by a point on from 0.3 to 2.5 times the height of the bench. The real slope in-
the GSI chart given in Fig. 10, and the rock mass class is re-defined cludes benches having a height of 10 m. Hence, the depth of DfL may
as “massive or sparsely jointed: in situ rock mass with few widely change from 3 m to 25 m based on the recommendation by Hoek
spaced discontinuities”. and Karzulovic (2000).
However, Rose et al. (2018) proposed a disturbance rating (DR)
for the determination of the depth of FDL. According to the rock
4. Engineering application of GSIED mass cases evaluated using DR, while DRs of strong rock mass are
about 1e3, DRs may increase up to 24 to 45 for weak rock masses
The novel complementary approach was applied to a real rock (Rose et al., 2018). The fully disturbed limit can be defined as a DR
slope case for an open pit mine in Türkiye studied by Ozcelik and percentage of overall slope height by Rose et al. (2018).
Sonmez (2020). The overall stability of the slope was close to fail- After explanations, DR of the tuff rock mass can be evaluated as
ure even if the failure did not occur completely. The depth of the in a range of 25e50 due to the weak rock mass properties. In this
Kisrakdere open pit base gradually has been increased by the ex- study, it assumes DR ¼ 37.5, which means 37.5% of UDL is equal to
cavations for production of the lignite coal. The cross-section DfL, for weak tuff rock mass. UDL was also considered as equal to the
through the general slope direction of the open pit mine benches ED of slope profiles at the initial stage of excavations, on 7 June
was drawn by using a digital elevation model (DEM) of excavation 2020 and September 1, 2020. The mobilized shear strength of the
stages at the initial and dated Jun 7, 2020 and September 1, 2020

Fig. 9. The rating approaches for determining SCRs based on surface condition ratings of RMR89, RMR76, and Q74 (after Bieniawski (1976, 1989) and Barton et al. (1974)).

163
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 10. Generalized form of scaled GSIED chart, and the plots for the critical points of the application case analyzed in this study.

fault surface was determined as c ¼ 2 kPa and f ¼ 10.2 from the The calculations of SRsf using Eq. (4b) for both cases are summa-
back analysis of bench stability by Ozcelik and Sonmez (2020). rized in Table 2. While the roughness of discontinuity surfaces was
Unfortunately, there was no data about piezometric surfaces along defined as smooth and slightly rough (rating of Rr ¼ 3), the
the slope profile. It was observed that an audit pit, excavated in weathering of discontinuity surfaces was defined as moderately to
front of the toe, was filled by groundwater within a short time (see highly weathered (rating of Rw ¼ 2). The thickness of the infilling
Fig. 11). Therefore, instead of considering the drained condition, the was less than 5 mm and filled with soft infilling material (rating of
pore water pressure ratio was assumed to have an average value of infilling Rf ¼ 2). By summing of ratings of roughness, weathering,
ru ¼ 0.1 by using possible pore water pressure acting along the and infilling, R ¼ 7 was determined according to the proposal given
failure surface. in Fig. 9. Therefore, SCRs ¼ 38.9 when it is standardized by
As soon as the deformations and cracks were noticed by engi- considering the maximum rating of SCRs as 100 (see Table 2). For
neers, the excavation operations along these benches were imme- the defined critical points on the slope profile, the calculation of
diately stopped as the first reaction. The monitoring station GSIED, including the parameters RSCALE-A and RSCALE-B, is listed in
network was established to investigate the slope displacements. Table 2. As seen in Table 2, the GSIED¼30 m and GSIED¼60 m are ob-
The vectors of monitoring stations are shown on the cross-section tained as 43.5 and 38.2 for the critical point of B1 (~30 m slope
given in Fig. 11. The displacements including upward components height above the point of B1) and B2 (~30 m slope height above the
were observed at the K5 station placed at the toe of the slope. On point of B2), respectively. The GSIED¼100 m was also determined as
the other hand, the inclination of displacements having downward 34.5 for the critical points of A1, A2, and A3 which have about 100 m
components increases towards the top of the slope at the stations slope height between the toe and top of the general slope profile.
K21, K44, S01, K28, and K27 (see Fig. 11). Therefore, these The values of GSIED decrease from 43.5 to 34.5 depending on the
displacement vectors indicate a possible circular failure surface change in the ED from ~30 m to ~100 m.
within the rock mass. As mentioned above, the tension cracks While stability analyses of circular surfaces were performed by
behind the N1 fault and the top of the slope were also mapped. using Bishop's simplified method (Bishop, 1955), the non-circular
Therefore, in addition to circular failure, the combined type of combined type of failure surface was analyzed by Janbu's simpli-
failure is also possible starting from the N1 fault and then passing fied method (Janbu, 1954, 1973). The factor of safety (FOS) with the
through the rock mass by a circular surface towards the toe of the potential failure surfaces for the analyzed critical points from the
slope. The overall stability condition of the slope is still far from initial stage of excavations to September 1, 2020 was given in
failure at the initial stage of excavations and on 7 June 2020 tension Fig. 12. The overall FOS for the slope profile decreases from 2.14 to
cracks were reported upon evaluation by using critical points B1 1.059 depending on the increase in open pit mine depth. The FOS of
and B2. Then, the slope became close to failure on September 1, slope on September 1, 2020 was determined by the circular failure
2020 after occurrence of tension cracks, which was also analyzed by surfaces and the combined failure surfaces as 1.141 and 1.059,
considering three critical points A1, A2, and A3. While point A1 is respectively.
the toe of the general slope profile on September 1, 2020, points A2 In addition to the observed displacements along the N1 fault in
and A3 are the deepest points of the tension cracks behind N1 Fig. 11, the inclinations of the displacement vectors of the moni-
faults. toring stations are more coherent with the combined failure surface
While the ED as slope height above critical points B1 and B2 was which starts along the N1 fault, then continues as circular towards
about ED ¼ 30 m and ED ¼ 60 m, respectively, ED was about 100 m critical point A1 as compared with a completely circular failure
for critical points A1, A2, and A3 (see Fig. 11). Jointing input surface. Therefore, it can be considered that the FOSA1 seems more
parameter (K) of rock mass was determined as 11 by using K ¼ 3.3/S. acceptable with respect to the factor of safety equal to 1.141, which

164
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 11. (a) The cross-section of a real slope exaction project including vectors of monitoring stations, and (b) views from the real slope case study.

Table 1
Transitional disturbance factor proposed by Rose et al. (2018) and the preferred form of the relationship in this study.

Transitional disturbance factor relationship proposed by Rose et al. (2018) The slightly improved transitional disturbance factor relationship in this study
9 9
D ¼ 1 ðfor depth of FDLÞ >
= D ¼ Df ðfor depth of DLÞ >
=
 
ln H  ln Hi > ln H  ln Hi
D¼ ðfor depth from FDL to UDLÞ ; D¼ >
Df ðfor depth from DL to UDLÞ ;
decay G
decay: ln(FDL/H) G: ln(DfL/H)
* UDL is a depth of approximately equal to slope height * UDL is a depth of approximately equal to slope height

Table 2
Determination of GSIED based on the EDs for the analyzed critical points.

Critical point ED (m) sf (¼ED/100) K SRsf¼0.6 SCRs RSCALE-A RSCALE-B GSIED¼30 m

B1 ~30 0.3 11 59.1 38.9 19.8 23.7 3.5

B2 ~60 0.6 11 47 38.9 18.9 19.2 38.2
A1, A2 and A3 ~100 1 11 38 38.9 18.3 16.2 34.5

is above 1, but close to the limit equilibrium condition. On the other mass movement mode. As mentioned before, the pore water
hand, tension cracks and deformations behind the N1 fault can be pressure ratio includes uncertainties due to the lack of
explained by the retrogressive failures following the combined groundwater-surface measurements in drill holes along the slope

165
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

Fig. 12. Results of slope stability analyses of the defined critical points on slope profiles (a) at the initial stage of excavations, (b) on 7 June 2020, and (c) on September 1, 2020.

profile. Therefore, the calculated FOS for the critical points labeled literature to overcome the limitations of RMR on identification of
as B1 and B2 may be evaluated as the conservative assumption in the weak to very weak rock masses. Therefore, the original GSI lines
terms of groundwater condition. In other words, the FOS may should be satisfied as much as possible when performing a quan-
slightly increase for low pore water pressure because of the un- titative approach. A series of equations derived by interpolation in
certainties related to the groundwater surface. Since providing the this study yield values for GSI, which was identical to those
limit equilibrium condition of the slope profile is not the aim of this observed in the original GSI lines. The values of GSI can be deter-
study, a value assumption is made for ru only for the evaluation of mined with ±5% error by using the developed equations. To over-
the slope profile that is prone to instability. come the limitation on the use of scale effect in terms of ED, a new
scaled GSI concept based on the scaled SR value was introduced to
the quantification of original GSI lines on the basic GSI chart. For
5. Conclusions
this, the scale factor (sf) ranging from 0.2 to 1 depends on EDs such
as slope height, the width of the tunnel, and the width foundation,
Quantification studies on the expert-based visual basic GSI chart
which is defined by using the representative elementary volume of
have been an attractive research topic in rock engineering. Due to
rock mass including joint pattern. Then, the scale factor (sf) was
the assumptions of every quantification effort on the basic GSI
adapted to the general formulation of structure rating which is
chart, different GSI values are obtained even if GSI evaluation is
denoted by SRsf.
performed on the same rock mass. However, it should be empha-
Before these improvements, the quantified-based GSI obtained
sized that every study is based on the assumptions underlying its
from any method in literature has been used as a unique parameter
results.
independent of the ED. However, in this study, GSI was transformed
As stated by Hoek et al. (2013), redrawing the GSI lines as fully
into a scale-dependent parameter based on the ED, which is of
parallel to each other was a minor modification to the basic GSI
great need in engineering applications. The improvements pro-
chart published by Hoek and Marinos (2000). Transforming original
posed in this study were applied to a real slope case study which
GSI lines into fully parallel GSI lines reveals slightly different GSI
was close to failure. Following the assessments of stability results, it
values, especially for those lower than 50 when compared to the
can be concluded that the new proposals have the capability to
original GSI values. It is noted that GSI was introduced to the
166
H. Sonmez, G. Dagdelenler, Y. Ozcelik et al. Journal of Rock Mechanics and Geotechnical Engineering 17 (2025) 154e167

unify almost all the quantification efforts in the existing literature. Geomechanics Symposium, ARMA 13-672, San Francisco, USA.
Hoek, E., Brown, E.T., 2019. The HoekeBrown failure criterion and GSIe2018 edition.
J. Rock Mech. Geotech. Eng. 11, 445e463.
Declaration of competing interest ISRM, 1981. In: ISRM Suggested Methods: Rock Characterization, Testing and
Monitoring. Pergamon Press, London, p. 211.
The authors declare that they have no known competing Janbu, N., 1954. Application of composite slip circles for stability analysis. In: Pro-
ceedings of the Fourth European Conference on Stability of Earth Slopes,
financial interests or personal relationships that could have pp. 43e49.
appeared to influence the work reported in this paper. Janbu, N., 1973. Slope stability computations. Embankment Dam Engineering -
Casagrande Volume. John Wiley and Sons, New York, pp. 47e86.
Marinos, P., Hoek, E., 2000. GSI: a geologically friendly tool for rock mass strength
Acknowledgments estimation. In: Proceedings of the GeoEng2000: an International Conference on
Geotechnical & Geological Engineering. International Society for Rock Me-
The authors would like to thank the responsible companies KIAS chanics, Melbourne, pp. 1422e1442.
Marinos, P., Hoek, E., 2001. Estimating the geotechnical properties of heterogeneous
and ODAS for coal production in the Kisrakdere open pit mine for rock masses such as flysch. Bull. Eng. Geol. Environ. 60, 85e92.
their permission to use the slope case study. Thoughts about the Marinos, V., 2017. A revised geotechnical classification GSI system for tectonically
use of scale effect on structure rating were presented at the “Rock disturbed rock masses, such as flysch. Bull. Eng. Geol. Environ. 19, 1e14.
Marinos, V., Carter, T.G., 2018. Maintaining geological reality in application of GSI for
Mass Characterization with Emphasis in Rock Slope Hazards" design of engineering structures in rock. Eng. Geol. 239, 282e297.
workshop organized by commission C38 at the 3rd European Milne, D., Hadjigeorgiou, J., Pakalnis, R., 1998. Rock mass characterization for un-
Regional Conference of IAEG (3rd EUROENGEO2020) held in Ath- derground hard rock mines. Tunn. Undergr. Space Technol. 13 (4), 383e391.
Morelli, G.L., 2017. Alternative qualification of the geological strength index chart
ens. We would like to thank both revisers for their valuable com-
for jointed rocks. Geotech. Jeol. Eng. 35, 2803e2816.
ments to improve the quality of the study. Ozcelik, Y., Sonmez, H., 2020. Slope Stability Interim Assessment Report of the R&D
Project for Developing a Safe Sustainable Mining System in the North Kisrak-
References dere Lignite Open Pit Mine Having License Number S:549. HT-TTM (in Turkish).
Palmstrom, A., 2005. Measurements of and correlations between block size and
rock quality designation (RQD). Tunn. Undergr. Space Technol. 20, 326e377.
Barton, N.R., Lien, R., Lunde, I., 1974. Engineering classification of rock masses for the Pettifer, G.S., Fookes, P.G., 1994. A revision of the graphical method for assessing the
design of tunnel supports. Rock Mech. 6 (4), 189e239. excavability of rock. Q. J. Eng. Geol. 27, 145e164.
Bieniawski, Z.T., 1976. Rock mass classification in rock engineering. In: Proceedings Priest, S.D., Hudson, J., 1976. Discontinuity spacing in rock. Int. J. Rock Mech. Min.
of the Symposium on Exploration for Rock Engineering. Balkema, Cape Town, Sci. Geomech. Abstr. 13 (5), 135e148.
pp. 97e106. Priest, S.D., 1993. Discontinuity Analysis for Rock Engineering. Chapman & Hall,
Bieniawski, Z.T., 1989. Engineering Rock Mass Classifications. John Wiley and Sons, p. 473.
p. 237. Rose, N.D., Scholz, M., Burden, J., King, M., Maggs, C., Havaej, M., 2018. Quantifying
Bishop, A.W., 1955. The use of the slip circle in the stability analysis of slopes. transitional rock mass disturbance in open pit slopes related to mining exca-
Geotechnique 5, 7e17. vation. In: Proceedings of the Slope Stability 2018eXIV International Congress
Cai, M., Kaiser, P.K., Uno, H., Tasaka, Y., Minami, M., 2004. Estimation of rock mass on Energy and Mineral Resources, pp. 1273e1288.
deformation modulus and strength of jointed hard rock masses using the GSI Russo, G., 2009. A new rational method for calculating the GSI. Tunn. Undergr.
system. Int. J. Rock Mech. Min. Sci. 41, 3e19. Space Technol. 24, 103e111.
Cunha, A.P., 1993. Scale Effects in Rock Masses 93. Balkema, Rotterdam. Schlotfeldt, P., Carter, T.G., 2018. A new and unified approach to improved scalability
Deere, D.U., 1964. Technical description of rock cores for engineering purposes. Rock and volumetric fracture intensity quantification for GSI and rock mass strength
Mech. Rock Eng. 1, 17e22. and deformability estimation. Int. J. Rock Mech. Min. Sci. 110, 48e67.
Dinc, O.S., Sonmez, H., Tunusluoglu, C., Kasapoglu, K.E., 2011. A new general Sonmez, H., Ulusay, R., 1999. Modifications to the geological strength index (GSI)
empirical approach for the prediction of rock mass strengths of soft to hard rock and their applicability to stability of slopes. Int. J. Rock Mech. Min. Sci. 36 (6),
masses. Int. J. Rock Mech. Min. Sci. 48 (4), 650e665. 743e760.
Duran, A., 2016. Rock mass assessmentdwhat goes wrong? APSSIM 2016ePM Sonmez, H., Ulusay, R., 2002. A discussion on the Hoek-Brown failure criterion and
dight. Australian Centre for Geomechanics, Perth. suggested modification to the criterion verified by slope stability case studies.
Farahmand, K., Vazaios, I., Diederichs, M.S., Vlachopoulos, N., 2018. Investigating the Yerbilimleri 26, 77e99.
scale-dependency of the geometrical and mechanical properties of a moder- Xia, K., Chen, C., Wang, T., Pang, H., Liu, H., 2022. Quantification of the GSI and D
ately jointed rock using a synthetic rock mass (SRM) approach. Comput. Geo- values in the HoekeBrown criterion using the rock quality designation (RQD)
tech. 95, 162e179. and discontinuity surface condition rating (SCR). Bull. Eng. Geol. Environ. 81 (1),
Hoek, E., Brown, E.T., 1980. Empirical strength criterion for rock masses. J. Geotech. 21.
Engng. Div., ASCE 106 (GT9), 1013e1035.
Hoek, E., 1983. Strength of jointed rock masses: 1983 Rankine Lecture. Geo-
technique 33 (3), 187e223.
Hoek, E., 1994. Strength of rock and rock masses. ISRM News J 2 (2), 4e16.
Hoek, E., Brown, E.T., 1997. Practical estimates of rock mass strength. Int. J. Rock Dr. Harun So € nmez is a professor in the Department of
Mech. Min. Sci. 34 (8), 1165e1186. Geological Engineering at Hacettepe University. He spe-
Hoek, E., Karzulovic, A., 2000. Rock mass properties for surface mines. Slope Sta- cializes in rock mechanics, the characterization of jointed
bility in Surface Mining. In: Hustralid, W.A., McCarter, M.K., van Zyl, D.J.A. (Eds.), rock masses, and the stability and design of deep slopes.
Littleton, Colorado: Society for Mining, Metallurgical and Exploration. SME), His research also covers soil mechanics, earthquake-
pp. 59e70. triggered soil liquefaction, and landslide mapping for sus-
Hoek, E., Marinos, P., 2000. Predicting tunnel squeezing problems in weak het- ceptibility, hazard, and risk assessment. Dr. So€nmez has
erogeneous rock masses. Tunnels Tunn. Int. 32 (11), 45e51. extensive consulting experience in dam, tunnel, and slope
Hoek, E., Carranza-Torres, C.T., Corkum, B., 2002. Hoek-Brown failure criterion-2002 design. Additionally, he has contributed to various engi-
edition. Proceedings of the 5th North American Rock Mechanics Symposium 1, neering geological projects with national organizations
267e273. Toronto, Canada. like Türkiye Coal Enterprises Authority. His research has
Hoek, E., Carter, T.G., Diederichs, M.S., 2013. Quantification of the geological led to the development of numerous geotechnical soft-
strength index chart. In: Proceedings of the 47th US Rock Mechanics/ ware tools for slope stability and other analyses.

167