- Email: [email protected]
Engineering geological properties and durability assessment of the Cappadocian tuff
ENGHNEEREN6 GEOL©GY ELSEVIER
Engineering Geology47 (1997) 175-187
Engineering geological properties and durability assessment of the Cappadocian tuff T. Topal *, V. Doyuran Department of Geological Engineering, Middle East Technical University, 06531, Ankara, Turkey Received 25 June 1996; accepted 1 February 1997
Abstract
Some of the unique landforms, the so-called "fairy chimneys", which were formed within the Cappadocian tuff of central Turkey, were used as dwellings in the past and contain culturally valuable wall paintings. However, these structures are undergoing chemical and physical deterioration due to atmospheric effects. For conservation studies, an understanding of the engineering geological properties of the tuff is essential. In this study, emphasis is placed on both material and mass properties of the tuff. These properties are evaluated for the assessment of rock durability. This study shows that the Cappadocian tuff is almost fresh, with local discoloration, is moderately weak to very weak, and has low unit weight, very high porosity, and high deformability. Discontinuity surveys revealed two dominant joint sets, which not only controlled the formation but also control the structural stability of the fairy chimneys. Various methods used for the durability assessment of the Cappadocian tuff indicate poor to very poor durability. Due consideration must be given to strength reduction due to moisture, poor to very poor durability, and the adverse effects of joints on the structural stability of the fairy chimneys. © 1997 Elsevier Science B.V.
Keywords: Cappadocia; Durability; Engineering geology; Tuff
1. Introduction
The Cappadocia region comprising Nev~ehir, Kayseri, and Ni~de provinces of Central Anatolia (Fig. 1), is one of the attractive tourist sites of Turkey due to its spectacular and unique landforms, fairy chimneys, and historical heritages. The fairy chimneys developed within the tufts are the results of weathering and differential erosion. Some of the fairy chimneys were dwellings and were inhabited during the Byzantine times. Some
* Correspondingauthor. Fax: +90 312 2101263; e-mail: [email protected] 0013-7952/97/$17.00© 1997ElsevierScienceB.V. All fights reserved. PII S0013-7952 (97) 00017-3
such dwellings contain invaluable wall paintings, which also provide historical attraction. Weathering and erosion, however, are still active and these natural processes, along with human activities, threaten the stability of the chimneys. Studies towards conservation of these valuable historical and touristic phenomena necessitate the determination of the engineering geological (material and mass) properties and the durability of the tufts. In this study, material and mass properties of the tuff are determined. The average pore diameter, saturation coefficient, wet-to-dry strength ratio, static rock durability index, index of rock durability, and slake-durability index of the tuff are used
T. Topal. ['~ Doyuran
176
"~/Io-
/
Engineering Geology 47 (1997) 175 187
N
to KJr$¢hir I!
KAYSERi ,AREA NEVSEHiR 3917
Mustafapasa
&
Erciyes
da~,
AKSARAY 0 .-~
3268 A
Hasan da(]l
I0~
3O
I
i
Km
2963 &
i
MeLcndiz d(:l~u
,~
N i GDE
h I
~.:.
... '-.-' ..... <.-~ ,.
Fig. I. Location map of the study area.
for the durability assessment of the rock, and of the fairy chimneys. Engineering geological properties and the assessed durability of the tuff are re-evaluated for the purpose of conservation of the fairy chimneys. The study is carried out within the Urgtip-G6reme region (Fig. 1) due to the occurrences of typical fairy chimneys in this area (Fig. 2).
2. Site geology The Cappadocia region is generally underlain by thick and extensive deposits of the volcanosedimentary sequence (Miocene-Pliocene) of the Orgtip formation (Pasquare, 1968). Although the formation comprises a number of well-distinguished members, in the study area only the Kavak
and Tahar members contain the fairy chimneys. Both members are characterized by non-welded tufts. Since most of the dwellings and wall paintings are found within the fairy chimneys of the Kayak member, our study is focused on this member. The Kavak member represents the product of the first intermittent volcanic activity which produced the LlrgOp formation. Thus, this member constitutes the basal portion of the volcanosedimentary sequence. The chaotic arrangement of pumice fragments within the unit suggests that it is an ash flow tuff deposited in a lacustrine environment. It is dirty white to pink, and contains phenocrysts of plagioclase, quartz, biotite and opaque minerals. Various rock fragments and pumice are also commonly observed. In the matrix, volcanic glass shards are rather common (Topal
T. Topal, F. Doyuran / Engineering Geology 47 (1997) 175-187
177
to assess anisotropy. These directions are expected to represent the material characteristics of the tuff perpendicular and parallel to layering. The laboratory tests, performed according to ISRM (1981), RILEM (1980) and TSE (1978), included determination of the index and some mechanical properties of the intact rock specimens. The field observations involved assessment of the rock mass properties, such as discontinuity characteristics and weathering grade. The description of rock material and mass characteristics is based on Anon (1977), BSI (1981), and ISRM (1981).
3.1. Material properties of the Cappadocian tuff The Cappadocian tuff is dirty white to pink, consisting mostly of lapilli size (2-64 mm) phenocrysts and rock fragments embedded within tuffaceous matrix. The tuff is almost fresh with very local discoloration around biotites and some rock fragments. The results of index and mechanical tests conducted by the authors and those of other authors are given in Table 1.
3.2. Mass properties of the Cappadocian tuff Fig. 2. Typical view of fairy chimneys with cap rocks.
and Doyuran, 1994; METU, 1987). The volcanic glass shards are partly altered to smectite. In this paper, the term "Cappadocian tuff" is used to denote the Kavak member.
3. Engineering geological properties of the Cappadocian tuff Evaluation of the engineering geological characteristics of the Cappadocian tuff is based on field observations and laboratory studies. Since the fairy chimneys are essentially formed within ignimbritic levels of the Cappadocian tuff, the material and mass properties of these levels will be considered in the following sections. For the laboratory studies, core samples extruded from 22 oriented block samples are used. Although the Cappadocian tuff does not show distinct layering within ignimbritic levels, cores from the oriented block samples are taken in the vertical and horizontal directions
Within the thickly bedded and almost horizontally layered Cappadocian tuff, the joints constitute the major discontinuities. In the study area, the joints are locally well-developed and can easily be traced for a considerable distance. Some, however, can only be traced due to the alignment of vegetation supported by moisture along the persistent joints. A number of scan line surveys, spot measurements, and field observations were carried out in the field to determine the orientation, spacing, aperture, persistence, infilling, surface roughness, water seepage and weathering state of the joints. The orientation of the joints (i.e., strike and dip) were determined using geological compass. In the field, 1905 joint orientation measurements were taken at 150 locations (Fig. 3). In order to determine the dominant sets, the joint orientation measurements were evaluated using a computer program called "Rockware" (Rockware, 1993). For the pole plots, a Schmidt diagram with lower hemisphere projection was
178
T. Topal, IL Doyuran / Engineering Geology 47 (1997) 175 187
Table 1 Correlation table for the material properties of the Cappadocian tuff Material properties
This study No. of samples
Unit weight (kN m 3) dry saturated Effective porosity (%) Water absorption by weight (%) under atmospheric pressure by weight (%) under pressure by weight (%) Saturation coefficient Uniaxial compressive strength (MPa) dry-vertical (MPa) dry-horizontal (MPa) saturated-vertical (MPa) saturated-horizontal (MPa) Point load strength, I,~5o (MPa) dry saturated Schmidt rebound hardnessa: dry Indirect (Brazilian) tensile strength: dry (MPa) Slake Durability (%): 2nd cycle P-wave velocityc (m s 1) dry-vertical dry-horizontal saturated-vertical saturated-horizontal S-wave velocityd (m s ~) dry-vertical dry-horizontal saturated-vertical saturated-horizontal ) Modulus of elasticity, E (GPa) dry-vertical dry-horizontal Poisson's ratio, v dry-vertical dry-horizontal Dynamic modulus of elasticity (GPa) dry-vertical dry-horizontal saturated-vertical saturated -horizontal Dynamic Poisson's ratio dry-vertical dry-horizontal saturated-vertical saturated-horizontal Internal friction angle% ~ (°)
Erguvanll and Y~zer (1977)
Erdogan (1986)
METU (1987)
De Witte et al. (1988)
Erguvanh et al. (1989)
Mean_+ SD
10 10 10
13.60+0.11 17.75_+0.08 38.29_+0.38
10
21.60_+0.27
10 10
28.08 _+0.46 0.78_+0.01
10 10 5 5
6.53_+0.67 4.87 _+0.43 2.16+0.34 0.93 -+0.29
20 20 20 10
0.48_+0.10 0.13_+0.04 27_+3.3 0.69 -+0.08
2
84_+0.17
I0 5 10 5
2082.46 _+84.10 1973.86_+41.60 1388.87_+85.00 1264.20 _+38.90
10 5 10 5
1124.67 _+77.30 1046.06_+47.60 759.43 _+41.20 724.18_+68.70
10 10
3.08 _+0.49 3.26 _+0.45
10 10
0.20+0.02 0.19_+0.01
10 5 10 5
4.47+0.53 3.90+0.29 2.64+0.27 2.33+0.34
10 5 10 5 10
0.29 _+0.03 0.30+0.02 0.28 + 0.02 0.25 _+0.06 28 + 2.95
11.50 28
15.10 19.20 28.76
13.90 -32.72
25
27.50
--
14.02
15.90
42
19.20 28.76 27.50
0.64
5.5
0.60
6.5
6.07
3.0
--
0.40
--
40-48 0.7
6.5 3.0
20 b 0.70
-0.83
2058
1.40
2.4
4.2
3.9
40
49
2.40
13
19
T. Topal, V. Doyuran / Engineering Geology47 (1997) 175-187
179
Table 1 (Continued)
Material properties
Cohesion e, C ( M P a ) Linear strain ( m m m - 1) vertical horizontal Average pore diameter (inn) Permeability (m s -1) T h e r m a l conductivity coefficient
This study
Erguvanh and Erdo~an METU De Witteet al. Yiizer (1977) (1986) (1987) (1988)
No. of samples
M e a n 4- SD
10
2.28 4-0.15
1.20
0.9
7 7 4 ---
0.55 4-0.03 0.51 +0.07 0.11 4-0.01 ---
-0.07 --0.32
---0.00 0.43
--
--
--
1.8
E r g u v a n h et al. (1989)
1.9
m m
0.00 0.43
(kcal/hm°C) T h e r m a l expansion coefficient (×
10 -6
6.3
6.3
I/°C)
Capillary water a b s o r p t i o n coefficient ( g / m 2. min °'s) Drying rate at 20°C and 40% R H ( g / m 2. h °'5) Water retention at 20°C (%)
.
.
.
.
187
.
.
.
.
444
.
.
.
.
4.85
"L-type Schmidt h a m m m e r is used. bField measurement. °For P-wave measurements, 63-kHz transducers are used ( N e w Sonicviwer, 5217A-OYO). dFor S-wave measurements, 100-kHz transducers and high voltage generators are used. ~ and C values are obtained f r o m triaxial compression tests, using strength envelope.
adopted. The contour diagram was then obtained from the pole plots (Fig. 4). The diagrams yield two dominant joint sets: N78°E/85-90°NW and N-S/90 °. The spacing of the joints generally ranges between 0.1 and 20m (moderately wide to extremely wide) (Anon, 1977) and their persistence exceeds 10m (high to very high persistence) (ISRM, 1981 ). Studies performed on 1:35 000 scale aerial photographs have shown that some of the joints may be traced for a distance of 800 m. The apertures of most joints fall within narrow to tight (0-20mm) categories (Anon, 1977). The joint surfaces are generally rough planar. Some narrow joints contain clayey infilling, but most are free of infilling material. During field inspections no groundwater seepage was noted through the joints. The tufts are faintly to slightly weathered. Due to the steep topography, slope wash, and lack of vegetation the weathering products are soon removed. Along the persistent joints, discoloration and/or surface staining is commonly observed. Spalling of exposed surfaces due to alternate
freezing-thawing, and wetting-drying is rather common.
4. Durability assessment of the Cappadocian tuff The durability of a stone is a measure of its ability to resist weathering and so retain its original size, shape, strength, and appearance over an extensive period of time (Bell, 1980, 1993; Sims, 1991). There are several means of assessing the durability of a stone. In this study, average pore diameter, saturation coefficient, wet-to-dry strength ratio, static rock durability index, index of rock durability, and slake-durability index are used. 4.1. Average pore diameter Average pore diameter is considered to be an important parameter for the freeze-thaw durability of stones. Larsen and Cady (1969) stated that the critical pore size is 5 prn, below which pore water
180
T Topal, K Doyuran / Engineering Geology 47 (1997) 175 187
Fig. 3. Map showing the distribution of joints in the Cappadocian tuff (after Topal and Doyuran, 1995).
cannot be drained out of the stone. Therefore, stones having an average pore size less than 5 ~tm are susceptible to frost damage. The average pore diameter of the Cappadocian tuff was calculated from intrusion data of the mercury porosimeter. The test results yielded a value of 0.11 ~tm (Topal, 1995), which suggests that the Cappadocian tuff is susceptible to frost damage.
4.2. Saturation coefficient Saturation coefficient (S) of a stone is the ratio between the natural capacity of a stone to absorb water after complete immersion under atmospheric pressure for a definite time, and the total pore volume that is accessible to water, i.e., S = (water absorption)/(effective porosity)
T. Topal, V. Doyuran / Engineering Geology 47 (1997) 175-187
181
8
o
~l
o °o °~
o o
_
e
| u
eP oo IJ
m
(A)
O,Zto 10~
!
?
lOto
i
tB)
20
20to 4,01 •l O t o 80Z 8 0 to l ~ O I t 160 ~o 2,10Z
Fig. 4. The pole plot (A) and the contour (B) diagrams of the joints in the Cappadocian tuff.
It is a dimensionless coefficient, mostly given as a percentage. A stone with a very high saturation coefficient may be deteriorated by freeze-thaw activity (Schaffer, 1972; RILEM, 1980). Therefore, this test can be useful for giving information on the freeze-thaw durability of a stone. The value of the saturation coefficient can mostly vary between 0.4 and 0.95 (BRE, 1983). A satura-
tion coefficient greater than 0.8 indicates low durability (susceptible to frost activity) (Hirschwald, 1912; TSE, 1977). However, many stones have saturation coefficients in the range of 0.66-0.77. In this range, the saturation coefficient gives an unreliable guide (Anon, 1975; BRE, 1983). The saturation coefficient of the Cappadocian tuff is found to be 0.78. Although this value is
182
T Topal, V. Doyuran / Engineering Geology 47 (1997) 175 187
close to the upper bound of the unreliable range mentioned above, it is also very near to the frost susceptibility boundary, which is 0.8. Adopting a conservative approach, the Cappadocian tuff may be considered to be frost susceptible based on the saturation coefficient. 4.3. Wet-to-dry strength ratio Swelling and non-swelling clays tend to attract water when exposed to moisture and this moisture significantly reduces the strength of a stone. Winkler (1986, 1993) suggested that the wet-todry strength ratio based on the modulus of rupture, or the uniaxial compressive strength, or the tensile strength, is a good and rapid method of testing the durability of a stone for use as a durability index. The general stone qualities using wet-to-dry strength ratios are presented by Winkler (1993) as follows: 80-90 good and safe; 70-80 further testing required; 60-70 unsafe, for frost and hygric forces; < 60 very poor quality, clay present. In this study, the durability index of the Cappadocian tuff is evaluated based on the saturated and dry uniaxial compressive strength of the rock. The wet-to-dry strength ratio of the tuff is determined as 33 and 19% in the vertical and horizontal directions, respectively. These ratios reveal that the Cappadocian tuff has very poor durability. 4. 4. Static rock durability index Fookes et al. (1988) suggested the use of static and dynamic rock durability indicators for the durability assessment of stone. Since the dynamic one is applicable only where dynamic loading conditions exist, in this study only the static rock durability is considered. The static durability indicator (RDI~) requires four engineering tests expressed as follows:
absorption by weight under atmospheric pressure (BSI, 1975); SGssd=saturated and surface dried relative density (BSI, 1975; CIRIA/CUR, 1991). The water absorption test result is multiplied by an arbitrary factor of 5 in the above equation to bring the magnitudes of the variables into equivalent terms, and to emphasize its importance in assessing the durability of rock. The point load strength is used to give an assessment of the static strength of the material, and is especially useful where material is not subject to dynamic loading. The magnesium sulphate soundness test is included to assess the ability of the material to resist some of the cyclic physical weathering processes, such as salt crystallization, heating-cooling, and possibly freezing-thawing (Fookes et al., 1988). A tentative estimation of the potential durabilities of rocks based on the static rock quality indicators is given in Table 2. For the calculation of the rock durability index of the Cappadocian tuff, Is~50~, is taken as 0.3 MPa (see Table 1). The magnesium sulphate soundness test is not performed on the Cappadocian tuff. However, the sodium sulphate soundness test performed on the tuff indicates complete collapse of the tuff (100% weight loss) after four cycles (Topal and Doyuran, 1996). Considering the fact that the magnesium sulphate soundness test is much more severe than the sodium sulphate soundness test (ASTM, 1990), a value of 100% is taken for SST values in the calculations. The water absorption (WA) of the tuff is 21.60%. The relative density of the tuff is 17.75 kN m -3. Use of these values in the formula given above reveals that the static durability indicator of the Cappadocian tuff is around - 12, which corresponds to poor rock class.
Table 2 T e n t a t i v e static d u r a b i l i t y e s t i m a t i o n o f r o c k s (after F o o k e s et al., 1988)
RDIs = (I~5o), - 0.1 ( SST + 5WA))/(SG~d)
RDI~
Potential durability
where I~tso),=average dry and saturated pointload index (ISRM, 1985); SST=magnesium sulphate soundness test (weight loss (%), at the fifth cycle) (Hosking and Tubey, 1969); WA=water
> 2.5 2.5 to 1 - 1 to - 3 < - 3
Excellent Good Fair Poor
T. Topal, 11. Doyuran / Engineering Geology 47 (1997) 175-187
4.5. Index of rock durability Assessment of rock durability through index properties is suggested by Rodrigues and Jeremias (1990). Swelling strain (linear strain), porosity, and compressive strength are used in the durability evaluation. The index of rock durability equation has the strength factor in the numerator and the summation of swelling strain and porosity in the denominator. Its mathematical expression is as follows: IRD = (R/R,)/(n + 2a) where IRD = index of rock durability; R = ultimate compressive strength of the rock (MPa); R,= 1 MPa (to turn the index adimensional); n = porosity (%); a = the mantissa of the swelling strain (e) when expressed as a x 10 -4. According to Rodrigues and Jeremias (1990) and Rodrigues (1993), an IRD of less than 2 indicates low durability, while values greater than 10 signify materials having good to excellent durability. They indicated that values between 2 and 10 need a large number of results to clarify this zone. Considering the compressive strength, effective porosity, and linear strain of the Cappadocian tuff, the index of rock durability is found to be 0.13 in the vertical direction and 0.10 in the horizontal direction. These values reveal that the Cappadocian tuff has low durability. 4.6. Slake-durability index The slake-durability index test is generally used to assess the resistance offered by a rock sample to weakening and disintegration when subjected to two standard cycles of wetting and drying (ISRM, 1981). The slake-durability index (Id2) is calculated using the following formula: /dE = [(C-- D)/(A - D)] × 100% where A = mass of dry drum plus dry sample; C = mass of dry drum plus retained portion of the dry sample after the second cycle of the test; D = mass of dry drum. In this study, the test is performed in accordance with ISRM (1981). Based on the test results on
183
two sets of samples, the slake-durability index of the Cappadocian tuff is found to be 84%. The durability classification given in ISRM (1981) and Goodman (1989) reveals that the Cappadocian tuff has medium durability.
5. G e n e r a l e v a l u a t i o n
5.1. Material characteristics In this section, the field observations and laboratory test results will be evaluated to better uncle rstand the engineering geological properties of theCappadocian tuff and its durability. The Cappadocian tuff has a very low unit weight and a very high porosity (Anon, 1979). The uniaxial compressive strength of the tuff in the vertical direction is slightly higher than that in the horizontal direction. Although the definitions for rock strength, especially at lower values are significantly different in various standards (Hencher, 1993), based on Anon (1977) and BSI (1981) recommendations, the Cappadocian tuff is moderately weak to weak when dry and weak to very weak when saturated. The point load strength index (Is~5o~) values of the Cappadocian tuff do not significantly change for the axial and diarnetral tests. It is customary to convert Is{5o~to an equivalent unconfined compressive strength by multiplying by a correlation factor, k. A wide scatter of k ranging between 8 and 54 is observed in the literature (Anon, 1972, 1977; Broch and Franklin, 1972; Bieniawski, 1975; Beawis et al., 1982; Foster, 1983; ISRM, 1985), although there is a preponderance of values between 16 and 24 (Norbury, 1986). In general, weak rocks like shale and siltstone give low k values, whereas strong rocks give high k values. In this study, however, a correlation factor of 13.6 is obtained if average I~5o} values are considered for the dry Cappadocian tuff in a vertical direction. This low figure is in good agreement with the correlation factor of the weaker rocks given in the literature. The sonic velocity, elastic constants, and linear strain values of the Cappadocian tuff do not show
184
T. Topal. V. Doyuran / Engineering Geology 47 (1997) 175-187
significant changes in the vertical and horizontal directions. This suggests that the tufts show no distinct anisotropy. However, a slight change of the strength of the tuff is noted in vertical (6.53 MPa) and horizontal (4.87 MPa) directions. The durability assessments of the Cappadocian tuff are summarized in Table 3. The average pore diameter, the wet-to-dry strength ratio, the static rock durability index, and the index of rock durability indicate a poor to very poor durability and a frost-susceptible rock for the Cappadocian tuff. The saturation coefficient also roughly indicates frost-susceptible rock. On the other hand, the slake-durability index yields medium durability. Comparison of the various durability classification methods reveals that, except for the slake-durability index, all the other methods generally indicate poor to very poor durability. In the case of the slake-durability classification, 16% weight loss is considered to be medium durability. It is the authors' idea that such a reduction in weight cannot be tolerated for the purpose of conservation since a few millimetres of surface erosion may cause loss of some valuable historical phenomena (e.g., inscription, carving, painting, etc.) on the stone surfaces. Therefore, the slake durability classification needs to be revised for the conservation studies. Based on the measurements of De Witte (1995), the field performance of the tuff may also be considered to be of poor durability. The rate of erosion of the fairy chimneys due to natural and artificial deterioration was measured by graduated stainless steel reference pins (10 mm in diameter and 100 mm long) driven into the tuff. The rate of erosion is found to be 4 mm year-1 between
1987 and 1992. The salt crystallization and freeze thaw tests performed on the tuff also reveal very low durability (Topal and Doyuran, 1996). Therefore, except for the slake-durability index, all the above mentioned durability assessment methods may be considered to be good indicators for the durability assessment of tufts.
5.2. Influence of structural features on fairy chimneys The joints do not only control the formation of the fairy chimneys but also their structural stability (Topal, 1995; Topal and Doyuran, 1995). This problem is particularly important for churches hewn within the fairy chimneys, such as the churches in the GOreme Open Air Museum. Steeply dipping joint sets are the main causes of rockfalls. Such joints are especially critical for churches located adjacent to a valley. In the case of the G/Sreme Open Air Museum, the joints run both parallel and nearly perpendicular to the adjacent valley. Those running parallel to the valley are generally widened due to atmospheric phenomena and loss of lateral support. The problem of structural instability is further complicated by the presence of artificial cracks formed due to the man-made excavations (rock-hewn) (Fig. 5). Considering the fact that the engineering properties of the discontinuities are adversely affected by water, the stable block may become potentially unstable during winter and spring seasons since moisture reduces the shear strength parameters along the discontinuities. The joints and artificial cracks are also places where rainwater seeps and causes discoloration of the wall paintings of the
Table 3 Durability assessment of the Cappadocian tuff using various methods Durability method
Durability of the Cappadocian tuff
Remarks
Average pore diameter (Larsen and Cady, 1969) Saturation coefficient(TSE, 1977) Wet-to-dry strength ratio (Winkler, 1986) Static rock durability index (Fookes et al., 1988) Index of rock durability (Rodrigues and Jereimas, 1990) Slake-durability Index (ISRM, 1981)
frost susceptible frost susceptible very poor durability poor durability low (poor) durability medium durability
mercury intrusion data is used very close to frost susceptiblerange uniaxial compressivestrength is used magnesium sulphate soundness value is estimated weight loss after 2nd cycleof the test is used
1". Topal, K Doyuran / Engineering Geology 47 (1997) 175-187
185
conservation studies. Thus, moisture of any kind should be avoided in and around the rock-hewn churches. The joints and the artificial cracks may be sealed by proper mixture of tilling materials having properties similar to that of the tuff. The unstable blocks at the fairy chimneys and adjacent to the valley walls must be identified and stabilized by non-corrosive type rock bolts.
6. Conclusions
Fig. 5. Structural collapse in a rock-hewn dwelling. Note collapsed section (upper left) and exposed living chamber (foreground).
churches by chemically altering the biotites and the rock fragments present within the Cappadocian tuff. The water also significantly reduces the uniaxial compressive strength of the tuff. This reduction ranges between 67% in vertical and 81% in horizontal directions. Such strength reduction will obviously create structural instability problems. The erosion of the exposed outer surface of the tuff (about 4 mm year-l; De Witte, 1995) gradually reduces the wall thickness of the rock-hewn churches. The reduced wall thickness together with reduced strength due to saturation make rockhewn churches more susceptible to structural collapse. The significant reduction in the strength of the tuff due to moisture, poor to very poor durability, and the adverse effects of joints on the structural stability need to be considered during
The investigation of the engineering geological properties of the Cappadocian tuff reveals that the tuff is almost fresh with local discoloration, and is moderately weak to very weak. It has a low unit weight, very high porosity, and high deformability. The engineering properties of the tuff do not show significant changes in vertical and horizontal directions. The uniaxial compressive strength of the tuff is significantly reduced due to saturation. The durability assessment methods suggest that the tuff has poor to very poor durability. The joints and/or the artificial cracks are the main causes of structural instability for the rock-hewn churches. Deterioration of tufts may be reduced by avoiding moisture and sealing of discontinuities. For structural stability of the churches, application of noncorrosive type rock bolts to secure unstable blocks is recommended.
Acknowledgment This study is financially supported by METU Research Fund Project (AFP). Thanks are due to Dr. Eddy De Witte from the Royal Institute for Cultural Heritage (Belgium), for the measurement of the average pore diameter of the tuff by mercury porosity measurements.
References ASTM, Designation C88-83, 1990. Standard test method for soundness of aggregates by use of sodium sulfate or
186
Z Topal. Id Doyuran / Engineering Geology 47 (1997) 175 187
magnesium sulfate. Annual Book of ASTM Standards, Vol. 04.02, pp. 39-43. Anon,1972. The preparation of maps and plans in terms of engineering geology. Q. J. Eng. Geol. 5,293-382. Anon,1975. Testing porous building stone: Stone Handbook. Architects J. 13, 337-339. Anon,1977. The description of rock masses for engineering purposes. Q. J. Eng. Geol. 10, 355--389. Anon,1979. Classification of rocks and soils for engineering geological mapping. Part 1. Rock and soil materials. Bull. Int. Assoc. Eng. Geol. 19, 364- 371. BSI, 1975. Testing aggregates: Part 2. Methods for determining the physical properties. British Standards Institution (BS 812), 20 pp. BSI, 198 l. Code of Practice for Site Investigations. British Standards Institution (BS 5930), 147 pp. Beawis, F.C., Roberts, F.I., Minskaya, L., 1982. Engineering aspects of weathering of low grade metapelites in an arid climatic zone. Q. J. Eng. Geol. 15, 29-45. Bell, F.G., 1980. Engineering Geology and Geotechnics. Butterworth and Co. Ltd., London, 497 pp. Bell, F.G., 1993. Engineering Geology. Blackwetl Scientific Publications, London, 359 pp. Bieniawski, Z.T., 1975. The point load test in geotechnical practice. Eng. Geol. 9, 1 -11. Broch, E., Franklin, J.A., 1972. The point load strength test. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 9, 669 697. BRE, 1983. The selection of natural building stone. Building Research Establishment Digest 290, 8 pp. CIRIA/CUR, 1991. Manual on the use of rock in coastal and shoreline engineering. Construction Industry Research and Information Association, CIRIA Special Publication 83, CUR Report 154, 607 pp. De Witte, E., 1995. Conservation of the Gtreme rock. Proc. Int. Seminar on the Safeguard of the Rock-Hewn Churches of the Gtreme Valley, ICCROM, pp. 109 124. De Witte, E., Terfve, A., Koestler, R.J., Charola, A.E., 1988. Conservation of the G6reme rock: preliminary investigations, Proc. 6th. Int. Cong. on Deterioration and Conservation of Stone, Torun, Poland, pp. 346 355. Erdo~an, M., 1986. Nev~ehir-Orgtip y~resi ttlflerinin malzeme jeolojisi aqismdan ara~tlnlmasl, i.T.U. Maden Fak.. Doktora Tezi, 100 pp. Ergnvauh, A.K., Ytizer, A.E., 1977. Past and present use of underground openings excavated in volcanic tufts at Cappadocia area. Proc. on Rock Storage, Oslo, Norway, pp. 15-17. Erguvanh, K., Yorulmaz, M., ~lli, F., Ahunbay, Z., Erdo~an M., 1989. GOreme yaplsal koruma ve sa~lamla~tirma projesi, E1 Nazar kilisesi, i.T.O., 46 pp. Fookes, P.G., Gourley, C.S., Ohikere, C., 1988. Rock weathering in engineering time. Q. J. Eng. Geol. 21.33 57. Foster, I.R., 1983. The influence of core sample geometry on the axial point load test.. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 20, 291 295. Goodman, R.E., 1989. Introduction to Rock Mechanics, 2nd ed. John Wiley and Sons, New York, 562 pp.
Hencher, S.R., 1993. Conference summary. The Engineering Geology of Weak Rock. Balkema, 504 pp. Hirschwald, J., 1912. Handbuch der bautechnischen Genteinsprtifung. Borntraeger, Berlin. Hosking, J.R., Tubey, W., 1969. Research on low-grade and unsound aggregates. Road Research Laboratory Report, LR 293, Crowthorne (UK), 30 pp. ISRM, 1981. Rock Characterization, Testing and Monitoring. Pergamon Press, Oxford, 211 pp. ISRM,1985. Suggested method for determining point load strength. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 22, 51 60. Larsen, T.D., Cady, P.D., 1969. Identification of frost susceptible particles in concrete aggregates, National Cooperative Research Program, Report 66, Washington, DC, Highway Research Board, 62 pp. METU, 1987. Investigation of the mechanisms of stone deterioration for the purpose of conservation of GOreme tufts, Progress report, 61 pp. Norbury, D.R., 1986. The point load test. Site investigation practice: assessing BS 5930. Geol. Soc. Eng. Geol. Special Publ. 2, 423 pp. Pasquare, G., 1968. Geology of the Cenozoic volcanic area of Central Anatolia, Memoire. Serie. 8, Vol. 9, Fasc. 3, Roma, pp. 57-201. RILEM,1980. Recommended tests to measure the deterioration of stone and to assess the effectiveness of treatment methods. Commission 25-PEM. Mater. Struet. 13, 175-253. Rockware, 1993. STEREO 3-D orientation analysis and plotting. Version 3, Rockware Scientific Software, USA, 56 pp. Rodrigues, J.D., Jeremias, F.T., 1990. Assesment of rock durability through index properties. Proc. 6th. Int. Congress IAEG, Amsterdam, Balkema-Rotterdam, Vol. 4, pp. 3055 3060. Rodrigues, J.D., 1993. Stone conservation in Portugal: problems, methods and means. Int. Workshop on Conservation of Architectural Surfaces: Stones and Wall Covering, Venice, pp. 147 154. Schaffer, R.J., 1972. The weathering of natural building stones. Dept. Sci. Ind. Res. Building Research Spec. Rept. No. 18, 149 pp. Sims, I., 1991. Quality and durability of stone for construction. Q. J. Eng. Geol. 24, 67-73. Topal, T., 1995. Formation and deterioration of fairy chimneys of the Kavak tuff in Orgiip-Gtreme area (Nev~ehir-Turkey). Ph.D. Thesis, Middle East Technical University, Ankara, Turkey, 250 pp. Topal, T., Doyuran, V., 1994. Effect of deterioration on the chemical and physical properties of G6reme tufts. Proc. 7th. Int. Cong. of IAEG, Lisboa, Vol. 5, pp. 3527-3536. Topal, T., Doyuran, V., 1995. Effect of discontinuities on the development of fairy chimneys in the Cappadocia region (Central Anatolia-Turkey). Turkish J. Earth Sci. (TISIBITAK) 4, 49-54. Topal, T., Doyuran, V., 1996, Ultrasonic testing of artificially weathered Cappadocian tuff, LCP-95 Cong. on Preservation
T. Topal, V. Doyuran / Engineering Geology 47 (1997) 175-187 and Restoration of Cultural Heritage, Montreux, pp. 205-211. TSE, 1977. Dotal yap~ ta~lan, Tilrk Standartlarl EnstitOsO, TS2513, 6 pp. TSE, 1978. Dotal yap1 ta~lanmn muayene ve deney metodlan. Tiirk Standartlarl EnstitOsO, TS699, 16 pp.
187
Winkler, E.M., 1986. A durability index for stone. Bull. Assoc. Eng. Geol. 23, 344-347. Winkler, E.M., 1993. Discussion and reply on "The durability of sandstone as a building stone, especially in urban environments". Bull. Assoc. Eng. Geol. 30, 99-101.
Engineering Geology47 (1997) 175-187
Engineering geological properties and durability assessment of the Cappadocian tuff T. Topal *, V. Doyuran Department of Geological Engineering, Middle East Technical University, 06531, Ankara, Turkey Received 25 June 1996; accepted 1 February 1997
Abstract
Some of the unique landforms, the so-called "fairy chimneys", which were formed within the Cappadocian tuff of central Turkey, were used as dwellings in the past and contain culturally valuable wall paintings. However, these structures are undergoing chemical and physical deterioration due to atmospheric effects. For conservation studies, an understanding of the engineering geological properties of the tuff is essential. In this study, emphasis is placed on both material and mass properties of the tuff. These properties are evaluated for the assessment of rock durability. This study shows that the Cappadocian tuff is almost fresh, with local discoloration, is moderately weak to very weak, and has low unit weight, very high porosity, and high deformability. Discontinuity surveys revealed two dominant joint sets, which not only controlled the formation but also control the structural stability of the fairy chimneys. Various methods used for the durability assessment of the Cappadocian tuff indicate poor to very poor durability. Due consideration must be given to strength reduction due to moisture, poor to very poor durability, and the adverse effects of joints on the structural stability of the fairy chimneys. © 1997 Elsevier Science B.V.
Keywords: Cappadocia; Durability; Engineering geology; Tuff
1. Introduction
The Cappadocia region comprising Nev~ehir, Kayseri, and Ni~de provinces of Central Anatolia (Fig. 1), is one of the attractive tourist sites of Turkey due to its spectacular and unique landforms, fairy chimneys, and historical heritages. The fairy chimneys developed within the tufts are the results of weathering and differential erosion. Some of the fairy chimneys were dwellings and were inhabited during the Byzantine times. Some
* Correspondingauthor. Fax: +90 312 2101263; e-mail: [email protected] 0013-7952/97/$17.00© 1997ElsevierScienceB.V. All fights reserved. PII S0013-7952 (97) 00017-3
such dwellings contain invaluable wall paintings, which also provide historical attraction. Weathering and erosion, however, are still active and these natural processes, along with human activities, threaten the stability of the chimneys. Studies towards conservation of these valuable historical and touristic phenomena necessitate the determination of the engineering geological (material and mass) properties and the durability of the tufts. In this study, material and mass properties of the tuff are determined. The average pore diameter, saturation coefficient, wet-to-dry strength ratio, static rock durability index, index of rock durability, and slake-durability index of the tuff are used
T. Topal. ['~ Doyuran
176
"~/Io-
/
Engineering Geology 47 (1997) 175 187
N
to KJr$¢hir I!
KAYSERi ,AREA NEVSEHiR 3917
Mustafapasa
&
Erciyes
da~,
AKSARAY 0 .-~
3268 A
Hasan da(]l
I0~
3O
I
i
Km
2963 &
i
MeLcndiz d(:l~u
,~
N i GDE
h I
~.:.
... '-.-' ..... <.-~ ,.
Fig. I. Location map of the study area.
for the durability assessment of the rock, and of the fairy chimneys. Engineering geological properties and the assessed durability of the tuff are re-evaluated for the purpose of conservation of the fairy chimneys. The study is carried out within the Urgtip-G6reme region (Fig. 1) due to the occurrences of typical fairy chimneys in this area (Fig. 2).
2. Site geology The Cappadocia region is generally underlain by thick and extensive deposits of the volcanosedimentary sequence (Miocene-Pliocene) of the Orgtip formation (Pasquare, 1968). Although the formation comprises a number of well-distinguished members, in the study area only the Kavak
and Tahar members contain the fairy chimneys. Both members are characterized by non-welded tufts. Since most of the dwellings and wall paintings are found within the fairy chimneys of the Kayak member, our study is focused on this member. The Kavak member represents the product of the first intermittent volcanic activity which produced the LlrgOp formation. Thus, this member constitutes the basal portion of the volcanosedimentary sequence. The chaotic arrangement of pumice fragments within the unit suggests that it is an ash flow tuff deposited in a lacustrine environment. It is dirty white to pink, and contains phenocrysts of plagioclase, quartz, biotite and opaque minerals. Various rock fragments and pumice are also commonly observed. In the matrix, volcanic glass shards are rather common (Topal
T. Topal, F. Doyuran / Engineering Geology 47 (1997) 175-187
177
to assess anisotropy. These directions are expected to represent the material characteristics of the tuff perpendicular and parallel to layering. The laboratory tests, performed according to ISRM (1981), RILEM (1980) and TSE (1978), included determination of the index and some mechanical properties of the intact rock specimens. The field observations involved assessment of the rock mass properties, such as discontinuity characteristics and weathering grade. The description of rock material and mass characteristics is based on Anon (1977), BSI (1981), and ISRM (1981).
3.1. Material properties of the Cappadocian tuff The Cappadocian tuff is dirty white to pink, consisting mostly of lapilli size (2-64 mm) phenocrysts and rock fragments embedded within tuffaceous matrix. The tuff is almost fresh with very local discoloration around biotites and some rock fragments. The results of index and mechanical tests conducted by the authors and those of other authors are given in Table 1.
3.2. Mass properties of the Cappadocian tuff Fig. 2. Typical view of fairy chimneys with cap rocks.
and Doyuran, 1994; METU, 1987). The volcanic glass shards are partly altered to smectite. In this paper, the term "Cappadocian tuff" is used to denote the Kavak member.
3. Engineering geological properties of the Cappadocian tuff Evaluation of the engineering geological characteristics of the Cappadocian tuff is based on field observations and laboratory studies. Since the fairy chimneys are essentially formed within ignimbritic levels of the Cappadocian tuff, the material and mass properties of these levels will be considered in the following sections. For the laboratory studies, core samples extruded from 22 oriented block samples are used. Although the Cappadocian tuff does not show distinct layering within ignimbritic levels, cores from the oriented block samples are taken in the vertical and horizontal directions
Within the thickly bedded and almost horizontally layered Cappadocian tuff, the joints constitute the major discontinuities. In the study area, the joints are locally well-developed and can easily be traced for a considerable distance. Some, however, can only be traced due to the alignment of vegetation supported by moisture along the persistent joints. A number of scan line surveys, spot measurements, and field observations were carried out in the field to determine the orientation, spacing, aperture, persistence, infilling, surface roughness, water seepage and weathering state of the joints. The orientation of the joints (i.e., strike and dip) were determined using geological compass. In the field, 1905 joint orientation measurements were taken at 150 locations (Fig. 3). In order to determine the dominant sets, the joint orientation measurements were evaluated using a computer program called "Rockware" (Rockware, 1993). For the pole plots, a Schmidt diagram with lower hemisphere projection was
178
T. Topal, IL Doyuran / Engineering Geology 47 (1997) 175 187
Table 1 Correlation table for the material properties of the Cappadocian tuff Material properties
This study No. of samples
Unit weight (kN m 3) dry saturated Effective porosity (%) Water absorption by weight (%) under atmospheric pressure by weight (%) under pressure by weight (%) Saturation coefficient Uniaxial compressive strength (MPa) dry-vertical (MPa) dry-horizontal (MPa) saturated-vertical (MPa) saturated-horizontal (MPa) Point load strength, I,~5o (MPa) dry saturated Schmidt rebound hardnessa: dry Indirect (Brazilian) tensile strength: dry (MPa) Slake Durability (%): 2nd cycle P-wave velocityc (m s 1) dry-vertical dry-horizontal saturated-vertical saturated-horizontal S-wave velocityd (m s ~) dry-vertical dry-horizontal saturated-vertical saturated-horizontal ) Modulus of elasticity, E (GPa) dry-vertical dry-horizontal Poisson's ratio, v dry-vertical dry-horizontal Dynamic modulus of elasticity (GPa) dry-vertical dry-horizontal saturated-vertical saturated -horizontal Dynamic Poisson's ratio dry-vertical dry-horizontal saturated-vertical saturated-horizontal Internal friction angle% ~ (°)
Erguvanll and Y~zer (1977)
Erdogan (1986)
METU (1987)
De Witte et al. (1988)
Erguvanh et al. (1989)
Mean_+ SD
10 10 10
13.60+0.11 17.75_+0.08 38.29_+0.38
10
21.60_+0.27
10 10
28.08 _+0.46 0.78_+0.01
10 10 5 5
6.53_+0.67 4.87 _+0.43 2.16+0.34 0.93 -+0.29
20 20 20 10
0.48_+0.10 0.13_+0.04 27_+3.3 0.69 -+0.08
2
84_+0.17
I0 5 10 5
2082.46 _+84.10 1973.86_+41.60 1388.87_+85.00 1264.20 _+38.90
10 5 10 5
1124.67 _+77.30 1046.06_+47.60 759.43 _+41.20 724.18_+68.70
10 10
3.08 _+0.49 3.26 _+0.45
10 10
0.20+0.02 0.19_+0.01
10 5 10 5
4.47+0.53 3.90+0.29 2.64+0.27 2.33+0.34
10 5 10 5 10
0.29 _+0.03 0.30+0.02 0.28 + 0.02 0.25 _+0.06 28 + 2.95
11.50 28
15.10 19.20 28.76
13.90 -32.72
25
27.50
--
14.02
15.90
42
19.20 28.76 27.50
0.64
5.5
0.60
6.5
6.07
3.0
--
0.40
--
40-48 0.7
6.5 3.0
20 b 0.70
-0.83
2058
1.40
2.4
4.2
3.9
40
49
2.40
13
19
T. Topal, V. Doyuran / Engineering Geology47 (1997) 175-187
179
Table 1 (Continued)
Material properties
Cohesion e, C ( M P a ) Linear strain ( m m m - 1) vertical horizontal Average pore diameter (inn) Permeability (m s -1) T h e r m a l conductivity coefficient
This study
Erguvanh and Erdo~an METU De Witteet al. Yiizer (1977) (1986) (1987) (1988)
No. of samples
M e a n 4- SD
10
2.28 4-0.15
1.20
0.9
7 7 4 ---
0.55 4-0.03 0.51 +0.07 0.11 4-0.01 ---
-0.07 --0.32
---0.00 0.43
--
--
--
1.8
E r g u v a n h et al. (1989)
1.9
m m
0.00 0.43
(kcal/hm°C) T h e r m a l expansion coefficient (×
10 -6
6.3
6.3
I/°C)
Capillary water a b s o r p t i o n coefficient ( g / m 2. min °'s) Drying rate at 20°C and 40% R H ( g / m 2. h °'5) Water retention at 20°C (%)
.
.
.
.
187
.
.
.
.
444
.
.
.
.
4.85
"L-type Schmidt h a m m m e r is used. bField measurement. °For P-wave measurements, 63-kHz transducers are used ( N e w Sonicviwer, 5217A-OYO). dFor S-wave measurements, 100-kHz transducers and high voltage generators are used. ~ and C values are obtained f r o m triaxial compression tests, using strength envelope.
adopted. The contour diagram was then obtained from the pole plots (Fig. 4). The diagrams yield two dominant joint sets: N78°E/85-90°NW and N-S/90 °. The spacing of the joints generally ranges between 0.1 and 20m (moderately wide to extremely wide) (Anon, 1977) and their persistence exceeds 10m (high to very high persistence) (ISRM, 1981 ). Studies performed on 1:35 000 scale aerial photographs have shown that some of the joints may be traced for a distance of 800 m. The apertures of most joints fall within narrow to tight (0-20mm) categories (Anon, 1977). The joint surfaces are generally rough planar. Some narrow joints contain clayey infilling, but most are free of infilling material. During field inspections no groundwater seepage was noted through the joints. The tufts are faintly to slightly weathered. Due to the steep topography, slope wash, and lack of vegetation the weathering products are soon removed. Along the persistent joints, discoloration and/or surface staining is commonly observed. Spalling of exposed surfaces due to alternate
freezing-thawing, and wetting-drying is rather common.
4. Durability assessment of the Cappadocian tuff The durability of a stone is a measure of its ability to resist weathering and so retain its original size, shape, strength, and appearance over an extensive period of time (Bell, 1980, 1993; Sims, 1991). There are several means of assessing the durability of a stone. In this study, average pore diameter, saturation coefficient, wet-to-dry strength ratio, static rock durability index, index of rock durability, and slake-durability index are used. 4.1. Average pore diameter Average pore diameter is considered to be an important parameter for the freeze-thaw durability of stones. Larsen and Cady (1969) stated that the critical pore size is 5 prn, below which pore water
180
T Topal, K Doyuran / Engineering Geology 47 (1997) 175 187
Fig. 3. Map showing the distribution of joints in the Cappadocian tuff (after Topal and Doyuran, 1995).
cannot be drained out of the stone. Therefore, stones having an average pore size less than 5 ~tm are susceptible to frost damage. The average pore diameter of the Cappadocian tuff was calculated from intrusion data of the mercury porosimeter. The test results yielded a value of 0.11 ~tm (Topal, 1995), which suggests that the Cappadocian tuff is susceptible to frost damage.
4.2. Saturation coefficient Saturation coefficient (S) of a stone is the ratio between the natural capacity of a stone to absorb water after complete immersion under atmospheric pressure for a definite time, and the total pore volume that is accessible to water, i.e., S = (water absorption)/(effective porosity)
T. Topal, V. Doyuran / Engineering Geology 47 (1997) 175-187
181
8
o
~l
o °o °~
o o
_
e
| u
eP oo IJ
m
(A)
O,Zto 10~
!
?
lOto
i
tB)
20
20to 4,01 •l O t o 80Z 8 0 to l ~ O I t 160 ~o 2,10Z
Fig. 4. The pole plot (A) and the contour (B) diagrams of the joints in the Cappadocian tuff.
It is a dimensionless coefficient, mostly given as a percentage. A stone with a very high saturation coefficient may be deteriorated by freeze-thaw activity (Schaffer, 1972; RILEM, 1980). Therefore, this test can be useful for giving information on the freeze-thaw durability of a stone. The value of the saturation coefficient can mostly vary between 0.4 and 0.95 (BRE, 1983). A satura-
tion coefficient greater than 0.8 indicates low durability (susceptible to frost activity) (Hirschwald, 1912; TSE, 1977). However, many stones have saturation coefficients in the range of 0.66-0.77. In this range, the saturation coefficient gives an unreliable guide (Anon, 1975; BRE, 1983). The saturation coefficient of the Cappadocian tuff is found to be 0.78. Although this value is
182
T Topal, V. Doyuran / Engineering Geology 47 (1997) 175 187
close to the upper bound of the unreliable range mentioned above, it is also very near to the frost susceptibility boundary, which is 0.8. Adopting a conservative approach, the Cappadocian tuff may be considered to be frost susceptible based on the saturation coefficient. 4.3. Wet-to-dry strength ratio Swelling and non-swelling clays tend to attract water when exposed to moisture and this moisture significantly reduces the strength of a stone. Winkler (1986, 1993) suggested that the wet-todry strength ratio based on the modulus of rupture, or the uniaxial compressive strength, or the tensile strength, is a good and rapid method of testing the durability of a stone for use as a durability index. The general stone qualities using wet-to-dry strength ratios are presented by Winkler (1993) as follows: 80-90 good and safe; 70-80 further testing required; 60-70 unsafe, for frost and hygric forces; < 60 very poor quality, clay present. In this study, the durability index of the Cappadocian tuff is evaluated based on the saturated and dry uniaxial compressive strength of the rock. The wet-to-dry strength ratio of the tuff is determined as 33 and 19% in the vertical and horizontal directions, respectively. These ratios reveal that the Cappadocian tuff has very poor durability. 4. 4. Static rock durability index Fookes et al. (1988) suggested the use of static and dynamic rock durability indicators for the durability assessment of stone. Since the dynamic one is applicable only where dynamic loading conditions exist, in this study only the static rock durability is considered. The static durability indicator (RDI~) requires four engineering tests expressed as follows:
absorption by weight under atmospheric pressure (BSI, 1975); SGssd=saturated and surface dried relative density (BSI, 1975; CIRIA/CUR, 1991). The water absorption test result is multiplied by an arbitrary factor of 5 in the above equation to bring the magnitudes of the variables into equivalent terms, and to emphasize its importance in assessing the durability of rock. The point load strength is used to give an assessment of the static strength of the material, and is especially useful where material is not subject to dynamic loading. The magnesium sulphate soundness test is included to assess the ability of the material to resist some of the cyclic physical weathering processes, such as salt crystallization, heating-cooling, and possibly freezing-thawing (Fookes et al., 1988). A tentative estimation of the potential durabilities of rocks based on the static rock quality indicators is given in Table 2. For the calculation of the rock durability index of the Cappadocian tuff, Is~50~, is taken as 0.3 MPa (see Table 1). The magnesium sulphate soundness test is not performed on the Cappadocian tuff. However, the sodium sulphate soundness test performed on the tuff indicates complete collapse of the tuff (100% weight loss) after four cycles (Topal and Doyuran, 1996). Considering the fact that the magnesium sulphate soundness test is much more severe than the sodium sulphate soundness test (ASTM, 1990), a value of 100% is taken for SST values in the calculations. The water absorption (WA) of the tuff is 21.60%. The relative density of the tuff is 17.75 kN m -3. Use of these values in the formula given above reveals that the static durability indicator of the Cappadocian tuff is around - 12, which corresponds to poor rock class.
Table 2 T e n t a t i v e static d u r a b i l i t y e s t i m a t i o n o f r o c k s (after F o o k e s et al., 1988)
RDIs = (I~5o), - 0.1 ( SST + 5WA))/(SG~d)
RDI~
Potential durability
where I~tso),=average dry and saturated pointload index (ISRM, 1985); SST=magnesium sulphate soundness test (weight loss (%), at the fifth cycle) (Hosking and Tubey, 1969); WA=water
> 2.5 2.5 to 1 - 1 to - 3 < - 3
Excellent Good Fair Poor
T. Topal, 11. Doyuran / Engineering Geology 47 (1997) 175-187
4.5. Index of rock durability Assessment of rock durability through index properties is suggested by Rodrigues and Jeremias (1990). Swelling strain (linear strain), porosity, and compressive strength are used in the durability evaluation. The index of rock durability equation has the strength factor in the numerator and the summation of swelling strain and porosity in the denominator. Its mathematical expression is as follows: IRD = (R/R,)/(n + 2a) where IRD = index of rock durability; R = ultimate compressive strength of the rock (MPa); R,= 1 MPa (to turn the index adimensional); n = porosity (%); a = the mantissa of the swelling strain (e) when expressed as a x 10 -4. According to Rodrigues and Jeremias (1990) and Rodrigues (1993), an IRD of less than 2 indicates low durability, while values greater than 10 signify materials having good to excellent durability. They indicated that values between 2 and 10 need a large number of results to clarify this zone. Considering the compressive strength, effective porosity, and linear strain of the Cappadocian tuff, the index of rock durability is found to be 0.13 in the vertical direction and 0.10 in the horizontal direction. These values reveal that the Cappadocian tuff has low durability. 4.6. Slake-durability index The slake-durability index test is generally used to assess the resistance offered by a rock sample to weakening and disintegration when subjected to two standard cycles of wetting and drying (ISRM, 1981). The slake-durability index (Id2) is calculated using the following formula: /dE = [(C-- D)/(A - D)] × 100% where A = mass of dry drum plus dry sample; C = mass of dry drum plus retained portion of the dry sample after the second cycle of the test; D = mass of dry drum. In this study, the test is performed in accordance with ISRM (1981). Based on the test results on
183
two sets of samples, the slake-durability index of the Cappadocian tuff is found to be 84%. The durability classification given in ISRM (1981) and Goodman (1989) reveals that the Cappadocian tuff has medium durability.
5. G e n e r a l e v a l u a t i o n
5.1. Material characteristics In this section, the field observations and laboratory test results will be evaluated to better uncle rstand the engineering geological properties of theCappadocian tuff and its durability. The Cappadocian tuff has a very low unit weight and a very high porosity (Anon, 1979). The uniaxial compressive strength of the tuff in the vertical direction is slightly higher than that in the horizontal direction. Although the definitions for rock strength, especially at lower values are significantly different in various standards (Hencher, 1993), based on Anon (1977) and BSI (1981) recommendations, the Cappadocian tuff is moderately weak to weak when dry and weak to very weak when saturated. The point load strength index (Is~5o~) values of the Cappadocian tuff do not significantly change for the axial and diarnetral tests. It is customary to convert Is{5o~to an equivalent unconfined compressive strength by multiplying by a correlation factor, k. A wide scatter of k ranging between 8 and 54 is observed in the literature (Anon, 1972, 1977; Broch and Franklin, 1972; Bieniawski, 1975; Beawis et al., 1982; Foster, 1983; ISRM, 1985), although there is a preponderance of values between 16 and 24 (Norbury, 1986). In general, weak rocks like shale and siltstone give low k values, whereas strong rocks give high k values. In this study, however, a correlation factor of 13.6 is obtained if average I~5o} values are considered for the dry Cappadocian tuff in a vertical direction. This low figure is in good agreement with the correlation factor of the weaker rocks given in the literature. The sonic velocity, elastic constants, and linear strain values of the Cappadocian tuff do not show
184
T. Topal. V. Doyuran / Engineering Geology 47 (1997) 175-187
significant changes in the vertical and horizontal directions. This suggests that the tufts show no distinct anisotropy. However, a slight change of the strength of the tuff is noted in vertical (6.53 MPa) and horizontal (4.87 MPa) directions. The durability assessments of the Cappadocian tuff are summarized in Table 3. The average pore diameter, the wet-to-dry strength ratio, the static rock durability index, and the index of rock durability indicate a poor to very poor durability and a frost-susceptible rock for the Cappadocian tuff. The saturation coefficient also roughly indicates frost-susceptible rock. On the other hand, the slake-durability index yields medium durability. Comparison of the various durability classification methods reveals that, except for the slake-durability index, all the other methods generally indicate poor to very poor durability. In the case of the slake-durability classification, 16% weight loss is considered to be medium durability. It is the authors' idea that such a reduction in weight cannot be tolerated for the purpose of conservation since a few millimetres of surface erosion may cause loss of some valuable historical phenomena (e.g., inscription, carving, painting, etc.) on the stone surfaces. Therefore, the slake durability classification needs to be revised for the conservation studies. Based on the measurements of De Witte (1995), the field performance of the tuff may also be considered to be of poor durability. The rate of erosion of the fairy chimneys due to natural and artificial deterioration was measured by graduated stainless steel reference pins (10 mm in diameter and 100 mm long) driven into the tuff. The rate of erosion is found to be 4 mm year-1 between
1987 and 1992. The salt crystallization and freeze thaw tests performed on the tuff also reveal very low durability (Topal and Doyuran, 1996). Therefore, except for the slake-durability index, all the above mentioned durability assessment methods may be considered to be good indicators for the durability assessment of tufts.
5.2. Influence of structural features on fairy chimneys The joints do not only control the formation of the fairy chimneys but also their structural stability (Topal, 1995; Topal and Doyuran, 1995). This problem is particularly important for churches hewn within the fairy chimneys, such as the churches in the GOreme Open Air Museum. Steeply dipping joint sets are the main causes of rockfalls. Such joints are especially critical for churches located adjacent to a valley. In the case of the G/Sreme Open Air Museum, the joints run both parallel and nearly perpendicular to the adjacent valley. Those running parallel to the valley are generally widened due to atmospheric phenomena and loss of lateral support. The problem of structural instability is further complicated by the presence of artificial cracks formed due to the man-made excavations (rock-hewn) (Fig. 5). Considering the fact that the engineering properties of the discontinuities are adversely affected by water, the stable block may become potentially unstable during winter and spring seasons since moisture reduces the shear strength parameters along the discontinuities. The joints and artificial cracks are also places where rainwater seeps and causes discoloration of the wall paintings of the
Table 3 Durability assessment of the Cappadocian tuff using various methods Durability method
Durability of the Cappadocian tuff
Remarks
Average pore diameter (Larsen and Cady, 1969) Saturation coefficient(TSE, 1977) Wet-to-dry strength ratio (Winkler, 1986) Static rock durability index (Fookes et al., 1988) Index of rock durability (Rodrigues and Jereimas, 1990) Slake-durability Index (ISRM, 1981)
frost susceptible frost susceptible very poor durability poor durability low (poor) durability medium durability
mercury intrusion data is used very close to frost susceptiblerange uniaxial compressivestrength is used magnesium sulphate soundness value is estimated weight loss after 2nd cycleof the test is used
1". Topal, K Doyuran / Engineering Geology 47 (1997) 175-187
185
conservation studies. Thus, moisture of any kind should be avoided in and around the rock-hewn churches. The joints and the artificial cracks may be sealed by proper mixture of tilling materials having properties similar to that of the tuff. The unstable blocks at the fairy chimneys and adjacent to the valley walls must be identified and stabilized by non-corrosive type rock bolts.
6. Conclusions
Fig. 5. Structural collapse in a rock-hewn dwelling. Note collapsed section (upper left) and exposed living chamber (foreground).
churches by chemically altering the biotites and the rock fragments present within the Cappadocian tuff. The water also significantly reduces the uniaxial compressive strength of the tuff. This reduction ranges between 67% in vertical and 81% in horizontal directions. Such strength reduction will obviously create structural instability problems. The erosion of the exposed outer surface of the tuff (about 4 mm year-l; De Witte, 1995) gradually reduces the wall thickness of the rock-hewn churches. The reduced wall thickness together with reduced strength due to saturation make rockhewn churches more susceptible to structural collapse. The significant reduction in the strength of the tuff due to moisture, poor to very poor durability, and the adverse effects of joints on the structural stability need to be considered during
The investigation of the engineering geological properties of the Cappadocian tuff reveals that the tuff is almost fresh with local discoloration, and is moderately weak to very weak. It has a low unit weight, very high porosity, and high deformability. The engineering properties of the tuff do not show significant changes in vertical and horizontal directions. The uniaxial compressive strength of the tuff is significantly reduced due to saturation. The durability assessment methods suggest that the tuff has poor to very poor durability. The joints and/or the artificial cracks are the main causes of structural instability for the rock-hewn churches. Deterioration of tufts may be reduced by avoiding moisture and sealing of discontinuities. For structural stability of the churches, application of noncorrosive type rock bolts to secure unstable blocks is recommended.
Acknowledgment This study is financially supported by METU Research Fund Project (AFP). Thanks are due to Dr. Eddy De Witte from the Royal Institute for Cultural Heritage (Belgium), for the measurement of the average pore diameter of the tuff by mercury porosity measurements.
References ASTM, Designation C88-83, 1990. Standard test method for soundness of aggregates by use of sodium sulfate or
186
Z Topal. Id Doyuran / Engineering Geology 47 (1997) 175 187
magnesium sulfate. Annual Book of ASTM Standards, Vol. 04.02, pp. 39-43. Anon,1972. The preparation of maps and plans in terms of engineering geology. Q. J. Eng. Geol. 5,293-382. Anon,1975. Testing porous building stone: Stone Handbook. Architects J. 13, 337-339. Anon,1977. The description of rock masses for engineering purposes. Q. J. Eng. Geol. 10, 355--389. Anon,1979. Classification of rocks and soils for engineering geological mapping. Part 1. Rock and soil materials. Bull. Int. Assoc. Eng. Geol. 19, 364- 371. BSI, 1975. Testing aggregates: Part 2. Methods for determining the physical properties. British Standards Institution (BS 812), 20 pp. BSI, 198 l. Code of Practice for Site Investigations. British Standards Institution (BS 5930), 147 pp. Beawis, F.C., Roberts, F.I., Minskaya, L., 1982. Engineering aspects of weathering of low grade metapelites in an arid climatic zone. Q. J. Eng. Geol. 15, 29-45. Bell, F.G., 1980. Engineering Geology and Geotechnics. Butterworth and Co. Ltd., London, 497 pp. Bell, F.G., 1993. Engineering Geology. Blackwetl Scientific Publications, London, 359 pp. Bieniawski, Z.T., 1975. The point load test in geotechnical practice. Eng. Geol. 9, 1 -11. Broch, E., Franklin, J.A., 1972. The point load strength test. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 9, 669 697. BRE, 1983. The selection of natural building stone. Building Research Establishment Digest 290, 8 pp. CIRIA/CUR, 1991. Manual on the use of rock in coastal and shoreline engineering. Construction Industry Research and Information Association, CIRIA Special Publication 83, CUR Report 154, 607 pp. De Witte, E., 1995. Conservation of the Gtreme rock. Proc. Int. Seminar on the Safeguard of the Rock-Hewn Churches of the Gtreme Valley, ICCROM, pp. 109 124. De Witte, E., Terfve, A., Koestler, R.J., Charola, A.E., 1988. Conservation of the G6reme rock: preliminary investigations, Proc. 6th. Int. Cong. on Deterioration and Conservation of Stone, Torun, Poland, pp. 346 355. Erdo~an, M., 1986. Nev~ehir-Orgtip y~resi ttlflerinin malzeme jeolojisi aqismdan ara~tlnlmasl, i.T.U. Maden Fak.. Doktora Tezi, 100 pp. Ergnvauh, A.K., Ytizer, A.E., 1977. Past and present use of underground openings excavated in volcanic tufts at Cappadocia area. Proc. on Rock Storage, Oslo, Norway, pp. 15-17. Erguvanh, K., Yorulmaz, M., ~lli, F., Ahunbay, Z., Erdo~an M., 1989. GOreme yaplsal koruma ve sa~lamla~tirma projesi, E1 Nazar kilisesi, i.T.O., 46 pp. Fookes, P.G., Gourley, C.S., Ohikere, C., 1988. Rock weathering in engineering time. Q. J. Eng. Geol. 21.33 57. Foster, I.R., 1983. The influence of core sample geometry on the axial point load test.. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 20, 291 295. Goodman, R.E., 1989. Introduction to Rock Mechanics, 2nd ed. John Wiley and Sons, New York, 562 pp.
Hencher, S.R., 1993. Conference summary. The Engineering Geology of Weak Rock. Balkema, 504 pp. Hirschwald, J., 1912. Handbuch der bautechnischen Genteinsprtifung. Borntraeger, Berlin. Hosking, J.R., Tubey, W., 1969. Research on low-grade and unsound aggregates. Road Research Laboratory Report, LR 293, Crowthorne (UK), 30 pp. ISRM, 1981. Rock Characterization, Testing and Monitoring. Pergamon Press, Oxford, 211 pp. ISRM,1985. Suggested method for determining point load strength. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 22, 51 60. Larsen, T.D., Cady, P.D., 1969. Identification of frost susceptible particles in concrete aggregates, National Cooperative Research Program, Report 66, Washington, DC, Highway Research Board, 62 pp. METU, 1987. Investigation of the mechanisms of stone deterioration for the purpose of conservation of GOreme tufts, Progress report, 61 pp. Norbury, D.R., 1986. The point load test. Site investigation practice: assessing BS 5930. Geol. Soc. Eng. Geol. Special Publ. 2, 423 pp. Pasquare, G., 1968. Geology of the Cenozoic volcanic area of Central Anatolia, Memoire. Serie. 8, Vol. 9, Fasc. 3, Roma, pp. 57-201. RILEM,1980. Recommended tests to measure the deterioration of stone and to assess the effectiveness of treatment methods. Commission 25-PEM. Mater. Struet. 13, 175-253. Rockware, 1993. STEREO 3-D orientation analysis and plotting. Version 3, Rockware Scientific Software, USA, 56 pp. Rodrigues, J.D., Jeremias, F.T., 1990. Assesment of rock durability through index properties. Proc. 6th. Int. Congress IAEG, Amsterdam, Balkema-Rotterdam, Vol. 4, pp. 3055 3060. Rodrigues, J.D., 1993. Stone conservation in Portugal: problems, methods and means. Int. Workshop on Conservation of Architectural Surfaces: Stones and Wall Covering, Venice, pp. 147 154. Schaffer, R.J., 1972. The weathering of natural building stones. Dept. Sci. Ind. Res. Building Research Spec. Rept. No. 18, 149 pp. Sims, I., 1991. Quality and durability of stone for construction. Q. J. Eng. Geol. 24, 67-73. Topal, T., 1995. Formation and deterioration of fairy chimneys of the Kavak tuff in Orgiip-Gtreme area (Nev~ehir-Turkey). Ph.D. Thesis, Middle East Technical University, Ankara, Turkey, 250 pp. Topal, T., Doyuran, V., 1994. Effect of deterioration on the chemical and physical properties of G6reme tufts. Proc. 7th. Int. Cong. of IAEG, Lisboa, Vol. 5, pp. 3527-3536. Topal, T., Doyuran, V., 1995. Effect of discontinuities on the development of fairy chimneys in the Cappadocia region (Central Anatolia-Turkey). Turkish J. Earth Sci. (TISIBITAK) 4, 49-54. Topal, T., Doyuran, V., 1996, Ultrasonic testing of artificially weathered Cappadocian tuff, LCP-95 Cong. on Preservation
T. Topal, V. Doyuran / Engineering Geology 47 (1997) 175-187 and Restoration of Cultural Heritage, Montreux, pp. 205-211. TSE, 1977. Dotal yap~ ta~lan, Tilrk Standartlarl EnstitOsO, TS2513, 6 pp. TSE, 1978. Dotal yap1 ta~lanmn muayene ve deney metodlan. Tiirk Standartlarl EnstitOsO, TS699, 16 pp.
187
Winkler, E.M., 1986. A durability index for stone. Bull. Assoc. Eng. Geol. 23, 344-347. Winkler, E.M., 1993. Discussion and reply on "The durability of sandstone as a building stone, especially in urban environments". Bull. Assoc. Eng. Geol. 30, 99-101.