RELATIONSHIP BETWEEN MICROMORPHOLOGICAL CHARACTERISTICS AND ENGINEERING PARAMETERS OF CALICHE (CALCRETE) İsmail Dinçer1*, Altay Acar1 & Donatello Magaldï2 1Çukurova University, Engineering and Architecture Faculty, Geological Engineering Department, Balcalı, 01330, Adana, Turkey 2University of L’Aquila, Engineering Faculty, Department of Structures, Water and Soil Engineering, Monteluco di Roio, 67040, L’Aquila, Italy *Corresponding author: İsmail Dinçer - Department of Geology Engineering, Faculty of Engineering-Architecture Çukurova University - 01330 Adana, Turkey - Fax: 90-322-3386126 - Tel: 90-322-3386715/15 - e-mail: idincer@mail.cu.edu.tr ABSTRACT: İ. Dinçer, A. Acar & D. Magaldi, Relationship between micromorphological characteristics and engineering parameters of caliche (calcrete). (IT ISSN 0394-3356, 2007). Caliches are described as secondary carbonate formations and calcareous, semiconsolidated aragonite or early diagenetic calcite for- ming in loose materials such as pebble, sand, silt, and soil under semi-arid and arid climatic regimes. The aim of this study is to identify and discuss effects of micro-morphological-mineralogical properties on engineering properties of caliche. Furthermore, correlations derived from mechanical, physical-index properties and micromorphological characteristics are evaluated and establish statistically predictable model for UCS. For this aim, samples were taken from different locations in cali- che outcrops occurring in the southern part of Turkey. Based on the statistical analyses, linear relationships were fitted between the Micromorphological index (MI) and each mechanical and physical-index properties. Based on these linear relationships among the engineering properties, uniaxial compressive strength, P- wave velocity, average Young’s modulus and point load index there are significant correlations with the micromorphological index (MI). Results suggested that strength of the caliche is controlled by micromorphological properties. Four micromorphological parame- ters which are effective on engineering properties of caliche were identified: microstructure, voids, coarse materials and matrix. Microstructure is the most important parameter to assess engineering behaviour of caliche followed by voids, coarse materials and matrix. Statistically most reliable models, which allow the estimating of UCS, are discussed in this study. The empirical equation, UCS = - 7.137 + 3.06x10-3 Vp + 0.38UW + 0.13MI where Vp = P-wave velocity; and UW = Unit weight is assumed to be the best and practical for engineering purposes. RIASSUNTO: İ. Dinçer, A. Acar & D. Magaldi, Relazioni tra caratteristiche micromorfologiche e proprietà geotecniche della crosta cal- carea (calcrete ). (IT ISSN 0394-3356, 2007). La costa calcarea (caliche o calcrete in inglese) è un accumulo carbonatico indurito semiindurito costituito da calcite e/o aragonite che si forma entro substrati geologici sciolti come le ghiaie, le sabbie e i limi oppure entro suoli di clima in genere arido o semiarido. Questa ricerca ha avuto come scopo l’identificazione e l’analisi delle relazioni esistenti tra le caratteristiche micromorfologiche e mine- ralogiche e le proprietà geotecniche della crosta calcarea affiorante nella Turchia meridionale.Al riguardo sono stati considerati modelli statistici di previsione della resistenza a compressione uniassiale (UCS) a partire dalle caratteristiche micromorfologiche.Sono state inoltre trovate relazioni significative non soltanto per l’UCS ma anche per altri parametri geo-ingegneristici quali la velocità delle onde P entro i provini, il modulo di Young e il valore risultante dalla prova di carico puntuale (Point Load Test). Questi risultati suggeriscono che le caratteristiche micromorfologiche e mineralogiche così come evidenziate dall’analisi in sezione sottile ed espresse da un Indice micromorfologico (MI) , condizionano le proprietà meccaniche della crosta. Le caratteristiche micromorfologiche che maggiormente appaiono influenzate da quelle micromorfologiche sono, nell’ordine, la microstruttura (micromass, secondo Stoops, 2003), la tipologia dei vuoti, la tipologia dei costituenti grossolani, la matrice. L’equazione empirica che appare maggiormente valida per stimare la resistenza a compressione uniassiale, principale indice delle pro- prietà meccaniche di una roccia,è la seguente: UCS = - 7.137 + 3.06x10-3 Vp + 0.38UW + 0.13MI, dove Vp = velocità delle onde P, UW= peso dell’unità di volume del materiale considerato e MI è il nuovo indice micromorfologico. Keywords: caliche, micromorphological characteristics, mechanical properties, physical-index properties. Parole chiave: caliche, caratteristiche micromorfologiche, proprietà meccaniche, proprietà geotecniche. Il Quaternario Italian Journal of Quaternary Sciences 20(2), 2007 - 137-150 1. INTRODUCTION AND BACKGROUND It is commonly accepted that engineering proper- ties of the rocks are controlled by their petrographic characteristics. ZORLU et al. (2004) concluded that packing density and packing proximity are most influencing in mechanical characteristics for sandstone. Also they reported that texture is more important than mineralogy in this respect as far as sandstones were concerned. TUGRUL and ZARIF (1999) correlated the mechanical, physical-index properties with petrographi- cal characteristics of granitic rocks. In the same way, they concluded that the influence of textural characteri- stics on the engineering properties appears to be more important than mineralogy. They also determined the types of contacts, grain (mineral) shape and size signifi- cantly influence the engineering properties of granitic rocks. KONCAGÜL and SANTI (1999) distinguished the fac- tors affecting hardness, strength and durability in rocks in three different groups such as effective in grained rocks, argillaceous rocks and effective in both rock types. They concluded that packing density, degree of bonding, type of cement and bonding material, per- meability, post-depositional factors, occurrence of 138 İ. Dinçer, A. Acar & D. Magaldi soft/soluble minerals and microfractures affect both uniaxial compressive strength (UCS) and the slake durability test (SDT) in a similar fashion. These factors cause both UCS and SDT results to increase or to decrease. JENG et al. (2004) found that porosity has more influence on the UCS than grain and matrix con- tent does. At the same time, few researchers establi- shed a statistical model to predict mechanical proper- ties of rocks using their petrographic characteristics (BELL, 1978; FAHY and GUCCIONE, 1979; SHAKOOR and BONELLI, 1991; ULUSAY et.al., 1994). Caliche adversely affects cut slope stability, earth fill material sensitivity, and ground stability at structure sites along the Motorway which is being constructed in the Southern 3 Turkey. About 90 % of active and potential slides in the region occur in the caliche depo- sits and colluvial deposits which have been derived mainly from caliche (YıLMAZER and Smith, 1992). At the same time caliche widely formed foundations of buil- dings in the southern part of Turkey and it was also used in the historical buildings as a dimension stone due to the ability to easily shape it. Caliches are the- refore very problematic materials for Turkey. In this study, the relationship between micro- morphological-mineralogical properties and rock index properties were examined for caliches. Correlations derived from mechanical, physical-index properties and micromorphological characteristics were evaluated to establish a statistical prediction of UCS. 2. MATERIALS 2.1 The caliche Thirteen samples were taken from different loca- tions of caliche outcrops in the southern part of Turkey (Fig 1). Caliche contains different levels for strength and deformability properties. However, only two levels were considered from a geotechnical point of view. The hardpan of the part shows rock properties whereas the soft pan at part below shows soil properties. Block samples (approx. 0.25x0.25x0.20 m) collected from hardpan (massive caliche) has low strength level. The caliche formation is widely distributed in the region on peneplain morphology occurring from east to west in the Adana Region. Caliches are described as a secondary calcium carbonate accumulation forming in loose materials such as gravel, sand, silt, and soil under semi-arid and arid climatic regimes. The caliche term was firstly used for the gravel and similar calcium carbonate cemented materials in south-western United States. Some condi- tions which are required for the formation of caliche are country or basement rocks to be of a carbonate type, the occurrence of carbonate 4 in the soils, a semi-arid to arid climate and a widespread water percolation down the soil profile. In Turkey, caliche occurrences are commonly observed in Aegean, Mediterranean, and Central Anatolia in young rock formations (ATABEY, et Figure 1 - Location map of the study area. Localizzazione della zona studiata. al., 1998). In southern of Turkey, formation of massive caliche was attributed to Pleistocene climatic fluctua- tions and a shallow lacustrine environment by KAPUR et al., (1993). The process is broadly similar to other sites around the Mediterranean. 2.2 Stratigraphic features of study area Different Tertiary aged lithostratigraphic units are outcropped in the study area. The Tertiary stratigraphy was defined as Pre-transgressive, transgressive and regressive deposits in succession (KAPUR et al., 1993). At the base of this succession, Paleozoic and Mesozoic aged lithostratigraphic units are outcropped (Fig 2). Cenozoic succession is mainly represented by Tertiary and Quaternary units. Tertiary units unconformably overlap Paleozoic and Mesozoic rocks (YETIS,, 1988). The transgressive deposits comprise the terrestrial and lacustrine formations (Oligocene-Lower Miocene). The transgressive cycle of the Miocene Sea (Burdigaliyen- Serravallian) consists of shallow water beach clastics, the reef carbonates, the deep marine and turbiditic for- mations. The uppermost of the Tertiary unit comprises the Handere formation, which is consist of clay stone, sand stone and very rarely gravel stone. It was covered by pedostratigraphic units of the Quaternary, the Villafrankian to Late Pleistocene High and Low terraces, in the study area (Ş ENOL et al, 1991) (Fig 3). This model is supported by DINÇ et al ., (1991), for the 139Relationship between micromorphological ... Mediterranean coastal areas with pedo- geomorpholo- gical criteria. It consists of young deposits of study area, which are calichified glacis and glacis or river type conglomerates. Red Mediterranean and Reddish Brown to Brown soils and massive caliches (calcretes) occur with discordance in the Pliocene and between the Mid and Late Pleistocene periods (KAPUR et al., 1993). 3. METHODS 3.1 Rock Mechanics Test Blocks samples (Fig 4 a and b) were taken from 13 different locations in the Adana Region. Using a core-drilling machine at the laboratory, NX and BX size core samples were prepared from caliche blocks (Fig 4 c and d). Unit weight (UW), apparent porosity (n), Schmidt rebound number (RN), shore hardness (SH), P- wave velocity (Vp), slake durability index (Id2), point load index (Is(50)), uniaxial compressive strength (UCS), Average Young’s Modulus (Eav) values of the samples were all measured according to International Society for Rock Mechanics Standards (ISRM, 1981a). Rock samples, having regular shapes, were used for unit weight and apparent porosity tests. Samples were sunk into water for 48 hours. Later, weight and dimensions of samples were measured with a scale having ±0.01 accuracy. After that, caliche samples were Figure 2 - Generalised lithostratigraphic sections of the study area (YETIŞ, 1988). Sezioni lito-stratigrafiche della zona studiata (YETIS, 1988). 140 dried for 24 hours at 105 °C and weighted. Determination of apparent porosity and unit weight was achieved using volume, dry weight and saturated wei- ght of the samples. The Schmidt Hammer tests were carried out in the field on large caliche blocks samples. All tests were performed by N type Schmidt Hammer following ISRM (1981b). The Schmidt Hammer tests were performed vertically on rock blocks with no discontinuity and cracks. Each time, twenty readings were taken and the upper ten values were averaged. Shore hardness values were determined using the C2 type shore schleroscope which was cleaned and cali- brated using calibration block. Test procedures inclu- ded 2.44 gr diamond tipped hammer drop freely on the caliche sample and carefully measuring and longing the rebounding height on the scale which ranged 0 to 140. Twenty readings were taken from each sample; and readings with the highest five and lowest five values were cancelled, and the remaining 10 readings were averaged according to ISRM (1978a). The pundit and two transducers (a transmitter and a receiver) were used in all tests. Three core samples were prepared (height/diameter 2-2.5) for each caliche samples. Both faces of drill cores were trimmed and smoothed so that the receiver and the transmitter could cover the faces tightly. The test was carried out according to ISRM (1978b). In the slake durability test, the apparatus com- bines the effects of both soaking and abrasion in order to evaluate the rate of weathering caused by water immersion. The test procedure was performed accor- ding to ISRM (1979) standards with ELE RM-310-2 test machine. Ten rock slumps with equal dimensions were prepared for the slake durability test. Rock slumps rota- ted in the steel mesh drum were partially immersed in water for ten minutes. Samples were subjected to three cycles and durability index (Id2) in which each cycle was calculated as a percentage ratio of final to initial dry weight of rocks in the drum after drying and wetting Figure 3 - Pedostratigraphic sections of the study area (ŞENOL et al., 1991). Sezioni pedo-stratigrafiche (ŞENOL et al., 1991). İ. Dinçer, A. Acar & D. Magaldi 141 cycles. In this study, the second cycle index (Id2) was used for evaluation of slake durability index of caliches. The point load test is an index test to determine the strength of intact rocks. The system of this test consi- sts of a hydraulic pump, a hydraulic jack, a pressure gauge and two steel points with standard dimensions. In the test, the rock sample is slowly loaded by activa- ting the hand pump until failure of sample. The failure load is read from gauge. The test can be applied to irre- gular and regular rock samples in field or laboratory. There are three test methods, which are diametrical test, axial test and block test (ISRM, 1985). In this study diametrical and axial tests were used in laboratory. Caliche core samples were prepared with length/dia- meter ratio for 1/1. Failure of rock samples was achie- ved within 10-60 seconds. The point load strength index Is(50) was calculated using the correction factor. The uniaxial compressive strength tests were performed according to the ISRM (1981c) standards. The test apparatus for the unconfined compressive tests consisted of a triaxial test machine (ELE ADR 2000) and a data acquisition system. NX and BX (length/diameter 2-2.5) size core samples were prepa- red for UCS. The loading rate was selected as 0.1 kN/s and failure of caliche samples was achieved within 5 – 10 minutes. At least five specimens were tested for each block samples. At the same time, ISRM (1981c) standard test procedure was used to determine the average Young’s modulus (Eav). 3.2 Micromorphological Analyses Thin sections (25 x 40 mm) of the samples were prepared for the microscopic observation. Some com- mon micro morphological features (microstructure, voids, coarse materials and matrix) were selected to evaluate micromorphological characteristics of caliche according the results and assumptions after several papers (MAGALDI, 1983; BULLOCK et al., 1985; STOOPS 1988; MAGALDI and TALLINI, 2000). 3.2.1 Microstructures Same microstructures classification was used after MAGALDI, 1983 (Fig 5): Undifferentiated; size, shape and arrangement of matrix have a homogeneous crust. It contains few if any voids, rock fragments and organic materials; Banded; laminated crust with different crystal size in individual laminae. Matrix and aggregates generally horizontally elongated and separated by planar voids. Nodular; groundmass consisting of many nodules. Brecciated; crust formed by irregular network of fractures and clasts. 3.2.2 Matrix The matrix of caliche consists mainly of calcite which is the cementing medium. The calcite material is composed of more or less similar crystalline or crypto- crystalline particles. The calcite crystals are less than 2 µm to 15 µm in size. Three types of calcite cement Relationship between micromorphological ... Figure 4 - Various views of the caliche outcrops and samples. (a) hardpan level of caliche, (b) morphological structures of caliches, (c) cored caliche blocks, (d) NX and BX size caliche core samples. Immagini degli affioramenti e dei campioni di crosta calcarea: (a) livello di hardpan entro la crosta calcarea, (b) microstruttura della cro- sta calcarea, (c) microstruttura di crosta calcarea con il foro di carotaggio, (d) campioni di crosta calcarea con dimensioni NX e BX. 142 occur in the matrix micrite, microsparite and sparite. Micrite (2 to 4 µm) occurs as cementing material between the skeletal grains and as a clear rim around the grains within the groundmass. Microsparite (4 to 10 µm) is formed by subhedral crystals. Microsparite occurs as drusy mosaic filling between the micritic rim- med grains and rock fragments. It is also observed in the following cases: a) hypo coatings around the voids and cavities; b) calcite crystals increasing away from the walls of the grain / channels or cavities; c) elonga- ted crystals perpendicular to the wall of the cavity/ channel or grain boundary. Sparite (10 µm and above) generally occurs as infilling crystals within channels and voids. In some caliche samples, two or more different matrix types were observed at the same time. This matrix type was identified as mixed type. 3.2.3 Voids Stoop’s classification (1988) is used for classifica- tion of voids and spaces not occupied by solid material. BREWER (1964) classified voids according to their morphology, considering them as individuals. His clas- sification was used by BULLOCK et al. (1985) with some small modifications. 3.2.4 Coarse Materials Size, shape, frequency and internal characteristics are used as main criteria for description of the coarse materials. Three main groups of coarse particles can be determined according to their composition and origin; (1) mineral grains, (2) rock fragments, (3) some coarse pedo features and components (nodule, organic matter, etc). 4. MICROMORPHOLOGICAL INDEX AND CORRELATION WITH ROCK- INDEX PROPERTIES Rock-Index properties of each sample were cor- related with their selected micromorphological features. Microstructure, voids, coarse materials and matrix were considered and each feature was quantified by a simple rating qualitatively related to degree of presumable strength of the sample. Voids and coarse materials are rated according to both modal size and frequency. We used a rating 2 to 6 for one feature; 1 to 5 for four fea- tures; 1 to 3 for one feature (Table 1). Overall rating ran- ges from 7 to 29 increasing with the strength of mate- rial. This rating (MI) was obtained for each caliche sam- ple. Table 2 shows the micromorphological (semi-quan- titative data) and engineering properties of caliche sam- ples. KONCAGÜL and SANTI (1999) concluded that index properties of rocks are controlled by mineralogical composition, degree of cementation and re-crystalliza- tion. Well cemented and completely recrystallizated samples have higher strength and durability than other samples. In order to investigate the influence of micro- morphological characteristics on mechanical and physi- cal-index properties of caliche samples, a number of correlations based on simple regression and some graphs were made. At the same time multiple linear regression variable selection analysis can be performed with a forward selection process, a backward elimina- tion process, or a stepwise selection process that is a combination of the forward and the backward methods. The stepwise selection method selects or eliminates the İ. Dinçer, A. Acar & D. Magaldi Figure 5 - Microstructure types of caliche samples. (a: Undifferentiated b: laminated c: Nodular d: Brecciated) Tipo di microstruttura dei campioni di crosta calcarea. (a:Indifferenziata, b:Laminata, c: Nodulare, d: Brecciata). 143Relationship between micromorphological ... variables as per given user probability criteria. A depen- dent variable or response variable is related to predictor or independent variable(s). The objective was to con- struct a regression model relating the dependent varia- ble, y, to independent variable(s). Stepwise forward selection procedure was used to select the best suita- ble regression model of the form to predict the uniaxial compressive strength (UCS) of caliches. y = βo + β1X1 + β2X2, + .... + βk Xk+ En, , (1) βo, β1 .... βk are the regression coefficients, and X1, X2 , Xk, are the selected independent variables. En is an error term representing the magnitude of y not accoun- ted for by other terms in the equation. 5. RESULTS 5.1 Geomechanical Properties Geomechanical properties of caliches are assu- med to be controlled by mineralogical composition, degree of cementation and recrystallization. Well cemented and completely recrystallizated samples have higher strength and durability than other samples. Table 2 shows the laboratory test results of caliche samples. Obtained values of unit weight of the studied caliche blocks vary from 14.96 kN/m3 to 22.00 kN/m3. Apparent porosity (n) is fluctuated between %16 and %35. Schmidt rebound values (RN) of caliches range between 14.50 and 39.80. Shore hardness (SH) of cali- ches varies from 4.20 to 22.50. Highest P- wave velo- city (Vp) was 1444 m/s, while the lowest was calculated as 375 m/s. The Slake durability test results (Id2) of cali- che blocks ranged from 65.36 to 97.53%. Point load index (Is(50)) varies from 0.53 to 1.91 MPa. The unconfi- ned compressive strength (UCS) of the samples ranges between 2.03 and 9.54 MPa. The value of average Young’s modulus (Eav) varies from 0.16 to 1.29 GPa. 5.2 Micromorphology Caliche has the following micromorphological features: undifferentiated, nodular and brecciated microstructures with 10YR 8/2, 5YR 5/6, 5YR 8/6, 10YR 7/4, 10 YR 8/6 and 10YR 5/4 colours; dark calcite matrix with micrite and rarely microsparite groundmass occurring in the voids and nodules; microsparite and sparite rarely occur- ring as drusy mosaic filling in the voids, serrate boundary between micrite and microsparite; very dense and irregularly shaped nodules, dark greyish and white in colour with microcrystalline ground mass. Edges of chambers, sim- ple packing voids, vughs, channels and vesicles type voids are rough-undula- ting, completely unaccommodated. Voids frequency increases up to % 50 in some samples; c/f-related distribu- tion pattern is close-single-open spa- ced porphyric; coarse materials consist of scattered angular grains of quartz, silica, tephra and limestone Their sizes range 5 µm to 1000 µm; large grains (>1000 µm) are rarely observed. Few orthic nodules and trace of plant root remains were observed. Some micro- scopic pictures under polarised light are given in Fig. 6. 5.3 Correlation of Micromorpholo- gical and Engineering Parameters of Caliche Mean values of index properties of microstructures types which are identi- fied for caliche samples, are given in fig 7. According to results undifferentiated microstructure has higher mean values than other types. But this type has similar mean values with brecciated with grain to grain contacts only for slake durability test. Microstructures types are probably the most important character controlling the mechanical behaviour of caliches. In a uniaxial compressive test stres- ses are likely to concentrate along the Sample No : Description : Characteristic Type Rating Sample Microstructure Undifferentiated 6 Banded 4 Nodular 2 Brecciated 2 Not: Brecciated with grain to grain contacts 6 Voids Size(µm) <50 5 50-100 4 100-500 3 500-1000 2 >1000 1 Frequency (%) <5 5 5-15 4 15-30 3 30-50 2 >50 1 Coarse Materials Size(µm) <50 5 50-100 4 100-500 3 500-1000 2 >1000 1 Frequency (%) <5 5 5-15 4 15-30 3 30-50 2 >50 1 Matrix Sparite 3 Mixed Type 3 Microsparite 2 Micrite 1 Total rating Table 1 - Micromorphological rating form. Punteggio attribuito alle caratteristiche micromorfologiche. 144 İ. Dinçer, A. Acar & D. Magaldi E ng in ee ri ng P ar am et er s M ic ro m o rp ho lo g ic al P ar am et er s S am p le V o id s C o ar se M at er ia ls N o g (R N ) (S H ) V p n I d 2 Is (5 0) U C S E av M ic ro st ru ct ur e M at ri x ( M I) (k N /m 3 ) (m /s n) (% ) (% ) (M P a) (M P a) (G P a) S iz e Fr eq ue nc y S iz e Fr eq ue nc y (µ m ) (% ) (µ m ) (% ) B L- 1 18 .5 1 28 .0 0 13 .5 4 68 1 28 .1 2 90 .0 8 1. 13 5. 63 0. 55 U nd iff er en tia te d 10 0- 50 0 < 5 50 -1 00 5- 15 M ix ed 25 Ty p e B L- 2 17 .9 3 24 .6 0 8. 58 49 1 28 .4 4 94 .2 9 0. 97 3. 37 0. 19 N od ul ar 50 0- 10 00 15 -3 0 < 50 < 5 M .S p ar ite 19 B L- 3 21 .5 5 27 .8 0 12 .1 3 11 46 21 .0 0 97 .5 3 1. 25 7. 85 0. 88 U nd iff er en tia te d < 50 15 -3 0 < 50 < 5 S p ar ite 27 B L- 4 14 .9 6 14 .5 0 4. 20 37 5 34 .7 9 65 .3 6 0. 53 2. 03 0. 16 N od ul ar 10 0- 50 0 30 -5 0 10 0- 50 0 < 5 M ic rit e 16 B L- 5 19 .1 7 31 .4 0 12 .0 4 50 9 23 .9 8 92 .7 6 0. 96 4. 52 0. 52 B re cc ia te d 50 0- 10 00 15 -3 0 50 -1 00 5- 15 M ic rit e 16 B L- 6 18 .7 0 35 .4 0 18 .0 8 10 85 26 .3 6 96 .9 2 1. 69 6. 28 0. 68 B re cc ia te d 50 -1 00 < 5 50 -1 00 15 -3 0 M .S p ar ite 24 (g ra in t o gr ai n co nt ac ts ) B L- 7 17 .3 2 19 .9 0 8. 40 43 6 31 .7 0 84 .8 2 0. 78 2. 99 0. 19 B re cc ia te d 10 0- 50 0 5- 15 50 0- 10 00 30 -5 0 M ic rit e 14 B L- 9 20 .3 3 26 .9 0 11 .0 0 75 8 24 .1 5 90 .5 2 1. 03 6. 50 0. 82 U nd iff er en tia te d 50 0- 10 00 < 5 < 50 < 5 M ix ed 26 Ty p e B L- 11 19 .8 4 26 .0 0 11 .8 0 72 7 23 .7 5 93 .5 2 0. 97 4. 90 0. 63 B re cc ia te d 10 0- 50 0 5- 15 50 0- 10 00 15 -3 0 M ic rit e 19 (g ra in t o gr ai n co nt ac ts ) B L- 12 15 .3 7 22 .6 0 10 .5 2 43 1 35 .4 9 92 .3 6 0. 95 2. 45 0. 19 N od ul ar 50 -1 00 30 -5 0 < 50 < 5 M ic rit e 18 B L- 14 22 .0 0 39 .8 0 22 .5 0 14 44 16 .2 3 96 .8 1 1. 91 9. 54 1. 29 U nd iff er en tia te d 10 0- 50 0 5- 15 < 50 < 5 M ix ed 26 Ty p e B L- 14 A 18 .8 7 22 .0 0 9. 60 71 3 26 .7 3 90 .1 5 0. 79 4. 32 0. 59 B re cc ia te d 10 0- 50 0 15 -3 0 > 10 00 30 -5 0 M ic rit e 16 (g ra in t o gr ai n co nt ac ts ) B L- 15 18 .2 6 27 .5 0 11 .4 6 47 7 26 .2 3 91 .9 0 0. 86 2. 65 0. 18 N od ul ar 10 0- 50 0 15 -3 0 10 0- 50 0 15 -3 0 M ic rit e 15 γ: U ni t w ei gh t, V P : P -w av e V el oc ity , R N : S ch m id t R eb ou nd N um b er , S H : S ho re H ar d ne ss , n: a p p ar en t p or os ity , I d 2 : S la ke d ur ab ili ty in d ex ( tw o cy cl e) , Is (5 0) : P oi nt L oa d I nd ex ( M P a) , U C S : U ni ax ia l C om p re ss iv e S tr en gt h, E a v : A ve ra ge Y ou ng 's M od ul us , M I: M ic ro m or p ho lo gi ca l i nd ex . Ta b le 2 - E ng in ee rin g an d M ic ro m or p ho lo gi ca l p ro p er tie s of s am p le s. P ro p rie tà t ec ni ch e e ca ra tt er is tic he m ic ro m or fo lo gi ch e d ei c am p io ni . edges of cracks, voids and coarse materials. A rock with these features easily fails at low stress rate. Therefore, while undifferentiated microstructures can withstand higher uniaxial compressive loads, breccia- ted type is weaker than other types. Nevertheless some brecciated samples showed high strength because the number of grain to grain contacts is probably higher, so the applied external force is distributed over larger con- tact surface. Mineralogy of bonding or cementing mate- rial is an important property that controls strength, hardness and durability (KONCAGÜL and SANTI, 1999). Quartz provides the strongest binding followed by cal- cite and iron minerals. Clay binding material is the weakest. Caliche consists of micrite, microsparite and spa- rite calcite matrix. Mean values of index properties are 145Relationship between micromorphological ... Figure 6 - Thin section views of samples. Sezioni sottili dei campioni. 146 given for matrix type of caliches in fig 5. Micrite values are normally lower than values of other types. Sparite matrix can withstand higher uniaxial compressive loads than micrite. Probably because micrite matrix occurs faster than sparite and micrite couldn’t be cemented sufficiently. Porosity has a significant effect on mechanical performance. PRICE (1960) and DUBE and SINGH (1972) reported that in sedimentary rocks all strength proper- ties decrease with increasing porosity. This happens because high porosity allows the stress-induced micro- fractures to propagate (HOWARTH and ROWLANDS, 1986). The low uniaxial compressive strength in caliche sam- ples is related to high frequency of voids. Correlation matrices summarizing the determina- tion coefficients of the index properties with significan- ce levels are given in Table 3. The relationships between UCS vs UW, UCS vs Vp, and UCS vs MI yiel- ded statistically significant correlations. Based on stati- stical analyses, linear relationships were fitted between the MI and some index property. Uniaxial compressive strength, P-wave velocity, average Young’s modulus and point load index have significant correlations with micromorphological index (MI). Instead slake durability index, Schmidt hardness, shore hardness and unit wei- ght have very weak correlation with MI. These findings suggested that micromorphological characteristics are controlling mechanical properties (as UCS) more than unit weight (UW) and durability (Id2). In order to determine how much the caliche strength is controlled by micromorphological properties MI was compared to Uniaxial Compressive Strength (UCS) which is the most widely used design parameter in rock engineering The correlations of thirteen samples is statistically significant with R2 = 0.73 (Fig 8). According to Fig 8, UCS increases with the increasing of MI. 5.4 Prediction of the uniaxial compressive strengt h of caliche UCS is the most important index property in rock mechanics. Testing procedure has been standardized by both The American Society for Testing and Materials Figure 7 - Mean values of rock-index properties of microstructure and matrix types (UD: undifferentiated, N: nodular, B: brecciated, BG: brecciated with grain to grain contacts, MC: microsparite, SP: sparite and MT: mixed type). Valori medi delle proprietà indici delle varie tipologie di microstruttura e matrice (UD: indifferenziate, N: nodulari, B: brecciate, BG: brecciate con contatti granulo-granulo, MC: microsparite, SP: sparite e MT: tipo misto). İ. Dinçer, A. Acar & D. Magaldi 147 Un-standardised coefficients Independent Variables β Std. error t Sig. Constant -5.742 2.131 -2.694 0.023 Vp 0.0045 0.001 5.171 0.000 UW 0.394 0.139 2.834 0.018 Constant -7.137 1.559 -4.578 0.001 Vp 0.003 0.001 4.025 0.003 UW 0.383 0.098 3.906 0.004 MI 0.133 0.040 3.336 0.009 Table 4 - Statistics of some parameters used to estimate UCS. Valori statistici di alcuni parametri usati per valutare la resisten- za a compressione uniassiale(UCS). Table 5 - Correlation and determination coefficients of the equations used to estimate UCS. Coefficienti di correlazione e di determinazione delle equazioni usate per valutare l’UCS. R R2 Adjusted R2 Standard error of the estimate 0.971a 0.943 0.932 0.589 0.987b 0.975 0.966 0.415 a. Predictors: (Constant), Vp, UW b. Predictors: (Constant), Vp, UW, MI Relationship between micromorphological ... Table 3 - Correlation matrix of engineering and micromorphological index with significance level. Correlazione di matrice dell’indice tecnico e micromorfologico con livello di significatività. UW RN SH Vp n Id2 Is(50) UCS Eav UW RN ,724(**) SH ,667(*) ,952(**) Vp ,805(**) ,771(**) ,839(**) n -,971(**) -,790(**) -,722(**) -,806(**) Id2 ,669(*) ,747(**) ,680(*) ,553(*) -,630(*) Is(50) ,648(*) ,910(**) ,956(**) ,894(**) -,682(*) ,655(*) UCS ,889(**) ,779(**) ,800(**) ,947(**) -,863(**) ,567(*) ,841(**) Eav ,882(**) ,742(**) ,777(**) ,927(**) -,873(**) ,506 ,782(**) 0.97(**) MI ,641(*) ,597(*) ,618(*) ,781(**) -,569(*) ,451 ,734(**) ,855(**) ,767(**) ** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). (ASTM) and The International Society for Rock Mechanics (ISRM). The tests are performed on well pre- pared core samples with a time consuming and expen- sive procedure. Preparing well core samples is rather difficult in soft rock like marl, caliches and others so many researchers employed simple models to estimate UCS by the rock index tests. Then a multiple linear regression analyses were considered to obtain equations for estimating UCS from several index properties and MI. The data set is given in Table 2. Regression analyses were performed in two stage using SPSS for Windows statistical software. In the fist stage, only the quoted index pro- perties (Table 4 and Table 5) were used as an indepen- dent variable to predict UCS: UCS = - 5,742 + 4.5x10-3 Vp + 0.39UW (2) In the second stage, all index properties and micromorphological index (MI) were used as indepen- dent variables and the stepwise forward technique for regression analyses are carried out to predict UCS. The best model for UCS was: UCS = - 7.137+3.06x10-3 Vp+ 0.38UW+0.13MI (3) The adjusted R2 values indicate that the above multiple regression models explain 93.2% in equation (2) and 96.6 % in equation (3) of the total variation, respectively (Table 5). Fig. 9 and Fig. 10 support the assumption of a linear relationship between the dependent variable UCS and the independent variables P- wave velocity (Vp), Unit Weight (UW) and micromorphological Index (MI). Regression equations have slopes equal to 1.0, high correlation coefficient and very low intercept values. The residual analyses confirmed that the assumption of a linear relationship is valid. 6. Conclusions The caliche occurs from east to west in the southern part of Turkey. Caliche adversely affects cut slope stability, earth fill material sensitivity and ground stability at structure sites along the Motorway which is being constructed in the Southern Turkey. Caliches are very problematic materials for study. Four micromorphological parameters which are significant on engineering behaviour of caliche were identified as follows: microstructure, voids, coarse 148 Figure 9 - Comparison of measured and the estimated UCS values for the model exclu- ded MI. Confronto tra i valori UCS misurati e quelli stimati con la relazione empirica, escludendo il parametro MI. Figure 10 - Comparison of measured and the estimated UCS values for the model with MI. Confronto tra i valori UCS misurati e quelli stimati con la relazione empirica, tenendo conto del parametro MI. İ. Dinçer, A. Acar & D. Magaldi Figure 8 - Micromorphological index (MI) ver- sus Uniaxial Compressive Strength (UCS) for thirteen samples. Indice micromorfologico vs. Resistenza alla compressione uniassiale (UCS) per i tredici campioni di crosta calcarea. 149 material and matrix. Microstructure is the most impor- tant parameter on engineering behaviour of caliche fol- lowed by voids, coarse materials and matrix. Micromorphological characteristics are more effective on mechanical properties than index properties and durability. Results were obtained from a limited number of samples but the employed methodology seems to be promising for reaching indicative values of some rock index properties. Based on these linear relationships among the engineering properties, uniaxial compressive strength, P-wave velocity, average Young’s modulus and point load index have significant correlations with a proposed micromorphological index (MI). The results suggested that strength of caliche is also controlled by micromorphological properties. Then the statistically most reliable model as UCS = - 7.137 + 3.06x10-3 Vp + 0.38UW + 0.13MI could be an accurate and less expensive estimation of UCS for engineering purposes based on micromorphogical as well geomechanic properties of the material. Lacking geomechanical properties UCS can be assessed with lesser precision by their simple relationship with MI. REFERENCES ATABEY E., ATABEY N. and KARA H. (1998) - Sedimen- tology of caliche (calcrete) occurrences of the Kırsehir region. Mineral Res. Expl. Bull., 120, 69- 80. BELL F.G., (1978) - The physical and mechanical proper- ties of the Fell Sandstone, Northumberland, England. Engineering Geology 12: pp. 1–29. BREWER R. (1964) - Fabric and mineral analysis of soils. J.Wiley Inc., New York. 470 pp. BULLOCK P., FEDOROFF N., JONGERIUS A., STOOPS G., TURSINA T. (1985) - Handbook for soil thin section descriptions. Waine Research Publications, Wolverhampton, England. DİNÇER U., SENOL M., KAPUR S., SARı M. (1991) - Catenary soil relationships in the Çukurova region, southern Turkey. CATENA 18, 185-196. DUBE AK, SINGH B. E. (1972) - Effect of humidity on ten- sile strength of sandstone. Journal of Mines Metals and Fuel; 20 (1):8-10. ENGIN C. KONCAGÜL, PAUL M. SANTI (1999) - Predicting the unconfined compressive strength of the Breathitt shale using slake durability, Shore hard- nesss and rock structural properties. International Journal of Rock Mechanics and Mining Sciences 36:139-153. FAHY MP, GUCCIONE MJ (1979) - Estimating strength of sandstone using petrographic thin- section data. Bull. Assoc. Eng. Geol.16:467-485. HOWARTH D.F., ROWLANDS J.C. (1986) - Development of an index to quantify rock texture for qualitative assessment of intact rock properties. Geotechnical Testing J. 9 (4), 169–179. INTERNATIONAL SOCIETY for ROCK MECHANICS (1978a) - Commission on standardization laboratory and field results. Suggested methods for determining hardness and abrasiveness of rocks. Int J Rock Mech Min Sci Geomech Abstr;15: 89–97. INTERNATIONAL SOCIETY for ROCK MECHANICS (1978b) - Suggested methods for determining sound velo- city. Int J Rock Mech Min Sci Geomech Abstr;15:53–8. INTERNATIONAL SOCIETY for ROCK MECHANICS (1979) - Commission on Standardization of Laboratory and Field tests, suggested methods for determining water content, porosity, density, absorption and related properties and swelling and slake durability index properties. Int. J. Rock Mech. Min.Sci. 16, 148–156. INTERNATIONAL SOCIETY for ROCK MECHANICS (1981a) - Rock Characterization Testing and Monitoring, ISRM Suggested Methods. International Society for Rock Mechanics. 211 pp. INTERNATIONAL SOCIETY for ROCK MECHANICS (1981b) - Suggested methods for determining hardnesss and abrasiveness of rocks, part 3. Commission on standardization of laboratory and field tests, pp 101–102. INTERNATIONAL SOCIETY for ROCK MECHANICS (1981c) - Suggested methods for determining the uniaxial compressive strength and deformability of rock materials. International society for rock mech. Commission on standardization of laboratory and field tests, pp 111–116 INTERNATIONAL SOCIETY for ROCK MECHANICS (1985) Suggested methods for determining point-load strength.Int J Rock Mech Min Sci; 22:53–60. KAPUR S., YAMAN S., GÖKÇEN S.L., YETIŞ C. (1993) - Soil Stratigraphy and Quaternary Caliche in the Misis Area of the Adana Basin, southern Turkey, Catena, 20, 431-445. JENG F.S., WENG M.C., LIN M.L., and HUANG T.H. (2004) - Influence of petrographic parameters on geotech- nical properties of tertiary sandstones from Taiwan. Engineering Geology 73, 71–91. MAGALDI D. (1983) - Calcareous Crust (Caliche) genesis in some mollisols and alfisols from southern Italy: A micromorphological approach. Soil Micromor- phology. V: 2 pp: 623-635. A.B Academic Publishers. MAGALDI D, TALLINI M (2000) - A micromorphological index of soil development for the Quaternary geo- logy research. CATENA. V: 41, pp: 261-276. PRICE NJ. (1960) - The compressive strength of coal measure rocks. Colliery Engineering; 37(437): 283- 92. SHAKOOR A, BONELLI RE (1991) - Relationship between petrographic characteristics, engineering index properties and mechanical properties of selected sandstones. Bull. Assoc. Eng. Geol. 28:55-71. STOOPS G., (1998) - Key to the ISSS “Handbook for Soil Thin Section Description”. Natuurwet. Tijdschr. 78: 193-203. ŞENOL M., KAPUR S., GÖKÇEN S.L. (1991) - Litho-pedolo- gical and sedimentological study of the Late Tertiary-Quaternary of the Adana-Balcalı area. Geosound, Adana 18, 79-91. TUGRUL A., and ZARIF I.H. (1999) - Correlation of minera- logical and textural characteristics with enginee- ring properties of selected granitic rocks from Turkey. Engineering Geology 51, 303–317. ULUSAY R., TURELI K., and IDER M.H., (1994) - Prediction of engineering properties of a selected litharenite sandstone from its petrographic characteristics Relationship between micromorphological ... using correlation and multi-variate statistical tech- niques. Engineering Geology, 37, 135-157 YETIŞ C. (1988) - Reorganisation of the Tertiary strati- graphyin the Adana Basin, Southern Turkey. Newsletter, Stratigraphie. 20 (1), 43-58. YıLMAZER I, SMITH I (1992) - Yumusakken ve sertken seviyelerinden olusan kalisin jeolojik ve jeoteknik özellikleri. Türkiye jeoloji kurultayı Bülteni S7, 145- 152. ZORLU K, Ulusay R, OCAKOGLU F, GOKCEOGLU C, SONMEZ H. (2004) - Predicting intact rock properties of selected sandstones using petrographic thin-sec- tion data. Int. J. Rock Mech. ACKNOWLEDGEMENTS This study was supported by TÜBÏTAK (Project No: 104Y189). The authors would l ike to thank the Çukurova University Scientific Research Project Unit for providing the financial support needed for this project (Project No: MMF-2004BAP-11 and MMF-2004D-19). 150 Ms. ricevuto il 5 luglio 2006 Testo definitivo ricevuto il 26 giugno 2007 Ms. received: July 5, 2006 Final text received: June 26, 2007 İ. Dinçer, A. Acar & D. Magaldi