Comparison of the Quaternary travertine sites in the Denizli extensional basin based on their depositional and geochemical data Academic research paper on "Earth and related environmental sciences"

ELSEVIER

Contents lists available at ScienceDirect

Sedimentary Geology

journal homepage: www.elsevier.com/locate/sedgeo

Comparison of the Quaternary travertine sites in the Denizli extensional ^^

basin based on their depositional and geochemical data^

Mehmet Özkula *, Sändor Kele b, Ali Gökgöz a, Chuan-Chou Shenc, Brian Jones d, Mehmet Oru^ Baykara a, Istvän Forizs b, Tibor Nemeth b, Yu-Wei Changc, Mehmet Cihat Al^ek a

a Pamukkale University, Department of Geological Engineering, TR-20070 Denizli, Turkey

b Hungarian Academy of Sciences, Research Centre for Astronomy and Earth Sciences, Institute for Geological and Geochemical Research, H-1112 Budapest, Budaörsi 45, Hungary c High-precision Mass Spectrometry and Environment Change Laboratory (HISPEC), Department of Geosciences, National Taiwan University, Taipei 10617, Taiwan, ROC d Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton AB T6G 2E3, Canada

ARTICLE INFO

ABSTRACT

Article history:

Received 23 November 2012 Received in revised form 30 May 2013 Accepted 31 May 2013 Available online 10 June 2013

Editor: J. Knight

Keywords: Travertine Quaternary Geochemistry Denizli Basin Western Turkey

In the Denizli Basin (Turkey), located in the western Anatolian extensional province, travertine and tufa deposition has been an ongoing process for at least 600,000 years. Travertine bodies, which are 30 to 75 m thick and each covers areas of 1 to 34 km2, are up to 1 km3 in volume.

Today, spring waters in this area have temperatures of 19 to 57 °C, are of the Ca-Mg-HCO3-SO4 type in the Pamukkale, Kelkaya and Pinarbaji areas and the Ca-Mg-SO4-HCO3 type at ^ukurbag. Thermal waters along the northern margin of the basin are generally hotter than those in the east-southeast and south. The 618O and 8D values of the spring waters indicate a meteoric origin. The average temperatures of the hydrothermal systems in the Denizli Basin appear to have decreased from Pleistocene to Holocene. Travertine, which formed from the hotter water, is more widespread than the tufa that formed in the cooler spring waters. Deposition of the travertine, which formed largely on slopes, in depressions, and along fissure ridges (mostly on northern basin margins), was controlled by the interplay between various intrinsic and extrinsic parameters. The travertines are formed largely of calcite with only minor amounts of aragonite in some of the vertically banded, crystalline crust, raft and pisoid travertines found in some of the northern sites. The aragonitic samples, rich in Sr, are typically found around the spring orifices and along the central axis of the fissure ridges.

The stable isotope values of the travertine found in the northwest and southeast parts of the basin are different. The 813C values of the northern travertine deposits are more positive (3.7 to 11.7%o VPBD) than those found in the south-southeast areas ( — 4 to 5.8% VPDB). In contrast, the travertine and tufa in the southeastern areas have higher 618O values ( —15.2 to —7.8% VPDB) than those of the northern areas ( —16.6 to — 4.8% VPDB). Available evidence indicates that spring activity and associated travertine precipitation in the Denizli Basin were controlled largely by tectonic activity rather than by climatic conditions.

© 2013 The Authors. Published by Elsevier B.V. All rights reserved.

1. Introduction

Tufa and travertine that form from spring waters are found throughout the world in many different depositional, climatic, and tectonic settings (Chafetz and Folk, 1984; Ford and Pedley, 1996; Hancock et al., 1999; Arenas et al., 2000; Andrews, 2006). Importance has been attached to these deposits because their depositional, geochemical, and iso-topic signatures can provide critical records of past palaeoenvironmental, palaeoclimatic, and tectonic conditions (Chafetz and Folk, 1984; Altunel and Hancock, 1993a; Guo and Riding, 1998; Minissale et al., 2002;

☆ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

* Corresponding author. Tel.: +90 258 2963404. E-mail address: mozkul@pau.edu.tr (M. Ozkul).

Andrews, 2006; Jones and Renaut, 2010). Although open to debate, tufa is herein considered as a deposit that typically contains remains of microphytes, macrophytes, invertebrates, and bacteria (Ford and Pedley, 1996). As such, it corresponds to 'meteogene travertine' as defined by Pentecost (2005). Although tufa generally has low carbon-isotopic values that range from —12 to — 4%» PDB (Arenas-Abad et al., 2010; Ozkul et al., 2010), some higher values have been reported (Andrews et al., 1997; Horvatincic et al., 2005). By comparison, travertine is treated as a hydrothermal deposit that is hard, crystalline, and less porous than tufa (Ford and Pedley, 1996; Pedley, 2009) and generally has more positive ô13Cvalues in the range of — 1to +10%» (Pentecost, 2005).

Many studies on travertine have been based on a single locality (e.g., Fouke et al., 2000; Kele et al., 2008, 2011) and little attempt has been made to integrate data from numerous springs into one model. Notable exceptions to this general assessment include the studies by Chafetz and Lawrence (1994), Minissale et al. (2002),

0037-0738/$ - see front matter © 2013 The Authors. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/! 0.1016/j.sedgeo.2013.05.018

and Sant'Anna et al. (2004). Chafetz and Lawrence (1994), for example, integrated information from six calcareous spring systems in U.S.A. and Italy to determine the relationships between the stable iso-topic composition of the calcareous spring deposits (i.e., travertine and tufa) and waters from which they were deposited. Similarly, Minissale et al. (2002) demonstrated that there were consistent differences in the geochemical and isotopic signatures of modern and late Quaternary travertines, thermal spring waters, and gas vents on the east and west sides of the Tiber Valley in central Italy.

Travertine deposits are commonly associated with normal or transtensional faults (e.g., Hancock et al., 1999; Brogi and Capezzuoli, 2009). The depositional architecture and geochemical characteristics of travertine deposits are controlled by the balance between extrinsic factors (e.g., climate, spring water composition) and intrinsic factors (e.g., spring flow patterns, biota). Separation of the extrinsic and intrinsic factors can be achieved by comparing different spring deposits found

in one basin. The Denizli extensional basin in western Anatolia, Turkey (Fig. 1A) is ideal for this purpose because it contains numerous carbonate spring deposits (mostly travertine), particularly along its northern boundary (Altunel and Hancock, 1993a; £akir, 1999; Ozkul et al., 2002; Uysal et al., 2009; Ozkul et al., 2010; Kele et al., 2011). Accordingly, this paper describes and compares the depositional, mineralogical, and geochemical features of six travertine deposits found in the Denizli Basin with the view of establishing the intrinsic and extrinsic factors that controlled their development. Conclusions reached from these comparisons can provide the basis for evaluating similar deposits throughout the world.

2. Geological setting

The Denizli Basin, located in the Western Anatolian Extensional Province of Turkey (Fig. 1A), is a graben that is bounded by normal

Fig. 1. (A) Geological map showing locations of travertine sites in the Denizli extensional basin (based on Sun, 1990 and Ozkul etal., 2002) that was examined in this study. Travertine sites on the map were numbered as: (1) Golemezli, (2) Pamukkale, (2a) ^ukurbag, (2b) Akkoy, (3) Ballik, (4) Kelkaya, (5) Kocaba$, (5a) Gurlelc, (6) Honaz, (6a) Degirmenler-Kayaalti and (6b) Karateke. Locations of the U-Th dated travertines are indicated with stars (see Table 1). A-A' and B-B' indicate locations of cross-sections shown in B and C. (B-C) Schematic cross sections along A-A' and B-B', respectively (see A).

Fig. 1 (continued).

faults along its northern and southern margins (Ko^yigit, 2005; Westaway et al., 2005; Kaymak^i, 2006; Al?i?ek et al., 2007). The basin includes the Pamukkale travertines and fossil counterparts (Altunel and Hancock, 1993a; §im§ek et al., 2000; Ozkul et al., 2002; Kele et al., 2011). Travertine bodies located along the northern margin (Fig. 1A) were deposited at the ends of normal fault segments (e.g., step-over zones; ^akir, 1999).

The Neogene basin fill and old bedrock that underlie the travertine deposits are exposed on the graben shoulders and in the mountainous horst areas (Fig. 1). The bedrock in the main catchment areas, exposed on Honaz (2571 m asl) and Babadag (2308 m asl) in the south and ^okelez (1840 m asl) in the north, is composed of schists and marbles that cover the Menderes Massif (Bozkurt and Oberhansli, 2001; Erdogan and Gungor, 2004) and the allochthonous Mesozoic limestone, dolomite, and gypsum of the Lycian Nappes, which tecton-ically overlie the Menderes Massif (Okay, 1989).

The Neogene fill in the Denizli Basin is formed from alluvial, fluvial and lacustrine deposits. The basin was initiated as a half graben in the late Early Miocene when deposition gradually evolved from alluvial fan settings into fluvial deposits and finally into lacustrine environments

(Al^ek et al., 2007). Deposition continued until the Late Pliocene. By the early Quaternary, there was a change in the regional tectonics and the Neogene Denizli half graben became a full graben as the Pamukkale Fault on the north margin became active (Fig. 1B, C). Quaternary deposits are evident on local fluvial terraces that become progressively younger towards the basin centre, and in the form of large travertine deposits that are widespread along the northern and southern margins of the graben. The Quaternary faults and fissures that are common in carbonate bedrock along the margins of the graben are natural pathways that allow meteoric waters to descend into the subsurface and hydrothermal fluids to come to the surface (Minissale et al., 2002; Dilsiz, 2006). These faults and associated fissures cut through the travertine bodies in some localities (Altunel and Hancock, 1993a,b; Van Noten et al., 2013).

The Denizli Basin is important in terms of its seismic activity as well as its geothermal potential (Aydan et al., 2005; Tan et al., 2008; Utku, 2009). The ancient city of Hierapolis at Pamukkale, for example, was damaged several times by earthquakes (Altunel and Barka, 1996; Piccardi, 2007) with magnitudes up to 6.0 (Hancock et al., 2000) that were triggered by normal faulting and extension of the basin (Kaypak

and Gokkaya, 2012). Most of these earthquakes had focal depths of 5 to 15 km. Local seismicity of the Denizli Basin strongly depends on the deep and shallow geothermal systems in the region (Kaypak and Gokkaya, 2012).

3. Sampling and analytical procedures

The travertine exposed at Golemezli, Pamukkale-Karahayit (including Çukurbag and Akkoy), Ballik, Kelkaya, Kocabaç and Honaz in the Denizli Basin was examined and sampled between 2006 and 2009. These sites are located in different parts of the basin. The quarry faces provided excellent opportunities for establishing the three dimensional depositional architectures of the deposits. Approximately 65 travertine samples were collected from these sites for mineralogi-cal, petrographical, and geochemical analyses (including stable isotopes), and radiometric dating.

In situ measurements of temperature (T), pH, Eh, and electric conductivity (EC) were carried out using a Hach-Lange HQ40d instrument. The free CO2 and alkalinity analyses were performed in the field using titrimetric methods. The ions (SO4, Cl, Ca, Mg, Na, K) and SiO2 were analyzed with ion chromatography (Dionex 1CS-100) and spectrophotometric (Hach DR4000 UV/Vis) methods, respectively. Sr analyses were performed with Optima 2100 DV ICP-OES instrument at the Water Chemistry Laboratory at Pamukkale University, Denizli. Water samples were collected in 100 ml glass bottles for stable isotope analyses of ô18O and 8D. Stable isotope compositions of the waters were determined at the Institute for Geochemical Research of the Hungarian Academy of Sciences (Budapest).

The travertine samples were examined by optical microscope using polished thin sections. Thirteen fracture samples, selected from different sites and various lithotypes, were mounted on stubs and sputter coated with a thin layer of gold before being examined on a JEOLJSM 6490 LV scanning electron microscope (SEM) at the Turkish Petroleum Corporation (TPAO) in Ankara, Turkey. The mineralogical composition of 48 samples was determined by the X-ray powder diffraction (XRD) technique using a Philips PW1730 diffractometer located at the Institute for Geological and Geochemical Research (Budapest, Hungary) with CuKa radiation at 45 kV and 35 mA. For this purpose, randomly oriented powders obtained from fresh unweathered samples were used, and semi-quantitative mineral composition was determined according

to the modified method of Bardossy et al. (1980). For some samples, layers formed from different materials were analyzed separately. If a sample was too small for normal preparation, it was attached directly to a steel slide for analysis. Estimates of the amount of aragonite and calcite in each sample produced by these analyses have a 5-10% error margin.

Cathodoluminescence examination was performed at the Institute for Geological and Geochemical Research, Hungarian Academy of Sciences (Budapest, Hungary) using a Reliotron type cold-cathode equipment that is attached to a Nikon Eclipse E600 optical microscope. The equipment was operated at 8 to 10 kV accelerating voltage with a 0.5 to 1.0 mA current. Photographs were taken with a Nikon Coolpix 4500 digital camera using automatic exposure and a defocused electron beam.

Elemental determinations were completed for 56 samples of recent (i.e., actively forming) and old travertines that came from different sites in the Denizli Basin. These analyses, done by the Acme Analytical Laboratory (AcmeLabs, Vancouver, Canada), utilized inductively coupled plasma mass spectrometer (ICP-MS) techniques. The total abundances of the major oxides and several minor elements are reported on a 0.1 g sample analyzed by ICP-emission spectrometry following a lithium metaborate/tetraborate fusion and diluted nitric digestion. Loss on ignition (LOI) is by weigh difference after ignition at 1000 °C. The TOT/C and TOT/S were determined by Leco. The determination of element composition was performed also by ICP-MS technique, following a lithium metaborate/tetraborate fusion and diluted nitric digestion.

Stable carbon and oxygen isotope values were obtained from 65 samples at the Institute for Geological and Geochemical Research, Hungarian Academy of Sciences, Budapest, Hungary. Carbon and oxygen isotope analyses of bulk carbonate samples were carried out using the conventional phosphoric acid method (H3PO4 digestion method at 25 °C) of McCrea (1950) and the continuous flow technique of Spotl and Vennemann (2003). Standardization was conducted using laboratory calcite standards calibrated against the NBS-18, NBS-19 standards. All samples were measured at least in duplicate and the mean values are given in the standard delta notation in parts per thousand (%) relative to VPDB (S13C) and VSMOW (618O) according to 6[%] = (Rsample / Rreference — 1) x 1000. Reproducibility is better than ±0.1%.

Table 1

Uranium and thorium isotopic compositions and 230Th ages for selected travertine samples from locations in the Denizli extensional basin by MC-ICP-MS, Thermo Neptune, at NTU. GL: Gölemezli, ÇB: Çukurbag, AK: Akköy, KB: Kocabas, GR: Gürlek, OT: Obruktepe (Karateke), KT: Karateke.

Sample no Weight (g) 238U (ppb) 232Th (ppt) S234U [230Th/238U] [230Th/232Th] Age uncorrected Age correctedc,e ö234Ujnitial

measureda activityc (ppmd) correctedb

GL-13 0.08970 426.4 ± 1.5 2698 ± 15 43.9 ± 4.3 1.0138 ± 0.0056 2645 ± 19 344,640 ± 18,049 344,490 ± 18,024 116 ± 13

GL14 0.05530 457.14 ± 0.53 44.034 ± 327 70.7 ± 2.0 1.097 ± 0.012 188.0 ± 2.4 613,067 ± ~ 611,230 ± ~ 398 ± ~

CB-12 0.05530 85.31 ± 0.24 126 ± 13 435.4 ± 5.6 0.3005 ± 0.0030 3350 ± 335 25,348 ± 308 25,321 ± 308 467.7 ± 6.0

AK-34 0.05660 93.07 ± 0.19 1259 ± 13 206.3 ± 4.3 0.1887 ± 0.0025 230.2 ± 3.8 18,483 ± 279 18,188 ± 314 217.2 ± 4.5

KB-20 0.08280 9.209 ± 0.071 3.3 ± 8.4 239 ± 14 0.714 ± 0.011 33,036 ± 84,544 90,471 ± 2645 90,463 ± 2645 308 ± 18

KB-21 0.11304 10.560 ± 0.082 96.9 ± 6.2 152 ± 13 0.6821 ± 0.0083 1228 ± 79 95,301 ± 2607 95,098 ± 2602 199 ± 17

KB-22 0.08418 15.53 ± 0.16 666.1 ± 9.1 180 ± 18 0.894 ± 0.011 344.0 ± 5.4 145,804 ± 6339 144,908 ± 6287 271 ± 28

GRL-1 0.05339 84.17 ± 0.60 124,905 ± 1401 246 ±10 1.196 ±0.030 13.31 ± 0.35 261,623 ± 25,496 231,517 ± 26,621 472 ±42

GRL-2b 0.05040 184.3 ± 1.1 144,009 ± 1783 275.9 ± 8.5 0.921 ± 0.021 19.47 ± 0.49 130,328 ± 5733 114,341 ± 10,011 381 ± 16

GRL-3 0.05720 159.73 ± 0.93 27,375 ± 206 282.3 ± 8.7 0.933 ± 0.011 89.9 ± 1.2 132,047 ± 3481 128,779 ± 3752 406 ± 13

GRL-4b 0.05798 205.7 ± 1.2 3993 ± 101 260.5 ± 9.3 0.8194 ± 0.0061 697 ± 18 109,032 ± 1989 108,652 ± 1988 354 ± 13

OT-1 0.10168 334.9 ± 1.7 66,492 ± 853 123.5 ± 7.3 0.904 ± 0.015 75.1 ± 1.5 168,619 ± 6919 164,143 ± 7014 196 ± 12

OT-2 0.08824 237.2 ± 1.2 21,452 ± 159 114.3 ± 6.4 0.9139 ± 0.0099 166.9 ± 2.0 177,093 ± 5441 175,061 ± 5430 187 ± 11

0T3 0.10113 269.31 ± 0.98 594.0 ± 7.3 76.1 ± 4.5 1.0997 ± 0.0052 8232 ± 104 543,661 ± 111,313 543,618 ± 111,258 354 ± 184

KT-1 0.08908 481.5 ± 1.7 1243 ± 10 74.5 ± 5.8 1.0952 ± 0.0053 7002 ± 59 519,458 ± 110,081 519,406 ± 110,011 324 ± 169

Chemistry was performed on March 20th, 2010 (Shen et al., 2003), and instrumental analysis on MC-ICP-MS (Shen et al., 2012). Analytical errors are 2ct of the mean.

Decay constants are 9.1577 x 10 6 yr

1 for 230Th, 2.8263 x 10 a S234U = ([234U/238U]activity - 1) x 1000. b S234Uinitial corrected was calculated based on

3yr-1 for U (Cheng etal., 2000), and 1.55125 x 10-10yr-1 for 238U (Jaffey et al., 1971).

Th age (T), i.e., 8234UinMa, = S234Umeasured x ex234 T, and T is corrected age. [230Th/238U]activity = 1 — ex230T + (8234Umeasured / 1000)[K2s0 / (^30 — W] (1 — e—(x23° — X234)T) where T is the age. d The degree of detrital 230Th contamination is indicated by the [230Th/232Th] atomic ratio instead of the activity ratio. e Age corrections were calculated using an estimated atomic 230Th/232Th ratio of 4 ± 2 ppm. The errors are arbitrarily assumed to be 50%.

Travertine samples were gently crushed, ultrasonicated, and dried before U/Th dating. About 50-100 mg for each subsample was selected for U-Th chemistry (Shen et al., 2008) in a class-10,000 geochemical clean room with class-100 benches at the High-Precision Mass Spectrometry and Environment Change Laboratory (HISPEC), Department of Geosciences, National Taiwan University. A triple-spike, 229Th-233U-236U, isotope dilution method (Shen et al., 2003) was employed to correct for instrumental fractionation and determine U/Th isotopic and concentration data (Shen et al., 2002). Measurements of U/Th isotopic abundances were performed on a multi-collector ICP-MS (MC-ICP-MS), Thermo Electron Neptune (Shen et al., 2012). Uncertainties in the U-Th isotopic data were calculated offline (Shen et al., 2002) at the 2ct level and include corrections for blanks, multiplier dark noise, abundance sensitivity, and contents of the four nuclides in spike solution. 230Th dates (before 1950 AD) were calculated using decay constants of 9.1577 x 10-6 yr-1 for 230Th and 2.8263 x 10-6 yr-1 for 234U (Cheng et al., 2000), and 1.55125 x 10-10 yr-1 for 238U (Jaffey et al., 1971).

4. Results

In this section, the travertine sites are described with age determination (Table 1) and the results of water chemistry (Table 2), mineralogical composition (Table 3), and geochemical signatures (Table 4).

4.1. Site descriptions

4.1.1. Golemezli (site 1)

This fossil site, located ~4 km northwest of Golemezli village, is found at elevations between 350 and 450 m on a southwest facing hillside (Fig. 1A). The deposit, which is ~35 m thick over an area of < 1 km2, is located at the west end of the NW-SE trending Akkoy fault (^akir, 1999). This fault segment forms the boundary between the Neogene deposits and metamorphic bedrocks. The volume of travertine is estimated to be ~0.035 km3.

Spaces that developed along the vertical to subvertical fault planes and extensional fissures in the metamorphic bedrocks have been filled by vein travertines that are up to a few metres thick (Fig. 2A). Multiple generations of emplacement are evident from different veins that commonly cross-cut each other (Fig. 2B, C). Downslope, the light green and honey coloured banded travertines pass into the bedded travertines. Vein travertines from this site yield U-Th ages of 344 ± 18 ka, and 613 ka (Table 1).

4.1.2. Pamukkale (site 2)

This huge travertine deposit, ~6.0 km long, is located on the hanging block of the northwest-trending Pamukkale Fault m (Fig. 1A) at elevations between 250 and 400 m (Altunel and Hancock, 1993a,b; Ekmek^i et al., 1995; §im§ek et al., 2000). In this study, Karahayit is regarded as a subsite of the Pamukkale travertine site because of depositional and geochemical differences in the precipitates and water temperatures (Gokgoz, 1994; §im§ek et al., 2000; §im§ek, 2003).

The Karahayit subsite, located northwest of Pamukkale (Fig. 1A, B), is characterized by thermal water with a temperature of ~53 °C (Table 2) that emerges from a fissure and precipitates red to brown travertine that is covered with microbial mats (Fig. 3A, B). Consequently, the local people call it 'Kirmizi Su' (meaning red water in Turkish).

The Pamukkale travertine, designated as a UNESCO world heritage site (§im§ek, 2003), is characterized by its dazzling white colour. The ancient Roman city of Hierapolis, built on this site, is at 365 m asl. The warm waters (~35 °C) that flow from four springs which are located along the Pamukkale Fault and connected fissures (Fig. 1A, B), are now directed to the slope aprons by a closed concrete channel system (Gokgoz, 1994; §im§ek et al., 2000; Kele et al., 2011). Precipitation

•s a

88 I I

CN CN PJ PJ

& CO CN CO O O O r-C3 C3 C5 C3

m ^ ^ ^

r- i- o CN

C3 C3 C5 C3

I I I I

^f cn ^f rn rn ^

t^ O in t^

co co co co

m m m m

I I I I

o co cn cn cn co co co I I I I

o t^ 1 -1

88 I I

CN CO CO

o co co m

co co co co

o o o o

co co o t^

m m ^ m

(N (N (N (N

CN t^ Cn CO <N

O CO o o

CO CO CO CO

^ a & &

(N (N (N (N PJ PJ PJ PJ

CN CN PJ PJ

■ 3 & \

ro ^ ^

2 U 3 P

ifl "cj 13 ^

o ^ ^ rn in CN

co co co

in co in

co cn co co co co i i i

o CN CO PO "Cf & c^ o ci c^ m CO o Ci^

cn r^ CO CO Ci^ un ^

CO o 00 co o^ ^^ cn c^

CO C3 m co uc CN 3 f^

cn LO CO Ci^ o 1 c^

CO CN CO CN cn cn cn C3 3

CO ^ O 8 <J1 o cc CO r^ CO CN 8 CO CN 8 CO O CN r^ m o t-^ CO 3 8 3 CN

m 3 CN LT o <J1 LT <J1 m m <ji CO m m CN 3 CN CN CN 3 8 cc

m m t^

00 CO ^

co co 1-

333 o o o

(N (N (N PJ PJ PJ

CN CN CN LO LO LO

£ m m ^ & >

CO Cu CO « :0

Mineralogical compositions of the travertine samples collected from different sites in the Denizli Basin based on XRD analyses. Ca: Calcite, Ar: Aragonite, Do: domite, Gy: gypsum, Q: quartz, Mi: Mica, Cl: Clorite, Ka: Kaolinite, KF: K-Feldspar, Gt: goethite Ra: rancieite, Sm: smectite, Il: illite, tr: trace amount.

Site name Sample no Travertine type Ca Ar Do Gy Q

Gölemezli GL-1 Vertically banded 100 tr

(1) GL-2 Vertically banded 100 tr

GL-6 Vertically banded 100 tr

GL-8 Bedded 100 tr tr

GL-9 Bedded 80 14 5

GL-10 Raft 95 3 tr

Pamukkalee PK-4 Crystalline crust 99 tr tr

(2a) PK-5 Crystalline crust 100

PK-5b Ooid 100 tr

PK-6 Crystalline crust 100

PK-7 Crystalline crust 100

PK-7b Pisoid 100

PK-8 Pisoid 99 tr

PK-9 Crystalline crust 95 3

PK-10 Crystalline crust 100

PK-11 Crystalline crust 100

PK-12 Crystalline crust 100 tr

PK-13 Crystalline crust 99 1

Çukurbag ÇB-6 Crystalline crust 100

(2b) ÇB-7 Raft 95 5

ÇB-9 Ooid 22 78

ÇB-10 Crystalline crust 100 tr

ÇB-11a Vertically banded 77-100 7-23

ÇB-12 brown Vertically banded 99 tr

CB-12 white Vertically banded 30 70

Akköy AK-8 Bedded 99 tr

(2c) AK-10 Bedded 98

AK-12 Raft and micrit 100

AK-15 Bedded 94 4

AK-45 Bedded 100 tr

Ballik BA-2 Bedded 90

(3) BA-3 Bedded 65 15

BA-4 Bedded 95 4

BA-5 Bedded 100 tr

BA-6 Bedded 100 tr

BA-7 Bedded 100 tr

BA-8 Bedded 100 tr

Kelkaya KL-1 Bedded 100

(4) KL-2 Bedded 100

KL-3 Bedded 100

Kocabaç KB-1 Vertically banded 25 75

(5) KB-2 Vertically banded 93 7

KB-6 Bedded 100 tr

KB-11 Bedded 98 2

Honaz KT-22 Oncolith 100

(6) KT-23 Tufa 100 tr

KT-24 Tufa 100 tr

from these spring waters (Table 2) has produced the vast white calcite deposits with their spectacular arrays of terraces and rimstone pools (Ekmekçi et al., 1995) (Fig. 3C, D). The white slopes are covered mostly by crystalline calcite, whereas lime mud (Fig. 3E), shrubs, pisoids, rafts, and coated gas bubbles are common in the terrace pools (Özkul et al., 2002; Jones and Renaut, 2008; Kele et al., 2011).

The first U-Th ages from the fossil travertine at Pamukkale, obtained by Altunel (1994,1996), ranged from 24 to 400 ka. More recently, a U-Th age of 55.4 ka was obtained from the vertically banded travertines along the scarp of the Pamukkale fault (Uysal et al., 2007).

The Çukurbag subsite is located near the slope apron, just in front of Jandarma spring, Pamukkale, which lies at an altitude between 250 and 300 m asl, and encompasses a relatively flat to gently concave area that is characterized by several fissure ridges that are variably oriented. The E-W trending Çukurbag fissure ridge (Fig. 4A), is ~400 mlong,10 mhigh,and40 mwide (Altunel, 1994). Past quarrying has exposed the central part of the ridge (Fig. 4A-C). Samples from the vertical bands that fill the central fissure space (Fig. 4B, C) yielded U-series ages that range from ~24.7 to 152 ka (Uysal et al., 2007, 2009). One sample from the banded travertines (Fig. 4B, C), collected during this study, yielded a U-Th age of 25.3 ka (Table 1 ).

Today, a mesothermal spring with a temperature of 57.1 °C (Table 2) is located at the east end of the ^ukurbag ridge. Travertine is presently being precipitated in small, shallow pools (a few metres wide) and narrow channels that are located around the spring orifice. Calcite rafts and coated gas bubbles around the pool margins are associated with thermophilic microbial mats (Fig. 4D, E).

The Akkoy subarea, situated ~1 km southeast of Akkoy town at ~250 m asl and southwest of the Karahayit-Pamukkale travertine plateau, is dominated by a large fissure ridge (Figs. 1, 5A). The northwest trending ridge, 1400 m long, 40 m high, and up to 800 m wide at its base, is also known as 'Karakaya Hill' or the 'Akkoy fissure ridge' (Altunel, 1994). A depression lies between the ridge and the Karahayit subarea to the north. Recent quarrying provides a three-dimensional view of the structure and depositional architecture of the Akkoy ridge (Fig. 5A-C). Secondary fissures are associated with the main fissure (Altunel, 1994). The fissure spaces in the ridges are filled with vertically banded/vein travertines (Fig. 5B), which become younger towards the central part of the fissure space (Altunel et al., 1993a,b). Some veins bifurcate towards the surface and cross cut the older bedded travertines (Uysal et al., 2009). Evidence of repeated fracturing and filling events are apparent in some of the fissures (Fig. 5C, D).

Table 4

Element and stable isotopic compositions of travertine samples from the different sites in the Denizli Basin.

Site no Sample ID Travertine type Ca Mg Fe Mn Sr S13C S18O S18O

(ppm) (%oV-PDB) (%V-PDB) (%V-SMOW)

Golemezli GL-1 Vertically banded 398,214 5040 < 280 232 1073 3.8 -16.0 14.5

(1) GL-2 Vertically banded 397,857 5040 350 387 1156 3.8 - 16.6 13.8

GL-3 Vertically banded 396,571 4560 1679 542 660 4.0 - 13.1 17.4

GL-4 Vertically banded 368,429 6960 2868 155 751 3.7 - 15.2 15.2

GL-5 Vertically banded -0.3 - 15.1 15.4

GL-6 Vertically banded 392,214 8160 629 155 1756 3.7 - 15.0 15.5

GL-7 Vertically banded 4.6 - 15.0 15.4

GL-8 Bedded 405,214 2640 769 77 769 5.0 - 13.9 16.6

GL-9 Bedded 329,857 20,940 4896 155 586 4.8 - 13.2 17.3

GL-10 Raft 340,929 5940 6365 155 484 4.0 - 15.4 15.1

Pamukkale PK-4 Crystalline crust 401,000 3240 560 <77 2991 11.3 - 9.3 21.3

(2) PK-5 Crystalline crust 6.1 - 10.7 19.9

PK-5b Pisoid 394,429 4080 280 <77 1971 6.6 - -

PK-6 Crystalline crust 5.9 - 10.2 20.4

PK-7 Crystalline crust 5.9 - 10.6 20.0

PK-7b Pisoid 396,643 4440 <280 <77 1973 6.4 - -

PK-8 Pisoid 402,571 3240 <280 <77 2219 5.8 - 11.1 19.5

PK-9 Crystalline crust 349,214 3840 2098 <77 1920 7.1 - 10.3 20.3

PK-10 Crystalline crust 400,786 3060 <280 <77 2346 8.2 - 10.5 20.1

PK-11 Crystalline crust 9.7 - 9.7 20.9

PK-12 Crystalline crust 396,214 3840 <280 <77 3028 11.5 - 9.0 21.6

PK-13 Crystalline crust 398,643 3720 839 <77 2715 11.7 - 9.1 21.5

Çukurbag ÇB-6 Crystalline crust 396,429 120 <280 <77 7392 4.3 - 15.6 14.8

(2a) ÇB-7 Crystalline crust 392,000 4320 2308 77 3771 5.8 - 14.0 16.5

ÇB-8 Raft and coated gas bubble 393,214 5760 1678 77 3113 5.5 - 13.3 17.2

ÇB-9 Ooid 390,571 2940 2378 77 2306 4.3 - 15.2 15.2

ÇB-10 Crystalline crust 398,500 2280 <280 1162 1285 4.7 - 4.8 26.0

ÇB-11 Banded 396,714 180 6822 5.2 - 15.8 14.6

ÇB-12a Banded 396,000 780 2238 5692 5.3 - 16.6 13.8

ÇB-13 Banded 398,214 1740 560 77 2215 5.1 - 11.1 19.5

Akkoy AK-1 Raft 401,571 4980 2308 155 1540 5.6 - 12.4 18.1

(2b) AK-3 Bedded 392,214 1980 7204 155 803 5.3 - 10.6 20.0

AK-4 Bedded 389,500 960 16,157 542 410 5.6 - 9.1 21.5

AK-5 Raft 386,500 1740 16,786 310 549 5.9 - 9.2 21.4

AK-6 Bedded 332,286 6120 13,009 232 1677 4.3 - 10.9 19.6

AK-8 Bedded 367,000 1860 37,209 232 705 8.0 - 10.3 20.3

AK-10 Bedded 388,929 1920 7414 542 615 6.2 - 10.8 19.8

AK-11 Bedded 380,786 1920 10,911 232 660 7.0 - 9.7 20.9

AK-12 Raft 391,714 2280 6085 232 548 7.3 - 10.8 19.7

AK-15 Bedded 391,286 1860 3007 77 564 7.8 - 8.6 22.1

Ballik BA-1 Crystalline crust 382,714 1560 <280 <77 1590 - 1.7 - 8.7 22.0

(3) BA-2 Dark coloured, manganiferous 276,286 3960 2238 27,648 190 - 0.4 - 8.9 21.8

BA-3 Bedded, reddish brown 350,071 4920 75,325 1471 250 - 3.3 - 8.5 22.1

BA-4 Bedded 401,857 1500 6225 774 796 0.6 - 7.8 22.8

BA-5 Bedded 384,785 2760 769 154 184 - 0.1 - 8.8 21.8

BA-6 Bedded 395,000 3240 3497 <77 274 - 0.6 - 11.1 19.5

BA-7 Bedded 394,214 3540 <280 77 398 0.3 - 11.4 19.2

BA-8 Bedded 396,643 3240 <280 <77 457 - 0.2 - 9.6 21.0

BA-9 Bedded 396,643 3240 <280 <77 458 - 0.2 - 9.6 21.0

BA-10 Bedded 397,929 2640 <280 <77 329 0.8 - 8.7 21.9

BA-11 Bedded 396,000 2280 490 <77 296 - 0.1 - 9.8 20.9

Kelkaya KL-1 Bedded 355,274 6091 <280 <77 2103 3.4 - 8.2 22.4

(4) KL-2 Bedded 361,420 3316 <280 <77 2181 5.8 - 8.1 22.6

KL-3 Bedded 361,849 3558 2098 <77 2330 3.2 - 8.9 21.7

Kocabaç KB-1 Vertically banded 5.6 - 13.5 17.0

(5) KB-2 KB-3 KB-4 Vertically banded Vertically banded Vertically banded 4.4 4.7 4.7 - 15.2 - 13.7 - 14.3 15.3 16.8 16.2

Honaz HO-1 Tufa, Karateke 387,221 5669 1748 <77 856 - 4.0 - 10.3 20.3

(6) HO-2 Tufa, Karateke 397,798 3437 699 <77 439 2.6 - 9.6 21.0

HO-3 Tufa, Karateke 397,071 3360 <280 <77 462 2.9 - 9.7 20.9

HO-4 Bedded tr., Obruktepe 398,643 2400 <280 <77 644 4.7 - 9.4 21.2

HO-5 Tufa, Degirmenler 379,928 4080 699 <77 975 - 0.9 - 9.4 21.2

HO-6a Tufa, Degirmenler 379,214 7380 769 <77 1053 - 0.4 - 9.2 21.4

HO-7a Tufa, Kayaalti - - - - - 1.8 - 10.0 20.5

HO-8a Tufa, Kayaalti 373,214 4140 1608 <77 844 0.1 - 8.3 22.3

a From HorvatinciC et al. (2005).

Bundles of raft flakes, earlier considered as bedded travertine (Uysal et al., 2009, their Fig. 2E), red mudstones rich in Fe-oxide, and angular clasts of older travertine and metamorphic bedrocks are found in some of the fissures (Fig. 5C-E). The white, horizontal ledges

(Fig. 5E) formed when the fissure spaces were filled with thermal waters. Some rafts were thickened due to subsequent encrustation (cf., Guo and Riding, 1998) whereas other rafts are now in vertical to subvertical positions.

Fig. 2. Field views of Golemezli travertine site. (A) Vein travertines with network appearance located along steeply dipping fault planes and fractures in metamorphic bedrock, exposed in quarry wall. U-Th dates are shown. View to northeast. (B) Cut surface showing cross cutting relationships between different generations of vein travertine. (C) Interpretation of surface shown in panel B, indicating vein generations 1, 2 and 3.

Bedded travertine, which dips at 10-20° away from the ridge crest, is present on both sides of the fissure axis (Fig. 5A, B). These beds are formed from alternating layers of crystalline shrubs, rafts, micrite, and coated gas bubbles (Fig. 5F-H). Palaeosols, up to 60 cm thick, and lithoclast interbeds are found in some parts of the succession. A sample from the uppermost part of the fissure ridge gave a U-Th age of 18.4 ± 0.3 ka (Table 2).

The Pamukkale travertine plateau, which includes the Karahayit, ^ukurbag and Akkoy subsites, has an average thickness of 50 m and covers an area of 11.8 km2. Thus, the volume of travertine is estimated to be about 0.6 km3.

4.1.3. Ballik (site 3)

Ballik, located ~5 km northwest of Kaklik, has travertine deposits that are exposed on a stepped southwest facing slope (Fig. 1A, C) that is 500 to 1000 m asl. It is the largest travertine site in the basin with travertine up to 120 m (average 75 m) thick that covers an area of 12.5 km2. The volume of travertine at this site is approximately 0.94 km3. The travertine deposit is dissected by northwest-trended normal faults and extensional fractures.

Today, there are about 50 quarries operating in this area. The cut surfaces in the quarries allow investigation of the spatial architecture of the depositional settings that developed as the travertines were being deposited. In the Ballik area, the travertine sequences are formed mostly of horizontally to subhorizontally bedded travertine, particularly in the lower and middle parts (Fig. 6A, B), that extends laterally for a few hundred metres. These successions are formed mainly of thin shrub layers, micritic laminae, coated bubbles, and locally included reed casts. Throughout the successions exposed in the quarries there are intercalations of fluvial conglomerate, sandstone,

and red and green mudstone (Fig. 6A, B), palaeosol horizons, and ero-sional surfaces (Ozkul et al., 2002). At Killik Tepe, south of this area, the horizontally bedded travertines at the base pass gradually upward into the low angle slope travertine that, in turn, passes upward into the tufa facies at the top (Fig. 6C, D). The slope facies is composed mainly of crystalline calcite, shrub, and micritic layers that are locally cut by erosional surfaces.

The uppermost tufa horizon, which can be followed laterally for several hundred metres, is formed from numerous mound and waterfall build-ups, each of them are convex upward and dominated by reed and bryophyte tufa. The mound facies, up to 25 m thick, typically have high porosity due to the presence of big cavities of a few metres in diameter and reed cast moulds (Fig. 6C, D).

4.1.4. Kelkaya (site 4)

This travertine at this site, located near the village of A^agidagdere, is at the foot of the north facing slope of the Kelkaya Hill at elevations between 520 and 750 m asl (Fig. 1A). Kelkaya Hill is formed from Mesozoic limestone bedrock. The travertine is ~50 m thick and covers an area of 0.75 km2. The estimated volume of travertine is 0.04 km3. The spring waters arise along a normal fault plane that dips to the north. Waterfall tufas with porous textures dominate the upper part of the site (Fig. 7A). In contrast, the horizontally bedded and low-angle slope travertines that dip to the north cover the lower part of the area (Fig. 7A). The terrace pools and their rims are clearly evident on the vertical wire-cut surfaces (Fig. 7B). In the quarry, the uppermost bench (~10 m high) is separated from the lower benches by a palaeosol and green claystone horizons.

Fig. 3. Field photographs from Karahayit-Pamukkale. (A) Thermal water, with a temperature of -53 °C, discharging from a pipe that has been placed in spring vent at Karahayit. Note reddish coloured travertine and green microbial algal mat on surface. (B) Surface of travertine partly covered by green microbial algal mat. (C) Stepped slope formed from travertine, slope apron in front of the Jandarma spring, Pamukkale. (D) Terraced slope travertines at Pamukkale, terrace pools filled with water. (E) Terrace pool lined by rimstone with lime mud on floor of pool.

4.1.5. Kocaba$ (site 5)

Travertines around the villages of Kocaba? and Gurlek, herein referred to as the Kocaba? site (Fig. 1A, C), are located on a relatively flat area that lies between 370 and 495 m asl. In this area, the travertine is -40 m thick and covers an area of -33.7 km2. The estimated volume of travertine is about 1.0 km3. Highway 320 (from Denizli to Afyon) passes through the site. The area, bounded to the north by the Ballik site, is formed mainly of several NW-trending inactive fissure ridges that are flanked by bedded travertines (Ozkul et al., 2002; Altunel and Karacabak, 2005). The main fissures in the ridges are filled by vertically banded travertines that are up to 16 m thick. Based on the observation from the quarry faces that are perpendicular to the main fissure axis, the bedded travertines (Fig. 8A), with thickness up to 50 cm, dip away at angles of 5-10°. Within the bedded travertine, bedding-parallel crystalline layers, up to 60 cm thick

(Fig. 8A, B), are formed almost entirely of calcite and/or aragonite (Table 3). The bedding-parallel crystalline layers (or veins) are commonly brecciated (Fig. 8B, C) and similar to the 'jigsaw puzzle' or 'crackle textures' found in the Akkoy fissure ridge near Pamukkale (Uysal et al., 2009). In the Ku?golu fissure ridge, immediately north of Highway 320, vertical and horizontal veins cut the bedded travertines and each other many times (Fig. 8E, F). U-Th dates obtained from these vertical and horizontal veins yielded ages of 90.5, 95.1, and 144.9 ka (Table 1). Thermoluminescence dates of 330 ± 30 and 390 ± 40 ka (Ozkul et al., 2004) and U-Th dates of 105 ± 9.8 to >400 ka (Altunel and Karacabak, 2005; De Filippis et al., 2012) have been obtained from samples collected from this travertine.

Near Gurlek (site 5a), -7 km west of Kocaba? (Fig. 1A), the horizontally bedded travertines are exposed at -370 m asl. The travertine benches are composed of light, medium and dark coloured horizons

Fig. 4. Field photographs of ^ukurbag. (A) Westward view of the Qjkurbag fissure ridge, partly exposed by quarrying. (B) Vertical banded and associated bedded travertine in central part of ridge with U-Th date. (C) Vertical bands expanding upward, eastern end of the quarry shown in panel A. (D) Recent coated gas bubbles, rafts and microbial mat on the pool surface around ^ukurbag hot spring, eastern end of ridge. (E) Recent to subrecent, rafts and associated coated gas bubbles.

(Fig. 9A) that are locally interbedded with palaeosols, claystone, and mudstones. Some of the medium and dark horizons contain numerous gastropods. Travertine samples from the quarries yielded U-Th ages of 231.5 ka, 128.8 ka, 114.3 ka and 108.7 ka (Table 1).

4.1.6. Honaz (site 6)

This area is (site 6) located at the west end of the east-west trending Honaz Fault, which forms the southern margin of the basin

(Fig. 1A, C) and is still seismically active today (Bozku§ et al., 2000). Sites at Degirmenler, Kayaalti and Karateke are included in the Honaz site. Both travertine and tufa are found at these localities (Fig. 9A-D). At Honaz, the travertine is 30 m thick and covers an area of 2.7 km2. The volume of travertine is estimated at 0.08 km3. At Kayaalti (subsite 6a), fossil tufa of Holocene age is exposed on a northwest facing cliff face (Fig. 9B) that is ~20 m high (Horvatincic et al., 2005). The upper part of this tufa body is formed from the

waterfall facies (Fig. 9B), with hanging tufa curtains and primary cavities (cf., Pedley et al., 2003; Ozkul et al., 2010). Below the waterfall, the distal slope is covered with subhorizontally bedded detrital tufas. Today, tufa precipitation is restricted to the Degirmenler subsite (Horvatincic et al., 2005). Older tufa and travertine samples (KT-2, OT-3), collected near Karateke (Fig. 9C), yielded U-Th ages of 519.4 ka and 543.6 ka (Table 1). Two other samples (OT-1, OT-2) from the travertine quarry at the Obruk Tepe nearby Karateke (Fig. 1A) yielded U-Th ages of 164.1 ka and 175.1 ka (Table 1).

4.2. Water chemistry

Today, travertine precipitating spring waters belong to four groups (I to IV) based on their physico-chemical features and locations in the basin: (I) Karahayit and ^ukurbag, (II) Pamukkale, (III) Kaklik cave and Kelkaya, and (IV) Honaz (Table 2). The reservoir rocks for the thermal waters in the Karahayit and ^ukurbag sites are Palaeozoic marble, whereas the main aquifers supplying warm and mineralized water to the Pamukkale thermal springs are Palaeozoic marbles and Mesozoic limestones (§im§ek et al., 2000).

Thermal waters from the springs and wells discharge at various temperatures. Relatively higher temperature thermal waters (32.957.1 °C) arise along the northern margin of the basin at Karahayit, Pamukkale and ^ukurbag, whereas the waters with lower temperature (18.7-23.7 °C) appear in and around Kaklik cave and Kelkaya to the southeast and Honaz to the south (Fig. 1A, Table 2).

The pH of waters in the study area varies from 6.0 to 7.4 (Table 2) and increases up to 7.8 towards the distal parts of discharge aprons at Pamukkale (Kele et al., 2011). The electrical conductivity (EC) values of the thermal waters vary between 2770 and 3090 ^S/cm in Karahayit and ^ukurbag, 2360-2400 ^S/cm at Pamukkale and 1769-1932 |jS/cm at Kelkaya-Kaklik. Honaz waters have lower EC values (685-1147 ^S/cm). Electrical conductivity, almost all ions, SiO2 and Sr values of the waters in the basin increase with increasing water temperature (Table 2). The waters in the springs at Pamukkale and Pinarba^i and associated wells are of the Ca-Mg-HCO3-SO4 type. In contrast, the waters from the ^ukurbag thermal spring, well KH-3 in Karahayit, Kaklik cave and Kelkaya are of the Ca-Mg-SO4-HCO3 type (Table 2). Basin-wide, the 618O and 6D values of thermal and cool waters are between — 9.0 and — 8.1% and between — 51.8 and — 58.9%», respectively (Table 2) (Forizs et al., 2011). The values indicate a meteoric origin for the waters with residence times of about 20-30 years (§im§ek, 2003). The R/RA and mantle-derived CO2 are higher in the eastern part of Buyuk Menderes Graben (Karaku? and §im§ek, 2013). The 613C analyses from the waters show that the CO2 required for travertine deposition was derived largely from the decomposition of carbonate rocks (Filiz, 1984), along with a substantial contribution of magmatic CO2.

4.3. Travertine mineralogy

At Pamukkale and ^ukurbag, aragonite, in varying amounts, is associated with the vertically banded, ooid, and coated gas bubble travertine (Table 3; Figs. 10,11,13).Awhite, compact and finely crystalline crust sample (^B-6) (Fig. 10A) formed entirely of aragonite (Fig. 10A-D) was found adjacent to the orifice of ^ukurbag spring, where mesothermal waters are discharged at a temperature of 57 °C.

Some of the vertically banded travertine from the ^ukurbag fissure ridge (Fig. 11A) is formed by fibrous aragonite and rhombohe-dral calcite (Fig. 11B-D), each being segregated into alternating laminae, as confirmed also by cathodoluminescence analysis. Several aragonite samples from the fine crystalline crust and vertically banded travertines show rosette forms that are radially arranged (Fig. 10B-D).

Subrecent aragonite ooids, up to 8 mm in diameter (Fig. 13A, B), are found around the orifice of ^ukurbag spring. The concentric cortical laminae (Fig. 13C, D), 3 to 13 |jm thick (Fig. 13E), are formed largely of aragonite needles that are up to 3 |jm long (Fig. 12F). The ooids are held in calcite cement (Table 3). Goethite (<2%) is present in some of the iron-rich samples.

Some of the samples from Golemezli contain up to 14% dolomite grains that are up 200 |jm long (Fig. 14). The dolomite is typically found with quartz and mica grains (Fig. 14A, B) in mm-scale layers, lenses, and pockets that lie between calcite laminae (Fig. 14A). Some of the grains were removed from their places leaving behind mouldic cavities (Fig. 14C, D). This association of the dolomite with the quartz and mica indicates that it is probably of detrital origin.

4.4. Travertine geochemistry

4.4.1. Major and trace elements

Based on 55 samples that were analyzed geochemically (Table 4), the Ca concentrations in the travertines range from 329,857 to 405,214 ppm whereas the Mg varies from 120 ppm at ^ukurbag to 8160 ppm at Golemezli (site 1). Calcite vein travertine at Golemezli yielded Mg values of 20,940 ppm. The minimum and maximum Sr values in calcite are from Ballik (184 ppm) and Pamukkale (3028 ppm), respectively. Aragonite at ^ukurbag (Pamukkale-site 2), which comes from the vertically banded travertine (^B-6) along the central fissure of the ^ukurbag ridge, contains up to 7392 ppm Sr (Table 4).

Fe and Mn values of the calcite are lower in the light coloured travertines (e.g., ~280 ppm Fe, 70 ppm Mn, in the white crystalline calcite travertine precipitated on the slope facies of Pamukkale). The Fe (37,209 ppm) is high in the reddish-brown calcite from the Akkoy travertine ridge (Table 4). The Mn content in the calcite is 1471 ppm from the dark brown coloured travertine from Ballik (site 3). In the same site, an abnormal Mn value of 27,640 ppm is from secondary enrichment in a cavity. Although there are some exceptions, basin-wide there is a positive correlation (R2 = 0.58) between Fe and Mn (Table 4).

Based on the XRD analysis of ~50 travertine samples, calcite is the most common mineral at all sites (Table 3). Aragonite and dolomite are, however, also present in the travertines found on the northern basin margin at ^ukurbag (Table 3, Figs. 10, 11) and Golemezli, respectively. Although micrite (grains < 4 |am long) dominate most deposits, the vertically banded travertine, crystalline crusts and secondary pore fills are formed largely of coarse spar calcite (Figs. 11D, 12). The white, recently precipitated crystalline crust travertine (Fig. 12A-D), precipitated on the slope in front of Jandarma spring at Pamukkale, is formed mainly of dendritic calcite (Fig. 12D) (Jones and Kahle, 1986; see also Kele et al., 2011). There, minor amounts of aragonite (< 1%) are also present on the distal part ofthe discharge apron. Some samples contain detrital minerals, including quartz, gypsum, mica, goethite, smectite and kaolinite in very small amounts (Table 3).

4.4.2. Stable carbon and oxygen isotope composition

The stable carbon (ô13C) and oxygen (ô18O) compositions of the travertine deposits, derived from 66 samples, vary from one site to another in the basin (Table 4, Fig. 15). The ô13C values range from - 4.0 to +11.7%» PDB, whereas S18O values range from -16.6 to — 7.8% PDB. Along the northern boundary (e.g., Golemezli, Pamukkale, Çukurbag, and Akkoy) and around Kocabaç, the recent and fossil travertines have high ô13C values ( + 3.7 to +11.7%). In contrast, the lowest ô13C values ( — 4.0%) are from the tufas that are located along the southern boundary (e.g., Honaz) where both positive and negative values of the ô13C have been measured. Similar 813C values ( — 3.3 to +0.8%) were obtained from Komurcuoglu quarry, which is located east of the Ballik site (Table 4, Fig. 15).

Fig. 6. Field photographs of Ballik site. (A) Horizontally bedded travertines (tr) at the bottom, overlain by green mudstones (gm) and sandstone (ss), west end of ^akmak quarry, southwest part of the area. (B) Close view of horizontally bedded travertines in A. (C) Tufa mound formed mainly of reed cluster with large cavities, quarry face is -5.5 m high. (D) Reed casts, in growth position, in tufa mound.

5. Interpretation of data

5.1. Variations in depositional setting of the travertine

The depositional architecture of the travertines varies from locality to locality in accord with the conditions that existed at each site when the water was actively flowing. At Ballik, the horizontally bedded travertines, exposed in the lower part of the succession, grade upward into the low angle slope facies and then to the tufa horizon. The uppermost tufa horizon is a dome-like structure formed from numerous mound and waterfall bodies, which have coalesced with each other in all directions (Ozkul et al., 2002). The mound facies is comparable to the 'reed mound facies' found in Rapolano Terme, Italy (Guo and Riding, 1998). The waterfall facies is similar to those found at other localities (Pedley et al., 2003; Ozkul et al., 2010). The lower part ofthe succession formed in a shallow lake or depression environment that was fed by warm thermal waters that were modified by rainwater. This part of the succession is comparable to the travertines found at Tivoli, near Rome (Chafetz and Folk, 1984; Minissale et al., 2002; Faccenna et al., 2008), Rapolano Terme in Tuscany, central Italy (Guo and Riding, 1998), and Süttó, which is located on the Danube River, ~60 km northwest of Budapest, Hungary (Sierralta et al., 2010).

The recent travertines at Pamukkale have developed largely on smooth and terraced slope environments that are fed by warm thermal springs (T = -34.5 °C at the spring orifice). Downslope variations in the thermal waters and associated precipitates are well documented (Ekmek^i et al., 1995; §im§ek et al., 2000; Kele et al., 2011). The slight downslope increase in the SD and S18Owater values is attributed to progressive evaporation (e.g., Fouke et al., 2000) and decrease in water temperature. The downslope increase in S13Ctravertine is linked to the amount of CO2 degassing (Kele et al., 2011).

Fissure ridges are one of the most prominent features in the region. They have been considered as important tools for tectonic investigations (Altunel and Hancock, 1993a,b; Hancock et al., 1999; Brogi and Capezzuoli, 2009; De Filippis etal., 2012) because local stress directions can be inferred from them (Hancock et al., 1999). At Pamukkale, for example, the east-west- and northwest-southeast-trending fissure ridges are products of local N-S and NE-SW stress directions, respectively.

Along the southern boundary of the basin, travertine and tufa deposits developed along the Honaz boundary fault. Travertines formed in these areas display slope and ridge-like depositional morphologies, whereas the tufa deposits are evident in the paludal and waterfall facies (Horvatincic et al., 2005). The coexistence of tufa and travertine at adjacent localities at the western end of the Honaz fault probably

Fig. 5. Field photographs of travertines in the Akkoy fisssure ridge showing which veins. Here bt: bedded travertine, lt: lithoclast, r: raft, and v: vein. (A) Cross-section view perpendicular to the central fissure axis. Bedded travertines gently dip away opposite sides with respect to the ridge axis in the middle. View from the northwest. (B) Fissure space, 3 m wide, bounded by vertical bands on both sides. (C) Multistage fissure fill with sharp contact. (D) Multistage vertical veins and raft clusters attached to the fissure wall. (E) Horizontal ledges (white) on fissure wall, resulting from progressive accumulation of rafts. (F) Bedded travertines composed of crystal shrubs. (G) Dendritic crystal shrub layers. (H) Coated gas bubbles in bedded travertines, north flank of Akkoy fissure ridge.

Fig. 7. Field photographs of Kelkaya site, situated at the foot of northern slope of the Kelkaya Hill. (A) Light coloured travertines were deposited on a gentle slope that prograded to the north. The uppermost bench (-12 mhigh) is underlain by apalaeosol layer (ps). Keltepe Hill, in the background, is composed of Mesozoic limestone. (B) Terrace pools and rims (r) on cut surface showing successive rims aligned on the same line (arrowed). Flow to the right.

resulted from shallow and deeply circulated waters in the fault damage zone (Brogi and Capezzuoli, 2009).

5.2. Occurrence of aragonite versus calcite

Precipitation of the calcite and aragonite found in these travertines was controlled by a complex set of interrelated parameters (Jones and Renaut, 2010). Aragonite and calcite precipitation depends on many factors including water temperature, growth inhibitors, supersaturation with respect to CaCO3 caused by CO2 degassing and/or evaporation (Chafetz et al., 1991), Sr content (Malesani and Vanucchi, 1975), Mg/Ca ratio (Folk, 1994), and various biological factors (Kitano, 1962; Busenburg and Plummer, 1986; Renaut and Jones, 1997; Pentecost, 2005; Jones and Renaut, 2010; Rodriguez-Berriguete et al., 2012). Based on a survey of spring deposits, Folk (1994) suggested that water temperature and the Mg:Ca ratio are the main controlling factors. Thus, he argued that aragonite is precipitated from waters with a temperature > 40 °C, whereas calcite precipitates when the temperature is <40 °C. There are, however, many exceptions. At Angel Terrace, Mammoth Hot Springs in Yellowstone National Park, for example, aragonite is precipitated at temperatures > 44 °C, whereas aragonite and calcite co-precipitate between 30 and 43 °C, and only calcite precipitates below 30 °C (Fouke et al., 2000). At Egerszalok, Hungary, where the water temperature is -70 °C, almost pure calcite is precipitated around a borehole (e.g., spring vent), whereas farther downslope, various amounts of aragonite (5-35%) are being precipitated where the water temperatures are 45-50 °C (Kele et al., 2008). Similarly, calcite was precipitated directly from waters with temperatures of >90 °C in Kenya (Jones and Renaut, 1995) and New Zealand (Jones et al., 1996). In natural settings, identifying the factor(s) that control the precipitation of these polymorphs is difficult because all the environmental factors are operating simultaneously.

In the Denizli Basin, calcite is the dominant mineral found at all travertine sites. Aragonite is found in two different settings that clearly illustrate the diverse settings where aragonite can develop.

• Some of the banded travertines, found along fault planes and exten-sional fissures, are formed from alternating calcite and aragonite laminae. Uysal et al. (2007) showed that the aragonite and calcite found at Pamukkale were original precipitates with no evidence that the calcite had formed by inversion of the aragonite. In other springs, aragonite precipitation has been attributed to periods of rapid CO2 degassing and the consequent high levels of supersaturation (Jones and Renaut, 2010). Occasionally, aragonites include needle aggregates, some of them may have microbial nuclei such as

bacteria and pollen, on which the needles grew (Guo and Riding, 1992). The aragonite in the fine crystalline crust- and vertically banded travertines examined in this study is formed from needle aggregates that do not appear to include any microbial nucleus (Fig. 10B-D). Spherulitic aragonite growth, without microbial nucleus, has been ascribed to high disequilibrium conditions (Jones and Renaut, 1995) or inorganic processes that include rapid CO2 degassing (Pentecost, 1990). • Aragonite ooids and crusts formed in and around the margin of the ^ukurbag hot spring pool. Similar aragonitic ooids, up to 5 cm in diameter, have also been reported from Tekkehamam (called as 'Tekke Ilica' in Richter and Besenecker, 1983), close to Saraykoy town southwest of the Denizli Basin, which is one of the main geothermal fields in the Denizli province. There, aragonitic ooids may have formed from the thermal waters (up to 99.7 °C) that are of Na-SO4-HCO3 type (§im§ek, 2003), high supersaturation level (Sc: 0.6-0.7) with respect to CaCO3, and rapid CO2 degassing (Ali Gokgoz, unpublished data).

5.3. Occurrence ofdolomite

Dolomite is rare in travertine. Barnes and O'Neil (1971) found dolomite in a few Californian travertines and trace amounts of dolomite have been found and in some of the Japanese (Kitano, 1963) and Italian hot springs (Folk, 1994). Minor amounts of dolomite were found with calcite and aragonite around hot spring orifices at Chemurkeu on the western shore of Lake Bogoria (Renaut and Jones, 1997).

In the Denizli Basin, dolomite was found in some of the banded travertine at Golemezli along with quartz and mica grains. There, the underlying bedrocks are formed from dolomite and dolomitic limestone. The association of the dolomite with the quartz and mica suggests that the dolomite is probably of detrital origin.

5.4. Strontium distribution at the travertine sites

In the Denizli Basin, the Sr in the hot waters and travertines ranges from 1.16 to 9.09 ppm and 184 to 7392 ppm, respectively (Tables 2,4). Travertines found at the northern sites (i.e., Golemezli, Yenice, Karahayit, Pamukkale, ^ukurbag and Akkoy) contain more Sr than the travertines found in the east, southeast and south localities (i.e., Ballik, Kelkaya and Honaz). The highest Sr contents in the northern sites are from the vertically banded- and crystalline crust travertines. The Sr in the modern crystalline calcite travertines at Pamukkale decreases significantly from the proximal to distal areas on the slope aprons (Kele et al., 2011). A

Fig. 8. Field photographs of crystalline vein travertines exposed in quarry faces around Kocaba?. (A) Light coloured, gently inclined crystalline beds, up to 60 cm in thickness, penetrated as an interlayer between the bedded travertines. Box labelled B indicates area shown in panel B. (B) Crystalline calcite bed in the inset area in A. (C) Brecciated crystalline vein travertine on the quarry face. (D) White crystalline vein beds penetrated along the bedding plane of the dark travertines. (E) Cross-cut relationship between the vertical and horizontal veins on quarry wall cut in fissure ridge, the Ku$golu area, north of the Highway 320. Box labelled F indicates area shown in panel F. (F) Close view of the cross-cut relationship, inset area in E. Many vertical and horizontal veins cut of each other repeatedly.

Fig. 9. Field photographs of the Gurlek and Honaz sites. (A) Light and dark coloured, horizontally bedded travertines with palaeosol horizon, quarry face (5 m high), near Gurlek, Kocaba$ showing U-Th date. (B) Waterfall facies with hanging tufa curtain (-20 m high), Kayaalti sublocality, west end of the Honaz fault zone, south of the basin. (C) Waterfall tufa exposure near the Karateke village, with a U-Th age of 520 ± 110 ka, west end of the Honaz fault zone. (D) Parallel bedded travertine sequence tilted to the southeast due to faulting, Gokpmar area, southwest boundary of the basin (see Fig. 1 for locations).

similar trend was reported from Angel Terrace, Mammoth Hot Springs (Yellowstone National Park, U.S.A.) (Fouke et al., 2000).

In central Italy, the Sr concentration of inactive travertines (up to 1500 ppm at Viterbo) and Sr concentration of present thermal springs are higher to the west of the Tiber Valley than those to the east (Minissale et al., 2002, their Fig. 8). The elevated concentration of Sr and SO4 in thermal springs west of the Tiber Valley is probably caused by the interaction of the circulating groundwater with the gypsum in the Triassic Burano Formation found at the base of the Mesozoic limestone sequence (Minissale et al., 2002).

In the Denizli Basin, the Late Triassic Kizilyer Formation (Gundogan et al., 2008), which is formed from evaporites and dolomites (Al^ek et al., 2003), crops out as a tectonic slice at the eastern end of the Honaz fault near Kizilyer, may be the source of Sr found in the travertines. The deeply circulated thermal waters in the basin could have interacted with the buried part of the Kizilyer Formation. Some Sr may also have come from the Neogene strata in the Denizli Basin that includes gypsum intercalations in the areas around Saraykoy and Golemezli (Al^i^ek et al., 2007).

5.5. Basinal variations of the stable isotopic composition

Stable isotopic compositions of travertines have been linked to regional fluid flow, active tectonics, and palaeoenvironmental and palaeoclimatic changes during the Late Quaternary (Chafetz and Lawrence, 1994; Guo et al., 1996; Hancock et al., 1999; Uysal et al., 2007, 2009; Sierralta et al., 2010; Kele et al., 2011). For example,

there are consistent differences in the geochemical and isotopic signatures ofmodern and fossil travertines, associated thermal spring waters, and gas vents found on the east and west sides of Tiber Valley (central Italy). On the west side of Tiber valley, 613C values of fossil travertines are higher than those to the east side (e.g., the Ancona-Anzio line). Fossil travertines to the east have characteristics typical of meteogene origin (e.g., tufa including abundant plant casts and organic impurities) (Minissale et al., 2002).

In the Denizli Basin, the stable isotope compositions of the travertine deposits show some variations based on the location within the basin. The 613C values of recent and fossil travertines found along the northern boundary (i.e., Pamukkale, Karahayit and Golemezli) yielded more positive 613C values than those from the other sites (Table 4, Fig. 15A). More positive 613C values like these have been attributed to the contribution of CO2 liberated by thermometamorphic processes associated with magmatic activity (Kele et al., 2011). The shift in the 613C values, up to + 11.7%» PDB, are thought to arise from the more rapid CO2-degassing that was associated with the fast flowing water on the steeper parts of the downslope sections (Kele et al., 2011). In contrast, the less positive and more negative values obtained from other precipitates possibly resulted from mixing between deeply sourced CO2 and soil-derived CO2 (Crosseyet al., 2006).

5.6. Palaeotemperature calculations

Calcite oxygen isotope palaeothermometry generally uses equations that implicitly assume equilibrium isotope fractionation during

Fig. 10. (A) Dense, white aragonite crust (sample ^B-6) from ^ukurbag hot spring orifice, Pamukkale. (B) SEM image of aragonite rosettes from the ^B-6 sample. Box labelled C indicates area shown in panel C. (C) Area formed from radially arranged aragonite crystals. (D) Aragonite rosette from vein travertine, ^ukurbag fissure ridge.

carbonate precipitation (e.g., Friedman and O'Neil, 1977 [Eq. (1)], Kim and O'Neil, 1997 [Eq. (2)]):

103 lnac_w = (2.78 x 106)/T2-2.89 (1)

103 lnac_w = 18030/T-32.42 (2)

where a = (S18Ocakite + 1000) / (S18Owater + 1000) and 103lnac-w « S18°calcite — S18Owater.

For these equations, palaeotemperature calculations are based on the S18Ocaicite values and the S18Owater. In the case of fossil travertine deposits, however, the S18Owater is not known and must therefore be inferred. This is further complicated by the fact that isotopic equilibrium is rarely maintained under natural conditions (e.g., Coplen, 2007; Kele et al., 2008; Demeny et al., 2010; Tremaine et al., 2011). At Pamukkale, for example, there is a systematic positive shift in the A(calcite-water) values downslope (Kele et al., 2011). For travertines precipitated around spring orifices, deviation from the equilibrium can cause the calculated temperature to be 8-9 °C lower than the true temperature (Kele et al., 2011). Most of the samples are composed of calcite. Some samples, like those from ^ukurbag and Kocaba? (Table 3), however, contain significant up to 100% aragonite (Table 3). Zhou and Zheng (2003), Kim and O'Neil (2005), and Kim et al. (2007) showed that the difference between calcite-water and aragonite-water frac-tionation can produce a bias of up to 5 °C (Kele et al., 2008). Herein, palaeotemperature calculations (Table 5) were based on the equilibrium equations of Friedman and O'Neil (1977) and Kim and O'Neil (1997) and modified using the observations of Kele et al. (2011). For these the calculations, we used S18Owater values of recent springs that are located closest to the travertine body and it was assumed that the oxygen isotope composition of the palaeosprings was similar to those of the current springs.

At Pamukkale, it was possible to test the reliability of the equilibrium equations (Eqs. (1), (2)) for palaeotemperature calculations, because there are recent deposits and the parent waters allow comparison of the measured and calculated temperatures of deposition (Kele et al., 2011, their Fig. 16). These samples showed that the use of the Friedman and O'Neil (1977) and Kim and O'Neil (1997) equilibrium equations produced underestimations of 4.2-7.1 °C compared to the real (observed) temperature (Table 5). This demonstrates that precipitation of the calcite was not in isotopic equilibrium with the spring waters.

The highest calculated temperatures came from Golemezli (51-73 °C), ^ukurbag (37-69 °C) and Kocaba? (47-56 °C), whereas medium temperatures came from Akkoy fissure ridge (26-45 °C) and Ballik (23-39 °C), and the lowest temperatures came from Honaz (22-30 °C) and Kelkaya (22-26 °C). These temperature ranges are based on all 818Otravertine values because it is generally impossible to precisely locate the orifice of palaeospring systems. For the fissure ridge systems, samples from the central part of the ridge should be located closest to the orifice and should provide the highest palaeotemperature data. In other words, samples providing the highest calculated palaeotemperatures (Table 5) could have been precipitated from the warmest waters (i.e., closest the orifice of the palaeospring).

Using the U-Th age of the travertines and comparing the temperature of the current springs with the calculated temperature values of the palaeosprings a slight overall decrease in the average temperature of the hydrothermal system seems to have taken placed between the Pleistocene and the Holocene. It should be noted, however, that the 818Owater values of the palaeosprings could have been 3% lower than the recent ones during glacial periods. If so, the calculated temperatures would be high.

5.7. Origin of the banded travertines

The banded (or vein) travertines, one of the most important components of the travertines in the Denizli Basin, typically fill spaces

Fig. 11. Hand specimen photo and SEM images of vertically banded travertine sample (^B-11a) from ^ukurbag fissure ridge, Pamukkale. (A) White and brown layers on cut surface formed from calcite and aragonite in different ratio, confirmed by XRD measurements. (B) Alternation of fibrous aragonite (ar) and rhombohedral calcite (ca) layers. Close views of (C) fibrous aragonite and (D) rhombohedral calcite crystals.

along the normal fault planes and fissure spaces that are found in the travertine ridges (Altunel and Hancock, 1993a,b; Hancock et al., 1999; Özkul et al., 2002). The best examples are found in the ^ukurbag and Akköy fissure ridges around Pamukkale, Gölemezli and Yenice to the north-northwest and at Kocaba? to the east. In some cases, they are found as veins that cut across bedded travertines and bedrocks (Rihs et al., 2000; Uysal et al., 2009). They are composed of calcite and/or aragonite.

The vein travertines with high Sr content (up to 6822 ppm, Table 4) are typically found closest to linear spring discharges along the fissures. The high Mg content (4560 to 8160 ppm) of the Gölemezli samples (Table 4) is probably related to the detrital dolomite that came from the underlying dolomitic bedrock. The ö13C values of the vein travertines, compiled from this study and others (Uysal et al., 2007; De Filippis et al., 2012), are highly positive ( + 3.7 to +5.8%) (Fig. 15A, B; Supplementary Table 1) and display a uniform distribution for each locality as in Gölemezli, ^ukurbag, Akköy and Kocaba? where the S18O values range from +13.8 to 19.5% (Fig. 15B). Similar ranges of values were obtained from vein deposits at Pamukkale and in the Gediz graben (Uysal et al., 2007, their Fig. 2). The positive ö13C values of the vein travertines have been attributed to a thermogene origin (Uysal et al., 2007, 2009). Data obtained in this study indicate that many of the other lithotypes are also of thermogene origin (Table 4, Fig. 15A, B). According

to Uysal et al. (2007), the banded travertines may have been generated by rapid precipitation from meteoric waters that were enriched in CO2 of deep origin and mobilised during intense seismic activity.

5.8. Geochronology of the travertine occurrences and possible influence of palaeoclimate on precipitation

Although the number of qualitative and quantitative data for the ages of the travertines in the Denizli Basin has increased (Altunel, 1996; Ozkul et al., 2004; Altunel and Karacabak, 2005; Erten et al., 2005; Uysal et al., 2007; Kappelman et al., 2008; Uysal et al., 2009; De Filippis et al., 2012), the age of the earliest travertine deposition and the complete chronology of different travertine occurrences remain debatable. Nevertheless, compilation all the age data from this study (Table 1) and previous works allows comparison between development of the springs in the Denizli with global and regional palaeoclimate trends (Fig. 16, Supplementary Table 2).

The travertine deposits are located at elevations ranging from 1092 m north of Ballik (site 3) to 370 m near Kocaba? (site 5). A travertine sample from the elevation of-1000 m at Ballik gave a thermoluminescence (TL) ageof510 ± 50 ka (Ozkul etal., 2004), whereas one from a lower elevation (-370 m) near Gurlek (site 5a) yielded an U-Th age of 108.7 ± 2.0 ka. Likewise, an U-Th age of the travertine sample OT-3

Fig. 12. (A) Front surface of rimstone dam formed from crystalline calcite, proximal slope in front of Jandarma spring, Pamukkale. Length of the hammer: 32 cm. (B) Microterrace pools developed on the proximal slope in front of Jandarma spring, Pamukkale. (C) Hand specimen of modern crystalline crust travertine developed perpendicular to the deposi-tional surface (sample PK-12). (D) SEM image from sample PK-12 shown in panel C. The calcite dendrites are in growth position.

from Obruk hill (510 m asl), near Karateke (site 6a) is 544 ± 111 ka. These ages indicate that the travertine deposition becomes younger from the graben's margins north and south towards the centre. It should be noted, however, that in some localities (e.g., Pamukkale) travertine deposition was active during many different time intervals (Fig. 16).

There is only limited number of palaeoclimate records for the Late Quaternary in Turkey (Nicoll and Kti^tikuysal, 2013). Thus, the high resolution records derived from the marine benthic record (Lisiecki and Raymo, 2005), the speleothems from the eastern Mediterranean region and China (Bar-Matthews et al., 1999, 2003; Wang et al., 2001, 2008), and the combined records of Peqiin and Soreq speleothems in Israel (Bar-Matthews et al., 2003) have been used in an effort to decipher the relationship between travertine deposition and palaeoclimate in the study area.

In this study, fifteen travertine samples have been U-Th dated (Table 1). The oldest age of 613 ka came from Golemezli (site 1), whereas the youngest age came from Akkoy ridge. Another vein sample from Golemezli yielded an age of 345 ± 18 ka, which corresponds to marine isotope stage (MIS) 10. The high error margin associated with this sample, however, means that it could partly overlap into MIS 9 (Fig. 16).

Fifty-eight age dates are available for Pamukkale (site 2) and its associated subsites (e.g., ^ukurbag and Akkoy) with those greater than 400 ka (Fig. 16, Supplementary Table 2) being obtained from some of the banded travertines and 'eroded sheet travertines' (Altunel, 1994, 1996). The U-Th dates spanning the 18 to 25 ka time interval match dry and cold period of MIS 2 (Uysal et al., 2009; De Filippis et al., 2012). This time period is characterized by 18O-enriched values in

cave records from the eastern Mediterranean region (Bar-Matthews et al., 1999, 2003; Fleitmann et al., 2009) and mainland China (Wang et al., 2001, 2008).

Eight samples with ages between 27.5 and 58.31 ka coincide with the younger 618O part of MIS 3. One bedded travertine sample from ^ukurbag with a U-Th age of 60 ± 1 ka (De Filippis et al., 2012) corresponds to the lower boundary of MIS 4.

Nineteen banded/vein samples from the fissure ridges at Pamukkale (Altunel and Karacabak, 2005; Uysal et al., 2007) yielded ages between 72.5 ± 0.9 and 124.3 ± 15 ka (Fig. 16, Supplementary Table 2). This group matches with MIS 5 as recorded from the Peqiin Cave stalagmite, north Israel (Bar-Matthews et al., 2003). In addition, four ages (e.g., 108.7 ± 2, 114 ± 10, 128.8 ± 3.8, 232 ± 27 ka) from the bedded travertines atGtirlek (site 5a) near Kocaba? (Table 1) indicate that precipitation took place during MIS 5 and MIS 7. This case implied that some veins at the Kocaba? area formed after surface travertine deposition had ceased at Gtirlek. According to data presently available, travertine deposition in the Denizli Basin took place during glacial and interglacial palaeoclimate periods (Fig. 16). These data indicate that travertine deposition during the late Quaternary was not strongly influenced by climatic variations.

6. Discussion

The depositional architecture and geochemical characteristics of the travertine deposits are controlled by the balance between extrinsic (e.g., tectonics, seasonal climatic variations) and intrinsic (e.g., composition and flow pattern of spring water) factors (Jones and Peng, 2012).

Fig. 13. Aragonite ooids from Qjkurbag, Pamukkale. (A) Hand specimen of the ooids, precipitated from bubbling mesothermal waters of 58 °C in ^ukurbag spring orifice. (B) SEM image of concentrically laminated ooids in different size. Some grains, ranging in shape from spherical through irregularly rounded, have been subjected to plastic deformation. (C) Thin section image of a single ooid displaying regular concentric lamination, formed in bubbling water. (D) Close view of concentric laminae. (E) Concentric laminae, up to 12 |jm thick, composed of aragonite needles. (F) Close view of aragonite needles with sizes of 2-3 |m.

These factors, which operate over all scales and vary through time, are collectively responsible for the tufa and travertine that form around spring vents (Guo and Riding, 1998; Minissale et al., 2002; Jones and Renaut, 2010). To separate the effect that each of these factors may have had on the development of ancient travertines is, however, difficult. One possible way of doing this is to compare different travertine deposits that occur in a single basin like the Denizli extensional basin (Fig. 1).

The Denizli Basin, which is one of the main geothermal provinces in Turkey (Ekmek^i et al., 1995; §im?ek, 2003; Mutlu et al., 2008), is still seismically active today (Ates and Bayulke, 1982; Altunel and Barka, 1996). In this region, the upper crust is cut by numerous water saturated cracks and fluid pressure is high (Kaypak and Gokkaya, 2012). High mountains to the north and south are the main recharge areas for the hydrothermal system (Ozler, 2000) and the deep- and shallow sourced waters in the basin are mainly of

meteoric origin (Table 2). As the thermal waters ascend to the surface, they mix with the shallow cool groundwater (Dilsiz, 2006; Crossey et al., 2009). The CO2 involved in travertine precipitation comes from thermometamorphic processes associated with magmatic sources in the area (e.g., Kele et al., 2011).

Thermal waters (Group I, II) that come to the surface along the northern boundaries (e.g., Karahayit, Pamukkale and ^ukurbag) are derived from deeply sourced waters, i.e. endogenic waters mixed with meteoric water in different ratios (Crossey et al., 2006; Forizs et al., 2011), and have high water temperature, high amounts of free CO2, and high saturation levels (Table 2). The waters in Karahayit, ^ukurbag and Kelkaya springs are below saturation level with respect to calcite (Table 2) but become saturated as CO2 degasses from the water while it flows downslope (Kele et al., 2011). In this part of the basin, travertines precipitated from hot waters of Group I, II and

Fig. 14. Thin section and SEM images of the vein travertine (sample GL-20) from Golemezli. (A) Thin calcite laminae with thickness of 1.0-1.5 mm below the sandy layer formed mainly of detrital dolomite, quartz, and mica. (B) Enlarged thin section image from the sandy field in A including detrital dolomite (d) arrowed. (C) Moulds appeared after removing a sand size grain. (D) A close view from one of the moulds in C.

fossil equivalents (i.e., Karahayit and Pamukkale) have higher 613C and Sr values than the others (Tables 2, 4), and are composed mostly of calcite with only small amount of aragonite locally (Table 3).

The spring waters in Group III and IV (Table 2) to the east, southeast, and south (i.e., Kaklik cave, Kelkaya and Honaz), with low temperature (~19-24 °C) and high discharge rates, are characterized by groundwater circulation and mixing at shallow depths (Gokgoz, 1998; Horvatincic et al., 2005; Dilsiz, 2006). Tufa and travertine precipitates that formed during different time periods are found along the southern boundary at the western end of the Honaz fault (Table 1). These variations in travertine and tufa precipitation may reflect temporal variations in water supply, water chemistry, mixture of deeply derived endogenic waters and epigenic waters, CO2 levels, tectonic/seismic activity and/or climatic controls (Jones and Peng, 2012).

In extensional and transtensional provinces, faults and associated fissures served as natural conduits for emerging thermal waters (Altunel and Hancock, 1993a; Hancock et al., 1999; §im?ek, 2003; Dilsiz, 2006; Brogi and Capezzuoli, 2009; De Filippis et al., 2012). Consequently, tectonic activity can significantly influence the deposition-al architecture of travertine precipitation at regional and local scales. At a local scale, various depositional morphologies exist (e.g., slope, waterfall, depression fill, fissure ridge, channel) during a particular period of time. Along boundary faults, the thermal springs emerge directly onto the slopes and lead to the formation of smooth and terraced slope deposits like those seen at Pamukkale and Kelkaya (Ozkul et al., 2002; Kele et al., 2011). In these settings, dendritic calcite is commonly precipitated as the CO2-rich spring water rapidly degases during their flow downslope (Kele et al., 2011; Jones and Peng, 2012). Similar modern and fossil examples of slope depositional systems have been reported from Mammoth Hot Spring in Yellowstone

National Park, Wyoming (Chafetz and Folk, 1984; Pentecost, 1990; Fouke et al., 2000), Rapolano Terme, central Italy (Guo and Riding, 1998), and the Denizli Basin (Ozkul et al., 2002; Kele et al., 2011).

Warm springs, like those at Ballik (site 3), which produced the largest travertine deposit over a volume of ~0.94 km3, are characterized by horizontally bedded travertines that were precipitated in depressions and/or large pools.

Similarly, warm spring waters issuing into depressions resulted in laterally extensive deposits, like those found at Bagni di Tivoli, east of Rome (Chafetz and Folk, 1984; Faccenna et al., 2008) and Rapolano Terme (Guo and Riding, 1998). These deposits have been described as 'shallow lake-fill travertines' (Chafetz and Folk, 1984). In those deposits, steepening (evolving from depression to slope facies) and/or levelling upward (evolving from slope to depression or from mound to depression depositional systems) are possible (Guo and Riding, 1998).

At Ballik, the travertine sequence gradually evolved from a depression depositional system, represented by horizontally bedded travertines (Ozkul et al., 2002), to a slope depositional system that was characterized by low angle smooth slope facies and waterfall or cascade facies (cf., Guo and Riding, 1998). The depression system that evolved upwards into slope deposition (cf., Guo and Riding, 1998) has been attributed to increased flow rates that caused an increase in deposition that promoted transformation of the area into a slope system (Jones and Renaut, 2010).

Fissure ridges are elongate, wedge like structures that formed as travertine was precipitated from hot waters that ascended along a fracture or fault plane (Altunel and Hancock, 1993a; Guo and Riding, 1999; Hancock et al., 1999; Atabey, 2002; Pentecost, 2005; Brogi and Capezzuoli, 2009; Selim and Yanik, 2009; De Filippis et al., 2012). The hydrostatic pressure needed to form the largest mounds

Fig. 15. (A) Basin scale distribution of stable carbon and oxygen isotope compositions of travertine samples in the Denizli Basin. (B) Stable carbon and oxygen isotope compositions of vein travertines. See Supplementary Table 1 for complete list of data. Data from this study (Table 4) are from Uysal et al. (2007) and De Filippis et al. (2012).

is considerable, approaching 7 kg cm-2 at ground level. These pressures, however, can be realised in artesian systems. Mound height must be limited by hydrostatic head. High supersaturation levels lead to rapid deposition around the vent and development of a steep mound (Pentecost, 2005).

The fissure ridges in the Denizli Basin are typically found at the ends of normal fault segments, which are step over or releasing zones, comprising a fracture network that are natural pathways for thermal waters (^akir, 1999; Ozkul et al., 2002; Altunel and Karabacak, 2005). These configurations are common along the northern boundary (e.g., Yenice and Pamukkale) and the Kocaba? area to the east (Fig. 1). In these areas, the physicochemical parameters like temperature, pCO2, saturation level with respect to CaCO3 of the present springs, and Sr are higher than those at other sites (Table 2). Maximum travertine deposition takes place along the central axis, possibly because of rapid CO2 degassing that maintains the high levels of supersaturation with respect to CaCO3. A hot spring, located at the eastern end of the ^ukurbag fissure ridge near Pamukkale (Fig. 4), for example, has the highest temperature, saturation levels, and Sr of all springs that were examined (Table 2). The formation of the banded travertines, however, is open to debate. Uysal et al. (2009) suggested that banded travertine formed during dry, cold periods and De Filippos et al. (2012) argued that they may have been formed from deep geothermal waters that had been released during seismic activity. Data compiled during this study, however, show that formation of the banded travertines cannot be attributed to specific climate conditions because they developed equally under cold, dry conditions and warm, wet conditions (Fig. 16). Some other

studies have also shown that vein travertines formed during warm periods, such as those that existed during MIS 1, MIS 5, and MIS 7 (Rihs et al., 2000; Kampman et al., 2012). In summary, the available data shows that travertine deposition in the Denizli Basin was probably caused by episodic fluxes in the deeply derived CO2 that was related primarily to seismicity as a consequence of neotectonic activity (Crossey et al., 2011).

The highest calculated palaeotemperatures came from deposits found in the northern (e.g., Gölemezli, ^ukurbag) and eastern (around the Kocaba? fissure ridges) areas, whereas medial temperatures came from the Akköy fissure ridge and Ballik sites, and the lowest temperatures came from Honaz and Kelkaya sites in the south (Table 5). Overall, there appears to have been a slight decrease in the average temperature of the hydrothermal system from the Pleistocene to the Holocene. Palaeotemperature values show a similar distribution with respect to those of some other parameters (i.e., stable isotope, pCO2, Sr, saturation level).

7. Conclusions

Quaternary travertine deposits of the Denizli extensional basin have been studied and compared in six locations with the view of determining the extrinsic and intrinsic factors that influenced their genesis. Travertine deposits in these locations, with thickness from 30 to 75 m, each cover an area between 1 and 34 km2. Up to 1 km3 of travertine volume is present in the largest deposit found at Ballik.

Today, spring waters found in the northern part of the Denizli Basin have the highest temperatures, electrical conductivity, dissolved CO2, Sr, and CaCO3 saturation levels. Collectively, the data indicate that these waters probably have deeper flow paths than elsewhere in the basin.

The travertines were precipitated in various depositional settings, including fissure ridges, slope systems, and depression systems. The fissure ridges, one of the main depositional morphologies, are restricted largely to the northern boundary. Present day travertine deposition continues on the terraced and smooth slopes at Pamukkale. The depression (or lake-fill) travertines are widespread around Ballik in the northeast. In this area, the depositional system evolved progressively from a horizontal depression, to low angle slope travertines, and finally to mound and waterfall systems at the top.

Although calcite is the dominant mineral at all sites, aragonite is present in some of the vertical banded, crystalline crust, raft and pisoid travertines from the ^ukurbag site (Pamukkale) in the north. The aragonite bearing samples rich in Sr are found mostly around the spring orifices. There are considerable variations in the stable isotope compositions of travertines throughout Denizli Basin. The ö13C values of travertines found along the northern boundary are more positive than elsewhere (up to + 12% PDB). Vein travertines precipitated in the fissure spaces and fractures of bedrocks, with a narrow range in ö13C values from + 3.7 to + 5.8% (PDB) for each locality, display more uniform conditions during the precipitation. In contrast, the S18O values have a wide range.

The travertines have been precipitated in warm and wet periods as well as in cold and dry periods. The spring activity and therefore travertine precipitation in the Denizli Basin is not related to climate and appears to be largely a function of tectonic activity. Palaeotemperature calculations, similar to the present case, show higher values for the spring waters at the northern locations. However, from Pleistocene to Holocene a slight overall decreasing has been recorded in the spring temperatures.

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.sedgeo.2013.05.018.

Acknowledgements

This work was supported by the Scientific and Technological Research Council ofTurkey (TUBITAK research grant of QAYDAG 106Y207) and the

Comprehensive table of palaeotemperature calculations for Denizli travertines including the stable oxygen isotope compositions of travertines and thermal springs.

Travertine Water T of recent

Site name 8180 Й180 Calc.1 Calc.2 Calc.3 Calc.4 Springs Traventine type Aragonite (Ж) (AGE) (ka)

(SMOW) (SMOW) T(°Q T(°C) T(°C) T(°C) T(°C)

Golemezli 14.5 -7.7 60 57 68 50 54,8 Vertically banded 611.230±°°

(1) 13.8 -7.7 65 62 73 54 Vertically banded

17.4 -7.7 43 41 51 35 Vertically banded

15,2 -7.7 56 53 64 46 Vertically banded

15.5 -7.7 54 52 62 44 Vertically banded

15.4 -7.7 54 52 62 45 Vertically banded

16.6 -7.7 47 45 55 39 Bedded

17.3 -7.7 43 45 51 35 Bedded

15.1 -7.7 56 54 64 46 Raft

Cukurbag 14.8 -8.3 54 52 62 45 58 Crytalline crust 100

(2) 16.5 -8.3 44 43 52 36 Raft 5

17.2 -8.3 40 39 48 33 Raft, coated gas bubble 78

15.2 -8.3 52 50 60 43 Ooid

14.6 -8.3 56 53 64 46 Vertically banded 7-23

13.8 -8.3 61 58 69 50 Vertically banded 25.321+308

19.5 -8.3 29 27 37 23 Vertically banded 70

Akkoy 18.1 -B.l 37 35 45 30 53 Raft

(3) 20 -B.1 27 26 35 22 Brown

21.5 -B.l 20 19 28 16 Bedded

21.4 -8.1 21 19 29 16 Raft and micritic layer

19.6 -8.1 29 28 37 23 Bedded

20.3 -8.1 26 24 34 20 Bedded 18.188±314

19.8 -8.1 28 27 36 22 bedded

20.9 -8.1 23 22 31 18 bedded

19.7 -8.1 29 27 37 23 Taft and micritic layer

22.1 -8.1 18 16 26 13 Bedded

Ballik 22 -8.1 18 16 26 14 22.8 Rimstone, Kaklik cave

(4) 21.8 -8.1 19 17 27 14 Dark coloured trav.

22.1 -8.1 18 16 26 13 Reddish-brown coloured

22.8 -8.1 15 13 23 11 Light coloured bedded

21.8 -8.1 19 17 27 14 Light coloured bedded

19.5 -8.1 30 28 38 24 Light coloured bedded

19.2 -8.1 31 30 39 25 Light coloured bedded

21 -8.1 23 21 31 17 Light coloured bedded

21 -8.1 23 21 31 17 Light coloured bedded

21.9 -8,1 19 17 27 14 Light coloured bedded

20.9 -8.1 23 22 31 18 Light coloured bedded

Kocabas 17 -8.8 39 37 47 31 23.7 Vertically banded 75 90.463+2645

(5) 15.3 -8.8 48 46 56 40 Vertically banded 7 144.908+6287

16.8 -8.8 40 38 48 32 Vertically banded

16.2 -8.8 43 42 51 35 Vertically banded

Kelkaya 22.4 -8,5 15 13 23 10 19.6 Bedded

(6) 22.6 -8.5 14 12 22 10 Bedded

21.7 -8,5 18 16 26 13 Bedded

Honaz 20.3 -8.8 22 21 30 17 20.2:24; 18.8 Karateke, passive tufa

(7) 21 -8.8 19 18 27 15 Karateke, passive tufa

20.9 -8.8 20 18 28 15 Karateke, Honaz

21.2 -8.8 18 17 26 14 Karateke, l;ik Tr. Quarry

21.2 -8.8 18 17 26 14 Degirmenler, recent tufa

21.4 -8.8 18 16 26 13 recent waterfall+moss

20.5 -8.8 21 20 29 16 Soft passive tufa

22.3 ¿o CO 14 12 22 10 Soft passive tufa

Pamukkale 19.9 -8.8 24 23 32 ,30.1 Recent crystalline crust recent

Pamukkale 19.3 -8.5 29 27 37 33.2 Recent crystalline crust recent

For the calculations the equilibrium equations of Friedman and O'Neil (1977) and Kim and O'Neil (1997) were used and was modified based from Kele et al. (2011). Temperature of recent thermal springs is also indicated for comparison, together with the travertine types, aragonite content and radiometric age of the deposits. Calculated maximum temperature values are highlighted with grey background. For the calculations the 81sO values of the following recent springs (data from Table 2) were used: (1) Golemezli ¡janhalp, (2) ^ukurbag spring, (3) Karahayit Belediyesi well, (4) Kaklikcave, (5) Pmarba$i warm well, (6) Kelkaya spring, (7) Pmarba$i spring, (8) Pamukkale, Jandarma spring, (9) Pamukkale, Beltes-2 spring. Calculated maximum temperature values are highlighted with grey background. Calc.1 is based on Friedman and O'Neil (1977). Calc.2 is based on Kim and O'Neil (1997). Calc.3 is modified based from Kele et al. (2011).

Fig. 16. Comparison of U-Th ages of the travertine deposits from the Denizli Basin with regional and global palaeoclimate records. High S1sO values are from stalagmites from Hulu and Soreq caves (Wang et al., 2001; Bar-Matthews et al., 2003). Age data from this study are combined with those from previous works (Altunel, 1996; Özkul et al., 2004; Altunel and Karabacak, 2005; Uysal etal., 2007,2009; De Filippis et al., 2012). Global glacial periods (white bars and even numbers) and interglacial periods (shaded bars and odd numbers) are shown by marine isotope stages (MISs) which are from Lisiecki and Raymo (2005).

National Office for Research and Technology of Hungary (NKTH, Hungary, project number: TR-10/2006). U-Th dating at the HISPEC was supported by the NSC grants (NSC99-2628-M-002-012, 100-2116-M-002-009, and 100-2116-M-002-009 to CCS). We are grateful to Dr. L.J. Crossey, an anonymous reviewer, and Dr. J. Knight (Editor) who critically commented on an earlier version of this manuscript.

References

Alçiçelk H., Özkul, M., Varol, B., 2003. Elementary sulphur formation in Kizilyer evapo-rites and fissure fill gypsum (Denizli, SW Anatolia). 14th International Petroleum and Natural Gas Congress and Exhibition of Turkey, Proceedings, pp. 86-94.

Alçiçelk H., Varol, B., Özkul, M., 2007. Sedimentary facies, depositional environments and palaeogeographic evolution of the Neogene Denizli Basin of SW Anatolia, Turkey. Sedimentary Geology 202, 596-637.

Altunel, E., 1994. Active Tectonics and the Evolution of Quaternary Travertines at Pamukkale, Western Turkey. (Unpublished Ph.D. Thesis) University of Bristol, United Kingdom (236 pp.).

Altunel, E., 1996. Pamukkale travertenlerinin morfolojik ozellikleri, yaçlan ve neotektonik onemleri. Maden Tetkik ve Arama Dergisi 118,47-64 (in Turkish).

Altunel, E., Barka, A., 1996. Evaluation of archaeoseismic damages at Hierapolis. Geological Bulletin of Turkey 39, 65-74 (in Turkish).

Altunel, E., Hancock P.L., 1993a. Morphology and structural setting of Quaternary travertines at Pamukkale, Turkey. Geological Journal 28,335-346.

Altunel, E., Hancock P.L., 1993b. Active Assuring, faulting and travertine deposition at Pamukkale, western Turkey. In: Stewart, I.S., Vita-Finzi, C., Owen, LA. (Eds.), Neotectonics and Active Faulting: Zeitschrift für Geomorphologie, Supplement, 94, pp. 285-302.

Altunel, E., Karacabak, V., 2005. Determination of horizontal extension from fissure-ridge travertines: a case study from the Denizli Basin, southwestern Turkey. Geodinamica Acta 18, 333-342.

Andrews, J.E., 2006. Paleoclimatic record from stable isotopes in riverine tufas: synthesis and review. Earth-Science Reviews 75,85-104.

Andrews, J.E., Riding, R., Dennis, P.F., 1997. The stable isotope record of environmental and climatic signals in modern terrestrial microbial carbonates from Europe. Palaeogeography, Palaeoclimatology, Palaeoecology 129,171-189.

Arenas, C., Gutiérrez, F., Osácar, C., Sancho, C., 2000. Sedimentology and geochemistry of fluvio-lacustrine tufa deposits controlled by evaporite solution subsidence in the central Ebro Depression, NE Spain. Sedimentology 47, 883-909.

Arenas-Abad, C., Vázquez-Urbez, M., Pardo-Tirapu, G., Sancho-Marcén, C., 2010. Fluvial and associated carbonate deposits. In: Alonso Zarza, A.M., Tanner, L.H. (Eds.), Carbonates in Continental Settings: Facies, Environments, and Processes: Development in Sedimentology, 61, pp. 133-175.

Atabey, E., 2002. Çatlak Sirt Tipi Laminali Traverten-Tufa Çokellerinin Oluçumu, Mikroskobik Ozellikleri ve Diyajenezi, Kirçehir, lç Anadolu. Maden Tetkik ve Arama Dergisi 123-124, 59-65 (in Turkish).

Ateç, R.C., Bayülke, N., 1982. The 19 August 1976 Denizli, Turkey, earthquake: evaluation of the strong motion accelerograph record. Bulletin of the Seismological Society of America 72,1635-1649.

Aydan, O., Kumsar, H., Tano, H., 2005. Multiparameter changes in the earth's crust and their relation to earthquakes in Denizli region of Turkey. In: Row, B., Horne, R., Stacey, R., Juliusson, E., Villaluz, A., Chen, C.-Y., Dastan, A., Li, K., Garner, L., Bolds, J., Polyakova, J. (Eds.), Proceedings World Geothermal Congress, Antalya, Turkey, 24-29 April 2005, pp. 1-10 (on CD).

Bárdossy, Gy, Bottyán, L., Gadó, P., Griger, Á., Sasvári, J., 1980. Automated quantitative phase analysis of bauxites. American Mineralogist 65,135-141.

Bar-Matthews, M., Ayalon, A., Kaufman, A., Wasserburg, G.J., 1999. The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq Cave, Israel. Earth and Planetary Science Letters 166, 85-95.

Bar-Matthews, M., Ayalon, A., Gilmour, M., Matthews, A., Hawkesworth, C.J., 2003. Sea-land isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implication for palaeorainfall during interglacial intervals. Geochimica et Cosmochimica Acta 67, 3181-3199.

Barnes, I., O'Neil, J.R., 1971. Calcium-magnesium solid solutions from Holocene conglomerate cements and travertines in the Coast Range of California. Geochimica et Cosmochimica Acta 35, 699-717.

Bozkurt, E., Oberhânsli, R., 2001. Menderes massif (Western Turkey): structural, metamor-phic and magmatic evolution — a synthesis. International Journal of Earth Sciences 89, 679-708.

Bozkuç, C., Kumsar, H., Ozkul, M., Hançer, M., 2000. Seismicity of active Honaz fault under an extensional tectonic regime. In: Dora, O.O., Ozgenç, 1., Sozbilir, H. (Eds.), Proceedings of International Earth Science Colloquium on the Aegean Region. Dokuz Eylül University, Izmir, Turkey, pp. 7-16.

Brogi, A., Capezzuoli, E., 2009. Travertine deposition and faulting: the fault-related travertine fissure-ridges at Terme S. Giovanni, Terme, Italy. International Journal of Earth Sciences 98, 931-947.

Busenburg, E., Plummer, L., 1986. A comparative study of the dissolution and crystal growth kinetics of calcite and aragonite. United States Geological Survey Bulletin 1578,139-168.

Çakir, Z., 1999. Along-strike discontinuity of active normal faults and its influence on Quaternary travertine deposition: examples from western Turkey. Turkish Journal of Earth Sciences 8, 67-80.

Chafetz, H.S., Folk R.L., 1984. Travertines: depositional morphology and the bacterially constructed constituents. Journal of Sedimentary Petrology 54, 289-316.

Chafetz, H.S., Lawrence, J.R., 1994. Stable isotopic variability within modern travertines. Géographie Physique et Quaternaire 48, 257-273.

Chafetz, H.S., Rush, P.F., Utech, N.M., 1991. Microenvironmental controls on mineralogy and habit of CaCO3 precipitates: an example from an active travertine system. Sedimentology 38,107-126.

Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmersom, Y., 2000. The half-lives of uranium-234 and thorium-230. Chemical Geology 169,17-33.

Coplen, T.B., 2007. Calibration of the calcite-water oxygen-isotope geothermometer at Devils Hole, Nevada, a natural laboratory. Geochimica et Cosmochimica Acta 71,3948-3957.

Crossey, L.J., Fischer, T.P., Patchett, P.J., Karlstrom, K.E., Hilton, D.R., Huntoon, P., Reynolds, A.C., 2006. Dissected hydrologic system at Grand Canyon: interaction between upper world and lower world waters in modern springs and travertine. Geology 34, 25-28.

Crossey, L.J., Karlstrom, K.E., Springer, A., Newell, D., Hilton, D., Fischer, T., 2009. Degassing of mantle-derived CO2 and 3He from springs in the southern Colorado Plateau region-neotectonic connections and implications for groundwater systems. Geological Society of America Bulletin 121,1034-1053.

Crossey, L.J., Karlstrom, K.E., Newell, D.L., Kooser, A., Tafoya, A., 2011. The La Madera Travertines, Rio Ojo Caliente, Northern New Mexico: investigating the linked system of CO2-rich springs and travertines as neotectonic and paleoclimate indicators.

New Mexico Geological Society Guidebook, 62nd Field Conference, Geology of the Tusas Mountains — Ojo Caliente, pp. 121-136.

De Filippis, L., Faccenna, C., Billi, A., Anzalone, E., Brilli, M., Ozkul, M., Soligo, M., Tuccimei, P., Villa, M., 2012. Growth of fissure ridge travertines from geothermal springs of Denizli Basin, western Turkey. Bulletin of the Geological Society of America 124,1629-1645.

Demény, A., Kele, S., Siklosy, Z., 2010. Empirical equations for the temperature dependence of calcite-water oxygen isotope fractionation from 10 to 70 °C. Rapid Communications in Mass Spectrometry 24,3521-3526.

Dilsiz, C., 2006. Conceptual hydrodynamic model of the Pamukkale hydrothermal field, southwestern Turkey, based on hydrochemical and isotopic data. Hydrogeology Journal 14, 562-572.

Ekmekçi, M., Gunay, G., Çimçek, §., 1995. Morphology of rimstone pools, Pamukkale, western Turkey. Cave Karst Science 22,103-106.

Erdogan, B., Gungôr, T., 2004. The problem of the core-cover boundary of the Menderes massif and an emplacement mechanism for regionally extensive gneissic granites, western Anatolia (Turkey). Turkish Journal of Earth Sciences 13,15-36.

Erten, H., Sen, §., Ozkul, M., 2005. Pleistocene mammals from travertine deposits of the Denizli basin (SW Turkey). Annales de Paléontologie 91, 267-278.

Faccenna, C., Soligo, M., Billi, A., De Filippis, L., Funiciello, R., Rossetti, C., Tuccimei, P., 2008. Late Pleistocene depositional cycles of the Lapis Tiburtinus travertine (Tivoli, Central Italy): possible influence of climate and fault activity. Global and Planetary Change 63, 299-308.

Filiz, S., 1984. Investigation of the important geothermal areas by using C, H, O isotopes. Seminar on the Utilization of Geothermal Energy for Electric Power Generation and Space Heating, 14-17 May 1984, Florence, Italy (Ref. No. EP/SEM.9/R.3).

Fleitmann, D., Cheng, H., Badertscher, S., Edwards, R.L., Mudelsee, M., Gokturk, O.M., Frankhauser, A., Pickering, R., Raible, C.C., Matter, A., Kramers, J., Tuysuz, O., 2009. Timing and climatic impact of Greenland interstadials recorded in stalagmites from northern Turkey. Geophysical Research Letters 36, L19707. http:// dx.doi.org/10.1029/2009GL040050.

Folk R.L., 1994. Interaction between bacteria, nannobacteria, and mineral precipitation in hot springs of Central Italy. Géographie Physique et Quaternaire 48, 233-246.

Ford, T.D., Pedley, H.M., 1996. A review of tufa and travertine deposits of the world. Earth-Science Reviews 41,117-175.

Forizs, I., Gôkgôz, A., Kele, S., Ozkul, M., Deâk, J., Baykara, M.O., Alçiçek, M.C., 2011. Comparison of the isotope hydrogeological features of thermal and cold karstic waters in the Denizli Basin (Turkey) and Buda Thermal Karst (Hungary). Central European Geology 54 (1-2), 115-119.

Fouke, B.W., Farmer, J.D., Des Marais, D.J., Pratt, L., Sturchio, N.C., Burns, P.C., Discipulo, M.K., 2000. Depositional facies and aqueous-solid geochemistry of travertine depositing hot springs (Angel Terrace, Mammoth Hot Springs, Yellowstone National Park, U.S.A.). Journal of Sedimentary Research 70, 565-585.

Friedman, I., O'Neil, J.R., 1977. Compilation of stable isotope fractionation factors of geochemical interest. United States Geological Survey KK1 -KK12.

Gokgoz, A., 1994. Pamukkale-Karahayit-Gôlemezli Hidrotermal Karstmm Hidrojeolojisi. (Unpublished Ph.D. thesis) Suleyman Demirel University, Isparta (263 pp.).

Gôkgôz, A., 1998. Geochemistry of the Kizildere-Tekkehamam-Buldan-Pamukkale geothermal fields. United Nations University, Geothermal Training Programme, Reports, Reykjavik, Iceland, pp. 115-156.

Gundogan, 1., Helvaci, C., Sôzbilir, H., 2008. Gypsiferous carbonates at Honaz Dagi (Denizli): first documentation of Triassic gypsum in western Turkey and its tectonic significance. Journal of Asian Earth Sciences 32, 49-65.

Guo, L., Riding, R., 1992. Aragonite laminae in hot water travertine crusts, Rapolano Terme, Italy. Sedimentology 39,1067-1079.

Guo, L., Riding, R., 1998. Hot-spring travertine facies and sequences, Late Pleistocene Rapolano Terme, Italy. Sedimentology 45,163-180.

Guo, L., Riding, R., 1999. Rapid facies changes in Holocene fissure ridge hot spring travertines, Rapolano Terme, Italy. Sedimentology 46,1145-1158.

Guo, L., Andrews, J., Riding, R., Dennis, P., Dresser, Q., 1996. Possible microbial effects on stable carbon isotopes in hot-spring travertines. Journal of Sedimentary Research 66,468-473.

Hancock P.L., Chalmers, R.M.L., Altunel, E., Çakir, Z., 1999. Travitonics: using travertines in active fault studies. Journal of Structural Geology 21, 903-916.

Hancock P.L, Chalmers, RM.L, Altunel, E., Çakir, Z., Becher-Hancock A., 2000. Creation and destruction of travertine monumental stone by earthquake faulting at Hierapolis, Turkey. In: McGuire, W.G., Griffiths, D.R., Hancock, P.L., Stewart, I.S. (Eds.), The Archaeology of Geological Catastrophes: The Geological Society, London. Special Publications, 171, pp. 1-14.

Horvatincic, N., Ozkul, M., Gôkgôz, A., Baresic, J., 2005. Isotopic and geochemical investigation of tufa in Denizli province, Turkey. In: Ozkul, M., Yagiz, S., Jones, B. (Eds.), Proceedings of 1st International Symposium on Travertine, Kozan Ofset Matbaacilik. San. ve Tic. Ltd. Std, Ankara, pp. 162-170.

Jaffey, A.H., Flynn, K.F., Glendenin, L.E., Bentley, W.C., Essling, A.M., 1971. Precision measurement ofhalf-lives and specific activities ofU-235 and U-238. Physical Reviews 4,1889-1906.

Jones, B., Kahle, C.F., 1986. Dendritic calcite crystals formed by calcification of algal filaments in a vadose environment. Journal of Sedimentary Petrology 56, 217-227.

Jones, B., Peng, X., 2012. Intrinsic versus extrinsic controls on the development of calcite den-drite bushes, Shuzhishi Spring, Rehai geothermal area, Tengchong, Yunnan Province, China. Sedimentary Geology 249-250,45-62.

Jones, B., Renaut, R.W., 1995. Noncrystallographic dendrites from hot-spring deposits at Lake Bogoria, Kenya. Journal of Sedimentary Research 65,154-169.

Jones, B., Renaut, R.W., 2008. Cyclic development of large, complex calcite dendrite crystals in the Clinton travertine, Interior British Columbia, Canada. Sedimentary Geology 203,17-35.

Jones, B., Renaut, R.W., 2010. Calcareous spring deposits in continental settings. In: Alonso Zarza, A.M., Taner, L.H. (Eds.), Carbonates in Continental Settings: Facies, Environments, and Processes.: Developments in Sedimentology, 61. Elsevier, pp. 177-224.

Jones, B., Renaut, R.W., Rosen, M.R., 1996. High-temperature (>90 °C) calcite precipitation at Waikite Hot Springs, North Island, New Zealand. Journal of the Geological Society 153, 481-496.

Kampman, N., Burnside, N.M., Shipton, Z.K., Chapman, H.J., Nicholl, J.A., Ellam, R.M., Bickle, M.J., 2012. Pulses of carbon dioxide emissions from intracrustal faults following climatic warming. Nature Geoscience 5, 352-358.

Kappelman, J., Al<icek M.C., Kazanci, N., Schultz, M., Ozkul, M., Sen, S., 2008. Brief communication: first Homo erectus from Turkey and implications for migrations into temperate Eurasia. American Journal of Physical Anthropology 135,110-116.

Karaku$, H., ¡jim^ek ¡¡¡., 2013. Tracing deep thermal water circulation systems in the E-W trending Buyuk Menderes Graben, western Turkey. Journal of Volcanology and Geothermal Research 252, 38-52.

Kaymak<i, N., 2006. Kinematic development and paleostress analysis of the Denizli Basin (Western Turkey): implications of spatial variation of relative paleostress magnitudes and orientations. Journal of Asian Earth Sciences 27, 207-222.

Kaypak, B., Gokkaya, G., 2012. 3-D imaging of the upper crust beneath the Denizli geothermal region by local earthquake tomography, western Turkey. Journal of Volcanology and Geothermal Research 211-212,47-60.

Kele, S., Demeny, A., Siklosy, Z., Nemeth, T., Maria, T.B., Kovacs, M., 2008. Chemical and stable isotope compositions of recent hot-water travertines and associated thermal waters, from Egerszalok, Hungary: depositional facies and non-equilibrium frac-tionations. Sedimentary Geology 211, 53-72.

Kele, S., Ozkul, M., Gokgoz, A., Forizs, I., Baykara, M.O., Al<i<ek M.C., Nemeth, T., 2011. Stable isotope geochemical and facies study of Pamukkale travertines: new evidences of low-temperature non-equilibrium calcite-water fractionation. Sedimentary Geology 238,191-212.

Kim, S.-T., O'Neil, J.R., 1997. Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochimica et Cosmochimica Acta 61, 3461 -3475.

Kim, S.-T., O'Neil, J.R., 2005. Comment on "An experimental study of oxygen isotope fractionation between inorganically precipitated aragonite and water at low temperatures" by G.-T. Zhou and Y.-F. Zheng. Geochimica et Cosmochimica Acta 69,3195-3197.

Kim, S.-T., Mucci, A., Taylor, B., 2007. Phosphoric acid fractionation factors for calcite and aragonite between 25 and 75 °C: revisited. Chemical Geology 246,135-146.

Kitano, Y., 1962. A study of polymorphic formation of calcium carbonate in thermal springs with emphasis on the effect of temperature. Bulletin of the Chemical Society of Japan 35,1980-1985.

Kitano, Y., 1963. Geochemistry of calcareous deposits found in hot springs. Journal of Earth Science, Nagoya University 11, 68-100.

Ko<yigit, A., 2005. The Denizli graben-horst system and the eastern limit of western Anatolian continental extension: basin-fill, structure, deformational mode, throw amount and episodic evolutionary history, SW Turkey. Geodinamica Acta 18, 167-208.

Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene-Pliestocene stack of 57 globally distributed benthic 818O records. Paleoceanography 20,1-17.

Malesani, P., Vanucchi, S., 1975. Precipitazione di calcite o di aragonite dalle acque termominerali in relazione alla genesi e all'evoluzione dei travertini: Accademia Lincell, Rendiconti Scienze fisica, matematica e naturale, 58, pp. 761-776 (in Italian).

McCrea, J.M., 1950. On the isotopic chemistry of carbonates and a paleotemperature scale. Journal of Chemical Physics 18, 849-857.

Minissale, A., Kerrick, D.M., Magro, G., Murrell, M.T., Paladini, M., Rihs, S., Sturchio, N.C., Tassi, F., Vaselli, O., 2002. Geochemistry of Quaternary travertines in the region north of Rome (Italy): structural, hydrologic and paleoclimatologic implications. Earth and Planetary Science Letters 203, 709-728.

Mutlu, H., Gulec, N., Hilton, D.R., 2008. Helium-carbon relationships in geothermal fluids of western Anatolia, Turkey. Chemical Geology 247, 305-321.

Nicoll, K., Ku<ukuysal, C., 2013. Emerging multi-proxy records of late Quaternary palaeoclimate dynamics in Turkey and the surrounding region. Turkish Journal of Earth Sciences 22,126-142.

Okay, A.I., 1989. Denizli'nin guneyinde Menderes masifi ve Likya naplarinin jeolojisi. Maden Tetkik ve Arama Dergisi 109,45-48 (in Turkish).

Ozkul, M., Varol, B., Al<i<ek, M.C., 2002. Depositional environments and petrography of the Denizli travertines. Bulletin of the Mineral Research and Exploration 125, 13-29.

Ozkul, M., Engin, B., Al<i<ek M.C., Koralay, T., Demirta$, H., 2004. Thermoluminescence dating of Quaternary hot spring travertines and some implications on graben evolution, Denizli, Western Turkey. 32nd International Geological Congress, August 20-28, 2004, Florence, Italy.

Ozkul, M., Gokgoz, A., Horvatincic, N., 2010. Depositional properties and geochemistry of Holocene perched springline tufa deposits and associated spring waters: a case study from the Denizli province, Western Turkey. In: Pedley, H.M. (Ed.), Tufas and Speleothems: Unravelling the Microbial and Physical Controls: The Geological Society, London. Special Publications, 336, pp. 245-262.

Ozler, H.M., 2000. Hydrogeology and geochemistry in the ^uruksu (Denizli) hydrothermal field, western Turkey. Environmental Geology 39,1169-1180.

Parkhust, D.L., 1995. User's guide to PHREEQC—a computer program for speciation, reaction path, advective transport, and inverse geochemical calculations. U.S. Geological Survey Water Resources Investigations Report 95-4227 (143 pp.).

Pedley, H.M., 2009. Tufas and travertines of the Mediterranean region: a testing ground for freshwater carbonate concepts and developments. Sedimentology 56,221-246.

Pedley, H.M., Ordonez, S., Gonzales-Martin, J.A., Garcia Del Cura, M.A., 2003. Sedimen-tology of Quaternary perched springline and paludal tufas: criteria for recognition, with examples from Guadalajara Province, Spain. Sedimentology 50, 23-44.

Pentecost, A., 1990. The formation of travertine shrubs: Mammoth Hot Springs, Wyoming. Geological Magazine 127,159-168.

Pentecost, A., 2005. Travertine. Springer Verlag (446 pp.).

Piccardi, L., 2007. The AD 60 Denizli Basin earthquake and the apparition of Archangel Michael at Colossae (Aegean Turkey). In: Piccardi, L., Masse, W.B. (Eds.), Geological Society, London, Special Publications, 273, pp. 95-105.

Renaut, R.W., Jones, B., 1997. Controls on aragonite and calcite precipitation in hot spring travertines at Chemurkeu, Lake Bogoria, Kenya. Canadian Journal of Earth Sciences 34, 801-818.

Richter, D.V., Besenecker, H., 1983. Subrecent high-Sr aragonite ooids from hot springs near Tekke Ilica (Turkey). In: Perty, T.M. (Ed.), Coated Grains. Springer-Verlag, Berlin, pp. 154-162.

Rihs, S., Condomines, M., Poidevin, J.L., 2000. Long-term behaviour of continental hydrothermal systems: U-series study of hydrothermal carbonates from the French Massif Central (Allier Valley). Geochimica et Cosmochimica Acta 64, 3189-3199.

Rodriguez-Berriguete, A., Alonso-Zarza, A.M., Cabrera, M.C., Rodriguez-Gonzalez, A., 2012. The Azuaje travertine: an example of aragonite deposition in a recent volcanic setting, N Gran Canaria Island, Spain. Sedimentary Geology 277-278,61-71.

Sant'Anna, L.G., Riccomini, C., Rodrigues-Francisco, B.H., Sial, A.N., Carvalho, M.D., Moura, C.A.V., 2004. The Paleocene travertine system of the Itaborai basin, Southeastern Brazil. Journal of South American Earth Sciences 18,11 -25.

Selim, H.H., Yanik, G., 2009. Development of the Cambazli (Turgutlu/MANISA) fissure-ridge-type travertine 744 and relationship with active tectonics, Gediz Graben, Turkey. Quaternary International 199,157-163.

Shen, C.-C., Edwards, R.L., Cheng, H., Dorale, J.A., Thomas, R.B., Moran, S.B., Weinstein, S.E., Hirschmann, M., 2002. Uranium and thorium isotopic and concentration measurements by magnetic sector inductively coupled plasma mass spectrometry. Chemical Geology 185,165-178.

Shen, C.-C., Cheng, H., Edwards, R.L., Moran, S.B., Edmonds, H.N., Hoff, J.A., Thomas, R.B., 2003. Measurement of attogram quantities of 231Pa in dissolved and particulate fractions of seawater by isotope dilution thermal ionization mass spectroscopy. Analytical Chemistry 75,1075-1079.

Shen, C.-C., Li, K.-S., Sieh, K., Natawidjaja, D., Cheng, H., Wang, X., Edwards, R.L., Lam, D.D., Hsieh, Y.-T., Fan, T.-Y., Meltzner, A.J., Taylor, F.W., Quinn, T.M., Chiang, H.W., Kilbourne, K.H., 2008. Variation of initial 230Th/232Th and limits of high precision U-Th dating of shallow-water corals. Geochimica et Cosmochimica Acta 72, 4201-4223.

Shen, C.-C., Wu, C.-C., Cheng, H., Edwards, R.L., Hsieh, Y.-T., Gallet, S., Chang, C.-C., Li, T.Y., Lam, D.D., Kano, A., Hori, M., Spötl, C., 2012. High-precision and high-resolution carbonate 230Th dating by MC-ICP-MS with SEM protocols. Geochimica et Cosmochimica Acta 99, 71 -86.

Sierralta, M., Kele, S., Melcher, F., Hambach, U., Reinders, J., van Geldern, R., Frechen, M., 2010. Uranium-series dating of travertine from Süttö: implications for reconstruction of environmental change in Hungary. Quaternary International 222,178-193.

§im$ek §., 2003. Hydrogeological and isotopic survey of geothermal fields in the Büyük Menderes graben, Turkey. Geothermics 32, 669-678.

§im$ek §., Günay, G., Elhatip, H., Ekmekci, M., 2000. Environmental protection of geo-thermal waters and travertines at Pamukkale, Turkey. Geothermics 29, 557-572.

Spötl, C., Vennemann, T.W., 2003. Continuous-flow isotope ratio mass spectrometric analysis of carbonate minerals. Rapid Communications in Mass Spectrometry 17, 1004-1006.

Sun, S., 1990. Denizli-U$ak arasmm jeolojisi ve linyit olanaklari (geology and lignite potential between Denizli and U$ak). Scientific report of the General Directorate of Mineral Research and Exploration of Turkey; No: 9985, Ankara, Turkey (in Turkish), p. 92.

Tan, O., Tapirdamaz, M.C., Yörük, A., 2008. The earthquake catalogues for Turkey. Turkish Journal of Earth Sciences 17,405-418.

Tremaine, D.M., Froelich, P.N., Wang, Y., 2011. Speleothem calcite formed in situ: modern calibration of 81sO and 813C paleoclimate proxies in a continuously-monitored natural cave system. Geochimica et Cosmochimica Acta 75, 4929-4950.

Utku, M., 2009. Etkinlik ve yigmsal etkinlik dönemlerine göre Denizli depremlerinin analizi. Maden Tetkik ve Arama Dergisi 138, 9-34 (in Turkish).

Uysal, I.T., Feng, Y., Zhao, J., Altunel, E., Weatherley, D., Karabacak, V., Cengiz, O., Golding, S.D., Lawrence, M.G., Collerson, K.D., 2007. U-series dating and geochem-ical tracing of late Quaternary travertine in co-seismic fissures. Earth and Planetary Science Letters 257, 450-462.

Uysal, I.T., Feng, Y., Zhao, J., I$ik, V., Nuriel, P., Golding, S.D., 2009. Hydrothermal CO2 degassing in seismically active zones during the late Quaternary. Chemical Geology 265, 442-454.

Van Noten, K., Claes, H., Soete, J., Foubert, A., Özkul, M., Swennen, R., 2013. Fracture networks and strike-slip deformation along reactivated normal faults in Quaternary travertine deposits, Denizli Basin, western Turkey. Tectonophysics 588,154-170.

Wang, Y.J., Cheng, H., Edwards, R.L., An, Z.S., Wu, J.Y., Shen, C.C., Dorale, J.A., 2001. A high-resolution absolute-dated Late Pleistocene monsoon record from Hulu cave, China. Science 294, 2345-2348.

Wang, Y., Cheng, H., Edwards, R.L., Kong, X., Shao, X., Chen, S., Wu, J., Jiang, X., Wang, X., An, Z., 2008. Millennial- and orbital-scale changes in the East Asian monsoon over the past 224,000 years. Nature 451,1090-1093.

Westaway, R., Guillou, H., Yurtmen, S., Demir, T., Scaillet, S., Rowbotham, G., 2005. Constraints on the timing and regional conditions at the start of the present phase of crustal extension in western Turkey, from observations in and around the Denizli region. Geodinamica Acta 18, 209-238.

Zhou, G.-T., Zheng, Y.-F., 2003. An experimental study of oxygen isotope fractionation between inorganically precipitated aragonite and water at low temperatures. Geochimica et Cosmochimica Acta 67, 387-399.