Characterization of building materials from the aqueduct of Antioch-on-the-Orontes (Turkey)

(1)

Internal

Geophysics

(Applied

Geophysics)

Characterization

of

building

materials

from

the

aqueduct

of

Antioch-on-the-Orontes

(Turkey)

Yacine

Benjelloun

a,

*

,

Julia

de

Sigoyer

a

,

Julie

Carlut

b

,

Aure´lia

Hubert-Ferrari

c

,

He´le`ne

Dessales

d

,

Hatice

Pamir

e

,

Volkan

Karabacak

f

a_{Universite´ Grenoble}_Alpes,_ISTerre,_CNRS,_BP_53,₃₈₀₄₁_Grenoble,_France

b_E´quipe_de_{ge´osciences}_marines,_Institut_de_physique_du_globe_de_Paris,_Sorbonne_Paris_Cite´,_{Universite´ Paris}_Diderot,_UMR_CNRS_7154,

75005Paris,France

c_University_of_Liege,_Unit_of_Physical_and_Quaternary_Geography,_Lie`ge,_Belgium d_UMR_8546,_Laboratoire_AOROC,_ENS,₇₅₂₃₀_Paris,_France

e_Mustafa_Kemal_University,_Archaeology_Department,_Hatay,_Turkey

f_Eskisehir_Osmangazi_University,_Department_of_Geology,₂₆₀₄₀_Eskisehir,_Turkey

1. Introduction

Themajorprogressaccomplishedinrecentyearsinthe fieldsofgeodesyandseismologyhasmadeitpossibleto clarify seismic hazard in numerous regions around the world.InordertodocumenttheQuaternaryseismicityofa given area,palaeoseismological studiesarenecessaryto characterizetheearthquakecycle,andtoprovide comple-mentary data to the instrumental approach for risk assessment.However,thesestudiesareespeciallycritical

whenthecurrentseismicitydoesnotreflectthelong-term activityofagivenarea,asitisthecaseoftheLevantfault. Thisfaultzonehasquiteanextensiveseismiccatalogue sincethebeginningofourera,whichstronglycontrasts withthemuchweakerseismicityduringthelastcentury (Ambraseys, 2009; Ambraseys and Jackson, 1998). The archaeological wealth of this region is a major assetto documentitsrecentseismicity,concurrently with tradi-tional palaeoseismologicalstudies.Several recent works have studied damaged historical sites in order to find tracesofearthquakesinthebuildingsandtoquantifythe associated deformation (Altunel et al., 2009; Galli and Galadini,2001; Meghraoui et al.,2003; Passchier et al., 2011; Volantetal., 2009). The aqueductsareespecially interesting buildings for archaeoseismic studiesas they

* Correspondingauthor.ISTerre,Domaineuniversitaire,1381,ruedela Piscine,38400Saint-Martin-d’He`res,France.

E-mailaddress:yacini@wanadoo.fr(Y.Benjelloun). ARTICLE INFO

Articlehistory:

Received25February2014

Acceptedafterrevision9December2014 Availableonlinexxx Keywords: Aqueduct Antioch Archaeoseismology Archaeology Brick Romanperiod ABSTRACT

TheRomanaqueductofAntioch-on-the-Orontes(Turkey),acitylocatednearthejunction betweentheactiveDeadSeafaultandtheEastAnatolianfault,hasbeendamagedseveral timesduetohistoricalearthquakes,asmentionedinancienttexts.Thetracesofrepairsare studied in order to identify their potential seismic origin. The deformations of the structure werecharacterisedthankstoaLIDARscan.Severalbricksweresampledon differentpartsofthecity’saqueducts,ontheoriginalstructureandonrepairedparts.The brickswerecharacterizedthroughapetrologicalapproach.14_C_and_{archaeomagnetism}

weretestedonthebricksinordertoconstraintheageoftheirproduction.Thesynthesisof allthedatashowedalocaloriginforthebricks,andledtotheidentificationofseveral manufacturing techniques and several types of production, thus, confirming the potentialityofthisapproachtodateandcharacterisepost-seismicrepairs.

!2015PublishedbyElsevierMassonSASonbehalfofAcade´miedessciences.

ContentslistsavailableatScienceDirect

Comptes

Rendus

Geoscience

w ww . sc i e nce d i re ct . co m

http://dx.doi.org/10.1016/j.crte.2014.12.002

(2)

commonlypresentasimple,linearpath,andwerequickly andsystematicallyrepairedafterbeingdamagedbecause of their critical role in providing the cities withwater. Ancientsources,suchasStrabo(XII.8.18,cf.Jones,1928,p. 515),alsoshowthattheRomanshadarealawarenessof seismic hazard and took that in consideration when planningtheirconstructions.

Inthisperspective,theRomancityof Antioch-on-the-Orontes (Antakya, Southern Turkey) (Fig. 1) offers the benefitofhavingawell-documentedhistoricalseismicity and information about the restoring phases of the aqueductofthecityafteritsdeteriorationbythe earth-quakes. Such restoring works, financed by the Roman

emperors, are mentioned by the Byzantine chronicler Malalas (243, 275–276,418–422, cf. Dindorf,1831) for threeearthquakesin37,115and526CE.Somepartsofthis aqueductarestillwellpreservednowadays.Theobjective ofthepresentstudyistocharacterisetherestoringphases and to evaluate if they could help us to clarify the earthquakecalendar.Thisstudyaimsatinvestigatingthe possible evolution of Roman materials and building techniques,aswellasitsimpactonthebuildingresistance to earthquakes. This latter issue is addressed by the analysisofbricks,astheyareanthropicmaterials.Looking atdifferentparameters,itispossibletogetaninsightinto the differentmanufacturingsteps.We presenthere the

Fig.1. (Colouronline.)Locationmapofthisstudy.Thethreecolouredcirclesrepresentthedamagedareasofthreehistoricalearthquakes,estimatedfrom thesourcesavailable.ThedatesaregivenCE.ThefaultsweredrawnaccordingtoAkyuzetal.(2006).

Y.Benjellounetal./C.R.Geosciencexxx(2015)xxx–xxx 2

(3)

firststepsofthiswork,whichconsistofcharacterizingand datingtheaqueduct’sbuildingmaterials.

2. Geologicalandarchaeologicalcontext 2.1. Localgeology

Thecity ofAntiochis locatedintheKarasuvalley, a depression set between two ranges of mountains, the AmanosonthewesternsideandtheAnatolianrangeon theeasternside.ThisriftisborderedbytwoNE–SWfault systems, which are part of the transition between the

LevantfaultandtheEastAnatolianfault.Atthispoint,the penultimatebigsegmentoftheLevantfaultcrossesthe Amikbasin,attheNorthwestofthecity,andjoinsthelast segmentwithaNE–SWorientation,knownastheKarasu fault(fig. 1)(Karabacakand Altunel,2013; Rojayet al., 2001). The surroundingrocks are seton a pre-Pliocene basement and can be subdivided into two groups: Palaeozoic igneous units topped by an allochthonous Mesozoicophioliticcomplex,andaLowerMiocene–Upper Eocenesedimentarysequencemadeoffoldedsandstones andlimestones(Fig.2)(Boulton,2009;Rojayetal.,2001). Theophioliteincludesserpentinisedperidotites,gabbros,

Fig.2.(Colouronline.)Mapoftheaqueductsandtheirgeologicalcontext.ThewhiteareasarecoveredbyQuaternarysedimentarydeposits.Thelocal lithologywastakenfromAksayetal.(2002).

(4)

diorite dykes, various basaltic lavas, boninites and plagiogranites(Robertson,2002).Theserocksarecovered withPlio-QuaternarysedimentsandQuaternaryvolcanic rocks (quartz tholeiite, olivine tholeiites, alkali basalts) (Alicietal.,2001;Bagcietal.,2008;Parlaketal.,2009). 2.2. Archaeologicalbackground

2.2.1. Theancientcityanditsaqueducts

Foundedin300BCbytheSeleucids,thecityofAntioch was famous in ancient times for its sophisticated and efficientsystemofwatersupply(Fig.2)(Kondoleon,2000). Drinkingwatermainlycamefromkarstsoverhangingthe plateauofDaphne,6km southofthecity.Theaqueduct system visible today was described by the Byzantine chroniclerMalalas(243, cf.Dindorf,1831).Followingan earthquakein37AD,Caligulabuiltbathsandanaqueduct supplyingthemwithwaterfromDaphnesprings.Malalas (275–276, cf. Dindorf, 1831) also writes that after the earthquakein115AD,TrajanandHadrianalsobuiltbaths withtheirassociatedaqueduct.Thesubsequentfateofthe aqueductsisunknown,althoughoneofthetwoaqueducts may have been used at the end of the 12th century accordingtoJohannesPhocas’testimony(Wilber,1938). Bothaqueducts(fig.2)wereidentifiedduring archaeolog-icalsurveysinthe1930s.Oneaqueduct(AQ1),supplying theNorthandthecentreofthecity,wasdatedtoCaligula’s reign,whiletheother(AQ2),locatedatahigherelevation and better designed, seemsto date backto Trajan and suppliedthesouthernpartofthecity(Wilber,1938). It shouldbenotedthatthosesurveyswerenotextensiveand havenotbeenupdatedsincethen.Duringafieldmissionin July2012,weidentifiedtwootherportionsofaqueduct (AQ4andAQ5),whichseemtobelongtoathirdsystem, higherthanthetwolinesdescribedabove.

2.2.2. Constructionmaterialssampled

Thankstoanextensive and almostcontinuous docu-mentationsincetheRomanperiod,weknowthatAntioch

anditssurroundingshavebeenrepeatedlystruckbymany earthquakes(Fig.1).Accordingtoancientsources,thecity wasstruckbynomorethanninesignificantearthquakes between0and500CE(Ambraseys,2009;Guidobonietal., 1994). These successiveearthquakes, together withthe effectsoftime,ledtoseveralrepairsthatarestillvisibleon theaqueducts.

Thedifferentbuildingmaterialsanalysedinthisstudy weresampledinJuly2012.Twelvebricksweresampledon fourdifferentportionsofaqueductidentifiedinthecity. AQ1andAQ2belongtothetwosystemsdescribedabove, whileAQ4andAQ5arepartofathirdhighersystem.Nine bricks come from thefirst buildingstages, while three weresampledonrepairedstructures,whichenabledusto comparethematerialsusedindifferentbuildingstages, and to check if the restoring bricks present peculiar features.Thename,location,sizeandbuildingcontextof eachbrickaresummarisedinTable1.

3. Methods

3.1. Constructionsurvey

ALIDARscanwascarriedoutonAQ4inordertohavea detailed topographic map of this portion of aqueduct, preliminarytoamorecomprehensivefieldstudy(fig.3). TheLIDARtechniqueisbasedonthemeasureofthetravel timeofalaserbeambetweenthesourcedeviceandagiven target.Scanningthetargetwithmultiplebeamsallowsone toproducea3Dscatterplotwithamillimetreresolution. AnextensiveLIDARsurveywasdonewiththeOptechILRIS 3Dsystem.Thescatterplotwasthengeneratedwiththe PolyWorks 10.0TM _software, _and _processed _with _the

TrimbleRealWorks7.1.1TM_software_at_the_University_of

Lie`ge (Belgium). Thanks to this processing, several measurements were done on the building: horizontal trackontheaqueduct,thicknessandheightofthedifferent architectural layers, deformations. Finally, the photo-graphstaken onthefield wereprojectedon thescatter

Table1

Names,locationandcontextofthestudiedsamples.AQ2remainsconsistinthreesuccessivepiles,withanorth–southorientation.Pile1,inthemiddle, supportsthesouthernpartofanarchandaportionofchannelabove.Pile2standsjustsouth.Thearchbetweenpiles1and2isnowdestroyed.Forthe locationsofAQ4samples,seealsoFig.3.

Sample Coordinates Context Dimensionsof sample

(length!thickness) incm

AQ1-1 36809.115N–36809.241E Origin,channelwall,justabovetheground 18!4 AQ2-1 36809.121N–36809.267E Origin,outsidewallofchannelabovethearch 30!3 AQ2-2 36809.121N–36809.267E SeeAQ2-1 30!3 AQ2-4 36809.121N–36809.267E SeeAQ2-1 30!3 AQ2-5 36809.121N–36809.267E Origin,arch 40!5.5 AQ2-8 36809.117N–36809.260E Origin,pile1,161mabovetheground 80!5 AQ2-9b 36809.117N–36809.260E Origin,pile1,justaboveAQ2-8 20!5.5 AQ2-9c 36809.117N–36809.260E SeeAQ2-9b 40!4 AQ2-10 36809.114N–36809.258E Origin,pile2,26mabovetheground 35!4.5 AQ4-1 36810.355N–36809.849E Repairstructureonacrack,westside,justsouth

ofabigtravertinedeposit,80cmbelowthechannel’stop

20!3.5 AQ4-4 36810.378N–36809.843E Repairstructureonacrack,westside,northernend

oftheaqueduct,2.6mabovethetheground

25!5.5 AQ5-1 36810.118N–36809.584E Repairstructureonacrack,3mabovetheground ? AQ5-2 36810.092N–36809.500E Origin,3mabovetheground ?!5

(5)

plotandortho-rectifiedtorealiseasketchoftheaqueduct, onwhichthemainstructuresweredrawn.

3.2. Petrologiccharacterisationofthebuildingmaterials 3.2.1. Macroscopicfeatures,texture,andmineralogy

Thedifferentbrickswerefirstcomparedonthebasisof macroscopicfeatures(Fig.4),suchasthesize,colour,and texture. Then, in order to study more precisely their manufacturing process, thin sections were madeat the Laboratory of Geologyof the E´cole Normale Supe´rieure (Paris) and characterized under an optical microscope LeicaDMRXP,equippedwithacameraLeicaDFC420.The natureofdifferentcomponentsobservedwasalsostudied withanFEGZeissSigmascanningelectronicmicroscope (Table2).

3.2.2. 2D-porosityandgranulometry

SEM picture mosaics were realised on most of the bricks, with a magnification of 500 and a resolution of 1024!768pixels.OntheSEMimages,theporesarethe only areas coloured in black, which makes it easy to compute then 2D-porosity using the JMicroVision soft-ware, (Roduit, 2007). To check the reliability of this method,3D-porositymeasurementswerealsocarriedout concurrentlyontwobricks(AQ4-4andAQ5-2),usingthe triple-weightmethod(NFEN1936,2007norm).

For the granulometry study, the brick thin sections AQ1-1,AQ2-4,AQ2-5,AQ2-8,AQ2-9b,AQ4-1,AQ5-1,and AQ5-2werefullyphotographedinpolarisedlightwitha magnification of 1.6 and a resolution of 2592!1944 pixels. The pictures were then processed usingtheJMicroVisionsoftware.Theareaandequivalent circular diameterof the grains arethen estimated. The percentageofareacoveredbythegrainsisthenplottedin diagrams(Figs.5and6).

3.3. Dating

3.3.1. Archaeomagnetism

In order to constrain the age of production of the differentbricks,magneticpalaeointensitiesmeasurements

werecarriedoutoneightsamples(AQ1-1,AQ2-1,AQ2-5, AQ2-9b,AQ2-10,AQ4-1,AQ5-1,andAQ5-2)atthe Saint-Maur-des-Fosse´sobservatory, incollaboration with Yves GalletandMaxime LeGoff(IPGP,Paris).These measure-mentsweredonewithtwotriaxemagnetometers,whose operatingprincipleisdescribedinLeGoffandGallet(2004)

(Table3).Low-fieldthermomagneticcurvesacquiredwitha KLYAgicoequippedwithaCS2ovenupto5808ConAQ1-1, AQ2-5,AQ2-9b,AQ2-10,AQ4-1,AQ4-4,AQ5-1,andAQ5-2 allowedustogaininsightsintothemagneticmineralogy and stability of the magnetic phases during thermal treatments.Inaddition,thecomposition,sizeandtexture offerri/ferromagneticphasesweredescribedusingtheBSE modeduringSEMobservationsonAQ1-1,AQ2-5,AQ2-9b, andAQ2-10thinsections.

3.3.2. Carbon14

Charcoalfragmentswereextractedfromthemortarof samplesAQ1-1,AQ2-9,AQ2-10,andAQ4-4,anddatedby Aeon Laboratories (aeonlaboratories.com) with the 14_C

method, to constrain the building age of the different architecturalunits.Theradiocarbonageswereconverted intocalendarageswiththeINTCAL09calibration(Heaton etal.,2009)usingOxcalprogram(Ramsey,2009). 4. Results

4.1. DamagesreportedonAQ4

Fig.3showsasyntheticsketchofthesurveyrealised withtheLIDARscanofAQ4.Theaqueductroughlyfollows anorth–southline,withtwodirectionchangesthatmake it concave toward the west. The first change in the northernparthasbeenmadebythehandofmantojoin theothersideoftheaqueductacrosstheriver.Bycontrast, thechangeinthesouthern part(redarrow inFig.3)is associatedwithacrack,hiddenbyahugeaccumulationof travertinethatprecipitated becauseof aleakage;thisis interpretedastheresultofadeformationsufferedbythe aqueductafteritsconstruction.Thewholesouthernpartof theaqueductsufferedananticlockwiserotationofabout 158, andthelayerindicated by a redline on Fig.3 has

Table2

Nature ofthetempergrains inthebricks.Cpx=clinopyroxene,opx=orthopyroxene,amphi=amphibole,plagio=plagioclase,K–Fd=K-feldspar,ol basalt=olivinebasalt,brlava=brownlava,carbpeb=carbonatepebble.‘‘o’’meansanoccurrenceoftheelement.

AQ1-1 AQ2-1 AQ2-2 AQ2-4 AQ2-5 AQ2-8 AQ2-9b AQ2-9c AQ2-10 AQ4-1 AQ4-4 AQ5-1 AQ5-2

Cpx o o o o o o o o o o o o Opx o o o o o o o o o Amphi o o o o o o o o o o o Olivine o o o o o o o o o o o Plagio o o o o o o o o o o Quartz o o o o o o o o o o K–Fd o o o o o Calcite o o o o o o o o o o Chromite o o o o Olbasalt o o o o o o o o o o o o o Gabbro o o o o o o o o o o o Brlava o o o o o Diorite o o o o o Andesite o o o o o o Granite o o o o o o

(6)

sufferedacollapseof10to15cm,southofthetravertine deposit.After thisdamage,water wasno longer ableto

flownorthwardasitdidbefore.Anotherfracturecanbe seenatthenorthernendoftheaqueduct.

4.2. Petrologyofthebricks

4.2.1. Microscopictextureandmineralogy

Allthebricksshowtheirowncharacteristicsintermsof texture, which may reflect different qualities of fabric (Fig.4).AQ2-1,AQ2-2,andAQ4-1,havearather homoge-neousgranulometryofthetemperandshowagoodquality ofclay-tempermixing.Onthecontrary,AQ2-4,AQ2-8and AQ2-10containvarioussizesofgrainsandthemixingwas donelesscareful,asshowedbythepresenceofseveralclay

Table3

Magneticpalaeointensities recordedin the bricks. The means were computedonthreefragmentsforeachbrick,exceptAQ2-5thatshowed greatervariabilityandforwhichsixfragmentswereused.

Sample Meanintensity(mT) Standarddeviation

AQ1-1 59.9 2.3

AQ2-9b 58.5 0.8 AQ2-10 57.2 1.0

AQ4-4 58.1 0.8

AQ5-2 54.8 2.8

Fig.3. (Colouronline.)InterpretativesketchesoftheLIDARscanforthesiteAQ4.a:horizontalprofileoftheaqueduct;b:photographoftheeasternsideof theaqueduct;c:schematicsketchoftheeasternsideoftheaqueductprojectedonaverticalplane.Thefeaturesweredigitalisedfromtheortho-rectified photographsoftheaqueductwithArcGISTM_._The_image_corresponds_to_a_{north–south}_projection._The_red_and_black_broken_lines_show_the_original_position

ofthesouthernpartoftheaqueductbeforedamage.

(7)

lumpswithouttemper.Theotherbrickspresent interme-diate qualities of fabric. Fig. 5 presents the grain area percentagesestimatedbyimageprocessing.Mostofthe brickshaveavaluecloseto30%,butsomebricks,like AQ1-1, AQ4-1and AQ5-2,exhibit lowervalues, whileAQ2-8 showsremarkablyhighgraincontent,largerthan40%.

ThemineralogyofthetemperisreportedinTable2.Itis composedofvariousfossils,suchasnummulites(fromEocene limestones),andisolatedmineralsandfragmentsof rocks, whichcanbeidentified.Mostofthemcomefrombasaltsand gabbros. All the bricks exhibit quite similar mineralogical contents. The temper includes rounded and angular frag-ments,thelatteroftenhavingveryacuteandreflexangles. 4.2.2. Porositymeasurements

The2D-porosity resultsareshowninFig.5.The 3D-porositymeasurementscarriedoutonAQ4-4andAQ5-2

providedvaluesof28%and16%,respectivelythatconfirm the2D-porosityanalysisdoneonthephotograph.The 2D analysisseemsthereforetogivereliableporosityestimates. The2D-porosityvaluesarequitesimilaramongthebricks. Still,AQ2-4,AQ4-1andAQ5-1exhibithighervalues,closerto 20%,whileAQ2-1andAQ2-2showlowerones,below10%. 4.2.3. Granulometry

TypicalgranulometrydiagramsareshowninFig.6.The curvescanbedividedintotwoparts.Thefirstoneincludes thesmallergrainsthatclearlyfollowaGaussianpattern, whilethebiggestgrainslieabovetheGaussiancurve.Itis noticeablethatthegraindiameterforwhichthegrainarea percentageis thehighestis thesamefor almostallthe bricks, around 40

m

m, except AQ2-9b, which contains more grains of a lower diameter, close to 20

m

m. The coarsergrainsinmostofthebricksfollowaregularpattern

Fig.4.(Colouronline.)Macroscopicphotographsofelevenfragmentsofthebricksamples.Thefragmentswerecutandpolishedduringthepreparationof thethinsections.

(8)

thatlookslikeaparabola(e.g.,AQ2-4).Severalpointsdonotfit onthisparabola,between200and700

m

m,andcanrepresent asignificantarea,asinAQ2-9b(fig.6).Theaveragediameterof these outliers is also different between the bricks. The maximumgraindiameterforallthebricksdoesnotexceed 3mm,anditisevenlowerthan1mmforAQ2-9bandAQ4-4. 4.3. Dating

4.3.1. Archaeomagnetism

Over the 24 samples measured, 6 did not provide resultsthat fitthe qualitycriteria, suchasdescribed in

Gallet and Le Goff (2006). AQ2-5 and AQ4-1 showed anomalousresultsthatcanbeexplainedbyanunsuitable mineralogy, with an important amountof large titano-magnetites.AllsuitableresultsaredisplayedinTable3.All thesamplesprovidedsimilarintensities,withameanof 57.6"1.7

m

T.Theobtainedintensityvaluesare tooclose, consideringtheassociateduncertainties,toreallydistinguish subgroupsamongthebricks.

Thankstothereferencecurveestablishedforceramics from theNishapur site in Iran, at the same latitudeas Antioch(GeneveyandGallet,2013),but1700kmmoreto theeast(Fig.7),themeasuredintensitiescanbeconverted into ages. As the reference curve is not established in Antioch, someerrors may be generated,in particular a systematic time lag due to a westward drift of the geomagnetic secular variation has been recently under discussion(Aubertetal.,2013;Livermoreetal.,2013).The timeintervalthatweobtainwiththiscorrelationranges from the5th centuryBCE tothe1st century CE, which covers the construction age of AQ1 estimated by the archaeologists(see Section2.2.1).Another less probable possibilitywouldbelaterthanthe4thcenturyCE.Asthe brickswerelikelyproducedduringtheRomanperiod,after the 1st century BCE, the production age is mostly constrained by thelowerlimit of theintensityinterval, towardthemid-1stcenturyCE.

4.3.2. 14_C_ages

Three charcoalsextractedfromthemortarsofAQ2-9 and AQ2-10 providedsimilar ages of 0"50year CE(2

s

uncertainty) that would match a construction during Caligula’sreign.AsecondcharcoalfromAQ2-9mortargives amucholderage,between500and400BCE,whichcouldbe duetotheuseofoldwoodintheconstruction.Thecharcoal fromAQ1-1mortarprovidedanageof800"100year.The lastcharcoal,fromAQ4-4mortar,givesaquasi-modernage, certainlydubiousasthesamplewastoolighttobeanalysed properly.

5. Discussion

5.1. Relevanceofthedifferentparametersobservedonthe bricks

Thecolourofabrickdependsmainlyontheclayandthe firingprocess.Claydepositednearthealterationareawill showaredtobrowncolour,whilepurersedimentaryclay will exhibit white to yellow colours (Velde and Druc, 1999).Thefiringprocesscansignificantlyalterthisoriginal colour:ahigh-temperature,oxidisingfiringwillgiveared colour,whileareducingfiringwillmakethebrickdarker. Thisreducingatmospherecanbeobtainedbyreducingthe temperatureandputtingwetorganicmatterinthekiln. Differences in thecolour intensitymayalsoresultfrom differentpositionsinthefurnace(LeNy,1985).

Themacroscopicormicroscopictexture,aswellasthe porosity,maygiveinformationonthecarewithwhichthe brick maker worked on his paste. For instance, a homogeneousspatialdistributionoftheporesandgrains, with sometimes a preferential orientation of grains,

Fig.5.(Colouronline.)Amountofgrainsandporosityofthebricks.The grains/totalarearatios(greybars)wereestimatedwiththinsection microphotographs,andthe2Dporosities(redbars)withSEMimages.The twoorangeellipsesindicate3D-porosityvalues.ThevaluesforAQ2-9b andAQ4-4mustbeconsideredwithcautionbecauseofthesmallamount ofmaterialonthethinsection.Theuncertaintyis"2%.

Fig.6.Graindistributionofthreerepresentativebricks.Thethreebricks arerepresentativeofthethreetypesidentifiedinthisstudy:typeA (AQ2-4,black),typeB(AQ4-1,grey),andtypeC(AQ2-9b,white).Thescale usedforthediameterislogarithmic.

(9)

indicatesacarefulmixtureoftheclayandthetemper.Itis also possible to study the selection of the temper by lookingatthenature,angularity,andgranulometryofthe grains.AccordingtoVeldeandDruc(1999),theGaussian patternforthefinergrainsinthegrainsizediagrams(fig. 8)correspondstothe‘‘naturaltemper’’,i.e.thegrainsthat wereoriginallypresentintheclayresource.Therefore,the peakoftheGaussiancurve,i.e.thegrainsizecoveringthe biggest area, can be used to characterise and compare differentclayresources.Thecoarsergrainswereaddedby thebrickmakertogivebettermechaniccoherencetothe claypaste. Thosegrainscanalsofollowregularpatterns like a parabola, which corresponds to an uniform distributionof thedifferentgrainsizes, ornot and it is thenpossibletoseeifapeculiarsizerangewaschosenby thebrickmaker.Thefactthatthemaximumgrainsizedoes not exceed 3mm can indicate the use of sieving. The anthropic addition of grains is also evidenced by the coexistenceofnaturallyincompatiblemineralsinthesame brick,likequartzand olivine,and thehighangularityof manygrainsthatcanresultofcrushing.

5.2. Originofthebuildingmaterials

Therockfragmentsusedastemperareverysimilarto the geological outcrops around Antioch, and several identifiedrockscanbecomparedtothedifferentunitsof theophioliticcomplex:basalticlavas,gabbros,diorites.Rare plagiogranitesmayhavebeenused.Thefossilsfoundinthe

bricks match with those found in the local limestones. Limestones were not only used for construction (we observedseveralblocksofsimilarlimestoneinthebuilding), butalsoastemperandfluxingagent.Allthesefactssupport thehypothesisofalocalproductionforthebricks. 5.3. Typologyforthebricks

5.3.1. Selectionofrelevantcomparisonparameters

ThebricksAQ2-1toAQ2-5,sampledinthearchofAQ2, areallverysimilaronmanyaspects.Theywereproduced using the same manufacturing process; they certainly belongtothesameproduction,andwereprobablymade bythesameceramist. Thebricksofthis coherentgroup present variability for some studied parameters, which give an idea of the heterogeneity inside one set. It is noticeable thattheporosity is quiteuneven withinthis group,andthattheirmineralcontentmaynotbestrictly similar.Therefore,theporosityandtheoccurrenceofsome peculiarminerals/rockswillnotbeusedtodiscriminate the different productions. By contrast, the amount of temper, the granulometry, and the colour are very homogeneous within this group, and may certainly be usedasrelevantparameters.

5.3.2. Comparisonandtypology

Afirstgroupofbricks(AQ2-1toAQ2-5)ischaracterised byahighproportion(30%)oftemper,aselectionof middle-sizedgrains(around300

m

m),a3-mmsieving,andarather

45 50 55 60 65 70 -500 0 500 1000 1500

57.6 µT

Terminal date

for production

Year

Pa

la

eo

in

te

ns

it

y (

µT

)

AQ2-9

AQ2-10 AQ1-1 14_{C ages}

Fig.7.(Colouronline.)ReferencecurveofthevariationsofthemagneticfieldforthesiteofNishapurThepalaeointensitydataandtheirdatingweretaken

from(GeneveyandGallet,2013).Themeanofthedataobtainedfromthebricksisindicatedbythehorizontalredlineandthestandarddeviationbythe

horizontalredband.Thetimeintervalscorrespondingtothoseintensityresultsareshownwithtwoverticalgreybands.Thereliableconstructionage intervalsobtainedfrom14_C_dating_on_charcoals_from_the_mortar_are_indicated_by_the_horizontal_black_bars.

(10)

oxidisingfiring(typeA).AQ2-10andprobablyAQ5-1can beaddedtothisgroupA.Asecondgroup(typeB)canbe identified,whichcouldincludeAQ1-1andAQ4-1,andis characterisedbyaverysmallamountoftemper,finegrains (around200

m

m),a3-mmsieving,andanoxidisingfiring. Finally, a third group (typeC) differs from the type A because it presents no grain above 1mm, and could includeAQ2-9band AQ4-4.Theydo notshowa similar distributionoffinegrains,thoughthepeakoftheGaussian curveoccursat40

m

mforAQ4-4likefortheotherbricks, against20

m

mforAQ2-9b.So,thesetwo bricksseemto havefollowedthesamemanufacturingprocess,butthey differbytheirclayresourceandmaynotbefromthesame workshop,AQ4-4havingasimilarclaytypetotheother bricks contrary toAQ2-9b. There remainthree isolated bricks,AQ2-8,AQ2-9c,andAQ5-2,alldifferent.Thethree brickssharetheparticularityofhavingbeenfiredundera reducingatmosphere,asshownbytheirbrownishcolour, butdoesnotpresentsimilargrainsizedistribution:while AQ2-8hasahugeamount(over40%)ofunsortedtemper, AQ5-2contains coarse temperin very small proportion (20%).

5.4. Datingofthebuilding

The archaeomagnetism results suggest that all the brickswereprobablymanufacturedbeforethe2ndcentury CE.This ageisconsistentwiththe14Cageobtainedfor AQ2-10smortar,andwithaconstructionduringCaligula’s reign, as claimed by Malalas (243). However, this age disagrees with the few archaeological data available, which claim that AQ2 was built by Trajan in the 2nd centuryCE.ForAQ1-1,weobtainedamuchlaterageforthe constructionthatcouldmatchwiththelatertimeinterval suggestedbyarchaeomagnetism(from500 to1200CE). Whilethearchaeological dataclaim that AQ1 wasbuilt during Caligula’s reign, the 14_C _data _on _AQ1-1 _mortar

suggeststhatthisportionofAQ1wasbuiltorrebuiltinthe MiddleAges.

6. Conclusion

Wehaveprovidedafirstworkonthisbuilding,which hasbeenverypoorlystudiedsofar.OurLidarstudyproved thattheaqueductwasdeformedafteritwasbuilt,which resultedinimportantleaksalongthecracks,atiltofthe wholestructure,andareversewaterflowcausingmassive travertineprecipitation.Thebricksstudyhasallowedusto characterise different productionsand toidentify some manufacturing techniques, like the use of sieving. The temperanalysisprovesalocaloriginforthebricks.Mostof theagesobtainedforthebricksdatedby archaeomagnet-ismandthe14_C_method_on_mortar_are_coherent,_except_for

one very young14_C _age._{Nevertheless,} _the _ages _remain

poorlyaccurate.Aslongaswelackasolidtimeframe,it seems difficult to link these repairs and the cracks observedon theaqueducts withhistorical earthquakes. Supplementary data is needed to better constrain the chronology of thebrick production, thebuilding of the differentaqueductsandtheirrepairs.Adetailed archaeo-logical study of the building and thermoluminescence

dating on the bricks could refine the chronology of the constructionandbrick production.Nevertheless,we have developedthroughthisworkpromising methodstostudy ancientbuildingdeformationsandmaterials,whichcouldbe appliedtootherarchaeologicalsitespronetoseismichazard. Acknowledgements

Thisresearchwasmostlyfounded bytheENS(Paris, France),theMistralprojectofINSU,andtheUniversityof Lie`ge(Belgium).WethankHaticePamirforenablingthe samplinginAntioch.ISTerreispartofLabexOSUG@2020 (ANR10LABX56).

References

Aksay,A.,etal.,2002.1:500,000ScaledGeologicalMapsofTurkey,Hatay

Region.In:GeneralDirectorateofMineralResearchandExploration,

Ankara,Turkey.

Akyuz,H.S.,Altunel,E.,Karabacak, V.,Yalciner,C.C.,2006. Historical

earthquakesactivityofthenorthernpartoftheDeadSeafaultzone,

southernTurkey.Tectonophysics426,281–293.

Alici,P.,Temel,A.,Gourgaud,A.,Vidal,P.,Gu¨ndogdu,M.N.,2001.

Quater-narytholeiitictoalkalinevolcanismintheKarasuValley,DeadSea

RiftZone,SoutheastTurkey:Sr-Nd-Pb-Oisotopicandtrace-element

approachestoCrust-MantleInteraction.Int.Geol.Rev.43,120–138.

Altunel,E.,Meghraoui,M.,Karabacak,V.,Akyuz,S.H.,Ferry,M.,Yalciner, C.,Munschy,M.,2009.Archaeologicalsites(TellandRoad)offsetby

theDeadSeaFaultintheAmikBasin,southernTurkey.Geophys.J.Int.

179,1313–1329.

Ambraseys,N.N.,2009.EarthquakesintheMediterraneanand

Middle-East:amultidisciplinarystudyofseismicityupto1900.Cambridge

UniversityPress,Cambridge,UK.

Ambraseys,N.N.,Jackson,J.A.,1998.Faultingassociatedwithhistorical

andrecentearthquakesintheEasternMediterraneanregion.

Geo-phys.J.Int.133,390–406.

Aubert,J.,Finlay,C.C.,Fournier,A.,2013.Bottom-upcontrolof

geomag-neticsecularvariationbytheEarth’sinnercore.Nature502,219–223.

Bagci,U.,Parlak,O.,Hoeck,V.,2008.Geochemistryandtectonic

environ-mentofdiversemagma generationsformingthecrustalunitsof

theKizildag(Hatay)Ophiolite,southernTurkey.Turk.J.EarthSci.

17,43–71.

Boulton,S.J.,2009.RecordofCenozoicsedimentationfromtheAmanos

Mountains, southernTurkey: Implicationsfor the inception and

evolutionoftheArabia-Eurasiacontinentalcollision.Sediment.Geol.

216,29–47.

Dindorf,L.A.,1831.IoannesMalalasChronographia.In:Corpus

Scrip-torumHistoriaeByzantinae32Bonn.1828–1897.

Gallet,Y.,LeGoff,M.,2006.High-temperaturearcheointensity

measure-mentsfromMesopotamia.EarthPlanet.Sci.Lett.241,159–173.

Galli,P.,Galadini,F.,2001.Surfacefaultingofarchaeologicalrelics.A

reviewofcasehistoriesfromtheDeadSeatotheAlps.Tectonophysics

335,291–312.

Genevey,A.,Gallet,Y.,2013.Theexcavationandtheabsolutechronology.

In:Collinet,A.,Rante,R.(Eds.),NishapurRevisited:Stratigraphyand

CeramicsoftheQohandez.OxbowBooks,Oxford,UK, pp.28–38.

Guidoboni,E.,Comastri,A.,Traina,G.,1994.Catalogueofancient

earth-quakesintheMediterraneanareauptothe10thcentury.Istituto

NazionalediGeofisica,Roma,Italy.

Heaton,T.J.,Blackwell,P.G.,Buck,C.E.,2009.ABayesianapproachtothe

estimationofradiocarboncalibrationcurves:theIntCal09

method-ology.Radiocarbon51,1151–1164.

Jones,H.L.,1928.Strabo:Geography.Books10–12.HarvardUniversity

Press,Cambridge,USA.

Karabacak,V.,Altunel,E.,2013.EvolutionofthenorthernDeadSeaFault

ZoneinsouthernTurkey.J.Geodyn.65,282–291.

Kondoleon,C.,2000.Antioch:thelostancientcity.PrincetonUniversity

Press,Princeton,USA.

LeGoff,M.,Gallet,Y.,2004.Anewthree-axisvibratingsample

magne-tometerforcontinuoushigh-temperaturemagnetization

measure-ments:applicationstopalaeo-andarchaeo-intensitydeterminations.

EarthPlanet.Sci.Lett.229,31–43.

LeNy,F.,1985.Lesfoursdetuiliersgallo-romains:me´thodologie,e´tudes

technologiques,typologieetstatistique,chronologie.E´ditiondela

Maisondessciencesdel’homme,Paris,France.

(11)

Livermore,P.W.,Hollerbach,R.,Jackson,A.,2013.Electromagnetically

drivenwestwarddriftandinner-coresuperrotationinEarth’score.

Proc.Natl.Acad.Sci.USA110,15914–15918.

Meghraoui,M.,Gomez,F.,Sbeinati,R.,VanderWoerd,J.,Mouty,M.,Darkal, A.N.,Radwan,Y.,Layyous,I.,AlNajjar,H.,Darawcheh,R.,Hijazi,F., Al-Ghazzi,R.,Barazangi,M.,2003.Evidencefor830yearsofseismic

quies-cencefrompalaeoseismology,archaeoseismology,andhistorical

seismic-ityalongtheDeadSeafaultinSyria.EarthPlanet.Sci.Lett.210,35–52.

Norme,N.F.E.N.,1936.2007.Me´thodes d’essaidespierresnaturelles.

De´terminationdesmassesvolumiquesre´elleetapparente etdes

porosite´souvertesettotales.E´dAFNOR.

Parlak,O.,Rizaoglu,T.,Bagci,U.,Karaoglan,F.,Hoeck,V.,2009.Tectonic

significanceofthegeochemistryandpetrologyofophiolitesin

South-eastAnatolia,Turkey.Tectonophysics473,173–187.

Passchier,C.,Wiplinger,G.,Su¨rmelihindi,G.,Kessener,P.,Gu¨ngo¨r,T., 2011.Romanaqueductsasindicatorsofhistoricalactivefaultsinthe

Mediterraneanbasin.In:2ndINQA-IGCP-567International

Work-shop onActive Tectonics, Earthquake Geology,Archaeology and

Engineering, Corinth,Greece.

Ramsey,C.B.,2009.Bayesiananalysisofradiocarbondates.Radiocarbon

51,337–360.

Robertson,A.H.F.,2002.Overviewofthegenesisandemplacementof

MesozoicophiolitesintheEasternMediterraneanTethyanregion.

Lithos65,1–67.

Roduit,N.,2007.JMicroVision:unlogicield’analysed’images

pe´trogra-phiquespolyvalent(The`sededoctorate`ssciences).Universite´ de

Gene`ve.

Rojay,B.,Heimann,A.,Toprak,V.,2001.Neotectonicandvolcanic

char-acteristicsoftheKarasufaultzone(Anatolia,Turkey):thetransition

zonebetweentheDeadSeatransformandtheEastAnatolianfault

zone.Geodin.Acta14,1–17.

Velde,B.,Druc,I.C.,1999.Archaeologicalceramicmaterials:originand

utilization.Ed.SpringerVerlag.

Volant,P.,Levret,A.,Carbon,D.,Scotti,O.,Combescure,D.,Verdel,T.,Piant, A.,Laurent,P.,2009.Anarchaeo-seismologicalstudyoftheNıˆmes

Romanaqueduct, France: indirect evidencefor an M>6seismic

event? Nat.Hazards49,53–77.

Wilber,D.N.,1938.TheplateauofDaphne.Thespringsandthewater

systemleadingtoAntioch.In:Stillwell,R.(Ed.),

Antioch-on-the-OrontesII:TheExcavationsof1932–1936.PrincetonUniversityPress,