lable at ScienceDirect

Journal of Arid Environments 74 (2010) 861–869

Contents lists avai

Journal of Arid Environments

journal homepage: www.elsevier .com/locate/ jar idenv

Abandonment ages of alluvial landforms in the hyperarid Negev determinedby luminescence dating

N. Porat a,*, R. Amit a, Y. Enzel b,c, E. Zilberman a, Y. Avni a, H. Ginat d, D. Gluck a,b

a Geological Survey of Israel, 30 Malkhe Israel St., Jerusalem 95501, Israelb Institute of Earth Sciences, The Hebrew University of Jerusalem, Jerusalem 91904, Israelc Department of Geography, The Hebrew University of Jerusalem, Jerusalem 91905, Israeld Ma’ale Shaharut High School, Kibbutz Yotvata, Arava Valley, Israel

a r t i c l e i n f o

Article history:Received 4 February 2009Received in revised form9 October 2009Accepted 31 October 2009

Keywords:Alluvial fansDatingDesertGeomorphic surfacesIsraelLuminescenceNegevTerraces

* Corresponding author. Tel.: þ972 2 5314298; fax:E-mail address: naomi.porat@gsi.gov.il (N. Porat).

0140-1963/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.jaridenv.2009.10.018

a b s t r a c t

Dating the time of abandonment of geomorphic surfaces in the arid mid latitudes is necessary for studiesranging from tectonics, landscape evolution and paleoclimate. It has often been hampered by the limitedmaterial suitable for conventional isotopic methods and the uncertainties inherent in cosmogenicradionuclide methods. We propose luminescence dating as a suitable method for dating the time ofabandonment of aggradational geomorphic surfaces in the hyperarid regions. We dated the top of suchsurfaces with different age ranges and geomorphic settings in a sequence of alluvial fans in NahalShehoret in the southern Negev and in a sequence of terraces in Nahal Ze’elim and in Nahal Zin, twoadjacent drainage basins in the Judean Desert, Israel. Samples were collected from beneath the gypsichorizon at a depth of 0.3–0.7 m, below which sand grains do not penetrate. Depositional ages for theuppermost beds of the landforms, which are proxies for abandonment, range from w90 ka to w5 ka. Inall three basins, the ages are in morphostratigraphic order and agree well with relative age estimatesbased on soil chronosequences and on Lake Lisan levels. Abandonment ages for an individual alluvial fancluster within �10–20%, therefore it is possible to distinguish between surfaces with ages differing bymore than 20%. Thus, in hyperarid areas, the luminescence methods can be used for surface dating.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Determining ages of alluvial fans in hyperarid areas is difficultbecause datable material, whether organic matter for radiocarbonor pedogenic carbonate for U-series disequilibrium dating, isscarce. This in the past hindered research on Quaternary landscapeevolution, fluvial responses to climate change, palae-oenvironmental and palaeoseismological studies, and assessmentof phases and rates of aggradation and incision (e.g., Bull, 1991).Dating landforms in semi-arid and arid regions has often beenbased on the limited application of isotopic dating methods such asU–Th on pedogenic carbonates (Sharp et al., 2003) or on strati-graphic correlations with independently dated landforms anddeposits such as lacustrine terraces, ash layers, marine shorelinesand basalt flows (e.g., Bull, 1991; Harvey and Wells, 2003;McDonald et al., 2003; Wells et al., 1987). Semi-quantitative andrelative dating methods have included morphostratigraphy, degree

þ972 2 5380688.

All rights reserved.

of soil development (Amit et al., 1993; Birkeland et al., 1991;Harrison et al., 1993; McDonald et al., 2003; Rockwell et al., 1994)and clast rubification (Helms et al., 2003). These methods requirecalibrations against known ages which in the hyperarid regions arefrequently absent (Noller et al., 2000).

In situ cosmogenic radionuclides have been applied to datesurfaces of aggradation landforms (e.g., Gosse and Phillips, 2001;Granger and Muzikar, 2001; Schildgen et al., 2002). While prom-ising, these methods provide model ages that rely on parametersthat are partially known: erosion rates, nuclide inheritance andthe complex history of the measured clasts (e.g., Clapp et al., 2000;Hancock et al., 1999; Liu et al., 1996; Matmon et al., 2003;Putkonen and Swanson, 2003; Repka et al., 1997; Watchman andTwidale, 2002).

In the past decade, luminescence methods (Aitken, 1998), whichdate the last exposure of mineral grains to sunlight, have greatlyimproved the prospects of dating the aggradation phases of alluvialfans and terraces (Rittenour, 2008 and references therein).However, dating the time of alluvial fan abandonment – the age ofthe geomorphic surface – still remains a challenge. While terracesand alluvial fans have been dated in arid to humid regions

Fig. 1. Location map. Squares denote the studied wadis (Nahal). The 50 mm, 100 mmand 200 mm isohyets are shown.

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869862

(Rittenour, 2008), sediments close to surfaces and unconformitieswere generally avoided as they could have been disturbed bypedoturbation. Even if not observed today, pedoturbation couldhave introduced grains with reset or partially reset luminescencesignal into the soil profile (Bush and Feathers, 2003; Bateman et al.,2003, 2007; Mee et al., 2004), resulting in an underestimation ofsediment age.

Detailed infrared stimulated luminescence (IRSL) dating ofa Late Pleistocene alluvial fan in Nahal (meaning ephemeralstream; hereafter N.) Shehoret located in the hyperarid region ofsouthern Israel, found that the top w1.5 m of the alluvial fanaccreted within a relatively short time period – between 67 ka and60 ka – compared to the underlying alluvial units which gave agesof 75–100 ka (Porat et al., 1997). Most likely, the age of sediment atthe top of a terrace is the time of the final event of deposition justprior to abandonment of the alluvial fan, and it should provide thebest estimate for the time of the surface formation.

Pedoturbation in hyperarid regions is minimal, as demonstratedby IRSL ages from the same Late Pleistocene alluvial fan inN. Shehoret (Porat et al., 1996). The eastern part of the fan waspartially buried at w35 ka due to faulting, but the original surfacecould still be detected at some depth in trenches by the buried regsoil profile. Samples collected from the exposed and buried fansurface, 30–50 cm below the top of the reg soil profile, gave iden-tical luminescence ages of 56 ka (Amit et al., 2002; Porat et al.,1996), implying that the exposed surface retained its integrity overthat time period and was not modified more than the buriedsurface with regard to the luminescence age of the sand grains.These results suggest that grains with reset luminescence signalsdo not infiltrate from the surface down below the gypsic soilhorizons (at a depth of 30–70 cm) even over tens of thousands ofyears of exposure and that the luminescence ages of samplescollected from beneath that horizon represent deposition ages.

These combined observations; that the uppermost part of allu-vial fans is deposited rather rapidly, and that in hyperarid regionssurfaces are essentially impenetrable to sand grains, prompted usto exploit this to the benefit of dating geomorphic surfaces usingluminescence. The ages would represent the last exposure ofmineral grains to sunlight and would provide the time of the lastdeposition of sediments underlying the surface. When the reg soilprofile is still fully intact, implying no substantial erosion, theluminescence ages are somewhat older than surface abandonment,but close in time to the actual event. Multiple sampling from thesame surface will provide the timing of this abandonment.

Here we report ages of well preserved alluvial fans and terracesurfaces from the hyperarid (<75 mm/year) region of the Negev,Israel (Fig. 1). Independent age control for verifying our results wasbased on (a) morphostratigraphy; (b) well dated lacustrine shore-lines; and (c) the degree of soil development on the dated surfaces.

2. Study sites

The study sites are located in the southern Negev and JudeanDesert (Fig. 1). The climate in this region is hyperarid with meanannual rainfall ranging from 30 mm in N. Shehoret to 75 mm in N.Ze’elim (Horowitz, 2001). Most precipitation falls in autumn tospring during very short, high intensity storms, with dry conditionsmost of the time (Barzilai et al., 2000). At present there is novegetation on the terraces and fans because of high surface andsubsurface salinity. The absence of calcic horizons in these soils,despite abundant Ca sources, indicates that vegetation coverneeded for their formation was limited throughout soil formationhistory (Amit et al., 2006).

The N. Shehoret site consists of a succession of faultedalluvial fan surfaces that have been studied in detail for

palaeoseismological purposes (Amit et al., 1999; 2002; Enzel et al.,1996; Gerson et al., 1993; Porat et al., 1996; 1997; 2009). Theextensive alluvial surfaces are well-defined in aerial photographsand in their reg soil characteristics (Amit et al., 1996). They providean excellent case-study for testing the range of ages measuredunder a single alluvial fan surface and for checking the consistencyof ages against the alluvial surface morphostratigraphy and soilchronosequences (Amit et al., 1996).

N. Zin is located near the southwestern margin of the Dead Seawhile N. Ze’elim drains into the Dead Sea further to the north. Thereaches of these two wadis are controlled by the levels and the

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869 863

proximity of the Dead Sea and its precursor Lake Lisan (e.g.,Bowman, 1997; Bookman et al., 2006; Gluck, 2001). The fieldstratigraphic relations indicate that in these two wadis, the highestand oldest alluvial fans formed soon after the regression of LakeLisan began, from its highest stand of w170 m below sea level at28,000–25,000 yr ago (Bookman et al., 2006). All other alluvialsurfaces are stratigraphically younger and systematically incisedwith decreasing relative age.

All the studied alluvial surfaces are characterized by gypsic-salicreg soils (Amit and Gerson, 1986; Amit et al., 1993) and showprogressive development with age. These soils were used to estimatethe relative age of the abandoned stable surface and to provide anapproximate age control. The names of the surfaces used here wereassigned locally and no regional correlation is signified by them.

3. Field work and laboratory procedures

In hyperarid regions, pedoturbation is minimal and the silt andclay dust particles that are deposited on top of the surfaces do notinfiltrate below the gypsic horizon (0.3–0.7 m). For luminescencedating, these particles can be avoided by sampling immediatelybelow that horizon and by using only sand-size particles in analyses(Porat et al., 1996; 1997). Past moisture contents in these regions iswell constrained; the presence of salt and gypsum horizons in thereg soils indicates that moisture was always very limited (Amitet al., 2006; Enzel et al., 2008), probably similar to present daymeasured values of 1–3%. U-series disequilibrium is not significantunder such dry conditions.

In all study areas the terrace surfaces have very low gradients,precluding any catenary influences. Well preserved areas on thealluvial fans and terraces were selected on the basis of the pres-ervation and completeness of the reg soil profile. Sediment samplesfor luminescence dating were collected immediately below thegypsic horizon, in trenches dug for palaeoseismological research, orfrom specially excavated pits.

Due to tectonic activity, in places in N. Shehoret, some alluvialfan surfaces are buried by younger fans; the older buried fans wereexposed in trenches excavated into the younger surfaces (Amitet al., 2002) and sampled at the same stratigraphic position, belowthe gypsic horizon. Up to six samples were collected from each

Table 1Average luminescence ages for fan and terrace surfaces in the studied drainage basins.majority of samples were collected in trenches dug for paleoseismic research (Amit et al., 2region, heights above the stream bed are not shown. At Nahal Zin and Nahal Ze’elim, sampSoil age estimates for both wadis are latest Pleistocene to Early Holocene (Amit, 1986; Glusupplementary Table S1 available online.

Surface Height abovestream bed (m)

No. of agedeterminations

Average age(ka)

Nahal Shehoret alluvial fansMiddle Pleistocene (?) terrace 1 >231Qa1* 3 87� 2Qa1 6 62� 5Qa2 6 27� 5Qa3 6 14� 2

Qa4 3 4.6� 0.7Nahal Zin terracesTLi 60 1 22.9� 1.0T0 30 1 17� 1T1 25 1 17� 1T3 18–20 1 14� 1T4 8 1 15� 1Nahal Ze’elim terraces1 65 4 23� 17 25–30 1 18� 18/9 20–25 1 16� 112 10–15 1 10� 1

surface (Table 1 and Table S1), depending on the extent of thesurface and its preservation. In total 27 luminescence ages wereobtained on either quartz or alkali feldspar (see below) from sixalluvial fan surfaces.

In N. Zin and N. Ze’elim the terraces are more limited in extentthan in N. Shehoret. After establishing that four samples froma single surface provide very similar luminescence ages (Table S1),one sample was collected from each well preserved terrace. Toevaluate the degree of signal resetting, a modern sample was takenfrom a shallow stream running on top of one of the terraces. Thisgave an age of 40 years, confirming that sediment transported ontop of a terrace is indeed well bleached. Altogether seven samplesfrom four terraces were dated from N. Ze’elim and five samplesfrom five terraces from N. Zin (Table 1 and Table S1).

Although the sediments are mainly gravels, there was alwayssufficient sand-size quartz or alkali feldspar (KF) present in lensesor beds. The quartz in the N. Shehoret alluvial fans derives fromrecently eroded Precambrian granites and Paleozoic sandstonesoccurring nearby. As a consequence, the quartz has a weak OSLsignal and low sensitivity (Pietsch et al., 2008), resulting in a largescatter of the equivalent dose (De) measurements. In contrast, thequartz in the N. Zin and N. Ze’elim terraces is mostly aeolian dust,brought by dust storms from North Africa, deposited on slopes andthen washed into the fluvial system (Haliva-Cohen, 2004). Its OSLsignal is strong and is highly suitable for dating. Thus, the choicebetween measuring quartz and KF depended on their lumines-cence properties, abundance and level of signal saturation, and ingeneral KF was measured in N. Shehoret and quartz in N. Zin andN. Ze’elim.

The grains were extracted from the sediments using routinelaboratory procedures (Porat, 2007). After sieving for the requiredgrain size and dissolving carbonates with 10% HCl, quartz wasfurther cleaned from heavy minerals and most feldspars withmagnetic separation (Porat, 2006), followed by hydrofluoricetching. KF was isolated using a 2-stage heavy liquid separation(Porat, 2002). Quartz was measured with the single aliquotregenerative dose (SAR) protocol (Murray and Wintle, 2000) usingthe optically stimulated luminescence (OSL) signal, while KF wasmeasured with the single aliquot added dose protocol (SAAD;Duller, 1994; Porat et al., 1996) using the infrared stimulated

Most measurements at Nahal Shehoret were carried out on alkali feldspar and the002). Soil age estimates are from Amit et al. (1996). Due to the tectonic nature of the

les were collected from excavated pits and measurements were carried out on quartz.ck, 2001). Field and laboratory data, dose rates and ages for all samples are listed in

Soil age estimate and other age control Comments

Middle Pleistocene Age underestimationLate Middle Pleistocene to early Late PleistoceneEarly Late Pleistocene to Late Pleistocene Some faulted and buried

Most faulted and buriedLatest Pleistocene to early Holocene.Agrees with fault scarp degradation modeling

Some faulted and buried.

Middle to late Holocene Some faulted and buriedComments:Following shortly after Lake Lisan regression at 25 ka

Following shortly after Lake Lisan regression at 25 ka

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869864

luminescence (IRSL) signal. A few old KF sample (De> 200 Gy)were also measured with the multiple aliquot added dose protocol(MAAD) and IRSL or thermoluminescence (TL) signals.

Dose rate determinations were based on field measurements ofthe g and cosmic dose with a portable gamma scintillator, and oncalculations of the a and b dose rates from the concentrations of theradioelements U, Th and K. The U and Th in the sediment weremeasured by Inductively Coupled Plasma (ICP) Mass Spectrometry(MS) while the K in the sediment and the extracted KF wasmeasured either by ICP Atomic Emission Spectroscopy (AES) orAtomic Absorption Spectroscopy (AAS; Porat and Halicz, 1996).Average water contents over the lifetime of the sediments wereestimated at 1–3%. Internal dose rate was calculated from the Kcontents in the KF-rich extract.

3.1. Partial bleaching

Partial bleaching (or resetting) of the luminescence signals iscaused by insufficient exposure to sunlight during transport of thequartz or KF grains. It is prevalent in fluvial settings, probably dueto the high sediment load and rapid transport during intense floodstypical of arid regions, that generate the most intense sedimenttransport. Much effort has been placed in identifying such condi-tions and devising measurement protocols for luminescence datingthat can identify the most bleached grain population (e.g., Olleyet al., 2004; Arnold and Roberts, 2009). Partial bleaching is recog-nized by the distribution of De values measured for a large numberof aliquots. It is particularly evident when the aliquots are small orwhen single grains are measured (Olley et al., 2004), but whenpartial bleaching is considerable, it can even be identified from

c

Equivalent dose 20 30 40 50

Rel

ativ

e Pr

obab

ility

a

Equivalent dose (Gy)

Equivalen

0 4 8 12 16

Rel

ativ

e Pr

obab

ility

0.0 0.4

00

2000

4000

OSL

(cts

per

0.6

0 s) d

Fig. 2. OSL results. a–c. Single aliquot cumulative probability density functions with individufor (large) single aliquots measured from w10 modern sediment samples from active channeto 13 Gy, and with dose rates of 1.5–2.5 Gy/ky the ages range from a few hundred to a fewvalues< 1 Gy, with ages up to several hundred years. b. Sample ZM-7 from N. Zin showing ty(De¼ 23.5� 6.3 Gy). The minimum age model (Galbraith et al. 1999) gives a De of 20.0� 2.4small error (De¼ 33.1�2.4 Gy). d. Natural OSL signal of a large aliquot from sample ZM-2 frresponse curve for the same aliquot. Recycling ratio (the two dose points at w8 Gy) is 1.02

measurements of a small number of large aliquots (Porat et al.2009). Several approaches have been used to circumvent thisproblem and calculate the De in partially bleached samples, eachisolating the young population and rejecting the old De values bydifferent means (e.g., Olley et al., 1999; Galbraith et al., 1999; Lepperet al., 2000).

KF extracted from modern sediments recently deposited inN. Shehoret show residual IRSL signals that vary substantiallybetween depositional environments. The active channel sedi-ments carry residual De values of 1–4 Gy, equivalent to 500–2000years, and recently abandoned bar sediments may have a residualage of 5000 years (Porat et al., 2001). Evidently, the sand grainsin these coarse sediments are not always fully bleached at thetime of deposition. Fig. 2a shows the De distribution of all single(large) aliquot measurements of modern sediment samples fromactive channels in the hyperarid parts of Israel. The distribution ispositively skewed, as expected from fluvial sediments, and thescatter is high. However, a substantial number of modernsamples have De values of less than one Gy (Fig. 2a, inset). Whenthe minimum age model (Galbraith et al., 1999) is used tocalculate the De of the best-bleached aliquots, a value of0.6� 0.4 Gy is obtained. This model isolates aliquots with lowerDe values that contain grains that were well bleached at the timeof deposition.

For some of the surfaces, the samples’ single aliquot dosedistribution is positively skewed, with a distinct peak at theyounger De values and a tail of older values (Fig. 2b). Based on theresults for the modern analogues, we used the minimum age model(Galbraith et al., 1999) to calculate the De values for the alluvial fanand terrace sediments.

Equivalent dose 20 40

ytilibaborP evitaleR

10 30

t dose (Gy)0.8 1.2

b

Stimulation Time (s)40 80 120

0 8 16 24 320.0

4.0

8.0

Beta Dose (Gy)

xT/xL

al De values and errors for each aliquot plotted in rank order. a. Compilation of De valuesls in desert regions (most samples are from Porat et al., 2001). Residual De values are upthousand years. Inset: The same data showing the large population of aliquots with De

pical De distribution of fluvial sediments with a main peak and a tail of higher De valuesGy c. Sample ZEL-5 from N. Ze’elim, showing a normal dose distribution with relativelyom N. Zin (measured on Risø DA-12 equipped with a filtered halogen bulb). Inset: Dose.

Fig. 3. Comparison between ages obtained for pairs of quartz and alkali feldsparextracted from the same sample. Quartz was measured by the single aliquot regen-erative dose protocol (Murray and Wintle, 2000) and alkali feldspar by the singlealiquot added dose protocol (Duller, 1994). Some samples are from other surfaces notpresented here. Squares – N. Paran (central Negev); circles – southern Arava. Dashedline is the 1:1 ratio.

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869 865

For older alluvial fans, any indication of partial resetting of theOSL signal at the time of deposition has been masked by the largeDe values. Fig. 2c presents De values for 19 large aliquots from N.Ze’elim sample ZEL-5, measured on quartz using SAR with a meanof 33.1�2.4 Gy. The relatively large errors on individual measure-ments potentially mask any partial bleaching. However, the dosedistribution is normal and the scatter small and there are noaliquots with substantially higher De values.

4. Age results

The OSL signal in the quartz samples from N. Zin and N. Ze’elimis dominated by the fast component (Fig. 2d), implying that the SARprotocol is suitable for measuring the De values. These do not varyas a function of preheat temperatures within the range of220–260 �C, therefore all measurements were carried out using thistemperature range. Recycling ratios were mostly within 10% ofunity (Fig. 2d), showing that the test dose corrects appropriately forsensitivity changes.

Concerns about anomalous fading of KF (Huntley and Lamothe,2001) in N. Shehoret were addressed by comparing pairs of quartzand KF extracts from the same sample. There was no systematicdifference between the quartz and KF ages and the data are mostlydistributed around the 1:1 line (Fig. 3). As no systematic underes-timation was observed, the KF ages were considered to be reliable.

The luminescence ages range from w4 ka to >230 ka, at theupper limit of the luminescence dating method. Field and laboratorydata for individual samples are listed in the Table S1A–C. Averagedages for individual landform surfaces in the different study areas arepresented in Table 1, together with the available age control.

4.1. N. Shehoret

Six alluvial fan surfaces were dated, mostly by SAAD and theIRSL signal (Table S1), and their average ages range from>230 ka to4.6 ka (Fig. 4 and Table 1). In samples for which both quartz and KF

were measured, it appears that the IRSL ages on KF ages aresomewhat older than the OSL ages on quartz (Table S1; Porat et al.2009), and overall the KF ages are stratigraphically more consistent.

The oldest alluvial fan at the highest morphostratigraphicposition, with a well-developed gypsic horizon and advancederosion of the surface attesting to its antiquity (Amit et al., 1995),was dated by MAAD IRSL and TL of KF to 231�22 ka (Table S1).Both signals were close to saturation and this age is considereda minimum estimate for this surface, which could be of EarlyPleistocene age.

The upper sediments of the Qa1* surface, positioned slightlyabove Qa1 (see below), were dated in Trench T-8 from 85� 4 ka to91�6 ka, with an average of 87�2 ka (Fig. 4, Table S1). Similarages, in the range of 90–100 ka, were obtained for samplescollected at the depth of the Qa1 surface in trenches T-6 and T-17further to the east (Fig. 4; Porat et al., 1997) and this time periodprobably represents an early stage of fan accumulation.

Surface Qa1, the most extensive alluvial fan in the Shehoret area,was dated at four locations (trenches T-2, T-6, T-10, T-17) and theages for six samples range from 70� 6 ka to 56� 4 ka, with anaverage of 62� 5 ka (Fig. 4). The ages increase upfan from east towest, expressing the diachronous nature of this alluvial surface.Similar ages, of 61�9 ka to 64�11 ka were obtained from gravelbeds below the base of the reg soil profile (Porat et al., 1997).

Surface Qa2 is exposed over only a small area on the westernpart of the N. Shehoret alluvial fan, where was dated by IRSL to34� 3 ka (Fig. 4). A faulted and buried surface located further to theeast had similar reg soil characteristics (Amit et al., 1996) and wasthus correlated to the exposed Qa2 surface. It was dated in trenchesT-18 and T-19 from 24� 5 ka to 30� 7 ka (Table S1). The averageIRSL age for Qa2, for both the exposed and buried surface, is27�4 ka.

Surface Qa3 was dated at five different localities (trenches T-6,T-19, and several channel outcrops) resulting in a fairly tight agerange of 12� 4 ka to 17� 3 ka, averaging 14� 2 ka. Here again theages increase upfan. A sample from surface Qa4, which coversa limited area and is exposed along the modern channel, was datedto 5.1�0.5 ka. This young surface was also detected in trench T-19,buried under sub-recent fluvial sediments, and there it was datedto 4.1�0.8 ka (Table S1; Fig. 4). Thus, the average age for surfaceQa4 is 4.6� 0.7 ka.

4.2. N. Ze’elim and N. Zin

The rapid fall of Lake Lisan levels after about 25 ka (Bookmanet al., 2006.) gives a chronological constraint for alluvial fan surfaceformations in N. Zin and N. Ze’elim basins (about 45 km apart),since both would have responded to the same receding shorelines.The young, shallow reg soils enabled us to collect samples froma depth of only 0.3–0.35 m, and our consistent ages demonstratethat the surfaces were not disturbed by pedoturbation.

Fourteen alluvial fan surfaces were mapped at N. Ze’elim(Fig. 5a; Bowman, 1975; Amit, 1986) and analyzed for stages of soilformation over time, while determining indicators for relative agedating (Amit and Gerson, 1986). All these surfaces postdate theinitial receding of Lake Lisan from its highest stand at about 25 ka.Four un-eroded surfaces, with well preserved reg soil profiles, wereselected for sampling (in pits; Fig. 5a; Table S1). A Lake Lisan beachridge and the correlated terrace Qa1, were dated by four sampleswith ages ranging from 24� 2 ka to 22� 2 ka, averaging 23�1 ka(Table S1). The elevation and age of the beach ridge sample fall onthe Lake Lisan level curve presented in Fig. 6, and the terrace agescorrespond well to the initial regression of Lake Lisan after 25 ka.On this uppermost terrace an end-of-Pleistocene to Holocene soildeveloped (Amit, 1986).

Fig. 4. Oblique aerial photo of the alluvial fan surfaces at N. Shehoret with some of the luminescence ages (white, in ka) of the surfaces. Ages of the buried surfaces are not shown.Trenches T-2 to T-19 are shown as short dashed black lines.

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869866

Alluvial fan surface Qa7 yielded an age of 18� 2 ka, indi-cating that surfaces Qa1–Qa7 were deposited within about5 ky. Surfaces Qa8/9 and Qa12 yielded ages of 18� 2 ka and10�1 ka, respectively (Table S1). The reg soil profiles show

22.9 1.0

Qa1Qa3 Qa4

Qa6

450 5001000

-180

-200

-220

-240

Hei

ght (

m)

N

50 550

N. Ze’elim+

350 400 4500 50 100

-200

WNW

T1T3

T0

22.2 1.9TLi

-220

-240

-260

Hei

ght (

m) 18.1 2.1

N. Zin

18.2 1.7+

+

+

a

b

Fig. 5. Sequences of terraces in two basins draining into the Dead Sea, with OSL ages (in kalabeled Qa1 to Qa12. b. a cross section at N. Zin (modified from Gluck, 2001). The different

minor changes between surfaces indicating that the timeelapsed since the deposition of the sequence of alluvial fansdid not vary much to allow for any substantial differencesbetween the surfaces.

17.5 1.5

18.0 2.510.3 1.4Qa7

Qa9 Qa10Qa12

800 900 1000

S

750 850 950

+

++

500 550 600

ESE

T4 T6

15.2 2.0

14.2 1.6+

+

). a. a cross section at N. Ze’elim (modified from Amit, 1986). The different terraces areterraces are labeled TLi, T0 to T6. The lowermost terraces are not shown on figure.

Age (ka)0 20 40 60 80 100

ytilibaborP evitaleR

Qa4 Qa3 Qa1 Qa1*Qa2

Fig. 7. Distribution of ages of alluvial fans in N. Shehoret showing the average age foreach surface. There is very little overlap between different alluvial fans.

Fig. 6. Ages of the terraces from N. Zin (squares) and N. Ze’elim (diamonds) plottedagainst their topographic heights. Note the similar trends. The curve on the rightdepicts the fall of the Lake Lisan level during that time span (modified from Bookmanet al., 2006). The oldest sample from N. Ze’elim is from a Lake Lisan beach ridge(marked in black).

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869 867

At the lower reaches of N. Zin, close to the Dead Sea, a series often alluvial fan terraces and terrace remnants was mapped (Fig. 5b;Gluck, 2001). Similar to the N. Ze’elim alluvial fan, these surfacespostdate the Lake Lisan regression at ca. 25 ka. Five different, wellpreserved surfaces were sampled from pits at an area where theuppermost terrace minimally incises the highest level of Lake Lisanshoreline. Its age should be somewhat younger than the initial lakeregression after 25 ka (Gluck, 2001). The 22� 2 ka age obtained forthis terrace (TLi,, Table S1, Fig. 5b), is therefore in agreement withthe predicted age. The ages of the lower terraces decreasesystematically according to morphostratigraphy from 18� 2 ka forT0 to 15� 2 ka for T4 (Figs. 5b and 6). A younger alluvial fan (Qf1)deposited by a small tributary on top of terrace T4 also yielded anage of 14�1 ka. Davis et al. (2009) present OSL ages from terrace T0

located further downstream on N. Zin. Their ages, of 16� 3 ka and18� 3 ka, are in a good agreement with our results. The degree ofsoil formation on these terrace surfaces indicates that they allformed during the end of the Pleistocene and the Holocene(Gluck, 2001).

5. Discussion

The luminescence ages for alluvial fan and terrace surfacespresented here can be compared to other age indicators for thesame surfaces. At N. Shehoret the ages of the alluvial fans conformvery well to the morphostratigraphy and correlate to earlier ageestimations based on soil indices such as the degree of desertpavement formation, thickness of gypsic horizon and intensity ofsalic horizon (Table 1; Amit et al., 1995; 1996). The age of Qa3 fan,14� 2 ka, which formed as a result of faulting, accords well withfault scarp degradation modeling age estimates of<15 ka (Enzelet al., 1996). Multiple ages from a single alluvial surface yielda discrete and narrow age range and cluster to within �10–20% ofthe average. Ages from single alluvial fans are distinguishable andrarely overlap ages from younger or older fans (Fig. 7). The scatterwithin a single alluvial surface results from the complex processesand the time span needed to form an alluvial fan.

The ages for the wadi terraces at N. Ze’elim and N. Zin are mostlyconsistent with the morphostratigraphic framework. Fig. 6compares the luminescence ages from the main terraces to theLisan Lake level curve (Bookman et al., 2006) which is based onradiocarbon and U–Th ages for the same time span (25–10 ka).Terrace ages should be younger than the lake level for the sameheight, and their elevation higher than the contemporaneous lakelevel. The ages for the uppermost terrace, formed soon after LakeLisan receded, conform well to the age of the initial rapid drop inlake level at w25 ka, attesting to the reliability of the luminescenceages. The luminescence chronologies of the terraces at both wadisare remarkably similar and the dated terrace surfaces closely followthe trend of lake level fall shown by Bookman et al. (2006). Thecluster of ages at 22–23 ka and at 18–16 ka corresponds to episodeswhen lake levels remained more or less constant (Fig. 6). It appearsthat the terraces formed roughly 1–5 ky after the lake receded andwere abandoned with a lag of a few thousand years.

The samples with oldest ages from these sequences of terraces(22–24 ka) do not show any evidence for partial bleaching and thesource of quartz must have been wind blown dust deposited in thedrainage basins and soon after washed into the wadis (Haliva-Cohen, 2004). However, as Lake Lisan shorelines receded, canyonsdeveloped within the soft sediments of the Lisan Fm., and LatePleistocene sediments were recycled into the terraces with insuf-ficient exposure to sunlight during transport. This is gleaned fromthe growing scatter in the De distribution down the morphos-tratigraphy, and the larger fraction of poorly bleached aliquots inthe younger terraces (Table S1). While the sediments in the terracesdo not suffer from bioturbation, and as such still reflect the time ofdeposition, the fluvial depositional environment requires thatsmall aliquots and single grains should be measured if accuratedepositional ages are to be obtained.

6. Conclusions

We have shown that luminescence dating of geomorphicsurfaces can be used in hyperarid areas, where other dating tech-niques are not always applicable. Luminescence dating of the top ofwell preserved surfaces produced systematic chronologies whichare in agreement with morphostratigraphy, reg soil formation andwith known events such as drops in the Lisan lake level. All thew40 luminescence ages obtained in this context provided robustage sequences. Physically correlated alluvial fan surfaces in N.Shehoret yielded similar ages and the series of alluvial fans gaveages consistent with their site-specific morphostratigraphy. Theluminescence chronostratigraphy for sequences of terraces in twowadis in the Dead Sea basin, controlled by the same base level, ishighly correlated. The ages agree well with prior age estimations

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869868

based on lake levels, degree of soil development and fault scarpdegradation modeling.

Although the luminescence methods date the age of the lastexposure of mineral grains to sunlight, and therefore give the age ofthe last deposition of the sediments underlying the surface, in thehyperarid regions these ages are close in time to the actual event oflandform abandonment.

Advances in luminescence dating techniques, such as singlegrain measurement, will provide more precise and accurate ages.Surface dating in hyperarid regions will greatly benefit from theregular use of the luminescence methods.

Acknowledgements

This research was partially supported by grant No. 72/96 fromthe Israeli Science Foundation to NP, YE and RA and a grant fromIsraeli Atomic Energy Commission to NP. We thank R. Madmon fortechnical support in the field and S. Zeffren for laboratory work. Wethank three anonymous reviewers whose comments improved themanuscript.

Appendix. Supplementary data

Supplementary data associated with this article can be found inthe online version at doi:10.1016/j.jaridenv.2009.10.018.

References

Aitken, M.J., 1998. An Introduction to Optical Dating. Oxford University Press,Oxford.

Amit, R., 1986. Shattered gravel in desert reg soil – the effect of salts on the natureand rate of the weathering process. Ph.D. Thesis, The Hebrew University ofJerusalem (in Hebrew; English. Abstract), 238pp.

Amit, R., Gerson, R., 1986. The evolution of Holocene reg (gravelly) soils in deserts –an example from the Dead Sea region. Catena 13, 59–79.

Amit, R., Gerson, R., Yaalon, D.H., 1993. Stages and rate of gravel shattering processby salts in desert reg soils. Geoderma 57, 295–324.

Amit, R., Harrison, J.B.J., Enzel, Y., 1995. Use of soils and colluvial deposits inanalyzing tectonic events – the southern Arava Rift, Israel. Geomorphology 12,91–107.

Amit, R., Harrison, J.B.J., Enzel, Y., Porat, N., 1996. Soils as a tool for estimating ages ofQuaternary fault scarps in a hyperarid environment – the southern Arava valley,the Dead Sea Rift, Israel. Catena 28, 21–45.

Amit, R., Zilberman, E., Porat, N., Enzel, Y., 1999. Relief inversion in the Avrona playaas evidence of large-magnitude historical earthquakes, southern Arava Valley,Dead Sea Rift. Quaternary Research 52, 76–91.

Amit, R., Zilberman, E., Enzel, Y., Porat, N., 2002. Paleoseismic evidence for timedependency of seismic response on a fault system in the southern Arava Valley,Dead Sea rift, Israel. Geological Society of America Bulletin 114, 192–206.

Amit, R., Enzel, Y., Sharon, D., 2006. Permanent quaternary aridity in the southernNegev, Israel. Geology 34, 509–512.

Arnold, L.J., Roberts, R.G., 2009. Stochastic modelling of multi-grain equivalent dose(De) distributions: Implications for OSL dating of sediment mixtures. Quater-nary Geochronology 4, 204–230.

Barzilai, E., Enzel, Y., Amit, R., 2000. Constructing synthetic time-series of rainfallevents for environmental modeling in a hyperarid environment, SouthernArava, Israel. International Association of Hydrological Sciences (IAHS) Publi-cation 261, 29–42.

Bateman, M.D., Frederick, C.D., Jaiswal, M.K., Singhvi, A.K., 2003. Investigation intothe potential effects of pedoturbation on luminescence dating. QuaternaryScience Reviews 22, 1169–1176.

Bateman, M.D., Boulter, C.H., Carr, A.S., Frederick, C.D., Peter, D., Wilder, M., 2007.Preserving the palaeoenvironmental record in Drylands: bioturbation and itssignificance for luminescence-derived chronologies. Sedimentary Geology 195,5–19.

Birkeland, P.W., Machette, M.N., Haller, K.M.,1991. Soils as a tool for applied Quaternarygeology. Utah Geological and Mineral Survey Misc. Publication, 91–93.

Bookman, R., Bartov, Y., Enzel, Y., Stein, M., 2006. Quaternary lake levels in the DeadSea basin: two centuries of research. In: Enzel, Y., Agnon, A., Stein, M. (Eds.),New Frontiers in Dead Sea Paleoenvironmental Research: Geological Society ofAmerica Special Paper, 401, pp. 155–170.

Bowman, D., 1975. River terraces in the Dead Sea area – morphology and genesis.Ph.D., dissertation, The Hebrew University of Jerusalem (in Hebrew, Englishsummery).

Bowman, D., 1997. Geomorphology of the Dead Sea Western Margin. In: Niemi, T.M.,Ben Avraham, Z., Gat, J.R. (Eds.), The Dead Sea: The lake and its setting. OxfordUniversity Press, New York, pp. 217–225.

Bull, W.B., 1991. Geomorphic Response to Climatic Changes. Oxford University Press.Bush, D.A., Feathers, J.K., 2003. Application of OSL single-aliquot and single-grain

dating to quartz from anthropogenic soil profiles in the SE United States.Quaternary Science Reviews 22, 1153–1159.

Clapp, E.M., Bierman, P.R., Schick, A.P., Lekach, J., Enzel, Y., Caffee, M., 2000. Differingrates of sediment productions and sediment yield determined by cosmogenicisotopes and conventional methods in extremely arid environments, Nahal Yael,Israel. Geology 28, 995–998.

Davis, M., Matmon, A., Zilberman, E., Porat, N., Gluck, D., Enzel, Y., 2009. Bathymetryof late Pleistocene Lake Lisan determines stream incision in response toclimatically induced lake level fall. Geomorphology 106, 352–362.

Duller, G.A.T., 1994. Luminescence dating using feldspars: a test case from southernNorth Island, New Zealand. Quaternary Science Review (Quaternary Geochro-nology) 13, 423–428.

Enzel, Y., Amit, R., Porat, N., Zilberman, E., Harrison, J.B.J., 1996. Estimating the agesof fault scarps in the Arava, Israel. Tectonophysics 253, 305–317.

Enzel, Y., Amit, R., Dayan, U., Crouvi, O., Kahana, R., Ziv, B., Sharon, D., 2008. Theclimatic and physiographic controls of the eastern Mediterranean over the latePleistocene climates in the southern Levant and its neighboring deserts. Globaland Planetary Change 60, 165–192.

Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., Olley, J.M., 1999. Opticaldating of single and multiple grains of quartz from Jinmium rock shelter,northern Australia: Part I. Experimental design and statistical models.Archaeometry 41, 339–364.

Gerson, R., Grossman, S., Amit, R., Greenbaum, N., 1993. Indicators of faulting eventsand periods of quiescence in desert alluvial fans. Earth Surface Processes andLandforms 18, 181–202.

Gluck, D., 2001. The landscape evolution of the southwestern Dead Sea basin andthe paleoseismic record of the southwestern marginal fault of the Dead Seabasin and of the Carmel Fault during the Late Pleistocene and Holocene. M.Sc. Thesis, The Hebrew University of Jerusalem, (In Hebrew; Englishabstract).

Gosse, J.C., Phillips, F.M., 2001. Terrestrial in situ cosmogenic nuclides: theory andapplication. Quaternary Science Reviews 20, 1475–1560.

Granger, D.E., Muzikar, P.F., 2001. Dating sediment burial with in situ-producedcosmogenic nuclides: theory, techniques and limitations. Earth and PlanetaryScience Letters 188, 269–281.

Haliva-Cohen A., 2004. Mineralogy, petrography and isotopic ratios of fine-graindetritial sediments of the Dead Sea Group: origin and transportation. M.Sc.thesis, The Hebrew University of Jerusalem, and GSI Report GSI/28/04, 88 p.(Hebrew, English abstract).

Hancock, G.S., Anderson, R.S., Chadwick, O.A., Finkel, R.C., 1999. Dating fluvialterraces with 10Be and 26Al profiles: application to the Wind River, Wyoming.Geomorphology 27, 41–60.

Harvey, A.M., Wells, S.G., 2003. Late quaternary variations in alluvial fans sedi-mentologic and geomorphic processes, Soda Lake basin, eastern Mojave Desert,California. GSA Special Paper 368, 207–230.

Harrison, J.B.J., McFadden, L.D., Weldon II, R.J., 1993. The influence of colluvialdeposition on rates of soil development in the transverse ranges, SouthernCalifornia. Israel Journal of Earth-Sciences 41, 139–154.

Helms, J.G., McGill, S.F., Rockwell, T.K., 2003. Calibrated, Late Quaternary age indicesusing clast rubification and soil development on alluvial surfaces in Pilot KnobValley, Mojave Desert, Southeastern California. Quaternary Research 60,377–393.

Horowitz, A., 2001. The Jordan Rift Valley. Balkema, Rotterdam – Boston, UnitedStates.

Huntley, D.J., Lamothe, M., 2001. Ubiquity of anomalous fading in K-feldspars andthe measurement and correction for it in optical dating. Canadian Journal ofEarth Sciences 38, 1093–1106.

Lepper, K., Larsen, N.A., McKeever, S.W.S., 2000. Equivalent dose distribution anal-ysis of Holocene eolian and fuvial quartz sands from Central Oklahoma. Radi-ation Measurements 32, 603–608.

Liu, B., Phillips, F.M., Pohl, M.M., Sharma, P., 1996. An alluvial surface chronologybased on cosmogenic 36Cl dating, Ajo Mountains (Organ Pipe Cactus NationalMonument), southern Arizona. Quaternary Research 45, 30–37.

Matmon, A., Cruvi, O., Enzel, Y., Bierman, P., Larsen, J., Porat, N., Amit, R., Caffee, M., 2003.Complex exposure histories of chert clasts in the Late Pleistocene shorelines of LakeLisan, southern Israel. Earth Surface Processes and Landforms 28, 493–506.

McDonald, E.V., McFadden, L.D., Wells, S.G., 2003. Regional response of alluvial fansto the Pleistocene–Holocene climatic transition, Mojave Desert, California. GSASpecial Paper 368, 189–206.

Mee, A.C., Bestland, E.A., Spooner, N.A., 2004. Age and origin of Terra Rosa soils inthe Coonawarra area of South Australia. Geomorphology 58, 1–25.

Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using improvedsingle-aliquot regenerative-dose protocol. Radiation Measurements 32, 57–73.

Noller, J.S., Sowers, J.M., Lettis, W.R. (Eds.), 2000. Quaternary Geochronology –Methods and Applications. American Geophysical Union Reference Shelf 4.AGU, Washington DC.

Olley, J.M., Caitcheon, G.G., Roberts, R.G., 1999. The origin of dose distribution influvial sediments, and the prospect of dating single grains from fluvialdeposits using optically stimulated luminescence. Radiation Measurements30, 207–217.

N. Porat et al. / Journal of Arid Environments 74 (2010) 861–869 869

Olley, J.M., Pietsch, T., Roberts, R.G., 2004. Optical dating of holocene sedimentsfrom a variety of geomorphic settings using single grains of quartz. Geomor-phology 60, 337–358.

Pietsch, T.J., Olley, J.M., Nanson, G.C., 2008. Fluvial transport as a natural lumines-cence sensitiser of quartz. Quaternary Geochronology 3, 365–376.

Porat, N., 2002. Analytical procedures in the Luminescence dating laboratory.Geological Survey of Israel Technical Report TR-GSI/2/2002 (in Hebrew),44 p.

Porat, N., 2006. Use of magnetic separation for purifying quartz for luminescencedating. Ancient TL 24, 33–36.

Porat, N., 2007. Analytical procedures in the Luminescence dating laboratory.Geological Survey of Israel Technical Report TR-GSI/2/2007 (in Hebrew),p. 42.

Porat, N., Halicz, L., 1996. Calibrating the luminescence dating laboratory. GeologicalSurvey of Israel Current Research 10, 111–116.

Porat, N., Wintle, A.G., Amit, R., Enzel, Y., 1996. Late Quaternary earthquake chro-nology from luminescence dating of colluvial and alluvial deposits of the Aravavalley. Israel. Quaternary Research 46, 107–117.

Porat, N., Amit, R., Zilberman, E., Enzel, Y., 1997. Luminescence dating of fault-related alluvial fan sediments in the southern Arava valley, Israel. QuaternaryGeochronology (Quaternary Science Review) 16, 397–402.

Porat, N., Zilberman, E., Amit, R., Enzel, Y., 2001. Residual ages of modern sedimentsin an hyperarid region. Israel. Quaternary Science Review 20, 795–798.

Porat, N., Duller, G.A.T., Amit, R., Zilberman, E., Enzel, Y., 2009. Recent faulting in thesouthern Arava, Dead Sea Transform: Evidence from single grain luminescencedating. Quaternary International 199, 34–44.

Putkonen, J., Swanson, T., 2003. Accuracy of cosmogenic ages for morains. Quater-nary Research 59, 255–261.

Repka, J.L., Anderson, R.S., Finkel, R.C., 1997. Cosmogenic dating of fluvial terraces,Fremont River, Utah. Earth and Planetary Science Letters 152, 59–73.

Rittenour, T., 2008. Luminescence dating of fluvial deposits: applications togeomorphic, paleoseismic and archaeological research. Boreas 37, 613–635.

Rockwell, T.K., Vaughan, P.R., Bickner, F., Hanson, K.L., 1994. Correlation and ageestimates of soils developed in marine terraces across the San Simeon faultzone, Central California. GSA Special Paper 292, 151–166.

Schildgen, T., Detier, D.P., Bierman, P., Caffee, M., 2002. 26Al and 10Be dating of LatePleistocene and Holocene fill terraces: a record of fluvial deposition and inci-sion, Colorado Front Range. Earth Surface Processes and Landforms 27, 773–787.

Sharp, W.D., Ludwig, K.R., Chadwick, O.A., Amundson, R., Glaser, L.L., 2003. Datingfluvial terraces by 230Th/U on pedogenic carbonate, Wind River Basin, Wyom-ing. Quaternary Research 59, 139–150.

Watchman, A.L., Twidale, C.R., 2002. Relative and ‘absolute’ dating of land surfaces.Earth Science Reviews 58, 1–49.

Wells, S.G., McFadden, L.D., Dohrenwend, J.C., 1987. The influence of Late Quater-nary climatic change on geomorphic and pedogenic processes on a desertpiedmont, Eastern Mojave Desert, California. Quaternary Research 27, 130–146.