MEMOIR GEOLOGICAL SOCIETY OF INDIA No.42, 1999, pp. 273-295 MAYANK JAIN AND OTHERS

Alluvial and Aeolian Sequences Along the River Luni, Barmer District: Physical Stratigraphy and Feasibility of Luminescence Chronology Methods

INTRODUCTION

Fluvial response to climatic, eustatic, and tectonic processes is significant for understanding the stratigraphic record, palaeoenvironments, and palaeocIimate. Earlier studies have highlighted the impact of such changes on landscape development in western Indid during the Quaternary, and it has MAYANK JAW S.K. TANDON, S.C. BHAIT, A.K. SINGHVI' and S H MISIIRA? ~ ~ ~been ~noted that changes in climate have influenced the relative dominance of fipartment of Geology, University of Delhi, ~ e ~ h i - 1007 10 aeolian and fluvial processes in the region (e.g. Merh and Chamyal, 1997; Kar, 'Physical Research Laboratory. Navarangpura, Ahmedabatl- 380 009 1995 and references therein). Despite this, integrated studies on the stratigraphic, 'Deccan College , Punc - 41 1 006 sedimentological, and chronological aspects of the Quaternary alluvial sequences of Rajasthan are lacking. So far, few studies have been carried out Abstract on Indian semi-arid fluvial basins (for e.g. Allchin et al. 1978; Kale and I n the lower Luni basin (Barmer district), Quaternary sediments (?), reaching Rajaguru, 1987; Merh and Chamyal, 1997; Pant and Chamyal, 1990; Tandon a maximum thickness o f -300 m. unconformahly ovcrlic a basement o f et al. 1997; Zeuner, 1950). Precamhrian rhyolites and granites. Stratigraphic measurements and facies Present study attempts to establish a physical stratigraphy for the descriptions o f six sections near Sindari, Bhuka, Karna and Khudala have bcen Quaternary alluvial sequences, exposed along the banks of River Luni, in the made to establish a physical stratigraphy o f exposcd a l l u v i d and acolian main outcrop belt between Karna and Khudala, Barmer district (Fig.1). As sequcnccs. This enabled identification o f lhrce types o f lithological scqucnccs termed as Type-], Typc-Il and Type-Ill sequences. Facies analysis o f these these aeolian and alluvial facies, and associated paleosols/calcretes are important sequcnces indicatc deposition i n arid/scmi-arid environments. tools for paleoenvironmental/ paleoclimatic reconstruction, a feasibility study Type-I SindarilBhuka scquencc represcnts deposits o f semi-arid gravel-sand towards luminescence chronology of these sequences has also been attempted. bedload streams and slackwater dcposits with incipient pcdogcncsis i n wcakly Changes in the fluvial styles, paleosols and relative dominance of aeolian devclopcd floodplains. Significant aeolian and fluvially rcworkcd ncolian deposits are recognised i n 'Qpe-I1 scqucnccs. Thcsc conlprisc implcmcntifcrous pcbbly dynamism have been used as proxy indicators of ambient climatic conditions sandstonesand rcworked-acolian/acoliansediments punctuatcd hy gravels deposited during the depostion of these sequences. by scmi-stable episodic flows with signilicant lateral transport. Pottery bearing pcbhly coarsc sands and ovcrbank silts arc idcntificd i n Type-111 scqucnces. A feasibility study o f lumincscence chronology using opticAly stimulated lumincscence (OSL) mcthotls has hecn auemptcd. Coarse grain quartz and fclspars (100 - 150 pm) were analyscd by additive dose, rcgcneration and Australian slide tcchniqucs. Analyses o f modern Luni channel sand sample gave an age o f 75 ycars indicating that thc assumption o f prc-depositional zeroing o f lumincsccncc signal i n quartz is rcasonablc. Despilc this. thc an~rlyscsproved n ~ n - t r i v i aon l account o f : 1 ) mid-term fading i n fclspars. 2) coniplcx non-linear growth curve o f quartz and 3) disc to disc variability i n thc lumincscencc output o f identical, short shine norm;~liscdaliquots o r quart/ separates. Nonetheless. Australian slidc method has cnablcd a first ordcr eslimatc o f ages iri [he Type [I scqucnccs. Thcse ages of around 70 ka arc consistent with thc prcscnce o f bliddlc Paleolithic artcfacts in them. O n the basis o f lumincsccncc rncasurcments and archacologic;~levidcnccs. Type 1. Type IIand Typc Illscquenccs arc tcntntively assigned to E x l y to Middle Pleistoccnc, early to middle Latc Plcistoccne and Holoccne rcspec~ivcly.Type I1 (< 100 l a ) scquencc5 reflect prcporidcrance o f aridity and dominance o f ncolian dynamism with possible climalic amelioration around 70 kt. T! pe 1 scqucnccs arc probably older than 400 ka and rcprcscnt deposition undcr rclntivcly wetter conditions. Type 111 pottcry bcarinp pebbly coarse sands rcpl-cscnr wcttcr climatc i n mid- Holoccnc (- 5 ka). and providc important constraint on the ;~gco f incision1 exhumat~ono f present Luni gorges ill the study arcn.

GEOLOGIC SElTINC

Pre-Quaternary stratigraphy of Rajasthan consists of Archean Banded Gneissic Complex, Proterozoic Aravalli Supergroup, Delhi Supergroup, postDelhi effusives and intrusives represented by Malani Igneous Suite, and Manvar Supergroup (Pareek, 1984). In the study area, Quaternary (?) continental deposition has largely occurred over a basement of Precambrian rhyolites and granites. These deposits reach a maximum thickness of up to 100 m in the lower Luni basin as inferred from the available sub-surface data. Ho\vever, in the absence of any chronological markers, inclusion of these sub-surface deposits in the Quaternary is tentative. River Luni and its tributaries constitute the major drainage system in the Thar. Luni emerges from the Aravallis, flows westward t i l l Tilwara (near Balotra) where it takes a sharp southerly bend, and debouches finally into the Arabian sea (Fig. 1). In the study area, Luni flows through rhyolite hills and dunal sands. In general, fluvially formed landscape includes hill - rocky/gravelly pediment - buried pediment - older alluvial plain - younger alluvial plain river bed (Kar 1995). Other important geomorphic features of the area are sand dunes which had peak accumulation around 14 ka t i l l 10 ka and stabilized during the Holocene (Chawla et al. 1992). Satellite imagery indicate prominent

STRATIGRAPHY 8r FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 275

Fig.1. Outline map of river Luni. Study area lies in the shaded sectors along the river between Balotra and Sindari.

palaeodrainages, notable amongst which is the Lik palaeodrainage in the study area (Kar, 1988).

FACIES DESCRIPTIONS

Detailed documentation of six sections (Fig. 2) was camed out in the outcrop belt between Karna and Khudala, along the banks of the River Luni. Depositional environments were determined on the basis of detailed documentation of facies, facies associations, and spatial integration of facies variations (Reading, 1986; Miall, 1997). Vertical graphic facies logs at different sections and bedding diagrams between the sections in critical areas were prepared for understanding the two dimensional organisation of facies. Measurements for bedding diagrams were done using a meter tape, and taking ambient water level in the river as the base level. Grain size, mineralogy and roundness were estimated by visual methods. Lithofacies characteristics, facies associations and contact relationship between the lithological sequences (in some sections) allow their classification under three broad categories (Fig.2) : ( I ) Type I sequence: Sindari and Dhuka (BH I and BH 11) sequcncrs.

(2) D p e I1 sequences: Thcsc comprise scqucnces in Karna section. Bhuka section (BH I ) and Khudala section. (3) Type 111sequences: Surface Gravel exposures near Karna. upper part of Bhuka section ( B H I ) , younger silts in the Karna sequence, upper part of Khud.11a section.

STRATIGRAPHY & FEASIBLITY OF LUMINESCENCE CHRONOLOGY METHODS 277

Different lithofacies/facies associations in these sequences are described low:

B*" ;Type I sequence F

sin&ri/Bhuka type sequence

4

This sequence is exposed along the banks of Luni for more than 12 km kStretch between Sindari and Bhuka (Fig 2). It consists of alternations of iconglomerates, pebbly sandstone and sandstone interbedded with mudstones !which show varying degree of incipient pedogenesis. The sequence is generally dominated by a heterolithic facies consisting of reddish brown mudstone and sandstone. Internal and lateral organisation of this facies is varied, and includes both sandstone-dominated and mudstonedominated intervals. Lateral facies and thickness variations are pronounced. In the Bhuka-Lohida stretch, a locally well developed braid channel facies association is present within the heterolithic member. Depending on the extent of development of this member, the heterolithic facies occurs either as a composite unit or is divisible into an upper and a lower unit. Incipient pedogenesis in the heterolithic facies is indicated by profuse rhizocretions, mottling, gleying and reddish brown colours Laterally this facies grades into vertisols in which the intercalated sand layers are either absent or occur in a very minor proportion. At Bhuka, the heterolithic facies is very well developed and is more sand dominated. In some sections, lithofacies in the upper part of the section include large scale trough cross-bedded rhizocretionary pebbly gritty sandstone. These, in turn, are overlain by poorly cross-bedded sheets of disorganised conglomerates. Some evidences of syndepositional tectonism have been found in the Type I sequences (Jain, unpublished data).

MAYANK JAlN AND OTHERS

ill sorted with angular to sub-rounded clasts of quartz, subrounded clasts locally derived caliche nodules, very angular rhyolite and other rock fragmen Clasts are supported in a matrix of gritty coarse sand composed of quartz felspar and rock fragments. This unit comprises cosets of large scale plan cross beds with minor channel fills in its upper parts. In places, for iqtance ih the Khudala section, this unit shows fining up from pebbly gritty sands to dominantly coarse sands. In the Karna section, cross-bedded pebbly sandstone shows interfingering relationship with thick (up to 1.5 m) mottled, rhizocretion calcrete nodule bearing, red to brown muds/silts. Laterally, this unit cant abundant dark coloured carbonate concretions (black pebbles) which disso in dilute HCI leaving behind a black residue. Whole unit is moderately cement Here the sandstone has yielded in-situ flakes which show prepara and retouching diagnostic of Middle Palaeolith, implying an age of between 40-200 ka. These artefacts do not show any sign of reworking or long distance transportation. Several flakes and ostrich egg shells are also found on top of the overlying sandy unit. I11 sorted nature of the sandstone, presence of locally derived detritus and sheet like disposition of individual beds represent episodic high magnitude sheet flow events. Very angular clasts of rhyolite indicate pediment derived detritus from the nearby rhyolitic hills. Fining upward trend results from waning flows and associated muds are deposited under slackwater conditions. Black pebbles are a striking feature of many calcretes. They are common in carbonates from aeolian (Ward, Folk and Wilson, 1970), lacustrine (Freytet, 1973) and pedogenic (Estaban and Klappa, 1983) settings and probably represent input of stablised organic matter from vegetated hinterland (Strasser and Davaud, 1983; Strasser, 1984). Well sorted nledium to fine sands/ calcrete nodule rich nzassive gravel association

Type I1 sequences

Type TI sequences are identified in Karna, Khudala, and Bhuka (BH I) sections. These sequences are dominated by pebbly sandstones (implementiferous), well sorted medium to fine sands, silts and calcrete bearing gravels.The general trend in these sequences is - presence of a lower cross bedded pebbly coarse sandstone (with associated pedogenised muds) which is followed up by a sand dominated system, at times with a few thin intermittent calcrete gravel sheets. A brief discussion of different lithofacies and facies associations is given as follows. Cross bedded pebbly very coarse sandstone

This is the lowest unit exposed in the Khudala and the Kama sections. Sandstone is disposed as a sheet and shows lateral variation in thickness. It is

Medium to fine sands with interbedded gravels overlie the implementiferous pebbly sandstone unit in the Karna section. Sands are well sorted and show planar bedding. Some intervals show gritty and coarse sand grains of quartz, felspars and rock fragments disseminated in a matrix of fine and medium sand. Dominant mineralogy of sand is quartz, felspars and micas. Conspicuous banding related to bedding concordant cementation in alternate beds is observed in the sands. Profuse calcrete nodule and rhizocretion development is also seen along the cemented bands. Locally, these sands lack cementation. Massive gravels with abundant reworked calcrete nodules are disposed as sheets or lensoidal units within the sands. These contain pebbles of caliche and rock fragments (-lcm dia.) within a matrix of grit and coarse sands comprising subrounded to subangular quartz, unaltered and angular felspar, rock fragments and caliche. In the Khudala section, the sand unit (overlying the pebbly sandstone) is massive, moderately cemented and pink. Here again a thin discontinuous band

MAYANK JAIN AND OTHERS

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 279

of calcrete gravel can be seen. Large troughs can be seen in the gravel where its thickness increases. Mud balls and few pebbles occur towards the base of the unit. Weak calcretisation is also observed. Well sorted nature of the sands with minor disseminated grit and coarse sand particles indicate reworking of aeolian sands by sheet flows. Reddening and calcretization indicate incipient pedogenesis. Patchy occurrences of noncemented well sorted sands probably represent aeolian deposits. Gravels indicate intermittent and episodic relatively larger magnitude semi-stable flows (with dominant lateral inputs). Sand-silt alternations

Sand-silt alternations overlie the sand member in the Khudala section. This is a thick unit consisting of alternation of well sorted fine to mcdium sands and laminated silts. Silts are generally 5-15 cm thick; however, they decrease in thickness and occur less frequently towards the top. Thickness of sands is variable and they show indistinct cross stratification. Sands are weakly cemented to non-cemented and comprise quartz, felspar and mica. Silts show micas along the laminations. Towards the base there are very small, (2-3 m lateral extent), pockety occurrences of gravel which comprise large broken angular silt fragments and caliche nodules in a matrix of medium to fine sands. Well sorted nature of sands represznt fluvially reworked aeolian deposits. Silts are deposited either as flood plain sediments or as slack water deposits. Pockety gravel occurrences indicate flow concentration in narrow localised zones. In general, flows are incompetent to carry coarse bed-load or there is a paucity of coarse sediments in the source area. All the evidences indicate preponderence of aridity. In addition to these, Type I1 sequence is also present at Bhuka (BH I, Fig.2) where it occurs as interbedded cross bedded pebbly very coarse sandstone and calcretized reworked aeolian sands. Nature of these deposits is similar to the deposits in the Karna section. Luminescence chronology (discussed later) together with the presence of Middle Paleolithic artefacts in the implementiferous, cross-bedded, pebbly, very coarse sandstone (Karna section) suggests early-middle Late Pleistocene time frame for these sequences.

Pottery-bearing pebbly gritty very coarse sands

This unit is exposed along the road from Bhuka to Karna, about 4 km away from Karna village, and extends discontinuously along the river bank from Karna till about four km upstream of Sindari section (Fig.3). It generally overlies the SindariBhuka type sequence, In contrast to older pebbly sandstones, this unit bears pottery fragments, animal bones and gastropod shells. Tentatively, the pottery has been assigned to the chalcolithic period. Thickness of this unit is about 1Sm. It is generally ill sorted, and comprises of very angular clasts of vein quartz, felspars, basic rocks, rhyolite and rounded to subangular caliche clasts in a matrix of gritty coarse sand to medium sand. Quartz in the coarse sand fraction is subangular to subrounded. Few sandstone clasts up to 10 cm long and 5cm wide are also present.These are derived from

N-8

Mean = 1

Kamothal

N

-

15 =

Meon

Type 111sequences

Type I11 sequences are observed at Karna, Bhuka (BH I) and Khudala. Dominant facies recognised are: pottery bearing pebbly, gritty, very coarse sands, silts, sand-silt deposits and well sorted medium to fine sands. These sequences are not very well developed in the study area. Individual units may be present unconformably over the older type I or type I1 sequences.

Fig.3. Outcrops and paleocurrent direction ( r o w d~a;ornrns)o f [he (Holocene) pottery bearing sands. Exposures are marked as shaded arcas with dashcd boundaries.

MAYANK JAlN AND OTHERS

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 281

the heterolithic facies (Type I sequence) and are especially abundant at places where this unit overlies the Type I sequence. Three sub-units have been observed within this unit. At the base there is a planar cross-bedded sub-unit which is followed up by small channel fills (about 30 cm deep and 1 m wide). Thickness of cross-beds ranges from 3050 cm and they show alternating foreset layersof pebbles, pebbly coarse sand and coarse sand. At the contact between the two, a silty layer with variable thickness may be present. The topmost sub-unit may be a normally graded layer grading from pebble size (0.5-3 cm) at the base to coarse sand size towards the top. Pebble rich foreset layers show clast supported framework as the fines are winnowed out resulting in an open work texture. Unabraded gastropod shells are present within the foreset layers. Animal bones are absent in the exposures near Lohira (Fig.3). Cementation is lacking in this unit and the paleocurrent directions are generally directed towards south and southwest (Fig.3). Bones collected from this unit belong mostly to domesticated species and their flourine-phosphate ratios are around one. However, one of the bones belongs to wild species and shows anomalously low flourine-phosphate ratios (Table 1). The fauna includes a species of Equus and Antelope. The antelope is slightly larger than modern form. Flourine-phosphate ratios increase with time and hence can be used as a first order indicator of the age of the bone. Anomalously low ratio in the wild species could be either because of the young age of the sample or due to the fact that it is a horn. These ratios in general indicate that the samples are of Chalcolithic age. Table 1. Identification and Flourine-Phosphate ratios of the bones found in the Type 111 pebbly coarse sands. Bone No.

Lab. No.

1 2 3 4 5 6 7 8 9 10 II

626 627 628 629 630 63 1 632 633 634 635 636

Identification Guzellu bennetrr Ros indicrts Bubalirls brrbulis Bos indicus Bos indicus 8 o . ~indicus Cupra Itircrrs Bubulius buhulis Bos indicus Bos indicus Equus sp.

l00FlP,O, 0.i40 1.06 1.06 1.14 1.05 08 1.47 0 9-15 1 7-1

Exposed at lower elevation areas in the Karna section, silty fine sari unconformably overlie the Late Pleistocene deposits. These silts have yieldd+ recent pottery. Predominantly fine-grained nature of these deposits indicate overbank sedimentation. .f

Coarse sanddsilty fine sands

-%A These deposits are observed in Bhuka section ( ~2: BH i ~I, western bank)$,# In the basal part, these deposits show three alternations of coarse sand and silty; fine sand. The succeeding units consist of silty fine sand and fine sand; some: disseminated coarse sands are also present in patches. Upper units (about, I .3 m thick) are gastropod bearing and there are two prominent calcretized; horizons within them. Occasional isolated pebbles are present along some horizons defining the bases of various sub-units. Sandslsilts are weakly cemented to non-cemented. These deposits have not yielded any artefacts. Well-sorted nature and fine-medium sand size indicates reworking of aeolian sands by episodic low magnitude flows. Coarse sands indicate relatively high magnitude sheet wash events.

Well sorted nzedium to fine sands

Well sorted medium to fine grained sands form the uppermost unit in the Khudala section, Karna section, and Bhuka section (BH I, eastern bank). These deposits show weak calcretization. Gastropod shells are seen in the gullied sections within this unit (like in the Khudala section). Microliths are found on the tops of some of these stablised sand covers. These sands represent a dunal sand cover over the alluvial deposits and are probably of Terminal Pleistocene or Early Holocene age (Fig.2). This is plausible as sand dunes in the Thar had a peak mobility between 14 ka and 10 ka (Chawla et al. 1992). It is also possible that some of these sands, especially in the Karna and Bhuka section might represent the Holocene arid phase. However, nothing can be said unequivocally in the absence of data on their ages. Presence of Chalcolithic and recent pottery, and an extrapolated age for the topmost aeolian sand cover from rest of the Thar suggests Holocene to Terminal Pleistocene time frame for the Type 111 sequences.

1 -17

'These deposits represent episodic sheet wash events, perhaps in the interdunal areas, with predominantly locally derived detritus. Presence of cross bedding indicates semi-stable flows. Presence of pottery and domestic animal bones indicates habitation in the surrounding areas.

LUMINESCENCE CHRONOLOGY

Basis

Luminescence chronology method is based on the assumption that daylight exposure during weathering and transport of the mineral grains reduces their geologically acquired luminescence signal to a small residual value. On burial

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 283

of sediment further daylight exposure ceases and re-accumulation of luminescence signal begins as a result of irradiation from the surrounding environment. Acquired luminescence over the depositional time when divided by annual rate of luminescence acquisition gives the age of the sediment. Sediments transported by wind experience full daylight exposure which reduces its geologically acquired luminescence to a small residual value. On the other hand, an estimation of residual luminescence signal becomes imperative especially in case of fluvially transported sediments where the net daylight flux to the mineral grains is attenuated to varying degree (Berger and Lauternauer, 1995). Consequently additional effort may be requ~redto ascertain the extent of pre-depositional photobleaching in the sediment; otherwise it may lead to over-estimation of ages. Following characteristics of aridlsemiarid systems, however, make alluvial deposits suitable for dating by luminescence chronology methods: 1 . low annual precipitation and prolonged dry seasons; 2. emergence of bar surfaces and stream bed during low-stage and ephemeral conditions respectively; 3. local reworking of fluvially deposited sediments by wind and vice-versa.

These conditions of prolonged exposure to sunlight would culminate in almost complete photobleaching of mineral grains. Mineralogy and alluvial motif of Quaternary alluvial sequences in Luni basin suggest deposition in arid1 semiarid climate. Further, the sand collected from the Luni river bed during the month of December gave a GLSL (quartz) age of 75232 years (Table 2) supporting the assumption of pre-depositional zeroing of GLSL signal in quartz. Table 2. Equivalent dose and GLSL age estimates on quartz grains. SL-Australian h d e method; ADDAdditive dose method; SE- Saturating exponential fit; CU-Cubic fit: LIN- Straight line fit

Sample

97-TR- 138 97-TR- 137 97-TR- 135 97-TR- 130 97-TR- I28 97-TR- 126 97-TR-13 1

Method 1 Fit

SL; CU SL; CU SL; SE SL; SE SL; SE ADD; SE ADD; LIN

Equivalent Dose 259 t 65 114+22 225 k 62 68 + 15 727 t 219 388 k 89 0. 16 ? 0.04

lntensityscaling 0.8 + 0.9 + 0.9 1.3 t 0.3 +

0.13 0.10 0.15 0.09 0.07

Dose Rare

Age (ka)

2 7 k0-1 1.6 + O . l 2.3 + 0 3 1.5 ?0.1 2.2 2 0 7 1.9 + 0 1 2.1 k 0 1

72 215 98229 45+ 1 I 3242 103 208+50 0.075 50.032

9548

284

MAYANK JAlN AND OTHERS

felspar and quartz respectively). Quartz was further treated with 40% HF 80 minutes followed by conc. HCI for 30 minutes to remove the alpha skin remnant felspars. Alpha skin etching was not carried out in case of felsp due to limited availability of the sample; however, a correction for alpha d was appl~edwhile estimating the annual dose. Both OSL (5 sec) and TL (u 400°C) measurements were done sequentially on the same set of aIiquo assuming that 5 sec OSL will not cause significant depletion of the TL sig (Duller and Botter-Jenson, 1993). Experiments were carried out to check validity of this hypothesis under laboratory conditions. Due to uncertaint the degree of pre-depositional bleaching of TL signal in both quartz and fels it's results are not discussed here. IRSL ( Infra-red stimulated luminescence) on felspars was done using TemT484 diodes emitting at 880 nm. The detection system comprised a photon counting system interfaced with a PC with a multiscanning card. The optical detection window comprised BG 39 coupled to UV 340 filters. Felspars were checked for two weeks anomalous fading test. In quartz, GLSL (Green light stimulated luminescence) measurements were done using a filtered Tungsten lamp using a set of GG 420 and SWP interface filters. Detection window comprised of Schott BG39+U340 filter spectra. Purity of quartz was confirmed using a short shine of 0.3 sec IRSL. All luminescence measurements were carried out in commercial reader Risoe TA-DA 15. Elemental concentrations of Uranium and Thorium were measured by thick source ZnS (Ag) alpha counting. Estimation of potassium was done using gamma counting. A water content of 15% was used for calculation of annual dose. Errors were computed using the method of Aitken (1985). Laboratory beta irradiation source was "'SrIq"Y. U, Th and K concentrations in the sample are given in Table 3. Short shine normalization was used for quartz GLSL and felspar IRSL, and both additive dose and regeneration methods were tr~ed.In view of early saturation in the GLSL signal of quartz and its sensitivity change in regeneration method, 'Australian Slide' technique (Prescott et a1.1993) was used (F1g.4ae). Modern sample: 97-TR- 131 (Fig.5) was also analysed to ascertain the degree of photobleaching in the present day Luni channel sand. Results

Methodology

Luminescence dating experiments viz. TL (Thermally stimulated luminescence) and OSL (Optically stimulated luminescence) were done on 105 - 150 pm size quartz and felspar grains. These were extracted following a sequential pre-treatment of sample with 1 N HCI and 30% H.O,, - . followed by sieving, Frantz magnetic separation and sodium-polytungstate heavy liquid separation of the two mineral fractions (< 2.58 gmlcc and >3-.58 gm/cc for

Quartz GLSL showed a saturation like behaviour of luminescence growth curve leading to difficulty in extrapolation for obtaining equivalent dose. In order to circumvent this problem, regeneration technique was tried. However, observed sensitivity changes in regeneration method could lead to underestimation of ages. Therefore, as a reasonable compromise between the two, 'Australian slide' technique with optimum intensity scaling was used for calculation of equivalent dose. In sample 97-TR-126 only additive dose method could be used as the sample amount was insufficient for carrying out the 'slide' technique.

MAYANK JAIN AND OTHERS

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS

281

L I N E A R EXTRAPOLATION

/ 7

INTEGRAL(O.2-4.91 DOSE BETA(SECOND1 S*c

F i g 5 GLSL Shine down curves and additive dose growth curve 0197-TR- 13 1. This is a modem Luni channel sand sample.

Fig.4(a-e). GLSL Growth curves by Australian slide method. Triangles represent additive dose points. Squares represent regeneration dose points.

Fig.6 shows a saturating exponential behaviour of quartz-GLSL in additive dose technique. Quartz depicts a plateau region feigning saturation above 60 Gy (Chawla et al. 1992). Our data suggests saturated exponential growth even at the total absorbed doses in excess of about 'natural+ 1200' Gy.This implies that the sample is placed in zone I1 of the quartz TL growth curve (Chawla et al. 1992). Hence, the saturating exponential fit only provides a minimum age estimate for this sample. Additive dose technique on 97-TR-131 (Fig. 3, a modern Luni sand sample, gave an age of 75232 years. For other samples slide technique was used (Fig.4 a-e), as it enabled reconciliation with difficulties faced in both additive dose and regeneration methods. Samples 97-TR- 135 and 97-TR-130 show an underestimation of the ages (discussed in detail later). The saturating behaviour of the additive dose points (Fig. 4 d,e), however, indicates that these samples are quite old. This dichotomy arises from the fact that in the samples from SindariIBhuka type sequence, the additive dose growth curve is near saturating even up to high applied beta doses (1 200 Gy), whereas the regeneration curves are generally cubic in nature (up to about 900 Gy). This implies a dose dependent sensitivity change In quartz and, therefore, intensity scaling and exponenrial fitting in the slide curve are probably not appropriate and underestim..re the dose shift. Present data, however, is insufficient to clearly understand the reason for lower ages in the quartz. Apparent age inversion in the samples 97-TR-138 & 97-TR-137 (Fig.4 a,b) and 97-TR-128 ( F i g . 4 ~ & ) 97-TR-126 (Iyig. 6) can be reconciled

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 287

MAYANK JAIN AND OTHERS

EXPONENTIAL REGRESSION

1

CXPONCNTYIL REGRESSION

TlME ( 0 . 2 - 4 . 8 S a c )

.-

HAT URAC N $ S O min

............

N + IOOtnln

N

+ 2 0 0 m l n -.-.-.

N+400min

CXPONCNTIAL REGRESSION

--.-.-.-.

N + 3 0 0 m l n -.

-

+

Fig.6. GLSL Shine down (a) and additive dose growth curve (b) of 97-TR-126.

the quartz. Apparent srge inversion in the samples 97-TR-138 & 97-TR-137 (Fig.4 a,b) and 97-TR-128 (Fig.4~)& 97-TR-126 (Fig. 6) can be reconciled with by considering the error range. 97-TR-128 and 97 -TR-126 can only be used as the minimum age estimates for these rocks. Scatter in quartz data also posed a problem. This could be due to either intrinsic o r extrinsic reasons. To understand this, mean and standard deviation of short shine normalised 'Nat. GLSL' values for each sample were calculated (Table 5). The general range of standard deviation is around 6-17. However, samples 131 and 135 show a higher percentage of scatter in luminescence output for identical short shine (SS) normalised discs. This is interesting as 'SS' normalised Nat. GLSL values for any sample should ideally be devoid of any scatter. As a preliminary hypothesis, we suggest that short shine samples only, most easy to bleach light sensitive signal of some grains, whereas, the longshine samples significantly larger number of traps. Felspars exhibit saturating exponential additive dose growth curve (Fig.7a). Negligible sensitivity change in regeneration method (Fig.7b) suggests that sensitivity change is not a serious problem. This is also confirmed by near concordance of the ages in additive, regeneration and slide methods (Table 4). Quartz GLSL and felspar IRSL ages are summarised in Tables 3 and 4 respectively. Comparison of quartz GLSL and felspar IRSL ages shows that IRSL ages are consistently under-estimated. This perhaps, is an artefact of U340 + BG-39 filter detection window or an evidence of mid-term fading

/

100

,

200

0..

3 3 1 2, 300

Fig.7. (a) IRSL shine down curves and additive dose growth curve of 97-TR-137, (b) TRSL shine down curve and growth curve of 97-TR-137 by regeneration method.

DISCUSSION

Depositional environment and luminescence chronology

Type I Sindari 1 Bhuka type alluvial sequence is dominated by heterolithic facies and cross-bedded pebbly gritty sandstone facies, which represent deposits ofsemi-arid gravel-sand bedload streams with slackwater deposits in weakly developed floodplains. Heterolithic facies largely records sheet flow events of varying competence and coeval silts and muds deposited under slack water conditions. Medium

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 289 Tablc 3. Concentration of U.Th and K in the studied samples. * indicates assumed K content in samples which are from the same lithofacies. Sample

Location

Wpm)

97-TR-131 97-TR- 138 97-TR-137 97-TR- 130 97-TR- 126 97-TR- 128 97-TR-135 97-TR-132 97-TR-134

RIVER BED KARNA KARNA SlNDARl SlNDARl SINDARI RHUKA BHUKA BHUKA

1.8 -t 0.6 2.6 + 1.2 2.1 + 0.4 1.2 + 0.2 0.6 + 0.3 0.5 2 0 . 7 2.1 .- 0.8 2.1 + 0.4 1.4 + 0.5

TUppm) 6.7 + 16.4 + 2.8 + 2.3 + 4.3 + 10.8 + 4.0 + 3.7 + 6.5 +

2.0 5.8 1.2 0.7 0.9 2.3 2.8 1.4 1.6

Method I Fit

Equivalent Dose

GY) 97-TR- 137

ADD. SE REG: SE SL; SE

97-TR- 135

ADD; SE REG; SE SL; SE

97-TR- 134

ADD; SE

133

97-TR- 132

ADD: SE

141 2 4 3

97-TR- 130

ADD; SE REG; SE SL; SE

71 + 2 9 95+ 3 78+ 6

97-TR- 128

ADD ; SE REG; SE SL: SE

100 + 35 96+ 4 96 + 10

.

DoseRate (GyW

1.5 1.3 I 1.2 1.6 1.6* 1.8 1.8' 1.8*

Age (ka)

61+ X 712 5 622 3

2.2

+ (0.2)

282 4 32 r 3 282 3

7 0 2 17 70217 86 + 15

2.9 r (0.3)

24+ 6 24 + 6 29-t 6

3.0 2 (0.2)

45 + I2

2

35

2.9

+ (0.2)

48

15

2 l ( O . )

35+14 4623 38% 4

+ (0.3)

34 + I ? 33 -t 3 332 5

2.9

Table 5. Scatter in short shine normalised Nat. GLSL values of quartz grains Sample

K%

Tablc 4. Equivalent dose and IRSL age estimale on felspar grains. REG-Regeneration method, explanation of other symbols same as that in Fig. 3. Sample

MAYANK JAlN AND OTHERS

scale cross-stratified pebbly gritstone/coarse sandstone and cross-stratified gritty very coarse-medium sandstone, associated with the heterolithic facies represent fluvial aggradation on higher parts of the alluvial plain, whereas, lower active channel belt is dominated by large scale planar cross bedded bar gravels. Disorganised pebble conglomerates are generally associated with the braid channel deposits and represent high magnitude episodic sheet flood events. Reddening, mottling, and profuse rhizocretionary activity in the heterolithic facies indicates incipient pedogenesis under emergent conditions. The alluvial motif, presence of unaltered large felspar fragments, calcareous rhizocretions indicate semi-arid conditions for these fluvial depositional systems. Quartz GLSL for the lower units in the Sindari sequence has provided

Standard Deviation

minimum ages of 2O81t5O ka and 324-+103 ka. Apparent age inversions in dates can be reconciled with by considering the large errors associated with them. The dates from the upper part of the SindarilE3huka sequence are 98229 ka and 4521 1 ka. Similar alluvial motif in the whole sequence from top to bottom, and apparent absence of any major hiatuses in the sequence are difficult to reconcile with a long duration of sediment accumulation from 45ka 198 ka to older than 324 ka. Also, the nature of sedimentation in these sequences is in sharp contrast to those of the younger 'Type 11' KarndKhudala sequences. KarnalKhudala type sequences are less cemented and largely dominated by aeolian sand or reworked aeolian sand (e.g reworked sands: Khudala section-90% ; Karna section-60%.), whereas, SindariIBhuka sequence shows 1) a higher degree of pedogenesis and reddening and 2) well established fluvial systems (such as sandy-gravelly braided stream deposits) suggesting deposition under relatively wetter climatic regime. Distinct climatic regimes and depositional milieu for these Type I and Type I1 lithological sequences suggest different time frames for them. Therefore, 45+11 ka and 98k29 ka dates require a reassessment. Evidences of semi-aridity (large unaltered felspars, rhizocretions, sheet flood deposits), luminescence dates (discussed later) and the absence of Middle Paleolithic artefacts allow us to tentatively assign an Early to Middle Pleistocene time period for the Type I Sindarimhuka sequence. Stratigraphic and facies analysis of the Type I1 sequences suggest preponderance of semi-aridity during last the 100 ka. Sequences at Karna and Khudala represent a prominent cross-bedded pebbly very coarse sand (with associated slackwaterlfloodplain deposits) at the base. This is followed by a period of dominance of aeolian processes resulting in deposits of well sorted medium to fine sands or sandjsilt intercalations. Abundant dispersed or locally concentrated coarse sand and grit particles in otherwise well sorted fine-medium grained sands suggest reworking of aeolian sands in episodic sheetwash events. The reworked aeolian sand record is punctuated at different levels by calcrete bearing massive gravels that represent deposition from semi-stable episodic flows. characteristic of the hydro-meteorological conditions that prevail in such areas. Lateral transport is significant as intraformational detrital constituents such as reworked calcrete nodules are abundant. The dominance

MAYANK JAIN AND OTHERS STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 29

of aeolian deposits, in comparison to the SindariIBhuka sequence, suggests a1 apparent climatic shift towards relatively more arid conditions. Large lateral extent of the pebbly sandstone, presence of coarse bedload and presence of in-situ artefacts allude to its deposition in relativel! wetter climate during the Late Pleistocene. This, however, is followed b! a period of increased aeolian dynamism punctuated by semi-stable episodic flows. The lower pebbly sandstone in Karna section has been dated to 72+ 15 ka This date is consistent with the in-situ Middle Paleolithic artefacts in the gravel The overlying sands have given an age of 95k28 ka. These two dates art stratigraphically consistent considering their error ranges. In some sections (e.g BH I, KH; Fig. 2) aeolian deposits, which are possibl] of terminal Pleistocene or Early Holocene age, are recognised in the Type I1 sequences. However, the first record of fluvial activity during the Holocene a: inferred from the studied sequences is around mid-Holocene (- 5 ka). Prominen pottery and animal bone bearing pebbly sands (near Karna), tentatively datec to 5 ka on the basis of Chalcolithic pottery, suggests sheet wash in the interduna areas and reflects climatic amelioration during the Holocene. Predominantl~ southward paleocurrent direction inferred from this unit indicates a paleo drainage from north during this time. This unit also places an importan constraint on the age of present Luni incision/exhumation (of gorges) in thi: area as it stratigraphically overlies the older Karna and Bhuka type sequences and perhaps also the aeolian sand cover. We suggest that pottery bearing sand: represent the beginning of re-integration of Luni drainage system during thc Holocene. In the studied segments of Luni river, choking of gorges possibl) occurred during the phase of peak aeolian activity, and they were exhumed b~ present Luni later than around 5 ka (however, it can not be ruled out that aitua incision, and not exhumation, of these gorges may have occured around thit time). A major palaeodrainage of river Lik, to the north of Tilwara (Fig.3), ha: been suggested earlier by Kar (1988). According to him, morphologica characteristics of the river indicate that it was of considerable size in the pas and had a far wider catchment compared to the present.This ancient Lik drainagt system possibly met present Luni river channel at Nausar, Mekrana an( Bhimarlai with a southward directed drainage (Kar 1988a). Our observation: on the Type I11 pottery bearing pebbly sands suggests that perhaps the Lik drainage was still prominent during early-mid Holocene and was responsible for their deposition. However, later the Luni river occupied its present course and the Lik drainage became obliterated due to aridity. In addition to these, recent pottery bearing silts and sands in the Type 11. sequence in Karna section represent floodplain deposits of the present Lun river. In general, the Holocene pottery bearing sequences indicate humar habitation in the area for the last five thousand years and re-establishment ol relatively wetter conditions around mid-Holocene (-5 ka).

Tal,le 6. Summary of major lithofacies characteristics. archaeo-stratigrapliy and luminescence dates of Quaternary alluvial sequences, Luni basin. 'L' represents age based on luminescence dating, .A. represents speculated age based on the presence of archeological evidences, 'P' stands for pottery ArchaeoStratigraphy

LithologylMajor facies

TYPC

Type Ill: KundaldKarnd Rhuka 1 (eastern bank)

Type I l sequence in KarndKhudala sections

Age

Recent

Recent(P)

Pottery bearing pebbly gritty very coarse sands (pottery, animal bones. gastropods)

Chalcolithic

-5 ka (P)

Dunal sand cover (medium to fine sands)

hlicroliths on top of some sand dunes

Older than -5 ka (A)

Siltslsilty finesands (pottery) Type Ill sequence near Karna

-

95+28ka (L)

Well sorted medium to fine sands/ sand-silt alternation Planar cross bedded pebbly gritty coarsc sandstone (Implement~ferous in Karna sequence)

Stiddle Paleolithic

* Upper hetrrol~thicfacies

Type I S~ndartl Bhuka sequence

Medium scale cross bedded pebbly gl-itty coarse sandstone

Large-scale planar cross bedded pcbhly gr~ttycoarse sandstone

Table 7. Typology of gra\~els/conglomerates.and their associated facie.; i n rhe Type I , Type I 1 and Type I11 sequences TyF

Diagnostic properties

Accoc~atcd facies

-

Rernarks on age

Pottery bearing pebbly gritty very coarse sands

Pebbly gl-~ttysands; tndictinct largescale planar cross bedding; angular detr~tus;abundant transported caliche; presence of pottery, animal bones and gastropod shclls

Dunal \and< ( ')

Holocene on the bail\ of Chalcollth~cpottery

Irnplcmentiferous cross bedded pebbly very coarse randstonc

Pebbly gl-itty very coarse sandstonc large-scale cross-bedding: tran.;portcd caliche nodules: very angulal- rhyolite and other rock fragments

Assoc~ated\\ ~ t h calcreticed mottled mudc

middle Late Pleistocene on the hnc~rof OSL d a t ~ n gand PI-esencsof hl~ddleP;IICOIIIIIIC amfact\

Older Conglomerates

Large-ecale crors bcdd~ng(-lm); i\olated boulderc and pebhlcs preccnt; outs~redrhyol~teclasts. pebble 71rcd qunttr. leld\par.\, !rock f ~ ~ g r n c nand t s cal~chc,rli~ck d~scordanrrI11coc1-ctionc ( 2 - 5 crn din): absence ol'a~tcfacts

He!erolith~cI.iilzc

-

Older than :?J+IOikn

294

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 293

increase in aridity in the less than 100ka sequences with possible wetter interludes at -70 ka and mid Holocene (- 5 ka).

In summary, the alluvial record of the last 100 ka represents palaeoenvironments of a semi-arid plain affected to varying degrees by aeolian sand spreads and fluvial reworking of aeolian sand sheets with wetter interludes around 5 ka and 70 ka. The older SindariIBhuka type sequence was deposited under a relatively wetter climate and is possibly of Early to Middle Pleistocene age. Tables 6 and 7 give a summary of archaeo-stratigraphy and a comparison of gravels/conglomerate typology with associated facies in these sequences respectively.

ACKNOWLEDGEMENTS I*

Authors express their gratitude to Dr. P.P. Joglekar of Deccan College for identifying the animal bones and Dr. A. Kshirsagar (Deccan College) for the Fluorine-Phosphate ratios of the bones from the pottery bearing gravel at Kaa* Prof. S.N. Rajaguru is thanked for discussions and Ms. Sonali Naik for assistance in the field. Mayank Jain acknowledges UGC for supporting his research work through JRF, and Department of Geology, University of Delhi for providing infrastructural support. PKL, Ahmedabad is thanked for giving permission to use the 'luminescence chronology laboratory'. Shri N. Juyal is thanked for providing support while carrying out luminescence dating. S.K. Tandon, S.C. Bhatt, Sheila Mishra and A.K. Singhvi thank the Department of Science and Technology, Govt. of India, for support provided through DST coordinated program no.ESS/CA/A3-08/92B.

CONCLUSIONS

On the basis of lithofacies characterisation, luminescence dating and archaeological evidences in the exposed aeolian-alluvial sequences in the Luni basin, Barmer district, the following conclusions can be drawn. 1. These lithological sequences show considerable variations in facies characteristics, particularly, in the relative importance of alluvial and aeolian deposits. 2. Type I SindariIBhuka sequence shows a complex facies organisation and has yielded luminescence dates which require reassessment. These sequences are inferred to have formed under semi-arid but relatively wetter conditions compared to late Pleistocene sequences. They are possibly older than 400 ka and may be placed in-Early to Mid-Pleistocene (?). 3. Type I1 Karna and Khudala sequences (< 100 ka ) are dominated by aeolian deposits (aeolian and fluvially reworked aeolian sands). GLSL date of 72+l5 ka together with the presence of in-situ Middle Paleolithic artefacts in the.implementiferous pebbly sandstone (Karna section) suggest an early to middle Late Pleistocene age for these sequences. 4. Type I11 sequences are assigned a tentative age of mid-Holocene (-5 ka) on the basis of Chalcolithic pottery. The Holocene pottery bearing pebbly sands in these sequences indicates climatic anlelioration around 5 ka.These are possibly the deposits of Lik paleodrainage in the area during the wetter phase of Holocene. 5 . Exhumation/incision of gorges by Luni river in the study area is later than 5 ka. 6. In view of incomplete understanding of stratigraphy and chronology of these sequences, a complete paleoclimatic reconstruction is difficult. Nevertheless, semi-ariditylaridity in the Quaternary is inferred from all the aeolian-alluvial sequences in the study area. There is an apparent increase in aridity in the less than 100 ka sequences with possible wetter interludes at -70 ka and mid Holocene (- 5 ka).

MAYANK JAIN AND OTHERS

References

.

AITKEN, M.J. (1985). "Thermoluminescence dating", Academic Press, London. ALLCHIN, B., GOUDIE, A.S.,, and HEGDE,K. (1978). The Prehistory and Paleogeography of the Great Indian Desert. Academic press, London, 370p. BERGER G.W., LAUTERNAUER, J.L. (1987). Preliminary fieldwork for thermoluminescence dating of Fraser river delta, British Columbia. Geol. SUN.Can. paper, 87-IA, pp. 900-904. CHAW, S., DHIR, R.P. and SINGHVI, A.K. (1992). Thermolumincsccnce chronology of Sand Profiles in the Thar desert and their implications, Quaternary Science Reviews,~.11, pp. 25-32. DULLER, G.A.T., and BOTTER-JENSON, L. (1993). Luminescence from potassium felspars stimulated by infrared and green light. Radiation Protection Dosimetry, v.47, pp.683-688. ESTABAN, C.M. and KLAPPA, C.F. (1983). Subaerial exposure environment. In: Carbonate depositional environments, (Eds. P.A. Scholle, D.G. Bebout & C.H. Moore), Mem. Am. Ass. Petrol. Geol., No.33, pp.1-63. FREYTET, P. (1973). Petrography and paleoenvironment of carbonate continental deposits with particular reference to Upper Cretaceous and Lower Eocene of Languedoc (Southern France), Sed. Geol, v.10, pp. 25-60. KALE,V.S. and RAJAGURU, S.N. (1987). Late Quaternary alluvial history of the north-western Deccan upland region. Nature, v.325, pp.612-614. KAR,A. (1988). The Lik river of the Rajasthan desert : Its source, courses and present condition. In: Desertification , monitoring and control, (ed. A.K. Tiwari), Scientific publishers, Jodhpur, pp.249-255. KAR,A. (1995). Geomorphology of arid western India, In: Quaternary environments and Geoarchaeology of India, (Eds. S.Wadia, R.Korisetter and V.S. Kale) Mem.Ceol. Soc. India. N0.32, pp. 168-190. MEW, S.S., and CHAMYAL, L.S. (1997). The Quaternary geology of Cujarat alluvial plains. Ind. Nat. Sci. Acad., v.63, A, No.1, pp 1-98. ~ ~ I A A.D. L L ,(1997). The Geology of fluvial deposits: Sedimentary facies, basin analysis, and petroleum geology, Springer-Verlag, pp. 1-582. R.K., ~ ~ ~ C H A M L.S. Y (1990). A L , Late Quaternary sedimentation pattern and terrain evolution PANT, in the Mahi river basin, Gujarat, India, Proc. Ind. Nat. Sci. Acad., v.56, pp.501 -51 1. PAREEK, H.S. (1984). Pre-quaternary geology and mineral resources of northwestern Rajasthan.

STRATIGRAPHY & FEASIBILITY OF LUMINESCENCE CHRONOLOGY METHODS 29.5

R,,o, M.S. (1996). Studies on physical basis oflumincscence geochronology and its applications, Unpublished Ph.D thesis, Department of Physics. Nagpur university. Nagpur, India. R~;AIIING, H.G. (Ed.), (1986). Scdimcntary environments and racies, 2 edn., Blackwell, Oxford SIRASSER, A. (1984). Black pcbble occurrcnce and genesis in Holoccnc carbonate sediments. Jour. Sed. Petrol., v.54, pp. 1097- 1109. S.I.RASSER, A., and DAVAUD, E. (1983). Black pcbblcs of the Purbcckian (Swiss and French Jura): litology, geochemsitry and origin. Eclog. Gcol. Helv.. v.76, pp.551-580. TANDON, S.K., SAREEN,B.K., RAO,M.S. and SINGHVI. A.K. (1997). Aggradation history and luminescence chronology of Late Quaternary semi-arid sequences of the Sabarmati basin, Guajrat, Western India, Paleogeography. Paleoclimatalogy. Paleoecology, v.128, pp.339-357. WARD.W.C., FOLK,R.L. and WILSON. J.L. (1970). Blackening of eolinite and caliche adjacent to saline lakes, Isla Mujeres, Quintana Roo, Mexico. Jour. Sed. Petrol.. v.40, pp.548-555. ZEUNER, F.E. (1 950). Stone age Plcistoccne chronology of Gujarat, Deccan College monograph, v.6, pp.46.