Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

PALAEO-06514; No of Pages 14

Contents lists available at SciVerse ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

j ourna l homepage: www.e lsev ie r .com/ locate /pa laeo

Mid–late Holocene monsoonal variations from mainland Gujarat, India: A multi-proxy study for evaluating climate culture relationship

Vandana Prasad a,⁎, Anjum Farooqui a, Anupam Sharma a,e, Binita Phartiyal a, Supriyo Chakraborty b,Subhash Bhandari c, Rachna Raj d, Abha Singh a

a Birbal Sahni Institute of Palaeobotany, Lucknow, Indiab Indian Institute of Tropical Meteorology, Pune, Indiac Department of Earth and Environmental Science, Kachchh University, Bhuj, Indiad Department of Geology, Faculty of Science, The MS University of Vadodara, Vadodara, Indiae Central University of Himachal Pradesh, India

⁎ Corresponding author. Tel.: +91 9839225422(mobE-mail address: prasad.van@gmail.com (V. Prasad).

0031-0182/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.palaeo.2013.05.025

Please cite this article as: Prasad, V., et al., Muating climate culture relationsh..., Palaeoge

a b s t r a c t

a r t i c l e i n f o

Article history:Received 2 April 2012Received in revised form 18 March 2013Accepted 20 May 2013Available online xxxx

Keywords:Mid-HoloceneClimate culturePaleoclimateHarappan civilizationMainland Gujarat

A multi-proxy study involving palynology, phytoliths, sedimentology, clay mineralogy, carbon isotopes andmagnetic mineralogy was carried out on Wadhwana Lake sediments from sub-humid zone of mainlandGujarat to determine the mid-Holocene climatic fluctuations, and its possible impact on the Harappanculture. The 1.03 m sediment profile of Wadhwana Lake shows five paleoclimatic phases. The study revealshigh lake stand during Phase I (~7500–5560 cal yr BP). A considerable cool and moist climatic condition dur-ing Phase I is inferred due to the presence of pollen belonging to wet evergreen taxa and high phytolith cli-matic index ‘Ic’. Later part of Phase I show gradual replacement of evergreen to deciduous pollen taxa,decrease in Ic value and dominance of smectite over kaolinite, indicating a reduction in wet climatic condi-tion due to decline in the precipitation and prevalence of seasonally dry climate. However, a large varietyof Madhuca, Meliaceae, Asteraceae pollen taxa, abundant fresh-water algae Chrozophora and Ceralia pollenin this period provides evidence of intensified arboriculture and agricultural activity. Low lake level anddry climate have been documented during Phase II (5560 to ~4255 cal yr BP) and are synchronous withthe lake records of western Indian region. High values of phytolith aridity index Iph, high primary minerals,increase in the δ13C values provide evidence for excessive dry climatic conditions at ~4255 cal yr BP. Phase IIIshows a gradual strengthening of SWmonsoon after ~3500 cal yr BP. Phase IV shows a short pulse of dry cli-matic conditions (~3238 to ~2709 cal yr BP) followed by somewhat similar to present day climate for PhaseV. The study concludes that onset of dry climate after 5500 cal yr BP is a regionally spread synchronous eventthat has been documented in several lake records of western India. It is surmised that the urbanization inHarappan civilization in North West India was coincident with the initial phase of declining rainfall ofmid-Holocene. The emergence of cultural complexity of Harappan civilization should be seen as an initial adap-tation to earliest phase of environmental deterioration and its subsequent decline is probably linked with thechanging seasonality pattern and excessive dry climate of later phase (~4200–4255 cal yr BP) of mid-Holocene.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

The Asian summer monsoon system is known to determine theclimate of the subtropics and tropics of South East Asia. The differen-tial precipitation on the Indian peninsula is attributed to the recedingSW monsoon from south to north and east to west. Though monsoonis a stable atmospheric phenomenon, considerable variability doesexist in the spatial and temporal rainfall patterns. Monsoonal fluctua-tions can have devastating impact upon the human society over a pe-riod of time. Rapid environmental changes in the past have caused

ile); fax: +91 522 2740008.

rights reserved.

id–late Holocene monsoonal vography, Palaeoclimatology, P

cultural changes, severe famine, and human displacements leadingto collapse of ancient civilizations (see de Menocal, 2001; Weiss andBradley, 2001; de Haug et al., 2003; Haberle and David, 2004;Wengxiang and Tungsheng, 2004). The onset of mid-Holocene aridityin the tropical and subtropical belts is attributed to the steepest de-cline of solar insolation at northern hemisphere around 6 kyr BP(Steig, 1999) causing major cooling of North Atlantic at 5.9 kyr BP(Bond et al., 1997). The synchronous records of environmental dete-rioration leading to widespread aridification post-6th millennia BPcan be traced to the Afro-Asiatic monsoonal belt, currently the desertareas of North Africa, South Asia, and China. Hence, paleoclimatic re-cords contradict the climate culture hypothesis as aridification inthese regions preceded the cultural complexity in the ancient civiliza-tions of mid-Holocene times (Brooks, 2006).

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

2 V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

The mid-Holocene climatic changes are coincident with theemergence of highly organized civilizations from the Afro-Asiaticmonsoonal belt such as Egypt, Mesopotamia, Indus–Saraswati andin northern China regions that form the bulk of the deserts today(Brooks, 2006). Harappan culture during 5th millennium BP in NWIndia is one of the most extensively studied, highly urbanized civili-zations of South Asia. Its archeological remains are widespread in thearid regions straddling the borders of India and Pakistan. Since thisculture flourished along the Indus river valley, it is also referred toas the Indus valley civilization. The Harappan civilization, had itsroots in earlier cultures. The earlier food producing communities(~9000 cal yr BP) established on the western highland of Baluchistannear Mehrgarh, later developed into village farming communities andpastoral camps (Possehl, 2002). During ~6300–5200 cal yr BP, furtherdevelopment in the agricultural practices and settled life led to thegeographical expansion of what is known as the early Harappancommunities that probably descended from the highlands to theplains of greater Indus valley. A rapid transition to urbanization dur-ing 4600–4500 cal yr BP resulted in a complex and stratified matureHarappan society with large urban centers (e.g. Harappa, Mohenjo-Daro,Kalibangan) mostly along the Indus and Ghaggar–Hakra rivers (Possehl,2002). The post-Harappan phase (4200–3000 cal yr BP) saw thedeurbanization of Harappan civilization that involved abandoning ofurban cities and establishment of small village-like communities. Lothal,a major mature Harappan archeological site from Gulf of Cambay marksthe southern most extent of Harappans in mainland Gujarat. Severalpost-Harappan sites can be seen i.e. Malvan near Surat and other adjoin-ing areas. There are a number of post-Harappan sites near Orsang Riverclose to Dhadhar.

Major urban settlements of Harappans near Indus and Ghaggar–Hakra river system indicate dependence of Harappan farmers on twowater sources, local rains aswell as river discharge. Two distinct sourcesof rainfall provide rains to north western region of India. Summer rainsassociatedwith the ‘SWmonsoon’, were initiated by the seasonal north-wardmovement of the Intertropical Convergence (ITC) due towarmingof Asian continents during summer (Wright et al., 2008). Mid-latitudecyclones, that originate in the Mediterranean, travel eastwards acrossSouthwest Asia and providewidespreadwinter rains to the northwest-ern part of India (Wright et al., 2008).

Decline of the Harappan civilization is often attributed to a region-al weakening of Afro-Asiatic monsoonal climate of mid-Holocene.Considerable disagreement exists in the paleoclimate records ofIndia and the emergence of highly complex Harappan culture.Harappan civilization was at its peak when large part of India wasexperiencing arid climatic conditions (Enzel et al., 1999; Gupta etal., 2003; Staubwasser et al., 2003).

The desert and desert margin areas of western India are sensitiveto precipitation related to the subtlest climatic changes. Palynologicalproxies have long been studied to assess the climate–culture linkedchanges from sedimentary profiles of several lakes (Didwana,Lunkaransar, Sambhar, Nal Sarovar) from desert and desert marginareas of western Rajasthan (Singh, 1971; Singh et al., 1972, 1973,1974, 1990; Prasad and Enzel, 2006). These authors argued that wetconditions due to higher winter precipitation during 6th millenniaBP, were the cause of establishment of village farming and agricultur-al base to the Harappan culture. Decline of early Harappan communi-ties in the plains of NW India during 4th millennia BP was due toweak monsoonal conditions (Singh et al., 1990). Precipitation recon-struction studies (Singh et al., 1972, 1974), were later modifiedthrough transfer function analysis (Bryson and Swain, 1981; Swainet al., 1983) provided better evidence of mid-Holocene aridity inwestern India, however, use of both uncalibrated and calibrated ra-diocarbon dates may have induced some error in age estimates inthese studies (Madella and Fuller, 2006). Combined geochemical,sedimentological and palynological studies showed that fall in lakelevels and desiccation preceded the de-urbanization or collapse of

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

Harappan civilization (Enzel et al., 1999). Archeological studies, furtherquestioned the climate–culture relationship (Misra, 1984; Ratnagar,1987, 2000; Possehl, 1997a, 1999, 2002). However, shifting agriculturalpractices due to changing climatic conditions have been linked to thedecline of complex society of Harappan culture (Madella and Fuller,2006). Though the distinct changing pattern in theHarappan culture re-quires more careful study of the environmental deterioration, paleocli-matic studies so far do not satisfactorily explain the contradictorybehavior of climate and Harappan culture flourishing at the sametime. The stimulus propelling the widespread establishment of villagefarming of early Harappan community into a well-organized matureHarappan society with peak of urbanization, specialization in agricul-tural practices, water storage techniques, focussing human activitynear river system and the rapid de-urbanization during later stages of5th millennia BP (Possehl, 2002), needs to be investigated in detail.

Known paleoclimatic records are mostly from lake sediments ofarid regions of Rajasthan (Singh, 1971; Singh et al., 1972, 1973,1974, 1990; Bryson and Swain, 1981; Swain et al., 1983; Wasson etal., 1984; Rai, 1990; Rai and Sinha, 1990; Sundaram and Pareek,1995; Kajale and Deotare, 1997; Deotare et al., 1998; Enzel et al.,1999; Roy, 2003). In contrast, apart from lake sediments from NalSarovar (Prasad et al., 1997; Prasad and Enzel, 2006) and estuarinesediments from Kothikhad (Prasad et al., 2007), which is in thelower reaches of Mahi River basin, about 60 km WNW of WadhwanaLake meager information exists from mainland Gujarat, a major cen-ter of mature Harappa (Lothal). Dhadhar River, which flows throughthe central part of the mainland Gujarat, has been studied in detailby Rachna Raj (2004, 2007) and the major sedimentary facies oflate Pleistocene deposits have revealed evidence of paleodrainagesand provided important link between the sub-humid Narmada Riverbasin in the south and semi-arid Mahi River basin in the north. Thepresent study on the Wadhwana Lake adds meaningfully to the gapof knowledge which exists from mainland Gujarat, a major center ofmature Harappa (Lothal).

We carried out multiproxy studies involving palynology,phytoliths, clay mineralogy, magnetic mineralogy and carbon isoto-pic studies fromWadhwana Lake sediments, located in a climaticallysensitive region lying in the transitional climatic zones, sub-humidsouth and semi-arid north (Fig. 1A), showing better sensitivity toeven minor changes in temperature and precipitation. Being locatedin the central part of the mainland Gujarat, this lake is ideal to docu-ment and assess the mid–late Holocene precipitation relatedchanges for the regional correlation as well as to evaluate the climateculture relationship.

2. Location and climate

On the basis of average rainfall, the Gujarat region has been cli-matically divided into humid, subhumid, semiarid, arid and ex-tremely arid zones (Fig. 1A). Dhadhar River basin lies in thetransitional area of sub-humid and semi-arid climatic zones towardssouth and north respectively (Fig. 1A), where the average annualrainfall varies from 800 mm to 1400 mm. Dhadhar is an independentriver which originates near Shivrajpur at 22°30′N latitude and 73°45′E longitude in the Aravalli range, and flows between two main riversof mainland Gujarat i.e. Mahi River to its north and Narmada river toits south and debouches into Gulf of Cambay (Fig. 1B). It shows broadestuarine the mouth, with the width of estuary being ~5 km(Fig. 1B). It has several small tributaries (Vishwamitri, Bukhi,Ranghav, Rupa, Khadi etc.) that are only active during monsoonmonths. In the area between Dhadhar and Orsang Rivers, a numberof lakes and ponds have been reported (Rachna Raj, 2004). The IRSFCC imagery of the area between Dhadhar and Orsang also showstraces of paleochannels, cut off meander, meander scars, ox-bowetc. This indicates that the ponding in this region is due to channelshifting, probably as a result of uplift along Orsang fault in the east

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

3V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

and Narmada–Son fault in the south (Rachna Raj, 2004). About40 km SE of Vadodara city, lies Wadhwana Lake (22° 11′ and 73°29′) which falls in Dabhoi Taluka of Vadodara district. It occupiesan area of 6 km2. The lake receives rains mainly from the south-western monsoon in the month of July, August and September, dur-ing this time the lake is fresh and relatively deep. The average tem-perature ranges from a maximum of 40 °C during summer day timeto a minimum of 12 °C during winter nights.

3. Material and methods

In the month of May, when the lake was almost dry, in the north-west part, which was b1 km from the road, a spot was selected fromwhere the maximum cut section on the dried lake was cleared forpreparing lithology and sampling (Fig. 2). 37 samples were collectedfrom 1.03 m sediment profile.

3.1. Phytolith studies

About 10 g of dry sediment was analyzed for phytolith extraction.Sediments were treated with 36% HCl for the removal of carbonate,and organic matter oxidation was done using 40% H2O2. Clay was re-moved by treating sediment samples with Calgon solution. Finallyheavy liquid treatment (480 g CdI2 and 500 g KI in 400 ml distilledwater) with a density of 2.3 g cm3, was centrifuged at 1000 rpm for5 min for the separation of phytoliths. Extracted phytoliths weremounted on glass slides in glycerol, which allows a three-dimensional observation while permanent slides were made in Canadabalsam. For each sample, more than 200 phytoliths were counted, at500×magnification. Phytoliths were classified according to the classifi-cation of Mulholland and Rapp (1992), Fredlund and Tieszen (1994),Alexandre et al. (1997), Barboni et al. (1999), Runge (1999), Wang(2003), and Li et al. (2010). Six different types of grass short silicacells (GSSC) were identified i.e. collapsed saddle, short saddle, trape-zoids, cross, bilobate and rondel (Figs. 3 and 4). The rondel and trape-zoid GSSC dominantly occur in the subfamily Pooideae mostlyadapted to cool and considerably high soil moisture conditions(Tieszen et al., 1979; Livingstone and Clayton, 1980). Cross and bilobateGSSC occur predominantly in the Panicoideae grass subfamily chiefly C4grasses adapted to warm and humid climate with high soil moistureconditions (Twiss, 1980; Scott, 2002). In Indian plains the dominanceof Panicoideae grasses (Pennisetum, Cenchrus, Andropogon, Sorghum)in moderate to high precipitation regions during south-west monsoon(July–September) is a common feature. Short saddle GSSC is producedin large quantity by grasses of subfamily Chloridoideae dominantly C4,characteristic of dry and warm climate where the soil moistureconditions are extremely low (Tieszen et al., 1979; Livingstone andClayton, 1980). The Chloridoideae grasses (Chloris, Lintonia, Cynodon,Sporobolus) that dominate in the desert and desertmargins in thewest-ern Indian region also occur in open land areas of Indian central plainsduring pre-monsoonwarmand dry climatic conditions or in drought af-fected regions.

Collapsed saddle, oblong bilobate, cross and bilobate with conicalprojections and blocky structures on one side are derived frommembersof the Bambusoideae subfamily (Prasad et al., 2011). Bambusoideaegrasses occur in moist and shady places close to water source (Wang,2003; Li et al., 2010).

Multifaceted, blocky polygon phytolith (Fig. 3) structures alongwith circular rugose types are derived from the epidermal cells andwoods of ligneous dicotyledon trees (Wang and Lu, 1993; Bremondet al., 2005; Lu et al., 2006; Gu et al., 2008; Li et al., 2010).

3.2. Phytolith climatic index

Climatic index ‘Iph’ is an aridity index and was first demonstratedon marine cores (Diester-Haas et al., 1973). It is a ratio of chloridoid

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

versus chloridoid and panicoid phytolith morphotypes (Twiss,1992). In the present study low soil moisture conditions due to lowmonsoonal activity are represented by the ratio of chloridoid (saddle)phytolith morphotypes versus panicoid and bambusoid along withchloridoid (dumbbell, cross, collapsed saddle, rondel) and areexpressed by a range of 0–1 values. Climatic index ‘Ic’ (winter precip-itation index) — in the present study is described by the ratio offestucoid phytoliths versus panicoid, chloridoid, bambusoid andfestucoid phytolith morphotypes. Twiss (1987, 1992) described cli-matic index ‘Ic’ to estimate the relative proportion of C3 grasses inthe American Great Plains. Ic is the ratio of pooid versus panicoid,chloridoid and pooid phytoliths. High Ic values indicate abundanceof C3 Pooideae grasses, indicative of cool climatic conditions. Hencetheir presence indicates high latitudinal (Twiss, 1992) or high altitu-dinal conditions in the tropical realm (Tieszen et al., 1979). More re-cent studies have shown that C3 grasses prefer cool and wetterclimate (Edwards and Smith, 2010). In the present study, winter pre-cipitation index has been represented by ratio of festucoid phytolith(rondel and trapezoids) versus panicoid, bambusoid and festucoidphytolith morphotypes.

Summer precipitation index — warm and humid climate lovinggrass genera and species of Panicoideae and Bambusoideae dominatein Indian plains during the summer monsoon (SW monsoon) season.A summer precipitation index has been described, that involves ratioof panicoid and bambusoid (dumbbell, cross, collapsed saddle) phy-tolith morphotypes versus festucoid, bambusoid and panicoid phyto-lith morphotypes.

3.3. Palynological studies

For palynological study, 10 g of air dried soil sample was treatedwith 10% potassium hydroxide (KOH) on sand bath for 5 min andsieved through 150 mesh size. The filtrate was then acetolysed follow-ing Erdtman (1943). The samples were then passed through 600mesh size and the residue (>10 μm)was collected for the palynologicalstudy. Approximately, 200–250 pollen grainswere counted per sample.The pollen spectrum (Fig. 5) is the percentage of the pollen sum (arbo-real and non-arboreals). The percentage of aquatic taxa given in the pol-len spectrum was calculated separately against the total sum ofnon-arboreals in the respective sediments. Thecamoebian data is alsopresented as percentage of the total thecamoebian count (~200)(Fig. 5). In the pollen spectra the arboreals were grouped into wetevergreen/semi-evergreen-moist deciduous (indicating wetter cli-mate) and dry deciduous according to Champion and Seth (1968). Tostudy the climate-induced evolution of vegetation in the region ratioof pollen belonging to wet evergreen/moist deciduous and dry decidu-ous taxa has been used to decipher seasonality and precipitation condi-tions. Percentage of algal palynomorphs has been considered todecipher fluctuation in the lake level whereas the ratio of arboreal/non-arboreal pollen provided evidence of the extent of forest cover.

3.4. Mineral magnetism

The universal presence of magnetic minerals and sensitivity to cli-matic conditions enables mineral magnetism (also known as environ-mental magnetism/rock magnetism) to be applied in a wide range ofgeographic locations and depositional environments. Magnetic pa-rameters provide robust, rapid and economic measure of qualitativeand in some cases quantitative constrains on the magnetic mineralspresent and their association under different climate conditions anddepositional environments (Bloemendal et al., 1988; Maher andThompson, 1992; Verosub and Roberts, 1995). Standard techniqueswere used for sample preparation (Walden, 1999). The sampleswere air dried and packed in standard non-magnetic plastic bottlesof 10 cm3. Several magnetic parameters were studied on these sam-ples. Magnetic susceptibility (χ) at low (0.47 kHz) frequency was

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

4 V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal variations from mainland Gujarat, India: A multi-proxy study for eval-uating climate culture relationsh..., Palaeogeography, Palaeoclimatology, Palaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

5V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

determined on Bartington Susceptibility Meter (Model MS2) with adual sensor. Anhysteric remanent magnetization (ARM) was inducedin the samples using a Molspin AF demagnetizer (with an ARM at-tachment) in a constant biasing field of 0.1 mT superimposed on adecaying alternating field (a.f.) with a peak field of 100 mT at adecay rate of 0.001 mT per cycle. The susceptibility of ARM (χARM)was calculated by dividing the mass specific ARM by size of the bias-ing field (0.1mT = 79.6 A/m; Walden, 1999). Isothermal remanentmagnetization (IRM) was induced in the samples at different fieldstrengths of 20, 50, 70, 100, 200, 300 up to 1500 mT and back fieldsup to 300 mT using ASC Scientific IM 10–30 Impulse Magnetizer.The remanences were measured in a Minispin magnetometer ofMolspin Ltd. The interparametric ratios that were used are SIRM/χand χARM/SIRM. The isothermal remanence induced at 1500 mT wasconsidered as the saturation isothermal remanent magnetization(SIRM). S-ratio was calculated by the expression (− IRM−300 mT/SIRM1800 mT). χARM/SIRM and SIRM/χ were calculated to determinegrain size. For magnetic mineralogy the IRM acquisition was impartedon all samples. The analyses were carried out in theWadia Institute ofHimalayan Geology, Dehradun, India.

The initial magnetic susceptibility (χ), that is the bulk representa-tion of the ferromagnetic content (along with the paramagnetic com-ponent) fluctuates between 14 and 115 × 10−8 m3/kg with anaverage value of 45.27 × 10−8 m3/kg for 37 levels. The IRM accusa-tions also show the contribution of magnetite which is the primarymineral contributing to the remanence in these sediments. As highervalues of χlf implies a close relationship between the erosional pro-cesses and increasing concentration of detrital input from the catch-ment (Williamson et al., 1998) probably during the comparativelywarmer phases, we have considered high values of concentration de-pendent parameters and higher S-ratio as indicators of increasedmonsoonal activity in the region while the low values may be periodsof less rainfall, anoxic conditions or dry periods.

3.5. Sediment texture, clay mineralogy and loss on ignition

Textural (grain size) analysis was performed on 10 g air driedsample. The samples were treated with different chemicals to removeexcess salts and carbonate, organic carbon and iron manganese coat-ings following the procedure outlined in Jackson (1956) and Knuze(1965). The samples were then sieved with a 53 μ to remove sandfraction and silt and clay fraction determined using pipetting(Tanner and Jackson, 1947; Day, 1965). The clay thus separated wasutilized for the mineralogical study by preparing the slides for X-rayanalysis. The separated clay was also saturated with KCl, CaCl2 andglycol and treated at different temperatures to differentiate the clayminerals. For loss on ignition determination ~5 g,−200 mesh samplepowder was taken in quartz crucibles and weight loss was measuredat 110 °C, 550 °C and 950 °C so as to determine the moisture, organiccarbon and carbonate carbon present respectively in the sample. Thistechnique is rapid and provides fairly consistent results comparedwith results obtained using carbon analyzer and therefore widelyused (Dean, 1974; Bengtsson and Enell, 1986).

3.6. Carbon isotopes

The carbon isotope ratio (13C/12C) of sediment organic matter pro-vides an important means to study the biological processes of thealgal and planktonic community of lakes that contributes to the or-ganic component of the sediment (Doner, 2003). In general, the car-bon isotope ratio of organic matter derived from lacustrine primaryproduction depends on the availability of dissolved CO2 mediated by

Fig. 1. A. Map of north west Indian subcontinent showing climatic zones and the archeolog(indicated by squares), sites discussed in the paper are identified by the following abbrevNal Sarovar; P, Pushkar Lake; S, Sambhar Lake; W, Wadhwana (modified after Madella and

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

photosynthesis and respiration. Ambient environmental conditions ad-ditionally influence growth and productivity of lacustrine algaethrough the control of nutrient availability, light, climate and watertemperature (Engstrom and Wright, 1990). Hence these parametersleave their signature in the carbon isotope ratio of the lake organicmatter. Lake algae δ13C values are typically −28‰, similar to the C3land plants, while wetland (C4) plants have much higher values,−9‰.

Carbon isotopic analysis was done on sediment organic mattertaken at 3 cm interval down the core. The samples were dissolvedwith 2 N hydrochloric acid for 2 h on a steam bath and left for 8 hat room temperature to remove carbonates. The samples werethen rinsed several times with deionized water until the solutionturned neutral. Then the samples were oven dried at 50 °C. Themass spectrometric analysis was carried out using a Delta V Pluscontinuous flow isotope ratio mass spectrometer interfaced witha 1112 Flash Elemental Analyzer. The standard deviation of repli-cate analysis of an internal standard was 0.3‰. The isotopicvalues were reported using the standard delta notation relativeto VPDB. The analyses were carried out at the Indian Institute ofTropical Meteorology, Pune.

3.7. Radiocarbon ages

Three AMS radiocarbon ages are obtained from the NSF-Arizona AMSradiocarbon dating laboratory. The ages obtained are 5565 ± 75, at adepth of 70 cm, 3550 ± 118, at 45 cm depth and 1297 ± 19 cal yr BPat a depth of 1 cm (Fig. 2). An age-depth model is developed by interpo-lating the ages between the dated horizons assuming constant sedimenta-tion rate (Fig. 6). Based on this the lake sediment spans a time period of7500- to 1297 cal yr BP.

4. Results

Based on pollen, phytoliths, mineral magnetic and clay-minerologicaldata the 1.03 m profile of Wadhwana pond was divided into 5 zones.

4.1. Zone 1 (103–70 cm)

The zone starts with the calcretized silty sand at the base followed by2–3 cm thick sandy layer in the upper part (Fig. 2). Fine lamination canbe observed in the clayey silt horizon (Fig. 2). This zone constitutesabout 20–35% palynomorphs of evergreen and moist deciduous treetaxa like Canarium, Manilkara/Sapota type, Madhuca, Meliaceae,Mimusops, Palaquium and Sideroxylon (Fig. 7). The lower part of thezone shows evergreen pollen taxa Palaquium, Canarium and Manilkara/Sapota, abundance of ‘gyrogonium’ of macroalgae Chara and algalpalynomorphs Crucigenia (Fig. 7). The most characteristic feature ofthis zone is the presence of charcoal specks that occur in large proportionin the lower part of the zone (Fig. 4). The phytolith recovery shows dom-inance of tree forms, phytolith morphotypes rondel, trapezoids, fanshaped and collapsed saddle belonging to festucoid and bambusoidgrasses (Figs. 3 and 4). Highly carbonized dendritic rods phytolith, be-longing to wheat and barley occur profusely in this zone.

In the latter part of the zone evergreen pollen taxa decline sharplyand are replaced by semi-evergreen to moist deciduous pollen taxasuch as Mimusops, Madhuca, Terminalia, Mangifera indica, Holoptelea,Anacardiaceae and large proportion of non-arbor pollens representedby members of Asteraceae, Amaranthaceae, Chenopodiaceae, Poaceaeand Cyperaceae (Figs. 5 and 7). The gradual decline of festucoid andincrease in chloridoid grass phytolith morphotypes (C4) is significant(Fig. 4). Further, the uppermost part of the zone shows highpercentage of VAM-Fungi (premier soil colonizers) and other terrestrial

ical sites of the Harappan civilization along with paleoclimatic records from lake sitesiations: BM, Bap–Malar playa; D, Didwana Lake; H, Harappa; L, Lunkaransar Lake; N,Fuller, 2006). B. Locality map of mainland Gujarat showing sample location.

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

6 V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

fungal ascospores and chytrid spores (saprobes), high percentagethecamoebians comprised of Centropyxis aculeatea ‘aculeata’, Centropyxisaculeata ‘spinosa’, Arcella vulgaris, Arcella megastoma, Arcella, discoides,Cyclopyxis kahlii and Trigonopyxis arcula (Figs. 5 and 7).

The initial magnetic susceptibility (χ), shows comparatively highand fluctuating values in this zone, variations in χ coincides with theχARM variation along the entire profile. In general, high but fluctuat-ing values of concentration dependent parameters (χ, χARM); lowback field values (IRM-300 mT) and fluctuating χARM/SIRM valuesare observed (Fig. 8). The clay mineral content in this zone varies be-tween 21 and 56%. In the clay fraction smectite is the most dominantmineral varying between 66 and 88% followed by kaolinite, quartzplus other primary minerals and chlorite (Fig. 8). It is interesting tonote that as the clay fraction increases, the kaolinite and chloritecontents also increase at the expense of smectite and other primaryminerals and vice-versa. As the region is largely dominated by theDeccan basalt with associated Vindhyan and Aravali supergroup for-mations, distribution of smectite is the most commonly observedclay mineral in this semi-arid region of western India (Sharma etal., 1994). It is important that along with the increase in clay mineral

Fig. 2. Lithology of Wadhwana showing radiocarbon dates a

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

content, an increase in organic carbon (org. C) content is also noticed(Fig. 8).

4.2. Zone 2 (70–57 cm)

This zone is characterized by silt dominated facies and is palynolog-ically barren. Phytolith recovery is also very low represented by theshort saddle Chloridoideae form and few phytolith morphotypes offestucoid grasses (Figs. 3 and 4). Mineral magnetic parameters alsoshow a marked change from 70 cm onwards (Fig. 8). Stable and lowconcentration dependent parameters (χ, χARM); and comparativelyhigh IRM (backfield); while stable and comparatively low χARM/SIRMratios are seen in Zones 2–4 indicative of the presence of large MD(multidomain) magnetite (Fig. 8). This zone shows a decreasing trendin clay content. Although the smectite is themost abundant clayminer-al in the bulk clay fraction but its overall concentration is significantlyreduced. Considerably decreased values of chlorite, kaolinite and δ13Care observed. Palygorskite, a claymineral that represents extreme arid-ity, is recorded in an extremely low amount in this zone (Fig. 8).

nd photograph of the same trench sediment succession.

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

Fig. 3. A. Generalized picture showing distribution of organic matter under microscope (scale bar 100 μm), B–C. cuticles showing dendritic long silica cells (scale bar 10 μm), D–K.light microscopy images of selected phytoliths (scale bar 10 μm), D. rondel, E. trapezoid, F. collapsed saddle, G. short saddle, H. bilobate, I. fan shaped, J. cross, and K. multifacetedblock.

7V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

4.3. Zone 3 (57–30 cm)

This zone consists of silt and silty clay facies. The lower part showsreappearance of moist deciduous arboreal Manilkara, Madhuca,Terminalia, Mangifera indica, and Azadirachta indica. Pollen grains ofMadhuca and Azadirachta show moderate increase in this zone (Figs. 5and 7). The exotic tree taxa Casuarina, Acacia, Pinus, Hibiscus, andMalvaceae also occur in low percentage in this zone (Figs. 5 and 7).Poaceae pollen occurs in moderate number and is represented byCerealia pollen (Figs. 5 and 7). Pteridophytic spores and thecamoebiansrepresented by Trigonopyxis arcula and Cyclopyxis kahlii, Centropyxis

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

aerophila, Euglypha and Nebela occur in low numbers however algalplanktons like Botryococcus and shallow water indicators such asPseudoscizae including aquatic fungi were observed in high propor-tion in the lower part of this zone (Figs. 5 and 7). Significant increasein the panicoid (bilobate and cross) and bambusoid type and fewshort saddle phytolith morphotypes is observed (Figs. 3 and 4).The zone shows moderate increase in tree forms and festucoid phy-tolith morphotypes along with charcoal specks (Fig. 4). χ, χARM andχARM/SIRM trends are similar to the Zone 2 (Fig. 8). In this zone, amore or less similar clay and silt content is a characteristic feature(Fig. 8).

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

Fig. 4. Chart showing the phytolith distribution along with lithology.

8 V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

4.4. Zone 4 (30–21 cm)

This zone is silt dominated showing extremely low pollen andphytolith content and no marked difference in the magnetic sus-ceptibility is seen than the preceding zone while a rise in cARMand SIRM/c, towards the top of the zone and S-ratios shows an in-crease at the start of the zone and then a decrease at the upperpart in the zone (Figs. 4, 5 and 8). The sediment texture of thiszone is somehow similar to the Zone 2 where the silty sand faciesis prominent as also reflected by the increasing trend seen in thesand and silt contents (Fig. 8). Clay content is decreasing and sois the organic C content (Fig. 8). A complete removal of chloriteand a diminishing trend in kaolinite is a pronounced feature(Fig. 8).

4.5. Zone 5 (21–0 cm)

This zone starts with the 2–3 cm thick sand interval that showshigh proportion of phytolith morphotypes, however, pollen is pres-ent in very low amount due to preservation factors in the sandy fa-cies (Figs. 4 and 5). The phytoliths show increased values ofpanicoid and bambusoid phytolith morphotypes (Figs. 3 and 4).The uppermost part of the zone shows the presence of extremelylow arboreal pollen represented by Madhuca and Toddalia asiatica(Figs. 5 and 7). Non-arboreals occur in moderate amount and arerepresented by Chrozophora, Eclipta, Asteraceae and Ceralia pollen(Figs. 5 and 7). It also shows large proportion of aquatic and terres-trial fungi (Fig. 7). Magnetic parameters are similar to Zone 1showing a gradual increase in the concentration dependent param-eters, decreasing backfield IRM (Fig. 8). The lower part of thissub-zone is sandy and gradually becomes more clayey in the up-permost part (Fig. 8). The moisture and organic C are also showinga trend similar to the clay abundance (Fig. 8). An increase in kaoliniteand reappearance of chlorite suggest relatively humid and warmconditions.

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

5. Discussion

5.1. Paleoclimatic phases and its correlation with salt lakes of Rajasthan

On the basis of the proxy data and chronology, five climaticphases (Phase I–V) during ~5565–1297 cal yr BP were interpreted(Fig. 9) each corresponding to the 5 zones of Wadhwana Lake de-scribed in the Results section. Lower part of the Phase I (~7500 to~6795 cal yr BP) has been interpreted as a considerably wet phasewith abundant freshwater algae supporting high lake stand (Fig. 9).Fine clay bands associated with sand layers during this phase indi-cate deposition under enhanced run off condition where coarsersandy material that settled initially followed by fine silt and clayfraction, is suggestive of strong monsoon regime. The presence ofchlorite and kaolinite in relatively higher abundance in the lowerpart of the Zone 1 (Figs. 8 and 9) further indicates higher rainfallconditions.

The high values of concentration dependant parameters in this zonecan be attributed to high sediment supply to the basin due to higherprecipitation (Dearing et al., 1996; Shankar et al., 2006; Warrier et al.,2011). The high S-ratio can also be linked to higher precipitation re-gimes as in the tropical regions this is a good proxy for paleomonsoon(Basavaiah and Khadkikar, 2004). The high proportion of C3 grassphytoliths and abundant plant tissue remains in the sand lenses reiter-ate deposition of sand under considerably wet climatic conditions. MAPof C4 optimum is drier than the C3 optimumbyover 500 mm year−1 in-dicating preference of C3 grasses for cool and wet climate, instead ofcool and dry (Edwards and Smith, 2010). Abundance of festucoid (C3)grass phytoliths and lower proportion of panicoid (C4) grass phytolithsin Phase I raise the value of phytolith index Ic indicating higher winterprecipitation activity but lower summer precipitation (Fig. 9) Thiswould have resulted more months with higher precipitation efficiencyand spread of cool and moist climatic conditions during Phase I(Fig. 9). Desert playa lakes at Didwana and Lunkaransar from Rajasthanshow considerably higher rainfall (almost twice the current level) in de-sert during 7500–4800 cal yr BP. The presence of high proportion of

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

Fig. 5. Chart showing the pollen diagram along with lithology.

9V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

ecologically sensitive Artemisia pollen (Singh et al., 1990) during thisperiod indicates higher winter precipitation activity as the main causefor high lake level and precipitation efficiency (Singh et al., 1974,1990; Bryson and Swain, 1981; Wasson et al., 1984; Enzel et al.,1999). In these studies, Artemisia has been taken as the indicator forwinter precipitation. Artemisia is a herbaceous weed that flourisheswhen annual rainfall exceeds 50 cm. This weed flowers during winters,and usually grows on stabilized dunes of semiarid-arid zones. Since thepresent study is confined to subhumid climatic zone, hence Artemisiapollen is not recorded.

The Phase I of Wadhwana Lake can be correlated with the upper-most part of Phase 4 of Lunkaransar and Didwana (Singh et al., 1974)and reinterpreted zones D5 in part and D4 of Didwana (Singh et al.,1990). The excessive wet period of Phase I (~7500–6795 cal yr BP) ofWadhwana Lake can be correlated with the wettest periods ~7200–6100 cal yr BP, of Nal Sarovar, ~7200–6000 cal yr BP of Lunkaransarand ~6700–5900 cal yr BP of Didwana lakes respectively. The wet cli-matic condition in Didwana and Lunkaransar lakes is attributed to thehigher winter precipitation activity (Bryson and Swain, 1981; Singh etal., 1990). Didwana and Lunkaransar lake records show thatwinter pre-cipitation continued in the northern region ~4800 cal yr BP (Brysonand Swain, 1981; Singh et al., 1990). Further, the upper part of Phase I

Fig. 6. Graph showing age versus depth of Wadhwana Lake.

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

(~6795–5565 cal yr BP) in Wadhwana (Fig. 9) shows change in thepalynological assemblage from evergreen to moist deciduous trees, de-crease in thewinter precipitation index Ic, and increase in smectite con-tent that is indicative of decline of winter precipitation activity andgradual shift from wet climate to more seasonal monsoonal climatethat correlates well with the Nal Sarovar lake records (Prasad andEnzel, 2006). After the wet phase, decrease in algal content indicateslow lake level in Wadhwana Lake around 5565 cal yr BP, thus signify-ing the onset of dry climate that continued, during Phase II (5565 to~4255 cal yr BP) in the lake succession (Fig. 9). High Iph phytolithindex and high salinity conditions during this phase further confirmlow precipitation and high evaporating conditions. However, the ab-sence of any evaporites indicates that moisture conditions weremaintained annually and the lake did not reach excessive dryingphase. Similar conditions can be observed in Lunkaransar and Didwanalake records after ~4700 cal yr BP. Likewise, the low values of the con-centration dependant mineral magnetic parameters also indicate a lowinput of sediment to the basin perhaps as a result of little or no precip-itation leading to conditions where the magnetic minerals are liable todissolution (Fig. 9). Decrease in χ and S-ratios and also a relative de-crease in kaolinite content and almost an absence of chlorite suggestlow weathering in the lake catchment due to decrease in precipitation(Fig. 9). An increase in primary mineral abundance in the clay fractionindicates limited chemical alteration over the mechanical mode,which is also supported by a more or less homogeneous texture of thesediments. The sharp rise in δ13C in the later part of this phase too indi-cates moisture deficient condition (Fig. 9).

The drop in Wadhwana Lake level at 5565 cal yr BP (Fig. 9)can be correlated with the drying out trend of Nal Sarovar(~6100–5400 cal yr BP), Lunkaransar (5700 cal yr BP) and Didwana(~5900 cal yr BP) lakes (Prasad and Enzel, 2006) (Fig. 10). Similardrying trend can be seen in the Sambhar Lake sequence at 5590–5290 cal yr BP (Singh et al., 1974). This drying out trend of the lakescan be correlated with the records of dune formation activity in aridand semi-arid regions indicating a major shift towards dry climaticconditions of NW India during 5200 cal yr BP (Enzel et al., 1999).High phytolith aridity index (Iph) ~4200 cal yr BP of Wadhwanacan be correlated with the other regional excessive dry climatic recordsduringmid-Holocene. Reduction in humidity in Garhwal Himalayas be-gins around 4000–3500 cal yr BP (Phadtare, 2000) and in peninsularIndia around 4500–3500 cal yr BP (Caratini et al., 1994). NorthwesternArabian Sea records indicate the beginning of aridity around 4800 and4200 cal yr BP (Gupta et al., 2003; Staubwasser et al., 2003).

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

Fig. 7. Light microscopic photomicrographs of palynomorphs (all scale = 10 μm). A and B. Palaquium ellipticum, C. Mimusops elengi, D. Melia azadirachta, E. Madhuca indica, F.Mangifera indica, G. Holoptelea, H. Canarium, I. Manilkara/Sapota, J and K. Sideroxylon, L. Toddalia asiatica, M. Terminalia, N. Casuarina, O. Pinus, P. Acacia, Q. Chenopodiaceae, R.Chrozophora, S. Hibiscus, T. Ipomoeae, U. Portulaca, V. Asteraceae, W. Cassia, X and Y. Unidentified aquatic pollen, Z. Cyperaceae, AA. Typha, AB. Poaceae, AC. Cerealia, AD. Eichhornia,AE. Lemna, AF. and AG. Osmunda spore, AH. Monolete spore, AI. Selaginella microspores, AJ. Ophioglossum, AK. Lycopodium, AL-AO. Selaginella microspores, AP. Adiantum type, AQ.Chara gyrogonium, AR. Pseudoschizae, AS. and AT. algal cysts, AU. Tilletia, AV. Cleistothecia, AW. Fungal spores, AX. and AY. Crucigenia, AZ. and BA. Abundant Ascospores, BB-BE.Fungal spores, BF-BJ. Fungal ascospores, BK-BP. Fungal spores of various types, BQ. Chytridiales (aquatic fungi).

10 V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal variations from mainland Gujarat, India: A multi-proxy study for eval-uating climate culture relationsh..., Palaeogeography, Palaeoclimatology, Palaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

Fig. 8. Graph showing clay mineralogy organic carbon, salinity, mineral magnetism and carbon isotope.

11V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

During Phase III (~4000–3238 cal yr BP) low lake level persistedduring this phase as indicated by the high salinity conditions(Fig. 8). However, the presence of algal component is indicative ofmoderate increase in the lake level (Fig. 9) but that did not reach tohigh stand as in Phase I. Increased summer precipitation, phytolithindex and low Ic values (winter precipitation index) show moderateincrease in SWmonsoon along with intermittent winter precipitation(Fig. 9). The increasing trend in kaolinite in this phase suggests a shiftfrom drier to relatively moist climatic conditions. It can be correlatedwith other records at Itola (Singh et al., 2007) and Kothiyakad (Prasadet al., 2007) of mainland Gujarat. It is observed that during this phasethere was change in the seasonality regime due to the influence of SWmonsoonal activity. An increase in kaolinite and reappearance ofchlorite suggest a relatively more humid and warm condition. PhaseIII is followed by short pulse of weak monsoonal condition definedas Phase IV (~3238–2709 cal yr BP) (Fig. 9).

The period ~3238–1297 cal yr BP represents Phase V. Based onthe pollen and phytolith data Phase-V indicates more or less presentday climatic pattern with fluctuating SW monsoonal activity butnear cessation of winter precipitation activity.

Fig. 9. Compilation of Smectite versus Kaolinite, phytolith index, pollen index, inferred lakeHarappan culture.

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

5.2. Climate culture relationship

The archeological data shows transition from haphazard pastoral andvillage farming communities of early Harappan (5200–4500 cal yr BP) toculturally advanced well developed urbanized centers of matureHarappan (4500–3900 cal yr BP) community (Possehl, 2002). Thetransition from early to mature Harappan community and its subse-quent decline was rather abrupt and took place quite rapidly (Possehl,2002). Harappan civilization was thought to have flourished duringthe higher rainfall monsoonal climate and collapsed during the aridphase of mid-Holocene (Singh, 1971; Singh et al., 1974; Bryson andSwain, 1981; Agrawal, 1982; Swain et al., 1983) though it was criticizedin many archeological studies (Misra, 1984; Ratnagar, 1987, 2000;Shaffer and Liechtenstein, 1989; Paddaya, 1994; Possehl, 1997a,b,1999; Fuller and Madella, 2001). Further, high resolution paleoclimaticstudies (Enzel et al., 1999; Prasad and Enzel, 2006) indicate that theonset of the mid-Holocene aridity proceeds the urbanization phase ofHarappan civilization, thus raising serious doubts on climate culture hy-pothesis. Large number of pre-Harappan sites shows richmacrobotanicalremains indicating extensive agricultural activity that would have

level, rainfall, seasonality and paleoclimatic phases along with archeological phases of

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

Fig. 10. Comparison of the Wadhwana lacustrine record with other lake records from Rajasthan and Kachchh.

12 V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

provided subsistence base and surplus food to the expanding Harappancultural practices (Costantini, 1983). The subsistence economy ofpre-Harappans was largely based on the winter–spring crop package(wheat, barley, peas, lentils, grass pea and linseed) (Meadow, 1996;Tengberg, 1999; Fuller and Madella, 2001; Fuller, 2002; Weber, 2003;Pokharia et al., 2011). Thus, high winter precipitation activity prior tothe rise of Harappan civilization as inferred by earlier studies musthave largely contributed in the spreading of winter/spring harvestedcropping system (Fuller, 2002; Weber, 2003; Pokharia et al., 2011). Inthe present study, dominance of wheat phytoliths during ~Phase I(Fig. 4) is also suggestive of cultivation of wheat much before the ma-ture Harappan phase (4500–3900 yr BP). Introduction of millets andpulses and non-staple summer crop package bymature Harappan is in-dicative of changing pattern of agricultural practices and subsistence re-gime probably due to changing rainfall pattern (Madella and Fuller,2006; Pokharia et al., 2011). High resolution paleoclimatic data emerg-ing from the region is suggestive of declining winter precipitation after5th millennia BP (Bryson and Swain, 1981; Singh et al., 1990). In thepresent study decrease in the festucoid phytolith morphotypes during~6795–5565 cal yr BP is indicative of decline in the winter precipita-tion activity. Changing pattern of palynoassemblage from wet ever-green to deciduous during this period (Fig. 5) is also suggestive ofshift from wet climate to seasonally dry climatic conditions as a re-sponse to reduced precipitation activity (Fig. 9). It is thus surmisedthat decrease in winter precipitation would have resulted in theunpredictability in the rainfall pattern and would have created condi-tions unsuitable for the winter cropping. Multicropping subsistencestrategy with the use of drought tolerant varieties and cultivation nearriverine sources probably indicates agricultural intensification effortsby mature Harappans to combat environmental stresses during region-ally spreading aridity conditions during ~4800–4200 cal yr BP. Fluviallandscape studies carried out near Ghagghar–Hakra and Indus rivershow large scale flood events prior to mature Harappan phase and con-siderable decrease in theflood intensity due toweakmonsoonal climateduring mature Harappan phase (4600–4200 cal yr BP), stimulated

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

intensive agricultural activity in the low lying, river interfluve re-gions (Wright et al., 2008; Giosan et al., 2012). It has been discussedthat the there is an inverse, co-varying trend in rainfall seasonality inthe north western region, as yearly monsoon precipitation dimin-ishes, winter rainfall increases proportionately (Wright et al.,2008). Further the innovative methods in agricultural activities dur-ing mature Harappan phase are an adaptation to changing rainfallpattern. Multicropping strategy with the introduction of summercrops (millets, gram), fruit cultivation and non-staple food cropslike sesame and cotton and cultivation along the riverine sourcesstrongly indicates agricultural practices in order to maximize agri-cultural yield to mitigate environmental deterioration (Madellaand Fuller, 2006; Pokharia et al., 2011). The presence of deciduousgenera like Madhuca, Terminalia, Mangifera, as well as large numberof Chrozophora and Ceralia pollen in the later part of Zone 1 suggestincreased arboriculture and agricultural activities at the end of the5th millennium BP (Fig. 5). These may not necessarily provide thetrue climatic signals based on disparity in the paleoclimatic andarcheological data record. However, it indicates more substantive ar-boriculture as well as agricultural activities, and thus suggests diver-sification in food production probably due to a shift in the rainfallpattern. Later part of Phase II at ~4255 cal yr BP indicates extremelymoisture deficient condition during the late Harappan phase (Fig. 9).Archeobotanical remains of summer crops in late Harappan sitessuggest increasing reliance on two cropping seasons (Madella andFuller, 2006; Pokharia et al., 2011). We decipher that shifting agricul-tural strategies due to climatic instability might have been the driv-ing force for urbanization. Persistent moisture deficient conditionsdue to weak rainfall activity and changing seasonality pattern duringlater stages of 4th millennium BP caused agricultural droughts anddeurbanization of Harappan culture. Viewed separately, paleoclimaticand archeological datamay not provide a holistic picture, however an in-tegrated approach is warranted that is likely to provide a better under-standing regarding the role of physical environment on early Harappancivilization. However, more detailed archeological and paleoclimatic

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

13V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

data sets from far and wide regions from western India are required forgreater understanding of the climate–culture relationship.

6. Conclusion

1. A multi-proxy study from Wadhwana Lake, Dhadhar River basinprovided evidence of mid–late Holocene precipitation fluctuationsfrom the sub-humid–semi-arid transitional climatic zone of main-land Gujarat.

2. Dominant winter precipitation due to active western disturbancesduring mid-Holocene not only provided rains to Indus Himalayanfoothills and Rajasthan area but also in the mainland Gujarat re-gion that forms the desert margin today.

3. Pollen and phytolith studies show prevalence of winter precipita-tion activity during ~7500 to ~6795 cal yr BP in mainland Gujarat.

4. Multi-proxy study indicates low seasonality, excessive wet climateand high lake stand during ~7500 to ~6795 cal yr BP in WadhwanaLake.

5. Widespread wet climatic conditions due to enhanced winter pre-cipitation activity provided suitable climatic conditions to earlyHarappan to establish winter cropping system and agriculturalbase in the north western part of the Indian subcontinent.

6. Based on pollen turn over and phytolith data, a considerable de-crease in the winter precipitation activity, low lake stand ofWadhwana and onset of dry climatic condition at ~5565 cal yr BPis envisaged

7. The urbanization in the mature Harappan phase can be seen as anadaptation to declining winter precipitation and prevalence of dryclimatic conditions (~5565–4200 cal yr BP).

8. Excessive moisture deficient conditions after ~4200 cal yr BPresulted in agricultural drought and deurbanization of Harappansfrom larger part of north western India.

Acknowledgments

This study was supported in part by grant no. SR/S4/ES-49/2003,SR/S4/ES21/Baroda and SR/FTP/ES-11/2000 from Department of Sci-ence and Technology, New Delhi. We are thankful to Directors ofBSIP, Lucknow and IITM, Pune for the encouragement. We thankProf L. S Chamyal, MS University of Vadodara, for help during thefield work. Constructive comments and suggestions from Prof I. BSingh, Lucknow University proved helpful in the preparation of thismanuscript. Wadia Institute of Himalayan Geology, Dehradun, Indiais thanked for providing paleomagnetic laboratory facilities.

NSF-Arizona AMS radiocarbon dating laboratory is thanked forproviding AMS dates.

References

Agrawal, D.P., 1982. The Indian Bronze Age cultures and their metal technology. In:Wednforf, F., Close, A.E. (Eds.), Advances in World Archaeology, vol. 1. AcademicPress, New York, pp. 213–264.

Alexandre, A., Meunier, J.D., Lezine, A.M., Vincens, A., Schwartz, D., 1997. Phytoliths: in-dicators of grassland dynamics during the late Holocene in intertropical Africa.Palaeogeography, Palaeoclimatology, Palaeoecology 136, 213–219.

Barboni, D., Bonnefille, R., Alexandre, A., Meunier, J.D., 1999. Phytoliths aspaleoenvironmental indicators, West Side Middle Awash Valley, Ethiopia.Palaeogeography, Palaeoclimatology, Palaeoecology 152, 87–100.

Basavaiah, N., Khadkikar, A.S., 2004. Environmental magnetism and its application to-wards palaeomonsoon reconstruction. The Journal of Indian Geophysical Union 8(1), 1–14.

Bengtsson, L., Enell, M., 1986. Chemical analysis. In: Berglund, B.E. (Ed.), Handbook ofHolocene Palaeoecology and Palaeohydrology. John Wiley and Sons Ltd., Chichester,pp. 423–451.

Bloemendal, J., Lamb, B., King, J.W., 1988. Palaeoenvironmental implications ofrockmagnetic properties of late Quaternary sediment cores from the eastern equa-torial Atlantic. Paleoceanography 3, 61–87.

Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., DeMenocal, P., Priore, P., Cullen,H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in North AtlanticHolocene and glacial climates. Science 278, 1257–1266.

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

Bremond, L., Alexandre, A., Hély, C., Guiot, J., 2005. A phytolith index as a proxy of treecover density in tropical areas: calibration with leaf area index along a forest–savannatransect in southeastern Cameroon. Global and Planetary Change 45, 277–293.

Brooks, N., 2006. Cultural responses to aridity in the Middle Holocene and increasedsocial complexity. Quaternary International 151, 29–49.

Bryson, R.A., Swain, A.M., 1981. Holocene variations of the monsoon rainfall in Rajasthan.Quaternary Research 16, 135–145.

Caratini, C., Bentaleb, I., Fontune, M., Morzadec-Kerfourn, M.T., Pascal, J.P., Tissot, C., 1994.A less humid climate since ca. 3500 yr BP from marine cores off Karwar, westernIndia. Palaeogeography, Palaeoclimatology, Palaeoecology 109 (2–4), 371–384.

Champion, H.G., Seth, S.K., 1968. A Revised Survey of the Forest Types of India. Managerof Publications, Delhi, India.

Costantini, I., 1983. The beginning of agriculture in the Kachi Plain: the evidence ofMehrgarh. In: Allchin, B. (Ed.), South Asian Archaeology 1981. Cambridge Univer-sity Press, Cambridge, pp. 29–33.

Day, P.R., 1965. Particle fractionation and particle-size analysis. In: Black, C.A. (Ed.), Methodsof soil analysis. Agronomy, 9. Am. Soc. of Agron, Inc., Madison, Wisconsin, pp. 545–577.

de Haug, G.H., Günther, D., Peterson, L.C., Sigman, D.M., Hughen, K.A., Aeschlimann, B.,2003. Climate and the collapse of Maya civilization. Science 299, 1717–1731.

de Menocal, P.B., 2001. Cultural responses to climate change during the late Holocene.Science 292, 667–673.

Dean Jr., W.E., 1974. Determination of carbonate and organic matter in calcareous sed-iments and sedimentary rocks by loss on ignition: comparison with other methods.Journal of Sedimentary Petrology 44, 242–248.

Dearing, J.A., Hay, K.L., Baban, S.M.J., Huddleston, A.S., Wellington, E.M.H., Loveland, P.J.,1996. Magnetic susceptibility of soil: an evaluation of conflicting theories using anational data set. Geophysical Journal International 127 (3), 728–734.

Deotare, B.C., Kajale, M.D., Kshirsagar, A.A., Rajaguru, S.N., 1998. Geoarchaeological andpalaeoenvironmental studies around Bap–Malar playa, district Jodhpur, Rajasthan.Current Science 75 (3), 316–320.

Diester-Haas, L., Schrader, H.J., Thiede, J., 1973. Sedimentological and paleoclimatologicalinvestigations of two pelagic ooze cores off Cape Barbas, North-West Africa. MeteorForschungsergebnisse, Deutsche Forschungsgemeinschaft, Reihe C 16, 19–66.

Doner, L., 2003. Late-Holocene paleoenvironments of northwest Iceland from lake sed-iments. Palaeogeography, Palaeoclimatology, Palaeoecology 193, 535–560.

Edwards, E.J., Smith, S.A., 2010. Phylogenetic analyses reveal the shady history of C4grasses. PNAS 107 (6), 2532–2537.

Engstrom, D.R., Wright Jr., H.E., 1990. Chemical stratigraphy of lake sediments as a re-cord of environmental change. In: Haworth, E.Y., Lund, J.W.G. (Eds.), Lake Sedi-ments and Environmental History. University of Minnesota Press, pp. 11–67.

Enzel, Y., Ely, L.L., Mishra, S., Ramesh, R., Amit, R., Lazar, B., Rajaguru, S.N., Baker, V.R.,Sandler, A., 1999. High-resolution Holocene environmental changes in the TharDesert, northwestern India. Science 284, 125–128.

Erdtman, G., 1943. An Introduction to Pollen Analysis. Chronica Botanica Co., WalthamMass, US.

Fredlund, G.G., Tieszen, L.L., 1994. Modern phytolith assemblages from the NorthAmerican Great Plains. Journal of Biogeography 21, 321–335.

Fuller, D.Q., 2002. Fifty years of archaeobotanical studies in India: laying a solid founda-tion. In: Settar, S., Korisettar, R. (Eds.), Indian archaeology in retrospect. Archaeologyand Interactive Disciplines, vol. III. Manohar Publishers, New Delhi, pp. 247–363.

Fuller, D.Q., Madella, M., 2001. Issues in Harappan archaeobotany: retrospect and prospect.In: Settar, S., Korisettar, R. (Eds.), Indian archaeology in retrospect. Protohistory, vol. II.Manohar Publishers, New Delhi, pp. 317–390.

Giosan, L., et al., 2012. Fluvial landscapes of the Harappan civilization. Proceedings ofthe National Academy of Sciences. http://dx.doi.org/10.1073/pnas.1112743109.

Gu, Y.S., Pearsall, D.M., Xie, S.C., Yu, J.X., 2008. Vegetation and fire history of aChinese site in southern tropical Xishuangbanna derived from phytolithand charcoal records from Holocene sediments. Journal of Biogeography35, 325–341.

Gupta, A.K., Anderson, D.M., Overpeck, J.T., 2003. Abrupt changes in the Asian south-west monsoon during the Holocene and their links to the North Atlantic Ocean.Nature 421, 354–357.

Haberle, S.G., David, B., 2004. Climates of change: human dimensions of Holocene en-vironmental change in low latitudes of the PEPII transect. Quaternary International118–119, 165–179.

Jackson, M.L., 1956. Soil Chemical Analysis-Advanced Course. Published by the AuthorDepartment of Soils, University of Wisconsin, Madison 6, Wisconsin.

Kajale, M.D., Deotare, B.C., 1997. Late Quaternary environmental studies on salt lakes inwestern Rajasthan, India: a summarised review. Journal of Quaternary Science 12(5), 405–412.

Knuze, G.W., 1965. Pre-treatment for mineralogical analysis. In: Black, C.A., et al. (Ed.),Methods of soil analysis. Agronomy, 9. Am. Soc. of Agron, Inc., Madison, Wisconsin,pp. 568–577.

Li, R., Carter, J.A., Xie, S., Zou, S., Gu, Y., Zhu, J., Xiong, B., 2010. Phytoliths and microcharcoalat Jinluojia archeological site in middle reaches of Yangtze River indicative ofpaleoclimate and human activity during the last 3000 years. Journal of ArchaeologicalScience 37, 124–132.

Livingstone, D.A., Clayton, W.D., 1980. An altitudinal cline in tropical African grassfloras and its paleoecological significance. Quaternary Research 13, 392–402.

Lu, H.Y., Wu, N.Q., Yang, X.D., Jiang, H., Liu, K.B., Liu, T.S., 2006. Phytoliths as quantita-tive indicators for the reconstruction of past environmental conditions in China I:phytolith-based transfer function. Quaternary Science Reviews 25, 945–959.

Madella, M., Fuller, D.Q., 2006. Palaeoecology and the Harappan civilisation of SouthAsia: a reconsideration. Quaternary Science Reviews 25, 1283–1301.

Maher, B.A., Thompson, R., 1992. Palaeoclimatic significance of the mineral magneticrecord of the Chinese loess and paleosols. Quaternary Research 37, 155–170.

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025

14 V. Prasad et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2013) xxx–xxx

Meadow, R., 1996. The origins and spread of agriculture and pastoralism in northwesternSouth Asia. In: Harris, D.R. (Ed.), The Origins and Spread of Agriculture and Pastoralismin Eurasia. UCL Press, London, pp. 390–412.

Misra, V.N., 1984. Climate, a factor in the rise and fall of the Indus civilization. In: Lal,B.B., Gupta, S.P. (Eds.), Frontiers of the Indus Civilization. Books and Books, NewDelhi, pp. 461–489.

Mulholland, S.C., Rapp Jr., G., 1992. A morphological classification of grass silica-bodies.In: Rapp Jr., G., Mulholland, S.C. (Eds.), Phytolith Systematics: Advances in Archae-ological and Museum Science. Plenum Press, New York, pp. 65–89.

Paddaya, K., 1994. Investigation of man–environment relationships in Indian archaeology:some theoretical considerations. Man and Environment XIX, 1–28.

Phadtare, N.R., 2000. Sharp increase in summer monsoon strength 4000–3500 cal yr BPin the central Himalaya of India based on pollen evidence from Alpine peat. Quater-nary Research 53, 122–129.

Pokharia, A.K., Kharakwal, J.S., Rawat, R.S., Osada, T., Nautiyal, C.M., Srivastava, A., 2011.Archaeobotany and archaeology at Kanmer, a Harappan site in Kachchh, Gujarat:evidence for adaptation in response to climate variability. Current Science 100(12), 1833–1846.

Possehl, G.L., 1997a. Climate and the eclipse of the ancient cities of the Indus. In: Dalfes,H.N., Kukla, G., Weiss, H. (Eds.), Third Millennium BC Climate Changes and OldWorld Collapse. Springer, Heidelberg, pp. 193–243.

Possehl, G.L., 1997b. The transformation of the Indus civilization. Journal of World Pre-history 11, 425–472.

Possehl, G.L., 1999. Indus Age. The Beginnings. University of Pennsylvania Press,Philadelphia.

Possehl, G.L., 2002. The Indus Civilization: A Contemporary Perspective. Altamira Press,Lanham.

Prasad, S., Enzel, Y., 2006. Holocene paleoclimates of India. Quaternary Research 66,442–453.

Prasad, S., Kusumgar, S., Gupta, S.K., 1997. A mid to late Holocene record ofpalaeoclimatic changes from Nal Sarovar: a palaeodesert margin lake in westernIndia. Journal of Quaternary Science 12, 153–159.

Prasad, V., Phartiyal, B., Sharma, A., 2007. Evidence of enhanced winter precipitationand the prevalence of a cool and dry climate during the mid to late Holocene inmainland Gujarat, India. The Holocene 17 (7), 889–896.

Prasad, V., Strömberg, C.A.E., Leaché, A.D., Samant, B., Patnaik, R., Tang, L., Mohabey,D.M., Ge, S., Sahni, A., 2011. Late Cretaceous origin of the rice tribe provides evi-dence for early diversification in Poaceae. Nature Communication 2, 480. http://dx.doi.org/10.1038/ncomms1482.

Rachna Raj, 2004. Fluvial response to Late Quaternary tectonic changes in the DhadharRiver basin, mainland Gujarat. Journal of the Geological Society of India 64, 666–676.

Rachna Raj, 2007. Late Pleistocene fluvial sedimentary facies, the Dhadhar River Basin,Western India. Quaternary International 159, 93–101.

Rai, V., 1990. Facies analysis and depositional environment of Pokaran saline Rann, DistrictJaisalmer, Rajasthan, India. Journal of the Geological Society of India 36, 317–322.

Rai, V., Sinha, A.K., 1990. Geological evolution of Kuchaman Lake, district Nagaur, Rajasthan.Journal of the Geological Society of India 35, 137–142.

Ratnagar, S., 1987. Pastoralists in the prehistory of Baluchistan. Studies in History 3,137–154.

Ratnagar, S., 2000. The End of the Great Harappan Tradition. Manohar Publishers, NewDelhi.

Roy, P.D., 2003. Mineralogical, Geochemical and Palaeoclimatic Characterization of theThar Desert Playas, Rajasthan, (India). [PhD thesis] Karlsruhe University 121.

Runge, F., 1999. The opal phytolith inventory of soils in central Africa — quantities,shapes, classification, and spectra. Review of Palaeobotany and Palynology 107,23–53.

Scott, L., 2002. Grassland development under glacial and interglacial conditions in south-ern Africa: review of pollen, phytolith and isotope evidence. Palaeogeography,Palaeoclimatology, Palaeoecology 177, 47–57.

Shaffer, J.G., Liechtenstein, D.A., 1989. Ethnicity and change in the Indus valley culturaltradition. In: Kenoyer, J.M. (Ed.), Old Problems and New Perspectives in SouthAsian Archaeology. Wisconsin Archaeological Reports, Madison, pp. 117–126.

Shankar, R., Prabhu, C.N., Warrier, A.K., Vijay kumar, G.T., Sekar, B., 2006. A multi-decadalrock magnetic record of monsoonal variations during the past 3,700 years from atropical Indian tank. Journal of the Geological Society of India 68, 447–459.

Sharma, J.P., Shyampura, R.L., Sehgal, J., 1994. Soils of Gujarat. National Bureau of SoilSurvey. Special Publication, 29.

Singh, G., 1971. The Indus Valley culture seen in the context of postglacial climatic andecological studies in north-west India. Archaeology and Physical Anthropology ofOceania 6, 177–189.

Please cite this article as: Prasad, V., et al., Mid–late Holocene monsoonal vuating climate culture relationsh..., Palaeogeography, Palaeoclimatology, P

Singh, G., Joshi, R.D., Singh, A.B., 1972. Stratigraphic and radiocarbon evidence for theage and development of three salt lake deposits in Rajasthan, India. Quaternary Re-search 2, 496–505.

Singh, G., Joshi, R.D., Singh, A.B., 1973. Pollen-rain from the vegetation of northwestIndia. New Phytologist 72, 191–206.

Singh, G., Joshi, R.D., Chopra, S.K., Singh, A.B., 1974. Late Quaternary history of vegeta-tion and climate in the Rajasthan Desert, India. Philosophical Transactions of theRoyal Society of London Series B 267, 467–501.

Singh, G., Wasson, R.J., Agrawal, D.P., 1990. Vegetational and seasonal climatic changessince the last full glacial in the Thar Desert, north-west India. Review of Palaeobotanyand Palynology 64, 351–358.

Singh, V., Prasad, V., Chakraborty, S., 2007. Phytoliths as indicators of monsoonal variabil-ity during mid–late Holocene in mainland Gujarat, western India. Current Science 92(12), 1754–1759.

Staubwasser, M., Sirocko, F., Grootes, P.M., Segl, M., 2003. Climate change at the4.2 ka BP termination of the Indus Valley civilization and Holocene Asian monsoonvariability. Geophysical Research Letters 30 (8) (7(1–4)).

Steig, E.J., 1999. Mid-Holocene climate change. Science 286, 1485–1487.Sundaram, R.M., Pareek, S., 1995. Quaternary facies and palaeoenvironment in north

and east of Sambhar Lake, Rajasthan. Journal of the Geological Society of India46, 385–392.

Swain, A.M., Kutzbach, J.E., Hastenrath, S., 1983. Estimates of Holocene precipitation forRajasthan, India, based on pollen and lake level data. Quaternary Research 19,1–17.

Tanner, C.B., Jackson, M.L., 1947. Monographs of Sedimentation times for soil particlesunder gravity or centrifugal acceleration. Soil Science Society of America Proceed-ings 12, 50–55.

Tengberg, M., 1999. Crop Husbandry at Mili Qalat, Makran, SW Pakistan (4000–2000BC). Vegetation History and Archaeobotany 8, 3–12.

Tieszen, L.L., Senyimba, M.M., Imbamba, S.K., Troughton, J.H., 1979. The distribution ofC3 and C4 grasses and carbon isotope discrimination along an altitudinal and mois-ture gradient in Kenya. Oecologia 37, 337–350.

Twiss, P.C., 1980. Opal phytoliths as indicators of C3 and C4 grasses. Geological Societyof America Abstracts 12, 17.

Twiss, P.C., 1987. Grass-opal phytoliths as climatic indicators of the Great Plains Pleis-tocene. In: Johnson, W.C. (Ed.), Quaternary environments of Kansas: Kansas Geo-logical Survey Guidebook Series, 5, pp. 179–188.

Twiss, P.C., 1992. Predicted world distribution of C3 and C4 grass phytoliths. In: Rapp Jr.,G., Mulholland, S.C. (Eds.), Phytolith systematics. Emerging issues. Advances in Ar-chaeological and Museum Science I. Plenum Press, New York.

Verosub, K.L., Roberts, A.P., 1995. Environmental magnetism: past, present and future.Journal of Geophysical Research 100, 2175–2192.

Walden, J., 1999. Sample collection and preparation. In: Walden, J., Oldfield, F.,Smith, J. (Eds.), Technical Guide, No. 6. Quaternary Research Association,London, pp. 63–88.

Wang, W.M., 2003. Climate indexes of phytoliths from Homo erectus' cave deposits inNanjing. Chinese Science Bulletin 48, 2005–2009.

Wang, Y.J., Lu, H.Y., 1993. The Study of Phytolith and its Application. China Ocean Press,Beijing (228 pp.).

Warrier, A.K., Sandeep, K., Harshavardhana, B.G., Shankar, R., Pappu, S., Kumar, A.,Prabhu, C.N., Gunnell, Y., 2011. A rock magnetic record of Pleistocene rainfall var-iations at the Palaeolithic site of Attirampakkam, Southeastern India. Journal of Ar-chaeological Science 38, 3681–3693.

Wasson, R.J., Smith, G.I., Agrawal, D.P., 1984. Late Quaternary sediments, minerals, andinferred geochemical history of Didwana Lake, Thar Desert India. Palaeogeography,Palaeoclimatology, Palaeoecology 46, 345–372.

Weber, S.A., 2003. Archaeobotany at Harappa: indications for change. In: Weber, S.A.,Belcher, W.R. (Eds.), Indus Ethnobiology. New Perspectives from the Field. LexingtonBooks, Lanham, pp. 175–198.

Weiss, H., Bradley, R.S., 2001. Archaeology. What drives societal collapse? Science 291,988.

Wengxiang, W., Tungsheng, L., 2004. Possible role of the “Holocene Event 3” on the col-lapse of Neolithic cultures around the central plain of China. Quaternary Interna-tional 117, 153–166.

Williamson, D., Jelinowska, A., Kissel, C., Tucholka, P., Gibert, E., Gasse, F., Massault, M.,Taieb, M., Van Campo, E., Wieckowski, E., 1998. Mineral magnetic proxies of erosion/oxidation cycles in tropical marr-lake sediments (Lake Tritivakely, Madagascar):palaeoenvironmental implications. Earth and Planetary Science Letters 155, 205–219.

Wright, R.P., Bryson, R., Schuldenrein, J., 2008. Water supply and history: Harappa andthe Beas regional survey. Antiquity 82, 37–48.

ariations from mainland Gujarat, India: A multi-proxy study for eval-alaeoecology (2013), http://dx.doi.org/10.1016/j.palaeo.2013.05.025