The global character of the time interval known asthe Little Ice Age (LIA) is at present relatively wellestablished. The forcing mechanisms behind thiscooling interval, however, are still elusive. Investiga-tions in annually-laminated sediments have shownthat varved sediments are among the best climaticarchives to tackle these questions.

Proglacial Lago Frías in northern Patagonia is fedby the Tronador ice cap (3554 m). Previous investi-gations have shown that this glacier has reactedsensitively to climate change during distinct episodessuch as the Late Glacial-Holocene transition (Ariz-tegui et al., 1997; Hajdas et al, 2003), and the Me-dieval Warm Epoch and the LIA, with well identifiedmajor glacial advances between A.D. 1800-1850(Rabassa et al., 1979; Villalba et al., 1990). A mete-orological record for the 1969-1985 interval at theMascardi weather station, located at 20 km from theFrías Glacier, shows a mean annual temperature of7.6 °C, varying from 12.9 °C in January to 2.4 °C inJuly. Additionally, these data indicate a strong pre-cipitation gradient between Mascardi (1409 mm year-1) and the River Frías valley (4300 mm year -1).

During the twentieth century the El Niño/SouthernOscillation (ENSO) and ENSO-like phenomena havedominated climate variations in the Americas on in-terannual and decadal time scales, respectively. TheENSO impact on local climate has been well deter-mined using meteorological, historical and a dendro-chronological approach at the Frías valley. More re-cently, a sediment trap study in Lake Mascardicovering the 1992-1998 interval combined with me-teorological data have shown changes in sedimenta-tion rates that can be linked to ENSO climaticevents. Since both proglacial lakes, Frías and Mas-cardi, are fed by the Tronador ice cap, the laminatedsequence of Lago Frías can provide a continuousrecord of ENSO and ENSO-like variations throughtime.

Results of a multiproxy study of Lago Frías sedi-ments reflect variations in the transport of glaciallyderived clay and silt to the basin that can be directlylinked to changes in climate. Sedimentological evi-dence combined with a good chronological modelindicate variations in varves thickness showing twofrequencies centered at 16.4 and 10.5 years thathave been previously attributed to the solar cyclesand the Tropical Atlantic Sea Surface Dipole (TAD),respectively. The main frequency is, however, lo-cated between 2.5 and 3.0 years pointing towards adominant ENSO signal. Thus, the Lago Frías recordprovides new insights about the myriad of complexforcing mechanisms behind the cooling during theLIA in an area of the planet with a paucity of high-resolution climate records. It also points out the needto develop regional networks of well-understood pa-leoclimate archives such as lake systems with a reli-able chronology. They will be critical evaluating theimpact of the various mechanisms triggering envi-ronmental changes at different spatial and temporalscales.

REFERENCES

Ariztegui, D., Bianchi, M. M., Masaferro, J., Lafargue, E. & Ni-essen, F. (1997): Interhemispheric synchrony of Late-glacial climatic as recorded in proglacial Lake Mascardi,Argentina. Journal of Quaternary Sciences 12/4:333-338.

Hajdas, I., Bonani, G., Moreno, P. I. & Ariztegui, D. (2003):Precise radiocarbon dating of Late-Glacial cooling in mid-latitude South America. Quaternary Research 59/1:70-78.

Rabassa, J., Rubulis, S. & Suárez, J. (1979): Rate of formationand sedimentology of (1976-1978) push-moraines, FríasGlacier, Mount Tronador (41°10'S; 71°53"W), Argentina. InSchlüchter, Ch. (edt.), Moraines and Varves, Rotterdam,Balkema:65-79.

Villalba, R., Leiva, J. C., Rubulis, S., Suarez, J. & Lenzano, L.(1990): Climate, Tree-rings and Glacial Fluctuations in theRio Frías Valley, Rio Negro, Argentina. Arctic and AlpineResearch 22:215-232.

Dominant ENSO frequencies during the Little Ice Age (LIA)in the rainshadow of the Andes -

The laminated record of proglacial Lago Frías, northern Patagonia(Argentina)

Ariztegui, D., Bösch, P. & Davaud, E.

Institute Forel and Dept. of Geology and Paleontology, 13 rue des Marîchers, 1205 Geneva, Switzerland

The Andes represent the locus of continued plateconvergence from the Early Palaeozoic to the pre-sent day. However the early evolution of much ofthe proto-Andean margin remains poorly understood.This is because exposures of pre-Andean basementrocks are in many places extremely limited, eitherbeing obscured by later tectonic events along theconvergent margin or buried by the ubiquitous vol-canic cover. This problem is particularly acute in thenorthern Andes, where Precambrian basement is notexposed for over 2000 km along strike. This corre-sponds to the distance between the northern extentof the Arequipa-Antofalla block (Fig. 1), a Protero-zoic crustal block which has experienced Grenvillemetamorphism (Loewy et al. 2004) and the south-ernmost basement exposures in Colombia, the Pro-terozoic Garzón inlier (Restrepo-Pace et al. 1997).This zone, extending from 15º S in southern Peru to2º N in southern Colombia (Fig. 1) is also character-ized by substantial development of Andean forelandsediments to the east, so that the basement geologyperipheral to the orogen is not known with any de-gree of certainty.

However, in the Eastern Cordilleras of Peru andEcuador, Palaeozoic metasedimentary sequencesare well exposed, along with several generations ofcross-cutting Palaeozoic granitoid intrusions. Mostof these sequences, including the Marañon Complexin Peru and the Isimanchi Formation of the CordilleraReal in Ecuador are considered to be autochthonouswith respect to the Gondwanan margin. Hence theirheavy mineral assemblages, and in particular theirdetrital zircon populations, should reveal informationregarding their source areas - in this case theGondwanan margin of the northern Andes. Fur-thermore, granitic magmas which intrude these se-quences can carry inherited zircon during ascent.These inherited zircons can also provide source in-formation, particularly with respect to deeper crustal

levels which would otherwise be impossible to sam-ple.

This study incorporates U-Pb LA-ICPMS dating ofdetrital zircons from various Gondwanan margin se-quences in Peru and Ecuador. Additionally, conven-tional U-Pb zircon and U-Pb titanite dating was un-dertaken to constrain the timing of metamorphismand magmatism in the Eastern Cordillera of Peru.

Figure 1. Geology of South America illustrating tectonic prov-inces and region of interest in this study.

Evolution of the Gondwanan marginof the northern Andes

1Chew, D. M., 1Schaltegger, U., 2Kosler, J., 1Fontignie, D., 1Spikings, R., 1Miskovic, A.

1 Department of Mineralogy, University of Geneva2 Department of Earth Sciences, University of Bergen

U-PB LA-ICPMS DETRITAL ZIRCON DATA

All samples exhibit prominent peaks between c. 0.45– 0.65 Ga and c. 0.9 – 1.3 Ga, and display a minimalamount of detritus older than c. 2 Ga. Between athird and a half of all grains in each sample lie withinthe 0.45 – 0.65 Ga age range. Nearly all samplesexhibiting a prominent peak between 0.45 and 0.5Ga, which temporally overlaps with the onset of sub-duction-related magmatism in the Famantinian ter-rane (Fig. 1) of northern Argentina, the Arequipa ter-rane of Southern Peru (Fig. 1, Loewy et al. 2004),Venezuela (the Caparo Arc of Bellizzia & Pimentel,1994) and Colombia. Further prominent peaks arefound within the c. 0.5 – 0.65 Ga age range, agestypical of the Brasiliano / Pan African orogenic cy-cles. The nearest rocks of equivalent age within theAndean chain are the c. 535 Ma subduction-relatedgranitoids of the Sierra Pampeanas (Fig. 1) in north-ern Argentina (Rapela et al. 1998) or the Neopro-terozoic (c. 650-580 Ma) basement gneisses inVenezuela (Bellizzia & Pimentel 1994). Additionally,there are no rocks of equivalent age exposed on thewestern side of the Amazonian craton – the exten-sive Neoproteozoic tectonic provinces of Brazil (theBrasiliano orogenic cycle) are located on the easternside (Fig. 1). It is difficult to invoke the eastern Ama-zonian craton as a potential source because theminimal amount of old detritus (i.e. Archean andPaleoproterozoic) argues that the core of the Ama-zonian craton (Fig. 1) was not a significant source ofdetritus to the Gondwana margin during the Palaeo-zoic.

Significantly less detritus lies within the c. 0.7 –0.9 Ga age range, while between a third and a halfof all grains in each sample lie within the 0.9 – 1.3Ga age range, with the largest peaks commonly en-countered between 1 and 1.2 Ga. The Sunsas oro-gen of the SW Amazonian craton is invoked as apotential source region for the abundant c. 0.9 – 1.3Ga peaks observed. It is interpreted that the Sunsasorogen continues in a northwestern direction under-neath the thick pile of foreland sediments that aredeveloped to the east of the Northern Andes (Fig. 1),where it is likely to be contiguous with the c. 1 Gagneissic basement inliers in the Columbian Andes(Fig. 1, Restrepo-Pace et al. 1997).

We envisage that the basement to the Gondwa-nan margin to the northern Andes was composed ofa metamorphic belt of Grenvillian age, upon whichwas sited a late Neoproterozoic – Early Palaeozoic(subduction-related?) magmatic belt, similar to theSierra Pampeanas of northern Argentina.

TIMING OF EARLY PALAEOZOIC METAMOR-PHISM IN PERU

Existing age constraints on the timing of metamor-phism of the Eastern Cordillera of Peru (MarañonComplex) are very poor. The pronounced c. 480 Madetrital zircon peak observed in Marañon Complexrocks in this study implies that the metamorphismmust be post-480 Ma, and that the c. 600 Ma U-Pbzircon ages from granulitic gneisses in central Peruof Dalmayrac et al. (1980) are a result of inheritance.Additionally, we have constrained the timing ofmetamorphism in at least part of the Marañon Com-plex to Silurian or younger. High-grade metamorphicassemblages (c. 650º C, 11 kb) are developed withinorthogneisses which yield U-Pb magmatic zirconages of 442.4 ± 1.4 Ma (TIMS) and 444.2 ± 6.4 Ma(LA-ICPMS). The timing of metamorphism is cur-rently refined by a combination of Sm-Nd garnet, Ar-Ar hornblende and U-Pb titanite dating.

The U-Pb detrital zircon data imply that orogenicbelts equivalent in age to the Famantinian / Pam-pean and Grenville (= Sunsas) orogenies wereavailable to supply detritus to the Palaeozoic se-quences of the Northern Andes. Combined with evi-dence that metamorphism of the Eastern Cordilleraof Peru document is Silurian or younger, it appearsthe Early Palaeozoic evolution of the Gondwananmargin of the Northern Andes is similar to that ofboth northern Argentina and Colombia / Venezuela,with broadly comparable tectonic histories over sev-eral thousand kilometers along strike.

REFERENCES

Bellizzia, A. & Pimentel, N. 1994. Terreno Mérida: Un cinturónalóctono Herciniano en la Cordillera de Los Andes deVenezuela. V Simposio Bolivarano Exploración Petroleraen las Cuencas Subandinas, Memoria, p. 271-299.

Dalmayrac, B., Laubacher, G. Marocco, R. (1980) Géologiedes Andes Péruviennes. Caractères généraux del’évolution géologique des Andes péruviennes : Travaux etDocument de l’ORSTROM. v.122, 501pp.

Loewy, S.L., Connelly, J.N., Dalziel, I.W.D. 2004 An orphanedbasement block; the Arequipa-Antofalla basement of theCentral Andean margin of South America. Geological Soci-ety of America Bulletin. 116; 1-2, 171-187.

Restrepo-Pace P.A., Ruiz, J., Gehrels G., Cosca M. 1997 Geo-chronology and Nd isotopic data of Grenville-age rocks inthe Colombian Andes: new constraints for Late Proterozoic-Early Paleozoic paleocontinental reconstructions of theAmericas Earth and Planetary Science Letters 150 (3),427-441.

There are correlated along-strike variations ofmagma chemistry and mineralogy, and proportionsof mafic and more evolved magmas with regionalelevation (proxy for crustal thickness) of the Andeancordillera among frontal arc volcanoes of the ‘acces-sible’ Southern Volcanic Zone (SVZ; 33-41ES).These regional-scale patterns are extremely usefulfor defining testable hypotheses and identifying pos-sible solutions, but such variations cannot be inter-preted rigorously in terms of the processes andcomponents involved in magma genesis and differ-entiation. Consequently, there is a lively debate con-cerning the relative roles and contributions of thesubducted plate, subduction erosion, depth of mantlemelting, and magma interactions with the continentallithosphere in generating these regional patterns.

We have evaluated over 2000 major and traceelement analyses of volcanic rocks as functions of:(1) the related parameters of latitude, volcano eleva-tion, structural position, and segmentation of the arc,(2) the age and structure of the Nazca plate beingsubducted beneath this portion of the SVZ, and (3)the compositions of Oligo-Miocene granitoid ba-tholiths underlying the arc between 35E and 38ES.Our general conclusions are that magmatic varia-tions along the arc are products of both lower plateand upper plate contributions, and that local crustalcomposition may play as large a role as crustalthickness during magma contamination. We recog-nize the distinctive signature of high fluid-flux asso-ciated with subducted fracture zones between 36Eand 41E, but suggest that crustal overprints in-creasingly dominate over a detectable lower platesignature in evolved magmas north of 36E.

Along-strike magmatic patterns defined by Quaternary frontal arcvolcanoes of the Andean Southern Volcanic Zone (SVZ; 33-41ES),

Central Chile

*Dungan, M.A., *Sellés, D., *Spikings, R., **Hildreth, W., ***López-Escobar, L., ****Frey, F.,*****Davidson, J., *Rodríguez, C. & *Bachmann, O.

* Section of Earth Sciences, University of Geneva, 13 rue des Maraîchers, Geneva 1205, Switzerland** U.S. Geological Survey, 345 Middlefield Rd, Menlo Park, CA 94025 USA

*** Instituto GEA, Universidad de Concepción, Casilla 106, Concepción, Chile**** Department of Earth, Atmospheric, and Space Sciences, MIT, Cambridge, MA 02139 USA***** Department of Earth Sciences, University of Durham, DH1 3LE, South Road, Durham, UK

Section of Earth Sciences, University of Geneva, 13 rue des Maraîchers, Geneva 1205, Switzerland

U.S. Geological Survey, 345 Middlefield Rd, Menlo Park, CA 94025 USA

Instituto GEA, Universidad de Concepción, Casilla 106, Concepción, Chile

Department of Earth, Atmospheric, and Space Sciences, MIT, Cambridge, MA 02139 USA

Department of Earth Sciences, University of Durham, DH1 3LE, South Road, Durham, UK

Seismic tomographic images of subduction zonesare characterized by strong positive and negativeanomalies in the wedge above the slab. We havedeveloped a 2-D thermo-chemical numerical modelto simulate the dynamics of such anomalies duringsubduction. By using an internal consistentpetrological-thermomechanical approach we investi-gate heat, water and melt transport in a subductionzone including the propagation of partially moltenupwellings (cold plumes) originating from the sub-ducting slab. The model demonstrates two chemi-cally distinct types of plumes (mixed and unmixed),and various rigid body rotation phenomena in thewedge (subduction wheel, fore-arc spin, wedge pin-ball). These thermal-chemical features strongly per-turb seismic structure, and the results are compara-ble with seismic tomographic images of subductionzone. For example, high Qp anomalies (P-wave at-tenuation) have been found below Salar de Atacamabasin (Schurr et al, 2003), and seismic propertiessuggest a relatively cold and rigid block sank inlighter material. Inferring from our results we can ex-plain these anomalies as subduction “wheel” phe-nomenon: a fragment of dry mantle wedge sur-rounded by hydrated partially molten rocks. Thefragment is trapped between the plume and the sub-ducting slab, and is put into stable rotational motion.This is associated with significant (by 200-300 K)lowering of temperature that causes gradual appear-ance of strong positive seismic velocity anomalieswithin a rotating body.

Figure 1. Picture shows subduction “wheel” phenomenon,where part of a mantle wedge is surrounded by lighter materialof rising upwelling.

REFERENCE

Schurr, B., Asch G., Rietbrock A., Trumbull R., Haberland Ch.(2003): Complex patterns of fluid and melt transport in thecentral Andean Subduction zone reveald by attenuationtomography. Earth and Planet Sci. Let., 215(1-2), 105-119.

Dynamics of seismic structures above slabs:insight from thermal-chemical numerical models

1Gorczyk, W., 1,2Gerya, T. V., 1Connolly, J. A.d., 3Yuen, D. a. & 4Rudolph, M.

1 Geological Institute, Swiss Federal Institute of Technology (ETH - Zürich), CH-8092 Zurich, Switzerland

2 Institute of Experimental Mineralogy, Russian Academy of Sciences, 142432 Chernogolovka, Moscow, Russia

3 Department of Geology and Geophysics, University of Minnesota, Minneapolis, Minnesota 55455-0219, USA

4 Oberlin College, Oberlin, Ohio 44074, USA

The Quindó-Risaralda Basin and the northern part ofthe Cauca depression lie between the Western andCentral Cordilleras in Central Colombia (Figs. 1 and2). This zone is dissected by the SSW-NNE trendingRomeral fault system which defines the boundarybetween oceanic basement to the west and conti-nental basement to the east. At this latitude thesystem has a left-lateral strike-slip component. Alarge part of this subsiding area was filled by Tertiarycontinental sediments (Fig. 3). Following the cata-strophic earthquake of Armenia (Fig. 2) in 1999, adetailed geological study of the Quaternary infill ofthis region has been undertaken aiming at estab-lishing the mechanisms which may have led to theopening of this interandean depression.

The late Andean tectonic phase caused the fold-ing of the Serranía de Santa Barbara, thereby subdi-viding the basin into two parts (Figs. 2 and 3). Whilethe western part was infilled by the fluvial-lacustrinesediments of a paleo-Cauca River, volcaniclasticmass flows derived from the Ruiz-Tolima volcanicsystem in the Central Cordillera accumulated in theeastern part.

Fluvio-lacustrine Quaternary sediments depositedin the Cauca valley form the Zarzal Formation(Neuwerth et al., 2005). In the eastern part of thebasin they interfinger with the volcaniclastic massflow deposits forming the Cartago, Pereira andQuindío Fans (Fig. 2). Based on field and paly-nological data, a sedimentary model (Suter et al.,this symposium) has been proposed that shows thesynchroneity of the deposition of the Zarzal Forma-tion with that of the volcaniclastic fans. Within thelast million years, this area has been exposed to anintense tectonic activity (evidenced by the wide-spread occurrence of seismites, Neuwerth et al.,

2005) and the Cauca Valley underwent a higher rateof subsidence.

Figure1. Location map

The Quindío-Risaralda Basin (east of the Serraniade Santa Barbara, Fig. 2) was filled during the Qua-ternary by a sequence of stacked volcaniclasticmass flows. According to their stratigraphic succes-sion, lateral continuity, genesis and sedimentologicalparameters, several individual units can be distin-guished within these deposits (Guarin et al., 2005and this symposium). So far the origin of subsidencein the Quindío-Risaralda Basin can not be clearlyexplained. Nevertheless, a sum of observations linkthe active fault systems, the present-day drainagepattern and the distribution of the volcaniclasticunits. The basin and the volcaniclastic fans bear theimprint of three dominant major fault trends known ata more regional scale (Guarin et al., this sympo-sium). The interaction of this transpressional multipleactive system led to the formation of localized pull-apart basins that became depositional lows for thevolcaniclastic units.

Contribution of Quaternary sediments to the understanding of thetectonic history in Central Colombia: The volcaniclastic fans inQuindío-Risaralda and the Zarzal Formation in the Cauca Valley

Gorin, G., Guarin, F., Neuwerth, R., Suter, F., *Espinosa, A. & **Guzman, C.

Department of Geology-Paleontology, University of Geneva, Switzerland; Georges.Gorin, Fernando.GuariRalph.Neuwerth, Fiore.Suter@terre.unige.ch

* Faculdad de Ingeneria, University of Quindío, Armenia, Colombia; espinosaar@epm.net.co** Department of Geological Sciences, University of Caldas, Manizales, Colombia; cguzmanl@yahoo.com

Figure 2. Regional distribution of Pleistocene sediments in theCauca Valley, Risaralda and Quindío. Location of figure 3.

Similarly, field observations in the Zarzal Forma-tion show lots of evidence of extension and thenorthern part of the Cauca depression might also berelated to strike-slip movements. However, at thispoint, further datings are needed to refine the dy-namic interpretation of this basin during the last mil-lion years. The whole area needs to be further stud-ied from a tectonic and neotectonic point of view.

ACKNOWLEDGEMENTS

This research is supported by the Swiss NationalScience Foundation (grant no. 200020-107866/1).

Figure 3. Geological cross-section across the Cauca Valleyand the Quindío Fan (see location in figure 2). The QuindíoFan and the Zarzal Formation are the subject of this re-search.

REFERENCES

Guarin, F., Gorin, G.& Espinosa, A. (2005): A Pleistocenestacked succession of volcanic mass flows in Central Co-lombia: the Quindío-Risaralda Fan. Acta Vulcanologica (inpress).

Guarin, F., Gorin, G.& Espinosa, A. (2005): The interandeanQuindío-Risaralda basin in Central Colombia and its Pleisto-cene infill by stacked volcaniclastic mass flows derived fromthe Central Cordillera. 3rd Swiss Geoscience Meeting, Zurich2005.

Neuwerth, R., Suter, F., Guzman, C. & Gorin, G. (2005):Soft-sediment deformations in a tectonically active area :the Pleistocene Zarzal Formation in the Cauca Valley(Western Colombia). Sedimentary Geology (in press).

Suter, F., Neuwerth, R., Guzman, C. & Gorin, G. (2005): De-positional model for the Quaternary Zarzal Formation(Colombia) and its stratigraphic relationship with the vol-caniclastic mass flows derived from the Central Cordil-lera. 3rd Swiss Geoscience Meeting, Zurich 2005.

Due to the extreme latitudinal rainfall gradient inChile, reaching from hyper-arid conditions in theAtacama desert through semi-arid Mediterraneantype conditions in central Chile to extremely highrainfall values in the southern mountains, the terri-geneous sediment input to the ocean off Chile, andthus offshore sedimentation rates, increases signifi-cantly to the south. This pattern of increasing sedi-mentation rates is clearly revealed by a set of datedsediment cores recovered from the Chilean conti-nental margin between 24°S and 44°S.

Holocene sedimentation rates increase from ~5cm/kyr off arid northern Chile, to ~10 cm/kyr offsemiarid north-central Chile, to ~50 cm/kyr off humidcentral Chile, to ~100 cm/kyr at 41°S off year-roundhumid southern Chile, and finally to ~200 cm/kyr at~44°S where rainfall reaches extremely high values.Sedimentation rates were even higher during the lastglacial reaching from ~10 cm/kyr at 27°S, to ~50cm/kyr at 33°S, and to ~200 cm/kyr at 41°S. Thesepronounced north-south gradients at the Chileancontinental margin provide a unique opportunity toreconstruct late Quaternary paleoenvironmentalchanges both on the continent and in the ocean withtime resolutions ranging from Milankovitch-scale (i.e.~20,000 to ~100,000 year variability) off arid north-ern Chile to centennial-scale off year-round humidsouthern Chile.

All along the Chilean margin there are clear indi-cations for enhanced marine productivity during thelast glacial compared to early and mid–Holocenetimes. A detailed analyses of past planktic fora-miniferal faunal compositions off central and north-ern Chile reveals distinct periods with relatively lowproductivities before ~ 32,000 cal yr BP, highestproductivities between 32,000 and 16,000 cal yr BP,and lowest productivities after 16,000 cal yr BP to-wards the present. Interestingly, the onset and end-ing of these periods do not show any latitudinal pref-

erences. Relative variations in the paleoproductivitycalculated from accumulation rates (AR) of biogeniccompounds generally support these findings, al-though with some local differences, mainly duringthe last deglaciation.

The predominantly southward increase in marineproductivity, as it is also seen under present-dayconditions, is a continuous feature throughout thelast 40,000 years. This pattern implies that the samecirculation systems have been responsible for pro-ductivity variations off Chile over this time span. Thecomparatively consistent pattern of the planktic fo-raminiferal fauna across ~10° of latitude (24°S to33°S) over the last 40,000 years implies large-scalechanges in the oceanic circulation synchronouslyaffecting the whole area, i.e., changes in the positionand/or advection of the Peru Chile Current (PCC) /Antarctic Circumpolar Current (ACC) system. Duringthe last glacial, a more northerly position for the ACCprobably led to higher productivities off Chile com-pared to the Holocene. Such a northward shift of theACC brings this main macronutrient-source closer tothe PCC area, thus providing more nutrients and in-creasing productivity.

Differences in the latitudinal paleoproductivitypattern deduced from biogenic compounds can berelated to local differences in this region. At 33°Sand further south productivity was additionallyboosted by enhanced riverrine micronutrient supplydue to northward displacement of the SouthernWesterly belt (SW), resulting in a northward shift ofthe associated rainfall gradient. This setting led tothe highest productivities recorded in this part of thestudy area around the Last Glacial Maximum (LGM),when the main circulation systems (ACC and SW)were at their northernmost position. However, northof 33°S, the highest productivities are recorded be-fore and after the LGM, suggesting additional or dif-ferent factors controlling the paleoproductivity. Con-

The Late Quaternary paleoenvironment of Chileas seen from marine archives

*Hebbeln, D., *Kaiser, J., **Lamy, F., *Mohtadi, M. & *Stuut, J.-B.

* MARUM – Center for Marine Environmental Sciences, University of Bremen, Germany

** Geo-Forschungszentrum Potsdam, Germany

tinental records from northern Chile do not imply anyinfluence of the Southern Westerlies since latePleistocene times, which might explain the differentproductivity patterns north of 33°S. Within this area,besides the more northerly position of the ACC, asteeper hemispheric thermal gradient causingstronger meridional and zonal winds and strongercoastal upwelling led to generally higher productiv-ities during the last glacial. The period of highestproductivity in marine sediments during the deglaci-ation coincides well with the observed late glacialhumid phases in northern Chile and Bolivia. En-hanced precipitations onshore and flooding of theshelves might have launched additional nutrientsources during deglaciation, leading to the highestpaleoproductivities in this part of the PCC.

During Late Holocene times productivity increasedwithin the PCC, most likely due to another northwardshift in the position of these zonal circulation sys-tems. Such a shift towards more humid conditions innorthern and central Chile in the Late Holocene isindicated in various archives. In addition, an intensi-fication of ENSO events occurred in the last 7000years when rapid fluctuations in the composition ofthe planktic foraminifera fauna started to occur. Thetiming of the onset (~7000 cal yr BP) and the inten-sification (~5000 cal yr BP) of ENSO in the marinerecords off Chile are consistent with records fromother sites in the Eastern Pacific.

Alkenone-based reconstructions of sea surfacetemperatures (SST) reveal a common trend alongthe Chilean margin with glacial SSTs being 6°C(41°S) to 2°C (24°S) colder than SSTs during theHolocene climate optimum (HCO), when SSTs were1-2°C warmer than today. These data point tostronger-than-today SST gradients during the LGM(ACC/SW shifted equatorward) and relatively weakgradients during the HCO (poleward shift).

For the continental paleoenvironment, there isstrong evidence for precession-driven variations inthe Norte Chico (27°S) marked by increased on-shore precipitation related to relatively northward po-sitions of the Southern Westerlies during precessionmaxima (including the LGM) and vice versa. Theseprecessional cycles are evident in different sedi-mentological parameters that suggest substantialchanges of sediment provenance, weathering re-gimes in the source areas, and modes of sedimentinput to the ocean.

Further south, off central Chile (~33°S), highersedimentation rates allow a more detailed recon-struction of climatic evolution over the last ~30,000years. As with the Norte Chico record, sedimen-tological data clearly record changes in the relative

contribution of Andean versus Coastal Range sourcerocks. However, as increased precipitation has en-abled rivers originating from the Andes to cutthrough the Coastal Range, these data instead re-cord the relative dilution of the Andean signal by ad-ditional input of Coastal Range material during com-paratively humid intervals. The sedimentologicalrecords clearly suggest more humid conditions dur-ing the LGM, which is consistent with the NorteChico record. The deglaciation is characterized by atrend towards a more arid climate that reached itsmaximum during mid-Holocene times (8000-4000 calyr B.P.). The Late Holocene was marked by morevariable paleoclimates with a return to slightly morehumid conditions.

Off humid southern Chile (41°S), where extremelyhigh sedimentation rates allow a reconstruction ofterrigeneous sediment input on centennial time-scales, Holocene “long-term” trends are generallyconsistent with the results off central Chile. On mil-lennial to multi-centennial-scales, terrigeneous inputchanges reveal dominating bands of variability cen-tred on ca. 900 and 1500 years suggesting signifi-cant rainfall changes and thus latitudinal shifts of theSouthern Westerlies on these time-scales. RecentODP drilling off southern Chile now offers thechance to extend the very high resolution recordsinto the last glacial. Under these glacial conditions,ODP Site 1233 was located in an ideal position tomonitor Patagonian ice-sheet (PIS) extent variationsby recording compositional changes in the regionalterrigeneous sediment input to the continental mar-gin. A rapid retreat of the PIS at ~17.5 kyr B.P. isclearly documented in a significant decrease of seasurface salinities at Site 1233 suggesting large in-puts of melt-water. Thereafter, the record of Site1233 becomes more and more dominated by rainfallchanges.

Combining the available datasets, it can be de-duced that paleoenvironmental variability acrossboth onshore and offshore Chile mainly depends onthe large scale oceanic/atmospheric circulation in theSE Pacific, which in turn is coupled to global climatevariability. While especially the ACC – SW system isclosely connected to Antarctic climate variability, it isthe south driving most of the paleoenvironmentalvariability in Chile. However, north of ~30°S the im-pact of the tropical system gains influence and has astrong impact in northern Chile.

The coastal lowlands of Ecuador, representing theouter Andean forearc between 1°N and 3°S, arecomposed of distinct structural blocks separated bymajor NNE and WNW oriented faults (Puerto Cayo,Canande and Chongòn-Colonche faults; Fig.1).Although their allochthonous derivation has beenproven (Roperch et al. 1987, Luzieux et al. 2005),the origin and the timing of accretion of the individualoceanic remnants is still poorly known. We presentnew paleomagnetic and 40Ar/39Ar data derived fromCretaceous basaltic lavas and their Late Cretaceoussedimentary cover rocks, which define the origin ofthe individual, fault-bounded blocks within theEcuadorian forearc.

Mafic, oceanic plateau derived rocks of the PiñonBlock (PB) are exposed in the southwestern part ofthe forearc (Fig. 1). Gabbros from volcanic basementwere sampled and separated. Hornblendes weredated by the 40Ar/39Ar step heating method at 88±1.6Ma, corresponding to the cooling of the plateaumaterial. This in not in accordance with earlierstratigraphic correlations of a pre-Turonian age(Benitez 1995, Jaillard et al. 1995). Ourpaleomagnetic investigation, additionally supportedby new micropaleontological results from theoverlying Calentura and (?) Cayo fms. indicate that afirst 30-40° clockwise rotation has taken placebetween 70 and 75 Ma, suggesting that the PBexperienced a major tectonic event during the MiddleCampanian. The remaining 10-20° rotation into thepresent position was probably achieved during theEocene final accretion of the blocks. The shallowpaleomagnetic inclination values measured inbasement and sediments point to an equatoriallatitude origin of the plateau sequence.

The San Lorenzo Block (SLB), which is in faultedcontact with the PB (Fig. 1), is composed of a maficoceanic plateau basement sequence, which isoverlain by Campanian (on biostratigraphic control)

island arc lavas and volcanoclastic deposits ofCampanian-Maastrichtian age (Cayo Fm.).Measured declinations within the basement and arclavas show a clockwise rotation of 60-65° sincecooling, and a mean value of about 25° clockwiserotation since the deposition of the Cayo Fm.Paleomagnetic inclinations show little variation andagain strongly support an equatorial position duringthe crystallisation of the basalts. The very similarvalues and timing of rotations in comparison with thePB strongly suggest a common origin and evolutionof both blocks.

The paleomagnetic data acquired from thecoastal blocks prove an equatorial latitude during theformation of the magmatic basement. Bothbasement sequences yield similar geochemicaloceanic plateau signatures (Kerr et al. 2002).Stratigraphic differences in their sedimentary covermay be accounted for by their different locations onthe single plateau. The present investigation alsoconstrains in both blocks a major tectonic rotationtaking place between 70-75 Ma. Alps.

The blocks in all probability have their commonorigin in the Caribbean large igneous province (alsoreferred to as the Caribbean Colombian OceanicPlateau, CCOP), which extruded from theGalapagos Hotspot (Spikings et al. 2001). Wesuggest that all oceanic basement blocks present inthe Ecuadorian forearc originally occupied marginalparts of the CCOP that fragmented during collisionwith the South American margin during the LateCretaceous. The first, Late Creaceous clockwiserotation in both blocks occurred during thecontemporaneous accretion of the Pallatanga Blockagainst the South American margin at 75-70 Ma(Vallejo et al. 2005).

Cretaceous block rotations in the coastal forearc of Ecuador:paleomagnetic, chronstratigraphic evidences, and implications for

the origin and accretion of the blocks

Luzieux, L.D.A., *Heller, F., **Spikings, R., Winkler, W. & Vallejo, C.F.

Geologisches Institut, ETHZ, Zürich, Switzerland

* Institut für Geophysik, ETHZ, Zürich, Switzerland

** Institut de Minéralogie, University of Geneva 4, Geneva, Switzerland

ACKNOWLEDGEMENTS

This work was supported by Swiss ScienceFoundation grants no. 2-77193-02 and 2-77504-04

REFERENCESBenitez, S. (1995). Evolution géodynamique de la côte

équatorienne pendant le Tertiaire. Geol. Alpine, 71, 3-163.Jaillard, E., Ordenez, M., Benitez, S., Berrones, G., Jimenez,

N., Montenegro, G., and Zambrano, I. (1995). Basindevelopment in an accretionary, oceanic-floored fore-arcsetting: Southern coastal Ecuador during Late Cretaceous -late Eocene time. Journal of South American EarthSciences, 9(1-2), 131-140.

Kerr, A., Aspden, J., Tarney, J., Pilatasig, L. (2002). The natureand provenance of accreted oceanic terranes in westernEcuador; geochemical and tectonic constraints. Journal ofthe Geological Society London 159, 577-594.

Luzieux, L., Heller, F., Winkler, W., Vallejo, C., Spikings, R.(2005). Cretaceous rotations of the coastal blocks ofEcuador 1°N-3°S: paleomagnetic evidence andimplications for the origin and accretion of the blocks.Extended Abstracts, 6th ISAG Meeting, Barcelona.

Roperch, P., Megard, ,F., Laj, C., Mourier, T., Noblet, C.(1987). Rotated oceanic blocs in western Ecuador.Geophysical Research Letters, 14, 558-561.

Spikings, R.A. Winkler, W. Seward, D. Handler, R. (2001).Along-strike variations in the thermal and tectonic responseof the continental Ecuadorian Andes to the collision withheterogeneous oceanic crust. Earth and Planetary ScienceLetters 186, 57-73.

Vallejo, C., Spikings, R., Winkler, W., Hochuli, P., Luzieux, L.(2005). Geochronology and provenance analysis ofbasement and clastic cover sequences within the northernWestern Cordillera, Ecuador. Extended Abstracts, 6th ISAGMeeting, Barcelona.

Figure 1. Position of the blocks in the Ecuatorian forearc in relation with the actual subduction zone. Only the PB and SLB are dis-cussed in the present work. Faults - CCF: Chongòn-Colonche Fault, PCF: Puerto Cayo Fault, CF: Canande Fault, ChTSZ: Chimbo-Toachi Shear Zone, CPF: Calacali-Pallatanga Fault Zone, MSZ: Mulaute Shear Zone. Blocks (Costa) – SEB: Santa Elena Block, PB:Piñon Block, SLB: San Lorenzo Block. PED: Pedernales-Esmeraldas Block. Units (Cordillera Ocidental) – PaB: Pallatanga Block,MB: Macuchi Block, PiU: Pilaton Unit, NB: Naranjal Block, DB: Rio Desgracia Block.

The Eastern Bolivian lowlands are situated at thetransition between the tropical-humid and the morearid subtropical climate regime. They should there-fore be particularly suitable for the reconstruction ofpast climate changes. At least throughout the Qua-ternary, weathering products from the Andean Pre-cordillera have been deposited in the Eastern Boliv-ian lowlands. Transport, accumulation, as well assubsequent erosion of these sediments, include flu-vial, aeolian and denudative processes, which areprimarily controlled by climate and tectonics. Ac-cordingly, past climate changes can be expected tohave influenced type and intensity of the prevailinggeomorphic process, and they should be recorded inthe stratigraphy and sedimentology of the foreland.

Here, we present preliminary results from usingpaleosol-sediment-sequences as archives for the re-construction of landscape evolution and paleocli-matic conditions in Eastern Bolivia. Absolute chro-nologies are being established by 14C and OSLdating, while analysis of remote sensing data en-ables the extension of local interpretations to a moreregional scale.

The piedmont in Eastern Bolivia is bordered bythe Andean Precordillera and the alluvial plain of theRío Grande. A series of small alluvial fans builds upthe piedmont. Several sequences along the pied-mont, which have been exposed due to the lateralerosion of the Río Grande, have been investigated(see Fig. 1 for an example outcrop at Cabezas). Al-ternating periods between geomorphological activity(gravels, sands, erosion channels) and stability (soilformation) have been identified .

Detailed sedimentological and stratigraphicalstudies along the outcrops enable us to define char-

acteristic stratigraphical units. These form the basisfor regional correlation and interpretation.

In general, coarse fluvial gravels and sands formthe basis of the investigated profiles. They wereprobably deposited during the Last Glacial Maximum(LGM) or Late Glacial (LG) and grade into fluvialsand and silts. A well-developed paleosol has devel-oped at the end of the Pleistocene/ Early Holoceneat ~10 ka BP (all ages are uncalibrated 14C years). Amajor erosional event then seems to have caused apronounced hiatus before rapid (fluvio-aeolian)sedimentation and formation of a new, flat accumu-lation surface occurred between ~8 and 4 ka BP.Since ~4 ka BP, soil formation indicates surface sta-bility. Base level changes resulting from large-scaleriver shifts of the Río Grande may have played a rolein ending the depositional history of the piedmont atthat time.

Based on the concept of landscape activity andstability, several conclusions regarding landscapeevolution in Eastern Bolivia can be drawn: Forestcover may have been absent or significantly reducedduring or before the LGM/ LG, thus allowing for themobilization and sedimentation of the observedcoarse fluvial deposits at the basis of the investi-gated sequences. This interpretation would be cor-roborated by findings from lake sediments in Bolivia,indicating dry conditions (Burbridge et al. 2004,Mayle et al. 2000). However, the limited number ofage dates does not allow the precise temporal cor-relation of these sediments. The Pleistocene/ Holo-cene transition in Eastern Bolivia seems to be char-acterized by relatively stable, wet conditions,favouring soil formation and probably extensive for-est cover. In contrast, the Early Holocene hiatus andthe subsequent rapid accumulation point to extreme

Paleosol-sediment-sequences along the Andean piedmont(Eastern Bolivia) and their implications

for Late Quaternary landscape evolution

May, J.-H. & Veit, H.

Geographical Institute, University of Bern, Hallerstr. 12, 3012 Bern, Switzerland

geomorphological instability. This might be due toreduced forest cover under prevailing dry conditions.This interpretation may be corroborated by appar-ently reduced accumulation rates in Bolivian lakes(Burbridge et al. 2004, Mayle et al. 2000). Dry con-ditions have also been proposed by Servant et al.(1981) for the Mid-Holocene, based on widespreadaccumulation of fluvial sands and subsequent for-mation of dune fields. Even though modern aeolianactivity can locally be observed in the study area, theLate Holocene transition to more humid conditionsseems to be well documented: Recent soil formationat the top of our sequences are consistent with pol-len findings from Bolivian lakes in the lowland (Bur-bridge et al. 2004, Mayle et al. 2000) and from lakesediments on the Altiplano (Baker et al., 2001b).

REFERENCES

Baker, P. A., Rigsby, C. A., Catherine, A., Seltzer, G. O., Fritz,S. C., Lowenstein, T. K., Bacher, N. P. & Veliz, C. 2001a:Tropical climate changes at millennial and orbital times-cales on the Bolivian Altiplano. Nature 409, 698 - 701.

Baker, P.A., Rigsby, C.A., Seltzer, G.O., Fritz, S.C., Lowen-stein, T.K., Bacher, N.P., Veliz, C., 2001b. The history ofSouth American tropical precipitation for the past 25,000years. Science 291: 640-643.

Burbridge, R. E., Mayle, F. E. & Killeen, T. J. 2004: Fifty-thousand-year vegetation and climate history of NoelKempff Mercado National Park, Bolivian Amazon. Quater-nary Research 61(2), 215-230.

Mayle, F. E., R. Burbridge & Killeen, T. J. 2000. Millennial-Scale Dynamics of Southern Amazonian Rain Forests.Science 290 (5500), 2291-2294.

Servant, M., Fontes, J.-C., Rieu, M. & Saliege, J.-F. 1981 :Phases climatiques arides holocènes dans le sud-ouest del'Amazonie (Bolivie). C.R. Acad. Sc. Paris, Serie II 292,1295-1297.

Figure 1. The picture shows the paleosol-sediment-sequence at Cabezas, E-Bolivia. Several stratigraphical units can be distin-guished by sedimentological characteristics and have been 14C-dated (uncalibrated ages are shown here). Source of figure: Jan-Hendrik May

To this date, the question of why and how a pla-teau-type orogen formed with crustal thickening atthe leading edge of western South America remainsone of the hotly debated issues in geodynamics.During the Cenozoic, the Altiplano and Puna pla-teaux of the Central Andes (average elevation some4 km, with an extent of 400 x 2000 km) developedduring continuous subduction of the oceanic Nazcaplate in a convergent continental margin setting – asituation that is unique along the 60,000 km of con-vergent margins around the globe. The key chal-lenge is to understand why a first-order mechanicalinstability of the later plateau extent developed alongthe central portion of the leading edge of SouthAmerica only, as well as why and how this featuredeveloped only during the Cenozoic, although thecycle of Andean subduction had been ongoing sinceat least the Jurassic. Moreover, it would appear thatonly rarely has this style of orogeny occurred inEarth history, another example probably being theCretaceous North American Laramides. Since the1980s, a plethora of models has been published thatattempt to find a solution to this ‘geodynamic para-dox’ (Allmendinger et al., 1997). A major integratedresearch program hosted by Berlin and PotsdamEarth science research units has attempted to un-ravel the mechanisms underlying plateau formationat a convergent continental margin.

Deep geophysical data across the Central Andesbetween 20°S and 24°S (ANCORP'96 and associ-ated geophysical studies; ANCORP working group2003) indicate the widespread presence of partialmelts or metamorphic fluids at mid-crustal level un-der the plateau between the Cordilleras boundingthe latter. From structural balancing studies, thesefluids or melts are associated with decoupling of up-per crustal shortening and lower crustal thickening.Based on similar indications from the distribution ofmagmatism it has been argued commonly that in fact

upper plate weakening from widespread heating andpartial melting may have been the key to under-standing its widespread shortening behind the vol-canic arc (Isacks, 1988; Allmendinger et al., 1997).In addition, changes in plate convergence are usu-ally considered to have been responsible in tuningthe changes in the upper plate system. While theavailable wealth of geophysical data would seem tolend support to the role of melts and fluids in upperplate orogeny, the sensitivity of elastic, thermal, andconductivity parameters as registered by most geo-physical imaging techniques, may overemphasizethis role.

We therefore analyzed the temporal and spatialevolution of deformation to provide better constraintsfor the identification of key mechanisms (see Elger etal., 2005, and Oncken et al., 2005 for details). An-other feature unique to the Central Andes, is thecomplete preservation of syntectonic volcanics andsediments throughout the orogen and at its marginsallowing high spatial and temporal reconstruction ofthe deformation history. Isotopic age-dating on thesedeposits as well as seismic-sequence analysis dem-onstrate that the Southern Altiplano crust was de-formed with a complex partitioning of deformationbetween various subunits that were partly synchro-nized. The general acceleration of shortening rateshows only a weak link to plate convergence ratesfor the early stages. In contrast, our results showthat the differential velocity between upper plate ve-locity and oceanic plate hinge and slab rollback ve-locity is crucial in determining amount and rate ofshortening as well as their lateral variability at theleading edge of the upper plate (see suggestions byRusso and Silver 1996, Heuret and Lallemand,2005). This first order control is tuned by factors af-fecting the strength balance between the upper platelithosphere and the plate interface of the Nazca andSouth American plates. These factors particularly in-

What controls long-term orogeny at a convergent continentalmargin?

The Andean case

Oncken, O. and SFB 267 research group

GeoForschungsZentrum Potsdam, Telegrafenberg, 14473 Potsdam, Germany

clude a stage of reduced slab dip accelerating short-ening (33 and 20 Ma) and an earlier phase of highertrenchward sediment flux reducing plate interfacecoupling with slowed shortening and enhanced slabrollback (45 and 33 Ma). The latter was not con-trolled by the Cenozoic climatic evolution (cf. Lamband Davis, 2003) but rather by construction and ero-sion of local relief during early stages of shortening.As a consequence of the interaction of these pa-rameters at the site of the southern hemisphereglobal arid belt, a plateau accumulating significantshortening developed in the backarc of the SouthAmerican volcanic arc – as opposed to most othercircum-Pacific margins that focus deformation in theforearc. The combination of these parameters (inparticular differential trench-upper plate velocityevolution, high plate interface coupling from lowtrench infill, and the lateral distribution of weak zonesin the upper plate leading edge) was highly uncom-mon during the Phanerozoic leading to very fewplateau style orogens at convergent margins like theCenozoic Central Andes in South America or theLaramide North American Cordillera.

REFERENCES

Allmendinger, R.W., Jordan, T.E., Kay, S.M. and Isacks, B.L.,1997. The evolution of the Altiplano-Puna Plateau of theCentral Andes. Annual Reviews Earth Planet Sciences, 25:139-174.

ANCORP-Working Group, 2003. Seismic imaging of an activecontinental margin - the Central Andes (ANCORP '96).Journal of Geophysical Research, in press.

Elger, K., Oncken, O. and Glodny, J., 2005. Plateau-style ac-cumulation of deformation - the Southern Altiplano. Tec-tonics, in press.

Heuret, A., Lallemand, S., 2005. Plate motions, slab dynamicsand back-arc deformation. Physics of the Earth and Plane-tary Interiors, 149, 31-51.

Isacks, B.L., 1988. Uplift of the Central Andean Plateau andbending of the Bolivian Orocline. Journal of GeophysicalResearch, 93(B4): 3211-3231.

Lamb, S. and Davis, P., 2003. Cenozoic climate change as apossible cause for the rise of the Andes. Nature (London),425(6960): 792-797.

Oncken, O., D. Hindle, J. Kley, K. Elger, P. Victor, K.Schemmann, 2005. Deformation of the Andean upper platesystem – facts, fiction, and constraints for modellers. In:Deformation of the Andes from top to bottom (eds:O.Oncken, G. Franz, H.J. Götze, M. Strecker, P. Wigger, P.Giese, V. Ramos, G. Chong), Frontiers in Earth Sciences,Springer, (submitted)

Russo, R.M. and Silver, P.G., 1996. Cordillera formation, man-tle dynamics, and the Wilson cycle. Geology, 24(6): 511-514.

Theoretical models for the development of moun-tainous topographies have animated discussionsabout the magnitudes and rates of growth and/or de-cay of relief in response to changes in climate, rockuplift and exposure of bedrock lithologies. The west-ern margin of southern Peru and northern Chile rep-resent such a mountain belt for which one would an-ticipate substantial relief growth. This is the casebecause this part of the Andes has been an activeplate-boundary zone as documented by crustalshallow seismic activities along the forearc and vol-canic eruptions since the Oligocene. In addition, theoccurrence of >1000 m deep valleys (e.g., Lluta andAzapa valleys) that dissect pediplains several thou-sands of km2 wide suggest that this part of the An-des has also been the location of substantial topog-raphic modifications.

This paper, however, suggests that in the Andesof northern Chile, the location of substantial reliefgrowth has been limited to ca. 20 km-long segmentsalong the major rivers since the Late Pliocene at thelatest. These segments are knickzones in the longi-tudinal stream profiles, separating the stream pro-files in two graded sections. In each of these sec-tions, the valley and channel morphologies revealdistinct trends in the upstream direction that corre-late with the channel gradient and the distance fromthe knickzones. In the flat segments beneath theknickzones, the valleys are occupied by a 250 and500 m-wide river belt, and a >400 m-wide floodplain.Towards the knickzones, the valleys narrow to sev-eral tens of meters, and the valley floor becomes oc-cupied by single mixed bedrock-alluvial channels. Inthe area of the knickzones, gravel bars are to largeextent absent, and the bedrock is exposed on thechannel floor. Above the knickzones, the valleys areseveral tens to hundreds of meters wide, and theriver belt occupies the entire valley floor and is madeup braided channels and gravel bars.

In the lower reaches where the gradient is flat, thepresence of braided alluvial channels indicatessediment bypass. The narrowing of the valleys to-wards the knickzones, the increase in the channelgradients and the mixed bedrock-alluvial nature ofchannels imply that these valley segments are in aphase of valley lowering. In the area of theknickzones, exposure of bedrock implies ongoingdissection into bedrock. Above the knickzones, how-ever, the presence of braided channels with longitu-dinal gravel bars implies sediment bypass. Hence,the narrowing of the valleys towards the knickzones,the increase in the channel gradients and the asso-ciated change in the channel morphologies from al-luvial to bedrock channels imply that these valleysare in a phase of headward erosion. Von Rotz et al.(2005) suggested that headward erosion was initi-ated in response to baselevel lowering in the coastalarea that occurred somewhen after 7.5 Ma. Theknickzones then represent the erosional fronts thathave shifted headward since that time. They sepa-rate the lower reaches that have adjusted to the low-ered baselevel from the headwaters that still recordthe Late Miocene situation.

The pattern of fluvial incision rates is based onelevation differences between the valley floor andthe 2.7 Ma-old Lauca ignimbrite that was depositedat the bottom of the Lluta valley at various locations(Wörner et al., 2002; García & Hérail., 2005). In theflat segment, the Lauca ignimbrite rests at the bot-tom of the Lluta valley at ca. 15 km distance from thePacific coast, and at 50 m above the valley floor 40km farther inland. This implies no significant down-cutting since the Pliocene. Similarly, concentrationsof multiple terrestrial cosmogenic nuclides (TCNs,10Be, 26Al, 21Ne; Kober et al., 2005) imply extremelylow erosion rates of <1 m/My for the pediplains adja-cent to these valleys due to the hyperarid climate. Inthe area of the knickzoned, the occurrence of the

Scale of relief growth in the Andes of northern Chile

Schlunegger, F., *Zeilinger, G. & *Kober, F.

Institute of Geological Sciences, University of Bern, Switzerland

*Department of Earth Sciences, ETH Zurich, Switzerland

Lauca ignimbrite between 400 and 700 m above therecent valley floor implies incision rates >250 m/Mybetween the Pliocene and the presence. On the ad-jacent pediplains, erosion rates have occurred at theorder of ca. 1 m/My according to concentrations ofTCNs (Kober et al., 2005). In the headwaters abovethe knickzones, the post 2.7 Ma depth of incision is<150 m, resulting in incision rates between 50-60m/My since that time. Similar magnitudes of erosionrates are measured with TCN analyses for the hill-slope interfluves (Kober et al., 2005).

The geomorphic data imply that the 20 km-wideknickzones in the longitudinal stream profiles ofnorthern Chilenean rivers represent headward shift-ing erosional fronts. There, rates of valley downcut-ting exceed erosion rates on adjacent pediplains bytwo orders of magnitudes. Hence, these segmentsare locations of substantial relief growth. Headwardincision was initiated in response to baselevel low-ering in the Late Miocene. Since that time, no sub-stantial changes in baselevels are recorded in thetopography. Indeed, beneath the knickzones, thevalley elevations have been stable since 2.7 Ma atthe latest as indicated by sediment bypass in thesesegments and the occurrence of the 2.7 Ma-oldLauca ignimbrite at the coast and in the Lluta valley.Also, there is no evidence of substantial erosion onthe adjacent pediplains. Similarly, the channel mor-phologies suggest no substantial fluvial dissectionand relief growth above the knickzones. Additionally,

identical erosion rates in valleys and on pediplainsand hillslope interfluves imply a nearly steady localrelief since the Late Pliocene at the latest. Hence,the only geomorphic recorders of relief growth arethe knickzones, and therefore, the scale of reliefgrowth has been limited to 20 km-long segmentssince the Late Pliocene. This also indicates that inthe Andes of Northern Chile, crustal uplift of theWestern Escarpment, and possibly of the Altiplano,has ceased or has probably been negligible for theobserved timescale.

REFERENCES

García, M. & Hérail, G. (2005): Fault-related folding, drainagenetwork evolution and valley incision during the Neogene inthe Andean Precordillera of Northern Chile. Geomorphol-ogy 65, 279-300

Kober, S. F., Ivy-Ochs, S., Schlunegger, F., Wieler, R., Baur,H. & Kubik, P.W. (2005): Denudation rates and topography-driven precipitation threshold in northern Chile: multiplecosmogenic nuclide data and sediment yield budgets.Gemorphology: submitted.

Von Rotz, R., Schlunegger, F. & Heller, F. (2005): Assessingthe age of relief growth in the Andes of northern Chile:Magneto-polarity chronologies from Neogene continentalsections. Terra Nova, in press.

Wörner, G., Uhlig, D., Kohler, I. & Seyfried, H. (2002): Evolutionof the West Andean Escarpment at 18°S (N. Chile) duringthe last 25 Ma: uplift, erosion and collapse through time.Tectonophysics 345, 183-198.

The Western Cordillera of Ecuador consists of al-lochthonous terranes, which accreted to the SouthAmerican margin during the Late Cretaceous andEarly Tertiary. Each allochthonous terrane is com-posed of a mafic oceanic basement, overlain by pre-,syn- and post-accretionary Late Cretaceous andTertiary sedimentary rocks. Transcurrent displace-ments along major N-S trending faults, resulting invariable clockwise rotations, have produced a com-plex juxtaposition of tectono-stratigraphic units. Wepresent both, provenance analyses of the sediments,which overlie the mafic basement, as well as U/Pbzircon (SHRIMP) and 40Ar/39Ar datings of basementrocks between 0º 20’ N to 2º S. Based on this datawe propose a tentative geological history of parts ofthe Western Cordillera.

The Pallatanga Terrane is exposed along theeastern border of the Cordillera (Fig. 1) and is sepa-rated from the continental margin by a deformedsuture zone (the Pujilí fault zone). Mafic basementrocks (Pallatanga Unit) yield oceanic plateau geo-chemical signatures and probably form part of theCaribbean Oceanic Plateau, implying a Late Creta-ceous age (Spikings et al., 2001; Kerr et al., 2002).The Macuchi Terrane is located along the westernborder of the cordillera, and is separated from thePallatanga Terrane by the regional-scale Chimbo-Toachi dextral shear zone. Geochemically, the ba-saltic-andesitic basement of the Macuchi Terrane(Macuchi Unit) has an island arc affinity (Hughes andPilatasig, 2002). 40Ar/39Ar analyses of plagioclasecrystals yield a plateau age of 42.62 ± 1.3 Ma, whichis consistent with the Middle Eocene age determinedin the sedimentary rocks associated with this unit(Eguez, 1986).

The San Juan Unit hosts ultramafic and gabbroicmafic cumulates, which yield an oceanic plateaugeochemical signature, similar to the Pallatanga Unit(Mamberti et al., 2004). U/Pb (SHRIMP) analyses of

homogeneous zircons extracted from a gabbro in theSan Juan Unit yielded an age of 87.1 ± 0.8. Ma. TheU/Pb age is interpreted as the crystalllisation age ofthe gabbros. Zircons from a fractionated juvenilepegmatite, which forms part of the Pallatanga Unit,yielded a U/Pb (SHRIMP) zircon age of 85.5±0.6,which we interpret to be the time of crystallisation oflate, fractionated melts that gave rise to the Palla-tanga Unit. The San Juan and Pallatanga units havevery similar ages, suggesting that they probably rep-resent a single plateau sequence that hosts both ul-tramafic cumulates and younger fractionated se-quences, similar to what is observed in theCaribbean Colombian Oceanic Plateau sequence inColombia (Kerr et al., 2002).

Associated with the Pallatanga Terrane is the RioCala volcanic arc system, which comprises variousvolcanic and volcaniclastic units (Natividad, Mulaute,Pilatón units). 40Ar/39Ar analyses of plagioclasecrystals extracted from basalts of the Natividad Unityield a plateau age of 64.3±0.4 Ma, interpreted asthe time of crystallization subsequent to eruption.

Heavy mineral (HM) assemblages of volcanic richsandstones overlying the Pallatanga Unit (Mulauteand Pilatón units), are characterized by high per-centages (90 - 95%) of volcanic derived minerals,suggesting that they were derived from a volcanicsource, distant from the continent. However, a sig-nificant quantity of material derived from continentalcrust occurs in the Saguangal Unit. The latter strati-graphically overlies the Mulaute and Pilaton units.Single grain geochemistry of clinopyroxenes (CPX)extracted from these units indicate that their maficsources were tholeiitic and formed in a subductionzone, suggesting an oceanic island arc origin. Thejuxtaposed Campanian - Maastrichtian NatividadUnit is dominated by volcanic derived minerals. Thecomposition of CPX extracted from basalts and

Geochronology and provenance analyses ofallochthonous terranes of the Western Cordillera, Ecuador

Vallejo, C., Winkler, W., *Spikings, R., **Hochuli, P. A., & Luzieux, L.

Geological Institute, Department of Earth Sciences, ETH Zurich, Switzerland

* Department of Mineralogy, Section des Sciences de la Terre, University of Geneva, Switzerland

** Paläontologisches Institut, University of Zurich, Switzerland

sandstones in this unit indicates a tholeiitic island arcsetting.

In conclusion, similar U/Pb crystallization agesfrom the San Juan and Pallatanga units suggest thatthey represent ultramafic and fractionated compo-nents of a single oceanic plateau sequence, whichcrystallized during 88 – 85 Ma. These ages partlyoverlap with those determined for the Caribbeanplateau (e.g. 91 - 88 Ma; Sinton et al. 1998). Heavyminerals, single grain geochemistry and sedimentaryfacies indicate that the associated Rio Cala islandarc units were derived from an island arc volcanicsource, which was separated from continental detri-tal input until deposition of the Saguangal Unit.40Ar/39Ar dating indicates that the island arc was ac-tive during the Maastrichtian-Danian. However, mi-crofossil data suggest it may be as old as the Cam-panian (Boland et al. 2000). Furtheron, U/Pb zircondating (LA-ICPMS) in sedimentary rocks of the Sa-guangal Unit yield a maximum age of 56 Ma for thedetrital zircons, which suggests that the Rio Cala is-land arc system was already accreted to the conti-nent during the Late Paleocene.

The Macuchi Terrane is interpreted as a volcanicarc produced by subduction of oceanic crust belowthe already accreted Pallatanga oceanic plateauelements. The activity of this arc probably started inthe Paleocene and continued till the Late Eocene.

ACKNOWLEDGMENTS

This project was supported by Swiss Science Foun-dation Grants no. 2-77193-02 and 2-77504-04.

Figure 1. Regional geology of Ecuador, including the allochtho-nous terranes of the Western Cordillera and coastal region.

REFERENCES

Boland, M., Pilatasig, L., Ibandango, C., McCourt, W., Aspden,J., Hughes, R., & Beate, B. (2000). CODIGEM-BGS, Quito,Informe 10.

Hughes, R & Pilatasig, L. (2002). Cretaceous and Tertary ter-rane accretion in the Cordillera Occidental of the Andes ofEcuador. Tectonophysics 345, 29-48.

Eguez, A. (1986). Evolution Cenozoique de la Cordillere Occi-dentale septentrionale d’Equateur (0 15’ S- 01 10’ S), lesmineralisations associetees. Doc. Thesis, UPMC, Paris,116 pp. (unpublished)

Kerr, A.C., Aspden, J.A., Tarney, J., Pilatasig, L..F. (2002). Thenature and provenance of acrreted oceanic terranes in west-ern Ecuador: geochemical and tectonic constraints. J. ofGeol. Soc. 159: 577-594.

Mamberti, M., Lapierre, H., Bosch, D., Jaillard, E., Hernandez,J. & Polve, M. (2004). The Early Cretaceous San JuanPlutonic Suite, Ecuador: a magma chamber in an oceanicplateau. Can. J. Earth Sci. 1: 1237-1258.

Spikings, R.A. Winkler, W. Seward, D. Handler, R. (2001).Along-strike variations in the thermal and tectonic responseof the continental Ecuadorian Andes to the collision withheterogeneous oceanic crust. Earth and Planetary ScienceLetters 5742: 1-17.

Located at 55°S in the island of Tierra del Fuego,Lago Fagnano occupies the deepest continental pull-apart basin forming part of a chain of tectonic de-pressions along the Fagnano-Magallanes left lateralrift system. This tectonic lineament crosses theSouthern Andes dividing the southern Nova Scotiaplate from the northern South American plate. Fag-nano is the biggest (~110 km long), southernmostnon-ice covered lake in the world close to Antarctica.More than 800 km of geophysical data were ac-quired in a recent seismic survey of Lago Fagnano(March, 2005) combining simultaneously a high-resolution pinger 3.5 kHz single-channel with 1 in3

airgun multi-channel systems. These data provide aunique opportunity to look at the most recent lacus-trine sediments with high-resolution, while imagingthe oldest sediments at the same time. Preliminaryinterpretations show that the lake is divided into twosub-basins: a deep eastern sub-basin (~200 m waterdepth), and a shallower western sub-basin (~100 m).The seismic survey penetrated more than severaltens of meters of sediments, exhibiting both lacus-trine and glacial provenance, probably comprisingthe LGM and the Holocene. A quick observation ofthe seismic reflectors indicates the presence of sev-eral neo-tectonic structures affecting even the mostrecent sedimentary package including some mega-turbidites. Thus, the geophysical survey providescritical information to evaluate the tectonic impact onthe sedimentation, and to choose the appropriatecoring sites to obtain a meaningful climatic recon-struction.

Given that Tierra del Fuego is secluded between twooceans, several environmental factors including thewesterly winds, the Southern Ocean circumpolarflow, and the South Pacific gyre directly influence itsregional climate and can be tracked in the sedimen-tary record of the lake. Today, the magnitude andintensity of the westerly winds dominate Patagonia’sclimate. They are the main transporters of moisturefrom the Pacific Ocean towards the continent trig-gering a seasonal precipitation pattern that shows aconspicuous latitudinal change. Continental recordsfrom southernmost Patagonia, thus, are ideally situ-ated to follow after the changing behaviour of thedifferent components of the climate system and, toinvestigate high-latitude climate variability throughtime.

Gravity cores from both Lago Fagnano sub-basinsreveal a regular alternation of light and dark laminaewith abundant diatom content. Ultra-high resolutionX-ray fluorescence micro-profiles show fluctuationsat mm scale in major and trace elements that mayindicate seasonal influx changes into the basin.These core data will provide a unique record ofchanges in regional climate with decadal resolution.The comparison of these results with other marineand continental archives will improve our under-standing of the forcing mechanisms behind climatechange that can be further used to validate the out-come of ocean and atmospheric climatic models forthe Southern Hemisphere. We are also using theinterpreted seismic profiles to identify piston-coringsites, for field operations anticipated in 2006.

Reconstructing past and present environmental processes in thesouthernmost tip of the Andean cordillera - The lacustrine record of

Lago Fagnano, Tierra del Fuego

1Waldmann, N., 1Ariztegui, D., 2Anselmetti, F. S., 3Austin Jr., J. A., 4Moy, C. M., & 4Dunbar, R.

1 Department of Geology and Paleontology, University of Geneva, 13 Rue des Maraichers, 1205 Geneva, Switzerland2 Geological Institute, Swiss Federal Institute of Technology ETHZ, Zürich, Sonneggstrasse 5, 8092 Zurich, Switzerland

3 Institute for Geophysics, University of Texas at Austin, TX 78759, Texas, USA4 Department of Geological and Environmental Sciences, Stanford University, CA 94305, California, USA

Western Peru including the Western Cordillera andthe coastal margin is made up mainly of the CoastalBatholith and the Precambrian to Palaeozoic Are-quipa Massif. The Coastal Batholith of Peru consistsof over one thousand different plutons ranging in agefrom Late Jurassic to Late Cretaceous and from Eo-cene to Early Miocene (Pitcher, 1985). It forms acomplex but well-defined linear feature which followsnearly the entire Peruvian coast, extending throughisolated plutons into Ecuador and Chile. The grani-toidic petrology of the individual intrusive bodiesshows only minor variations. The climatic zone runsparallel to the Andes and therefore almost the entirecoastal region is subject to the same very dry cli-mate. Hence many parameters that might have aneffect on erosion rates such as variations in climateand in rock type are tightly constrained and uniformwithin the whole area. This makes Western Peru anexquisite region to study the evolution of a coastalmargin in an area of active subduction. Additionallythe Nazca Ridge located on the oceanic Nazca Plateis being subducted in an essentially east-west direc-tion underneath the South American Plate. There isa clear trend towards a south east passage of thecollision zone over the last 10-12 Myrs (Hampel,2002; Pilger, 1981). The Peruvian coastal margintherefore offers unique possibilities to study the in-teraction of the different processes involved..

The tools best suited to constrain the evolution ofthe Peruvian coastal margin and the Western Cor-dillera as well as the processes induced by theNazca Ridge are low temperature geochronometerssuch as fission-track and (U-Th)/He. Samples col-lected at several locations in the Western Cordilleraand the Peruvian Coastal margin between Piura(Lat. 5°30S) and Tacna (Lat. 18°S) at altitudes fromsea level to 3500 m have been analyzed. Coolinghistories determined through modelling of the fission

track data show two distinctly different types ofcooling histories.

a.) Rocks from Precambrian or Palaeozoic suitesshow simple slow cooling. Significant changes in therate of cooling within these models can be correlatedwith known major tectonic events like the Peruvian-and the Incaic Phase. In the models from central andnorthern Peru only the Incaic event is observed. Thisimplies that the older Peruvian phase was weak orabsent in these areas. The models from the verysouth show a change in cooling which can be corre-lated with the Peruvian Phase. There is no indicationof an event, which can be associated with the IncaicPhase.

b.) The modelled cooling histories from theCoastal Batholith show a phase of very rapid coolingjust after intrusion which was as high as 60°C/Ma.This cooling slows down and eventually becomes in-distinguishable from the slow cooling observed in thePrecambrian and Palaeozoic rocks. In most modelsa significant slow down of the cooling rate is ob-served at various times. These changes are thoughtto represent the adjustment of the high geothermalgradient of the intrusives to the surrounding gradientof the “old” crust and are not believed to representtectonic events. In the majority of the models a rela-tively long period of quiescence with slow continuousexhumation follows. Depending on the variable ageof the intrusions and the rates of cooling, the dura-tion of the subsequent period of quiescence varies.

A renewed increase in cooling rates is observedbetween 8-1 Ma in various samples at altitudes be-tween 95 and 2895 m. In most cases these changesare outside the APAZ (Apatite Partial Anealing Zone)and therefore inconclusive. In samples from centralPeru, however these renewed phases of rapid ex-humation are supported by He data. It is plausible

The Peruvian coastal margin,evidence from low-temperature thermochronology

1Wipf, M., 1Seward, D., 2Stirling, C., 3Zeilinger, G. & 3Schlunegger, F.1

1 Geological Institute, ETH-Zentrum, 8092 Zürich, Switzerland2 Institute for Isotope Geology and Mineral Resources ETH-Zentrum, 8092 Zürich, Switzerland

3 Institute of Geological Sciences, University of Berne, 3012 Berne, Switzerland

that the event seen is real as the Cordillera Blancalocated east of these sites began uplifting at ~ 6 Ma.

In general the He ages of fast cooled samples likethe Coastal Batholith tend to reproduce nicely andsupport the modelling of cooling histories based onfission-track data. Apatite He ages from the slowlycooled Precambrian and Palaeozoic suites with po-tentially complex cooling histories and extensive pe-riods spent within the HePRZ produce a huge scatterof ages.

An impact caused by the subduction of the NazcaRidge is seen in the change of apparent apatite fis-sion track- and He-ages occurring over today’s posi-tion of the ridge. The younger ages in the north arethe result of enhanced erosion due to surface upliftcaused by the passing of the ridge. The very youngHe ages determined in the vicinity of the knickpointsin the Lunahuana and the Pisco valley are inter-preted to be the result of enhanced river incision.Data for a possible Late Neogene tectonic event re-corded by the modelled cooling histories of centralPeru is difficult to interpret. A correlation with thepassing of the Nazca Ridge is plausible but ex-tremely difficult to verify.

In conclusion it can be said that no significant ero-sion has been taking place in the entire coastal mar-gin. The modelled cooling histories are in goodagreement with notions that the area has been drysince at least 23 Ma possibly 37 Ma (Dunai et al.2005) or maybe even since the Jurassic as proposed

by (Hartley 2005). An increase in erosion has oc-curred in the coastal areas uplifted by the subductingNazca Ridge. This increase is relatively small how-ever and has lead to younger apparent apatite fis-sion-track- and He-ages. It is however not bigenough to reveal the timing of the ridge induced up-lift itself. At higher altitudes the enhanced erosionwas big enough to expose very young He ages.

REFERENCES

Dunai, T.J., Gonzalez Lopez, G.A., and Juez Larre, J. (2005):Oligocene-Miocene age of aridity in the Atacama Desertrevealed by exposure dating of erosion-sensitive landforms,Geological Society of America (GSA). Boulder CO UnitedStates. 2005., 321-324 p.

Hampel, A. (2002): The migration history of the Nazca Ridgealong the Peruvian active margin: a re-evaluation: Earthand Planetary Science Letters, v. 203, p. 665-679.

Hartley, A.J. (2005): What caused Andean uplift, 6th Interna-tional Symposium on Andean Geodynamics (ISAG 2005):Barcelona, p. 824-827.

Pilger, R.H. (1981): Plate Reconstructions, Aseismic Ridges,and Low-Angle Subduction beneath the Andes: GeologicalSociety of America Bulletin, v. 92, p. 448-456.

Pitcher, W.S. (1985): A multiple composite batholith, in Ather-ton, M.P., Cobbing, E.J., and Beckinsale, R.D., eds., Mag-matism at a Plate edge: The Peruvian Andes: Glasgow,Blackie, p. 93-101.

Figure 1. Geomorphological setting of the research location inthe Encierro Valley. The dashed lines indicate terminal mo-raines, the dotted lines pronounced lateral moraines. Num-bers behind sample labels are the exposure ages in ka.

Glaciers sensitively record climate changes, be-cause their mass balance depends on both tem-perature and precipitation. Moraines are thereforepotentially valuable archives for Quaternary climatereconstructions. The possibility to determine thedeposition age of moraines has, however, oftenbeen limited due to the lack of organic material forradiocarbon dating. This is especially true for aridhigh mountain areas, like e.g. the Central Andes.There, little is known about the glacial history, al-though the position between the tropical and ex-tratropical circulation systems (westerlies vs. tropicalinfluence) makes them a key area for studying Qua-ternary climate. Huge moraines north and south ofthe so-called ‘Arid Diagonal’, have been interpretedto document periods of substantially increased pre-cipitation (Amman et al. 2001), but insufficient datingcontrol prevented detailed interpretations.

Recent progress in developing surface exposuredating (SED) (Gosse & Phillips 2001) has provided anew tool, which now enables to address questionsconcerning- the age of glacial advances north and south of

the Arid Diagonal,- their value as archive for shifting atmospheric cir-

culation systems and concurrent precipitationchanges, and

- their potential correlation with well-known north-ern hemispheric cold periods during deglaciation(e.g. the Oldest, Older and Younger Dryas).

So far, we have obtained first exposure ages fromthe Encierro Valley (~29°S), Northern Chile, justsouth of the Arid Diagonal (Zech et al., submitted)(Fig. 1). A very prominent moraine stage (M-II)documents an extensive glaciation, which reachedan altitude of ~3500 m, resulting in an equilibriumline altitude (ELA) depression of ~1000 m. A largelateral moraine (LM in Fig. 1) corresponds to this ad-

Exposure dating in the Central Andes:

Quaternary glacier and climate reconstruction

Zech, R., *Kull, Ch. & Veit, H.

Institute of Geography, University of Bern, Switzerland

* PAGES IPO, Sulgeneckstrasse 38, 3007 Bern, Switzerland

Figure 1. Geomorphological setting of the research location inthe Encierro Valley. The dashed lines indicate terminal mo-raines, the dotted lines pronounced lateral moraines. Num-bers behind sample labels are the exposure ages in ka.

vance and has also been sampled. Several reces-sional moraines (M-III, IV and V) can be found in thevalley bottom and probably document climate fluc-tuations during deglaciation.

The exposure ages from M-II B and LM indicatethat the prominent glacial stage occurred at ~14.0 ±1.4 ka BP. Too young exposure ages can be ex-plained with post-depositional exhumation and insta-bility of boulders, so that in the absence of obviousoutliers the oldest exposure age should generally beconsidered the best estimate for the deposition age(Zech et al., in press). The oldest sample (EE71 onM-II A) may either be an outlier due to inheritance(pre-exposure) or, more likely, document a previous,more extensive glaciation than M-II B.

Deglaciation of the nowadays ice-free valley oc-curred after a final glacial advance at ~11.6 ± 1.2 kaBP.

Previous modeling has shown that a temperaturedepression of ~5.5°C and a precipitation increase of~550 mm/a (presently: ~300 mm/a) has been neces-sary to create the most prominent moraine stage M-II B (Kull et al. 2002). Today, the study area is mainlyinfluenced by the westerlies and winter precipitation.It has therefore been postulated that a glacial ad-vance could have occurred synchronous with thetemperature minimum during the last glacial maxi-mum (LGM) at ~18-20 ka BP. Not only low tem-peratures, but also a northward shift of the wester-lies, deduced from marine sediments off-shore Chileand pollen records, would have favored this sce-nario. However, the glacier-climate model could notdetermine any pronounced seasonality of the in-creased precipitation, and it can therefore not be ex-cluded that austral summer advection of humid airmasses from the Andean east-side might haveplayed an important role for the observed glacialstages at ~14.0 and 11.6 ka BP. There is increasingevidence for major climatic changes in tropical andsubtropical South American climate at these times(e.g. oxygen isotope records from stalagmites in SE-Brazil, from ice-cores on Vulcan Sajama and Illimani,from lake sediments on the Altiplano, etc.), but hith-erto both proxy interpretation and correlation are amatter of ongoing debate.

We conclude that- extensive glacial advances in the Encierro Valley

occurred during the Lateglacial,- these advances required substantially increased

precipitation, which we speculate might originatefrom the tropical circulation system, and

- that the timing of glacial advances resembles thecold event stratigraphy in the northern hemi-sphere (Oldest Dryas, Younger Dryas), but thatabsolute uncertainties (~10%) need to be re-duced for corroboration.

Further exposure age samples are currently beingprepared and results are anticipated for Octo-ber/November. These ongoing studies are focusing:- on the refinement of the glacial chronology in a

valley close to the presented Encierro Valley (atleast six glacial stages),

- on the spatial extent of the identified glacial ad-vances: Comparison with exposure age chro-nologies from Bolivia, as well as from south of theEncierro Valley are expected to yield further in-formation on past precipitation changes andsources.

REFERENCES

Ammann, C., Jenny, B., Kammer, K. & Messerli, B. (2001):Late Quaternary Glacier response to humidity changes inthe arid Andes of Chile (18-29°S). Palaeogeography, Pa-laeoclimatology, Palaeoecology 172: 313-326.

Gosse, J. C. & Phillips, F. M. (2001): Terrestrial in situ cos-mogenic nuclides: theory and application. Quaternary Sci-ence Reviews 20: 1475-1560.

Kull, C., Grosjean, M. & Veit, H. (2002): Modeling Modern andLate Pleistocene Glacio-Climatological Conditions in theNorth Chilean Andes (29–30 °). Climatic Change 52: 359-381.

Zech, R., Glaser, B., Sosin, P., Kubik, P. W. & Zech, W. (2005):Evidence for long-lasting landform surface instability onhummocky moraines in the Pamir Mountains from surfaceexposure dating. Earth and Planetary Science Letters: inpress.

Zech, R., Kull, C. & Veit, H. (2005): Late Quaternary GlacialHistory in the Encierro Valley, Northern Chile (29°S), de-duced from 10Be Surface Exposure Dating. Palaeogeogra-phy, Palaeoclimatology, Palaeoecology: submitted.

.