Complete Dissertation

VU Research Portal

Lines of Time Kaandorp, R.J.G.

2007

document version Publisher's PDF, also known as Version of record

Link to publication in VU Research Portal

citation for published version (APA) Kaandorp, R. J. G. (2007). Lines of Time: Seasonality, climate and environments of the Miocene Pebas Formation in western Amazonia derived from chemical records in molluscan growth-bands.

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

E-mail address: [email protected]

Download date: 29. Sep. 2021 VRIJE UNIVERSITEIT

Lines of Time

Seasonality, climate and environments of the Miocene Pebas Formation in western Amazonia derived from chemical records in molluscan growth-bands

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad Doctor aan de Vrije Universiteit Amsterdam, op gezag van de rector magniﬁcus prof.dr. L.M. Bouter, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de faculteit der Aard- en Levenswetenschappen op dinsdag 16 oktober 2007 om 15.45 uur in het auditorium van de universiteit, De Boelelaan 1105

door

Ronald Johannes Gerardus Kaandorp

geboren te Alkmaar promotoren: prof.dr. J.E. van Hinte prof.dr. D. Kroon copromotoren: dr. G.M. Ganssen dr. H.B. Vonhof voor Brad, Soraya, mijn familie ISBN 978-90-9022201-1

Netherlands Research School of Sedimentary Geology (NSG) Publication n ◦: 20071016

Lines of Time — Seasonality, climate and environments of the Miocene Pebas Formation in western Amazonia derived from chemical records in molluscan growth-bands. [Ph.D. thesis, Vrije Universiteit Amsterdam] In Dutch: Tijdslijnen — Seizonaliteit, klimaat en landschappen van de Miocene Pebas Formatie, westelijk Amazonia, uit chemische data in groeibanden van molluskenschelpen. [ac. proefschrift, Vrije Universiteit Amsterdam]

Cover photo’s: satellite picture from NASA World Wind, molluscs photographed by Saskia Kars

This project was funded by WOTRO, the Netherlands Foundation for the Advancement of Tropical Research, residing under the Netherlands Organization for Scientiﬁc Research (NWO). reading committee: prof.dr. G.J. Boekschoten prof.dr. H. Hooghiemstra dr. C. Hoorn prof.dr. M. Räsänen prof.dr. B.R. Schöne

Contents

Dankwoord xiii

Samenvatting xv

Resumen xxi

1 Introduction and summary 1 1.1 Aim of the present work ...... 1 1.2 The geography of Amazonia ...... 2 1.3 Amazonia’s seasonal variation of climate ...... 6 1.4 Neogene sedimentation history of the Amazon Basin ...... 8 1.5 Pebas Formation ...... 8 1.5.1 Stratigraphy ...... 8 1.5.2 Environments of Lago Pebas ...... 12 1.5.3 Marine inﬂuences ...... 13 1.6 Isotope records from accretionary growth structures ...... 17 1.6.1 Isotope records from molluscs ...... 17 1.6.2 MicroMill ...... 17 1.6.3 Mass spectrometry ...... 18 1.7 Amazonian climate ...... 20 1.7.1 Present-day climate ...... 20 1.7.2 Low latitudes ...... 20 1.7.3 Amazonian Miocene climate ...... 22 1.8 Main conclusions of this work ...... 23 1.8.1 The Miocene climate of Amazonia ...... 23 1.8.2 Ecological implications of molluscan stable geochemical records . . 24 1.8.3 Aquatic landscapes in the Miocene of western Amazonia ...... 25

2 Seasonal variation reﬂected in Amazonian freshwater bivalve A. trapesialis 27 2.1 Introduction ...... 28 2.2 Experiment, material and methods ...... 29 2.2.1 Monitoring project ...... 29

vii 2.3 Results ...... 31 2.3.1 Playa Cocha water ...... 31 2.3.2 Isotope composition of aragonite shells ...... 32 2.4 Discussion ...... 35 2.4.1 Seasonal stable isotope variation of water ...... 35 2.4.2 Equilibrium precipitation of shell aragonite ...... 35 2.4.3 A. trapesialis growth rates ...... 39 2.5 Conclusions ...... 39

3 Growth rates and stable isotopes in juvenile Anodontites trapesialis molluscs 41 3.1 Introduction ...... 41 3.2 Material and methods ...... 42 3.3 Results ...... 42 3.3.1 Bottom water temperature record ...... 42 3.3.2 Isotope records ...... 43 3.4 Discussion ...... 44 3.4.1 Juvenile growth rates ...... 44 3.4.2 Extrapolation of the growth in the dry season ...... 45 3.5 Conclusions ...... 46

4 Seasonal Amazonian rainfall variation in the Miocene Climate Optimum 47 4.1 Introduction ...... 47 4.2 Methods ...... 49 4.3 Discussion ...... 52 4.4 Conclusions ...... 52

5 Ecological implications of Miocene western Amazonian bivalves 55 5.1 Introduction ...... 56 5.1.1 Pebas Formation ...... 56 5.1.2 Molluscs ...... 57 5.2 Material and methods ...... 58 5.3 Results ...... 59 5.3.1 Quality of fossil material ...... 59 5.3.2 Stable-isotope composition of Pebas Formation bivalves ...... 59 5.4 Discussion ...... 66 5.4.1 Stable oxygen isotope characteristics of Amazonian continental waters ...... 70 5.4.2 Variation of carbon isotopes in host water ...... 71 5.4.3 Variation of carbon isotopes in bivalves ...... 71 5.4.4 Coupled carbon and oxygen isotope cyclicity ...... 73 5.4.5 Consistency of isotope signals ...... 74 5.4.6 Trace element variation ...... 74 5.5 Conclusions ...... 80 5.6 Appendix ...... 81

6 Aquatic landscapes in the Miocene of western Amazonia 85 6.1 Introduction ...... 85 6.2 Material and methods ...... 86 6.3 Description of the section and facies interpretation ...... 89 6.4 Molluscan composition and diversity ...... 92 6.5 Carbon and oxygen isotope geochemistry ...... 97 6.6 Growth band chemistry ...... 98 6.7 Discussion ...... 100 6.8 Conclusions...... 106

References 109

Appendices 120

Appendix AÐ Isotope data of modern molluscs 121

Appendix B Ð Miocene shells: isotope data and minor elements 125

List of Figures

1.1 Geographic setting of Amazonia ...... 4 1.2 Amazon River and Amazon Basin ...... 4 1.3 Conﬂuence Nanay and Amazon rivers ...... 5 1.4 Amazon environments ...... 5 1.5 Seasonal ITCZ movement ...... 6 1.6 Climate of Iquitos ...... 7 1.7 Stratigraphic ranges of Pebas Formation outcrops ...... 9 1.8 Drawings of typical Pebas molluscs ...... 9 1.9 Amazon Basin: interpreting the past ...... 10 1.10 Grainsize distribution Los Chorros ...... 11 1.11 Overview of sites ...... 11 1.12 Distribution of Pebas molluscan faunas ...... 12 1.13 Inferred distribution of assemblages ...... 13 1.14 Palaeoenvironmental impression of Lake Pebas ...... 14 1.15 Molluscan Sr isotope ranges ...... 16 1.16 MicroMill ...... 18 1.17 Milling thin lines with a thick drill ...... 19 1.18 Finnigan MAT252 Mass spectrometer ...... 19 1.19 Seasonal change Amazonian rainwater ...... 20 1.20 Location of the Playa Cocha and Mazán sites ...... 22

2.1 Location map of sampling site ...... 28 2.2 Marking and cutting of shells ...... 30 2.3 Playa Cocha water data ...... 33 2.4 Stable isotope records of A. trapesialis ...... 34 2.5 Predicted isotope values and growth rates ...... 37

3.1 Water temperatures and precipitation ...... 43 3.2 Isotope records vs growth, determining anchor points ...... 44 3.3 Growth ﬁt, using 3 anchor points...... 45 3.4 Temperature change reﬂected in mollusc oxygen isotopes ...... 46

xi 4.1 Seasonal ITCZ movement ...... 48 4.2 Isotopic change of rainwater ...... 49 4.3 Location map ...... 50 4.4 Oxygen isotope records ...... 51

5.1 Location map of sample sites...... 56 5.2 Photographs of shells ...... 58 5.3 Sampling of growth lines in two directions ...... 60 5.4 Fe record from growth records of living shell ...... 61 5.5 Isotope records of modern analogues ...... 62 5.6 Isotope profiles of analyzed Pebas Formation molluscs ...... 63 5.7 Isotope profiles of analyzed Pebas Formation molluscs ...... 64 5.8 Isotope profiles of analyzed Pebas Formation molluscs ...... 65 5.9 Isotope profiles of analyzed Pebas Formation molluscs ...... 66 5.10 Isotope profiles of analyzed Pebas Formation molluscs ...... 67 5.11 Isotope records against distance from umbo ...... 68 5.12 Overview of measured Pebasian molluscs ...... 69 5.13 Bjerrum plot of the concentrations of carbonate species at different pH . . . 72 5.14 Minor element records from growth increments ...... 75 5.15 Minor element records from growth increments ...... 76 5.16 Minor element records from growth increments ...... 77 5.17 Minor element records from growth increments ...... 78 5.18 Concentrations of minor elements dissolved in Amazon/Playa Cocha water 79 5.19 Stratigraphic columns of the outcrops ...... 82

6.1 Map of the surroundings of Santa Rosa ...... 88 6.2 Stratigraphic log of the Santa Rosa de Pichana section ...... 90 6.3 Examples of found ichnological characteristics ...... 91 6.4 Diversity indices in the section of Santa Rosa de Pichana ...... 93 6.5 Isotope ratios of the Pebasian shells and indications of typical isotope signatures of marine and Amazonian river shells...... 98 6.6 Stable isotope values of whole shells ...... 99 6.7 Stable isotope values along growth bands ...... 101 6.8 Stable isotope values along growth bands ...... 102 6.9 Effect of marine incursions ...... 106 6.10 Depositional and faunistic characteristics ...... 107 6.11 Interpretation of the Santa Rosa de Pichana section ...... 108 Dankwoord

In de nazomer van 1997 sprak ik mijn buurman (en latere broer) Hubert Vonhof op het ter- ras van Café Schaeper in de Tweede Hugo de Grootstraat. Hubert, met veel enthousiasme vertelde je dat je samen met Frank Wesselingh een voorstel voor een promotieproject aan het schrijven was dat je naar WOTRO ging sturen. De promovendus zou een half jaar in Zuid-Amerika veldwerk gaan doen met de bedoeling om klimaatssignalen uit mollusken te gaan halen. Je enthousiasme werkte aanstekelijk en ik gaf je te kennen dat ik het project, indien gehonoreerd, graag zou willen uitvoeren. WOTRO ging akkoord met het voorstel en uiteindelijk getuigt dit proefschrift dat ik het geluk heb gehad het project te mogen doen. Dank Frank Wesselingh, voor je hulp met de determinatie van recente en fossiele schelpen en het gebruik van enkele door jou verzamelde fossielen. Ik heb met zeer veel plezier met jullie samengewerkt en genoten van de discussies over "Pebas". Gerald Ganssen hartelijk dank voor je hulp bij het opzetten van het experiment in Playa Cocha en het veraangena- men van de tijd in Iquitos. Jan van Hinte, dank voor je hulp en logistiek in Suriname. Ik weet, Jan, dat je het zeer jammer vond dat de Surinaamse kant van het veldwerk niet geworden is wat het had moeten zijn. Dick Kroon, je bent de catalisator geweest toen de snelheid van afronding van het proefschrift terugliep en je hebt altijd met veel enthousiasme de voortgang in de gaten gehouden. Heren, dank aan jullie allen voor de tijd die jullie aan de totstandkoming van dit proefschrift hebben besteed. Thanks to the reading committee: Bert Boekschoten, Carina Hoorn, Henry Hooghiem- stra, Matti Räsänen and Bernd Schöne. Thanks to Gareth Davies, David Dettman, Franz Fürsich, Mathias Harzhauser, Carina Hoorn, Patrick De Deckker, Matti Räsänen and Werner Ricken for their (peer) reviews of earlier versions of the papers. Gracias a Lidia Romero por las logísticas en Perú. Andrés Marmól gracias por tu ayudo en el campo de Iquitos, por encontrar las conchas vivas en Río Itaya y Playa Cocha. Cahuide del Busto, gracias por todo tu trabajo. Sin tu ayudo en el campo por mas que un año, el experimento no hubiera sido posible. Grácias a José y Rusbel Arimuya. Sin ustedes tampoco era posible de hacer el trabajo en el campo. Ustedes sabian donde encontrar las aﬂoramien- tos del Mióceno, donde se fueron antes con Frank. Mi familia peruana Soraya, Rosa, Karla y Paulo, gracias por su amor y hospitalidad, Pepe y Rusbel, por su compañerismo en Iquitos. Frank Peeters, dank voor je hulp met de mathematische kanten van het proefschrift, voor de vruchtvolle ideeën waar je me op bracht en voor de mentale voorbereiding op de verdediging, alweer een aantal jaren geleden. Veel van de foto’s van schelpen in dit proef-

xiii schrift zijn gemaakt door Saskia Kars, dankjewel Sas. Wynanda Koot en Bouk Laçet dank jullie voor de preparatie van de dunne doorsneden van de schelpen. Soraya y Glenda, gracias su ayuda con el resumen. Ton en Riet voor het rode potlood... Pour "le sentier de traverse": Ilhem, Dominique, Céline, Matthieu. Dank aan de mensen van de NSG-tijd, eerstens mijn vriend Anco, voor je vertrouwen, Sierd, ook voor dat van jou, collega Paul, Anna en Alwien, Charlotte en Lieke, Dick, Henk, Marjolein, Simon en Gerald. Dank voor de tijd op de VU buiten de promotie om: Aafke, Alex, Anco, Ane, Anna, Auberge in den Koele Blonde, Aude, Bart (Hendriks), Bram, Bert, Bogusia, de Houtza- gerij, Dick (alweer), Emma, Frank (peef), Geert-Jan, GeoVUsie, Gerald, Glenda, Jan, Jens, Jochem, John, Jop, Jos, José, José, Karen, Katja, Liviu, Maarten, Margo, Margot, Mark, Martin, Orson, Paul, Patrizia, Ralph, Sandra, Sandra, Sandrine, Simon, Simon, Stefan, Suzanne, en Els. Frank, Kees, Bernd. Karin en Tibor, en iedereen die ik vergeten ben. Dank aan Thomas, Vincent en het internet voor de hulp met LATEX. Buiten de VU: Carla en Jorik, Lucky en Gert-Jan, Mayra en Hubert. Soraya, Maas en mijn ouders, dank voor jullie engelengeduld. De Famliën. En voor dat van jou Brad, vooral als jij nu eindelijk eens op de computer wilde spelen, terwijl ik zat te tikken. Samenvatting

Het vochtige regenwoud rondom Iquitos, een stad in het noordoosten van Peru wordt beschouwd als één van de meest soortenrijke bossen ter wereld (Kalliola and Puhakka, 1993). De laatste jaren is de Neogene ontwikkeling van het Amazone Bekken en zijn regenwoud een zeer belangrijke focus in biodiversiteitstudies en studies naar duurzaam landgebruik. De opmerkelijke soortenrijkdom van Amazonia lijkt te worden gecontroleerd door een com- plexe interactie van factoren zoals de veranderingen in brongebieden, rivierlopen, de dis- tributie van bodemsoorten, klimaatsverandering, tektoniek en menselijke activiteit. Wegen- bouw en steeds versnellende ontbossingen in het gebied zijn een grote bedreiging voor het voortbestaan van de bossen (Kalliola and Paitán Flores, 1998). Om de impact die menselijke activiteit op het ecosysteem van de laaglanden van Amazonia heeft beter te kunnen beoordelen moet men ook de natuurlijke variabiliteit begrijpen; de paleogeografische en pa- leoecologische evolutie van de regio door de geologische tijd. De uitstekend gepreserveerde fossiele mollusken die in het gebied worden gevonden zijn bestudeerd om een inzicht te kunnen geven in de regionale evolutie in het Mioceen, en wel de periode tussen 16 en 12 miljoen jaar geleden: tussen het late Vroeg Mioceen en het vroege Laat Mioceen. In het gebied zijn verschillende Nederlanders actief geweest. De groep van de profes- soren van der Hammen en Hooghiemstra van de Universiteit van Amsterdam heeft zeer veel studies naar Holocene klimaatsveranderingen gedaan in de Andes van Colombia. Ca- rina Hoorn, ook uit die groep, heeft veel pionierswerk gedaan op het gebied van de paleo- geografie en paleoecologie van het Amazone Bekken. De oevers van de Amazone Rivier en diens zijrivieren bevatten vele ontsluitingen (onbegroeide sedimenten en gesteenten). Be- halve recente sedimenten zijn er langs de rivieroevers ook gesteenten ontsloten van Miocene ouderdom. Met pollen uit die ontsluitingen en de sedimentologie heeft Hoorn haar studies aan Mioceen Amazonia verricht (Hoorn, 1994b). Tijdens één van haar veldwerken in het gebied is Frank Wesselingh mee geweest die aan een paleolontologische studie begon naar de schelpen die in deze sedimenten zijn gevonden. Dit resulteerde later in een promotieon- derzoek van Wesselingh (Scripta Geologica 133). Nadat Hubert Vonhof m.b.v. strontiu- misotopen kon vertellen waar het water vandaan kwam waarin de schelpen leefden (Vonhof et al., 1998) werd er een aanvraag gedaan om de paleoklimatologie van het gebied te bestud- eren, wat uiteindelijk heeft geresulteerd in dit proefschrift. Het Amazone Bekken lag in het Mioceen vrijwel op dezelfde geografische breedte als nu, in de tropen. Een vergelijking tussen het huidige en toenmalige klimaat is derhalve

xv op zijn plaats. De grafiek in figuur 1.6 op pagina 7 laat zien dat in Iquitos de gemiddelde maandtemperatuur door het jaar heen nauwelijks varieert, terwijl er grote veranderingen zijn in de gemiddelde maandelijkse neerslag. Als gevolg van dit gegeven is er voor gekozen om voor de studie van het Miocene klimaat in westelijk Amazonia te kijken naar de (seizoens- gebonden) neerslaghoeveelheid. Er is bekeken in hoeverre de samenstelling van het regenwater per seizoen veranderde. Nu is het onmogelijk om de samenstelling van regenwater die reeds gevallen is, te meten, zeker niet van neerslag zo’n 12 tot 16 miljoen jaar geleden. Daarvoor is er een omweg nodig. De chemische samenstelling van het oppervlaktewater (van de rivieren en meren) in het Amazone gebied verandert door het jaar heen. Dit komt door de veranderende neerslag en de verdamping. Metingen aan chemie van het oppervlaktewater zegt dus iets over de chemie van de neerslag die het oppervlaktewater voedt. De neerslaghoeveelheid en de isotopensamenstelling van neerslag blijken door het jaar heen grote veranderingen te ondergaan (dit is te zien in animaties op de web-site van het IAEA, het Internationale Atoom Energie Agentschap te Wenen). De gebieden rond de evenaar vallen onder het regiem van de Inter-Tropische Convergentie Zone, de ITCZ. De ITCZ is een band rond de evenaar waar door opwarming door de zon luchtmassa’s omhoog worden gebracht. De ITCZ verandert door het jaar heen van positie. Dit komt door de seizoensveranderingen. De ITCZ volgt de baan van de zgn. calorische evenaar. Op de oceanen ligt deze evenaar daar waar de zon op het middaguur recht boven staat, maar op de continenten, door de slechte geleiding van de aarde, kan deze calorische evenaar verder van de geografische evenaar afliggen (zie figuur 1.5 op pagina 6). Als gevolg van de stijging van luchtmassa’s op de calorische evenaar wordt over het aardoppervlak lucht van elders aange- zogen. Hierdoor ontstaan de zgn. passaatwinden. In het Amazonegebied heeft dit tot gevolg dat als de ITCZ ten noorden van de geografische evenaar ligt er een zuidoost passaat droge lucht aanvoert en er sprake is van een droog seizoen, terwijl er sprake is van een nat seizoen als de ITCZ ten zuiden van de geografische evenaar ligt en de noordoost passaat vochtige lucht vanaf de Atlantische Oceaan tot aan de oostelijke flanken van de Andes brengt. De (zuurstof)-isotopensamenstelling van neerslag is zeer verschillend. Die is namelijk onder andere afhankelijk van de isotopensamenstelling van het brongebied (verdampings- gebied), van de neerslag, de geografische breedte, de hoogte, de afstand van het neerslagge- bied en het brongebied, de temperatuur en de hoeveelheid neerslag die er valt. In het droge seizoen behoudt de waterdamp in de lucht boven Amazonia, en de neerslag die daaruit valt, ongeveer de isotopensamenstelling van haar brongebied, de waarde die de Atlantische Oceaan heeft. Alles wat in het droge seizoen op de route van de Atlantische Oceaan naar de Andes onderweg neerslaat, verdampt weer. De chemische samenstelling en de hoeveelheid waterdamp in de lucht blijven zo hetzelfde. In het natte seizoen is de neerslaghoeveelheid veel groter dan de verdamping. Doordat de waterdamp niet, of minder aangevuld wordt, vindt er fractionering van de isotopensamenstelling plaats. De zwaardere isotopen zullen eerder uitregenen en eenmaal op het aardoppervlak weer moeilijker verdampen. Hierdoor wordt de waterdamp boven het regenwoud naar het westen toe steeds lichter van samenstelling (zie figuur 1.19). De samenstelling van oppervlaktewater verandert door het jaar heen dus door de veranderende samenstelling van het regenwater. Omdat hoeveelheid en samenstelling van de neerslag aan elkaar gekoppeld zijn, is dit ook te zien in het oppervlaktewater. De verdamping van oppervlaktewater heeft echter een versterkende werking op de isotopensamenstelling van het water. In het natte seizoen is het water isotopisch gezien lichter van samenstelling door isotopisch lichtere neerslag, in het droge seizoen zorgt verdamping ervoor dat het oppervlakte water dat toch al isotopisch zwaar is door de samenstelling van het regenwater nog zwaarder wordt. Jaarlijkse variatie van de (zuurstof)isotopensamenstelling van het water wordt zodoende vrij groot, een effect dat voornamelijk groot is in meertjes in het overstromingsgebied van de Amazone. In hoofdstuk 2 komt dat tot uitdrukking. Een jaar lang zijn er watermonsters genomen van water uit Playa Cocha, een meertje in het overstromingsgebied van de Amazone, die later op isotopensamenstelling zijn gemeten. Duidelijk is te zien dat in het natte seizoen, als Playa Cocha deel uitmaakt van de hoofd- stroom van de Amazone Rivier, de zuurstof -(èn koolstof)isotopen veel lichter zijn dan in het droge seizoen, als Playa Cocha opdroogt (verdampt) tot een klein vijvertje, alleen nog gevoed door lokale regenbuien. Behalve dat er watermonsters genomen zijn, is er in Playa Cocha een jaar lang een experiment geweest met levende schelpen (zie hoofdstukken 2 en 3. De in Playa Cocha gevonden mollusken, Anodontites trapesialisgeheten, werden gemerkt en daarna in plastic boodschappenmandjes (gevuld met bodemsedimenten uit het meertje) geplaatst. De boodschappenmandjes werden met drijvers eraan bevestigd in het meer gezet, zodat ze teruggevonden konden worden in het natte seizoen. Het was de bedoeling om elke twee maanden één van de manden uit het water te lichten. Helaas is gebleken, dat toen het water van de Amazone Rivier Playa Cocha overspoelde de manden te zwaar waren om naar boven gehaald te kunnen worden. Nadat het water weer gezakt was, bleken alle gemerkte schelpen te zijn verdwenen. In de uitstroomopening van Playa Cocha bleken nog een tiental juveniele schelpen te zitten, die niet met het water mee de rivier afgedreven waren . Een volwassen ongemerkte schelp werd uit het water gehaald en bemonsterd, eve- nals een schelp van ongeveer een jaar oud (hoofdstuk 3). De afdeling gesteentebewerking van het Instituut der Aardwetenschappen heeft de schelpen in kunstmatige hars gegoten en er dunne doorsneden van gemaakt. Deze dunne doorsneden werden langs de as van snel- ste groei uitgezaagd. Vervolgens zijn de dunne doorsneden onder de microscoop van een zgn. Micromill gelegd (Figuur 1.16, pagina 18). Een MicroMill is een computergestuurde tandartsboor. De beelden die een camera, door de microscoop van de MicroMill, van de dunne doorsnede maakt worden op een computerscherm geprojecteerd. De doorsnede van de schelp die zodoende te zien is, vertoont allerlei groeilijnen. De met de muis uitgezette lijnen parallel aan deze groeilijnen geven het boorpad aan die de boor aflegt. Wanneer de boor langs de uitgezette lijn zijn pad aflegt, verpoedert deze de schelp. Het poeder, van cal- ciumcarbonaat, wordt vervolgens geanalyseerd met behulp van een massaspectrometer. De verkregen waarden van zuurstof- en koolstofisotopen uit de schelpen van Anodontites trapesialis zijn vergeleken met de waarden van het water in Playa Cocha. Er is gebleken dat de zuurstofisotopen in schelpen, volgens een uit de literatuur bekende formule, (Grossman and Ku, 1986), de zuurstofisotopenwaarden van het Playa Cocha water zeer goed weergeven. Ook is gebleken dat de koolstofisotopenwaarden van de schelpen overeen komen met die van het in het water opgeloste HCO3, bicarbonaat, wat door de schelpen gebruikt wordt om aragoniet te maken. Er is echter wel een verschil met de uit de literatuur (Romanek et al., 1992) bekende fractionatie bij abiotische neerslag van aragoniet. De isotopencurven zijn uitgezet in hoofdstuk 2 tegen de afstand van de groeilijn en de umbo, het scharnier- punt van de schelp. Hierdoor is het mogelijk om groeisnelheden te bepalen. Het blijkt dat volwassen Anodontites trapesialis schelpen tijdens het droge seizoen 56 micrometer en tijdens het natte seizoen 32 micrometer per dag groeien. Schelpen groeien tijdens hun eerste levensfase echter veel sneller. Daardoor is het mogelijk om snelle fluctuaties in de samenstelling van het de schelp omringende water vast te leggen. De juveniele schelp die na een jaar uit Playa Cocha werd gehaald hebben we ook doorgemeten (hoofdstuk 3. Omdat we echter niet weten wanneer de schelp precies is geboren, (maar waarschijnlijk ongeveer ten tijde van het begin van ons experiment), hebben we een groeicurve gemaakt op basis van drie bekende punten in de isotopencurve: de collectiedatum, de datum van uitstroming van Amazonewater uit Palaya Cocha zeven maanden eerder, en de datum van instroming van Amazonewater, ongeveer vijf maanden weer daarvoor. De chemie van de groeilijnen uit de eerste maanden van het leven van de juveniele Anodontites trapesialis (in het droge seizoen eind 1998) laat een grote en snelle fluctuatie zien in de isotopencurves. Helaas hebben we uit deze tijd slechts enkele watermonsters die we moeilijk kunnen vergelijken met de schelp. Er werd echter ook in die periode elk kwartier de temperatuur van het Playa Cocha water genomen. In de isotopencurves van de schelp en de temperatuurgrafiek zitten grote overeenkomsten. Het laatste gedeelte van het droge seizoen wordt redelijk gereflecteerd in de isotopencurve. Het eerste gedeelte van de isotopencurve van de schelp is echter niet te correleren met de temperatuurgrafiek. De groeisnelheden die we hebben kunnen berekenen lopen van 0.45 mm in de eerste dag terug naar 0.16 mm op de laatste dag van groei, ongeveer een jaar later. Met deze grote snelheden zijn schelpen in staat om regenbuien in een kleine watermassa te kunnen vastleggen, zoals de regenbui van 8 en 9 januari 1999 terug te vinden is in de isotopencurve. Nu eenmaal bekend is hoe veranderingen in de chemische samenstelling van het omringende water worden vastgelegd in schelpen, is het dus mogelijk om met behulp van isotopen metingen aan groeilijnen in fossiele schelpen iets te kunnen zeggen over de veranderingen in chemische samenstelling van het water waar deze schelpen in leefden, en daarmee is het mogelijk om iets te zeggen over veranderingen in de neerslag in die tijd. De schelpen uit de Pebas Formatie, sedimenten die tussen het late Vroeg Mioceen en het vroege Laat Mioceen zijn afgezet, zijn daarom op isotopensamenstelling doorgemeten. Een verrassende uitkomst van metingen aan schelpen die ongeveer 16 miljoen jaar geleden geleefd hebben, is dat ze zeer sterk overeenkomen met de isotopencurven van moderne schelpen die we verzameld hadden uit Playa Cocha, en uit de Itaya Rivier, een zijrivier van de Amazone (hoofdstuk 4). Zoals vermeld is de isotopenvariatie in een schelp afhankelijk van de verandering in isotopensamenstelling en de hoeveelheid neerslag. Zoals hierboven vermeld is de isotopenwaarde van neerslag van verschillende factoren afhankelijk. Echter al deze factoren verschillen niet wezenlijk met de huidige. Zelfs de sedimenten waar de fossiele schelpen (Diplodon aff. longulus) in gevonden zijn, lijken door de afwisseling van zand en kleilaagjes, op afzettingen die in Playa Cocha werden neergelegd ten tijde van ons experiment. De conclusie is dan ook dat de seizoensveranderingen 16 miljoen jaar geleden zeer vergelijkbaar met die van nu moeten zijn geweest en dat de regenvalpatronen daar de groot- ste factor in zijn. Niet alle schelpen uit de Pebas Formatie laten een dergelijke uitslag zien. Er zijn schelpen die niet, zoals die van hoofdstuk 4, eenzelfde verandering laten zien als de moderne schelpen, maar die een signaal laten zien met een veel lagere amplitude (hoofdstuk 5). Deze schelpen hebben later in de Pebas geleefd dan die besproken in hoofdstuk 4. Een verklaring hiervoor zou kunnen zijn dat toen deze schelpen in "de Pebas" groeiden er andere klimatologische omstandigheden heersten, er een minder uitgesproken afwisseling van droge en natte seizoenen was. Er is echter ook een andere verklaring mogelijk die on- dersteund wordt vanuit de paleontologische invalshoek. Wanneer water zich in een groot meer bevindt en daar een grote verblijftijd heeft, dan wordt dat water vermengd met water dat eerder of later in datzelfde meer is terecht gekomen. Zodoende worden wateren met verschillende isotopische samenstellingen met elkaar vermengd, wat als gevolg heeft dat de extreme waarden van het vermengde water, de amplitude van het isotopensignaal door de tijd heen af zullen nemen (hoofdstuk 5). Dit wordt vervolgens weer opgeslagen in de groei-incrementen van de schelpen die dus een lagere amplitude laten zien in hun isotopencurve. Er blijkt nog wel een jaarlijkse variatie in regenval te zijn, dus is er nog wel degelijk sprake van een weerpatroon dat vergelijkbaar is met het huidige, door de ITCZ gestuurde regenpatroon. Gezien het feit dat de groeisnelheid in het droge en natte seizoen van elkaar verschilt (zie hoofdstuk 2), is het niet mogelijk om voor de tijdsintervallen uit het Mioceen waarin deze schelpen hebben geleefd te bepalen of de lengte van de droge en natte seizoenen hetzelfde is als de huidige lengte van deze seizoenen. Als opgelost CO 2 gas in het water niet wordt aangevuld door rottend plantenmateriaal, door chemische erosieproducten of door nieuw instromend water van elders, dan kan het in equilibrium komen met atmosferisch CO2 gas. Uit de schelpen hebben we kunnen halen dat de wateren "de Pebas" niet in equilibrium waren met atmosferisch gas, zoals zeewater dat bijvoorbeeld wel is. De kool- 13 stof isotopenwaarde van zeewater heeft een typische δ CDIC waarde van ongeveer 0 tot +2 . De minimale waarden in de koolstof isotopencurve van de Pebas schelpen komen op minder dan Ð10 in het natte seizoen. Deze waarden pleiten dus vóór een constante aan- voer van rottend plantenmateriaal en tegelijkertijd tégen een mariene invloed in het gebied. Door de vondst van structuren in sedimenten die duiden op (micro-) getijdenwerking, en die van graafsporen zouden een argument kunnen zijn vóór een marien milieu in "de Pebas". We hebben één sectie uit de Pebas Formatie, de Santa Rosa sectie, extra uitgelicht, om te kijken of in de sedimenten en schelpen uit deze ontsluiting een antwoord ligt opgesloten op de vraag of de Pebas nu een zeewater of een zoetwater milieu was, (hoofdstuk 6). De sedimenten zijn representatief voor de reeds bekende sedimenten uit de Pebas Formatie. Of de sedimenten in de sequenties nu het gevolg zijn van water dat zich cyclisch heen en weer bewoog (trans- en regressies, hoofdstuk 6) of dat er een uitbouwende delta vanuit de Andes het (geleidelijk dalende) Pebas Bekken opvulde, waarbij de verschillende lobben zich door de tijd heen verplaatsten (hoofdstuk 1, feit is, dat daar waar het waterniveau boven de sedimenten op zijn hoogst was, we schelpen hebben kunnen vinden en daar waar er sprake was van zeer ondiep water, er geen schelpen werden gevonden. Hoe hoger de waterstand (hoe dieper de sedimenten onder water lagen) des te beter zijn de schelpen bewaard gebleven. Zu- urstofarme omstandigheden in de diepere wateren zullen hier zeer waarschijnlijk een grote rol in hebben gespeeld. Behalve die van schelpen uit een kleine ontsluiting nabij de kleine nederzetting "Buenos Aires" in zuidoostelijk Colombia, laten alle isotopensignalen van de onderzochte schelpen uit de Pebas Formatie zien dat ze in een zoetwater milieu leefden. Zo ook dus de schelpen uit Santa Rosa. Graafsporen, en wel die van de (Glossifungites ichnofacies duiden echter op mariene milieus, omdat deze sporen momenteel alleen door mariene fauna worden gegenereerd. Omdat er in het Amazone gebied rond Iquitos en Letitia, in zuidoost Colombia zoetwater "zee"koeien, zoetwaterdolfijnen en zoetwaterpijlstaartroggen voorkomen, is het zeer wel mogelijk dat deze gravende fauna ook een zoetwaterminnend familielid heeft (gehad). Resumen

La selva baja situada en las proximidades de Iquitos, ciudad situada en el noreste de Perú, representa una de las formaciones más ricas en biodiversidad de la Tierra (Kalliola and Puhakka, 1993). En la cuenca amazónica ha cobrado especial importancia el conocimiento de la biodiversidad y el uso sostenible del paisaje, esto es por lo que recientemente se han llevado a cabo estudios sobre el desarrollo del Neógeno y su selva. Se ha observado que la riqueza en especies en el Amazonas parece estar controlada por la mayor o menor inciden- cia de algunos factores relacionados con cambios en las cuencas hidrológicas, los cauces, la distribución de los suelos, cambios en el clima, la tectónica y ciertas actividades humanas (construcciones de carreteras y planes de desarrollo (Kalliola and Paitán Flores, 1998)). Para entender mejor el impacto humano y las presiones del desarrollo económico se pre- cisa información sobre la variabilidad del medioambiente y la evolución paleo-geográfica y paleo-ecológica de la región en el tiempo geológico. Con este trabajo presentamos la evolución regional del Mioceno, en concreto, el periodo comprendido entre el final del Mioceno Temprano y el inicio del Mioceno Tardío (16 y 12 millones años atrás). Para ello se han estudiado fósiles de moluscos bien preservados encontrados en la región. La amazonía cuenta con estudios relevantes sobre la climatología de los Andes colombinos realizados por científicos holandeses, como es el grupo de investigación liderado por van der Hammen y Hooghiemstra de la Universidad de Ámsterdam. En este grupo destaca la investigadora Carina Hoorn, cuyos estudios han sido pioneros en el área de paleo-geografía y paleo-ecología de la cuenca amazónica. Hoorn llevó a cabo sus estudios usando polen de los afloramientos (sedimentos y rocas) procedente de los bancales del río Amazonas y sus ríos secundarios (Hoorn, 1994b). Bajo el mismo estudio de campo, Frank Wesselingh desarrolló estudios paleontológicos en las conchas procedentes de los mismos afloramientos, los cuales forman parte de su tesis doctoral (Scripta Geológica 133). En la misma línea de investigación, los estudios realizados por Hubert Vonhof desvelaron la procedencia del agua en la que vivieron esas conchas (llevado a cabo mediante el estudio químico de los isótopos de estroncio). En base a los anteriores trabajos, Wesselingh y Vonhof observaron la necesidad de llevar a cabo el estudio de la paleo-climatología de la región, lo cual ha dado lugar a esta tesis doctoral. Teniendo en cuenta que la situación geográfica (longitud) de la cuenca amazónica no ha cambiado desde el Mioceno (situada en la selva tropical), hemos comparado el clima durante el Mioceno y la era actual. El resultado de este estudio se presenta en el gráfico 1.6 de la página 7, el cual muestra que en Iquitos la temperatura

xxi media mensual no presenta fluctuaciones significativas durante el año, pero sí las precipitaciones medias. Esta es la razón por la que se decidió cuantificar la precipitación y los cambios en los isótopos de oxigeno de la misma durante las estaciones secas y húmedas de la amazonía oeste. Sin embargo, esta metodología no puede ser aplicable en el caso del Mioceno por la imposibilidad de obtener precipitación ocurrida 12 - 16 millones de años atrás, por lo que debe ser estudiada indirectamente. Una manera de evaluar la composición química de la precipitación en el pasado consiste en estudiar las aguas superficiales (lagos y ríos) de la región amazónica, ya que se ven sometidas a fluctuaciones durante el año debido a las diferentes cantidades de precipitación y evaporación. La cantidad de precipitación y su composición isotópica cambian significativamente a lo largo del año, como puede obser- varse en las gráficas proporcionadas por la Agencia Internacional del Átomo y la Energía (IAEA) (http://www.iaea.org/). Las áreas cercanas al ecuador quedan comprendidas entre el régimen de la zona de convergencia intertropical (ZCIT). La ZCIT es un cinturón alrededor de la línea ecuatorial en donde gracias a la acción del sol las masas de aire se elevan. La posición de la ZCIT varía con el ciclo estacional siguiendo la posición del sol. En el con- tinente la ZCIT es más amplia debido a su menor conductancia calorífica, y en los océanos más estrecha y cercana al ecuador geográfico (figura 1.5 en página 6. Debido a la elevación de las masas de aire en el ecuador, las masas de aire de otras áreas se acercan a la superficie de la tierra lo que resulta en la formación de vientos alisios. Cuando la ZCIT esta situada en el norte del ecuador geográfico, en el área amazónica los vientos alisios del sureste son secos. En cambio, cuando la ZCIT esta situada en el sur del ecuador geográfico, los vientos alisios proceden del océano Atlántico proporcionando humedad a los flancos del este de los Andes. La composición de isótopos de oxigeno en las precipitaciones es altamente variable dependiendo de la composición isotópica del área de origen (área de evaporación) de la precipitación, la longitud, la altura, la distancia desde el área de formación de la precip- itación, la temperatura y la cantidad de la precipitación. Durante la estación seca el vapor de agua en la selva amazónica y la precipitación que se deriva de éste mantienen la misma composición isotópica de la fuente inicial, es decir, la misma que el océano Atlántico. Las pérdidas de agua desde el océano hasta las montañas son debidas a la evapotranspiración. Sin embargo, durante la estación de las lluvias (estación húmeda) estas pérdidas de agua son despreciables frente a la enorme cantidad de precipitación lo que produce una frag- mentación de la composición isotópica del vapor de agua, en la cual los isótopos pesados se depositan fácilmente, siendo difícil su evaporación. Esta es la razón por la que el peso del vapor de agua disminuye desde el océano hasta la cordillera de los Andes (ver figura 1.19. Estos cambios en la composición de las lluvias afectarán igualmente a la composición de las aguas superficiales a lo largo del año. El efecto de la evaporación sobre la composición isotópica del agua superficial se observa de la siguiente manera: durante la estación húmeda el agua tiende a ser menos pesada y durante la estación seca tiende a ser más pesada debido a la composición de la lluvia. Esta variación anual de la composición isotópica del oxigeno en el agua superficial tuvo un mayor efecto en lagos pequeños (estanques) situados en el área de inundación del Amazonas (demostrado en el capítulo 2). También se llevó a cabo un muestreo de aguas durante un año en playa Cocha, lago de reducidas dimensiones en el área de inundación del Amazonas. Las mediciones de la composición isotópica mostraron que durante la época de lluvias este lago es parte de la corriente principal del río, ésto es porque los isótopos de oxigeno y carbono son menos pesados en comparación con la época seca, cuando playa Cocha prácticamente se evapora y sólo es alimentada por lluvias y que- bradas locales. Para completar este estudio, se realizó un experimento con conchas vivas en playa Cocha que duro un año (ver capítulos 2 y 3). Estos moluscos encontrados en playa Cocha (llamados Anodontites trapesialis se marcaron, se depositaron en cestas rellenadas con sedimentos procedentes de las profundidades del lago y se dejaron en el fondo. Para poder encontrar fácilmente las cestas durante la época de lluvias, se colocaron boyas asoci- adas a cada cesta. El diseño de muestreo consistió en sacar a la superficie una cesta cada dos meses, sin embargo no fue posible porque las cestas eran demasiado pesadas. Esto ocurrió durante la época de las lluvias cuando el fondo de playa Cocha media más de cinco metros de profundidad. Después de las lluvias, cuando el nivel de agua fue menor, observamos que casi la totalidad de las conchas marcadas habían desaparecido. Tan solo diez conchas juveniles permanecieron en las cestas al no poderse ir con el agua del río, una concha con un año de edad y una concha adulta (denominada "la abuela") (capítulo 3). A continuación, las conchas fueron tratadas con resina artificial (preparación de las muestras por el Instituto de Ciencias de la Tierra) para obtener secciones lo suficientemente pequeñas paralelas al eje de crecimiento máximo. Las secciones fueron analizadas mediante un microscopio Mi- croMill (figura 1.16, página 18). El MicroMill tiene un taladro de dientes. Este microscopio muestra las imágenes de los incrementos de crecimiento de las conchas y las transmite al ordenador. Una vez marcadas las líneas paralelas que separan cada incremento, el taladro las pule y el residuo de carbonato cálcico es analizado con el espectrómetro de masas. Los valores obtenidos de los isótopos de oxigeno y carbono de las conchas fueron comparados con los del agua, dando como resultado que los valores de oxigeno en las conchas, modifi- cados mediante una formula conocida en la bibliografía, fueron similares a los encontrados en el agua de playa Cocha. Los valores de carbono en el residuo sólido tras el pulimento resulto comparable con los valores de los isótopos del carbono del bicarbonato disuelto en el agua. El bicarbonato es usado por los moluscos para producir aragonito, que es el material de su concha, sin embargo se observó la existencia de una fragmentación en este mineral conocida como precipitación abiótica del aragonito (Romanek et al., 1992). En el capítulo 2 se han obtenido las curvas isotópicas y la distancia entre el incremento y el umbo (el punto de la bisagra), con lo que hará posible la determinación de la rapidez de crecimiento de las conchas. De este modo se observó que un espécimen adulto de Anodontites trapesialis crece 56 micrómetros por día en la época seca y 32 micrómetros por día en la época de lluvias. También se observó que las conchas crecieron más rápido en su fase inicial de vida, lo que permitió detectar precipitaciones aisladas. En el caso de la concha juvenil emphAnodontites trapesialis, recogida transcurrido un año (capítulo 3), el procedimiento para obtener la curva de crecimiento se realizó en base a tres puntos conocidos de la curva isotópica: la fecha de entrada del río en el lago, la fecha de salida del río (cinco meses después) y la fecha de muestreo de la concha. Esta metodología fue elegida ya que no se conocía exactamente cuando nació el molusco, aunque creemos que fue cuando se inició el experimento. Los resultados del análisis químico de los incrementos en los primeros meses de su vida (en la época seca del año 1998) mostraron una fluctuación grande y rápida. Desafortunadamente, no se contó con una gran cantidad de muestras de agua en esa época, lo que dificultó el análisis comparativo. Sin embargo, se dispuso de medidas de la temperatura del agua cada 15 minutos que fueron comparadas con las curvas isotópicas de la concha. Los resultados mostraron una buena correlación durante la última parte de la curva de la estación seca y los índices de crecimiento calculados decrecieron desde 0.45 milímetros por día en el primer día de vida hasta 0.16 milímetros el día de su recolección (aproximadamente un año más tarde). Estos índices de crecimiento permitieron detectar lluvias aisladas, como por ejemplo la ocurrida durante el 8 y 9 de enero de 1999, registrada en la curva isotópica. Hoy en día es posible conocer cómo quedan registrados los cambios ambientales químicos en conchas, lo que permite la reconstrucción de los cambios en la precipitación. De este modo, estudiando las conchas fósiles del Mioceno podremos conocer los cambios de la lluvia Miocena. Los moluscos fósiles de la Formación Pebas encontrados en los afloramientos de Mazán fueron medidos por su composición isotópica. Los resultados de estos fósiles de 16 millones años atrás fueron muy similares a los que se observan ahora en playa Cocha y río Itaya, río secundario del Amazonas (capítulo 4). Como se ha descrito anteriormente, la variación isotópica de una concha depende de la cantidad y la composición de la precipitación, esta última dependiendo de varios factores que fueron constantes hasta la fecha actual. Además, los sedimentos donde hemos encontrado fósiles de la especie Diplodon aff. longulus (afloramiento de Mazán) tienen similares características que los encontrados en playa Cocha (capas de arena y arcilla alternadas). Por eso se ha llegado a la conclusión de que los cambios estacionales 16 millones años atrás fueron parecidos a los actuales, y que además, la variabilidad temporal de las precipitaciones fue el factor más importante. Los fósiles de la Formación Pebas más antiguos 4 mostraron ciclos con la misma amplitud que las cochas de playa Cocha (era actual) y los fósiles mas jóvenes ciclos con menor amplitud (capítulo 5). Una posible explicación es que estos fósiles vivieron en un régimen climatológico diferente con menos cambios entre las estaciones secas y húmedas. El agua en lagos grandes tiene largos tiempos de residencia y está mezclada con otras aguas que llegan al lago antes o después, por lo que la composición isotópicas será mixta y las máximas y mínimas de los valores isotópicos disminuirán con el tiempo (capítulo 5 (reflejado en las curvas isotópicas registradas por los incrementos de crecimiento de las conchas). Sin embargo, aunque la amplitud de la curva isotópica es menor podemos concluir que la variación estacional de la precipitación es similar a la registrada en la actualidad (régimen de ZCIT), pero no fue posible determinar la duración exacta de las estaciones en el Mioceno debido a la variación del índice de crecimiento entre estaciones (capítulo 2). Por otro lado, si el CO 2 disuelto en el agua no se renueva (descomposición de la materia orgánica, erosión, productos quími- cos procedentes de otras aguas), el gas permanecerá en equilibro con el CO 2 atmosférico. El estudio de los fósiles reveló que las aguas de "Pebas" no estaban en equilibrio, como ocurre con el agua de mar. El valor isotópico del carbono del agua de mar (δ 13C ) tiene un valor típico entre 0 +2 . Los mínimos obtenidos en las curvas de los isótopos de carbono fueron menores que -10 durante la estación húmeda, lo que indica que hubo una provisión constante de material vegetal descompuesto. Este valor ratifica que la región no estuvo bajo la influencia marina, a pesar de que existen estudios en "Pebas" en los que se han encontrado vestigios de micro-mareas en estructuras sedimentológicas. Para completar este estudio sacamos una sección representativa de la Formación Pebas, afloramiento próx- imo a Santa Rosa, con objeto de ver si en los sedimentos o en las conchas se encontraban pruebas de la existencia de agua dulce o marina (capítulo 6). Donde el nivel del agua sobre los sedimentos fue más alto encontramos fósiles, pero no donde el agua alcanzó niveles más bajos. Este hecho fue observado con independencia de si las secuencias de los sedimentos fueron cíclicas (trans- y regresiones, capítulo 6) o si los sedimentos formaron un delta en la cuenca de Pebas (esta cuenca esta bajando lentamente de nivel, capítulo 1). Además, los fósiles presentaron mejor conservación donde el nivel del agua fue más alto. Las condi- ciones anaeróbicas de las aguas profundas fueron importantes para ese proceso. Por último, observamos que todas las señales isotópicas de las conchas fósiles de Formación Pebas y los moluscos de Santa Rosa mostraron un medioambiente rico en agua dulce, a excepción de las conchas fósiles de uno de los afloramientos próximo al pueblo de Buenos Aires en el sureste de Colombia. Gingras et al. (2002b) argumentan que la existencia de productos orgánicos descompuestos (bioturbation) de la actividad de la icnofauna (Glossifungites) en los sedimentos de la Formación Pebas son indicativo de un ambiente marino. Sin embargo, en mi opinión es bastante probable que la icnofauna haya tenido algún descendiente procedente de agua dulce, ya que hoy en día en la región de Iquitos y Leticia en Colombia podemos encontrar también delfines, rayas y manatíes.

Chapter 1

Introduction and summary

Abstract In this thesis the excellently preserved fossil bivalves from the Miocene Pebas For- mation in northeastern Peru and southernmost Colombia are examined for their suitabil- ity as recorder of seasonal change in precipitation. Stable isotope records from growth increments of modern bivalves, which were monitored during a one year experiment, accurately reﬂect the seasonal change of ambient water chemistry. Extrapolating these results to the Miocene fossils indicates that rainfall patterns are strikingly similar to today and that the endemic fauna of the Pebas Formation exists (almost) entirely of freshwater molluscs. The isotope records indicate an environment consisting of a wetland system with (interconnected) lakes and ﬂoodplains.

1.1 Aim of the present work

The humid rain forest surrounding the city of Iquitos in northeastern Peru is recognized to be one of the most species-rich forests in the world. The remarkable species richness in Amazonia, is the result of a complex interplay of factors such as the long-term dynamics of catchment areas and river courses, soil distribution, climate change, tectonics and human activity. This extraordinarily diverse area is currently under severe stress by road construction and accelerated deforestation pressure (Kalliola and Paitán Flores, 1998). To better scale the impact of human activity on lowland Amazonian ecosystems, one must understand the natural variability; i.e. the paleogeographical and paleoecological evolution of this region through geologic time. In the footprints of previous work by Carina Hoorn (1994b) and PAUT (Proyecto Ama- zonia de Universidad Turku: Kalliola and Puhakka, 1993), the Climazonia Project was developed to improve Neogene paleogeographical and paleoclimatological reconstructions for western Amazonia. The project is a joint effort based on molluscs, (i.e. projects of Wes- selingh, National Museum of Natural History, Leiden and University of Turku, Finland; Vonhof and Kaandorp, Vrije Universiteit Amsterdam) and combines paleontological (Wes- selingh, 1993; Wesselingh et al., 2002) and isotope geochemical studies of fossil molluscan 2 Introduction and summary shells (Vonhof et al., 1998, 2003); Chapter 4, (published as Kaandorp et al. 2005); with sedimentological studies of the depositional environment Chapter 5, (published as Kaandorp et al. 2006) and Chapter 6, that is adapted from Wesselingh et al., 2006b.

In recent years, the Neogene development of the Amazon Basin and its rainforest biota has become a major topic in biodiversity and sustainable land-use studies. In this study we focus on the isotope geochemistry of the excellently preserved fossil molluscs in the area. A process study for proxy validation has been conducted which gave us an insight into the paleoclimatic and paleoenvironmental evolution of western Amazonia during the period between 16 and 12 million years (Ma), i.e. the early Late and the late Early Miocene. The abundant fossil bivalves from the Miocene Pebas Formation, found in outcrops along the Amazon River and its tributaries, hold information of past climate and environments. In order to investigate this in Playa Cocha, an Amazonian floodplain lake, a monitoring experiment with living shells was carried out for a period of a year (Chapter 2, published as Kaandorp et al. 2003, and Chapter 3). The Modern bivalve Anodontites trapesialis studied in Chapter 2 is a várzea-type floodplain dweller; Triplodon corrugatus, studied in Chapter 4, was collected in the black-water river Río Itaya, a tributary of the Amazon River, which is not a floodplain environment. As the rainfall pattern in Amazo- nia is unique and important for shaping the landscapes in the modern Basin, it probably has played a comparable significant role in the past. To obtain rainfall patterns from the past and paleoclimatological and paleoenvironmental indications, a commonly accepted technique is to make use of isotope records from carbonates, but in this regard so far little has been done in Amazonia. We also compared the isotope geochemistry of the monitored molluscan shells with that of ambient water. For this purpose the water of Playa Cocha was sampled twice a month. By comparing the results of the process study with the geochemistry of fossil bivalves from the Pebas Formation (Chapters 4, 5 and 6) we were able to reconstruct the paleoclimate and paleoenvironments of western Amazonia. A key target for the reconstructions is the stability of the seasonal climatic cycle during Miocene uplift phases of the Andes mountain range. This study is supported by the Foundation for the Advancement of Tropical Research (WOTRO), residing under NWO (the Netherlands Organization for Scientific Research), under grant W76-195.

1.2 The geography of Amazonia

The Amazonian rainforest is one of the largest in the world and is almost entirely covered with thick tropical forests, cut through by large rivers. Only the rims of the Amazon Basin (Amazonia) are mountainous: the Andes in the West, the Guyana Shield in the North, the Brazilian Shield in the South and some inselbergs in the Northwest (Figure 1.1). The Amazonian Basin is relatively ﬂat, with only little topography that can be seen at 1.2 The geography of Amazonia 3

arches such as the Iquitos Arch or Iquitos fore bulge (Roddaz et al., 2006), or at the white sand hills in the surroundings of Iquitos (Kauffman et al., 1998; Linna et al., 1998). The Peruvian segment of rainforest alone has a surface area of approximately 770 000 km2, (almost the size of the UK, Spain and the Netherlands together) of which the major part is defined as lower rainforest, (80Ð600m above sea level, (Kalliola and Puhakka, 1993)). The Amazon River is more than 6 500 km long and drains a catchment area of 6 - 7 million km2, discharging an annual mean of ∼106 000 m 3/s at Manacapuru (just upriver of the confluence of the Solimões River and Rio Negro, (Zhakarova et al., 2006). It transports a mean of ∼ 3 - 3.5 million tons of sediment per day (Meade et al., 1985). The seasonal rainfall fluctuation in the Amazon Basin has resulted in the formation of distinct landscapes in modern Amazonia; terra firme, land that is never flooded, and the floodplains, land that is seasonally or (semi-) permanently flooded. According to the chemical composition and the sediment load of the flood-water the floodplains are divided in white-water inundation forest (várzea), and black-water inundation areas (igapó), hosts of species-rich and highly adapted floodplain forests (De Simone et al., 2003). The temporally inundated (várzea-type) forests occupy about 8% of Peruvian Amazonia, (Salo et al., 1986), almost twice the size of the Netherlands. An additional 78 000 km 2 of Brazilian land along the mainstream of the Amazon is subject to flooding. The two types of floodplains undergo different sedimentation rates. While the white-water Amazon River contains a suspended load of 100 mg/l, the black-water river Río Negro contains only 6 of this amount. White- water rivers (Sioli, 1984) originate from the Andes Mountains and the pre-Andean zone and contain a sediment load consisting of clay-minerals whereas the soils of the igapó are the result of strongly weathered and lixiviated Tertiary sediments (De Simone et al., 2003). The big Andean rivers usually form braided streams in the foot hill zone (500-100 m a.s.l.) and start meandering further down stream. As erosion and aggradation continually destroy and reform fluvial landforms, the rivers undergo rapid spatial and temporal change, (Kalliola et al., 1991). Newly formed grounds are colonized by new plants; going inland from the river, older meander bars are vegetated by older forest and the vegetation will slowly transforms into a terra firme forest. (The major part of western Amazonian forest grows on fluvial deposits). These floodplain forest communities, growing on the várzea adjacent to the Amazon River and its tributaries, are subject to large flood pulses that last several months yearly. The trees in these forest communities are facing long periods of flooding, up to 230Ð270 days in central Amazonia. Many várzea trees are deciduous, some are evergreen as are most of the igapó trees (De Simone et al., 2003 and references therein). The Modern bivalve Anodontites trapesialis studied in Chapter 2 is a várzea-type floodplain dweller; Triplodon corrugatus, studied in Chapter 4, was collected in the black-water river Río Itaya, a tributary of the Amazon River, which is not a floodplain environment. As the rainfall pattern in Amazonia is unique and important for shaping the landscapes in the modern Basin, it probably has played a comparable significant role in the past. To obtain rainfall patterns from the past and paleoclimatological and paleoenvironmental indications, a commonly accepted technique is to make use of isotope records from carbonates, but little has been done in Amazonia. Abundant fossil bivalves from the Miocene 4 Introduction and summary

72°W 66°

Guyana Shield C

o Inselbergs 0° r (Mesa de Iguaje and others)

d Pebas Manaus i Leticia l Manacapuru l Iquitos e A M A Z O N I A r a 12°S d e Lima los Andes

Brazilian Shield

(And

es M es ountain

72° s)

Figure 1.1: Geographic setting of Amazonia

Pebas Formation, found in outcrops along the Amazon River and its tributaries, should hold

80° 40°

ATLANTIC OCEAN Venezuela

Colombia

Negro Ecuador MANAUS Río Amazonas Silmões zonas NAUTA Ama Marañon eira Parus Mad Tapajos

Uca Xingu yali Tocatins Peru

Bolivia PASSIFIC Brazil OCEAN Paraguay Tropic of Capricorn

Figure 1.2: The Amazon River and its basin: between the confluence of the Andean (white-water) rivers Ucayali and Marañon, just northeast of the town of Nauta, in Peruvian Amazonia, until the border of Peru and Brazil (at the town Tabatinga), the river is called Río Amazonas. The Brazilians call the river Solimões. From the confluence of the Solimões River with the Río Negro near the city of Manaus to the outflow into the Atlantic Ocean, the river is called Río Amazonas again. 1.2 The geography of Amazonia 5

Figure 1.3: Conﬂuence of black-water river Río Nanay and white-water Río Amazonas in the wet season. (See boats for scale).

information of past climate and environments.

b d a c e

Figure 1.4: The Amazon Basin and its environments. (a) Terra firme forest covers approx. 90% of the rainforest in the Amazon Basin. (b) Floodplain lake at low-water (rooted marginal vegetation, Photo W. Crampton) (c) Swamps in flooded forest at low-water (Photo N. Lovejoy). (d) Floating macrophytes raft (’floating meadow’, Photo W. Crampton) (e) Floodplain lake at high-water (floating marginal vegetation, Photo W. Crampton, http://www.flmnh.ufl.edu/ucamara/habitats.htm). 6 Introduction and summary

1.3 Amazonia’s seasonal variation of climate

The atmosphere above the Amazon Basin is the core of earth’s largest and most intensive land-based convective center, which is best developed during Southern Hemisphere (austral) summer. This convection results in rainfall amounts that make the Amazon Basin/Amazonia one of the wettest regions on earth (Figure 1.5; Marengo and Nobre, 2001).

L L 0° ITCZ 0° "dry" "wet" ITCZ 10° H H H L H

ITCZ in ITCZ in July January

Figure 1.5: The Inter-Tropical Convergence Zone (ITCZ) is a low-pressure band in the tropics where air masses converge from both hemispheres at the boundary between northeasterly and southeasterly trade winds. The ITCZ follows the insolation maxima, migrating to the warmest surface areas throughout the year attaining its northernmost displacement in June-July, and southernmost in December-January. H stands for atmospheric high pressure, L stands for low pressure.

The region has a distinct seasonal precipitation cycle; the rainfall pattern is controlled by a monsoonal system driven by the seasonal migration of the ITCZ, the Inter-Tropical Con- vergence Zone (Marengo, 1998), a low-pressure band in the tropics where air masses converge from both hemispheres at the boundary between northeasterly and southeasterly trade winds. The ITCZ follows the insolation maxima, migrating to the warmest surface areas throughout the year attaining its northernmost displacement in June-July, and southernmost in December-January. A contracted equatorial continental air mass, together with activity of the eastern Paciﬁc and western Atlantic ITCZ, produces southeastern trade winds over Amazonia. This results in an Amazonian dry period during boreal summer with low river water levels in Peruvian Amazonia in June-September. During austral summer the ITCZ lies south of Amazonia, attracting northern hemisphere northeastern trade winds. Convec- tive activity of an expanded continental air mass over Amazonia in this season causes a wet regime in the region, with maximum precipitation and inundation of várzea in October- May. Northeastern trade winds moving humid air meet the ITCZ in the northwestern part of Amazonia. Elevation of the air mass due to the Andean topography will produce maximum rainfall in April-June. In the central part of the Basin the maximum rain season has a peak during March-May (Marengo, 1998). Depending on the distance from the insolation maxima at the equinox, areas between the maximal latitudes where the ITCZ transpires may expect one or two wet and dry sea- 1.3 Amazonia’s seasonal variation of climate 7 sons annually. Thus, changes in the position of the ITCZ over the continent can have a local impact on the number of seasons in a particular area. For instance, the ITCZ moves twice a year over Suriname leading to a long and a short annual "rain-time", whereas the city of Iquitos is subject to only one wet season (Figure 1.6) because the orography of western Amazonia and the eastern Andes forces the ITCZ to stay over Iquitos for a large part of the year.

350 30°C (mm)

300 25 Monthly mean rainfall (mm) 250 20

200 Monthly mean temperature °C 15 150

10 100

5 50

0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure 1.6: The annual climate variability of Iquitos. The monthly mean temperature variation is less than daily temperature variation. The annual precipitation shows a cycle with a peak in March and a drop in June.

Longer term changes in rainfall are related to the El Niño Southern Oscillation (ENSO). Driest periods in northern and central Amazonia, with low Amazon River levels, are related to the low ENSO phase that is associated with the El Niño phenomenon, whereas positive rainfall anomalies relate to the high ENSO phase, linked to La Niña. The drought in northern and central Amazonia during the El Niño event of 1998 (the start of our monitoring experiment, Chapter 2) is claimed to be the most severe in the last 118 years (Marengo and Nobre, 2001), although there is some controversy on this in the literature (Williams et al., 2005). The extremely high water levels we measured in the vicinity of Iquitos were related to the 1998-1999 La Niña event. The climatic oscillation results in a distinct wet-dry season cyclicity to which rainforest taxa are strongly adapted (Salo et al., 1986; Puhakka and Kalli- ola, 1995; De Simone et al., 2003). On an even longer time-scale, changes in seasonality of the local climate are linked to the insolation variation that is driven by the Milankovitch cycles (of which the precession cycle (∼ 26kyr) is the most important). Data obtained from speleothems in northeastern Brazil (Wang et al., 2004) suggest that southward expansion of Northern Hemisphere ice sheets 8 Introduction and summary and northern sea-ice cover caused a southward shift in the position of the ITCZ. In its turn this more southern position of the ITCZ may result in a modulation of heat and salt transport by ocean currents and water vapor export from the Atlantic to the Paciﬁc. Wang et al. (2004) conclude that the currently semi-arid southeastern border of Amazonia experienced relatively frequent, dramatic and abrupt changes in the available moisture during the Pleistocene.

1.4 Neogene sedimentation history of the Amazon Basin

The uplift of the northeastern Andes played a key role in the paleogeographic development of northern South America (Hoorn et al., 1995). With a low Andes Mountain range during the late Oligocene to early middle Miocene the Amazon Basin was drained by a paleo-Orinoco to the North into the Proto-Cariaco basin, Amazonia was a riverine drainage system. After infill of sediments derived from the uplifted Andes the base level of the rivers in the Miocene Amazon Basin was altered. The drainage of the Miocene Amazon Basin had become less efficient, and a large system of (connected) lakes and wetlands came into exis- tence. Between the late Miocene and the Holocene the Andes attained their present configu- ration (Hoorn et al., 1995). The Amazon River established its easterly course to the Atlantic Ocean at ∼ 8 Ma BP and the basin became well drained again, but with drastically changed watercourses and paleogeography. Preservation of invertebrate fossils in river systems is poor. However,preservation of molluscs proved to be excellent in the Miocene lake sediments of the Pebas Formation. In figure 1.9 an interpretation of the sedimentation history of paleo-Amazonia is visualized.

1.5 Pebas Formation

1.5.1 Stratigraphy In 1869, W.M. Gabb described a remarkable fauna of molluscan fossils from the vicinity of Pebas, a town in Peruvian Amazonia (Figure 1.11), of which all but one of the species were new. Because nearly all species are extinct, the assemblage shows hardly any similarity to any modern South American fauna. Because of the exceptional preservation of the fossils Gabb believed that they are of relatively recent age. He described several taxa whose Mod- ern relatives live in freshwater, others are related to extant marine molluscs and yet others to be associated with environments of variable salinities. Nutall’s monograph (1990) on the Neogene molluscs from western Amazonia and adjacent regions, clariﬁed much of the systematic uncertainties that complicated age estimation and environmental interpretation. Through detailed region-wide stratigraphic correlation Nuttall assigned a Middle Miocene age to the western Amazonian fauna and proposed a depositional environment resembling a continually shifting pattern of streams, swamps and lakes of varying salinity, with marine inﬂuence from the north (Caribbean). With pollen data 1.5 Pebas Formation 9

Time Pollen Zone Vene- Localities, Epoch Brazil Peru bivalves measured (Ma) (Hoorn, 1994a) zuela

Los Chorros, Nuevo Horizonte 10 Late Late Miocene Grimsdalea * P. tenuis, D. longulus, A. capax A. capax, D. longulus 12 San Antonio, * Santa Elena P. tenuis (2x), Crasso- 14 P. erectus, retitriletes * Pebas, Santa Rosa D. longulus Pebas Formation Formation Pebas Pebas Formation Formation Pebas Middle Miocene

16 Solimões Formation * D. aff. longulus Psiladiporites - 18 Crototricolpites Indiana, Mazan

Retricolporites 20 Early Miocene

Verrutricolporites Cham- bira? Mariñame sand unit

Figure 1.7: Stratigraphic ranges of Pebas Formation outcrops. White bars represent palynologic stratigraphic ranges (Hoorn, 1994b). Shorter ranges indicated by black bars are based on molluscan zona- tions of Wesselingh et al. (2002). The time scale is that of Berggren et al. (1995).

Hoorn (1993; 1994a) reﬁned Nuttall’s age assignments of the Pebas Formation by dividing it in three pollen zones, (Psiladiporites-Crototricolpites, Crassoretitriletes, and Grimsdalea Zones, Figure 1.7) spanning an age between late Early Miocene and early Late Miocene (∼ 17Ð10 Ma). Parts of the last two zones are subdivided in Ostracoda zones by Muñoz Tor-

b d

Figure 1.8: Drawings of some typical Pebas faunal elements from Vonhof et al. (1998). (a) ?Nas- sarius reductus (Vermeij and Wesselingh, 2002), height 1.1 cm; (b) Melongena woodwardii (Roxo de Oliveira, 1924), height 3.0 cm; (c) Glabertryonia glabra Wesselingh 2006a, height 0.24 cm; (d) Pachy- don trigonalis Nuttall 1990, height 0.6 cm; (e) Pachydon cf. P. ovalis Nuttall 1990, height 0.8 cm. The collection is ﬁled at the Nationaal Natuurhistorisch Museum, Leiden, the Netherlands. All shells were collected at Buenos Aires, Amazonas dept., Colombia. 10 Introduction and summary

25 Ma

Nascent NE Andes

15 Ma

Nascent Amazon

Modern NE Andes

Submarine fan

Modern Amazonia

Figure 1.9: Amazon Basin: interpreting the past. Adapted from Hoorn 2006. Impressions of the Amazon Basin 25 Ma ago, 15 Ma ago and at Present ©2006 Scientific American Inc., Ron Miller res et al. (2006). Hoorn (1994a) deduced a fluviolacustrine depositional environment with marine influence. Wesselingh et al. (2002) added the La Tagua Beds of southern Colombia to the Formation that are supposed to be older than the sediments from the Psiladiporites- Crototricolpites Zone in Peru, extending the age span to ∼ 19Ð10 Ma (Wesselingh et al., 2002). In that paper the taxonomic composition and palaeoecological signature of molluscan faunas from the Miocene Pebas Formation of Peruvian Amazonia are assessed. The Pebas fauna is almost entirely made up of extinct, obligate aquatic taxa, and is dominated in numbers of species and specimens by endemic cochliopine hydrobiid gastropods and pachydontine corbulid bivalves. The geographic distribution of the Pebas Formation and contemporaneous non-marine 1.5 Pebas Formation 11

Figure 1.10: Adapted from Vonhof et al. (2004). Grainsize distribution of a composite section at Los Chorros (Colombia, for location see Figure 5.1). Over the total height of the section samples were analyzed at 50 cm stratigraphic intervals. Grain size distributions per sample were contoured to produce this representation of shifting grain size distribution with stratigraphic height. Lignite layers are marked by black horizontal bars. The grey bar at 20 m in the section is not a lignite but a bone bed with large amounts of plant fossils. Grain size distribution between lignite beds clearly shows coarsening-upwards trends, interpreted to reflect deposition of prograding coastal lobes. Occasional shifts to much finer grain size at 50 cm below the lignites is interpreted to reflect shielded shallow water (lagoonal?) deposition. These fine grained zones generally have a mottled appearance.

74°W 73°W 72°W 71°W 70°W 02°S Colombia Putumayo River

Western Amazonia

Mazán River Na Peru 03°S po Pebas Riv er

Sta Rosa Mazán Indiana Amazon River Iquitos ColombiaLos Chorros San Playa Cocha Antonio Sta Elena 04°S Itaya River Nuevo Horizonte Nauta Peru Yav ari River Brazil 05°S

Figure 1.11: Fieldwork for this thesis has been conducted in 1998 and 1999 in the area between Iquitos in north-eastern Peru and Leticia in southernmost Colombia. Black dots represent sites where fossil bivalves have been collected during ﬁeld campaigns of Wesselingh in 1993 and 1997, and Kaandorp in 1998 and 1999. faunas is shown in Figure 1.12. The Formation encompasses predominantly turquoise-blue smectitic clays; immature, feldspar-rich, usually grey silts and sands, and brown-black or- 12 Introduction and summary

Figure 1.12: Distribution of Pebas molluscan faunas and the probable maximum limits of the Pebas system (after, Wesselingh et al. (2002). ganic clays and lignites. At its type locality it is characterised by 3-7 m thick, mainly coarsening-upwards (CU) cycles (Räsänen et al., 1998; Vonhof et al., 2004; Wesselingh, 2006b, Figure 1.10). The lower part of the Pebas Formation is extremely rich in fossils, often beautifully preserved, i.e. molluscs, ostracods, wood fragments as well as remains of ﬁsh, amphibians and reptiles. In Brazil, Pebas Formation deposits have usually been included in the Solimões Group/Formation, an informal unit containing an excess of western Brazil- ian Amazonian Cainozoic strata (e.g. Maia et al. 1977), but some authors (Costa, 1981; Petri and Fulfaro, 1983) have indeed distinguished the Pebas Formation as such in Brazil. Deposits formerly referred to as "Terciario inferior Amazonico" in Colombian Amazonia, were attributed to the Pebas Formation by Hoorn (1994b). The La Tagua Beds (Nuttall, 1990; Hoorn, 1994b) are included in the Pebas Formation as well. The Curaray Forma- tion of eastern Ecuador (Tschopp, 1953; Baldock, 1982) is the time equivalent of the Pebas Formation.

1.5.2 Environments of Lago Pebas

In what kind of environment lived the Pebas molluscs? Despite the assumed presence of marine taxa (the gastropod genera Turbonilla and Mesalia and the bivalve Corbula), their marine origin in lowland Amazonia, far away from modern coastal areas, was questioned. Gabb’s 1869 work was the first of some fifteen papers dealing with the remarkable fossil molluscs from western Amazonia. Different authors came up with diverse interpretations of the environment that sustained these unusual faunas (freshwater, brackish and/or marine settings) and speculated on the origin of possible marine influence: (i.e. via Pacific, the Caribbean and eastern Atlantic through the course of the present-day Amazon River valley). 1.5 Pebas Formation 13

Figure 1.13: Inferred distribution of assemblages adapted from Wesselingh et al. (2002). Note that near-coastal dysoxic settings (e.g., in channel scours, indicated by a dashed line) may contain the Pachydon obliquus assemblage, that otherwise dominates the deepest parts of the Pebas system. "Shelf" ∼ 5Ð10 m deep, lake bottom below wave base ∼ 10Ð?30m deep.

Molluscan faunas found in Miocene deposits in Western Amazonia have been exten- sively studied by Nuttall (1990) and Wesselingh (2006a) and are found to be predominantly aquatic, endemic and extinct (Wesselingh et al., 2002). The character of the fauna (Figure 1.8) is typical of long-lived lake environments. Pebas fauna assemblages (Figure 1.13) are predominantly lacustrine, but rare ﬂuvial assemblages occur. Sedimentological data and molluscan carbon isotope signatures support such a lacustrine environment of the sampled sequences, (Chapters 5 and 6). The Middle Miocene section at Santa Rosa de Pichana (Figures 1.7 and 1.11, Chapter 6, which is adapted from Wesselingh et al., 2006b), with a characteristic succession of coarsening upward sedimentary sequences that represent typical depositional settings within the Pebas system of western Amazonia (Räsänen et al., 1998; Gingras et al., 2002a; Vonhof et al., 2003, c.f. Figure 1.10). These sequences might either point toward a repeatedly prograding delta front, base level variations or switching coastal lobes in a subsiding setting. The isotope signatures of bivalves in this section indicate fresh water conditions. Its faunistic character is very similar to the Pebas Formation sampled by Wesselingh et al. (2002) based on 285 samples used in Wesselingh et al (2002). Only perimarine taxa (that are rare in the Pebas fauna anyway) are lacking in the Santa Rosa samples. In total, 54 identiﬁable mollusc species were encountered (Appendix 1 of Wesselingh et al., 2006b). The studied fauna can be characterized as predominantly endemic and freshwater.

1.5.3 Marine inﬂuences The discovery of presumed Miocene tidal deposits in Acre, Brazil (Räsänen et al., 1995) lead to the interpretation by Webb (1995) that Miocene western Amazonia was part of an intracontinental seaway occupying all major sedimentary basins of inland South America. This set the stage to an ongoing discussion on the salinity of the depositional environment 14 Introduction and summary

Figure 1.14: A palaeoenvironmental impression of Lake Pebas from Wesselingh et al., 2002. The Andes to the left, the Guyana shield in the upper right-hand corner. The width of the block diagram is ∼ 1,500 km. Abbreviations: FBZ - Foreland Basin Zone, PCZ - Pericratonic Zone. of the Pebas system. Although tidal rythmites are regarded as unequivocal evidence of marine conditions in sedimentary basins (Mazumder and Arima 2005 and references therein), Webb’s views roused severe criticism by Hoorn (1996); Marshall and Lundberg (1996); Paxton et al. (1996). Yet, more tidal deposits were discovered since then in Peruvian Ama- zonia (Räsänen et al., 1998; Gingras et al., 2002a), in the upper Pebas Formation and Nauta Formations (Rebata H. et al., 2006) and Ipururo and Madre de Dios Formations (upper Miocene, Hovikoski et al. 2005). The claim by the latter authors that the Upper Miocene sediments are deposited under saline conditions is disputed by a comment of Hoorn et al. (2006). Hovikoski et al. (2005) proposed a connection with the open sea in the south, recently also suggested by Roddaz et al. (2006). The upper Pebas Formation is fossil-barren and only ichnofauna and sedimentological data (Gingras et al., 2002b; Rebata H. et al., 2006) give evidence of a large, long-lived inland sea with muddy subtidal ﬂats, muddy paleosoils, secondary tidal channels and run-off tidal creeks, muddy tidal point-bar deposits, sandy point bar deposits and sandy shorefaces. Although some of the found ichnofauna is (also) known from freshwater environments, the Modern equivalents of the majority of the found associations are representatives of brackish and marine environments. Unfortunately an outcrop containing both sedimentological and ichnofaunal indicators 1.5 Pebas Formation 15

of saline conditions and molluscs has not been discovered yet in the lower Pebas Formation. The sedimentological features that are proposed to be indicators for marine tidal structures can also be found in floodplain deposits, however, and (micro-) tidal waves are observed in large lakes (Ayers et al., 1956). When the east-west extension of a lake is large enough, the difference of timing of attraction by the moon on both ends will result in a tidal wave. The large extend of the Pebas Formation indicates a lake or system of connected lakes that is larger than the Great Lakes in the US. With the proper dimensions an embayment with an open connection to the sea can get an amplification of tidal motions without the necessity of having saline water (Sztanó and de Boer, 1995). The ostracod fauna from the Pebas For- mation was studied by Whatley et al. (1998), who proposed a low energetic epicontinental marine environment, based on the taxonomic composition and exceptional preservation of the ostracod fauna. The predominantly endemic molluscan faunas lower in the Pebas Formation are in disagree- ment with the proposed interior seaway environment by Räsänen et al. (1995) and Webb (1995), (e.g. Nuttall, 1990; Wesselingh, 1993; Whatley et al., 1998; Muñoz Torres et al., 1998; Vonhof et al., 1998; Wesselingh et al., 2002; Vonhof et al., 2003, Chapters 4 and 5). Based on strontium, carbon and oxygen isotopes, as well as on fauna composition, Vonhof et al. (1998, 2003) concluded that the mollusc-bearing part of the Pebas Formation had been deposited in a lacustrine system. Andes run-off was the dominant water source in Miocene Western Amazonia, though there was occasional influx of waters from cratonic catchments and one stratigraphic level with increased salinity (i.e. Buenos Aires, a small outcrop on the banks of Cotuhe River, Colombia). This level belongs to Hoorn’s Grimsdalea Zone (Hoorn, 1994b). A Caribbean (northerly) origin of the incursions as proposed by Nuttall (1990) and Hoorn (1994b; see also Hoorn et al., 1995) was supported by these faunal (see also Wesselingh et al., 2006) and isotope data. Compilations of Pebas molluscan taxa autecological signatures, such as provided by Nuttall (1990); Wesselingh et al. (2002) and Vonhof et al. (1998) show an admixture of (inferred) ecological preferences. Almost all species from the Pebas Formation are extinct, and several of the genera dominating the fauna are extinct or have survived as relicts only (i.e. manatees, dolphins, stingrays, molluscs, see also Wesselingh, 2000), limiting their actualistic ecological use. This has complicated the insight into the environmental system in which these faunas lived. Were they endemic during the Miocene or living in seawater? The ichnofauna might have been endemic as well. From the mollusc-bearing part of the Pebas Formation it can be concluded that occasionally marine incursions reached the area, which is also reflected in (pollen) floras. A stratigraphic database of Western Amazonian molluscan strontium isotope ratios is constructed by (Vonhof et al., 1998, 2003). By analyzing 87Sr/86Sr ratios of molluscan carbonate one can reconstruct the 87Sr/86Sr composition of the waters in which the molluscs lived (Figure 1.15). Since seawater and freshwater 87Sr/86Sr ratios are distinct, these Sr-isotope data can, with some assumptions, be used as a paleo-salinometer. The molluscan isotopic signatures suggest fresh water environments in Miocene Western Amazonia predominantly derived from Andean run-off, with occasional peri-marine incursions that result in a maximal 3 PSU (practical salinity unit) salinity. 16 Introduction and summary

0.7092 Buenos Aires 5.0 psu salinity line (cratonic) 0.7090 a 18 psu salinity line (cratonic) oligohaline water 0.7088 mesohaline and polyhaline water

S 0.7086

S N. Horiz 35 psu seawater

7 0.7084 8 Porvenir and Los fresh water 5.0 psu salinity line 0.7082 Chorros

0.7080 Pebas 0.5 psu salini and Indi- ana ty line 0.7078 Sta Rosa

0.0 5.0 10.0 15.0 20.0 25.0 30.0 age (Ma) 0.7092

b 0.7090 mesohaline and polyhaline water 0.7088 oligohaline water

S 0.7086

S N. Horiz 35 psu seawater

7 0.7084 8 Porvenir and Los 5.0 psu salinity l 0.7082 Chorros 0.5 psu salini fresh water ty line ine 0.7080 Pebas and Indi- ana 0.7078 Sta Rosa

0.0 5.0 10.0 15.0 20.0 25.0 30.0 age (Ma)

Figure 1.15: Molluscan 87Sr/86Sr ranges of Pebas Formation outcrops plotted over their full palynostratigraphic range (from Vonhof et al., 2003). Superimposed are the paleosalinity fields for this part of the Neogene calculated on the basis of a marine 87Sr/86Sr reference curve (black line; data from Howarth and McArthur (1997)) and three different Andean freshwater end member compositions. (a) Andean freshwater end member at a S87Sr/86Sr ratio of 0.7074 and a Sr concentration of 0.0416μmol/L. Two short broken lines around the data from the Buenos Aires outcrop group indicate salinity fields in a seawater cratonic freshwater mixture, which only applies to this specific outcrop (see text). (b) Similar to "a", but calculated with Andean freshwater composition equal to that of the modern Napo River. (87Sr/86Sr = 0.7059 at 0.0416μmol/L Sr; Palmer and Edmond, 1992). 1.6 Isotope records from accretionary growth structures 17

Sr isotope analyses show that during the deposition of the Miocene sediments in the area of the towns of Pebas and Santa Rosa, (Crassoretitriletes Zone) the Pebas-environment was divided in several compartments. The relative stability of 87Sr/86Sr ratios in the Grimsdalea Zone (Vonhof et al., 1998, 2003) points toward a rather large interconnected body of water and thus supports a geologically long-lived lake, "Lago Pebas", as was already suggested by the composition of molluscan assemblages of Wesselingh et al. (2002). At ∼8 Ma BP the Amazon River established its modern (easterly) course into the At- lantic Ocean, drastically changing the watercourses and paleogeography of the Amazon Basin. Although this event marked the end of widespread lacustrine environments in western Amazonia, it allowed for the establishment of terrestrial biota, eventually leading to the modern species-rich Amazonian rainforests.

1.6 Isotope records from accretionary growth structures

Stable isotope records obtained from inorganically and organically precipitated carbonates have been widely used for paleo-climatic and paleo-environmental reconstructions. I used stable isotope records of molluscs.

1.6.1 Isotope records from molluscs Isotope records from accretionairy growth structures from molluscs provide detailed information in small windows of time such as well-preserved seasonal variation in δ 18Oor temperature (Dutton et al., 2002). Although some molluscs, like members of the pearl mussel family Margaritiferidae, (such as Margaritifera margaritifera (Linnaeus, 1758)) are reported to live more than 100 years (Comfort, 1957) and accrete at low rates, other freshwater mussels accrete CaCO3 at high rates. Anodontites trapesialis (Chapters 2 and 3) grows 30Ð 600 μm/day, in comparison with, for instance, speleothems (1Ð500 μm/year, Treble et al., 2003; Vaksa et al., 2003). Consequently bivalves can be adequate high-resolution recorders of rapid change in water chemistry, especially in their juvenile stage when growth rates are highest (Chapter 4). High variation in host water chemistry (δ 18O), which is reﬂecting changes in the composition and amount of precipitation, makes fresh water molluscs not the most suitable paleo- thermometers, but the variation in host water chemistry itself is well documented in the shells and not overprinted by temperature variation (Figure 1.6) as would be the case at higher latitudes (Dettman et al., 1999, Dutton et al, 2002, Ricken et al., 2003).

1.6.2 MicroMill Just like trees the shells of molluscs (in our case bivalves) have an accretionary growth structure: increments between growth lines. In order to observe the chemical change in the shell in time, these increments (or growth bands) are to be sampled. First the molluscs are put in an artificial resin, then they can be cut into thin sections without breaking them. All 18 Introduction and summary increments can be observed stacked on top of each other. The thin sections are put under a microscope with an attached automated dental drill: the MicroMill TM. This is a very accurate apparatus with which lines ∼ 20 μm apart can be milled. As result a time-resolution per sample of ∼ 2 weeks has been acquired in the adult portion of our Recent Playa Cocha molluscs (controlled part during our experiment) and a daily resolution in the juvenile part of A. trapesialis. This time resolution is still insufficient to reflect diurnal temperature variation. The water body of Playa Cocha varied widely between the dry and wet seasons during our experiment (1998-1999) and a shift in temperature of 3 ◦C occurred between the dry and the wet season. The higher temperatures in the dry season and lower temperatures in the wet season have dampened the amplitude of the annual 8 δ 18O variation by 1. More stable and larger water bodies like the Amazon River or ancient Lago Pebas are expected to have (had) less annual temperature variation.

Figure 1.16: The Mercantek MicroMill, which is capable of milling at high precision. The microsampler has a ﬁxed dental drill and a computer-controlled micro-positioning stage. A video system supplies a real-time image of the thin section on the computer screen. Software allows complex drill paths to be drawn, and subsequently drilled at variable depths for microsampling.

1.6.3 Mass spectrometry Water samples were analyzed on a Finnigan Delta+ mass spectrometer with a GASBENCH 18 13 preparation device, at a 1SD precision of ∼ 0.15 for δ OW and 0.1 for δ CDIC. 18 13 δ OW us reported in V-SMOW (Vienna Standard Mean Ocean Water), δ CDIC in V- PDB (the Vienna-Peedee belemnite standard). For stable isotopes measurements in carbonates a MAT 252 isotope ratio mass spectrometer (Figure 1.18) was used. Minor elements analyses were performed on a Varian Liberty ICP-AES (inductively coupled plasma-atomic-emission spectrometer) after digestion of ∼ 1.5 mg of powdered 1.6 Isotope records from accretionary growth structures 19

100 µm drill

sample

Figure 1.17: Although the dental drill has a thickness ∼ 100 μm, a 20 μm line spacing can be milled.

Figure 1.18: The MAT 252 is equipped with an automated carbonate extraction line (Kiel II de- 18 13 vice) and tuned to analyze < 10 μg of CaCO3 for δ Oand δ C.Samples were digested in concen- trated orthophosphoric acid at 80◦C. A routinely analysed carbonate standard was reproducible within 0.09 for δ18O and 0.05 for δ13C(1σ) Both δ18O and δ13C are reported relative to V-PDB (Vienna-Peedee belemnite).

sample in 1 N HNO3 and subsequent dilution to an ∼0.1 N HNO3 sample solution. Con- centrations of Sr, Ba and Fe are reported in parts per million (ppm). Conservative ICP-AES detection limits are set at 40 ppm for Fe. A standard test experiment indicated a reproducibility (1 σ) of better than 5% for the method used. 20 Introduction and summary

1.7 Amazonian climate

1.7.1 Present-day climate The monsoonal control of the regional rainfall pattern due to movement of the ITCZ is expressed in the seasonal change in the isotopic composition of δ 18O in precipitation (Figure 1.19), which on its turn changes the isotopic composition of run-off water and molluscan shells. In Chapters 2 and 3 the results of a one year experiment with living shells are discussed (Figure 1.20a). Every two weeks water samples were taken for isotope measurements on the Finnigan Delta. This experiment has been conducted to see to what extend the isotope signatures in molluscs are comparable with that of their environment.

18 δ Op,dry = -10.5 ‰ 18 δ Op,wet = -33.7 ‰

18 δ Op,dry = -0.38 ‰ δ18O = -19.6 ‰ p,wet 18 δ Op = +0.7 ‰

Andes Mt range

Amazon Basin Atlantic (Iquitos, Peru, 100m a.s.l.) Ocean

Figure 1.19: Seasonal change in oxygen isotope composition of precipitation in Amazonia, modiﬁed after Grootes et al., (1989). The isotopic composition in the wet season is much lower than in the dry season. This phenomenon is reﬂected in the composition of Amazonian surface water and on its turn in the molluscs living in it.

The results of this experiment show that the seasonal isotopic change in the ambient water is reﬂected in the growth increments of the molluscs. With the results from the experiment taken into account the isotope signals in fossil molluscan shells are interpreted in Chapters 4 and 5.

1.7.2 Low latitudes As is demonstrated by Savin (1977) and Martínez and del Río (2002) the sea surface temperature in the low latitudes has changed only a few degrees since the Miocene. This knowl- 1.7 Amazonian climate 21 edge paves the way for reconstructions of past hydrological cyclicity. If annual temperature variation is limited, the CaCO3 in bivalves can be regarded as an adequate recorder of oxygen isotope change in its host water, which in its turn reflects oxygen isotope variation in precipitation/evaporation. The amplitude in seasonal cycles that is preserved in bivalve growth increments is not only similar to the variation in rainwater chemistry. Depending on the residence time, i.e. the through-flow of river water, ground water and the size of the watershed, the amplitude of seasonality in rainwater is dampened (Leopoldo et al., 1992). The longer the residence time of water in a watershed, the lower the amplitude in host water δ18O variation will be in comparison with the amplitude of rainwater δ 18O variation. The size of a watershed and the residence time of water in the water body are also influencing the apparent duration of wet and dry seasons. Defining these seasons by the amount of rainfall is not the same as those defined by the water level of the water body. In Playa Cocha for instance the isotopic composition was related to the water level, which was dependent on the presence or absence of Amazon River water (Chapter 2). The sudden inundation of the floodplain lake by the Amazon was not related to an increase in local rainfall. Retreat of the Amazon River due to a decrease in its water level controlled the onset of evapora- tive regimes in the cut-off floodplain lake. The local rainy season had already terminated a month earlier. A good example of this phenomenon is the record from Triplodon corrugatus that we collected from Itaya River (Figure 1.20b). This small black-water tributary of the Amazon River is ∼ 100 km in length and has a relatively small watershed that consists of purely lower tropical rainforest. Because of the short residence time, in comparison with the mainstream of the Amazon River, the duration of the wet season, as is recorded in the growth increments of bivalve shell, is much shorter than the dry season. Unfortunately, shell accre- tion by bivalves is not performed at a constant rate. From higher latitudes it is known that cessation of shell growth occurs at low water temperatures. Thermally controlled shut down is reported by Dettman et al. (1999); Dutton et al. (2002) and references therein; Ricken et al., (2003) and Verdegaal et al. (2005). Decreased growth rates are suggested to be a result of seasonal bio-rhythms such as spawning Jones et al. (1983) and nutrient limitation (Dut- ton et al., 2002 and references therein). A. trapesialis in Playa Cocha (Chapter 2) showed reduced growth rates, but no shut down of growth during wet season. If timing of growth rates is lacking, as is the case in studies on fossil shells, the duration of the wet season as reflected by host water may appear too short. Because the fossil molluscs from the Pebas Formation are extinct, it is impossible to deduct any variation in the growth rates during the year, or to reconstruct the duration of wet and dry seasons. Assuming a single dry and wet seasonal cycle annually, it is possible to make reliable assumptions on the duration of the lifespan of bivalves. Even in cases where the amplitude of the oxygen isotope cycles is reduced to nil, pronounced δ 13C cycles in the shells can be of help (Figure 5.11c, on page 68). 22 Introduction and summary

1.7.3 Amazonian Miocene climate

Owing to efﬁcient rainwater recycling, wet-season rainwater in western Amazonia has ex- ceptionally negative δ18O values that are reﬂected in the low oxygen isotope ratios (δ 18O) of living freshwater bivalves (Chapters 2 and 3).

Napo River Bellavista Diplodon aff. longulus, ~16 Ma

Nanay River A m azon R Mazán RiverCollection site IquitosIquitos iver Mazán Indiana Playa Cocha experiment site Airport

Amazon River Triplodon corrugatus collected alive Sept. 1998 a Iquitos b Collection site at Itaya River 0 10 km

Figure 1.20: (a) Location of the Playa Cocha experiment site; an Amazon River ﬂoodplain lake in the vicinity of the city of Iquitos. The runway of Iquitos’ airport shown in the photo is 2.5 km long. (b) Location of the Mazán and Itaya rivers collection sites (Chapter 4). Three fossil Diplodon aff. longulus were collected at the conﬂuence of Mazán and Napo rivers, several living specimens of Triplodon corrugatus were collected for comparison.

Using the results of our experiment with living molluscs (Paragraph 1.7.1), past Ama- zonian seasonality (Chapter 4) and aquatic environments (Chapter 5) were reconstructed through the analysis of fossil molluscs. A key target for the reconstructions is the stability of the climatic conditions during Miocene uplift phases of the Andes mountain range. Comparisons of modern and fossil bivalve δ 18O data show many similarities, both in absolute values as well as in the seasonal variation recorded in growth increments (Chapter 4, the fossils used in this Chapter were dated at ∼ 16 Ma, at the time of the Miocene Climate Op- timum, MCO). This documents that the Amazonian hydrological cycle, transporting humid air from the Atlantic to Amazonia, was already present 16 million years ago, including the wet and dry seasons as we know them today (Chapter 2). Consequently, the shielding effect of the Andes mountain range, which is of importance for this climate system, was already in place. However, the contribution of high-altitude precipitation to natural waters in our study area is limited, because rainfall intensity is highest in the Amazonian lowlands, we expect that a possible lower Andean Mountain Range in the Miocene (Steinmann et al., 1999) might not have had much inﬂuence on the isotopic composition of waters in our study area. The extremely low isotope values in parts of the molluscs suggest that the Andes during the Middle Miocene acted as a barrier to atmospheric vapour. 1.8 Main conclusions of this work 23

In summary modern and fossil δ 18O isotope data in molluscs show that:

¥ Modern and Miocene Amazonian conditions are strikingly similar: cyclicity and absolute values of δ18O records are the same. For modern as well as Miocene waters, the amount of precipitation must have been the controlling parameter for seasonal isotopic variation. We conclude that seasonal migration of the ITCZ and the intensity of rainfall in the MCO are comparable to today. Possibly higher temperatures during the MCO in our study area are irresolvable in the molluscan δ 18O records.

¥ Wider implications are that the Amazonian hydrological cycle will most likely not be highly affected by globally warmer climates than today considering potential warming of Earth’s climate in future.

¥ The extent of Western Amazonian rainforest in the Middle Miocene was not as large as today, since most of the area was occupied by lakes and wetlands (Nutall, 1990; Wesselingh et al., 2002). Our study shows that climatic conditions required to sustain highly diverse tropical forests and várzea environments also existed during the Miocene Climate Optimum.

1.8 Main conclusions of this work

1.8.1 The Miocene climate of Amazonia

In summary modern and fossil δ 18O isotope data in molluscs show that:

1.8.2 Ecological implications of molluscan stable geochemical records Wesselingh et al. (2002) concluded, based on molluscan data and a review of other data, that the Miocene sediments in western Amazonia, known as the Solimões Formation in Brazil and as the Pebas Formation in northeastern Peru (Chapter 4), were deposited in a large long-lived lake system and associated swamps. This lake system was occasionally reached by diluted marine incursions. Because the Miocene molluscan fauna is endemic and extinct we have assessed the ecological implications of the stable isotope data of both oxygen and carbon from bivalves from the Pebas formation. Oxygen isotope signals in riverine host waters show more seasonality than lacustrine waters (Chapter 5). The strontium isotope analyses performed by (Vonhof et al., 2003) indicate that these swamps and lakes were interconnected with at times a compart- mentalized basin. The Sr isotope records of outcrops in eastern Peru representing part of the Crassoretitriletes Zone of Hoorn (1993) with a duration of ca. 1 Myr cannot be correlated to the west Brazilian borehole data of the same time interval. Yet, outcrops representing earlier and later time intervals can be correlated with each other, suggesting temporal subdivision of the Pebas in some (several) small subbasins (Vonhof et al., 2003). In summary (conclusions of Chapter 5): Stable isotope profiles of Miocene bivalves from the Pebas Formation of western Amazonia provide considerable information about their biotope preferences and growth regimes and yield wider implications for the nature of the Pebas ecosystem as a whole. The new data accord with those discussed in Chap- ter 4 thus support a monsoon system throughout the Middle and Late Miocene. Different δ18O amplitudes between the specimens studied are correlated with their different biotopes, as interpreted from faunal assemblages. Pachydontine bivalve isotope profiles, characterized by low amplitudes and irregular or absent cyclicity, are interpreted as representing lacustrine living modes. These shells were accompanied by very low numbers of pearly freshwater mussels, which dominated fluvial-influenced environments in the Pebassystem, where Pachydontinae were scarce or absent. These unionoids yield isotope signals comparable to those of modern Amazonian floodplain unionoids, characterized by relatively high amplitude and regular seasonal variation.

¥ Lower seasonal δ18O amplitude for lake dwellers compared with river dwellers is interpreted as caused by the longer residence time of lake water, which buffered seasonal δ18O of rainfall in the area. Therefore, molluscs from lakes are less suitable than molluscs from rivers for accurate reconstructions of the seasonal variation of precipitation and evaporation.

¥ Intra-annual growth rate variation in fossil bivalves could not be assessed because the relative length of the wet and dry season during the Miocene could not be established independently. The studied fossil pearly freshwater mussels would have reached ages of 6-10 or more years. An age determination for Pachydon is more complicated because of the low seasonal variation in the skeletal chemical record. An expected long lifespan, based on growth line counts from Pachydontinae (4-15 years), would be conﬁrmed if the variation in [Sr] and δ 13C is regarded as a seasonal chemical 1.8 Main conclusions of this work 25

variation. ¥ Surface waters were not in carbon isotopic equilibrium with the atmosphere. The δ13C records of all measured Pebasian bivalves have minima below -10 (similar to modern Amazonian freshwater shells), which argues for a significant terrestrial carbon input in the system and simultaneously against seawater influence. ¥ Although only limited molluscan trace element data have been available until now, considerable seasonal variation in trace element concentrations is evident. The variation observed is tentatively interpreted as reflecting seasonal change in host water chemistry. No clear indications were found for any kinetic control on molluscan trace element distribution. Seasonally variable trace element chemistry of the host water inhibits the use of trace elements ratios as paleothermometry tools (cf. Klein et al., 1996). ¥ Isotopic signals are frequently used in climate reconstructions of the Miocene record (Savin et al., 1975; Zachos et al., 2001). On their basis, the Middle-Late Miocene (16.5-12.5 Ma) global climate is believed to represent a cooling trend that culminated in the Pleistocene ice ages. The isotopic signal of the mollusc shells from Miocene Amazonia does not confirm or deny this scenario, probably due to the relative insensitivity of the tropical latitudes to global climate change. ¥ Although orbitally forced variation in ITCZ migration is likely to have taken place in the Middle- Late Miocene, the near-equator position of our sampling sites prevents us from confirming that this variation affected the amplitude of molluscan δ 18O profiles. Furthermore, our data set does not provide clear evidence that major changes in seasonality took place in the Middle-Late Miocene. However, we can conclude that a climatic system with comparable seasonality to that of today was already in place during the late Early Miocene, which is particularly important because it enhances our knowledge about the climatic parameters that ruled Amazonia during the Miocene and confirms that, at the time, the climatic conditions needed for a tropical rainforest were in place.

1.8.3 Aquatic landscapes in the Miocene of western Amazonia The sequences found at the outcrop near Santa Rosa de Pichana might either point toward a repeatedly prograding delta front, base level variations or switching coastal lobes in a subsiding setting. The conclusions from Wesselingh et al. (2006b) (Chapter 6):

¥ A common type of succession of transgressive, highstand and regressive/ prograding intervals is found in depositional sequences of the non-marine Miocene Pebas Forma- tion of western Amazonia. Mollusc faunas are almost entirely limited to the transgressive and highstand intervals. In the transgressive to highstand intervals a succession of assemblages is found, from the Thiaridae/Pulmonata assemblage, through the Try- onia and Tall-Dyris assemblages in the early transgressive intervals and Small-Dyris 26 Introduction and summary

assemblage in the late transgressive intervals, to the Pachydon obliquus assemblage in the highstand interval. Reworking during the early transgressive phase was common. With increasing water depth, the preservation of the molluscs improved. Diversity increased to its zenith in the Small-Dyris assemblage and then decreased towards the Pachydon obliquus assemblage. ¥ Dissolved oxygen levels in the water probably played a major role in the determination of fauna successions, with dysoxia becoming important during highstand intervals and dominant in the early regressive/prograding intervals. Molluscs and stable isotope proﬁles show no indications of elevated salinities, not even on seasonal time scales. Thalassinoid ichnofossils of the Glossifungites ichnofacies at the base of sequences may represent up to lower mesohaline salinities, but also may reﬂect evolution of lowered salinity tolerances for the constructors (crustaceans and possibly polychaetes). ¥ The Santa Rosa de Pichana section appears representative for large parts of the Pebas Formation. Based on molluscan and stable isotope evidence we think that the system consisted of predominantly freshwater swamp to lacustrine conditions in a long-lived lake/wetland system at sea level and open to marginal marine settings. However, discrepancies with ichnofossil data from the same stratigraphic intervals, which consistently indicate elevated salinities, are in need of further study.

Outlook

Lately, there is a lot of interest in the timing of the uplift of the Andes and the influence of the Andes on Amazon climate. Because ice-cores are not old enough and sofar the dating of old (Eocene-Miocene) speleothems is not yet possible, isotope studies on molluscs might give more insight in this subject. Isotope records from fossils closer tho the east flanks of the Andean Cordilleras and further away from the equator (towards Bolivia and Colombia) might give some answers on when the tropical seasonality that is governing the Amazon region was established and the influence of the Andean Cordilleras on it. Not only mollsucs would be suitable for answering this questions; the isotope records that can be obtained from increments in the enamel of teeth could be helpful as well. Chapter 2

Seasonal stable isotope variations of the modern Amazonian fresh water bivalve Anodontites trapesialis

Abstract

In a ﬂoodplain lake of the Amazon River near the city of Iquitos, north-eastern Peru, a one-year monitoring experiment was conducted, during which water samples and living bivalves (Anodontites trapesialis) were collected with the aim to investigate seasonal δ18O variation in and fractionation between bivalve aragonite and host water. Both host water and molluscan growth increments show more than 8 seasonal variation in δ18O. In the ﬂoodplain lake under study the δ18O variation of the water is controlled by contrasting dry and wet season evaporation - precipitation regimes. Molluscan δ18O appears to be in equilibrium with host water. Although an approximately 4.0 offset occurs, δ13C records of water and bivalves are in good agreement, suggesting that both δ18O and δ13C of the shells of fresh water bivalve Anodontites trapesialis are good recorders of (paleo-) environmental conditions. δ13C of Dissolved Inorganic Carbon (DIC) is governed by plant growth and/or by changes in aquatic chemistry, affecting the DIC pool. Keywords— Amazonas Peru; Bivalvia; fresh water; stable isotopes; evaporation; atmospheric precipitation

This Chapter is based on Based on Kaandorp, Vonhof, Del Busto, Wesselingh, Ganssen, Marmól, Romero Pittman, and van Hinte 2003. Seasonal stable isotope variations of the modern Amazonian fresh water bivalve Anodontites trapesialis. Palaeogeography, Palaeoclimatology, Palaeoecology 194 (4): 339-354. 28 Seasonal variation reﬂected in Amazonian freshwater bivalve A. trapesialis

2.1 Introduction

With approximately 6 000 000 km 2 the Amazon River has the world’s largest catchment area. The major part of the Amazon Basin belongs to the lower rainforest (80-600 m a.s.l.) that has little variation in topography (Kalliola and Puhakka, 1993). The Guyana Shield in the North, the Brazilian Shield in the South and the Andes mountain range in the West border the basin. The Amazon River drains the basin to the East into the Atlantic Ocean. The climatic records of the cities of Iquitos and Pucallpa in Peruvian Amazonia (Figure 73°15'W 73°05'W

Cuenca (Ec.) 10 km Amazon 3°05'S Amazon Iquitos

Río MarañónYurimaguas Río N Trujillo Pucallpa P E R U Ucayali Río Nanay Playa Cocha South LIMA Pacific Cusco 3°45'S Ocean Iquitos

Arequipa Peru 500 km

ío Itaya R

Figure 2.1: Location map of Playa Cocha, the sampling site, a ﬂoodplain lake near the Amazon River, 5 km east of Iquitos.

2.1) show a yearly average temperature of 26 ◦C and relatively humid conditions (80-90%).

Pronounced seasonal variation in rainfall leads to the formation of two distinct landscapes in Western Amazonia: i) terra firme, land that is never flooded, and ii) floodplains that are permanently or temporary inundated because of seasonal fluctuation in water level of main rivers and their tributaries. The floodplains cover around 50 000 to 60 000 km 2 and comprise swamps, lakes and seasonally inundated forests (e.g. Hoorn, 1994a). Near Iquitos the water level of the Amazon River can vary 10 m between wet and dry season (Kalliola and Puhakka, 1993). Wet season in Peruvian Amazonia starts in October and lasts until May, the remaining four months are known as dry season. Flooding of flood plain lakes (locally called "Cochas") may occur some time after the initiation of the wet season. Flood- plain lakes are separated from the main river channel during the dry season when the water level of the main rivers drops. During dry season (June until September), when floodplain lakes are disconnected from rivers, their main water sources are direct rainfall and creeks that drain the local basin or terra firme forest. 2.2 Experiment, material and methods 29

In general, the δ18O of rainwater is related to environmental parameters such as latitude, altitude, distance to coast, surface air temperature and the amount of precipitation (Dansgaard, 1964). In the Amazonian hydrological cycle, water vapour is transported from its Atlantic source by prevailing easterly winds towards the Andes mountain range, which blocks Paciﬁc water vapour from entering Amazonia (Salati and Vose, 1984). There is a large difference in the δ 18O values of atmospheric water vapour and rainwater between the wet and dry season, with high values in the dry season and low values in the wet season. According to Grootes et al. (1989) and Grootes (1993) rainfall exceeds evapotranspiration 2-3 times in the wet season resulting in a westward depletion of δ 18O in rainwater. Dry season precipitation is recycled by evapotranspiration and therefore not depleted; its values stay close to those of the source (Atlantic ocean water). However, Rozanski et al. (1993) conclude (based on ﬁve selected IAEA/WMO global network stations) that west of Manaus (Brazil) the continental gradient of δ 18O in rainwater is very small or absent in the wet season. In their work very low values of δ 18O are mostly attributed to the amount effect.

Fresh water unionids often form discrete growth increments in their aragonitic shell (Figure 2.2), which can be recognized in cross sections and sampled at high temporal resolution. Dettman and Lohmann (1993), Abell et al. (1995), Jones and Quitmyer (1996), Veinnot and Cornett (1998), Dettman et al. (1999) and Wurster and Patterson (2001) showed that seasonal environmental variation is adequately recorded in the isotopic composition of mollusc shells. On the other hand (Fastovsky et al., 1993) argued that some freshwater bivalves precipitate aragonite in disequilibrium with their host water. This was refuted by Dettman et al. (1999) on the argument that Fastovsky and coworkers compared shell δ18O values to environmental conditions that occurred later than the time of shell growth. In the present study we monitored a population of Anodontites trapesialis over a 13-month period. These are unionid bivalves that are common in Amazonian ﬂoodplain lakes. To determine whether growth increments of the aragonite shell of Anodontites trapesialis were precipitated in equilibrium with its host water, we analyzed the isotopic composition of 18 13 18 13 both aragonite (δ Oar and δ Car) and host water (δ OW and δ CDIC). With this experiment we aim to test the applicability of bivalve growth incremental stable isotope proﬁles for the reconstruction of past Amazonian aquatic environments.

2.2 Experiment, material and methods

2.2.1 Monitoring project The studied floodplain lake, locally named "Playa Cocha" lies in an inner bend of the Ama- zon River, at 3◦45.536’S, 73◦10.791’W, 5 km southeast of the city of Iquitos, in Peruvian Amazonia (Figure 2.1). We monitored a population of living molluscan bivalves (Anodon- tites trapesialis (Lamarck, 1819)) during a complete dry-wet season cycle (26-09-1998 until 21-10-1999) in which lake water level changed 5 meters. At the beginning of the experi- 30 Seasonal variation reflected in Amazonian freshwater bivalve A. trapesialis ment the size of 24 specimens was measured. They were marked (Figure 2.2) and put back in their natural environment to be able to monitor growth rates at final collection. One marked specimen was collected on 15-11-1998 (specimen 981115) another on 31-01-1999 (specimen 990131) after which collection had to be discontinued as Amazon River water in the floodplain lake rose higher than 3.5m. In July, when water level had dropped again, all other marked specimens of A. trapesialis had disappeared. They might have floated away, as freshwater mussels in the Mississippi sometimes do (Cvancara, 2000). On 21-10-1999, a small number of unmarked living specimens of A. trapesialis were collected in Playa Cocha, of which one was used for this study (specimen 991021). For the duration of the experiment bottom- and surface water temperatures were measured by hand every two weeks; bottom water temperatures were also measured at 15-minute intervals with a SEAMON mini temperature recorder (until 02-05-1999). The pH was measured using pH paper (0.2 accu- 18 racy); filtered water samples (0.45 μm mesh) were taken for isotope analyses (δ OW and

marked 26-09-1998

collected 31-01-1999 } 1 cm

b 1 cm

Figure 2.2: (a) Marked specimen of Anodontites trapesialis, specimen 990131, collected 31-01-99. The mark was scratched in the periostracum of the shell on 26-09-98. Between marking and collection it had grown 8mm along the axis of maximum accumulation (marked with ’}’). (b) Line drawing of a thinsection along the black line in a; this thin section was cut for later microsampling. The lines show sampling trajectory. 2.3 Results 31

13 δ CDIC Table 1). Analytical method To acquire high-resolution stable isotope curves of successive growth increments of A. trapesialis we used a Merchantek MicroMill TM Sam- pler (Dettman and Lohmann, 1995) on which thin sections of about 300 μm thickness were mounted. These thin sections were cut perpendicular to growth lines along the axis of maximum carbonate accumulation (Figure 2.2). The microsampler has a ﬁxed dental drill and a computer-controlled micro-positioning stage. A video system supplies a real-time image of the thin section on the computer screen. Software allows complex drill paths to be drawn, and subsequently drilled at variable depths for microsampling. Lines were drilled parallel to molluscan growth banding, with different line spacing for the individual molluscs (80, 130 and 200 μm). The aragonite powders recovered from each line were used to obtain 18 13 stable isotope data (δ Oar and δ Car). Stable isotope analyses of shell aragonite were performed at the Vrije Universiteit Amsterdam on a Finnigan MAT 252 mass spectrometer equipped with an automated preparation line (Kiel II type). The reproducibility of a routinely analyzed carbonate standard (NBS 19) is better than 0.09 for both 13C and 18O (1 SD). Results of the isotope measurements on A. trapesialis specimens are printed in the Appendix starting on page 121)

Water samples were analyzed at the Alfred Wegener Institut Bremerhaven and at the Free University of Amsterdam on Finnigan Delta+ mass spectrometers with GAS- 18 BENCH II preparation device, at a 1SD precision of ∼ 0.15 for δ OW and 0.1 for 13 18 18 δ CDIC. δ OW is reported in V-SMOW (Vienna Standard Mean Ocean Water), δ Oar, 13 13 δ Car and δ CDIC in V-PDB (Vienna PeeDee Belemnite). Shell material will be ﬁled at Naturalis National Museum of Natural History Leiden.

2.3 Results

2.3.1 Playa Cocha water In situ measured water level and bottom water temperature in Playa Cocha varied 5 meters and 6◦C respectively during the monitoring period (Figure 2.3). The lowest water level measured in the lake was 45 cm (15-11-1998) when its size was reduced to about 200 m in length at 5 to 10 m width. Amazon River water entered Playa Cocha on 26-01-1999. The highest water level was recorded on 08-05-1999 at 547 cm. At the beginning of July the ﬂoodplain lake was disconnected again from the river, after which water level gradually dropped to 80 cm. During the monitoring period, bottom water temperatures ranged from 25.0 ◦C to 31.0◦C, with lowest temperatures during high water level and highest temperatures during low water level. From 26-09-1998 to 02-05- 1999 bottom water temperatures were also monitored with a SEAMON-mini temperature recorder at 15-minute intervals. At 20-10-1999 when the recorder was retrieved it was found 15 cm deep in the sediment. Daily variations up to 3 ◦C were recorded on normal days in the dry season (with an observed maximum variation of 5.3 ◦C after a heavy rain 32 Seasonal variation reﬂected in Amazonian freshwater bivalve A. trapesialis

13 18 ◦ ◦ ◦ Date Day Depth δ CDIC δ OW pH pH T( C) T ( C) T ( C) (day-mo-yr) (m) (PDB) (SMOW) Surface Bottom Surface Bottom Air 26-09-98 1 0.50 -1.84 0.5 6.3 6.1 28 13-10-98 17 0.50 Ð5.36 2.1 6.3 6.1 31.2 02-11-98 36 0.70 Ð4.79 Ð1.1 27.8 15-11-98 49 0.45 Ð5.06 Ð1.1 29.3 15-12-98 79 0.54 Ð4.51 Ð0.3 6.4 6.4 29 29 32 31-01-99 126 2.64 Ð12.18 Ð7.1 6.1 5.8 27.5 27.2 15-02-99 141 3.94 Ð13.32 Ð7.3 6.4 5.8 29 26.4 29 02-03-99 156 4.29 Ð14.01 Ð8.4 6.3 6.1 28 26.7 28.5 16-03-99 170 4.14 Ð13.88 Ð8.6 6.4 6.1 28 27 28.5 31-03-99 185 4.24 Ð14.48 Ð8.6 6.4 6.1 27 27 29 16-04-99 202 4.73 Ð14.41 Ð9.0 6.1 6.1 25 25 25 08-05-99 223 5.47 Ð15.14 Ð8.8 6.1 5.8 26.4 26.3 31.4 24-05-99 239 4.84 Ð15.32 Ð9.3 6.1 5.8 25.5 25.3 26 08-06-99 254 3.93 Ð15.16 Ð8.7 6.1 5.8 27.5 26 28.5 26-06-99 272 3.36 Ð14.69 Ð8.0 6.1 5.8 28.2 26.9 29.8 14-07-99 290 1.80 Ð13.72 Ð6.5 6.4 6.1 27.4 26.1 25.5 31-07-99 307 1.15 Ð11.68 Ð4.6 6.4 6.4 26.9 28.4 31.8 15-08-99 322 1.16 Ð10.16 Ð3.5 6.4 6.4 26.8 26.6 22.5 02-09-99 339 1.10 Ð7.67 Ð1.7 6.4 6.4 32.2 26.9 32.8 16-09-99 354 1.00 Ð5.88 Ð0.5 6.4 6.4 29 29 27 01-10-99 369 0.80 Ð5.60 0.1 6.4 6.4 33 31 33.5 15-10-99 383 0.80 Ð7.43 Ð0.2 6.4 6.4 29 28.4 3 20-10-99 388 0.75 Ð4.50 0.1 6.4 6.4 31.2 29 3 Rainwater 19-09-98 Ð1.2 5.2 15-05-99 Ð19.6 5.2 26-06-99 Ð5.8 5.2 27-07-99 Ð0.4 5.2 02-09-99 Ð1.8 5.2 20-09-99 Ð1.6 5.2 Amazon River 20-10-99 20 Ð13.51 Ð7.1 20-10-99 4 Ð11.41 Ð7.2 20-10-99 0 Ð13.64 Ð7.1

Table 2.1: Water data of Playa Cocha shower), and less than 0.5◦C in the wet season. Hand-measured temperatures on sampling days (around noon) were within 1 ◦C of diurnal mean temperatures. During the monitoring 18 period the oxygen isotope composition (δ OW ) varied from +2.1 to Ð9.3 [V-SMOW]. 13 18 δ CDIC, which ranged from Ð1.0 to Ð15.3 [V-PDB] co-varied with δ OW (Table 18 13 1, Figure 2.3b). Lowest δ OW and δ CDIC values clearly correspond with high water 18 13 levels in Playa Cocha; during low water levels δ OW and δ CDIC values were highest. Rainwater samples from Iquitos ranged from -0.4 to -19.6 [V-SMOW] and can be found in Table 1.

2.3.2 Isotope composition of aragonite shells Three specimens of A. trapesialis were selected for microsampling and subsequent analyses 18 13 of δ Oar and δ Car (Table 2.1). All were collected alive; the ﬁrst specimen, collection date 15-11-1998 at the end of the dry season (Figure 2.4a) and the second specimen, collected on 31-01-1999 at the very beginning of the wet season (Figure 2.4b), were marked on 2.3 Results 33

6 32 a dry seasonwet season dry season 31

5 30

4 28

temperature (°C) 27

3 26

2 24

23 temperature bottom water (°C) 1 water depth (m) 22

21 isolated lakepart of the Amazon River isolated lake 0 20 6 -18 b water depth (m) -16

5 ‰ [V-PDB] -14 δ13 CDIC -12 4 -10

3 -8

-6

2 18 δ Owater -4

-2 1 water depth (m) 0

0 2 26-09-98 13-10-98 02-11-98 15-11-98 15-12-98 31-01-99 15-02-99 02-03-99 16-03-99 31-03-99 16-04-99 08-05-99 24-05-99 08-06-99 26-06-99 14-07-99 31-07-99 15-08-99 16-09-99 01-10-99 15-10-99 20-10-99

Figure 2.3: (a) Water depth and bottom water temperature of ﬂoodplain lake Playa Cocha in the period of 26-09-1998 Ð 21-10-1999. Collection dates are given in table 1. (b) isotope composition of ﬂoodplain lake Playa Cocha during the same period. δ18O values are in vs V-SMOW and δ13C values are in pe vs V-PDB. Higher water levels correlate with lower isotope values.

26-09-1998. Both specimens appear to have grown considerably in the short time interval between marking and eventual collection (Figure 2.2). The third specimen was collected on 21-10-99 in dry season at the end of our monitoring experiment (Figure 2.4c). It was unmarked, so that no other direct time marker than the date of collection is available. Within 18 the time frame described above, growth incremental variation of δ Oar was in phase with 18 seasonal δ OW variation in Playa Cocha. 34 Seasonal variation reﬂected in Amazonian freshwater bivalve A. trapesialis

a 2 collected 15-11-98 linespacing 0 ~ 130µm VII -2

-4

B] B] d18O -6 Marked I III 26-09-98

[V-PD -8 ‰ -10 d13 VI -12 C II IV -14 umbo ventral margin -16 120 100 80 60 40 20 0 sample numbers 2 collected 31-01-99 b linespacing 0 VII ~ 200µm -2 -4 B] B] d18O -6

[V-PD -8 I

‰ III Marked d13C -10 26-09-98 VI -12 IV -14 II umbo ventral margin -16 c 80 70 60sample numbers50 40 30 20 10 0 linespacing 2 ~ 80µm collected 21-10-99 0 VII -2

B] B] -4 D d18O

-6 V-P [ -8 I ‰ III d13 -10 C

-12 IV VI VIII -14 II umbo ventral margin -16 160 140 sample120 numbers100 80 60 40 20 0

Figure 2.4: Shell stable isotopes of three specimens of Anodontites trapesialis, collected on different dates. The juvenile parts of (b) and (c) are not completely drilled. Specimens (a) and (b) were sampled 5 m apart, (c) was sampled 50 m ’upstream’. Curves are aligned with carbon peak IV.).

13 The δ Car curves of the three specimens are correlated by assigning the numbers IÐVII to corresponding inﬂection points (Figure 2.4). This correlation implies that all three specimens are of the same generation. Number VIII, representing 1999 wet season, is shown in Figure 2.4c only, because both other specimens had been collected before that time. 18 13 δ Oar and δ Car co-vary, except for peak IV. Compared to the adult parts of the shells all three A. trapesialis specimens show relatively little variation in the isotope composition 18 of the juvenile part (left of number I in Figure 2.4), where δ Oar values range between −2 13 18 and Ð5 and δ Car between Ð6 and Ð8 . The longest juvenile δ Oar record is that of specimen 981115 (shown in Figure 2.4a), its relatively high values suggest that this specimen may have grown about 6 cm in a single dry season. Both other specimens that 2.4 Discussion 35

18 are not sampled completely towards the umbo, show a similar δ Oar pattern, suggesting that all are from the same generation of A. trapesialis (presumably from the dry season of 1996).

2.4 Discussion

We performed this study in order to evaluate the accuracy with which bivalve shells record seasonal/climatological signals in their shell chemistry. First, we consider the seasonal changes in oxygen and carbon isotopes of Playa Cocha water. Secondly, the accuracy of the A. trapesialis isotope signal in its shell is examined, in relation to possible fractionation processes. Finally, growth rate variations of A. trapesialis will be discussed.

2.4.1 Seasonal stable isotope variation of water As shown in Figure 2.3b the isotopic variation of Playa Cocha water correlates with the seasonal change in water depth. When the Amazon River flooded Playa Cocha in the wet season, the water composition became that of Amazon River water, characterized by low 18 13 values of δ OW and δ CDIC. After the wet season the water level dropped and eventually the floodplain lake was disconnected from the Amazon River. Once disconnected 16 18 from the river, selective evaporation of O isotopes resulted in an enrichment of OW . 18 Highest δ OW values were found in September/October, when Playa Cocha water level 18 had dropped to its lowest water depth (see table 1). The Amazon River showed a δ OW of 18 Ð7.4 in October 1999, while strongly evaporated Playa Cocha water δ OW was Ð0.1 . The amount of evaporation cannot be calculated, because rainwater input has not been monitored and rain showers are likely to affect a small water body like Playa Cocha during 13 18 dry season. Changes in δ CDIC in Playa Cocha co-varied with changes in δ OW . Playa 13 Cocha water had a δ CDIC value of Ð2 on 26 September 1998, when it was isolated from the Amazon River, and -15 in May 1999, when it was flooded. Growth of plants and algae in the dry season floodplain lake could be responsible for this dry season enrichment because plants preferably extract 12C from the Dissolved Inorganic Carbonate (DIC) pool (e.g. Cole et al., 1994). Degassing of CO2, resulting in shifting isotopic balances in the DIC pool and an increase in δ 13C, may have had an effect as well (e.g. Usdowski and Hoefs, 1990; Zhang et al., 1995; Luz et al., 1997; Barkan et al., 2001). The conditions under which this experiment was conducted do not allow for a quantification of these fractionation mechanisms.

2.4.2 Equilibrium precipitation of shell aragonite To investigate if A. trapesialis precipitate their aragonitic shells in equilibrium with host 18 13 water we compared δ Oar and δ Car with calculated host water isotopic equilibrium calciﬁcation values. Oxygen isotope fractionation is temperature dependent. Several temperature equations exist (e.g. Urey, 1947; Epstein et al., 1953; Emiliani, 1955) with which 36 Seasonal variation reﬂected in Amazonian freshwater bivalve A. trapesialis the oxygen isotopic composition of shells can be predicted, when the δ 18O of host water and ambient temperature are known. We have applied the relationship between fractionation factor and temperature of Dettman et al. (1999) that is based on data from Grossman and Ku (1986): 2.559 · 106 103lnα = +0.715 equation 2.4.2.1 T 2 K where TK is temperature in Kelvin and α is the fractionation between water and aragonite: 1000 + δ18 αaragonite = Oar equation 2.4.2.2 water 18 1000 + δ Owater

18 18 To convert δ Oar(PDB) to δ Oar(V −SMOW) we used the equation of Gonﬁantini et al. (1995):

18 18 δ Oar(V −SMOW) =1.03091(1000 + δ Oar(PDB)) − 1000 equation 2.4.2.3

With the limited amount of time markers (date of marking and date of collection) in the specimens studied, considerable uncertainty must be accounted for when compar- 18 13 18 ing predicted isotope values with actual δ Oar and δ Car in the shells. The δ Oar of specimens 981115 and 990131, collected before the 1999 wet season, can only be compared with predicted values for the 1998 dry season; only specimen number 991021, that was sampled in October 1999, spans the complete time range of the experiment (Figure 2.3). For 18 specimen 991021, the overall δ Oar curve shape of its last seasonal cycle (marked VII 18 and VIII in Figure 2.4c) fits reasonably well with the curve of δ Opredicted in Figure 18 2.5a. To refine the time control on specimen 991021, we have lined up the δ Oar curve 18 by tying it at two points to the δ Opredicted curve in Figure 2.5a. The first point, marked 18 "A", is at a rapid δ Oar decrease presumably caused by the entry of the Amazon River in Playa Cocha (26 January 1999). The second point, marked "B", is at the inflection point 18 where δ Oar start increasing again, probably related to the disconnection of Playa Cocha from the Amazon River at the beginning of 1999 dry season ( 05-07-1999). Growth rates of specimen 991021 are assumed to be constant for the period 26-09-1998 to point A, for the period between points A and B and for the period after point B until 21-10-1999. Be- tween tie-points A and B, the 1999 wet season is represented by a relatively stable plateau 18 18 18 of δ Oar and δ Opredicted values indicating low variability in temperature and δ OW . This plateau is the best interval to see if A. trapesialis builds its shell in isotopic equilib- δ18 18 rium. Values of Oar- δ Opredicted are less than 0.2 %, indicating (near) equilibrium 18 calcification of Oar. 18 More variable values in the dry season prelude an exact comparison between δ Oar and 18 13 δ OW but are in general agreement. The analysis of δ CDIC strips all DIC out of a 2.4 Discussion 37 sample; as a result the obtained data is total DIC. Molluscs will probably not use total DIC for calcification, but only one of the carbonic species, being bicarbonate, like marine uni- cellular algae E. Huxleyi do (Buitenhuis et al. 1999 and references therein). To compare 13 13 δ Car with δ Cbicarbonate we first calculated the concentrations of the carbonic species

a dry season ABwet season dry season stagnant water flowing water stagnant water 2 δ18Opredicted

0 δ18Oar 21/10/99

δ18Oar 31/01/99 -2 * δd181 Oar 15/11/98 -4

in ‰ [PDB] -6

-8 ~ 05-Jul-1999 disconnected from Amazon River

-10

-12 26-Jan-1999 connection with Amazon River 0 50 100 150 200 250 300 350 400 day #

b 26-Sep-1998 AB 20-Oct-1999 0

-2 - δ13CHHCO - 4.0‰ 3 -4 ** δ13Car 21/10/99 -6 δ13Car 31/01/931/01/99

δ 13 Car 15/11/98 -8 in ‰ [PDB]

-10 ~ 05-Jul-1999 disconnected from Amazon River ~ 05-Jul-1999 disconnected from

-12 26-Jan-1999 Amazon River connection with -14 0 50 100 150 200 250 300 350 400 ABday #

56 µm/day 56 µm/day c (specimen 990131) 30 µm/day (specimen 991021) (specimen 991021)

Figure 2.5: (a) Predicted aragonite oxygen isotope composition ofwater samples plotted vs. that of A. 44 trapesialis. Dashed plot symbols represent small samples that gave less than 0.9 volts ( CO2) and so do not comply with the reported typical reproducibility. Apparent mismatch of specimen 990131 may be due to cessation of growth during the ﬂooding of Playa Cocha. (b) Predicted aragonite carbon isotope composition of water samples plotted vs. that of A. trapesialis. Lacking actual pH data we calculated these two points using an estimated pH of 6.4. (c) Mean growth rates in μm per day for dry season ’98 and wet and dry season ’99. 38 Seasonal variation reﬂected in Amazonian freshwater bivalve A. trapesialis

with the help of temperature and pH. The diffusion of CO 2(g) into water forms four main species of DIC giving the following net reaction:

KCO2 K1 + K2 + 2− CO2(g) + H2O←→ H2CO3←→ H + HCO3←→ 2H + CO3 equation 2.4.2.4

Concentrations of carbonic species are controlled by the pH of the water. As the water in Playa Cocha had a pH between 5.8 and 6.4 dissolved CO 2 was dominant or equal to − 2− − bicarbonate (HCO3 ) and no CO3 was present. The concentrations of HCO3 and CO2(aq) can be obtained using the following equation (Clark and Fitz, 1997):

+ − [H ][HCO3 ] K1 = [H2CO3] and −4 2 pK1 =1.1 · 10 TC − 0.012TC +6.58 equation 2.4.2.5 ◦ st where TC is temperature in C. Because of the 1 dissociation of carbonic acid, H2CO3 ↔ + − + − H +HCO3 , we assume [H ]=[HCO3 ]. In the temperature range at hand the isotopic fractionation of dissolved CO2 relative to is between 8.5 and 9 according to the equation of (Mook and de Vries, 1998): −9866 ε − = +24.12 equation 2.4.2.6 CO2(aq) /HCO 3 TK

− The ratio of CO2(aq) and HCO3 concentrations and the fractionation between them yields 13 13 δ − δ C and CCO2(aq) for Playa Cocha water. HCO3 13 Based on the curve and the measured δ Car of specimen 991021 for the period between days #126-388 (after point "A", Figure 2.5) we have calculated the fractionation 13 2 between and δ Car of A. trapesialis showing a constant depletion of 4.0 (±0.7 %,R = 0.95) 13 1000 + δ C ( ) αshell = ar PDB =0.9660 bicarbonate 13 1000 + δ C − HCO3 (PDB) or 13 13 δ Car,shell = δ C − − 4 equation 2.4.2.7 HCO3 In a similar experiment on the fractionation between inorganic aragonite and bicarbonate by Romanek et al. (1992) an enrichment of 2.7 was found. The 6.7 difference between their and our experiment may result from the incorporation of metabolic carbon by the mollusc. Metabolic concentration of bicarbonate in the extrapallial ﬂuid and brooding of young may have played a role as well (Dettman et al., 1999). The other specimens (98115 and 990131) have only a few data points in the monitored period that can be used 13 to compare δ Car with but a similar offset of 4.0 is evident. Specimen 991021 shows a larger offset in the period prior to tie point "A". We ascribe that to the fact that the specimen was collected 50 m from the spot where the other two specimens and the water 2.5 Conclusions 39

samples were taken. An observed thermocline in the lake in this period suggests a strati- 13 ﬁcation that could lead to different δ CDIC values within the ﬂoodplain lake. Romanek et al. (1992) describe inorganic aragonite-bicarbonate fractionation as independent of temperature. Grossman and Ku (1986) describe decreasing 13C enrichment with increasing temperature in biogenic aragonite. The temperature variation in our dataset is too low to resolve whether the disequilibrium between the δ 13C of predicted inorganic aragonite and A. trapesialis has a temperature related component or not. As described by Dettman et al. (1999) an offset towards more negative values in shell aragonite is most likely due to the incorporation of metabolic carbon.

2.4.3 A. trapesialis growth rates The growth rates are determined by measuring along the axis of maximum carbonate accumulation along the outside part of the shell. The most important control on growth rate is the general ontogenetic decrease. Growth prior to the precipitation of carbon peak I (juvenile stage, Figure 2.4 was 60 mm for specimen 981115. The shape of this part of the curve and its values suggest that this juvenile stage is precipitated in one dry season. Assuming a half-year duration of this season this results in 333 μm of growth per day. Specimen 991021 grew 90 mm prior to peak I resulting in a growth rate of 500 μm/day. After peak I the growth rate quickly decreases to less than 60 μm/day. The rates of shell growth during the monitoring period are shown in Figure 2.5c. The first part is derived from specimen 990131 and reflects the period of growth between the mark, placed on 26-09-1998, and the ventral margin (collection date, 31-01-1999, Figure 2.3). Assuming a constant growth rate, the 7 mm that was precipitated during the 125-day period results in a growth rate of 56 μm/day. The fact that the mollusc did not register the chemistry change when the Amazon River flooded Playa Cocha five days prior to its collection might suggest a growth inter- ruption or a lack of sampling precision. To determine the growth rates between tie-points A and B and final collection, we used specimen 991021. The timeframe was obtained by using our correlations of carbon peaks (Figure 2.5) and our tie-points A, 31-01-1999, and B 05-07-1999 (between 26th of June and 14th of July). Between tie-points A and B (day # 126 and 281) specimen 991021 has grown 4.6 mm, resulting in a growth rate of about 30 μm/day during this period, Figure 2.5c). In the last period, between tie-point B, and collection date, growth was 6 mm, which relates to 56 μm/day assuming a constant growth rate. The increase in growth rate in the period after day #281 (tie-point B) is more dominant than the general ontogenetic decrease in growth Lower growth rates during the flooding of the Amazon River can have several reasons. Turbid water, lower nutrient levels, or brooding might take place.

2.5 Conclusions

¥ Pronounced seasonal cyclicity in rainfall patterns leads to strong isotopic variation 40 Seasonal variation reﬂected in Amazonian freshwater bivalve A. trapesialis

in Western Amazonian ﬂoodplain lakes. In Playa Cocha, a ﬂoodplain lake near the 18 Peruvian city of Iquitos, the oxygen isotope composition of water (δ OW ) varied between 0.5 and Ð9.3 in the dry and wet season respectively. The carbon isotope 18 composition of Playa Cocha water co-varied with δ OW between Ð1.8 in the dry season and Ð15.3 in the wet season. The δ 18O of rainwater samples collected varied between Ð19.6 (15 May 1999) and Ð0.4 (27 July 1999).

18 ¥ The ﬂoodplain lake dweller A. trapesialis precipitates a shell with δ Oar values well 18 described by the equation of Dettman et al. (1999), recording seasonal δ OW variation at high resolution in its growth increments.

13 ¥ A. trapesialis δ Car values have a constant offset of 6.7 more negative than predicted inorganic aragonite 13C described by Romanek et al. (1992), most likely due to metabolic effects. ¥ Generally, growth rates of A. trapesialis decrease from juvenile to adult growth stages. Marked growth rate changes occur within the wet - dry season cycle. Measured growth rates are higher in the dry season than in the wet season possibly due to turbid water, lower nutrient levels, or brooding during the wet season.

18 13 ¥ δ Oar and δ Car of modern A. trapesialis are high-resolution monitors of sea- 18 18 sonal variation of δ OW and δ O − in their host water. This paves the way HCO3 for the use of fossil Amazonian bivalves to document and understand past seasonal environmental change in Western Amazonia. Chapter 3

Growth rates and stable isotopes in juvenile Anodontites trapesialis molluscs

Abstract

In this chapter we show that juvenile Anodontites trapesialis has a much higher growth rate than its adult equivalent. This allows for reconstruction of seasonal and unique climate (weather) events particularly in the juvenile shell. Here the results of a monitoring experiment in the ﬂoodplain lake of Playa Cocha are given which highlight that unique rainfall events are recorded in the juvenile specimen. Keywords— Amazonas Peru; Bivalvia; fresh water; stable isotopes; atmospheric precipitation

3.1 Introduction

Mollusc shell growth bands are suitable recorders of inter- and intra-annual changes of host water chemistry and temperature. Their oxygen isotope record is the result of growth in ambient host water with a certain isotopic composition and of fractionation due to a certain temperature. In this chapter a calibration experiment is used to determine the growth rate of a juvenile mollusc shell, which is unknown. In Chapter 2, we documented that growth rates of three-year-old adult specimens of the bivalve A. trapesialis differ between ∼ 56μm/day in the dry season and ∼ 30μm/day in the wet season. It is thought that in order to escape predation, (the so called "predation window" as previously described for ﬁsh, e.g. Claessen et al. 2002; Dörner and Wagner 2003), juvenile molluscs grow signiﬁcantly faster than adult specimens. Theoretical molluscan growth curves that are exponential and asymptotic are proposed by Mutvei et al. (1994) and Schöne (2003). This in turn favours high resolution sampling and thus short events in the local weather are recorded in the molluscan shell chemistry. 42 Growth rates and stable isotopes in juvenile Anodontites trapesialis molluscs

3.2 Material and methods

For the calibration experiment, a juvenile A. trapesialis has been collected in floodplain Playa Cocha waters on the 21st of October, 1999, at the same time as the three-year-old adult specimen of the same species was retrieved that we discussed in Chapter 2. The juvenile specimen was only slightly smaller than the adult specimen, indicating that during the first year growth is significantly higher than in the subsequent years. The results of the chemistry of water samples were shown in Chapter 2. A Seamon mini temperature recorder (Hugrún Inc., Reykjavik, Iceland) was used in the floodplain lake to record bottom water temperature every 15 minutes for a period of 7 months. We used a Merchantek MicroMillTM (see Dettman and Lohmann, 1995; Spötl and Mat- tey, 2006, and references therein) to sample aragonite powder from growth increments for isotope measurements (see also Chapters 1 and 2). With the MicroMill sampling device a density of 50 milled lines per mm can be reached.

3.3 Results

3.3.1 Bottom water temperature record

Maximal diurnal temperature variations as observed in Playa Cocha during the dry season of 1998/1999 (when it was a small water body), were up to 5 ◦C and a few tens of a degree during the wet season, when Playa Cocha was part of the mainstream of the Amazon River (Figure 3.1). The large changes in temperature were caused by cooling of the pond during heavy rain showers (Chapter 2). Usually variation was less than 3 ◦C in the dry season. The entrance of the Amazon River is documented in the temperature record (Figure 3.1), fitting its reported entry in the floodplain lake on January 26 th, 1999. In the three weeks after the entrance of the Amazon River in Playa Cocha, the record shows a few cooling steps in Amazon water and then a small daily variation in temperature of less than a degree. Primar- ily, the daily temperature variation is dampened because of the trough-flow of river water. In late March the recorded daily temperature variation is limited and since the beginning of April there is no variation. A possible explanation for the absence of daily temperature variation since April is that sediment covering the temperature recorder may have buffered temperature variations. Sedimentation was fast during the flooding by the Amazon River, because we found the temperature recorder to be covered under 20 cm of sand in October. Every 15 minutes (96 times per day) a temperature measurement was recorded. To subtract the daily temperature fluctuations of night and day from the record, a 96 point moving average has been applied to the record of Figure 3.1. This enabled a comparison between the isotope record of the juvenile specimen (Figure 3.2) with the high-resolution temperature record. 3.3 Results 43

34°C

26°C

Temperature bottom water water bottom Temperature 22-09-1998 22-10-1998 21-11-1998 21-12-1998 20-01-1999 19-02-1999 21-03-1999 20-04-1999

Figure 3.1: Playa Cocha bottom water was measured every 15 minutes for more than 7 months (in grey, the black line represents a 96 moving average to subtract daily temperature ﬂuctuations). On January 26th Amazon water entered the ﬂoodplain lake, which resulted in a decrease in the daily temperature variation. The recorder, recovered in October 1999, was buried under 20 cm of sand. Sediments covered the recorder most likely since the beginning of April. The data of the precipitation record (October 1, 1998-January 31, 1999) is from Iquitos Airport (SPQT), location 15 km west-southwest from the site. Large rain showers result in a temperature drop in Playa Cocha water.

3.3.2 Isotope records

The specimen did not exist prior to the experiment campaign. Before measuring the isotope composition of the growth lines in the shell only one date, (one anchor point) was known: the collection date October 21st,1999. Looking at the isotope record two more anchor points (dates) can be determined: the entrance of the Amazon River on January 26 th and its reported exit on July 5th. Oxygen isotope values vary between Ð2.10 and Ð8.45 (see Table A-10 in the Ap- pendix on page 122. The values for carbon isotopes vary between Ð3.67 and Ð14.97 . The drop of 5 in the δ 18O-record, caused by the entrance of Amazon water in the floodplain lake occurs synchronously with a drop of 8 in the δ 13C record (Figure 3.2). Before these significant drops synchronous cyclicity is observed for the oxygen and carbon isotope records, after the drop both isotope records show a gradual increase in values. The isotope values can be found in Appendix A, Table A-10, on page 122. At first sight the isotope curve shows similarity with the 96 point moving average temperature profile: relatively flat lines after the first anchor point, (dated on January 26 th), due to the entrance of the Amazon River and similar spiky records prior to the anchor point representing the end of the dry season of 1998. 44 Growth rates and stable isotopes in juvenile Anodontites trapesialis molluscs

0 3 1 -2 26-01-1999: entrance Amazon River

-4 Collection date 21-10-1999

-6

in ‰ [PDB] -8 C

13 2

O,δ -10 18

δ 05-07-1999: exit Amazon River δ18O -12 δ13C

-14

-16 0.00 20.00 40.00 60.00 80.00 100.00 120.00 distance from the umbo (mm)

Figure 3.2: The isotope records of a juvenile A. trapesialis. The entrance of Amazon water into the ﬂoodplain lake on January 26th (anchor point 1) dampened the temperature signal (Figure 3.1) and resulted in a large shift in the stable isotope records. From then onward the δ18O values increased gradually whereas the the δ13C values remain relatively constant until the ﬂoodplain lake and the Amazon are disconnected. The δ13C record shows a rapid increase on July 5th (anchor point 2). δ18O values are back to the values of the beginning of the record, δ13C are still slightly lower than the mean of the previous dry season.

3.4 Discussion

3.4.1 Juvenile growth rates The growth rate of the three-year-old adult A. trapesialis measured over the longest axis of growth has been determined to be resp. 30 and 56 μm/day for the wet and dry season in mean distances per day, (i.e. in its 3rd year of growth, Chapter 2, Figure 2.4 on page 34). In the first year, bivalves tend to grow at much higher rates (Mutvei et al., 1994; Schöne, 2003), probably to escape the predation window as quickly as possible. Therefore the growth pattern of the juvenile specimen is most likely different from that of the adult, i.e. growth rate might be less affected by seasonal change. Because the date of the entrance of Amazon river water into Playa Cocha, the date of disconnection with the Amazon River and the date of collection of the juvenile are known (resp. Anchor Points 1, 2 and 3, Figure 3.2), the growth rate of the juvenile specimen can be determined. We have used a second order polynomial curve fit through the three anchor points resulting in Figure 3.3. Dates and size of the mollusc are now fitted with each other. Sizes and isotope values are now linked to dates (Figure 3.4). 3.4 Discussion 45

The result of our model shows a decrease in the growth rate from 0.45 mm on the ﬁrst day to 0.16 mm on the collection day, which is still almost 3 times faster than the growth rate of the three-year-old adult in the dry season.

3.4.2 Extrapolation of the growth in the dry season With the curve fit we were able to allocate further dates to the individual isotope samples by extrapolation in between the anchor points. The isotope records and temperature record can now be correlated to inspect if finer links between temperature and the isotope records exist (Figure 3.4). Prior to December 5th there is no correlation between the isotope records and temperature, between December 5th and Anchor Point 1 (January 26th) there is some similarity. Al- though some of the peaks in the isotope records are also to be found in the temperature record, not all of them can be found back. While expecting a negative correlation (a temperature rise should turn into isotopic drop in the record), there appears to be a positive correlation. A temperature rise, leading to evaporation and outgassing results in a "heavier" residue water (higher δ18O values, Chapter 2). A temperature drop in the record sometimes goes together with a rain shower, such as the ones on December 23 and January 8 and 9. Water is mixed with rain water and depending on the amount of rain water and its isotopic composition, drops in temperature may result in a rise or a fall in the isotopic trend. The rain shower of December 23rd is not reflected in the isotope record although the one in Jan-

130 3 120 Collection date 21-10-1999: 112369µm

110 05-07-1999: 90508µm 2 100 exit Amazon River

60 26-01-1999: 41611µm 1 50 entrance Amazon River length (mm from umbo) 40 Y = M0 + M1*x + M2*x 30 M0 = -5.23112e+05

20 M1 = 28.546 Fitted date of birth: M2 = -0.00038934 10 19-10-1998

0 date 01-11-1998 09-02-199820-05-1999 28-08-1999 06-12-1998

Figure 3.3: In Figure 3.2 three anchor points are determined. With the MicroMill camera and software the distance between the sampled lines and the umbo is measured. With time and distance (growth) linked for this three anchor points a second order polynomial curve ﬁt has been conducted to obtain a growth model. 46 Growth rates and stable isotopes in juvenile Anodontites trapesialis molluscs

Figure 3.4: A 96 moving average (temperature measurements were conducted every 15 minutes) has been applied to the temperature data from Figure 3.1 in order to subtract the diurnal ﬂuctuation (upper panel). The temperature record extends from 25-09-1998 until 02-05-1999. The isotope record of a juvenile A. trapesialis is plotted in the lower panel. The entrance of Amazon water (marked as anchor point 1 from Figure 3.3 in the ﬂoodplain lake dampened the temperature signal and resulted in a large shift in the stable isotope record. The growth model of Figure 3.3 has been used to link the tamperature record with the isotope record. uary is. The last unknown parameter with which the matching could have been completed is missing unfortunately: a high time resolution isotopic composition record of the host water.

3.5 Conclusions

¥ In order to escape the so called "predation window", the growth rates of juvenile Anodontites trapesialis are3Ð5times the growth rate of adult A. trapesialis in the third year. ¥ In this study a sampling resolution of 1 sample per 2 days has been conducted, in some parts of the record even 1 sample per day. ¥ The very high growth rates in juvenile bivalves record changes in water chemistry as evidenced in the isotope response to a single rain shower on January 8 and 9, 1999. Chapter 4

Seasonal Amazonian rainfall variation in the Miocene Climate Optimum

Abstract Modern and fossil freshwater bivalves from north-eastern Peru are investigated to reconstruct seasonal rainfall patterns in Miocene Amazonia. Oxygen isotope variation in incremental growth bands of fossil bivalves reﬂects past hydrological conditions in the Miocene Climate Optimum (MCO), when the world was warmer than today. A calibration experiment was conducted on a modern bivalve. Modern river dwelling Triplodon corrugatus shows large amplitudinal changes in δ18O, which mirror the seasonal variation in rainfall as a result of the annual migration cycle of the Inter Tropical Conver- gence Zone (ITCZ). Growth incremental oxygen isotope records of Miocene Amazonian Diplodon aff. longulus bivalves show strikingly similar patterns. This suggests that the seasonal migration of the ITCZ and the intensity of the hydrological cycle in the MCO were comparable to today. The implications are that humid climate conditions sufﬁcient to sustain a rainforest ecosystem already existed ∼ 16 Ma ago. Keywords— Western Amazon Basin, Pebas Formation, Miocene Climate Optimum, stable isotopes, bivalves, precipitation, seasonal variations.

4.1 Introduction

Rainforest systems are sustained by speciﬁc environmental conditions characterized by large amounts of rainfall. The Amazonian rainforest is one of the major rainforests in

This Chapter is based on Kaandorp, Vonhof, Wesselingh, Romero Pittman, Kroon, and van Hinte 2005. Seasonal Amazonian rainfall variation in the Miocene Climate Optimum. Palaeogeography, Palaeoclimatology, Palaeoecology 221 (1-2): 1-6. 48 Seasonal Amazonian rainfall variation in the Miocene Climate Optimum the world and today its rainfall is controlled by a monsoonal system driven by the seasonal migration of the ITCZ (Figure 4.1 Marengo, 1998). The climatic oscillation results in a distinct wet-dry season cyclicity to which rainforest taxa are strongly adapted. Pollen records show that the rainforest existed in the Miocene (Hoorn, 1994a), but it is not clear if environmental conditions were similar to today. In this paper we show that the seasonal patterns of rainfall in the warmest period of the past 35 million years, ∼ 16 Ma ago, in the middle of the Miocene Climate Optimum (MCO, Flower, 1999; Zachos et al., 2001), were exactly the same as today by using δ 18O signatures of growth bands in the shells of bivalves. The largest part of the warm and wet Amazon tropical rainforest is located on terra firme, land that is never inundated. A smaller part, called várzea is permanently or tem- porarily inundated and consists of floodplains, floating meadows, swamps and flood basin meadows (Hoorn, 1994a).

L L 0° ITCZ 0° "dry" "wet" ITCZ 10° H H H L H

ITCZ in ITCZ in July January

Figure 4.1: The ITCZ is a low-pressure band in the tropics where air masses converge from both hemispheres at the boundary between northeasterly and southeasterly trade winds. The ITCZ follows the insolation maxima, migrating to the warmest surface areas throughout the year attaining its northernmost displacement in June-July, and southernmost in December-January. A contracted equatorial continental air mass, together with activity of the eastern Paciﬁc and western Atlantic ITCZ, produces southeastern trade winds over Amazonia. This results in an Amazonian dry period during boreal summer with low river water levels in Peruvian Amazonia in June-September. During austral summer the ITCZ lies south of Amazonia, attracting northern hemisphere northeastern trade winds. Convective activity of an expanded continental air mass over Amazonia in this season causes a wet regime in the region, with maximum precipitation and inundation of várzea in October-May. Northeastern trade winds moving humid air meet the ITCZ in the northwestern part of Amazonia. Elevation of the air mass due to the Andean topography will produce maximum rainfall in April-June. In the central part of the Basin the maximum rain season has a peak during March-May (Marengo, 1998). H stands for atmospheric high pressure, L stands for low pressure.

Seasonal variation in the amount of precipitation results in several meters of river level variation is a precondition for the várzea landscape, with its highly diverse and adapted (semi-) aquatic biota. The Andean Cordillera in the west of the continent trap atmospheric moisture, which originates from evaporated Atlantic Ocean water, in the Amazon Basin. Air masses rain out when rising against the eastern Andean slopes (Grootes et al., 1989; Salati 4.2 Methods 49

and Vose,1984; Hooghiemstra and van der Hammen, 1998) providing Amazonia with direct rainfall and runoff water. Rainfall variation in Amazonia, forced by the seasonal shifting of the ITCZ, is reﬂected in the oxygen isotopic composition (δ 18O) of precipitation (Dans- gaard, 1964; Grootes et al., 1989; Rozanski et al., 1993, Global Network for Isotopes in Pre- cipitation and Isotope Hydrology Information System, (available at http://isohis.iaea.org/)). Dry season precipitation in June-August is low; all atmospheric water vapour loss due to precipitation is replenished by means of evapotranspiration. Consequently no depletion of δ18O occurs during this period, and rainwater retains an oceanic signature throughout the basin. Convective showers during the wet season (November-April) exceed evapotranspiration by a factor of two to three, depleting the water vapour of its heavy isotopes during transport from Atlantic to Andes (Grootes et al., 1989; Figure 4.2). This seasonal be- haviour of isotopic composition of precipitation is recorded in Andean ice cores (Grootes et al., 1989; Thompson et al., 1995). At lower altitude temporal and spatial changes of δ18O in lakes and rivers are potentially documented in bivalves.

18 δ Op,dry = -10.5 ‰ 18 δ Op,wet = -33.7 ‰

18 δ Op,dry = -0.38 ‰ δ18O = -19.6 ‰ p,wet 18 δ Op = +0.7 ‰

Andes Mt range

Amazon Basin Atlantic (Iquitos, Peru, 100m a.s.l.) Ocean

Figure 4.2: Seasonal change in oxygen isotope composition of precipitation in Amazonia, modiﬁed after Grootes et al., (1989). Precipitation in Iquitos in 1998 and 1999 was measured in this study.

4.2 Methods

Discrete growth increments of freshwater unionoid shells can be recognized in cross sections and sampled at high temporal resolution. It has been shown that these growth increments adequately record seasonal variation in the isotopic composition of ambient water (Dettman et al, 1999). For north-western Amazonia we have shown in a monitoring 50 Seasonal Amazonian rainfall variation in the Miocene Climate Optimum experiment (Chapter 2) that the bivalve Anodontites trapesialis (Lamarck, 1819) from an 18 Amazon ﬂoodplain lake builds its aragonitic shell (δ Oar) in isotopic equilibrium with 18 18 its host water (δ OW ). The δ OW is predominantly controlled by rainfall patterns and evaporation. Seasonal temperature variation is negligible. To further support the use of bivalves as recorders of environmental conditions a Triplodon corrugatus (Lamarck, 1819) was collected alive from a clayey substrate in Itaya River (03 ◦47’08"S; 73◦17’08"W), a small tributary of the Amazon, south of Iquitos, Peru on September 23 rd 1998 (Figure 4.4). Oxygen isotopes derived from micro-sampled growth increments of T. corrugatus show a large amplitudinal change from -3.9 to -10.1 (Figure 4.4a, Table A-11 on page 123) representing at least 13 annual cycles each containing a wet season (low δ 18O values) and a dry season (high δ18O values) (Grootes et al., 1989; Chapter 2). Results of isotope measurements are listed the Appendices in Tables A-11, B-27 and B-30. As modern bivalves

Napo River 73°00'W 73°15'W

Mz In 3°30'S Mazán River Amazon a c N

Iq b 10 km 3°45'S

Figure 4.3: Location map. (a) Miocene D. aff. longulus bivalves from an outcrop in the right bank of the Napo River at the conﬂuence with the Mazán River. (b) The sampling site of modern T. corrugatus in the Itaya River. (c) Monitoring site of A. trapesialis bivalves (Chapter 2) . The city of Iquitos and the villages of Indiana and Mazán are indicated by Iq, In and Mz respectively. are evidently good recorders of seasonal patterns, fossil shells are expected to record past seasonality. The extraordinary well preserved fossil bivalves from the widely distributed Miocene Pebas Formation in Amazonia (Wesselingh et al., 2002; Vonhof et al., 2003) give an excellent opportunity to reconstruct Miocene seasonal rainfall patterns from stable isotope records from bivalve growth increments. Palynological analyses of Pebas Formation sediments (Hoorn, 1994b) provided evidence for the presence of várzea type of biota during the Middle Miocene. The Pebas Formation consists of sands, organic clays, lignites and blue smectite clays found in 3-7 m thick coarsening-up cycles. Interpolation of pollen ages from nearby Indiana outcrops (Psiladiporites-, Crototricolpites zones: Hoorn, 1994a), indi- 4.2 Methods 51

cate an age of late Early to Middle Miocene, corresponding to an age of 15-17 Ma on the geological time scale (Berggren et al., 1995). Here we evaluate the climatic conditions that sustain a várzea-type rainforest system in the Miocene Climate Optimum. Fossil specimens of Diplodon aff. longulus (Conrad, 1874, this form is more convex than D. longulus and is characterized by two or three prominent rounded ventrolateral ridges) were collected from sediments of the Miocene Pebas Formation at the conﬂuence of the Mazán and Napo rivers (co-ordinates 03◦29’58"S; 73◦05’45"W, Figure 4.3). This outcrop is part of the Indiana outcrop group. Sr isotopes of bivalves from this outcrop indicate a fresh water environment fed by Andean run off (Vonhof et al., 2003). The bivalves are excellently preserved. Analyses with a Micro-raman spectrometer show

a direction of shell growth b c umbo ventral margin

0 juvenile adult juvenile adult juvenile adult

-4

-8 O, ‰, V-PDB] O, ‰, V-PDB] 18 δ [ -12 50 70 90 110 130 150 60 70 80 90 30 40 50 60 70 80 [mm from umbo]

Figure 4.4: Oxygen isotope records of samples from bivalves taken along the axis of maximum shell growth. (a) Living T. corrugatus sampled in Itaya River. (b) and (c) Two fossil D. aff. longulus specimens sampled near Mazán. V-PDB is Vienna PeeDee Belemnite. Growth increments were micro-sampled with a resolution of 9 lines per mm. (White lines in the enlargement show examples of micro sampling milling lines). Powders recovered from each line were used to obtain oxygen stable 18 isotope data (δ Oar) performed at Vrije Universiteit Amsterdam on a Finnigan MAT 252 mass spectrometer equipped with an automated preparation line. The long-term reproducibility of a routinely 18 18 analyzed carbonate standard (NBS 19) is better than 0.09 for δ O (1 SD). δ Oar values are re- 18 ported vs. Vienna-PeeDee Belemnite standard (V-PDB); Water values (δ OW ) are reported against Vienna Standard Mean Ocean Water (V-SMOW).

that all bivalves have retained their aragonitic mineralogy. Trace element analyses from similarly well-preserved Pebas Formation bivalves from other localities showed no evidence for diagenetic alteration (Vonhof et al., 1998; 2003). The bivalves were sampled in a single bed containing abundant paired valves, indicating minimal transport prior to deposition. The oxygen isotope records of the fossil mollusc shells are strikingly similar to modern A. trapesialis (Chapter 2) and T. corrugatus (Figure 4.4, table 1). Fossil Diplodon bivalves show at least 8 successive δ18O cycles with values ranging from -3.7 to -10.5 and -1.9 to -10.5 (Figure 4.4c). These data are attributed to dry/wet seasonal change in host water oxygen isotope composition. Diminished growth rates towards the ventral margin, indicated by a condensed annual cyclic signal, are observed in both the modern and fossil shells. 52 Seasonal Amazonian rainfall variation in the Miocene Climate Optimum

4.3 Discussion

In order to interpret the Miocene δ 18O signals, we need to assess to what extent the various environmental parameters (Dansgaard, 1964) responsible for the oxygen isotope composition of rainwater in the Miocene may have differed from today. These parameters are: latitude, altitude, coastal distance or "continental effect", surface air temperatures, water vapour source and the amount of precipitation. The latitudinal effect has not changed, Amazonia already straddled the equator during the Middle Miocene (Scotese, 2001), and low annual variation of surface air temperature like today is expected. The altitude effect is responsible for the most depleted values of δ 18O in Amazonian precipitation. Typically the lowest values are found in the Andean ice cores (Grootes et al., 1989; Thompson et al., 1995). However the contribution of high-altitude precipitation to natural waters in our study area is limited, because rainfall intensity is highest in the Amazonian lowlands. Therefore we expect a possible lower Andean Mountain Range in the Miocene (Steinmann et al., 1999) not to have had much inﬂuence on the isotopic composition of waters in our study area. The extremely low isotope values in parts of the molluscs demonstrate that the Andes during the Middle Miocene must have acted as a barrier to atmospheric vapour. The mechanism caus- ing the "continental effect" (Dansgaard, 1964; Rozanski et al., 1993) is a gradual removal of moisture by condensation from air masses moving inland, coupled with a preferential removal of heavy isotopes during the condensation process. A longer travel distance would lead to lower δ18O of the remaining vapour. Paleogeographical maps of Miocene northern South America (Hoorn et al., 1995) imply no differences in the pathway of vapour travelling from the Atlantic Ocean water to Western Amazonia. Although surface air temperatures on the continent during the Miocene Climate Optimum are not well known, there are good indications that low-latitude sea surface temperatures (SST) were comparable to today (Savin, 1977; Martínez and del Río, 2002). At comparable SST’s the amount of evaporated source water should be the same, although the source waters of MCO Amazonian atmospheric moisture might have been slightly more negative than today: MCO calculated global ma- 18 rine water δ OW value (-0.8, V-SMOW, (Lear et al., 2000)) is lighter than its modern 18 equivalent (δ OW = -0.28), but this is negligible with respect to the observed changes in δ18O of the molluscs.

4.4 Conclusions

¥ For modern as well as Miocene waters, the amount of precipitation must have been the controlling parameter for seasonal isotopic variation. Based on the observation that cyclicity and absolute values of δ 18O of Modern and Miocene molluscs are strikingly similar, we conclude that seasonal migration of the ITCZ and the intensity of rainfall in the MCO are comparable to today. Possible higher temperatures during the MCO 18 in our study area are irresolvable in the molluscan δ OW records. ¥ Wider implications are that the Amazonian hydrological cycle is most likely not affected by globally warmer climates than today considering potential warming of 4.4 Conclusions 53

Earth’s climate in future. ¥ The extent of Western Amazonian rainforest in the Middle Miocene was not as large as today, since most of the area was occupied by lakes and wetlands (Nutall, 1990; Wesselingh et al., 2002). Our study shows that climatic conditions required to sustain highly diverse tropical forests and várzea environments existed in the Miocene Climate Optimum. 54 Seasonal Amazonian rainfall variation in the Miocene Climate Optimum Chapter 5

Ecological implications from geochemical records of Miocene western Amazonian bivalves

Abstract

Stable-isotope profiles through successive growth increments of Miocene bivalves from western Amazonia reveal paleobiological characteristics, such as biotope preferences, longevity, and ontogenetic development. Two groups of bivalves, a fluvial and a lacustrine group, are recognized. The fluvial group is composed of pearly freshwater mussels whose isotope profiles are characterized by clearly recognizable cyclicity of considerable amplitude, similar to those from pearly freshwater mussels studied from the modern Amazon region. The lacustrine group, dominated by pachydontine bivalves, is characterized by very low amplitude and irregular isotope signals. Additional trace element analyses show seasonal variation in phase with stable-isotope cycles. Seasonal stable isotope cycles show growth rates and longevity in bivalves from the fluvial group but are not clearly expressed in the lacustrine group. The Miocene Pebas ecosystem of western Amazonia can be characterized as a wetland system of connected shallow lakes, swamps, and tributaries. The stable isotope signatures of fluvial bivalves are good recorders of prior seasonal climate changes, in contrast with those of lacustrine bivalves. Keywords— Amazonia, Bivalvia, Colombia, Miocene, Paleobiology, Pebas Formation, Peru, Stable isotopes

This Chapter is based on Kaandorp, Wesselingh, and Vonhof 2006. Ecological implications from geochemical records of Miocene western Amazonian bivalves, Journal of South American Earth Sciences 21(1-2): 54-74. 56 Ecological implications of Miocene western Amazonian bivalves

5.1 Introduction

Ongoing research on high-resolution growth band isotope geochemistry of Amazonian bivalves has provided records of present and past seasonality and water chemistry dynamics (Wesselingh et al., 2002; Chapter 2, 4). Environments in equatorial continental interiors are subject to low seasonal cyclicity with respect to temperature (Crowley et al., 1986). Isotopic variation therefore is dominantly attributed to annual variations in precipitation regimes and host water chemistry. Pearly freshwater mussels living in ﬂoodplain lakes in Peruvian Amazonia (Figure 5.1) adequately record host water chemistry (Chapter 2). Well-preserved pearly freshwater mussels from the Miocene Pebas Formation of western Amazonia provide evidence of seasonal precipitation/evaporation variation that was similar to that of today (Chapter 4). In this article, we investigate the use of stable isotope signatures to distinguish among biotopes (and, by inference, depositional environments) and address paleobiological implications, such as growth regimes and longevity. We aim to improve interpretations of predominant depositional settings within the Pebas Formation, about which considerable scientiﬁc debate exists.

74°W 73°W 72°W 71°W 70°W 02°S Colombia Putumayo River

Western Amazonia

Mazán River Napo River Peru 03°S Pebas

Sta Rosa Mazán Indiana Amazon River Iquitos ColombiaLos Chorros San Playa Cocha Antonio Sta Elena 04°S Itaya River Nuevo Horizonte Nauta Peru Yav ari River Brazil 05°S

Figure 5.1: Location map of the sampling sites in northeastern Peru.

5.1.1 Pebas Formation The Pebas Formation (a.k.a., Solimões Formation in Brazil) covers an area of over 1 000 000 2 km in Colombian, Peruvian, Ecuadorian, and Brazilian Amazonia (Wesselingh et al., 2002) and is mainly of late Early-early Late Miocene age (Hoorn, 1993, 1994a; Figure 1.7 on page 5.1 Introduction 57

9). The formation is characterized by a predominance of blue smectitic clays, silts, and fine- grained sands, commonly deposited in coarsening-upward sequences and often capped by lignites. Fossils such as bivalves, gastropods, plant material (in lignite layers), and pollen are common. Interpretations of the (predominant) depositional setting of the Pebas Formation include a series of shallow lakes, streams, and swamps of varying salinities (Nuttall, 1990); a river- (floodplain) lake environment with rare marine influence (Hoorn, 1993; Hoorn, et al., 1995); and a predominant marginal marine environment (Gingras et al., 2002). Previous work on molluscan paleontology and geochemical signatures has been interpreted to reflect a long- lived lake system with swamps, only rarely reached by a diluted marine incursion (Vonhof et al., 1998, 2003; Wesselingh et al., 2002). All shell material analyzed in this study is believed to represent the strictly freshwater environment of the Pebas system. Different biotopes assigned to the molluscs in this study (fluvial, lacustrine) are based on the analysis of faunal assemblages presented by Wesselingh et al. (2002). Herein, we refer to this Miocene ecosystem simply as Pebas and by its derived adjective Pebasian.

5.1.2 Molluscs

Representatives of two groups of bivalves from the Pebas Formation were used in the present analyses, including three species of pearly freshwater mussels [Unionoidea; Diplodon aff. longulus (Conrad, 1874), Diplodon longulus Conrad, 1874, Anodontites capax (Conrad, 1874)] and two species of pachydontine corbulids (Pachydon tenuis Gabb, 1869, Pachy- don erectus Conrad, 1874). For comparison, two modern species were used: Triplodon corrugatus (Lamarck, 1819) from Itaya River and Anodontites trapesialis (Lamarck, 1819) collected in Playa Cocha (Figure 5.1). The bivalves are portrayed in Figure 5.2. Pearly freshwater mussels are semi-infaunal filter feeders restricted to freshwater ecosystems, such as lakes, streams, and rivers. These shells are found in low abundances throughout the Pebas Formation; they are common in the floodplain lake and lake-margin facies but rare in the lake assemblages. In the Pebas For- mation, the Pachydontinae form a radiation represented by six genera with approximately 19 species that are numerically dominant and almost entirely endemic. These shallow bur- rowers are commonly found preserved in life positions. Pachydon obliquus, Gabb, 1869, the most common species of Pachydon in the Pebas Formation, dominates the faunas of dysoxic, organic-rich, clayish intervals (Wesselingh et al., 2002). Only one pachydontine corbulid bivalve lives today: Anticorbula fluviatilis (Adams, 1860). It lives in freshwater habitats of the central and lower Amazon region, as well as in rivers draining the Guyana Shield all the way down the upper reaches of estuaries (Wesselingh et al., 2002 and references therein). 58 Ecological implications of Miocene western Amazonian bivalves

5.2 Material and methods

The sampling localities of thirteen fossil Pebas and two modern shells (reﬂecting potential modern analogues of Pebasian environments) that have been microdrilled and analyzed, are indicated in Figure 5.1. Depositional contexts of the localities are described in the Ap- pendix. Microdrill sampling techniques are discussed in Chapter 2. Stable isotope analyses of shell aragonite were performed at the Vrije Universiteit Amsterdam on a Finnigan MAT 252 mass spectrometer equipped with an automated (Bremen-type) preparation line. The reproducibility of a routinely analyzed carbonate standard (NBS 19) is better than 0.09 for δ18O and 0.05 for δ13C (1 SD). Minor and trace element analyses were performed on a Varian Liberty ICP-AES after digestion of ∼ 100-200 μg of powdered sample in 1 N

c. a. b. f.

e. d.

Figure 5.2: (a) Triplodon corrugatus (Lamarck, 1819), Itaya River, Loreto, Peru. Leg. R.J.G. Kaan- dorp, 1998. Width 167 mm. (b) Diplodon longulus (Conrad, 1874), Santa Rosa de Pichana, Loreto, Peru. Pebas Formation, Middle Miocene. Leg. F.P.Wesselingh, 1996. Width 91 mm. (c) Diplodon aff. longulus (Conrad, 1874), Mazán, Loreto, Peru. Pebas Formation, late Early-early Middle Miocene. Leg. F.P. Wesselingh, 1996. Width 77 mm. (d) Anodontites capax (Conrad, 1874), Los Chorros, Amazonas, Colombia. Pebas Formation, late Middle-early Late Miocene. Leg. M.C. Hoorn, 1991. Width 91 mm. (e) Anodontites trapesialis Lamarck 1819, Playa Cocha, Iquitos, Loreto, Peru. Leg. R.J.G. Kaandorp, 1998. Width 130 mm. (f) Pachydon erectus Conrad, 1871, Santa Rosa de Pichana, Loreto, Peru. Pebas Formation, Middle Miocene. Leg. F.P. Wesselingh, 1996. Width 49 mm (g) Pachydon tenuis Gabb, 1869, Nuevo Horizonte, Loreto, Peru. Pebas Formation, late Middle-early Late Miocene. Leg. F.P. Wesselingh, 1991. Width 48 mm. 5.3 Results 59

HNO3 and subsequent dilution to ∼ 0.1 N HNO3 sample solution. Concentrations of Fe, Ba, and Sr in CaCO3 are reported in parts per million (ppm). A standard test experiment indicated reproducibility (1 SD) of better than 5% for the method used. The samples are stored at the National Museum of Natural History Naturalis, Leiden, the Netherlands.

5.3 Results

5.3.1 Quality of fossil material

The excellent preservation of Pebas Formation molluscs was documented using trace element concentrations and SEM photography (Vonhof et al., 2003). For this study, we performed additional Raman probe analyses, which gave an aragonite signal throughout the shells that indicated no alteration toward calcite has occurred. Some of the analyzed shells even contained remains of their periostracum. Another preservation test was performed on a D. aff. longulus sample (from Mazán, Figure 5.3a, b); stable isotope records from two different growth directions (nacreous and pris- matic layers) were established to test the consistency of the isotope patterns and amplitude of changes. All isotope excursions in both transects show high similarity in the number of cycles and amplitudinal change, which strongly indicates that isotope variation is not a secondary (diagenetic) effect (Figure 5.3b). Increased Fe concentrations in some parts of the specimens studied are not interpreted as indications of diagenetic alteration but as a result of dysoxia in bottom sediments. This interpretation is in accordance with elevated Fe concentrations in living Anodontites sp. from the Amazon River (Figure 5.4).

5.3.2 Stable-isotope composition of Pebas Formation bivalves

Isotope data of the analyzed bivalves are summarized in Table 1 and shown in Figure 5.3, Figure 5.5, and Figures 5.6Ð 5.10. These data include thirteen shells from the Pebas Formation and three shells from different modern environments (riverine, floodplain lake, and lacustrine) for comparison. Two modern bivalves representing fluvial and fluviolacustrine environments show distinct isotope signatures (Figure 5.5). The δ 18O cyclicity in the fluvial T. corrugatus is well defined, with comparatively large amplitude, and cyclicity is pronounced in the fluviolacustrine A. trapesialis as well. Significant variation occurs in the stable isotope profiles of the 13 analyzed Pebas Formation bivalves (Figures 5.3, Figure 5.6 — 5.10). Analogous to the modern shells, growth incremental δ 18O cycles are interpreted to reflect δ18O variation of the host water between successive wet (lower values) and dry seasons (higher values; Chapter 2). 60 Ecological implications of Miocene western Amazonian bivalves

Pachydon tenuis Nuevo Horizonte 91FW70, Figure 5.6. The δ18O record ranges from approximately Ð1 to Ð3, and δ 13C ranges from approximately Ð9 to Ð13, showing no cyclicity, which makes life age estimation for the shell based on the oxygen/carbon isotope record impossible (Figure 5.5e).

Diplodon longulus Nuevo Horizonte 91FW70, Figure 5.6. From the same layer, this specimen has very pronounced cyclicity in both δ 18O and δ13C. Ten regular δ18O cycles are interpreted with approximately 2 amplitudes. Well-deﬁned δ13C cycles of variable amplitude (2-5) are nonsynchronous with δ 18O cycles. The δ13C cycles consistently peak slightly earlier. Average δ 13C values for both Diplodon and

growth direction 1 0

a b c d e f g h i j k l mno p q -2

-4

-6

Umbo Ventral margin O in ‰ V-PDB -8 18 δ

-10 C, 13 δ -12

-14

-16 0 20 40 60 80 100 120 140

growth direction 1 δ13C 4 mm δ18O growth direction 2 { 0 a b c d e hi jklmnop -2

-4

growth direction 2 -6

-8 O in ‰ V-PDB

18 -10 δ

C, -12 13 δ

-14

-16 200 180 160 140 120 100 80 60 40 20 0 sample numbers

Figure 5.3: The isotope records from a single specimen of Diplodon aff. longulus, sampled in two different growth directions, show a high similarity, indicating that isotope variation is not a secondary (diagenetic) effect. Sampling in the ﬁrst growth direction results in a higher resolution in the latest stage; sampling along the second growth direction obtains a higher resolution in the juvenile stage. The majority of bivalves in this study are sampled against the second growth direction, from ventral margin to umbo. The isotope records start from the umbo, the juvenile part, on the left side, and end at the ventral margin on the right. All carbon and oxygen stable isotope values obtained from molluscan growth increments are plotted against the Vienna PeeDee Belemnite standard. For comparative reasons, all molluscan isotope records are plotted with the same scale on the y-axis, values between ∼ 16 and 0, unless stated differently. 5.3 Results 61

Pachydon from Nuevo Horizonte are Ð11. However, δ 18O averages differ markedly:Ð 2 for Pachydon and Ð8 for Diplodon.

Anodontites capax Los Chorros 89CH, Figure 5.6. Observed variation in the δ 18O record can be interpreted as six or seven cycles. The carbon isotope proﬁle is more irregular, though for the adult part (below increment number 120) variation is correlated with δ18O variation, albeit with a slight lagging.

Anodontites capax San Antonio 96FW810, Figure 5.7 The amplitude of δ18O cycles increases from less than 1 in juvenile stages to ∼ 4 in adult stages. The δ13C cycles are better developed throughout and are in phase with oxygen isotope cycles in the adult part of the shell. The δ 13C amplitude varies but is usually around 3.

Diplodon longulus Santa Elena 96FWSE, Figure 5.7 Oxygen isotope cycles are unclear in the juvenile half of the shell (samples 36-67). Three cycles are clear in the adult part of the shell. The more pronounced δ 13C cyclicity is in

350 2

-2 250 -4 δ18O -6 in ‰ [PDB] [Fe] in ppm -8 150 δ13C -10

-12 50 -14 Fe 0 -16 121 61 1 sample numbers

Figure 5.4: Minor element Fe record (thick lines) from the growth increments of Living Anodontites sp. (Playa Cocha) compared with isotope data (δ13C: open dots, δ18O: black dots). Left axis: ppm/Ca, right axes: δ13C, δ18Oin versus PDB, x-axis: isotope sample numbers. Not all samples have been analyzed for iron. 62 Ecological implications of Miocene western Amazonian bivalves phase with δ18O cycles in the adult part. If the two δ 13C cycles in the juvenile part represent one year each, then up to ﬁve years of growth are present in the record.

Pachydon tenuis #1 Santa Rosa de Pichana (96FWSR), Figure 5.8

No δ18O cycles are evident from the proﬁle, which renders the age determination of this specimen uncertain. The range of δ 18O values is low (typically 2-3). Irregular δ 13C cycles with very large amplitudes (up to 9) are found in the juvenile half of the analyzed shell (growth increment numbers 100-170, x-axis). Irregular δ 13C variation (amplitude typically 2-3) in the adult half of the shell cannot be interpreted in terms of seasonality.

0 A Triplodon corrugatus, Itaya River -2 a b c d e f g h i jk l m no p q r s -4

-6

-8

-10 O in ‰ V-PDB 18 δ -12 C, 13

δ -14 juvenile adult -16 300 250 200 150 100 50 0

+2 B Anodontites trapesialis, Playa Cocha

a b c d -4

O in ‰ V-PDB -10 18 δ C, 13 δ juvenile adult -16 80 70 60 50 40 30 20 10 0

Figure 5.5: Isotope records of modern analogues. (a) Triplodon corrugatus, a riverine bivalve, (b) Anodontites trapesialis, a ﬂoodplain lake bivalve. Values are expressed against V-PDB in (y-axis). Numbers on the x-axes are sample numbers. The number of sample lines of Anodontites trapesialis exceeds the number of growth lines observed. Oxygen records are plotted as black dots, carbon records are plotted as open circles. Interpreted seasonal cycles indicated with letters. 5.3 Results 63

Diplodon longulus Santa Rosa de Pichana (96FWSR), Figure 5.8

Five regular full δ18O cycles and the beginning of a sixth are apparent. The δ 13C cycles are somewhat less regular and in striking antiphase with the oxygen isotope cycles. The amplitude of both δ18O and δ13C is low (typically 1-2).

Pachydon tenuis, Nuevo Horizonte 0

-2 -4 -6 -8 -10 O in ‰ V-PDB

18 -12 δ

C, -14

13 juvenile adult δ -16 45 40 35 30 25 20 15 10 5 0

Diplodon longulus, Nuevo Horizonte 0 δ13C -2 δ18 -4 O -6 a b c d e f g h i j k l mn -8

O in ‰ V-PDB -10 18 δ -12 C,

13 -14

δ juvenile adult

-16 250 200 150 100 50 0 Anodontites capax, Los Chorros 0 a b? c? d e f -2 -4 -6 -8 O in ‰ V-PDB

18 -10 δ

C, -12 13

δ -14 juvenile adult

-16 180 160 140 120 100 80 60 40 20 0 sample numbers

Figure 5.6: Isotope proﬁles of analyzed Pebas Formation molluscs (legend as in Figure 5.5) 64 Ecological implications of Miocene western Amazonian bivalves

Pachydon erectus Santa Rosa de Pichana (98RK65c), Figure 5.9 Five δ18O cycles of low amplitude are interpreted, though one is uncertain (c in Figure 5.9). Irregular δ13C cycles lag the δ18O cycles. Amplitude ranges from 4 (juvenile) to 2 (adult stages). The proﬁles of P. tenuis #2 from the same layer (98RK65c) are similar.

Diplodon aff longulus #3 Mazán (99RKMz3), Figure 5.10 The δ18O and δ13C cycles are well developed and in phase. Slightly irregular δ 18O amplitudes range from approximately 3-4.5. From the oxygen isotope record, we estimate that eight dry seasons are present (a-h) in Figure 5.10) and thus that the clam lived for at least eight

Anodontites capax, San Antonio 0 a? b? c? d e f g? h i j k l? m n -2

-4

-6

-8 O in ‰ V-PDB 18

δ -10 C,

13 -12 δ

-14 juvenile adult -16 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0

δ13C δ18O Diplodon longulus, Sta Elena 0

-2 a? b? c d e -4

-6

-8

O in ‰ V-PDB -10 18 δ -12 C, 13 δ -14 juvenile adult -16 70 60 50 40 30 20 10 0

sample numbers

Figure 5.7: Isotope proﬁles of analyzed Pebas Formation molluscs (legend as in Figure 5.5) 5.3 Results 65

years. To visualize the growth rates, the distance of sampled lines of three specimens relative to the umbo are plotted in mm in Figure 5.11. Similar to the monitored recent A. trapesialis(Chapter 2), the isotopic cycles of the two fossil bivalves are more condensed toward the ventral margin, which demonstrates reduced growth rates in the adult stages. General differences exist between the isotope signatures of unionoid and pachydontine bivalves. The unionoids show clearly developed, usually regular cyclicity in both δ 18O and δ13C values. Within the Pachydontinae, cyclicity of both oxygen and carbon is irregular or even absent. Their oxygen isotope range is generally very small but has overall higher values than that for unionoids, whereas the pachydontine amplitude of δ 13C is erratic. In general, the δ13C values of unionoids increase and those of Pachydon decrease during on-

Diplodon longulus, Santa Rosa de Pichana 0

-2

-4 a b c d e -6

-8 O in ‰ V-PDB 18

δ -10

C, -12 13 δ -14 juvenile adult -16 70 60 50 40 30 20 10 0 sample numbers

δ13C δ18O

Pachydon tenuis, Santa Rosa de Pichana 0

-2

-4

-6

-8 O in ‰ V-PDB

18 -10 δ

C, -12 13 δ -14 juvenile adult -16 180 160 140 120 100 80 60 40 20 0 sample numbers

Figure 5.8: Isotope proﬁles of analyzed Pebas Formation molluscs (legend as in Figure 5.5) 66 Ecological implications of Miocene western Amazonian bivalves togeny.

5.4 Discussion

Absolute isotope values and cyclicity are quite similar in all measured bivalves from the three different pollen zones (Figure 1.7). The lack of an isotope trend in the different stratigraphic occurrences and the similarity with modern values suggest that the Pebas paleoenvi- ronment was climatically relatively stable and not very different than that of today. Unionoid

Pachydon tenuis, Santa Rosa de Pichana 0

-2

-4

-6

-8

O in ‰ V-PDB -10 18 δ -12 C, 13

δ -14 juvenile adult -16 45 40 35 30 25 20 15 10 5 0 sample numbers

δ13C δ18O

Pachydon erectus, Santa Rosa de Pichana 0 a b c? d e -2

-4

-6

-8

-10

O in ‰ V-PDB -12 18 δ -14 C,

13 juvenile adult δ -16 200 180 160 140 120 100 80 60 40 20 0

sample numbers

Figure 5.9: Isotope profiles of analyzed Pebas Formation molluscs (legend as in Figure 5.5) 5.4 Discussion 67 shells show regular, high-amplitude cyclicity in both δ 13C and δ18O , whereas cyclicity in pachydontines is at best erratic and sometimes absent (98HV21, 98RK65c, 96FWSR, 91FW70, Figures 5.6 — 5.10). This finding should not be attributed to taxonomical group- ing per se but rather appears to reflect differences in the biotope preferences of both groups.

Pachydon tenuis, Indiana I 0

-2

-4

-6

-8 δ13C -10 O in ‰ V-PDB δ18O 18

δ -12

C, -14

13 juvenile adult δ -16 25 20 15 10 5 0

Diplodon aff. longulus 2, Mazan 0

a b c d e f g hi -2

-4

-6

-8

O in ‰ V-PDB -10 18

δ -12

C, -14 13

δ juvenile adult -16 180 160 140 120 100 80 60 40 20 0

Diplodon aff. longulus 3, Mazan 0

a b c d e f g h -2

-4

-6

-8

O in ‰ V-PDB -10 18

δ -12

C, -14 13

δ juvenile adult -16 350 300 250 200 150 100 50 0

sample numbers

Figure 5.10: Isotope proﬁles of analyzed Pebas Formation molluscs (legend as in Figure 5.5) 68 Ecological implications of Miocene western Amazonian bivalves

Pachydontine bivalves are predominant in lacustrine settings. Unionoid bivalves are more common in ﬂuviolacustrine environments but also live in lacustrine conditions (Wesselingh

0 (a) direction of shell growth

-2 -4 -6 -8 O in ‰ V-PDB

18 -10 δ

C, -12 13 δ -14 -16 50 70 90 110 130 150

0 (b) δ13C -2 δ18O -4 -6 -8 -10 O in ‰ V-PDB 18

δ -12

C, -14 13 δ -16 15 17 19 21 23 25 27 29 31 33 35

-6 -8 -10

O in ‰ V-PDB -12 18 δ

C, -14 13 δ -16 31 36 41 46 51 56 61 66 71 [mm from umbo]

Figure 5.11: Modern and fossil examples of isotope records plotted against the distance from the umbo (in mm, x-axis). (a) Modern Triplodon corrugatus of Figure 5.5 (Itaya River, Iquitos Peru), (b) fossil Pachydon tenuis #1, (c) fossil Diplodon longulus of Figure 5.8 (Santa Rosa). Both modern and fossil shells show a decrease of wavelength toward the ventral margin, indicating decreased growth rates toward adult stages. 5.4 Discussion 69 et al., 2002). To assess the biological processes and settings from the stable isotope signatures of the bivalves, we ﬁrst address the incorporation of physical environmental signals in these signatures.

70°W r e iv R

o Chorros Los y a m

Colombia u t u P

71°W er er

Brazil

Peru

Riv ari Sta Elena

Sta Rosa Yav Pebas 72°W Pachydon tenuis

Amazon River Peru Itaya River

iver

R Indiana Playa Cocha po

73°W a Los Chorros, ~ 11 Ma Chorros, Los N Anodontites capax

r Mazán e Iquitos Western Western iv

Amazonia R

n San á Antonio z

a Nuevo Nauta M Horizonte 74°W 05°S 02°S 03°S 04°S Diplodon longulus Santa Elena, ~ 12 Ma Santa Elena, Diplodon longulus

Pachydon tenuis C O 13 18 Indiana, δ δ Santa Rosa de Pichana, ~ 14 Ma Santa Rosa de Pichana, Pachydon erectus Base Sta Rosa longulus aff. Diplodon

Mazan and Indiana, ~ 16 Ma Mazan and Indiana, Pachydon tenuis

Base Sta Rosa Diplodon longulus Mazan 3, 0‰ longulus -16‰ aff. San Antonio, ~ 12 Ma San Antonio, Diplodon Anodontites capax Nuevo Horizonte, ~ 11 Ma Horizonte, Nuevo Mazan 2 0‰ -16‰ Pachydon tenuis Pachydon longulus umbo ventral margin umbo ventral aff. 0‰

-16‰

O in ‰ V-PDB ‰ in O C,

δ Diplodon 18 13 Mazan 1 0‰ -16‰

Figure 5.12: Overview of Pebasian molluscs, oldest outcrops at the bottom, youngest at the top. Isotope graphs: all y-axes show δ13C and δ18O, scale from 0 to Ð16; all x-axes: sample numbers (not mm from umbo), from left to right represent from umbo to ventral margin. See Figure 5.1 for locations. 70 Ecological implications of Miocene western Amazonian bivalves

5.4.1 Stable oxygen isotope characteristics of Amazonian continental waters

In the Amazonian hydrological cycle, water vapor is transported from its Atlantic source by prevailing easterly winds toward the Andes mountain range, which prevents Pacific water vapor from entering Amazonia (Salati and Vose, 1984). Seasonal movement of the In- tertropical Convergence Zone (ITCZ) results in alternating wet and dry seasons (Figure 4.1 on page 48). There is a distinct difference between the wet and dry season δ 18O values of rainwater. High values prevail in the dry season and low values in the wet season, which is reflected in the δ18O profiles of freshwater bivalves (Chapter 2). Rivers and lakes are fed by precipitation and reflect the seasonal variation of δ 18O of precipitation. Especially when the residence time of surface waters exceeds a seasonal cycle, the amplitude of δ 18O variation is reduced (Burgman et al., 1987). Therefore, the seasonal δ 18O amplitude in a lake should be lower than that of its tributaries. When the dimensions of a lake, or a system of interconnected lakes, are large enough, further dampening of the δ 18O signal can occur because large water bodies in steady-state conditions create their own atmosphere, with average isotopic and relative humidity values that can be substantially different from those in surrounding areas. Precipitation from this local atmosphere does not comply with the seasonal variation observed in the surrounding land areas (Fontes and Gonfiantini, 1970). If evaporation exceeds precipitation, large differences in the isotopic composition between lakes and their confluents can develop. For example, more than 10 enrichment of δ 18O in Lake Titicaca was found relative to its tributaries (Fontes et al., 1979). These processes will cause molluscan shells from lakes and rivers to record different isotope profiles, even if they live in close proximity to one another and in similar climatic conditions. It can generally be expected that lacustrine molluscs exhibit dampened seasonal δ 18O variation compared with river molluscs, possibly with increased absolute values if evaporation exceeds precipitation. In the modern molluscs, this trend is reflected in the 6 seasonal variation of the river dweller T. corrugatus and the floodplain inhabitant A. trapesialis. Extreme evaporation during dry seasons in isolated Amazon floodplain lakes probably caused the ∼ 4 higher dry season δ18O values in A. trapesialis compared with the dry season values of T. corrugatus from a nearby tributary river. Due to this biotope effect, molluscs from lakes are less suitable than molluscs from rivers for accurate reconstructions of the seasonal variations of precipitation and evaporation. Variation in the amplitude of molluscan δ 18O is not uniquely a biotope effect. The strikingly similar δ 18O records of modern T. corrugatus and fossil D. aff. longulus from the oldest part studied in the Pebas Formation (Figure 4.4a, b, Figure 5.10) suggest that an annually migrating ITCZ that forced seasonality already existed in the late Early Miocene and that its intensity is comparable to that of today (Chapter 4). Changes in the orbitally forced annual latitudinal range of the ITCZ, as observed in Holocene climate records (Seltzer et al., 2000; Haug et al., 2001), must be expected to have occurred in the Miocene as well. In this study, we lack the stratigraphic resolution (Milankovitch scale) needed to address this option. Due to the near-equatorial latitude of our sample sites, however, we expect this orbital forcing to have had a limited effect on changes in the seasonal δ18O amplitude in the molluscs studied. Our data set does not enable us to see the global 5.4 Discussion 71

cooling trend that coincides with a major growth in the eastern Antarctic ice sheet in the period between 16.5 and 12.5 Ma, as is observed in bottom-dwelling deep-sea foraminifera (Zachos et al., 2001), mainly because of the relative insensitivity of tropical climates to global temperature change. Climate modeling experiments with quadrupled CO 2 levels (Renssen et al., 2004) show a temperature increase of only 1-2 ◦C in the tropics in contrast to 5-8◦C in higher latitudes. For our study, global climate signatures are further overprinted by the biotope effect. In the absence of riverine bivalves in our sample set from the younger part of the Pebas Formation, we are unable to isolate the climate signal in the δ 18O range of shells from that stratigraphic interval.

5.4.2 Variation of carbon isotopes in host water − Dissolved inorganic carbon (DIC) in water consists of three carbonic species: CO 2, HCO3 2− (bicarbonate), and CO3 , and their relative proportions depend on the pH of the water (Fig- ure 48). Its isotopic composition has been used to distinguish carbon originating from a variety of sources (Mook et al., 1974). In freshwater, DIC available to molluscs for shell building may enter the water through tributaries or ground water discharge, uptake of atmospheric CO2,CO2 in soils (almost entirely derived from decomposed organic matter), or dissolution of carbonates or chemical weathering of silicates Pawellek and Veizer (1994). Although the contribution of different sources of DIC may vary seasonally, the dominant source, by far, to surface water in Amazonia is decomposed, dead plant material. Sinks of DIC in water are: (1) photosynthesis by plants and algae (Pawellek et al., 2002); (2) outgassing of CO2 into the atmosphere, due to partial overpressure of CO 2 in a 13 water body; and (3) carbon ﬁxation in CaCO 3 (molluscs, ostracods, etc.). The δ C value of DIC is determined by the interplay of these sources and sinks. For example, seasonally changing ﬂuxes of decomposed plant material (or the relative contribution of C 3 and 13 C4 plants) can shift the δ CDIC. Likewise, photosynthesis results in an enrichment of the 13C of DIC because plants preferentially extract 12C from the DIC pool (e.g., (Cole et al., 1994)). Furthermore, outgassing of CO 2 into the atmosphere results in a shifting − CO2/HCO3 balance in the DIC pool and an increase in d13C because the fractionation be- − tween CO2 and HCO3 is approximately 8 (Usdowski and Hoefs, 1990; Pawellek and Veizer, 1994; Zhang et al., 1995; Luz et al., 1997; Barkan et al., 2001). CO 2 outgassing into or uptake from the atmosphere is partly controlled by the pH of the water. Particu- larly, in acidic waters (pH around 6.4), small shifts in the pH will change the concentrations 13 of bicarbonate and CO2(aq) and their respective δ C compositions dramatically. Unlike freshwater, marine water is buffered and has a pH of 8. As a result, 95% of the carbonic species appear in the form of bicarbonate (Figure 5.13), and marine δ 13C varies between 0 and 2 only.

5.4.3 Variation of carbon isotopes in bivalves Bivalves are generally believed to use bicarbonate for shell growth. Several authors men- tion disequilibria between bicarbonate and molluscan shell δ 13C (Fastovsky et al, 1993; 72 Ecological implications of Miocene western Amazonian bivalves

Dettman et al., 1999). Our work on A. trapesialis (Chapter 2) shows a constant fractionation of Ð4 between the δ13C of bicarbonate of the host water and the incorporated carbon in the shell of molluscs. Data by (Abell and Hoelzmann, 2000) suggest a similar fractionation in African freshwater molluscs. This fractionation tentatively can be used to reconstruct the δ13C variation of DIC of previous freshwater bodies. Applying this fractionation of Ð4 between bicarbonate and molluscan aragonite, the δ 13C in bivalves from seawater or water bodies in equilibrium with the atmosphere (with a long residence time) should be close to Ð4 (see the δ13C values of molluscs from Pebasian incursion levels in Vonhof et al., 1998). The low δ 13Cvalues observed in the analyzed shells suggest that continuous remineralization of organic carbon prevented Pebasian waters from equilibrating with the atmosphere. This observation is of relevance to the ongoing debate about Pebas Formation paleosalinity (Hoorn, 1994a; Räsänen et al., 1995; Webb, 1995; Vonhof et al., 1998, 2003; Gingras et al., 2002b; Wesselingh et al., 2002). Because typical marine δ 13C values of DIC range between 0 and 2, low molluscan δ 13C values and high amplitude seasonal cyclicity are at odds with seawater conditions.

- 2- [H2CO3] [HCO3 ] [CO3 ] 100

30°C 25°C 10°C % of Total Carbon 0 3 4 5 6 7 8 9 10 11 12 13 pH [H CO ] [HCO -] [CO 2-] 100 2 3 3 3

35‰ 5‰ 0‰ salinity % of Total Carbon

0 3 4 5 6 7 8 9 10 11 12 13 pH

Figure 5.13: Bjerrum plot of the concentrations of carbonate species at different pH. Left: Plotted for salinities of 0, 5, and 35 at a temperature of 25◦C. Right: Plotted for 10, 25, and 30◦Cat 0 salinity. 5.4 Discussion 73

5.4.4 Coupled carbon and oxygen isotope cyclicity 13 18 Numerous environmental factors are reﬂected in δ CDIC and δ O signals of the molluscan shell. Three types of coupled δ 13C and δ18O cyclicity are found in our fossils:

¥ δ18O and δ13C cycles in phase, as found in the modern ﬂoodplain dweller A. trapesialis, A. capax from San Antonio, and all D. aff. longulus from Mazán.

¥ Antiphase δ18O and δ13C cyclicity, as is evident in D. longulus from Nuevo Hori- zonte and Santa Rosa de Pichana.

¥ δ18O and δ13C cycles partly in phase and partly out of phase. This type is best visible in A. capax from Los Chorros (Figure 5.6) and D. longulus from Santa Elena (Figure 5.7).

In-phase cyclicity of δ18O and δ13C in modern A. trapesialis from the Amazon floodplain has previously been attributed to increased photosynthetic activity and outgassing of CO 2 in dry season floodplain lakes (Chapter 2). Radical environmental changes between wet and dry seasons (e.g., a rapidly flowing river during the wet season and a stagnant floodplain lake in the dry season) cause the strong correlation between δ 18O and δ13C in these specimens of A. trapesialis. The strikingly similar isotope pattern observed in several specimens of D. longulus from the Mazán outcrops suggests that very similar changes in seasonal environmental conditions occurred for these shells. Less radical environmental change takes place in strictly riverine habitats of Amazonia, as shown by the modern T. corrugatus data. The δ18O variation is comparable to that in A. trapesialis, which is governed by regional rainfall variation. Seasonal δ13C variation, however, is more attenuated and shows little correlation to δ18O cycles. This finding appears to accord with relatively stable environmental conditions for T. corrugatus; its habitat remains riverine in both the wet and dry seasons. Antiphase δ18O and δ13C cycles have not been observed in modern shells in this study but do occur in fossil material. This finding again suggests seasonal environmental change that affects the DIC pool. One possibility could be increased stratification and remineralization of organic matter in bottom water during the dry season. Another option might be seasonal switching of water sources between more local and more distal parts of the drainage basin with different C3/C4 plant distributions. Poor correlation between δ18O and δ13C is found in all Pachydontinae studied (with the possible exception of P. erectus from Santa Rosa, which shows up to 3 variation in δ 18O, Figure 5.8). Most Pachydontinae combine a relatively flat δ 18O pattern with significant variation in δ 13C. A flat δ18O pattern may occur if the Pachydontinae accreted shell material only during times in successive years when the chemical conditions were optimal. Seasonal growth stops can be caused by a variety of processes (e.g., growth cessation during unfavorable summer and/or winter conditions, spawning, monthly lunar cycles; Fastovsky et al., 1993; Abell et al., 1995). However, no indications of extensive or repeated growth stops in the growth banding of these Pachydontinae have been found, which makes growth in specific seasons of successive years unlikely. Because it is also unlikely that the analyzed pachydontine 74 Ecological implications of Miocene western Amazonian bivalves specimens all grew in less than one season, their uniform δ 18O pattern suggests that their habitat had little seasonal δ18O variation. This claim is in good agreement with the inferred lacustrine ecology of the Pachydontinae (Wesselingh et al., 2002), because lakes are most likely to have reduced seasonal δ 18O variations. Cyclic δ13C variation in these Pachy- dontinae is often comparable to the (seasonal) δ 13C range observed in other genera, which seems in good agreement with the preceding assumption that the longevity of Pachydon sp. exceeded one year in the Pebas system.

5.4.5 Consistency of isotope signals In four cases, we measured more than one shell from a single bed. Three Diplodon specimens from Mazán (98RKMz#1, 2, 3) show very similar isotope signatures in their shells, as do two Pachydon specimens from Santa Rosa de Pichana (level 98RK65c). However, the signatures of D. longulus and P. tenuis from level 96FWSR (Santa Rosa de Pichana) differ. Both have low amplitude δ18O variation, but the well-defined cyclicity in Diplodon is lacking in Pachydon. Most obvious are the differences between D. longulus and P. tenuis from a complex and condensed layer (level 91FW70, Nuevo Horizonte). Although δ 13C values have comparable ranges, the δ 18O records differ approximately 7. As in Santa Rosa, well-defined cyclicity in Diplodon is lacking in Pachydon. The co-occurrence of well- preserved shells with a lacustrine signal and a fluvial signal within a single bed points to the proximity of these depositional environments. Together with the indications of general shallow depositional conditions and swamps from sedimentary structures, rooted lignites, and common Mauritia pollen, we take this as evidence of an amalgamous lake/wetland depositional setting for the Pebas Formation, similar to that proposed by Nuttall (1990) but with predominantly freshwater rather than brackish chemistry. The shallowness did not interfere with the continuity of the ecosystem, as can be witnessed by the geographic widespread and continuous stratigraphic presence of a variety of endemic molluscan and ostracod lineages (Whatley et al., 1998; Wesselingh et al., 2002).

5.4.6 Trace element variation To evaluate possible seasonality in molluscan trace element concentrations, we measured Sr, Ba, and Fe from four fossil specimens (three different genera) at variable sample distances but in direct comparison with stable isotope data from the same samples (Figure 5.14 — 5.17). Strontium, a relatively conservative element in modern Amazonian freshwater, is believed to vary mainly with the precipitation/evaporation balance of the waters in which the shells live, as is particularly clear in Figure 5.15, in which the clearly cyclic (seasonal) δ18O pattern of two specimens of D. longulus is in phase with the variation in Sr concentration ([Sr]) in the shell. The season of highest evaporation (= high δ 18O values) coincides with highest Sr concentrations, which indicates that the [Sr] in the shells is proportional to the [Sr] in the water. From this shell, we also learn that considerable growth rate reduction through its ontogeny appears to have had little effect on the [Sr]; the amplitude of variation is comparable in the juvenile and full-grown shells. This ﬁnding is at odds with a kinetic 5.4 Discussion 75

control on [Sr] observed in a previous study of marine shells (Stecher et al., 1996). The interpretation of [Sr] variation in the other genera is less straightforward, mainly because the specimens do not exhibit a clear cyclic δ 18O pattern. Several general observations can be made: The Pachydon specimen (Figure 5.16) shows considerable [Sr] variation that seems in antiphase with δ13C variation in the juvenile part of the shell. It is unclear what drives this antiphase relationship. The Anodontites specimen (Figure 5.14) has the lowest [Sr]

Anodontites capax, Los Chorros 1700 0

1600 -2

[Sr] (ppm) 1500 -4 1400 -6 1300

1200 -8

1100 -10

1000 O in ‰ V-PDB

-12 18 900 Sr δ -14 800 C, 13 towards umbo towards ventral margin 700 -16 δ 182 162 142 122 102 82 62 42 22 2

450 0

400 -2

350 -4

300 [Ba] (ppm) -6 250 -8 200 -10 150 O in ‰ V-PDB

-12 18

100 δ Ba -14 50 C, 13

0 -16 δ 182 162 142 122 102 82 62 42 22 2

2500 0

-2 2000 -4

[Fe] (ppm) 1500 -6

-8

1000 -10 O in ‰ V-PDB

-12 18 500 δ

Fe -14 C, 13

0 -16 δ 182 162 142 122 102 82 62 42 22 2

sample numbers

Figure 5.14: Minor elements Sr, Ba, and Fe records (thick lines) from growth increments of Anodon- tites capax (Los Chorros) compared with isotope data (δ13C: open dots, δ18O: black dots). Left axis: ppm/Ca, right axes: δ13C, δ18Oin versus PDB, x-axis: isotope sample numbers. Not all samples have been analyzed for elements. 76 Ecological implications of Miocene western Amazonian bivalves with an unclear relation to both δ 13C and δ18O.

Diplodon longulus, Santa Rosa de Pichana

1600 -6

[Sr] (ppm) 1500 -7

1400 -8

1300 -9 O in ‰ V-PDB 1200 -10 18 Sr δ

1100 -11 C, 13 δ towards umbo towards ventral margin 1000 -12 66 61 56 51 46 41 36 31 26 21 16 11 6 1

-6 140

-7 120

[Ba] (ppm) 100 -8

80 -9

-10 O in ‰ V-PDB

40 18 δ -11

20 Ba C, 13

0 -12 δ 66 61 56 51 46 41 36 31 26 21 16 11 6 1

200 -6

180

-7 160

140 -8 [Fe] (ppm) 120 100 Fe -9

80 -10

60 O in ‰ V-PDB 18

40 δ -11 20 C,

0 -12 13 66 61 56 51 46 41 36 31 26 21 16 11 6 1 δ

sample numbers

Figure 5.15: Minor elements Sr, Ba, and Fe records (thick lines) from growth increments of Diplodon longulus (Santa Rosa) compared with isotope data (δ13C: open dots, δ18O: black dots). Left axis: ppm/Ca, right axes: δ13C, δ18Oin versus PDB, x-axis: isotope sample numbers. Not all samples have been analyzed for elements.

Iron typically is an element that yields information about the oxygenation state of the water and bottom sediments. In the presence of oxygen, dissolved Fe will be removed from the water and unavailable to be trapped in the shells. Consequently, at lower oxygen levels, Fe is released in the water column and can be incorporated in the shells. This redox behavior of Fe is observed in a modern specimen of Anodontites sp. from a ﬂoodplain environment 5.4 Discussion 77

in which episodic dysoxia is common (Figure 5.14) and seems present in the Anodontites and Pachydon specimens and the D. aff. longulus from Mazán. These shells show low base levels of [ Fe] with occasional spikes that presumably represent brief periods of lower bottom water oxygenation. The amount of [Fe] data from the Santa Rosa D. longulus is too limited to yield clear patterns but sufﬁcient to show generally low [Fe] values. Barium is associated with particulate organic matter in the water column. Classic (paleo-

Pachydon tenuis #2, Santa Rosa de Pichana

1800 -4

1700 -6 1600 -8 1500 [Sr] (ppm) Sr 1400 -10

1300 O in ‰ V-PDB

-12 18 δ 1200

-14 C,

1100 13 δ towards umbo towards ventral margin 1000 -16 171 161 151 141 131 121 111 101 91 81 71 61 51 41 31 21 11 1

350 -4

300 -6

250 -8 200 Ba [Ba] (ppm) -10 150 -12 100 O in ‰ V-PDB 18 δ 50 -14 C, 13

0 -16 δ 171 161 151 141 131 121 111 101 91 81 71 61 51 41 31 21 11 1

700 -4

600 -6

500 -8 [Fe] (ppm) 400 -10 300 -12 200 O in ‰ V-PDB

100 -14 18 Fe δ

0 -16 C,

171 161 151 141 131 121 111 101 91 81 71 61 51 41 31 21 11 1 13 δ

sample numbers

Figure 5.16: Minor elements Sr, Ba, and Fe records (thick lines) from growth increments of Pachydon tenuis #2 (Santa Rosa) compared with isotope data (δ13C: open dots, δ18O: black dots). Left axis: ppm/Ca, right axes: δ13C, δ18Oin versus PDB, x-axis: isotope sample numbers. Not all samples have been analyzed for elements. 78 Ecological implications of Miocene western Amazonian bivalves ceanographic) applications of [Ba] data from the sediment column suggest increased [Ba] reﬂects increased organic particle ﬂuxes caused by surface water productivity (e.g., Stecher et al., 1996). During remineralization of organic matter, Ba, whether adsorbed or in the form of barite, is dissolved in the bottom water and incorporated in the shells. A similar interpretation of [ Ba] data in Pebas shells is conceivable. In D. aff. longulus from Mazán, a seasonal Ba cycle is evident (Figure 5.17), which suggests maximum productivity dur-

Diplodon aff. longulus #2, Mazan 3000 0 161 151 141 131 121 111 101 91 81 71 61 51 41 31 21 11 1

-2 2500 Sr

-4 2000

-6

1500

-8

[Sr] (ppm) O in ‰ V-PDB 1000 -10 18 δ

500 C,

-12 13 δ

towards umbo towards ventral margin 0 -14

350 0 161 151 141 131 121 111 101 91 81 71 61 51 41 31 21 11 1

300 Ba -2

250 -4

200 -6

[Ba] (ppm) 150 -8 O in ‰ V-PDB 18

100 -10 δ C,

50 -12 13 δ

0 -14

100 0 161 151 141 131 121 111 101 91 81 71 61 51 41 31 21 11 1

90 -2 80

70 -4

60 [Fe] (ppm) -6 50

-8 40 O in ‰ V-PDB 30 -10 18 δ 20 Fe C, -12 10 13 δ

0 -14 sample numbers

Figure 5.17: Minor elements Sr, Ba, and Fe records (thick lines) from growth increments of Diplodon aff. longulus (Mazán) compared with isotope data (δ13C: open dots, δ18O: black dots). Left axis: ppm/Ca, right axes: δ13C, δ18Oin versus PDB, x-axis: isotope sample numbers. Not all samples have been analyzed for elements. 5.4 Discussion 79

ppm waterdepth (cm)

2000 600

1800 10*[Ba] ppm [Fe] ppm [Sr] ppm 500 1600 depth (cm)

1400 400

1200

1000 300

800

200 600

400 100

200

0 0 26-09-1998 13-10-1998 11-02-1998 15-11-1998 15-12-1998 31-01-1999 15-02-1999 03-02-1999 16-03-1999 31-03-1999 16-04-1999 08-05-1999 24-05-1999 08-06-1999 26-06-1999 14-07-1999 31-07-1999 15-08-1999 02-09-1999 16-09-1999 01-10-1999 15-10-1999 20-10-1999

Figure 5.18: Concentrations of minor elements dissolved in Amazon/Playa Cocha water. note: Bar- ium concentrations are exaggerated 10 times.

ing dry season ﬂoodplain conditions. For the other specimens, [Ba] does not correlate with δ13C or other chemical data. Generally high [Ba] in Pebas shells in comparison with [Ba] in marine taxa may pertain to relatively high rates of remineralization of particulate organic matter in their biotope. The analyzed P. tenuis from Santa Rosa shows the highest [Ba] of all, which may relate to its suggested tolerance of dysoxic bottom water (Wesselingh et al., 2002). The preceding discussion shows that trace element patterns of Pebasian shells are not always easily interpreted. [Sr] exhibits seasonal cyclicity most clearly in at least one of the shells; growth rate (kinetic) effects seem to play a subordinate role. The positive correlation of [Sr] with δ18O values bolsters the wet-dry season interpretation of the δ 18O cycles. Large seasonal changes in the trace element chemistry of the host water have been observed in modern Amazonian ﬂoodplain environments (Kaandorp, unpublished data) which inhibit the use of molluscan trace element ratios as paleothermometry tools (cf. Klein et al., 1996).

Episodic redox gradients at the sediment water interface are believed to result in Fe peaks in the molluscs. An important consequence of this observation is that the traditional use of elevated Fe concentrations to detect diagenetic alteration is of limited use for shells from this kind of environment. The potentially complex behavior of Ba in an environment 80 Ecological implications of Miocene western Amazonian bivalves so rich in organic matter seems to limit its use as a simple paleoproductivity proxy, as it can be in marine environments (Stecher et al., 1996), though seasonal Ba variation is suggested to occur in one of the analyzed shells (Figure 5.15).

5.5 Conclusions

Stable isotope profiles of Miocene bivalves from the Pebas Formation of western Amazo- nia provide considerable information about their biotope preferences and growth regimes and yield wider implications for the nature of the Pebas ecosystem as a whole. The new data accord with those discussed in Chapter 4 thus support a monsoon system throughout the Middle and Late Miocene. Different δ 18O amplitudes between the specimens studied are correlated with their different biotopes, as interpreted from faunal assemblages. Pachy- dontine bivalve isotope profiles, characterized by low amplitudes and irregular or absent cyclicity, are interpreted as representing lacustrine living modes. These shells were accompanied by very low numbers of pearly freshwater mussels, which dominated fluvial- influenced environments in the Pebas system, where Pachydontinae were scarce or absent. These unionoids yield isotope signals comparable to those of modern Amazonian floodplain unionoids, characterized by relatively high amplitude and regular seasonal variation. ¥ Lower seasonal δ18O amplitude for lake dwellers compared with river dwellers is interpreted as caused by the longer residence time of lake water, which buffered seasonal δ18O of rainfall in the area. Therefore, molluscs from lakes are less suitable than molluscs from rivers for accurate reconstructions of the seasonal variation of precipitation and evaporation. ¥ Intra-annual growth rate variation in fossil bivalves could not be assessed because the relative length of the wet and dry season during the Miocene could not be established independently. The studied fossil pearly freshwater mussels would have reached ages of 6-10 or more years. An age determination for Pachydon is more complicated because of the low seasonal variation in the skeletal chemical record. An expected long lifespan, based on growth line counts from Pachydontinae (4-15 years), would be confirmed if the variation in [Sr] and δ 13C is regarded as a seasonal chemical variation. ¥ Surface waters were not in carbon isotopic equilibrium with the atmosphere. The δ13C records of all measured Pebasian bivalves have minima below Ð10 (similar to modern Amazonian freshwater shells), which argues for a significant terrestrial carbon input in the system and simultaneously against seawater influence. ¥ Although only limited molluscan trace element data have been available until now, considerable seasonal variation in trace element concentrations is evident. The variation observed is tentatively interpreted as reflecting seasonal change in host water chemistry. No clear indications were found for any kinetic control on molluscan trace element distribution. Seasonally variable trace element chemistry of the host water 5.6 Appendix 81

inhibits the use of trace elements ratios as paleothermometry tools (cf. Klein et al., 1996).

¥ Isotopic signals are frequently used in climate reconstructions of the Miocene record (Savin et al., 1975; Zachos et al., 2001). On their basis, the Middle-Late Miocene (16.5-12.5 Ma) global climate is believed to represent a cooling trend that culminated in the Pleistocene ice ages. The isotopic signal of the mollusc shells from Miocene Amazonia does not conﬁrm or deny this scenario, probably due to the relative insensitivity of the tropical latitudes to global climate change.

¥ Although orbitally forced variation in ITCZ migration is likely to have taken place in the Middle-Late Miocene, the near-equator position of our sampling sites prevents us from confirming that this variation affected the amplitude of molluscan δ 18O profiles. Furthermore, our data set does not provide clear evidence that major changes in seasonality took place in the Middle-Late Miocene. However, we can conclude that a climatic system with comparable seasonality to that of today was already in place during the late Early Miocene, which is particularly important because it enhances our knowledge about the climatic parameters that ruled Amazonia during the Miocene and confirms that, at the time, the climatic conditions needed for a tropical rainforest were in place.

5.6 Appendix

Two living shells reflecting potential depositional environments within the Pebas Forma- tion were used for comparison. Triplodon corrugatus (Lamarck, 1819) was collected alive from Itaya River (03◦47’08"S, 73◦17’08"W), a tributary of the Amazon River (Chapter 4). The Itaya River is a small black water river (∼ 100 km in length) that drains a small tropical rainforest area and is fed by local rainwater. Specimens of the pearly freshwater mussel Anodontites trapesialis (Lamarck, 1819) were monitored for a period of 13 months in Playa Cocha, an Amazon floodplain lake near Iquitos (03 ◦45’32"S, 73◦10’47"W) (Chap- ter 2). The shells adequately record changes in host water oxygen isotopes. Carbon isotopes showed an offset of 4 lower than the bicarbonate of the host water. During the experiment, the floodplain lake was part of the Amazon River in the wet season (February-July) but separated from the Amazon and subsequently subject to evaporation in the remainder of the year. Five bivalve species from the Pebas Formation are used in measurements: An- odontites capax (Conrad, 1874), Diplodon longulus (Conrad, 1871), Diplodon aff. longulus (Conrad, 1871), Pachydon tenuis Gabb, 1869, and Pachydon erectus Conrad, 1871. The shells were collected during field campaigns in 1993 and 1997 by Wesselingh and 1998 and 1999 by Kaandorp, unless stated otherwise. Zonal names refer to the palynological zonation of Hoorn (1994b). Material is from the following localities: 82 Ecological implications of Miocene western Amazonian bivalves

Los Chorros (89CH) 3◦47’S; 70◦20’W, Grimsdalea Zone, ∼ 11 Ma (Figure 5.19). The analyzed A. capax from this outcrop was collected by Hoorn in 1991 in a layer containing fauna that yielded a ﬂuvial or ﬂuviolacustrine assemblage dominated by pearly freshwater mussels and the gastropod Aylacostoma browni (Etheridge, 1879). Many freshwater mussels in the layer possessed both valves, indicating minimal transport.

Nuevo Horizonte (91FW70), 4◦05’S; 73◦25’W, Grimsdalea Zone, ∼ 11 Ma (Figure 5.19). A P.tenuis and D. longulus were sampled from a layer containing an assemblage of a very diverse and well-preserved endemic lacustrine Pebas fauna, together with a few reasonably well-preserved marginal marine molluscs (Melongena woodwardii (Roxo de Oliveira, 1924) and Nassarius reductus (Vermeij and Wesselingh, 2002), as well as abundant but poorly preserved foraminifers (Ammonia and Elphidium spp.) and poorly preserved ﬂuvial snails (Aylacostoma browni (Etheridge, 1879)). Because of their excellent preservation, the analyzed molluscs are considered to belong to the lacustrine faunal component of this layer

Legend Los Chorros I Nuevo Horizonte Santa Rosa Mazan II de Pichana sand

silt

clay and silt

lignite

organic rich clay 89CH 99RKMZ concretions 91FW70 shells

roots Indiana I San Antonio/ irregular contact Itaya I

not exposed 98HV21 M mottling 96FW810 burrow

bone 96FWSR

wood remains

98RK65

Figure 5.19: Stratigraphic columns of the outcrops. Sample numbers on the right indicate the beds from which the bivalves were collected. 5.6 Appendix 83 n 63 21 65 68 45 42 231 286 168 189 338 170 164 197 174 2: level of co- 2 C 13 δ Min max range -15.28 -7.97 7.31 -14.32 -10.30 4.02 -11.23 -5.83 5.40 -10.79 -8.70 2.09 -11.19 -3.89 7.30 -11.75 -6.82 4.93 -13.10 -9.21 3.89 -14.10 -1.12 12.98 -10.29 -5.38-12.43 4.90 -6.06 6.37 -14.88 -5.41 3.10 -11.06 -6.83 4.23 -11.55 -3.03 8.53 -10.63 -5.44 5.19 -10.47 -5.08 5.39 2 R 0.00 0.06 0.19 0.38 0.81 0.05 0.00 0.27 0.29 0.70 0.27 0.00 0.32 0.01 0.25 (C vs. O) O max. 18 δ min max range ampl. -7.39 -3.98 3.41 3.4 -6.79 -1.39 5.40 3 -8.63 -6.32 2.31 2 -7.07 -4.96 2.11 2 -2.99 -0.80 2.19 2.2 -9.91 0.37 10.29 9 -7.89 -4.78 3.10 2.5 -5.81 -3.40 2.42 2.4 -8.56 -2.66 5.90 5 -4.87 -1.33 3.54 3 -5.35 -1.09 4.26 3 -10.07 -6.40 3.67 3.4 -10.68 -3.92 6.76 6 -10.55 -3.69 6.86 6.5 -10.51 -1.91 8.61 5 3 9 19 ? ? ? 3 5 5 3 8 11 13 4.5 > < 4.5? > Cycles ? L L L L L L L FL FL Env. ML/FL ML/FL ML/FL ML/FL ML/FL Mazán Mazán Mazán Sta Rosa Sta Rosa Location Sta Rosa Indiana I Sta Rosa Sta Elena Itaya River Los Chorros San Antonio Playa Cocha Nuevo Horizonte Nuevo Horizonte #1 #3 #2 #1 #2 longulus longulus longulus Name A. capax P. tenuis P. tenuis A. capax P. erectus D. longulus D. longulus D. longulus P. tenuis P. tenuis aff. aff. aff. A. trapesialis T. corrugatus D. D. D. F810 89CH Sample 981115 98RK65 98RK65 98HV21 91FW70 91FW70 96FWSE 96FWSR 96FWSR 991021-19 99RKMz 1 99RKMz 2 99RKMz 3 Table 5.1: Isotope data ofmax.: the minimum measured and bivalves. maximum values Notes: of env: all isotope inferred measurements environment performed from on sediment the and/or growth faunal increments context of (Appendix); a min single and specimen; R variance between carbon and oxygen stable(shore isotopes; zone, n: intra-lacustrine number swamp, ofand measurements backswamp, performed river interdistributary channels). on one bay, shell. and L: outer lacustrine, deltaic ML: marginal environments), lacustrine FL: ﬂuviolacustrine (ﬂoodplain lakes 84 Ecological implications of Miocene western Amazonian bivalves

(see Table 5.1).

Santa Elena, (96FWSE) 3◦52’S; 71◦23’W, Grimsdalea Zone (Hoorn, 1994b), ∼ 12 Ma (no column). The section, 500 m west of the landing stage (harbor) of Santa Elena, yields both lignitic clay layers with ﬂuvially inﬂuenced faunas and sandy clays with lacustrine faunas. The exact origin of the sampled D. longulus is uncertain because it was collected from a surface scree.

San Antonio (96F810) 4◦02’S; 73◦23’W, Grimsdalea Zone, ∼ 12 Ma (Figure 5.19). According to the classification of Wesselingh et al. (2002), a bulk sample from this outcrop has been attributed to the "tall-Dyris assemblage," which represents upper foreshore lacustrine depositional environments. However, the combination of common A. capax (on which isotope measurements are performed) and the paucity of Pachydontinae indicate this sediment was deposited in the vicinity of a fluvial influence. The shell bed from which the sample was collected overlaid an erosive discontinuity, and shells are randomly oriented and size sorted, which indicates transport.

Santa Rosa, (98RK65, 96FWSR) 3 ◦40’S; 71◦46’W, Crassoretitriletes Zone, ∼ 14 Ma (Figure 5.19). Four specimens sampled from two different layers were measured: P. erectus and P. tenuis from 98RK65 and D. longulus and P. tenuis specimens from 96FWSR. Both layers contain doublets of bivalves and yield typical lacustrine assemblages.

Indiana (98HV21) 3◦30’S; 73◦02’W Psiladiporites-Crototricolpites Zone, ∼ 16 Ma (Figure 5.19). Slightly decalciﬁed Pebasian pachydontines and cochliopines (gastropods) in combination with well-preserved doublets of D. aff. longulus in blackish ﬁne-grained sediment indicate an interdistributary bay to delta front environment. A P. tenuis has been sampled and measured.

Mazán (99RKMZ1,2,3) 3◦30’S; 73◦06’W, Psiladiporites-Crototricolpites Zone, ∼ 16 Ma (Figure 5.19). Three specimens of D. aff. longulus were collected from sandy clay on top of a lignite at the conﬂuence of the Mazán and Napo rivers near the village of Mazán. The lagoonal muds and peat are correlated with the base of sequence B from Räsänen et al. (1998). Chapter 6

Aquatic landscapes in the Miocene of western Amazonia

Abstract The Miocene Pebas Formation from the section Santa Rosa de Pichana (Loreto, Peru) was investigated. Its sedimentary facies, molluscan communities and taphonomy were studied, and in addition the stable isotopes of both entire shells and growth bands in bivalves were analyzed. Three sequences, comprising a succession of transgressive, maximum ﬂooding and regressive/prograding intervals, are documented. Molluscs are most common in the transgressive/highstand intervals and are almost absent in regressive/prograding intervals. The fauna is dominated by endemic Pebasian species, such as Pachydon and Dyris spp. The nature of the deposits as well as the availability of oxygen varied in a predictable way within each of the sequences and determined the nature of the assemblages. Highest diversity was reached in the late transgressive phase before the development of dysoxia that was widespread during the late highstand and early regressive/ prograding phase. In contrast to ichnofossils found in the section, the mollusc and isotope data show no indications of elevated salinities. This discrepancy is interpreted to result either from temporal separation of the ichnofossils and the mollusc fossils or from evolution beyond usual ecological tolerances of taxa that produced these ichnofossils into freshwater settings. Key words — Miocene, Amazonia, paleoecology, sedimentology, stable isotopes

6.1 Introduction

Ecological conditions during deposition of the Pebas Formation (Miocene, western Ama- zonia) have been the subject of scientiﬁc debate ever since the publication of the ﬁrst fossil

This chapter is adapted from Wesselingh, Kaandorp, Vonhof, Räsänen, and Renema 2006b. The nature of aquatic landscapes in the Miocene of western Amazonia: an integrated palaeontological and geochemical approach, Scripta Geologica, 133: in press. 86 Aquatic landscapes in the Miocene of western Amazonia faunas from these strata by Gabb (1869). Interpreted depositional settings ranged from fluviolacustrine (see: Figure 1.14 on page 14) to long-lived lake to a seaway (see Wesselingh et al., 2002 for references). Only recently two papers were published that interpreted the same successions very differently. Gingras et al. (2002) interpreted the common occurrence of mesohaline-polyhaline settings during deposition of the Pebas Formation based on ichnofossils and tidal sedimentary structures. In contrast, Wesselingh et al. (2002) argued for a long-lived system of predominantly freshwater lakes and swamps with only very limited marine influence, up to oligohaline at its maximum; conclusions derived from the molluscan faunas, strontium, carbon and oxygen isotope signatures in the shells. The conflicting interpretations were partially based on the same sections. Subsequently, more information has become available about possible depositional settings in the Pebas Formation (Vonhof et al., 2003; Wesselingh, 2006a, b; Chapters 4 and 5), adding to rather than resolving the confusion. Wesselingh et al, (2006b) aims to document environmental settings during deposition of Pebas Formation intervals in a single outcrop (Santa Rosa de Pichana, Loreto, Peru). This outcrop contains a succession of predominant characteristic alternating fining and coarsening up sedimentary sequences (Figure 6.1). It is almost certainly the locality known as ’Pichua’ or ’Pichana’ (see Nuttall, 1990) from which some of the oldest published faunas from the Pebas Formation originate (Woodward, 1871; Conrad, 1871, 1874). The section of Santa Rosa is located in the Middle Miocene Crassoretitriletes zone of Hoorn 1993 and in Molluscan Zone MZ7 of Wesselingh et al. (2006a), and is of similar age as the outcrops near Pebas. The locality contains some of the most regularly developed fining and coarsening up stratigraphic sequences that are very common throughout the Pebas Formation and is considered to be a representative example of Pebas Formation deposits. This Chapter will focus on finding evidence for elevated salinity in the Pebas Formation in molluscan chemistry.

6.2 Material and methods

The exposure at Santa Rosa de Pichana is located on the west bank of the Amazon (3 ◦40’04"S, 71◦45’58"W; Figure 6.1), south of the conﬂuence of the Rio Pichana and the Amazon. Here, the Pebas Formation strata are exposed continuously over approximately 1 km length. The section was logged and sampled c. 200 m south of this conﬂuence. We visited the outcrop on a number of occasions; Wesselingh and Räsänen (in 1996), Kaandorp and Vonhof (in 1998), and Räsänen and Gingras (in 1999). Räsänen performed sedimentary facies descriptions for the upper 16 m of the outcrop; Vonhof and Kaandorp made crude lithological descriptions of the lowest 3 m c. 50 m southward. We have applied the term ’sequence’ throughout this paper, although the sequences described herein might also be termed ’parasequence’, especially when comparing to the sequence terminol- ogy applied by Wesselingh et al. (2006b). A brief assessment of ichnological content was conducted only; it is likely that in some intervals ichnofossils may have been overlooked. Mollusc samples were collected by Wesselingh (usually taken at c. 20 cm intervals, typi- 6.2 Material and methods 87

Wear — 1. No traces of abrasion. 2. Few traces of abrasion at vulnerable parts such as the umbo and shell edges of bivalves and the apex of gastropods. 3. Umbones/edges worn; abrasion (corrugation or polishing) also apparent on other parts of the shell’s surface. 4. Umbones/apex heavily worn or entirely eroded, surface worn to strongly worn (usually corrugate). 5. Only strongly worn fragments.

Fragmentation — 1. All or almost all shells entire (apart for shells obviously broken during sampling/washing). 2. Whole shells more common than damaged shells and fragments. 3. Whole shells about as common as damaged shells and fragments. 4. Damaged shells and fragments more common than whole shells. 5. Fragments/damaged shells only.

Dissolution marks — 1. No dissolution marks. 2. Superﬁcial etching marks rare. 3. Surface (partially) etched, few, usually shallow dissolution holes. 4. Surface etched to strongly etched, dissolution holes common and of variable depth. 5. Surface strongly etched with common very deep dissolution holes.

Periostracum — 1. Common (>10% of shells). 2. Occur (5-10% of shells). 3. Rare (1-5% of shells). 4. Very rare (<1% of shells). 5. Absent.

Shell integrity — 1. Shell surface robust, commonly (partially) translucent. 2. Shell surface robust but dull. 3. Shell surface dull, parts may be scratched. 4. Shell surface soft, easily scratchable. 5. Shells fall apart when touched.

Colour mixing — 1. All shells and fragments of the same colour. 2. One dominant colour, rarely additional colours. 3. One dominant colour, common additional colours or two colours common. 4. One or two common colours and several additional colours. 5. Various different colours, none predominant.

Mixed preservation — 1. All shells and fragments with about the same colour, shell integrity, wear and dissolution status. 2. Almost all shells and fragments with about the same colour, shell integrity, wear and dissolution status or slight variation in preservation characteristics only. 3. Some variation in colour, shell integrity, wear and/or dissolution status. 4. Common variation in colour, shell integrity, wear and/or dissolution status. 5. Strong variation in colour, shell integrity, wear and dissolution status.

Table 6.1: Taphonomic criteria. 88 Aquatic landscapes in the Miocene of western Amazonia

Figure 6.1: Map of the surroundings of Santa Rosa. cally containing 1 kg of sediment and indicated with F numbers), and Kaandorp and Vonhof (RK numbers, containing 200-300 grams).

The samples that consisted of unconsolidated sedimentary rock were washed (minimum sieve mesh 1 mm), and the washing residues were qualitatively assessed for taphonomic parameters (Table 6.2). The scores were summed in a taphonomic index to provide a crude indication of the quality of preservation of samples; low scores indicate good preservation, high scores indicate poor preservation of faunas. The species-sample matrix was subjected to a variety of analyses, using PC-ORD (Mc- Cune and Medford, 1999). The taxonomy follows Wesselingh (2006a). Simpson’s diversity index, Shannon’s diversity index and evenness (Etter, 1999) were computed. The sample- species matrix was subjected to cluster analysis (both the raw and log transformed (y = log (x+1)) data, using Bray-Curtis similarity and ﬂexible β clustering methods (β = -0.75). Indi- cator species analyses (Dufrêne and Legendre, 1997) were run using raw data and excluding rare species (species occurring only in a single sample). Non-metric multidimensional scal- ing (NMDS; Etter, 1999) was run using Bray-Curtis similarity index of the log-transformed data set in order to investigate the spatial dimensions of variation between samples and between species. Two types of stable oxygen and carbon isotope analyses were performed: analyses of whole or substantial parts of whole shells (for methodology, see Vonhof et al., 1998); and isotope analyses of successive growth bands within shells (for methodology, see Chapter 4). Whole-shell analyses were performed on 133 shells. Two bivalves from level 98RK65 and two bivalves from the interval F532-F537 (indicated by 96FWSR) were cut, and isotopic analyses were performed along growth bands. The isotope data of the four 6.3 Description of the section and facies interpretation 89

bivalves and trace element data of two of them (see: Chapter 5 are discussed below in the light of the present study. are discussed below. Two additional 87Sr/86Sr ratios measured on shells from this section were available from Vonhof et al. (2003).

6.3 Description of the section and facies interpretation

The studied section is over 19 m thick (Figure 6.2). Some meters of section exposed at the top could not be reached for study. The section yields three complete, 4-6.5 meter thick depositional sequences (II-IV), as well as the top of a fourth (I) and the base of a fifth (V). The lowest complete sequence (Sequence II in Figure 6.2) starts at a transgressive surface of erosion (TSE) developed in probable backshore lagoonal muds of sequence I. The transgressive interval is predominantly composed of mudstone, but contains a thin basal sand layer and furthermore sand lenses, organic debris and shells. In the base (samples 98RK65 and 98RK66), strongly abraded shells and shell fragments as well as paired Pachydon bivalves were found, indicating that these basal layers contain both in situ and reworked specimens. The fine-grained maximum flooding interval (98RK69) contains paired Pachydon obliquus specimens and also dispersed organic debris. This interval turns gradually to a prograding low-energy wave-influenced bay / lake-margin sequence topped by backshore lagoonal mudstones with some rhizoliths.

Sequence III starts with an evident wave ravinement surface representing the TSE. From this level an assemblage assigned to the Glossifungites-ichnofacies (composed of Tha- lassinoides burrows, Figure 6.3a) is protruding into the underlying lagoonal mudstones of sequence II. The Glossifungites ichnofacies is an omission surface developed in hard or firm substrate (Bromley, 1990). The TSE is covered by transgressive deposits represented by reworked lignite grading laterally in the section into layered organic mudstone. The base of this layer has a lag of wood debris (including a tree trunk) and passes laterally into a c. 1 m deep channel. Also, the top of the lignite/organic mudstone is somewhat bioturbated and contains a thin mollusc lag deposit, made up of terrestrial snails and snail fragments (Peba- siconcha immanis). The lower part of the fining up transgressive interval contains the trace fossil Asterosoma (Figure 6.3b) and Planolites reburrowed by Chondrites (Figure 6.3c). The unit is topped by massive sandy siltstones. The maximum flooding interval contains massive mudstone with abundant shells (samples F532-F537) and some carbonate concretions. The fauna is very well preserved and dominated by various Pachydon species that are often found in situ, or with valves in close proximity, indicating some bioturbation, but no profound physical disturbance of the depositional environment. The maximum flooding interval grades upwards into a well developed, massive shoreface sand interval of almost 2 m thickness that is topped by a regressive backshore peat horizon. In the lower shoreface interval a thin organic layer penetrated by a minor Glossifungites ichnofacies with Thalassi- noides was observed together with some thin shell strings. The shoreface is capped by one of the thickest, well-developed pedogenic horizons encountered in the Pebas Formation. 90 Aquatic landscapes in the Miocene of western Amazonia

Figure 6.2: Stratigraphic log of the Santa Rosa de Pichana section with facies interpretations, depositional sequences and the location of mollusc samples. Key: peb = pebbles; b = bioturbated contact, g = gradual contact, i = interlayered contact, o = oxidized horizon, s = sharp contact, IMF = interval of maximum ﬂooding, RPI -= regressive/prograding interval, TI = transgressive interval, TSE = transgressive surface of erosion, G = Glossifungites,A=Asterosoma,P=Psilonichnus,Ar=Arenicolites, Gy = Gyrolithes,Pl=Planolites, @ indicates level with Amazoniconcha immanis, roman numbers refer to sequences.

Sequence IV is initiated with a well-developed Glossifungites ichnofacies penetrating through the entire pedogenized horizon of sequence III. Also, a horizontal deposited large Psilonichnus, with rhythmical sediment inﬁll, was observed penetrating the shoreface deposits. Similar burrows were previously interpreted to represent tubular tidalites, that is, open burrows that trapped tidal lamination (Gingras et al., 2002). The wave ravinement 6.3 Description of the section and facies interpretation 91

a c

Figure 6.3: Examples of ichnological characteristics of the Pebas Formation in the area of the Santa Rosa de Pichana section. (a) Sand-ﬁlled Thalassinoides (arrow) in sandy silt from the Santa Rosa de Pichana (from c. 11.5 m height in the section). (b) Asterosoma (indicated by arrows) from the lower part of the section displayed at Indiana (Loreto, Peru). Field of view 14 cm. (c) Chondrites reburrowing Planolites from Santa Rosa de Pichana (from c. 7.5 m height in the section). Arrows indicate parts of the image where small Chondrites in cross section are just visible as small dots within the Planolites.

surface is covered with a 40 cm thick, shell-bearing lag (F538-F539) topped by redeposited organic matter, which is also covered by a thin shell layer (F540). The shell layers in this interval are composed of lenses dominated by oriented gastropods (principally Dyris hershleri and D. lintea). In this transgressive interval the sandstones are clearly bioturbated, and grade upward into to the massive mollusc-rich mudstones of the maximum ﬂooding interval (F542), in which in situ specimens of Pachydon obliquus are common. The lower shoreface of the regressive interval is made up of interlayered mudstone and sandstone with abundant starved ripples and organic debris. Gyrolithes, Arenicolites and Planolites burrows are recorded in this interval. Within the upper shoreface, the sediments are mildly deformed and locally bioturbated at the top. These sands are ultimately capped by a terrestrial regressive 92 Aquatic landscapes in the Miocene of western Amazonia peat horizon.

The erosional transgressive base of sequence V is initiated with bioturbated sandy siltstones, the lower contact of which is lightly burrowed. These transgressive deposits grade into the mudstones of the maximum ﬂooding interval, which yielded two shell-bearing layers with in situ Pachydon obliquus (level F541).

The three complete sequences (II-IV), as well as the incomplete sequence V all have transgressive bases with reworked organic debris and shell bioclasts and are overlain by transgressive sandstones. The base of sequence III is interpreted to represent a channel cutting 1 m into the substrate. This channel contains reworked bone and teeth, as well as wood fragments and even logs. Laterally, the channel grades into reworked or winnowed lignite that contains giant terrestrial gastropods (Pebasiconcha immanis). The coarse-grained, shell-rich base of the cycles grade rapidly into mudstone intervals. These yield relatively high quantities of shells as well as dispersed organic matter. The shell content in the overlying mudstones is comparatively low. The mudstones grade upwards into a siltstone-ﬁne- grained sandstone alternation that in turn grades into massive sands in cycles III and IV. A coarsening-up trend deﬁnes the major part of these cycles (apart for the very base and top).

6.4 Molluscan composition and diversity

In total, 54 identifiable mollusc species were encountered (Appendix 1). In species numbers, the Cochliopidae (28 species, 51%) dominate, with the Pachydontinae (17 species, 31%) as the second most abundant group. Pachydon obliquus is the dominant species (47% in abundance). The fauna is dominated by endemic Pebas species (84% of species), even more so in terms of abundance (95%; for definition of endemicity, see Wesselingh, 2006a). Non-endemic freshwater species occur in low numbers (0.9% of the fauna). Even in the sample where these taxa are most common (98RK65) they make up only 6%. Fluvial (non- endemic) freshwater taxa are rare, but they do occur in 8 of the 16 samples. Terrestrial taxa (2 species) are very rare. Some juvenile shells of Corbula cotuhensis, a species that may represent saline (brackish) conditions, were found in two samples. The studied fauna can be characterised as predominantly endemic and freshwater. Diversity indices of samples are shown in Table 6.4. The variation of diversity indices in the section is shown in Figure 6.4. The samples of Sequences II (98RK65-98RK69) and III (F532-F537) show a similar trend of successive increase in diversity, followed by a decrease (Figure 6.4). The lowermost sample of Sequence II (98RK65) breaks with this trend, but it includes various taxa are interpreted as being reworked from other assemblages, explaining its higher then expected diversity (see below). In general, diversity is intermediate in the lower transgressive part of sequences, peaks at the late transgressive interval and drops at the maximum flooding, when faunas become dominated by Pachydon obliquus. Molluscan communities and taphonomic characteristics 6.4 Molluscan composition and diversity 93

Figure 6.4: Diversity indices in the section of Santa Rosa de Pichana. Key: H = Shannon’s diversity index; D = Simpson’s diversity index; grain-size in lithological column as in Figure 6.2. A, B, C refer to mollusc clusters (see text).

Three clusters are deﬁned based on the cluster analyses. Different analyses (using raw data and log transformed data, including or excluding rare species) produced very similar clusters with the exception of sample F542 that either falls in cluster A or in cluster C. In NMDS plots, the clusters do not show overlap in both the 2 and 3 dimensional solution (residual stress =0.09 and 0.05 respectively). Sample F542 plotted closer to cluster A samples than to cluster C samples. The three clusters described below result from analyses on raw data excluding rare species. Table 6.4 contains rank abundance data for the clusters, Table 6.4 summarizes the diversity indices of the clusters and taphonomic characteristics of the samples are provided in Table 6.4. 94 Aquatic landscapes in the Miocene of western Amazonia

SH D E F541 6 0.13 0.04 0.07 F542 22 0.87 0.32 0.28 F540 13 1.16 0.55 0.45 F539 12 1.15 0.57 0.46 F538#2 13 1.35 0.63 0.52 F538#1 11 1.23 0.59 0.51 F537 12 0.92 0.41 0.37 F536 31 1.56 0.59 0.46 F535 38 2.28 0.81 0.63 F533 32 2.12 0.73 0.61 F532 30 1.75 0.62 0.52 98RK69 12 0.99 0.38 0.40 98RK68 18 1.53 0.60 0.53 98RK67 16 1.96 0.73 0.71 98RK66 35 1.49 0.55 0.42 98RK65 30 1.93 0.73 0.57

Table 6.2: Diversity indices. Key: S = species richness (number of species in sample); H = Shannon’s diversity index = - (Pi· ln(Pi)) where Pi = importance probability in element i (= fraction of species i in sample); D = Simpson’s diversity index (= 1- (Pi· Pi)); E = evenness (= H / ln (richness)).

Cluster A — The seven samples from this cluster contain 57 species. The dominant species is Pachydon obliquus (57%: Table 6.4). Various delicate and small species of Dyris occur only in samples from cluster A. The cluster has the highest diversity and evenness (Table 4). Indicator species with very high indicator values are Dyris microbispiralis (In- dicator Value of 100), D. hauxwelli (90), D. gracilis (85), Onobops communis (84) and ?Cochliopina hauxwelli (86). Nineteen species have indicator values over 50. The mollusc composition of samples in cluster A are very similar to samples attributed to the Small- Dyris assemblage of Wesselingh et al. (2002), with the exception of assumed reworked taxa, such as Sheppardiconcha tuberculifera, Anodontites capax, Tryonia scalarioides tu- berculata and Dyris lintea in sample 98RK65. Shells in cluster A samples are generally well preserved. Wear, fragmentation and integrity of shells are intermediate. The presence of shells and shell fragments of different colour is rare. There are few dissolution marks and periostracum remains are relatively rare. The Small-Dyris assemblage, to which these samples are attributed, lived in little agitated lake bottoms (Wesselingh et al., 2002). Judging from the common occurrence of charophytic oogenia, these waters must have been clear. The (modest) variation in mixed preservation styles may, apart for reworking in the basal samples of lacustrine successions, indicate either depths between fair weather and storm wave base or relatively low deposition rates leaving shells exposed prolonged times near the lake ﬂoor.

Cluster B — The four samples in this cluster contain only 25 species. There are three common species, Dyris hershleri (32% of the specimens), Pachydon obliquus (27%) and Tryonia minuscula (25%). Indicative species are Dyris hershleri (indicator value of 100), D. lintea (75), Tryonia minuscula (59) and Pachydon tenuis (56). Shannon’s diversity index is low, but Simpson’s index is high, as is the measure of evenness. Samples from this cluster contain relatively few species, none of which is dominant. The mollusc content of the 6.4 Molluscan composition and diversity 95 Pachydon amazonensis Pachydon obliquus Mytilopsis sallei Tryonia minuscula Dyris tricarinata Pachydon tenuis 2.8 3.7 3.5 2.1 1.4 78.1 Dyris hershleri Pachydon tenuis Dyris tricarinata Dyris lintea Pachydon obliquus Tryonia minuscula Mytilopsis sallei Pachydon amazonensis 4.8 3.5 32.2 26.7 24.8 2.7 2.4 1.4 Dyris hauxwelli Pachydon obliquus Dyris lintea Ostomya papyria Pachydon ledaeformis Sheppardiconcha tuberculifera Toxosoma eboreum Onobops communis Pachydon amazonensis Dyris tricarinata Tryonia minuscula Mytilopsis sallei Table 6.3: Table 6.4 Rank abundance of species that make up more 1% of the sum within the three clusters; abundance in %. 2.1 1.6 1.5 1.2 2.6 2.6 2.8 3.3 Cluster ABC 56.9 7.6 4.7 4.0 96 Aquatic landscapes in the Miocene of western Amazonia

HDEn Average Cluster A 1.71 (0.47) 0.62 (0.16) 0.50 (0.12) 6 Average Cluster B 1.22 (0.09) 0.58 (0.03) 0.49 (0.04) 4 Average Cluster C 1.11 (0.69) 0.43 (0.26) 0.42 (0.23) 6 Average all clusters 1.40 (0.55) 0.55 (0.19) 0.47 (0.15) 16

Table 6.4: Table 6.4Diversity indices for the clusters (standard deviation in brackets). Key: H = Shannon’s diversity index; D = Simpson’s diversity index; E = Evenness; n = number of samples; formulas in Table 6.4.

Sample cluster wear f dm gd cm mp per sum F541 C 2 3 3 4 2 2 1 16 F542 C 2 2 3 3 2 2 1 14 F540 B 3 3 4 3 1 3 3 17 F539 B 3 2 2 3 2 2 2 14 F538#2 B 4 2 3 3 2 3 2 17 F538#1 B 5 2 4 4 1 3 2 19 F537 C 1 2 1 1 1 1 1 7 F536 A 2 1 1 2 1 2 2 9 F535 A 2 3 2 2 2 3 1 14 F533 A 3 3 3 4 1 5 2 19 F532 A 3 3 2 3 2 2 2 15 98RK69 C 2 2 3 2 1 2 3 12 98RK68 C 2 3 2 4 1 3 1 15 98RK67 C 2 3 2 2 1 2 2 12 98RK66 A 3 2 2 3 2 2 2 14 98RK65a A 3 3 4 4 2 4 2 20

Table 6.5: Table 6.4 Taphonomic parameters. Key: f = level of fragmentation in sample; dm = dissolution marks (holes); gd = generalised dissolution or shell integrity; cm = colour mixing; mp = mixed preservation styles; per = periostracum. Criteria outlined in Table 6.2. samples in this cluster resembles that of samples assigned to the Tryonia and the Tall-Dyris assemblages of Wesselingh et al. (2002). Cluster B samples are moderately well preserved. Abrasion of the shells is prominent and dissolution pits are common. Furthermore, the integrity of the shell carbonate is often somewhat compromised; shells that are ’soft’ and easily scratched are common. Periostracum remains are common. Possibly, periostracum was more prominently developed in order to withstand (episodic) lowered pH conditions, as indicated by the common dissolution holes in the shells. Tall-Dyris / Tryonia assemblages represent nearshore (foreshore, beach and back swamp) environments (Wesselingh et al., 2002). These settings experienced the strongest physical disturbance (erosion and transport from waves, currents and rivers) in the Pebas system. The four samples attributed to this cluster are from an organic-rich interval with shell lenses and beds containing transported bioclasts (samples F538#1 and F538#2 represent two shell lenses at the same height in the section). In sample 98RK65, taxa typical of the Tall-Dyris and Tryonia assemblages (Anodontites capax, Sheppardiconcha tuberculifera) were found to have similar preserva- 6.5 Carbon and oxygen isotope geochemistry 97

tion characteristics (mostly fragments that are strongly worn, contain dissolution holes and whose carbonate is softened) as the samples attributed to cluster B. These shells and fragments have clearly been transported into cluster A faunas, as is also suggested by isotope evidence (see below). Apparently, fast deposition took place during deposition of cluster B faunas, enabling the preservation of the organic periostracum.

Cluster C — The four samples attributed to this cluster yield 33 species. Samples from this cluster are characterised by the dominance (78%) of Pachydon obliquus. There are no species indicative for this cluster; P. obliquus has the highest indicator value (IV = 29). Diversity and evenness are lowest of the three clusters; there are few rare species. The mollusc composition in Cluster C samples is very reminiscent to samples assigned to the Pachydon obliquus assemblage by Wesselingh et al. (2002). Shells in cluster C samples are in general well preserved. They are usually little worn, and mixed preservation styles are rare, with the exception of carbonate integrity. Two samples (F541 and 98RK68) yield shells whose carbonate is softened, which are mixed with shells that are durable and well preserved. In general, the amount of fragmentation is slightly higher than in samples from the other associations. The Pachydon obliquus assemblage represents dysoxic muds deposited below storm wave base on lake bottoms (Wesselingh et al., 2002). Slightly elevated fragmentation levels may point to higher densities of predators and/or lower deposition rates leaving shells exposed longer to predators. Softening of shells may represent interstitial low pH values due to breakdown of organic matter within the clays. Mixtures of such soft shells, and shells that are well preserved and hard, may indicate commonly low sedimentation rates for these intervals. The level containing solely Amazoniconcha immanis (base of sequence III at c. 5.00 m) should be attributed to the Thiaridae-Pulmonata assemblage of Wesselingh et al. (2002). The snail lived in terrestrial environments, presumably on forest ﬂoors (Wesselingh and Gittenberger, 1999).

6.5 Carbon and oxygen isotope geochemistry

Stable oxygen and carbon isotope data of entire shells, or larger portions of shells are shown in Appendix 2 and Figure 6.6. A positive correlation exists between δ 13C and δ18O (Figure 6.6a) with the R2 of 0,5217 (n=133). Maximum values of δ 13C and δ18O are around 2 and 0 , respectively; minimum values are slightly below -11 . Average values are -6.8 and - 5.7 , respectively. The isotope averages for the three clusters are given in Table refcluster and shown in Figure 6.6b-e. Clusters B and C have no overlap, but both overlap with cluster A. The δ13C and δ18O isotope ratios of one Dyris lintea specimen (respectively +1.88 and -0.42, Appendix 2, Figures 6.6 and 6.5) could be interpreted as indicative of seawater. How- ever the co-occurrence with pearly fresh water mussels that cannot withstand the slightest elevated salinity requires another explanation. Within the Pebas fauna, taxa attributed to Cluster B, including pearly freshwater mussels, are found halfway between riverine and marine values (Figure 6.6). Therefore, we conclude that the stable isotope signatures of 98 Aquatic landscapes in the Miocene of western Amazonia

-15,00 -10,00 -5,00 0,00 5,00 5,00 sea

Modern floodplain 0,00 lakes

-5,00

-10,00 Modern river Pebasian freshwater shells Pebasian endemics -15,00

Figure 6.5: Isotope ratios of the Pebasian shells and indications of typical isotope signatures of marine and Amazonian river shells. Pebasian shells from cluster B, which include typical freshwater taxa such as pearly fresh water mussels, have the highest values, whereas the endemics from the A and especially C clusters yield very low values. A strong enrichment of δ13C isotope ratios and a slight elevation of δ18O isotope ratios are explained by shallow, possibly seasonally isolated habitats, were outgassing, evaporation and plant growth modified the original river isotope signatures. the Pebasian shells cannot be explained from mixing models involving a marine end member. The isotope ratios almost certainly reflect lake systems, where the original river water signals have been modified by, amongst others, outgassing, equilibration with atmospheric 13 CO2 and plant growth, leading to strongly elevated δ C signatures, evaporation and proximity to marine source water of precipitation, leading to elevated δ 18O signatures (Vonhof et al., 1998, 2003; Chapter 5). Isotope signatures from growth bands in modern Peruvian floodplain lake pearly freshwater mussels representing the dry season actually closely re- semble the values found in the B-cluster shells from the Pebas Formation.

6.6 Growth band chemistry

Two shells (Pachydon tenuis and Diplodon longulus) from the interval comprising levels F532-F536 have low isotope ratios (δ 18O average c. -7, δ13C below -10; Figures 6.7 and 6.8). The two specimens from 98RK65 (Pachydon tenuis and P. erectus) yield slightly 6.6 Growth band chemistry 99

13 13

18 18

13 13

18 18

13 13

18 18

Figure 6.6: (a) Stable isotope values of whole shell analyses from Santa Rosa de Pichana. (b-f) Stable isotope values for shells attributed speciﬁc cluster. 100 Aquatic landscapes in the Miocene of western Amazonia

Cluster n avg δ13CSDavgδ18OSD A 73 -6.86 2.11 -5.53 2.32 A- 60 -7.40 1.51 -6.13 2.01 B 33 -5.47 1.82 -4.46 1.23 B+ 46 -5.16 2.14 -3.98 1.50 C 27 -8.35 1.42 -7.56 0.76

Table 6.6: Average isotope ratios for the identified clusters. In sample 98RK65 an admixture of fauna elements attributable to cluster A (Pachydon obliquus) and to cluster B (the other measured species) was found (see below). Subtracting these latter species from cluster A resulted in cluster A-, and adding them to cluster B resulted in Cluster B+ (see also Figure 6.4b-f). Key: n = number of analyzed samples; avg = average; SD = standard deviation. higher isotope values (average δ 18O c. 4; δ13Catc. -9). Poorly developed cycles of δ18O in Pachydon are, irregular and of a low amplitude (typically 2 - 3). Pachy- don specimens show an average decreasing amplitude during growth. Cyclicity of δ 13Cis irregular. In the juvenile stages of two of the three specimens, variations are very large (up to 9). A conclusion of Chapter 5 is that the flat oxygen isotope profiles were con- sistent with freshwater biotopes in water bodies with prolonged residence times (lakes). Low amplitude cyclicity in δ18O that is well visible in the Diplodon specimen is in antiphase with regular δ13C cycles of low amplitude (c. 1-2). In Chapter refchecol the anti-phase δ13C-δ18O cycles in the Diplodon specimen are explained to reflect seasonal environmental change affecting the dissolved inorganic carbon (DIC) pool. Either increased stratification and remineralization of organic matter in bottom water during the dry season or seasonal switching of water sources between more local and more distal parts of the drainage basin with different C3/C4 plant distributions could explain the anti-phase signal in Diplodon. Additional trace-element analyses on the Diplodon longulus and Pachydon tenuis specimens from level F532-F537 confirmed seasonality interpretations from the oxygen record in Diplodon, but added no insights into the age of the Pachydon specimen (Chap- ter 5). The irregular occurrence of iron spikes was interpreted as recurring episodic lowering of dissolved oxygen in the water column. High barium concentrations in Pachydon tenuis were suggested to relate to the tolerance of this species to low oxygen concentrations in the bottom waters (Chapter 5; Wesselingh, 2006b). The relative flat and irregular oxygen and carbon isotope profiles in the Pachydontinae were explained to reflect little and unpre- dictable variation in the aquatic chemistry, indicating long residence times of the water. The low to very low δ13C and δ18O values were found to be incompatible with any substantial marine influence in the system.

6.7 Discussion

The faunistic character of the counted samples from Santa Rosa de Pichana is very similar to that of the faunistic character of 285 samples from Pebas Formation exposures in northeast Peru and southeast Colombia (Wesselingh et al., 2002). The stable isotope ratios measured 6.7 Discussion 101 in the Santa Rosa de Pichana shells also are representative for the Pebas Formation (Vonhof et al., 1998, 2003). The sequences in Santa Rosa are well developed (thick), but otherwise comply with the dominant type of sedimentary sequence found in the Pebas Formation (Räsänen et al., 1998; Gingras et al., 2002b; Vonhof et al., 2003). The studied sequences contain three intervals, a transgressive, maximum ﬂooding and regressive/prograding interval.

Pachydon tenuis, Santa Rosa de Pichana 0

-2

-4

-6

-8

O in ‰ V-PDB -10 18 δ -12 C, 13

δ -14 juvenile adult -16 45 40 35 30 25 20 15 10 5 0 sample numbers

δ13C δ18O

Pachydon erectus, Santa Rosa de Pichana 0 a b c? d e -2

-4

-6

-8

-10

O in ‰ V-PDB -12 18 δ -14 C,

13 juvenile adult δ -16 200 180 160 140 120 100 80 60 40 20 0

sample numbers

Figure 6.7: Stable isotope values along growth bands in two bivalves from Santa Rosa de Pichana. (a) Pachydon tenuis and (b) P. erectus from level 98RK65. 102 Aquatic landscapes in the Miocene of western Amazonia

Transgressive interval — The base of the transgressive interval is formed by organic- rich clays or (partially) reworked peat, commonly containing molluscs that may be concen- trated in lenses, as well as sand intervals. In one case a channel is developed in the basal transgressive interval. The base of the transgressions are usually characterised by burrowed ﬁrmgrounds. The Thalassoides type of burrowing suggests substantial elevation of salinity (up to mesohaline; Gingras et al., 2002b). Upwards, the sediments grade into blue-grey massive or bioturbated fossiliferous mudstones or silty, ﬁne-grained sands. The presence of Asterosoma burrows just below level F532 again suggests elevated salinities for this interval. Terrestrial snails occur in the basal lignites. In the overlying/intervening sandy interval,

Diplodon longulus, Santa Rosa de Pichana 0

-2

-4 a b c d e -6

-8 O in ‰ V-PDB 18

δ -10

C, -12 13 δ -14 juvenile adult -16 70 60 50 40 30 20 10 0 sample numbers

δ13C δ18O

Pachydon tenuis, Santa Rosa de Pichana 0

-2

-4

-6

-8 O in ‰ V-PDB

18 -10 δ

C, -12 13 δ -14 juvenile adult -16 180 160 140 120 100 80 60 40 20 0 sample numbers

Figure 6.8: Stable isotope values along growth bands in two bivalves from Santa Rosa de Pichana. (a) Diplodon longulus and (b) P. tenuis from the interval in which samples F532-F537 were taken. 6.7 Discussion 103

low diversity faunas occur that are attributed to the Tall-Dyris and Tryonia assemblages, with indications of considerable reworking. Indications for (episodically) low pH values, also indicating freshwaters, exist in the form of common dissolution marks and softening of shell carbonate. Stable oxygen and carbon isotope ratios show large variations, but ratios are in general less negative then those of the late transgressive and highstand faunas. Strontium isotope ratios of two specimens from level F538#1 (Dyris lintea and Pachydon tenuis; Vonhof et al., 2003) show typical values of Andean freshwater sources and exclude any substantial mixing with seawater for that interval. This implies that relatively high carbon and oxygen isotope ratios can only be explained by plant growth and/or evaporation in combination with outgassing of CO2, similar as seen in Anodontites living in present-day Amazonian ﬂoodplain lakes (Chapter 2). The presence of juvenile specimens of Corbula cotuhensis in samples 98RK65 and 98RK66 indicates that the species, which is representative of marginal marine settings (Wesselingh, 2006a), managed to become settled, but apparently did not survive to adulthood. The rare occurrence of such one-year old marginal marine species elsewhere in the Pebas Formation shows that the system was open to marine settings during the early phase of transgression. The upper part of the transgressive interval is usually a mollusc-rich siltstone layer that yields faunas attributed to the Small-Dyris assemblage sensu Wesselingh et al., 2002. Faunas are diverse and well preserved. The faunas lived in little agitated bottoms, between the fair-weather and storm wave base or below the latter. Judging from the common occurrence of charophytic oogenia, these waters must have been clear. The shelly mud intervals can also directly overlay the basal lignite/peat, in which case considerable quantities of shells can be mixed in from the Tall-Dyris and Tryonia assemblages (level 98RK65).

Interval of maximum flooding — These intervals are dominated by dark grey mudstones with dispersed organic matter and in situ mollusc faunas (dominated by Pachydon obliquus). The Small-Dyris assemblages of the late transgressive intervals grade into the Pachydon obliquus assemblage of the maximum flooding intervals. Diversity drops, as the faunas become dominated by P. obliquus. Preservation of the molluscs is good, apart for variation in the shell integrity. The Pachydon obliquus assemblage is though to represent dysoxic lake bottom communities (Wesselingh et al., 2002), below storm wave base. Low stable isotope values exclude the possibility of substantial marine influence during deposition of this interval.

Regressive/prograding interval — The dark grey mudstones of the highstand interval grade into a succession of dark grey and blue mudstones, and laminated, fine-grained sand lenses and laminae, with rare burrows (Gyrolithes, Arenicolites, Planolites). Sand intercalations increase upward, and sandstone layers and sometimes massive sandstone layers dominate the upper part of the regressive/prograding interval, reflecting the shallowing of the shoreface. The shore eventually became emerged, allowing the development of a soil in the upper part of sequence III. The regressive/prograding interval of sequence III is unusually thick for the Pebas Formation sequences. Shells are almost entirely absent in regressive/prograding intervals. In the studied section a few fossil molluscs were observed 104 Aquatic landscapes in the Miocene of western Amazonia only in sequence III, but could not be sampled. In other sections, the regressive/prograding interval is usually devoid of molluscs, and judging from the excellent preservation of fine- scale sedimentary structures, burrowing faunas (including corbulid bivalves) must have been extremely rare or absent. In one section (Santa Teresa I, Loreto, Peru: locality data in Wes- selingh et al., 2002), we observed thin strings or pavements consisting of juvenile Pachydon specimens preserved in butterfly position (valves lying in plane attached through ligament). These juveniles apparently represent single settlement events followed by mass mortality in the same year. The presence of Gyrolithes in the lower shoreface of sequence IV might indicate strong salinity fluctuations.

A number of indications exist that oxygen levels in waters in the Pebas system played a crucial role in the determination of faunal distributions. Low oxygen levels are indicated by common preservation of organic matter, possible presence of chemosymbiosis in Pachydontinae (Wesselingh, 2006b), absence of oxyphyllic taxa such as Corbicula spp., low carbon isotope ratios and uncommon elevated barium and iron concentrations in shell carbonate (Kaandorp et al., 2006). Oxygen levels seem to have been high in the transgressive phase, but dropped during highstand intervals. With decreasing depositional depths upwards in the sequences, oxygenation also should have increased. Little agreement exists over the predominant salinity regimes in the Pebas system. Based on ichnofossil assemblages and the presence of tidal sedimentary structures, Gingras et al. (2002) argued for predominantly oligohaline-mesohaline salinities during deposition of the Pebas Formation. Based on molluscs, strontium and stable isotopes, Wesselingh et al. (2002) and Vonhof et al. (2003) argued for almost exclusively freshwater settings in the Pebas Formation with only very few oligohaline incursion levels. We have investigated possible salinity variations on a range of scales in Santa Rosa de Pichana. Oligohaline-polyhaline conditions throughout are excluded in intervals bearing molluscs, based on strontium isotope ratios, low oxygen and carbon isotope ratios and the presence of strict freshwater taxa in low numbers in half of the samples. Oxygen and carbon isotope ratios are particularly low in faunas from the late transgressive and maximum flooding intervals. The relatively high, but nevertheless negative, oxygen isotope ratios found in mollusc assemblages in the basal part of the transgressive horizons might indicate some marine influence. However, strontium isotope measurements, low carbon isotope ratios and the presence of strict freshwater species at the same levels rule elevated salinities out. The clearest indications for possible elevated salinities are from Thalassinoides-demarcated Glossifungites surfaces in firmgrounds at the base of sequences and from Asterosoma burrows in sands in the basal transgressive interval. Molluscs are absent in the latter sands, but mollusc faunas from strata directly overlying the Asterosoma-bearing sands show an admixture of shells representative of fluvial, marginal lacustrine and lacustrine settings, while lacking marine species, with a single exception. In the base of sequence II some juvenile specimens of Cor- bula cotuhensis were found, indicating the ability of settlement of marginal marine species there. The presence of juveniles only indicates that the species were able to settle, but died before reaching adulthood. This is a clear sign that the Pebas system was open to marginal marine settings during at least the initial phase of transgression. Fully-grown marginal ma- 6.7 Discussion 105

rine species do occur in the Pebas Formation in younger stratigraphic intervals (Vermeij and Wesselingh, 2000; Wesselingh et al., 2002; Wesselingh, 2006a). The apparent incompatible salinity indications can be explained in two ways. The first possibility is that Thalassinoides were emplaced at elevated salinities under depositional conditions where carbonate faunas did not live or were not fossilized. In that case, marine incursions might be preserved only in association with the flooding surfaces, after which the embayed water rapidly freshened due to poor circulation and accumulation of river waters. The second possibility is that the animals responsible for these burrows (burrowing shrimps) adapted to freshwater settings, in a manner similar to the corbulid pachydontine bivalves (Wesselingh, 2006b). A very similar situation comprising marine ichnofossils in deposits with freshwater molluscs and ostracods has been described lately from the Paleocene Fort Union Formation of the United States (Belt et al., 2005). Here, three freshwater ostracod and two freshwater bivalve species, one of which a Pachydon, occur in deposits that also yield a diverse array of marine ichnogenera. Belt et al. 2005 showed that, in the Fort Union Formation, the different assemblages were deposited in separate time intervals, and concluded that a very low gradient coastal lowland, susceptible to sea level variations, explains the occurrence of both freshwater and marine indicators in successive layers, like the first scenario outlined above for the Pebas Formation. A difference with the Pebas system is the high diversity, and especially endemicity and stratigraphic continuity, of the mollusc and ostracod faunas in the latter system. To date no indications have been found that these faunas could cope with elevated salinities and it is unlikely that they could have taken shelter in ’normal’ (fluvial) freshwater environments during marine ingressions (Figure 6.9). The evolutionary continuity of the endemic Pebasian faunas (especially from c. 17 to c. 9 Ma: Wesselingh and Salo (2006); Wesselingh et al. (2006a)) in western Amazonia argues against repeated basin-wide establishment of mesohaline or normal marine settings. On the other hand, the sporadic occurrence of (typically) brackish-water trace fossils is an argument indicating that the fauna did not evolve into stable infaunal, fresh-water niches in Miocene Amazonia; that is, if depositional conditions were consistently fresh and burrowing animals evolved into increasingly fresh-water environments, then their success should be recorded in the sedimentary record as pervasive and dominant fabrics. Likewise the relatively rare occurrence of highly burrowed deposits suggests the depositional system was not consistently brackish. This leaves us with the possibility that the brackish-water character of the depositional system has been previously overstated (Gingras et al., 2002b), but that regular brackish-water incursions did occur. This possibility would imply that fossil shells represent only ’snapshots’ of the entire depositional (including geochemical) system. Pos- sible salinity variations on a seasonal scale in shell-bearing intervals have been excluded by isotope work in successive growth bands of bivalves including the specimens from Santa Rosa de Pichana (Chapter 5). In general, waters of the Pebas system must have been non-acidic, given the common and very delicate preservation of carbonate fossils. In Chapter 5 the very large δ 13C variation observed in some Pachydon specimens (Figures 6.7 and 6.8) is attributed to pH values of the water at around pH 6.4. Episodic lowering of pH may have occurred in environments during the early stages of the transgressive phases, as indicated by the presence of dissolu- 106 Aquatic landscapes in the Miocene of western Amazonia

M F M

(a) (b)

F F E M

Figure 6.9: Effect of marine incursions. (a) In a low-gradient coastal plain, fluvial faunas can become restricted towards the basin margins during maximum marine transgressions (b). In the case of the Pebas system, a highly diverse endemic freshwater fauna existed between the marine and fluvial freshwater settings (c). Such endemic faunas are not adapted for living in fluvial environments. If substantial marine transgressions occur, the endemic faunas would be expected to become squashed between marine and fluvial regimes (d). Key: M = marine faunas; E = endemic (freshwater) faunas; F = fluvial and fluviolacustrine faunas. tion holes on shells. In shallow waters with abundant plant production, the degradation of organic matter may produce interstitial lowering of pH, dissolving carbonate shells (Aller, 1982). Dissolution may also result from bacterial degradation, but this type of dissolution often provides a corroded surface, not large dissolution holes that characterise some of the samples from Santa Rosa de Pichana. Also, local drainage of black water creeks or rivers from swampy backlands may have caused lowering of pH values in the water, as may have substantial rainwater input in shallow (isolated) marshes and ponds. Furthermore, pH values may have been depressed in the vicinity of cratonic rivers, but no indication was found in the Santa Rosa section for these. A model for the depositional sequences and faunal development of the Santa Rosa de Pichana sequences is illustrated in Figure 6.10 and a full interpretation of the section is given in Figure 6.11.

6.8 Conclusions

Regressive/ Prograding molluscs absent interval

Highstand Pachydon obliquus

Transgressive small Dyris interval 0 tall Dyris/ Tryonia

Thiaridae/Pulmonata waterdepth - + - + -10 -2 f o m f o m oxygen diversity preservation δ13C/δ18O ichnosal molisosal

Figure 6.10: Schematic representation of depositional and faunistic characteristics in sequences from Santa Rosa de Pichana. Salinity indications from combined molluscan/isotope data and from ichnofossil occurrences are also outlined. The height of a sequence (in this case 5 m) ranges between 4 and 6.5 m in Santa Rosa. In other Pebas outcrops such sequences are often only 2.5-4 m thick. Key: f = freshwater; o = oligohaline; m = mesohaline; ichnosal = salinity estimates based on ichnofossil data; molisosal = salinity estimates based on mollusc and isotope data. Other abbreviations explained under Figure 6.2.

onia and Tall-Dyris assemblages in the early transgressive intervals and Small-Dyris assemblage in the late transgressive intervals, to the Pachydon obliquus assemblage in the highstand interval. Reworking during the early transgressive phase was common. With increasing water depth, the preservation of the molluscs improved. Diversity increased to its zenith in the Small-Dyris assemblage and then decreased towards the Pachydon obliquus assemblage. ¥ Dissolved oxygen levels in the water probably played a major role in the determination of fauna successions, with dysoxia becoming important during highstand intervals and dominant in the early regressive/prograding intervals. Molluscs and stable isotope proﬁles show no indications of elevated salinities, not even on seasonal time scales. Thalassinoid ichnofossils of the Glossifungites ichnofacies at the base of sequences may represent up to lower mesohaline salinities, but also may reﬂect evolution of lowered salinity tolerances for the constructors (crustaceans and possibly polychaetes). ¥ The Santa Rosa de Pichana section appears representative for large parts of the Pebas Formation. Based on molluscan and stable isotope evidence we think that the system consisted of predominantly freshwater swamp to lacustrine conditions in a long-lived lake/wetland system at sea level and open to marginal marine settings. However, discrepancies with ichnofossil data from the same stratigraphic intervals, which consistently indicate elevated salinities, are in need of further study. 108 Aquatic landscapes in the Miocene of western Amazonia

Figure 6.11: Interpretation of the Santa Rosa de Pichana section. Key: HSI = highstand interval; TrI = transgressive interval; RPI = regressive/prograding interval; div = diversity; pres = preservation; average isotope values for samples are plotteed in the last column; Depth ranges from left to right: below storm wave base, between fair weather and storm wave base, above fair-weather wave base, terrestrial; grain-size in lithological column as in Figure 6.2. References

Abell, P. I., Amegashitsi, L., Ochumba, P., 1995. The shells of Etheria elliptica as recorders of seasonality at Lake Victoria. Palaeogeography, Palaeoclimatology, Palaeoecology 119, 215Ð219. Abell, P. I., Hoelzmann, P., 2000. Holocene palaeoclimates in northwestern Sudan: stable isotope studies on molluscs. Global and Planetary Change 26, 1Ð12. Adams, H., 1860. Description of a new genus of freshwater bivalve Mollusca, belonging to the family Corbulidae, from the collection of Hugh Cuming, Esq. Proceedings of the Zoological Society of London 28, 203. Ayers, J. C., Anderson, D. V., Chandler, D. C., Lauff, G. H., 1956. Currents and water masses of Lake Huron. In: Great Lakes. Vol. 1 of Res. Inst. Univ. Mich. Technical Papers. University of Michigan, p. 217. Baldock, J. W., 1982. Geology of Ecuador: explanatory bulletin of the National geological map of the Republic of Ecuador. Barkan, E., Luz, B., Lazar, B., 2001. Dynamics of the carbon dioxide system in the Dead Sea. Geochimica et Cosmochimica Acta 65 (3), 355Ð368. Belt, E. S., Tibert, N. E., Curran, H. A., Diemer, J. A., Hartman, J. H., Kroeger, T. A., Harwood, D. M., 2005. Evidence for marine inﬂuence on a low-gradient coastal plain: ichnology and invertebrate paleontology of the lower Tongue River Member (Fort Union Formation, middle Paleocene), western Williston Basin, U.S.A. Rocky Mountain Geol- ogy 40, 1Ð24. Berggren, W. A., Kent, D. V., Swisher, C. C., Aubry, M. P., 1995. A revised Cenozoic geochronology and chronostratigraphy. In: Geochronology time scales and global stratigraphic correlation. Vol. 54 of SEPM Special Publication. Society for Sedimentary Geol- ogy, pp. 129Ð212. Buitenhuis, E. T., de Baar, H. J. W., Veldhuis, M. J. W., 1999. Photosynthesis and calciﬁca- tion by Emiliania huxleyi (Prymnesiophycaea) as a function of inorganic carbon species. Journal of Phycology 35, 949Ð959. 110 REFERENCES

Burgman, J. O., Calles, B., Westman, F., 1987. Conclusions from a ten year study of oxygen- 18 in precipitation and runoff in Sweden. In: Isotope techniques in water resources development. Vol. 1987. IAEA, Vienna, pp. 579Ð590.

Claessen, D., van Oss, C., de Roos, A. M., Persson, L., 2002. The impact of size-dependent predation on population dynamics and individual life history. Ecology 83 (6), 1660Ð1675.

Clark, I. D., Fitz, P., 1997. Environmental isotopes in hydrogeology. CRS Press, Lewis Publishers.

Cole, J. J., Caraco, N. F., Kling, G. W., Katz, T. K., 1994. Carbon dioxide supersaturation in the surface waters of lakes. Science 265, 1568Ð1570.

Comfort, A., 1957. The duration of life in molluscs. Proceedings of the Malacological So- ciety of London 32, 219Ð241.

Conrad, T., 1871. Descriptions of new fossil shells of the upper Amazon. American Journal of Conchology 6, 192Ð198.

Conrad, T. A., 1874. Remarks on the Tertiary clay of the Upper Amazon with descriptions of new shells. Proceedings of the Academy of Natural Sciences, Philadelphia 26 (1874), 25Ð32.

Costa, E. V., 1981. Revisão gastropodes fosseis da localidade do Tres Unidos, Formacao Pebas, Plioceno do Alto Amazonas, Brasil. In: Anais II Congresso Latino-Americano Palaeontologia. Porto Alegre, pp. 635Ð649.

Crowley, T. J., Short, D. A., Mengel, J. G., North, G. R., 1986. Role of seasonality in the evolution of climate during the last 100 million years. Science 231, 579Ð584.

Cvancara, A. M., 2000. Floating mussels in the Upper Mississippi River, Minnesota and their implications for dispersal in Paleontology and Archeology. Central Plains Archeol- ogy 8 (1), 143Ð148.

Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus XVI, 436Ð461.

De Simone, O., Junk, W. J., Schmidt, W., 2003. Central Amazon ﬂoodplain forests: root adaptions to prolonged ﬂooding. Russian Journal of Plant Physiology 50 (6), 848Ð855.

Dettman, D. L., Lohmann, K. C., 1993. Seasonal change in Paleogene surface water: δ18O: fresh-water bivalves of western North America. In: Swart, P. K., Lohmann, K. C., McKenzie, J., Savin, S. (Eds.), Climate change in continental isotopic records; Geophys- ical Monograph. Vol. 78. American Geophysical Union, Washington, pp. 153Ð163.

Dettman, D. L., Lohmann, K. C., 1995. Microsampling carbonates for stable isotope and minor element analyses: physical separation of samples on a 20 micrometer scale. Journal of Sedimentary Research A65, 566Ð569. REFERENCES 111

Dettman, D. L., Reische, A. K., Lohmann, K. C., 1999. Controls on the stable isotope composition of seasonal growth bands in aragonitic fresh-water bivalves (unionoidae). Geochimica et Cosmochimica Acta 63 (7/8), 1049Ð1057. Dörner, H., Wagner, A., 2003. Size-dependent predator-prey relationships between perch and their ﬁsh prey. Journal of Fisher Biology 62, 1021Ð1032. Dufrêne, M., Legendre, P., 1997. Species assemblages and indicator species: the need for a ﬂexible asymmetrical approach. Ecological Monographs 67, 345Ð366. Dutton, A. L., Lohmann, K. C., Zinsmeister, W. J., 2002. Stable isotope and minor element proxies for Eocene climate of Seymour Island, Antarctica. Paleoceanography 17 (2), 6Ð1 Ð 6Ð13. Emiliani, C., 1955. Pleistocene temperatures. Journal of Geology 63, 538Ð578. Epstein, S., Buchsbaum, R., Lowenstam, H. A., Urey, H. C., 1953. Revised carbonate-water isotopic temperature scale. Bulletin of the Geological Society of Amercia 64, 1351Ð1326. Etheridge, R., 1879. Notes on Mollusca collected by C. Barrington Brown, Esq., A.R.S.M., from the Tertiary deposits of Solimoes and Javary Rivers, Brazil. Quaterly Journal of the Geological Society of London 35, 82Ð88. Etter, W., 1999. Community analysis. In: Harper, D. (Ed.), Numerical Palaeobiology. John Wiley and Sons Ltd., Chichester, pp. 285Ð360. Fastovsky, D., Arthur, M., Strater, N., Foss, A., 1993. Freshwater bivalves (Unionidae), disequilibrium isotopic fractionation, and temperatures. Palaios 8, 602Ð608.

Flower, B. P., 1999. Warming without high CO2? Nature 399, 313Ð314. Fontes, J. C., Boulange, B., Carmouze, J. P., Florkowski, T., 1979. Preliminary oxygen- 18 and deuterium study of the dynamics of Lake Titicaca. In: Isotopes in Lake Studies. IAEA, Vienna, pp. 145Ð150. Fontes, J. C., Gonfiantini, R., 1970. Composition isotopique et origine de la vapeur d’eau atmospherique dans la region du Lac Leman. Earth and Planetary Science Letters 7, 325Ð 329. Gabb, W. M., 1869. Descriptions of fossils from the clay deposits of the Upper Amazon. American Journal of Conchology 4, 197Ð200. Gingras, M. K., Räsänen, M., Ranzi, A., 2002a. The significance of bioturbated inclined heterolithic stratification in the southern part of the Miocene Silmões Formation, Rio Acre, Amazonia Brazil. Palaios 17, 591Ð601. Gingras, M. K., Räsänen, M. E., Pemberton, S. G., Romero Pittman, L., 2002b. Ichnology and sedimentology reveal depositional characteristics of bay-margin parasequences in the Miocene Amazonian Foreland Basin. Journal of Sedimentary Research 72 (6), 871Ð883. 112 REFERENCES

Grootes, P. M., 1993. Interpreting Continental Oxygen Isotope Records. In: Swart, P., Lohmann, K., McKenzie, J., Savin, S. (Eds.), Climate Change in Continental Isotopic Records. Vol. 78 of Geophysical monograph. American Geophysical Union, pp. 37Ð46.

Grootes, P. M., Stuiver, M., Thompson, L. G., Mosley-Thompson, E., 1989. Oxygen isotope changes in tropical ice, Quelccaya, Peru. Journal of Geophysical Research 94, 1187Ð 1194.

Grossman, E., Ku, T.-L., 1986. Oxygen and carbon isotope fractionation in biogenic aragonite: temperature effects. Chemical Geology 59, 59Ð74.

Haug, G. H., Hughen, K. A., Sigman, D. M., Peterson, L. C., Röhl, U., 2001. Southward migration of the Intertropical Convergence Zone through the Holocene. Science 293, 1304Ð1308.

Hooghiemstra, H., van der Hammen, T., 1998. Neogene and Quaternary development of the neotropical rain forest: the forest refugia hypothesis, and a literature overview. Earth Science Reviews 44 (3/4), 147Ð183.

Hoorn, C., 1993. Marine incursions and the inﬂuence of Andean tectonics on the Miocene depositional history of northwestern Amazonia: Results of a palynostratigraphic study. Palaeogeography, Palaeoclimatology, Palaeoecology 105, 267Ð309.

Hoorn, C., 1994a. An environmental reconstruction of the palaeo-Amazon River system (Middle to Late Miocene, NW Amazonia). Palaeogeography, Palaeoclimatology, Palaeoecology 112, 187Ð238.

Hoorn, C., 1994b. Miocene palynostratigraphy and paleoenvironments of northwestern Amazonia: Evidence for marine incursions and the inﬂuence of Andean tectonics. Phd thesis, University of Amsterdam.

Hoorn, C., 1996. Technical comments: Miocene deposits in the Amazonian Foreland Basin. Science 273, 122Ð123.

Hoorn, C., 2006. Birth of the mighty Amazon. Scientiﬁc American May 2006, 52Ð59.

Hoorn, C., Aalto, R., Kaandorp, R. J. G., Lovejoy, N. R., 2006. Miocene diurnal tidal rhythmites in Madre de Dios, Peru: COMMENT. Geology , e98Ðe99, doi:10.1130/G22115.1.

Hoorn, C., Guerrero, J., Sarmiento, G., Lorente, M., 1995. Andean tectonics as a cause for changing drainage patterns in Miocene northern South America. Geology 23, 237Ð240.

Hovikoski, J., Räsänen, M., Gingras, M., Roddaz, M., Brusset, S., Hermoza, W., Romero Pittman, L., Lertola, K., 2005. Miocene semidiurnal tidal rhythmites in Madre de Dios, Peru. Geology 33 (3), 177Ð180. REFERENCES 113

Howarth, R., McArthur, J., 1997. Statistics for Strontium Isotope Stratigraphy: A Robust LOWESS Fit to the Marine Sr-Isotope Curve for 0 to 206 Ma, with Look-up Table for Derivation of Numeric Age. The Journal of Geology 105, 441â A¸ù S456].

Jones, D. S., Quitmyer, I. R., 1996. Marking Time with Bivalve Shells: Oxygen Isotopes and Season of Annual Increment Formation. Palaios 11, 340Ð346.

Jones, D. S., Williams, D. F., Arthur, M. A., 1983. Growth history and ecology of the At- lantic surf clam Spisula solidissima (Dillwyn), as revealed by stablle isotopes and annual shell increments. Journal of Experimental Marina Biology and Ecology 73, 225Ð242.

Kaandorp, R. J. G., Vonhof, H. B., Del Busto, C., Wesselingh, F. P., Ganssen, G. M., Mar- mól, A. E., Romero Pittman, L., van Hinte, J. E., 2003. Seasonal stable isotope variation of the Amazonian fresh water bivalve Anodontites trapesialis. Palaeogeography, Palaeo- climatology, Palaeoecology 194 (4), 327Ð342.

Kaandorp, R. J. G., Vonhof, H. B., Wesselingh, F. P., Romero Pittman, L., Kroon, D., van Hinte, J. E., 2005. Seasonal Amazonian rainfall variation in the Miocene Climate Optimum. Palaeogeography, Palaeoclimatology, Palaeoecology 221 (1-2), 1Ð6.

Kaandorp, R. J. G., Wesselingh, F. P., Vonhof, H. B., 2006. Ecological implications from geochemical records of Miocene Western Amazonian bivalves. Journal of South Ameri- can Earth Sciences 21 (1-2), 54Ð74.

Kalliola, R., Paitán Flores, S., 1998. Geoecología y desarollo Amazónico. Estudio integrado en la zona de Iquitos, Perú. Vol. A II 114 of Annales Universitatis. Turun Yliopisto, Turku.

Kalliola, R., Puhakka, M., 1993. Geografía de la selva baja Peruana. In: Kalliola, R., Puhakka, M., Danjoy, W. (Eds.), Amazonia Peruana. PAUT and ONERN, Turku, pp. 9Ð21.

Kalliola, R., Salo, J., Puhakka, M., Rajasilta, M., 1991. New site formation and colonizing vegetation in primary succession on the Western Amazon ﬂoodplains. The Journal of Ecology 79 (877-901).

Kauffman, S., Arce, G. P., Marquina, R., 1998. Suelos de la zona de Iquitos. In: Kalliola, R., Flores Paitán, S. (Eds.), Geoecología y desarollo Amazónico. Estudio integrado en la zona de Iquitos, Perú. Vol. A II 114 of Annales Universitatis. Turun Yliopisto, Turku, pp. 139Ð229.

Klein, R. T., Lohmann, K. C., Thayer, C. W., 1996. Bivalve skeletons record sea-surface temperature and δ18O via Mg/Ca and 18O/16O ratios. Geology 24 (5), 415Ð418.

Lamarck, J. B. P. A., 1819. Histoire naturelle dans animaux sans vertèbres. Vol. 6. Déterville and Verdière, Paris. 114 REFERENCES

Lear, C. H., Elderﬁeld, H., Wilson, P. A., 2000. Cenozoic deep-sea temperatures and global ice volumes from Mg/Ca in benthic foraminiferal calcite. Science 287, 269Ð272. Leopoldo, P. R., Martinez, J. C., Mortatti, J., 1992. Estimation using 18O of the water residence time in small watersheds. In: IAEA (Ed.), Isotope Techniques in Water Resources Development. Vol. 1991. IAEA, Vienna, pp. 75Ð84. Linna, A., Kauffman, S., Wesselingh, F., Kalliola, R., 1998. Heterogeneidad edáﬁca de la zona de Iquitos: origen y comprensión de sus propiedades. In: Kalliola, R., Flores Paitán, S. (Eds.), Geoecología y desarollo Amazónico. Estudio integrado en la zona de Iquitos, Perú. Vol. A II 114 of Annales Universitatis. Turun Yliopisto, Turku, pp. 461Ð480. Luz, B., Stiller, M., Talma, S., 1997. Carbon dynamics in the Dead Sea. In: Niemi, T. M., Ben-Avram, Z., Gat, J. R. (Eds.), The Dead Sea: the lake and its setting,. Vol. 36 of Oxford Monographs on Geology and Geophysics. Oxford University Press, Oxford, pp. 184Ð192. Maia, R. G., Godoy, H. K., Yamaguti, H. S., de Moura, P. A., Da Costa, F. S., de Holanda, A. M., Costa, J., 1977. Projeto de Carvao no Alto Solimoes. Relatorio Final. Tech. rep., CPRM-DNPM. Marengo, J., 1998. Climatología de la zona de Iquitos, Perú. In: Kalliola, R., Flores Paitán, S. (Eds.), Geoecología y desarollo Amazónico. Annales Universitatis Turkuensis. Turun Yliopisto, Turku, pp. 35Ð57. Marengo, J. A., Nobre, C. A., 2001. General characteristics and variability of climate in the Amazon Basin and its links to the global climate system. In: McClain, M. E., Victoria, R. L., Richey, J. E. (Eds.), The biochemistry of the Amazon Basin. Oxford University Press, p. 357. Marshall, L. G., Lundberg, J. G., 1996. Technical comments: Miocene deposits in the Ama- zonian Foreland Basin. Science 273, 123Ð124. Martínez, S., del Río, C. J., 2002. Late Miocene molluscs from the southwestern At- lantic Ocean (Argentina and Uruguay): a palaeobiogeographic analysis. Palaeogeogra- phy, Palaeoclimatology, Palaeoecology 188, 167Ð187. Mazumder, R., Arima, M., 2005. Tidal rhythmites and their implications. Earth-Science Reviews 69, 79Ð95. Meade, R. H., Dunne, T., Richey, J. E., De M. Santos, U., Salati, E., 1985. Storage and remobilization of suspended sediment in the lower Amazon River of Brazil. Science 228, 488Ð490. Mook, W. G., Bommerson, J. C., Staverman, W. H., 1974. Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide. Earth and Planetary Science Letters 22, 169Ð176. REFERENCES 115

Mook, W. G., de Vries, J. J., 1998. Environmental isotopes in the hydrological cycle: Prin- ciples and Applications. In: Mook, W. G. (Ed.), Introduction. Vol. 1 Theory, methods review. IAEA, Vienna, p. 271.

Muñoz Torres, F., Whatley, R., van Harten, D., 1998. The endemic non-marine Miocene ostracod fauna of the Upper Amazon Basin. Revista española de Micropaleontología 30, 89Ð105.

Muñoz Torres, F., Whatley, R. C., van Harten, D., 2006. Miocene ostracod (Crustacea) biostratigraphy of the upper Amazon Basin and evolution of the genus Cyprideis. Journal of South American Earth Sciences 21 (1-2), 75Ð86.

Mutvei, H., Westermark, T., Dunca, E., Carell, B., Forberg, S., Bignert, A., 1994. Methods for the study of environmental changes using the structural and chemical information in o molluscan shells. Bulletin de l’Institut océanographique, Monaco, n ø spécial 13 , 163Ð 186.

Nuttall, C., 1990. A review of the Tertiary non-marine molluscan faunas of the Pebasian and other inland basins of north-western South America. Bulletin British Museum (Natural History) 45, 165Ð372.

Pawellek, F., Frauenstein, F., Veizer, J., 2002. Hydrochemistry and isotope geochemistry of the upper Danube River. Geochimica et Cosmochimica Acta 66 (21), 3839Ð3854.

Pawellek, F., Veizer, J., 1994. Carbon cycle in the upper Danube and its tributaries: 13 δ CDIC constraints. Israel Journal of Earth Sciences 43, 187Ð194. Paxton, C., Crampton, W., Burgess, P., 1996. Technical comments: Miocene deposits in the Amazonian Foreland Basin. Science 273, 123.

Petri, S., Fulfaro, V. J., 1983. Universidade de Sao Paulo, Sao Paulo.

Puhakka, M., Kalliola, R., 1995. Floodplain vegetation mosaics in Western Amazonia. Bio- geographica 71 (1), 1Ð14.

Räsänen, M., Linna, A., Irion, G., Hernani, L. R., Huaman, R. V., Wesselingh, F., 1998. Geología y geoformas de la zona de Iquitos. In: Kalliola, R., Paitán, S. F. (Eds.), Geoe- cología y desarollo Amazónico: estudio integrado en la zona de Iquitos, Perú. Turun Yliopisto, Turku, pp. 59Ð137.

Räsänen, M., Linna, A., Santos, J., Negri, F., 1995. Late Miocene tidal deposits in the Amazonian foreland basin. Science 269, 386Ð390.

Rebata H., L. A., Räsänen, M. E., Gingras, M. K., Vieira, Jr., V., B., M., Irion, G., 2006. Sedimentology and ichnology of tide-inﬂuenced Late Miocene successions in western Amazonia: The gradational transition between the Pebas and Nauta formations. Journal of South American Earth Sciences 21 (1-2), 96Ð119. 116 REFERENCES

Renssen, H., Beets, C. J., Fichefet, T., Goosse, H., Kroon, D., 2004. Modeling the climate response to a massive methane release from gas hydrates. Paleoceanography 19, PA2010, doi:10.1029/2003PA000968.

Roddaz, M., Brusset, S., Baby, P., Hérail, G., 2006. Miocene tidal-inﬂuenced sedimentation to continental Pliocene sedimentation in the forebulge-backbulge depozones of the Beni- Mamore foreland Basin (northern Bolivia). Journal of South American Earth Sciences 20 (4), 351Ð368.

Romanek, C. S., Grossman, E. L., Morse, J. W., 1992. Carbon isotopic fractionation in synthetic aragonite and calcite: Effects of temperature and precipitation rate. Geochimica et Cosmochimica Acta 56 (1), 419Ð430.

Roxo de Oliveira, R. G., 1924. Breve noticea sobre os fosseis terciario do Alto Amazonas. Boletim do Servicio geoligico-mineralogico Brasil 11, 41Ð52.

Rozanski, K., Araguas-Araguas, L., Gonﬁantini, R., 1993. Isotopic patterns in modern global precipitation. In: Swart, P., Lohmann, K., McKenzie, J., Savin, S. (Eds.), Cli- mate Change in Continental Isotopic Records. Geophysical Monograph: 78. American Geophysical Union, pp. 1Ð36.

Salati, E., Vose, P. B., 1984. Amazon Basin: A system in equilibrium. Science 225, 129Ð 138.

Salo, J., Kalliola, R., Häkkinen, I., Mäkinen, Y., Niemelä, P., Puhakka, M., Coley, P. D., 1986. River dynamics and the diversity of Amazon lowland forest. Nature 322, 254Ð258.

Savin, S. M., 1977. The history of the earth’s surface temperature during the last 100 million years. Annual Review of Earth and Planetary Sciences 5, 319Ð355.

Savin, S. M., Douglas, R. G., Stehli, F. G., 1975. Tertiary marine paleotemperatures. Bul- letin of the Geological Society of America 86, 1499Ð1510.

Schöne, B., 2003. A ’clam-ring’ master-chronology constructed from a short-lived bivalve mollusc from the northern Gulf of California, USA. The Holocene 13 (1), 39Ð49.

Scotese, C. R., 2001. Atlas of Earth History. Vol. 1, Paleogeography of PALEOMAP Project. Arlington, Texas.

Seltzer, G., Rodbell, D., Burns, S., 2000. Isotopic evidence for late Quaternary climatic change in tropical South America. Geology 28 (1), 35Ð38.

Sioli, H., 1984. The Amazon and its main afﬂuents: hydrography, morphology of the river courses, and river types. In: Sioli, H. (Ed.), The Amazon. Limnology and landscape ecology of a mighty tropical river and its basin. Dr. W. Junk, pp. 127Ð165. REFERENCES 117

Spötl, C., Mattey, D., 2006. Stable isotope microsampling of speleothems for palaeoenvironmental studies: A comparison of microdrill, micromill and laser ablation techniques. Chemical Geology 235, 48Ð58. Stecher, H. A., Krantz, D. E., Lord, C. J., Luther, G. W., Bock, K. W., 1996. Profiles of strontium and Barium in Mercenaria mercenaria and Spisula solidissima shells. Geochimica et Cosmochimica Acta 60, 3445Ð3456. Steinmann, M., Hungerbühler, D., Seward, D., Winkler, W., 1999. Neogene tectonic evolution and exhumation of the southern Andes: a combined stratigraphy and fission-track approach. Tectonophysics 307, 255Ð276. Sztanó, O., de Boer, P. L., 1995. Basin dimensions and morphology as controls on amplification of tidal motions (the Early Miocene North Hungarian Bay). Sedimentology 42, 665Ð682. Thompson, L. G., Mosley-Thompson, E., Davis, M. E., Lin, P. N., Henderson, K. A., Cole- Dai, J., Bolzan, J. F., Liu, K. B., 1995. Late glacial stage and Holocene tropical ice core records from Huascarán, Peru. Science 269, 46Ð50. Treble, P., Shelley, J. M. G., Chappell, J., 2003. Comparison of high resolution sub-annual records of trace elements in a modern (1911-1992) speleothem with instrumental climate data from southwest Australia. Earth and Planetary Science Letters 216, 141Ð153. Tschopp, H. J., 1953. Oil explorations in the Oriente of Ecuador, 1938-1950. Bulletin of the American Association of Petroleum Geologists 37, 2303Ð2347. Urey, H. C., 1947. The thermodynamic properties of isotopic substances. Journal of the Chemical Society 1947, 562Ð581. Usdowski, E., Hoefs, J., 1990. Kinetic 13C/12C and 18O/16O effects upon dissolution and outgassing of CO2 in the system CO2-H2O. Chemical Geology 80, 109Ð118. Vaksa, A., Bar-Matthews, M., Ayalon, A., Schilman, B., Gilmour, M., Hawkesworth, C. J., Frumkind, A., Kaufman, A., Matthews, A., 2003. Paleoclimate reconstruction based on the timing of speleothem growth and oxygen and carbon isotope composition in a cave located in the rain shadow in Israel. Quaternary Research 59 (2), 182Ð193. Veinnot, G. I., Cornett, R. J., 1998. Carbon isotopic disequilibrium in the shell of the freshwater mussel Elliptio complanata. Applied Geochemistry 13, 49Ð57. Verdegaal, S., Troelstra, S. R., Beets, C. J., Vonhof, H. B., 2005. Stable isotopic records in unionid shells as a paleoenvironmental tool. Netherlands Journal of Geosciences 84 (4), 403Ð408. Vermeij, G. J., Wesselingh, F. P., 2002. Neogastropod molluscs from the Miocene of western Amazonia, with comments on marine to freshwater transitions in mollusks. Journal of Paleontology 76, 265Ð270. 118 REFERENCES

Vonhof,H., Kaandorp, R., Romero Pittman, L., Guerrero, J., Palacios Moncayo, O., Wessel- ingh, F., 2004. Isotope studies of fossil shells give insight in the Miocene paleogeography of Western Amazonia. In: Dávila, J., Carlotto, V., Chalco, A. (Eds.), XII Congreso Pe- ruano de Geología. Resúmenes Extendidos. Sociedad Geológica del Perú. Publicación o Especial Nø 6, Lima, pp. 556Ð559. Vonhof, H. B., Wesselingh, F. P., Ganssen, G. M., 1998. Reconstruction of the miocene western Amazonian aquatic system using molluscan isotopic signatures. Palaeogeogra- phy, Palaeoclimatology, Palaeoecology 141, 85Ð93. Vonhof, H. B., Wesselingh, F. P., Kaandorp, R. J. G., Davies, G. R., van Hinte, J. E., Guer- rero, J., Räsänen, M., Romero Pittman, L., Ranzi, A., 2003. Paleogeography of Miocene Western Amazonia: Isotopic composition of molluscan shells constrains the inﬂuence of marine incursions. GSA Bulletin 115 (8), 983Ð993. Wang, X., Auler, A. S., Edwards, R. L., Cheng, H., Cristall, P. S., Smart, P. L., Richards, D. A., Shen, C.-C., 2004. Wet periods on northeastern Brazil over the past 210 kyr linked to distant climate anomalies. Nature 432, 740Ð743. Webb, S. D., 1995. Biological implications of the Middle Miocene Amazon Seaway. Sci- ence 269, 361Ð362. Wesselingh, F., Räsänen, M., Irion, G., Vonhof, H., Kaandorp, R., Renema, W., Romero Pittman, L., Gringas, M., 2002. Lake Pebas: A paleo-ecological reconstruction of a Miocene long-lived lake complex in Western Amazonia. Cainozoic Research 1 (1-2), 35Ð81. Wesselingh, F. P., 1993. On the systematics of Miocene aquatic molluscs from Los Chorros (Colombia). Msc thesis, Vrije Universiteit Amsterdam. Wesselingh, F. P., 2000. On relict hydrobiid species in Brazilian Amazonia (Gastropoda, Prosobranchia, Hydrobiidae). Basteria 64, 129Ð136. Wesselingh, F. P., 2006a. Molluscs from the Miocene Pebas Formation of Peruvian and Colombian Amazonia. Scripta Geologica 133, 19Ð290. Wesselingh, F. P., 2006b. Evolutionairy ecology of the Pachydontinae (Bivalvia, Corbulidae) in the Pebas lake/wetland system (Miocene, western Amazonia). Scripta Geologica 133, 395Ð417. Wesselingh, F. P., Hoorn, M. C., Guerrero, J., Räsänen, M. E., Renema, W., Gingras, M., 2006a. The stratigraphy and regional structure of Miocene depostis in western Amazonia (Peru, Colombia and Brazil), with implications for Late Neogene landscape evolution. Scripta Geologica 133, 291Ð361. Wesselingh, F. P., Kaandorp, R. J. G., Vonhof, H. B., Räsänen, M., Renema, W., 2006b. The nature of aquatic landscapes in the Miocene of western Amazonia: an integrated palaeontological and geochemical approach. Scripta Geologica 133, 363Ð393. REFERENCES 119

Wesselingh, F. P., Salo, J. A., 2006. A Miocene perspective on the evolution of the Amazon biota. Scripta Geologica 133, 439Ð458. Whatley, R., Muñoz Torres, F., van Harten, D., 1998. The ostracoda of an isolated Neogen saline lake in the Western Amazon Basin. Bulletin Du Centre De Recherches Exploration Production Elf, Mémoire 20, 231Ð245. Williams, E., Antonia, A. D., Antonia, V. D., Almeida, J. M. d., Suarez, F., Liebmann, B., Malhado, A. C. M., 2005. The Drought of the Century in the Amazon Basin: An Analysis of the Regional Variation of Rainfall in South America in 1926. Acta Amazonica 35 (2), 231Ð238. Woodward, H., 1871. The Tertiary shells of the Amazons valley. Annals and Magazin natural History, London 7, 59Ð64, 101Ð109. Wurster, C. M., Patterson, W. P., 2001. Seasonal variation in stable oxygen and carbon isotope values recovered from modern lacustrine freshwater molluscs: Paleoclimatological implications for sub-weekly temperature records. Journal of Paleolimnology 26, 205Ð 218. Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rythms, and aber- rations in global climate 65 Ma to Present. Science 292, 686Ð693. Zhakarova, E. A., Kouraev, A. V., Cazenave, A., Seyler, F., 2006. Amazon River discharge estimated from TOPEX/Poseidon altimetry. Comptes Rendus Geosciences 338 (3), 188Ð 196. Zhang, J., Quay, P. D., Wilbur, D. O., 1995. Carbon isotope fractionation during gas-water exchange and dissolution of CO2. Geochimica et Cosmochimica Acta 59 (1), 107Ð114. 120 REFERENCES Appendix AÐ Isotope data of modern molluscs

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O # δ C δ O # δ C δ O 1 -3.33 -2.70 24 -13.43 -7.88 47 -7.48 -3.58 70 -6.50 -2.68 93 2 -3.10 -2.64 25 -11.67 -8.52 48 -7.44 -3.36 71 -6.22 -2.19 94 -8.94 -3.71 3 -3.33 -1.40 26 -8.46 -6.45 49 -7.30 -3.37 72 -6.34 -1.92 95 -9.21 -3.79 4 -2.56 -0.32 27 -5.91 -4.72 50 -7.24 -3.15 73 -6.43 -2.04 96 -8.41 -3.52 5 -2.43 -0.78 28 -5.21 -4.47 51 -7.22 -3.24 74 -6.14 -1.94 97 -9.81 -3.86 6 -2.13 -1.15 29 -5.59 -4.82 52 -7.09 -3.13 75 -6.38 -2.10 98 -10.04 -3.88 7 -1.31 -2.40 30 -6.89 -5.53 53 -7.17 -3.17 76 -6.28 -2.03 99 -9.04 -3.73 8 -2.76 -4.37 31 -6.57 -4.79 54 -7.03 -3.06 77 -6.49 -2.69 100 -9.76 -3.89 9 -10.58 -8.60 32 -6.60 -4.34 55 -7.07 -3.25 78 -6.93 -3.13 101 -9.42 -3.92 10 -10.11 -8.24 33 -6.76 -4.24 56 -6.89 -2.91 79 -7.31 -3.10 102 -9.70 -3.98 11 -7.82 -5.48 34 -7.14 -4.17 57 -6.90 -3.06 80 -7.37 -3.14 103 -9.72 -4.41 12 -12.10 -2.81 35 -7.45 -4.08 58 -6.90 -3.02 81 -7.36 -3.18 104 13 -11.59 -2.46 36 -7.40 -3.88 59 -6.83 -3.02 82 -7.05 -3.12 105 -10.35 -4.36 14 -9.26 0.60 37 -7.46 -3.94 60 -6.81 -2.47 83 -6.95 -3.02 106 -10.30 -4.25 15 -8.67 -0.81 38 -7.54 -3.79 61 -6.84 -2.81 84 -8.02 -3.29 107 -10.55 -4.35 16 -8.91 -1.21 39 -7.54 -3.83 62 -6.82 -2.72 85 -8.59 -3.70 108 -8.37 -3.49 17 -9.61 -2.01 40 -7.48 -3.76 63 -7.17 -3.54 86 -8.44 -3.60 109 -10.21 -4.23 18 -9.98 -2.95 41 -7.47 -3.65 64 -7.19 -3.79 87 -8.46 -3.65 110 -9.98 -4.16 19 -10.36 -3.52 42 -7.40 -3.45 65 -6.69 -2.83 88 -8.42 -3.56 111 -10.14 -4.21 20 -11.98 -5.15 43 -7.28 -3.50 66 -6.15 -2.27 89 -8.85 -3.82 112 -10.18 -4.25 21 -12.06 -5.06 44 -7.31 -3.54 67 -6.01 -2.44 90 -9.26 -3.83 113 -9.96 -4.04 22 -12.87 -6.13 45 -7.31 -3.51 68 -6.16 -2.66 91 -9.36 -3.83 114 -10.37 -4.07 23 -12.64 -5.91 46 -7.42 -3.54 69 -6.15 -2.57 92 -8.66 -3.62

Table A-7: Anodontites trapesialis sampled in Playa Cocha on 15-11-1998

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O # δ C δ O # δ C δ O 1 -1.84 -2.45 10 -11.29 -8.43 19 -9.02 -6.29 28 -7.07 -3.89 37 -7.16 -3.67 2 -2.24 -2.95 11 -8.30 -6.49 20 -6.69 -3.91 29 -7.17 -3.84 38 -7.30 -3.55 3 -3.01 -3.03 12 -11.64 -3.53 21 -6.93 -4.54 30 -7.54 -4.15 39 -7.05 -3.43 4 -3.97 -2.45 13 -9.46 -0.51 22 -7.75 -5.32 31 -8.10 -4.11 40 -6.95 -3.37 5 -4.05 -0.78 14 -11.22 -2.96 23 -7.98 -5.49 32 -8.31 -3.75 41 -6.94 -3.00 6 -3.53 -0.29 15 -14.46 -6.45 24 -6.86 -4.65 33 -8.35 -3.84 42 -7.48 -3.23 7 -2.73 -0.96 16 -14.91 -7.21 25 -6.71 -4.24 34 -8.30 -3.87 8 -1.46 -2.97 17 -13.32 -8.19 26 -6.83 -4.34 35 -7.12 -3.85 9 -4.30 -5.27 18 -11.75 -9.37 27 -6.86 -3.79 36 -6.96 -3.94

Table A-8: Anodontites trapesialis sampled in Playa Cocha on 31-01-1999 122 Appendix AÐ Isotope data of modern molluscs

13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O # δ C δ O 1 -5,45 -2,14 21 -9,44 -7,84 41 -12,21 -9,22 61 -7,56 -3,85 2 -6,05 -2,31 22 -11,49 -8,65 42 -12,11 -9,06 62 -7,42 -3,46 3 -6,23 -2,59 23 -7,38 -6,00 43 -12,49 -8,93 63 -7,64 -3,49 4 -6,49 -2,38 24 -10,62 -2,50 44 -12,16 -8,39 64 -7,68 -3,48 5 -7,16 -2,83 25 -11,11 -1,20 45 -10,32 -6,72 65 -7,90 -3,49 6 -9,24 -4,35 26 -10,73 0,35 46 -6,67 -3,31 66 -8,14 -3,57 7 -10,73 -5,48 27 -9,63 0,37 47 -6,91 -3,49 67 -8,31 -3,76 8 -11,84 -6,22 28 -9,21 -0,10 48 -7,12 -3,63 68 -8,37 -3,53 9 -12,30 -6,53 29 -9,38 -0,73 49 -7,24 -3,87 69 -7,79 -3,40 10 -12,46 -6,77 30 -9,87 -1,35 50 -7,08 -4,31 70 -7,55 -3,42 11 -12,57 -7,41 31 -10,70 -1,91 51 -7,31 -5,04 71 -7,43 -3,32 12 -12,27 -9,51 32 -11,26 -2,49 52 -7,45 -5,03 72 -7,42 -3,27 13 -12,16 -9,91 33 -10,99 -2,82 53 -7,68 -5,11 73 -7,54 -3,60 14 -12,38 -9,84 34 -11,31 -3,29 54 -7,73 -5,11 74 -7,90 -3,54 15 -11,80 -9,74 35 -11,38 -3,79 55 -7,72 -5,24 75 -7,48 -3,26 16 -10,30 -8,70 36 -11,99 -4,28 56 -7,32 -5,32 76 -6,74 -3,28 17 -1,45 -2,24 37 -12,17 -4,76 57 -7,31 -5,07 77 -6,59 -3,39 18 -1,12 -1,98 38 -12,81 -5,95 58 -7,40 -4,67 19 -1,18 -2,66 39 -14,10 -7,19 59 -7,20 -4,31 20 -1,27 -3,28 40 -13,52 -7,96 60 -7,38 -4,01

Table A-9: Anodontites trapesialis sampled in Playa Cocha on 21-10-1999

13 18 13 18 13 18 # δ C δ Omm#δ C δ Omm#δ C δ O mm*) 1 -8,25 -2,14 112,37 55 -13,29 -6,08 72,64 109 -5,51 -3,5 34,37 2 -8,2 -2,1 112,33 56 -13,32 -6,14 72,29 110 -4,86 -3,2 33,15 3 -8,24 -2,17 112,3 57 -13,35 -6,23 71,94 111 -4,82 -3,24 31,94 4 -8,3 -2,15 112,26 58 -13,34 -6,22 71,59 112 -4,58 -3,27 31,39 5 -8,91 -2,4 112,23 59 -13,46 -6,29 71,24 113 -5,14 -4,02 30,98 6 -9,25 -2,44 112,19 60 -13,52 -6,36 70,89 114 30,5 7 -9,15 -2,38 112,16 61 -13,71 -6,52 70,54 115 -5,28 -4,21 28,92 8 -8,88 -2,48 111,35 62 -13,6 -6,71 70,28 116 -6,24 -4,48 28,03 9 -8,63 -2,47 111,11 63 -13,44 -6,42 69,89 117 -6,22 -4,37 27,1 10 -8,49 -2,71 109,76 64 -13,46 -6,63 69,57 118 -5,4 -3,85 26,7 11 -8,47 -2,75 109,16 65 -13,37 -6,51 69,16 119 -5,2 -3,55 25,45 12 -8,79 -3,09 108,46 66 68,7 120 -4,86 -3,38 24,36 13 -8,95 -2,92 108,04 67 -13,38 -6,33 67,88 121 -4,66 -3,54 23,48 14 -8,69 -2,58 106,97 68 -13,48 -6,3 66,99 122 -4,52 -3,44 22,56 15 -9,22 -2,19 106,14 69 -13,6 -6,41 65,6 123 21,58 16 -9,43 -2,28 105,22 70 -13,62 -6,52 64,59 124 -6,66 -4,3 20,65 17 -9,51 -2,32 105,15 71 -13,63 -6,53 62,81 125 -7,21 -4,68 20,08 18 -9,37 -2,46 105,13 72 -13,68 -6,74 61,3 126 -6,17 -4,15 19,45 19 -9,37 -2,52 105,05 73 -13,48 -6,79 59,44 127 -6,31 -4,27 18,8 20 -9,63 -2,54 105,02 74 58,23 128 -5,25 -3,86 18,19 21 -9,83 -2,42 104,99 75 -13,73 -7,1 57,2 129 -4,74 -3,57 17,56 22 -10,02 -2,56 104,96 76 -13,75 -7,27 55,9 130 -4,55 -3,34 16,98 23 -10,32 -3,09 103,53 77 -13,87 -7,57 54,7 131 -4,9 -3,54 16,42 24 -10,65 -2,92 101,71 78 -13,95 -7,52 52,97 132 -5,32 -3,73 15,85 25 -10,87 -3,24 100,01 79 -14,31 -7,6 51,93 133 -5,05 -3,71 15,12 26 -10,9 -3,41 99,29 80 -14,43 -7,66 51,03 134 -4,83 -3,53 14,57 27 -11,21 -3,71 98,13 81 -14,79 -7,9 50,62 135 -5,06 -3,7 14,06 28 -11,43 -3,9 97,37 82 -13,89 -7,35 49,93 136 13,21 29 -11,46 -3,97 97,17 83 -14 -7,9 48,9 137 -5,62 -4,17 12,65 30 -11,81 -4,09 96,74 84 -13,76 -7,67 48,18 138 12 31 -12,07 -4,15 94,56 85 -13,85 -7,78 47,6 139 -5,82 -4,23 11,67 32 -11,74 -4,4 93,2 86 -13,66 -7,83 47,2 140 11,14 33 -12,09 -4,7 91,1 87 -13,65 -7,88 46,81 141 -5,87 -4,22 10,53 34 -12,77 -4,69 90,51 88 -13,5 -7,59 46,36 142 -5,53 -3,9 9,96 35 -13,28 -4,68 90,24 89 -13,61 -7,4 45,81 143 -6,27 -4,52 9,43 36 -13,33 -4,47 89,41 90 -13,57 -7,42 45,12 144 -7,15 -4,81 8,77 37 -13,28 -4,58 88,59 91 -13,75 -7,59 44,89 145 -7,16 -4,75 8,24 38 -13,36 -4,55 87,8 92 -13,67 -7,64 44,7 146 -6,41 -4,2 7,74 39 -13,38 -4,64 87 93 -13,68 -7,82 44,39 147 -7,2 -5,08 7,21 40 -13,19 -5,11 85,85 94 -13,45 -7,97 44,08 148 -6,59 -4,91 6,73 41 -13,2 -5,51 84,25 95 -13,48 -7,89 43,69 149 6,28 42 -13 -5,46 82,63 96 -13,28 -7,91 42,9 150 -7,01 -5,03 5,8 43 -13,12 -5,53 81,82 97 -13,4 -7,6 42,5 151 -5,92 -4,47 5,27 44 -13,12 -5,88 80,72 98 -13,2 -7,81 41,93 152 4,84 45 -13,13 -5,93 79,62 99 -11,32 -8,45 41,61 153 -3,67 -2,7 4,24 46 -13,32 -5,99 78,52 100 -5,81 -4,19 41,46 154 -3,71 -2,41 3,64 47 -13,43 -5,85 77,42 101 -5,4 -4,19 40,9 155 -3,76 -2,28 3,27 48 -13,3 -5,94 76,32 102 -9,12 -6,1 40,12 156 -3,72 -2,22 2,63 49 -13,12 -6 75,76 103 -5,36 -3,34 38,89 157 -3,98 -2,84 2,2 50 -13,12 -6,06 75,22 104 -5,12 -3,07 37,38 158 -3,8 -2,78 1,42 51 -13,13 -6,12 74,66 105 -5,14 -3,27 37 159 1,06 52 -13,24 -6,01 74,11 106 -7,14 -4,79 36,09 160 0,65 53 -13,28 -6,05 73,55 107 -7,19 -4,99 35,73 54 -13,31 -6,11 73 108 -4,1 -2,63 35,22

Table A-10: Juvenile Anodontites trapesialis sampled in Playa Cocha on 21-10-1999. *)mm from umbo 123

13 18 13 18 13 18 # δ C δ Omm # δ C δ Omm # δ C δ O mm*) 1 153,00 -11,14 -5,23 97 108,95 -12,78 -5,74 193 86,60 -13,06 -5,57 2 152,87 -10,63 -5,15 98 108,63 -12,80 -5,78 194 86,38 -13,18 -5,74 3 152,74 -11,22 -5,44 99 108,32 -12,82 -6,06 195 86,17 -13,09 -5,87 4 152,61 -10,92 -6,28 100 108,00 -12,81 -6,35 196 85,96 -12,91 -6,12 5 152,48 -10,89 -6,10 101 107,76 -12,86 -7,51 197 85,75 -12,70 -6,12 6 152,35 -10,95 -7,18 102 107,51 -12,93 -7,55 198 85,46 -12,54 -6,24 7 152,22 -10,93 -5,29 103 107,27 -13,01 -7,32 199 85,17 -12,31 -6,47 8 152,09 -10,94 -5,21 104 107,02 -13,06 -7,74 200 84,89 -12,29 -6,53 9 151,96 -11,24 -5,65 105 106,78 -12,93 -7,89 201 84,60 -12,39 -6,63 10 151,84 -10,30 -7,88 106 106,53 -13,12 -7,90 202 84,31 -12,48 -7,03 11 151,40 -11,43 -5,56 107 106,29 -13,03 -8,04 203 84,03 -12,45 -7,25 12 151,30 -10,55 -5,04 108 106,04 -12,99 -8,07 204 83,74 -12,37 -7,28 13 151,19 -10,74 -5,82 109 105,80 -13,07 -7,70 205 83,45 -12,34 -7,60 14 151,09 -10,65 -4,37 110 105,55 -13,05 -7,39 206 83,17 -12,39 -8,01 15 150,98 -11,32 -4,87 111 105,31 -13,02 -7,15 207 82,88 -12,62 -8,65 16 150,88 -11,24 -4,85 112 105,06 -12,98 -7,07 208 82,67 -12,97 -9,56 17 150,77 -11,21 -5,28 113 104,82 -12,99 -6,64 209 82,46 -13,12 -10,12 18 150,67 -10,98 -6,41 114 104,57 -13,01 -6,45 210 82,25 -12,72 -9,97 19 150,18 -10,59 -5,20 115 104,33 -13,06 -6,28 211 82,04 -12,21 -9,97 20 149,68 -11,50 -6,13 116 104,08 -12,53 -6,03 212 81,83 -11,61 -9,13 21 149,19 -11,01 -6,05 117 103,84 -12,05 -5,64 213 81,62 -11,62 -8,46 22 148,25 -11,23 -6,64 118 103,51 -12,58 -5,96 214 81,41 -11,68 -8,30 23 148,10 -10,83 -6,48 119 103,19 -12,74 -6,10 215 81,20 -12,62 -8,21 24 147,95 -11,11 -5,34 120 102,86 -12,90 -6,27 216 80,99 -12,72 -6,53 25 147,79 -10,82 -6,51 121 102,54 -13,11 -6,39 217 80,78 -12,71 -5,95 26 147,64 -10,78 -6,78 122 102,21 -13,23 -6,60 218 80,57 -12,66 -5,80 27 146,50 -10,96 -6,32 123 101,89 -13,13 -6,70 219 80,07 -12,35 -5,94 28 146,32 -11,50 -4,74 124 101,56 -13,56 -7,59 220 79,56 -12,07 -6,09 29 146,14 -11,22 -4,86 125 101,24 -13,79 -8,39 221 79,06 -11,88 -6,25 30 145,97 -10,84 -5,98 126 100,91 -13,74 -8,65 222 78,56 -11,95 -6,29 31 145,79 -10,98 -5,08 127 100,58 -13,70 -9,06 223 78,05 -12,21 -6,10 32 145,61 -12,04 -5,79 128 100,26 -13,80 -8,29 224 77,55 -12,52 -6,12 33 144,72 -11,85 -5,81 129 99,93 -13,53 -7,90 225 77,05 -12,24 -6,17 34 144,64 -12,00 -6,34 130 99,61 -13,21 -7,10 226 76,54 -12,07 -6,26 35 144,56 -10,73 -6,48 131 99,28 -13,43 -6,74 227 76,04 -12,36 -6,02 36 144,48 -11,50 -4,34 132 98,96 -13,57 -6,24 228 75,54 -12,54 -5,72 37 144,41 -10,52 -5,87 133 98,63 -13,32 -6,10 229 75,03 -12,77 -5,66 38 143,30 -10,44 -5,67 134 98,31 -12,66 -5,93 230 74,53 -12,80 -5,30 39 143,11 -10,66 -6,53 135 97,98 -12,80 -6,02 231 73,58 -12,88 -5,08 40 142,92 -10,91 -5,25 136 97,78 -13,15 -5,86 232 72,63 -13,04 -4,58 41 142,73 -10,93 -6,86 137 97,57 -13,31 -5,66 233 71,68 -12,51 -5,10 42 142,54 -11,21 -6,14 138 97,37 -13,22 -5,41 234 71,38 -11,97 -5,46 43 142,35 -11,46 -6,57 139 97,17 -12,86 -5,42 235 71,09 -12,05 -5,62 44 142,08 -11,44 -6,46 140 96,96 -12,58 -5,78 236 70,79 -12,38 -5,23 45 141,81 -10,71 -7,22 141 96,76 -12,62 -6,05 237 70,49 -12,33 -5,67 46 141,54 -10,92 -5,88 142 96,56 -12,67 -6,05 238 70,20 -11,99 -5,87 47 141,37 -11,88 -6,17 143 96,35 -12,44 -5,95 239 69,90 -11,98 -6,06 48 141,21 -11,12 -5,63 144 96,15 -12,21 -5,86 240 69,60 -12,23 -6,07 49 141,05 -11,23 -8,08 145 95,95 -12,21 -5,88 241 69,30 -13,06 -5,58 50 140,88 -11,30 -8,66 146 95,74 -12,40 -6,01 242 69,01 -13,33 -5,58 51 140,72 -12,32 -6,95 147 95,54 -12,13 -6,26 243 68,71 -12,73 -5,82 52 140,56 -11,76 -6,08 148 95,34 -12,04 -6,76 244 68,41 -12,13 -6,10 53 138,63 -11,59 -6,93 149 95,14 -12,34 -7,52 245 68,12 -12,00 -6,21 54 138,39 -10,32 -7,82 150 94,86 -12,49 -8,00 246 67,82 -12,04 -6,38 55 138,15 -12,29 -6,12 151 94,80 -12,56 -8,61 247 67,23 -12,37 -6,36 56 137,91 -11,99 -6,14 152 94,74 -12,38 -8,65 248 66,64 -12,45 -6,44 57 137,68 -12,25 -6,65 153 94,68 -12,42 -8,75 249 66,05 -12,93 -6,95 58 137,44 -10,67 -8,12 154 94,62 -12,66 -9,15 250 65,46 -13,34 -7,35 59 136,08 -11,92 -6,46 155 94,56 -12,84 -9,03 251 64,87 -13,63 -7,29 60 135,35 -11,89 -6,18 156 94,51 -13,01 -8,49 252 64,65 -14,27 -7,64 61 134,71 -11,57 -5,88 157 94,47 -12,87 -7,70 253 64,44 -14,21 -8,10 62 134,06 -11,09 -6,33 158 94,43 -12,83 -7,47 254 64,22 -14,32 -8,32 63 130,00 -10,95 -8,24 159 94,39 -12,68 -7,06 255 64,00 -14,08 -8,46 64 128,86 -11,35 -7,28 160 94,36 -12,71 -7,00 256 63,79 -14,23 -8,81 65 127,73 -11,59 -6,24 161 94,32 -12,63 -6,01 257 63,57 -14,12 -9,10 66 126,59 -10,77 -7,59 162 94,28 -12,30 -5,39 258 63,29 -13,71 -9,04 67 125,45 -11,29 -5,45 163 94,25 -12,02 -5,52 259 63,00 -13,28 -9,15 68 124,32 -11,74 -7,08 164 94,21 -12,02 -5,26 260 62,72 -12,68 -9,42 69 123,18 -11,90 -5,88 165 94,17 -12,12 -5,61 261 62,43 -12,28 -9,43 70 122,04 -11,40 -6,05 166 94,13 -12,13 -5,70 262 62,15 -12,23 -9,61 71 120,91 -12,40 -6,49 167 94,10 -12,12 -5,92 263 61,86 -11,94 -9,46 72 119,77 -12,16 -5,85 168 94,06 -12,35 -6,02 264 61,58 -11,81 -9,69 73 118,63 -12,72 -7,69 169 93,81 -12,48 -6,44 265 61,14 -12,11 -10,68

Table A-11: Triplodon corrugatus (Río Itaya). *)mm from umbo 124 Appendix AÐ Isotope data of modern molluscs

13 18 13 18 13 18 # δ C δ Omm # δ C δ Omm # δ C δ O mm*) 74 117,50 -12,59 -7,72 170 93,57 -12,71 -6,32 266 60,70 -10,91 -9,13 75 116,36 -13,21 -8,66 171 93,32 -12,46 -6,20 267 60,26 -10,42 -8,74 76 115,97 -13,80 -8,60 172 93,07 -12,73 -6,30 268 56,42 -11,56 -7,83 77 115,57 -13,56 -7,97 173 92,82 -12,56 -5,46 269 56,34 -11,80 -7,04 78 115,18 -13,67 -7,20 174 92,58 -12,72 -4,69 270 56,26 -12,46 -6,53 79 114,79 -12,86 -5,64 175 92,33 -12,58 -4,28 271 56,18 -12,65 -6,63 80 114,39 -13,42 -6,12 176 92,08 -12,83 -4,94 272 56,10 -12,59 -6,66 81 114,00 -13,25 -5,99 177 91,83 -12,91 -5,25 273 56,02 -12,30 -5,28 82 113,68 -12,97 -6,61 178 91,59 -12,28 -5,30 274 55,94 -12,14 -5,58 83 113,37 -12,23 -5,96 179 91,34 -11,61 -5,68 275 55,86 -12,02 -5,41 84 113,05 -12,46 -6,21 180 91,09 -11,67 -5,71 276 55,78 -11,94 -5,43 85 112,74 -12,80 -7,05 181 90,84 -12,78 -4,91 277 55,70 -12,30 -5,48 86 112,42 -13,14 -8,26 182 90,60 -12,90 -4,67 278 55,62 -12,52 -5,38 87 112,11 -13,68 -9,73 183 90,35 -12,70 -5,07 279 55,54 -12,46 -5,25 88 111,79 -13,11 -9,17 184 90,10 -12,25 -5,27 280 55,46 -12,63 -5,56 89 111,47 -12,02 -8,22 185 89,85 -12,17 -5,71 281 55,38 -12,70 -5,42 90 111,16 -11,69 -7,41 186 89,61 -12,01 -6,19 282 55,30 -12,16 -3,92 91 110,84 -11,93 -6,78 187 89,36 -11,93 -6,33 283 54,44 -13,26 -5,46 92 110,53 -12,84 -6,57 188 89,11 -12,13 -6,30 284 52,46 -12,59 -5,76 93 110,21 -13,00 -7,08 189 87,46 -12,24 -6,15 285 51,98 -12,51 -5,98 94 109,89 -12,78 -6,51 190 87,23 -12,23 -5,64 286 51,50 -12,41 -6,43 95 109,58 -12,78 -5,54 191 87,02 -12,43 -5,58 96 109,26 -12,97 -6,12 192 86,81 -12,79 -5,66

Table A-12: Triplodon corrugatus (Río Itaya). *)mm from umbo. continuation Appendix B Ð Miocene shells: isotope data and minor elements

13 18 13 18 # δ C δ OFeBaMgSr# δ C δ OFeBaMgSr 2 -6,63 -1,09 336,57 529,49 1009,04 97 -9,91 -4,07 83,88 915,86 3 -6,11 -1,11 98 -9,35 -3,89 4 -5,57 -1,37 159,47 246,43 911,73 99 -9,50 -4,16 55,86 105,84 35,71 921,10 5 -6,06 -2,18 100 -9,52 -4,35 6 -7,04 -2,40 284,24 395,09 994,87 101 -9,32 -4,66 48,71 102,16 37,20 905,55 7 -7,20 -2,51 102 -9,25 -4,78 8 -8,20 -2,97 389,75 451,04 878,03 103 -9,44 -4,86 56,34 99,67 46,98 891,73 9 -8,34 -3,11 104 -9,65 -4,90 10 -8,88 -3,41 40,01 302,36 105 -9,57 -4,95 11 -9,15 -3,40 106 -9,11 -5,31 12 -9,31 -3,64 1679,97 283,54 107 -8,24 -5,34 13 -9,40 -3,82 108 -7,52 -5,35 14 -9,03 -3,53 330,99 68,76 165,25 853,51 109 -7,12 -5,35 67,11 111,70 47,38 928,90 15 -9,68 -3,93 110 -7,22 -5,22 16 -8,10 -2,85 160,48 55,60 142,70 111 -7,25 -5,09 17 -7,59 -2,43 112 -7,46 -4,37 95,05 87,65 145,32 975,42 18 -7,19 -2,25 286,10 14,87 259,96 923,23 113 -7,43 -4,47 19 -7,11 -2,09 114 -7,77 -4,08 20 -7,92 -2,41 310,76 57,18 157,88 883,91 115 -7,90 -3,76 21 -7,73 -2,16 116 -7,96 -2,99 22 -8,28 -2,31 560,30 14,89 176,14 869,17 117 -7,56 -2,57 23 -8,79 -2,32 118 -7,59 -2,66 24 -7,83 -2,22 1013,83 349,69 820,71 119 -7,58 -2,72 25 -8,53 -2,41 120 -7,42 -2,65 26 -8,42 -2,63 306,90 76,19 122,86 832,66 121 -7,45 -2,74 27 -8,71 -2,74 122 -7,52 -2,88 28 -9,09 -2,84 221,60 59,75 122,89 816,84 123 -7,60 -2,90 29 -9,56 -2,90 124 -7,48 -2,79 30 -8,93 -2,70 261,54 32,91 186,98 818,41 125 -7,52 -2,93 31 -9,54 -2,90 126 -7,51 -2,98 32 -8,81 -3,10 357,86 68,15 33,12 825,27 127 -7,22 -3,01 33 -8,72 -3,14 128 -6,89 -2,89 34 -8,71 -3,39 138,05 45,34 51,33 789,36 129 -6,77 -2,92 35 -8,95 -3,43 130 -7,13 -3,17 36 -8,27 -3,70 131 -7,02 -3,22 37 -8,23 -3,64 132 -7,01 -3,36 48,27 122,33 51,11 972,42 38 -8,55 -3,68 133 -7,02 -3,31 40 -8,80 -3,71 850,86 68,05 52,64 872,39 134 -7,36 -3,28

Table B-13: Anodontites capax (Los Chorros, 89CH). 126 Appendix B Ð Miocene shells: isotope data and minor elements

13 18 13 18 # δ C δ OFeBaMgSr# δ C δ OFeBaMgSr 42 -8,96 -3,67 134,59 79,42 93,99 937,53 135 -7,95 -3,09 43 -8,40 -3,27 136 -8,58 -3,18 44 -8,71 -3,46 70,49 110,69 46,15 970,08 137 -8,85 -3,01 45 138 -8,98 -2,80 46 -8,39 -3,22 64,63 96,87 61,60 1000,36 139 -9,42 -2,89 47 -8,56 -3,28 140 -10,14 -3,11 65,93 123,69 85,12 1027,60 48 -8,65 -3,07 71,11 63,50 82,72 945,61 141 -10,12 -2,90 49 -8,71 -2,83 142 -9,75 -2,89 50 -9,01 -3,26 45,24 111,68 46,17 933,31 143 -9,42 -2,73 51 -8,84 -3,28 144 -9,18 -2,82 60,11 115,14 59,98 995,38 52 -8,89 -3,07 48,56 99,71 45,50 933,30 145 -9,15 -2,66 53 -9,01 -3,00 146 -9,41 -2,65 54 -8,84 -2,88 54,51 105,09 41,16 917,93 147 -9,46 -2,72 55 -8,80 -3,01 148 -9,32 -2,69 56 -8,66 -3,02 56,88 104,75 40,37 906,67 149 -9,11 -2,88 57 -8,68 -2,95 150 -8,60 -3,11 58 -8,59 -2,78 60,04 96,32 41,42 881,36 151 -8,25 -3,01 59 -8,48 -2,92 152 -7,23 -3,17 41,26 126,13 55,22 934,07 60 -8,49 -2,80 520,88 169,17 958,81 153 -7,25 -2,98 61 -8,78 -2,82 154 -6,79 -3,35 45,82 113,62 59,71 922,49 62 -9,02 -2,48 710,96 100,93 985,98 155 -6,81 -3,42 63 -8,81 -2,58 156 -6,99 -3,44 64 -8,64 -2,55 157 -7,08 -3,20 65 -9,76 -2,50 506,74 125,72 224,56 996,21 158 -7,45 -3,28 51,38 99,82 48,56 902,24 66 -9,09 -2,61 610,93 54,91 171,88 1042,26 159 -7,65 -3,16 67 930,85 152,25 180,86 1003,86 160 -7,45 -3,07 68 -9,96 -2,70 372,97 65,39 172,34 1029,05 161 -7,13 -2,83 69 -10,25 -3,16 117,30 106,48 100,73 976,76 162 -6,96 -2,84 70 -10,23 -3,27 165 -6,83 -2,60 43,61 66,21 76,29 895,23 71 -10,17 -3,20 51,27 71,46 109,84 951,72 166 -6,83 -2,69 72 -9,99 -3,22 167 -6,62 -2,64 73 -10,11 -3,33 48,17 63,83 109,71 945,73 168 -6,16 -2,29 74 -10,31 -3,35 169 -6,08 -2,37 75 -10,43 -3,45 93,48 68,76 105,14 932,04 170 -6,27 -2,39 76 -10,14 -3,30 171 -6,04 -2,06 77 -10,47 -3,43 617,50 70,45 140,26 927,31 172 -6,02 -2,22 78 -10,30 -3,54 173 -5,96 -2,11 79 -9,74 -3,23 88,11 100,07 49,97 883,10 174 -5,73 -2,03 80 -9,50 -3,69 141,61 39,08 163,94 883,54 175 -5,29 -1,87 81 -8,66 -3,84 78,33 79,25 61,86 868,87 176 -5,28 -2,49 82 -8,35 -3,97 177 -5,08 -1,86 49,57 91,75 70,28 943,93 83 -8,89 -3,97 68,55 60,53 65,94 841,37 178 -5,21 -1,85 84 -9,06 -4,05 179 -5,21 -1,69 85 -9,14 -3,17 74,06 66,31 48,28 857,84 180 -5,35 -1,69 86 -9,34 -3,37 181 -5,20 -1,62 87 -9,12 -4,05 44,08 66,91 24,82 821,74 182 -5,24 -1,79 88 -8,93 -3,25 183 -5,31 -1,74 89 -9,15 -4,00 63,56 83,24 25,17 889,54 184 -5,36 -1,61 90 -8,90 -3,63 185 -5,29 -1,18 91 -8,62 -3,68 72,03 29,86 58,67 869,96 186 -5,34 -1,13

Table B-14: Anodontites capax (Los Chorros, 89CH). continuation

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O# δ C δ O#δ C δ O 1 -11,55 -1,75 10 -12,45 -1,11 19 -12,55 -1,37 28 -9,33 -1,32 37 -10,91 -1,78 2 -11,55 -1,68 11 -12,97 -1,10 20 -12,30 -1,33 29 -9,38 -1,33 38 -11,26 -1,87 3 -11,81 -1,18 12 -13,10 -1,01 21 -11,67 -1,10 30 -9,32 -1,20 39 -11,97 -2,08 4 -12,06 -0,96 13 -11,60 -0,88 22 -10,75 -0,95 31 -9,50 -1,29 40 -11,23 -2,35 5 -12,39 -0,89 14 -11,43 -0,80 23 -10,34 -1,10 32 -9,21 -1,18 41 -11,46 -2,34 6 -12,42 -0,85 15 -11,65 -0,86 24 -9,84 -1,08 33 -9,42 -1,19 42 -12,39 -2,83 7 -12,21 -1,02 16 -12,42 -0,82 25 -9,79 -1,10 34 -9,94 -1,30 43 -11,02 -2,93 8 -10,71 -0,83 17 -12,88 -2,63 26 -9,60 -1,09 35 -10,59 -1,51 44 -9,70 -2,99 9 -12,40 -0,83 18 -12,09 -0,86 27 -9,39 -1,30 36 -11,03 -1,84 45 -10,55 -2,88

Table B-15: Pachydon tenuis (Nuevo Horizonte, 91FW70). 127

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O# δ C δ O#δ C δ O 1 -10,16 -7,88 48 -8,74 -7,55 96 -10,80 -9,19 143 -12,15 -7,73 190 -14,56 -8,25 2 -9,59 -7,06 49 -7,97 -7,23 97 -10,76 -9,33 144 -11,48 -7,07 191 -12,53 -7,18 3 -9,42 -7,00 50 -8,24 -7,31 98 -10,54 -9,01 145 -11,23 -6,99 192 -14,31 -8,14 4 -9,75 -7,45 51 -9,31 -7,68 99 -11,00 -9,33 146 -10,84 -7,01 193 -13,98 -7,76 5 -10,05 -7,78 52 -9,82 -8,19 100 -11,67 -9,37 147 -10,50 -7,29 194 -13,38 -7,98 6 -10,14 -8,02 53 -9,71 -8,45 101 -12,73 -9,95 148 -10,89 -7,50 195 -13,43 -8,07 7 -10,17 -8,08 54 -9,51 -8,88 102 -12,55 -8,21 149 -10,66 -7,73 196 -13,17 -7,74 8 -10,07 -8,19 55 -10,20 -8,76 103 -12,99 -8,01 150 -9,84 -7,80 197 -12,46 -7,38 9 -9,58 -7,63 56 -10,49 -8,20 104 -12,61 -8,04 151 -10,03 -8,40 198 -12,23 -7,22 10 -10,47 -7,83 57 -8,23 -6,40 105 -12,97 -7,93 152 -9,99 -8,60 199 -12,48 -7,47 11 -10,65 -7,69 58 -10,99 -7,88 106 -12,16 -7,89 153 -10,17 -8,70 200 -12,50 -7,29 12 -10,73 -8,05 59 -10,88 -8,35 107 -10,89 -7,50 154 -9,94 -8,61 201 -11,89 -7,39 13 -10,76 -8,40 60 -10,71 -8,86 108 -10,73 -7,60 155 -9,78 -8,30 202 -11,82 -7,43 14 -10,65 -8,74 61 -10,78 -9,50 109 -10,41 -7,93 156 -10,21 -8,13 203 -11,65 -7,50 15 -11,05 -8,61 62 -11,02 -9,12 110 -10,45 -8,14 157 -10,51 -8,18 204 -11,89 -7,74 16 -11,25 -8,78 63 -10,23 -8,41 111 -10,54 -8,50 158 -10,69 -8,14 205 -11,97 -7,91 17 -11,19 -8,54 64 -9,22 -7,54 112 -9,94 -8,80 159 -10,86 -8,51 206 -11,90 -7,84 18 -10,91 -8,33 65 -9,00 -7,45 113 -9,63 -8,85 160 -11,65 -9,14 207 -12,56 -7,86 19 -11,08 -8,15 66 -9,27 -7,69 114 -9,49 -8,81 161 -12,78 -9,38 208 -12,61 -8,10 20 -11,71 -9,16 67 -9,45 -7,94 115 -9,62 -8,80 162 -13,92 -9,19 209 -12,36 -8,32 21 -10,64 -7,76 68 -9,71 -7,88 116 -10,10 -8,80 163 -14,41 -8,79 210 -12,57 -8,32 22 -10,36 -7,89 69 -9,89 -8,32 117 -10,49 -8,68 164 -15,25 -8,65 211 -13,36 -8,54 23 -9,64 -7,44 70 -9,74 -8,28 118 -10,71 -8,32 165 -15,24 -7,93 212 -12,04 -8,82 24 -9,46 -7,56 71 -9,65 -8,54 119 -11,07 -8,31 166 -15,28 -7,86 213 -11,45 -8,78 25 -9,44 -7,90 72 -9,44 -8,46 120 -11,33 -8,24 167 -14,07 -7,60 214 -11,76 -9,06 26 -9,76 -8,01 73 -9,08 -8,48 121 -12,02 -8,35 168 -13,15 -7,31 215 -12,32 -9,15 27 -9,56 -7,86 74 -9,22 -8,51 122 -12,85 -8,58 169 -13,04 -7,21 216 -12,03 -9,08 28 -9,74 -8,18 75 -9,53 -8,44 123 -13,26 -8,59 170 -13,16 -6,99 217 -12,22 -8,87 29 -9,89 -8,20 76 -9,79 -8,53 124 -13,03 -8,58 171 -12,93 -7,11 218 -12,47 -9,10 30 -9,97 -8,49 77 -10,79 -7,64 125 -11,13 -8,07 172 -13,33 -7,28 219 -12,79 -9,06 31 -9,44 -8,96 78 -11,12 -7,69 126 -10,63 -7,90 173 -13,05 -7,76 220 -12,58 -9,07 32 -9,25 -9,18 79 -11,28 -7,62 127 -10,46 -8,16 174 -11,32 -7,63 221 -12,49 -9,08 33 -9,25 -9,47 80 -11,25 -7,56 128 -10,15 -8,32 175 -10,75 -7,71 222 -12,86 -8,89 34 -9,18 -9,65 82 -11,38 -7,54 129 -10,02 -8,47 176 -11,30 -7,77 223 -13,09 -8,54 35 -9,31 -9,92 83 -11,50 -7,53 130 -9,87 -8,94 177 -11,16 -8,00 224 -13,01 -8,04 36 -9,46 -10,07 84 -10,14 -7,56 131 -9,40 -9,18 178 -9,89 -7,49 225 -11,65 -7,66 37 -9,71 -9,62 85 -10,04 -7,65 132 -9,43 -9,50 179 -9,48 -8,41 226 -11,49 -8,36 38 -10,03 -9,50 86 -10,30 -8,02 133 -9,74 -9,63 180 -9,79 -8,64 227 -10,48 -7,58 39 -10,34 -8,59 87 -10,42 -8,06 134 -10,07 -9,64 181 -10,48 -8,56 229 -11,34 -7,99 40 -10,71 -8,93 88 -10,47 -8,33 135 -10,33 -9,68 182 -11,15 -8,53 230 -11,26 -8,07 41 -10,91 -9,07 89 -10,49 -8,49 136 -10,24 -9,33 183 -11,52 -8,63 231 -10,76 -7,65 42 -11,13 -9,33 90 -10,72 -8,81 137 -10,63 -9,10 184 -12,09 -8,62 232 -10,31 -7,62 43 -11,21 -9,06 91 -10,29 -9,04 138 -11,23 -8,62 185 -12,99 -8,72 233 -10,21 -7,87 44 -11,49 -8,69 92 -9,88 -9,37 139 -11,84 -8,22 186 -12,83 -8,85 45 -11,36 -8,34 93 -9,42 -9,59 140 -12,35 -7,80 187 -13,07 -9,03 46 -11,04 -8,51 94 -9,84 -9,40 141 -12,38 -7,67 188 -13,56 -8,60 47 -9,76 -7,99 95 -10,30 -9,13 142 -12,20 -7,39 189 -13,97 -8,55

Table B-16: Diplodon longulus (Nuevo Horizonte, 91FW70).

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O # δ C δ O # δ C δ O 1 -6,82 -4,33 14 -10,53 -6,50 27 -8,57 -6,35 40 -8,33 -5,06 53 -9,39 -5,28 2 -7,58 -4,70 15 -9,96 -5,83 28 -7,30 -5,59 41 -9,23 -5,07 54 -8,74 -5,07 3 -7,69 -5,18 16 -9,47 -5,57 29 -8,02 -5,71 42 -9,47 -5,06 55 -9,00 -5,01 4 -7,57 -5,04 17 -9,08 -5,45 30 -7,40 -4,62 43 -9,82 -5,49 56 -8,91 -5,04 5 -7,84 -4,59 18 -8,30 -5,41 31 -10,60 -4,35 44 -10,01 -5,66 57 -9,36 -5,12 6 -7,46 -5,13 19 -8,61 -5,95 32 -11,75 -3,98 45 -10,02 -5,23 58 -8,43 -4,76 7 -7,42 -5,35 20 -9,47 -6,67 33 -11,36 -3,98 46 -10,04 -5,37 59 -9,24 -4,93 8 -7,76 -5,44 21 -9,67 -7,04 34 -10,54 -4,28 47 -10,73 -5,54 60 -9,83 -5,05 9 -10,39 -6,16 22 -10,02 -7,27 35 -9,98 -5,13 48 -10,14 -5,54 61 -10,02 -5,16 10 -11,13 -6,69 23 -10,43 -7,39 36 -9,84 -5,06 49 -10,39 -5,43 62 -9,88 -5,09 11 -10,71 -6,98 24 -9,88 -6,67 37 -9,83 -5,62 50 -10,58 -5,38 63 -9,97 -5,09 12 -10,64 -7,15 25 -9,24 -6,37 38 -9,71 -5,61 51 -10,52 -5,53 64 -10,21 -5,10 13 -10,34 -6,90 26 -9,00 -6,12 39 -9,26 -5,31 52 -10,14 -5,47 65 -10,37 -5,03

Table B-17: Diplodon longulus (Santa Elena, 96FWSE). 128 Appendix B Ð Miocene shells: isotope data and minor elements

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O# δ C δ O#δ C δ O 1 -9,44 -3,96 35 -8,37 -3,13 69 -8,05 -3,24 103 137 -9,52 -4,33 2 -8,57 -3,50 36 -8,58 -3,74 70 -7,42 -3,06 104 -8,66 -3,10 138 -9,56 -4,57 3 -9,16 -3,56 37 -9,98 -4,68 71 -8,51 -3,21 105 -8,74 -3,06 139 -9,18 -4,45 4 -9,65 -3,91 38 -11,02 -4,55 72 -8,66 -3,16 106 -8,88 -3,05 140 -9,14 -4,46 5 -10,03 -4,61 39 -11,23 -4,42 73 -9,16 -3,32 107 -8,67 -3,33 141 -7,72 -4,13 6 -9,94 -4,87 40 -11,06 -4,28 74 -9,29 -3,25 108 -10,45 -3,75 142 -7,28 -4,08 7 -10,52 -5,29 41 -10,47 -3,84 75 -8,61 -2,99 109 -10,36 -3,69 143 -7,10 -4,16 8 -10,45 -4,74 42 -8,61 -2,75 76 -8,80 -3,04 110 -10,01 -3,92 144 -7,49 -4,57 9 -8,75 -3,86 43 -8,00 -1,73 77 -8,28 -2,93 111 -9,07 -3,98 145 -8,26 -4,39 10 -7,55 -2,83 44 -8,67 -1,83 78 -8,34 -3,08 112 -10,29 -4,10 146 -8,18 -4,47 11 -8,42 -2,50 45 -8,41 -2,16 79 -7,23 -3,02 113 -8,33 -4,38 147 -8,30 -4,51 12 -9,32 -3,66 46 -7,74 -2,69 80 -7,60 -3,00 114 -8,59 -4,40 148 -9,61 -4,43 13 -10,31 -3,95 47 -9,30 -4,06 81 -7,93 -3,16 115 -9,88 -4,47 149 -8,91 -4,48 14 -9,63 -3,42 48 -10,43 -4,48 82 -8,34 -3,37 116 -9,08 -4,48 150 -8,73 -4,21 15 -9,03 -3,38 49 -10,14 -4,39 83 -8,61 -3,34 117 -8,58 -4,68 151 -8,46 -4,50 16 -8,64 -3,73 50 -9,99 -4,53 84 -8,88 -3,39 118 -7,41 -4,36 152 -8,93 -4,45 17 -9,11 -4,32 51 -10,00 -4,45 85 -8,12 -3,48 119 -8,45 -4,35 153 -8,25 -3,67 18 -7,74 -3,55 52 -9,37 -4,09 86 -8,22 -3,51 120 -7,89 -4,08 154 -8,60 -4,26 19 -8,46 -3,74 53 -9,25 -3,75 87 -8,23 -2,85 121 -8,28 -3,90 155 -7,32 -4,38 20 -11,06 -5,06 54 -9,22 -3,57 88 -8,72 -2,76 122 -8,30 -3,79 156 -7,68 -4,16 21 -10,95 -4,71 55 -9,04 -3,30 89 123 -9,25 -3,97 157 -6,90 -4,22 22 -10,11 -3,98 56 -8,17 -2,94 90 -9,71 -2,86 124 -9,88 -4,11 158 -7,28 -4,07 23 -8,92 -3,63 57 -7,34 -2,49 91 -9,89 -3,15 125 -9,99 -4,26 159 -7,87 -3,87 24 -6,80 -2,89 58 -6,98 -2,17 92 -9,05 -3,44 126 -9,16 -4,35 160 -7,66 -3,87 25 -6,35 -1,74 59 -7,30 -1,98 93 -8,89 -3,81 127 -8,71 -4,47 161 -8,43 -3,93 26 -7,27 -1,92 60 -7,31 -1,77 94 -9,15 -4,37 128 -7,65 -4,73 162 -8,21 -4,49 27 -7,47 -2,93 61 -7,07 -1,42 95 -9,91 -4,90 129 -7,90 -4,02 163 -7,30 -3,84 28 -8,63 -2,99 62 -6,68 -1,45 96 -10,44 -5,17 130 -8,26 -4,04 164 -7,90 -4,19 29 -9,41 -2,91 63 -6,69 -1,39 97 -9,56 -4,44 131 -8,34 -3,75 165 -7,46 -4,09 30 -9,10 -3,03 64 -7,38 -1,70 98 -5,83 -5,03 132 -8,50 -3,72 166 -7,88 -4,20 31 -8,20 -2,97 65 -7,31 -1,99 99 -8,81 -5,99 133 -9,39 -4,13 167 -7,89 -4,34 32 -7,35 -2,54 66 -7,55 -2,05 100 -8,59 -6,39 134 -9,24 -4,14 168 -7,69 -4,26 33 -7,89 -2,36 67 -8,25 -3,20 101 -8,97 -6,74 135 -9,40 -4,10 169 -7,87 -4,35 34 -8,30 -2,29 68 -9,55 -3,66 102 -9,04 -6,79 136 -9,64 -4,20 170 -8,19 -4,35

Table B-18: Anodontites capax (San Antonio, 96FW810).

13 18 13 18 13 18 # δ C δ Omm# δ C δ Omm# δ C δ O mm*) 1 Ð8.93 Ð7.30 35.00 23 Ð9.31 Ð8.12 32.35 45 Ð10.60 Ð6.96 27.04 2 34.90 24 Ð9.30 Ð8.37 32.18 46 Ð10.79 Ð7.26 26.70 3 Ð9.07 Ð8.63 34.79 25 Ð8.76 Ð8.30 32.02 47 Ð10.79 Ð7.23 26.14 4 Ð9.08 Ð8.56 34.69 26 Ð9.05 Ð8.27 31.86 48 Ð10.45 Ð7.39 25.58 5 Ð9.03 Ð8.61 34.58 27 Ð8.94 Ð8.38 31.70 49 Ð10.25 Ð7.62 25.01 6 34.48 28 Ð9.90 Ð7.96 31.53 50 Ð9.91 Ð7.97 24.45 7 Ð9.99 Ð7.67 34.38 29 Ð10.07 Ð7.65 31.37 51 Ð9.42 Ð8.00 24.07 8 34.27 30 Ð10.14 Ð7.27 31.21 52 Ð9.43 Ð8.04 23.69 9 Ð9.95 Ð6.82 34.17 31 Ð10.06 Ð6.87 31.05 53 Ð9.45 Ð7.84 23.32 10 Ð9.12 Ð7.14 34.06 32 Ð10.29 Ð6.95 30.89 54 Ð9.34 Ð8.20 22.94 11 Ð9.62 Ð7.59 33.96 33 30.72 55 Ð9.20 Ð8.09 22.56 12 Ð9.67 Ð7.92 33.85 34 Ð9.94 Ð6.95 30.56 56 Ð8.92 Ð8.09 22.18 13 Ð9.11 Ð8.14 33.75 35 30.40 57 Ð9.02 Ð7.98 21.81 14 33.65 36 Ð9.77 Ð7.37 30.06 58 Ð9.50 Ð7.94 21.43 15 Ð8.83 Ð8.39 33.54 37 Ð9.69 Ð7.81 29.73 59 Ð9.73 Ð7.95 21.05 16 Ð8.84 Ð8.24 33.44 38 Ð9.46 Ð7.89 29.39 60 Ð10.03 Ð7.64 20.51 17 Ð8.70 Ð7.65 33.33 39 Ð9.17 Ð7.95 29.05 61 19.97 18 33.23 40 Ð9.32 Ð7.98 28.72 62 Ð10.34 Ð7.80 19.43 19 Ð9.16 Ð7.07 33.13 41 Ð9.55 Ð7.95 28.38 63 Ð10.61 Ð7.57 18.89 20 Ð9.24 Ð7.28 32.93 42 Ð10.39 Ð7.65 28.05 64 Ð9.72 Ð7.04 18.35 21 Ð9.40 Ð7.42 32.74 43 Ð10.74 Ð6.57 27.71 65 Ð9.69 Ð7.44 17.75 22 Ð9.35 Ð7.92 32.54 44 Ð10.22 Ð6.32 27.37 66 Ð9.51 Ð7.49 17.15

Table B-19: Diplodon longulus (Santa Rosa, 96FWSR). *) mm from umbo 129

13 18 #mmδ C δ OFeBaMgSr 1 71,000 -12,29 -7,53 2 70,914 -12,29 -7,49 3 70,828 -11,82 -6,74 4 70,742 -12,48 -6,71 5 70,656 -12,83 -6,74 6 70,570 -12,68 -7,02 7 70,484 -12,54 -6,87 8 70,398 -12,45 -7,13 9 70,312 -12,54 -7,04 10 70,226 -12,78 -6,78 11 70,140 12 70,054 -13,57 -6,75 13 69,968 -13,70 -6,83 14 69,882 -13,73 -6,90 15 69,795 -13,00 -6,80 37,06 153,48 72,25 1382,71 16 69,709 -12,12 -6,39 17 69,623 -12,19 -6,55 18 69,537 -12,58 -7,04 19 69,451 -12,57 -7,07 20 69,365 -12,63 -7,02 21 69,279 -12,79 -7,39 22 69,193 -12,68 -7,82 23 69,107 -12,20 -7,89 87,36 153,98 34,47 1475,42 24 69,021 -12,15 -7,78 25 68,935 -12,48 -7,35 26 68,849 -13,21 -6,78 27 68,763 -13,21 -6,62 28 68,660 29 68,557 -12,98 -6,48 30 68,454 -11,98 -6,34 58,55 165,01 169,87 1355,39 31 68,350 -12,31 -6,41 32 68,247 187,04 177,10 431,72 1731,98 33 68,144 -13,02 -6,63 34 68,041 -12,92 -6,65 35 67,938 -12,31 -6,44 36 67,835 -12,16 -6,46 37 67,731 -12,11 -6,12 38 67,628 -12,05 -6,00 132,18 193,53 246,99 1290,16 39 67,525 -12,48 -6,23 40 67,422 -13,41 -6,84 101,64 217,12 181,44 1360,08 41 67,319 -13,23 -6,86 42 67,216 -12,13 -6,49 74,32 223,28 197,48 1413,37 43 67,112 -11,68 -6,65 44 67,009 -11,56 -7,26 148,66 258,01 431,72 1559,78 45 66,906 -12,21 -7,04 46 66,803 -11,70 -6,68 70,86 309,33 259,78 1750,56 47 66,666 -11,28 -6,52 48 66,650 -11,60 -6,93 134,99 289,94 258,17 1625,59 49 66,599 -11,90 -6,93 50 66,548 -12,13 -7,16 51 66,497 -11,92 -7,24 52 66,167 -11,61 -7,86 53 65,960 -11,31 -7,75 54 65,752 -12,17 -7,05 55 65,545 -14,68 -6,22 72,95 209,39 86,31 1378,50 56 65,421 -13,60 -6,39 57 65,298 -12,69 -6,20 58 65,174 -11,78 -6,16 59 65,050 -11,57 -6,09 60 64,926 -10,89 -6,41 61 64,803 -11,05 -6,66 67,36 170,99 87,57 1343,95 62 64,679 -11,50 -6,43 63 64,555 -11,24 -6,07 656,09 110,77 227,10 1478,30 64 64,431 -11,51 -6,15 65 64,308 -11,64 -6,09 66 64,184 -11,49 -6,18 67 64,060 -11,49 -6,20 68 63,936 -11,50 -6,05 69 63,813 -10,76 -5,79 56,31 204,32 51,31 1318,10 70 63,689 -10,29 -5,93 71 63,565 -10,06 -5,94 72 63,394 -9,86 -6,09 65,61 204,53 158,30 1244,87 73 63,222 -10,00 -6,00 74 63,051 -10,28 -6,01 85,21 190,27 45,69 1126,13 75 62,880 -12,00 -6,07 76 62,709 -11,21 -5,95 64,43 148,61 7,56 1094,49 77 62,537 -10,81 -5,71 78 62,366 -11,54 -5,66 79 62,195 -12,42 -5,59 80 62,024 -12,10 -5,70

Table B-20: Pachydon tenuis #1 (Santa Rosa, 96FWSR). *)mm from umbo 130 Appendix B Ð Miocene shells: isotope data and minor elements

13 18 #mmδ C δ OFeBaMgSr 81 61,852 -12,02 -5,66 82 61,681 -11,96 -5,81 83 61,510 -11,01 -5,54 87,76 193,64 56,29 1079,54 84 61,338 -11,11 -5,57 85 61,167 -12,26 -5,40 86 60,996 -13,73 -5,73 130,39 208,78 53,86 1049,90 87 60,825 -13,22 -5,64 88 60,653 -12,31 -5,75 89 60,482 -11,73 -5,68 90 60,311 -11,50 -5,87 91 60,140 -11,78 -6,02 92 59,968 -12,00 -6,00 64,46 189,73 37,25 1269,57 93 59,797 -12,09 -6,05 94 59,648 -12,24 -6,46 146,07 187,83 291,73 1409,05 95 59,499 -12,12 -6,45 96 59,350 -11,95 -6,56 60,28 177,70 68,35 1468,52 97 59,201 -11,47 -6,35 98 59,052 -11,59 -6,45 75,64 183,13 97,68 1502,08 99 58,903 -11,65 -6,16 100 58,754 -12,28 -5,84 101 58,605 -14,05 -6,33 102 58,456 -13,54 -6,70 103 58,307 -13,16 -6,23 104 58,158 -11,74 -5,90 105 57,965 -10,39 -5,61 63,57 239,57 44,56 1316,38 106 57,771 -9,18 -5,69 107 57,578 -6,95 -5,32 37,76 222,55 64,28 1279,36 108 57,384 -6,08 -5,16 109 57,191 -6,90 -5,47 115,71 222,64 111,30 1195,03 110 56,998 -6,95 -5,65 111 56,804 -6,01 -5,36 112 56,611 -5,62 -5,38 113 56,417 -5,41 -5,13 114 56,224 -5,64 -4,93 31,37 246,98 51,62 1235,77 115 56,031 -5,96 -4,83 116 55,837 -7,18 -4,78 117 55,644 -8,85 -5,29 118 55,450 119 55,257 -10,78 -5,98 120 54,900 -11,27 -6,29 60,10 286,65 96,65 1347,02 121 54,543 -10,73 -6,11 122 54,186 -10,51 -6,17 123 53,829 -10,18 -5,97 34,78 304,86 43,72 1329,66 124 53,472 -9,77 -5,75 125 53,115 -10,43 -6,03 80,07 262,38 41,69 1299,34 126 52,758 127 52,401 -10,75 -6,05 301,36 314,87 181,97 1325,82 128 52,117 -10,44 -6,10 129 51,833 -9,80 -6,24 130 51,549 -9,84 -6,19 131 51,265 -9,13 -6,32 132 50,980 -8,27 -6,25 133 50,696 -8,20 -6,34 564,53 326,95 830,96 1309,98 134 50,412 -8,14 -6,16 135 50,128 -8,58 -6,49 136 49,615 -8,79 -6,31 137 49,103 -9,77 -6,61 94,42 305,18 168,45 1283,39 138 48,590 -9,92 -6,45 139 48,077 -10,64 -6,72 140 47,565 -13,80 -6,46 141 47,052 -14,88 -6,58 142 46,895 -13,74 -6,33 143 46,737 -12,41 -6,09 206,63 243,64 334,30 1414,94 144 46,580 -12,30 -6,44 145 46,422 -11,63 -6,73 43,47 318,61 64,04 1469,23 146 46,119 -11,05 -6,66 147 45,816 -10,19 -7,09 162,65 296,13 302,74 1369,98 148 45,514 -10,23 -7,34 149 45,211 -9,78 -7,20 150 44,908 -9,40 -6,78 151 44,605 -8,90 -6,44 198,26 326,98 336,00 1342,56 152 44,243 -8,96 -6,48 153 43,881 -9,07 -6,37 154 43,519 -9,56 -6,76 155 43,157 -9,93 -6,49 156 42,795 -10,14 -6,61 157 42,381 -10,43 -6,50 158 41,967 -11,12 -6,22 43,80 322,90 45,72 1369,70 159 41,554 -11,79 -6,37 160 41,140 -12,74 -6,81

Table B-21: Pachydon tenuis #1 (Santa Rosa, 96FWSR). *)mm from umbo continuation 131

13 18 #mmδ C δ OFeBaMgSr 161 40,726 -13,19 -7,18 57,07 328,33 47,96 1452,27 162 40,312 -13,21 -6,98 163 39,998 -12,41 -6,61 164 39,683 -11,26 -6,44 165 39,369 -11,39 -6,34 166 39,054 -11,73 -6,35 167 38,740 -9,71 -5,84 168 37,806 -6,68 -5,48 169 36,872 -7,62 -5,65 170 35,937 -7,98 -6,03 171 35,003 -10,97 -6,49 67,52 331,93 43,76 1306,86 172 33,768 -11,64 -6,37

Table B-22: Pachydon tenuis #1 (Santa Rosa, 96FWSR). *)mm from umbo continuation 13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O# δ C δ O#δ C δ O 1 -11,06 -5,71 10 -9,75 -3,46 19 -9,75 -4,23 28 -8,11 -4,29 37 -7,38 -4,42 2 -10,90 -5,81 11 -10,56 -3,65 20 -9,36 -4,44 29 -8,28 -4,14 38 -7,05 -4,29 3 -10,03 -4,41 12 -9,97 -3,91 21 -9,05 -4,28 30 -8,55 -4,25 39 -6,98 -4,46 4 -10,17 -4,35 13 -9,79 -4,17 22 -8,90 -4,31 31 -8,63 -4,18 40 -6,83 -4,32 5 -9,59 -4,04 14 -9,73 -4,07 23 -8,70 -4,09 32 -8,89 -4,08 41 -6,98 -4,55 6 -9,54 -3,48 15 -9,18 -4,24 24 -8,77 -4,10 33 -8,39 -4,31 42 -7,23 -4,52 7 -9,57 -3,60 16 -8,95 -4,34 25 -8,76 -4,12 34 -8,41 -4,46 8 -9,55 -3,60 17 -8,58 -4,19 26 -8,26 -4,26 35 -8,18 -4,58 9 -9,37 -3,40 18 -8,45 -4,33 27 -8,00 -4,25 36 -7,56 -4,33

Table B-23: Pachydon tenuis #2 (Santa Rosa, 98RK65(Pt)). *)mm from umbo

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O# δ C δ O#δ C δ O 1 -10,04 -3,91 45 -8,62 -3,29 82 -9,02 -4,65 120 -9,53 -2,67 161 -7,83 -2,99 2 -9,32 -3,25 46 -8,47 -3,15 83 -8,68 -4,53 122 -8,47 -3,14 162 -8,42 -2,95 10 -8,54 -1,59 47 -8,60 -3,17 84 -8,09 -4,73 123 -9,15 -2,70 163 -8,00 -3,09 11 -8,53 -1,57 48 -8,69 -3,04 85 -8,64 -4,54 124 -9,14 -3,09 164 -8,70 -3,11 12 -8,45 -1,45 49 -9,08 -3,31 86 -8,40 -4,87 125 -8,76 -2,85 165 -8,04 -3,03 13 -8,52 -1,55 50 -8,44 -3,05 87 -8,33 -4,64 126 -8,30 -2,98 166 -8,61 -3,10 14 -8,40 -1,33 51 -8,53 -2,73 88 -7,98 -4,36 127 -8,22 -2,88 167 -8,72 -3,12 15 -8,42 -1,67 52 -8,53 -2,75 89 -8,12 -4,14 128 -8,10 -2,85 168 -8,29 -3,28 16 -8,07 -1,83 53 -8,80 -2,75 90 -8,56 -4,24 129 -7,76 -2,71 169 -7,51 -3,26 17 -8,18 -2,00 54 -8,87 -2,68 91 -8,49 -4,17 130 -7,93 -2,64 170 -7,92 -3,58 18 -8,12 -1,61 55 -9,47 -2,67 92 -8,50 -3,98 131 -7,94 -2,59 171 -7,16 -3,61 19 -8,16 -1,64 56 -9,72 -2,75 94 -8,42 -3,74 135 -7,38 -2,42 172 -6,84 -3,85 20 -8,08 -1,69 57 -10,03 -2,82 95 -8,40 -3,17 136 -7,08 -2,59 173 -7,08 -3,72 21 -8,05 -1,90 58 -10,43 -2,77 96 -8,50 -3,13 137 -7,32 -2,46 174 -7,16 -3,51 22 -7,79 -1,95 59 -10,44 -2,52 97 -8,21 -2,93 138 -7,34 -2,58 175 -7,05 -3,46 23 -7,49 -1,99 60 -10,63 -2,53 98 -8,39 -3,00 139 -6,99 -2,41 176 -7,71 -3,19 24 -7,63 -1,99 61 -10,55 -2,51 99 -8,73 -2,97 140 -6,96 -2,55 177 -7,46 -3,27 25 -7,90 -2,10 62 -10,37 -2,31 100 -8,06 -2,97 141 -7,13 -2,43 178 -7,34 -3,05 26 -8,32 -2,09 63 -10,21 -2,41 101 -7,73 -2,88 142 -6,81 -2,41 179 -7,32 -2,87 27 -8,41 -2,13 64 -10,13 -2,56 102 -7,91 -2,86 143 -6,78 -2,21 180 -7,46 -2,89 28 -8,34 -2,20 65 -9,91 -2,40 103 -8,39 -2,69 144 -6,81 -2,33 181 -7,28 -2,98 29 -9,13 -2,30 66 -10,35 -2,62 104 -8,68 -2,75 145 -6,84 -2,26 182 -7,12 -3,19 30 -9,17 -2,32 67 -10,48 -2,82 105 -8,72 -2,88 146 -6,67 -2,32 183 -6,96 -3,09 31 -9,07 -2,45 68 -10,40 -3,22 106 -7,96 -3,11 147 -6,40 -2,13 184 -6,73 -3,20 32 -8,96 -2,32 69 -10,14 -3,27 107 -7,73 -3,01 148 -5,65 -2,23 185 -6,95 -3,04 33 -9,14 -2,49 70 -10,14 -3,53 108 -8,08 -3,00 149 -5,68 -2,73 186 -6,72 -3,25 34 -9,18 -2,37 71 -10,07 -3,55 109 -7,78 -2,92 150 -5,49 -2,24 187 -7,17 -3,29 35 -9,30 -2,45 72 -9,91 -3,58 110 -8,44 -3,01 151 -5,46 -2,09 188 -6,93 -3,10 36 -9,03 -2,71 73 -10,12 -3,88 111 -8,73 -2,88 152 -6,10 -2,26 189 -6,98 -3,07 37 -9,17 -2,73 74 -9,87 -4,03 112 -8,41 -2,72 153 -5,44 -2,60 190 -7,10 -3,24 38 -9,03 -2,63 75 -9,72 -3,90 113 -8,81 -2,59 154 -8,46 -2,29 191 -7,12 -3,20 39 -9,27 -2,69 76 -9,77 -4,10 114 -9,17 -2,72 155 -9,07 -2,22 192 -7,12 -3,48 40 -8,76 -2,86 77 -9,60 -4,23 115 -8,62 -2,59 156 -10,28 -2,58 193 -7,04 -3,06 41 -8,19 -2,94 78 -9,51 -4,26 116 -9,49 -2,69 157 -9,77 -2,64 194 -7,24 -3,32 42 -8,06 -2,90 79 -9,20 -4,18 117 -10,10 -2,70 158 -8,57 -2,56 195 -7,20 -3,41 43 -7,93 -3,02 80 -8,91 -4,25 118 -10,20 -2,74 159 -8,91 -2,87 196 -6,82 -3,33 44 -8,36 -3,14 81 -8,78 -4,52 119 -9,92 -2,75 160 -7,47 -3,00 197 -5,96 -3,69

Table B-24: Pachydon erectus (Santa Rosa de Pichana, 98RK65c(Pe)).

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O# δ C δ O#δ C δ O 1 -11,19 -6,87 6 -7,56 -5,84 10 -9,47 -5,93 14 -8,88 -6,03 18 -3,88 -4,98 2 -10,52 -7,05 7 -8,58 -5,75 11 -9,96 -6,20 15 -6,38 -5,22 19 -5,92 -5,30 3 -10,55 -7,07 8 -9,99 -7,08 12 -10,40 -6,14 16 -7,87 -5,90 20 -5,63 -5,10 4 -7,94 -5,90 9 -9,75 -6,06 13 -9,85 -6,38 17 -4,04 -4,96 21 -6,05 -5,20 5 -8,71 -5,76

Table B-25: Pachydon tenuis (Indiana,98HV21). 132 Appendix B Ð Miocene shells: isotope data and minor elements

13 18 13 18 13 18 13 18 13 18 # δ C δ O # δ C δ O # δ C δ O # δ C δ O # δ C δ O 1 Ð7,07 Ð5,43 41 Ð8,86 Ð6,25 81 Ð10,16 Ð6,71 121 Ð6,02 Ð3,35 161 Ð9,81 Ð6,92 2 Ð7,13 Ð5,68 42 Ð8,23 Ð5,00 82 Ð10,03 Ð6,84 122 Ð6,27 Ð3,27 162 Ð9,64 Ð6,87 3 Ð8,11 Ð6,32 43 Ð6,89 Ð4,18 83 Ð10,52 Ð7,02 123 Ð7,00 Ð3,49 163 Ð9,78 Ð6,89 4 Ð6,62 Ð5,08 44 Ð7,15 Ð5,01 84 Ð10,23 Ð6,99 124 Ð7,13 Ð3,45 164 Ð9,76 Ð6,76 5 Ð6,38 Ð4,91 45 Ð7,77 Ð6,04 85 Ð10,41 Ð6,93 125 Ð7,33 Ð3,59 165 Ð9,34 Ð6,17 6 Ð7,49 Ð6,63 46 Ð8,12 Ð6,97 86 Ð10,15 Ð7,18 126 Ð7,29 Ð3,63 166 Ð10,55 Ð6,99 7 Ð6,54 Ð4,71 47 Ð8,46 Ð7,41 87 Ð9,34 Ð7,45 127 Ð7,42 Ð3,63 167 Ð10,56 Ð7,12 8 Ð6,34 Ð4,75 48 Ð8,24 Ð7,75 88 Ð9,13 Ð7,54 128 Ð7,26 Ð3,55 168 Ð10,32 Ð7,11 9 Ð6,96 Ð7,16 49 Ð9,05 Ð7,67 89 Ð9,56 Ð7,59 129 Ð7,07 Ð3,75 169 Ð9,89 Ð6,80 10 Ð8,19 Ð7,90 50 Ð9,51 Ð7,49 90 Ð9,99 Ð7,79 130 Ð7,07 Ð3,64 170 Ð9,95 Ð6,99 11 Ð7,54 Ð6,55 51 Ð10,56 Ð6,22 91 Ð8,86 Ð7,48 131 Ð6,82 Ð3,78 171 Ð10,13 Ð7,06 12 Ð5,87 Ð3,45 52 Ð7,24 Ð3,87 92 Ð9,56 Ð7,68 132 Ð7,06 Ð3,80 172 Ð10,68 Ð7,02 13 Ð7,07 Ð5,85 53 Ð6,64 Ð3,11 93 Ð9,62 Ð7,71 133 Ð7,28 Ð3,98 173 Ð10,37 Ð7,04 14 Ð8,02 Ð6,65 54 Ð7,39 Ð5,16 94 Ð9,55 Ð7,64 134 Ð6,94 Ð3,72 174 Ð10,23 Ð6,96 15 Ð7,51 Ð5,44 55 Ð7,96 Ð5,72 95 Ð10,28 Ð7,82 135 Ð7,25 Ð3,86 175 Ð10,45 Ð7,01 16 Ð7,94 Ð8,13 56 Ð9,33 Ð6,57 96 Ð10,48 Ð8,01 136 Ð7,38 Ð3,63 176 Ð10,64 Ð7,03 17 Ð5,24 Ð4,58 57 Ð10,71 Ð7,10 97 Ð11,16 Ð8,01 137 Ð8,00 Ð3,60 177 Ð9,38 Ð6,52 18 Ð6,61 Ð6,83 58 Ð10,72 Ð7,29 98 Ð10,84 Ð7,96 138 Ð8,13 Ð3,66 178 Ð9,68 Ð6,43 19 Ð5,88 Ð5,89 59 Ð11,33 Ð7,24 99 Ð10,70 Ð8,21 139 Ð8,05 Ð3,64 179 Ð10,39 Ð6,76 20 Ð5,71 Ð7,15 60 Ð10,16 Ð7,10 100 Ð10,61 Ð7,69 140 Ð7,82 Ð3,78 180 Ð10,40 Ð6,83 21 Ð7,92 Ð8,20 61 Ð11,22 Ð6,52 101 Ð10,76 Ð7,00 141 Ð7,99 Ð3,91 181 Ð10,27 Ð6,85 22 Ð6,60 Ð5,70 62 Ð11,06 Ð6,34 102 Ð9,61 Ð6,29 142 Ð8,47 Ð4,25 182 Ð9,80 Ð6,71 23 Ð4,63 Ð3,78 63 Ð11,61 Ð5,20 103 Ð9,48 Ð5,95 143 Ð8,48 Ð4,23 183 Ð9,86 Ð6,63 24 Ð6,61 Ð6,68 64 Ð11,32 Ð4,58 104 Ð8,20 Ð5,90 144 Ð8,60 Ð4,22 184 Ð10,39 Ð6,39 25 Ð7,69 Ð6,90 65 Ð10,15 Ð3,90 105 Ð8,85 Ð5,23 145 Ð8,89 Ð4,37 185 Ð10,00 Ð6,63 26 Ð7,02 Ð5,55 66 Ð9,44 Ð3,73 106 Ð8,99 Ð5,36 146 Ð9,33 Ð4,41 186 Ð10,25 Ð6,44 27 Ð5,27 Ð5,14 67 Ð8,72 Ð3,63 107 Ð9,51 Ð5,45 147 Ð9,16 Ð4,35 187 Ð10,72 Ð6,50 28 Ð7,72 Ð7,71 68 Ð9,06 Ð4,12 108 Ð9,22 Ð5,33 148 Ð9,54 Ð4,50 188 Ð10,58 Ð6,54 29 Ð7,93 Ð7,32 69 Ð9,85 Ð4,24 109 Ð9,39 Ð5,19 149 Ð9,80 Ð4,61 189 Ð11,00 Ð6,21 30 Ð6,19 Ð6,16 70 Ð9,82 Ð4,40 110 Ð7,92 Ð4,66 150 Ð9,67 Ð4,61 31 Ð7,16 Ð6,18 71 Ð9,61 Ð4,06 111 Ð6,98 Ð4,12 151 Ð9,63 Ð4,82 32 Ð6,07 Ð3,01 72 Ð9,38 Ð4,18 112 Ð7,61 Ð4,28 152 Ð9,48 Ð5,16 33 Ð7,19 Ð5,73 73 Ð9,57 Ð4,38 113 Ð6,96 Ð3,83 153 Ð9,28 Ð5,25 34 Ð8,85 Ð6,14 74 Ð9,70 Ð4,80 114 Ð6,46 Ð3,55 154 Ð9,17 Ð5,89 35 Ð8,77 Ð4,51 75 Ð10,42 Ð5,08 115 Ð6,41 Ð3,50 155 Ð8,93 Ð5,91 36 Ð6,45 Ð2,41 76 Ð10,80 Ð5,45 116 Ð6,81 Ð3,46 156 Ð8,84 Ð6,01 37 Ð7,01 Ð2,66 77 Ð11,07 Ð5,74 117 Ð7,20 Ð3,39 157 Ð8,62 Ð6,08 38 Ð7,54 Ð4,04 78 Ð10,55 Ð6,08 118 Ð7,53 Ð3,35 158 Ð9,13 Ð6,68 39 Ð7,77 Ð5,16 79 Ð10,25 Ð6,38 119 Ð6,90 Ð3,08 159 Ð9,05 Ð6,51 40 Ð8,97 Ð6,34 80 Ð9,91 Ð6,48 120 Ð5,65 Ð3,26 160 Ð9,27 Ð6,53

Table B-26: Diplodon longulus (Mazán, 99RKMz1).

13 18 13 18 # δ C δ O # δ C δ OBaFeMnSr 1 -6,35 -5,38 83 -6,80 -5,04 319,38 37,57 983,20 2578,02 2 -6,07 -5,35 84 -6,73 -5,14 3 -6,75 -7,27 85 -6,79 -4,52 4 -6,77 -7,42 86 -6,72 -3,76 326,68 -87,33 895,93 2467,86 5 -6,38 -6,35 87 -7,10 -4,30 6 -5,58 -5,75 88 -7,17 -4,38 7 -5,70 -6,39 89 -7,76 -4,89 292,61 -34,43 757,36 2323,71 8 -6,20 -7,22 90 -7,93 -4,74 9 -6,47 -7,89 91 -7,46 -4,51 10 -6,83 -8,80 92 -7,07 -4,72 305,89 31,76 730,46 2467,20 11 -7,01 -9,19 93 -7,45 -4,76 12 -7,14 -9,28 94 13 -7,25 -9,08 95 -7,74 -5,01 290,91 18,12 612,02 2321,26 14 -6,80 -8,13 96 15 -6,38 -6,35 97 -8,29 -5,89 16 -5,95 -4,81 98 -8,91 -6,39 270,27 33,98 444,13 2107,63 17 -5,46 -4,33 99 -9,05 -6,49 18 -5,44 -4,55 100 -8,97 -6,91 263,10 0,00 380,51 2056,72 19 -5,69 -5,14 101 -9,02 -6,97 20 -5,87 -5,06 102 -9,62 -7,73 260,11 -52,02 277,45 1942,17 21 -7,20 -6,63 103 -9,69 -7,32 253,75 19,90 398,03 2083,46 22 -7,35 -7,06 104 23 -7,24 -7,37 105 -9,46 -6,97 24 -6,38 -6,87 106 -9,68 -6,99 272,30 31,45 537,41 2117,28 25 -6,39 -5,39 107 -9,40 -6,48 26 -6,29 -5,06 108 -9,86 -6,48 27 -6,35 -5,57 109 -9,74 -5,86 275,17 52,64 690,32 2190,58 28 -6,43 -6,20 110 -9,36 -6,10 29 -6,79 -6,98 111 -9,29 -6,19 30 -7,42 -7,99 112 -9,22 -6,45 259,02 76,47 872,03 2177,00

Table B-27: Diplodon longulus (Mazán, 99RKMz2), continuation (1). 133

13 18 13 18 # δ C δ O # δ C δ OBaFeMnSr 31 -7,58 -8,36 113 -8,80 -6,21 32 -7,45 -8,22 114 -8,80 -7,28 33 -7,57 -8,20 115 -8,80 -7,47 280,39 45,41 897,92 2277,15 34 -7,49 -7,85 116 -8,66 -7,96 35 -7,12 -6,86 117 -8,57 -8,46 283,22 10,00 623,07 2237,40 36 -7,50 -6,20 118 -8,71 -9,07 267,70 30,02 467,85 2039,03 37 -7,58 -5,51 119 -8,31 -8,44 38 -3,03 -6,09 120 -8,83 -8,74 234,80 21,72 329,95 1922,84 39 -6,71 -3,99 121 -8,68 -8,59 40 -6,39 -3,69 122 -8,38 -9,13 41 -6,33 -4,00 123 -8,52 -9,03 223,78 35,93 236,10 1813,89 42 -7,07 -4,33 124 -8,44 -8,88 43 -7,33 -5,38 125 -8,43 -8,89 44 -7,73 -5,80 126 211,36 29,49 199,89 1701,55 45 -7,97 -6,29 127 -9,28 -9,12 46 -8,24 -6,53 128 -9,38 -9,46 47 -7,95 -6,27 129 -9,35 -9,48 207,04 19,58 334,34 1685,67 48 -8,16 -6,15 130 -9,73 -9,50 49 -8,67 -6,17 131 -9,93 -9,84 50 -9,22 -6,61 132 -10,01 -9,85 207,02 0,00 276,69 1654,14 51 -8,76 -5,92 133 -10,15 -10,08 52 -7,82 -4,62 134 -10,02 -10,29 214,66 0,00 341,09 1683,07 53 -8,01 -4,27 135 -10,15 -10,49 54 -7,50 -4,32 136 -10,20 -10,55 55 -7,18 -4,36 137 -10,51 -10,27 194,22 0,00 466,34 1606,06 56 -6,91 -4,64 138 -10,47 -10,19 57 -7,04 -4,64 139 -10,42 -10,31 58 -7,17 -5,38 140 -10,36 -10,02 205,65 0,00 388,30 1587,69 59 -7,16 -5,15 141 -11,10 -10,37 60 -7,70 -5,84 142 -11,66 -10,58 206,02 0,00 268,21 1527,62 61 -7,92 -6,23 143 -11,23 -10,05 62 -7,95 -6,22 144 -11,55 -10,05 206,48 0,00 298,56 1614,17 63 -8,14 -6,76 145 -11,00 -9,93 64 -8,65 -7,10 146 -11,14 -9,49 210,82 0,00 294,17 1729,04 65 -8,60 -7,30 147 -10,11 -8,59 66 -8,84 -7,23 148 -10,93 -8,43 226,79 0,00 396,55 1966,84 67 -8,47 -6,62 149 -10,10 -6,80 246,24 0,00 567,09 2148,99 68 -8,41 -6,45 150 -9,01 -4,75 270,91 0,00 972,19 2322,88 69 -8,46 -6,27 151 -6,92 -3,69 70 -8,38 -5,79 152 -8,18 -4,09 271,59 0,00 1030,98 2145,02 71 -8,10 -5,38 153 -8,79 -4,14 72 -7,74 -5,14 154 -8,80 -4,65 282,86 0,00 841,61 2169,51 73 -7,82 -4,89 155 -9,42 -5,17 74 -7,98 -4,73 156 -9,32 -5,78 252,08 88,97 451,02 2035,15 75 -7,37 -4,29 157 -9,31 -6,14 76 -7,58 -4,30 158 -9,77 -6,58 77 -7,25 -4,17 159 -9,58 -6,81 210,79 66,56 261,82 1888,23 78 -7,03 -4,16 160 -9,63 -6,82 79 -6,93 -4,25 161 -9,85 -180,70 -522,39 -693,24 -864,09 -1034,94 80 -7,71 -4,63 162 -9,55 -181,09 -524,18 -695,73 -867,27 -1038,82 81 -7,56 -4,94 163 -9,30 -181,61 -526,22 -698,52 -870,82 -1043,13 82 -7,80 -5,31 164 -9,15 -5,93 165 -8,64 -6,10 203,84 8,93 177,06 1938,70 166 -8,28 -6,17 167 -8,85 -5,97 168 -8,44 -5,63 202,97 -29,29 150,66 1895,79 169 -7,94 -5,58 170 -8,35 -5,65 171 -8,34 -5,68 206,38 -63,06 88,86 1883,20 172 -8,17 -5,81

Table B-28: Diplodon longulus (Mazán, 99RKMz2), continuation (2).

13 18 13 18 13 18 13 18 # δ C δ Omm# δ C δ Omm # δ C δ Omm # δ C δ Omm 1 -8,31 -8,76 80,00 86 -11,69 -8,03 73,18 171 -9,75 -7,95 56,94 256 -8,18 -5,96 43,95 2 -8,39 -8,34 79,92 87 -11,83 -7,94 73,12 172 -9,43 -7,61 56,76 257 -8,24 -5,89 43,86 3 -8,04 -7,91 79,84 88 -11,53 -8,20 72,60 173 -9,41 -8,37 56,53 258 -8,36 -6,15 43,76 4 -8,14 -7,46 79,77 89 -11,42 -8,28 72,24 174 -9,08 -8,09 56,37 259 -8,25 -6,05 43,64 5 -8,08 -7,04 79,69 90 -11,07 -8,11 71,95 175 -8,79 -8,40 56,13 260 -8,05 -5,97 43,57 6 -7,59 -6,14 79,61 91 -10,46 -8,21 71,79 176 -9,17 -8,49 55,91 261 -7,79 -5,66 43,46 7 -7,26 -5,09 79,53 92 -10,47 -7,53 71,58 177 -9,41 -8,60 55,78 262 -7,41 -5,52 43,36 8 -7,19 -4,57 79,46 93 -9,82 -6,46 71,45 178 -9,37 -8,68 55,68 263 -7,08 -5,18 43,23 9 -6,35 -3,86 79,38 94 -9,24 -4,92 71,34 179 -9,22 -8,65 55,59 264 -6,97 -5,00 43,12 10 -6,07 -3,73 79,30 95 -8,32 -3,87 71,27 180 -9,23 -8,68 55,47 265 -7,17 -4,89 43,01

Table B-29: Diplodon longulus (Mazán, 99RKMz3). 134 Appendix B Ð Miocene shells: isotope data and minor elements

13 18 13 18 13 18 13 18 # δ C δ Omm # δ C δ Omm # δ C δ Omm # δ C δ Omm 11 -6,40 -4,93 79,22 96 -7,91 -3,86 71,12 181 -9,25 -8,64 55,23 266 -6,51 -4,49 42,90 12 -7,11 -5,52 79,15 97 -7,86 -4,42 71,07 182 -9,38 -8,56 55,04 267 -7,22 -5,52 42,84 13 -7,15 -5,55 79,07 98 -7,83 -4,25 70,96 183 -9,97 -8,76 54,76 268 -6,87 -4,61 42,78 14 -8,05 -6,45 78,99 99 -8,07 -4,44 70,84 184 -10,20 -8,53 54,52 269 -6,81 -4,55 42,67 15 -8,45 -6,73 78,91 100 -8,69 -4,60 70,69 185 -10,58 -8,82 54,27 270 -6,83 -4,55 42,56 16 -8,76 -6,87 78,83 101 -8,93 -5,15 70,24 186 -10,56 -8,87 54,10 271 -6,77 -4,52 42,44 17 -9,08 -7,16 78,76 102 -10,00 -6,03 69,90 187 -10,68 -9,01 53,95 272 -6,76 -4,32 42,33 18 -8,97 -7,07 78,68 103 -10,27 -6,43 69,53 188 -11,75 -8,73 53,85 273 -6,83 -4,31 42,22 19 -8,79 -6,61 78,60 104 -10,49 -7,36 69,22 189 -11,60 -8,93 53,77 274 -6,90 -4,36 42,10 20 -8,82 -6,43 78,52 105 -10,46 -7,53 68,98 190 -11,27 -8,87 53,70 275 -7,01 -4,56 41,98 21 -8,95 -5,86 78,45 106 -10,89 -7,63 68,79 191 -12,28 -8,84 53,55 276 -7,03 -4,41 41,88 22 -9,36 -6,08 78,37 107 -11,57 -8,09 68,54 192 -12,25 -10,51 53,42 277 -7,37 -4,51 41,77 23 -9,59 -5,57 78,29 108 -11,08 -7,95 68,32 193 -12,31 -8,73 53,28 278 -7,32 -4,44 41,65 24 -9,45 -5,41 78,21 109 -10,62 -8,05 68,18 194 -11,76 -8,44 53,11 279 -7,15 -4,20 41,53 25 -7,22 -3,91 78,14 110 -9,93 -8,02 67,94 195 -11,08 -8,41 52,91 280 -7,23 -4,27 41,42 26 -7,08 -3,95 78,06 111 -9,75 -7,59 67,75 196 -10,59 -8,28 52,70 281 -7,09 -4,26 41,29 27 -6,84 -4,15 77,98 112 -10,48 -7,22 67,63 197 -10,78 -8,42 52,53 282 -7,05 -4,26 41,18 28 -7,01 -3,54 77,90 113 -10,05 -6,80 67,32 198 -11,18 -8,30 52,41 283 -7,12 -4,29 41,08 29 -7,10 -4,99 77,82 114 -8,87 -6,17 67,03 199 -11,96 -8,38 52,30 284 -7,14 -5,40 41,00 30 -7,43 -5,74 77,71 115 -8,25 -5,71 66,72 200 -12,43 -8,43 52,18 285 -7,04 -4,40 40,81 31 -7,55 -5,86 77,59 116 -7,90 -5,41 66,41 201 -11,45 -8,10 51,97 286 -7,03 -4,21 40,61 32 -7,91 -6,07 77,54 117 -8,66 -6,05 66,04 202 -11,80 -8,34 51,72 287 -7,14 -4,42 40,42 33 -8,12 -6,64 77,49 118 -9,83 -7,26 65,82 203 -11,75 -8,38 51,59 288 -7,23 -4,38 40,23 34 -8,52 -6,71 77,46 119 -10,60 -8,29 65,46 204 -11,92 -8,34 51,42 289 -7,32 -4,43 40,04 35 -8,84 -6,82 77,41 120 -10,93 -8,77 65,09 205 -12,01 -8,47 51,22 290 -7,36 -4,21 39,84 36 -8,49 -6,72 77,33 121 -10,87 -8,79 64,95 206 -12,11 -8,52 51,08 291 -7,70 -4,59 39,65 37 -8,65 -6,50 77,26 122 -10,63 -9,23 64,77 207 -12,21 -8,40 50,88 292 -7,68 -4,38 39,46 38 -8,82 -5,93 77,21 123 -10,62 -8,77 64,68 208 -12,23 -8,42 50,63 293 -7,66 -4,35 39,26 39 -8,60 -5,91 77,15 124 -10,65 -9,02 64,57 209 -12,18 -8,32 50,65 294 -7,78 -4,34 39,07 40 -7,99 -4,78 77,08 125 -11,08 -9,13 64,38 210 -12,22 -8,28 50,51 295 -7,92 -4,41 38,88 41 -7,45 -4,24 77,05 126 -11,42 -9,50 64,27 211 -12,30 -8,36 50,35 296 -8,07 -4,37 38,68 42 -7,55 -4,31 77,02 127 -11,31 -9,24 64,14 212 -12,14 -8,29 50,18 297 -8,12 -4,90 38,49 43 -7,24 -3,77 76,99 128 -11,50 -9,54 64,03 213 -12,27 -8,16 50,06 298 -8,11 -4,47 38,30 44 -6,68 -3,21 76,98 129 -11,35 -9,77 63,90 214 -12,27 -7,97 49,93 299 -7,97 -4,30 38,11 45 -6,06 -2,53 76,96 130 -11,02 -9,87 63,72 215 -11,97 -7,80 49,71 300 -7,87 -4,48 37,91 46 -6,12 -2,19 76,91 131 -11,12 -9,87 63,24 216 -11,67 -7,85 49,60 301 -7,70 -4,35 37,72 47 -6,60 -1,97 76,85 132 -11,27 -9,95 62,96 217 -11,27 -7,68 49,47 302 -7,58 -4,53 37,53 48 -7,03 -1,91 76,82 133 -10,97 -9,68 62,76 218 -10,92 -7,66 49,24 303 -7,65 -4,22 37,33 49 -7,20 -2,59 76,79 134 -11,20 -9,48 62,49 219 -10,92 -7,58 49,08 304 -7,71 -4,54 37,14 50 -7,01 -1,94 76,75 135 -11,32 -9,44 62,32 220 -10,85 -7,71 48,95 305 -7,70 -4,44 36,95 51 -7,00 -1,99 76,71 136 -11,96 -9,33 62,19 221 -10,89 -7,47 48,79 306 -7,77 -4,48 36,75 52 -6,75 -2,15 76,64 137 -12,41 -9,53 62,03 222 -10,78 -8,63 48,63 307 -8,05 -4,39 36,56 53 -6,41 -3,53 76,58 138 -12,08 -9,33 61,86 223 -10,31 -7,08 48,51 308 -8,97 -4,69 36,37 54 -6,76 -2,66 76,53 139 -11,24 -9,01 61,69 224 -10,56 -7,29 48,32 309 -7,13 -4,65 36,18 55 -6,75 -3,66 76,49 140 -10,34 -8,59 61,46 225 -10,14 -7,08 48,10 310 -7,30 -4,77 35,98 56 -7,50 -4,56 76,42 141 -9,57 -8,08 61,30 226 -10,08 -7,42 47,96 311 -8,17 -4,83 35,79 57 -7,83 -5,03 76,39 142 -9,50 -7,80 61,15 227 -9,59 -7,01 47,75 312 -8,14 -5,04 35,60 58 -8,10 -5,22 76,38 143 -9,81 -6,69 60,95 228 -9,52 -7,36 47,57 313 -7,35 -5,49 35,40 59 -8,81 -5,77 76,34 144 -10,05 -5,85 60,74 229 -9,08 -6,93 47,35 314 -7,80 -5,64 35,21 60 -9,09 -6,14 76,34 145 -10,84 -5,94 60,51 230 -8,63 -6,69 47,21 315 -7,99 -5,91 35,02 61 -8,95 -6,17 76,31 146 -10,43 -5,93 60,37 231 -8,08 -6,24 47,03 316 -7,60 -6,20 34,82 62 -8,92 -6,50 76,30 147 -9,60 -5,76 60,30 232 -7,77 -6,17 46,87 317 -7,45 -6,36 34,63 63 -8,32 -5,91 76,29 148 -9,79 -4,75 60,18 233 -7,32 -5,86 46,74 318 -7,49 -6,39 34,44 64 -8,99 -6,18 76,13 149 -10,60 -5,70 60,06 234 -7,23 -5,92 46,62 319 -7,58 -6,55 34,25 65 -9,15 -6,13 76,19 150 -9,68 -5,76 59,88 235 -7,22 -5,71 46,49 320 -7,64 -6,64 34,05 66 -8,93 -5,83 76,04 151 -9,46 -5,78 59,76 236 -7,43 -6,01 46,34 321 -7,70 -6,47 33,86 67 -8,37 -5,24 75,94 152 -9,07 -5,86 59,64 237 -7,82 -6,27 46,27 322 -7,80 -6,44 33,67 68 -7,94 -5,14 75,88 153 -8,82 -6,00 59,50 238 -8,10 -6,53 46,18 323 -8,19 -6,82 33,47 69 -8,25 -5,21 75,84 154 -8,85 -6,10 59,33 239 -8,60 -7,16 46,00 324 -8,57 -7,07 33,28 70 -7,40 -4,57 75,71 155 -8,90 -6,20 59,21 240 -8,47 -6,74 45,93 325 -8,34 -6,62 33,09 71 -7,33 -4,93 75,34 156 -8,90 -6,13 59,11 241 -8,61 -6,81 45,76 326 -8,57 -6,90 32,89 72 -8,65 -5,25 75,23 157 -9,05 -6,15 58,97 242 -8,65 -6,83 45,62 327 -8,83 -7,03 32,70 73 -7,26 -4,72 74,77 158 -9,45 -6,11 58,86 243 -8,66 -6,58 45,48 328 -9,41 -7,43 32,51 74 -9,14 -6,58 74,70 159 -10,49 -6,38 58,75 244 -8,83 -6,88 45,34 329 -9,22 -7,24 32,32 75 -7,62 -4,65 74,63 160 -10,58 -6,12 58,65 245 -8,92 -6,85 45,21 330 -9,82 -7,58 32,12 76 -7,97 -5,66 74,61 161 -10,77 -6,42 58,56 246 -8,92 -6,81 45,14 331 -9,85 -7,64 31,93 77 -9,41 -7,00 74,55 162 -10,44 -6,28 58,45 247 -8,74 -6,60 45,03 332 -9,97 -7,93 31,74 78 -9,01 -7,16 74,47 163 -10,90 -6,50 58,35 248 -8,83 -6,74 44,90 333 -9,71 -8,29 31,54 79 -9,20 -7,33 74,34 164 -10,42 -6,68 58,18 249 -8,14 -5,91 44,75 334 -9,68 -8,67 31,35 80 -9,11 -7,34 74,22 165 -9,69 -6,77 58,11 250 -8,03 -5,86 44,61 335 -10,35 -8,73 31,16 81 -9,27 -7,36 73,99 166 -9,79 -7,24 57,95 251 -7,78 -5,54 44,49 336 -10,62 -8,54 30,96 82 -9,51 -7,19 73,93 167 -9,71 -7,28 57,84 252 -7,73 -5,61 44,36 337 -10,98 -8,80 30,77 83 -11,08 -7,83 73,75 168 -9,57 -7,50 57,70 253 -7,84 -5,66 44,27 338 -10,97 -8,96 30,58 84 -11,24 -7,94 73,54 169 -9,87 -7,69 57,48 254 -7,89 -5,74 44,15 339 -10,55 -9,29 30,39 85 -11,40 -8,05 73,34 170 -9,66 -7,60 57,18 255 -8,00 -5,83 44,05 340 -10,66 -8,62 30,19 341 -10,82 -8,25 30,00

Table B-30: Diplodon longulus (Mazán, 99RKMz3, continuation).