Age of depositional and weathering events in Central Amazonia

Age of depositional and weathering events in Central Amazonia

Quaternary Science Reviews 170 (2017) 82e97 Contents lists available at ScienceDirect Quaternary Science Reviews journal homepage: www.elsevier.com/...

3MB Sizes 0 Downloads 89 Views

Quaternary Science Reviews 170 (2017) 82e97

Contents lists available at ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Age of depositional and weathering events in Central Amazonia Lucy Gomes Sant'Anna a, b, *, Emílio Alberto do Amaral Soares c, Claudio Riccomini b, d, Sonia Hatsue Tatumi e, Marcio Yee e ~o Paulo, Av. Arlindo Bettio 1000, CEP 03828-000, Sa ~o Paulo, SP, Brazil School of Arts, Sciences and Humanities, University of Sa ~o Paulo, Av. Prof. Luciano Gualberto 1289, CEP 05508-010, Sa ~o Paulo, SP, Brazil Institute of Energy and Environment, University of Sa c vio Jorda ~o Ramos, 6200, CEP 69077-000, Institute of Exact Sciences, Department of Geosciences, Federal University of Amazonas, Av. Gen. Rodrigo Octa Manaus, AM, Brazil d ~o Paulo, Rua do Lago 562, CEP 05508-080, Sa ~o Paulo, SP, Brazil Institute of Geosciences, University of Sa e ~o Paulo, Campus Baixada Santista, Avenida Saldanha da Gama 89, Ponta da Praia, CEP 11030-400, Sa ~o Paulo, SP, Brazil Federal University of Sa a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 November 2016 Received in revised form 14 June 2017 Accepted 15 June 2017

In the last three decades, several studies have been devoted to understanding the role of Late Pleistocene eHolocene climate changes in the Amazonia lowlands environment. However, most of these studies used data obtained from sedimentary deposits (lakes, swamps, and colluvium) located away from the central plain or on the edges of the Amazonia region. This article integrates optically stimulated luminescence and accelerated mass spectrometry 14C ages with sedimentological and geomorphological data obtained during this study or compiled from the ~ es-Amazon River. literature for fluvial and lacustrine deposits of the central alluvial plain of the Solimo The age data allow us to present a chronological framework for the Late PleistoceneeHolocene deposits and conclude that (i) the dryness of the LGM in central Amazonia lowlands is recorded by the formation of fluvial terraces and their weathering to pedogenic hematite between 25.3 ka and 17.7 ka; (ii) floodplain deposition was contemporaneous with terrace weathering and occurred in a context of decreased water volume in fluvial channels, lowering of river base level and sea level, and isostatic ~ es-Amazon River rebound of the continent; and (iii) lateral and mid-channel fluvial bars in the Solimo have a minimum age of 11.5 ± 1.5 ka, and their deposition responded to increased precipitation at the beginning of the Holocene. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Amazonia Alluvial plain OSL dating Iron oxide Pleistocene LGM

1. Introduction Investigations of the climate of the Amazonia region during the Quaternary have primarily focused on the Late PleistoceneeHolocene boundary, when relevant worldwide changes occurred in the last glacial-interglacial cycle. These studies have mainly used data from lakes, swamps, and colluvium located far from the central plain (Fig. 1). Research on the central plain has been focused on Holocene floodplain lakes in the lower course of ~ es-Amazon River (e.g. Moreira et al., 2014). Because few the Solimo studies have incorporated data from fluvial deposits of the main rivers of the Amazonia lowlands, there is a lack of chronological data from fluvial deposits that assist in the correlation between

* Corresponding author. School of Arts, Sciences and Humanities, University of ~o Paulo, SP, Brazil. S~ ao Paulo, Av. Arlindo Bettio 1000, CEP 03828-000, Sa E-mail address: [email protected] (L.G. Sant'Anna). http://dx.doi.org/10.1016/j.quascirev.2017.06.015 0277-3791/© 2017 Elsevier Ltd. All rights reserved.

fluvial dynamics in the lowlands of Central Amazonia and climate change during the LGM, as has already been highlighted by several authors (Heine, 2000; Turcq et al., 2002; Vimeux et al., 2009; Irion and Kalliola, 2010; Baker and Fritz, 2015). Several Quaternary depositional units have been identified in ~es-Amazon River in the alluvial plain and channel of the Solimo Central Amazonia (Irion et al., 1995; Mertes et al., 1996; Latrubesse and Franzinelli, 2002; Rossetti et al., 2005; Irion et al., 2011; Rozo et al., 2012). However, the set of ages available for these units has led the authors to establish an almost exclusively Holocene sedimentary history. Recently, Soares et al. (2010) and Gonçalves et al. (2016) dated Pleistocene to Holocene sediments deposited by or ~ es-Amazon River in Central under the influence of the Solimo Amazonia, but the authors focused their approach on morphostratigraphic units and fluvial terraces and did not correlate them with previous important sedimentological studies (e.g., Latrubesse and Franzinelli, 2002). Thus, the information is still disjointed, with no clear relationship between geomorphological, sedimentological,

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

83

 Rise, Fig. 1. Location of the study area and sites mentioned in the text (locations given according to original references). 1. Amazon fan, ODP site 942 (Maslin et al., 2011). 2. Ceara ODP sites 925 (2a) and 926 (2b) (Harris and Mix, 1999). 3. Pata Lake (Colinvaux et al., 1996a; Santos et al., 2001; Barbosa et al., 2004; Cordeiro et al., 2011; D'Apolito et al., 2013). 4. Dune system (Carneiro Filho et al., 2002). 5. Coari Lake (Horbe et al., 2011). 6. Calado Lake (Behling et al., 2001). 7. Profile 1 (Franzinelli, 2011). 8. Site NSK-50 (Ferreira, 2013). 9. Caraj as (Sifeddine et al., 2001). 10. Carajas (Absy et al., 1991). 11. Carajas (Hermanowski et al., 2012a, b). 12. Interfluve of Madeira and Purus Rivers (Bertani et al., 2015). 13. Katira  Creek (Van der Hammen and Absy, 1994). 14. Aguas Emendadas (Ledru et al., 1998). 15. Salitre (Ledru et al., 1998).

and chronological data, which has prevented a clear understanding of paleoclimate records in the Quaternary plain. In this paper, we present new optically stimulated luminescence (OSL) and accelerated mass spectrometry (AMS) 14C ages for surface and subsurface samples of several alluvial plain and channel de~ es-Amazon River in its stretch between the posits of the Solimo confluences with Purus (west) and Negro (east) Rivers. We have integrated the new ages with previously published data (dating and sedimentology of samples whose locations are reported) in order to construct the most complete chronological framework currently possible for the Quaternary fluvial sediments in Central Amazonia. Instead of using a geomorphological approach that highlights the river terraces, we focus on the fluvial sedimentary deposits and their evolution. The focus on depositional units is justified by the existence of clear interactions between current depositional processes and floodplains already formed on the banks of the Sol~es-Amazon River, such as the current lacustrine deltas in lakes imo located on the floodplain (e.g., Mertes et al., 1996). In addition, it is likely that the PleistoceneeHolocene fluvial terraces defined by Gonçalves et al. (2016) include deposits from different sedimentary environments not yet fully identified due to the extensive vegetation cover and difficult access. Based on the collated chronological framework and sedimentological data, we discuss the (i) relationship among fluvial dynamics, generation of river terraces, and weathering; (ii) age interval for floodplain deposition and formation of ria lakes in Central Amazonia; (iii) correlation with other regional data; and (iv) possible climatic and sea-level influences on fluvial dynamics during the last glacial maximum (LGM).

2. Distribution of depositional units Fig. 2 shows the distribution of depositional units of the Sol~es-Amazon River in the area between its confluence with the imo Purus and Negro Rivers, as well as the alluvial deposits of the Negro River in its lower stretch. This figure incorporates depositional units previously described by other authors and own geologic data from fieldwork focused on facies analyses. Previously published OSL and radiocarbon ages for these deposits (for which geographical coordinates of dated samples are available) are integrated into our description of the units. ~es-Amazon River 2.1. Solimo Based on geomorphological and sedimentological properties, ~es-Amazon River were grouped in an the deposits of the Solimo alluvial plain, occurring in both margins of the river and where two main depositional units have been identified, and in an alluvial channel, which houses the several units identified in the current river channel. The substrate of the Quaternary depositional units is formed by ~o Formation (Caputo et al., 1972) the Upper Cretaceous Alter do Cha and the Miocene Novo Remanso Formation (Dino et al., 2012; Soares et al., 2015). 2.1.1. Alluvial plain Fluvial and lacustrine depositional units have been identified in ~es-Amazon River. the alluvial plain at both margins of the Solimo The oldest fluvial depositional unit in the alluvial plain currently

84

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

~es-Amazon River with the Purus (west) and Negro (east) Rivers (after Latrubesse and Fig. 2. Distribution of Quaternary alluvial deposits in the region of confluence of the Solimo Franzinelli, 2002; Soares, 2007; Soares et al., 2010; Reis et al., 2006; Rozo et al., 2012). The vast blue area corresponds to the current water bodies (mainly rivers and lakes) and their bottom sediments. Locations of dated deposits and relations among depositional units are indicated. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

~esoccurs as terraces that are not annually flooded by Solimo Amazon water, which are locally named “terra firme” (e.g., Irion et al., 2011), and correspond to the upper fluvial terrace of Soares et al. (2010) and Gonçalves et al. (2016). The unit has been attributed to the Pleistocene (e.g., Irion et al., 1995; Latrubesse and Franzinelli, 2002). Soares et al. (2010) dated two fluvial terraces of this unit and obtained OSL ages of 65.2 ± 8.8 ka (site NS-198) and 60.7 ± 6.6 ka (site NS-230) at the right margins of Manacapuru Lake ~ es-Amazon River, respectively (Fig. 2). Gonçalves and the Solimo ~eset al. (2016) studied this unit on the right bank of the Solimo Amazon River and expanded the OSL ages of this unit to between 240 ± 16 ka and 51 ± 5 ka. The fluvial deposits have been described as formed by a scroll-dominated plain, which presents a welldeveloped scroll morphology and is composed of silt-clay sediments (Latrubesse and Franzinelli, 2002). Scroll bar is a fluvial deposit formed over a point bar platform adjacent to the convex bank of a meandering river, and its deposition initiates with a flood episode (Nanson, 1980). The formation of a scroll bar is dependent of concave bank erosion of a meander bend, as the widened channel promotes decrease of water flow velocity, which invades the platform by discharging sediments in a ridge construction (van de Lageweg et al., 2014). Several flood episodes lead to the deposition of scrolls quasi-parallel to the curved channel and to the generation of a typical ridge-and-swale topography, which records the pulsed lateral river migration. A general grain size decreasing upward has been usually recognized from point bars, deposited by traction load at the base, to scroll bars composed of fine sand and silt carried by fine traction and suspended load and growing up and laterally by almost exclusively suspended load (Nanson, 1980; van de Lageweg et al., 2014). The Pleistocene unit has also been described as formed by point bars built by alternating sand and mud layers that produce inclined heterolithic stratification (Soares, 2007; Gonçalves et al., 2016). In scroll and point bars, sand and mud layers are weathered and display yellowish to reddish and whitish to reddish colors, respectively (Soares, 2007). In the alluvial plain, the Pleistocene scroll bar deposits border a floodplain unit, informally named an “impeded” floodplain by Latrubesse and Franzinelli (2002), which defines a region of very ~ es-Amazon River and flat relief along both margins of the Solimo corresponds largely to the intermediate terrace of Soares et al. (2010) and Gonçalves et al. (2016). Latrubesse and Franzinelli (2002) described mainly vertically accreted grey to grey-green muddy sediments, frequently mottled with orange or yellow colors, bioturbated and locally containing trunk remains dated by radiocarbon to between 493 and 288 cal yr BP (310 ± 50 14C yr BP) and 1057-798 cal yr BP (1030 ± 50 14C yr BP). OSL ages of between 30.9 ± 8 ka and 19.1 ± 6.3 ka were obtained by Gonçalves et al. (2016) in deposits composed of greyish to brownish silty mud layers intercalated with rare meter thick greyish sand layers. Mertes et al. (1996) defined this unit as a mixed floodplain composed of scroll bars, overbank deposits, floodplain channels, and lakes. The floodplain may be partially inundated by annual floods and has been locally reworked or has received local deposition. Disruption of levees and coalescence of splay deltas have led to the formation of shallow round or irregularly shaped lakes over ~esthe floodplain, whose sediments are supplied by the Solimo Amazon annual floods (Latrubesse and Franzinelli, 2002). These lakes have been named “v arzea” lakes (Irion et al., 2009), dishshaped lakes (Hess et al., 2003), and floodplain lakes (Rozo et al., 2012; Moreira et al., 2013), which is the name used in this paper. Good examples of such lakes in the study area are Cabaliana and ~ es-Amazon River (Hess Padre Lakes in the left margin of the Solimo , Grande, Pesqueiro, Miraja , and Janauaca et al., 2003), and also Inaja ~ es-Amazon River. Lakes, all located in the right margin of the Solimo

85

Some lakes with high sediment loads, such as the central lake of Marchantaria Island, have varve deposits that may reach up to 6 m thick and represent 2000 years of deposition (Irion et al., 2009). Ria lakes, which also occur over the alluvial plain in Central Amazonia, are typically elongated lakes, many with a length of up to 1 km and a depth of just a few meters (usually less than 5 m) (Gourou, 1950; Müller et al., 1995; Irion et al., 2011). Representative examples in the study area include Manacapuru, Miriti, and Calado ~es-Amazon River. NeverLakes, at the left margin of the Solimo theless, the Preto da Eva River, about 75 km east of the city of Manaus, is the best exemplar of a ria in Central Amazonia (Müller et al., 1995; Irion et al., 2009; Rozo et al., 2012). Behling et al. (2001) studied the sedimentary record (1190 cm long) of Calado Lake, located to the east of Manacapuru and Cabaliana Lakes. Wood material at the core base provided an age of 9472-9141 cal yr BP (AMS age of 8330 ± 50 yr BP), and the material at the top has a modern age. The lake sedimentation was initiated by increasing water levels after 7700 14C yr BP (early Holocene) and was preceded by fluvial tributary deposition in the Late Pleistocene. The transition from a fluviatile to lacustrine environment was gradual, as indicated by the alternation of sand (fluvial) and dark detrital mud (lake) deposits. The lake is annually controlled by the water level of ~es-Amazon River and still has active sedimentation. the Solimo 2.1.2. Alluvial channel The depositional units of the inner alluvial channel are point and scroll bars, mid-channel longitudinal bars, levees, and crevasse splay deposits (Latrubesse and Franzinelli, 2002; Hess et al., 2003; Mertes and Dunne, 2008; Franzinelli, 2011; Rozo et al., 2012, and our own field data). Two deposits located along the right margin of ~ es-Amazon River, recorded as the lower fluvial terrace the Solimo level, yielded OSL ages of 34.5 ± 4.4 ka (site NS-238) and 7.5 ± 0.9 ka (site NS-242) (Soares et al., 2010). On the western side of Careiro Island, a scroll bar deposit furnished an OSL age of 3.4 ± 0.6 ka (Rozo et al., 2012; sample D1). Gonçalves et al. (2016) determined an OSL age range of between 18.3 ± 4 ka and 2 ± 1 ka for the deposits of the alluvial channel that were grouped at the lower terrace level and described as formed by intercalations of meter thick dark-brown silty layers and decimeter thick light-grey sandy layers.  do Ariaú River 2.2. Parana  do Ariaú Graben (GPA) was formed by neotectonic The Parana ~ es-Amazon alluvial plain (Soares and activity in the Solimo Riccomini, 2003; Soares 2007). Two fluvial terrace levels (upper and lower) occur in the GPA, and both are floodplain deposits composed of suspension clay interlayered with point bar sand. The GPA deposits were formed by the secondary meandering system of  do Ariaú River, developed under the influence of the the Parana ~ es-Amazon River (Soares et al., 2001 and Soares et al., 2007). Solimo Sediments lodged in the GPA yielded OSL ages of 65 ± 2.4 ka (upper terrace, site NS-28A) and 13e8.9 ka (lower terrace, site NS-143), which are very similar to the OSL ages obtained for the terraces ~es alluvial plain and channel (Soares et al., 2010). in the Solimo Soares et al. (2010) interpreted the age variation from 65.2 to 7.5 ka as recording lateral accretion events controlled by the channel ~ es-Amazon River, which is also observed in migration of the Solimo the Paran a do Ariaú River deposits. 2.3. Negro River The main depositional feature of the lower course of the Negro River is the Anavilhanas Archipelago, with more than 400 islands and a neotectonic origin (Franzinelli and Igreja, 2002; AlmeidaFilho and Miranda, 2007). Islands are formed by upstream large

86

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

channel bars (up to 420 km2) that grade downstream to very long and narrow deltaic bars (one of them 13 km2). Bars are composed of yellow-gray clay and silt, and very rarely contain finely laminated, reddish mottled fine sand (Franzinelli and Igreja, 2002). No chronological information is available for the Anavilhanas Archipelago. The Cachoeira do Castanho (GCC) and Cacau do Pirera (GCP) Grabens, located in the Negro River alluvial plain, have also a neotectonic origin and serve as restricted flooded zones with predominant suspension sedimentation (Soares and Riccomini, 2003; Soares, 2007). In the GCC, a deposit dated by OSL yielded an age of 44.7 ± 1.8 ka (site NS-74B) and was interpreted as marking the beginning of sedimentation in the Negro River (Soares et al., 2010). 3. Obtaining new OSL and AMS

14

C ages

3.1. Sampling Six samples for OSL dating were collected from outcrops in the ~ es-Amazon River (Table 1). alluvial plain and channel of the Solimo Four of them are from the Pleistocene scroll-dominated plain, two being from the left margin (sites PD-109 and PD-111) and two from the right margin (sites PD-16 and PD-41). The fifth sample is from a floodplain deposit (site PD-03) located at the right margin of Sol~es-Amazon River, in front of its junction with the Negro River. imo The sixth sample was collected in a laterally accreted scroll bar deposit (site PD-48) of the alluvial channel, at the left margin of the ~es-Amazon River. All samples were collected in 2008 during Solimo the low water level season (November). Ten surface and subsurface samples were collected for AMS 14C ~es-Amazon River dating (Table 1): eight samples from the Solimo ~ es-Amazon and two from the Negro River. Samples from the Solimo

River include three from a surface point bar deposit (site NS-208) and one from a levee (site PD-119) at the right margin of the river, as well as four samples of lacustrine delta deposits (one from a marginal levee and three from a delta front). A marginal levee of a distributary channel in the delta plain was sampled in an outcrop (site PD-66) at the Cabaliana floodplain lake at the left margin of ~ es-Amazon River. Subsurface samples from delta front were Solimo obtained by manual boring in the Cabaliana floodplain lake (site PD-67C) and in the Caapiranga ria lake (site MO-01). The other two samples for AMS 14C dating were collected in channel bars at the right side of the Anavilhanas Archipelago, Lower Negro River. One sample is from site NS-205, located at the south end of a large channel bar, and the other is from site NS-206, situated at the northernmost area of a deltaic bar. Each sample for AMS 14C dating ~ es-Amazon and Negro Rivers was collected in 2003 from the Solimo or 2008 during the low water level season (November). 3.2. OSL dating The OSL dating was performed at the Laboratory of Glasses and Dating of the Faculty of Technology of S~ ao Paulo (FATECeSP), Brazil. Of the six samples sent for dating, only one (sample PD-16) had an insufficient amount of quartz. The other five samples were dated by the single-aliquot regenerative-dose (SAR) procedure (Murray and Wintle, 2003), using total regeneration of quartz crystals. The accumulated dose analysis was recorded with a TL/OSL Automated System, Model 1100-series Daybreak Nuclear Instruments Inc. The annual dose analysis was performed using a Canberra Inspector Portable Spectroscopy Workstation, equipped with an NaI(Tl) detector. All samples were collected with aluminum tubes 50 cm in

Table 1 Depositional settings and locations of dated samples.

River

Alluvial deposit

Depositional unit

Scroll-dominated plain

Floodplain

Alluvial plain

Floodplain lake

Solimões

Marginal levee of distributary channel in the delta plain

Delta front Ria lake Scroll bar Alluvial channel

Levee Point bar

Negro

Alluvial channel

Channel bar Deltaic bar

Location 61°23'45.65" W 2°57'1.70" S 61°23'45.65" W 2°57'1.70" S 60°21'34.21" W 3°50'2.68" S 60°17'43.21" W 3°33'51.70" S 59°51'17.25" W 3°16'32.92" S 60°55'37.30" W 3°21'46.14" S 60°50'32.37" W 3°18'53.77" S 60o15'36.89" W 3o24'30.89" S 60°05'50.08" W 3°18'12.86" S 60°39'8.23" W 3°20'41.93" S 60°35'37.07" W 3°20'33.74" S 60°46'23.90" W 2°44'55.0" S 60°47'2.79" W 2°44'27.21" S

Site

Sample

Outcrop (O) Manual boring (MB)

Dating method

PD-109

PD-109

O

OSL

PD-111

PD-111

O

OSL

PD-16

PD-16

O

OSL

PD-41

PD-41

O

OSL

PD-03

PD-03

O

OSL

PD-66

PD-66B

O

14C

PD-67C (93.0-95.0 cm) PD-67C (150.5-150.9 cm) MO-1G (31.0-43.0 cm)

MB

14C

MB

14C

MB

14C

PD-48

PD-48

O

OSL

PD-119

PD-119D

MB

14C

NS-208

NS-208A NS-208B NS-208C

O O O

14C 14C 14C

NS-205

NS-205A

O

14C

NS-206

NS-206A

O

14C

PD-67C MO-01

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

length and 5 cm in diameter, and were properly wrapped in the field in order to avoid any contact with natural light before being sent to the laboratory. 3.3. AMS

14

C dating

AMS 14C age determinations were performed by Beta Analytic Inc. Of the analyzed samples, eight provided ample carbon for accurate measurements, but two (PD-119D and NS-205A) were not suitable for AMS dating due to low amounts of carbon. Samples were fragments of wood, plant, and charred material, and were pretreated according to the standard acid/alkali/acid pretreatment protocols at Beta Analytic. The AMS 14C ages obtained in this study and conventional 14C ages compiled from the literature were changed to calendar year before present (cal yr BP, 2-sigma) using IntCal13 (Reimer et al., 2013; Stuiver et al., 2017). 4. Dating results ~es-Amazon River 4.1. Solimo 4.1.1. Scroll-dominated alluvial plain Dated deposits are located relatively far from the current ~es-Amazon River. Sites PD-109 and PD-111 channel of the Solimo are located about 65 km to the north of the channel, and site PD41 is located about 25 km to the south (Fig. 2). These deposits sustain rounded hills and are composed of alternating layers of mud and sand showing white (5Y 8/1) color with discontinuous and irregularly distributed red (2.5 YR 4/8) and yellow (10 YR 8/8) blends, resulting in a mottled sediment. Dated sand layers are dominantly red with little white portions, are medium to fine grained, contain a low amount of clay, and exhibit massive structure (Fig. 3). The results obtained for the three samples containing enough amount of quartz to be dated by OSL (SAR protocol) are listed in Table 2. Samples yielded ages varying between 17.7 ± 2 ka and 25.3 ± 3 ka. The two younger samples (17.7 ± 2 ka and 18.6 ± 2.1 ka) are from the left margin (sites PD-109 and PD-111). The older sample (25.3 ± 3 ka) is from the right margin (site PD-41). 4.1.2. Floodplain The floodplain deposit (site PD-03) dated by OSL (SAR protocol) is composed of unweathered dark greenish gray clay that yielded an age of 20.3 ± 2.95 ka (Fig. 2, Table 2). 4.1.3. Fluvial bars Site PD-48 (Fig. 2) is a laterally accreted scroll bar deposit of the ~es-Amazon alluvial channel located at the left margin of the Solimo River. The dated level is in an intermediate position of the profile and is composed of non-weathered dark gray sand that furnished an OSL (SAR protocol) age of 11.5 ± 1.5 ka (Table 2). The point bar deposit (site NS-208) at the right margin of the ~es-Amazon River shows large lateral extension (at least Solimo 217 m along the margin) cropping out during the low water level season. Silty clay layers with planar lamination occur from the base to the intermediate portion of the profile (Fig. 3). All layers exhibit green to greenish gray colors, indicating their non-weathered condition. Three black organic-rich lamina (up to 1 cm in thickness) containing plant remains are interlayered in the lower portion of the profile. Upwards laterally discontinuous very fine sand lamina (2e3 mm in thickness) exhibiting crossed lamination and a brown color are interlayered with clay layers. A laterally extensive and almost tabular (15e25 cm in thickness) sand bed showing convolute structure occurs at the middle portion, and

87

massive clay predominates toward the top of the profile. The three organic samples (A, B, and C; Fig. 3) dated by AMS 14C provided modern ages (Table 3), indicating that the organic material was still alive after 1950s CE. Sample NS-208A is at the base of the profile and yielded the lowest pMC (percent modern carbon) value, being followed by the stratigraphically intermediate sample, NS-208B, and the upper sample, NS-208C. Considering that the atmospheric pMC has been decreasing since 1965 (Hua et al., 2013) and the pMC of samples increases toward the top, it seems reasonable to assume that they deposited between 1950 and 1965.

4.1.4. Lacustrine delta deposits Two lacustrine delta deposits were sampled in the Cabaliana ~es-Amazon River floodplain lake at the left margin of the Solimo (Table 1, Fig. 2). One deposit corresponds to a marginal levee (site PD-66) of the Paran a do Piranha, a distributary channel in the delta plain that feeds the south border of the lake, and consists of 6-mthick greenish brown clay containing organic fragments close to the base (Fig. 3). Sample PD-66B (containing wood material) furnished an age of 425-0 cal yr BP (Table 3). The other deposit is the delta front in the central portion of the lake (site PD-67C) that is composed of mostly dark gray clay and fine sand layers (with heterolitic stratification) interlayered in the lower part of the package and yellowish brown silty clay layers predominating towards the top (Fig. 3). Samples PD-67C (93e95 cm) and PD-67C (150.5e150.9 cm), both containing wood material from dark gray clay, provided ages of 618-342 cal yr BP and 759-656 cal yr BP, respectively (Table 3). The delta front in the eastern part of the Caapiranga ria lake (site ~es-Amazon River MO-01), located at the right margin of the Solimo (Fig. 2), contains black organic-rich lamina interlayered with greenish brown clay at the intermediate portion of the bored package (Fig. 3). The upper lamina (sample MO-1G) yielded an age of 898-676 cal yr BP (Table 3).

4.2. Mid-channel island of the Negro River The deltaic bar deposit (site NS-206; Figs. 2 and 3) sampled at the right side of the Anavilhanas Archipelago of the Lower Negro River is characterized by a 5-m-thick yellowish gray clay package capped by a 0.2-m-thick tabular black organic layer. An age of 913707 cal yr BP was obtained by AMS 14C dating of the organic layer, which contains charred material (Table 3). Vegetated soil at the top of outcrop stabilizes the mid-channel island.

5. Data integration into a chronological framework The new OSL and AMS 14C ages obtained in this study were integrated with previously published dating results (Table 4) for which geographic coordinates of analyzed samples were possible to obtain. The locations of these samples are included on the maps shown in Figs. 1 and 2. The integration of all these data allowed the determination of a chronological framework for the depositional ~es-Amazon River in its and weathering events related to the Solimo stretch between the junctions of the Purus and Negro Rivers (Fig. 4). OSL SAR protocol ages are preferentially employed in the discussion of the chronological framework. OSL MAR (multiple aliquot regenerative-dose) ages are mentioned only in the absence of any other data, due to the procedure limitations related to the use of multiple aliquots and no corrections for sensitivity changes in the analyzed grains (Murray and Wintle, 2000; Hilgers et al., 2001).

88

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

Fig. 3. Columnar sections of deposits dated by OSL and AMS. Locations of profiles are indicated in Fig. 2. The depth at which each sample for dating was collected is indicated in the profiles.

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

89

Table 2 ~es-Amazon River. OSL ages by SAR protocol obtained for surface samples from alluvial deposits of the Solimo

Sample

Laboratory number

Th (ppm)

U (ppm)

K (%)

Floodplain

PD-109 PD-111 PD-16 PD-41 PD-03

2290 2291 2287 2288 2286

6,449 ± 0,232 7,330 ± 0,264 12,113 ± 0,436 9,791 ± 0,352

1,786 ±0,220 2,476 ± 0,211 3,355 ± 0,302 3,367 ± 0,136

0,931 ± 0,135 1,032 ± 0,150 1,618 ± 0,235 2,165 ± 0,314

Mean Age paleodose (years) (Gy) 2.100 ± 210 40,6 17.700 ± 2.000 2.400 ± 230 44,6 18.600 ± 2.100 sample with insufficient quartz crystals 3.600 ± 350 90,8 25.300 ± 3.000 4.000 ± 380 81,2 20.300 ± 2.950

Scroll bar

PD-48

2289

10,401 ± 0,374

3,261 ±0,064

1,699 ± 0,246

3.540 ± 300

Alluvial deposit

Depositional unit

Alluvial plain

Scrolldominated plain

Alluvial channel

Annual dose (μGy/ano)

44,2

11.500 ± 1.500

Table 3 ~ es-Amazon and Negro Rivers. AMS 14C ages obtained for surface and sub-surface samples from alluvial deposits of the Solimo

River

Alluvial deposit

Depositional unit

Floodplain lake Alluvial plain Solimões

Negro

Alluvial channel

Sample

Laboratory number

Material

Measured radiocarbon age (BP or pMC*)

Conventional radiocarbon age (BP or pMC*)

Calibrated age (yr BP)

PD-66

PD-66B

261896

wood

220 ± 40

220 ± 40

425-0

261898

wood

570 ± 40

470 ± 40

618-342

261897

wood

760 ± 40

760 ± 40

759-656

261889

wood

261890 261891 261892

plant plant plant

261894

charred material

Channel bar

NS-205

PD-67C (93.0-95.0 cm) PD-67C (150.5-150.9 cm) MO-1G PD-119D NS-208A NS-208B NS-208C NS-205A

Deltaic bar

NS-206

NS-206A

Ria lake

Alluvial channel

Marginal levee of distributary channel in the delta plain

Site

Delta front

PD-67C

Levee

MO-01 PD-119

Point bar

NS-208

5.1. Scroll-dominated alluvial plain 5.1.1. Deposition Several authors have attributed a Middle to Late Pleistocene age to the upper fluvial terrace level in the area, which crops out at the ~es-Amazon River (e.g., left and right margins of the Solimo Latrubesse and Franzinelli, 2002; Irion et al., 2011; their Fig. 2.7; and Irion and Kalliola, 2010; their Fig. 11.3). The deposition of a Late Pleistocene scroll-dominated plain (the oldest scroll-dominated plain identified by Latrubesse and Franzinelli, 2002) by the Sol~es-Amazon River was confirmed by the OSL (MAR protocol) imo ages of 65.2 ± 8.8 ka (site NS-198) and 60.7 ± 6.6 ka (site NS-230) published by Soares et al. (2010). Middle Pleistocene OSL (SAR protocol) ages were recently presented by Gonçalves et al. (2016) that expanded the age range to 240 ± 16 ka and 51 ± 5 ka. The upper fluvial terrace deposited by the Ariaú River in the Paran a do Ariaú Graben (GPA) furnished an OSL age (SAR protocol) of 65 ± 2.4 ka, which led Soares et al. (2010) to interpret both ~es-Amazon and Ariaú Rivers) as deposition (in the Solimo contemporaneous, which is supported here. 5.1.2. Weathering OSL dating of chemically weathered deposits yield stratigraphically inconsistent ages due to changes of both dose rate and equivalent dose (Jeong et al., 2007). The OSL ages reported in this paper for the weathered deposits of the Pleistocene scrolldominated plain range from 25.3 ka to 17.7 ka (Table 4) and are

860 ±40 830 ±40 canceled 127.4 ± 0.6* 124.9 ± 0.6* 129.7 ± 0.6* 126.8 ± 0.6* 126.9 ± 0.6* 128.0 ± 0.6* canceled 890 ± 40

880 ± 40

898-676

913-707

younger than the stratigraphic ages previously published for this unit (Soares et al., 2010; Gonçalves et al., 2016). Due to the deep weathering of deposits dated in this paper, the obtained OSL ages are interpreted as dating of the weathering event that affected the older fluvial unit in the study area. 5.2. Floodplain The new OSL age (SAR protocol) of 20.3 ka (Table 4) obtained for a non-weathered floodplain deposit (site PD-03) located at the ~es-Amazon River indicates that the right margin of the Solimo deposition of this unit goes back to the Late Pleistocene. This OSL age agrees with the OSL age interval of 30.9 ± 8 ka and 19.1 ± 6.3 ka furnished by Gonçalves et al. (2016) for their intermediate terrace, which is largely formed by the floodplain deposits. Therefore, the deposition of the fluvial sediments that define a very flat relief ~ es-Amazon River (the “impeded” along both margins of the Solimo floodplain of Latrubesse and Franzinelli, 2002) is not restricted to the Holocene as previously described (e.g., Irion et al., 1999; Latrubesse and Franzinelli, 2002) but goes back to the Late Pleistocene. The fluvial terrace dated to 34.5 ± 4.4 ka (site NS-238) by Soares ~ eset al. (2010) (Table 4, Fig. 2), at the right margin of the Solimo Amazon River was recorded by these authors to be the lower and younger terrace level in the area. However, in this paper, this deposit is attributed to the floodplain unit due to its very welldeveloped plan-parallel stratification. In addition, the age of

90

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

Table 4 Integration of OSL and 14C ages: (a) OSL and AMS 14C ages obtained during this study; (b) OSL ages by Soares et al. (2010); (c) 14C ages by Latrubesse and Franzinelli (2002); (d) OSL age by Rozo et al. (2012); (e) 14C ages by Behling et al. (2001), and (f) OSL age by Ferreira (2013).

River

Alluvial deposit

DeposiƟonal unit

Scroll-dominated plain

Floodplain lake Ria lake

Solimões

Alluvial plain

Floodplain Marginal levee of distributary channel

Delta front

Lake FluviaƟle to lacustrine

Alluvial channel Ariaú Negro

Age MAR (ka) (b) 65.2±8.8 60.7±6.6

14

C ConvenƟonal age (years)

Calibrated age (cal yr BP)

17.7±2.0 (a) 18.6±2.1 (a) 25.3±3.0 (a) 20.3±2.95 (a) 34.5±4.4 220±40BP (a)

425-0

PD-67C (93.0-95.0 cm) PD-67C (150.5-150.9 cm) MO-1G LC-B1 (162 cm)

470±40BP (a)

618-342

760±40BP (a)

759-656

(a)

830±40BP 280±30BP (e)

898-676 452-155

LC-B1 (1190 cm)

8330±50BP (e)

Deltaic bar

NS-206A

Point bar

Age SAR (ka)

PD-66B

Floodplain Floodplain

Scroll bar

Mid-channel island

Floodplain Midchannel islands

NS-198 NS-230 PD-109 PD-111 PD-41 PD-03 NS-238

NS-208A NS-208B NS-208C LF-05 LF-06 LF-08 LF-09 NS-268 NS-242 PD-48 D1 NSM-50k NS-28A NS-20A NS-143 NS-74B

Point bar

Floodplain

Sample

34.5 ± 4.4 ka (Soares et al., 2010) agrees with the maximum age of 38.9 ka for the floodplain presented by Gonçalves et al. (2016). Müller et al. (1995, their Fig. 3) presented a high resolution (3.5 kHz) sonic profile obtained in a cross-section transect close to the mouth of the Manacapuru ria lake in which three major acoustic units were recognized. The authors attributed the basal ~ es Formation, which is Miocene in age and oldest unit to the Solimo and occurs only to the west of the Purus Arch (Latrubesse et al., 2010; Gross et al., 2011). Thus, considering that Pleistocene deposits were already represented at the borders of Manacapuru Lake (e.g., Latrubesse and Franzinelli, 2002; Soares et al., 2010), it is here suggested that the basal unit of the seismic profile of Manacapuru Lake published by Müller et al. (1995) corresponds to the MiddleLate Pleistocene scroll-dominated plain (Fig. 5). According to the profile of Müller et al. (1995), the top of the Pleistocene scroll unit is eroded and unconformably overlain by a unit informally named “A.” Unit B is Holocene in age and overlays Unit A. Based on data gathered in this study, we suggest that Unit A may be correlated in age with the floodplain deposits and that Unit B is a sedimentary filling coeval to the deposits of Cabaliana Lake (Fig. 5). The OSL age (SAR protocol) of 13e8.9 ka yielded by the lower fluvial terrace of the GPA (site NS-143, Soares et al., 2010) suggests

9472-9141 (a)

1.6±0.1 (b) 11.5 ± 1.5 (a) 3.4 ± 0.6 (d) 11.9±1.18 (f) 65.0±2.4 (b) 65.0±2.4 (b) 8.9-13.0 (b) 44.7±1.8 (b)

124.9 ± 0.6 pMC 126.8±0.6 pMC (a) 128.0±0.6 pMC (a) 990±50BP (c) 310±50BP (c) 1000±40BP (c) 1030±50BP (c)

1045-785 493-288 975-796 1057-798

880±40BP (a)

913-707

1.3±0.4 7.5±0.9

55.3±16.3 45.6±14.0 7±2 48.6±12.3

that the floodplain deposition of the Ariaú River crossed the PleistoceneeHolocene boundary (Fig. 4). 5.3. Lake formation In this study, the dated sediments of the Cabaliana floodplain lake and Janauaca ria lake are located near the top of the lacustrine layers, and yielded ages of the last millennium for both the marginal levee and delta front (Table 4). Thus, in the study area, a lacustrine basal level age is currently available only for the Calado ria lake, dated to 9472-9141 cal yr BP (AMS age of 8330 ± 50 yr BP) by Behling et al. (2001). 5.4. Alluvial channel The laterally accreted scroll bar deposit (site PD-48) at the left ~es-Amazon River is a remarkable depositional margin of the Solimo feature that truncates the floodplain and records both the end of deposition on the alluvial plain and the beginning of sedimentation in the alluvial channel. The OSL age of 11.5 ± 1.5 ka obtained at the upstream part of this bar deposit indicates the PleistoceneeHolocene boundary as the minimum age for closure of the main

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

91

~es-Amazon River between the confluences with the Purus (west) and Negro (east) Rivers. Fig. 4. Chronological chart of Late PleistoceneeHolocene depositional units of the Solimo

Fig. 5. Stratigraphic interpretation of the schematic diagram showing the major acoustic units presented by Müller et al. (1995) in Manacapuru Lake.

depositional phase in the floodplain. This closure was due to the ~ es-Amazon River in its current channel entrenching of the Solimo between the confluences with the Purus and Negro Rivers, which rzea corridor” described by Irion et al. (2011). corresponds to the “va An equivalent OSL (SAR protocol) age was obtained by Ferreira ~es(2013) for a top layer of a mid-channel bar in the Solimo Amazon River just after its confluence with the Negro River. The agreement of both ages might suggest that the entrenchment of the ~es-Amazon River led to a concomitant sedimentation of bars Solimo at the margins and mid-channel.

Younger ages for this event, ranging from 9 ka to 7 ka, which were previously suggested (e.g., Irion and Kalliola, 2010; Ferreira, 2013) may be due to the location of sample collection. As the bars migrate eastward (e.g., Rozo et al., 2012), a trend toward younger ages would be expected. In Careiro Island, scroll bars dated by OSL yielded ages of 7.5 ± 0.85 ka and 3.4 ± 0.6 ka (Rozo et al., 2012). The two fluvial terraces dated to 7.5 ± 0.9 ka (site NS-242) and 1.6 ± 0.1 ka (site NS-268) by Soares et al. (2010) (Table 4, Fig. 2), ~es-Amazon River, both situated at the right margin of the Solimo were recorded by these authors as being located in the lower and

92

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

younger terrace level. Their ages show the deposition of the alluvial channel bars in the Holocene. Laminated silt-clay deposits located at the left margin of the ~es River and dated to the Holocene by Latrubesse and Solimo Franzinelli (2002, their profiles 6, 8, and 9) were attributed by these authors to their “impeded floodplain unit” (Latrubesse and Franzinelli, 2002; their Fig. 5) but are situated in their “channel dominated floodplain unit.” Thus, considering the similarities of these deposits with our NS-208 profile in terms of location, sedimentological characteristics, and age, we propose that these deposits are better grouped with our alluvial channel unit. 5.5. Negro River According to Soares et al. (2010), sedimentation in the Negro River started in the Late Pleistocene, dating back to 44.7 ka (Fig. 3), as restricted floodplain deposition in the Cachoeira do Castanho Graben (GCC). In this study, the only radiocarbon age obtained in the Anavilhanas Archipelago (site NS-206) places the deposition of deltaic bars in the Holocene, which agrees with previous interpretations (e.g., Franzinelli and Igreja, 2002; Latrubesse and Franzinelli, 2005). 6. Implications for environmental changes related to the LGM in Central Amazonia 6.1. LGM, aridity, and weathering The peak of the LGM was recently identified as occurring from 27.540 to 23.340 yr (Hughes and Gibbard, 2015), but its range is usually given as extending from ~23 to 19 ka (Shakun et al., 2012). The LGM is responsible for significant changes in the Late Pleistocene, including the lowering of sea level by about 120 m (Austermann et al., 2013; Lambeck et al., 2014), and reduction in arboreal vegetation (e.g., Anhuf et al., 2006) due to increased aridity and decreased temperature of landmasses. Many data sets and interpretations concerning the influence of this event in Amazonian lowlands have been generated from the study of sedimentary profiles recovered in small lakes and colluvial deposits, located far from the central plain or on the edges of the Amazonian region (Fig. 1). Several authors agree that there was a decrease in temperature of between 5  C and 6  C during the LGM in the Amazon (Colinvaux et al., 1996a; Bush et al., 2001). However, interpretations with regard to rainfall are still controversial (Sylvestre, 2009), with some suggesting maintenance of humidity levels during the reduction in temperature (e.g., Colinvaux et al., 1996a) and others positing reduction of both precipitation and temperature (e.g., Van der Hammen and Absy, 1994; Sifeddine et al., 2001) during the LGM. Among the lakes that have been studied, Lake Pata is the closest of the study area; because it presents sedimentological similarities with fluvial deposits in the study area, it is detailed here. Lake Pata (0160 N and 66 410 W) is located in the Hill of Six Lakes, northwest Amazonia region, at an altitude of 300 m in the headwaters of the Negro River. This lake is about 680 km northwest from site PD-111 (Fig. 1). It is a small lake (about 150,000 m2) subdivided into four lobes. The lake developed on a thick lateritic crust, has its water level controlled by precipitation, and receives low amounts of detrital sediment and nutrients from runoff of the surrounding area. The water column contains organic pigments provided by internal cycling of nutrients (Cordeiro et al., 2011). A distinctive feature in the sedimentary columns is a yellowish mineral-rich clastic layer in the middle portion located between organic-rich layers, as noted in the east (core LPTIII, Santos et al., 2001), center (core LPTIV, Barbosa et al., 2004), north-central

(cores Ia and Ib, Bush et al., 2004 and core II, D'Apolito et al., 2013), and north (core LPTV, Cordeiro et al., 2011) parts of the lake. In the western lobe, the clastic layer forms the basis of the sampled profile and is covered by organic sediment (core LPTVI, Barbosa et al., 2004). In the east (core LPTIII) and central (core LPTIV) parts of the lake, the clastic layer is about 20e25 cm thick, and has sand-sized grains, a dark brown (10 YR 3/3) color, and low values of water content and total organic carbon (TOC). This sandy layer was dated to ~18,000 yr BP and overlies an apparent depositional hiatus that occurred between 25,700 yr BP (25,690 ± 950 14 C yr BP; 31,285e27,900 cal yr BP in this paper) and 18,000 yr BP (17,410 ± 155 14C yr BP; 21,477e20,600 cal yr BP in this paper) (Santos et al., 2001). In the central-north part of the lake (core II, D'Apolito et al., 2013), the clastic layer is described as a yellowish nodular clay that contains sand and nodules, and is overlain by a 3-cm-thick greyish clay. In the clastic layer, the interval between 20 and 18 cal ka BP corresponds to the lowest lake level, with high concentration of spores, very low concentrations of pollen exhibiting high degradation, lack of algae, and low values of d15N. To the north (core LPTV), the mineral-rich layer is clayey with a dark yellowish brown (10 YR 4/6) color and exhibits low values of water content, TOC, and chlorophyll derivates, in addition to an increase of d13C values related to contributions from C4-type plants. Iron concentration shows a gradual increase starting at 33,800 cal yr BP, with maximum values between 28,000 cal yr BP and 20,000 cal yr BP due to the presence of gossan fragments provided by physical erosion of lateritic crusts surrounding the lake. Cordeiro et al. (2011), studying the LPTV core on the north edge of the lake, concluded that a widespread lowering of water level occurred between 26,300 cal yr BP and 15,300 cal yr BP. The origin of the sandy to clayey clastic sediment in Lake Pata has been attributed to a erosive event in the catchment area of the lake due to episodic and torrential rainfall, typical of a dry climate, when the lake would have reached its lowest water level during the LGM (Santos et al., 2001; Barbosa et al., 2004; Cordeiro et al., 2011; D'Apolito et al., 2013). As discussed by Cordeiro et al. (2011), the lithologic data and biogeochemical proxies obtained for the lacustrine sediments record alterations in the temperature and precipitation between 26,300 cal yr BP and 15,300 cal yr BP, when cold and dry weather prevailed, with the lowest precipitation levels occurring at 22,000 cal yr BP, refining the previous interpretations for the LGM in this area (Bush et al., 2004; Colinvaux et al., 1996a, 1996b, 2000). The fluvial deposits that make up the highest terraces of the study area consist of scroll bars (Latrubesse and Franzinelli, 2002; Soares et al., 2010). Scroll bar formation is related to critical flood events (Nanson, 1980) and channel expansion by erosion of the concave margin (van de Lageweg et al., 2014). These fluvial deposits have been assigned to the Pleistocene by several authors (e.g., Latrubesse and Franzinelli, 2002; Irion et al., 2011) and recently yielded OSL ages ranging from 240 ka to 51 ka (Gonçalves et al., 2016). In the present study, OSL ages between 25.3 ka and 17.7 ka were obtained for weathered deposits of this terrace level, which contain irregularly distributed pedogenic hematite overprinting both sand and mud layers, resulting in a typical mottled sediment. Temperature and water pressure are the two main factors controlling the formation of goethite and hematite. Low temperatures and plenty of water in the environment favor the formation of goethite, whereas raising the temperature and reducing water availability leads to the formation of hematite. Thus, for the same temperature, the formation of these two oxides is strongly controlled by the availability of water in the environment (Cornell

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

and Schwertmann, 1996), which has led several authors to use the ratio of pedogenic goethite to hematite as a proxy of paleoprecipitation. Different surface deposits have been addressed, including soils and paleosols (e.g., Balsam et al., 2011; Maxbauer et al., 2016), loess deposits associated with paleosols (Balsam et al., 2004), and marine sediments (e.g., Zhang et al., 2007; Harris and Mix, 1999). In such cases, the predominance of hematite is representative of a decrease in water availability in the environment and therefore a decrease in rainfall and a period of aridity. Using abundance of goethite (goethite/ (goethite þ hematite)) in marine sediments of the Amazon fan, Harris and Mix (1999) recognized the increase in rainfall in the deglacial stages of the glacial-interglacial cycles in the last 1 myr. These authors also recognized that, during this time, the maximum rainfall (indicated by the maximum concentrations of goethite) precedes sea-level highstands, and the minimum rainfall (minimum concentrations of goethite) predates sea-level lowstands. Harris and Mix (1999) included hematite in the analysis they did with goethite because they concluded that both minerals were eroded at the same time from soils in the Amazon basin in times of increased rainfall and deposited as terrigenous marine sediments in the Amazon fan. However, careful observation of the graphs showing variations of goethite and hematite concentrations presented by these authors (Harris and Mix, 1999; their Fig. 4) reveals a small offset between the peaks of maximum concentration of these minerals, with the peak of hematite being slightly older than that the goethite, which is clearly visible at ~880,000 yr, 790,000 yr, 695,000 yr, and 340,000 yr. These short time lags, ranging from 1300 to 6100 years, could suggest that increased rainfall recorded by the concentration of goethite is not exactly concurrent with the rainfall indicated by hematite. This time difference seems reasonable because, on the continent, (i) these minerals require different water availability in the environment for crystallization, and (ii) erosion and fluvial transport by the Amazon River continued even with the decrease in rainfall during glacial periods and low sea levels, as recorded by the presence of terrigenous sediments of these phases throughout the drill core analyzed by Harris and Mix (1999). Thus, it seems plausible to conclude that the weathering events responsible for hematite formation, as well as erosion and river transport of this oxide to the Atlantic Ocean, have occurred in lower rainfall conditions and before the genesis, erosion, and transport of goethite. The data showing variations of goethite and hematite concentrations presented by Harris and Mix (1999, their Fig. 4) are not clear for the last 40,000 yr, yet the relationship between goethite, rainfall, and rising sea level posited during earlier times were also extended to the LGM by these authors. The dominance of hematite in the deposits of MiddleeLate Pleistocene scroll-dominated plain, prominent in outcrops sampled ~es-Amazon River (sites PD-41, PDon both margins of the Solimo 109, and PD111, Fig. 2) suggests that the same paleoclimatic conditions interpreted for other continental and marine deposits may have affected the deposits of this fluvial unit. The dominance of hematite in these deposits dated from 25.3 ka to 17.7 ka points to a dry phase in the center of the Amazon plain, in the region of the ~es-Amazon and Negro Rivers, during the junction of the Solimo LGM. Therefore, the data gathered in this study fill a gap, already claimed by previous authors (e.g. Baker and Fritz, 2015), regarding paleoclimatic data associated with a chronological framework in the center of the Amazon plain, and suggest that a dry phase during the LGM may have occurred in this area. Regionally, the interpretation of aridity during the LGM (ca. 22 ka) in the center of the Amazonian lowlands presented in this paper from the study of fluvial deposits is in agreement with evidence previously obtained from other geological records, including:

93

(i) Black clay soil dated 18,500 ± 150 14C yr BP (22,682e21,941 cal yr BP), whose top is eroded and covered by a yellow-to red-colored colluvium in the Katira Creek valley (Van der Hammen and Absy, 1994). s area, (ii) A depositional hiatus in two small lakes in the Caraja recorded by a gap in the 14C ages, between 22 ka and 13 ka BP (31,897e21,189 cal yr BP and 15,185e14,178 cal yr BP), in an organic lacustrine layer that exhibits mud cracks in its top (Sifeddine et al., 2001). (iii) Brown silty clay and organic clay layers forming a banded unit deposited in shallow water in a small swamp in the Caraj as area and dated 24,390e23,650 cal yr BP and 11,708e11,259 cal yr BP; minerals (gibbsite, hematite and goethite) eroded from iron crust surrounding the water body compose the silty clay layer (Hermanowski et al., 2012a, 2012b). (iv) Top layers (depths between 250 and 300 cm) of a dune system in the Negro River floodplain, at the upper course of this river (northern Brazil), whose thermoluminescence ages include the LGM (22.8 ± 2.1 ka to 22.0 ± 1.9 ka) (Carneiro Filho et al., 2002). (v) Interruption in organic deposition, described as a hiatus, in several Brazilian lacustrine sites (Ledru et al., 1998). The hiatus is identified by a gap in 14C ages and is usually accompanied by an abrupt change of lithology, with the appearance of terrigenous sediment eroded from the nearby source area that replaces the lacustrine deposition of organic matter  s, and Aguas (Salitre, Caraja Emendadas sites), or without a lithological change but with reduced sedimentation rates (Serra Negra site). The hiatus lasted 9000 years at the Caraj as site and dry climate was indicated between 24,000 and 17,000 14C yr BP, including the LGM. (vi) Reduction in sedimentation rate, which remained constant and did not completely ceased in the Pata Lake, between 24,000 and 12,000 yr BP (Colinvaux and De Oliveira, 2000), or between 30,000 and 18,000 yr BP (Turcq et al., 2002). (vii) Appearance of a clastic layer interrupting the organic lacustrine deposition in the center and eastern portions of Pata Lake (Santos et al., 2001; Barbosa et al., 2004; and D'Apolito, et al., 2013), dated to ~ 18,000 yr BP and overlying an apparent depositional hiatus that occurred between 25,700 yr BP and 18,000 yr BP (31,285e27,900 cal yr BP) and 18,000 yr BP (21,477e20,600 cal yr BP) (Santos et al., 2001). The interval between 20 and 18 cal ka BP corresponds to the lowest lake level (D'Apolito et al., 2013). (viii) The occurrence of a nodular yellowish layer (D'Apolito et al., 2013) and increasing iron content (Cordeiro et al., 2011) in sediment profiles in Pata Lake dated to the LGM also indicate conditions favorable to iron higher concentrations at that time. (ix) One of the time intervals related to a reduction in the Amazon River outflow to the Atlantic Ocean (21e18 ka yr BP) and maximum aridity in the Amazon Basin (20.5e17 ka yr BP) (Maslin et al., 2011). Moreover, our study indicates that the area under arid conditions during the LGM extended farther west-northwest than previously proposed by Cheng et al. (2013). 6.2. LGM, base level change, and floodplain deposition Van der Hammen and Hooghiemstra (2000) estimated a reduction of rainfall of 50%e60% at the Katira site between ca. 28,000 and 13,000 yr BP and about 30%e50% in the Amazon Basin during the LGM as compared to the present. This significant

94

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

reduction in rainfall would have led to the incision of fluvial deposits between 28,000 and 13,000 yr BP. Irion and Kalliola (2010, their Fig. 11.4) estimated the average ~ es-Amazon River to be more than 40 m drawdown of the Solimo compared to the current level at the city of Manaus, and a slope to be 7.6 cm/km in the reach between Manaus and Santarem, during the LGM. The chronological framework presented here supports the previous interpretation proposed by Van der Hammen and Hooghiemstra (2000) and Irion and Kalliola (2010) on the incision of fluvial deposits between 28,000 and 13,000 yr BP. However, Van der Hammen and Hooghiemstra (2000) concluded that deposition ceased during the incision stage, resuming only after 13,000 yr BP, which is not supported by the chronological data presented in this paper. The OSL age obtained for site PD-03 shows that floodplain deposition occurred during the LGM, which would ~ eshave been due to the lowering of the base level of the Solimo Amazon River. According to our data, this lowering of the base level formed the upper river terrace, whose erosion and weathering occurred concurrently with the deposition of the floodplain during the LGM. This interpretation is supported by the agreement of the OSL ages of 20 ka yielded by the unaltered deposits at site PD-03 and the weathered deposits dated between 25 ka and 17 ka. The ages presented here for the floodplain (our data for site PD03 and the age range of Gonçalves et al., 2016) indicate that the floodplain is a key depositional unit in the understanding of the influence of the LGM in Central Amazonia.

~es-Amazon River and sea-level 6.3. Central channel of the Solimo change The OSL ages of 11.5 ka (this study) and 11.9 ka (Ferreira, 2013) obtained for lateral and longitudinal bars, respectively, of the Sol~es-Amazon River, show that fluvial sedimentation in the curimo rent channel of the river was active at the Pleistocene-Holocene boundary. The agreement of these ages also suggests synchronicity in the sedimentation of lateral and longitudinal bars in the current ~es-Amazon River. In addition, the chronochannel of the Solimo logical result obtained in a lateral fluvial bar (site PD-48) points to a significant regional change in the depositional regime of the Sol~es-Amazon River, with a minimum age of 13.1 ka, when imo deposition in lateral bars, and probably also in levees, led to the ~ estruncation of the floodplain and the entrenchment of the Solimo Amazon River in its current trough. Thus, the deposition of the lateral bars, longitudinal bars, and probably also the levees, occurs from the Pleistocene-Holocene boundary and is still a current process, as already pointed out by previous authors (e.g., Mertes et al., 1996; Latrubesse and Franzinelli, 2002; Irion and Kalliola, 2010; Rozo et al., 2012). The only data on the age of onset of bar ~es-Amazon River is deposition in the current channel of the Solimo presented by Gonçalves et al. (2016), who pointed to the end of deposition of their intermediate terrace (corresponding to our floodplain) in 19.1 ka, and the beginning of deposition of their lower terrace that corresponds to the fluvial deposits of the current ~es-Amazon River, at 18.3 ka. Taking into account channel of Solimo the ages obtained by Gonçalves et al. (2016), deposition in the ~es-Amazon River would have started current channel of the Solimo near the end of the LGM and was not preceded by erosion of floodplain deposits in the prior 800 years (19.1e18.3 ka). However, one should notice that, regionally, the time interval between 19.1 ka and 18.3 ka agrees with the following: (i) the estimated age for the erosion phase that affected the top of the organic deposit dated at 18,500 ± 150 14C yr BP

(22,682e21,941 cal yr BP) at the Katira site (Van der Hammen and Absy, 1994); (ii) the lowest lake level in Pata Lake, which occurred between 20 and 18 cal ka BP (D'Apolito et al., 2013); (iii) the end of a long time of hiatus (22 kae13 ka BP, Sifeddine et al., 2001 and Hermanowski et al., 2012a) in the deposition of organic sediment in lakes and swamps in the Caraj as area. The OSL ages of 11.5 ka (this study) and 11.9 ka (Ferreira, 2013) were obtained for outcropping portions of lateral and longitudinal ~ es-Amazon bars emergent during the lower phase of the Solimo ~ esRiver during the annual dry season. The thalweg of the Solimo Amazon River had a maximum depth of 30 m from 1978 to 1998 in the area immediately upstream of the confluence with the Negro River (Franzinelli, 2011), located 26 km NE of site PD-48. Due to the barchanoid geometry of the river bars (Almeida et al., 2016), the ages of 11.5 ka (this work) and 11.9 ka (Ferreira, 2013) provide a minimum age for the start of construction of these bars; therefore, deeper layers, submerged during the entire year, should provide older ages. Thus, it is unclear whether the gap of time of about 4 ka recorded by the minimum age of 17 ka obtained for the floodplain (site PD-03) and maximum age of 13 ka for the scroll bar (site PD48) represents non-deposition and erosion or if there was continuity of sedimentation that has not yet been dated due to the difficulty of sampling portions of the bars that are submerged all year. Evidence supporting a depositional hiatus and erosion is given by Müller et al. (1995), who described an acoustic Unit A of probable Pleistocene age with indications of erosion at the top in the Manacapuru and Preto da Eva ria lakes. Evidence supporting the continuity of deposition includes the OSL ages (with a small gap) reported by Gonçalves et al. (2016) and perhaps the suggestion of Latrubesse and Franzinelli (2002) that the floodplain deposits may underlie and constitute the inner part of the longitudinal bars in the mid-channel, although this obviously does not eliminate the possibility of prior erosion. In summary, it is still uncertain whether the time lapse of ca. 4 ka is a time of non-deposition. More studies in the area are needed to confirm or refute this hypothesis. 6.4. Ria lakes and sea level change Changes in the sedimentary environment of the Sol~eseAmazon River leading to the formation of ria lakes in Cenimo tral Amazonia due to the Holocene sea level rise have been already described by Irion et al. (2009, 2011). Ria lakes are typically elongated lakes with variable length (many with lengths up to 1 km but some up to 100 km) and a depth of just a few meters (usually less than 5 m) in the smaller lakes (Irion et al., 2011). Good examples of ria lakes in Central Amazonia are on the Preto da Eva, Urubu, and ~ Rivers (located about 75 km, 130 km, and 225 km east of Uatuma the city of Manaus, respectively) (Müller et al., 1995; Irion et al., 2009; Rozo et al., 2012). In the study area, Manacapuru Lake is the best known (Müller et al., 1995). Two hypotheses have been suggested for the origin of the ria lakes. One hypothesis attributes their genesis to the drowning of lower courses of tributaries with low sediment loads due to rising sea level in the Holocene; thus, their elongated shape is considered to be inherited from channels eroded during the LGM, when sea level was lower (Müller et al., 1995). However, a tectonic origin for these elongated lakes has been proposed by several authors (Sternberg, 1950; Dumont, 1993; Franzinelli and Igreja, 2002; Latrubesse and Franzinelli, 2002; Bemerguy et al., 2002, Ibanez et al., 2014a; Bertani et al., 2015). Ages previously obtained for ria lakes in the central part of Amazonia corroborate the interpretation of Irion et al. (2009, 2011),

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

given that only Holocene ages were obtained, with a maximum age of 9000 14C yr BP for the basal lacustrine deposit at Calado Lake (Behling et al., 2001) and an age of 10,237e9914 cal yr BP ~ es-Amazon River (8975 ± 59 14C yr BP) for the avulsion of the Solimo that formed Coari Lake (Horbe et al., 2011). However, Bertani et al. (2015) dated the beginning of formation of a paleoria lake, located in the interfluve of the Madeira and Purus Rivers, at 22,000 cal yr BP and related the tributary detachment to neotectonic activity. Given the existence of continuous organic sedimentation and no evidence of erosion at 22,000 cal yr BP, Bertani et al. (2015) suggested that the LGM did not leave traces in Central Amazonia. However, it should be noted that there is no information about why this detachment occurred at 22,000 cal yr BP, during the LGM. Is this a coincidence? Did the lowering of sea level and the reduction of water flow in river channels have some influence on tectonic movement within the continent? Recent (but undated) vertical crustal movements were related ~ River basins, just east to the development of the Urubu and Uatuma ~es-Amazon and Negro Rivers. The of the junction of the Solimo crustal movements, which caused lateral stream migration in the opposite sense to the Amazon River flow as well as channel sinuosity changes and topographic alterations, were attributed to flexural subsidence at the confluence area, resulting from surface water loading (Ibanez et al., 2014b). The crustal oscillation is an instantaneous response to the hydrological loading cycle and may affect the entire Amazon basin (Guimar~ aes et al., 2012). This effect may extend back to the Miocene (Ibanez et al., 2014b). Considering that there was an abrupt change in the hydrological regime in the last glacial-interglacial cycle, and taking into account the advent of drought conditions with reduced water volume in the river drainage network in the central part of Amazonia, it is possible to assume that the continent has responded with isostatic rebound. In the beginning of the Holocene, the steady increase of moisture in the Amazon Basin and the 40% increase in Amazon River discharge (Maslin and Burns, 2000) are both related to a higher volume of water in the Amazonian river channels, which could led to the isostatic readjustment of the continent by flexural subsidence and closure of the mouth of tributaries channels. It should be noted that the continuous sedimentation record between nearly 22,000 cal yr BP and 6000 cal yr BP in the paleoria studied by Bertani et al. (2015) conflicts with evidence of erosion at the top of the organic soil dated at 18,500 ± 150 14C yr BP (22,682e21,941 cal yr BP) (Van der Hammen and Absy, 1994) in the Katira Creek valley, located about 115 km southeast. This discordant information probably results from sampling, as the lacustrine sediments studied by Bertani et al. (2015) were cored at the mouth of a tributary of the Madeira River, an area that could be favorable to continued flooding over time, and probably is more protected from changes in local water level. In addition, the location sampled by Bertani et al. (2015) is in the defined transition region between savanna or cerrado vegetation and drier rain forest for the LGM (Van der Hammen and Hooghiemstra, 2000), fitting into the pre axis dicted vegetation gradient for the Katira-Porto Velho-Humaita during the LGM (Anhuf et al., 2006). 7. Conclusions ~esDetermining the chronology of fluvial deposits of the Solimo Amazon system is important to the understanding of the dynamics of this river, the largest on Earth, and the influence of climate change on its environment. The geomorphological, sedimentological, and chronological data collated in this paper for the deposi~es-Amazon River between its junctions with tional units of Solimo the Purus and Negro Rivers have been used to provide the most comprehensive chronological framework for this area.

95

OSL data suggest the Late Pleistocene (between 25 ka and 17 ka) as a time for exposure and weathering of fluvial deposits forming the Pleistocene scroll-dominated plain and contemporaneous floodplain. This weathering occurred under conditions of restricted water availability leading to pedogenic hematite formation, a mineral that imparts a typical red color to weathered deposits. The recognition of the dominance of hematite in alluvial deposits dated from 25 ka to 17 ka fills a gap of paleoclimate and chronological data in the central Amazonia lowlands and supports the set of regional paleoclimate records obtained from lakes, swamps, and colluvium, which points to dry conditions in this central area during the LGM. The decrease in rainfall and water volume in the river channels led to continental isostatic rebound and the formation of fluvial terraces. However, the decrease in the volume of water in river channels was not enough to completely interrupt river transport, which allowed the deposition of a floodplain due to the lowering of river base level and sea level. In the ocean, there was deposition of terrigenous sediments containing hematite. The data obtained in the present study suggest that Irion and Kalliola (2010) may be correct in proposing that the beginning of the Holocene was marked by changes in Amazonia lowlands, which led to the formation of ria lakes. Data collated in this paper suggest that increased rainfall and volume of water in rivers, isostatic adjustment of the continent, and deposition of lateral and mid~es-Amazon River, which led to channel fluvial bars in the Solimo the closure of deposition in the floodplain and formation of the ria lakes over its flat relief, are linked processes in the beginning of the Holocene. In this paper, this age is set at 11.5 ± 1.5 ka, the PleistoceneeHolocene boundary, as a minimum age. Acknowledgments ~o de This research was supported by the FAPESPdFundaça  Pesquisa do Estado de Sa ~o Paulo, and the first author is Amparo a thankful for the Grants 2007/58319-1 and Undergraduate fellowships 2009/53198-7 and 2010/01807-7. CR is a research fellow of CNPq, Brazil, Grant #307871/2010-0. We thank the journal editor, Dr. Neil Roberts, and an anonymous reviewer who provided critical review. References Absy, M.L., Cleef, A., Fournier, M., Martin, L., Servant, M., Sifeddine, A., Ferreira da Silva, M., Soubies, F., Suguio, K., Turcq, B., Van der Hammen, T., 1991. Mise en vidence de quatre phases d’ouverture de la fore ^t dense dans le sud-est de e res anne es. Premie re comparaison avec l’Amazonie au cours des 60000 dernie gions tropicales. Comptes Rendus Acade mie Sci. Paris,. 312 (Se rie II), d’autres re 673e678. Almeida, R.P., Galeazzi, C.P., Freitas, B.T., Janikian, L., Ianniruberto, M., Marconato, A., 2016. Large barchanoid dunes in the Amazon River and the rock record: implications for interpreting large river systems. Earth Planet. Sci. Letters 454, 92e102. Almeida-Filho, R., Miranda, F.P., 2007. Mega capture of the rio Negro and formation ^ nia, Brazil: evidences in a SRTM of the Anavilhanas Archipelago, central amazo digital elevation model. Remote Sens. Environ. 110, 387e392. Anhuf, D., Ledru, M.P., Behling, H., Cruz Jr., F.W., Cordeiro, R.C., Van der Hammen, T., Karmann, I., Marengo, J.A., De Oliveira, P.E., Pessenda, L., Sifeddine, A., Albuquerque, A.L., Da Silva Dias, P.L., 2006. Paleoenvironmental change in amazonian and african rainforest during the LGM. Palaeogeogr. Palaeoclimatol. Palaeoecol. 239, 510e527. Austermann, J., Mitrovica, J.X., Latychev, K., Milne, G.A., 2013. Barbados-based estimate of ice volume at Last Glacial Maximum affected by subducted plate. Nat. Geosci. 6, 553e557. Baker, P.A., Fritz, S.C., 2015. Nature and causes of Quaternary climate variation of tropical South America. Quat. Sci. Rev. 124, 31e47. Balsam, W., Ji, J., Chen, J., 2004. Climatic interpretation of the Luochuan and Lingtai loess sections, China, based on changing iron oxide mineralogy and magnetic susceptibility. Earth Planet. Sci. Lett. 223, 335e348. Balsam, W.L., Ellwood, B.B., Ji, J., Williams, E.R., Long, X., El Hassani, A., 2011.

96

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97

Magnetic susceptibility as a proxy for rainfall: worldwide data from tropical and temperate climate. Quat. Sci. Rev. 30, 2732e2744. Barbosa, J.A., Cordeiro, R.C., Silva Filho, E.V., Turcq, B., Gomes, P.R.S., Santos, G., Sifeddine, A., Albuquerque, A.L.S., Lacerda, L.D., Hausladen, P.A., Tims, S.G., Levchenko, V.A., Fifield, L.K., 2004. 14C-AMS as a tool for the investigation of mercury deposition at remote Amazon location. Nuclear Instruments & Methods in Physics Research. Section B. Beam Interact. Mater. Atoms 223e22, 528e534. Behling, H., Keim, G., Irion, G., Junk, W., Mello, J.N., 2001. Holocene environmental changes in the central Amazon basin inferred from Lago Calado (Brazil). Palaeogeogr. Palaeoclimatol. Palaeoecol. 173, 87e101. Bemerguy, R.L., Costa, J.B.S., Hasui, Y., Borges, M.D., Junior, A.V.S., 2002. Structural geomorphology of the Brazilian Amazon region. In: Klein, E.L., Vasquez, M.L., ~es a  Geologia da Amazo ^nia, vol. 3. Sociedade Rosa-Costa, L.T. (Eds.), Contribuço m, pp. 201e207. Brasileira de Geologia Núcleo Norte, Bele Bertani, T.C., Rossetti, D.F., Hayakawa, E.H., Cohen, M.C.L., 2015. Understanding amazonian fluvial rias based on a late pleistoceneeholocene analog. Earth Surf. Process. Landforms 40, 285e292. Bush, M.B., Stute, M., Ledru, M.P., Behling, H., Colinvaux, P.A., De Oliveira, P.E., Grimm, E.C., Hooghiemstra, H., Haberle, S., Leyden, B.W., SalgadoLabouriau, M.L., Webb, R., 2001. Paleotemperature estimates of the lowland americas between 30 degrees south and 30 degrees north at the last glacial maximum. In: Markgraf, V. (Ed.), Interhemispheric Climate Linkages. Academic Press, New York, pp. 293e306. Bush, M.B., De Oliveira, P.E., Miller, M.C., Moreno, E., Colinvaux, P.A., 2004. Amazonian paleoecological histories: one hill, 3 watersheds. Palaeogeogr. Palaeoclimatol. Palaeoecol. 214, 359e393. fica Caputo, M.V., Rodrigues, R., Vasconcelos, D.N.N., 1972. Nomenclatura estratigra da Bacia do Amazonas. In: XXVI Congresso Brasileiro de Geologia. Sociedade m, pp. 36e46. Brasileira de Geologia, Bele Carneiro Filho, A., Schwartz, D., Tatumi, S.H., Rosique, T., 2002. Amazonian paleodunes provide evidence for drier climate phases during the late pleistoceneeholocene. Quat. Res. 58, 205e209. Cheng, H., Sinha, A., Cruz, F.W., Wang, X., Edwards, R.L., d’Horta, F.M., Ribas, C.C., Vuille, M., Stott, L.D., Auler, A.S., 2013. Climate change patterns in Amazonia and biodiversity. Nat. Commun. 4, 1411. Colinvaux, P.A., De Oliveira, P.E., 2000. Palaeoecology and climate of the Amazon basin during the last glacial cycle. J. Quat. Sci. 15, 347e356. Colinvaux, P.A., De Oliveira, P.E., Moreno, J.E., Miller, M.C., Bush, M.B., 1996a. A long pollen record from lowland Amazonia forest and cooling in glacial times. Science 274, 85e88. Colinvaux, P.A., Liu, K.-B., De Oliveira, P.E., Bush, M.B., Miller, M.C., SteinitzKannan, M., 1996b. Temperature depression in the lowland tropics in glacial times. Clim. Change 32, 19e33. Colinvaux, P.A., De Oliveira, P.E., Bush, M.B., 2000. Amazonian and neotropical plant communities on glacial time-scales: the failure of the aridity and refuge hypotheses. Quat. Sci. Rev. 19, 141e169. Cordeiro, R.C., Turcq, B., Sifeddine, A., Lacerda, L.D., Silva Filho, E.V., Gueiros, B., dua, E.O., Patchinelam, S.R., 2011. Biogeochemical Potty, Y.P., Santelli, R.E., Pa indicators of environmental changes from 50 Ka to 10 Ka in a humid region of the Brazilian Amazon. Palaeogeogr. Palaeoclimatol. Palaeoecol. 299, 426e436. Cornell, R.M., Schwertmann, U., 1996. The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses. VCH, Weinheim, 573 pp. D'Apolito, C., Absy, M.L., Latrubesse, E.M., 2013. The Hill of Six Lakes revisited: new data and re-evaluation of a key Pleistocene Amazon site. Quat. Sci. Rev. 76, 140e155. Dino, R., Soares, E.A., Antonioli, L., Riccomini, C., Nogueira, A.C., 2012. Palynostratigraphy and sedimentary facies of Miocene fluvial deposits of the Amazonas Basin. Braz. J. S. Am. Earth Sci. 34, 61e80. Dumont, J.F., 1993. Types of lakes as related neotectonics in Western Amazonia. In: sio Internacional do Quaterna rio da Amazo ^ nia. PICH/INQUA, Manaus, Simpo pp. 99e102. ^ncia dos sedimentos holoce ^nicos da Ferreira, M.P., 2013. Geocronologia e provenie ^ncia dos rios Negro e Solimo ~ es, AM. MSc. Dissertation. University of S~ conflue ao Paulo, S~ ao Paulo, Brazil, 81 pp.  gicas da conflue ^ncia dos rios Negro e Franzinelli, E., 2011. Características morfolo ~ es (Amazonas, Brasil). Rev. Bras. Geosci. 41, 587e596. Solimo Franzinelli, E., Igreja, H., 2002. Modern sedimentation in the lower Negro River, amazonas state, Brazil. Geomorphology 44, 259e271. Gonçalves Jr., E.S., Soares, E.A.A., Tatumi, S.H., Yee, M., Mittani, J.C.R., 2016. Pleis~es-amazon fluvial system between tocene-holocene sedimentation of solimo the tributaries Negro and Madeira, central Amazon. Braz. J. Geol. 46, 167e180. ~es geogr ^nia. Revista Brasileira de GeoGourou, P., 1950. Observaço aficas na Amazo grafia. Ano XII 2, 3e165. Gross, M., Piller, W.E., Ramos, M.I., Paz, J.D.S., 2011. Late Miocene sedimentary en~es Formation; Brazil). J. S. Am. vironments in south-western Amazonia (Solimo Earth Sci. 32, 169e181. ~es, G.N., Blitzkow, D., Matos, A.C.O.C., Almeida, F.G.V., Barbosa, A.C.B., 2012. Guimara Analysis of the crust displacement in Amazon basin. In: Kenyon, S., Pacino, M.C., Marti, U. (Eds.), Geodesy for Planet Earth, vol. 136. International Association of Geodesy Symposia, Buenos Aires, Argentina, pp. 885e891. Harris, S.E., Mix, A.C., 1999. Pleistocene precipitation balance in the Amazon Basin recorded in deep sea sediments. Quat. Res. 51, 14e26. Heine, K., 2000. Tropical South America during the Last Glacial Maximum: evidence from glacial, periglacial and fluvial records. Quat. Int. 72, 7e21.

Hermanowski, B., Costa, M.L., Carvalho, A.T., Behling, H., 2012a. Palaeoenvironmental dynamics and underlying climatic changes in southeast Amas, Brazil) during the late Pleistocene and Holocene. zonia (Serra Sul dos Caraja Palaeogeogr. Palaeoclimatol. Palaeoecol. 365e366, 227e246. Hermanowski, B., Costa, M.L., Behling, H., 2012b. Environmental changes in southeastern Amazonia during the last 25,000 yr revealed from a paleoecological record. Quat. Res. 77, 138e148. Hess, L.L., Melac, J.M., Novo, E.M.L.M., Barbosa, C.C.F., Gastil, M., 2003. Dual-season mapping of wetland inundation and vegetation for the central Amazon basin. Remote Sens. Environ. 87, 404e428. Horbe, A.M., Behling, H., Nogueira, A.C.R., Mapes, R., 2011. Environmental changes in the western Amazonia: morphological framework, geochemistry, palynology ^ncias 83, 863e874. and radiocarbon dating data. An. Acad. Bras. Cie Hilgers, A., Murray, A.S., Schlaak, N., Radtke, U., 2001. Comparision of quartz OSL protocols using Lateglacial and Holocene dune sands from Brandenburg, Germany. Quat. Sci. Rev. 20, 731e736. Hua, Q., Barbetti, M., Rakowski, A.Z., 2013. Atmospheric radiocarbon for the period 1950-2010. Radiocarbon 55, 2059e2072. Hughes, P.D., Gibbard, P.L., 2015. A stratigraphical basis for the last glacial maximum (LGM). Quat. Int. 383, 174e185. Ibanez, D.M., Riccomini, C., Miranda, F.P., 2014a. Geomorphological evidence of recent tilting in the central Amazonia region. Geomorphology 214, 378e387. Ibanez, D.M., Riccomini, C., Miranda, F.P., 2014b. Is the current stress state in the Central Amazonia caused by surface water loading? J. S. Am. Earth Sci. 55, 19e28. Irion, G., Kalliola, R., 2010. Long-term landscape development processes in Amazonia. In: Hoorn, C., Wesselingh, F.P. (Eds.), Amazonia, Landscape and Species Evolution: a Look into the Past. Wiley-Blackwell, Chichester, pp. 185e197. Irion, G., Müller, J., Mello, J.N., Junk, W.J., 1995. Quaternary geology of the Amazonian lowland. Geo-Marine Letters 15, 172e178. Irion, G., Müller, J., Keim, G., Mello, J.N., Junk, W.J., 1999. The late quaternary river and lake development in Central Amazonia. In: International Symposium on Hydrological and Geochemical Processes in Large-scale River Basins. Manaus, Brazil, Extended Abstract (10pp) in the Abstract Volume (CD-ROM). Irion, G., Müller, J., Morais, J.O., Keim, G., Mello, J.N., Junk, W.J., 2009. The impact of Quaternary sea level changes on the evolution of the Amazonian lowland. Hydrol. Process. 23, 3168e3172. Irion, G., Mello, J.A.S.N., Morais, J., Piedade, M.T.F., Junk, W.J., Garming, L., 2011. Development of the Amazon Valley during the middle to late Quaternary: sedimentological and climatological observations. Amaz. Floodplain For. Ecol. Stud. 210, 27e42. Jeong, G.Y., Cheong, C.S., Choi, J.H., 2007. The effect of weathering on optically stimulated luminescence dating. Quat. Geochronol. 2, 117e122. Lambeck, K., Rouby, H., Purcell, A., Sun, Y., Sambridge, M., 2014. Sea level and global ice volumes from the last glacial maximum to the Holocene. PNAS 111, 15296e15303. Latrubesse, E.M., Franzinelli, E., 2002. The Holocene alluvial plain of the middle Amazon River, Brazil. Geomorphology 44, 241e257. Latrubesse, E.M., Franzinelli, E., 2005. The late Quaternary evolution of the Negro River, Amazon, Brazil: implications for island and floodplain formation in large anabranching tropical systems. Geomorphology 70, 372e397. Latrubesse, E.M., Cozzuol, M., Silva-Caminha, S.A.F., Rigsby, C.A., Absy, M.L., Jaramillo, C., 2010. The late Miocene paleogeography of the Amazon basin and the evolution of the Amazon River system. Earth Sci. Rev. 99, 99e124. Ledru, M.P., Bertaux, J., Sifeddine, A., Suguio, K., 1998. Absence of last glacial maximum records in lowland tropical forests. Quat. Res. 49, 233e237. Maslin, M.A., Burns, S., 2000. Reconstruction of the Amazon Basin effective moisture availability over the past 14,000 years. Science 290, 2285e2287. Maslin, M.A., Ettwein, V.J., Wilson, K.E., Guilderson, T.P., Burns, S.J., Leng, M.J., 2011. Dynamic boundary-monsoon intensity hypothesis: evidence from the deglacial Amazon River discharge record. Quat. Sci. Rev. 30, 3823e3833. Maxbauer, D., Feinberg, J.M., Fox, D.L., 2016. Magnetic mineral assemblages in soils and paleosols as the basis for paleoprecipitation proxies: a review of magnetic methods and challenges. Earth Sci. Rev. 155, 28e48. Mertes, L.A.K., Dunne, T., Martinelli, L.A., 1996. Channel-floodplain geomorphology ~es-Amazon River, Brazil. GSA Bull. 108, 1089e1107. along the Solimo Mertes, L.A.K., Dunne, T., 2008. Effects of tectonism, climate change, and sea-level change on the form and behavior of the modern Amazon River and its floodplain. In: Gupta, A. (Ed.), Large Rivers: Geomorphology and Management. Wiley, Wiltshire, pp. 115e144. Moreira, L.S., Moreira-Turcq, P., Kim, J.H., Turcq, B., Cordeiro, R.C., Caquineau, S., Viana, J.C.C., Brandini, N., 2013. Holocene paleoenvironmental reconstruction in the eastern amazonian basin: Comprido lake. J. S. Am. Earth Sci. 44, 55e62. Moreira, L.S., Moreira-Turcq, P., Kim, J.H., Turcq, B., Cordeiro, R.C., Caquineau, S., , J.S., 2014. A mineralogical and organic Mandengo-Yogo, M., Sinninghe Damste geochemical overview of the effects of Holocene changes in Amazon River flow on three floodplain lakes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 415, 152e164. Müller, J., Irion, G., Mello, J.N., Junk, W., 1995. Hydrological changes of the Amazon during the last glacial-interglacial cycle in central Amazonia (Brazil). Naturwissenschaften 82, 232e235. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single- aliquot regenerative-dose protocol. Radiat. Meas. 32, 57e73. Murray, A.S., Wintle, A.G., 2003. The single aliquot regenerative dose protocol: potential for improvements in reliability. Radiat. Meas. 37, 377e381.

L.G. Sant'Anna et al. / Quaternary Science Reviews 170 (2017) 82e97 Nanson, G.C., 1980. Point bar and floodplain formation of the meandering Beatton River, northeastern British Columbia, Canada. Sedimentology 27, 3e29. Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey, C., Buck, C.E., Cheng, H., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., , C., Heaton, T.J., Hogg, A.G., Hughen, K.A., Haflidason, H., Hajdas, I., Hatte Kaiser, K.F., Kromer, B., Manning, S.W., Niu, M., Reimer, R.W., Richards, D.A., Scott, E.M., Southon, J.R., Turney, C.S.M., van der Plicht, J., 2013. IntCal13 and MARINE13 radiocarbon age calibration curves 0-50,000 years cal BP. Radiocarbon 55, 1869e1887. Reis, N.J., Almeida, M.E., Riker, S.L., Ferreira, A.L., 2006. Geologia e recursos minerais gico do Brasil, do Estado do Amazonas. Escala 1:1.000.000. CPRM - Serviço Geolo Manaus. es, A.M., 2005. New geological framework for the Rossetti, D.F., Toledo, P.M., Go Western Amazonia: implications for biogeography and evolution. Quat. Res. 63, 78e89. Rozo, M.G., Nogueira, A.C.R., Truckenbrodt, W., 2012. The anastomosing pattern and the extensively distributed scroll bars in the middle Amazon River. Earth Surf. Process. Landforms 37, 1471e1488. Santos, G.M., Cordeiro, R.C., Silva Filho, E.V., Turcq, B., Lacerda, L.D., Fifield, L.K., Gomes, P.R.S., Hausladen, P.A., Sifeddine, A., Albuquerque, A.L.S., 2001. Chronology of the atmospheric mercury in Lagoa da Pata Basin, Upper Rio Negro Region of Brazilian Amazon. Radiocarbon 43, 801e808. Shakun, J.D., Clark, P.U., He, F., Marcott, S.A., Mix, A.C., Liu, Z., Otto-Bliesner, B., Schmittner, A., Bard, E., 2012. Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation. Nature 484, 49e54. Sifeddine, A., Martin, L., Turcq, B., Volkmer-Ribeiro, C., Soubies, F., Cordeiro, R.C., Suguio, K., 2001. Variations of the Amazonian rainforest environment: a sedimentological record covering 30,000 years. Palaeogeogr. Palaeoclimatol. Palaeoecol. 168, 221e235.  sitos pleistocenos da regia ~o de conflue ^ncia dos rios Negro Soares, E.A.A., 2007. Depo ~ es, porç~ ~o e Solimo ao oeste da Bacia do Amazonas. PhD Thesis. University of Sa ~o Paulo, Brazil, 255 pp. Paulo, Sa sitos quaSoares, E.A.A., Suguio, K., Barros, D.S., Santos, W.H.D., 2001. Os depo rios na conflue ^ncia dos rios Negro e Solimo ~ es, municípios de Iranduba e terna sio de Geologia da Amazo ^nia. SBG/NO, Manacapuru, Amazonas. In: VII Simpo m, CD-Rom. Bele ^nica e sedimentaça ~o quatern Soares, E.A.A., Riccomini, C., 2003. Neotecto aria na

97

~o do baixo Rio Negro, Amazonas. In: VIII Simpo sio de Geologia da regia ^nia. SBG/NO, Manaus, CD-Rom. Amazo ~o pleistocena nas Soares, E.A.A., Riccomini, C., Nogueira, A.C.R., 2007. Sedimentaça ~ es tecto ^nicas da regia ~o de conflue ^ncia dos rios Negro e Solimo ~es, depresso rio. Amazonas. In: XI Congresso da Associaç~ ao Brasileira de Estudos do Quaterna m, Boletim de Resumos. ABEQUA, Bele Soares, E.A.A., Tatumi, S.H., Riccomini, C., 2010. OSL age determinations of Pleisto^ncias 83, 691e699. cene fluvial deposits in Central Amazonia. An. Acad. Bras. Cie Soares, E.A.A., Dino, R., Soares, D.P., Antonioli, L., Silva, M.A.L., 2015. New sedimentological and palynological data from surface Miocene strata in the central Amazonas Basin area. Braz. J. Geol. 45, 337e357. ^ nicos na planície amazo ^ nica? Rev. Bras. Geogr. 4, Sternberg, H.O.R., 1950. Vales tecto 3e26. Stuiver, M., Reimer, P.J., Reimer, R.W., 2017. CALIB 7.1 [WWW program] at http:// calib.org, accessed 2017-6-12. Sylvestre, F., 2009. Moisture pattern during the last glacial maximum in south America. In: Vimeux, F., Sylvestre, F., Kodri, M. (Eds.), Past Climate Variability in South America and Surrounding Regions: from the Last Glacial Maximum to the Holocene. Developments in Paleoenvironmental Research, vol. 14, pp. 3e27. ~es Filho, F.F.L., Albuquerque, A.L.S., Turcq, B., Cordeiro, R.C., Sifeddine, A., Simo Abr~ ao, J., 2002. Carbon storage in Amazonia during the last glacial maximum: secondary data and uncertainties. Chemosphere 49, 821e835. van de Lageweg, W.I., van de Dijk, W.M., Baar, A.W., Rutten, J., Kleinhans, M.G., 2014. Bank pull or bar push: what drives scroll-bar formation in meandering river? Geology 42, 319e322. Van der Hammen, T., Absy, M.L., 1994. Amazonia during the last glacial. Palaeogeogr. Palaeoclimatol. Palaeoecol. 109, 247e261. Van der Hammen, T., Hooghiemstra, H., 2000. Neogene and Quaternary history of vegetation, climate, and plant diversity in Amazonia. Quat. Sci. Rev. 19, 725e742. Vimeux, F., Sylvestre, F., Kodri, M., 2009. Past climate variability in south America and surrounding regions: from the last glacial maximum to the Holocene. Dev. Paleoenviron. Res. 14, 418. Zhang, Y.G., Ji, J.F., Balsam, W., Liu, L., Chen, J., 2007. High resolution hematite and goethite records from ODP 1143, South China Sea: Co-evolution of monsoonal ~ o over the past 600,000 years. Earth Planet. Sci. Letters precipitation and El Nin 264, 136e150.