Chapter 36 Evolution of the Amazon Basin in Ecuador with Special Reference to Hinterland Tectonics: Data from Zircon Fission-Track and Heavy Mineral Analysis

The retro-arc Andean Amazon Basin (AAB), a foreland basin located east of the evolving Andes, has been an active sediment depocentre since the late Early Cretaceous (Aptian-Albian). Heavy mineral analysis of the proximal shallow marine and continental deposits in the Sub-Andean Zone reveals an overall trend from ultrastable zircon, tourmaline and rutile (ZTR) dominated assemblages, to more complex heavy-mineral suites with metamorphic and mafic volcanic signatures. This reflects successive derivation from shallow to deep crustal rocks and, subsequently, from accreted oceanic terranes during formation of the proto-Andes. In parallel with heavy mineral analysis, detrital zircons were dated using fission-track methodology. The measured lag times of the zircons range from 400 to 0 Ma. Approximately zero lag times generally represent fast cooling events in the source region, while long lag times are the result of slow cooling or recycling in the source regions. In the present study, zero lag times combined with lithological and mineralogical data sets have also been used to identify volcanic events in the hinterland which, in turn, establishes the stratigraphic age of the enclosing sediments more precisely. Changes in source areas are identified by changes in lag times, which coincide with variations in the composition of the heavy mineral suites. During early stages of basin evolution (Aptian-Campanian), lag times of detrital zircons, occurring in the dominantly ultrastable ZTR mineral suite, allow the discrimination of sediment input from two different continental basement sources. Zircons, with very long lag times (≈400 Ma), indicate input from the Amazonian Craton regions at the eastern margin of the basin, while comparatively short lag times (0–60 Ma) in zircons infer coeval sediment flux from the primordial Cordillera Real to the west. Higher up in the succession, a switch in provenance to dominantly westerly sources to the evolving Cordillera Real occurred, where progressively higher grade metamorphics were eroded. In the uppermost, Miocene to Pliocene sections, the appearance of mafic minerals documents new sources associated with the uplift of accreted basement terranes of the Cordillera Occidental and the reworking of arc volcanics. Keywords provenance analysis heavy minerals zircon fission-track ages exhumation geodynamics Andes Ecuador 1 Introduction Six discrete topographic and tectono-stratigraphic regions can be identified in northern Ecuador ( Fig. 1 ). These are from west to east: (1) the coastal lowlands (Costa), consisting of accreted oceanic plateau and arc series together with their Late Cretaceous to Neogene sedimentary cover; (2) the Cordillera Occidental, assembled as a tectonic patchwork of accreted Cretaceous oceanic basement, volcanic arcs and sedimentary successions; (3) the Inter-Andean Depression, formed during the last 6–5 Ma due to differential uplift and thrusting of the cordilleras; (4) the Cordillera Real comprising a tight tectonic stack of Palaeozoic and Mesozoic, medium- to high-grade metamorphosed plutonic, volcanic and sedimentary rocks; (5) the Sub-Andean Zone comprising a Jurassic volcanic arc basement and Cretaceous to Palaeogene cover rocks; (6) the Oriente, which is part of the upper Amazon Basin where the Proterozoic basement of the South American plate is overlain by fragmentary Palaeozoic and Mesozoic rocks. These include sedimentary and volcanic successions with mid-Cretaceous to Neogene, dominantly siliciclastic, cover (e.g., Baldock, 1982 ; Balkwill et al., 1995 ; Rivadeneira and Baby, 1999 ). The mid-Cretaceous to Neogene sedimentary series is common to both the Sub-Andean Zone and the Oriente and here is referred to as the AAB, which formed in a retro-arc foreland basin system ( Fig. 1B ) ( Willett et al., 1993 ; DeCelles and Giles, 1996 ). The Sub-Andean Zone is subdivided into a proximal, highly deformed stack of tectonic slivers between the Sub-Andean Fault to the west and the Cosanga Fault to the east, referred to as the Sub-Andean Thrust Belt, ( Figs. 1B and 2 ) and an outer, less deformed push-up, commonly called the Napo Uplift, in the north and the Cutucú Uplift in the south ( Figs. 1 and 2 ). Both units form the transition between the over 4000 m high Cordillera Real and the Oriente in front of the Andean Front Fault with an elevation of less than 400 m ( Fig. 2 ). The observed stratigraphic relationships imply that the Sub-Andean Zone and the Oriente originally represented a composite basin system that was initiated during the Middle Cretaceous. Since then and until recent times, the basin received clastic material from both the Amazon Craton (i.e., the Guyana Shield) and the Andean cordilleras. However, proximal parts of the basin were successively eliminated as sedimentation sites and became tectonically integrated into the orogenic wedge by eastward-directed thrust propagation (see below). The Northern Andes in Ecuador and Colombia are thought to be the product of repeated accretion of oceanic terranes, and continental compression (e.g., Jaillard et al., 1995 ). During the Early Cretaceous (approximately 140–120 Ma) possibly various small continental and oceanic plates were accreted during the tectono-metamorphic Peltetec event in the Cordillera Real ( Litherland et al., 1994 ). The Pallatanga and Piñon terranes of the Cordillera Occidental and Costa are assumed to have been accreted during the latest Cretaceous and Eocene, respectively (e.g., Egüez, 1986 ; Hughes and Pilatasig, 2002 ; Kerr et al., 2002 ). The younger history of the Ecuadorian Andes is related to the NNE movement of the Northern Andean Block after about 15 Ma ( Hungerbühler et al., 2002 ; Winkler et al., 2004 ). Displacement of the block is accompanied by approximately E–W oriented shortening and right-lateral displacement along regional scale strike-slip faults (e.g., Ego et al., 1996 ; Winkler et al., 2004 ). With this regional setting in mind, the main goal of the present study was to carry out integrated heavy mineral and fission-track analysis (FTA) of detrital zircon and apatite to determine the source of the sediments of the AAB, and to reveal stages of hinterland exhumation. 2 The Evolution of the Andean Amazon Basin The sediments of the AAB lie unconformably on a basement composed of Proterozoic metamorphic and plutonic rocks, Palaeozoic-Early Jurassic sediments and arc volcanics. The Misahualli volcanic arc series spans from Early Jurassic to Early Cretaceous (approximately 190–130 Ma) ( Romeuf et al., 1995 ; Spikings et al., 2001 ; Ruiz, 2002 ). The Abitagua granite and the equivalent Zamora granite in the southern Sub-Andean Zone are considered to represent arc-related intrusions ( Litherland et al., 1994 ; Spikings et al., 2001 ; Ruiz, 2002 ). An angular unconformity exists between the oldest AAB sediments, the Hollin Formation and the basement ( Fig. 3 ) ( Tschopp, 1953 ; White et al., 1995 ; Shanmugam et al., 2000 ). The Aptian-Albian succession, up to 170 m thick, comprises coarse, porous, often tabular cross-bedded quartz-arenites characterised by the presence of blue quartz. The facies trend shows a transgressive sequence from alluvial braid plain to tidal plain and shallow marine conditions ( White et al., 1995 ; Shanmugam et al., 2000 ). The Middle Albian to Campanian Napo Group ( Fig. 3 ) consists of organic-rich shales, limestones and sandstones, which were deposited in a restricted paralic-neritic sea ( Mello et al., 1995 ; Jaillard, 1997 ; Vallejo et al., 2002 ). The many stratigraphic gaps that exist in the Turonian to Campanian section of the Sub-Andean Zone were the result of tectonic movements and wedge-top erosion ( Vallejo et al., 2002 ). The analysis of seismic lines in the Oriente reveals that large-scale, basement-involving transpressive tectonic inversions affected the area since the Turonian ( Balkwill et al., 1995 ; Rivadeneira and Baby, 1999 ). The deposition of redbeds in the Maastrichtian-Paleocene Tena Fm. ( Fig. 3 ) marks an extreme change towards prevailing continental conditions in a fluvial environment. During this time-span, the AAB developed a typical triangular retro-arc foreland profile with a maximum thickness (approximately 1000 m) in the west (i.e., in the Sub-Andean Zone) wedging out to ∼150 m toward the easternmost Ecuadorian Oriente ( Dashwood and Abbotts, 1990 ). A major erosional unconformity separates the Tena Fm. from the conglomeratic, Eocene Tiyuyacu Fm. ( Fig. 3 ). A shift of the depocentre towards the east is described by Christophoul et al. (2002a) . This succession was deposited in fluvial (lower member) and alluvial-fan (upper member) systems—both sourced from the west in the Andean Cordillera. The marginal marine deposits of the ∼250 m thick Oligocene Orteguaza Fm. paraconformably overlap the Tiyuyacu Fm. ( Christophoul et al., 2002a ), and pass up into the 250–450 m thick continental sediments of the Chalcana Fm. The overlying Neogene fluvial-alluvial Arajuno and Chambira formations ( Fig. 3 ) are separated by a minor erosional unconformity ( Christophoul et al., 2002b ; Ruiz, 2002 ). The sandy and silty Curaray Fm. is considered to represent the contemporaneous estuarine system toward the east (e.g., Baldock, 1982 ). The Neogene series had a clear western source with transport directions turning parallel to the Andean Cordillera in distal position in the Oriente foredeep ( Christophoul et al., 2002b ) and the shifting of the depocentres towards the east. The Late Pliocene alluvial fan deposits of the Mesa/Mera formations are restricted to the Pastaza Depression, the most important outlet draining the Cordillera Real and the Inter-Andean Depression. Notably, in the Sub-Andean Thrust Belt ( Litherland et al., 1994 ) ( Fig. 1 ), the older horizons of the AAB series, the Hollin Fm., the Napo Group and the Tena Fm. together with the arc basement (Misahualli Fm. and Abitagua granite) are present as a steeply westward-dipping tectonic slice ( Fig. 1B ). This indicates that, by the end of the Palaeocene, proximal parts of the AAB had already been eliminated through shortening, i.e., were integrated into the orogenic wedge ( DeCelles and Giles, 1996 ). From the Miocene onward (Arajuno Fm.), sedimentation was directed towards the Pastaza Depression between the Napo and Cutucú uplifts on which sediment bypass occurred ( Christophoul et al., 2004 ). This suggests the emergence of these highs since that time. Due to prevailing transpressive shortening, this uplift was probably driven in positive flower structures ( Rivadeneira and Baby, 1999 ). 3 Methods—Theory and Practice Heavy mineral analysis and detrital zircon fission-track (ZFT) analysis were performed on the same samples, ranging from sandstones to fine conglomerates and minor siltstones. Details on samples including universal transverse mercator (UTM) coordinates are given in Table 1 . Both methods were also conducted on a sand sample from the modern Napo River near to Puerto Misahualli, as a control. The known drainage area of this river provides a reliable assessment of the sand composition derived from ancient sedimentary and magmatic source rocks. Stratigraphic ages based on previous work in the AAB ( Tschopp, 1953 ; Faucher and Savoyat, 1973 ; Bristow and Hoffstetter, 1977 ; Baldock, 1982 ; Balkwill et al., 1995 ; White et al., 1995 ; Jaillard, 1997 ; Rivadeneira and Baby, 1999 ; Zambrano et al., 1999 ; Christophoul et al., 2002a, 2002b ; Barragán et al., 2005 ) were augmented by the dating of syn-depositional volcanic events during the present study. 3.1 Heavy Mineral Analysis The heavy mineral preparation procedure followed the standard methods described by Mange and Maurer (1992) . Rocks were crushed and the carbonate fraction was dissolved in 10% acetic acid, while organic material and carbonate acetates were removed by the repeated addition of hydrogen peroxide. The heavy minerals were separated from the 0.063 to 0.4 mm sieve fraction using bromoform (density 2.88–2.9), and the residues were mounted in piperine (refraction index 1.67). Ideally, a minimum of 200 grains were counted under the petrographic microscope, using the mid-point ribbon counting method ( Van der Plas, 1962 ) to provide grain-size independent frequency percentage distributions. Where one particular mineral species was dominant, up to 300 grains were counted in order to encounter the rare components. Special attention was also paid to the presence of idiomorphic zircons, prismatic and inclusion-rich apatite and pseudo-hexagonal biotite, as these are potential indicators of contemporaneous volcanic influx ( Weaver, 1963 ; Winkler et al., 1985 ), critical for the correct interpretation of the meaning of lag times of the detrital ZFT ages. 3.2 Detrital Zircon Age Studies The detrital zircons were separated from 24 horizons for fission-track analysis. Technical details are reported in Ruiz (2002) and Ruiz et al. (2004) . 3.2.1 The concept of lag time Zeitler et al. (1986) first used the concept of lag time with fission-track ages of detrital zircons in a study of the Neogene molasse-type sediments from northern Pakistan. The lag time ( L g ) is defined as the difference between the time of closure of a geochronological system in the source region and the time of sedimentation in the basin ( Fig. 4 ). The closure temperature is variable and dependent on the cooling/exhumation rates but for this study a closure temperature of 260°C for ZFT ( Brandon et al., 1998 ) has been used throughout. In applying detrital thermochronology with emphasis on the concept of lag time, several points should first be noted ( Ruiz et al., 2004 ). Generally, in an exhuming single tectonic block, the first material eroded contains the oldest signature. This implies that ages derived from such a block are younger as the pile is eroded and deposited into the basin ( Garver et al., 1999 ) ( Fig. 4 ). Long lag times may represent either a slowly exhuming source region or reworking of older sediments. The distinction between these two can often be clarified by identifying the heavy minerals associated with the zircons. Short lag times, on the other hand, clearly indicate rapid exhumation of the source rocks, and as they approach a zero value, they become positive markers of active tectonics in the hinterland. Zero lag times are also indicative of input from contemporaneous volcanics into a basin, which may be confirmed by mineral and lithological associations. Thus, by tracking the variation in lag time in clastic sediments overtime, the observed changes can be linked to events in the hinterland. During a phase of exhumation, the first erosional products will yield inherited thermochronological ages from an earlier phase—these ages cannot constrain the exhumation rates associated with the new phase. The new phase is recorded only when the rocks at depth (10 km?) pass through the closure temperature. Once the newly recorded tectonic phase is at the surface, it is possible to estimate the exhumation rate in the hinterland by dividing the estimated depth of closure ( Z c ) by the lag time ( L g ): Exh= Z c / L g in mm·a −1 ( Garver et al., 1999 ) on the assumption that the time of transport ( t e ) from erosion site to basin is negligible ( Fig. 4 ). The rock cover above the temperature of closure T c before this new phase is termed the thermochronological “dead zone”. 3.2.2 Detrital age curves In the present study, the majority of sample ages failed the Pχ 2 -test ( Green, 1981 ), a value that is suggestive of the presence of multiple populations ( Table 2 ). Statistical separation of the grains ( Brandon, 1992, 1996 ) revealed three to five populations ( P n ) at some horizons. In order to visualise the change in the lag time upwards in the section, populations ( P 1 – P n ) were then joined according to their rank forming the detrital, D 1 – D n curves ( Fig. 5 ). Their change of gradient upwards within the stratigraphic column is a gauge of different events in the source regions. Ruiz (2002) and Ruiz et al. (2004) presented a series of five theoretical paths illustrating the possible trends that may occur in any sequence ( Fig. 6 ). Inflection points along a D n curve, depicting increases or decreases in lag times, indicate changes in the exhumation of the source regions and/or flipping of source regions. The D 1 curve is the easiest and safest to interpret because it hosts the population with the shortest lag time and, consequently, the most probable record of rapid cooling events in the nearby hinterland. The five paths are briefly described below with reference to Fig. 6 . The type 1 path most likely occurs in connection with a change of source region. The older ages may also be those derived from the upper thermochronological “dead zone” of a newly exhuming block in the hinterland where rocks from a renewed event have not yet reached the surface as explained above. Such examples can be identified and/or verified through the combination of heavy mineral studies. A type 2 path may suggest the erosion of a thick pile of volcanic rocks that were extruded over a short period of time as, for example, plateau basalts, or the erosion of a succession that underwent extremely rapid exhumation. Ages decrease upwards, but with increasing lag time the characteristic pattern of a type 3 path appears. Such a path represents waning of the exhumation assuming a constant depth of closure, and might follow the slowing down of an orogenic phase. The type 4 path describes a constant lag time through the stratigraphic record along an iso-lag time line parallel to the stratigraphic correlation line 1/1 ( Fig. 6 ). This path characterises “steady-state cooling” within the source region. The type 5 path is characterised by a decreasing lag time up-section, which implies increasing cooling/exhumation in the source region. This is typically associated with orogenic growth phases (e.g., Garver et al., 1999 ). 4 Results of Heavy Mineral Provenance Analysis Our data set represents the first integrated application of heavy mineral analysis for tracing the provenance of the detrital material in the Ecuadorian Andean domain. The results are summarised in Fig. 7 and Table 3 . The highly variable heavy mineral suites indicate a complex provenance from changing lithologies with time (see Mange and Maurer, 1992 , for a compilation of diagnostic minerals of specific rock types). With regard to the preservation potential of the heavy minerals in the sediments, we have positive arguments for excluding high-level diagenetic or metamorphic overprint. According to apatite fission-track (AFT) modelling in rocks of the Napo uplift ( Ruiz, 2002 ), from where the majority of our samples is derived, the volcanic basement (Misahualli Fm.) and, consequently, the sedimentary cover series were—since the start of deposition in the middle Cretaceous—never buried to a depth greater than the temperature-equivalent of 110°C (i.e., approximately 3–4 km at an assumed geothermal gradient of 30°C/km). Obviously, the sediments remained in a middle-grade diagenetic stage inferring only minor diagenetic modification of the heavy mineral assemblages. On the other hand, due to the high mineralogical and textural maturation stage of the quartz-arenites of the mid-Cretaceous Hollin Fm., a possible enrichment of stable minerals, i.e., the zircon/tourmaline/rutile (ZTR) association, may be invoked ( Hubert, 1962 ). However, the striking similarity with the mineral assemblages observed in the immature sandstones of the Napo Group suggests a primary source rock-induced signal for both formations. 4.1 Stratigraphic Trends of Heavy Minerals The Hollin Fm. and the Napo Group are dominated by the ZTR group, accompanied by other Ti-bearing minerals (brookite, anatase and sphene) as well as by minor amounts of monazite (rarely xenotime), and few clinopyroxenes of diopsidic composition (simply referred to as diopside) ( Fig. 7 ). In addition, the Hollin Fm. and the lower Napo Group are characterised by the abundance of pink and remarkably dark-brownish, often zoned, zircons. The dark-brownish colour of these zircons suggests long-lasting radioactive damage, i.e., they may represent “old” grains as confirmed by the fission-track measurements (see below). Rounded and rounded zoned zircons make up more than 94%, often 98–100% of the zircon population. Muscovite is abundant. The marine Napo Group also contains many diagenetic phosphatic and glauconite grains (not quantified in Fig. 7 and Table 3 ). With the strong dominance of the ZTR group, this assemblage indicates derivation of the clastic material primarily from shallow granitic continental crust comprising pegmatites (cassiterite), with minor additions from basic rocks or skarns (signalled by diopside), and polycyclic sediments. In the uppermost Napo Group and in the Tena Fm., apatite becomes common. In sample 98GR37 (Tena Fm., Fig. 7 ), the presence of inclusion-rich apatites suggests a volcanic source. However, no fission-track dating of associated zircons was obtained from this sample. Significantly, the first occurrence of metamorphic grains (garnet and chloritoid, most likely derived from medium-grade metamorphic shales) is observed in the Tena Fm. Up-section an increasing number of other metamorphic grains (epidote group, kyanite, sillimanite) is obvious ( Fig. 7 ). These generally medium- to high-grade metamorphics-derived associations commonly contain muscovite and chlorite. The metamorphic grains are accompanied by variable proportions of the ZTR group, representing associated granitic and reworked sedimentary rocks in the source areas. Several abrupt changes in the stratigraphic succession, depicted in Fig. 7 , suggest repeated switching of source terranes, i.e., variable activation or abandonment of sources and/or associated palaeodrainage systems. A quantitative increase from low/medium- to high-grade metamorphic assemblages in the stratigraphic column (e.g., samples 98GR81—99GR44 and 99GR81—00GR02—00GR03) points to accelerated exhumation of deeper, metamorphic crustal levels in the source areas. Increasingly diverse heavy mineral associations occur up-section from the Miocene Chalcana Fm. (00GR02) and the Miocene Arajuno Fm. (00GR04). The additional but variable appearance of augite, hypersthene, olivine, diopside, chromite and (green and brown) hornblendes points to the presence of basic to intermediate as well as ultrabasic source rocks. In the lower Tiyuyacu Fm. (99GR36) and Arajuno Fm. (00GR04) prismatic apatites and apatites with abundant strings of inclusions are present. They occur together with many volcanic rock fragments (in 99GR36) and pseudo-hexagonal biotite (in 00GR04). The inferred, contemporaneous volcanic influx is corroborated by the ZFT analysis revealing 0 Ma lag times (see below). The modern Napo River's main catchment area lies in the Cordillera Real between latitudes 1°S and 30°S. Its sand is rich in medium-grade metamorphic minerals (approximately 60% clinozoisite and minor zoisite, Fig. 7 ) and granitic rock fragments. Associated hornblende and augite may be derived either from minor intermediate to basic volcanic rocks or from reworking of the Miocene-Pliocene series, drained by its tributary, the Rio Arajuno, flowing along the Pastaza Depression towards the Rio Napo. Hence, the observed mineral association mirrors the rocks encountered in the drainage basin. In summary, the provenance of detrital material reflects an overall trend from shallow continental granitic crustal (Hollin Fm. and Napo Group) to medium- to high-grade metamorphic continental sources, with signatures from the latter appearing first in the Maastrichtian Tena Fm. This trend mirrors the tectonic evolution of the Cordillera Real, where increasingly deeper metamorphic levels of the nascent orogen were exhumed by differential uplift and unroofing. Furthermore, transport from a discrete, important source region, the Amazon Craton (i.e., the Guyana Shield), providing detritus during the Cretaceous, is suggested tentatively by the presence of “old” zircon grains. The relatively high abundance of rounded brookite, anatase and sphene, and the minor but continuous presence of monazite/xenotime, diopside and cassiterite in these sediments, could also be indicative of a cratonic source. In the Miocene sequences, the appearance of minerals from ultrabasic and basic volcanics indicates exposure of mafic rocks in the Cordillera Occidental that accreted in the Maastrichtian ( Vallejo et al., 2006 ). Alternatively, a part of these minerals also may have been reworked from arc volcanics in the cordilleras. This new supply into the basin coincides with the eastward shift of the depocentre in Arajuno Formation time (Miocene), and the opening of a fluvial conduit from the Cordillera Occidental across the Cordillera Real into the Pastaza Depression, as it exists today. 5 Results of Detrital Zircon Fission-Track Study The ages of the analysed zircon populations from the full succession range from 579±65 Ma to 22.9±1.2 Ma ( Table 2 ), and a general younging trend of the P 1 and P 2 populations is observed upwards ( Fig. 8 ). Lag times vary erratically throughout the section but it is interesting to note that the D 1 and D 2 curves are often parallel. This implies the possibility of a large-scale regional pattern overprinting the differently cooling/exhuming regions. The third population, P 3 in Fig. 8 , in sediments older than Maastrichtian, has ages older than 450 Ma. Many rounded and zoned pink zircons from the Hollin Fm. and Napo Group often had strong radiation damage and are un-datable because they are too old (track density too high to count). This leads to an immediate bias in this data set, but was not considered important for the present study because the identification of new proximal tectonic phases was the immediate goal of the project. The extremely long lag time is indicative of long, slow exhumation in the range of 0.02 mm·a −1 , typical of cratonic regions in Brazil ( Harman et al., 1998 ). The abundance of the associated ZTR group in the Late Cretaceous succession points to a probable source in the Guyana Shield and its Palaeozoic cover-rocks to the east of the basin. The old zircon-age population declines abruptly at the end of the Cretaceous when a major switch to prevailing supply from the Cordillera Real is inferred. In the Hollin Fm. and lower Napo Group, lag times measured in population P 1 range from 63 to 0 Ma, suggesting exhumation in the source regions in the order of 0.1–1 mm·a −1 . Such exhumation rates—an order faster—are not compatible with a derivation of the detrital material from the Amazon Craton, but rather suggest the initiation of a primordial Cordillera Real to the west of the basin. This is corroborated by the exhumation ages of the P 1 zircons, which correlate with the previous Peltetec tectonic event ( Fig. 8 ), when the northern South American continental margin was deformed ( Litherland et al., 1994 ). Our observation stands in clear contrast with sedimentary facies reconstructions in oil exploration boreholes (e.g., White et al., 1995 ; Barragán et al., 2004 ) in the Oriente whereby a sole supply to the early AAB from the Amazon Craton is interpreted. From Late Cenomanian to Coniacian, a strong type 5 path is evident in the D 1 and D 2 paths ( Fig. 8 ), implying an increase in exhumation in the source region that is interpreted as the initiation of a tectonic event. A minimum exhumation rate of 1 mm ·a −1 using a maximum lag time of 6 Ma (error on 98RS06, Table 2 ) is estimated ( Garver et al., 1999 ). A very short lag time is evident in the D 1 path through to the middle Palaeocene (a type 4 path) and is correlated with the incoming of low- to medium-grade metamorphic minerals ( Figs. 7 and 8 ) in the Tena Fm. This detrital record most likely reflects the initiation and closure of the Pallatanga event, associated with the accretion of the oceanic Pallatanga Terrane in the Ecuadorian forearc ( Cosma et al., 1998 ; Lapierre et al., 2000 ; Hughes and Pilatasig, 2002 , Vallejo et al., 2006 ), and coincides also with the disappearance of the presumably Guyana Shield derived old zircons. The short lag time, the loss of older ages and the incoming of metamorphic minerals are strong indications that the new source is to the west of the AAB, i.e., in the growing Cordillera Real. This is confirmed by (1) a switch in palaeocurrent direction from an easterly to a dominant westerly source for the Tena Fm. (e.g., Tschopp, 1953 ; Balkwill et al., 1995 ), resulting in lower input from the eastern basin margins; (2) a major change in the depositional environment from shallow marine (Napo Group) to continental (Tena Fm.); and (3) increased tectonic subsidence ( Thomas et al., 1995 ), suggesting an increase in orogenic load to the west of the basin. Evidence for the first volcanic input in the Lower Eocene sediments is provided by sharp, idiomorphic zircons and a tuffaceous cement in the conglomerate sample 99GR36 ( Figs. 5 and 7 ). The clast lithology at this horizon 99GR36 is unique, comprising dominantly sub-rounded radiolarian black cherts of 10–12 cm in diameter, which may have been derived from the earlier accreted (Maastrichtian) Pallatanga Terrane ( Ruiz, 2002 ). The zircons in the matrix yield a zero lag time for the P 1 population with an age of 51±5 Ma ( Table 1 ) and hence record the age of the host sediment. Because of the contemporaneous volcanic origin of the P 1 population, it is not considered in the detrital curves, and at this horizon P 2 becomes P 1 . This leads into an extreme type 1 path ( Fig. 8 ), implying a switch to a new hinterland. This path coincides with a dramatic change in the heavy mineral suite as evidenced by sample 99GR52, which contains a chloritoid-dominated group and less than 25% ZTR ( Fig. 7 ). The long lag time implies that the new metamorphic source terrane experienced an early exhumation at the end of the Palaeozoic but was eroded only during the Eocene. From Middle to Late Eocene, a major type 5 path defines the change in lag time from 184 Ma to approximately zero ( Fig. 3 ). This is recorded in both the D 1 and D 2 paths. During the Middle to Late Eocene, rates of exhumation in the source region were probably >1 mm ·a −1 , assessed from the approximately zero lag time in 99GR44 ( Table 3 ). The two sub-parallel D 1 and D 2 curves suggest that two different hinterland blocks were rapidly and contemporaneously denuded at this time. The continuous period of exhumation is supported by the appearance of higher grade metamorphic minerals, e.g., kyanite and sillimanite (99GR44) in increasing proportions ( Figs. 7 and 8 ), sourced from deeper crustal levels. Two main events are revealed in the evolution of the source terranes during deposition of the Tiyuyacu Fm.: (1) during the Early Eocene, a change of source terrane occurred, caused most probably by a radical tectonic rearrangement of the hinterland; (2) higher in the section, during the Middle-Late Eocene (44–36 Ma), exhumation of high-grade metamorphic terranes was dominant. We interpret this as a response to Middle to Late Eocene compression and unroofing of the metamorphic core in the Cordillera Real. The Oligocene detrital curves (Orteguaza and Chalcana Fms) begin with a type 1 path and represent a further change in source. The heavy mineral input changed from ZTR and high-grade index kyanite and sillimanite (99GR44) to one with dominantly medium-grade metamorphic suites, such as garnets, epidote and chloritoid (99GR81). The subsequent reappearance of kyanite and sillimanite-containing assemblages in 00GR02 and 00GR03 ( Fig. 7 ) parallels a clear shortening of lag times and is indicative of the re-activation of high-grade metamorphic rock source regions ( Fig. 8 ). A second period of volcanic activity is recorded in the Early Miocene (00GR04; Table 2 ). Pseudo-hexagonal biotites, hornblende, euhedral, inclusion-rich apatites and idiomorphic zircons in the basal Arajuno Fm., combined with indistinguishable ZFT and AFT ages of 23 Ma ( Ruiz, 2002 ), suggest supply from contemporaneous volcanics. Because of the volcanic components, the D 1 and D 2 paths, spanning the upper Chalcana and lower Arajuno Fms., depict two type 2 paths. These contain identical ZFT populations in both levels that are attributed to (1) the presence of similar zircon assemblages in both formations ( Fig. 7 ), and (2) the unconformable contact between the Chalcana and Arajuno Fms. This supports the interpretation of Christophoul et al. (2002a) , who suggested reworking of the Chalcana Fm. into the Arajuno Fm. The pattern of the detrital zircon ages for the Miocene to Pliocene Chambira and Mesa/Mera Fms and those from the modern Napo River sand follow a simple type 4 path for D 1 with constant lag time values of ∼30–40 Ma ( Fig. 8 ; Table 2 ). This lag time persisted, despite the dramatic change in the source region during the Pliocene, as evidenced by the flux of mafic minerals from the Cordillera Occidental in the Mesa/Mera Fms, and by the dominance of medium-grade metamorphic grains from the Cordillera Real in the modern Napo River sand ( Fig. 7 ). The constant lag time suggests that the entire hinterland had already been brought through the partial zircon-annealing zone by Oligocene times. Smaller, young events, recorded by AFT analysis at 15 Ma and younger ( Spikings et al., 2001 ), did not achieve sufficient erosion to bring lithologies with associated zircon ages to the surface. The young events included (1) thrust propagation involving the Napo and Cutucú Uplift areas, causing Neogene sedimentation to become localised in the Pastaza Depression and in front to the east of the uplifted areas, (2) enhanced exhumation in the northern Cordillera Real since 15 Ma ( Spikings et al., 2001 ), and (3) the formation of the Inter-Andean Depression since 6–5 Ma ( Winkler et al., 2004 ). 6 Conclusions Integrated heavy mineral and detrital ZFT analysis has proved instrumental in the reconstruction of the evolution of the mid-Cretaceous to Recent retro-arc AAB foreland basin and adjacent source terranes ( Fig. 8 ). Major events identified are: (1) Aptian to Turonian sediment supply to the nascent marginal and shallow marine AAB from both the Amazon Craton to the east and the primordial Cordillera Real to the west. The latter source area was established during the preceding Peltetec event that also caused deep erosion of the volcanic basement (Misahualli Fm.) before initiation of the AAB. Exhumation in the primordial Cordillera Real occurred at a rate of 0.1–1 mm ·a −1 , which contrasts with the exhumation rates of at least one order lower in the Amazon Craton. (2) Exhumation rates increased in the Cordillera Real from Turonian to Paleocene (⩾1 mm ·a −1 ), but supply continued from the Amazon Craton until the Maastrichtian. Turonian and younger inversion caused transpressive deformation in the Oriente Basin. Rapid exhumation and unroofing of the Cordillera Real is indicated by the first appearance of metamorphic minerals in the basin since the Maastrichtian, simultaneously with the termination of supply from the Amazon Craton. This development was related to the accretion of the oceanic plateau Pallatanga Terrane along the Ecuadorian forearc during the Maastrichtian-Palaeocene ( Luzieux et al., 2006 ; Vallejo et al., 2006 ), dated thermochronologically from the Cordillera Real at approximately 75–60 Ma ( Spikings et al., 2001 ), indicating that the rapid exhumation of the Cordillera Real was driven by the collision of the oceanic terranes with the forearc. (3) The Eocene to Oligocene history of detrital supply to the basin is characterised by intervals of repeatedly switching palaeotransport paths carrying material from different source regions and progressive exhumation of deeper metamorphic levels ( Fig. 8 ). We ascribe this to a main phase of orogenic growth by vigorous tectonic uplift and unroofing in the Cordillera Real. (4) Following a pronounced Oligocene exhumation of high-grade metamorphic complexes in the Cordillera Real at 33–25 Ma ( Fig. 7 and 8 ), the Mio-Pliocene detrital path is characterised by approximately constant lag times of some 30–40 Ma (a path 4 iso-lag time), along with frequent changes of source complexes and drainage patterns. The activation of a new Cordillera Occidental source occurred during this time. The relatively long lag time, seen also in the zircons from the modern river sediments, is connected with a regional exhumation during the Eocene-Early Oligocene orogenic growth ( Fig. 8 ). Although younger events have been identified through AFT analysis ( Spikings et al., 2001 ), young zircon ages associated with them are not apparent because they have not yet been exhumed to the surface. Acknowledgements This work was supported by the Swiss Science Foundation Grants No. 21-050844.97 and 20-056794-99. The reviewers John Aspden and Roberto Barragán are thanked for their constructive comments on an earlier version of the manuscript. References Baldock, 1982 Baldock, J.W., 1982. Geología del Ecuador: Boletín de Explicación del Mapa geológico de la Républica del Ecuador. Dirección General de Geología y Minas, Quito, Ecuador, 70pp. Balkwill et al., 1995 H.R. Balkwill G. Rodrigue F.I. Paredes J.P. Almeida Northern part of the Oriente Basin, Ecuador: reflexion seismic expression of structures American Association of Petroleum Geologists Memoir 62 1995 559 571 Barragán et al., 2005 R. Barragán P. Baby R. Duncan Cretaceous alkaline intra-plate magmatism in the Ecuadorian Oriente Basin: geochemical, geochronological and tectonic evidence Earth and Planetary Science Letters 236 2005 670 690 Barragán et al., 2004 R. Barragán F. Christophoul H. White P. Baby M. Rivadeneira F. Ramírez J. Rodas Estratigrafia secuencial del Cretacio de la Cuenca Oriente del Ecuador P. Baby M. Rivadeneira R. Barragan La Cuenca Oriente: Geologia y Petroléo vol. 144 2004 Traveaux de l’Institut Français d’Etudes Andines 45 68 Brandon, 1992 M.T. Brandon Decomposition of fission-track grain-age distributions American Journal of Science 292 1992 535 564 Brandon, 1996 M.T. Brandon Probability density plot for fission-track grain-age samples Radiation Measurements 26 1996 663 676 Brandon, 1998 M.T. Brandon M.K. Roden-Tice J.I. Garver Late Cenozoic exhumation of the Cascadia accretionary wedge in the Olympic Mountains, NW Washington State Geological Society of America Bulletin 110 1998 985 1009 Bristow and Hoffstetter, 1977 Bristow, C.R., Hoffstetter, R., 1977. Ecuador. In: Hoffstetter, R. (Ed.), Lexique Stratigraphique International (2e éd.), vol. 5. Centre national de la recherche scientifique (CNRS), Paris, 410pp. Christophoul et al., 2002a F. Christophoul P. Baby C. Dávila Stratigraphic response to a major tectonic event in a foreland basin: the Ecuadorian Oriente basin from Eocene to Oligocene times Tectonophysics 345 2002 281 298 Christophoul et al., 2002b F. Christophoul P. Baby J.-C. Soula M. Rosero J. Burgos Les ensembles fluviatiles néogènes du bassin subandine d’Equateur et implications dynamiques Comptes Rendus Geoscience 334 2002 1029 1037 Christophoul et al., 2004 Christophoul, F., Burgos, J.D., Baby, P., Soula, J.-C., Bès de Berc, S., Dàvila, C., Rosero, M., Rivadeneira, M., 2004. Dinàmica de la Cuenca de antepaìs Oriente desde el Paleògeno. In: Baby, P., Rivadeneira, M.,Barragán, R. (Eds.), La Cuenca Oriente. Geología y Petróleo. Petroecuador, Quito-Ecuador, 295pp, ISBN 9978-43-859-9. Cosma et al., 1998 L. Cosma H. Lapierre E. Jaillard G. Laubacher D. Bosch A. Desmet M. Mamberti P. Gabriele Pétrographie et géochimie des unités magmatiques de la Cordillère occidentale d’Equateur (0°30′S): implications tectoniques Bulletin de la Société Géologique de France 169 1998 739 751 Dashwood and Abbotts, 1990 M.F. Dashwood I.L. Abbotts Aspects of the petroleum geology of the Oriente Basin, Ecuador J. Brooks Classic Petroleum Provinces vol. 50 1990 Geological Society of London Special Publication 89 117 DeCelles and Giles, 1996 P.G. DeCelles K.A. Giles Foreland basin systems Basin Research 8 1996 105 123 Ego et al., 1996 F. Ego M. Sébrier A. Lavenu H. Yepes A. Egüez Quaternary state of stress in the Northern Andes and the restraining bend model for the Ecuadorian Andes Tectonophysics 259 1996 101 116 Egüez, 1986 Egüez, A., 1986. Evolution Cénozoique de la Cordillère Septentrionale d’Equateur: les minéralisations associées. Unpublished Ph.D. thesis. Université Pierre et Marie Curie, Paris, 116pp. Faucher and Savoyat, 1973 B. Faucher E. Savoyat Esquisse géologique des Andes de l’Equateur Revue de Géographie Physique et de Géologie Dynamique 15 1973 115 142 Galbraith, 1981 Galbraith, R.F., 1981. On statistical models for fission-track counts. Mathematical Geology 13, 471–478; reply, pp. 485–488. Galbraith and Laslett, 1993 R.F. Galbraith G.M. Laslett Statistical models for mixed fission-track ages Nuclear Tracks Radiation Measurements 21 1993 459 470 Garver et al., 1999 J.I. Garver M.T. Brandon M. Roden-Tice P.J.J. Kamp Erosional denudation determined by fission-track ages of detrital apatite and zircon U. Ring M.T. Brandon S. Willett G. Lister Normal Faulting, Ductile Flow, and Erosion, Exhumation Processes vol. 154 1999 Geological Society of London Special Publication 283 304 Green, 1981 P.F. Green A new look at statistics in fission-track dating Nuclear Tracks and Radiation Measures 5 1981 77 86 Harman et al., 1998 R. Harman K. Gallagher R. Brown A. Raza Accelerated denudation and tectonic/geomorphic reactivation of the cratons of northeastern Brazil in the Late Cretaceous Journal of Geophysical Research 103 1998 27091 27105 Hubert, 1962 J.F. Hubert A zircon-tourmaline-rutile maturity index and the interdependence of the composition of heavy mineral assemblages with the gross composition and texture of sandstones Journal of Sedimentary Petrology 32 1962 440 450 Hughes and Pilatasig, 2002 R.A. Hughes L.F. Pilatasig Cretaceous and Tertiary terrane accretion in the Cordillera Occidental of the Andes of Ecuador Tectonophysics 35 2002 29 48 Hungerbühler et al., 2002 D. Hungerbühler M. Steinmann W. Winkler D. Seward A. Egüez D.E. Peterson U. Helg C. Hammer Neogene stratigraphy and Andean geodynamics of southern Ecuador Earth-Science Reviews 57 2002 75 124 Hurford and Green, 1983 A.J. Hurford P.F. Green The zeta calibration of fission-track dating Chemical Geology 1 1983 285 317 Jaillard, 1997 E. Jaillard Sintesis Estratigrafíca del Cretaceo y Paleogeno de la Cuenca Oriental del Ecuador 1997 Convenio ORSTOM-Petroproduccion Quito 164pp Jaillard et al., 1995 E. Jaillard M. Ordoñez S. Benítez G. Berrones N. Jiménez G. Montenegro I. Zambrano Basin development in an accretionary, oceanic-floored forearc setting: southern coastal Ecuador during late Cretaceous to late Eocene times A.J. Tankhard R. Suárez H.J. Welsink Petroleum Basins of South America vol. 62 1995 American Association of Petroleum Geologists Memoir 615 631 Kerr et al., 2002 A.C. Kerr J.A. Aspden J. Tarney L.F. Pilatasig The nature and provenance of accreted oceanic terranes in western Ecuador: geochemical and tectonic constraints Journal of the Geological Society of London 159 2002 577 594 Lapierre et al., 2000 H. Lapierre D. Bosch V. Dupuis M. Polvé R.C. Maury J. Hernandez P. Monié D. Yeghicheyan E. Jaillard M. Tardy B. Mercier De Lépinay M. Mamberti A. Desmet F. Keller F. Sénebier Multiple plume events in the genesis of the peri-Caribbean Cretaceous oceanic plateau province Journal of Geophysical Research 105 2000 8403 8421 Litherland, 1994 M. Litherland J.A. Aspden R.A. Jemielita The metamorphic belts of Ecuador British Geological Survey, Overseas Memoir 11 1994 147 Luzieux et al., 2006 L.D.A. Luzieux F. Heller R. Spikings C.F. Vallejo W. Winkler Origin and Cretaceous tectonic history of the coastal Ecuadorian forearc between 1°S–4°S: paleomagnetic, radiometric and fossil evidence Earth and Planetary Science Letters 249 2006 400 414 Mange and Maurer, 1992 M.A. Mange H.F.W. Maurer Heavy Minerals in Colour 1992 Chapman and Hall London 147pp Mello et al., 1995 M.R. Mello E.A.M. Koutsoukos W.Z. Erazo The Napo Formation, Oriente Basin, Ecuador: hydrocarbon source potential and paleoenvironmental assessment B.J. Katz Petroleum Source Rocks 1995 Springer-Verlag New York 167 181 Rivadeneira and Baby, 1999 M. Rivadeneira P. Baby La Cuenca Oriente: Estilo Tectónico, Etapas de Deformación y Caracteristicas Geológicas de los Principales Campos de Petroproducción 1999 Convenio ORSTOM-Petroproducción Quito 88pp Romeuf et al., 1995 N. Romeuf L. Aguirre P. Soler G. Féraud E. Jaillard G. Ruffet Middle Jurassic volcanism in the Northern and Central Andes Revista Geológica de Chile 22 1995 245 259 Ruiz, 2002 Ruiz, G., 2002. Exhumation of the northern Sub-Andean Zone of Ecuador and its Source Region: A Combined Thermochronological and Heavy Mineral Approach. Unpublished Ph.D. thesis. ETH-Zurich, 260pp ( http://e-collection.ethbib.ethz.ch/ ) Ruiz et al., 2004 G.M.H. Ruiz D. Seward W. Winkler Enhancing detrital radiogenic provenance studies towards understanding geodynamic development of hinterland orogens: an example using zircon fission-track analysis from the northern Ecuadorian Sub-Andean Zone Basin Research 16 2004 413 430 Shanmugam et al., 2000 G. Shanmugam M. Poffenberger J. Toro Alava Tide-dominated estuarine facies in the Hollin and Napo (“T” and “U”) formations (Cretaceous), Sacha Field, Oriente Basin, Ecuador American Association of Petroleum Geologists Bulletin 84 2000 652 682 Spikings et al., 2001 R.A. Spikings W. Winkler D. Seward R. Handler Along-strike variations in the thermal and tectonic response of the continental Ecuadorian Andes to the collision with heterogeneous oceanic crust Earth and Planetary Science Letters 186 2001 57 73 Thomas et al., 1995 G. Thomas A. Lavenu G. Berrones Evolution de la subsidence dans le Nord du bassin de l’Oriente équatorien (Crétacé supérieur à Actuel) Comptes Rendu Géodynamique 320 1995 617 624 Tschopp, 1953 H.J. Tschopp Oil exploration in the Oriente of Ecuador American Association of Petroleum Geologists Bulletin 37 1953 2303 2347 Vallejo et al., 2002 C. Vallejo P.A. Hochuli W. Winkler K. Von Salis Palynological and sequence stratigraphic analysis of the Napo Group in the Pungarayacu 30 well, Sub-Andean Zone, Ecuador Cretaceous Research 23 2002 845 859 Vallejo et al., 2006 C.F. Vallejo R. Spikings L. Luzieux W. Winkler D. Chew L. Page The early interaction between the Caribbean Plateau and the NW South American Plate Terra Nova 18 2006 264 269 Van der Plas, 1962 L. Van der Plas Preliminary note on the granulometric analysis of sedimentary rocks Sedimentology 1 1962 145 157 Weaver, 1963 C.E. Weaver Interpretative value of heavy minerals from bentonites Journal of Sedimentary Petrology 33 1963 343 349 White et al., 1995 H.J. White R.A. Skopec F.A. Ramirez J.A. Rodas G. Bonilla Reservoir characterization of the Hollin and Napo Formations, Western Oriente Basin, Ecuador A.J. Tankard S.R. Suárez H.J. Welsink Petroleum Basins of South America vol. 62 1995 American Association of Petroleum Geologists Memoir 573 596 Willett et al., 1993 S. Willett C. Beaumont Ph. Fullsack Mechanical models for the tectonics of doubly vergent compressional orogens Geology 21 1993 371 374 Winkler et al., 1985 W. Winkler G. Galetti M. Maggetti Bentonite im Gurnigel-, Schlieren- u. Wägital-Flysch: Mineralogie, Chemismus, Herkunft Eclogae geologicae Helveticae 78 1985 545 564 Winkler et al., 2004 W. Winkler D. Villagómez R.A. Spikings P. Abegglen St. Tobler A. Egüez The Chota Basin and its significance for the inception and tectonic setting of the Inter-Andean Depression in Ecuador Journal of South American Earth Sciences 19 2004 5 19 Zambrano et al., 1999 Zambrano, I., Ordoñez, M., Jiménez, N., 1999. Micropaleontología de 63 Muestras de Afloramientos de la Cuenca Oriental Ecuatoriana. Informe Tecnico No. 016-PPG-99, LABOGEO, Petroproducción, Guayaquil, 67pp. Zeitler et al., 1986 P.K. Zeitler N.M. Johnson N.D. Briggs C.W. Naeser Uplift history of the NW Himalaya as recorded by fission-track ages on detrital Siwalik zircons J. Huang Proceedings of the Symposium on Mesozoic and Cenozoic Geology in Connection of the 60th Anniversary of the Geological Society of China 1986 Geological Publishing House Beijing 481 496