Neogene erosion surfaces and the Andean uplift in northern Peru Open Access

McLaughlin (1924) recognized a sequence of erosional surfaces in central Peru, ranging in elevation from c. 3000 to c. 4800 m. These features still form the core of most studies of the geomorphology of the Peruvian Andes. Later studies by Myers (1976) and Wilson (2011, 2015) described features in the flank of the Western Cordillera in northern and central Peru. Wilson (2011, 2015) established a sequence of 20 surfaces and pediments in central Peru, extending from near sea-level to 4800–5000 m. Unpublished work by this author confirms that the same sequence can be recognized on the flank of the Eastern Cordillera in central Peru. The present paper forms part of a regional study of Neogene erosion surfaces in northern Peru and southern Ecuador and deals specifically with the Peruvian sector (Fig. 1). The Huancabamba Andes of NW Peru and southern Ecuador are commonly described as a continuation of the Western Cordillera of Peru or as a transition from the Central Andes to the Northern Andes. Although the Huancabamba Andes have many features in common with the Central Andes, they form a distinct structural entity and will be the subject of a separate report.

This author worked extensively in northern and central Peru as part of the British Geological Mission to Peru in the period 1963–71. That work showed the widespread development of a number of high-level erosional features (Cobbing et al. 1981). Subsequent desk studies involving the detailed examination of 1:100 000 topographic maps have greatly enhanced our understanding of the sequence of surfaces. The author subsequently carried out further fieldwork in the study area, which provided opportunities to confirm the presence of the same sequence of surfaces over large parts of the Andes.

The geological development of the region is underpinned by a number of major palaeotectonic units (Fig. 2), which can be described briefly as follows.

The Olmos Arch/Cordillera Real forms a NNE-trending unit of metamorphic rocks. Litherland et al. (1994) considered the Cordillera Real to be an accreted terrane and it seems reasonable to assume that the Olmos Arch had a similar history. Further accretions of oceanic crust, island arcs and the Amotape–Tahuin terrane led to the progressive westward growth of South America at the latitude of NW Peru and southern Ecuador.

The West Peruvian Trough has an infill of 5–7 km of Upper Jurassic–Cretaceous sediments, which underlie much of the Western Cordillera. The sediments pass westwards into a volcanic sequence intruded by the Coastal Batholith. The northern limit of the West Peruvian Trough coincides with that of the Marañon Geanticline (MGA), such that the basin sequence thins north of c. 7° S and grades into a relatively shallow NNE-trending sag basin lying between the MGA and the Olmos Arch/Cordillera Real and connecting with the Marañon Basin (Fig. 2).

The MGA was initially defined by Benavides-Cáceres (1956) as a major feature controlling the Cretaceous stratigraphy in northern Peru. It was subsequently identified elsewhere in central and northern Peru by Wilson (1963) and Wilson and Reyes Rivera (1964). The term Marañon Arch has since been used to describe that part of the MGA that has been uplifted to form the Eastern Cordillera. The MGA is a more extensive body with a core of Precambrian and Lower Paleozoic metamorphic rocks that was accreted to Gondwana in the Ordovician (Ramos 2018). It occupies the Eastern Cordillera and the adjacent part of the Western Cordillera (Fig. 2). Both the eastern and western boundaries are major faults. The eastern feature is the Subandean Fault Zone, separating the MGA from the folded Cenozoic and older sediments of the Subandean Belt. The western boundary is a hinge line that separates the MGA from the West Peruvian Trough (Wilson 1963). The hinge line probably corresponds to the Incaic Megafault (Benavides-Cáceres 1999), which controlled the emplacement and subsequent uplift of the Cordillera Blanca Batholith in northern Peru and extends as far north as Cajamarca. The hinge line is commonly obscured by basin sediments that were thrust onto the flank of the MGA in the Eocene, but generally lies near to the crest of the Western Cordillera. The MGA can be traced as far as c. 6.5° S, where it plunges abruptly to the NW (Fig. 2).

The influence of the MGA on the stratigraphy of the region began as early as the mid-Jurassic, when the Sarayaquillo red beds were being deposited in a basin bounded by the MGA and the Olmos Arch/Cordillera Real (Fig. 2). The influence of the MGA continued into the Cenozoic, as demonstrated by the fact that the infill of the Marañon Basin pinches out westwards and southwards against the MGA.

The orogenic history of the region includes a number of major events, beginning with the Peruvian phase (Senonian) of the Andean Orogeny, involving the Western Cordillera, which spread red bed clastics to the east. The main feature, however, was the Incaic Orogeny (Early Eocene), which led to the inversion of the West Peruvian Trough and thrusting of the basin sediments onto the flank of the MGA. Erosion of the deformed Mesozoic sediments was followed by a widespread volcanic phase in the Late Eocene–Miocene, which came to an end with the Quechua I Orogeny. The latter produced mainly gentle folding in the study area and was followed by a phase of erosion that led to the widespread development of a peneplain known as the Puna Surface. The uplift of this feature initiated the Neogene rise of the Andes.

The Peruvian cordilleras form part of the Central Andes, characterized by more or less extensive areas at >4000 m and maximum elevations of c. 4800 m, together with abundant magmatic activity in the Western Cordillera. The southern sector of the study area shows these characteristics and is thus considered to be part of the Central Andes, with a boundary as shown in Figure 3. The two cordilleras continue north of c. 6° S, although at a reduced height, and show a progressive swing to a more northerly strike. Neogene magmatic activity is absent north of that latitude in the study area. On this basis, the sector of the study area lying north of c. 6° S is regarded as part of the Northern Andes. Baby et al. (2018) similarly considered that the northern limit of the Central Andes in the Subandean Belt occurs where the NW–SE strike of the Huallaga Basin swings to a more northerly trend in the Santiago Basin (Fig. 3).

The Huancabamba Andes, named by Mourier et al. (1988) and lying immediately west of the study area, have commonly been regarded as transitional between the Central Andes and the Northern Andes. In fact, as the Huancabamba Andes and the Northern Andes are both more or less north-striking features dominated by the Olmos Arch/Cordillera Real uplift, the author suggests that the Huancabamba Andes are included in the Northern Andes, as indicated in Figure 3. It may be noted that Chew et al. (2008) also reached a similar conclusion. The individual morphotectonic components can be briefly described as follows.

This range consists mainly of Jurassic–Cretaceous sediments that were folded and thrust during the Incaic Orogeny of the Early Eocene and unconformably overlain by Cenozoic volcanic formations. Regionally, the Western Cordillera contains extensive areas >4000 m, with significant surfaces at up to c. 4800 m. Within the study area, the Central Andes contains only a few ridges and peaks at c. 4700 m, although there are more substantial remnants at 3800–4000 m. The Central Andes sector of the Western Cordillera is characterized by abundant mineral deposits, particularly associated with the Miocene metallogenic belt (Noble and McKee 1999).

The high-level surfaces and the Miocene mineralization are restricted to the Central Andes. Although the Western Cordillera continues north of c. 6° S, as far as its termination against the Olmos Arch, there is an abrupt decrease in elevation north of Hualgayoc, where, as explained earlier, that part of the cordillera is considered to be part of the Northern Andes. The Miocene mineralization also ends abruptly in the same area. It may be noted, however, that both the high-level surfaces and the Miocene mineralization are, in effect, displaced c. 50 km to the SW. Work initiated in that area (Wilson, unpublished data) indicates scattered remnants of a surface at c. 3800 m along the Olmos Arch, plus a number of Miocene mineral deposits.

This range has a core of Precambrian–Lower Paleozoic metamorphic rocks corresponding to part of the MGA, with faulted inliers of Upper Paleozoic and Mesozoic formations. It is separated from the Western Cordillera by a NW–SE-trending fault system following the Marañon valley (Fig. 3) as far as the Bagua Basin, where it joins the north–south-trending fault system that forms the limit of the Olmos Arch. The eastern boundary of the Eastern Cordillera is a major NW–SE-striking fault system that juxtaposes the metamorphic complex of the Eastern Cordillera against the Mesozoic and Cenozoic sediments of the Subandean Belt.

The Eastern Cordillera has a similar topographic profile to the Western Cordillera, with remnants of surfaces at up to c. 4700 m in central Peru, but rarely exceeding c. 4400 m in the study area. Like the Western Cordillera, the Eastern Cordillera also declines in elevation towards the NW, with an average elevation of c. 2000–2500 m in the sector denominated as part of the Northern Andes, with only isolated ridges reaching c. 3500 m.

This unit lies largely outside the study area and the reader is referred to Baby et al. (2018) for details of its stratigraphy and structure. The Subandean Belt consists of Mesozoic sediments overlain by thick Cenozoic formations, the whole being thrown into a series of parallel folds and thrusts controlled by an underlying Triassic evaporite sequence.

Earlier studies (Wilson 2011, 2015) established that the Andean flanks have stepped profiles in central and northern Peru, which can be observed in the crests of the divides separating valleys. The ‘steps’ represent remnants of old erosion surfaces or pediments and are separated by the back-scarps of the individual features. The methodology involved the careful examination of modern topographic maps published by the Instituto Geografico Militar of Peru at a scale of 1:100 000 with the objective of finding sub-horizontal benches corresponding to the erosional features. The data, including elevations, were transferred to a base map that was then used to establish the distribution of the individual features. Although the topographic maps are not available for publication, the methodology can be shown on the available digital data, of which two examples are included in Figures 4 and 5a.

Figure 4 illustrates part of the Andean flank within the study area c. 100 km SSW of Cajamarca in the Otuzco Quad. As a result of intense erosion, the area does not have any sizeable remnants of the older surfaces, although ridges at c. 3800–4000 and c. 4200 m indicate that the La Oroya and Junin surfaces were developed there. The value of Figure 4 lies in illustrating pediments in the range 800–2600 m, fragments of which are preserved in the lower ridge tops. As is commonly the case, subsequent erosion means that most of the ridges do not display complete sequences of pediments. However, it is still possible to construct composite profiles by combining the data from all the ridge tops. The pediments recognized in the area of Figure 4 occur in the following ranges: c. 800 m, 900–1000 m, 1100–1200 m, 1300–1400 m, 1500–1600 m, 1700–1800 m, 1900–2000 m, 2100–2200 m, 2300–2400 m and 2500–2600 m. As revealed in Figure 4, those pediments are separated by distinct scarps. As the focus of the author's interest has been the early history of the Andean uplift, attention has been focused on the higher surfaces rather than the pediments, which nevertheless will be important elements in developing a more comprehensive view of the Andean uplift.

Figure 5a shows the topography of the Pelagatos Range and surrounding area at c. 8° S, in the Pallasca Quad. In the western part of the area it is clear that the ridge tops fall into natural groups at elevations of c. 3800–4000, c. 4100–4300, c. 4400–4500 and c. 4600–4700 m. Moreover, the individual groups are separated by scarps. The scarps between the feature at c. 4600–4700 m and the lower surfaces are very clear. The scarps between the c. 4100–4300 m and c. 4400–4500 m features are less dramatic, but can nevertheless be clearly recognized in the north-central part of the area (Fig. 5a). The SW corner of Figure 5a, immediately north of Conchucos, also clearly displays the separation between the c. 3800–4000, c. 4100–4300 and c. 4400–4500 m surfaces.

The situation is more complex in the eastern sector of Figure 5a, where the River Marañon and its tributaries have destroyed large areas of the surfaces. In these circumstances, it is necessary to carefully examine the divides between the tributaries to recognize erosional benches separated by scarps, which represent fragmentary remnants of erosion surfaces and pediments. It can be seen that there are groups of sub-horizontal features at c. 2900–3000, c. 3100–3300 and c. 3500–3600 m, separated by scarps. These features represent surfaces found throughout the cordilleras of central and northern Peru.

A total of 20 erosional features can be recognized in central and north-central Peru, ranging in elevation from c. 200 to c. 4800 m (Wilson 2011, 2015). It must be noted, however, that it is fairly common for individual profiles to lack at least one of the erosional features found in the complete sequence. This is a result of later erosion and particularly affects the pediments rather than the higher surfaces.

As this study is focused on the early history of the Andean Uplift, particular attention is paid to the higher surfaces found at elevations of 2800 m or higher. Although all the pediments found below that elevation are not described in this paper, it is important to realize that they can be recognized regionally in central and northern Peru, in both flanks of the Andes. They are important in establishing the episodic nature of the Andean uplift and also the apparently widespread correlation of the individual units' individual features. The profiles included in Figure 6 therefore include all the erosional features recognized in the specific areas.

Profiles B–B′ (c. 12° S) and A–A′ (7° S) in Figure 6 are representative profiles for the flank of the Western Cordillera in central and northern Peru, respectively, and demonstrate that the same sequence of erosional features can be recognized in both areas. Profiles from c. 10° S (Wilson 2015) reveal the same pattern, which is also confirmed by the author's unpublished studies from the Andean flank between 10 and 7° S. The fact that the same sequence of surfaces can be found for hundreds of kilometres along the strike of the Western Cordillera is interpreted as indicating that the individual features originally formed as essentially continuous surfaces.

Profile C–C′ (Fig. 6) illustrates a further profile of the Western Cordillera at c. 7° S within the study area. Profile D–D′ is from the Eastern Cordillera at c. 6° S. It is clear that the same sequence of erosional surfaces occurs in both the Western Cordillera and the Eastern Cordillera. Unpublished profiles from the Eastern Cordillera in central and northern Peru confirm the presence of a sequence of features along the length of that cordillera. The flanks of the two ranges are thus effectively mirror images of one another.

It is assumed that the features at specific elevations on either flank can be correlated in time, for the following reasons. First, in scattered areas at c. 7.5° S in the Eastern Cordillera, the surface at c. 4000 m extends across the crest of the range and appears to have formed by the coalescence of surfaces advancing from both the east and west. Also, the same feature is found on both sides of the Marañon valley just south of 7° S, giving the impression that they once formed a continuous feature from the Western Cordillera to the Eastern Cordillera before the incision of the River Marañon.

Second, the surfaces at c. 3000 and c. 3200 m can be followed from the flank of the Eastern Cordillera at c. 6.5° S westwards across the Utcubabmba valley and the River Marañon, then into the Western Cordillera. Although there is not a continuous path, the gaps are not great and are obviously the result of later erosion, such that it is apparent that we are dealing with features that were originally of uninterrupted regional extent. As these surfaces can be correlated along and across the cordilleras, it is assumed that the other features, both higher and lower, can also be correlated and that the surfaces at similar elevations can be regarded as time equivalents.

The longitudinal profiles (Fig. 7) show the same step-like pattern at higher elevations as can be observed in the cross-strike profiles. These longitudinal profiles are important in demonstrating that the general northwards decline in the elevation of the Andes is not due to a slight tilt of the highest feature (the Puna Surface), but rather to the progressive northward loss by erosion of the highest surfaces, which are widespread further south.

On the basis of the close similarity between the profiles in both central and northern Peru, it is concluded that the features at specific elevations in both areas can be correlated over considerable distances. Accordingly, it is suggested that the nomenclature established in central Peru is also applicable in northern Peru, with the high-level surfaces and their elevations being as follows: Puna Surface, 4700–4800 m; Punrun Surface, 4400–4600 m; Junin Surface, 4100–4300 m; La Oroya Surface, 3800–4000 m; Jauja Surface, 3400–3700 m; Upper Chacra Surface, 3100–3300 m; and Lower Chacra Surface, 2800–3000 m. Although the ages of these features are not well constrained, they appear to lie in the range c. 12 Ma (Lower Chacra Surface) to c. 18–17 Ma (Puna Surface). The subject of ages is addressed more fully below.

Studies in central Peru (Wilson 2011, 2015) have suggested that the Andes were uplifted uniformly as a block, with little or no tilting of the flanks. The case for this interpretation is strong for the eastern edge of the Andean Block, where high-level surfaces reach close to the boundary fault formed by the Subandean Fault System without showing any sign of tilting. A particularly compelling example is found in the Panao Quad, east of Huanuco, at latitude 10° S, where the La Oroya Surface extends to within c. 8 km of the Subandean Fault System without any discernible tilting, even though the sedimentary formations on the eastern side of the fault system are strongly deformed.

The situation at the western boundary of the Andean Block is less clear. In the study area, fragments of the La Oroya Surface reach to within c. 50 km of the coast and commercially acquired data indicate that a fault system lies just offshore. Also, in central Peru, pediments as high as c. 1600 m extend close to the coast, implying uplift as a result of fault activity. By contrast, the western flank in southern Peru has been uplifted mainly by tilting. It is not possible to establish the relative importance of tilting and faulting in the Andean flank of the study area.

Although the Puna and Punrun surfaces are well preserved as far north as latitude 9° S, they occur in the study area only as isolated and generally small remnants.

Just south of 8° S in the Western Cordillera is found the Pelagatos Range, consisting of folded and thrust Cretaceous limestones. The crest of the range forms a ridge at 4700–4900 m (Fig. 5a). There is no sign of neotectonic activity in the area. The summit is flanked by concordant ridge tops at 4400–4600 m. The situation is similar to that observed in central Peru, where the Puna and Punrun surfaces are found at 4600–4800 m and 4400–4600 m, respectively. The Pelagatos features are therefore considered to be remnants of those surfaces. Further north, just south of latitude 7.5° S and 13 km NE of Cajabamba, Cerro Rumi Rumi is formed by a single ridge rising to c. 4500 m, flanked by ridge tops at c. 4200 m, which, in turn, are separated by scarps from ridges at c. 3900 m (Fig. 8). The relations between this group of concordant ridges and their elevations suggest that they represent remnants of the Punrun, Junin and La Oroya surfaces, respectively.

A scattering of peaks rising above the general level of the Eastern Cordillera in the study area are interpreted as remnants of the Puna and Punrun surfaces. First, just south of 8° S in the Tayabamba Quad, in an area characterized by ridge tops at c. 4100–4300 m, two peaks reaching 4600 m occur (Fig. 9). The latter are remnants of a surface that can be traced southwards along the Eastern Cordillera into central Peru, where it forms the Junin Surface (Wilson 2015). The features at c. 4400–4600 m are thus probably small remnants of the Punrun and Puna surfaces, which, in the Eastern Cordillera of central Peru, occur at 4400–4500 m and 4600–4800 m, respectively. Second, just south of 6° S, near the town of Bolivar, is an area characterized by ridge tops mainly at 3800–4000 m and 4100–4300 m, but which are surmounted by ridges at 4400 m and a peak at c. 4700 m (Fig. 10). The situation is similar to that outlined for the Tayabamba area and the features are therefore correlated with the Puna, Punrun and Junin surfaces, respectively.

This surface was originally recognized by McLaughlin (1924) in the neighbourhood of Lake Junin in central Peru and was further described by Wilson (2011) as a mature feature at an average elevation of c. 4200 m, separated by scarps from the Punrun and La Oroya surfaces. A surface with these same characteristics is found in both the Western Cordillera and Eastern Cordillera within the study area and can be confidently correlated with the Junin Surface.

The most extensive areas of the Junin Surface in the Western Cordillera are found in the southern sector of the study area, where they occur as two belts separated by the higher terrains forming the crest of the cordillera in this area (e.g. the Pelagatos Range). To the east of this topographic axis the Junin Surface forms a dissected plateau c. 20 km wide, which extends to the canyon of the River Marañon (Fig. 11). The feature can be followed southwards beyond the limit of the study area, where it forms a dissected plateau on the eastern side of the Cordillera Blanca.

Another belt of the Junin Surface is found west of the topographic axis of the Western Cordillera and is preserved in the area south of 8° S (Santiago de Chuco Quad) as a plateau c. 20 km wide at an elevation of 4100–4300 m (Fig. 12). Further north is a well-preserved remnant of the Junin Surface in the area east of Cajabamba, where it occurs as multiple concordant ridges at c. 4200 m, separated by scarps from both the Punrun and La Oroya surfaces (Fig. 8).

The most northerly examples of the Junin Surface in the Western Cordillera are found at c. 7° S, in the area between Cajamarca and Celendin, where the La Oroya Surface, widespread in this area, is surmounted by isolated groups of ridge tops at c. 4100 m, which are ascribed to the Junin Surface.

The northernmost remnants of the Junin Surface in the Eastern Cordillera occur, as in the Western Cordillera, at c. 7° S (Fig. 10), where features ascribed to the Puna and Punrun surfaces are separated by scarps from a group of concordant ridge tops at c. 4200 m, interpreted as remnants of the Junin Surface. The 4200 m feature continues northwards into the headwaters of the River Utcubabmba (Fig. 13). The main area of the Junin Surface in the Eastern Cordillera occurs southwards from c. 7.5° S, where it comprises a plateau-like feature at c. 4200 m, which forms the Continental Divide in this area (Fig. 9). The feature is separated by scarps from higher surfaces found just south of 8° S (Fig. 9) and from the La Oroya Surface to the east. Reconnaissance data show that the surface continues along the Eastern Cordillera southwards from the study area, maintaining a constant elevation in the range 4100–4300 m. It extends into central Peru and is extensively preserved in the flanks of the Eastern Cordillera and in the type area around Lake Junin.

The La Oroya Surface was described by Wilson (2011) as a mature surface lying at 3800–4000 m in both the Western Cordillera and Eastern Cordillera of central Peru. The current study has shown it to be well represented in northern Peru.

The La Oroya Surface can be found throughout the Western Cordillera as far north as c. 6.5° S (Fig. 12), typically occurring as concordant ridges or more extensive areas of mature erosion surface at c. 3800–4000 m. It is separated by scarps from the Jauja Surface and is surmounted by features corresponding to the Junin Suface. The best examples of the La Oroya Surface in the Western Cordillera of northern Peru include: (a) the area between the Chusgon and Marañon drainages at c. 7.5–8° S (Fig. 6); (b) the area between the Cajamarca–Crisnejas and Marañon drainages at c. 7.5–7° S (Fig. 12); and (c) the Yanacocha–Hualgayoc area lying immediately north of the city of Cajamarca (Fig. 14). Smaller areas of concordant summits at c. 3800–4000 m are found distributed along the Pacific flank of the Western Cordillera (Fig. 12).

The La Oroya Surface in areas (a) and (b) is typically represented by gently undulating terrain at c. 3900 m, which can be envisaged as part of a regionally extensive surface subsequently uplifted to its current altitude without undergoing significant internal deformation. The northern part of area (a) conforms to that pattern and, although NW–SE-striking faults have been mapped (Longo 2005), they are bevelled by the La Oroya Surface, which they clearly predate. However, in the southern sector, including the Yanacocha epithermal gold deposit, the La Oroya Surface has been subjected to faulting and flexuring. Detailed studies by Longo (2005) show that the Yanacocha Volcanic Complex was erupted onto a planar feature recognized by this author as the La Oroya Surface and that there was faulting during and immediately after the volcanic activity. The faults are mainly on a NW–SE strike and led to the formation of features such as the La Quinua Graben (Longo 2005).

Data included in Longo (1997) indicate that structural warping has also affected the western and southern flanks of the Yanacocha area. Thus, in the western sector of Yanacocha, the Lower Andesite Unit of the Yanacocha Volcanic Complex is found mantling a surface at 3500–3600 m, which the author considers to be a down-warped portion of the La Oroya Surface. In the southern sector, the main body of the San Jose ignimbrite occupies what Longo (1997) interpreted as an erosional feature named the Otuzco Trough. In fact, the feature is a portion of the La Oroya Surface that has been tilted down to the south and affected by later erosion before being mantled by the ignimbrite. The nature of the down-dip termination of the monocline remains to be established. However, it is possible that part of the Cajamarca valley may be a graben in which the La Oroya Surface has been modified by later phases of erosion to produce the current distribution of surfaces. By contrast, the magmatic activity in Hualgayoc and the surrounding area, 20–30 km north of Yanacocha, has not resulted in any obvious structural effects on the La Oroya Surface.

The northern limit of the La Oroya Surface in the Western Cordillera is formed by an abrupt NE–SW-trending feature, the back-scarp of the Jauja Surface (Fig. 14), which gives way northwards to remnants of the Chacra surfaces. The La Oroya Surface presumably at one time extended further north than currently, but the fact that no outlier has been found suggests that it had only a modest northwards extension. The implications of the fact that the original northern limit of the Puna, Punrun and Junin surfaces coincides with that of the La Oroya Surface, and that the Miocene metallogenic belt terminates in the same area, are considered in the Discussion.

The most northerly remnant of the La Oroya Surface in the Eastern Cordillera is found at c. 6° S. The main body of the surface is found south of 6.5° S and continues to the limit of the study area (Fig. 12). It occurs as a gently rolling, mature surface lying in the range 3800–4000 m and is separated by scarps from the Jauja and Junin surfaces. The relationships between the surfaces can be appreciated by reference to Figure 13.

Between 7° S and 7.5° S the La Oroya Surface straddles the crest of the Eastern Cordillera, separating the Huallaga and Marañon drainages (Fig. 12). It appears probable that the surface developed as a result of erosion from both the east and west, the resulting features coalescing at the crest of the range, eliminating the Junin and higher surfaces. It may be noted that there is a degree of asymmetry in the Eastern Cordillera, with the higher ground commonly being found to the west of the drainage divide. Also, the eastern flank of the cordillera has much better developed areas of the Jauja and Chacra surfaces than the western flank, which is characterized by relatively short, steeply graded tributaries of the River Marañon that have destroyed much of the pre-existing surfaces. The asymmetry is presumably the result of the much greater rainfall over the eastern flank of the cordillera.

Small remnants of the La Oroya Surface are found on the higher flanks of the Marañon valley, backed by higher ground (Fig. 5b). The author interprets the situation as indicating that the La Oroya Surface was formed by a valley separating the two cordilleras and that the proto-Marañon drained into the area occupied by the Bagua Basin.

The Jauja Surface was named by Wilson (2011) as a mature erosional feature found at c. 3400–3700 m in central Peru. It has been recognized throughout the study area, displaying the same characteristics as observed in central Peru.

The Jauja Surface is fairly well preserved in the Western Cordillera south of c. 6.5° S (Fig. 12). Good examples can be observed in the southern part of the study area at c. 8° S, where the surface has a width of c. 40 km and reaches to within c. 50 km of the coastline. The surface is separated by scarps from the La Oroya and Chacra surfaces. Extensive remnants of the surface are also recognized further north at c. 6.5–7.5° S, where it forms a perimeter around the La Oroya Surface (Fig. 6). Small remnants of the latter are found rising above the Jauja Surface, indicating that the latter grew by erosion and destruction of the La Oroya Surface.

The surface forms the Continental Divide in the area south of Cajamarca, where a ridge at c. 3600 m separates the River Cajamarca and River Jequetepeque drainages (Fig. 12). The divide is usually higher in this part of northern Peru. In fact, the divide between the River Cajamarca and the River Marañon lies consistently above 3800 m. The history of drainage development in this area would make an interesting study, with the possibility of the River Cajamarca originally draining to the Pacific.

The northern limit of significant areas of the Jauja Surface occurs at c. 6.5° S. From that latitude northwards there is only a small remnant of the Jauja Surface, lying 10–25 km north of Cutervo (Fig. 12), where it forms a ridge at 3400–3500 m separated by scarps from the Chacra surfaces. Nevertheless, the presence of that remnant indicates that the Jauja Surface possibly extended at least to c. 6° S prior to the erosion of the Chacra surfaces.

It has already been established that the Marañon valley existed at the time of erosion of the La Oroya Surface and scattered remnants on the shoulders of the valley confirm that it continued to act as the main component of the drainage of northern Peru in Jauja Surface time.

The Jauja Surface is well preserved in various parts of the Eastern Cordillera and in the Cordillera del Tigre that separates the Bagua Basin from the Santiago Basin. The surface is particularly well represented at c. 6° S, where it occurs in the ridge tops on the flanks of the Upper Utcubamba valley (Fig. 13). The ridges lie at c. 3400–3700 m and extend for up to 30 km. This area also confirms the relationships between the Jauja Surface and other features. Thus clearly defined scarps separate it from the La Oroya and Junin surfaces at latitude 6.45° S, whereas clear scarps also act as boundaries with the Chacra surfaces at c. 6.5° S. The Jauja Surface is also preserved in ridge tops on the eastern flank of the Eastern Cordillera at c. 7–8° S, where the relationships with other surfaces can also be confirmed.

The occurrence of the Jauja Surface in the mountainous belt separating the Bagua and Santiago basins is particularly interesting. Thus the crest of the Cordillera del Tigre, forming the eastern boundary of the Bagua Basin, is formed by sinuous ridge tops at c. 3500 m that have the characteristics of the Jauja Surface (Fig. 15). Moreno et al. (2020) considered the surface forming the summits of the Cordillera del Tigre to be the equivalent of the unconformity separating the San Antonio fanglomerates from the underlying Sambimera Formation, which they dated at c. 13 Ma. Part of this unconformity was uplifted to become the Cordillera del Tigre, which provided some of the clastic infill corresponding to the San Antonio Formation (Moreno et al. 2020). This phase was ended by gentle folding of the basin sediments at c. 8 Ma.

The same unconformity is found in the western part of the Bagua Basin, where the Miocene fanglomerates lie directly on Cretaceous formations. The western boundary of the basin is formed by the Olmos Arch, where features resembling the Jauja Surface have been identified by the author. It therefore appears that the latter area presents a mirror image of the Cordillera del Tigre, which forms the eastern limit of the basin.

It seems likely that, prior to the progressive uplift of the basin and its margins, the San Antonio Formation and/or the Paleogene sediments may have extended significantly beyond their present limits. It is therefore difficult to determine which parts of the Jauja Surface have always been a subaerial feature and which parts have been exhumed after burial beneath Miocene sediments.

The Chacra surfaces were originally recognized by McLaughlin (1924) and were collectively described as the Valley Surface by later authors (Wilson and Reyes Rivera 1964). Wilson (2011) subsequently divided the feature into the Upper and Lower Chacra surfaces at elevations of c. 3100 and c. 2900 m, respectively.

The Chacra surfaces in the Western Cordillera are found as a fringe around the perimeter of the Jauja Surface and as remnants in the crests of the divides separating the various rivers draining to the Pacific. The remnants currently reach to within 40–50 km of the coast, but no doubt had a greater extent at the time of their formation. The surfaces are also well represented in the valley system extending from Cajamarca to Huamachuco (Fig. 12) and in the area surrounding Celendin, where they are commonly covered by a thin sheet of the San Jose ignimbrite.

At the northern end of the Cajamarca valley, the Chacra surfaces are buried by the San Jose ignimbrite. The unconformity forming the base of the ignimbrite lies on the north bank of the River Chonta, a tributary to the River Cajamarca, and lies at elevations in the range c. 2800–3200 m for c. 10 km. The unconformity is considered to represent the buried portions of the Chacra surfaces. Abrupt changes in the elevation of the unconformity from c. 3000 to c. 3200 m and from c. 3200 to c. 3400 m are interpreted as the back-scarps of the Lower and Upper Chacra surfaces, respectively.

The Chacra surfaces are also well represented in the Eastern Cordillera, where they occur on both sides of the summit area and comprise fairly extensive remnants in the area north of 7° S (Fig. 12). At about that latitude the surfaces lie within the Utcubamba valley (Fig. 13) and clearly show that this valley, like the Marañon, was well established by Chacra time.

The Chacra surfaces are also found on the flanks of the Cordillera del Tigre. The Chacra surfaces and the pediments that succeed them chronicle the episodic uplift of the Cordillera del Tigre. A particularly clear example occurs on the southern flank of the Bagua Basin, where the periclinal nose of a large anticline developed in Cretaceous sediments shows small remnants of those pediments (Fig. 15) It appears that the anticline rose as a result of episodic uplift during the Neogene, with the pediments being cut in the intervening tectonically neutral periods. It seems probable that the anticline was originally mantled by Neogene sediments, which were later stripped off during the course of the uplift. The pediments identified in this example match the sequence established by the author in central and northern Peru (Wilson 2015) and further demonstrate the geographical extent of the episodic uplift phenomenon.

Although the pediments are not dealt with in this paper, they are nevertheless an important feature in the growth of the Andean Cordilleras throughout central and northern Peru in so far as they confirm that episodic uplift continued after the development of the high-level erosion surfaces. In addition, they indicate that once uplift was initiated in the area north of c. 6° S, it proceeded at the same rhythm as in the area to the south, as indicated by the following characteristics.

The flank of the Cordillera del Tigre and the southern margin of the Bagua Basin are etched by generally small remnants of pediments in the range c. 1200–2600 m, which punctuated the gradual rise of those areas (Fig. 15). The Andean flank in the Otuzco area (Fig. 4) shows a very similar sequence of pediments in the range c. 800–2600 m. In both cases the pediments are surmounted by the Chacra surfaces, indicating a general correlation between the two areas. Thus, although uplift was delayed north of c. 6° S, once initiated it led to the development of the same sequence of pediments as in the Central Andes. It also appears that the rate of uplift was the same in both regions, the difference being in the date at which uplift was initiated.

Although the identification and mapping of the individual erosion surfaces in the Peruvian Andes is relatively straightforward, the establishment of their ages is much more problematic for the following reasons: there has been no regional study aimed specifically at defining the ages of the surfaces; much of the available data have been acquired in an essentially ad hoc manner without reference to the erosion surfaces; and dating has been carried out over several decades using different methods and it can be difficult to reconcile some of the data acquired.

The author has made various attempts to estimate the ages of the erosion surfaces in central Peru (Wilson 2011, 2015), but the results have not been satisfactory, partly due to a lack of data and partly due to errors in interpretation. The present study has benefited from the work of Longo (1997) and Moreno et al. (2020), such that, although problems remain, we can now derive ages for the higher surfaces with more confidence. Nevertheless, while the following estimates provide a more accurate representation of the ages of individual surfaces, it is considered prudent to give an age range for each feature rather than a single date.

This study has not revealed any age data specific to the Puna and Punrun surfaces. The Puna Surface was eroded in response to uplift associated with the Quechua I Orogeny, dated at c. 17 Ma (Benavides-Cáceres 1999). In fact, that date is not well constrained, such that the implied age for the surface may also be questioned. Thus it is difficult to accept an age of c. 16 Ma for the Puna Surface if more recently acquired age data on some of the lower surfaces are correct. The La Oroya Surface can be dated at c. 15–14 Ma, implying that the development of the Junin, Punrun and Puna surfaces was achieved in only 1–2 Myr. This is clearly impossible. It seems more reasonable to assume ages of c. 18–17 Ma for the Puna Surface and c. 17–16 Ma for the Punrun Surface. These estimates still need to be used with caution. There is obviously an urgent need to carry out thermochronological studies on well-defined areas of the Puna and Punrun surfaces to assist in resolving the enigma.

The work of Baby et al. (2018) and Eude et al. (2013) in the Eastern Cordillera and Subandean Belt of northern Peru has revealed important new information on the uplift and cooling history of these morphotectonic units, with particular relevance to their early history. Thus Baby et al. (2018) found that the propagation of the thrust wedge began to the east of the Eastern Cordillera in the interval 30–24 Ma, whereas in the interval 24–17 Ma the deformation advanced into the area of the Subandean Belt as a result of the impingement of the Eastern Cordillera on the Huallaga Basin, with one structure having begun to undergo uplift/cooling at c. 23 Ma. Also, Baby et al. (2018) and Eude et al. (2013) established that some of the main structures in the Huallaga Basin were initiated at c. 17–16 Ma. It seems possible that these events may have been associated with the initiation of episodic uplift of the Andean Block, but they do not offer direct data on the age of the Puna Surface.

Any erosion surfaces developed as a result of early movements would have been largely destroyed by the Quechua I Orogeny and subsequent erosion of the Puna Surface. However, the stratigraphy of the Bagua Basin may shed some light on the possible erosion of the Eastern Cordillera at c. 24–23 Ma. Thus Moreno et al. (2020) described the Sambimera Formation of the Bagua Basin as having a lower unit of mainly mudstones and an upper unit of sandstones with conglomeratic intercalations. The palaeocurrent data indicate derivation from the east. The age data show a range from c. 31–28 Ma to c. 15–12 Ma for the whole of the Sambimera Formation. It may therefore be speculated that the coarse clastic sediments of the Upper Sambimera Formation began to accumulate at c. 23 Ma and that they were derived from the early uplift of part of the Eastern Cordillera. A comparable situation can be recognized c. 20 km north of Cajamarca, where Noble et al. (1990) described the Chala Sequence as a small area of clastic sediments and intercalated tuffs dated at c. 22 Ma and lying unconformably on Cretaceous formations. It may therefore be speculated that the sediments were derived from the erosion of early uplifts in the Eastern Cordillera.

No new data has been acquired in the course of this study. Probably the best information on the age of the Junin Surface comes from the Cerro de Pasco area of central Peru, where epithermal mineralization is associated with a diatreme, dated at c. 15.4 Ma (Baumgartner et al. 2008), which penetrates an erosion surface at c. 4300 m near the type area for the Junin Surface. A similar situation has been described in the nearby Colquijirca area, where Bendezú et al. (2008) studied a diatreme dated at c. 12.7–12.4 Ma that penetrates the same surface. On this basis, an age of c. 16–15.5 Ma is assigned to the Junin Surface.

Our knowledge of this surface has benefited greatly from the detailed study of the Yanacocha area by Longo (2005). The Lower Andesite Member of the Yanacocha Volcanic Complex, dated at c. 14.3 Ma, lies on a widespread erosion surface at 3800–4000 m, which corresponds to the La Oroya Surface. The latter must therefore have a minimum age of c. 14.5 Ma. Elsewhere in the Yanacocha area the surface bevels the Fraile Formation, dated at c. 15.5 Ma (Longo 2005). These data suggest that the age of the La Oroya Surface lies in the range 15–14.5 Ma.

However, data from the Hualgayoc area, c. 30 km north of Yanacocha, suggest a younger age for the surface. Thus Viala and Hattori (2021), studying zircon crystals derived from intrusive rocks in that area, did not find rocks older than the diorite stock of Cerro Corona, dated at c. 14.5 Ma, which is bevelled by a surface at c. 4000 m, implying an age of c. 14 Ma or less for that surface. The surface is in continuity with the La Oroya Surface in the Yanacocha area, which underlies a volcanic unit dated at 14.3 Ma (Longo 1997).

Montgomery (2012), in a study of the Lagunas Norte area, c. 100 km south of Yanacocha, dated a surface at c. 4000 m, which he named the ‘Chicama Pediment’, at c. 17 Ma. The author correlates this feature with the La Oroya Surface and has no explanation for the significant difference in age when compared to the Yanacocha dates. Noble and McKee (1999) mention that metalliferous veins in the Quiruvilca mining district, lying c. 20 km SW of Lagunas Norte, have been dated at c. 15 Ma. These veins are bevelled by a surface at c. 4000 m, which the author correlates with the La Oroya Surface.

It is quite difficult to explain the broad range of dates for what appears to be the same erosional feature in various parts of northern Peru and it is obvious that more work is required to confirm that all the features at c. 4000 m are in fact part of the La Oroya Surface, and that they do bevel the units from which the ages are being obtained. It is suggested that a provisional age of c. 15–14 Ma is assigned to the La Oroya Surface.

Moreno et al. (2020), working in the Bagua Basin area, dated the unconformity at the base of the San Antonio fanglomerates at c. 13 Ma and correlated it with a surface forming the crest of the nearby Cordillera del Tigre at c. 3500 m. This feature is, in fact, a remnant of the Jauja Surface. The age data pertaining to the unconformity are as follows. Moreno et al. (2020) dated detrital zircons from near the base of the San Antonio Formation at c. 11.5 Ma, whereas units from near the top of the underlying Sambimera Formation yielded detrital zircon ages of c. 13.9 Ma and an age of c. 15.5 Ma from a tuffaceous sandstone. On the basis of these dates, the author suggests an age of 13.5–13 Ma for the Jauja Surface.

The ages of the Chacra surfaces can be assessed by their relationship with the San Jose ignimbrite, dated at 11.4–11.0 Ma (Longo 1997). Probably the best example occurs near the town of Celendin, where a surface at c. 3200 m is mantled by the ignimbrite. In the area immediately east of Cajamarca is an extensive sheet of the San Jose ignimbrite lying on Cretaceous and Paleogene units. The unconformity at the base of the ignimbrite varies in elevation from c. 2800 to 3200 m and is interpreted as representing the Upper and Lower Chacra surfaces. On this basis, an age of 12–11.5 Ma is suggested for the Lower Chacra Surface. There is no further data applicable to the Upper Chacra Surface, although the author suggests an age in the range 12.5–12 Ma subject to further study.

The surface ages that can be assigned to the erosion surfaces on the basis of the available data are as follows: Puna Surface, c. 18–17 Ma; Punrun Surface, c. 17–16 Ma; Junin Surface, c. 16–15.5 Ma; La Oroya Surface, c. 15–14 Ma; Jauja Surface, c. 13.5–13 Ma; Upper Chacra Surface, c. 12.5–12 Ma; and Lower Chacra Surface, c. 12–11.5 Ma.

Figure 16 is a plot of elevation v. age for the surfaces discussed in this paper, plus a date for one of the lower pediments, derived from the fact that the Fortalesa Ignimbrite, dated at 5.6 Ma, rests on the 1400 m pediment in the Fortalesa valley of north-central Peru (Wilson 2015). Although the plot suggests a slight decrease in the rate of uplift for the last 6–5 Ma, this does not materially affect the overall pattern, which suggests an average rate of c. 0.28 mm a⁻¹ for the whole interval 17–0 Ma. The apparently constant rate of uplift does not imply a regular periodicity to the episodes of uplift, nor a uniform amount of uplift corresponding to each episode. It does, however, indicate that the episodes formed a continuous pattern throughout much of the Neogene and presumably continue at the present day. In fact, the Central Andes have been undergoing continuous orogeny since c. 18–17 Ma.

Although much work using conventional and thermochronological techniques has been dedicated to the question of dating the various stages of Andean uplift, sampling has commonly been carried out in an ad hoc manner. Now that specific erosional features can be recognized throughout much of central and northern Peru, it will be possible to focus age-related studies on specific surfaces. Although there is reasonable confidence in correlating specific surfaces over considerable distances in central and northern Peru, it is important to obtain data to confirm those correlations. This should apply to correlations along-strike and also across the Andean belt.

Examples of the type of studies required can be found within the study area. Isotope studies could be applied to the Negritos ignimbrite, lying c. 30 km north of Cajamarca, dated by Noble et al. (1990) at c. 8 Ma and resting directly on a fairly widespread surface at c. 4000 m, which is correlated with the La Oroya Surface. Also, near Celendin, remnants of the Upper Chacra Surface are mantled by the San Jose ignimbrite, outcrops of which lying immediately east of Cajamarca have been dated at c. 11.4 Ma (Noble et al. 1990). Palaeoelevation data from these two areas could help in refining estimates of the rates of uplift in the Late Miocene.

An important objective of future work should be the possible extension of the phenomenon of episodic uplift along the Andean chain to the north and south. Work on northwestern Peru and southern Ecuador (Wilson in prep.) confirms that that region was subjected to episodic uplift. A sequence of erosional surfaces and pediments can be recognized that is very similar to that established in Peru, indicating that the Olmos Arch/Cordillera Real uplift has had a comparable history to the Central Andes. Valuable insights into the rise of the Ecuadorian Andes have been provided by Spikings et al. (2010), but further work focused on specific surfaces would be most useful.

In the case of southern Peru, episodic uplift has been established in a general way by the work of Tosdal et al. (1984) and Quang et al. (2005). Wilson (2009) described a stepped profile in the Andean flank near Ica, while later studies by Schildgen et al. (2009) and Thouret et al. (2017) considered the Ocoña–Cotahuasi river system, which has deeply incised the Andean flank, with a view to constraining the timing and mechanism of the uplift of the Altiplano plateau. Future work could aim to establish a complete sequence of erosional features in the Andean flank throughout southern Peru. The occurrence of Neogene volcanic units within some of the valley systems could facilitate dating of some of the individual features. It may also be pointed out that Tosdal et al. (1984) stated that ‘essentially contemporaneous pediment formation and bevelling of the western slopes of the cordillera are recognized in transects from southernmost Peru … to as far south as the Copiapo area (26°S to 29°S) of north-central Chile’.

To date, most of the studies of erosional features in the Andes have been directed towards the Pacific flank. This report and the author's unpublished studies show that the eastern flank of the Andes in northern and central Peru is a mirror image of the Pacific flank, with the same sequence of surfaces and pediments. An examination of topographic maps of the eastern flank of the Andes in southern Peru and northern Bolivia suggests that those areas also have stepped profiles, although it is not possible to say what relationship those surfaces may have with the features recognized further north.

It seems that a preliminary case can be made that episodic uplift has affected much of the Central Andes, including that sector incorporating the Altiplano. It is therefore interesting that the work of Kar et al. (2023) confirms that the Altiplano had already reached an elevation of 1.8–2.4 km by the Late Miocene, followed by rapid uplift by a further c. 1.5 km since c. 10 Ma. This suggests a different scenario from that emerging from the author's studies and, in fact, it has been thought for some time that the uplift of the Altiplano began earlier than in areas further north. It is suggested that further study of the erosional features of both flanks of the Andes from central Peru southwards would be a potentially fruitful field of research. The results could help to constrain the possible uplift mechanisms involved and any changes therein, both along-strike and in time.

The Central Andes, which extend as far north as c. 6° S, are characterized by the early initiation of episodic uplift at c. 18–17 Ma and the occurrence of erosion surfaces at elevations of c. 4000 m or higher. The widespread remnants of a common sequence of erosional features ranging in elevation from <600 m to >4000 m indicates that the uplift was uniform and simultaneous over a very large area. In fact, the Central Andes rose as a block. Although the Western Cordillera is underlain by a crust with an average thickness of c. 60 km, there is no evidence of significant Neogene contraction or shortening within the Andean Block. The Western Cordillera is also associated with a belt of Miocene mineralization that ends abruptly at c. 6° S. Although the individual cordilleras continue north of 6° S, the initiation of episodic uplift occurred later than in the Central Andes and they are significantly lower than in the Central Andes itself.

By contrast, the Olmos Arch has remnants of surfaces at c. 3800 m (Wilson in prep.), suggesting that uplift was initiated earlier than in the area immediately north of the Central Andes. There are also signs of Miocene mineralization in the Olmos Arch. In effect, the axis of uplift and magmatism was offset c. 50 km from the Central Andes to the Olmos Arch. This offset may reflect a difference in the plate tectonic pattern prevailing in the Miocene from that of today. It appears that the effects of subduction were initially absorbed by the Olmos Arch, such that the area immediately to the east remained undisturbed until c. 14–13 Ma. As that time coincides with the collision of the Inca Plateau with South America at the approximate latitude of northern Peru (Rosenbaum et al. 2005), it may be speculated that that event initiated the spread of uplift into that hitherto undeformed area.

Despite the important differences between the Central Andes and the area immediately north of c. 6° S, both regions were subjected to the same pattern of episodic uplift. Confirmation of this view can be summarized as follows. The flank of the Cordillera del Tigre and the southern margin of the Bagua Basin are etched by generally small remnants of pediments in the range c. 1200–2600 m, which punctuated the gradual rise of those areas (Fig. 15). The Andean flank in the Otuzco area (Fig. 4) shows a very similar sequence of pediments in the range c. 800–2600 m. In both cases the pediments are surmounted by the Chacra surfaces, indicating a general correlation between the two areas. Thus, although uplift was delayed north of c. 6° S, once initiated it led to the development of the same sequence of pediments as in the Central Andes. It also appears that the rate of uplift was the same in both regions, the difference being in the date at which uplift was initiated.

Different parts of the study area possibly have different uplift mechanisms and it is therefore convenient to consider them separately. The Western Cordillera and Eastern Cordillera are discussed in the framework of the Central Andes, extending to c. 6° S, followed by comments on the part of the study area lying beyond that latitude.

The Central Andes, consisting of the Western Cordillera and Eastern Cordillera, extend as far as c. 6° S. The Western Cordillera is underlain by thickened crust, with the axis of thickening lying beneath the higher reaches of the cordillera. There is a gradual thinning of the crust from c. 70 km in southern Peru (Assumpção et al. 2013) to c. 55 km at 6° S and c. 50 km at 5° S (Condori et al. 2017). The Western Cordillera is in isostatic equilibrium (Fukao et al. 1989) and it is generally thought that crustal thickening, whether resulting from underplating or crustal flow, is the factor controlling uplift.

Although the Eastern Cordillera reaches virtually the same elevation as the Western Cordillera, it is underlain by a crust c. 40 km thick (Fukao et al. 1989) and is not in isostatic equilibrium. Fukao et al. (1989) suggested that the mass of the Eastern Cordillera may be supported by the pressure exerted by the westward displacement of the Brazilian Shield. Alternatively, Beck and Zandt (2002) proposed that the Brazilian Shield in fact underthrusts the Eastern Cordillera and thus supports the Eastern Cordillera at its present elevation. Neither of these hypotheses explains why the early uplift of both the Western Cordillera and Eastern Cordillera should end at c. 6° S. There is no obvious reason why the movement of the Brazilian Shield should not affect the area north of 6° S.

The author suspects that the MGA is a more important factor than hitherto considered, forming as it does a rigid block underlying all of the Eastern Cordillera and part of the Western Cordillera (Fig. 2). A possible mechanism controlling Andean uplift could be the gradual accumulation of compressive stresses within the whole of the Andean Block, including the thickened crust and the MGA, until it reaches a level that triggers an abrupt episode of uplift and release of pressure, followed by a repetition of the process. In this scenario, it is envisaged that the MGA is attached to the crustal root underlying the Western Cordillera, such that when uplift is triggered, the Eastern Cordillera rises in tandem with the Western Cordillera. Support from the Brazilian Shield would clearly help the process. This hypothesis explains the virtually identical Neogene histories of the Western Cordillera and Eastern Cordillera better than earlier suggestions. One observation of interest is that the initial uplift of the Puna Surface occurred at a time when crustal thickening was just beginning, yet both cordilleras rose simultaneously by the same amount. Moreover, the uplift extended only as far as c. 6° S and the limit of the MGA. It is difficult to escape the idea that the presence of the MGA was a factor in the uplift, even though the details of the mechanism remain to be established.

The area lying north of c. 6° S, ascribed to the Northern Andes, remained undeformed until c. 13 Ma. Prior to that date it probably still constituted part of the Cenozoic Marañon Basin, limited by uplifts in the Central Andes and also in the Olmos Arch (Wilson, unpublished data). It seems possible that the Olmos Arch was absorbing the compressive stresses generated by subduction, sheltering the area immediately to the east. As the faulting and subsequent uplift of that area at c. 13 Ma coincides with the estimated time of docking of the Inca Plateau, that event may have triggered the uplift of the Cordillera del Tigre and surrounding structures. As emphasized earlier, once initiated, the uplift of the area north of c. 6° S proceeded at the same rhythm as the area to the south.

Whatever uplift mechanism is proposed must also explain the fact that the uplift was episodic, simultaneous and of the same amount across the whole of the Andean Belt. Stepped profiles indicative of episodic uplift can also be recognized in the Huancabamba Andes (Wilson, unpublished data) and also in the Subandean Belt (unpublished data). Moreover, the rate of uplift seems to have remained constant at an average of 0.27–0.29 mm a⁻¹ over an interval of c. 17 Ma (Fig. 16) and across areas of radically different underlying geology, regardless of whether the region overlies a normal or a flat-slab subduction regime.

The Inca Plateau was postulated by Gutscher et al. (1999) to be an oceanic plateau with estimated dimensions of c. 500 km (north–south) by c. 150 km (east–west), which collided with the South American plate at the latitude of northern Peru and southern Ecuador at c. 12–10 Ma. Rosenbaum et al. (2005) calculated that the collision occurred at c. 13 Ma. The feature was subducted and is now thought to lie 600–700 km east of the trench (Gutscher et al. 1999) in the heart of the Amazon Basin (Fig. 17). The subduction of the Inca Plateau is believed to have initiated the flat-slab regime, which ended the main phase of magmatism at Yanacocha at c. 11 Ma.

A number of structural events in northern Peru and southern Ecuador may be due to the docking and subduction of the Inca Plateau. Thus although the main structuring of the Subandean Belt is generally considered to be the result of the Quechua II Orogeny, dated at c. 10–8 Ma (Benavides-Cáceres 1999), uplift began earlier in the study area. As a result, the Neogene Bagua Basin, which originally connected with the Marañon Basin, was separated at c. 13 Ma (Moreno et al. 2020) by the faulting, erosion and subsequent uplift of the Cordillera del Tigre and associated structures. This structuring does not correlate with any of the various phases of the Quechua Orogeny recognized in the Peruvian Andes, but does coincide with the apparent date of docking of the Inca Plateau, which is therefore deemed to be the cause of the tectonism.

Further south in the study area, the faulting and flexuring of the La Oroya Surface in the Yanacocha area affected the Lower Andesite Unit of the Yanacocha Volcanic Complex, dated at c. 14 Ma (Longo 2005), and could therefore be associated with the docking of the Inca Plateau. Also in the study area, the half-graben of the intermontane San Marcos and Cajabamba basins were initiated in the Mid-Miocene (Rousse et al. 2003) and are possibly also the result of the docking and subduction of the Inca Plateau. It may also be noted that the southern limit of the intermontane basins at c. 8° S coincides with the approximate southern edge of the Inca Plateau as it began to be subducted (Gutscher et al. 1999).

The following conclusions can be drawn on the basis of the data and interpretations presented here.

1. The sequence of surfaces and pediments established in central Peru, where they total 20 separate features, can be recognized across the whole width of the Andes in northern Peru.

2. This study focused on the seven highest surfaces, ranging in elevation from c. 2800 to c. 4800 m. The older, higher surfaces occur only as small, isolated remnants as a result of the effects of later erosion.

3. The four surfaces lying at c. 3800 m or higher are not found north of c. 6–6.5° S in the study area. The distribution of the remnants suggests that all the higher features reached that latitude, but did not extend further north.

4. The latitude c. 6–6.5° S is also associated with the following factors.

   • There is significant thinning of the crust north of 6.5° S, as indicated by Fukao et al. (1989).

   • The northern limit of the MGA, a major palaeotectonic feature underlying the Eastern Cordillera and part of the Western Cordillera, coincides with this latitude.

   • The belt of Miocene magmatism and mineralization, so characteristic of the Western Cordillera, ends abruptly at c. 6.5° S.

5. By contrast, the area north of c. 6° S is characterized by a later initiation of uplift. Thus, prior to c. 13 Ma, while the Central Andes had already experienced 1000–1200 m of uplift, the area in question remained undisturbed.

6. On the basis of these observations, it is suggested that the Central Andes terminate at c. 6° S. Baby et al. (2018) similarly placed the northern limit of the Central Andes in an area between the Santiago and Huallaga basins.

7. Given these conclusions, the plate tectonic pattern of the Miocene may have been different from that prevailing today. Be that as it may, the cessation of the main phase of magmatic activity in the Central Andes at c. 11 Ma is interpreted as marking the onset of flat-slab subduction in northern Peru, probably associated with the subduction of the Inca Plateau.

8. Regardless of any changes in the plate tectonic pattern and in the subduction regime, the average rate of uplift remained constant at c. 0.29 mm a⁻¹ from the Mid-Miocene onwards.

9. The uplift has effectively been continuous from the Mid-Miocene, implying ongoing orogeny.

10. The Andean uplift is mainly the result of a thickened crust lying beneath the Western Cordillera, reaching c. 60 km at 10° S.

11. Although the Western Cordillera is in isostatic equilibrium, the Eastern Cordillera is underlain by a crust only c. 40 km thick, yet reaches virtually the same elevation as the Western Cordillera. It has been suggested that pressure exerted by the Brazilian Shield may support the Eastern Cordillera, but that would not necessarily explain the common uplift history and elevation of the two cordilleras. A possible explanation is that, as the Western Cordillera rises because of progressive crustal thickening, the rigid mass of the MGA, which underlies both the Eastern Cordillera and part of the Western Cordillera, rises in tandem with it.

12. The mechanism leading to crustal thickening remains to be resolved, although underplating seems more likely than the extrusion of crustal material from the Bolivian Orocline. It is nevertheless difficult to explain the fact that episodic uplift appears to have occurred simultaneously and to the same degree over very large areas.

13. There is no clear explanation for the episodic nature of Andean uplift. It may result from a more or less regular process of underplating, although this would not explain episodic uplift in areas distant from the axis of crustal thickening. Alternatively, while crustal thickening is proceeding in a regionally compressive regime, there may be a build-up of pressure that reaches a point which triggers an abrupt uplift and release of pressure, followed by a repeat of the process.

The author is grateful for constructive reviews by the Editor Philip Hughes and Patrice Baby, which materially improved the paper.

JJW: conceptualization (lead), investigation (lead), methodology (lead), validation (lead), visualization (lead), writing – original draft (lead), writing – review and editing (lead).

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The author declares that he has no financial interest or personal relationships that could have appeared to influence the work reported in this paper.

All data generated and analysed during this study are included in the published article.