Site 1227 Ash Layers

Site 1227 is located on the upper slope of the southeastern edge of the Trujillo Basin (Fig. F1) at a water depth of 427.5 m (D'Hondt, Jørgensen, Miller, et al., 2003; Pouclet et al., 1990). The cores from this site contain one Type 1 ash layer (1.3 cm), two Type 2 layers (0.5 and 1.3 cm), and nine Type 3 layers (Table T4). In addition to these layers and accumulations, 12 Type 4 layers have been recognized. Type 1 ash layers for this hole consists of ~85%–90% volcanic glass with 5%–10% quartz and feldspar (no mafic minerals are present), and there is no rounding of grains and little or no biogenic component present (Fig. F4; Table T3). According to the stratigraphic section (Fig. F5), this layer occurs in Miocene sediments, whereas the two Type 2 layers occur in Pliocene sediments. Type 3 ash occurs predominantly within Pleistocene sediments and within one Holocene, one Pliocene, and two Miocene layers as well. Type 4 ash layers occur within Pliocene and Miocene sediments with the exception of one Holocene layer.

Site 1228 Ash Layers

Site 1228 is located on the Peru shelf at the southeastern edge of the Salaverry Basin (Fig. F1) at a water depth of 252 m (D'Hondt, Jørgensen, Miller, et al., 2003; Pouclet et al., 1990). The cores from this site contain 52 Type 1 ash layers, 38 of which were analyzed for whole-rock geochemistry (layer thickness ranges from 0.5 to 13.5 cm) (Table T3), and 28 Type 2 ash layers, 8 of which contain some dolomite (layer thickness ranges from 0.7 to 14 cm). Layers containing Type 3 ash are dispersed throughout the entire Pleistocene–Pliocene cored sections. In addition, 11 Type 4 layers are present in Cores 201-1228A-5H through 9H (Table T4). Type 1 ash layers from this site consist mostly of quartz and feldspars (usually 85%–90%, varying between quartz and feldspar) and contain, on average, 0%–2% mafic, pleochroic minerals (biotite or hornblende) and 0%–5% volcanic glass. Seven of the Type 1 layers have no round quartz; two contain 50% and 70% volcanic glass, four contain high pumice (40%–80%), and one contains low glass and pumice (Fig. F4; Table T3). These Type 1 ash layers are shown in the stratigraphic section (Fig. F5) as occurring within Pliocene sediments, whereas Type 2 ash layers occur within Holocene, Pleistocene, and Pliocene sediments, Type 3 and Type 4 ashes occur within Pleistocene and Pliocene sediments.

Site 1229 Ash Layers

Site 1229 is located on the Peru shelf at the southeastern edge of the Salaverry Basin slightly northwest of Site 1228 (Fig. F1) (D'Hondt, Jørgensen, Miller, et al., 2003; Pouclet et al., 1990). This site is the most landward site of the three sites in this study region and was drilled at a water depth of 150.5 m (D'Hondt, Jørgensen, Miller, et al., 2003). The cores from this site contain 15 Type 1 ash layers, 8 of which are considered borderline and 14 of which were analyzed for whole-rock geochemistry. In addition, there are 6 ash laminae of Type 1 composition (Table T3) and 92 Type 2 ash layers (9 pods, 1 lens, and 1 laminae set), 24 Type 3 ash occurrences (2 pods), and 23 Type 4 layers that contain varying amounts of dolomite (3 are borderline) (Table T4). Type 1 ash layers for this site consist mostly of quartz and feldspars (usually 85%–92% varying between quartz and feldspar) and contain, on average, 0%–2% mafic, pleochroic minerals (biotite or hornblende), and 0%–3% volcanic glass (Table T3; Fig. F4). These Type 1 ash layers are shown in the stratigraphic section (Fig. F5) as occurring within Pleistocene and Pliocene sediments, whereas Type 2 ash layers occur within Holocene, Pleistocene, and Pliocene sediments and Type 4 layers occur within Pleistocene sediments (with the exception of one layer in interval 201-1229A-18H-2, 127–128.8 cm [~138.4 mbsf] that occurs just below the Pleistocene/Pliocene boundary).

Glass Geochemistry of Type 1 Ash Layers

Geochemical analysis of glass shards was conducted on a select group of ash layers from each Leg 201 site. We attempted to analyze at least 10 glass shards from each ash layer; however, fewer analyses were conducted for some layers because of the very fine grained nature of some samples and the plucking of grains during polishing of the microprobe mounts. The results of these analyses are displayed on total alkalis vs. silica (TAS) and potassium vs. silica plots (Figs. F6, F7) for both the full set of glasses analyzed and the average value for each ash layer, respectively.

Overall, Leg 201 glasses display predominantly rhyolitic (–70 wt% SiO₂) chemical compositions (Table T5) and plot mostly in the subalkaline (tholeiitic) series on a TAS diagram (Fig. F6). However, two samples have andesitic compositions (most likely due to hydration of glass, indicated by low total values), four samples have trachyte–trachydacitic compositions, and one sample has a dacitic composition (Fig. F6; Table T5). Additionally, the dacitic sample and one other sample (which falls on the boundary between dacite and rhyolite) have very low alkali contents (<4 wt%) and two of the rhyolitic samples have very high silica values (>80 wt%).

Average glass compositions form two distinct groupings in TAS space. Five samples from Hole 1228A have total alkali of 8.33–8.56 wt% and SiO₂ of 69.8–71.6 wt%. We designate these samples geochemical Group 1, whereas the remaining four Hole 1228A samples and all Hole 1227A and 1229A samples (geochemical Group 2) have lower total alkalis (7.29–8.06 wt%) and higher silica (72.9–74.2 wt%). One exception is a Hole 1228A sample that has elevated SiO₂ (81.3 wt%) and may not reflect a primary ash composition (Table T5; Fig. F7).

On average, the rhyolitic glasses plot in the high-K calc-alkaline and shoshonite fields in the K₂O vs. SiO₂ space (Figs. F6, F7), which complements the Pouclet et al. (1990) geochemical evaluation of glasses from the same region (Leg 112). The average values for the most representative major element oxides are plotted against silica and their trends are displayed on variation diagrams in Figure F8. Additionally, major element oxide regions for glass values from Leg 112, the CVZ, and offshore Central America are included on the variation diagrams for the purpose of possible source region comparisons.

Major element data for glasses within Group 1 display clear and almost linear trends with increasing CaO and K₂O and decreasing FeO, Na₂O, TiO₂, and MgO (which is not quite as distinct as most of the trends) as SiO₂ increases. Major element data for glasses in Group 2 do not display clear linear trends with increasing SiO₂ content. The data for these glasses tend to cluster around one particular value as SiO₂ increases per major element oxide (Fig. F9). Additionally, the data for this group can be further subdivided into two separate groups for Al₂O₃, CaO, MgO, and Na₂O, with each grouping around one value per major element oxide with increasing SiO₂.

All Leg 201 glasses have major element oxide values that plot on variation diagrams (Fig. F9) within the range of Leg 112 samples, with the exception of the one sample from Hole 1228A that has elevated SiO₂. Most major element oxide data also fall within the CVZ region with the exception of Al₂O₃ (Fig. F8). All Leg 201 glass values plot outside the Central American region, with the exception of Na₂O and K₂O, in which a minimal number plot within this region.

Major Element Oxides

Leg 201 ash layers have predominantly andesitic to dacitic (58.8 < SiO₂ < 64.5 wt%) whole-rock chemical compositions (Table T6) and plot in the subalkaline (tholeiitic) series in TAS space (Fig. F9). However, seven samples have trachybasaltic to basaltic trachyandesitic (49.5 < SiO₂ < 53.5 wt%) compositions. These values are most likely low due to hydration evidenced by the relatively high (>10 wt%) loss on ignition (LOI) values (Fig. F9). Four of the dacitic samples are distinguished from the rest, in that two of them have high silica values (66.4 and 69.2 wt%) and the other two have the lowest total alkali values (3.79 and 4.55 wt%) (these low values are due to low sodium values, 1.71 and 2.44 wt%, respectively) (Fig. F9). The Hole 1229A sample with the lowest sodium has a high concentration of silica (68.4 wt%) as well. Additionally, six samples with average silica values lie on or just above the boundary between the trachyandesite/trachydacite fields and andesite-dacite fields (one sample from Hole 1227A plots well within the trachyandesite field). All ash layer samples have whole-rock compositions that fall in the medium-K or high-K calc-alkaline fields, with the exception of the seven hydrated samples discussed above (Fig. F9).

Whole-rock major element oxide data are plotted vs. silica in Figure F10. Additionally, major element oxide regions for whole-rock values from the CVZ, the NVZ, and the SVZ are included on the variation diagrams for the purpose of possible source region comparisons.

The hydrated ash layers plot to the left side (low silica) on the variation diagrams. The intermediate (andesitic and trachyandesitic) and acidic (dacitic and trachytic) samples plot together and display regular and continuous trends with increasing Al₂O₃, Na₂O, K₂O, and TiO₂, as SiO₂ increases for Hole 1228A samples. These element oxides decrease with increasing SiO₂ for Hole 1229A samples, with the exception of TiO₂, for which a trend is unapparent. CaO, Fe₂O₃, and MgO decrease as SiO₂ increases for samples from both sites; apparent trends are less obvious for P₂O₅ with increasing SiO₂.

Samples from Holes 1228A and 1229A have P₂O₅ values between 0.12 and 2.65 wt% and 0.10 and 1.36 wt%, respectively (average values = 1.09 and 0.49 wt%, respectively), which in some samples is considerably greater than that reported by the USGS and Le Maitre (1976) for andesitic to dacitic whole-rock standard compositions (0.15–0.63 wt%). The high P₂O₅ values for these samples can possibly be attributed to diagenetic alteration, in the form of secondary apatite and phosphate, since fluoroapatite and phosphate concretions have been documented throughout Leg 201 cores (D'Hondt, Jørgensen, Miller, et al., 2003).

Nearly all Leg 201 samples (with the exception of the hydrated rocks) have major element oxide values that plot within the field of published CVZ values (Fig. F10) within the CVZ region. One sample from Hole 1228A plots outside of the CVZ region for Al₂O₃ and Na₂O, one Hole 1229A sample plots outside the CVZ field for Fe₂O₃, Na₂O, and TiO₂. The Hole 1227A sample plots outside the CVZ array for TiO₂. The NVZ region overlaps that of the CVZ for all major element oxides with the exception of Fe₂O₃ and MgO (Fig. F10).

Rare Earth Elements

Chondrite-normalized rare earth element (REE) plots for Andean volcanic zones and for all Type 1 ash layers are shown in Figure F11. The REE pattern for most ash layers steadily decreases from the light REE (LREE) to middle REE (MREE). Heavy REE (HREE) displays a nearly flat pattern (Fig. F11). These REE patterns are most similar to those of the CVZ (Fig. F11). REE of the SVZ and NVZ are more HREE enriched and thus display a less steeply sloping pattern (Fig. F11). Based on these data our samples have more of an affinity to CVZ type ashes than SVZ and NVZ.

Two ash layers, one from Site 1227 and one from Site 1228, display REE patterns that are anomalous. These layers display steeply sloping REE patterns, and the layer from Site 1228 is slightly more enriched in HREE than the layer from Site 1227 (Fig. F11).

REE patterns are commonly used to indicate processes occurring during magma generation (Richards and Villeneuve, 2001; Monzier et al., 1999; Trumbull et al., 1999; Droux and Delaloye, 1996; Matteini et al., 2002; Dorendorf et al., 2000; Rollinson, 1993). As previously noted, the principal magma source for the andesitic volcanoes of the CVZ is interpreted to be partial melting of an asthenospheric wedge between the overriding continental South American plate and the descending oceanic Nazca plate (de Silva and Francis, 1991; Thorpe and Francis, 1979; Baker and Francis, 1978; Hanus and Vanek, 1978). The moderate depletion of HREE with respect to LREE displayed by all Leg 201 ash layers (Fig. F11) indicates either partial melting of a lower crust or mantle source region with garnet as a residual phase or precipitation of garnet in a deep-seated magma body (Richards and Villeneuve, 2001; Monzier et al., 1999; Trumbull et al., 1999; Droux and Delaloye, 1996; Matteini et al., 2002; Dorendorf et al., 2000; Rollinson, 1993). The suppressed negative Eu anomaly (with respect to CVZ samples) could be due to effective plagioclase enrichment in the ashes owing to winnowing of more dense ferromagnesian phases during airborne transport. However, the occurrence of Eu as Eu³⁺ has been noted by Richards and Villeneuve (2001) to frequently occur within fairly oxidized and hydrous arc magmas, resulting in a weak or lacking negative Eu anomaly as well. The excessive depletion of HREE in the two samples from Sites 1227 and 1228 might be a result of excessive garnet precipitation and sequestration at depth. Alternatively, this REE pattern is similar to those of adakites and Archean tonalite, trondhjemite, and granodiorite (TTG) suites present within southern Peru (Condie, 2005; Martin, 1993, 1999) samples.

Ash layer correlations between sites were based on lithology, stratigraphic position, and major element geochemistry and were refined through REE geochemistry. Most ash layers from Sites 1228 and 1229 are lithologically similar, with the primary difference being the abundance of volcanic glass or pumice, and half of Site 1229 ash layers correlate with Site 1228 ash layers even upon comparing volcanic glass and pumice abundances (Table T3). However, some of these ash layers do not correlate well stratigraphically.

As previously discussed, most Leg 201 ash layers display the same general major element oxide trends between sites for both glass and whole-rock analyses. However, no ash layers correlate perfectly based on major element oxides. Because of the lack of definite correlations of ash layers based on lithology, stratigraphic position, and major element oxide concentrations, REE geochemistry was used to refine possible correlations between Leg 201 sites; these correlations are displayed in Figure F12. This figure shows that three ash layers correlate between Sites 1228 and 1229 and the REE pattern for one ash layer from Site 1228 is similar to that of the one layer at Site 1227 (Fig. F12), in that both display steep REE patterns.

Many studies have recorded the explosive volcanic cycles of land eruptions through documentation of ash layers in deep-sea sediments (Prueher and Rea, 2001; Pouclet et al., 1990; Paterne et al., 1990; Ledbetter and Sparks, 1979; Kennett et al., 1977; Donnelly, 1976). The thickness and number of Type 1 ash layers have been summed for each Leg 201 site in this study and have been plotted per half million year time increments (Fig. F13) for the purpose of depicting the explosive volcanic cycles for the Central Andes. One limitation in utilizing Leg 112 sedimentation rate curves to depict ash occurrence per half million years from Leg 201 is that the sedimentation rate curves from Leg 112 only extend to 120 mbsf, whereas Type 1 ash is present in Leg 201 Sites 1228 and 1229 cores to 152.74 and 159.55 mbsf, respectively. Additionally, another limitation is that Leg 112 sedimentation rate curves are constructed based upon the first and last occurrence of either diatoms or nannoplankton, which may yield large time errors (Fig. F13).

This summation does not take into account accumulations, pods, very thin laminae (<1 mm), or diagenetically altered layers and therefore provides minimal estimations at best. Additionally, all Leg 201 sites have been affected by sedimentary hiatuses (Fig. F5), and some sections of core display evidence of possible slump deposits and/or turbidites. These features could be a significant factor in the disappearance of ash layers from the marine record. Prevailing wind direction and marine currents are two limiting factors that should also be taken into account when recording volcanic cycles from ash within deep-sea sediments (Pouclet et al., 1990). Because the closest known volcanic source is >400 km from the study region (Fig. F1), it is apparent that wind direction and, to a smaller extent, ocean current may be responsible for depositing ash far offshore. Figure F3 displays the southwestern trade winds that circulate north-northwestward along the Chile coast. These winds appear to follow the Chilean littoral up to the Peru coast, where their direction takes on a southeasterly path that parallels the Peruvian coast. These winds, coupled with the coriolis effect, cause a deflection in the Humboldt Current, which circulates northward along the South American western coast to the surrounding area of the Equator, thus severing the current from the Peruvian coast (Garcia, 1994). This suggests that volcanic ash erupted into the atmosphere from northern Chile and/or southern Peru would have been transported into the study region by the prevailing trade winds along western South America.

Figure F13 displays ash layer thickness and the number of ash layers per half million year time increments. Explosive eruption cycles for the Andean region have been deduced from these data. Our record of volcanic cycles indicates that explosive activity was less intense during the Miocene, in which one ash layer (1.3 cm) was deposited, compared to that of the Pliocene and Pleistocene, which experienced most of the explosive volcanic activity in which 52 ash layers (total thickness equal to 208.6 cm) and 14 ash layers (total thickness equal to 122.1 cm) were deposited, respectively. These data are consistent with the previous study of Pouclet et al. (1990); however, these data indicate that explosive activity during the Pliocene and Pleistocene was more intense than previously reported. Additionally, the total thickness of primary ash within all three Leg 201 sites is equal to 332.0 cm, which is ~24 times as much primary ash reported by Pouclet et al. (1990; 14 cm) within the previously occupied sites of Leg 112. However, Pouclet et al. (1990) report three Miocene volcanic phases for the northern and southern sites of Leg 112. Here only one Miocene ash layer is observed within cores from the northern site (Hole 1227A; Pouclet et al. [1990] Site 684), and no Miocene ash is observed for Leg 201 southern sites. This discrepancy is due to two factors: first, only one Miocene ash layer is present within the northern site studied both here (Hole 1227A) and previously (Site 684); the other two layers reported by Pouclet et al. (1990) were in cores from northern sites that were either not examined during this study or were not reoccupied during Leg 201. Second, the southern sites studied by Pouclet et al. (1990) that contained Miocene ash layers were not reoccupied during Leg 201.