Centennial-scale variations in diatom productivity off Peru over the last 3000 years

Abstract

The Peruvian coastal upwelling is one of the most productive systems in the global ocean, with important impacts on the carbon cycle. Primary productivity there displays strong variations at the interannual to decadal timescales. However, down-core investigations rarely reach sufficient temporal resolution to assess the response of productivity to climatic variations at these timescales beyond the instrumental and historical periods. We here analyzed diatom assemblages, sea-surface temperatures (SSTs), and nitrogen and organic carbon contents on a laminated sediment core from the Peruvian continental shelf to trace variations in regional productivity over the last 3000 years. Our record provides evidence for different climatic and oceanic conditions with more humid and less productive conditions older than 2500 cal. yr BP and drier and more productive conditions younger than 2500 cal. yr BP. The last 2500 years also present much stronger centennial-scale variability with the occurrence of six intervals with higher total diatom abundances and stronger percentages in upwelling-related diatom species, representative of intensified productivity, congruent to lower percentages in benthic diatoms, indicative of reduced rainfall. These six periods were synchronous to intervals of enhanced Walker circulation, suggesting a strong imprint of the Pacific zonal circulation on productivity variations off Peru. Our record also demonstrates that SSTs did not vary in phase with productivity, arguing against the idea of regional SSTs controlled by the upwelling intensity, but were rather in agreement to SST records off southern Chile, suggesting that Peruvian SST variations were largely controlled by oceanic currents at southern high latitudes.

Keywords

diatoms laminae late Holocene productivity Peruvian upwelling system sea-surface temperatures

Introduction

The Peruvian Upwelling System (PUS) is one of the most productive regions in the global ocean (Barber and Smith, 1981). High phytoplankton productivity is sustained by the coastal upwelling of cold and nutrient-rich subsurface waters (Huyer et al., 1987). Intense phytoplankton productivity in turn supports a rich ecosystem in terms of biomass (Brink et al., 1983), especially high fish abundances (Carr et al., 2002). Rapid carbon export in coastal upwelling systems such as the PUS contributes to carbon sequestration in the deep ocean (Antia et al., 2001), which has been suggested as one of the factors driving changes in atmospheric CO₂ concentrations over the late Quaternary (McElroy, 1983). The mean conditions of the PUS are modulated at the interannual timescale by El Niño Southern Oscillation (ENSO), which strongly impacts productivity and carbon export. El Niño events cause a decrease in productivity (Cowles et al., 1977) whereby La Niña events are marked by strong productivity off Peru (Aronés et al., 2009) and northern Chile (Thomas et al., 2001). Tracing past changes in productivity off Peru could thus enable to improve the prediction of natural greenhouse gas emissions and its response to changes in ENSO mean state and activity.

Most reconstructions of past productivity off Peru focused on the glacial–interglacial cycles at the millennial timescale (Schrader and Sorknes, 1991) or the Holocene at the centennial to millennial timescales (Chazen et al., 2009; Makou et al., 2010). More recently, some effort was done to investigate climate and productivity changes during the last millennium at the decadal timescale (Salvatteci et al., 2014; Sifeddine et al., 2008). The records studied at the centennial to millennial timescales were obtained from the analysis of laminated sediment cores which were sampled at regular steps, regardless of the sediment structure, and thus certainly mixed contrasted environmental conditions. Indeed, sedimentary laminations preserved on the continental shelf off Peru, alternating between opal-rich terrigenous-poor and opal-poor terrigenous-rich millimeter-thick layers (Bull and Kemp, 1996), and thus give access to past changes in mean environmental conditions and the resulting productivity.

Although diatoms are the main primary producers in coastal upwelling systems all year round (De Mendiola, 1981) and that their preserved assemblages reflect environmental and productivity changes (Abrantes et al., 2007), they have been rarely used to infer past productivity changes whereby previous studies were essentially based on geochemical tracers (Higginson and Altabet, 2004; Salvatteci et al., 2014; Schrader and Sorknes, 1990). We here studied a sediment core off Peru at the lamination-scale, respecting the lithology of the core, to provide a record of total productivity, diatom assemblages, and sea-surface temperatures (SSTs) with the main objective to explore the response of regional productivity, especially in the PUS, to climatic variations at the decadal timescale. We focus on the last 3000 years, during which sea-level (Siddall et al., 2003) and insolation (Berger and Loutre, 1991) changes can be disregarded, to have access to the natural climatic variability expected under conditions similar to those prevailing today.

Study area

The PUS is one of the four major upwelling systems associated with eastern boundary currents (Carr and Kearns, 2003). The Peru–Chile Current (PCC, or Humboldt Current (HC)) and the Peru–Chile Coastal Current (PCCoastalC, and labeled PCcC in Figure 1) flow northward along the southwestern American coast from the Southern Ocean to the Eastern equatorial Pacific Ocean (Figure 1a). The strength and persistence of this current and of the resulting upwelling cells are intimately coupled with the northwestward blowing trade winds (Fiedler and Talley, 2006; Kessler, 2006). The Equatorial Undercurrent flows eastward under surface waters and is deflected southward where it contributes to the formation of the Peru–Chile UnderCurrent (PCUC). Similarly, the Southern Subsurface CounterCurrent (SSCC) generates the Peru–Chile CounterCurrent (PCCC). At the latitude off Peru, the PCCC and PCUC are oxygen-depleted and nutrient-rich and source the waters upwelled along the coast, sustaining high productivity (Chavez and Barber, 1987; Toggweiler et al., 1991; Wyrtki, 1981). These currents especially bring Equatorial Subsurface Waters (ESSW), which are the main source of nutrients on the Peruvian and Chilean margins (Strub et al., 1998). The upwelling of subsurface waters drives lower SSTs in the PUS compared with surrounding tropical waters (Huyer et al., 1987) (Figure 1b). The phytoplankton community dwelling in this highly productive environment is dominated by diatoms, especially the neritic bloom-forming Chaetoceros spp. (Avaria and Muñoz, 1987; Cowles et al., 1977).

Figure 1.

Maps of oceanographic parameters used in this study. Core location and comparative sites are represented by red star and black squares, respectively (1 – Moon Lake, 2 – GeoB 3313-1, 3 – GeoB7186-3). (a) Chlorophyll a concentrations from Aqua MODIS averaged from 2002 to 2013 winters (http://oceancolor.gsfc.nasa.gov). Surface and subsurface currents are marked by solid black and dashed gray lines, respectively; EUC: Equatorial Undercurrent; PCC: Peru–Chile Current; PCcC: Peru–Chile coastal Current; PCCC: Peru–Chile CounterCurrent; PCUC: Peru–Chile UnderCurrent; SEC: South Equatorial Current; SSCC: Southern Subsurface CounterCurrent (adapted from Mollier-Vogel et al., 2012) (b) Sea-Surface Temperature from World Ocean Atlas 2009 averaged from 1955 to 2006 (https://www.nodc.noaa.gov; Locarnini et al., 2010).

Off Peru, productivity varies seasonally in response to changes in upwelling intensity and water mixing. The maximum in upwelling intensity occurs during austral winter (i.e. in June, July, and August) (Messié et al., 2009). However, because of greater and deeper mixing in winter, productivity is higher during austral spring (Chavez et al., 1996; Pennington et al., 2006). In addition to seasonal variations, productivity and oxygen contents within the PUS are very sensitive to ENSO (Philander, 1990). During El Niño events, warmer surface waters are observed offshore Peru despite the occurrence of upwelling-favorable winds (Strub et al., 1998) and upwelling is then restricted to a narrower area along the coast (Chavez and Barber, 1987). During these events, the nutricline deepens and leads to the upwelling of warm waters with moderate nutrient contents. El Niño events are known to drive dramatic reductions in productivity, biomass, and changes in specific composition (Cowles et al., 1977) compared with the mean conditions observed off Peru. On the opposite, La Niña events are characterized by enhanced upwelling conditions and increased abundances of bloom-forming species such as the diatom Chaetoceros spp. La Niña events also strongly impact the continent, intensifying droughts in the Peruvian coastal desert, whereas rainfall only occurs during strong El Niño events (Wells, 1990).

Materials and methods

Core description and age model

Sediment piston core M772-003-2 was collected at 15°05′S, 75°44′W, on the Peruvian margin, during cruise M772 in 2008. The core is located within the PUS, at the shallow water depth of 271 m and in the core of the Oxygen Minimum Zone, under conditions favorable to the formation of laminae (Chazen et al., 2009). The core is composed of a succession of laminae measuring between 1 mm and 1 cm (6 mm on average) and appearing alternatively light and dark on x-ray images (Figure 2). Several laminae of similar density are grouped in thicker units. For example, a light-laminated interval, characterized by the dominance of light laminae, is visible between 49 and 62 cm (Figure 2). The light laminae contain dark olive to olive clays or silty clays, high quantities of microfossils, predominantly diatom frustules, and low quantities of dense terrigenous material. In contrast, dark laminae are characterized by dark olive to gray silty clays or silts, lower quantities of microfossils, and higher amounts of terrigenous material. Dark layers are therefore denser than lighter layers and appear darker on x-ray radiographies.

Figure 2.

Photo images (left) and x-ray radiographies (right) of the two sections of piston core M772-003-2 studied in this paper. The gray shaded areas represent the extension of the light-laminated intervals discussed in the paper. The red triangles indicate the position of the radiocarbon ages.

Our high-resolution analyses of diatom assemblages and geochemical contents focus on the last 3000 years, corresponding to the top part (1.83 m) of the core. The age model is based on six radiocarbon measurements, with the oldest age being measured just below the studied interval, on humic acids extracted from 1 g of dried and homogenized sediment per sample at the Leibniz Laboratory for Isotope Research and Dating in Kiel. We considered a reservoir age equal to 511 ± 278 years (Ortlieb et al., 2011). The corrected dates were calibrated with the program CLAM 2.2, using the MARINE13 calibration curve (http://calib.qub.ac.uk/marine/; Table 1). The program CLAM 2.2 was also used to provide the best fit between radiocarbon ages through a Bayesian method (Blaauw, 2010) using a ‘smooth spline’ age model (Figure 3). The resulting age model enabled to estimate the mean duration of a lamination as 11 years, since 247 laminae were observed on the core, which covers 2634 years. Laminae do not allow the identification of individual El Niño or La Niña events and rather represent the mean conditions prevailing in the PUS over a decade, thus allowing centennial-scale paleoceanographic reconstructions.

Table 1.

Radiocarbon ages on piston core M772-003-2.

Depth (cm)	Radiocarbon age (yr BP)	Error (years)	Age ranges at 95% confidence intervals (yr BP)	Calibrated age (cal. yr BP)
8	1100	±25	524–631	578
22	1455	±20	816–948	882
48	2135	±20	1525–1671	1598
90	2630	±25	2112–2289	2201
145	3260	±25	2843–3028	2936
208	3945	±25	3691–3870	3781

Figure 3.

(a) Age model and (b) sedimentation rates of piston core M772-003-2. The points represent the age ranges corresponding to each radiocarbon measurement. The bold line corresponds to the age model used by CLAM 2.2 to estimate the age of each sample. The error range in age estimations is represented by the gray shaded area.

Methodology

The x-ray radiographies of the cores were obtained using an x-ray image-processing tool (Migeon et al., 1999) while the geochemical composition of the sediment was estimated with the AVAATECH XRF core-scanner at EPOC. Before the analysis, the sediment surface was flattened and covered with Ultralene film. The core sections were scanned at a resolution of 200 µm at two different levels of energy (10 and 30 keV), giving information on down-core variations of major elements such as Silicon (Si) and Iron (Fe). These non-destructive high-resolution analyses enabled to capture millimetric laminae based on their color (pictures), density (x-ray radiography), and elementary composition (XRF). These results were used to count, identify, and sample the laminae for additional micropaleontological and geochemical analyses.

For diatom analysis, three slides were mounted per sample using the procedure described in Rathburn et al. (1997). Diatom identification was achieved on an Olympus BX-51 phase contrast microscope at a magnification of ×1000 and following the counting rules described in Crosta and Koç (2007). A minimum of 300 valves were counted per sample. Diatoms were identified to genus or to sub-species group level, and the relative abundance of each species was determined as the fraction of the diatom species against total diatom abundance in the sample. Diatom identification was based on Sundström (1986), Hasle and Syvertsen (1996), Moreno et al. (1996), Sar et al. (2001, 2002), Sarno et al. (2005), and Sunesen et al. (2008). Total diatom assemblages include 104 species, among which Chaetoceros resting spores (CRS), Skeletonema spp., and Thalassionema nitzschioides var. nitzschioides are dominant. Species were grouped based on their ecological preferences (Abrantes et al., 2007; Fleury, 2015; Romero et al., 2001).

Total nitrogen (TN) contents were determined on 8–15 mg of dried, ground, and homogenized bulk sediment using a Carlo-Erba CN analyzer 2500. The internal consistency of our measurements was continuously checked using several calibrated laboratory standards. Their reproducibility, based on in-house and international standard replicates, was ±0.01%. In parallel, organic carbon contents were measured with a LECO C-S 125 analyzer on 80 to 100 mg of dried and homogenized sediment after calcium carbonate was removed with hydrochloric acid. Their reproducibility reached ±0.5%.

Alkenone contents were measured at the Institute of Geosciences in Kiel. Alkenones were extracted from 0.5 to 1 g of freeze-dried and homogenized sediment through Accelerator Solvent Extraction with the Dionex ASE200 of the Institute of Geosciences. The extraction was performed using a solution of dichloromethane and methanol (9:1) as solvent. The extracted components were analyzed with a double column multidimensional gas chromatograph system (Agilent 6890N) with hydrogen as a carrier gas. Concentrations in C_37:2 and C_37:3 alkenones (di- and tri-saturated ketones respectively) were quantified by normalizing them with two internal standards (Skagerrak and Standard C₂₇C₃₆), using the method described in Blanz et al. (2005). SSTs were estimated using the calibration proposed by Müller et al. (1998).

Results

Sediment structure

The core is composed of a succession of laminae measuring between 1 mm and 1 cm (6 mm on average) and appearing alternatively light and dark on x-ray images (Figure 2). Several laminae of similar density are grouped in thicker units. We observed six intervals composed of a succession of laminae dominated by light sediments on the x-ray images (Figure 2): between 148 and 145 cm, 96 and 89 cm, 82 and 69 cm, 62 and 49 cm, 40 and 31 cm, and between 21 and 4.5 cm from bottom to top, corresponding respectively to the time intervals 2960–2930, 2300–2200, 2100–1900, 1800–1550, 1350–1150, and 900–500 cal. yr BP, respectively. These periods are further called ‘light-laminated intervals’ and labeled as intervals 6 to 1 from the oldest to the most recent. The definition of ‘light-laminated intervals’ is based on the combination of several criteria: light color, high total diatom abundances and Si/Fe ratio values, low Fe content and density.

Diatom assemblages

Total diatom abundances present mean values of ~211 million valves/g of dry sediment between 3050 and 2500 cal. yr BP as well as a minimum at around 2550 cal. yr BP (Figure 4a). Diatom abundances increased 2500 years ago and present a baseline around ~310 million valves/g of dry sediment over the last 2500 years, as well as large peaks up to ~900 million valves/g of dry sediment in six intervals: 2960–2930, 2300–2200, 2100–1900, 1800–1550, 1350–1150, and 900–500 cal. yr BP. Each of these periods with higher total diatom abundances corresponds to the light-laminated intervals defined on the x-ray images. The magnitude of the increase differs from one interval to the other and is much higher during light-laminated 3 (i.e. between 1800 and 1550 cal. yr BP).

Figure 4.

Diatom assemblages in piston core M772-003-2 (Peru margin, 15°S) over the last 3000 years: (a) total diatom abundances and relative abundances of (b) Chaetoceros resting spores (CRS group), (c) Skeletonema spp., (d) tropical oligotrophic diatoms, (e) benthic diatoms. The red triangles indicate the radiocarbon ages. The gray shaded areas represent the intervals of increased diatom productivity, coincident with the dominance of light laminations in the sediment, as seen on the x-ray radiographies. The gray shaded areas represent the light-laminated intervals characterized by light sediments, high total diatom abundances, high Si/Fe ratios, and low Fe contents.

These variations in diatom abundances are accompanied by large changes in the composition of the assemblage. The main group consists of CRS (CRS group) and is considered as the sum of Chaetoceros debilis, Chaetoceros didymus, and Chaetoceros radicans, and an unidentified species. The unidentified species is represented by small oval-shaped resting spores such as those produced by Chaetoceros debilis but without any ornamentation on the valves. The percentages of both the unidentified species and Chaetoceros debilis vary in parallel (not shown), suggesting they have the same ecological preferences. The CRS group represents the main contributor to the total assemblage over the last 3000 years with mean percentages reaching 30% over the whole record (Figure 4b). The highest percentages of the CRS group are found during intervals of increased diatom abundances, especially during the period between 2300 and 1550 cal. yr BP, which includes light-laminated intervals 5 to 3 (Figure 4).

Skeletonema spp. represents the second group of diatoms in terms of abundances in the studied record. It includes Skeletonema costatum, the most abundant Skeletonema species in our record, and Skeletonema tropicum. This group displays strong increases over the light-laminated intervals (Figure 4c). The highest values occur during light-laminated interval 3 (between 1800 and 1550 cal. yr BP), where it contributes largely to the strongest peak in total diatom abundances. Two other large peaks in Skeletonema spp. are found in light-laminated intervals 6 (2960–2930 cal. yr BP) and 5 (2300–2200 cal. yr BP). Light-laminated interval 5 also stands out with high total diatom abundance. The variations in total diatom abundances were mostly driven by changes in Skeletonema spp. abundances.

In contrast, the percentages of tropical oligotrophic taxa displayed an opposite trend to the groups mentioned above (Figure 4d). This group here includes Fragilariopsis doliolus, Thalassiothrix longissima, Thalassionema bacillare, Thalassionema frauenfeldii, and Thalassionema nitzschioides var. parva as defined at the regional scale with Principal Component Analyses performed on the studied core as well as on trigger cores from the Peruvian margin (Fleury, 2015). Percentages in tropical oligotrophic diatoms increased between 3050 and 2500 cal. yr BP and decreased between 2500 and 2000 cal. yr BP, before reaching a minimum over light-laminated intervals 4 and 3 (i.e. 2100–1900 and 1800–1550 cal. yr BP respectively) (Figure 4d). They increased again between 1550 and 1100 cal. yr BP, except a sharp drop during light-laminated interval 2. Subsequently, percentages in tropical oligotrophic diatoms decreased until 900 cal. yr BP, followed by a third minimum observed during light-laminated interval 1 (900–500 cal. yr BP). The youngest part of the record was marked by a final increase in the contribution of the tropical oligotrophic diatom group.

Benthic species, which thrive on the substratum or live attached to it in coastal regions (Hasle and Syvertsen, 1996), are presented in Figure 4e. The dominant species in this group are Cocconeis spp. and Delphineis karstenii with minor contributions from Surirella fastuosa var. fastuosa and Trachyneis aspera. The total percentage of benthic diatoms remained low compared with the other groups, varying in general between 2% and 6%, although higher values (8% to 9%) were found between 2800 and 2600 cal. yr BP (Figure 4e). All light-laminated intervals were marked by a decrease in these diatoms. Relative abundances of the benthic diatoms thus displayed opposite variations compared with the total diatom abundances. On the opposite, benthic diatoms varied in parallel with tropical oligotrophic diatoms at the centennial timescale, the percentage of both groups decreasing in all light-laminated intervals.

In summary, diatom assemblages show a contrast between the periods older and younger than 2500 cal. yr BP. The older interval is characterized by lower diatom abundances and higher percentages in tropical oligotrophic and benthic diatoms compared with the last 2500 years. Shorter scale changes are materialized by six intervals of increased total diatom abundances, CRS, and Skeletonema spp. percentages as well as decreased contributions of tropical oligotrophic and benthic diatoms. All groups materialize these centennial-scale variations, suggesting that these changes were related to a parameter controlling the whole diatom community. However, tropical oligotrophic diatoms display an overwhelming millennial cycle, indicating that this group may have responded to additional environmental parameters.

Geochemical proxies

Iron (Fe) and Silicon (Si) contents measured with XRF core scanning display a similar pattern during the last 3000 years (Figure 5a and b). The highest values were observed between 3050 and 2500 cal. yr BP before a decrease found between 2500 and 2100 cal. yr BP. The last 2100 years were marked in both elements by variations of smaller amplitude around a stable mean value (25,000 counts per second (cps) for Fe and 12,500 cps for Si). All six light-laminated intervals are characterized by a decrease in the Fe and Si contents of the sediment. The Fe content especially varies in parallel with the percentage of benthic diatoms (Figure 4e).

Figure 5.

Geochemical records in piston core M772-003-2 (Peru margin, 15°S) over the last 3000 years: (a) Fe; (b) Si; (c) Si/Fe; (d) opal (Mollier-Vogel, 2012); (e) TN; (f) C_org; (g) total alkenone abundance; (h) alkenone-derived SSTs. The red triangles indicate the radiocarbon ages. The gray shaded areas represent the light-laminated intervals characterized by light sediments, high total diatom abundances, high Si/Fe ratios, and low Fe contents.

Si has both a lithogenic and a biogenic origin (diatoms, radiolarian, and several species of sponges) whereas Fe is mostly found in the lithogenic fraction of the sediment. In order to evaluate the variations in biogenic Si, we calculated the Si/Fe ratio (Figure 5c). The Si/Fe ratio was higher during the light-laminated intervals and lower in between these intervals. The low-resolution opal content record follows a very similar pattern over the last 3000 years (Figure 5d) here supporting the use of Si/Fe as a proxy for biogenic silica. The Si/Fe record is also in agreement with the total diatom abundance record both in timing and amplitude, indicating that diatoms were the main opal vector to the sea-floor in the PUS.

Organic carbon (C_org) and TN contents are high throughout the record, with C_org varying between 3% and 7.5% (Figure 5f) and TN reaching values between 0.3% and 0.8% (Figure 5e). They both displayed the same variations, especially large decreases during the light-laminated intervals. In contrast, such decreases were not observed during the oldest part of the record, where organic matter contents remained high. The resulting C_org/TN (not shown) ratio was between 8.5 and 11 over the whole record, indicating that the organic matter was predominantly of marine origin (Meyers, 1997). Such a result is consistent with previous studies on marine sediment cores from the Peruvian margin, all indicating C_org/TN ratios between 8 and 13 (Diaz-Ochoa et al., 2009; Gutiérrez et al., 2009; Mollier-Vogel, 2012; Morales et al., 2006; Wolf, 2002). Organic matter contents generally varied in opposition to indicators of siliceous productivity (Si/Fe, opal, and total diatom abundances) while they changed in parallel to Fe content.

Total alkenone concentration remained around 10,000 ng/g of sediment over the record, except during all light-laminated intervals where it decreased drastically, especially during intervals 3 to 1 (Figure 5g). Its variations mimic changes in organic matter contents (Figure 5e and f) and are opposite to total diatom abundances (Figure 4a). We inferred SSTs using the alkenone unsaturation measurements (Sachs et al., 2000 and references therein). SSTs increased between 3050 and 2700 cal. yr BP, reaching a maximum of ~23.3°C (Figure 5h). This was followed by a sharp decrease between 2700 and 2600 cal. yr BP and SSTs remained stable around 22.3°C until 1800 cal. yr BP. Then, SSTs showed a second increase lasting approximately 400 years. Finally, SSTs decreased continuously over the last 1400 years. In contrast with other geochemical proxies, SSTs did not display any change in phase with the light- or dark-laminated intervals (Figure 4d).

Discussion

Although our record allows for a decadal temporal resolution, the strongest changes are observed at the centennial to multi-centennial timescales. Inter-laminae variations are evident all through the records, but the differences between two successive laminae are of much lower amplitude than the changes observed between dark-laminated and light-laminated intervals. Weak inter-laminae fluctuations are caused by the succession of several laminae of similar composition (dark or light laminae) in our record (Figure 2), as described by Brodie and Kemp (1994) on another core from the Peruvian OMZ. The lack of alternating dark and light laminae suggests that the conditions in the PUS were relatively stable over successive decades. On the opposite, the alternation of dark- and light-laminated intervals represents strong changes at the centennial to millennial timescales. We thus focus on variations at these timescales in the following discussion.

Productivity variations at the millennial to centennial timescales

Millennial changes in productivity

Diatoms are a major biogenic component of coastal upwelling sediments and have been largely used to derive information on paleoproductivity variations and, hence, on upwelling activity variations on tectonic, orbital, and millennial timescales (e.g. Abrantes, 1991a, 1991b, 2000; Bernárdez et al., 2008; Lange et al., 1998; Romero et al., 2006). Off Peru, only a few studies have focused on diatom assemblages and mostly at low temporal resolution for the last climatic cycles (De Vries and Schrader, 1981; Schrader, 1992; Schrader and Sorknes, 1990, 1991).

Our data reveal high diatom abundances throughout the record, varying between 200 and 900x10⁶ valves/g of sediment (300 10⁶ valves per gram of sediment on average). Our diatom abundances are one to two orders of magnitude higher than the maximum abundance reported for the southern part of the HC off Chile (Romero and Hebbeln, 2003), but further offshore and at greater depths (outer shelf and upper slope), and thus certainly outside the main upwelling productive cell. They are however in agreement with diatom fluxes recorded in shelf cores off central and southern Peru over the last millennium (Gutiérrez et al., 2009). Our results are also of similar magnitude to values reported over the shelf off Namibia in the Benguela Current (Romero, 2010; Schuette and Schrader, 1981) and over the Antarctic continental shelves, where diatoms are the main primary producers (Bárcena et al., 2002; Crosta et al., 2008; Leventer, 1991; Stickley et al., 2005). Similarly, opal content in piston core M772-003-2 (35% on average) reaches values similar to the ones observed in other systems where diatoms are dominant (Crosta et al., 2005; Romero, 2010). Therefore, the huge diatom abundances and high opal content observed over the Peruvian shelf must result from strong primary production in the euphotic layer coupled to rapid export and burial of biogenic material helped by the shallow water depth. High export and burial efficiency is confirmed by the great percentage of Skeletonema spp., a dissolution sensitive diatom genus, in down-core M772-003-2 sediments. As previously observed in the laminated sediments of the Gulf of California (Schrader et al., 1980) and in the Benguela Current (Schuette and Schrader, 1981), the preservation of such weakly silicified diatoms must be related to extremely high productivity events associated with rapid burial.

Mean diatom abundances were lower in the 3000–2500 cal. yr BP interval (211 million valves/g of sediment) than in the 2500–1600 and the 1600–400 cal. yr BP interval, during which mean total diatom abundances are equal to 408 and 318 million valves/g of sediment respectively (Figure 6a). Although the former interval is only 500 years long, total diatom abundances presented very little variance. This interval is also characterized by high and stable Fe content, except for a very brief drop in the light-laminated interval 6 (2960–2930 cal. yr BP) (Figure 5a), coincident with high relative abundances of benthic diatoms (Figure 4e) and tropical oligotrophic diatoms (Figure 4d). These data suggest a period of increased precipitation on the nearby continent and a reduction of the Peruvian upwelling and expansion of the oligotrophic realm. The reduction of the upwelling between 3000 and 2500 cal. yr BP lessened nutrient fluxes to the surface and stabilized the surface layer of the ocean, generating conditions more favorable to the coccolithophorids, the second producers in the PUS after diatoms (De Mendiola, 1981). Indeed, alkenone concentrations were higher older than 2500 cal. yr BP (Figure 5g), indicating an increase in coccolithophorid production, triggered by the persistence of conditions more favorable to coccolithophorids over this period, that is, less water mixing at the surface (Nanninga and Tyrrell, 1996) and lower nutrient contents (Cavender-Bares et al., 2001) compared with the conditions favorable to diatoms (Ziveri et al., 1995). The contribution of coccolithophorids to total productivity was estimated through the normalization of total alkenone concentrations to organic matter contents. The contribution of alkenone concentrations normalized to TN (chosen here as presenting a much higher resolution than C_org) was globally higher in the 3000–2500 cal. yr BP interval than in the 2500–400 cal. yr BP interval although a pluri-decadal drop is observed at ~2700 cal. yr BP (Figure 6b). These results further suggest enhanced coccolithophorid production at the expense of diatom production during the 3000 to 2500 cal. yr BP period. Even though diatom abundances decreased over this period, abundances over 100 million valves/g of sediment indicate that an upwelling cell remained active. We propose that the increase in alkenone content in the 3000–2500 cal. yr BP period records a seasonal reduction of the intensity and/or extent of the upwelling whereby more recent periods (2500–400 Cal yr BP) are marked by a more perennial upwelling.

Figure 6.

(a) Total diatom abundances, (b) total alkenone abundance normalized to total nitrogen, and (c) benthic diatom percentages measured in piston core M772-003-2 (Peru margin, 15°S) over the last 3000 years; (d) zonal SST gradient in the equatorial Pacific Ocean (Conroy et al., 2009); (e) salinity reconstruction on sediments from Moon Lake, North Dakota, USA, 46°N (Laird et al., 1996); (f) SSTs measured on piston core M772-003-2, Peru margin, 15°S (this study); (g) SSTs measured on core GeoB 3313-1, Chile margin, 41°S (Lamy et al., 2002); (h) SSTs measured on core GeoB7186-3, Chile margin, 44°S (Mohtadi et al., 2007). The black lines in Figure 6f and g represent the smoothing of the corresponding lines. The red triangles indicate the radiocarbon ages. The gray shaded areas represent the light-laminated intervals characterized by light sediments, high total diatom abundances, high Si/Fe ratios, and low Fe contents.

Centennial changes in productivity

Mean total diatom abundances have been higher since 2500 cal. yr BP. Six periods of drastic increase in total diatom abundances are observed in piston core M772-003-2 over the last 2500 years (Figure 6a, gray bands). They correspond to the light-laminated intervals observed on the pictures and x-ray images of the core. These variations can result from changes in diatom productivity and/or in opal preservation. Indeed, pore-water silica concentration controls the solubility of opal and any change in dissolved silica may drive changes in the preservation of diatom frustules (Archer et al., 1993). However, well-preserved diatom frustules dominate the assemblages throughout the record, indicating relatively little transformation of the frustules once they were buried in the sediment. Light-laminated intervals were also characterized by strong increases in relative abundances of Skeletonema spp. and CRS (Figure 4b and c), two rapid blooming species groups, and by higher opal concentrations (Figure 5d). These observations suggest that these six periods were characterized by greater siliceous productivity boosted by more intense upwelling conditions, bringing more nutrients to the PUS.

Surprisingly, these intervals were also marked by lower organic matter contents (i.e. organic carbon, TN, and alkenone concentrations; Figure 5e –g). As Peruvian sediments are either terrigenous or biogenic (Agnihotri et al., 2008), total diatom abundances and total organic carbon content are diluted in the same way by variations in the delivery of riverine terrigenous material (Agnihotri et al., 2008; Rein et al., 2005). Thus, a dilution effect is certainly not an explanation for the differences observed between biogenic opal and organic matter. An alternative explanation could be provided by changes in the composition of the sediment. Previous studies from the Peruvian shelf indicated that the terrigenous mud laminae have higher organic carbon concentrations than the diatom ooze laminae (Kemp, 1994; Patience et al., 1990). These authors proposed that the organic matter originally associated with diatoms would be more prone to degradation compared with the biogenic material produced by other organisms (Patience et al., 1990), this process being reinforced by the high porosity (low density) of opal-rich sediments.

The contribution of alkenone concentrations normalized to TN congruently decreased during intervals of increased siliceous productivity (Figure 6a), suggesting reduced contributions of coccolithophorids during these periods. These organisms need lower amounts of nutrients compared with diatoms (Ziveri et al., 1995); therefore, decreased contributions of coccolithophorids support the idea of intensified upwelling and productivity within the light-laminated intervals.

High- and low-latitude forcings on the PUS

Low-latitude control on productivity variations at the centennial timescale

The centennial-scale alternation between periods of high and low productivity occurred in phase with the alternation of dry and humid periods traced with biotic and terrigenous indices. First, proxies of siliceous productivity (total diatom abundances, opal content) and upwelling intensity (percentages of Skeletonema spp. and CRS) vary in opposition with percentages in benthic diatoms (Figure 6c). These organisms thrive attached to the substratum in sub- to inter-tidal environments (Hasle and Syvertsen, 1996). The presence of these diatoms at the core location indicates transport from the land coast, driven by runoff (Sánchez et al., 2012). Even though the southern Peruvian coast is hyperarid today (Garreaud et al., 2009), studies on continental deposits have shown that this region received rainfall during the strongest El Niño events (Beresford-Jones et al., 2009; Vargas et al., 2006). Benthic diatoms can thus be here considered as a proxy for intense rainfall. Decreased benthic diatom percentages within the light-laminated intervals indicate that conditions were probably drier when productivity increased. This interpretation is coherent with the observation that periods of lower (higher) benthic diatom abundances (Figure 4e) correspond to decreases (increases) in silicoclastic material delivery, as indicated by decreased (increased) Fe content (Figure 5a). The observed negative correlation between benthic diatoms and productivity indices is thus coherent with the idea of variations in the PUS driven by changes in the mean state of ENSO. We consider the six light-laminated intervals, marked by enhanced productivity and reduced rainfall, as dominated by La Niña–like mean conditions while the rest of the record displays El Niño–like mean conditions, especially the period older than 2500 cal. yr BP, when both benthic diatom percentages and Fe content reached their highest values.

To test the hypothesis of a control by the mean state of ENSO, we compared our results with a record of the zonal SST gradient in the equatorial Pacific Ocean (Figure 6d; Conroy et al., 2009). The zonal SST gradient was stronger over the ‘Medieval Warm Period’ (MWP; 950–600 cal. yr BP; Stuiver et al., 1995), which is materialized in our record by light-laminated interval 1. Two other smaller amplitude increases in the zonal SST gradient are observed around 1050 and 1200 cal. yr BP, and both are synchronous with increases in total diatom abundances (Figure 6a). The zonal SST gradient in the Pacific Ocean thus increases in phase with productivity off Peru. A strong zonal SST gradient, indicative of strengthened Walker circulation, is observed today during La Niña events (Julian and Chervin, 1978). Increased zonal SST gradients in the Pacific Ocean over the MWP thus indicate the predominance of La Niña–like mean conditions during this period and to a lesser extent at ~1050 and ~1200 cal. yr BP. We note an anti-correlation between records of the zonal SST gradient and benthic diatom relative abundances (Figure 6c and d) suggesting dry conditions during periods of increased zonal SST gradients in the Pacific Ocean.

Our results indicate that La Niña–like conditions also prevailed during the RWP (1900–1550 cal. yr BP; Lamb, 1995) (light-laminated interval 3 in Figures 4 –6) in agreement with previous studies (Routson et al., 2011; Salvatteci et al., 2014). Conversely, light-laminated interval 2 (1350–1150 cal. yr BP) occurred during the Dark Ages Cold Period (DACP; 1150–1500 cal. yr BP; Lamb, 1995), contrasting with previous studies inferring low productivity during this period (Salvatteci et al., 2014).

We note that all intervals of increased productivity in the PUS coincide with increases in salinity in Moon Lake, located in the Great Plains of North America (Figure 6e; Laird et al., 1996), where La Niña events are marked with dry summers today (Hu and Feng, 2001), supporting some periods of sustained La Niña–like conditions off Peru during the DACP. The highest salinity in Moon Lake occurred during the 3000–2500 cal. yr BP period, which is dominated by El Niño–like conditions in our record. However, such high salinity values, based on lacustrine diatoms, were biased by poor diatom preservation in Moon Lake between 4700 and 2200 yr BP (Laird et al., 1996) and should be taken with great care. In addition, low salinities in Medicine Lake (Juggins et al., 1994) and Devils Lake (Fritz et al., 1991) between 5000–4500 and 2500 yr BP and a high lake level at Waldsea Lake around 3000 yr BP (Last and Schweyen, 1985) suggest the occurrence of wet summers in the Great Plains over the period earlier than 2500 yr BP, supporting the hypothesis of El Niño–like conditions at that time. The occurrence of three light-laminated intervals earlier than 2000 cal. yr BP (light-laminated intervals 4, 5, and 6 in Figures 4 –6) suggests that the succession of El Niño– and La Niña–like periods observed by Salvatteci et al. (2014) over the last 2000 years already prevailed as soon as 3000 cal. yr BP.

Some of the light-laminated intervals discussed in this paper, which we interpreted as resulting of dominant La Niña–like conditions, occurred when the frequency of El Niño events increased according to the results published by Moy et al. (2002). For example, the record published by Moy et al. (2002) indicates a maximum in El Niño activity over the MWP. This observation contradicts our record as El Niño events should promote lower primary productivity (Pennington et al., 2006) and increased runoff for the strongest events (Ortlieb and Macharé, 1993). However, the record published in Moy et al. (2002) traces the frequency of individual El Niño events while our record is a reconstruction of the mean conditions in the PUS. Even though most studies assume that a La Niña–like mean state is not favorable for the development of El Niño events (Clement et al., 1999), model studies indicate that the frequency of El Niño events can increase when the zonal SST gradient is stronger in the Pacific Ocean, indicating a La Niña–like mean state (Fedorov, 2002). A shallow thermocline in the Eastern equatorial Pacific, typical of La Niña–like conditions (Fiedler et al., 1992), leads to a strong thermocline slope in the Pacific Ocean. A strong thermocline slope potentially generates high amounts of available potential energy (Goddard and Philander, 2000), which are potentially converted into kinetic energy. Under such conditions, the thermocline is strongly sensitive to the Kelvin waves generated in the western equatorial Pacific if the trade winds are sufficiently weakened to enable the occurrence of westerly wind bursts (Fedorov, 2002). This mechanism could explain how intense El Niño events can be more frequent over periods dominated by La Niña–like mean conditions. Reconstructions of past changes in the intensity of the trade winds and in the depth of the thermocline at the centennial timescale are needed to test this hypothesis.

High-latitude control on SST variations at the multi-centennial timescale

SSTs are generally used as tracers of upwelling intensity in coastal upwelling systems (e.g. Chapman and Shannon, 1987) because upwelling brings cold subsurface waters to the surface (Huyer et al., 1987). As such, we would expect a decrease in SSTs when an increase in diatom abundances points to conditions of increased upwelling intensity. Indeed, a reduction in the upwelling intensity and extent, characteristic of El Niño events, is expected to be accompanied with the advection of warmer waters from the Eastern equatorial Pacific to the Peruvian margin (Barber and Chávez, 1986), leading to an increase in SSTs off Peru (Quinn et al., 1987). However, the SST record from piston core M772-003-2 (Figure 6f) did not vary in opposition with total diatom abundances from the same core (Figure 6a). SSTs off Peru were therefore controlled by other factors than the intensity of the upwelling over the last 3000 years. An alternative hypothesis would be that alkenone-based SSTs might have been biased by strong changes in seasonal contrast but our data do not allow for the reconstruction of these seasonal contrasts. Further analyses are required to test this hypothesis.

We here suggest that changes in the temperature of the water masses fueling the PUS were pivotal to explain our SST record. Indeed, surface waters off Peru are driven by the PCC or HC (Fiedler and Talley, 2006). This current originates from the mid-latitude Pacific Ocean off Chile (Strub et al., 1998), where the South American continent blocks the Antarctic Circumpolar Current (ACC). Parallel pluri-centennial variations in SSTs off Peru (this study, see the smoothed signal indicated by the black line in Figure 6f) and SSTs off southern Chile (Figure 6g; Lamy et al., 2002, see the smoothed signal indicated by the black line), where the PCC originates, support the idea of Peruvian SSTs being primarily controlled by the temperature of the PCC. SSTs off Peru and Chile decrease between 2800 and 2200 cal. yr BP, increase between 2200 and 1600–1400 cal. yr BP, and decrease over the last millennium. The decrease in SSTs off Peru observed over the last 1400 years occurs in parallel with a decrease in SSTs measured further south off Chile (Figure 6h; Mohtadi et al., 2007), pointing to a regional control on SSTs. SST records from the southeastern Pacific suggest a warming PCC between 2200 and 1400 cal. yr BP before a cooling trend over the last millennium.

A stronger and colder PCC is observed when the ACC intensifies (Lamy et al., 2002), which occurs in response to the intensification of the southern westerly winds (McDermott, 1996). Past changes in the position and intensity of the southern westerlies were driven by the extent and intensity of the zonal atmospheric cells (Bertrand et al., 2014). Mohtadi et al. (2007) suggested that the temperature of the PCC and the intensity of the Walker circulation were both controlled by latitudinal shifts of the South Pacific Anticyclone. They indicated a northward shift of the ACC through a decrease in SSTs off Chile and proposed that this shift intensified the Walker circulation through a northward shift of the South Pacific Anticyclone. However, neither SSTs off Chile (Figure 6g and h) nor SSTs off Peru (Figure 6f) vary in phase with any of the proxies we used to trace the intensity of the upwelling in the PUS (Figure 6a). Indeed SST records indicate continuous surface water cooling, probably as a result of a continuous northward shift of the ACC, while records of the intensity of the upwelling demonstrated strong oscillations. These observations argue against the hypothesis of a direct link between the ACC and the Walker circulation and rather suggest that past changes in the strength and temperature of the PCC were mainly driven by atmospheric circulation in the polar region without any major impact on the Walker cell.

SSTs off the Antarctic Peninsula (Etourneau et al., 2013) and atmospheric temperatures over Antarctica (Masson et al., 2000) decreased over the last millennium, indicating the intensification of the southern polar cell. The strengthening of the polar cell caused the latitudinal contraction of the southern westerlies wind belt (SWWB; Bertrand et al., 2014; Garreaud et al., 2013; Lamy et al., 2010). A contraction of the SWWB leads to the strengthening of the southern westerlies at the core of the SWWB, as observed today (Garreaud et al., 2009) and in the past (Lamy et al., 2010). Stronger westerly winds at the core of the SWWB intensify the ACC over the last millennium, leading to decreased SSTs off Chile and Peru. On the opposite, SSTs off the Antarctic Peninsula increased between 1800 and 1000 cal. yr BP (Etourneau et al., 2013), indicating weakened circulation over Antarctica. Under these conditions, the SWWB expanded, leading to a weakened ACC and warming PCC, as shown by an increase in SSTs off Chile between 1800 and 1000 cal. yr BP (Figure 6h). The coherence of SSTs off Peru with SST and temperature records from Antarctica and the Southern Ocean suggest that multi-centennial variations of SSTs off Peru were mainly forced by climatic changes in the southern high latitudes rather than changes in upwelling intensity driven by the alternation of El Niño–like and La Niña–like mean conditions.

Conclusion

The multiproxy study of piston core M772-003-2, conducted at the lamination-scale and decadal resolution, indicates that the PUS response over the last 3000 years was better expressed at the centennial to millennial timescales compared with the decadal timescale. Centennial-scale changes in productivity were driven by tropical climatic variations. Past changes in the Walker circulation were the main processes controlling upwelling intensity in the PUS and runoff in coastal Peru, with increased diatom productivity and decreased rainfall being observed over periods of intensified Walker circulation. Such intervals of increased productivity, considered as dominated by La Niña–like mean conditions, yet occurred over periods of increased El Niño event frequency. We thus propose that the Walker circulation and the frequency of El Niño events were disconnected at centennial timescales over the last 3000 years. Model studies suggested a control of El Niño events by parameters besides the depth of the thermocline and the intensity of trade winds (Fedorov, 2002), but there is at the moment no record suitable to test this hypothesis. Multi-centennial SST variations were remotely controlled by the intensity of the ACC. This high-latitude control on SSTs prevailed over ENSO-related SST variations.

Footnotes

Acknowledgements

Cruise M772 was carried out as part the German collaborative research program ‘Climate-biogeochemistry interactions in the tropical ocean’ (). We thank Stéphanie Desprat for providing help with CLAM program and for constructive discussions on the age model. Isabelle Billy provided help with XRF core scanning. We thank Karine Charlier and Loïc Thiao-Layel for helping with the measurement of nitrogen isotopic ratios. Elfi Mollier-Vogel provided the opal data. We thank Silvia Koch for helping with the analysis of the alkenones. Vincent Hanquiez prepared the map displayed in Figure 1. We thank Dr Matthias Hüls from the Leibniz Laboratory for Isotope Research and Dating in Kiel for providing the ¹⁴C ages.

Funding

The research leading to these results received funding from the European Union’s Seventh Framework Programme (FP7/2007–2013) under Grant 243908, ‘Past4Future, Climate change – Learning from the past climate’ and the German Research Foundation through Collaborative Research Centre 754 ‘Climate-Biogeochemistry Interactions in the Tropical Ocean’ ().

References

Abrantes

(1991a) Increased upwelling off Portugal during the last glaciation: Diatom evidence. Marine Micropaleontology 17(3): 285–310. DOI: 10.1016/0377-8398(91)90017-Z.

Abrantes

(1991b) Variability of upwelling off NW Africa during the latest Quaternary: Diatom evidence. Paleoceanography 6(4): 431–460. DOI: 10.1029/91PA00049.

Abrantes

(2000) 200 000 yr diatom records from Atlantic upwelling sites reveal maximum productivity during LGM and a shift in phytoplankton community structure at 185 000 yr. Earth and Planetary Science Letters 176(1): 7–16. DOI: 10.1016/S0012-821X(99)00312-X.

Abrantes

Lopes

Mix

. (2007) Diatoms in Southeast Pacific surface sediments reflect environmental properties. Quaternary Science Reviews 26: 155–169. DOI: 10.1016/j.quascirev.2006.02.022.

Agnihotri

Altabet

Herbert

. (2008) Subdecadally resolved paleoceanography of the Peruvian margin during the last two millennia. Geochemistry, Geophysics, Geosystems 9: Q05013. DOI: 10.1029/2007GC001744.

Antia

Koeve

Fischer

. (2001) Basin-wide particulate carbon flux in the Atlantic Ocean: Regional export patterns and potential for atmospheric CO2 sequestration. Global Biogeochemical Cycles 15(4): 845–862. DOI: 10.1029/2000GB001376.

Archer

Lyle

Rodgers

. (1993) What controls opal preservation in tropical deep-sea sediments? Paleoceanography 8(1): 7–21. DOI: 10.1029/92PA02803.

Aronés

Ayón

Hirche

. (2009) Hydrographic structure and zooplankton abundance and diversity off Paita, northern Peru (1994 to 2004) – ENSO effects, trends and changes. Journal of Marine Systems 78(4): 582–598. DOI: 10.1016/j.jmarsys.2009.01.002.

Avaria

Muñoz

(1987) Effects of the 1982–1983 El Niño on the marine phytoplankton off northern Chile. Journal of Geophysical Research 92: 14369–14382. DOI: 10.1029/JC092iC13p14369.

10.

Barber

Chávez

(1986) Ocean variability in relation to living resources during the 1982–83 El Niño. Nature 319: 279–285. DOI: 10.1038/319279a0.

11.

Barber

Smith

(1981) The role of circulation, sinking, and vertical migration in physical sorting of phytoplankton in the upwelling center at 15 S. In: Richards

(ed.) Coastal Upwelling. Washington, DC: American Geophysical Union, pp. 366–371.

12.

Bárcena

MÁ

Isla

Plaza

. (2002) Bioaccumulation record and paleoclimatic significance in the Western Bransfield Strait. The last 2000 years. Deep Sea Research Part II: Topical Studies in Oceanography 49(4): 935–950. DOI: 10.1016/S0967-0645(01)00132-1.

13.

Beresford-Jones

Lewis

Boreham

(2009) Linking cultural and environmental change in Peruvian prehistory: Geomorphological survey of the Samaca Basin, Lower Ica Valley, Peru. Catena 78(3): 234–249. DOI: 10.1016/j.catena.2008.12.010.

14.

Berger

Loutre

(1991) Insolation values for the climate of the last 10 million years. Quaternary Science Reviews 10: 297–317. DOI: 10.1016/0277-3791(91)90033-Q.

15.

Bernárdez

González-Álvarez

Francés

. (2008) Palaeoproductivity changes and upwelling variability in the Galicia Mud Patch during the last 5000 years: Geochemical and microfloral evidence. The Holocene 18(8): 1207–1218. DOI: 10.1177/0959683608096596.

16.

Bertrand

Hughen

Sepúlveda

. (2014) Late Holocene covariability of the southern westerlies and sea surface temperature in northern Chilean Patagonia. Quaternary Science Reviews 105: 195–208. DOI: 10.1016/j.quascirev.2014.09.021.

17.

Blaauw

(2010) Methods and code for ‘classical’age-modelling of radiocarbon sequences. Quaternary Geochronology 5(5): 512–518. DOI: 10.1016/j.quageo.2010.01.002.

18.

Blanz

Emeis

Siegel

(2005) Controls on alkenone unsaturation ratios along the salinity gradient between the open ocean and the Baltic Sea. Geochimica et Cosmochimica Acta 69(14): 3589–3600. DOI: 10.1016/j.gca.2005.02.026.

19.

Brink

Halpern

Huyer

. (1983) The physical environment of the Peruvian upwelling system. Progress in Oceanography 12(3): 285–305. DOI: 10.1016/0079-6611(83)90011-3.

20.

Brodie

Kemp

AES

(1994) Variation in detrital fluxes and formation of laminae in late Quaternary sediments from the Peruvian coastal upwelling zone. Marine Geology 116: 385–398. DOI: 10.1016/0025-3227(94)90053-1.

21.

Bull

Kemp

(1996) Composition and origins of laminae in late Quaternary and Holocene sediments from the Santa Barbara Basin. Geological Society, London, Special Publications 116(1): 143–156. DOI: 10.1144/GSL.SP.1996.116.01.13.

22.

Carr

Kearns

(2003) Production regimes in four Eastern Boundary Current systems. Deep Sea Research Part II: Topical Studies in Oceanography 50(22): 3199–3221. DOI: 10.1016/j.dsr2.2003.07.015.

23.

Carr

Strub

Thomas

. (2002) Evolution of 1996–1999 La Niña and El Niño conditions off the western coast of South America: A remote sensing perspective. Journal of Geophysical Research: Oceans (1978–2012) 107(C12): 29–31. DOI: 10.1029/2001JC001183.

24.

Cavender-Bares

Karl

Chisholm

(2001) Nutrient gradients in the western North Atlantic Ocean: Relationship to microbial community structure and comparison to patterns in the Pacific Ocean. Deep-Sea Research I 48: 2373–2395. DOI: 10.1016/S0967-0637(01)00027-9.

25.

Chapman

Shannon

(1987) Seasonality in the oxygen minimum layers at the extremities of the Benguela system. South African Journal of Marine Science 5: 51–62. DOI: 10.2989/025776187784522162.

26.

Chavez

Barber

(1987) An estimate of new production in the equatorial Pacific. Deep Sea Research 34: 1229–1243. DOI: 10.1016/0198-0149(87)90073-2.

27.

Chavez

Buck

Service

. (1996) Phytoplankton variability in the central and eastern tropical Pacific. Deep Sea Research II 43: 835–870. DOI: 10.1016/0967-0645(96)00028-8.

28.

Chazen

Altabet

Herbert

(2009) Abrupt mid-Holocene onset of centennial-scale climate variability on the Peru-Chile Margin. Geophysical Research Letters 36: L18704. DOI: 10.1029/2009GL039749.

29.

Clement

Seager

Cane

(1999) Orbital controls on the El Nino/Southern Oscillation and the tropical climate. Paleoceanography 14(4): 441–456. DOI: 10.1029/1999PA900013.

30.

Conroy

Restrepo

Overpeck

. (2009) Unprecedented recent warming of surface temperatures in the eastern tropical Pacific Ocean. Nature Geoscience 2: 46–50. DOI: 10.1038/ngeo390.

31.

Cowles

Barber

Guillen

(1977) Biological consequences of the 1975 El Niño. Science 195: 285–287. DOI: 10.1126/science.195.4275.285.

32.

Crosta

Koç

(2007) Diatoms: From micropaleontology to isotope geochemistry. Developments in Marine Geology 1: 327–369. DOI: 10.1016/S1572-5480(07)01013-5.

33.

Crosta

Denis

Ther

(2008) Sea ice seasonality during the Holocene, Adélie Land, East Antarctica. Marine Micropaleontology 66(3): 222–232.

34.

Crosta

Crespin

Billy

. (2005) Major factors controlling Holocene δ13Corg changes in a seasonal sea-ice environment, Adélie Land, East Antarctica. Global Biogeochemical Cycles 19(4): GB4029. DOI: 10.1029/2004GB002426.

35.

De Mendiola

(1981) Seasonal phytoplankton distribution along the Peruvian coast. In: Richards

(ed.) Coastal Upwelling (Coastal and Estuarine Sciences, 1.). Washington, DC: American Geophysical Union, pp. 348–356.

36.

De Vries

Schrader

(1981) Variation of upwelling/oceanic conditions during the latest Pleistocene through Holocene off the central Peruvian coast: A diatom record. Marine Micropaleontology 6(2): 157–167. DOI: 10.1016/0377-8398(81)90003-7.

37.

Diaz-Ochoa

Lange

Pantoja

. (2009) Fish scales in sediments from off Callao, central Peru. Deep Sea Research II 56: 1124–1135. DOI: 10.1016/j.dsr2.2008.09.015.

38.

Etourneau

Collins

Willmott

. (2013) Holocene climate variations in the western Antarctic Peninsula: Evidence for sea ice extent predominantly controlled by changes in insolation and ENSO variability. Climate of the Past 9(4): 1431–1446. DOI: 10.5194/cp-9-1431-2013.

39.

Federov

(2002) The response of the coupled tropical ocean-atmosphere to westerly wind bursts. Quarterly Journal of the Royal Meteorological Society 128: 1–23.

40.

Fiedler

Talley

(2006) Hydrography of the eastern tropical Pacific: A review. Progress in Oceanography 69: 143–180. DOI: 10.1016/j.pocean.2006.03.008.

41.

Fiedler

Chavez

Behringer

. (1992) Physical and biological effects of Los Niños in the eastern tropical Pacific, 1986–1989. Deep Sea Research Part A. Oceanographic Research Papers 39(2): 199–219. DOI: 10.1016/0198-0149(92)90105-3.

42.

Fleury

(2015) Laminations sédimentaires et variabilité climatique et océanographique haute-fréquence sur la marge péruvienne. PhD thesis, University of Bordeaux, Bordeaux.

43.

Fritz

Juggins

Battarbee

. (1991) Reconstruction of past changes in salinity and climate using a diatom-based transfer function. Nature 352: 706–708. DOI: 10.1038/352706a0.

44.

Garreaud

Vuille

Compagnucci

. (2009) Present-day South American climate. Palaeogeography, Palaeoclimatology, Palaeoecology 281: 180–195. DOI: 10.1016/j.palaeo.2007.10.032.

45.

Garreaud

Lopez

Minvielle

. (2013) Large-scale control on the patagonian climate. Journal of Climate 26: 215–230. DOI: 10.1175/JCLI-D-12-00001.1.

46.

Goddard

Philander

SGH

(2000) The energetics of El Niño and La Niña. Journal of Climate 13: 1496–1516. DOI: 10.1175/1520-0442(2000)013<1496:TEOENO>2.0.CO;2.

47.

Gutiérrez

Sifeddine

Field

. (2009) Rapid reorganization of ocean biogeochemistry off Peru towards the end of the Little Ice Age. Biogeosciences 6: 835–848. DOI: 10.5194/bg-6-835-2009.

48.

Hasle

Syvertsen

(1996) Marine diatoms. In: Thomas

(ed.) Identifying Marine Diatoms and Dinoflagellates. San Diego, CA: Academic Press, 385 pp.

49.

Higginson

Altabet

(2004) Initial test of the silicic acid leakage hypothesis using sedimentary biomarkers. Geophysical Research Letters 31: L18303.

50.

Feng

(2001) Variations of Teleconnection of ENSO and interannual variation in summer rainfall in the Central United States. Journal of Climate 14: 2469–2480. DOI: 10.1175/1520-0442(2001)014<2469:VOTOEA>2.0.CO;2.

51.

Huyer

Smith

Paluszkiewicz

(1987) Coastal upwelling off Peru during normal and El Niño times, 1981–1984. Journal of Geophysical Research 92(13): 14,297–14,307. DOI: 10.1029/JC092iC13p14297.

52.

Juggins

Battarbee

Fritz

(1994) Diatom/salinity transfer functions and climate change: An assessment of methods and application to two Holocene sequences from the northern Great Plains, North America. In: Funnel

Kay

RLP

(eds) Paleoclimate of the Last Glacial/Interglacial Cycle. Swindon: NERC Earth Sciences Directorate, pp. 37–41 (NERC Spec. Publ. 94/2).

53.

Julian

Chervin

(1978) A study of the Southern Oscillation and Walker circulation phenomenon. Monthly Weather Review 106: 1433–1451. DOI: 10.1175/1520-0493(1978)106<1433:ASOTSO>2.0.CO;2.

54.

Kemp

AES

(1994) Laminated sediments from coastal and open ocean upwelling zones: What variability do they record? In: Summerhayes

Emeis

Angel

. (eds) Upwelling in the Ocean: Modern Processes and Ancient Records. Chichester: Wiley & Sons, pp. 239–257.

55.

Kessler

(2006) The circulation of the eastern tropical Pacific: A review. Progress in Oceanography 69(2): 181–217. DOI: 10.1016/j.pocean.2006.03.009.

56.

Laird

Fritz

Grimm

. (1996) Century-scale paleoclimatic reconstruction from Moon Lake, a closed-basin lake in the northern Great Plains. Limnology and Oceanography 41(5): 890–902.

57.

Lamb

(1995) Climate, History and the Modern World. 2nd Edition. New York: Routledge.

58.

Lamy

Kilian

Arz

. (2010) Holocene changes in the position and intensity of the southern westerly wind belt. Nature Geoscience 3: 695–699. DOI: 10.1038/ngeo959.

59.

Lamy

Rühlemann

Hebbeln

. (2002) High and low-latitude climate control on the position of the southern Peru-Chile Current during the Holocene. Paleoceanography 17: 1028–1038. DOI: 10.1029/2001PA000727.

60.

Lange

Romero

Wefer

. (1998) Offshore influence of coastal upwelling off Mauritania, NW Africa, as recorded by diatoms in sediment traps at 2195m water depth. Deep Sea Research Part I: Oceanographic Research Papers 45(6): 985–1013. DOI: 10.1016/S0967–0637(97)00103–9.

61.

Last

Schweyen

(1985) Late Holocene history of Waldsea Lake, Saskatchewan, Canada. Quaternary Research 24(2): 219–234. DOI: 10.1016/0033–5894(85)90008–0.

62.

Leventer

(1991) Sediment trap diatom assemblages from the northern Antarctic Peninsula region. Deep Sea Research Part A. Oceanographic Research Papers 38(8): 1127–1143. DOI: 10.1016/0198–0149(91)90099–2.

63.

Locarnini

Mishonov

Antonov

. (2010) World Ocean Atlas 2009, volume 1: Temperature. In: Levitus

(ed.) NOAA Atlas NESDIS 68. Washington, DC: U.S. Government Printing Office, 184 pp.

64.

McDermott

(1996) The regulation of northern overturning by Southern Hemisphere winds. Journal of Physical Oceanography 26(7): 1234–1255. DOI: 10.1175/1520-0485(1996)026<1234:TRONOB>2.0.CO;2.

65.

McElroy

(1983) Marine biological controls on atmospheric CO2 and climate. Nature 302: 328–329. DOI: 10.1038/302328a0.

66.

Makou

Eglinton

Oppo

. (2010) Postglacial changes in El Niño and la Niña behavior. Geology 38: 43–46. DOI: 10.1130/G30366.1.

67.

Masson

Vimeux

Jouzel

. (2000) Holocene climate variability in Antarctica based on 11 ice-core isotopic records. Quaternary Research 54(3): 348–358. DOI: 10.1006/qres.2000.2172.

68.

Messié

Ledesma

Kolber

. (2009) Potential new production estimates in four eastern boundary upwelling ecosystems. Progress in Oceanography 83: 151–158. DOI: 10.1016/j.pocean.2009.07.018.

69.

Meyers

(1997) Organic geochemical proxies of paleoceanographic, paleolimnologic and paleoclimatic processes. Organic Geochemistry 27: 213–250. DOI: 10.1016/S0146-6380(97)00049-1.

70.

Migeon

Weber

Faugeres

. (1999) SCOPIX: A new X-ray imaging system for core analysis. Geo-Marine Letters 18: 251–255. DOI: 10.1007/s003670050076.

71.

Mohtadi

Romero

Kaiser

. (2007) Cooling of the southern high latitudes during the Medieval Period and its effect on ENSO. Quaternary Science Reviews 26: 1055–1066. DOI: 10.1016/j.quascirev.2006.12.008.

72.

Mollier-Vogel

(2012) Peruvian Oxygen Minimum Zone Dynamics During the Last 18 000 Years. PhD thesis, University of Kiel, Kiel.

73.

Mollier-Vogel

Ryabenko

Martinez

. (2012) Nitrogen isotope gradients off Peru and Ecuador related to upwelling, productivity, nutrient uptake and oxygen deficiency. Deep Sea Research Part I. Oceanographic Research Papers 70: 14–25.

74.

Morales

MDC

Field

Mayor Pastor

. (2006) Variaciones de foraminíferos de los últimos 460 años en sedimentos laminados de la plataforma continental peruana. Boletín Sociedad Geológica de Perú 101: 5–18.

75.

Moreno

Licea

Santoyo

(1996) Diatomeas del golfo de California. La Paz, Mexico: Universidad Autónoma de Baja California Sur, SEP-FOMES/PROMARCO, 274 pp.

76.

Moy

Seltzer

Rodbell

. (2002) Variability of El Niño/Southern Oscillation activity at millenial timescales during the Holocene epoch. Nature 420: 162–165. DOI: 10.1038/nature01194.

77.

Müller

Kirst

Ruhland

. (1998) Calibration of the alkenone paleotemperature index U₃₇^K′ based on core-tops from the eastern South Atlantic and the global ocean (60° N-60° S). Geochimica et Cosmochimica Acta 62(10): 1757–1772. DOI: 10.1016/S0016-7037(98)00097-0.

78.

Nanninga

Tyrrell

(1996) The importance of light for the formation of algal blooms by Emiliania huxleyi. Marine Ecology Progress Series 136: 195–203. DOI: 10.3354/meps136195.

79.

Ortlieb

Macharé

(1993) Former El Niño events: Records from western South America. Global and Planetary Change 7(1): 181–202. DOI: 10.1016/0921–8181(93)90049-T.

80.

Ortlieb

Vargas

Saliège

(2011) Marine radiocarbon reservoir effect along the northern Chile-southern Peru coast (14–24°S) throughout the Holocene. Quaternary Research 75: 91–103. DOI: 10.1016/j.yqres.2010.07.018.

81.

Patience

Clayton

Kearsley

. (1990) An integrated biochemical, geochemical and sedimentological study of organic diagenesis in sediments from Leg 112. In: Proceedings of the ocean drilling project, scientific results, vol. 112. College Station, TX: Ocean Drilling Program, pp. 135–153.

82.

Pennington

Mahoney

Kuwahara

. (2006) Primary production in the eastern tropical Pacific: A review. Progress in Oceanography 69: 285–317. DOI: 10.1016/j.pocean.2006.03.012.

83.

Philander

(1990) El Niño, La Niña and the Southern Oscillation. San Diego, CA: Academic press, 289 pp.

84.

Quinn

Neal

Antunez de Mayolo

(1987) El Niño occurrences over the past four and a half centuries. Journal of Geophysical Research 92: 14449–14461. DOI: 10.1029/JC092iC13p14449.

85.

Rathburn

Pichon

Ayress

. (1997) Microfossil and stable-isotope evidence for changes in the late Holocene paleoproductivity and paleoceanographic conditions in the Prydz Bay region of Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology 131: 485–510. DOI: 10.1016/S0031–0182(97)00017–5.

86.

Rein

Lückge

Reinhardt

. (2005) El Niño variability off Peru during the last 20,000 years. Paleoceanography 20: PA4003. DOI: 10.1029/2004PA001099.

87.

Romero

(2010) Changes in style and intensity of production in the Southeastern Atlantic over the last 70,000 yr. Marine Micropaleontology 74(1): 15–28. DOI: 10.1016/j.marmicro.2009.11.001.

88.

Romero

Hebbeln

(2003) Biogenic silica and diatom thanatocoenosis in surface sediments below the Peru-Chile Current: Controlling mechanisms and relationship with productivity of surface waters. Marine Micropaleontology 48: 71–90. DOI: 10.1016/S0377–8398(02)00161–5.

89.

Romero

Hebbeln

Wefer

(2001) Temporal and spatial distribution in export production in the SE Pacific Ocean: Evidence from siliceous plankton fluxes and surface sediment assemblages. Deep-Sea Research I 48: 2673–2697. DOI: 10.1016/S0967–0637(01)00037–1.

90.

Romero

Kim

Hebbeln

(2006) Paleoproductivity evolution off central Chile from the Last Glacial Maximum to the Early Holocene. Quaternary Research 65(3): 519–525. DOI: 10.1016/j.yqres.2005.07.003.

91.

Routson

Woodhouse

Overpeck

(2011) Second century megadrought in the Rio Grande headwaters, Colorado: How unusual was medieval drought? Geophysical Research Letters 38(22): L22703. DOI: 10.1029/2011GL050015.

92.

Sachs

Schneider

Eglinton

. (2000) Alkenones as paleoceanographic proxies. Geochemistry, Geophysics, Geosystems 1(11): 1035. DOI: 10.1029/2000GC000059.

93.

Salvatteci

Gutiérrez

Field

. (2014) The response of the Peruvian Upwelling Ecosystem to centennial-scale global change during the last two millennia. Climate of the Past 10(2): 715–731. DOI: 10.5194/cp-10–715–2014.

94.

Sánchez

Lange

González

. (2012) Siliceous microorganisms in the upwelling center off Concepción, Chile (36°S): Preservation in surface sediments and downcore fluctuations during the past 150 years. Progress in Oceanography 92: 50–65. DOI: 10.1016/j.pocean.2011.07.014.

95.

Sar

Sunesen

Castaños

(2001) Marine diatoms from Buenos Aires coastal waters (República Argentina). I. Thalassiosiraceae. Nova Hedwigia 73: 199–228.

96.

Sar

Sunesen

Lavigne

(2002) The diatom genus Thalassiosira: Species from the northern San Matías Gulf (Río Negro, Argentina). Nova Hedwigia 74: 373–386.

97.

Sarno

Kooistra

WHCF

Medlin

. (2005) Diversity in the genus Skeletonema (Bacillariophyceae). II. An assessment of the taxonomy of S. costatum-like species with the description of four new species. Journal of Phycologia 41: 151–176.

98.

Schrader

(1992) Comparison of Quaternary coastal upwelling proxies off central Peru. Marine Micropaleontology 19(1): 29–47. DOI: 10.1016/0377–8398(92)90020-K.

99.

Schrader

Sorknes

(1990) Spatial and temporal variation of Peruvian coastal upwelling during the latest Quaternary. In: Suess

von Huene

Emeis

K-C

. Proceedings of the Ocean Drilling Program, Scientific Results. College Station, TX: Ocean Drilling Program, 112 pp.

100.

Schrader

Sorknes

(1991) Peruvian coastal upwelling: Late Quaternary productivity changes revealed by diatoms. Marine Geology 97: 233–249. DOI: 10.1016/0025–3227(91)90118-N.

101.

Schrader

Kelts

Curray

. (1980) Laminated diatomaceous sediments from the Guaymas Basin slope (central Gulf of California): 250,000-year climate record. Science 207(4436): 1207–1209. DOI: 10.1126/science.207.4436.1207.

102.

Schuette

Schrader

(1981) Diatom taphocoenoses in the coastal upwelling area off South West Africa. Marine Micropaleontology 6(2): 131–155. DOI: 10.1016/0377–8398(81)90002–5.

103.

Siddall

Rohling

Almogi-Labin

. (2003) Sea-level fluctuations during the last glacial cycle. Nature 423(6942): 853–858. DOI: 10.1038/nature01690.

104.

Sifeddine

Gutiérrez

Ortlieb

. (2008) Laminated sediments from the central Peruvian continental slope: a 500-year record of upwelling system productivity, terrestrial runoff and redox conditions. Progress in Oceanography 79: 190–197. DOI: 10.1016/j.pocean.2008.10.024.

105.

Stickley

Pike

Leventer

. (2005) Deglacial ocean and climate seasonality in laminated diatom sediments, Mac. Robertson Shelf, Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology 227(4): 290–310.

106.

Strub

Mesias

Montecino

. (1998) Coastal ocean circulation off Western South America (coastal segment). In: Robinson

Brink

(eds) The Global Coastal Ocean. New York: Wiley, pp. 273–313.

107.

Stuiver

Grootes

Braziunas

(1995) The GISP2 N18O climate record of the past 16,500 years and the role of the sun, ocean and volcanoes. Quaternary Research 44: 341–354. DOI: 10.1006/qres.1995.1079.

108.

Sundström

(1986) The marine diatom genus Rhizosolenia: A new approach to the taxonomy. PhD thesis, University of Lund, Lund.

109.

Sunesen

Hernández-Becerril

Sar

(2008) Marine diatoms from Buenos Aires coastal waters (Argentina). V. Species of the genus Chaetoceros. Revista de Biología Marina y Oceanografía 43: 303–326.

110.

Thomas

Blanco

Carr

. (2001) Satellite-measured chlorophyll and temperature variability off northern Chile during the 1996–1998 La Niña and El Niño. Journal of Geophysical Research 106(C1): 899–915. DOI: 10.1029/1999JC000052.

111.

Toggweiler

Dixon

Broecker

(1991) The Peru upwelling and the ventilation of the South Pacific thermocline. Journal of Geophysical Research 96: 20467–20497. DOI: 10.1029/91JC02063.

112.

Vargas

Rutllant

Ortlieb

(2006) ENSO tropical–extratropical climate teleconnections and mechanisms for Holocene debris flows along the hyperarid coast of western South America (17–24 S). Earth and Planetary Science Letters 249(3): 467–483. DOI: 10.1016/j.epsl.2006.07.022.

113.

Wells

(1990) Holocene history of the El Niño phenomenon as recorded in flood sediments of northern coastal Peru. Geology 18: 1134–1137. DOI: 10.1130/0091-7613(1990)018<1134:HHOTEN>2.3.CO;2.

114.

Wolf

(2002) Zeitlische Variationen im peruanischen Küstenauftrieb seit dem letzten glazialen Maximum-Steuerung durch globale Klimadynamik. PhD thesis, University of Kiel, Kiel.

115.

Wyrtki

(1981) An estimate of equatorial upwelling in the Pacific. Journal of Physical Oceanography 11: 1205–1214. DOI: 10.1175/1520–0485(1981)011.

116.

Ziveri

Thunell

Rio

(1995) Export production of coccolithophores in an upwelling region: Results from San Pedro Basin, Southern California Borderlands. Marine Micropaleontology 24(3): 335–358. DOI: 10.1016/0377–8398(94)00017-H.