The Chilean continental margin has distinct topographical, climatic, and oceanographic characteristics, modulating the primary productivity and chemical composition of the water column. This zone has been characterized as having several sites of high primary production (0.5−9.3 g C m^-2 d^-1; González et al., 1998; Daneri et al., 2000; Thomas et al., 2001) off Iquique (21°S), Antofagasta (23°S), Coquimbo (30°S) and Concepción (36°S), produced by the influx of waters enriched in nutrients forced by local winds. Differences can be established latitudinally. Upwelling is semi-permanent in the North of Chile (18–30°S) and seasonal in the south (Montecino and Lange, 2009). This productivity occurs close to the coast above the narrow continental shelf in the north, which allows the development of relevant fisheries (Escribano et al., 2004 and references therein). Toward the south, the continental shelf is wider (~36–37°S). The upwelling sustains high primary productivity rates (9.9–19.9 g C m^-2 d^-1) during the spring-summer period (Bernal et al., 1989; Fossing et al., 1995; Dellarosa, 1998; Daneri et al., 2000).

To the north (at 21°S) and off Peru, the OMZ is permanently present and can extend into the euphotic zone. In the case of northern Chile and southern Peru, it does not interface significantly with the benthic environment because of a narrow continental shelf (Helly and Levin, 2004). Off Central Peru (~11-14°S) and northern Chile (21-23°S), the OMZ extends over the shelf. It has been reported that during the 1997–98 warm ENSO event (El Niño), the appearance of the sediments was less reduced, in some cases, up to a depth of 10 cm, with positive redox potential values at the surface (Muñoz et al., 2004). The most evident response of the benthic fauna was a switch in species composition, deeper penetration of the fauna into the sediments, and increased body size of the organisms (Gutiérrez et al., 2006; Sellanes et al., 2007; Gutiérrez et al., 2008). Following this El Niño, the dissolved oxygen decreased gradually to reach normal low concentrations (< 0.5 ml L^-1), dropping well below 0.25 ml L^-1 due to repeatedly La Niña events during 1999, 2000, and 2001, reaching considerably lower oxygen concentrations than values reported in previous years (Gonzalez et al., 2007). The oxygen rise at the bottom directly affects biogeochemical cycles that can be observed several months after the event (Ulloa et al., 2001; Escribano et al., 2004).

We reanalyzed previously published information from six sedimentary records at four sites along the Chilean margin: Mejillones del Sur Bay (23°S), Inglesa Bay (27°S), Coquimbo Bay (30°S; Guanaqueros and Tongoy Bay), and Concepcion Bay (36°S) (Muñoz et al., 2012; Guiñez et al., 2014; Srain et al., 2015; Castillo et al., 2017; Muñoz et al., 2020; Valdés et al., 2021). The cores have been dated using short-lived radioisotopes (²¹⁰Pb, ²⁴¹Am) by gamma and alpha counting combined with AMS radiocarbon measurements. In the case of Concepción Bay, the geochronology obtained with lineal regression (Muñoz et al., 2012) was updated using the Clam 2.2 software. The background information for each sedimentary record is presented in Table 1 .

The sedimentological characteristics of the selected cores are summarized below for a better approach to proxy interpretation.

The core located at the northern site (23°S, Mejillones del Sur Bay) was composed of silts (11–15%) and clays (> 70%), presenting an increase in clays in recent times (~90%). The total organic (TOC) content was high, between ~4% in the base of the core, increasing to the surface, reaching 5.5–7%. The sediments were described as olive green, with low stratifications, only visible with X-Rays, and were recognized three sedimentary facies based on color and thickness. Therefore, low bioturbation was observed. The age lower limits for the units were 1880 and 1600 AD, respectively (Guiñez et al., 2014; Valdés et al., 2021).

At 27°S off Inglesa Bay, the sediments also exhibited laminations of alternating light and dark color bands, the latter corresponding to hemipelagic organic material. Four units were recognized based on these laminations. The age lower limits were 1826, 1311, and 1009 AD, with slumps between the last three units. The TOC content was between 4% and 7%, with decreasing values toward the present (Castillo et al., 2017).

The cores retrieved from the southern sites did not show lamination. Off ~30°S (Guanaqueros and Tongoy Bays) were composed of very fine sand and silt, with evidence of weak laminations and low bioturbation. They also contain debris from fish vertebrates and shells, commonly observed in sediments from the northern Chilean shelf areas (Milessi et al., 2005). The sediment colors varied between bays. Guanaqueros exhibited accentuated olive-green colors, whereas Tongoy displayed colors in the range of very dark grayish-brown to dark olive green. In this core, a slight increase in the mean grain size occurred in the last 1000 years owing to a major continental input, which also explains the colors found. In agreement with this, the TOC (%) was lower (1−1.5%) than that in Guanaqueros (< 5%) (Muñoz et al., 2020).

The cores from the shelf of Concepcion (36°S) were composed of dark olive green hemipelagic mud at the surface, followed by distinctive olive green layers alternating with grey clayey mud layers. Differences in colors were related to their composition; olive-green corresponded to low bulk density and higher TOC content, between 1% and 2%, while the gray layers were low in TOC and showed higher bulk densities. This composition is similar to that of the Coquimbo shelf, but the core of Concepcion has a thin gray layer at ~79 cm, corresponding to hemipelagic sediments. These sediments exhibited moderate bioturbation and were not typically laminated (Srain et al., 2015; Muñoz et al., 2020).

Water column denitrification has been indirectly inferred from δ¹⁵N in sedimentary organic materials (Altabet, 2005; Robinson et al., 2014). Under low subsurface oxygen in the water column (< 2 µmol O₂ L^-1), the partial NO₃ uptake and denitrification process resulted in increased values of δ¹⁵N-nitrate, imprinting higher δ¹⁵N in sinking organic particles. This signal has been used as a proxy for changes in oxic conditions at the bottoms (Lam et al., 2009; Farmer et al., 2021). Global δ¹⁵N variation has been associated with glacial cooling. It is described as a less intense suboxia in denitrification zones due to oxygen supply to the thermocline zone or reduced oxygen consumption due to reduced export production (Robinson et al., 2007 and references therein). This process marks periods of reduced denitrification associated with lower δ¹⁵N values during glacial periods and higher values during interglacial periods. This variability has been observed along the SE Pacific in shallower and deeper zones, establishing denitrification zones and their regional variations in surface sediments (De Pol-Holz et al., 2006; Robinson et al., 2007; De Pol-Holz et al., 2009; Mollier-Vogel et al., 2012). Here, we used this tool as a proxy for denitrification intensity, which describes the subsurface variability of oxygen content over the continental shelf.

A few redox-sensitive metals have been used to reconstruct the redox conditions of depositional environments. Under oxygen-deficient subsurface water conditions (reducing conditions), metals show a higher enrichment compared with normoxic conditions (Tribovillard et al., 2006; Bennett and Canfield, 2020; Sánchez et al., 2022). However, the availability of reactive organic material in the column may also control the accumulation and enrichment of redox-sensitive metals on the seafloor (Nameroff et al., 2002; McManus et al., 2006; Algeo and Li, 2020). In the SE Pacific, redox-sensitive metals, such as Mo, Re, and U, have demonstrated moderate to strong enrichment behavior, especially in coastal areas associated with the OMZ (Böning et al., 2004; Böning et al., 2005; Böning et al., 2009; Muñoz et al., 2012; Salvatteci et al., 2014; Valdés et al., 2014; Castillo et al., 2019; Valdés et al., 2021; Muñoz et al., 2022). Here, we used redox-sensitive elements (U and Mo) to interpret the oxidative changes on the bottoms related to oxygen availability, using TOC and stable isotope distribution for complementary interpretations.

The latitudinal distribution of δ¹⁵N does not describe a clear pattern ( Figure 2 ). The values vary between ~1.2 and ~13‰, being higher in northern areas (23–27°S). The range of variability of the isotope was narrow for Mejillones and Coquimbo cores, while the cores at 27°S and 36°S showed more variability. In the case of Concepcion, the period of lowest values is concomitant with that of the lowest values for TOC (i.e., ~400 AD), corresponding to sandy sediment below the gray layer. Off-Mejillones (23°S), the maximum values were at the oldest periods (~1400 AD), and in Inglesa Bay, the maxima were around 900−1100 AD. Meanwhile, at the southern sites (30°S−36°S), δ¹⁵N values peak during the recent period (~9‰−10‰).

Similarly, TOC content did not exhibit a clear latitudinal pattern. The maximum TOC values were in Mejillones (21°S) with high variability, showing an increasing trend in recent times; from 1950 to 2000, the TOC reached 6.5% (~0.5%/decade). Off Coquimbo (30°S), the TOC values were lower and similar amongst the different cores, ~1% and 3% in Tongoy and Guanaqueros Bay, respectively. Other sites showed maximum values between 1000 and 1400 AD, around 6% and 3% off Inglesa Bay and Concepcion, respectively. After this age, the TOC content tended to diminish (~3% and ~1%, respectively) by ~1950−1960. After this date, most sites evidence an increase, except for Inglesa Bay (27°S), which indicates a gradual decrease in the recent period from ~1500 AD.

For a better approach, δ¹⁵N related to the denitrification process was standardized to establish its variability at each site ( Figure 3 ). Standardization consisted of estimating the variability around the mean value at each core according to (value-average)/standard deviation. This approach resulted in positive and negative values that helped visualize the intensity of the processes at each site, independently of the latitudinal differences in the proxy concentrations. Figure 3 indicates that, between 0 and 500 AD in Concepcion (36°S), the standardized δ¹⁵N evolved from low positive to higher negative values, corresponding to less intense denitrification. In contrast, Coquimbo (30°S) showed high negative values that grew to small positive values, which means that the denitrification was increasing, being slightly more intense at the end of the period. There is no information to compare to northern sites.

Between 500 and 1100 AD, Inglesa Bay (27°S) and Concepcion (36°S) displayed a similar trend. The δ¹⁵N evolved from strongly negative or “neutral” to positive values; both sites indicated increasing denitrification, peaking at ~900−1000 AD.

Between 1100 and 1500 AD, denitrification was more intense in Mejillones (23°S, only data from ~1300 AD) and Inglesa Bay (27°S), diminishing to the southern sites (36°S) during the same period (there is a gap in information from Coquimbo). The cores of Concepcion showed low variability, low positive values around the mean value. On the other hand, the intensity in denitrification decreased toward 1500 AD in Mejillones, while in Inglesa Bay, this is less clear due to a gap of information in this period, only few data indicated small negative values occurred around 1500 AD, suggesting less intensity in this process.

Between 1500 AD and ~1900 AD, Mejillones showed the lowest intensity in denitrification (higher negative values). Similar to Coquimbo, mostly negative values indicated a diminishing denitrification process; however, only a few pieces of data were available for this period. For Concepcion, the values were also negative but close to the mean intensity. No information is available at Inglesa Bay for this period.

After 1900 AD, Mejillones (23°S) showed alternating positive and negative values with a magnitude similar to the other sites, indicating high variability in the intensity of denitrification. In Inglesa Bay (27°S), they were consistently negative or less intense denitrification. On the other hand, in Coquimbo (30°S), both cores showed an increasing trend to higher positive values indicating an intensified denitrification. Concepcion (36°S) evolved from negative values until ~1970 to very high positive values at the current time.

We further explored the relationship between Mo and TOC ( Figure 4 ) since Mo indicates reduced conditions closely related to organic matter remineralization in zones with high productivity. Also, we analyzed the relationship between U and Mo ( Figure 5 ) in order to characterize the flux of organic particles. There is a clear linear relationship between U and Mo, independent of the temporal scale of each core (see Table 1 ). This relationship was more robust in the northern zones (23–27°S; r2 = 0.78–0.81; p < 0.0001) and decreased towards the southern sites, where the relationship between U and Mo weakens but still remains significant (36°S; r² = 0.24−0.33). TOC was higher in Mejillones (23°S) with maximum values reaching 6–7%, and the relationship with Mo showed a negative correlation (r² = 0.45). To the south, the TOC values decreased to 3% at Concepción with a positive relationship with Mo ( Figure 4 ), but with a weaker value of correlation (36°S; r² = 0.13−0.26), probably caused by an increase in detrital inputs that spoilt the relationship between sedimented organic matter and metal precipitation.

These EFs were reported as the amount of metal over the crustal concentrations determined in each location (Muñoz et al., 2012; Castillo et al., 2017; Castillo et al., 2019; Valdés et al., 2021) ( Table 2 ). The use of enrichment factors considering local detrital supply has been criticized because several processes in sediment transport may reduce detrital sedimentation. It has been proposed that Me/Al is a better approach (Bennett and Canfield, 2020). However, after reanalyzing the data, the Mo/Al ratio (in µg g^-1/%) showed similar behavior to using the EFs, reaching higher values in Mejillones (20–70) and decreasing in southern areas (< 3). Considering the standard EF calculation, the enrichments were high in Mejillones (23°S), one order of magnitude higher (~50−300) than at the southern sites, which was accompanied by a decrease in TOC content. South of 23°S, the EF values peaked to ~25 in Coquimbo and were as low as 3 in Inglesa Bay. This enrichment is probably underestimated because it was calculated using a very high detrital value measured in aeolian dust, enriched in metals by the presence of phosphorites in the zone ( Table 2 and Figure 4 ).

This study used metal ratios to estimate the variability at each site, showing changes that resulted from the redox condition variability at the bottom over time along the margin.

Similar to δ¹⁵N, the metal concentrations were standardized. This information is not complete for all sites ( Figure 6 ), but a similar palpable trend of lower concentrations (below the mean value) of U and Mo was observed from ~1400 AD to the present, suggesting less suboxic conditions, except for Concepcion (36°S) at the southern tip of our study area where the OMZ was more intense between 1900 and 1970 AD, which was not observed in the northern areas. Between 1400 and 1600 AD, significant anoxic conditions in the north sites (23–27°S) were not observed at 36°S, indicating a less intense OMZ. The few data points off Coquimbo also reveals a less severe suboxic environment. Between 800 and 1400 AD, these suboxic conditions were somewhat variable. In the case of 27°S, less intense suboxic conditions covered the period spanning between around 1300 to 600 AD. Both elements suggest less anoxic conditions. Before 800 AD, only the Concepcion site benefits from the complete data. They indicate less intense suboxic conditions between 100 and 600 AD, peaking between 300 and 500 AD. A few data available in Coquimbo 30°S) suggest that this condition was the opposite in this period because U continued to show moderate intensification although mostly negative values of Mo, implying that the sulfidic conditions necessary for Mo precipitation were weak.

Differences in δ¹⁵N were established between the northern and southern sites of the Chilean margin, showing a slight latitudinal decrease to the south ( Figure 2 ). The variability in temporal scale at the northern sites (23°S-27°S) were ~3-4‰, in a similar range to that observed in Peru, with a maximum reaching 8–9‰ and a minimum of ~5‰ (Gutiérrez et al., 2009). At the southern sites, the δ¹⁵N was lower, between 10‰ and−5‰, except in the sand layer in the core off Concepcion (<3‰), establishing a latitudinal decreasing trend of δ¹⁵N toward higher latitudes. This trend could be interpreted as a dilution effect by the increasing input of organic matter from rivers, characterized by low δ¹⁵N (2–3‰)(Meyers, 1997; Calvert et al., 2001), which would be absent in the north (23–27°S). However, the organic matter in sediments along the SE Pacific has been described as predominantly marine (De Pol-Holz et al., 2009; Mollier-Vogel et al., 2012). The continental effect would be relevant only in the period corresponding to the sand layer of Concepcion´s core. A better explanation of this light latitudinal gradient may lie from differences in nitrate concentrations and the uptake and assimilation of this nutrient by phytoplankton along the margin (Robinson et al., 2007; De Pol-Holz et al., 2009; Sánchez et al., 2022). The denitrification process domain at shallower depths and any variability could be caused by differences in the utilization of nitrate, such that, north of ~30°S, this nitrate consumption is complete compared with the southern sites, where it could be partial (De Pol-Holz et al., 2009). This signal is translated to the sediments, where the shelf sediments represent the composition of the sinking organic matter favored by the high sedimentation rates (0.24 – 0.15 cm/yr; Muñoz et al., 2004). Therefore, diagenetic alterations would not be relevant for δ¹⁵N, this being a robust proxy in sedimentary records to reconstruct denitrification intensity (Robinson et al., 2007; De Pol-Holz et al., 2009). On the other hand, latitudinal increases from Peru to northern Chile were attributed to differences in upwelling and progressive δ¹⁵N enrichment associated with the southward flux by the PCUC, which explained the increased values from Peru to Mejillones Bay (23°S) (Vargas et al., 2007; Sifeddine et al., 2008; Gutiérrez et al., 2009). In contrast, the southern sites of Chile showed a decreasing trend that could be attributed to a change in the upwelling pattern, being seasonal off 36°S and almost permanent north of Coquimbo. Therefore, the influence of the denitrification zone would be reduced in the southern sites (Vargas et al., 2007; Sánchez et al., 2012).

Similarly, the TOC content ( Figure 2 ) slightly decreased at the southern sites. The highest values (4–6%) were found in Mejillones, dropping to 3–4% in Inglesa Bay and Coquimbo and off Concepcion, with values ranging between 2% and 3%, except in the sand layer (< 1%). This observation suggests differences in the continental runoff over the shelves, particularly along the coast of Concepcion. The rivers off Concepcion are the primary particle sources that reduce the TOC in the bulk content. However, this contradicts previous analysis of δ¹³C, which indicates that the organic carbon source is controlled by marine productivity (Hebbeln et al., 2000; De Pol-Holz et al., 2009). However, it has increased the influence of freshwater species on the diatom compositions preserved in sediments in recent times (Sánchez et al., 2012).

The sedimentation rates (SR) estimated in the sediment cores of this study were in the range of values reported previously. Off Concepcion, it is higher (0.27 cm/yr) than the north of 30°S (0.04 and 0.11 cm/yr, in Mejillones and Coquimbo, respectively). However, the organic carbon accumulation rate shows an opposite trend. Off the Mejillones Bay, the value is ~105 gC/cm²yr, and in the southern sites, it is ~38.9 – 24.50 gC/m²yr. Therefore, the differences in organic carbon (OC) accumulation at our sites could result from differences in the detrital runoff. However, it could also respond to enhanced organic preservation in the north because the OMZ is thicker and less affected by seasonal variability (Valdés et al., 2004; Valdés et al., 2009). The buried OC results from organic matter that escapes oxidation referred to as burial efficiency (Middelburg et al., 1993), which depends on oxygen availability and sedimentation. Here, differences in SR should not be enough to explain differences in OC accumulation. Therefore, the differences in its accumulation could be attributed to oxygen availability and primary productivity fluxes at the bottom. Other factors relevant to C preservation include surface area reactivity, grain size composition, and organic carbon reactivity (Burdige, 2007; Katsev and Crowe, 2015), which should be considered but have not been evaluated. The OC content variability over time followed similar patterns at all sites. Considering a larger timescale (> 2000 cal BP) for Coquimbo and Concepcion (Muñoz et al., 2012; Muñoz, et al. 2020), the TOC content showed a slowly decreasing trend after 1400 AD toward the beginning of the Common Era. However, after the 1970s ( Figure 2 ), it rapidly increased, being markedly high in Mejillones and Concepcion, corresponding to the north and south ends of our study area. This change is coincident with the climate shift of the 1970s (Jacques–Coper and Garreaud, 2014), described as a warming shift on the east coast of South America concomitant with a weakening of the subtropical South Pacific High (SPSH). It would be condusive of a decreased organic flux by diminished primary productivity. Notwithstanding, an intensification of coastal upwelling was described for Mejillones (Vargas et al., 2007) and the Peruvian continental margin (Gutiérrez et al., 2011; Salvatteci et al., 2014), explaining the enhanced TOC trend, although others identified this trend earlier, around 1878 (Flores-Aqueveque et al., 2015). These studies identified enhanced wind stress in particular years, suggesting an interplay between inter-annual and decadal variability, which also had an incidence in the manifestation of the 1970s shift (Jacques–Coper and Garreaud, 2014) and would also explain the increased TOC off Concepcion. Even with increased runoff with low metabolizable organic material contributing to enhancing the organic content of these sediments, the δ¹³C still indicates a marine phytoplankton source (Sánchez et al., 2012). These authors state that the increased river inputs would be a nutrient source that overcompensates for the diminished nutrient inputs from the impoverished upwelling.

The range of the mass accumulation rates was not highly dissimilar. Despite the contribution of rivers in the south that support fresh organic C (highly reactive matter) to shallow depths, phytoplankton sedimentation dominates. Therefore, they maintain the isotopic composition of the settled organic material. The δ¹⁵N in the sediments is related to denitrification, linked to the available oxygen content over the bottom. The standardized δ¹⁵N data show better the variability of this process over the shelf despite the narrow range of the isotopic values ( Figures 2 , 3 ). In general, less intense denitrification, peaking at 300 and 600 AD off Concepción (36°S) and Inglesa Bay, respectively, occurred predominantly in a cold period (Dark Age Cold Period, DACP). However, lower values were also observed in the previous warm period (Roman Warm Period, RWP). Thus less intense denitrification was set before 800 AD; afterward, all sites exhibited enhancement of denitrification (more positive values), peaking during ~1100-1400 AD. The maximum values corresponded predominantly to a warm climate in the medieval climate anomaly (MCA) and the current warm period, while minimum values were during the cold period between 1600 and 1750 AD when the little ice age (LIA) arose. During this cold period, denitrification was less intense, except off Coquimbo (30°S), which has remained positive but more intense in the current time. Previous periods (< 1400 AD) at 23°S were insufficient for comparison. Thus, attenuation of the denitrification process during warmer periods (after 1400 AD) and a less intense OMZ was observed, implying a higher variability of oxygen conditions in the bottom waters. In the last 50 years, the data indicated a more intense denitrification, suggesting a more reduced environment, similar to most sites along the Chilean margin (except at 27°S). These conditions agree with the increase in C accumulation in recent sediments and more reduced conditions on the shelf caused by the intensification of the OMZ, promoting also intensified denitrification. Thus, changes in denitrification could be related to the variability in the upper limit of the OMZ.

U and Mo enrichments (authigenic precipitations) in reduced sediments have been widely used to interpret past redox conditions. U accumulation rate is highly correlated with the organic flux to the ocean bottom, where the precipitation of this element within the sediments could be microbially mediated (Zheng et al., 2002). Still, a substantial fraction can also be lost to the water column during sediment perturbations that affect the redox conditions and release U in the porewater. The U cycling would be similar to Fe cycling, having a similar redox reduction boundary (Zheng et al., 2002). Instead, Mo precipitation requires a sulfidic environment with concentrations >11 µM (Erickson and Helz, 2000). However, in sediments below suboxic waters, Mo-sulfide and Fe-sulfide coprecipitate at very low sulfide concentrations ~ 0.1 µM (Zheng et al., 2000), producing authigenic metal precipitation. Otherwise, Mo could also form intermediate complexes with dissolved organic matter in the bottom waters and sediments in suboxic conditions, contributing to sequestering Mo during early diagenesis (Wagner et al., 2017). This process could be relevant in places where the high primary productivity promotes higher organic flux to the bottoms, as described for the HCS; thus, several pathways can enhance Mo concentrations over the lithogenic background within the sediments. Instead of Mo, for U authigenic enrichment, an oxygen penetration of < 1 cm is necessary to maintain suboxic conditions within the sediments and U in a precipitated form (Morford and Emerson, 1999). Therefore both elements help determine the intensity of the redox changes, establishing suboxic and anoxic conditions.

In the Chilean margin, higher EFs were estimated for Mo over U concentrations, which showed a positive and significative correlation ( Figures 4 , 5 ), except at Concepcion, probably due to enhanced input of detrital material (r² = 0.24-0.33). This situation indicates that the sulfide concentrations were sufficient for Mo precipitation, which are between < 10 and 30 µM off Concepcion (36°S; Muñoz et al., 2012b), and reports for Coquimbo (30°S) indicate similar values of ~10 µM and ~80 µM (Muñoz, unpublished data). There are no sulfide records off Mejillones (23°S) and Inglesa Bay (27°S); however, in both sites, the correlation between metal enrichment and sulfate reduction intensity is poor, and the intensity of this process does not control the metal enrichment by itself (Valdés et al., 2014; Castillo et al., 2019). Thioploca mats on the bottoms probably reduce sulfide availability, diminishing metal-sulfide formation (Valdés et al., 2014). Instead, the metal transport during the Fe reduction/oxidation cycle has been proposed as a relevant mechanism for Mo enrichment (Valdés et al., 2014). Off-Perú, models of Fe shuttle for Mo enrichments would apply for the Chilean margin, even in shallower zones, where Fe has a crucial control over metal accumulation below a nitrogen-rich OMZ (Scholz et al., 2017; Scholz, 2018). The relationship between Mo and U enrichment is presented in Figure 7 . It is based on the scheme proposed by Algeo and Tribovillard (2009) that establishes the range for metal enrichment in sulfidic, anoxic, and suboxic conditions. The U/Mo relation exceeds three times the seawater values in Mejillones, indicating a crucial particulate shuttle for metal enrichment here, but not in the other sites, which mostly fall in the suboxic condition area ruled by oxygen content in the bottoms and organic fluxes.

The organic fluxes measured in sediments are highly dependent on primary productivity and constitute an essential carrier phase for metal accumulation in surface sediments in the Chilean margin and Peru in deeper zones (McManus et al., 2006; Scholz, 2018). This relationship is negative in Mejillones and positive in the southern sites ( Figure 4 ), suggesting enhanced organic carbon preservation and reduced detrital siliciclastic input in Mejillones. In restricted basins, authigenic enrichment can limit Mo accumulation because the dissolved phase is determined by circulation. This situation diminishes Mo precipitation, producing a lower Mo/Al value related to higher TOC content within the sediments (Tribovillard et al., 2012), explaining the negative correlation. However, this would not apply to upwelling zones; instead, the negative relationship between TOC and Mo in Mejillones (23°S) could result from increased oxygenation. The presence of Fe oxides promotes Mo accumulation during the recycling of oxides, as discussed earlier, and reduces the organic content by remineralization ( Figure 4 ). Besides, it is essential to note that the concentrations of Mo and EFs are very high in Mejillones compared with the bays south of 23°S. Therefore, although the exact mechanism of this metal enrichment has not been established yet, the Fe shuttle is a plausible explanation (Valdés et al., 2014; Castillo et al., 2019).

The temporal evolution of U and Mo ( Figure 6 ) suggests that in Mejillones Bay (23°S), an increasing trend of oxygenation was sustained from ~1600 AD to 1900 AD. While at the southern sites, the trend has been maintained since 1400 AD. No records were obtained at 23°S before 1300 AD. Therefore, while the southern sites (30−36°S) showed cycles of more or less extended periods of suboxic conditions of ~600 years, in the northern areas (23−27°S), this periodicity would be reduced to ~400 years. On the other hand, a reducing condition is observed in all cores but diminishes over time.

Comparing the standardized data of metals and δ¹⁵N records, they were relatively in phase until ~1900 AD, suggesting more or less intensity in the OMZ and denitrification. Following this period, an increase in oxygenation does not imply a decrease in denitrification. For example, after 1970, in most sites (except for Inglesa Bay), the less intense OMZ coincided with an intensified denitrification, which appears contradictory. In that case, other variables as the mixing of water mass properties can be considered relevant to adjust the frequency and seasonality of the OMZ upper limit, consistent with the increased occurrence of strong El Niño events after the climate shift of the 70s (Capotondi and Sardeshmukh, 2017).