Spatial Distribution of Dissolved Methane Over Extreme Oceanographic Gradients in the Subtropical Eastern South Pacific (17° to 37°S)

Abstract

Methane (CH₄) is one of the most powerful greenhouse gases with the capacity to influence the Earth's radiative budget as well as contribute to atmospheric chemistry. Natural oceanic production makes up to ∼4% of the overall global CH₄ emissions, however, there is uncertainty around the accuracy of this value due to a lack of accurate measurements. Such is the case in the Subtropical Eastern South Pacific Ocean (SESP), a region with pronounced chlorophyll-a and oxygen gradients, which in turn affect the microbial CH₄ cycling. This study was conducted during spring-summer (2014–2016) in the SESP. The region (∼17°–37°S/71°–110°W) is separated into (i) eutrophic, (ii) mesotrophic, and (iii) oligotrophic areas, according to oceanographic and biogeochemical criteria. The SESP presents high CH₄ zonal variability with levels ranging from 0.63 to 33.4 nmol L⁻¹, corresponding to 29% and 1,423% saturation, respectively. High CH₄ concentrations (>1,000% saturation) are observed in the narrow eutrophic area subjected to coastal upwelling. These conditions clearly differ to those observed in the extended oligotrophic subtropical gyre (∼100% saturation). Furthermore, CH₄ also tends to accumulate in the mesotrophic area (with upto 1,423% saturation), where oceanographic conditions as stratification, mesoscale eddies and island mass effect could trigger the presence of a microbial biomass that may be able to induce CH₄ regeneration. The CH₄ efflux is estimated to be between 0.13 and 19.1 µmol m⁻² d⁻¹ (mean ± SD = 4.72 ± 4.67) and the SESP has an emission rate of ∼87.9 Gg CH₄ yr⁻¹.

Key Points

• CH₄ spatial distribution of in the subtropical eastern South Pacific shows strong heterogeneity

• Maxima in CH₄ accumulations occur in the mesotrophic area associated with eddies followed by coastal upwelling

• The subtropical eastern South Pacific region contributes with 87.9 Gg of CH₄ per year toward the atmosphere

Plain Language Summary

Methane is a potent greenhouse gas that escapes from different natural and anthropogenic sources into the atmosphere and thus accelerates climate change. Atmospheric CH₄ concentrations have risen 2.5 times since the beginning of the Industrial age. While much of this increase is attributed to human activities, natural sources can contribute between 35% and 50% of global CH₄ emissions, Aquatic environments as estuarine, coastal, and open sea systems make up to ∼4% of the overall global CH₄ emissions. Subtropical Eastern South Pacific Ocean (SESP), a region with different coastal (upwelling) and open sea (subtropical gyres) ecosystems and pronounced photosynthetic biomass and oxygen gradients. There, the majority of CH₄ is created by microorganisms whose activities depend on both oxygen and organic matter variables. The SESP presents high CH₄ spatial variability with levels ranging from 29% to 1,423% saturation (100% is equivalent to an equilibrium between the ocean and atmosphere). Thus, most of this region exchanges CH₄ to the atmosphere at rates between 0.13 and 19.1 µmol m⁻² d⁻¹ (mean ± SD = 4.72 ± 4.67) with a regional emission rate of ∼87.9 Gg CH₄ yr⁻¹. This value could highlight the role of the ocean as emitting gases and emphasize same processes at basin scale.

1 Introduction

Methane (CH₄) is the most important greenhouse gas after CO₂, globally accounting for ∼17% of the global radiative forcing of all greenhouse gases (IPCC, 2013). Also, it is emitted by both natural and anthropogenic sources. Natural sources include wetlands, gas hydrates, termites, wildfires, as well as fresh and oceanic waters. In the ocean, CH₄ is actively recycled, as microbial communities drive various metabolic pathways, leading to the production and/or consumption of CH₄. Additionally, different physical processes lead to gas transport and exchange across the air-sea interface (Sarmiento & Gruber, 2006).

Indeed, CH₄ concentration and distribution in the ocean is primarily driven by physical (e.g., temperature and salinity) and biogeochemical conditions. The latter includes microbial activities, which in turn depend on the trophic conditions (e.g., organic matter availability) as well as dissolved oxygen levels (Naqvi et al., 2010; Reeburgh, 2007). Moreover, particulate and dissolved organic matter act as a carbon source and an electron donor for various forms of respiration and fermentative processes that drive CH₄ regeneration (production). Meanwhile, oxygen acts as the universal electron acceptor and establishes if aerobic or anaerobic biogeochemical processes are taking place (Kiene, 1991).

All of these physical and biogeochemical variables are highly variable in the subtropical eastern South Pacific Ocean (SESP). In the coastal regions, upwelling brings cold and saline waters with high nutrient levels to the surface waters, which triggers nutrient fertilization, high chlorophyll-a (Chl-a) availability, as well as the presence of oxygen deficient waters (Naqvi et al., 2010; Thiel et al., 2007). In contrast, the subtropical gyre contains warm waters with significantly higher oxygen levels and undetectable nutrient and Chl-a levels (Lepère et al., 2009; Ras et al., 2008; Troncoso et al., 2018). Given these conditions, it is likely that CH₄ levels will respond to gradients, as it is closely related with microalgal dynamics (Bogard et al., 2014; Leon-Palmero et al., 2020), as well as with oxygen deficiency as a result of anaerobic methanogenesis, which in turn is fueled by organic matter (Reeburg, 2007).

The waters in the SESP are dynamic and follow a large scale anticyclonic circulation around the center of the subtropical gyre, with the Humboldt Current System (HCS) limited to the eastern section (Talley et al., 2011). At subtropical or midlatitudes (off northern and central Chile), the HCS alternates between surface-equatorward and subsurface-poleward currents that carry different water masses into the stretch of water within a few hundred kilometres off Chile (Strub et al., 1998). Superficially, the SESP basin is dominated by cold waters of sub-Antarctic origin (Sub-Antarctic water “SAAW”) in the southern region; and to the north by warm and more saline waters of subtropical origin (Subtropical water mass “STW”), which is formed around the Subtropical Front south of 5°S, occupying a large area of the Subtropical Pacific gyre (Wong & Johnson, 2003). The relative presence of these water masses are marked by the Subtropical Convergence (STC) of the South Pacific (30°S–40°S). The SAAW subducts below the STW at around 30°S–40°S, leading to Eastern South Pacific Intermediate Water (ESPIW) representing a shallow salinity minimum or SSM (Leth et al., 2004). This modified SAAW presents remarkably low salinity values (∼33.2) due to high spring precipitation and freshwater input, probably due to the input of summer glacial meltwater from the south (Karstensen, 2004).

Below the surface layer, a relatively saline, cold water mass of equatorial origin (ESSW) carries water with very low oxygen and high NO₃⁻ and PO₄³⁻ levels southward via the Peru-Chile Undercurrent (Blanco et al., 2001; Huyer et al., 1991; Silva et al., 2009). The oxygen deficiency of the ESSW, markedly distinct to the SAAW and the Antarctic intermediate water (AAIW), creates a notorious oxygen minimum zone (OMZ) (Fuenzalida et al., 2009), with near anoxic conditions at its core (∼100 m depth) at 20°S (Canfield et al., 2010). During its journey to the south, the OMZ is mixed and oxygenated, with hypoxic conditions occurring at ∼36°S (Carrasco et al., 2017; Fuenzalida et al., 2009). The ESSW reaches the ocean surface through coastal upwelling along the north and central coast of Chile (Strub et al., 1998; Thiel et al., 2007).

In addition, several Archipelagos are located in the SESP, for example, Desventuradas Islands, including the San Felix (26°15′S, 80°W) and San Ambrosio (26°20′S, 79°55′W) islands, Juan Fernández Archipelago with the Robinson Crusoe (33°40′S, 78°50′W), and Alejandro Selkirk (33°45′S, 80°45′W) Islands as well as the Rapa Nui Island (27°7′10″S, 109°21′W). These islands are all distributed from ∼600 to 3,500 km off the Chilean coast. It is currently understood that high biological production (high Chl-a levels) occurs around the Juan Fernandez archipelago (Andrade et al., 2014; Pizarro et al., 2006). This phenomenon, known as Island Mass Effect or IME (Doty & Oguri, 1956) seems to occur due to a nutrient increase in the euphotic zone, which in turn triggers biological productivity as a consequence of the interactions between mesoscale swirls, wind and island topography (Andrade et al., 2014; Hasegawa et al., 2009; Sangrà et al., 2001). The IME could induce natural iron enrichment from lithogenic material and complex mesoscale circulations that stimulate microbial activities, such as those associated with CH₄ cycling, as is observed in the Kerguelen Islands (Blain et al., 2001; Farías et al., 2015).

Finally, mesoscale structures (eddies, filaments, and fronts) should be considered in greenhouse gas cycling, as they are involved in the zonal export of high nutrient coastal waters, which also contain organic matter, and/or phytoplankton biomass (Gruber et al., 2011). These structures represent unique and relatively isolated environments with distinctive biological communities and chemical conditions (Cornejo et al., 2016; Silyakova et al., 2020; Weller et al., 2013). The SESP is characterized by cyclonic (CEs) and anticyclonic (AES) eddies, particularly in the coastal transition zone that connects the coastal upwelling zones with the oligotrophic oceanic waters (Hormazabal et al., 2004). These eddies have typical diameters of 150–300 km and an eddy lifespan >30 days (with mostly long lived anticyclonic eddies); these structures impose a zonal impact on the heat and salt budgets through lateral turbulent fluxes (Chaigneau & Pizarro, 2005; Chaigneau et al., 2011). Then, CEs and AEs, which have a water mass structure typical of their formation region, propagate eastward with movement velocities of few cm s⁻¹, from a combination of mean flow advection and self-propagation.

In addition, seamounts, ocean ridges, and oceanic islands in the SESP could also produce or intensify mesoscale activity, thereby increasing biological production in the surrounding area (Wang et al., 2018). Despite having several important marine ecosystems, the SESP is lacking oceanographic exploration, especially in terms of investigating CH₄ distribution and exchange with the atmosphere. Thus, this study aims to analyze the CH₄ distribution at the basin scale, in order to clarify whether these described oceanographic processes interfere with distribution patterns, and additionally improve understanding around potential production or consumption processes.

2 Material and Methods

2.1 Study Area

Sampling was carried out during three cruises on board the Chilean R.V. Cabo de Hornos; the CIMAR 21 cruise between Caldera (27°S; 70.88°W) and near Rapa Nui Island (27.04°S; 109.31°W) from October 11th to November 11th, 2015, with a total of 19 oceanographic stations, including 15 stations with CH₄ data. The CIMAR 22 cruise, which implemented two main zonal transects, the first one from the coast of Caldera (27°S, 70.87°W) to the Desventuradas islands (DI) (26.16°S, 80.12°W); the second corresponding to a zonal transect from Valparaiso (32.97°S, 72.41°W) to the Juan Fernandez Archipielago (JFA) (33.63°S, 78.85°W) from the 12^th of October to the 15^th of November 2016, with 31 oceanographic stations, including 26 stations with CH₄ measurements.

Additional cruises were included with the objective of describing the surface CH₄ distribution; the Galathea cruise off northern Chile and Peru took place from 8^th February to 4^th March 2007; and from the Front Upwelling.

Cruise (FDS) off central Chile (36.30°S–36.45°S/73.06°W–73.30°W) on board the R.V. Abate Molina carried out from the 2^nd to the 7^th February 2014. On the FDS cruise, two transects were set, including 27 oceanographic stations with 14 stations including CH₄ data. This covered a narrow extension (100 km) off the Chilean coast and complemented the previously mentioned cruises, as it had an appropriate spatial coverage and synoptic scale. Figure 1 presents the location of all these transects and sampling stations, and includes topographic (islands) and bathymetric characteristics. More information about these cruises is included in the supplementary Table S1.

Figure 1

(a) Map showing the location of sampled stations during the CIMAR 21 cruise (blue circle) between Caldera (27°00′S; 70°52′W) and near Rapa Nui Island (27°10′S, 105°34′W) from October 11^th to November 11^th, 2015, and the CIMAR 22 cruise (green square) with two main zonal transects; one from the coast Caldera (27°S/70.87°W) to San Félix island (SF) (26.16°S/80°12′W), and another from Valparaiso (32.97°S/72.41°W) to the Juan Fernandez Archipielago (33.63°S/78.85°W), from October 12^th to November 15^th, 2016. The location of the sample stations near the islands for Rapa Nui Island (b) San Felix (c) and, Alejandro Selkirk (d). Information for each transect is indicated in the supplementary Table S1.

2.2 Sampling

On the CIMAR 21 and 22 cruises, continuous temperature, and salinity profiles were obtained with a CTD Seabird 911 plus; whereas dissolved oxygen was taken with a Seabird SBE43 oxygen sensor mounted on a rosette system and calibrated with Winkler titration. In the FDS, a Seabird SBE25 CTD was used. In addition, triplicate surface samples for CH₄, nutrients and Chl-a were obtained using 10 L Niskin bottles mounted on a rosette. CH₄ samples were collected in 20 mL gas chromatography (GC) vials. Subsequently, the samples were preserved through the addition of 50 µL of saturated HgCl₂ (Tilbrook & Karl, 1995) and immediately sealed with a gray butyl-rubber septum and an aluminum cap to avoid the formation of air bubbles, then stored in darkness until analysis. Discrete nutrient samples (NO₃⁻, NO₂⁻, and PO₄⁻³) were taken in triplicate from seawater. The samples were filtered with 0.7 µm grade GF/F filters (Watman®) and stored at −20°C until analysis. For Chl-a, between 0.25 and 1 L (depending on the area) of seawater was filtered (in triplicate) through 0.7 μm, GF/F glass fiber (Watmann) and then immediately frozen at −20°C.

2.3 Chemical Analysis

NO₃⁻, NO₂⁻, and PO₄³⁻ analysis were performed according to colorimetric techniques (Grasshoff et al., 1983) using a an AutoAnalyzer (SEAL Analytical AA3). However, during the CIMAR 21 cruise, given the extremely low (nanomolar) nutrient concentrations that were expected in the oligotrophic SESP region, the AutoAnalyzer was connected to a 500 mm Liquid Waveguide Capillary Cell (LWCC-3050, World Precision Instrument), as recommended for submicromolar levels in NO₃⁻ and PO₄³⁻ (Troncoso et al., 2018). Chl-a samples were analyzed using fluorometry (Turner Designs 10AUTM) according to standard procedures (Parsons et al., 1984).

Dissolved CH₄ was analyzed through static-headspace equilibration (McAuliffe, 1963) for gas chromatography (GC). In this regards, 5 mL of ultrapure Helium was added into GC vials, generating a headspace for the equilibration between both phases. Once an equilibrium was reached, an automatic gas injection takes place (2007–2018, Agilent 7697A autosampler), on a Schimadzu 17A gas chromatograph with a Flame ionization detector and capillary column (Restek RT QS, 0.53 mm3 × 30 m), operated at 30°C temperature with a flow of 2.6 mL min⁻¹. The gas carrier is ultrapure N₂ (99.999%).

Four point calibration curves were carried out for each sample set using a certified NOAA primary standard with a similar composition to the atmosphere (1,863.4 ± 0.3 ppb for CH₄) (Bullister et al., 2016). In addition, two standard gas mixtures provided by Air Liquide (USA) with a certified concentration of 1 and 5 ppm, and zero air (synthetic air without CH₄ tracers) were used. Two standards were used daily to assess the performance of GC. The detector response was linear within these concentration ranges. The precision of this method was around 5.5%, estimated from the coefficient of variation based on triplicate analysis (Wilson et al., 2018).

2.4 Data Analysis

The daily CH₄ sea-air flux (F, μmol m⁻² d⁻¹) was calculated as:

(1)

where k_w is the gas transfer velocity (m d⁻¹), C_w is the ambient concentration of dissolved gas (nmol L⁻¹) measured in the waters of the mixing layer (ML). For this study, the C_w was averaged from the CH₄ concentration measured at the surface (1 m depth) and at 10 m depth. The C* is the gas concentration at equilibrium with the atmosphere at its last previous contact with the atmosphere, and is estimated using the CH₄ solubility equation (Wiesenburg & Guinasso, 1979). Historical atmospheric CH₄ values are taken from the global monthly mean register from the NOAA/ESRL program (available online at https://www.esrl.noaa.gov/gmd/dv/data/).

Gas transfer velocity (K_w) was calculated according to the parameterization from Wanninkhof (1992), or W92 updated in 2014 (Wanninkhof et al., 2014). It is considered to be the best parameterization for the study area according to the wind regime (Farías, Besoain, et al., 2015b). Indeed, a comparison was made with Nightingale's parameterization, and there was almost no difference (less than 4.5%). However, when W92 was compared with the parametrization proposed by Liss and Merlivat (1986), the air-sea gas exchange was underestimated by 30%.

where U² is the wind speed (cm s⁻¹), and Sc is the Schmidt number for CH₄; more specifically it refers to the relationship between viscosity and the diffusion coefficient of CH₄ in water, which is dependent on the seawater's temperature and salinity. For CH₄, the Schmidt number as a function of temperature (°C) was revised by Sarmiento and Gruber (2006) as:

For CIMAR 21 and 22, the wind speed and direction was measured on board and normalized to 10 m height using the relationship in Garratt (1977). For the FDS cruise, wind speed and direction was based on a six hourly register, obtained from a permanent meteorological station located at Carriel Sur (http://www.meteochile.gob.cl/), a coastal weather station that meets with international standards. For air-sea gas exchange, the average wind speed over the previous seven days was used. During CIMAR 22, the sea level anomaly (SLA) and the distribution of mesoscale eddies was analyzed. Data was extracted from the AVISO data visualisation portal (https://aviso.oceandatalab.com/) onto a 0.25° × 0.25° longitude/latitude grid to obtain SLA map on a 1 day basis (October 16th) and mesoscale eddies were identified using Chaigneau et al. (2009) methodology.

Spearman correlations (Rho) were calculated between oceanographic variables (T°C, S, DO, Chl-a, nutrients), along with CH₄ in the surface layer and throughout the entire water column (sup-500 m depth), in order to determine the probable physical and biogeochemical variables that control the zonal variation of CH₄. The threshold value for statistical significance was set at p < 0.05 and p < 0.01. Procedures for several pairwise comparisons between average values were carried out using a Kruskal-Wallis test, with the prior application of a Shapiro-Wilks test to confirm normal distribution.

3 Results and Discussion

3.1 Biogeochemistry and Oceanography of the SESP

The Chilean coast (excluding Patagonia) is characterized by an almost completely absent continental shelf accompanied by a very steep continental slope. Indeed, Chile has one of the most narrow continental shelves worldwide (less than 10 km) (Paris et al., 2016), except for between 35° and 37°S where the continental shelf is much wider.

The SESP basin is characterized by an intricate and deep bathymetry, including the Atacama trench, which is interrupted by islands, archipelagos, and seamounts (Figure 2). Also, the SESP has pronounced gradients in terms of phytoplankton biomass. Previous findings have reported from hyperoligotrophy at the center of the subtropical gyre with levels of surface Chl-a lower than 0.03 mg m⁻³ and undetectable levels of nutrients (or nutrient depletion) (Raimbault et al., 2008), to eutrophic areas with coastal upwelling and Chl-a levels as high as 10 mg m⁻³ and nutrient levels two orders of magnitude higher than in the subtropical gyre (Claustre et al., 2008; Testa et al., 2018).

Figure 2

Study area showing the location of sample stations superimposed over an image of the mean surface Chl-a concentrations during the austral spring period (color data denotes Chl-a concentration expressed in µg L⁻¹ and systematized according to the color bar). The cruise tracks used in this study are indicated. Bathymetric and topography features are shown in the subtropical ESP region. The black line delimits the position of subtropical convergence (STC). Stations located near islands are indicated.

In addition to the differences in trophic levels and nutrient availability in the SESP, marked differences also exist in the water mass structures along the SESP. The hydrological features in the water column (0–500 m) are mainly associated with the distribution of three water masses previously described by Strub et al. (1998), Silva et al. (2009), and Llanillo et al. (2012). These are the Subtropical Water (STW; σ_θ = 25–25.5 kg m⁻³, T > 20°C; S = 34.9), ESPIW (σ_θ = 25.75–26.2 kg m⁻³, T = 12.5°C; S < 34.3), and Equatorial Subsurface Water (ESSW; σ_θ = 26.25–27.25 kg m⁻³, T = 12.5°C; S = 34.9). Additionally, at depths below 500 m, the influence of AAIW (σ_θ > 27.2 kg m⁻³) is observed. Within the study region, each water mass proportion may influence the CH₄ distribution.

In the region covering the area from 17° to 37°S, a transition between sub-Antarctic and subtropical regions is detected, where saline and warm STW is distinguished from the colder and fresher sub-Antarctic water (SAAW). This separation also coincides with the presence of subtropical convergence (30° and 40°S; see Figure 2). According to the Ɵ-S diagram (Figure 3a), at 27°S (during 2015), the STW is confined to the upper water column of the coastal waters, with σ_θ= 25–25.75 kg m⁻³, that corresponds to the surface waters down to ∼30–50 m depth, while westward (within the oligotrophic gyre) it is persistent up to 250 m depth. Below the STW, a thin layer of the ESPIW is only present within the eutrophic and mesotrophic sections, demarcated by the SSM. Indeed, the SSM is present as shallow as 30–85 m depth, which corresponds to the coastal section and deepens toward the open ocean (80–200 m depth). The SAAW is understood to subduct below the STW near 35°S, creating a tongue of low salinity water (SSM), which is associated with the ESPIW (Schneider et al., 2003; Silva et al., 2009). Once subducted, the SSM continues on a northward path to depths of 100–220 m, moving between two more saline water masses, that is, the STW and the ESSW(Karstensen, 2004; Leth et al., 2004). The presence of the SSM gives the water column a pronounced stratification (Figures 3b, 3f and 3j). This stratification may enhance particle accumulation and the degradation of organic matter. In fact, Whitmire et al. (2009) reported that particulate backscattering at 700 nm, which is associated with high particulate accumulation, coincides with a secondary Chl-a maximum present at the base of the euphotic zone and upper oxycline, where the SSM is also located.

Figure 3

θ-S diagram from the surface to 500 m depth in the Subtropical Eastern South Pacific Ocean (SESP) region, with data taken in different transects during the CIMAR 21 at 27°S (a, b, c, d) and CIMAR 22 at 27°S (e, f, g, h) and 33°S (i, j, k, l). The θ-S plots are related with stratification (b, f, j), oxygen (c, g, k), and methane (d, h, i) for each described transect. Subtropical Water (STW), Eastern South Pacific Intermediate Water (ESPIW) and Equatorial Subsurface Water (ESSW) water masses are indicated.

Figure 3 (continued)

θ-S diagram from the surface to 500 m depth in the Subtropical Eastern South Pacific Ocean (SESP) region, with data taken in different transects during the CIMAR 21 at 27°S (a, b, c, d) and CIMAR 22 at 27°S (e, f, g, h) and 33°S (i, j, k, l). The θ-S plots are related with stratification (b, f, j), oxygen (c, g, k), and methane (d, h, i) for each described transect. Subtropical Water (STW), Eastern South Pacific Intermediate Water (ESPIW) and Equatorial Subsurface Water (ESSW) water masses are indicated.

The ESPIW has an attenuated presence as it mixes with waters from immediately above and below. Thus, a smoothening of the Ɵ-S diagram was observed in 2016, at the same latitude (Figure 3e). As a consequence, the water column becomes less stratified and oxygen levels also increase (Figure 3g; Table 1). These circumstances may be attributed to the 2015–2016 El Niño event (see Section 3.2.2), which appears to impact stratification (Figure 3f) and oxygen distribution (Figure 3g).

Table 1. Ranges, Average and Standard Deviation Values of Oceanographic Variables Measured in the Sampled Water Column in the Eutrophic, Mesotrophic, and Oligotrophic Zones

	Eutrophic area		Mesotrophic area		Oligotrophic area
Northerm region (∼27°S) CIMAR 21 and 22
Long. and distance to the coast	∼71.2°W (>50 km)		∼82°W (1,000 km)		∼109°W (1,500 km)
CIMAR 21: 1, 3/CIMAR 22: 1,4		CIMAR 21: 6, 9, 11, 13/CIMAR 22: 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 17, 18, 19		CIMAR 21: 16, 18, 20, 24, 26, 30, 40, 46, 59, 64, 66, 71, 79
Variables	Range	Mean ± SD	Range	Mean ± SD	Range	Mean ± SD
T (°C)	10.76–16.16/9.90–15.89	13.62 ± 1.70/13.64 ± 2.63	6.73–17.94/5.80–18.14	13.28 ± 3.72/14.83 ± 3.72	6.53–21.46	17.05 ± 4.34
S	34.53–34.83/34.42–34.78	34.75 ± 0.09/34.62 ± 0.10	34.19–34.92/34.29–35.01	34.59 ± 0.19/34.65 ± 0.19	34.18–36.00	35.22 ± 0.58
O₂ (µM)	7.69–252.9/7.32–255.2	118.9 ± 102.4/152.5 ± 132.1	13.20–257.9/18.13–257.5	162.3 ± 89.54/181.8 ± 86.31	77.82–243.5	219.2 ± 29.97
Chl-a (mg m⁻³)	0.19–0.82/0.65–1.56	0.45 ± 0.23/1.10 ± 0.64	0.10-0.74/0.01–1.73	0.37 ± 0.16/0.56 ± 0.33	0.01–0.43	0.11 ± 0.09
NO₃ (µM)	3.27–31.35/4.87–36.42	14.16 ± 7.81/11.30 ± 8.17	0.32–39.92/0.33–30.44	14.38 ± 13.06/7.13 ± 8.13	0.02–33.45	4.74 ± 7.68
NO₂ (µM)	0.60–4.43/0.01–0.27	0.60 ± 0.97/0.10 ± 0.09	0.02–0.66/0.02–0.51	0.11 ± 0.11/0.05 ± 0.07	0.04–0.30	0.09 ± 0.04
PO₄ (µM)	0.98–2.78/0.50–2.16	1.95 ± 0.66/1.19 ± 0.73	0.13–3.01/0.16–2.94	0.99 ± 0.76/0.82 ± 0.71	0.01–2.35	0.42 ± 0.53
CH₄ (nM)	3.27–10.90/3.70–16.09	6.82 ± 2.14/9.14 ± 4.16	1.31–6.10/1.35–24.40	3.10 ± 1.32/9.63 ± 5.46	0.68–5.95	2.33 ± 1.03
Central Chile (∼33°S) CIMAR 22
Longitude and distance to the coast	∼72.3°W		∼80.7°W
Stations	CIMAR 22: 39, 40		CIMAR 22: 26, 27, 29, 30, 32, 33, 34, 35, 36, 37, 38
Variables	Range	Mean ± SD	Range	Mean ± SD
T (°C)	7.56–15.06	12.29 ± 2.36	7.03–16.98	13.72 ± 3.09
S	34.13–34.74	34.66 ± 0.16	34.03–34.71	34.31 ± 0.12
O₂ (µM)	12.48–257.4	147.6 ± 96.44	12.90–269.4	182.3 ± 89.90
Chl-a (mg m^−³)	0.06–4.43	1.25 ± 1.66	0.002–4.21	0.70 ± 0.80
NO₃ (µM)	0.43–26.33	7.34 ± 7.16	0.43–26.33	7.34 ± 7.16
NO₂ (µM)	0.29–0.71	0.47 ± 0.14	0.21–0.87	0.33 ± 0.11
PO₄ (µM)	0.71–2.31	1.64 ± 0.52	0.18–2.31	0.81 ± 0.65
CH₄ (nM)	6.61–13.10	9.51 ± 2.25	1.01–33.49	4.98 ± 4.05

Considering oxygen distribution, the ESSW has low oxygen concentration and a well-known oxygen minimum zone (OMZ). The upper limit (oxycline) is considered to be the shallowest in the open ocean (30–70 m depth) (Fuenzalida et al., 2009), and in its core it even reaches anoxia (Revsbech et al., 2009). This water mass is relatively more saline and nutrient rich, and it moves poleward via the Peru-Chile Undercurrent, with σ_θ = 26.5 kg m⁻³ at ∼200 m depth (Fuenzalida et al., 2009). Very low oxygen values trigger anaerobic microbial processes, such as denitrification and sulfate reduction in the water column (Canfield et al., 2010; Farías et al., 2009). Under these conditions, anaerobic methanogenesis is expected to be triggered. Although these processes were not investigated in this study, previous metagenomics and meta-transcriptomics studies reveal the occurrence of anaerobic methanogenesis in the OMZ (Ulloa et al., 2012). Based on this principle, levels of CH₄ super saturation are associated with the upper limit of the OMZ and advected by the ESSW with the possibility that anaerobic methanogenesis is taking place (Figures 3d and 3h).

Thus, levels of dissolved CH₄ depend on water mass distribution (Figures 3d, 3h and 3l), and redistribution, as observed during the El Nino Southern Oscillation (ENSO)-El Niño event in 2015–2016. In essence, CH₄ accumulates in less dense water masses that are associated with a marked stratification (in the subsurface), which coincides with the upper oxycline (Figures 3d and 3h). To describe the spatial variability of CH₄, all biogeochemical and oceanographic features are considered and analyzed over a zonal vertical cross section (0–500 m depth) along two transects at ∼27° and ∼33°S. Conversely, the coastal transect (at 36.5°S) was analyzed for the sole purpose of calculating the CH₄ exchange across the air-sea interface. Remarkably, although all cruises took place in the austral spring-summer, an inter-annual variability should be considered due to the 2015–2016 El Niño event.

3.2 Distribution of CH₄ and Oceanographic Variables on the ∼27°S Transect

Figure 4 shows cross sections of T°C, S, oxygen, Chl-a, NO₃⁻, PO₃³⁻, NO₂⁻, and CH₄ during the CIMAR 21 cruise positioned at ∼27°S. This transect includes several stations covering the widest longitudinal range, from the coast (70.88°W) to near the Rapa Nui Island (109°W). The boundary between eutrophic and mesotrophic conditions is located around 72.5°S, whereas for mesotrophic and oligotrophic conditions it is at 82°S. These boundaries are based on the surface Chl-a and nutrient concentrations, as described in the following (see Table 1).

Figure 4

Vertical cross sections of different variables over a zonal transect at 27°S (CIMAR 21 cruise), between 71° and 105°W, illustrating (a) temperature; (b) salinity; (c) oxygen (µmol L⁻¹); (d) chlorophyll-a (µg L⁻¹); (e) nitrate (µmol L⁻¹); (f) phosphate (µmol L⁻¹); (g) nitrite (µmol L⁻¹); and (h) methane (nmol L⁻¹).

3.2.1 The CIMAR 21 Cruise

During CIMAR 21, T°C and S show an increase from the coast to the most western sample stations, ranging from 6.53°C to 21.46°C and 34.19–35.00, respectively. In general, T°C and S decrease with depth; as observed in the water mass distributions in this study. In fact, the sample transect indicates the presence of the STW, with a high S and T°C, which progressively occupy greater depths toward the center of the ESP gyre. This contrasts with the colder and fresher waters present on the Chilean coast from subequatorial origin (ESSW) (Figures 4a and 4b). Oxygen concentrations decrease from almost 258 to 7.68 µmol L⁻¹, with mean ± SD values of 118 ± 102, 162 ± 90 and 219 ± 30 µmol L^-1, for the eutrophic, mesotrophic, and oligotrophic areas, respectively, revealing an increase in oxygen with greater distance from the coast (Table 1). The strip of water adjacent to the coast shows a vertical oxygen distribution with significantly higher levels in the surface and photic layer, which decreases to less than 7.7 μmol L⁻¹ (<1% saturation) as shallow as 75 m depth, producing a very shallow and sharp upper oxycline. Below the oxycline, oxygen remains constant down to 300 m. In contrast, in the mesotrophic section, the oxycline is deeper, and found at 100–150 m (Figure 4c). Below 100–150 m, oxygen remains relatively constant at levels lower than 13.20 µmol L⁻¹ down to 150–350 m (the OMZ's core). Offshore, the thickness and meridional extent of the OMZ decreases until it finally vanishes eastward at ∼85°W (oxygen > 100 µmol L⁻¹). Whereas moving southward, the zonal extension of the OMZ strongly diminishes from 1,200 km (20°S) to 25 km (30–35°S) (Fuenzalida et al., 2009; Pizarro, 1999), and by 36°S the ESSW encroaches onto the continental shelf, occupying from 30–50 m depth and covering the sediments.

Phytoplankton biomass varies intensely, from 0.010 to 0.82 mg m⁻³, with mean ± SD values decreasing one order of magnitude from eutrophic to oligotrophic waters (Table 1; Figure 4d). The lowest values coincide with the hyper-oligotrophic South Pacific Subtropical Gyre, with the clearest and depleted nutrient waters of the world's ocean (Claustre et al., 2008) and characterized by extremely low surface Chl-a concentrations (<0.03 mg m⁻³) (Ras et al., 2008). In contrast, the coastal waters show the highest Chl-a values and coincide with areas of coastal upwelling, since it fertilizes the surface waters with a significant amount of nutrients (Testa et al., 2018), thus increasing phytoplanktonic activity and resulting in surges in Chl-a values. Remarkably, the Chl-a maximum deepens as it moves toward the gyre (Figure 4d). This is due to a change in the phytoplanktonic composition, that is, from large diatoms in the coastal area to small picophytoplankton (cyanobacteria, such as Prochlorococcus and eukaryotes) toward the subtropical gyre, as reported by Grob et al. (2007) and Ras et al. (2008).

NO₃⁻ and HPO₃²⁻ varied between 0.024 and 39.92 µmol L⁻¹ (mean ± SD = 14.96 ± 13.71) and 0.010–3.01 µmol L^–1 (mean ± SD = 0.99 ± 0.77), respectively. The oligotrophic waters are almost fully depleted between the surface and 150–200 m depth, with persistently low concentrations below this depth (Table 1; Figures 4e and 4f). As expected, nutrient concentrations increase under mesotrophic conditions (mean ± SD = 14.38 ± 13.06 for NO₃⁻ and 0.99 ± 0.76 for PO₄³⁻), whereas in the eutrophic area, NO₃⁻ remains constant (mean ± SD = 14.16 ± 7.81) but HPO₃²⁻significantly increases (mean ± SD = 1.95 ± 0.66). This observation can be attributed to the presence of very low oxygen levels that triggers nitrogen loss through denitrification and anammox (Farías et al., 2009; Thamdrup et al., 2006). In the coastal area, nutrient concentration increases significantly, consistent with more shallow nutriclines and oxyclines toward the coast (Figure 3g). This clearly indicates the presence of coastal upwelling, which results from the maximum alongshore wind stress at this latitude, which drives upwelling throughout the spring-summer period (Thomas et al., 2001).

CH₄ concentration ranges from 0.68 (29% saturation) to 10.90 nmol L⁻¹ (436% saturation) (mean ± SD = 3.00 ± 8.59 nmol L⁻¹); with maximum values at the upper limit (oxycline) of the OMZ, as observed at the coastal station (St. 3), and minimum values in the surface waters of the oligotrophic region (St. 18) (Table 1; Figure 4h). Indeed, mean CH₄ concentration is three times higher in the coastal zone than in the oligotrophic zone, and it decreases with distance from the coast, consistent with Chl-a and nutrient distribution (Figure 4g). With regard to the vertical CH₄ distribution, despite the absence of a proper continental shelf, the area of maximum CH₄ accumulation is in the coastal zone (6.82 ± 2.14 nmol L⁻¹). CH₄ is noted to accumulate in the mesotrophic region above a low salinity tongue (SSM); this causes a pronounced subsurface stratification in the region that coincides with the highest CH₄ values (Figure 3b). The existence of a CH₄ maximum that occurs at a distance from the coastal zone has not yet been reported in the SESP. A limited number of studies exist that investigate zonal CH₄ distribution in the SESP. These include a transect at 17°S analyzed by Yoshikawa et al. (2014), however the study does not include the coastal zone nor the influence of the SSM, as the SSM has disappeared at this latitude. It is important to note that the ESPIW occurs as a less saline water mass between two more saline water masses, thus giving a strong stratification. In turn, this reduces the vertical mixing and subsequently results in more favorable conditions for the accumulation of organic matter or phytodetritus (particles). This shallow stratification coincides with the oxycline (upper limit of the OMZ); thus it is expected that CH₄ accumulation occurs through either anaerobic respiration in microniches, or by aerobic methylotrophy through the association of phytoplankton and bacterioplankton (Bogards et al., 2014; Rehder et al., 1999; Repeta et al., 2016). In situ CH₄ production has been reported by Weller et al. (2013), which leads to CH₄ super saturation below the mixed layer under subsurface stratification and low wind stress conditions; these occur during phase II of the phytoplankton bloom in a subtropical mesoscale eddy (Lagrangian experiment). It is possible that the lysis of phytoplankton cells and the release of organic substrates could lead to CH₄ production via the methylotrophic pathway (Weller et al., 2013).

At intermediate depths below the highly stratified layer, CH₄ appears to be consumed, likely to coincide with the core of the ESPIW. The apparent CH₄ depletion may be due to poor advection of CH₄ rich waters through the ESPIW or aerobic CH₄ oxidation as this water mass moves northward (see Section 3.3). Finally, in the oligotrophic zone (west of 83°W), CH₄ is homogeneously distributed with saturation levels close to equilibrium with the atmosphere.

3.2.2 The CIMAR 22 Cruise

During the CIMAR 22 cruise, the same transect was sampled including more sampling stations but only covering the eutrophic and mesotrophic zone (upto 82°W) (Table 1, Figure 5). The T°C and S range from 5.80°C to 21.46°C and 34.18–36.00, respectively. Zonally, both variables show a coastward increase that correspond to water mass distributions, as well as thermoclines and haloclines that deepen toward the east (Figures 5a and 5b). At 83°W, the SSM is present at depths exceeding 200 m, and the OMZ is most clearly demarcated in the eutrophic and mesotrophic areas (Figure 5c). Oxygen concentrations fluctuate from 257 to 7.32 µmol L⁻¹ with a considerable area of water column occupied by waters with less than 50 µmol L⁻¹ of oxygen; mean ± SD values by section reveal an increase in oxygen with greater distance from the coast, with values of 152 ± 132 and 181 ± 86 µmol L⁻¹ for eutrophic, mesotrophic section, respectively (Table 1).

Figure 5

Vertical cross sections of different variables over a zonal transect at 27°S (CIMAR 22 cruise), between 71 and 80.5°W, illustrating (a) temperature; (b) salinity; (c) oxygen (µmol L⁻¹); (d)chlorophyll-a (µg L⁻¹); (e) nitrate (µmol L⁻¹); (f) phosphate (µmol L⁻¹); (g) nitrite (µmol L⁻¹); and (h) methane (nmol L⁻¹).

Chl-a show ranges from 0.01 to 1.56 mg m⁻³ with mean ± SD values slightly higher than in CIMAR 21 (Table 1). Chl-a is heterogeneously distributed, with high values at the coastal band, as well as at certain stations in the mesotrophic area, such as Sts. 8, 9, and 10 and near the JFA (Figure 5d). According to remote sensing data of sea level anomalies (Section 3.4), this pattern coincides with the presence of mesoscale eddies. NO₃⁻ and HPO₃²⁻ range between 0.02–36.42 and 0.01–2.94 µmol L⁻¹, respectively (Table 1), and the means values decrease among eutrophic, mesotrophic, and oligotrophic zones (Table 1). Both nutrients profiles also show an irregular distribution, indicating a movement of nutriclines toward the surface at Sts. 6 and 7, which contrasts with the surrounding waters (Figures 5e and 5f). This suggests that there are varying responses to mesoscale processes, such as eddies and fronts (see Section 3.4). Finally, CH₄ levels fluctuate between 0.68 (29% saturation) and 24.4 nmol L⁻¹ (1,094% saturation). Maximum values are found in both the surface and subsurface layer at the mesotrophic stations Sts. 7, 8, and 11 (Figure 5h), where a Chl-a maxima is also observed (Figure 5d).

A comparison between two consecutive years (CIMAR 21 and 22) at both eutrophic and mesotrophic zones reveals that in 2015, the cross vertical distributions for the majority of the sampled variables differ from those sampled in the 2016 (CIMAR 22). By October 2015, sea surface temperatures in the Niño 3.4 region showed a strong positive temperature anomaly, along with important atmospheric changes (https://climatedataguide.ucar.edu/). These changes are comparable with the 1997–1998 events, for which reason it was referred to as El Niño Godzilla (Cai et al., 2020). The effects of this process were observed off Chile from autumn to the end of December 2016 (during CIMAR 22), whereas considerably different oceanographic conditions were observed in early 2015 (Farías et al., 2020).

The surface layer is more saline and warmer in 2016 compared with 2015, also 2016 presented relatively higher oxygen levels in the OMZ core, indicating a redistribution of water masses between these consecutive years (Figures 3a and 3e). Additionally, the mean values indicate a decrease in both NO₃⁻ and PO₄³⁻ (Table 1). It is already established that ENSO, especially in its warm phase, can modify the physical, chemical, and biological dynamics of the eastern South Pacific (Escribano et al., 2004; Graco et al., 2017; Gutiérrez et al., 2008). For example, coastal waters off Peru under ENSO conditions have reported: (i) an increase of sea surface temperature and a progressive deepening of the 15°C isotherm in the water column, (ii) changes in water mass distribution with SSW dominating the surface waters, (iii) a decrease in primary production, mainly associated with a decrease in the nutrient supply to surface waters, and (iv) a deepening of the upper boundary limit of the OMZ to more than 100 m depth.

However, Chl-a and CH₄ increases are more heterogeneous during 2016; this is an unexpected result and it may be an artifact of the available data, since in 2016 the sample stations are closer together which improves the detection of mesoscale variability. The presence of mesoscale eddies and their effect on the accumulation of Chl-a and CH₄ is discussed in Section 3.4. However, it remains unresolved if the 2016 El Niño event may have caused increased mesoscale activity, which may have had an effect on several variables, including CH₄ accumulation. It is widely accepted that coastal waters off central Chile respond differently to those off Peru and northern Chile, due to a more efficient oceanic teleconnection in the north with respect to central-southern Chile, where altered Kelvin wave signals can have an impact, as well as the presence of varied oceanographic and hydrographic conditions (Montecinos & Gomez, 2010).

3.3 CH₄ Distribution, Oceanographic and Biogeochemical Variables from the 33°S Transect

Figure 6 shows cross sections of T°C, S, oxygen, Chl-a, NO₃⁻, HPO₃²⁻, NO₂⁻ and CH₄ at ∼33°S during CIMAR 22. T°C and S range from 7.03°C to16.98°C and 34.03–34.74, respectively, with less variation compared to at 27°S. This indicates a reduced influence from STW, as well as an increased presence of SAAW, and in particular the ESPIW (observed through SSM) is much shallower and has a greater extension (Figure 6b), with an influence reaching as far as 83°S. This occurs due to the fact that the water mass is formed adjacent to the subtropical convergence (Karstensen et al., 2004).

Figure 6

Vertical cross section of different variables over a zonal transect ∼33°S/(CIMAR 22) between 72.74° and 80.78°W (a) temperature; (b) salinity; (c) oxygen (v); (d) chlorophyll-a (µg L⁻¹); (e) nitrate (µmol L⁻¹); (f) phosphate (µmol L⁻¹); (g) nitrite (µmol L⁻¹); and (h) methane (nmol L⁻¹).

Oxygen fluctuates from 12.5 to 266 µmol L⁻¹, and a cross section of the area (Figure 6c) shows a westward deepening of the oxycline. Oxygen distributions at 33°S reveal a gradual oxygenation of the OMZ's core and a more narrow vertical expansion of the OMZ compared to at 27°S transect (Figure 6c). This is mainly due to the ventilation effect that impacts the AAIW in the OMZ as it moves equatorward, as described by Carrasco et al. (2017). The latitudinal gradient creates an ideal study site for thresholds within oxygen-sensitive processes. Chl-a values are heterogeneously distributed (Figure 6d), ranging between 0.013 and 4.43 mg m⁻³, with increased values further from the coast compared to those observed in St. 34 at ∼79.2°W. This transect is characterized by the presence of the JFA and various seamounts (Figure 2), which make it difficult to analyze biogeochemical variables, given the effect of mesoscale eddies and the IME (Section 3.4). NO₃⁻ and HPO₃²⁻ concentrations varied between 0.43–26.33 and 0.18–2.45 (Figures 6e and 6f), respectively. In the surface water, there is a relative consumption of both nutrients, except in Sts. 39 and 38, where a marked surface accumulation of these nutrients is observed. This pattern also corresponds with a salinity and temperature increase, which could indicate the presence of mesoscale structures, since these stations are located far from upwelling areas.

Also, dissolved CH₄ is heterogeneously distributed, for example, in the northern transect from CIMAR 22, levels fluctuate between 1.17 (45% saturation) and 33.5 (1,423% saturation) and there is high CH₄ accumulation throughout the water column, specifically in the mesotrophic zone and also in the eutrophic zone. In the latter zone, high CH₄ concentrations are expected; and they seem to respond to diffusion-advection behavior where CH₄ is partially produced by diagenesis in the sediments, providing a source to the water column. Furthermore, coastal upwelling appears to bring CH₄-rich subsurface waters from the sediments toward the mixed layer (Naqvi et al., 2010; Weber et al., 2019). However, similar to the previous transects, high CH₄ accumulation occurs in the mesotrophic zone around the JFA and within eddies (St. 37). Finally, this transect and those from 27°S report CH₄ subsaturation (Table 1). Most of the subsaturation occurs subsuperficially around the SSM core. This pattern could be attributed to CH₄ consumption, since this depletion occurs between two layers with a relatively high concentration of CH₄ (Figure 6g). Since the ESPIW is closer to the formation zone, this depletion may be also due to the advection of CH₄ poor waters from the formation site that is fed by the SAAW. Surface waters from the eastern South Pacific within the latitudinal band, between 40°S and 60°S, show slightly less CH₄ saturations than the atmospheric equilibrium (Bates et al., 1996; Kelley & Jeffrey, 2002), suggesting that SAAW have low CH₄ concentrations. Another plausible explanation for this is that CH₄ oxidation occurs, while the ESPIW moves. Remarkably, this zonal CH₄ depletion coincides with NO₂⁻ accumulation (Figures 4h and 4g), which was previously described by Cornejo et al. (2012) as meridional maximum NO₂⁻ accumulation. The ability of marine ammonia-oxidizing bacteria to also oxidize CH₄ has been early reported (Jones & Morita, 1983; Ward, 1987). This metabolic versatility of aerobic ammonium oxidizer species as those widely distributed in the eastern South Pacific (Ward et al., 2011) may explain the coexistence of CH₄ depletion and NO₂⁻ accumulation at the core of an isolated water mass as the ESPIW.

3.4 The Association Between Methane and Mesoscale Eddies and the Island Mass Effect

The coastal transition zone off Chile extends from the coast to ∼600–1,000 km offshore, and from 19° to 39°S, and is prone to the formation and the propagation of the mesoscale eddies (Hormazabal et al., 2004). Eddies are mainly associated with the baroclinic instability of coastal currents and the westward propagation of Rossby waves from the ocean eastern boundary (Chaigneau et al., 2009). In addition, the zone between 29° and 39°S off central Chile is characterized by high eddy kinetic energy and a strong but variable equatorward wind stress, creating different types of mesoscale eddies (Correa-Ramirez et al., 2007; Hormazabal et al., 2004). In general, these mesoscale eddies, mainly generated in the spring summer upwelling season, have a long lifespan (>3 months) and are able to travel over long distances. These circumstances eventually introduce nutrient-rich, oxygen poor coastal waters into the oligotrophic waters of the subtropical gyre (Correa-Ramirez et al., 2007; Morales et al., 2012).

During the CIMAR 22 cruise, which took place during El Niño conditions, an increased eddy activity was detected using remote sensing data of sea level anomalies (Figure 7). Remarkably, Sts. 11 and 37 have the highest detected levels of CH₄ accumulation in the surface and subsurface water, which correspond with the presence of two AES. In addition, over the northern transect, CEs are present close to Sts. 7, 8, and 9, where an increase in Chl-a and CH₄ was detected in the surface and subsurface waters. This suggests that potential mechanisms are in place to input nutrients into the surface waters, as negative salinity and temperature anomalies are observed compared to the surrounding water. Probably, the presence and intensity of various eddies that characterize the SESP are able to explain the detection of a heterogeneous CH₄ distribution within a mesotrophic area.

Figure 7

Anticyclonic (continuous line) and cyclonic (dotted line) eddies detected from altimetry (sea level anomalies) during October 2016 in the Subtropical Eastern South Pacific Ocean. The location of the sample stations from the CIMAR 22 cruise is overlaid on the map.

Both CEs and ACs have different thermohaline vertical structures, and these differences are mainly attributed to the mechanisms involved in the eddy formation (Chaigneau et al., 2011). At the same time, these eddies have an effect on water biogeochemistry and biological productivity, as they produce intense vertical and horizontal physical processes that can result in localized subduction, upwelling, advection (particle transport), stirring, and/or mixing in the water column (Chelton et al., 2011; McGillicuddy et al., 2016). In addition, both eddy types generate submesoscale vertical transport and a lateral density front (along the periphery), leading to a Chl-a accumulation, such as in the eddies off the Chilean coast (Wang et al., 2018). This enhances the export of organic matter stimulating nitrogen loss via anammox, as observed offshore of the Peruvian coast (Callbeck et al., 2017), as well as via denitrification (and N₂O accumulation), as observed offshore of Central Chile (Cornejo et al., 2016). As eddies affect the nitrogen cycle, they could also be affecting the CH₄ cycle.

Considering the IME, results from the extremely oligotrophic waters surrounding Rapa Nui island showed Chl-a levels upto 0.19 mg m⁻³ and CH₄ upto 4.05 nmol L^-−1, with the highest results at St. 46 located to the north of Rapa Nui, compared to the neighboring stations (Sts. 40 and 59) (Figures 8a and 8b). The CH₄ maximum coincides with a deep Chl-a and NO₃⁻ maximum at around 100 m depth (Figure 8c). Indeed, there is a positive correlation between both variables (R² = 0.77, supplementary material), which suggests that the increase in microalgal biomass could be favoring the biosynthesis of methylated organic compounds such as methylphosphonate, among other compounds that are consumed by heterotrophic bacteria, generating CH₄ and PO₄³⁻ (Karl et al., 2008; Lenhart et al., 2016; Rakowski et al., 2015; Repeta et al., 2016). On the other hand, the increase in microbial biomass in an extreme oligotrophic zone such as Rapa Nui island, indicates that this island exerts a strong influence on the distribution of phytoplankton.

Figure 8

Vertical cross sections of methane (nmol L⁻¹); chlorophyll-a (µg L⁻¹) and nitrate (µmol L⁻¹) for stations located close to the islands of (1) Rapa Nui (a, b, c) San Félix (d, e, f) and Alejandro Selkirk (g, h, i).

The IME was originally described by Doty and Oguri (1956), and has been associated with the interaction of currents induced by topographic elevations, which generate circulation patterns, for example, Taylor columns and eddies. These, in turn, trigger the upsurge of subsurface nutrients into the euphotic layer. Likewise, other authors suggest that there is a lithogenic input of micronutrients, such as iron, from land. These come either directly from continental runoff or through underground aquifers, and may fertilize the area (Blain et al., 2001; Bucciarelli et al., 2001; Palacios, 2002; Perissinotto et al., 2000), as reported for the Kerguelen Islands, where CH₄ accumulation is observed in the water column (Farías et al., 2015).

Another similar case occurs on the border between the oligotrophic region and the mesotrophic zone, for example, at the Sts. 17, 18, and 19 around the San Félix Island (26°17′S/80°07′W), which is part of the DI Archipelago, approximately 972 km from the coast (Figures 8d–8f). Chl-a and CH₄ values vary between 0.1 and 1.20 ug L⁻¹ and 3.21 and 8.31 nmol L⁻¹, respectively; although CH₄ shows an heterogeneous distribution with a maximum level at St. 17, this difference may be due to the influence of the island's topography and mesoscale events that may be influencing this island. CH₄ accumulations are observed in the surface waters at the three stations surrounding the San Félix Island (Figure 8d). Similar phenomena seem to occur in the JFA (Sts. 26, 27, 29, 30, 32), particularly near the Alejandro Selkirk island (33°48; 80°50°W) (Figures 8g–8i). The IME in the JFA has previously been well described (e.g., Andrade et al., 2014); a local increase in phytoplankton associated with the islands leads to the formation of a Chl-a plume, with concentrations one order of magnitude higher (∼1 µg L⁻¹; Sts. 27, 30, and 32) than the surrounding oligotrophic waters (<0.1 µg L⁻¹). For example, at stations 26, 27, 29, and 30, CH₄ maxima are observed in surface waters (upto 7.64 nmol L⁻¹) and coincide with peaks of Chl-a (upto 1.89 µg L⁻¹), whereas NO₃⁻ is being consumed (Figure 8i). In fact, a well-defined positive linear regressions between Chl-a and CH₄ are observed (supporting Figures S1a, S1c, S1e, S1g, S1i, and S1k) with determination coefficients (R²) fluctuating between 0.30 and 0.77. In addition, linear regressions between NO₃⁻ and CH₄ are variable ([Figures S1b, S1d, S1f, S1h, S1j, and S1l] positive and negative) denoting different physical (and biological) mechanisms that may create an IME.

These observations indicate that a CH₄ maximum could be favored by the increase in phytoplanktonic biomass, resulting in a stimulus for the microbial community that partially recycles CH₄ (Bogard et al., 2014). Phytoplankton and bacterioplacton biomass are responsible for the abundant production of methylated organic compounds, such as dimethylsulfoniumpropionate (DMSP) or methylphophonate (MPh) (Damm et al., 2010; Stefels, 2000), which can subsequently be metabolized to methanethiol and then to CH₄ by marine bacteria such as Roseobacter, which predominate in oligotrophic waters (Kiene et al., 2000; Moran et al., 2003). Other studies have also reported that the CH₄ production in oligotrophic waters is associated with nitrogen limitation (Damm et al., 2010; Stefels, 2000), as reported in the SESP (Knapp et al., 2016).

3.5 Surface Distribution and Air-Sea Methane Fluxes at the Subtropical ESP, and Correlations Between Biogeochemical and Oceanographic Variables

Figure 9 shows the spatial distribution of CH₄ saturation in surface waters within the study area (from ∼17° S to 37° S), and from Chilean and Peruvian coast to 110ºW. In order to construct the surface distribution, all existing surface CH₄ data from the region is considered (including nonpublished data). These measurements were obtained from seven cruises between 1989 and 2016; mostly carried out during austral spring and summer (supplementary Table S1). Surface CH₄ saturation in the SESP fluctuates from 93% to 844% saturation (mean ± SD = 142.6 ± 114.5). Incidences of high CH₄ supersaturations >800% are found in the shallow coastal area (less than 100 m depth). In Figure 8, three foci or hot spots with high CH₄ supersaturation are observed over the limited number of continental shelves off the west coast of Latin America; that is, off Concepcion (36°S), off Valparaiso (32°S), and off central Peru (15°S).

Figure 9

Surface methane saturation (%) based on current (white spots) and previous data (Black spots) for the subtropical eastern South Pacific region.

Although scarce, the presence of continental shelves highlights the importance of these areas as CH₄ hotspots, as they are focus points for high anaerobic decomposition of organic matter in sediments, which correspond with coastal upwelling events that lead to high organic matter production as well as a continuous or seasonal degassing toward the atmosphere. However, high values of CH₄ supersaturation are not only associated with the shallow coastal zone under the influence of coastal upwelling. Remarkably, values as high as 1,423% are found in the mesotrophic area, which corresponds to the so-called, coastal transition zone (Hormazabal et al., 2004). This zone, extending from the coast to ∼600–1,000 km off shore, and from 19° to 39°S, is prone to the creation and propagation of mesoscale eddy activity (Figure 7) that prompts nutrient and biomass accumulation. This, in turn, triggers CH₄ accumulation, as observed near St. 11 and St 37, where ACs are present.

In addition, islands and bottom irregularities, such as several seamounts, generate CH₄ accumulation (Figure 9). Oligotrophic waters present slightly under saturated methane levels; ∼98% and close to the equilibrium, which agrees with those values found by Bates et al. (1996), Yoshida et al. (2011), and Yoshikawa et al. (2014) for the central Pacific basin; however, in the presence of islands, CH₄ levels increase with respect to the surrounding waters.

Table 2 shows Spearman correlations among CH₄, T°C, S, oxygen, Chl-a, and nutrients, considering data throughout the entire water column. The results include data from CIMAR 21 (Table 2a), and indicate that CH₄ levels were significantly and negatively correlated with T°C and S, positively correlated with Chl-a, NO₃⁻, and PO₄³⁻, and no correlation was observed with oxygen. This implies that more CH₄ was accumulated or produced in cooler, saltier, and nutrient rich waters, which agrees with the gradients present in the coastal and open ocean.

Table 2. Spearman Correlations of Oceanographic Variables (Including CH₄) Measured in the Water Column for CIMAR 21 and CIMAR 22

	CIMAR 21 (0–500 m)							CIMAR 22 (0–500 m)
	S	O₂ (µM)	Chl-a (mg m⁻³)	NO₃ (µM)	NO₂ (µM)	PO₄ (µM)	CH₄ (nM)	S	O₂ (µM)	Chl-a (mg m⁻³)	NO₃ (µM)	NO₂ (µM)	PO₄ (µM)	CH₄ (nM)
T	0.944^a	0.587^a	−0.096	−0.887^a	−0.098	−0.851^a	−0.188^b	0.445^a	0.746^a	0.426^a	−0.770^a	−0.149^b	−0.754^a	0.038
	0.000	0.000	0.226	0.000	0.171	0.000	0.014	0.000	0.000	0.000	0.000	0.028	0.000	0.586
	200	200	160	197	197	200	169	231	132	179	213	216	216	212
S		0.459^a	−0.086	−0.815^a	−0.108	−0.788^a	−0.148		−0.372^a	−0.109	−0.085	−0.534^a	−0.071	0.021
		0.000	0.279	0.000	0.132	0.000	0.054		0.000	0.145	0.217	0.000	0.296	0.758
		200	160	197	197	200	169		132	179	213	216	216	212
O₂			0.430^a	−0.680^a	0.069	−0.573^a	0.002			0.844^a	−0.792^a	0.220^b	−0.831^a	0.055
			0.000	0.000	0.339	0.000	0.975			0.000	0.000	0.011	0.000	0.537
			160	197	197	200	169			96	129	131	131	127
Chl-a				0.056	0.435^a	0.079	0.377^a				−0.412^a	0.172^b	−0.377^a	0.086
				0.489	0.000	0.320	0.000				0.000	0.022	0.000	0.267
				157	157	160	139				172	175	175	168
NO₃^−					0.122	0.823^a	0.209^a					−0.017	0.791^a	−0.092
					0.088	0.000	0.006					0.797	0.000	0.192
					197	197	169					222	222	203
NO₂^−						0.190^a	0.349^a						−0.019	0.199^a
						0.008	0.000						0.782	0.004
						197	169						225	206
PO₄^3−							0.228^a							−0.131
							0.003							0.061
							169							206

• ^a The correlation is significant at the 0.01 level (bilateral).
• ^b The correlation is significant at the 0.05 level (bilateral).

However, a lack of significant correlations in the CIMAR 22 data, indicate high levels of heterogeneity in the eutrophic and mesotrophic zone of the SESP (Table 2). For example, no correlations are observed between CH₄ and Chl-a, and nor with others variables. Although various studies report correlation between CH₄ and Chl-a (Yoshikawa et al., 2014; Zindler et al., 2013), this seems to occur on a basin scale, but not in regions with high heterogeneity due to the previously described mesoscale processes. The same pattern was observed using the estimates of correlations for the central Chile coastal region, where no significant correlation was observed except for with physical variables, such as temperature, salinity and NO₂⁻.

Surprisingly, CH₄ strongly correlates with NO₂⁻. This nutrient is rarely present in seawater (only in submicromolar levels), but it can accumulate in the areas known as the primary and secondary NO₂⁻ maximum. The first one is associated with aerobic ammonium oxidation that occurs at the base of the euphotic zone (Zafiriou et al., 1992), while the second one takes places within the oxygen minimum layer, and exists due to dissimilative NO₃⁻reduction and denitrification in cases where the water column is functionally anoxic (Thamdrup et al., 2012). Thus, the strong positive correlations found between CH₄ and NO₂⁻, and the lack of negative correlations that were expected to occur (Table 2), seem to be associated with the primary NO₂⁻ maximum together with the maximum stratification layer and oxycline. Also, there is a secondary fluorescence peak where marine photosynthetic picocyanobacteria dominate, Prochlorococuss and Synechococuss, which are adapted to low light conditions (Bouman et al., 2006). Recently, Bižić et al. (2020) attributed the production of CH₄ with primary production driven by photoautotrophs, including the previously mentioned cyanobacteria, introducing further complexity in the identification of marine methanogenic pathways.

Finally, in order to estimate a marine CH₄ budget and to determine the role of the SESP region (if it is a source or sink of CH₄ to the atmosphere), the region is separated into three longitudinal zones according to the trophic criteria (Table 1). Based on data of the surface CH₄ concentration, obtained during CIMAR 21, CIMAR 22, and those from the FDS and Galathea cruise (unpublished data), along with Wanninkhof´s parameterization (Wanninkhof, 1992) and the shipboard wind speed, the air-sea CH₄ fluxes range from −0.19 to 19.12 µmol m⁻² d⁻¹ (mean ± SD = 4,72 ± 4,67). Remarkably, even compared with the most commonly used parameterizations from Wanninkhof (2014) and Nightingale et al. (2000), there are no significant differences (less than 5% difference), thus the W92 is applied because it is more commonly used and easily compared with other studies, such as those from the subtropical ESP (Bates et al., 1996; Kelley & Jeffrey, 2002; Yoshida et al., 2011; Yoshikawa et al., 2014).

Except for the three stations located within the oligotrophic region, all the estimated CH₄ fluxes are positive (i.e., from the ocean to air, or efflux into the atmosphere) with a maximum value found at St. 11, where there is an ACs eddy at the limit of mesotrophic region (∼80°W). When effluxes are separated by trophic area, mean values decrease substantially from the coast in an offshore direction, with values of 7.77 ± 4.71, 5.22 ± 4.48, and 0.40 ± 0.65 µmol m⁻² d⁻¹ for the eutrophic, mesotrophic, and oligotrophic region, respectively (Table 3). CH₄ fluxes in the eutrophic zone, defined as the coastal area within 50 km from the coast, have the highest mean efflux, as expected for coastal areas. In fact, the global CH₄ emissions calculated previously by Bange et al. (1994) and Borges et al. (2016), and even more recently by Weber et al. (2019), revealed that approximately 60%–75% of the total oceanic CH₄ flux comes from coastal environments. Although coastal upwelling areas comprise only 1% of the total area of the ocean, they represent major sites of CH₄ outgassing due to vertical advection, which enhances CH₄ emissions; as observed in the NW and SW Arabian Sea, the coast off Oregon, and off the Canary Islands and Namibia (SW Africa) (Kock et al., 2008; Monteiro et al., 2006; Owens et al., 1991; Rehder et al., 2002; Sudheesh et al., 2020). At ∼36.5°S off central Chile, CH₄ air-sea fluxes average 10.94 ± 7.48 µmol m⁻² d⁻¹, based on a 12-year time series that included monthly sampling (Farías et al., 2020). This indicates an exponential response in marine CH₄ emissions, as the coastal systems become shallower and sediments are at a closer proximity. Thus, CH₄ is released from the seafloor and can escape into the atmosphere prior to oxidation (Borges et al., 2016; Weber et al., 2019).

Table 3. Regional Emission of Methane From the Ocean to the Atmosphere (in Gg per Year) Mostly Based on Spring-Summer time

CH4 efflux	Unit	Eutrophic zone	Mesotrophic zone	Oligotrophic zone	Total SESP	Global St ocean	Global ocean
Area	km^2	1,35,000	24,55,000	29,65,095	55,55,095	9,29,00,000	36,10,00,000
mean ± SD efflux	µmol m^−2 d^−1	7.77 ± 4.71	5.22 ± 4,48	0.4 ± 0.65
Range	µmol m^−2 d^−1	1.40–14.37	1.51–19.12	−0.19–2.15
N		11	30	11
Annual mean ± SD	µmol yr^−1	3.83E + 14 ± 2.32E + 14	4.68E + 15 ± 4.01E+15	4.33E + 14 ± 7.03E + 14
Annual	Gg yr^−1	6.13 ± 3.71	74.84 ± 64.23	6.93 ± 11.26	87.89 ± 75.41	1,469.9	5,711.73
mean ± error

• Air-sea fluxes were calculated using transfer velocities from Wanninkhof (1992) and the regional area was based on distance among latitudes and longitudes, while global ocean area was based on Levitus (1982).

Using the ocean area extension for each defined zone, and its average CH₄ flux (Table 3), an overall annual CH₄ flux is calculated. For coastal or the eutrophic zone, the net CH₄ exchange is about 6.12 Gg CH₄ yr⁻¹, however, when this estimate is made for the mesotrophic and oligotrophic zone, the net annual CH₄ exchange is 64.23 and 6.92 Gt CH₄ yr⁻¹, respectively. In total, this adds up to 88.7 Gg for the total area of the SESP, about 5.56 × 10⁶ km² (from 17° to 37°S and from the coast up to 110°W). It is important to note that, except for CO₂, there is a lack of general information on the air-sea gas exchange in the SESP, and inclusive for the whole south Pacific basin, as there are fewer available studies compared to the Atlantic Ocean (for example, through the Atlantic Meridional Transect program) and the Indian Ocean (Oudot et al., 2002; Owens et al., 1991; Rhee et al., 2009). Some CH₄ budgets have been estimated for the central Pacific by Bates et al. (1996) and Kelley and Jeffrey (2002). For example, in the central Pacific between the 15°–30°S and 30–45°S latitudinal bands, Bates et al. (1996) estimated a CH₄ exchange of 76 and 41 Gg y⁻¹, respectively over an ocean area of 92.9 × 10⁶ Km²; whereas Kelley and Jeffrey (2002) conducted a latitudinal transect around ∼ 77°W for an ocean area of 3.4 × 10⁸ Km² and found a mean CH₄ efflux of 0.38 ± 0.32 µmol m⁻¹ d⁻¹ and an overall CH₄ annual flux of 800 ± 600 Gg CH₄ yr⁻¹. If this weighted annual flux is extrapolated through the global ocean extension, the net CH₄ emission is 5,711 Gg y⁻¹ (or 5.71 Tg y⁻¹).

Compiled data on CH₄ made by Saunois et al. (2020), based on bottom up reservoir approaches, indicated that the oceanic emissions of biogenic CH₄ is around 6 (4–10) Tg y⁻¹. The range of uncertainty includes an increased budget that is presented by Bange et al. (1994), with a global source of 11–18 Tg CH₄ yr⁻¹, including a majority contribution from coastal regions, and low estimates of 0.2–3 Tg CH₄ yr⁻¹ (Bates et al., 1996). These are based on basin-wide observations mostly in oligotrophic regions. The latter estimates are an order of magnitude less than the estimated budget for the open ocean (saturation anomaly of ∼0.04) and the continental shelf (saturation anomaly 0.2) (see Rhee et al., 2009).

In this study, which is also based over a basin scale, the CH₄ emission is one order of magnitude higher than those reported by Bates et al. (1996). The study area has an unexpectedly higher CH₄ emission in the coastal transition zone or mesotrophic zone due to the existence of complex bathymetry, islands, seamounts, and intense mesoscale activities, which also include the narrow coastal system. This emphasizes the importance of both the upwelling regime in the coastal zone, and also of mesoscale processes, as a significant CH₄ source, which also trigger considerable CH₄ accumulations. It is important to note that most of the data was obtained in the spring-summer, period which is expected to be slightly higher than in the autumn-winter period in the South Pacific Ocean (Bates et al., 1996). Thus, it is expected that over an annual basis these results would give a slight overestimation.

4 Conclusions

In general, CH₄ accumulation occurs within the low density water masses as STW, and in the upper layer of the ESPIW, associated with high stratification and the oxycline. CH₄ accumulates in eutrophic areas subjected to coastal upwelling, and in mesotrophic areas related with intense mesoscale activities. Mesoscale structures, such as eddies and others associated with the island mass effect, as observed in JFA, cause considerable spatial CH₄ variability in the ESP region and therefore do not present any clear relationship between CH₄ concentration and Chl-a levels.

The SESP acts as important net source of CH₄ toward the atmosphere, with efflux varying from 0.13 and 19.1 µmol m⁻² d⁻¹, the highest CH₄ emission rates take place in the band of coastal waters and in the mesotrophic zone, with emission rates that are one order of magnitude higher than in oligotrophic waters. The mesotrophic area had relatively high CH₄ effluxes, due to the island mass effect as well as other mesoscale processes, which trigger surface and subsurface CH₄ accumulation.

This study contributes toward improving estimates of CH₄ emissions within marine ecosystems, and further investigates the feedback between emissions and anthropogenic forcing (e.g., climate change, stratification, eutrophication, among others). The regional CH₄ emission budget of ∼5.71 Tg y⁻¹ is greater than previous estimates over the scale of the entire basin, highlighting the need to improve the spatial resolution in sampling methods in order to capture processes that lead to CH₄ production and effluxes, especially in the coastal transition zone that links the coastal zone widely recognized to have a much greater level of CH₄ emissions compared to the open ocean (at the subtropical gyre's core).

Acknowledgments

Data was collected during the CIMAR 21 cruise (CONA CIMAR-21 Islas C21I 15-06y C21I 15-12) the Cimar 22 cruise (CONA CIMAR-21 Islas C22I 16-05y C22I 16-07). The authors thank the Servicio Hidrográfico y Oceanográfico de la Armada (SHOA) for the funds granted and to the Chilean Army, Armada de Chile, for providing the AGS 61 Cabo de Hornos. They acknowledge the Center for Climate and Resilience Research (1511009) and the Instituto Milenio en Socio-Ecología Costera SECOS (ICM 2019_015 ICM-ANID). We appreciate Dr. Morales's invitation to participate in the FDS cruise off central Chile. L. F was also supported by FONDECYT N1161138. Finally, the authors are very grateful of Cheryl Kelley, Tim Bates, and Chisato Yoshikawa who provided the CH₄ data for developing Figure 9.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Open Research

Data Availability Statement

The data archiving is underway and will be available at www.pangaea.de. https://doi.pangaea.de/10.1594/PANGAEA.922967