INTRODUCTION
The calibration of radiocarbon (14C) ages of terrestrial materials requires a good record of the well-mixed atmospheric14C concentration for the last 50,000 yr. This record has been built based on tree-ring, varved lake sediments, and speleothem data (Reimer et al. Reference Reimer, Bard, Bayliss, Beck, Blackwell, Bronk, Buck, Cheng, Lawrence, Friedrich, Grootes, Guilderson, Haflidason, Hajdas, Hatté, Heaton, Hoffman, Hogg, Hughen, Kaiser, Kromer, Manning, Niu, Reimer, Richards, Scott, Southon, Staff, Turney and van der Plicht2013). For marine samples, however, a record of past surface ocean 14C activity is available derived from planktonic foraminifera and warm water corals, showing a mean difference of ~400 14C yr from that of the atmosphere (Hughen et al. Reference Hughen, Baillie, Bard, Beck, Bertrand, Blackwell, Buck, Burr, Cutler, Damon, Edward, Fairbanks, Friedrich, Guilderson, Kromer, McCormac, Manning, Bronk Ramsey, Reimer, Reimer, Remmele, Southon, Stuiver, Talamo, Taylor, van der Plicht and Weyenmeyer2004). This difference is known as the marine reservoir age or “R” (Stuiver et al. Reference Stuiver, Pearson and Braziunas1986; Stuiver and Braziunas Reference Stuiver and Braziunas1993) and it is the result of the mixing of the surface ocean with older, deeper waters (Broecker Reference Broecker1987; Broecker et al. Reference Broecker, Klas, Clark, Bonani, Ivy and Wolfli1991; Stuiver and Braziunas Reference Stuiver and Braziunas1993). A critical complication arises since the rate of mixing can be very different in different parts of the ocean, leading to spatial differences in the value of R, like those found in high latitudes (Heier-Nielsen et al. Reference Heier-Nielsen, Heinemeier, Nielsen and Rud1995; Berkman and Forman Reference Berkman and Forman1996) or zones with strong upwelling regimes like the southeast Pacific (Southon et al. Reference Southon, Oakland and True1995). It is therefore mandatory to have an accurate (as much as possible) knowledge of the regional 14C age difference from that of the global mean or what is called ∆R (Stuiver and Braziunas 1986; for a complete recent review see Alves et al. Reference Alves, Macario, Ascough and Bronk Ramsey2018). This value can be positive (older age than global surface ocean waters) like in upwelling regions, or negative (younger) like in highly stratified systems (Ascough et al. Reference Ascough, Cook and Dugmore2005; Hinojosa et al. Reference Hinojosa, Moy, Prior, Eglinton, McIntyre, Stirling and Wilson2015). There have been considerable efforts in trying to assess ∆R values associated to different oceanic regimes, given their importance in paleoceanographic reconstructions and age model accuracy. For example, in the Gulf of California, Goodfriend and Flessa (Reference Goodfriend and Flessa1997) calculated a ∆R of 550±108 and 408±122 14C yr for the northern and central gulf samples (all cited values were updated using the latest marine calibration curve [Reimer et al. Reference Reimer, Bard, Bayliss, Beck, Blackwell, Bronk, Buck, Cheng, Lawrence, Friedrich, Grootes, Guilderson, Haflidason, Hajdas, Hatté, Heaton, Hoffman, Hogg, Hughen, Kaiser, Kromer, Manning, Niu, Reimer, Richards, Scott, Southon, Staff, Turney and van der Plicht2013]), respectively. This high reservoir age corresponds to a region with strong upwelling of deeper waters. Culleton et al. (Reference Culleton, Kennett, Ingram, Erlandson and Southon2006) studied intra-shell marine reservoir corrections from the Santa Barbara Channel region, and found modern ∆R mean values of 366±35 and 278±45 14C yr, with ranges of 180 and 240 14C yr respectively. This variability was also associated with the presence of coastal upwelling of deeper waters, together with seasonal inputs of 14C derived from terrestrial runoff. Southon et al. (Reference Southon, Kashgarian, Metivier and Yim2002) presented ∆R data for other upwelling zones such as the Arabian Sea, with mean regional values ranging from 154±25 to 213±30 14C yr, and for the South China Sea, with values close to zero due to the presence of Pacific waters with high 14C content. Petchey et al. (Reference Petchey, Anderson, Zondervan, Ulm and Hogg2008) presented 31 reservoir correction ages for an open-ocean zone as the South Pacific subtropical gyre, with most of the values near zero due to the stable surface water conditions imposed by the ocean circulation. On the other hand, Hinojosa et al. (Reference Hinojosa, Moy, Prior, Eglinton, McIntyre, Stirling and Wilson2015) investigated 14C reservoir ages in the fjord zone of southwest New Zealand, obtaining a ∆R of 59±35 14C yr. This value was related to the influence of 14C-depleted subantarctic water into the fjords, which may lead to greater reservoir ages in the region. In a similar context, Heier-Nielsen et al. (Reference Heier-Nielsen, Heinemeier, Nielsen and Rud1995) found low reservoir ages for the Danish coastal zone with a mean ∆R value of 29±16 14C yr, derived from the North Atlantic and the North Sea.
Precise knowledge of ∆R is of particular importance for regions with large oceanographic and hydrographic contrasts like the coast of Chile. Located at the western south margin of the South American continent from ~18°S to ~56°S, it is characterized by (1) different upwelling-regime zones (permanent to seasonal) from the north (~18°S) to central-south (~40°S) Chile (Letelier et al. Reference Letelier, Pizarro and Nuñez2009); (2) increasing input of freshwater through precipitation and continental runoff toward the south (Torres et al. Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011); (3) the presence of fjords and glaciers that are remnants of the last Ice Age; and (4) the influence of Subantarctic waters (SAW) flowing northward associated to the Humboldt Current System (HCS) from around 44°S (Silva et al. Reference Silva, Rojas and Fedele2009). The consequent latitudinal variations of 14CO2 exchange between the atmosphere and the ocean caused by these factors, surely affects the 14C concentration (and subsequent age) of the surface waters along the Chilean coast.
Previous studies of ∆R along the Chilean coast have been carried out by Southon et al. (Reference Southon, Oakland and True1995) from archeological sites in northern Chile (19ºS). They compared 14C ages of materials of terrestrial and marine origin, obtaining a ∆R range of 135±110 14C yr for the period 100–340 AD, up to 245±85 14C yr for the period 340–530 AD. Similarly, pre-bomb 20th century shell measurements of Concholepas concholepas at 24°S and Tegula atra at 33°S show a ∆R of 175±41 and 315±79 14C yr, respectively (Taylor and Berger Reference Taylor and Berger1967). In contrast, Ingram and Southon (Reference Ingram and Southon1996) reported pre-bomb mytilid shells from Valparaíso (33°S) with a lower ∆R of 60±55 14C yr. and for the inner Patagonian fjord of Puerto Natales (51°S) a relatively large ∆R of 220±46 14C yr. By studying Holocene variations of ∆R, Ortlieb et al. (Reference Ortlieb, Vargas and Saliège2011) obtained a range of modern pre-bomb ∆R values for the northern Chilean coast between 20–23°S with a mean of 253±207 14C yr, showing high variability in ∆R associated to changes in upwelling regimes and/or by climatic oscillations like the ENSO cycle. Finally, Carré et al. (Reference Carré, Jackson, Maldonado, Chase and Sachs2016) and more recently Latorre et al. (Reference Latorre, De Pol-Holz, Carter and Santoro2017) have calculated ∆R based on paired terrestrial-marine 14C measurements from Holocene shell middens in central and northern Chile, and show that, despite recent relatively stable oceanographic conditions, ∆R in this area can be highly variable at millennial time-scales.
Here we provide new modern pre-bomb estimates of ∆R for the Chilean Continental Margin based on 14C measurements on marine shells from museum collections of known age. In particular, our data fills a large gap in the regional knowledge of ∆R along the eastern South Pacific Ocean, where no ∆R information exists between 33º and 51ºS.
REGIONAL SETTING
The general oceanographic setting of the eastern South Pacific is influenced by the Humboldt Current System (Figure 1). The northward Chile-Peru branch originates at about 45°S where the northern Circumpolar Current impinges against South America bifurcating into a northern component (Chile-Peru current) and a southern one (Cape Horn Current) (Silva and Neshyba Reference Silva and Neshyba1977; Strub et al. Reference Strub, Mesías, Montecino, Rutllant and Salinas1998). The Chile-Peru current advects (SAW) at the surface, a water-mass characterized by very low values of salinity (<34.5) and temperature (<15ºC). SAAW also enters the south Chilean fjords area where it mixes with continental waters (Silva and Vargas Reference Silva and Vargas2014) while it is transported southward by the Cape Horn Current. The mixing causes further freshening due to the large freshwater input from Patagonian glaciers (Silva et al. Reference Silva, Calvete and Sievers1997). The southward components of the Humboldt Current System, the Peru-Chile Countercurrent, transports Subtropical Surface Waters (SSW), while the Peru-Chile Undercurrent carries the Equatorial Subsurface Water (ESSW at the subsurface from about 10°S to 48°S (Silva and Neshyba Reference Silva and Neshyba1979; Toggweiler et al. Reference Toggweiler, Dixon and Broecker1991). This water mass is characterized by high nutrient and low oxygen content and it is implicated in the upwelling process occurring along the coast. At the north, upwelling is a perennial feature while south of 30°S it is a highly seasonal process (Letelier et al. Reference Letelier, Pizarro and Nuñez2009).
Figure 1 Surface ocean currents in the eastern South Pacific along the Chilean coast. The West Wind Drift feeds major currents flowing north/southward. Northward: the Humboldt Current and southward, the Cape Horn Current. The Perú-Chile Counter Current (PCCC) with its Undercurrent homologue (PCU). Subsurface and intermediate water masses are also shown (white: northward, black: southward). SSW: Sub Tropical Surface Water; SAW: Subantarctic Surface Water; EESW: Equatorial Subsurface Water; AAIW: Antarctic Intermediate Water.
South of 42°S, ice stream remnants of the last Ice Age (ca. 20,000 BP) supply large quantities of fresh-water and promote low salinity and cold temperatures along the coast (Acha et al. Reference Acha, Mianzan, Guerrero, Favero and Bava2004), while wind speed increases southward reaching a maximum of 20 ms–1 at the Polar Jet latitude band of near 40–50ºS (Torres et al. Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011). This has the important effect of increasing CO2 solubility at the atmosphere-ocean interface (Wanninkhof and McGillis Reference Wanninkhof and McGillis1999). In addition, our southern study region (south of 40°S) is a highly productive environment due to the input of continentally derived nutrients (like iron and silica) from freshwater runoff along the Patagonian Fjords, which contributes further to regionally CO2-undersaturated surface waters (Torres et al. Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011).
Torres et al. (Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011) described CO2 fluxes along the Chilean coast from 23°S to 56°S, comprising almost its entire latitudinal range. They found a strong relationship between CO2 outgassing and the upwelling-regime zones as far south as ~40°S because of the presence of CO2 supersaturated deeper waters at the surface, enhanced by heating at shallow depths. From ~42°S southward, cold coastal waters act as sink for CO2 by the stratification of the water due to the great input of cold freshwater (Dávila et al. Reference Dávila, Figueroa and Müller2002) and increased rates of photosynthesis that result respectively in high CO2 solubility and low pCO2 at these latitudes.
MATERIALS AND METHODS
Samples used in this study were provided by The Museum of Zoology of the University of Concepción (n=11), and the collection of the Laboratory of Malacology at the National Museum of Natural History in Santiago, Chile (n=26). We selected samples with labeled collection dates ranging between AD 1852 and 1960. Mollusks used in this work include gastropods and bivalves (Table 1). Bivalves selected were as follows: Arca sp., Diplodonta inconspicua, Ensis macha, Gary solida, Glycimerys ovatus, Perumytilus purpuratus, Petricola rugosa, Tagelus dombeii, Ostrea chilensis. Gastropods: Crepidula sp., Fisurella cumingi, Littorina peruviana, Nacella clypeater, Nacella deaurata, Nacella magellanica, Nacella Polaris, Purpura xanthostoma, Scurria lacerata, and Scurria zebrina. The available museum documentation that was present with the samples included: species name, sampling date and location of collection or broad geographical provenance (e.g. “Magallanes Region”). Our selected samples cover a geographical span from Arica in northern Chile (19°S) to Hoste Island at the southern tip of South America (55°S).
Table 1 Modern reservoir age deviations along the Chilean coast (19º–55ºS).
*Ortlieb et al. (Reference Ortlieb, Vargas and Saliège2011), **Taylor and Berger (Reference Taylor and Berger1967), ***Carré et al. (Reference Carré, Jackson, Maldonado, Chase and Sachs2016), ****Ingram and Southon (Reference Ingram and Southon1996).
#These samples were excluded from the calculation of the regional mean.
a δ13C values of this study were measured at the stable isotope facility of the University of California at Davis.
For each selected individual, we used a Dremel® rotary tool for cutting 20–30 mg fragments from either the ventral side of bivalves, or from the outer lip of gastropod shells. These fragments were then leached to 50% of their weight using hydrochloric acid (0.1N) in order to remove any potential secondary carbonate and rinsed with deionized water. Samples were then hydrolyzed with 0.8 mL of 85% phosphoric acid by injecting acid through the septum of an evacuated vacutainer for ca. 30–60 min at 70°C. The CO2 obtained from the samples was graphitized to 550°C by hydrogen reduction with an iron catalyst (Vogel et al. Reference Vogel, Southon, Nelson and Brown1984; Loyd et al. Reference Loyd, Vogel and Trumbore1991). 14C measurements were performed at the W.M. Keck Carbon Cycle AMS Laboratory of the University of California, Irvine and 14C ages were reported following the conventions of Stuiver and Polach (Reference Stuiver and Polach1977). ∆R values were calculated by subtracting the sample’s conventional 14C age from the Marine13 (Reimer et al. Reference Reimer, Bard, Bayliss, Beck, Blackwell, Bronk, Buck, Cheng, Lawrence, Friedrich, Grootes, Guilderson, Haflidason, Hajdas, Hatté, Heaton, Hoffman, Hogg, Hughen, Kaiser, Kromer, Manning, Niu, Reimer, Richards, Scott, Southon, Staff, Turney and van der Plicht2013) 14C age at the calendar year labeled for each individual (see Southon et al. Reference Southon, Oakland and True1995 for a graphical description). For samples with calendar age later than 1950 AD we used the 1950 value of Marine13 as the model age. Individual ∆R errors where calculated as the square root of the sum of the analytical 1-sigma error and the model error squared. We also provide regional error-weighed mean ∆R values for 5 macro zones along the Chilean coast and their associated standard deviation.
In order to maximize our sample availability, we included many individuals collected between 1950 and 1960 AD. We recognize that this might not be ideal since some bomb-derived 14C might be already present in the surface ocean (Ascough et al. Reference Ascough, Cook and Dugmore2005). This effect however should be very minor for the eastern South Pacific (see for example figure 4 in Druffel [Reference Druffel1981]; figure 1 in Druffel and Griffin [Reference Druffel and Griffin1995] and figure 3 in Guilderson et al. [Reference Guilderson, Schrag, Goddard, Kashgarian, Wellington and Linsley2000]) and we therefore chose to include results of technically post-bomb shells. As mentioned before, for consistency previously published ∆R values were recalculated using the Marine13 calibration curve.
RESULTS AND DISCUSSION
14C Ages
Table 1 shows the conventional 14C ages, collection date, location, and latitude for the 37 samples analyzed and data from previous studies in the region using known-age mollusk shells. Weighed mean ∆R values were calculated by proximity of location of the samples to establish a reliable and representative reservoir offset correction. All existing ∆R for the Chilean coast as a function of latitude are shown in Figure 2. Apparent 14C ages range from –650±20 to 48,800±1500 14C yr and ∆R from 25±30 to 260±3014C yr, with regional weighed means ranging from 52±47 to 196±128 14C yr (Figure 3). Anomalous 14C dates can be observed in five samples (shown in bold in Table 1), and were not considered for the calculation of the regional mean. Three were discarded because of ancient 14C dates (Arica 1948, Niebla 1911 and Niebla 1960 AD), one was not used for mean ∆R calculations due to potential 14C artifacts in the gastropod species Littorina peruviana (Valparaíso 1953), and one sample from Hoste Island (1954) was not used because of the clear influence of bomb-derived 14C (negative 14C ages). ANOVA statistical analysis of the set of mean ∆R ages from the Chilean coast reveal significant differences between all regions (p<0.001), with the exception of the northern ∆R value against 33°S mean which have no significant differences. Nevertheless, there is a clear latitudinal tendency of decreasing values along the coast exists as shown in Figure 2.
Figure 2 ∆pCO2 (difference in partial pressure of CO2 between the ocean with respect to the atmosphere, upper panel) and all modern ∆R values existing for the Chilean coast (lower panel). A clear decreasing pattern can be observed in CO2-saturation values from north to the south, similar to decreasing ∆R values. ∆pCO2 data extracted from Torres et al. (Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011). In solid diamonds Taylor and Berger (Reference Taylor and Berger1967), semi-filled hexagons Ingram and Southon (Reference Ingram and Southon1996), filled circles Ortlieb et al. (Reference Ortlieb, Vargas and Saliège2011) and black square Carré et al. (Reference Carré, Jackson, Maldonado, Chase and Sachs2016).
Figure 3 Mean regional ∆R values for the Chilean coast including previous studies. There are 5 regional values for the coast (gray) and one for the open ocean. Colors represent annual mean sea surface temperature obtained from the World Ocean Circulation Experiment (http://www.ewoce.org/). (Please see online version for color figures.)
Marine Species Habitat and Feeding Mode
Selection of mollusks species for calculating ∆R values is important because the source of the carbon from which their shells is formed can be derived from dissolved inorganic carbon (DIC) of the seawater as well as carbon derived from metabolic processes like feeding in different proportions (Tanaka et al. Reference Tanaka, Monaghan and Rye1986; Lorrain et al. Reference Lorrain, Paulet, Chauvaud L, Dunbar, Mucciarone and Fontugne2004; McConnaughey and Gillikin Reference Mc Connaughey and Gillikin2008). In fact, Gillikin et al. (Reference Gillikin, Lorrain, Bouillon, Willenz and Dehairs2006) showed that metabolic carbon varies from 10 to 35% in mollusks shells. Because of this, it is important to know the feeding habits and the environment where selected specimens lived. Ideally, chosen species for 14C dating should preferably be those with suspension-feeding habits, like bivalves, because they incorporate phytoplankton at the same time they absorb DIC (Tanaka et al. Reference Tanaka, Monaghan and Rye1986; Forman and Polyak Reference Forman and Polyak1997) and are considered to be in equilibrium with seawater DIC (Hogg et al. Reference Hogg, Higham and Dahm1998).
In contrast, gastropods may incorporate aged carbon from other sources like limestone, thus incorporating another unknown reservoir effect becoming unreliable for ∆R calculations (Dye Reference Dye1994). Pigati et al. (Reference Pigati, Rech and Nekola2010) performed a comprehensive analysis of the presence of “dead carbon” in small terrestrial gastropods of 3749 individual shells to test the reliability of 14C dating in these organisms. They found a nearly 78% of the samples without influence of the limestone problem, bringing reliable 14C dates even when the shells where collected at sites with strong influence of carbonate rocks. More recently, Macario et al. (Reference Macario, Alves, Carvalho, Oliveira, Bronk Ramsey, Chivall, Souza, Simone and Cavallari2016) did not find significant differences in 14C ages of terrestrial gastropods in the Brazilian coast. They analyzed the 14C content of individuals of the genera Megalobulimus and Thaumastus from the bomb-period (1948–2004). They found that the samples were not affected by the presence of limestone albeit their habitat was strongly influenced by carbonate rocks, and represented consistently the atmospheric 14C concentration during the time they lived. Despite Pigati et al. (Reference Pigati, Rech and Nekola2010) and Macario et al. (Reference Macario, Alves, Carvalho, Oliveira, Bronk Ramsey, Chivall, Souza, Simone and Cavallari2016) results seem encouraging for the inclusion of gastropods in 14C reservoir assessments, we consider that further analysis is needed in marine organisms before including these organisms in 14C age studies. Although there is no sufficient information to extrapolate this to the marine environment, this may be used as an indicator of the confidence of dating gastropods shells. Along the coast of Chile, there are marine sedimentary rocks distributed intermittently from the north down to around Chiloé (42°S, Paskoff Reference Paskoff2010) which might contribute to the addition of ancient carbon to the coastal mollusk shells if they are truly using it. According to the SERNAGEOMIN (Chilean Geology and Mining Service) Chilean Geologic map, two of our sample sites might present carbonate rocks: Antofagasta (23°S), which is located over a Pleistocene coastal marine sedimentary sequence of biogenic carbonates, and the Magallanes Strait (53°S), which presents a complex geological formation dominated by deposits of glacial and fluvial origin as well as Mesozoic-Cenozoic granite batholiths (Hervé et al. Reference Hervé, Quiroz and Duhart2009) but also has Cretaceous and Paleogene marine sedimentary sequences. In these locations, our data do not show any anomalously large ∆R estimates with the exception of Puerto Natales (see below).
The species used in this study were selected under the criteria of calendar year, favoring pre-bomb samples. In order to complete ∆R gaps along the Chilean coast, we chose to include samples that could present problems due to deposit feeding habits. Other studies have been performed dating deposit feeder shellfishes and carnivorous gastropods (Hogg et al. Reference Hogg, Higham and Dahm1998; Soares and Dias Reference Soares and Dias2006), and apparently there are no important differences attributable to the feeding way of the organisms, but results obtained from this mollusk type need to be analyzed with caution. At 33°S for example, ∆R values were exclusively obtained from deposit-feeders. This is not ideal since gastropods can scrape and take dead carbon from their feeding substrate. Notwithstanding, our results match the ∆R values obtained by Taylor and Berger (Reference Taylor and Berger1967) and Ingram and Southon (Reference Ingram and Southon1996) for the same locality and there is no indication of large amount of dead carbon influencing our ∆R estimations. We excluded a Littorina peruviana ∆R estimation from the regional mean because of some enriched 14C giving a negative ∆R value as is the case with the periwinkle Littorina littorea (Petchey et al. Reference Petchey, Ulm, David, McNiven, Asmussen, Tomkins, Richards, Rowe, Leavesley, Mandui and Stanisic2012). Freshwater input through rivers or runoff can bring CO2 derived from organic debris product of plants or soil decomposition, which can result in lower reservoir ages (Stuiver and Braziunas Reference Stuiver and Braziunas1993; Southon et al. Reference Southon, Kashgarian, Metivier and Yim2002). This result suggests that ∆R variability observed at 33°S, is to a great extent influenced by environmental variability like seasonal upwelling regimes or freshwater detritus-carbon input, whereas more southern reservoir ages are governed by low salinity and low temperature waters that favor CO2 exchange between the surface ocean and the atmosphere.
∆R and Regional Oceanography
According to Toggweiler et al. (Reference Toggweiler, Dixon and Broecker1991), ESSW upwells in the Peru and northern Chilean coastal region with the lowest ∆14C signal across the Pacific. We infer that this is causing the higher ∆R values found in the northern part of our study region down to 40°S as shown by our data. Permanent upwelling regimes off northern Chile (20–30°S) promote higher ∆R ages than those found in the south, where this process presents seasonal intermittency (Letelier et al. Reference Letelier, Pizarro and Nuñez2009). We attribute the large variability in ∆R values observed at mid-latitude coastal localities like Valparaíso (33°S, Figure 2) to this intermittency. New ∆R data calculated for this region (n=4) ranges from a minimum of 80±38 to a maximum of 260±30 14C yr bracketing well previous values published by Taylor and Berger (Reference Taylor and Berger1967), Ingram and Southon (Reference Ingram and Southon1996) and Carré et al. (Reference Carré, Jackson, Maldonado, Chase and Sachs2016). When analyzed together, all the data for this region give a mean ∆R of 148 14C yr with a large standard deviation of ±7814C yr. At the open ocean site of the Juan Fernández archipelago (between 78.5° and 80.5°W, ca. 800 km from the Chilean coast) a low ∆R of 40±38 14C yr compares well with the extensive dataset of Petchey et al. (Reference Petchey, Anderson, Zondervan, Ulm and Hogg2008) for the south Pacific Subtropical Gyre island system. We note that this is the first time that a ∆R value is given for this area.
The data for the central-south region of this study comprise the localities of Penco, San Vicente and Quiriquina Island (located inside the Bay of Concepción all located near 37°S, and showing a ∆R mean of 194±24 14C yr. This zone is characterized by strong seasonal upwelling of 14C-depleted subsurface waters like off Valparaíso, although in this case, the data clusters well around the mean (low standard deviation). These values are the first published for this region.
Further south around 42°S, ∆R ages decrease to a mean of 141±43 14C yr, ranging from 55±38 to 215±38 14C yr. Although climatologically this latitude south of the main northerly upwelling-favorable wind region, strong seasonal variability might include upwelling events (Letelier et al. Reference Letelier, Pizarro and Nuñez2009). The species analyzed from this region are mostly bivalves, with two exceptions (Crepidula sp. and Purpura xanthostoma), which are in good agreement with respect to their ∆R (170±34 and 180±34 14C yr, respectively) with the rest of the data. The variability observed is probably related to different water mass influences as suggested by the carbonate δ13C data. Broadly, this region is bathed by SAAW, probably advecting a surficial 14C-depleted signal originated at the Southern Ocean upwelling region. SAAW presents high δ13C values (>1‰, Kroopnick Reference Kroopnick1985) due to high productivity and CO2 consumption by phytoplankton. Primary productivity estimates in this region can reach values between 1 and 23 mg C m–3 hr–1 in winter and spring seasons, respectively (Iriarte et al. Reference Iriarte, González, Liu, Rivas and Valenzuela2007). Besides, episodes of upwelling of subsurface waters can be present at this latitude and may decrease the 13C isotopic signal because of organic matter remineralization at deeper waters. According to Silva and Neshyba (Reference Silva and Neshyba1979), ESSW can reach latitudes as far as 45°S and it is possible that it brings its lower δ13C signal (<1‰) to the surface. Those water masses isotopic signals may be affecting the 13C content that is being taken by coastal mollusk to build their shells. However, there is also a large input of fresh water by precipitation, which might contribute to a reduction of the reservoir effect (Stuiver and Braziunas Reference Stuiver and Braziunas1993; Southon et al. Reference Southon, Kashgarian, Metivier and Yim2002) of this region and to the south, as well as large tidal cycles that could encourage gaseous mixing between the surface ocean and the atmosphere (Ingram and Southon Reference Ingram and Southon1996). In summary, we expect a large ∆R variability at this part of the ocean on seasonal and interannual time-scales, as shown by the data.
Southward, between 51° and 54°S, we have calculated eight new ∆R values ranging from 25±30 to 60±30 14C yr. These values are significantly lower thanthe one reported by Ingram and Southon (Reference Ingram and Southon1996) of 220±46 14C yr for Puerto Natales (51°S). We note that our new data comes from open to relatively open water locations and therefore, the large difference might be the result of complex oceanographic processes acting upon ∆R and by the large difference between water-mass ∆R endmembers in this area. For instance, SAAW can occasionally mix with fjord waters (Silva et al. Reference Silva, Rojas and Fedele2009) lowering the14C concentration (increasing ∆R) inland. In addition, Puerto Natales is located at the Señoret Fjord which is far from the open ocean and adjacent to the large Montt gulf. There are indications that strong hypoxia develops at the bottom of the Montt gulf, which could imply that below the sill depth, waters could age considerably potentially influencing surface ∆R values (Robert B. Dunbar, pers. comm.). Besides, lower values of ∆R are expected from the uptake of atmospheric CO2 resulting from a greater solubility in seawater due to a lower temperature and salinity, CO2 subsaturation of surface waters and the lack of coastal upwelled-waters (Siani et al. Reference Siani, Paterne, Arnold, Bard, Métivier, Tisnerat and Bassinot2000; Torres et al. Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011). Therefore, we believe that these factors must be taken into account when choosing a ∆R value for marine 14C calibrations in materials from this area.
Based on our new and previously published data we therefore suggest five broad regional mean ∆R values based on predominant oceanographic processes occurring along the Chilean coast. These are: Northern upwelling system (18–30°S), Central upwelling system (30–33°S), Southern upwelling system (33–39°S), Northern Patagonia (39–43°S) and Southern Patagonia (43–55°S) (Figure 3).
∆R and pCO2
CO2 content of surface waters along the Chilean coast from 18° to about 40°S is intimately related to the upwelling of CO2-saturated subsurface waters (Latorre et al. Reference Latorre, De Pol-Holz, Carter and Santoro2017). In fact, ∆R values along the coast fit well with the mean ∆pCO2 (the difference in the ocean-atmosphere CO2 partial pressures, calculated such that positive values represent a larger ocean pCO2 and therefore a CO2 efflux to the atmosphere) data of Torres et al. (Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011).
From 40–42°S toward the pole, the combination of reduced upwelling, enhanced local biological productivity, the strengthening of wind stress at the surface and a greater solubility of CO2 in colder seawater, promotes major CO2 uptake into the ocean from the atmosphere, renewing its surface 14C content and thus, diminishing ∆R values of the surface (Siani et al. Reference Siani, Paterne, Arnold, Bard, Métivier, Tisnerat and Bassinot2000). In fact, wind speed has an important influence on gas transfer between the atmosphere and the ocean (Takahashi Reference Takahashi2001), because increasing of atmospheric transfer of CO2 to the ocean is directly related to higher wind speeds close to the surface. Along the Chilean coast, a trend of poleward increasing wind speed values has been reported according to Takahashi et al. (Reference Takahashi, Sutherland, Sweeney, Poisson, Metzl, Tilbrook, Bates, Wannikhof, Feely, Sabine, Olafsson and Nojiri2002) and Torres et al. (Reference Torres, Pantoja, Harada, González, Daneri, Frangopulos, Rutlant, Duarte, Rúiz-Halpern, Mayol and Fukasawa2011), ranging from 5.5 m/s at 19°S to 11 m/s at 54°S, with maximum reported values >20 m/s. In addition, this increase in speed is also related to a change of direction from upwelling favorable southerly winds inthe region between 18º to 40–42ºS to a net westerly direction south of it. In fact, Bard (Reference Bard1988) and Bard et al. (Reference Bard, Arnold, Mangeru, Paterne M, Labeyrie, Duprat, Mélières, Sonstegaard and Duplessy1994) reported that changes in mean wind speed of 50% can result in 100 14C yr lower ∆R values.
CONCLUSIONS
14C measurements on museum shells of known age were used to calculate new ∆R estimates for the Chilean coast, filling an important latitudinal gap from 33 to 53°S. In general, our estimates for the upwelling region north of 42ºS are in good agreement with previously published data. For this region ∆R weighed mean values range from 141±43 to 196±128 14C yr, thus we recommend using 180±27 14C yr as a reasonable choice for the calibration of marine material from the upwelling favorable wind section of the Chilean coast. For the southern Patagonian Fjord area we calculated a mean of 52±47 14C yr at 53ºS. These lower values represent the combined effect of reduced SST and sea surface salinity affecting the ∆pCO2 between the ocean and the atmosphere, with higher ∆pCO2 values being associated with higher ∆R (older waters). Finally, for the open ocean setting of the Juan Fernández archipelago, our data show a relatively minor i.e. 40±38 14C yr deviation from the global modern ocean-atmosphere difference of 400 yr. We believe these that these results will improve considerably our knowledge of the modern ∆R corrections needed for reliable calibration of 14C measurements of marine-derived materials for the Chilean coast.
ACKNOWLEDGMENTS
We thank the curator of the Museum of Zoology of the University of Concepción, Dr. Jorge Artigas, and the personnel of the Chilean National Museum of Natural History: curator Andrea Martínez, Dr. Sergio Letelier and Dr. Oscar Gálvez for the help provided with the samples. Funding was provided by FONDAP 15110009, Iniciativa Científica Milenio NC120066 and FONDECYT Grants #1140536 (RDP-H), #11100281 (RDP-H) and #1150763 (CL, RDP-H). We finally thank associate editor Quan Hua, Matthew Carré, and an anonymous reviewer for their helpful comments, which improved the quality of this manuscript considerably.
