A Holocene tephra layer within coastal aeolian deposits north of Caleta Olivia (Santa Cruz Province, Argentina)

In this paper we illustrate the stratigraphy, geochronology, and geochemistry (major, minor, trace elements and Sr-isotopes) of a Holocene tephra layer found within coastal sedimentary deposits north of Caleta Olivia (Santa Cruz Province, Argentina). The stratigraphic succession comprises beach deposits with basal erosive surface resting on the local substrate (“Formación Patagonia”) followed by a poorly developed paleosoil. The paleosoil is covered by a lenticular finegrained (Mdφ: 5.2, 0.027 mm), well sorted (σφ: 1.2) volcanic ash layer and aeolian sands. The geochemical composition of shard fragments points to an origin from the Hudson volcano, located in the southern Andes, ca. 400 km to the west. The geochemistry, Sr-isotopes and the radiometric constraints (younger than the age of the underlying marine layer dated at ca. 4,100 a cal BP) further allow correlating this tephra with the so-called H2 eruption (ca. 3,900 a cal BP). This finding is of interest owing to the poor preservation potential of tephra within the Late Holocene sedimentary deposits of the Atlantic coast of Patagonia and represents the first finding of H2 eruption in this area, improving our knowledge of the dispersion of the fine-grained distal deposit of the Hudson volcanic explosive activity, thus allowing a better estimate of the eruptive dynamics and the risks associated with the Hudson volcano.


Introduction
The reconstruction of volcanic ashes dispersion related to past explosive volcanic eruptions represents a fundamental tool for correctly estimating volcanic eruptive dynamics, frequency and associated risks (e.g., Stern, 2008;Fontijn et al., 2014;Rawson et al., 2015;Stern et al., 2015;Sulpizio et al., 2014;Naranjo et al., 2017;Del Carlo et al., 2018). Moreover, the study of distal tephra layers is an invaluable tool for correlating different natural archives (e.g., Zanchetta et al., 2008Zanchetta et al., , 2018aBlockley et al., 2014) and solving different issues in Quaternary stratigraphy, archaeology and palaeoclimatology (e.g., Lowe, 2011;Lowe et al., 2012;Prieto et al., 2013;Zanchetta et al., 2018bZanchetta et al., , 2019. Southern South America is an area where studies on distal tephrostratigraphy have been systematically addressed for the number of potentially active volcanoes (e.g., Stern, 2004). The Hudson volcano is a good case, which exemplifies the potential volcanic risk of the region. On 12-14 August 1991 the Hudson volcano erupted (Naranjo, 1991;Naranjo et al., 1993) affecting the central and southernmost part of the Patagonia. The ash of the Plinian eruption covered an area of ca. 100,000 km 2 ( Fig. 1) heavily damaging local people, cattle and economy (Scasso et al., 1994). The arid and windy climatic condition of the Patagonia region and its strongly eroded soil owing to overgrazing allowed a continuous remobilization of the volcanic ashes for years, prolonging their negative impact on the environment and economy (Inbar et al., 1995). This clearly indicates that the impact of the volcanic explosive eruptions is exacerbated in marginal land experiencing a strong human pressure. Archeological investigations also suggest that a different Plinian eruption from the Hudson volcano, the so-called H 1 , occurred ca. 7,750 a cal BP (Stern, 1991(Stern, , 2008Naranjo and Stern, 1998), impacted hunter-gatherers populations of southern Patagonia and Tierra del Fuego (Prieto et al., 2013).
Stratigraphic works in southern Patagonia and Tierra del Fuego (e.g., Daga et al., 2010Daga et al., , 2014Stern, 1991Stern, , 2008Naranjo and Stern, 1998;Stern et al., 2011Stern et al., , 2015Stern et al., , 2016Fontijn et al., 2014;Rawson et al., 2015;Naranjo et al., 2017;Del Carlo et al., 2018) have shown that recurrent explosive volcanic activity mantled the landscape with different tephra thickness also far from the volcanic centers. This unequivocally indicates that distal tephra dispersion represents a real and immanent natural hazard for these marginal lands, which host population and important economic activities. However, most of the researches on recurrence of tephra layers are restricted to the Andean to peri-Andean region (e.g., Daga et al., 2010Daga et al., , 2014Stern et al., 2011Stern et al., , 2015Stern et al., , 2016Fontijn et al., 2014;Weller et al., 2015Weller et al., , 2019Watt et al., 2013) and studies along the Atlantic coast of Patagonia are particularly scarce (Del Carlo et al., 2018). This can be related to the specific environment of this region characterised by dry and windy conditions with a steppic vegetation, which renders preservation of volcanic ashes very occasional. Indeed, as observed by Naranjo et al. (2017) the accumulation and preservation of tephra layers are mainly a function of the volume of the dispersed pyroclasts and the presence of retainer agents such as the presence of mature forests that prevent aeolian reworking of the fallen ash and lapilli. Therefore, in the Atlantic coast of Patagonia the finding of tephra is exceptional but can be very useful for complementing more proximal data giving information on ash dispersion and for a better estimation of the total volume of an eruption.
In this study, we report for the first time the finding of a tephra layer from Hudson volcano within raised coastal Holocene sedimentary deposits of the Golfo San Jorge, north of Caleta Olivia (Fig. 1). This finding gives further data to improve the knowledge on the dispersion of the product of this volcano and for reconstructing a tephra "lattice" at a regional scale for Quaternary stratigraphic work.

The study area
The study area is located within the Golfo San Jorge basin between 44 and 47° S and 66° and 71° W, in central Patagonia, southern Argentina (Fig. 1). The Golfo San Jorge is an intracratonic basin placed in between two structural highs, namely the Somuncurá Massif (also known as Norpatagónico Massif) to the north and the Deseado Massif to the south (Ramos and Ghiglione, 2008). These positive elements represent the pre-Cretaceous basement of the basin. The infilling of the basin consists of marine to continental sedimentary successions spanning from Cretaceous to Quaternary (Sylwan, 2001). Most of the outcrops in the studied area are dominated by the neogenic marine deposits of the Formación Patagonia and Formación Santacruz (e.g., Feruglio, 1950;Figari et al., 2002). During the Late Pliocene and the Quaternary successions of fluvial  Scasso et al., 1994); Dispersion for H 2 eruption (after Naranjo and Stern, 1998); Dispersion for H 1 eruption (after Naranjo and Stern, 1998;Stern, 2008), Dispersion from H 0 (after Weller et al., 2014). Isopachs for H 2 , H 1 , H 0 represent 10 cm, for H 3 continous line 0.1 cm, dotted line 10 cm (after Scasso et al., 1994 andNaranjo andStern, 1998 to glacio-fluvial gravels and sands formed extensive covers organized at different altitudinal level and paleovalley (often known as Rodados Patagónicos or Rodados Tehuelches, e.g., see the extensive review of Panza, 2002). Since the Middle Pleistocene, in the coastal area, successions of marine terraces with prevailing gravelly deposits developed during the phases of interglacial marine highstand (Rostami et al., 2000). A system of Holocene raised beaches started to form since ca. 6,000 a cal BP organised by successions of beach ridges and littoral terraces discontinuously preserved along the coast (e.g., Bini et al., 2014Bini et al., , 2018Codignotto et al., 1988;Ribolini et al., 2011;Schellmann and Radtke, 2010;Isla and Bujalesky, 2008;Isola et al., 2011;Zanchetta et al., 2012aZanchetta et al., , 2014. The local climate is dominated by the southern Westerlies (Garreaud et al., 2009) and can be considered temperate arid (mesothermal). The mean annual precipitations are ca. 230 mm, mostly concentrated during the cold season. The average annual temperature is 12.7 °C, with the maximum in January (19.1 °C) and minimum in July (6.6 °C). The coast is dominated by a meso-to-macrotidal regime (Isla and Bujalesky, 2008).

Materials and Methods
The stratigraphic section was described and measured during a field survey carried out in February 2010. The ash samples were wrapped in a plastic bags and transported to the laboratory. Then they were dried at room temperature, observed and described under a binocular microscope. A fraction of the bulk samples was dried and embedded in epoxy resin, coated with graphite and screened for glass shards using a scanning electron microscope (SEM) at the Earth Sciences Department of the University of Pisa. Energy-dispersive spectrometry (EDS) of glass shards was performed using an EDAX-DX micro-analyzer mounted on a Philips SEM 515 (operating conditions: 20 kV acceleration voltage, 100 s live time counting, 200-500 nm beam diameter, 2100-2400 shots s -1 , ZAF correction). The ZAF correction procedure does not include natural or synthetic standards for reference and requires analysis normalization at a given value (chosen at 100%). Detailed discussion on the SEM-EDS performance, inter-calibration trials and standards can be found in Marianelli and Sbrana (1998), Cioni et al. (1997), Caron et al. (2012), Vogel et al. (2010, Zanchetta et al. (2012b). According to the intercalibration, analytical precision is 0.5% for oxides abundances higher than 15 wt%, 1% for abundances around 5 wt%, 5% for abundances of 1 wt%, and less than 20% for abundances close to the detection limit (around 0.5 wt%). During this work, we used two standards whose chemistry has been accurately controlled by XRF analyses and certified (Albite in-house standard and CFA47, Cioni et al., 1997;Marianelli and Sbrana, 1998) and routinely analyzed in the Pisa laboratory in the last 15 years (e.g., Zanchetta et al., 2018b).
To avoid alkali loss by diffusion a window spot ca. 10 μm wide was used. In some cases, due to the different shapes and sizes of the glass shards a smaller beam size was used. The reduction of the size defocuses the beam and this may affect the analytical data due to the complex combination of element mobility and sample damage (Hunt and Hill, 2001). However, this potential bias is minimized by analyzing several shards from the same sample (Zanchetta et al., 2012b). Table 1 shows the data.
The concentrations of Li, Be, Sc, V, Cr, Co, Ni, Cu, Zn, Ga, Rb, Sr, Y, Nb, Mo, Cs, Ba, rare earth elements (REEs), Ta, W, Pb, Th and U were determined on a bulk sample (WP542) by inductively coupled plasma-mass spectrometry (Perkin-Elmer NexION 300x) at the Dipartimento di Scienze della Terra (University of Pisa). The complete sample dissolution was obtained by repeated treatment with concentrated HF+HNO 3 (suprapure grade) on a hotplate. International geochemical reference rocks were used in calibration and quality check. The correction procedure includes (i) blank subtraction, (ii) instrumental drift correction using internal standardization (Rh, Re, Bi at 20 ng/ml), and (iii) oxide-hydroxide interference correction. The analytical precision is better than 5% RSD (Relative Standard Deviation) for all elements, except Sc, Cr, Ni, Cu, Zn (10-15% RSD). Table 2 shows the data. Sr isotope composition was determined by thermal ionization mass spectrometry at the Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano (INGV, Naples, Italy), using a ThermoFinnigan Triton TI multicollector mass spectrometer on a glass shardrich sample. 0.2 g of sample were leached several times with cold 6 N HCl for 10 min, and then rinsed in pure Milli Q® H 2 O. After dissolution, Sr was separated by conventional ion-exchange chromatographic techniques. The Sr blank was on the order of 0.1 ng during the period of chemistry processing. For characterizing the origin of the tephra layer grain-size analyses were carried out between 3.5 and 9 φ (90-2 µm) at 0.5 φ intervals, with φ=-log 2 d, where d is the diameter in mm, using a Beckman Coulter Multisizer 4 at the Dipartimento di Scienze della Terra e Geoambientali, University of Bari. The median size (Mdφ) and the sorting parameters (Inman, 1952) of each grain-size distribution were calculated using a Fortran 90 code, available into the PYFLOW package (Dioguardi and Dellino, 2014).
A sample of the mitilidae Nacella (Patinigera) deaurata, with joint valves collected below the tephra in the coastal marine layer was selected for radiocarbon dating. The sample was cleaned in an ultrasonic bath with the addition of oxygen peroxide and then gently etched with diluted HCl to remove any recent carbonate encrustation. Radiocarbon measurement was carried out at the CIRCE laboratory of Caserta, Italy (Terrasi et al., 2007(Terrasi et al., , 2008 and calibrated using the Marine13 curve (Reimer et al., 2015). A better calibration would be obtained using local knowledge of the reservoir effect, which depends on different radiocarbon content of marine waters respect to a simple equilibration with atmospheric 14 CO 2 . Indeed, different mixing rate between superficial waters with the tendency of rapid equilibration with atmospheric 14 CO 2 and deeper waters, more depleted in 14 CO 2 , for the Southern Atlantic Ocean and, in particular, Patagonia, are not well constrained (Schellmann and Radtke, 2010). For instance, Cordero et al. (2003) have reported reservoir effect from different localities of the Patagonian coast between ca. 42° S and 50° S ranging between ca. 180 and 530 years. Table 3 shows the radiocarbon dating.

Stratigraphy and lithology of the deposits
The stratigraphic succession comprises three main lithostratigraphic units (LU, Fig. 2A). The lowest unit (LU-1) correspond to Miocene marine deposits (Formación Patagonia) representing the local substrate. This unit is separated by a planar erosive surface (marine ravinement surface) from LU-2, which is formed by beach deposits. The LU-2 shows 30-50 cm of a basal shelly layer then becomes sandier showing clear low angle stratification and frequent mollusk fragment (Fig. 2B) typical of beach deposits. Shells are mostly fragmented and only a few specimens preserved joined valves. One of this shells yielded a radiocarbon age of 3,967-4,216 a cal BP (Table 3). The top of this unit (ca. 20-30 cm) is massive and shows evidence of subaerial bioturbation (root traces) corresponding to a poorly developed soil (Inceptisol) (Fig. 2A, B). A tephra layer, in form of metric lens, lies just above the top of the soil. Above this soil aeolian coarse sands develops embedding the upper part of the tephra layer (LU-3; Fig. 2A, B). The basal part of the tephra is whitish with no evident stratification (massive) and appears almost composed by glass shards (Fig. 3A). These have different size and shape comprising bubble walls and micropumices (Fig. 3A). The shards are almost aphiric and rare quartz and plagioclase are present as loose crystals or crystal fragments. In the upper part, changes in color and evident stratification indicate reworking and mixing with the lithic material of the aeolian sands. The grain-size of the sample taken from the basal part, probably the less contaminated by reworking, shows that it is well sorted and showing a unimodal distribution with a median size value of 5.2 φ (0.027 mm; Fig. 3B), suggesting a "primary" deposition. On top of the LU-3 there is an accumulation of marine shells, bone remains and flinstones (shellmidden) and then the present, poorly developed, soil (LU-4; Fig. 2A, B).
(*) Collected in a coastal marine sediment located inmediately below the studied tephra deposit. The radiocarbon age was calibrated using the Marine13 curve in Calib 7.10 (Reimer et al., 2015). Considering the geographical position, the most probable source of this tephra should be located in volcanoes pertinent to the southern edge of the South Volcanic Zone (SVZ) of the Andes as Mentolat, Macá, Cay and Hudson (Stern, 2004;Stern et al., 1984), where the Hudson volcano is one of the most active during the Holocene (Naranjo and Stern, 1998). There have been at least 12 Holocene to Recent explosive eruptions recorded from the Hudson volcano (Naranjo and Stern, 1998) and the major explosive events have been called H 0 , H 1 , H 2 , and H 3 (e.g., Naranjo and Stern, 1998;Weller et al., 2014). They occurred at ca. 17,400 a cal BP (H 0 , Weller et al., 2014), at ca. 7,750 a cal BP (H 1 ) (e.g., Prieto et al., 2013), at ca. 3,900 a cal BP (H 2 ) (Naranjo and Stern, 1998), and in AD 1991 (H 3 ) (Naranjo and Stern, 1998).
In the K 2 O-SiO 2 diagram (Fig. 4B) the glass shards chemical composition lie in the general evolution trend of the Hudson volcanic products and well distinguishable from other nearby volcano products (e.g., Stern et al., 2016). Some of the most used chemical diagram for discriminating volcanoes of the SVZ, like Zr-SiO 2 and Ba-Sr (e.g., Stern et al., 2016;Weller et al., 2014Weller et al., , 2019, shows that the glass shards product are all compatible with Hudson volcano (Fig. 4C, D). Data show in figure 4B, C, D are not perfectly coincident with the chemical composition for Hudson volcano defined in the literature. However, it is important to consider that most of these fields have been defined using bulk rock data, rather than the composition of volcanic distal dust-size shard particles. The attribution to the Hudson volcano is straightforwardly confirmed by 87 Sr/ 86 Sr ratio versus Rb (Naranjo and Stern, 1998; Table 1; Fig. 5), which shows consistently the values within the compositional range of the Hudson volcano.
Our tephra is chemically distinct from H 3 , which shows a more basic chemical bimodality (Fig. 4A, e.g., Kratzmann et al., 2009), and is decisively older, because the successions is topped by a shell midden, which regionally stopped to form with the arrival of European (Castro, 2008). However, stratigraphy and radiocarbon chronology indicate that our tephra is younger than ca. 4,100 a cal BP (Table 3) and cannot be correlated with H 0 and H 1 eruptions. There is also a significant difference between glass shards chemical composition of H 1 tephra unit reported, for instance, by Del Carlo et al. (2018) and our data, which are more felsic (Fig. 4A, also considering the use of EDS in this works, compared to WDS). Therefore, it is reasonable to correlate the Caleta Olivia tephra layer with H 2 eruption. The more evolved character of our micro-analyses on shards compared to the original data produced by Naranjo and Stern, 1998, is related to the fact that our chemical data have been obtained on single glass shards and not on bulk samples (Fig. 4A). Further support for the proposed correlation is the 87 Sr/ 86 Sr ratio versus the Sr content, which is much more in agreement with H 2 than FIG. 5. A. 87 Sr/ 86 Sr versus Rb diagram, field for different volcanoes after Naranjo and Stern (1998); B. 87 Sr/ 86 Sr versus Sr diagram for the main Hudson explosive eruptions (after Weller et al., 2014). Burney and Reclus are from the Austral Volcanic Zone (Stern, 2008). Red point from Caleta Olivia tephra. the other major Hudson eruptions (Fig. 5;Naranjo and Stern, 1998). Overall, all the major, minor and trace element data support the notion that the Caleta Olivia tephra layer can be confidently correlated with Hudson volcano and H 2 eruption is the best candidate, according to radiocarbon constraint and Sr-isotopes and trace element. The finding of the Caleta Olivia tephra layer on top of a poorly developed paleosol suggests that deposition was facilitated by the presence of vegetation associated to the soil formation. The following developing of eolian deposition probably produced significant reworking of the tephra and the lentinticular shape visible today. An accumulation of reworked tephra and lithic aeolian sand may have locally preserved the tephra layers from subsequent dispersion. The general appearance of the outcrops can suggest that only the lower part (ca. 5-6 cm) may have experienced little or no reworking, as suggested by the fresh aspect and sharp edges of shards and the sorting (Fig. 3A, B). Therefore, 5 cm would a minimum possible estimated thickenss for the tephra layer, a value that needs to be considered with extreme care. However, a major important point is that this finding gives a clue to the extent of the original dispersion for the H 2 Hudson eruption pyroclastic material (Fig. 1).
We note that Naranjo and Stern (1998), reported isopachs reconstruction for H 2 eruption which are also more consistent with a dispersion toward E, in agreement with our finding, in contrast to H 1 and H 3 which show a clearer dispersion toward SE (Fig. 1, see also Stern, 2008). The different dispersion may have been the expression of prevailing wind condition at time of the eruption, and/or changing in wind direction that may occur during the the paroxysmal Plinian eruption stages (see Kratzmann et al., 2010) or the occurrence of the eruption in a different season.
We note, that dispersion of the Late-glacial H 0 tephra, is direct decisively toward NE (Weller et al., 2014), but Zanchetta et al. (2018b) found glass shards mixed with aeolian deposit with chemical composition compatible with H 0 well south of the dispersion proposed by Weller et al. (2014). This indicates the high mobility of reworked glass shards in a windy climate with bare soils. In addition, we have to note that despite our intensive fieldwork in the area this was the only tephra identified within sedimentary coastal deposits. This is probably related to the poor capacity of preserving tephra layer in this environment. It is obvious that there is an underrepresentation of volcanic ash layers over this part of the Atlantic coastal area and this can probably underestimate our perception of volcanic risk based on the past distribution of tephra layers.

Conclusion
The finding of the Holocene tephra layer interbedded in the aeolian deposit north of Caleta Olivia represents an unique information on past volcanic ash dispersion over Atlantic coast of Patagonia. In particular chemical, Sr-isotopic and radiocarbon dating data allow correlating this layer with the H 2 eruption from Hudson volcano (Naranjo and Stern, 1998). This allows documenting the extent to the east of the fine ash facies of that eruption, giving interesting information on potential dispersion of future explosive eruptions. This, in addition, contributes to the reconstruction of tephra lattice for this part of the Patagonian region for which information on the presence of tephra layers in the Quaternary successions is still scarce.