Importance of nanoparticles and colloids from volcanic ash for riverine transport of trace elements to the ocean: Evidence from glacial-fed rivers after the 2010 eruption of Eyjafjallajökull Volcano, Iceland

Highlights

• Icelandic glacial-fed rivers show high dissolved REE and HFSE concentrations.

• Volcanic ash, filter residues and river waters show similar REE and HFSE patterns.

• Volcanic ash particles < 450 nm control dissolved REE and HFSE concentrations.

• River water composition confirms that volcanic ashes contain natural nanoparticles.

Abstract

Volcanic ashes are often referenced as examples for natural nanoparticles, yet the particle size distribution < 1000 nm is only rarely documented. We here report results of a geochemical study of glacial-fed rivers, glacial surface runoff, glacial base flow, and pure glacial meltwater from southern Iceland, that had been sampled 25 days after the explosive eruptions at Eyjafjallajökull in 2010. In addition to the dissolved concentrations of rare earth elements (REE), Zr, Hf, Nb, and Th in the 450 nm-filtered waters, we also studied the respective filter residues (river particulates > 450 nm) and volcanic ash. In spite of the low solubilities and high particle-reactivities of the elements studied, most water samples show high dissolved concentrations, such as up to 971 ng/kg of Ce and 501 ng/kg of Zr. Except for the pure glacial meltwater and glacial base flow, all waters display the same shale-normalized REE patterns with pronounced light and heavy REE depletion and positive Eu anomalies. While such patterns are unusual for river waters, they are similar to those of the respective river particulates and the volcanic ash, though at different concentration levels. The distribution of dissolved Zr, Hf, Nb, and Th in the waters also matches that of filter residues and ash. This strongly suggests that in all 450 nm-filtered river waters, the elements studied are associated with solid ash particles smaller than 450 nm. This reveals that volcanic ash-derived nanoparticles and colloids are present in these glacial-fed rivers and that such ultrafine particles control the trace element distribution in the surface runoff. Subsequent to explosive volcanic eruptions, these waters provide terrigenous input from landmasses to estuaries, that is characterized by a unique trace element signature and that subsequent to modification by estuarine processes delivers a pulse of nutrients to coastal seawater in regions not affected by plume fall-out.

Introduction

Volcanic ashes resulting from explosive volcanic eruptions are often referenced as typical examples for natural nanoparticles (e.g., Handy et al., 2008). However, in studies of such ashes the size distribution in the < 1000 nm range is only rarely reported (for few exceptions see, e.g., Ammann and Burtscher, 1990, Overbeck et al., 1982, Reich et al., 2009, Rietmeijer and Mackinnon, 1997), although the toxicological effects of incorporation of nanoparticles (< 100 nm; e.g., Nowack and Bucheli, 2007) and colloids (< 1000 nm; e.g., Buffle, 2006, Lead and Wilkinson, 2006) on organisms can be severe (Nowack and Bucheli, 2007). If a volcanic ash plume reaches the ocean and (mafic) ash particles are deposited in seawater they act as an important source of nutrients such as iron and phosphate, which may increase bioproductivity and trigger phytoplankton blooms (Frogner et al., 2001, Gíslason et al., 2002, Jones and Gíslason, 2008, Langmann et al., 2010, Lin et al., 2011, Lindenthal et al., 2013, Olgun et al., 2013). Besides nutrients, ash plumes also contribute the full spectrum of associated trace elements to seawater. However, in contrast to aeolian transport, rather little is known about the role of river systems and estuaries for the transport of such volcanic ash-derived material, although they might deliver nutrients to coastal seawater also in regions not directly affected by fall-out from volcanic ash plumes.

It is particularly unclear, whether certain trace elements, such as Hf (Bau and Koschinsky, 2006) that is used as a tracer of marine water masses (e.g. Frank, 2002), reach the ocean associated with solid particulates (> 450 nm), with colloids/nanoparticles (10 kD–450 nm) or as truly dissolved compounds (< 10 kD). In a simplified approach, the relative susceptibilities of trace elements to mobilization during water–rock interaction are often estimated based on their respective ionic potential. The high-field strength elements (HFSE) with their high ionic charge to ionic size ratios are, therefore, considered rather immobile. Moreover, as the HFSE tend to hydrolyze easily, they are particle-reactive and in natural waters typically associated with particulate matter. Hence, they are usually characterized by very low dissolved (“dissolved” = in 200 nm- or 450 nm-filtered waters) concentrations unless they are bound to nanoparticles and colloids that are small enough to pass such filter membranes. For Ce (one of the REE) and Zr, for example, the “world average” is as low as 262 ng/kg and 39 ng/kg, respectively (Gaillardet et al., 2003), although this average value includes data from trace element-rich tropical rivers.

Following major explosive volcanic eruptions and the subsequent deposition of air-fall tephra in a catchment, river waters typically carry a high particulate load (Hayes et al., 2002). If HFSE in such river waters are bound to nanoparticles and colloids in the volcanic ash, their dissolved concentrations can be expected to be rather high despite filtration through 450 nm filter membranes. They remain in suspension and are transported downstream until the river water reaches the ocean and mixes with seawater in an estuary. The actual impact of such waters on the composition of seawater will depend on the availability of complexing ligands during processes operating in the estuary (e.g., Bau et al., 2013). However, if the volcanic ash does not comprise nanoparticles and colloids and the HFSE are bound to larger particulates (> 450 nm) instead, their dissolved concentrations are low. They may go numerous times through the cycle of sediment deposition during low-discharge periods and re-suspension during high-discharge periods within the river system, before they eventually reach the estuary. In this latter case, riverine deliverance of ash-derived HFSE to estuaries would rather be related to river discharge and follow a seasonal pattern, while in the former case explosive volcanism in the catchment would result in a major HFSE pulse.

There are several studies of Icelandic rivers that investigated the behavior of riverine particulates and the release of elements to the dissolved phase (e.g., Eiriksdottir et al., 2008, Georg et al., 2007, Gíslason et al., 2002, Jones et al., 2012a, Jones et al., 2012b, Oelkers et al., 2011, Oelkers et al., 2012, Pearce et al., 2010, Pearce et al., 2013, Pogge von Strandmann et al., 2006, Pogge von Strandmann et al., 2008). Eiriksdottir et al. (2008) point out that the river fluxes of the dissolved and particulate fraction are controlled by lithology, climate, relief, tectonics, glacial cover, vegetation, and the age of the rocks. In the present contribution, we focus on the effect of volcanic eruptions. While other studies showed the release of easily soluble metal salts and nutrients from Icelandic ash following eruptions at Eyjafjallajökull in 2010 (Alfredsson et al., 2012) and Grímsvötn in 2011 (Olsson et al., 2013), we focus on the behavior of the particle-reactive HFSE.

In the morning of April 14th 2010, eruptions at Eyjafjallajökull Volcano produced a 9 to 10 km high ash plume of benmoreitic composition (Gudmundsson et al., 2012) the. This initial event is referred to as phase I and was followed by three more phases. Phase II started on April 18th with a significant drop in the intensity of explosive activity and a small and ash-poor plume. On May 5th, phase III produced large volumes of fine-grained tephra of trachytic composition with a plume height of up to 10 km. The final phase (phase IV) started on May 18th with a progressively decline in activity before eruptions ceased around midnight on May 22nd. The ash plumes from the sub-glacial eruptions not only affected the immediate vicinity of the volcano, but also led to the closure of large parts of European airspace, and resulted in unprecedented disruption of air traffic in wide parts of Europe (Gudmundsson et al., 2012). Although the size and the composition of individual particles originating from these eruptions were studied using transmission electron microscopy and other microanalytical techniques (e.g., Sigmarsson et al., 2011), even the thorough study by Gíslason et al. (2011) did not provide information on the particle size distribution of < 1800 nm.

We sampled river waters in the vicinity of Eyjafjallajökull in June of 2010, to evaluate the concentrations and distribution of REE and several other HFSE (Zr, Nb, Hf, Th) in the 450 nm-filtrates and in the respective filter residues. Comparison of the chemical composition of volcanic ash from Eyjafjallajökull suggests that the studied trace elements are bound to nanoparticles and colloids present in the river water and that these ultrafine particles are derived from the volcanic ashes of the 2010 volcanic eruption. These results confirm the notion that volcanic ash particles are examples of natural nanoparticles. They further indicate that explosive volcanic eruptions will eventually result in a pulse of HFSE deliverance to estuaries, which is rather independent from seasonal patterns and which may deliver enhanced amounts of nutrients to coastal seawater not affected by plume fall-out.