element chemistry
Recently Published Documents
TOTAL DOCUMENTS
H-INDEX
Introduction of plumbole to f-element chemistry
f-element sandwich complexes bearing a η5-plumbole ligand are reported. Quantum chemical calculations suggest that this ligand retains its aromaticity upon coordination. The Er complexes show SMM behavior including magnetic hysteresis.
AbstractCorals nucleate and grow aragonite crystals, organizing them into intricate skeletal structures that ultimately build the world’s coral reefs. Crystallography and chemistry have profound influence on the material properties of these skeletal building blocks, yet gaps remain in our knowledge about coral aragonite on the atomic scale. Across a broad diversity of shallow-water and deep-sea scleractinian corals from vastly different environments, coral aragonites are remarkably similar to one another, confirming that corals exert control on the carbonate chemistry of the calcifying space relative to the surrounding seawater. Nuances in coral aragonite structures relate most closely to trace element chemistry and aragonite saturation state, suggesting the primary controls on aragonite structure are ionic strength and trace element chemistry, with growth rate playing a secondary role. We also show how coral aragonites are crystallographically indistinguishable from synthetic abiogenic aragonite analogs precipitated from seawater under conditions mimicking coral calcifying fluid. In contrast, coral aragonites are distinct from geologically formed aragonites, a synthetic aragonite precipitated from a freshwater solution, and mollusk aragonites. Crystallographic signatures have future applications in understanding the material properties of coral aragonite and predicting the persistence of coral reefs in a rapidly changing ocean.
<p>Mt Ngauruhoe is a 900 m high andesitic cone constructed over the last 2500 yr, and is the youngest cone of the Tongariro Massif. It was previously one of the most continuously active volcanoes in New Zealand, with ash eruptions having occurred every few years since written records for the volcano began in 1839. However, it has now been more than 30 yr since the last eruption. Eruptions in 1870, 1949, 1954 and 1974-1975 were accompanied by lava and block-and-ash flows. Detailed sampling of these historical lava and block-and-ash flows was conducted, including sampling from seven different lava flows erupted over the period June-September 1954 to investigate changes in magma geochemistry and crystal populations over short timescales, and to enable observed changes to be related back to known eruption dates. Mineral major and trace element chemistry highlights the importance of mixing between distinct basaltic and dacitic melts to generate the basaltic andesite whole rock compositions erupted. The basaltic end member can be identified from the presence of olivine crystals with Mg# 75-87, clinopyroxene cores with Mg# 82-92, and plagioclase cores of An80-90. The dacitic melt is identified by SiO2-rich clinopyroxene melt inclusions, clinopyroxene zoning with Mg# 68-76 and plagioclase rims of An60-70. Textural evidence from complex mineral zoning and large variability in the widths of reaction rims on olivine crystals suggests that mafic recharge of the more evolved system is frequent, and modelling of Fe-Mg inter-diffusion applied to the outermost rims of the clinopyroxene crystal population indicates that such recharge events have occurred weeks to months or even shorter prior to each of the historical eruptions, and thus likely trigger the eruptions.</p>
<p>Mt Ngauruhoe is a 900 m high andesitic cone constructed over the last 2500 yr, and is the youngest cone of the Tongariro Massif. It was previously one of the most continuously active volcanoes in New Zealand, with ash eruptions having occurred every few years since written records for the volcano began in 1839. However, it has now been more than 30 yr since the last eruption. Eruptions in 1870, 1949, 1954 and 1974-1975 were accompanied by lava and block-and-ash flows. Detailed sampling of these historical lava and block-and-ash flows was conducted, including sampling from seven different lava flows erupted over the period June-September 1954 to investigate changes in magma geochemistry and crystal populations over short timescales, and to enable observed changes to be related back to known eruption dates. Mineral major and trace element chemistry highlights the importance of mixing between distinct basaltic and dacitic melts to generate the basaltic andesite whole rock compositions erupted. The basaltic end member can be identified from the presence of olivine crystals with Mg# 75-87, clinopyroxene cores with Mg# 82-92, and plagioclase cores of An80-90. The dacitic melt is identified by SiO2-rich clinopyroxene melt inclusions, clinopyroxene zoning with Mg# 68-76 and plagioclase rims of An60-70. Textural evidence from complex mineral zoning and large variability in the widths of reaction rims on olivine crystals suggests that mafic recharge of the more evolved system is frequent, and modelling of Fe-Mg inter-diffusion applied to the outermost rims of the clinopyroxene crystal population indicates that such recharge events have occurred weeks to months or even shorter prior to each of the historical eruptions, and thus likely trigger the eruptions.</p>
<p>This thesis research focuses on clast and crystal-specific studies to investigate the pre- and syn-eruptive magmatic processes of two supereruptions in the TVZ: the 1.21 Ma Ongatiti (>500 km3) and the 1.0 Ma Kidnappers (~1200 km3), together with the smaller (~200 km3) 1.0 Ma Rocky Hill eruption from the Mangakino Volcanic Centre (MVC). Crystallisation histories determined through SIMS U-Pb dating of zircons reveal that the paired Kidnappers and Rocky Hill eruptions were products of a common magmatic system, which built over ~200 kyr, in the time break after the Ongatiti eruption. U-Pb age spectra from the Ongatiti show a protracted crystallisation history (over ~250 kyr), in which the majority of zircon crystallised ~100 kyr prior to eruption in a crystal mush. Zircons then ascended with melt during accumulation of the final erupted magma body in the shallow crust. Zircons remained stable in the melt dominant body but underwent little further crystallisation. Zircons from all three systems record common geochemical processes governed by the fractionating assemblage (predominantly plagioclase and amphibole). In particular, the MREE/HREE ratios and Sr concentrations of zircons from the Ongatiti record imply two contrasting source regions governed by different proportions of crystallising amphibole. The in-situ major and trace element chemistry of glass shards and crystals from the Kidnappers fall deposit reveal that magma within the Kidnappers was stored in three discrete bodies, which were systematically tapped during the early stages of eruption. Temperature and pressure (T-P) estimates from amphibole and Fe-Ti oxide equilibria from each magma type are similar and therefore the three magma bodies were adjacent, not vertically stacked, in the crust. Amphibole model T-P estimates range from 770 to 840 °C and 90 to 170 MPa corresponding to pre-eruptive storage depths of ~4.0-6.5 km. The systematic evacuation of the three independent magma bodies implies that there was tectonic triggering and linkage of eruptions. The termination of fall deposition and onset of the overlying ignimbrite emplacement marks the point of widespread caldera collapse and the catastrophic evacuation of a wider variety of melt during the Kidnappers eruption. Pumice compositions from the Kidnappers ignimbrite fall into three groups, two of which (KI-1 and KI-2) can be matched to bodies tapped during the fall phase of the eruption, with the addition of a further discrete batch of lower SiO2 (KI-3) magma. Core-rim textural and chemical variations in major crystal phases (plagioclase, amphibole and orthopyroxene) suggest each compositional group was sourced from a common mush but underwent a unique magmatic history during the development of melt-dominant bodies in the final stages prior to eruption. The field relationships and distinctive appearance of the Rocky Hill ignimbrite (~200 km3 DRE) and the underlying Kidnappers ignimbrite suggests that the two deposits are from distinct eruption events. However, major and trace element chemistry of matrix glass, coupled with the textural and chemical signatures of crystals suggests the magma erupted during the Rocky Hill was generated from the same source or mush zone as the Kidnappers. The two largest melt-dominant bodies (KI-1 and KI-2) within the Kidnappers were renewed, underwent mixing and incorporation of marginal material to form two magma types (RH-1 and RH-2) in the time break prior to the Rocky Hill eruption. Fe-Mg interdiffusion timescales in orthopyroxenes from the Kidnappers and Rocky Hill deposits suggest the establishment of the final melt-dominant bodies, through extraction of melt and crystals from a common mush, occurred within 1000 years, and peaked within centuries of each eruption. In addition, one discrete batch of Kidnappers melt has evidence for interaction with a lesser evolved melt within 50 yrs prior to eruption. This rejuvenation event was not the eruption trigger but may have primed the magma for eruption. The difference in timescales from common zones from both the Kidnappers and Rocky Hill orthopyroxene, recording the same processes reveal the time break between the two eruptions was ~20-40 years. This work highlights the rapidity of rejuvenation and renewal of the melt-dominant bodies within the Kidnappers/Rocky Hill magmatic system. The textural and in-situ compositional signatures of crystals from the Ongatiti ignimbrite imply the final erupted magma body was assembled from a thermally and chemically zoned mush, which extended to the base of the quartzofeldspathic crust (~15km). The mush was close to water saturation and was dominated by amphibole crystallisation. Melt and crystals (including the majority of zircons) were extracted from the mush and ascended to 4-6 km depths during the development of a crystal-rich (20-30%), but melt-dominant body. Significant crystallisation of plagioclase (and lesser proportions of orthopyroxene and amphibole) occurred in an event involving the gradual heating and/or increase of water in the rhyolite, from a broadly andesitic underplated magma. Homogeneous crystal rim and matrix glass compositions imply the final erupted volume of magma was effectively mixed through convection. Eu/Eu* values of whole-rock and matrix glass suggest little crystal-melt separation occurred in the melt-dominant magma body prior to eruption. This work has implications for understanding the generation, storage and eruption of large-scale silicic magma systems. The Ongatiti ignimbrite does not represent either an erupted mush, or a stratified magma chamber, suggesting an alternative model for the development of eruptible magma within large-scale silicic systems. The Kidnappers/Rocky Hill sequence records a complex interplay of multiple melt-dominant bodies, which were established and renewed on rapid timescales. The rapid timescales for the development of melt-dominant bodies and the systematic tapping of magmas in the Kidnappers/Rocky Hill system imply that tectonics may have had a strong external control on the eruptions at Mangakino.</p>
<p>This thesis research focuses on clast and crystal-specific studies to investigate the pre- and syn-eruptive magmatic processes of two supereruptions in the TVZ: the 1.21 Ma Ongatiti (>500 km3) and the 1.0 Ma Kidnappers (~1200 km3), together with the smaller (~200 km3) 1.0 Ma Rocky Hill eruption from the Mangakino Volcanic Centre (MVC). Crystallisation histories determined through SIMS U-Pb dating of zircons reveal that the paired Kidnappers and Rocky Hill eruptions were products of a common magmatic system, which built over ~200 kyr, in the time break after the Ongatiti eruption. U-Pb age spectra from the Ongatiti show a protracted crystallisation history (over ~250 kyr), in which the majority of zircon crystallised ~100 kyr prior to eruption in a crystal mush. Zircons then ascended with melt during accumulation of the final erupted magma body in the shallow crust. Zircons remained stable in the melt dominant body but underwent little further crystallisation. Zircons from all three systems record common geochemical processes governed by the fractionating assemblage (predominantly plagioclase and amphibole). In particular, the MREE/HREE ratios and Sr concentrations of zircons from the Ongatiti record imply two contrasting source regions governed by different proportions of crystallising amphibole. The in-situ major and trace element chemistry of glass shards and crystals from the Kidnappers fall deposit reveal that magma within the Kidnappers was stored in three discrete bodies, which were systematically tapped during the early stages of eruption. Temperature and pressure (T-P) estimates from amphibole and Fe-Ti oxide equilibria from each magma type are similar and therefore the three magma bodies were adjacent, not vertically stacked, in the crust. Amphibole model T-P estimates range from 770 to 840 °C and 90 to 170 MPa corresponding to pre-eruptive storage depths of ~4.0-6.5 km. The systematic evacuation of the three independent magma bodies implies that there was tectonic triggering and linkage of eruptions. The termination of fall deposition and onset of the overlying ignimbrite emplacement marks the point of widespread caldera collapse and the catastrophic evacuation of a wider variety of melt during the Kidnappers eruption. Pumice compositions from the Kidnappers ignimbrite fall into three groups, two of which (KI-1 and KI-2) can be matched to bodies tapped during the fall phase of the eruption, with the addition of a further discrete batch of lower SiO2 (KI-3) magma. Core-rim textural and chemical variations in major crystal phases (plagioclase, amphibole and orthopyroxene) suggest each compositional group was sourced from a common mush but underwent a unique magmatic history during the development of melt-dominant bodies in the final stages prior to eruption. The field relationships and distinctive appearance of the Rocky Hill ignimbrite (~200 km3 DRE) and the underlying Kidnappers ignimbrite suggests that the two deposits are from distinct eruption events. However, major and trace element chemistry of matrix glass, coupled with the textural and chemical signatures of crystals suggests the magma erupted during the Rocky Hill was generated from the same source or mush zone as the Kidnappers. The two largest melt-dominant bodies (KI-1 and KI-2) within the Kidnappers were renewed, underwent mixing and incorporation of marginal material to form two magma types (RH-1 and RH-2) in the time break prior to the Rocky Hill eruption. Fe-Mg interdiffusion timescales in orthopyroxenes from the Kidnappers and Rocky Hill deposits suggest the establishment of the final melt-dominant bodies, through extraction of melt and crystals from a common mush, occurred within 1000 years, and peaked within centuries of each eruption. In addition, one discrete batch of Kidnappers melt has evidence for interaction with a lesser evolved melt within 50 yrs prior to eruption. This rejuvenation event was not the eruption trigger but may have primed the magma for eruption. The difference in timescales from common zones from both the Kidnappers and Rocky Hill orthopyroxene, recording the same processes reveal the time break between the two eruptions was ~20-40 years. This work highlights the rapidity of rejuvenation and renewal of the melt-dominant bodies within the Kidnappers/Rocky Hill magmatic system. The textural and in-situ compositional signatures of crystals from the Ongatiti ignimbrite imply the final erupted magma body was assembled from a thermally and chemically zoned mush, which extended to the base of the quartzofeldspathic crust (~15km). The mush was close to water saturation and was dominated by amphibole crystallisation. Melt and crystals (including the majority of zircons) were extracted from the mush and ascended to 4-6 km depths during the development of a crystal-rich (20-30%), but melt-dominant body. Significant crystallisation of plagioclase (and lesser proportions of orthopyroxene and amphibole) occurred in an event involving the gradual heating and/or increase of water in the rhyolite, from a broadly andesitic underplated magma. Homogeneous crystal rim and matrix glass compositions imply the final erupted volume of magma was effectively mixed through convection. Eu/Eu* values of whole-rock and matrix glass suggest little crystal-melt separation occurred in the melt-dominant magma body prior to eruption. This work has implications for understanding the generation, storage and eruption of large-scale silicic magma systems. The Ongatiti ignimbrite does not represent either an erupted mush, or a stratified magma chamber, suggesting an alternative model for the development of eruptible magma within large-scale silicic systems. The Kidnappers/Rocky Hill sequence records a complex interplay of multiple melt-dominant bodies, which were established and renewed on rapid timescales. The rapid timescales for the development of melt-dominant bodies and the systematic tapping of magmas in the Kidnappers/Rocky Hill system imply that tectonics may have had a strong external control on the eruptions at Mangakino.</p>
<p>Mount Tauhara is the largest dacitic volcanic complex of onshore New Zealand and comprises seven subaerial domes and associated lava and pyroclastic flows, with a total exposed volume of ca. 1 km3. The dacites have a complex petrography including quartz, plagioclase, amphibole, orthopyroxene, clinopyroxene, olivine and Fe‐Ti oxides and offer an excellent opportunity to investigate the processes and timescales involved in assembling dacitic magma bodies in a continental subduction zone with in situ and mineral specific analytical techniques. Whole rock major and trace element data and Pb isotopes ratios define linear relationships indicating that the dacites are generated by mixing of silicic and mafic magmas. Two groups of samples define separate mixing trends between four endmembers on the basis of La/Yb ratios, 87Sr/86Sr ratios and Sr contents. The older Western and Central Domes have low 87Sr/86Sr (0.7042‐0.7046) and high LREE/HREE (LaN/YbN = 8.0‐11.5) and Sr (380‐650 ppm) compared to the younger Hipaua, Trig M, Breached and Main Domes, which have higher 87Sr/86Sr (0.7047‐0.7052) and lower LREE/HREE (LaN/YbN = 6.5‐7.5) and Sr (180‐400 ppm). In situ mineral major and trace element chemistry of mineral phases, as well as Sr and Pb isotope ratios of mineral separates have been used to: (i) fingerprint the origin of each crystal phase; (ii) constrain the chemistry of the four endmembers involved in the mixing events and; (iii) estimate the timing of mixing relative to eruption and the ascent rate of the dacitic magmas. The presence of quartz and analyses of quartz‐hosted melt inclusions are used to fingerprint the chemistry of the silicic endmembers, which is a rhyolitic melt with a major element chemistry similar to that of either the Whakamaru Group Ignimbrite melts (Western, Central and Trig M Domes) or intermediate between that of the Whakamaru and the Oruanui Ignimbrite melts (Hipaua, Breached and Main Domes). Similarly, Ba‐Sr concentrations and Sr isotopic signatures of plagioclase show that this phenocryst phase also predominantly crystallized from the rhyolitic melt. Variations in the Mg# and trace element chemistry of clinopyroxenes suggest they were formed both in the mixed dacitic melts and in a mafic endmember. The chemistry of the mafic endmembers have been traced using a combination of back‐calculated Sr melt concentrations from clinopyroxene with the highest Mg# in each sample group, and the linear trends between whole rock SiO2 content and most elements. These results indicate that dacites erupted from the Western and Central Dome were generated by the mixing of a high alumina basalt and a rhyolitic melt and Trig M Dome dacites were generated by the mixing of an andesite with a rhyolitic melt. Magmas erupted from Hipaua, Breached and Main Domes were also produced by the mixing of an andesitic melt and a rhyolitic body with a composition intermediate between that of the Whakamaru and the Oruanui melt bodies. Trace element data and 87Sr/86Sr ratios of amphibole demonstrate that it crystallized from the mixed dacitic melt. Thermobarometric conditions obtained from amphibole indicate that the magma mixing event that produced the dacites occurred within a magma chamber located at ca. 9 km depth and ca. 900°C with the exception of Trig M Dome which occurred deeper at 13 km and 950°C. Diffusion profiles of Ti in quartz and Fe‐Mg in clinopyroxene indicate the magma mixing events occurred < 6 months prior to eruption. Amphibole reaction rims show the magma to have ascended over 2‐3 weeks for each dome, with the exception of Main Dome where reaction rims were not present in the amphibole, suggesting the ascent rate was faster than 0.2 m/s (< 6 hours).</p>
<p>Mount Tauhara is the largest dacitic volcanic complex of onshore New Zealand and comprises seven subaerial domes and associated lava and pyroclastic flows, with a total exposed volume of ca. 1 km3. The dacites have a complex petrography including quartz, plagioclase, amphibole, orthopyroxene, clinopyroxene, olivine and Fe‐Ti oxides and offer an excellent opportunity to investigate the processes and timescales involved in assembling dacitic magma bodies in a continental subduction zone with in situ and mineral specific analytical techniques. Whole rock major and trace element data and Pb isotopes ratios define linear relationships indicating that the dacites are generated by mixing of silicic and mafic magmas. Two groups of samples define separate mixing trends between four endmembers on the basis of La/Yb ratios, 87Sr/86Sr ratios and Sr contents. The older Western and Central Domes have low 87Sr/86Sr (0.7042‐0.7046) and high LREE/HREE (LaN/YbN = 8.0‐11.5) and Sr (380‐650 ppm) compared to the younger Hipaua, Trig M, Breached and Main Domes, which have higher 87Sr/86Sr (0.7047‐0.7052) and lower LREE/HREE (LaN/YbN = 6.5‐7.5) and Sr (180‐400 ppm). In situ mineral major and trace element chemistry of mineral phases, as well as Sr and Pb isotope ratios of mineral separates have been used to: (i) fingerprint the origin of each crystal phase; (ii) constrain the chemistry of the four endmembers involved in the mixing events and; (iii) estimate the timing of mixing relative to eruption and the ascent rate of the dacitic magmas. The presence of quartz and analyses of quartz‐hosted melt inclusions are used to fingerprint the chemistry of the silicic endmembers, which is a rhyolitic melt with a major element chemistry similar to that of either the Whakamaru Group Ignimbrite melts (Western, Central and Trig M Domes) or intermediate between that of the Whakamaru and the Oruanui Ignimbrite melts (Hipaua, Breached and Main Domes). Similarly, Ba‐Sr concentrations and Sr isotopic signatures of plagioclase show that this phenocryst phase also predominantly crystallized from the rhyolitic melt. Variations in the Mg# and trace element chemistry of clinopyroxenes suggest they were formed both in the mixed dacitic melts and in a mafic endmember. The chemistry of the mafic endmembers have been traced using a combination of back‐calculated Sr melt concentrations from clinopyroxene with the highest Mg# in each sample group, and the linear trends between whole rock SiO2 content and most elements. These results indicate that dacites erupted from the Western and Central Dome were generated by the mixing of a high alumina basalt and a rhyolitic melt and Trig M Dome dacites were generated by the mixing of an andesite with a rhyolitic melt. Magmas erupted from Hipaua, Breached and Main Domes were also produced by the mixing of an andesitic melt and a rhyolitic body with a composition intermediate between that of the Whakamaru and the Oruanui melt bodies. Trace element data and 87Sr/86Sr ratios of amphibole demonstrate that it crystallized from the mixed dacitic melt. Thermobarometric conditions obtained from amphibole indicate that the magma mixing event that produced the dacites occurred within a magma chamber located at ca. 9 km depth and ca. 900°C with the exception of Trig M Dome which occurred deeper at 13 km and 950°C. Diffusion profiles of Ti in quartz and Fe‐Mg in clinopyroxene indicate the magma mixing events occurred < 6 months prior to eruption. Amphibole reaction rims show the magma to have ascended over 2‐3 weeks for each dome, with the exception of Main Dome where reaction rims were not present in the amphibole, suggesting the ascent rate was faster than 0.2 m/s (< 6 hours).</p>
<p>This thesis presents a sub-seasonally resolved, decade long record of snow pack chemistry from Evans Piedmont Glacier (EPG), southern Victoria Land coast, Antarctica. Snow chemistry measurements were made at ca. 20 analyses per year for stable isotope ratios [delta to the power of 18]O and [delta]D, major ions Ca+, Cl-, K+, Mg+, MS-, Na+, NO3-, SO42- by ion chromatography (IC), and major and trace element chemistry by inductively coupled plasma mass spectrometry (ICP-MS). Na, Mg, Al, Fe, Mn and Ba were measured by ICP-MS using a hydrogen flushed collision cell to reduce the formation of polyatomic ion interferences, whereas Ti, V, Cr, Ni, Cu, Zn, As, Rb, Sr, Y, Zr, Sb, Cs, Ba, La, Ce, Pb, Bi, Th and U were measured in non-collision cell mode to increase count sensitivity. ICP-MS analytical precision is typically 5 to 10 % (2 rsd) that is two orders of magnitude at minimum below natural variability (e.g. samples range between Na = 10 to 18031 ppb and Al = 5 to 3856 ppb). The presence of undigested mineral dusts in weakly acidified samples, however, complicates the measurement of elemental concentrations in snow samples by randomly entering the ICP-MS. Despite this, the range of sample concentrations (Zr = 3.0 to 5630 ppb) is still orders of magnitude higher than sample reproducibility. The dominant source regions of element chemistry transported to EPG snow are identified as marine (Na, Mg, SO4, Cl, K, As and Sr) and terrestrial derived aerosol (Al, Mn, Fe, Ba, Ti, V, Ni, Cr, Zn, Rb, Y, Zr, Cd, Sb, Cs, Ba, La, Ce, Pb, Th and U), with minor contributions from anthropogenic (V, Cr, Ni, Cu, Zn, As, Sb and Pb) and volcanic emissions (Bi, SO4 and K). This is based on both elemental ratio modelling and ICP-MS time resolved analysis that identifies elements present in particulate form (mineral dusts). EPG snow chemistry is related to measured meteorological conditions at nearby Cape Ross. Winter maxima of elemental concentrations is consistent with maximum winter wind speed and low precipitation rates. Furthermore, winter snow samples that are depleted in SO42- relative to other marine derived elements (e.g. Na), indicate the sea ice surface is an important source of marine aerosol transported to EPG in addition to an open ocean source. Annual maximum chemistry concentrations of terrestrial derived elements (e.g. Zr) are significantly correlated to maximum annual wind speed measured at Cape Ross (r2 = 0.68, p< 0.01). Lower correlation of marine derived chemistry (e.g. Na) and maximum wind strength reflects additional controls of source region and other meteorological parameters such as storm duration on marine derived chemistry. In contrast to elemental concentrations, elemental ratios are less sensitive to extreme wind conditions. Rather elemental ratios provide a more robust signature of changes in mean atmospheric circulation related to delivery of aerosol from different source regions and via different transport fractionation processes. Al/Na is controlled by variable delivery of terrestrial (Al) and marine (Na) aerosol to EPG, although the longer term trend is driven primarily by changes in Na. Al/Na is significantly higher between winter 2000 and summer 2006/07 with a mean value of Al/Na = 0.15 compared to Al/Na = 0.02 prior to 2000. Although sea ice extent was highly variable over this time period, there is no clear relationship between Al/Na and sea ice. Rather, Al/Na is significantly correlated to mean summer wind speed measured at Cape Ross (r2 = -0.51, p<0.01). This demonstrates the sensitivity of Al/Na to changes in the average transport of marine aerosol to EPG during summer, when an open ocean source is most proximal. The shift in Al/Na is also concurrent with a shift in the relationship between [delta]18O and d excess, indicative of a changing precipitation source region to EPG. The observed changes in EPG chemistry are concurrent with shifts in mean Southern Oscillation Index (SOI), a measure of the El Nino Southern Oscillation (ENSO) strength and polarity. Al/Na is low when SOI is predominantly negative (El Nino), associated with increased summer wind strength. This is in accordance with a strong Amundsen Sea Low, positioned directly north of the Ross Sea as previously reported during El Nino years. Although the establishment of a statistically significant relationship between SOI and EPG Al/Na ratios is inhibited by the brevity of this record, this study highlights the potential for the 180 m firn core also extracted from EPG to track long-term changes in SOI. Elemental chemistry of EPG provides a high resolution tool to reconstruct atmospheric circulation changes within the southern Ross Sea region.</p>
