Atmospheric mercury: Long-term (Late Pleistocene – Holocene) variations in mercury accumulation reconstructed using environmental archives

PhD THESIS - TESIS DOCTORAL MODALIDAD DE COMPENDIO DE ARTÍCULOS Marta Pérez Rodríguez PROGRAMA DE DOUTORAMENTO DE MEDIO AMBIENTE E RECURSOS NATURAIS FACULTADE DE BIOLOXÍA SANTIAGO DE COMPOSTELA 2017 ATMOSPHERIC MERCURY: LONG-TERM (LATE PLEISTOCENE - HOLOCENE) VARIATIONS IN MERCURY ACCUMULATION RECONSTRUCTED USING ENVIRONMENTAL ARCHIVES Don Antonio Martínez Cortizas, Catedrático de Universidade do Departamento de Edafoloxía e Química Agrícola da Facultade de Bioloxía da Universidade de Santiago de Compostela, DECLARA que ésta tese de doutoramento, titulada “Long-term (Late Pleistocene – Holocene) variations in mercury accumulation reconstructed using environmental archives”, foi realizada por Dona Marta Pérez Rodríguez baixo a miña supervisión e é idónea para ser presentada como Tese por compendio de Artigos. A doutoranda participou de xeito activo na investigación reflexeda nos artigos e a súa contribución foi decisiva para levar a cabo os traballos. Todos os coautores participantes nos artigos, tanto doutores como non doutores, están en coñecemento e aceptaron a inclusión das publicacións nesta Tese de Doutoramento. Ningún dos traballos foi ou será presentado por ningún dos coautores noutra tese de Doutoramento. Santiago de Compostela, a 10 de Maio de 2017 Don Antonio Martínez Cortizas Abstract The principal aim of this PhD research is to gain further insights into the variations in atmospheric Hg deposition over long timescales (late Pleistocene and Holocene) in the northern and southern hemispheres, by the use of environmental archives (peat and lake sediments). For this purpose, five archives from different areas of the world and with records that operate at different time scales were studied: Rano Aroi (Easter Island, Chile), Pinheiros (Estado de Minas Gerais, Brazil), Lago Hambre (South Patagonia, Chile), Limnopolar Lake (South Shetland Islands, Antarctica) and Sandhavn (South Greenland). These records have been studied previously by geochemical, paleoclimatic and palynological means, enabling the multi-proxy approach that is needed for the assessment of a subject with the complexity of Hg cycling. This work pays special attention to understanding the processes involved in Hg deposition and accumulation, the interactions between these processes and the spatial and temporal variability in their relative importance. Furthermore, the research included a methodological assessment that aimed to select a suit of relatively costand time-efficient methodologies, which is necessary for multicore/multi-site studies. Eight factors were identified as the main drivers of Hg concentration - through deposition-uptake and also through accumulation - in the peat and lake sediments studied: anthropogenic pollution, volcanic activity, organic matter decomposition, catchment processes, (lake) primary productivity, atmospheric Hg depletion events, vegetation type and direct and indirect climatic effects. Essentially, lake sediments and peatlands are environmental archives that can be used as records of atmospherically deposited Hg, thus providing a “picture” of the Hg cycle. However, the use of such archives as Hg atmospheric records is conditioned by factors that operate in different geographical and temporal scales. Thus, the Hg determined in the samples does not directly reflect atmospheric deposition, but results of changing combination different factors. Keywords: mercury cycle, peatlands, lake sediments, paleoenvironment, atmospheric deposition Resumen El principal objetivo de esta tesis doctoral es obtener información sobre las variaciones de la deposición atmosférica de mercurio a escalas largas de tiempo (Pleistoceno tardío y Holoceno) en los hemisferios Norte y Sur, usando archivos ambientales (turberas y sedimentos lacustres). Con este propósito, se estudiaron cinco archivos de diferentes partes del mundo y de diferentes escalas de tiempo: Rano Aroi (Isla de Pascua, Chile), Pinheiros (Estado de Minas Gerais, Brasil), Lago Hambre (Sur de la Patagonia, Chile), Lago Limnopolar (Islas Shetland del Sur, Antártida) y Sandhavn (Sur de Groenlandia). Estos registros han sido objeto de estudio en diferentes ámbitos como geoquímica, paleoclimatología, evolución de turberas y palinología, lo que ha permitido el enfoque multi-indicador necesario para la evaluación de un problema con la complejidad del ciclo del Hg. Se presta especial atención a entender cómo los procesos relacionados con la deposición y acumulación de Hg en los archivos ambientales, interaccionan y cambian su peso relativo en el tiempo y el espacio. Se incluye un aspecto metodológico para la aplicación de técnicas más baratas y de más rápidas para ayudar en estudios más amplios. Se identificaron ocho factores como los principales impulsores de la concentración de Hg – a través de la deposición-absorción y la acumulaciónen los registros estudiados: contaminación, actividad volcánica, descomposición de materia orgánica, procesos de captación, productividad primaria, eventos de agotamiento atmosférico, el tipo de vegetación y los efectos climáticos. Los sedimentos lacustres y las turberas son archivos ambientales que pueden utilizarse como registros de Hg depositado atmosféricamente, ya que proporcionan una “imagen” del ciclo de Hg pasado. Pero este uso está condicionado por factores que operan en diferentes escalas geográficas y temporales. Por lo tanto, el Hg determinado en las muestras no refleja directamente la deposición atmosférica, es el resultado de la combinación de diferentes factores. Palabras clave: ciclo del mercurio, turberas, sedimentos lacustres, paleoambiente, deposición atmosférica. Resumo O principal obxectivo desta tese doctoral é obter información sobre as variacións da deposición atmosférica de mercurio a escalas longas de tempo (Pleistoceno tardío e Holoceno) nos hemisferios Norte e Sur, usando arquivos ambientais (turberas e sedimentos lacustres). Con este propósito, estudáronse cinco arquivos de diferentes partes do mundo e de diferentes escalas de tempo: Rano Aroi (Illa de Pascua, Chile), Pinheiros (Estado de Minas Gerais, Brasil), Lago Hambre (Sur da Patagonia, Chile), Lago Limnopolar (Illas Shetland do Sur, Antártida) e Sandhavn (Sur de Groenlandia). Estes registros foron obxecto de estudo en diferentes ámbitos como xeoquímica, paleoclimatología, evolución de turberas e palinología, o que permitiu o enfoque multi-indicador necesario para a avaliación dun problema coa complejidad do ciclo do Hg. Préstase especial atención a entender como os procesos relacionados coa deposición e acumulación de Hg nos arquivos ambientais, interaccionan e cambian o seu peso relativo no tempo e o espazo. Inclúese un aspecto metodológico para a aplicación de técnicas máis baratas e de máis rápidas para axudar en estudos máis amplos. Identificáronse oito factores como os principais impulsores da concentración de Hg - a través da deposición-absorción e a acumulaciónnos registros estudados: contaminación, actividade volcánica, descomposición de materia orgánica, procesos de captación, productividad primaria, eventos de agotamiento atmosférico, o tipo de vegetación e os efectos climáticos. Os sedimentos lacustres e as turberas son arquivos ambientais que poden utilizarse como registros de Hg depositado atmosféricamente, e proporcionan unha “imaxe” do ciclo do Hg pasado. Pero este uso está condicionado por factores que operan en diferentes escalas xeográficas e temporais. Polo tanto, o Hg determinado nas mostras non reflicte directamente a deposición atmosférica, si non que é o resultado da combinación de diferentes factores. Palabras chave: ciclo do mercurio, turberas, sedimentos lacustres, paleoambiente, deposición atmosférica. 16 source. Recycling from surface reservoirs results in an effective lifetime of 1.6 years before transfer to soil and deep ocean pools. Due to its volatility as Hg0 and the redox transformations of Hg in the environment, deposition and release are influenced by multiple environmental factors including seasonality, vegetation cover, temperature, light, moisture, atmospheric turbulence and the presence of reactants. Several factors control Hg dynamics in different environments such as ice dynamics, ocean surface temperature, soil carbon content, riverine discharge, type of precipitation [rain vs. snow], and methylation rates. In addition, human activity has played a key role in increasing Hg abundance in the environment. A combination of natural and anthropogenic factors ensures that the biogeochemical cycle of Hg is particularly susceptible to global change. Thus, the use of environmental archives is essential to Hg research, as they provide a unique link between current and past Hg dynamics in the environment. Environmental or natural archives are those physical (natural) structures that due to their formation process and evolution can store proxies of environmental change. Such archives, such as peatlands, ice cores and lake sediments, have had an essential role in the reconstruction of atmospheric deposition of Hg on a local, regional, and global scale. Peatlands and lake sediments have a wide geographic distribution (peatlands cover more than 4 million km2 worldwide) and have proved to provide accurate Hg deposition records. Pollution is considered one of the main drivers behind Hg concentration in records worldwide that cover the last centuries or last millennia – mainly by the effect of Industrial or the Pre-Industrial activities, respectively. However, contributions from natural sources have also been recorded. Large differences in the estimated background Hg fluxes exist between different records, hampering our understanding of the emission sources, the atmospheric processes involved and the rates of exchange between terrestrial, oceanic and atmospheric pools. These differences in fluxes could be due to a variety factors that could affect Hg accumulation in lake sediments and peatlands, complicating their use as archives of atmospheric deposition. In lakes, these factors include the size and land-use of the watershed, Hg mobility in sediments, diagenesis, primary productivity and algal scavenging. In peatlands, on the other hand, the spatial variability in peat surface, vegetation type and decomposition rate are still 1. Summary 17 poorly understood which limits the use of Hg records from peat archives for the reconstruction of atmospheric deposition of Hg. In spite of several advances that have been made during the last decade to examine the underlying processes that occur in archives of Hg deposition, the number of reliable records is still very limited. Furthermore, most of the studies are from the Northern Hemisphere (mainly Europe and North America) and spanning the Holocene (last 11,000 years). However, the few available long-term studies showed that non-anthropogenic factors such as changes in climatic conditions or decomposition rate, can have strong influence in records of Hg accumulation. This implies that also in Holocene records an effort needs to be made to disentangle the anthropogenic and natural controls on Hg concentration. The principal aim of this PhD research is to gain further insights into the variations in atmospheric Hg deposition over long timescales (late Pleistocene and Holocene) in the northern and southern hemispheres, by the use of environmental archives (peat and lake sediments). For this purpose, five archives from different areas of the world and with records that operate at different time scales were studied: Rano Aroi (Easter Island, Chile), Pinheiros (Estado de Minas Gerais, Brazil), Lago Hambre (South Patagonia, Chile), Limnopolar Lake (South Shetland Islands, Antarctica) and Sandhavn (South Greenland). These records have been studied previously by geochemical, paleoclimatic and palynological means, enabling the multi-proxy approach that is needed for the assessment of a subject with the complexity of Hg cycling. This work pays special attention to understanding the processes involved in Hg deposition and accumulation, the interactions between these processes and the spatial and temporal variability in their relative importance. Furthermore, the research included a methodological assessment that aimed to select a suit of relatively costand time-efficient methodologies, which is necessary for multi-core/multi-site studies. There is a wide variety of concentrations and rates of Hg accumulation in the studied records. The enrichment values in some cases reached values of up to 800 times the minimum value, suggest the combination of different factors (natural and due to human activity) that have interacted and possibly changed over time. Eight factors were identified as the main drivers of Hg concentration in the peat and lake sediments studied here: anthropogenic pollution, volcanic activity, organic matter decomposition, catchment processes, (lake) primary 18 productivity, atmospheric Hg depletion events, vegetation type and direct and indirect climatic effects. These factors controlled Hg concentrations in the studied records through deposition-uptake and also through accumulation. As previously mentioned, pollution has been considered one of the main drivers controlling Hg concentrations in records that cover the last centuries or millennia – mainly showing the effect of industrial or pre-industrial activities, respectively. Previous studies from areas in proximity of the ones that are the subject of this thesis have pointed towards the importance of anthropogenic Hg enrichment, however only the Greenland peat record (Sandhavn, spanning the last ~700 cal. yr) provided a clear signal of Hg pollution. Several reasons may explain the lack of evidence of recent atmospheric pollution in most of the studied cores, such as the relatively low resolution for the last centuries in the Pleistocene records (Rano Aroi and Pinheiros) or the overlapping effect of other processes driving Hg concentrations (Limnopolar Lake and Lago Hambre). At Sandhavn, the results demonstrated that the Hg accumulation rate has steadily increased since the beginning of the 19th century, with maximum values of 9.3 μg m-2 yr-1 recorded ~1940, which is in agreement with previous records from Greenland. The accumulation rates are generally low in comparison with other records from Europe and North America probably because i) Sandhavn is located at a long distance from the major sources of pollution or ii) mid-latitudinal records have been affected more strongly by Hg enrichment due to the more intense peat mineralization. The detailed chronology of metal pollution provided by the detrending of the Pb isotopic ratio record (extracting the effect of geogenical mixing), indicated that widespread pollution started AD ~1740-1780, which is in agreement with previous studies using Pb that found that modern metal pollution in Greenland had begun prior to the 19th century; maximum pollution signatures occurred AD ~1960-1970. The isotopic composition of the Sandhavn peat record since the 19th century and the timing of Pb enrichment clearly points to the dominance of pollution sources from the USA. There is no direct evidence for the origin of Hg, but according to the Pb isotope results we would expect the main source of Hg contamination also to be North America. However, Hg has a complex behavior in the environment. Its relatively long residence time in the atmosphere (>1 year) favors long-range transport and homogenization at a hemispheric scale, making it 1. Summary 19 more difficult to determine its precise origin. The volcanic emissions produce a natural mobilization of Hg from geogenic reservoirs. However, there are regional differences in average Hg emissions. Unlike ice cores, peat and lake sediments are usually collected from lower altitudes, which often implies that enrichments in Hg due to volcanic eruptions are superposed on stronger variability in Hg influx by other processes that the relatively “clean” high altitude ice records. The Deception Island volcano significantly influenced the geochemistry of the Limnopolar Lake that is located at a distance of 30 km, mainly by increasing the volcanic vs. catchment inorganic fluxes during several episodes of the last 1600 yr. The chronology of volcanic eruptions of Deception Island and excursions of Hg concentration in Limnopolar Lake (Hg accumulation rates between 300 to 4900 m-2 yr-1 against a background ~9 µg m-2 yr-1) and recent tephra records are similar. Nevertheless, Hg concentration and volcanic activity seem to be decoupled from AD ~1800 onwards, which is not in fase of high volcanic activity in Deception Island. Therefore, the nearby volcano appears to be the main source of Hg in Limnopolar Lake, but other mechanisms have been involved. The effect of volcanic eruptions on the Hg concentrations in the Rano Aroi peatland (Easter Island) is less evident. In fact, there is no evidence, in the form of tephra layers, of recent volcanic activity in the peat cores. On a volcanic island, outgassing by active fumaroles, which does not produce tephra, could act as the primary Hg source and may be related to some of the very high peaks in Hg concentration detected (> 1000 ng g-1). Either way, other processes such as deposition liked to rainfall events probably have a stronger control on Hg dynamics at Rano Aroi. It has been argued that organic matter decomposition has a strong influence on Hg accumulation in peatlands, which may impose limitations to the use of peat bogs as reliable Hg atmospheric records. Nevertheless, the available literature is almost exclusively based on studies of Holocene peat deposits. Using the C/N ratio as an indicator of the degree of peat decomposition, the longer-term peat records presented here (Rano Aroi and Pinheiros) have a limited effect of peat decomposition on Hg concentrations. In these records, the main decay processes had completed by ca. 10-11 cal kyr BP in Pinheiros and 5 cal kyr BP in Rano Aroi. Only in Pinheiros the decay of organic matter in the youngest sections has affected 20 Hg content. In Rano Aroi, only the clear increase in Hg concentration at ~40 cal kyr BP may have responded to an extreme oxidation event driven by a drought period. The results indicated a two-fold increase in Hg concentrations. Both direct aerial deposition of Hg on the lake surface and remobilization of atmospheric inputs through fluxes from the catchment are to be considered in Hg reconstruction in lacustrine environments. A huge amount of pollutants have been deposited and stored in the catchment surface soils of many lake sites, mainly due to the industrial emissions during the last centuries. Thus, processes such as watershed erosion and other catchment processes may have a strong influence on Hg concentration in lake sediments or peatlands receiving water inputs from sources other than precipitation (usually as runoff from upslope area) such as Pinheiros and Rano Aroi. In both records, catchment erosion has influenced Hg concentration but in an opposite way. In Pinheiros the mineral matter erosion from the quartzitic catchment produced a dilution effect: on the organic matter to which Hg is bound and by inputs of low Hg-containing mineral matter. In Rano Aroi, the catchment erosion mobilized atmospherically deposited Hg and associated and retained to organometallic complexes (mostly with Fe) from the volcanic soils. Thus, superficial runoff driven by precipitation may transport Hg to the mire and amplify the effect of wet deposition during rainfall events. Similarly to Rano Aroi, in Limnopolar Lake the atmospherically deposited Hg (probably from Deception Island volcano) is accumulated in the catchment. But instead of stabilization of Hg through the formation of organometallic complexes, at Rano Aroi part of the gaseous Hg is deposited and stabilized on the ice-snow pack that covers the lake and its catchments during most of the year or during prolonged cold periods with permanent ice cover. Here, peaks in Hg concentrations are related to release of such accumulated Hg after thawing. Previous research in Lago Hambre has suggested that Hg scavenging in the water column and accumulation in the sediments are mainly controlled by fluxes from the catchment soils. However, the multirpoxy approach (with spectroscopic and geochemical data) enabled the identification of unequivocal proxies of primary productivity, which showed a strong dependency on total solar irradiance (TSI). For the last 4.5 cal kyr BP the strong variations in Hg accumulation corresponded to changes in TSI and associated changes in aquatic productivity. The accumulation of Hg was highest during drier periods when insolation and lake productivity were 1. Summary 21 high and erosion fluxes from the catchment were low, indicating that sediment Hg accumulation (and potential Hg methylation) in this highly productive lake was controlled by insolation and the related algae productivity. Atmospheric mercury depletion events (AMDEs) result from the oxidation of gaseous elemental Hg to highly reactive forms and subsequently rapid elimination of the oxidized Hg species through precipitation. In Limnopolar Lake, the combination of Hg emission by volcanic activity and amplification of the signal by the freezing and thawing of the lake and the basin, could explain the recorded changes in Hg content. But, AMDEs could also help to explain the extraordinary high levels of Hg found in some samples from the sediment. With the exception of a significant correlation between the Br/C molar ratio (used as a proxy of AMDEs) and high Hg accumulation peaks, there are no other data that support the link between Hg maxima and AMDEs. However, the proximity of the coast (< 2 km) and the cold periods during which most of the Hg peaks occurred are favourable conditions for the generation of AMDEs. Differences in vegetation type can affect the net deposition and sequestration of Hg in peatlands, however no systematic sampling was carried out on the vegetation that could potentially affect the accumulation of Hg in peatlands and lake sediments. Vegetation samples were only collected in the Easter Island field sampling campaign and Hg results show a wide range of Hg concentrations reaching 11,117 ng g-1, on an unidentified rush (Juncus) sample. It has been hypothesized that an increase in the abundance of rush, coupled to more humid conditions by ~20 cal kyr BP, is related to detected peaks in Hg. However, there is no evidence for the presence of rush in the mire catchment at present and the available pollen and macroremains data do not support this hypothesis. Climate has undoubtedly been the main driver of Hg accumulation in the records presented in this PhD thesis, with the exception of Sandhavn.In all the other records climate played a major role, both directly and indirectly. The broad influence of climate on the records presented here could be due to two main reasons, i) the remote location of the sampling sites limit the direct effect of anthropogenic activities and/or ii) the long time-span of some of the records implies that different periods of very distinct climatic conditions are recorded (e.g. the transition between dry and wet climates during the Late Pleistocene at Tropical and Subtropical latitudes such as in Rano Aroi and Pinheiros). 22 The peat records of Pinheiros and Rano Aroi provided an exceptional opportunity to examine how two main climatic periods –Late Pleistocene and Holocene– affected the Hg cycle at tropical and subtropical latitudes. The results of these long-term records (Rano Aroi and Pinheiros) suggest, as it is generally accepted by extensive monitoring, that wet deposition (i.e. the wash out of Hg by rainfall) is the main source of Hg to peatlands. This is in contrast with recent shortterm studies with Hg isotopes that indicate that gaseous and particulate Hg (i.e. dry deposition) are the main species deposited into peatlands. The long temporal perspective provided by the Pleistocene records might highlight the dominance of the longer-term wet deposition processes. Additionally, at the timing of the Heinrich event 1 (17 cal kyr BP) characterized by dry and very cold conditions, the Pinheiros peatland recorded an extraordinary high Hg concentration. A similar effect was found in the Rano Aroi record at the end of the Last Glacial Maximum (at 20 cal kyr BP) when humid conditions prevailed in Eastern Island. Both, colder and humid conditions would have favored Hg accumulation since Hg deposition is controlled by temperature and humidity variations. The increased in peat oxidation between ~40 and 42 cal kyr BP that led to an increase in Hg concentrations was also driving by climate conditions, more specifically long-term drought. In Lago Hambre, changes in primary productivity in the lake during the last ~4.5 kyr cal BP, and thereby indirectly also changes in Hg content in the sediments, were related to TSI, Total solar irradiance is obviously a parameter that is transferred to the Lago Hambre palaeoenvironmental record through climatic factors. The freezing and thawing of lake ice regulated the accumulation of Hg in Limnopolar Lake and in its catchment. Freeze-thawing processes are also directly controlled by climate. The ice-snow layer over the lake and the catchment may have acted as a Hg sink during 9 - 10 months per year. Periods of sustained general cold climatic conditions may have extended the effect of this Hg trap at least from years to decades. This is consistent with the high Hg concentrations and accumulation rates corresponding to phases of minima in insolation (colder periods) as the Wolf, Spörer, Maunder and Dalton minima. Regarding the methodological aspects, it appeared that spectroscopic analysis in combination with multivariate analysis, in particular Partial Least Squares statistics can be efficiently and reliably used to predict Hg concentrations in 1. Summary 23 minerogenic peat, at least within the same core. This approach can be used to reduce the costs of multicore approaches facilitating the study of spatial variability within and between mires. Essentially, lake sediments and peatlands are environmental archives that can be used as records of atmospherically deposited Hg, thus providing a “picture” of the Hg cycle. However, the use of such archives as Hg atmospheric records is conditioned by factors that operate in different geographical and temporal scales. Thus, the Hg determined in the samples does not directly reflect atmospheric deposition, but results of a changing combination of the eight different factors described in this thesis. 25 1. Resumen El mercurio (Hg) es un elemento natural de distribución global. Debido a la toxicidad de sus formas metiladas, tanto para la vida salvaje como para la salud humana, el Hg se considera un metal de preocupación ambiental. De acuerdo a numerosas organizaciones gubernamentales y para la salud, miles de millones de personas están bajo un alto riesgo de exposición al envenenamiento por Hg, principalmente por dietas ricas en pescado (con un alto contenido de metil mercurio) o a la exposición al Hg emitido por la minería artesanal o a pequeña escala de oro en países en vías de desarrollo. Diferentes procesos industriales como la combustión de gasolina, la producción de cemento, la fundición de metales o la minería de oro a gran escala, así como la propia minería de Hg, han emitido directamente a la atmósfera miles de toneladas de Hg desde 1850. Además de las emisiones de Hg causadas por la actividad humana, existen numerosos procesos naturales que emiten mercurio al medio ambiente como la actividad volcánica (volcanes y fuentes hidrotermales) o las emisiones de suelos naturalmente enriquecidos. Otros procesos reemisionan el Hg en el medio ambiente, como los incendios forestales y las emisiones oceánicas. Estos son consideradas fuentes mixtas i.e. naturales y de origen antrópico. Las fuentes primarias son aquellas que transfieren el Hg desde los depósitos de la litosfera a la atmósfera y después a la tierra y los océanos, es decir, incrementando la 32 Los eventos de agotamiento de mercurio atmósfera (atmospheric mercury depletion events o AMDEs – siglas en inglés –) resultan de la oxidación en la atmósfera de Hg gaseoso elemental a formas altamente reactivas y, posteriormente, la eliminación rápida de las especies de Hg oxidadas a través de la precipitación. En el caso de Limnopolar la combinación emisiones volcánicas y amplificación de la señal por los procesos de la cuenca pueden explicar por si mismos los altos contenidos de Hg encontrados. Pero, los AMDEs ayudarían a reforzar los extraordinarios valores encontrados en algunas muestras de los sedimentos. Con la excepción de una correlación significativa entre la relación Br/C (usada como indicados de los eventos de agotamiento de Hg en este registro) y los valores más elevados de Hg, no hay otros datos que apoyen la relación entre el Hg y los eventos de agotamiento. Sin embargo la proximidad a la costa (< 2km) y las condiciones frías en las que se produjeron los mayores picos de Hg indican un ambiente muy favorable para que se produzcan los AMDEs. Diferencias en el tipo de vegetación puede afectar a la deposición neta y retención del Hg en las turberas. Sin embargo en los registros incluidos en esta tesis doctoral no se llevó a cabo un muestreo sistemático de la vegetación que potencialmente podría haber afectado a acumulación de Hg en las turberas y sedimentos lacustres. Únicamente se recogieron muestras de vegetación en la campaña de muestreo de Isla de Pascua. Las determinaciones de Hg mostraron un rango muy amplio de concentraciones en estas muestras de vegetación, desde 342 ng g-1 a 11.117 ng g-1 correspondiendo este último valor a una muestra de un junco no identificado. Se especula con la posibilidad de que un incremento de la abundancia de los juncos, unido a condiciones más húmedas puedan ayudar a explicar el incremento de Hg alrededor del 20.000 cal BP y en general las mayores concentraciones de Hg en los periodos húmedos. Sin embargo no hay certeza de la presencia de juncos en la turbera y los datos disponibles de polen y macrorestos no ayudan a apoyar esta especulación. El clima ha sido indudablemente el principal impulsor de la acumulación de Hg en los registros presentados en esta tesis doctoral. Con la excepción de Sandhavn, en los otros registros el clima ha desempeñado un papel principal, tanto directa como indirectamente. La amplia influencia del clima en los registros aquí presentados podría deberse a dos razones principales: i) la ubicación remota de los sitios de muestreo limita el efecto directo de las actividades de origen antrópico y / o ii) el largo período de tiempo cubierto por algunos de los registros 1. Resumen 33 que incluyeron condiciones climáticas contrastadas (por ejemplo, la transición entre los climas secos y húmedos durante el Pléistoceno Tardío en las latitudes tropicales y subtropicales, como en Rano Aroi y Pinheiros). Los registros de turba de Pinheiros y Rano Aroi proporcionaron una oportunidad excepcional para examinar cómo dos periodos climáticos principales -el Pleistoceno Lateral y el Holocenoafectaron al ciclo del Hg en las latitudes tropicales y subtropicales. Los resultados de estos registros a largo plazo (Rano Aroi y Pinheiros) sugieren, como es generalmente aceptado por el monitoreo extensivo, que la deposición húmeda (es decir, el lavado de Hg por la lluvia) es la principal fuente de Hg a las turberas. Esto contrasta con estudios recientes dónde utilizando isótopos de Hg en registros a corto plazo se indica que el Hg gaseoso y particulado (es decir, deposición en seco) son las especies principales depositadas en las turberas. La larga perspectiva temporal proporcionada por los registros del Pleistoceno podría poner de relieve la dominancia de los procesos de deposición húmeda a largo plazo. Además, durante el evento Heinrich 1 (17.000 cal BP) caracterizado por condiciones secas y muy frías, la turbera de Pinheiros registró una concentración extraordinariamente alta de Hg. Un efecto similar se encontró en el registro de Rano Aroi al final del último máximo glacial (a 20.000 cal BP) cuando las condiciones húmedas prevalecieron en la Isla de Pascua. Ambos, condiciones más frías y húmedas habrían favorecido la acumulación de Hg ya que la deposición de Hg está controlada por variaciones de temperatura y humedad. El aumento de la oxidación de la turba entre ~ 40.000 y 42.000 cal BP que condujo a un aumento en las concentraciones de Hg también fue impulsado por las condiciones climáticas, más específicamente por un periodo largo de sequía. En el Lago Hambre, los cambios en la productividad primaria en el lago durante los últimos ~ 4.500 cal BP, e indirectamente también los cambios en el contenido de Hg en los sedimentos, se relacionaron con TSI, la irradiancia solar total; que es, obviamente, un parámetro climáticos. La congelación y descongelación del hielo del lago reguló la acumulación de Hg en el Lago Limnopolar y en su cuenca. Los procesos de congelacióndescongelación también son controlados directamente por el clima. La capa de hielo-nieve sobre el lago y la cuenca pudo haber actuado como sumidero de Hg durante 9-10 meses al año. Durante los períodos de condiciones climáticas frías largas se puedo haber extendido el efecto de trampa de Hg de años a décadas. Esto 34 es consistente con las altas concentraciones de Hg (y las tasas de acumulación) correspondientes a las fases de mínimos en insolación (períodos más fríos) como los mínimos de Wolf, Spörer, Maunder y Dalton. En cuanto a los aspectos metodológicos, parece que el análisis espectroscópico en combinación con el análisis multivariado, en particular PLS, puede utilizarse de manera eficiente y fiable para predecir las concentraciones de Hg en la turba minerogénica, al menos dentro del mismo testigo. Este enfoque puede utilizarse para reducir los costos de los enfoques multi-testigo que facilitando el estudio de la variabilidad espacial dentro y entre los turberas. Esencialmente, los sedimentos de los lagos y las turberas son archivos ambientales que pueden utilizarse como registros de Hg depositado atmosféricamente, proporcionando así una “imagen” del ciclo de Hg. Sin embargo, el uso de estos archivos como registros atmosféricos Hg está condicionado por factores que operan en diferentes escalas geográficas y temporales. Por tanto, el Hg determinado en las muestras no refleja directamente la deposición atmosférica, si no que es el resultado de una combinación cambiante de los ocho factores diferentes descritos en esta tesis. 35 2. Introduction 2.1. Mercury as a global pollutant Mercury (Hg) is a naturally occurring element and is found throughout the world. Its volatility at ambient temperature and an unpredictable behaviour in the environment, make Hg fate on the environment “one of the most insidiously interesting and scientifically challenging biogeochemical cycles at the Earth’s surface” (Fitzgerald and Lamborg, 2003). Furthermore, because of the toxicity of its methylated forms, both to wildlife and human health, Hg is considered as a metal of environmental concern (WHO, 1989). The attention to health effects derived from Hg pollution began with a wellknown poisoning accident in the 1950s, when Chisso Corporation’s chemical residues were released into the nearby Minamata Bay, Japan. The wastewater discharge produced an increased amount of organic and extremely toxic Hg forms in the bay (methylmercury, MeHg), consequently boosting their contents in local marine organisms (mainly fish and shellfish) and entering in the food chain. This resulted in devastating health effects to local population who consumed fish as main food source (Kurland et al., 1960). Prenatal or postnatal exposure to MeHg produces neurological impacts in adults and children, including sensory disturbance, difficulty in coordinating movements, and tremors, among others symptoms (this syndrome is known as Minamata disease, (Harada, 1995). With increasing awareness of environmental stewardship, an incidence of acute Hg 36 poisoning from industrial pollution like Minamata’s has become rare. However, the scale of chronic exposure to a lower dose of Hg as a result of global pollution or occupational hazard has grown (Ha et al., 2017). According to the Food and Agriculture Organization (FAO) and the World Health Organization (WHO), billions of people are at risk of high exposure to poisoning by Hg due to their diet rich in fish (high content of MeHg) (World Health Organization, 2011) or by exposure of Hg emitted by artisanal and small-scale gold mining, mainly in lowand middle-income countries (Veiga et al., 2006). Recent research estimated that 112000 tons of Hg have been emitted directly to the atmosphere since 1850 from by-product sources (fuel combustion, cement production, metal smelting, large-scale gold mining with non-Hg methods) and 720000 tons of Hg were mined during the same period for commercial use (Streets et al., 2011). Commercial Hg use includes: Hg-containing products (e.g., batteries) and manufacturing processes that involve Hg (e.g., vinyl chloride monomer production). Other studies include an additional 540000 tons of Hg from other commercial Hg uses and non-atmospheric discharges from chlor-alkali plants and mining (Horowitz et al., 2014). The effects of artisanal and small-scale gold mining are also remarkable. For example, it has been calculated that this activity has released about 2000 to 3000 tons of Hg to Brazilian Amazon environment since 1980’s (Malm, 1998). In last decades, these concerns have drawn the attention of the political agenda. Following the international actions, the European Union (EU) adopted in 2005 the “Community Strategy Concerning Mercury” (COM, 2005). The Strategy aimed to reduce Hg levels both in relation to human exposure and the environment, addressing most aspects of the Hg life cycle. It identified twenty priority actions to be undertaken, both within the EU and internationally. The degree of implementation of these actions was reviewed in 2010, and it was declared that the implementation of the Mercury Strategy was in an advanced stage (COM, 2005). However the text highlighted the necessity of further international actions for the coming years. Giving the global aspect of the Hg problem, internal EU legislation alone cannot guarantee effective protection of the population. In 2013 the EU signed the Minamata Convention on Mercury, spearheaded by the United Nations Environment Programme with the aim “to protect the human health and the environment from anthropogenic emissions and releases 2. Introduction 37 of mercury and mercury compounds” (UNEP, 2013). The European Commission adopted on February 2, 2016 a ratification package that will allow the EU to ratify the Convention once the legislative process is concluded. On December 1, 2016 Minamata Convention has 128 signatures and 35 ratifications (Minamata Convention on Mercury web). The Article 19 section (e) of the Minamata Convention, concerning to Research, development and monitoring, highlights the necessity of cooperating to develop and improve the Information on the environmental cycle, transport (including long-range transport and deposition), transformation and fate of mercury and mercury compounds in a range of ecosystems, taking appropriate account of the distinction between anthropogenic and natural emissions and releases of mercury and of remobilization of mercury from historic deposition (UNEP, 2013). It therefore stresses the need to investigate current sources of Hg and their impact on the environment and health; as well as studying the mercury cycle, in the present and the past. 2.2. Global biogeochemical mercury cycling 2.2.1. Chemical species of mercury The chemical symbol for mercury, Hg, is derived from the Latin name hydrargyrum, which means silver water. The name refers to the appearance of elemental mercury (Hg0) which is readily recognized as a silvery liquid at room temperature and has a high vapour pressure (Brown et al., 2008) for a heavy metal. In the natural environment, it can exist in the gaseous or liquid state. Gaseous elemental mercury (GEM) is the dominant form in the atmosphere. Most natural waters are nearly saturated, or even supersaturated with respect to the atmospheric Hg0 (Morel et al., 1998; Fitzgerald et al., 2007). However, other species of Hg occur naturally in the environment. Elemental Hg can be oxidized into divalent mercury (Hg II), in the atmosphere, and washed out by rainfall (Hall, 1995). Mercury divalent compounds, both organic and inorganic, exist in gaseous, dissolved and solid states, being Hg (II) much more prevalent in waters than in the atmosphere (Swartzendruber and Jaffe, 2012). Although it is not a chemical Hg species, particulate-bound mercury, or Hg (p), refers to Hg that is extracted from particles, either airborne or waterborne. It has been shown that the observed Hg (p) concentration depends on the size of particles that are collected 38 by the technique, for example most airborne measurements include only particles < 2.5 µm (aerodynamic diameter) (Swartzendruber and Jaffe, 2012). Finally, monomethylmercury and dimethylmercury (MMHg or Me-Hg and DMHg) are organic forms of Hg. They are organometallic compounds formed by of one/two methyl groups (CH3)/(CH3-)2 bonded to a Hg ion. Hg (II) can be readily methylated in aquatic systems. Mercury methylation appears to be predominately biotic, although some abiotic production is likely in natural waters (Benoit et al., 2002). Both methyl compounds are significantly toxic and Me-Hg can accumulate up the aquatic food chain and lead to high concentrations in predatory fish (National Research Council, 2000). 2.2.2. Sources and reservoirs of mercury: mercury cycle The traditional classification of Hg sources as natural, mixed or anthropogenic, is useful to stress the emission mechanisms to the environment. Volcanoes, geothermal vents and emission by natural enriched soils are considered purely natural emissions, while land emissions, forest fires and ocean emissions are considered as mixed sources because a significant fraction of their Hg burden was previously deposited including some anthropogenic Hg (Selin et al., 2008). Anthropogenic Hg can be sourced from a wide array of activities. For the year 2000 the largest sources, in order of importance, were: coal combustion, gold production, nonferrous metal smelting, cement production, waste incineration and caustic soda manufacturing (Pacyna et al., 2006). Other additional sources account for the emissions in relation to various uses of Hg as battery making or production of electrical lighting, wiring devices and electrical switches (Pacyna et al., 2006). However, research in the last decades has demonstrated that the Hg fluxes within and between different reservoirs (air, soils, oceans, ice, atmosphere) are also relevant for defining its global biogeochemical cycle (e.g. Mason and Sheu, 2002; Sunderland and Mason, 2007; Selin et al., 2008; Holmes et al., 2010; SmithDowney et al., 2010; Driscoll et al., 2013) (Figure 1). This led to the adoption of a different classification of Hg sources as primary and secondary sources, having into account not only the initial emission, but also the subsequent transferences to other compartments and the processes involved. Primary sources transfer Hg from long-lived lithospheric reservoirs to the atmosphere and then to land and oceans, i.e. increasing the global pool of Hg in 2. Introduction 39 surface reservoirs (Driscoll et al., 2013). They would include naturally originated Hg (e.g. from weathering of Hg-rich materials and volcanic emissions). Degassing carries out the natural mobilization of Hg from the geogenic reservoirs through volcanic emissions, and are estimated around 500 Mg Hg yr-1 (Selin et al., 2008), ranging between 80 and 600 Mg yr-1 (see references in Driscoll et al., 2013). Studies on Mt. Etna emissions (Sicily, Italy) determined that only 1% (by mass) is in of particulate form, and Hg0 is the main species (Bagnato et al., 2007). However, primary sources are nearly always augmented by human activities (mining, fuel combustion, waste incineration) (Mason et al., 1994; Mason and Sheu, 2002). Most Hg in the atmosphere is in the form of Hg0 emitted from primary sources, although there is also Hg (II) and Hg(p) that are released by fuel combustion (Selin et al., 2008). The Hg life time in the atmosphere is ~0.5 years (Selin et al., 2008), allowing a long transport from the sources to remote locations such as the Arctic and Antarctica (Lindqvist and Rodhe, 1985; Mason et al., 1994; Ebinghaus et al., 2002; Sigler et al., 2003; Ariya et al., 2004; Lindberg et al., 2007; Durnford et al., 2010; Dastoor et al., 2015; Steffen et al., 2015). After deposition, Hg can be reemitted from the surface reservoirs to the atmosphere, therefore constituting a secondary source. Mercury from secondary sources is then spread among and within ecosystems (Driscoll et al., 2013). Recycling from surface reservoirs results in an effective lifetime of 1.6 years against transfer to long-lived reservoirs in the soil and deep ocean (Selin et al., 2008). Due to its volatility as Hg0 and the redox transformations of Hg in environment, deposition and evasion are influenced by multiple environmental factors including seasonality, vegetative cover and its life cycle, temperature, light, moisture, atmospheric turbulence and the presence of reactants that change with the surfaces e.g. soil, water or snow (Zhu et al., 2016). In land surfaces, on which ~60% of the Hg is deposited (Mason and Sheu, 2002), meteorological parameters (e.g., solar radiation, soil/air temperature, atmospheric turbulence), soil substrate characteristics (e.g., Hg content, soil moisture, organic matter, porosity and microbial activity), and ambient air characteristics (e.g., Hg0 and O3 concentration) can influence the air–surface exchange of Hg0 (Zhu et al., (2016) and references there in). It is remarkable the role of solar radiation in controlling Hg emissions from soil substrates, enhancing Hg (II) reduction to Hg0 40 and facilitating its evasion (Gustin et al., 2002). By this mechanism –facilitated reduction from the Hg pool of the soilfor example, agriculture operations resulting that disturb the soil surface, as tilling, may be important sources of Hg to the atmospheric (Bash and Miller, 2007). Similarly, vegetation plays an important role as Hg pool. Isotopes measurements have shown that freshly deposited Hg – more reactive with respect to volatilization and methylation than the native Hgthat is not immediately volatilized, is associated preferentially with vegetation (Hintelmann et al., 2002). Biomass burning releases around 600 Mg yr-1 (annual average for the period 1997 – 2006), which is equivalent to 8% of the total Hg emission (Friedli et al., 2009). Other processes affecting land, as for example runoff, are sometimes neglected in global Hg models (Selin et al., 2008) although it was estimated as 40 Mg yr-1 for preindustrial times (Mason et al., 1994). The deep-ocean sedimentation is the ultimate sink of Hg (Selin, 2009; Mason et al., 2012). Atmospheric deposition is the dominant source of Hg to most remote water bodies (Mason et al., 1994; Selin et al., 2008) and similar to freshwater and land systems; Hg (II) can be deposited by dry and wet deposition and Hg0 as dry deposition (Selin, 2009). The contributions from other sources to open ocean regions are much smaller on a global scale (Mason et al., 2012). Inputs from rivers could be equivalent to 25 to 41% of atmospheric deposition in Surface Atlantic, North Pacific and the Mediterranean Sea, but unimportant in the North Atlantic or the rest of the Pacific (Sunderland and Mason, 2007). On the other hand, hydrothermal vents are estimated to contribute to less than 20% of atmospheric inputs (Lamborg et al., 2006; Mason et al., 2012). In surface waters, processes controlling the concentration of dissolved gaseous Hg directly regulate air-water Hg0 flux (Zhu et al., 2016). More than ~70% of the Hg deposited in the ocean is re-emitted to the atmosphere as Hg0 but also some as (CH3)2Hg (Mason and Sheu, 2002; Soerensen et al., 2010; Corbitt et al., 2011). Evasion as Hg0 and Hg removal from the surface ocean by particle scavenging reduce the pool of potentially bioavailable Hg(II) that may be methylated (i.e. toxic forms) and bioaccumulated into marine organism (Mason et al., 2012). Polar regions are areas in which Hg retention is of particular interest due to the high concentrations found in the biota – mainly in the Arctic (Muir et al., 1992; Atwell et al., 1998; Douglas et al., 2012; Kirk et al., 2012) and to a lesser extent 2. Introduction 41 in Antarctica (Bargagli et al., 1998; Bargagli, 2008; Carravieri et al., 2013) - and because of the predicted increase (~25%) in atmospheric Hg deposition in the Arctic by 2020 (AMAP, 2011). Since the observation of episodes of unexpected low Hg0 air concentrations – known as Atmospheric Mercury Depletion Events (AMDEs) (Schroeder et al., 1998)-, first in the Arctic air during springtime and later in Antarctica (Ebinghaus et al., 2002), there has been an increasing interest to characterize and study their impact in the global Hg cycle (e.g. Holmes et al., 2010). Halogens (mainly atomic bromine and bromine oxide) are thought to play essential roles in AMDEs (e.g. Ariya et al., 2002, 2004; Calvert and Lindberg, 2003; Simpson et al., 2007; Faïn et al., 2008), in the proposed mechanisms (i.e. photo-chemical activation under daylight and marine evasions under darkness (Faïn et al., 2008; Nerentorp Mastromonaco et al., 2016)). Although it is known that this process also occurs at temperate and low latitudes (Obrist et al., 2011), in the Arctic it is estimated that AMDEs alone are responsible for the deposition of up to 100 tons of Hg per year north of the polar circle (Ariya et al., 2004; Durnford and Dastoor, 2011) constituting a significant part of the region’s annual Hg deposition (i.e. Ariya et al., 2004). Between 20 – 50 % of the freshly deposited, highly reactive (Lalonde et al., 2002) Hg during AMDEs is returned from the snow cover to the atmosphere within 24 hours (see references values in Durnford et al., Figure 1. The global mercury cycle from AMAP (2013), adapted from Mason et al., 2012 1.2 Global mercury budgets Mercury is released to the environment from natural sources and processes and as a result of human activities. Once it has entered the environment, Hg cycles between major environmental compartments – air, soils and waters – until it is eventually removed from the system through burial in deep ocean sediments and mineral soils. Methylmercury, the most toxic and bioaccumulative form of Hg which presents most health risk to humans and wildlife, is mainly produced in aquatic ecosystems through natural bacterial processes. In order to provide a general framework for the discussions in Chapters 2 to 5 of this report, the following section presents a global Hg budget based on recent modelling work. Owing to its scale and chemical complexity, and the lack of detailed information for many parts of the ecosystem, globalscale models provide the most practical means of describing the global Hg cycle in a quantitative manner. A number of global atmospheric Hg models exist. But until recently, only one combined atmospheric-terrestrial-oceanic model has been available, the GEOS-Chem Mercury model (Strode et al., 2007; Smith-Downey et al., 2010; Mason et al., 2012). Recent GEOS-Chem model results, as described by Mason et al. (2012), represent the current ‘best estimate’ of the global Hg budget. Th e model is constrained and generally supported by empirical data on Hg concentrations and fl uxes in various environmental media, and represents a consensus which has not been challenged within the Hg scientifi c community. As with all such modelled budgets, large uncertainties exist regarding both the amounts of Hg ‘stored’ in the different environmental compartments and the fl uxes of Hg between these compartments (see Table 1.1). Most of these uncertainties are due to unknown or poorly known input parameters and process rates, as discussed in more detail in Chapter 3. Despite this fact, most global Hg models do not present uncertainty estimates on their mass balances and fl uxes. Sunderland and Mason (2007) reported that 90% confi dence intervals for GEOS-Chem estimates of most fl uxes (i.e., for rivers, atmospheric deposition, particle settling, lateral and vertical fl ows) were only 2to 4-times as large as the median or best estimate values. However, the estimates of evasion were less certain, with 90% intervals of 5–10 times the best estimate for diff erent ocean basins. Uncertainty analysis conducted as part of new global Hg modelling work by Qureshi et al. (2011; the World Multimedia Mercury Model, WorM3) suggested that 95% estimate dispersion ranges were over an order of magnitude for most global Hg inventories and fl uxes. However, the best estimates from Qureshi et al. (2011) compared well with other models. For example, Qureshi et al. (2011) estimated a net conversion of Hg0 (elemental mercury) to HgII (inorganic divalent mercury) in the atmosphere of 3000 t/y, with a 95% range of 400 to 12 400 t/y. Th is average compares favourably with the 6000 t/y estimate using GEOS-Chem (Selin et al., 2007). Th e calculated atmospheric residence time of Hg0 in WorM3 was 8.2 months with a 95% dispersion of 2.4 to 24 months, which also agrees well with other estimates of 8.4 to 20.4 months (Holmes et al., 2006; Selin et al., 2007). In general, good agreement (within a factor of three) was observed for the best estimates of most global Hg compartment inventories, chemical reaction rates and fl uxes, between WorM3 (Qureshi et al., 2011) and other spatially resolved global models including GEOS-Chem (Lamborg et al., 2002; Selin et al., 2008; Sunderland and Mason, 2007; Soerensen et al., 2010; Smith-Downey et al., 2010). Th e largest potential errors in the GEOS-Chem model, in the context of the aquatic Hg cycle, may concern air-water gas exchange, specifi cally: (i) the mechanisms of the redox reactions in surface oceans, as defi ned by the amount of reducible Hg present in surface oceans, and rate constants for reduction and oxidation of Hg species by various pathways; (ii) atmosphere-water Hg mass transfer processes as defi ned by wind velocity; and (iii) Hg species inter-conversion reactions in the atmosphere (Qureshi et al., 2011). It was estimated that these uncertainties may contribute more than errors in anthropogenic emission estimates to the total uncertainty in modelled atmospheric concentrations and deposition fl uxes. 80600 3700 20002950 (200%) 2000300600 17002800 (290%) 3200 380 <600 Geogenic Anthropogenic AnthropogenicBiomass burning Soil and vegetation Net Hg0 evasion Deposition to oceans Rivers Deposition to land / freshwater Geogenic particle removal net vertical transport mineral soils: 802000 (n/s) atmosphere: 5100 (300-500%) organic soils: 200600 (20%) deep waters: 221000 (11%) intermediate waters: 134500 (25%) surface ocean: 2900 (205%) Anthropogenic Natural Re-emissions / Re-mobilisation Figure 1.2. The global mercury budget. Source: adapted from Mason et al. (2012). Total inventories (numbers in white boxes) are in tonnes, and fluxes in tonnes per year. The percentage values in brackets are the estimated increases in inventories in the past 100 years due to anthropogenic activities. 2 Technical Background Report for the Global Mercury Assessment 2013 48 2016), as well as recent decreases in North America and Europe (e.g. Bindler et al., 2001b; Mast et al., 2010). Additionally, lake sediments have also recorded Hg deposited after long-range transport, in particular in lakes from remote areas; e.g. cores collected in northern Canada showed increasing Hg inputs from 1500’s and more rapidly after AD 1750 and 1900 (respectively) due to the effect of regionalglobal emissions in an area with no local industrial sources of Hg (Lockhart et al., 1995). Data from lake sediments from remote regions of Tibet show the current increasing trend in pollution in Asia with a start around the AD 1970s and 1980s due to the Asian Industrial revolution (Yang et al., 2010a). Contributions form natural sources have also been recorded. Mercury emissions from volcanic activity and forest fires have been registered in mountain lakes from the Northern Patagonian Andes, a region highly susceptible to volcanic activity because it is part of the Southern Volcanic Zone – it includes at least 60 historically and potentially active volcanoes (Ribeiro Guevara et al., 2010; Daga et al., 2016). Despite all of this, several investigations have also identified processes that may alter the concentration of Hg in lake sediments, complicating their use as archives of atmospheric deposition. The size and land uses of the watershed have been proved to have a relevant influence on Hg concentrations in the sediments (e.g. Fitzgerald et al., 2005; Engstrom et al., 2007; Drevnick et al., 2016). A recent extensive study (165 dated sediments cores from 138 lakes across western North America) on Hg accumulation rates showed that, for lakes not directly affected by point sources (smelter or mining activity), two groups were clearly discernible: i) lakes with little or not watershed perturbation and ii) lakes with watersheds with an extensive agricultural or urban cover (Drevnick et al., 2016). The disturbance of the watershed, by reducing Hg retention by soils and enhanced Hg deposition in urban areas, may be the main cause of this separation (Drevnick et al., 2016). Inputs by erosion and watershed runoff could be equal or higher than the direct atmospheric deposition, also in relatively small catchments (Fitzgerald et al., 2005). The effect of Hg mobility in sediments and diagenesis have been considered as likely processes since the fist works (e.g. Aston et al., 1973; Matty and Long, 1995). Lockhart and co-authors (Lockhart et al., 2000) tested Hg mobility using lake records from Canada with an independently known history of Hg inputs, finding a good agreement. Some studies also addressed in detail the role of organic 2. Introduction 49 matter and its digenesis – chemical and physical changes that occur within the sediment after deposition (Sanei and Goodarzi, 2006; Rydberg et al., 2008). In spite of the important role of soluble organic matter in Hg concentrations and the 20-25% of carbon loss after the first 10-15 years, not obvious loss of Hg was found over time (Rydberg et al., 2008). Finally, and not far from the role of the organic matter, within lake primary productivity has been shown to have a significant effect on Hg accumulation, mainly through faster rates of algal - scavenging and sedimentation (Outridge et al., 2007; Stern et al., 2009). These investigations suggest that these processes, which are climatically driven, could imply an overestimation of atmospheric Hg contributions to High Arctic lakes, since the increased productivity may enhance the accumulation of Hg in sediments without significant variations in atmospheric fluxes. 2.3.3. Peatlands Peatlands are wetlands formed by the accumulation of peat and have current peat-forming vegetation. Peat – dead and decaying plant material - accumulates because the net production of organic matter exceeds its decomposition. On their surface, atmospherically deposited elements, such as Hg, can be trapped. The continuous growth of the peat deposit allows the record formation. Similar to lakes, peatlands are world wide distributed: they cover over 400 million ha in about 180 countries (Parish et al., 2008). Thus, peatlands have been abundantly used as Hg deposition records; in North America (Norton et al., 1997; Benoit et al., 1998; Givelet et al., 2003; Roos-Barraclough et al., 2006; Outridge and Sanei, 2010; Outridge et al., 2011), Greenland (Shotyk et al., 2003), Scandinavia (Pheiffer-Madsen, 1981; Shotyk et al., 2003; Bindler et al., 2004; Steinnes and Sjøbakk, 2005; Rydberg et al., 2010a), Faroe Islands (Shotyk et al., 2005), United Kingdom (Coggins et al., 2006; Farmer et al., 2009), Germany (Biester et al., 2012), Switzerland (Roos-Barraclough et al., 2002), Belgium (Allan et al., 2013), Czech Republic (Zuna et al., 2012), France (Enrico et al., 2016), Spain (Martínez Cortizas et al., 1999, 2012; Corella et al., 2017), China (Tang et al., 2012; Li et al., 2016), Patagonia (Biester et al., 2002, 2003; Franzen et al., 2004) and New Zealand and Nova Scotia (Lamborg et al., 2002). Several investigations performed on peat cores were able to show the effect of industrial pollution, although the sources were not always clearly identified. For 50 example, the study of three peat cores from western Ireland showed maximum Hg accumulation occurring between AD 1950 and 1970, and concentrations were significantly lower than those found in other European peat records (Coggins et al., 2006). The authors related the chronology of Hg pollution with the influence of North American sources, due to long-range transport, and the lower concentrations with the prevailing westerly winds bringing clean, marine air to the sector. Another multisite study, using cores from different areas in Scotland, recorded Hg maxima at different times possibly reflecting local/regional influences until AD 1970s (Farmer et al., 2009). Similarly, regional anthropogenic Hg sources during and after the Industrial Revolution – coal burning and smelter Hg emissions – may have been responsible for the extremely high accumulation rates in peatland records from Eastern Belgium, which exceeded the pre-Industrial values by a factor of 63 (Allan et al., 2013), much higher than 4.2 fold found in Northern Spain (Martínez Cortizas et al., 2012) and 4 in Sweden (Bindler et al., 2004). Peat records also recorded other modifications in atmospheric Hg deposition. For example, an 8000 years reconstruction in Canadian peatlands showed the effect of biomass burning for agricultural activities by Native North Americans (Givelet et al., 2003). The increase in rainfall had a positive effect by increasing the deposition of Hg emitted to the atmosphere by different sources, as metal smelting from ancient China cultures (Tang et al., 2012) and volcanic emissions (Roos-Barraclough et al., 2002). An exceptional use of Hg records was the one developed by Martínez Cortizas and co-authors (Martínez Cortizas et al., 1999). Using a 4000 years record from Northwest of Spain they demonstrated that changes in temperature may have influenced Hg deposition and Hg thermal stability in peat. Cold climates seem to have promoted an increase in Hg accumulation as well as the preservation of Hg with low thermal stability. Conversely, lower accumulation and larger proportions of Hg with moderate to high thermal stability characterized warm climates. The findings have not only implications for the Hg cycle, but also Hg thermal lability was proposed as a paleotemperature proxy (Martínez Cortizas et al., 1999). However, there are still issues to be solved regarding the use of peat cores to reconstruct Hg atmospheric deposition. Some of them have been previously identified and revised (e.g. Benoit et al., 1998; Biester et al., 2007). One of the main issues concerns the spatial variability in the Hg accumulation 2. Introduction 51 within the same peatland. The studies (e.g. Bindler et al., 2004; Martínez Cortizas et al., 2012) show that within bog variability is high, and data for a single core reconstruction do not necessarily provides a representative estimation of atmospheric fluxes (although chronologies were quite similar). Bindler et al. (Bindler et al., 2004) suggest that, for a reliable reconstruction of Hg fluxes, it is the necessary to incorporate data from multiple sites or at least multiple cores. As the work done by Allan et al., (2013) shows, this approach can provide excellent results. However, the cost and logistics are sometimes an insurmountable problem. Related to the within bog variability it is the effect of vegetation type in Hg sequestration. Rydberg et al., (2010b) analysed for Hg living plants in a mire from northern Sweden with two different types of vegetation. They found differences in peat and living mosses depending on whether the area was pine-covered or had no tree vegetation and significant differences in Hg concentrations for two Sphagnum species. The effect of decomposition of the peat organic matter and its role on the release of Hg, is another issue that has been considered in several investigations. In peat sections dated to pre-anthropogenic times Hg enrichments were found to coincide with comparatively low C/N ratios, which indicate higher peat decomposition (Biester et al., 2003). The increase in peat humification promoted by dry conditions at the surface of the peatland, may lead to a 2-3 fold increase in Hg concentrations independent of the changes in atmospheric deposition (Martínez Cortizas et al., 2007). Because of this, (Biester et al., 2003) suggested that all peat properties affected by humification processes should be normalized to the same degree of humification using a “Mass Loss Compensation Factors”. On the contrary (Zaccone et al., 2009) did not found a major effect of the degree of decomposition in Hg concentrations, concluding that atmospheric deposition was the main driver of Hg fluxes during the last decades. Other processes that could modify the suitability of peatlands as archives of Hg deposition were discussed by Biester et al., (2007). As previously mentioned, one of the focus of their critical review was to understand the differences in Hg deposition rates between lakes and peat cores (3-5 times vs. 30-500 times above the natural background, respectively). The authors suggest that peat decomposition may result in lower background Hg accumulation. At the same time, the downward migration and accumulation of 210Pb at the limit of the water table could produce and underestimation of peat ages and Hg accumulation rates as well. 53 3. Justification and Objectives The previous sections have put in value the relevance of Hg for human health and the complexity of its global cycle in the environment. Similarly, the most commonly investigated environmental archives (ice-cores, lake sediments and peat records) have been proof to be useful to understand the Hg cycle in the past and in at present, and may also help to model future scenarios. However, following the recommendation of Blais et al., (2015), it is necessary to examine the underlying processes that occur in the archives to understand what is going on and unravel the true story. In relation to this topic, several advances have been made in the last decade but on a limited sample of records; i.e. most of the studies are from Northern Hemisphere records – Mainly Europe and North America - and are based on Holocene (last 11,000 years) records. The few longer term studies, as those of (e.g. Vandal et al., 1993; Roos-Barraclough et al., 2002; Hermanns and Biester, 2013a), showed that other factors than human activity could have a main role in controlling Hg accumulation. This leads to the need to consider whether some of the processes that control the accumulation of Hg in the archives, such as the effect of decomposition, climate and human activity, occur at long time scales. Finally, with advances in the use of environmental proxies and the incorporation of new techniques, it is necessary to look for tools that allow a better understanding of the studied system as well as facilitate multi core or broader studies. 54 The main objective of this PhD thesis is to gain insights into the variations in atmospheric Hg deposition over long time scales (late Pleistocene and Holocene) in the Northern and Southern hemispheres, through the use of environmental archives (peat and lake sediments). Within this general objective, this work pays special attention to understanding how the processes that control and modify the deposition and accumulation of Hg in the environmental archives i) change their relative weight over time ii) in different environments and iii) the relationships between them. Figure 2 describes the research line followed in this thesis through a backward glance, from the Hg determined in the samples and the multiproxy approach to the processes – and the interaction between them – on a given time scale and location. A minor methodological aspect addressed is the application of cheaper, and low time consuming, techniques to aid in multi-core/multi-site studies and what kind of information can be obtained from them. Figure 2. Research line followed in this Doctoral Thesis organic matter decomposition pollution vegetation climate volcanic activity algae scavenging soil erosion Lake sediments Peatlands geographic scale accumulation AMDEs deposition CONTROLLING FACTORS mercury concentration multiproxy approach time scale 55 4. Material and methods 4.1. Study sites To achieve the objectives, five archives from different areas of the word and different time scales were studied. Figure 3 shows the location and the type (lake sediments or peat) of the archives studied whereas Table 1 summarizes the main characteristics. All the records used in this PhD thesis have been subject to previous studies on different topics as geochemistry, paleoclimate, peatland development and palynology, providing detailed background information and allowing a multi proxy study. See more information in the corresponding chapters and institutional acknowledgments section for groups implicated in the research. Two of the cores have sections with ages dating back to the late Pleistocene, and both were collected in tropical or subtropical peatlands from remote areas. Rano Aroi (ARO) is a mire located in an ancient Pleistocene volcano crater in Easter Island (at 511 m a.s.l.). The two Rano Aroi cores studied here (ARO 08 02 and ARO 06 01, 4 and 13.9 m deep respectively) cover different temporal range (38.7 kyr BP-present day and 71.0-8.5 kyr BP, respectively) and were retrieved in different parts of the peatland, although less than 50 meters apart. Climatic conditions have been shown to be the main driver in the development of the peatland during the last ~71 kyr BP, as was also responsible for the increase in soil erosion from the small basin over the peatland during high precipitation 56 phases (Margalef et al., 2013, 2014). Pinheiros (PI) is a minerogenic, valley mire located in Serra do Espinhaço Meridional, state of Minas Gerais, Brazil. Although minerogenic, the location of the mire on the mountain summit (higher elevations at 1300 – 1320 m a.s.l.) and the rather small catchment area (~1.4 km2) makes it quite sensitive to variations in rainfall. Previous studies in the area also show that the climate played a main role in the development of the mire and in their geochemical properties (Horák-Terra et al., 2014). The studied core was sampled in 2010 to a depth of 3.24 m and covers the last ~57.0 kyr BP. Table 1. Summary of the characteristics of the environmental archives used in this work. Records Type of archive Age section Hemisphere manuscript nº Peatland Lake sediments Holocene Late Pleistocene North South Rano Aroi x - x x - x I Pinheiros x - - x - x II, III Lago Hambre - x x - - x IV, V Limnopolar Lake - x x - - x VI Sandhavn x - x - x - VII Two lake sediments from the South Hemisphere provided us an exceptional opportunity to complement the information obtained with the peat studies. Lago Hambre (LH) is a small (0. 013 km2) and deep lake located near the Strait of Magellan in southern Patagonia (Chile). A long sediment core was recovered from the deepest part of the lake in 2008 using a 5 m long piston corer and a 41 cm long gravity core was also taken, to recover the unconsolidated recent deposits (Hermanns and Biester, 2013a). Although the deeper sections of the lake sediments have late Pleistocene sections (reaching 17.0 cal kyr BP; Hermanns and Biester, 2013a), in the two papers presented here only the Holocene sections – last ~12.0 and 4.5 cal kyr BP (respectively) were use. Previous studies indicated that the input of inorganic and organic matter from the catchment soils were driven by climate and played a main role in the geochemistry of the lake, including the 4. Material and Methods 57 processes controlling Hg accumulation (Hermanns and Biester, 2013a; Hermanns et al., 2013). In addition, the sediments of the lake recorded four previously known volcanic eruptions of Reclus, Mt Burney and Hudson volcanoes. Limnopolar Lake (LIM) is a small lake (0.022 km2) located in Byers Peninsula, the westernmost part of Livingston Island (South Shetland Islands, Antarctica). The lake has a small catchment area of 0.58 km2 and is ice covered except for 2–3 months during the summer. The results of previous studies on the same core (LIM-03, 57 cm deep) indicate that volcanic activity, most probably of the nearby Deception Island volcano, has played a major role in the chemical and mineralogical composition of the lake’s sediments during the last ~1600 years recorded in the retrieved core (Martínez Cortizas et al., 2014). A short core (40 cm) was collected in 2008 from a small peatland nearby the Norse farmstead of Sandhavn (SH), located on the coast of Greenland approximately 50 km northwest of Cape Farewell (the most southerly point on the island). In a small rock depression isolated from the groundwater table, the Figure 3. Map showing the location of the study sites. Circles: Peatlands; Diamonds: Lake Sediments Pinheiros PI Limnopolar Lake LIM Sandhavn SH Rano Aroi ARO Lago Hambre LH 64 5.2. Pollution Human activity (i.e. pollution) has been considered one of the main drivers controlling Hg concentrations in records worldwide distributed that cover the last centuries or millennia – mainly showing the effect of Industrial or the PreIndustrial activities, respectively. The human processes, which released Hg in both periods, are usually different. While in the Pre-industrial time mining and smelting were the main human activities causing pollution (e.g. Martínez Cortizas et al., 1999; Cooke et al., 2009a, 2011; Corella et al., 2017), from AD 1850 several sources have been emitting Hg to the atmosphere as a byproduct (fuel combustion, cement production, metal smelting, large-scale gold mining), as well as commercial products (e.g. batteries) and manufacturing processes that involve Hg directly (e.g. vinyl chloride monomer production) (Streets et al., 2011). Investigations performed in areas near to the ones studied in this PhD attest for the presence of Hg pollution (see e.g. (Sun et al., 2006) for Antarctica, (Hermanns and Biester, 2013b) South Patagonia, (Lacerda et al., 1999) and Brazil; or human activities potentially causing of Hg pollution signal, as fires in Easter Island (Rull et al., 2015) or in South Patagonia (Morello et al., 2012). However only the Greenland record (Sandhavn) provided a clear signal of the effect of Hg pollution. Several reasons may explain the lack of evidence of recent atmospheric pollution in some of the studied cores. In the case of the Pleistocene records (Rano Aroi and Pinheiros), the time resolution of each peat section we have analysed is of a few hundred years and the pollution may has been averaged out, due to the sampling strategy or by low sampling resolution in Rano Aroi. For Limnopolar and Lago Hambre it is likely that other processes driving the Hg accumulation may overlap with the pollution signal. In Lago Hambre, we cannot rule out that the modern sections of the core (which reach AD 1850) do not cover the pollution period, stated as beginning at AD ~1900 in other studies in the area (Biester et al., 2002; Hermanns and Biester, 2013b). At Sandhavn, a previous multiproxy study already showed evidence for Pb enrichment beginning by AD ~1845 and, based on the timing of the events and other proxy data (long-distance pollen and crytotephras), it was suggested that the source area was most probably northeastern North America (Silva-Sánchez et al., 2015). The Hg accumulation rates (<1.0 - 10 μg m-2 yr-1, minimum and maximum) determined are in agreement with other Greenland records (Bindler et 5. General discussion 65 al., 2001a) but generally low in comparison with other records from Europe and North America (e.g. (Benoit et al., 1994; Bindler, 2003; Farmer et al., 2009). These observations may be explained in two ways: i) Sandhavn is located far from any major sources of pollution in North America and Europe or/and ii) by a lower peat decomposition (see the discussion section about organic matter decomposition). The chronology provided by Hg accumulation rates is similar to other studies in the area, although there may be slight mismatches. On one hand, the maximum value of Hg pollution agrees with previous data obtained on snow and ice from Summit Station at North Greenland (at ~1400 km) (Boutron et al., 1998; Faïn et al., 2009) and its chronology is close to the AD 1950 also found in a nearby peat record (~150 km northwest) (Shotyk et al., 2003). On the other, lake sediments records of lakes from Kangerlussuaq Fjord (~850 km northwest) point to an onset of Hg pollution at least by the late 19th century but possibly as early as the 17th century (Bindler et al., 2001a) (between ~50 years before to ~60 years later that in Sandhavn). The records also show a disagreement in the maxima of Hg pollution. The mismatch in the chronology of Sandhavn and the lakes from Kangerlussuaq Fjord is probably due to the meteorology of the ice margin (i.e. depletion events) that could drive Hg enrichment (Bindler et al., 2001a). This, together with the agreement with summit data suggests that our chronology of Hg pollution is more reliable. To get a better perspective of metal pollution in South Greenland, Hg chronology was compared with data from another long-range atmospheric metal pollutant (Pb) from the same record (Silva-Sánchez et al., 2015). The results show the same significant increase by the end of the AD ~1800s but also some differences as: i) the start of pollution (AD ~1845 for Pb and just before ~1800 for Hg) and ii) their peak values (at AD ~1980 and ~1940, respectively). The isotopic Pb residuals (i.e. the detrending between the trajectory of the unpolluted trend and the observed (polluted) values from the past two centuries, Manuscript VII) suggest an earlier beginning of pollution, between AD ~1740 – 1780, in agreement with other studies using Pb that found that modern metal pollution in Greenland begun prior to the 19th century (Massa et al., 2015). Simultaneous with Hg and Pb concentrations trend, isotopic Pb residuals indicate an increased in pollution from the beginning of AD ~1800. Finally, the highest Pb pollution signal indicated by the isotopes dates to the AD 1940s and 1970s, and agrees with data obtained from analyses of the Summit ice cores (Rosman et al., 1994; Faïn et al., 2009). 66 The fall in gasoline Pb consumption in the USA since AD 1970, declining ~80% by the early AD 1980s (Nichols, 1997), probably contributed to the pronounced increase in isotopic residual values from AD ~1979. Additionally, when more than two sources of Pb pollution are involved, the calculation of trends and residuals proposed by us seems to be a helpful and precise approach to determine changes in the chronology of the Pb isotopic signature. We do not have direct evidence for the origin of Hg, but according to the Pb isotope results we would expect the main source for Hg contamination to be also northern North America. However, Hg has a complex behavior in the environment. Its relatively long residence time in the atmosphere (1 year) favors long-range transport and homogenization at a hemispheric scale, making more difficult to determine its precise origin. The combined analyses of Pb and Hg may provide the means to assist further in the identification of such pollution sources in northern latitudes and to obtain a better perspective of metal pollution. 5.3. Volcanic activity Volcanic emissions produce a natural mobilization of Hg from geogenic reservoirs and are estimated around 500 Mg Hg yr-1 (Selin et al., 2008). However, there are regional differences in Hg average emissions. Mercury fluxes in South and Central America are up to 12 times higher than those in Africa for the period 1980 – 2000 (Nriagu and Becker, 2003); and fluxes from Mt Etna (Italy) could be equivalent to 5% of the estimated annual industrial atmospheric Hg released in the Mediterranean area (data from the period 2004-2006; Bagnato et al., 2007). Thus, at a regional scale, it has been shown that the volcanogenic Hg contribution is not trivial (Nriagu and Becker, 2003; Bagnato et al., 2007) and must be taken into account as a significant source. Nonetheless, evidence of Hg associated to volcanic eruptions in environmental records is not as common as expected. Ice cores from high elevations have proved to be the most reliable for this purpose. An ice core collected from the Upper Fremont Glacier (Wyoming, US), and representing the last 270-years (Schuster et al., 2002), provided a record of global (Tambora (at 1815) and Krakatu (at 1883); (Delmas et al., 1992; Cole-Dai et al., 2000)) and regional (Mount St. Helens (at 1980); (Nafth, 1993)) volcanic Hg emissions. But even in this case, and despite the proximity of Mount St. Helens (600 km) and the magnitude of the eruptions, volcanic events only contributed 6% of the total Hg vs. 52% of 5. General discussion 67 anthropogenic inputs (Schuster et al., 2002). In peat and lake sediments, usually from lower altitudes, enrichments in Hg due to volcanic eruptions are even scarcer and generally show a regional influence area, as observed in lakes located in the proximity of Southern Volcanic Zone, in Northern Patagonian (Guevara et al., 2010). Previous studies demonstrated the influenced of regional volcanism in two of the records studied in this PhD, Limnopolar Lake and Lago Hambre, but their effect on Hg concentrations were completely different. Lago Hambre record has several tephra layers corresponding to the eruptions of Reclus, Mt Burney and Hudson but no Hg signal was apparently linked to them (see (Hermanns and Biester, 2013a)). Only a Hg peak below the Mt Burney tephra could have been taken as an indication of a regional (Mt Burney ~300 km NE) volcanic event; but the lack of correlation of this peak in the Hg record of a peat core, sampled not far from the lake, excludes the volcanic activity as its cause (Hermanns and Biester, 2013a). The authors suggested that mechanisms linked to the fluxes of organic matter from the catchment to the lake may have been the main process controlling Hg accumulation (Hermanns and Biester, 2013a; Hermanns et al., 2013). The reevaluation of Lago Hambre Hg data (Manuscript V), based on the climatic and environmental reconstruction of the area (Manuscript IV), enabled to refine the previous hypotheses but also definitely confirmed the absence of a direct effect of volcanism on the accumulation of Hg (see Manuscript IV and V). Several tephra layers found in Limnopolar sediments attest for the effect of volcanic activity - probably of Deception Island volcano (30 km to the SE) – during the last ~1600 years (Toro et al., 2013). Volcanic eruptions have influenced the geochemistry of the lake by increasing the volcanic vs. catchment inorganic fluxes (Martínez Cortizas et al., 2014). Unlike Lago Hambre, volcanic eruptions have had an effect on the Hg record of Limnopolar Lake (Manuscript VI). Three of the identified tephra layers from Deception Island Volcano match reasonably well with five of the nine large peaks in Hg content (Hg accumulation rates between 300 to 4900 µg m-2 yr-1), specifically with tephras L4 (AD ~1300), L3 (AD ~1450 – 1470) and L1 (AD ~1840 – 1860). On the other hand, there was no Hg peak corresponding to tephra layer L2 (AD ~1570 – 1650) and there are no tephra layers associated to some Hg peaks (AD ~549, ~884, ~1715 and ~1758). Additionally, the chronology of volcanic eruptions of Deception Island (see references in Bartolini et al., 2014) and that of Limnopolar’s Hg main peaks and recent tephra records 68 is quite similar; although not for the most recent phase of high volcanic activity in Deception Island (from AD. ~1800 to the present), which does not correspond with the same number of Hg peaks (Figure 4 in Manuscript VI). A second order of Hg peaks (Hg accumulation rates higher than 15 µg m-2 yr-1) has no relation with tephra layers neither with volcanic eruptions in Deception Island (Bartolini et al., 2014). Thus, it seems evident that although volcanic activity played a role in the extremely high levels of Hg found in Limnopolar sediment samples, probably as a main Hg source, other mechanisms may have also be involved (see Manuscript VI). Finally, and somewhat more speculatively, is the effect of volcanic eruptions in the Hg accumulated at Rano Aroi peatland since, unlike Limnopolar and Lago Hambre, there is no evidence of recent volcanic activity in the peat cores (i.e. tephra layers). However, on an island containing more than 70 volcano craters (Baker et al., 1974; González-Ferrán et al., 2004), the possible degassing by active fumaroles that could act as primary Hg sources can not be ruled out; despite other processes are suggested to control the Hg deposition (see catchment effect and climate effect). Similarly, although we have no data to support it, the Hg peak at ~4.7 cal kyr BP (ARO 08 02) and/or at ~8.5 cal kyr BP (ARO 06 01) match reasonably well with the previously mentioned South Patagonia volcanic eruptions at ~4.2 cal kyr BP and ~7.7 cal kyr BP identified as Mt Burney (52ºS 72ºW) and Hudson (45ºS 72ºW), respectively (McCulloch and Davies, 2001; Stern, 2008) (see further discussion in Manuscript I relative the 0.5 kyr of mismatch between Hg peak and tephra layer). If this was true, i) both volcanoes would have been operating as large scale Hg sources given that a rise in Hg was found in Rano Aroi (Easter Island, ~4000 km) but not in peat and lake sediment records (South Patagonian (Biester et al., 2003; Hermanns and Biester, 2013a) and Manuscript VI) closer to the volcanoes (<300 km); ii) or other mechanisms would have been involved in enhancing Hg accumulation. 5.4. Organic Matter Decomposition Changes in superficial wetness or in the water table level played a critical role in the cycling of elements coupled to organic matter dynamics, as Hg. It has been suggested that factors associated to peat organic matter evolution and mass loss should control Hg accumulation in peatlands (Biester et al., 2003, 2006, 2007; Martínez Cortizas et al., 2007), which may impose some limitations to the use of 5. General discussion 69 bogs as reliable atmospheric Hg archives (Biester et al., 2007). Using the C/N ratio as indicator of degree of peat decomposition, it was found that Hg concentrations are higher in more decomposed peat due to the mass losses during decay (Biester et al., 2003) and could reach a 2-3 times increases at short term periods independent from atmospheric fluxes (Martínez Cortizas et al., 2007). On the contrary, other studies using absorbance measurements of alkaline peat extracts as an index of humification found higher Hg concentrations in poorly decomposed peat, which allowed to identify the atmospheric pollution input (e.g. Givelet et al., 2003; Shotyk et al., 2005; Bao et al., 2016). In the same line, studies on Hg retention by humic acids fractions indicate that the highest Hg concentrations occur in low humified peat and are associated to atmospheric pollution –supporting the use of bogs as archives of atmospheric Hg deposition (Zaccone et al., 2009). However, a recent comparison of different methods to determine the degree of peat decomposition concluded that the alkaline extract mainly reflects the formation of humic acids through humification and, to a lower extent, mass loss associated to mineralization (Biester et al., 2014), questioning some of the results mentioned above. The previous observations are almost exclusively based on Holocene peats, since only one of the cores from the mentioned studies covers part of the LatePleistocene, reaching 14 kyr (Biester et al., 2003). This could be a limitation to study a process (i.e. organic matter decomposition) that depends on time but also on climate changes. Additionally most of the previous studies are base on one decomposition proxy despite it has been suggested that more than one should be included to obtain a more representative assessment of long-term peatland evolution (Hansson et al., 2013). For this purpose multirpoxy approach of Rano Aroi (Easter Island) and Pinheiros (Brazil), which extend back to ~71 and ~57 kyrs respectively, are unique to address the long-term effect of organic matter decomposition and of major climate changes on Hg accumulation (Manuscript I and II). At high latitudes, organic matter degradation can be limited and small climatic variations could induce strong changes. Sandhavn has a detailed chronology of changes in decomposition during cold periods (mainly LIA, see pollution discussion) (Silva-Sánchez et al., 2015), providing an opportunity to get insights into the effect of short-term decomposition. On the other hand, Limnopolar Lake has a very limited input of organic matter, which may be an advantage to observe the effects of its changes in Hg accumulation. Detailed information about decomposition and proxies used are found in the manuscripts (I, II and VII) and previous studies (Margalef et al., 2013, 2014; Horák-Terra et 70 al., 2014; Silva-Sánchez et al., 2015). As a result of long-term processes, the degree of peat decomposition/mass loss, indicated by C/N ratios, show a rapid increased with depth in the upper 50 cm and 390 cm in Pinheiros and Rano Aroi (respectively) and almost stable values below (Supporting Information, Figure S4, Manuscript II and Figure 3, Manuscript I). This may indicate that the main decay of organic matter by decomposition - before its stabilization - was reached at 10-11 cal kyr BP in Pinheiros and 5 cal kyr BP in Rano Aroi. Changes in C/N ratio in older sections should respond to other factors than long-term changes in peat organic matter. In Pinheiros, the dominant sapric (i.e. highly decomposed) nature of the studied peat core is also indicated by the general low signal of the polysaccharide components (1030–1080 cm-1; FTIR data, Manuscript III), which is somewhat unusual for spectroscopic fingerprints of peat, but agrees with the Pleistocene age of most samples (only the 50 cm from 240 cm correspond to Holocene peat). However, the large abundance of quartz (main absorbance at 1084 cm-1) masked the O-H stretching from the polysaccharides signal, so no other FTIR based humification indices could be used to compare with the C/N ratio. Due to the stable values of long-term decomposition in the Pleistocene sections of both cores (Pinheiros and Rano Aroi), mass loss associated with organic matter peat decomposition can be ruled out as a major factor of Hg accumulation. In the case of Pinheiros, this was also confirmed by the information obtained through prediction models - base on geochemical (PCR-model) and FTIR data (PLSmodel). In these models, the regression coefficients, which are a measure of the relative weight of each factor involved in Hg content for the whole record, indicate that peat decomposition is the least important one (Figure 4 for Manuscript II and Figure 4 Manuscript III). Taking into account the chronology, the role of peat decomposition on Hg concentrations is limited to the Holocene section and almost negligible in peats with ages older than 10-11 cal kyr BP (Manuscript II). There is no correlation between the decline in organic matter decomposition and Hg concentration in peat sections younger than 5 cal kyr BP in Rano Aroi (Figure 5, Manuscript I). Mercury concentrations remain low and, as previously mention, the high peak at 4.7 cal kyr BP is probably related to volcanic activity. The cause of two small increases at 1.2 cal kyr BP and at the superficial sample is unknown, but the effect of peat decomposition is unlikely. Only the clear 5. General discussion 71 increase in Hg concentration at ~40 cal kyr BP may have responded to increased decomposition in Rano Aroi. An extreme oxidation event, as a consequence of a dry phase occurred between ~40-42 cal kyr BP (Margalef et al., 2013), resulted in a 2 fold increase in Hg concentrations (compared to the lower values of the peat core). The rates of peat accumulation, and the decomposition of organic matter at Sandhavn, were limited in the period ~AD 1400-1800 (Silva-Sánchez et al., 2015) – encompassing much of the Little Ice Age (LIA) – at least when compared with peatlands from mid-latitudes. Although this might have affected Hg accumulation, there is no apparent change and values remained low (peat background values). The increase in Hg accumulation rate just before AD 1800’s is in agreement with other records from Greenland (see Manuscript VII) and associated with to atmospheric pollution. The comparison of these findings with those from midlatitudes records that have been more affected by enrichments in Hg, driven by intense peat mineralization at short term phases (i.e. Martínez Cortizas et al., 2007), supports the suitability of cold high latitude environments as sensitive Hg peat archives, with reduced post-depositional processes as peat organic matter decomposition. There is not much information about organic mater and its role on Hg accumulation in Limnopolar Lake. The lake is ultraoligotrophic and the lake bottom is covered by a patchy carpet of the moss Drepanocladus longifolius (Mitt.) Broth ex Paris (Toro et al., 2013), which constitutes the main organic matter source. Thermo-desorption analysis showed variable temperature ranges of Hg release (Figure 3, Manuscript VI), which indicates differences in Hg binding strength. Similar Hg release characteristic have been observed in other sediments (Biester et al., 2000) where the peak at low temperatures (200ºC in Limnopolar) is thought to be due to relatively weakly bound Hg-OM (here mainly mosses), while with increasing degradation of the organic matter and decrease of the redox potential it appears to be bound to organo-sulfides or occurs as metacinnabar (HgS). The observation that sulphide-bound Hg appears only in the deeper/older sediments indicates that organic matter turnover is extremely slow in this lake due to the low temperatures. 72 5.5. The effect of the processes occurring in the catchment area Both direct aerial deposition of Hg on the lake surface and remobilization of atmospheric inputs through fluxes from the catchment are to be, necessarily, considered in Hg estimations (Schroeder and Munthe, 1998). A huge amount of pollutants have been deposited and stored in the catchment surface soils of many lake sites, mainly due to the industrial emissions during the last centuries/ decades. Thus, processes as watershed erosion appear to be the main vectors of terrigenous Hg fluxes toward the lake water column (Ouellet et al., 2009) and processes related with catchment land uses (e.g., agricultural, residential) have been proved to influence Hg concentrations in lake sediments (e.g., Drevnick et al., 2016). Nowadays, the release of pollutants stored in the soils has become one of the dominant sources for lakes and, in many cases, lake sediments do not show a decline or the decline is not as large as expected according to reductions in emission (e.g., Yang and Smyntek, 2014; Yang, 2015; Drevnick et al., 2016). Similarly to lakes, Hg concentrations in peatlands receiving water inputs from sources other than precipitation, usually as runoff from upslope areas through drainage from surrounding mineral soils, could be affected by catchment soil erosion. Four out of five Hg records studied in this PhD investigation (Rano Aroi, Pinheiros, Lago Hambre and Limnopolar), have been influenced by some kind of effect related to catchment soil erosion. Pinheiros mire is located on a quartzitic valley in Serra do Espinaço Meridional (Figure 1 and S1 in Manuscript II). Despite it is situated at the mountain summit the upper part of the small catchment area had a profound influence on the dynamics of the mire. Previous research on morphological, physical, chemical and elemental properties of five mountain mires of the area found that erosion of the catchment soils has played a main role on the genesis and evolution of the peatlands (HorákTerra et al., 2014). The authors found that fluxes of mineral mater (dominated by quartz and indicated by increases in Si) from the catchment resulted in a dilution of the organic matter content (Horák-Terra et al., 2014). The second component (Cp2) of the PCA performed by us using only data from Pinheiros, showed the same relation between elements (silicon vs. organic matter), pointing to the same process. The regression analysis suggests that catchment erosion had the largest (negative) effect on Hg content (Manuscript II). This is because quartzite is the dominant lithological material in the catchment, and it has negligible Hg contents. 5. General discussion 73 Thus, mineral matter fluxes from the catchment produced a double dilution effect: on organic matter to which Hg is bound and by inputs of low Hg-containing mineral matter. However, from the methodological point of view these results have another implication for the interpretation of Hg accumulation rates and other estimations obtained using peat bulk density (a property that in this case is controlled by the input of quartz) (see Table 1, Factor loadings, Manuscript II). Mercury accumulation rates are calculated using peat bulk density, and thus, they depend upon both local fluxes of mineral matter and degree of peat decomposition/mass loss. Increased fluxes of mineral matter from the catchment (which provide almost no Hg to the peat) increase the bulk density and consequently the estimated mass accumulation rates. The dilution effect on Hg concentrations due to the addition of allochthonous mineral matter with low Hg content does not compensate for the apparent increase in Hg accumulation rates due to the higher bulk density. This leads to a situation where, for example, the maximum Hg concentration in Pinheiros within the core (370 ng g-1 at 74 cm) has an estimated accumulation rate of 9.9 μg m-2 yr-1, which is similar to that of sample at 218 cm (9.6 μg m-2 yr-1) with a concentration 4.7-fold lower (79 ng g-1) but a bulk density 4.5-fold higher. Thus, the analysis of the processes affecting Hg contents on Pinheiros core was done on the concentrations and not on the accumulation rates. Rano Aroi is located in an ancient Pleistocene volcano crater and the crater slopes form its small catchment area (0.15 km2). From previous investigations, it is known that during high rainfall events there was an increase in soil erosion and enhanced fluxes of mineral matter from the small basin into the peatland (Margalef et al., 2013, 2014). Although a low release of Hg can occur from the catchment materials due to weathering (see discussion in Manuscript I), metal-hummus complexes found in (andic) volcanic soils (García-Rodeja et al., 2004), could bound Hg deposited from the atmospheric pool (Nóvoa-Muñoz et al., 2008; Peña-Rodríguez et al., 2012). Thus, mobilization of Hg due to soil erosion could have also taken place in Rano Aroi as Hg-humus complexes that were transported to the mire by superficial runoff (Manuscript I). Increases in Fe content during high rainfall periods (Margalef et al., 2013, 2014) and a higher Hg content during the former supports the effect of this mechanism. The comparison of Hg concentration with 80 higher than the expected value by the model) was found at the timing of Heinrich event 1 (H1), which was characterized by dry and very cold conditions (Lowell et al., 1995; Sagnotti et al., 2001). No clear reason was found for this peak, although the effect of multiple factors is the most likely explanation (Manuscript II). In Rano Aroi (Margalef et al., 2013, 2014), the drier-colder climatic conditions extended from ~71 to ~55 krys BP, when a transition started ending at ~43 cal kyr BP. After ~43 cal kyr BP the predominance of C3 vegetation indicates a more humid climate (Margalef et al., 2013, 2014). Similarly to Pinheiros, the drier phase was punctuated by negative excursions in δ13C that were interpreted as short wetter events, but the effect on Hg concentrations is completely different. Increases in Hg concentration seem to match – at least partially – with the wet events, suggesting enhanced Hg-wet deposition over the peatland and the catchment (see discussion section in Manuscript I). Additionally, during these events there was an increase in catchment soil erosion, driven by precipitation. In contrast to Pinheiros, in Rano Aroi the increased input of mineral fluxes from the catchment soils resulted in increases of Hg concentrations in peat, most probably due to inputs of metal (Hg)- humus complexes, which are typically abundant in volcanic soils (see the effect of the catchment). Between ~40 - 42 cal kyr BP a long-term drought resulted in increased peat oxidation that led to an increase in Hg concentrations (see the effect of organic matter decomposition). With the exception of the abrupt rainfall events, under prevalent humid conditions (from ~42 cal kyr BP to the present) there was not a general increase in Hg wet-deposition since the background values continued to be low. The ~20 cal krys BP Hg maximum corresponds to a cold phase occurred at the end of the Last Glacial Maximum, when humid conditions in Eastern Island prevailed (Margalef et al., 2014). Both, colder and humid conditions would have favored Hg accumulation in Rano Aroi since Hg deposition is controlled by temperature and humidity variations (Martínez Cortizas et al., 1999; Corella et al., 2017). The results of these long-term records (Rano Aroi and Pinheiros) suggest, as it is generally accepted by extensive monitoring (Gay et al., 2013), that wet deposition (i.e. the wash out of Hg by rainfall) is the main source of Hg to peatlands. In contrast to our results, recent short term studies (Enrico et al., 2016) performed in the French Pyrenees have found that gaseous and particulate Hg (i.e. dry deposition) 5. General discussion 81 are the main species deposited into peatlands. This apparent inconsistency most probably arises from the different time scales of the studies. The long temporal perspective provided by the Pleistocene records might highlight the dominance of the longer-term wet deposition processes. In Lago Hambre, total solar irradiance directly controlled primary productivity in the lake and indirectly the Hg content in the sediments. During the last ~4.5 cal kyr BP higher TSI corresponded to higher green algae production and Hg uptake by them, probably accompanied by increased methylation (Manuscript V). Additionally, periods of high TSI (and productivity) largely corresponded to periods of low mineral matter fluxes from the catchment (Figure 2, Manuscript IV), indicating relatively dry conditions due to weaker westerly winds at Lago Hambre. Our results confirm that there were not important inputs of Hg from the catchment during periods of high productivity and higher uptake by algae (Manuscript V). The freezing and thawing lake dynamics regulated the accumulation of Hg in Limnopolar Lake and its catchment, a process also directly controlled by climate. Annually, Limnopolar Lake is ice covered except for 2–3 months during the summer. The ice break-up in the lake has been observed to start in December near the lake outlets or inlets, however ice blocks or thin ice layers could persist until February (Toro et al., 2007) (see pictures in Figure 9.3 in Camacho et al., 2014). The ice-snow layer over the lake and the catchment may have acted as a Hg sink during 9 - 10 months per year. Periods of sustained general cold climatic conditions may have extended the effect of this Hg trap from years to decades. This is consistent with high Hg concentrations and accumulation rates corresponding to phases of minima in insolation (colder periods). The comparison of the TSI record (Steinhilber et al., 2009) (Figure 4, Manuscript VI), with Hg accumulation shows that six of eight Hg peaks (>270 µg m-2 yr -1) coincided with periods of low irradiance. Peaks at AD ~1300, ~1450 and ~1715, ~1820 correspond with the Wolf, Spörer, Maunder and Dalton minima; and peaks at AD ~550 and ~1750 coincided with relative low insolation. Even more, without considering the samples with extremely high accumulation (only those with <270 µg m-2 yr-1), the record also shows increases of Hg accumulation during periods of low irradiance (Figure 5, Manuscript VI). As indicated above, under such cold conditions it is likely that the lake and its basin remained snow/ice-covered for long periods of time, enhancing Hg accumulation in the snowpack. During thawing, the more material stored in 82 the ice the more transferred to the lake and its sediments. Finally, the lake snow/ ice-cover also promotes reducing conditions in the sediment and the longer the period is the more intense the reducing conditions will be. This may have also had an effect on the heterotrophic degradation of the organic matter in the lake, which should interact with Hg (Manuscript VI). 83 6. Conclusions Lake sediments and peatlands are environmental archives that can be used as records of atmospherically deposited Hg and thus providing a “picture” of the Hg cycle. However, this use - as Hg atmospheric records - is conditioned by factors which operate in a geographic and time scale. Thus, the Hg determined in the samples does not directly reflect the atmospheric deposition, but result of the factors effects. The sources and processes that affect Hg deposition and accumulation in the records varied depending on the location and the period of time studied. Additionally their final effect on the concentration of Hg might change along the evolution of the record or due to the interaction with other factors. Due to the long-residence time in the atmosphere, Hg sources could have a local, regional of global effects and in some cases is not simple determined the geographic scale of the main source. Sandhavn provided a good example of this issue. It is located in geographic position susceptible of North America and European Hg pollution sources, and the chronology of Hg alone does not allow determine which one is the main source are, despite the timing of Industrial trends in both areas are well-known and quite different. According to the lead isotopic results it would be expected the main source of Hg contamination to be also from US, but it does not prove it and hemispheric mixing and long-range transport might be driving. 84 In other cases, the effect of a local process may overlap regional or global Hg signals, such as the volcanic activity of Deception Island as a local Hg source over Limnopolar Lake sediments. The dynamics of freezing and thawing in the lake and its catchment show a high sensitivity by changes in insolation, probably because of its extreme polar location. Both facts limits information on long-term Hg transport to remote polar areas but constitutes a valuable model of local or regional environments. Similarly, the control that the aquatic productivity exerted on Hg accumulation in sediments on Lago Hambre, restricts its use as a reliable record of atmospheric Hg deposition but nevertheless it helps to understand the role of productive and eutrophic lakes in the Hg cycle. Specific characteristics of the records could modify also the effect of regional processes. While rainfall events dominate the Hg wet deposition in Rano Aroi record and amplified its effect through the erosion of the volcanic soils from the catchment; in Pinheiros they produce a strong dilution of Hg concentrations due the input of mineral matter from quartzitic catchment. The time scale covered by a record also determines the factors that may affect it, and therefore the information that it provides as an archive of atmospheric Hg deposition. In records covering long periods of time, such as Pinheiros and Rano Aroi (Late Pleistocene and Holocene) there is a wide range of processes that could be observed and that could modify Hg record. There is also a strong interaction between the processes and the system itself. Thus, in Pinheiros the results of rainfall precipitations under dry conditions, low peatland developed and scarce catchment vegetation is an increased the input of inorganic matter from the catchment producing the mentioned dilution effect on Hg concentrations. With a greater development of peatland and more vegetation covering the basin (during a prolonged humid period), the effect of the precipitations produces an increase of the concentration of Hg, linked to the wet deposition. The relative weight of the factors on Hg concentration also changes with the time scale. Long-term records studied here (Rano Aroi and Pinheiros) suggest that wet deposition (i.e. the wash out of Hg by rainfall) is the main source of Hg to peatlands, in contrast to recent short-term studies that indicate that gaseous and particulate Hg (i.e. dry deposition) are the main species deposited into peatlands. This apparent inconsistency most probably arises from the different time scales 6. Conclusions 85 of the studies. The long temporal perspective provided by the Pleistocene records might highlight the dominance of the longer-term wet deposition processes. Similarly, processes as organic matter decomposition that main driver of Hg content in Holocene reconstructions, seems to have a low overall weight at Pleistocene scale. Notwithstanding strong oxidative events caused by droughts periods may increase Hg concentrations up to 2 times. By the other hand, processes as AMDEs that has been shown a mainly role driving Hg atmospheric concentrations in annual or monthly records, its role a long term scales using records that account hundreds or thousands years is difficult to show and demonstrate. Essentially, peat and lake sediments are potential records of atmospheric Hg deposition. The recording of these depositions is highly conditioned by the local, regional, and global characteristics of archives, sources, and control factors. This is due to the dynamic behaviour of Hg in the atmosphere and in the environment. From the methodological point of view, the multiproxy approach has been shown a useful tool to obtain a complete interpretation of the systems and separates the different sources and factors that affect the Hg content in the record and thus, be able to obtain an accurate reconstruction of Hg atmospheric deposition. In cases of pollution sources it seems recommended the comparison of Hg record with other metals, which would help to establish the possible scale of pollution sources (local, regional or global). Lead Isotopic studies prove to be an invaluable support to determined the most probable dominant Hg source. Multivariate statistics (PCA, PCR and PLS) proved to be a very useful tool to complement multirpoxy approach. Specifically, the principal components analysis followed by principal components regression enabled us to determine the evolution of the weight of the latent processes governing the accumulation of Hg through time in Pinheiros record. This novelty approximation seems more suitable for long records and with a regular age-depth model. Spectroscopic data in combination with multivariate analysis, specifically PLS statistics can efficiently be use to predict Hg concentrations in minerogenic peat, at least within the same core. Although the model could be very precise in the prediction of Hg concentration, it might not reproduce the changes with 86 time/depth of the relative weights of the driving processes. Supplementing models with extra data (like δ13C ratios in our case) may be required to use the model for the interpretation of the underlying processes. Although more research is needed on different types of peatlands, this methodology (MIR-spectroscopy combined with PLS) may enable to reduce the cost of multicore approaches to study the spatial variability within mire or between mires (of a similar type and area) in Hg accumulation, and probably also other peat properties. 87 7. References Allan M., Le Roux G., Sonke J.E., Piotrowska N., Streel M., and Fagel N., 2013. Reconstructing historical atmospheric mercury deposition in Western Europe using: Misten peat bog cores, Belgium. Science of The Total Environment 442, 290–301 AMAP A., 2011. Assessment 2011: mercury in the Arctic (Oslao, Norway) AMAP/UNEP, 2013. Technical Background Report for the Global Mercury Assessment (Oslo, Norway) Amos H.M., Jacob D.J., Streets D.G., and Sunderland E.M., 2013. Legacy impacts of alltime anthropogenic emissions on the global mercury cycle. Global Biogeochemical Cycles 27, 410–421 Appelquist H., Ottar Jensen K., Sevel T., and Hammer C., 1978. Mercury in the Greenland ice sheet. Nature 273, 657–659 Ariya P.A., Khalizov A., and Gidas A., 2002. Reactions of Gaseous Mercury with Atomic and Molecular Halogens: Kinetics, Product Studies, and Atmospheric Implications. The Journal of Physical Chemistry A 106, 7310–7320 Ariya P.A., Dastoor A.P., Amyot M., Schroeder W.H., Barrie L., Anlauf K., Raofie F., Ryzhkov A., Davignon D., Lalonde J., et al., 2004. The Arctic: A sink for mercury. Tellus, Series B: Chemical and Physical Meteorology 56, 397–403 Asplund G. and Grimvall A., 1991. Organohalogens in nature. Environmental Science and Technology 25, 1346–1350 Aston S.R., Bruty D., Chester R., and Padgham R.C., 1973. Mercury in lake sediments: a possible indicator of Technological Growth. Nature 241, 450–451 Atwell L., Hobson K.A., and Welch H.E., 1998. Biomagnification and bioaccumulation of mercury in an arctic marine food web: insights from stable nitrogen isotope analysis. Canadian Journal of Fisheries and Aquatic Sciences 55, 1114–1121 Bagnato E., Aiuppa A., Parello F., Calabrese S., D’Alessandro W., Mather T.A., McGonigle 88 A.J.S., Pyle D.M., and Wängberg I., 2007. Degassing of gaseous (elemental and reactive) and particulate mercury from Mount Etna volcano (Southern Italy). Atmospheric Environment 41, 7377–7388 Baker P.E., Buckley F., and Holland J.G., 1974. Petrology and geochemistry of Easter Island. Contributions to Mineralogy and Petrology 44, 85–100 Bao K., Shen J., Wang G., Sapkota A., and McLaughlin N., 2016. Estimates of recent Hg pollution in Northeast China using peat profiles from Great Hinggan Mountains. Environmental Earth Sciences 75 Bargagli R., 2008. Environmental contamination in Antarctic ecosystems. Science of the Total Environment 400, 212–226 Bargagli R., Monaci F., Sanchez-Hernandez J.C., and Cateni D., 1998. Biomagnification of mercury in an Antarctic marine coastal food web. Marine Ecology Progress Series 169, 65–76 Bartolini S., Geyer A., Martí J., Pedrazzi D., and Aguirre-Díaz G., 2014. Volcanic hazard on Deception Island (South Shetland Islands, Antarctica). Journal of Volcanology and Geothermal Research 285, 150–168 Bash J.O. and Miller D.R., 2007. A note on elevated total gaseous mercury concentrations downwind from an agriculture field during tilling. Science of the Total Environment 388, 379–388 Beal S., Osterberg E.C., Zdanowicz C., and Fisher D., 2015. An ice core perspective on mercury pollution during the past 600 years. Environmental Science and Technology 49, 7641–7647 Benoit J.M., Fitzgerald W.F., and Damman A.W.H., 1994. Historical atmospheric mercury deposition in the mid-continental US as recorded in an ombrotrophic peat bog. Mercury Pollution: Integration and Synthesis 187–202 Benoit J.M., Fitzgerald W.F., and Damman A.W.H., 1998. The biogeochemistry of an ombrotrophic bog: Evaluation of use as an archive of atmospheric mercury deposition. Environmental Research 78, 118–133 Benoit J.M., Gilmour C.C., Heyes A., Mason R.P., and Miller C.L., 2002. Geochemical and Biological Controls over Methylmercury Production and Degradation in Aquatic Ecosystems. In Biogeochemistry of Environmentally Important Trace Elements, (American Chemical Society), pp. 19–262 Biester H., Gosar M., and Covelli S., 2000. Mercury Speciation in Sediments Affected by Dumped Mining Residues in the Drainage Area of the Idrija Mercury Mine, Slovenia. Environmental Science and Technology 34, 3330–3336 Biester H., Kilian R., Franzen C., Woda C., Mangini A., and Schöler H.F., 2002. Elevated mercury accumulation in a peat bog of the Magellanic Moorlands, Chile (53°S) - An anthropogenic signal from the Southern Hemisphere. Earth and Planetary Science Letters 201, 609–620 Biester H., Martinez-Cortizas A., Birkenstock S., and Kilian R., 2003. Effect of Peat Decomposition and Mass Loss on Historic Mercury Records in Peat Bogs from Patagonia. Environmental Science and Technology 37, 32–39 Biester H., Keppler F., Putschew A., Martinez-Cortizas A., and Petri M., 2004. Halogen Retention, Organohalogens, and the Role of Organic Matter Decomposition on Halogen Enrichment in Two Chilean Peat Bogs. Environmental Science and Technology 38, 1984–1991 7. References 89 Biester H., Bindler R., and Martínez Cortizas A., 2006. Mercury in mires. In Peatlands: Evolution and Records of Environmental and Climatic Changes, I. Martini, A. Martínez Cortizas, and W. Chesworth, eds. (Elsevier B.V), pp. 465–478 Biester H., Bindler R., Martinez-Cortizas A., and Engstrom D.R., 2007. Modeling the Past Atmospheric Deposition of Mercury Using Natural Archives. Environmental Science and Technology 41, 4851–4860 Biester H., Hermanns Y.-M., and Martinez Cortizas A., 2012. The influence of organic matter decay on the distribution of major and trace elements in ombrotrophic mires – a case study from the Harz Mountains. Geochimica et Cosmochimica Acta 84, 126–136 Biester H., Knorr K.-H., Schellekens J., Basler A., and Hermanns Y.-M., 2014. Comparison of different methods to determine the degree of peat decomposition in peat bogs. Biogeosciences 11, 2691–2707 Bindler R., 2003. Estimating the natural background atmospheric deposition rate of mercury utilizing ombrotrophic bogs in Southern Sweden. Environmental Science and Technology 37, 40–46 Bindler R., Renberg I., Appleby P.G., Anderson N.J., and Rose N.L., 2001a. Mercury Accumulation Rates and Spatial Patterns in Lake Sediments from West Greenland: A Coast to Ice Margin Transect. Environmental Science and Technology 35, 1736– 1741 Bindler R., Olofsson C., Renberg I., and Frech W., 2001b. Temporal trends in mercury accumulation in lake sediments in Sweden. Water, Air, and Soil Pollution 1, 343– 355 Bindler R., Klarqvist M., Klaminder J., and Förster J., 2004. Does within-bog spatial variability of mercury and lead constrain reconstructions of absolute deposition rates from single peat records? The example of Store Mosse, Sweden. Global Biogeochemical Cycles 18, 1–12 Blais J.M., Rosen M.R., and Smol J.P., 2015. Using natural archives to track sources and long-term trends of pollution: an introduction. In Environmental Contaminants. Using Natural Archives to Track Sources and Long-Term Trends of Pollution, J.M. Blais, M.R. Rosen, and J.P. Smol, eds. (Springer), pp. 1–3 Boutron C.F., Vandal G.M., Fitzgerald W.F., and Ferrari P., 1998. A forty year record of mercury in central Greenland snow in snow deposited Greenland reported from previous studies of Greenland snow by major Combined probably plagued by major contamination problems during estimated contributions from natural Hg sources. Goephysical Research Letters 25, 3315–3318 Brown A.S., Brown R.J.C., Corns W.T., and Stockwell P.B., 2008. Establishing SI traceability for measurements of mercury vapour. The Analyst 133, 946– 953 Brunke E.-G., Labuschagne C., Ebinghaus R., Kock H.H., and Slemr F., 2009. Total gaseous mercury depletion events observed at Cape Point during 2007andndash;2008. Atmospheric Chemistry and Physics Discussions 9, 20979–21009 Calvert J.G. and Lindberg S.E., 2003. A modeling study of the mechanism of the halogenozone-mercury homogeneous reactions in the troposphere during the polar spring. Atmospheric Environment 37, 4467–4481 Camacho A., Villaescusa J.A., Rochera C., and Jørgensen S.E., 2014. Chapter 9 - Modeling the Response of the Planktonic Microbial Community to Warming Effects 96 Lowell T. V, Heusser C.J., Andersen B.G., Moreno P.I., Hauser A., Heusser L.E., Schlüchter C.H., Marchant D.R., and Denton G.H., 1995. Interhemispheric correlation of late Pleistocene glacial events. Science 269, 1541–1549 Macdonald R.W., Harner T., and Fyfe J., 2005. Recent climate change in the Arctic and its impact on contaminant pathways and interpretation of temporal trend data. Science of the Total Environment 342, 5–86 Malm O., 1998. Gold mining as a source of mercury exposure in the Brazilian Amazon. Environmental Research 77, 73–78 Margalef O., 2014. The last 70 ky of Rano Aroi (Easter Island, 27 ° S) peat record: New insights for the Central Pacific paleoclimatology. Universitat de Barcelona Margalef O., Cañellas-Boltà N., Pla-Rabes S., Giralt S., Pueyo J.J., Joosten H., Rull V., Buchaca T., Hernández A., Valero-Garcés B.L., et al., 2013. A 70,000 year multiproxy record of climatic and environmental change from Rano Aroi peatland (Easter Island). Global and Planetary Change 108, 72–84 Margalef O., Martínez Cortizas A., Kylander M., Pla-Rabes S., Cañellas-Boltà N., Pueyo J.J., Sáez A., Valero-Garcés B.L., and Giralt S., 2014. Environmental processes in Rano Aroi (Easter Island) peat geochemistry forced by climate variability during the last 70kyr. Palaeogeography, Palaeoclimatology, Palaeoecology 414, 438– 450 Martínez Cortizas A., 2000. Archivos geoquímicos para la reconstrucción de los Paleoambientes cuaternarios: Ideas y Ejemplos. In II Jornadas Do Quaternário Da APEQ, (Porto), pp. 12–15 Martínez Cortizas A., Pontevedra-Pombal X., García-Rodeja E., Nóvoa-Muñoz J.C., and Shotyk W., 1999. Mercury in a Spanish Peat Bog: Archive of Climate Change and Atmospheric Metal Deposition. Science 284, 939–942 Martínez Cortizas A., Biester H., Mighall T., and Bindler R., 2007. Climate-driven enrichment of pollutants in peatlands. Biogeosciences 4, 905–911 Martínez Cortizas A., Peiteado Varela E., Bindler R., Biester H., and Cheburkin A., 2012. Reconstructing historical Pb and Hg pollution in NW Spain using multiple cores from the Chao de Lamoso bog (Xistral Mountains). Geochimica et Cosmochimica Acta 82, 68–78 Martínez Cortizas A., Rozas Muñiz I., Taboada T., Toro M., Granados I., Giralt S., PlaRabés S., Cortizas A.M., Muñiz I.R., Taboada T., et al., 2014. Factors controlling the geochemical composition of Limnopolar Lake sediments (Byers Peninsula, Livingston Island, South Shetland Island, Antarctica) during the last ca. 1600 years. Solid Earth 5, 651–663 Mason R.P. and Sheu G.R., 2002. Role of the ocean in the global mercury cycle. Global Biogeochemical Cycles 16, 1093 Mason R., Fitzgerald W., and Morel F., 1994. The biogeochemical cycling of elemental mercury: Anthropogenic influences. Geochimica et Cosmochimica Acta 58, 3191– 3198 Mason R.P., Choi A.L., Fitzgerald W.F., Hammerschmidt C.R., Lamborg C.H., Soerensen A.L., and Sunderland E.M., 2012. Mercury biogeochemical cycling in the ocean and policy implications. Environmental Research 119, 101–117 Massa C., Monna F., Bichet V., Gauthier É., Losno R., and Richard H., 2015. Inverse modeling of past lead atmospheric deposition in South Greenland. Atmospheric 7. References 97 Environment 105, 121–129 Mast M.A., Manthorne D.J., and Roth D.A., 2010. Historical deposition of mercury and selected trace elements to high-elevation National Parks in the Western US inferred from lake-sediment cores. Atmospheric Environment 44, 2577–2586 Matty J.M. and Long D.T., 1995. Early diagenesis of mercury in the Laurentian Great Lakes. Journal Great Lakes Research 21, 574–586 McCulloch R.D. and Davies S.J., 2001. Late-glacial and Holocene palaeoenvironmental change in the central Strait of Magellan, southern Patagonia. Palaeogeography, Palaeoclimatology, Palaeoecology 173, 143–173 McKinney M.A., Pedro S., Dietz R., Sonne C., Fisk A.T., Roy D., Jenssen B.M., and Letcher R.J., 2015. A review of ecological impacts of global climate change on persistent organic pollutant and mercury pathways and exposures in arctic marine ecosystems. Current Zoology 61, 617–628 Meyers P. and Lallier-Vergès E., 1999. Lacustrine records of changes in Late Quaternary continental environments and climates: an overview of sedimentary organic matter indicators. Journal of PalaeoLimnology 21, 345–372 Minamata Convention on Mercury web (UNEP) Morel F.M.M., Kraepiel A.M.L., and Amyot M., 1998. the Chemical Cycle and Bioaccumulation of Mercury. Annual Review of Ecology and Systematics 29, 543– 566 Morello F., Borrero L., Massone M., Stern C., Garcia-Herbst A., McCulloch R., ArroyoKalin M., Calas E., Torres J., Prieto A., et al., 2012. Hunter-gatherers, biogeographic barriers and the development of human settlement in Tierra del Fuego. Antiquity 86, 71–87 Muir D.C.G., Wagemann R., Hargrave B.T., Thomas D.J., Peakall D.B., and Norstrom R.J., 1992. Arctic marine ecosystem contamination. Science of The Total Environment 122, 75–134 Nafth D., 1993. Ice-core records of the chemical quality of atmospheric deposition and climate from mid-latitude glaciers, Wind River Range, Wyoming. Colorado Schoold of Mines, Golden, Colorado National Research Council, 2000. Toxicological effects of methylmercury Nerentorp Mastromonaco M., Gårdfeldt K., Jourdain B., Abrahamsson K., Granfors A., Ahnoff M., Dommergue A., Méjean G., and Jacobi H.W., 2016. Antarctic winter mercury and ozone depletion events over sea ice. Atmospheric Environment 129, 125–132 Nichols A.L., 1997. Lead in Gasoline. In Economic Analyses at EPA: Assessing Regulatory Impact, R.D. Morgenstern, ed. (Washington DC: Resources for the Future), pp. 49–86 Norton S.A., Evans G.C., and Kahl J.S., 1997. Comparison of Hg and Pb fluxes to Hummocks and Hollows of ombrotrophic Big Heath Bog and to Nearby Sargent Mt. Pond, Maine, USA. Water, Air, and Soil Pollution 100, 271–286 Nóvoa-Muñoz J.C., Pontevedra-Pombal X., Martínez-Cortizas A., and García-Rodeja Gayoso E., 2008. Mercury accumulation in upland acid forest ecosystems nearby a coal-fired power-plant in Southwest Europe (Galicia, NW Spain). Science of The Total Environment 394, 303–312 Nriagu J. and Becker C., 2003. Volcanic emissions of mercury to the atmosphere: Global 98 and regional inventories. Science of the Total Environment 304, 3–12 Obrist D., Tas E., Peleg M., Matveev V., Faïn X., Asaf D., and Luria M., 2011. Bromineinduced oxidation of mercury in the mid-latitude atmosphere. Nature Geoscience 4, 22–26 Ouellet J.F., Lucotte M., Teisserenc R., Paquet S., and Canuel R., 2009. Lignin biomarkers as tracers of mercury sources in lakes water column. Biogeochemistry 94, 123– 140 Outridge P.M. and Sanei H., 2010. Does organic matter degradation affect the reconstruction of pre-industrial atmospheric mercury deposition rates from peat cores? — A test of the hypothesis using a permafrost peat deposit in northern Canada. International Journal of Coal Geology 83, 73–81 Outridge P.M., Sanei H., Stern G. A., Hamilton P.B., and Goodarzi F., 2007. Evidence for control of mercury accumulation rates in Canadian High Arctic Lake sediments by variations of aquatic primary productivity. Environmental Science and Technology 41, 5259–5265 Outridge P.M., Rausch N., Percival J.B., Shotyk W., and McNeely R., 2011. Comparison of mercury and zinc profiles in peat and lake sediment archives with historical changes in emissions from the Flin Flon metal smelter, Manitoba, Canada. The Science of the Total Environment 409, 548–563 Pacyna E.G., Pacyna J.M., Steenhuisen F., and Wilson S., 2006. Global anthropogenic mercury emission inventory for 2000. Atmospheric Environment 40, 4048– 4063 Parish F., Sirin A., Charman D., Joosten H., Minayeva T., Silvius M., and Stringer L., 2008. Assessment on Peatlands, Biodiversity and Climate Change: Main Report (Kuala Lumpur and Wetlands International, Wageningen) Peleg M., Matveev V., Tas E., Luria M., Valente R.J., and Obrist D., 2007. Mercury depletion events in the troposphere in mid-latitudes at the Dead Sea, Israel. Environmental Science and Technology 41, 7280–7285 Peña-Rodríguez S., Pontevedra-Pombal X., Fernández-Calviño D., Taboada T., AriasEstévez M., Martínez-Cortizas A., Nóvoa-Muñoz J.C., and García-Rodeja E., 2012. Mercury content in volcanic soils across Europe and its relationship with soil properties. Journal of Soils and Sediments 12, 542–555 Petit J.-R., Briat M., and Royer A., 1981. Ice age aerosol content from East Antarctic ice core samples and past widn strngth. Nature 293, 391–394 Petit J.R., Jouzel J., Raynaud D., Barkov N.I., Barnola J.-M., Basile I., Bender M., Chappellaz J., Davis M., Delaygue G., et al., 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 429– 436 Pheiffer-Madsen P., 1981. Peat bog records of atmospheric mercury deposition. Nature 293, 127–130 Pirrone N., Cinnirella S., Feng X., Finkelman R.B., Friedli H.R., Leaner J., Mason R., Mukherjee A.B., Stracher G.B., Streets D.G., et al., 2010. Global mercury emissions to the atmosphere from anthropogenic and natural sources. Atmospheric Chemistry and Physics 10, 5951–5964 R CoreTeam, 2014. R: A Language and Environment for Statistical Computing. Rada R.G., Wiener J.G., Winfrey M.R., and Powel D.E., 1989. Recent increases in 7. References 99 atmospheric deposition of mercury to North-Central Wisconsin Lakes Inferred From Sediment Analyses. Archives of Environmental Contamination and Toxicology 18, 175–181 Rankin A.M., Wolff E.W., and Martin S., 2002. Frost flowers: Implications for tropospheric chemistry and ice core interpretation. Journal of Geophysical Research Atmospheres 107 Ribeiro Guevara S., Meili M., Rizzo A., Daga R., and Arribére M., 2010. Sediment records of highly variable mercury inputs to mountain lakes in Patagonia during the past millennium. Atmospheric Chemistry and Physics 10, 3443–3453 Rizzo A., Arcagni M., Campbell L., Koron N., Pavlin M., Arribére M.A., Horvat M., and Ribeiro Guevara S., 2014. Source and trophic transfer of mercury in plankton from an ultraoligotrophic lacustrine system (Lake Nahuel Huapi, North Patagonia). Ecotoxicology 23, 1184–1194 Roos-Barraclough F., Martinez-Cortizas A., Garcıa-Rodeja E., Shotyk W., García-Rodeja E., and Shotyk W., 2002. A 14 500 year record of the accumulation of atmospheric mercury in peat: volcanic signals, anthropogenic influences and a correlation to bromine accumulation. Earth and Planetary Science Letters 202, 435–451 Roos-Barraclough F., Givelet N., Cheburkin A.K., Shotyk W., and Norton S.A., 2006. Use of Br and Se in peat to reconstruct the natural and anthropogenic fluxes of atmospheric Hg: A 10000-year record from Caribou Bog, Maine. Environmental Science and Technology 40, 3188–3194 Rosman K.J.R., Chisholm W., Boutron C.F., Candelone J.P., and Hong S., 1994. Isotopic evidence to account for changes in the concentration of lead in Greenland snow between 1960 and 1988. Geochimica et Cosmochimica Acta 58, 3265– 3269 Rossmann R., 2010. Protocol to Reconstruct Historical Contaminant Loading to Large Lakes : The Lake Michigan Sediment Record of Mercury. Environmental Science and Technology 44, 935–940 Rull V., Cañellas-Boltà N., Margalef O., Sáez A., Pla-Rabes S., and Giralt S., 2015. Late Holocene vegetation dynamics and deforestation in Rano Aroi: Implications for Easter Island’s ecological and cultural history. Quaternary Science Reviews 126, 219–226 Rydberg J., Gälman V., Renberg I., and Mart A., 2008. Assessing the Stability of Mercury and Methylmercury in a Varved Lake Sediment Deposit. Environmental Science and Technology 42, 4391–4396 Rydberg J., Klaminder J., Rosén P., and Bindler R., 2010a. Climate driven release of carbon and mercury from permafrost mires increases mercury loading to sub-arctic lakes. Science of The Total Environment 408, 4778–4783 Rydberg J., Karlsson J., Nyman R., Wanhatalo I., Näthe K., and Bindler R., 2010b. Importance of vegetation type for mercury sequestration in the northern Swedish mire, Rödmossamyran. Geochimica et Cosmochimica Acta 74, 7116–7126 Sagnotti L., Macrí P., Camerlenghi A., and Rebesco M., 2001. Environmental magnetism of antarctic late pleistocene sediments and interhemispheric correlation of climatic events. Earth and Planetary Science Letters 192, 65–80 Sanei H. and Goodarzi F., 2006. Relationship between organic matter and mercury in recent lake sediment : The physical – geochemical aspects. Applied Geochemistry 21, 1900–1912 100 Schartup A.T., Balcom P.H., Soerensen A.L., Gosnell K.J., Calder R.S.D., Mason R.P., and Sunderland E.M., 2015. Freshwater discharges drive high levels of methylmercury in Arctic marine biota. Proceedings of the National Academy of Sciences 112, 11789–11794 Schroeder W.H. and Munthe J., 1998. Atmospheric mercury—An overview. Atmospheric Environment 32, 809–822 Schroeder W.H., Anlauf K.G., Barrie L. a, Lu J.Y., and Steffen A., 1998. Arctic Springtime Depletion of Mercury. Nature 394, 331–332 Schuster P.F., Krabbenhoft D.P., Naftz D.L., Cecil L.D., Olson M.L., Dewild J.F., Susong D.D., Green J.R., and Abbott M.L., 2002. Atmospheric Mercury Deposition during the Last 270 Years: A Glacial Ice Core Record of Natural and Anthropogenic Sources. Environmental Science and Technology 36, 2303–2310 Schuster P.F., Striegl R.G., Aiken G.R., Krabbenhoft D.P., Dewild J.F., Butler K., Kamark B., and Dornblaser M., 2011. Mercury Export from the Yukon River Basin and Potential Response to a Changing Climate. Environmental Science and Technology 45, 9262–9267 Schwander J., Stauffer B., and Sigg A., 1988. Air mixing in firn and the age of the air at pore closeoff. Annals of Glaciology 10, 141–145 Selin N.E., 2009. Global Biogeochemical Cycling of Mercury : A Review. Selin N.E., Jacob D.J., Yantosca R.M., Strode S., Jaeglé L., and Sunderland E.M., 2008. Global 3-D land-ocean-atmosphere model for mercury: Present-day versus preindustrial cycles and anthropogenic enrichment factors for deposition. Global Biogeochemical Cycles 22, 1–13 Shotyk W., 1996. Peat bog archives of atmospheric metal deposition: geochemical evaluation of peat profiles, natural variations in metal concentrations, and metal enrichment factors. Environmental Reviews 4, 149–183 Shotyk W., Goodsite M.E., Roos-Barraclough F., Frei R., Heinemeier J., Asmund G., Lohse C., and Hansen T.S., 2003. Anthropogenic contributions to atmospheric Hg, Pb and As accumulation recorded by peat cores from southern Greenland and Denmark dated using the 14C “bomb pulse curve.” Geochimica et Cosmochimica Acta 67, 3991–4011 Shotyk W., Goodsite M.E., Roos-Barraclough F., Givelet N., Le Roux G., Weiss D., Cheburkin A.K., Knudsen K., Heinemeier J., van Der Knaap W., et al., 2005. Accumulation rates and predominant atmospheric sources of natural and anthropogenic Hg and Pb on the Faroe Islands. Geochimica et Cosmochimica Acta 69, 1–17 Sigler J.M., Lee X., and Munger W., 2003. Emission and long-range transport of gaseous mercury from a large-scale Canadian boreal forest fire. Environmental Science and Technology 37, 4343–4347 Silva-Sánchez N., Schofield J.E., Mighall T.M., Martínez Cortizas A., Edwards K.J., and Foster I., 2015. Climate changes, lead pollution and soil erosion in south Greenland over the past 700 years. Quaternary Research 84, 159–173 Simpson W.R., Von Glasow R., Riedel K., Anderson P., Ariya P., Bottenheim J., Burrows J., Carpenter L.J., Friess U., Goodsite M.E., et al., 2007. Halogens and their role in polar boundary-layer ozone depletion. Atmos. Chem. Phys. 7, 4375–4418 Smith-Downey N. V., Sunderland E.M., and Jacob D.J., 2010. Anthropogenic impacts on 7. References 101 global storage and emissions of mercury from terrestrial soils: Insights from a new global model. Journal of Geophysical Research: Biogeosciences 115, 1–11 Soerensen A.L., Sunderland E.M., Holmes C.D., Jacob D.J., Yantosca R.M., Skov H., Christensen J.H., Strode S.A., and Mason R.P., 2010. An improved global model for air-sea exchange of mercury: High concentrations over the North Atlantic. Environmental Science and Technology 44, 8574–8580 Steffen A., Douglas T., Amyot M., Ariya P., Aspmo K., Berg T., Bottenheim J., Brooks S., and Cobbett F., 2008. A synthesis of atmospheric mercury depletion event chemistry in the atmosphere and snow. Atmospheric Chemistry and Physics 8, 1445–1482 Steffen A., Lehnherr I., Cole A., Ariya P., Dastoor A., Durnford D., Kirk J., and Pilote M., 2015. Atmospheric mercury in the Canadian Arctic. Part I: A review of recent field measurements. Science of the Total Environment 509–510, 3–15 Steinhilber F., Beer J., and Fröhlich C., 2009. Total solar irradiance during the Holocene. Geophysical Research Letters 36, 1–5 Steinnes E. and Sjøbakk T.E., 2005. Order-of-magnitude increase of Hg in Norwegian peat profiles since the outset of industrial activity in Europe. Environmental Pollution 137, 365–370 Stern C.R., 2008. Holocene tephrochronology record of large explosive eruptions in the southernmost Patagonian Andes. Bulletin of Volcanology 70, 435–454 Stern G.A., Macdonald R.W., Outridge P.M., Wilson S., Chételat J., Cole A., Hintelmann H., Loseto L.L., Steffen A., Wang F., et al., 2012. How does climate change influence Arctic mercury? The Science of the Total Environment 414, 22–42 Stern G. a, Sanei H., Roach P., Delaronde J., and Outridge P.M., 2009. Historical Interrelated Variations of Mercury and Aquatic Organic Matter in Lake Sediment Cores from a Subarctic Lake in Yukon, Canada : Further Evidence toward the AlgalMercury Scavenging Hypothesis. Environ. Sci. Technol. 43, 7684–7690 Streets D.G., Devane M.K., Lu Z., Bond T.C., Sunderland E.M., and Jacob D.J., 2011. Alltime releases of mercury to the atmosphere from human activities. Environmental Science and Technology 45, 10485–10491 Sun L., Yin X., Liu X., Zhu R., Xie Z., and Wang Y., 2006. A 2000-year record of mercury and ancient civilizations in seal hairs from King George Island, West Antarctica. Science of the Total Environment 368, 236–247 Sunderland E.M. and Mason R.P., 2007. Human impacts on open ocean mercury concentrations. Global Biogeochemical Cycles 21 Swain E.B., 1997. Recent Declines in Atmospheric Mercury Deposition in the Upper Midwest. Environmental Science and Technology 31, 960–967 Swain E.B., Engstrom D.R., Brigham M.E., Henning T.A., and Brezonik P.L., 1992. Increasing Rates of Atmospheric Mercury Deposition in Midcontinental North America. Science 257, 784–787 Swartzendruber P. and Jaffe D., 2012. Sources and Transport: In Mercury in the Environment, (University of California Press), pp. 3–18 Tang S., Huang Z., Liu J., Yang Z., and Lin Q., 2012. Atmospheric mercury deposition recorded in an ombrotrophic peat core from Xiaoxing’an Mountain, Northeast China. Environmental Research 118, 145–148 Thomas R.L., 1972. The Distribution of Mercury in the Sediments of Lake Ontario. 102 Canadian Journal of Earth Sciences 9, 636–651 Toro M., Camacho A., Rochera C., Rico E., Bañón M., Fernández-Valiente E., Marco E., Justel A., Marco E., Avedaño M.C., et al., 2007. Limnological characteristics of the freshwater ecosystems of Byers Peninsula, Livingston Island, in maritime Antarctica. Polar Biology 30, 635–649 Toro M., Granados I., Pla S., Giralt S., Antoniades D., Galán L., Cortizas A.M., Lim H.S., and Appleby P.G., 2013. Chronostratigraphy of the sedimentary record of Limnopolar Lake, Byers Peninsula, Livingston Island, Antarctica. Antarctic Science 25, 198–212 UNEP, 2013. Minamata Convention on Mercury - Text and Annexes. UNEP, 2013a 69 Vandal G.M., Fitzgerald W.F., Boutron C.F., and Candelone J.-P., 1993. Variations in mercury deposition to Antarctica over the past 34,000 years. Nature 362, 621– 623 Veiga M.M., Maxson P.A., and Hylander L.D., 2006. Origin and consumption of mercury in small-scale gold mining. Journal of Cleaner Production 14, 436–447 Weiss D.J., Kober B., Dolgopolova A., Gallagher K., Spiro B., Le Roux G., Mason T.F.D., Kylander M., and Coles B.J., 2004. Accurate and precise Pb isotope ratio measurements in environmental samples by MC-ICP-MS. International Journal of Mass Spectrometry 232, 205–215 Weiss H., Bertine K., Koide M., and Goldberg E.D., 1975. The chemical composition of a Greenland glacier. Geochimica et Cosmochimica Acta 39, 1–10 Weiss H. V, Koide M., and Goldberg E.D., 1971. Mercury in a Greenland Ice Sheet: Evidence of Recent Input by Man. Science 174, 692–694 WHO, 1989. Environmental Health Criteria 86: Mercury - Environmental Aspects (Geneva) Wiklund J.A., Kirk J.L., Muir D.C.G., Evans M., Yang F., Keating J., and Parsons M.T., 2017. Anthropogenic mercury deposition in Flin Flon Manitoba and the Experimental Lakes Area Ontario (Canada): A multi-lake sediment core reconstruction. Science of The Total Environment World Health Organization, 2011. Report of the joint FA (Geneva: World Health Organization) Yang H., 2015. Lake Sediments May Not Faithfully Record Decline of Atmospheric Pollutant Deposition. Environmental Science and Technology 151016145422005 Yang H. and Rose N.L., 2003. Distribution of mercury in six lake sediment cores across the UK. 304, 391–404 Yang H. and Smyntek P., 2014. Use of the mercury record in Red Tarn sediments to reveal air pollution history and the implications of catchment erosion. Environ. Sci.: Processes Impacts 16, 2554–2563 Yang H., Rose N.., Boyle J.., and Battarbee R.., 2001. Storage and distribution of trace metals and spheroidal carbonaceous particles (SCPs) from atmospheric deposition in the catchment peats of Lochnagar, Scotland. Environmental Pollution 115, 231– 238 Yang H., Battarbee R.W., Turner S.D., Rose N.L., Derwent R.G., Guangjian W., and Ruiqiang Y., 2010a. Historical Reconstruction of Mercury Pollution Across the Tibetan Plateau Using Lake Sediments. Environ. Sci. Technol. 44, 2918– 7. References 103 2924 Yang H., Engstrom D.R., and Rose N.L., 2010b. Recent Changes in Atmospheric Mercury Deposition Recorded in the Sediments of Remote Equatorial Lakes in the Rwenzori Mountains, Uganda. Environmental Science and Technology 44, 6570– 6575 Yang H., Turner S., and Rose N.L., 2016. Mercury pollution in the lake sediments and catchment soils of anthropogenically-disturbed sites across England. Environmental Pollution 219, 1092–1101 Zaccone C., Santoro A., Cocozza C., Terzano R., Shotyk W., and Miano T.M., 2009. Comparison of Hg concentrations in ombrotrophic peat and corresponding humic acids, and implications for the use of bogs as archives of atmospheric Hg deposition. Geoderma 148, 399–404 Zhang Q., Huang J., Wang F., Mark L., Xu J., Armstrong D., Li C., Zhang Y., and Kang S., 2012. Mercury distribution and deposition in glacier snow over Western China. Environmental Science and Technology 46, 5404–5413 Zheng J., 2015. Archives of total mercury reconstructed with ice and snow from Greenland and the Canadian High Arctic. Science of the Total Environment 509–510, 133– 144 Zhu W., Lin C.J., Wang X., Sommar J., Fu X., and Feng X., 2016. Global observations and modeling of atmosphere-surface exchange of elemental mercury: A critical review. Atmospheric Chemistry and Physics 16, 4451–4480 Zuna M., Ettler V., Š O., Mihaljevi M., Šebek O., and Mihaljevič M., 2012. Mercury accumulation in peatbogs at Czech sites with contrasting pollution histories. Science of The Total Environment 424, 322–330 105 8. Funding and Acknowledgments I would like to express my gratitude to all those research groups, projects and researchers that have provided the samples used in this doctoral thesis and access to laboratory facilities, through the collaboration with my supervisor Prof Antonio Martínez Cortizas and the group Ciencias do Sistema Terra of the Universidade de Santiago de Compostela. This research was made possible through funding provided by: The Ministerio Español de Ciencia e Innovación (Project CGL2010-20672), the Dirección General de Innovación y Desarrollo de Galicia (Project 10PXIB200182PR), Consellería de Cultura, Educación e Ordenación Universitaria da Xunta de Galicia (grants R2014/001, GPC2014/009 and GPC GI-1553) projects leaded by Prof Antonio Martínez Cortizas (USC). The Ministerio Español de Ciencia e Innovación (Project REN2000-0345-ANT) leaded by Dr Antonio Quesada (UAM). The Ministerio Español de Ciencia y Educación through the projects LAVOLTER (CGL2004-00683/BTE), leaded by Dr Conxita Taberner and Dr Santiago Giralt and project GEOBILA (CGL2007-60932/BTE) leaded by Dr Alberto Sáez (UB). Conselho Nacional deesenvolvimento Cientfico e Tecnolgico (CNPq), Brazil (Project Universal 14/2011482815/2001-6) leaded by Prof. Pablo Vidal-Torrado (ESALQ) The Deutsche Forschungsgemeinschaft (project DFG-BI 734/15-1) leaded by Prof Harald Biester (TUBs). Manuscript I 113 ~71 kyr of mercury deposition in the Southern Hemisphere recorded by Rano Aroi mire, Easter Island (Chile) Marta Pérez-Rodríguez 1*, Olga Margalef 2,3, J. Pablo Corella 4, Sergi Pla-Rabes3, Santiago Giralt5, Antonio Martínez Cortizas 1 * corresponding author: M. Pérez-Rodríguez; [email protected] 1 Departamento de Edafoloxía e Química Agrícola, Universidade de Santiago de Compostela, Santiago de Compostela, Spain 2 CSIC, Global Ecology Unit CREAF-CEAB-UAB, Spain 3 Ecological Research Center and Forestry Applications (CREAF), Campus de Bellaterra (UAB) Cerdanyola del Vallès, Barcelona, Spain 4 Department of Atmospheric Chemistry and Climate, Institute of Physical Chemistry Rocasolano, CSIC, Madrid, Spain 5 Institute of Earth Sciences Jaume Almera (ICTJA-CSIC), Barcelona, Spain Abstract The study of mercury accumulation in peat cores provides an excellent opportunity to improve the knowledge on mercury cycling and depositional processes at remote locations far from pollution sources. We analysed mercury concentrations in 150 peat samples on the cores ARO 06 01 and ARO 08 02 from Rano Aroi (Easter island, 27ºS) and in selected vegetation samples of present-day flora of the island in order to characterize the mercury cycling for the last ~71 kyr BP. Mercury concentration in plants ranges between 1470 and 11,175 ng g-1, corresponding the latter with a rush sample. The mercury record in ARO 08 02 and ARO 06 01 cores show values raging between 37-177 ng g-1 and 35200 ng g-1 respectively, except for two large maxima (910 and 1023 ng g-1) occurred at the end of the Last Glacial Maximum (~20 kyr cal BP) in both peat cores. Climate (both precipitation and temperature) seems to have controlled the abrupt increase in Hg deposition during this period. These values are higher than those recorded in most peat records belonging to the Industrial period, highlighting that natural factors played a significant role on Hg deposition - sometimes even higher than anthropogenic sources. Combined with a previous paleoenvironmental reconstruction, our new results suggest that wet deposition was the main process controlling Hg fluxes, both to the mire and to the catchment soils. Changes in mire vegetation, peat decomposition and volcanism had a secondary role in the mercury content in Rano Aroi. 114 1. Introduction Mercury (Hg) is a metal of environmental concern due to its high volatility, long atmospheric residence time (1-2 years) and toxicity of its methylated forms. It is released and dispersed in the atmosphere by natural emission sources like volcanoes, geothermal vents, and Hg-enriched soil as well as by anthropogenic activities such as mining, coal-fired plants and chloro-alkali plants (Schroeder and Munthe, 1998; Hylander and Meili, 2005; Horowitz et al., 2014). The transport and deposition in different environments on Earth depends on the Hg species. Elemental Hg (Hg0), the dominant species in the atmosphere, can be transported over long distance from emissions sources. It can be eventually oxidized, into divalent Hg (Hg+2) which is washed out by rainfall and deposited on the land surface (Hall, 1995; Fitzgerald and Mason, 1996; Schroeder and Munthe, 1998; Wang et al., 2014). Particulate Hg (Hgp) represents a minor fraction of total Hg in the atmosphere and can be dispersed over tens to hundreds of kilometers (Schroeder and Munthe, 1998). Natural archives such as mires, lake sediments and glacial ice have been widely used to reconstruct Hg accumulation at local, regional and global scale – see Amos et al., 2015; Biester et al., 2007; Cloy et al., 2008; Engstrom et al., 2014 and references therein. These archives have allowed the identification of processes and factors that control the deposition and accumulation of mercury over time. On most recent times, a factor between 3 and 5-fold increase in the deposition of Hg was observed in different parts of the world since the advent of the Industrial Revolution, suggesting a worldwide increase in the atmospheric Hg deposition (Lamborg et al., 2002; Amos et al., 2015). This anthropogenic input overlaps the signal driven by natural processes of deposition and accumulation. Organic matter degradation and mass loss related with peat evolution enhance Hg accumulation in peatlands (Biester et al., 2003, 2004). Shortand long-term climate oscillations seem to play an important role in Hg cycling through i) controlling the re-emission of part of the accumulated Hg (Martínez Cortizas et al., 1999); ii) inducing Hg depletion events in polar environments during glacial periods (Jitaru et al., 2009); iii) producing algal scavenging in lakes (Outridge et al., 2007; Kirk et al., 2011) or also by iv) releasing Hg from permafrost mires in warm periods (Rydberg et al., 2010a). Finally, processes such as volcanism and fires may play a role in releasing Hg into the atmosphere or in the landscape but the effects are only visible in very specific cases (i.e. Roos-Barraclough et al., 2002; Schuster et al., 2002; Ribeiro Manuscript I 115 Guevara et al., 2010; Daga et al., 2016; Corella et al., 2017). Nevertheless, most of these considerations are almost exclusively based on Holocene records, and only a few studies extend their conclusions back to the Pleistocene. In contrast to the Northern Hemisphere, there is a limited amount of quite heterogeneous information available for the Southern Hemisphere. The works focused on Hg reconstruction in the South Hemisphere are almost exclusively from Patagonian area. Most of these studies are based on lake sediments (Ribeiro Guevara et al., 2010; Hermanns et al., 2012; Hermanns and Biester, 2013; Daga et al., 2016) and a few on peat (Biester et al., 2003; Franzen et al., 2004). In a recent work we studied Hg accumulation in Pinheiros (18º S, 43ºW) (PérezRodríguez et al., 2015), a Pleistocene age (last ~57 kyrs) tropical mire located in a valley in Serra do Espinhaço Meridional (state of Minas Gerais, Brazil). In this work, we found that three of the four main processes controlling mercury concentration depended on climate: wet and dry mercury deposition (rainfall and dustfall, respectively) as well as local catchment soil erosion owing to precipitation events (Pérez-Rodríguez et al., 2015). The effect of long-term peat decomposition, the only autogenic process identified, was confined to the Holocene section of the peat (Pérez-Rodríguez et al., 2015). Although some of the mentioned works provided information from Pleistocene sections (~11.2, 14.6 and ~57 kyr cal BP, respectively [Biester et al., 2003; Franzen et al., 2004; Pérez-Rodríguez et al., 2015, 2016]), more long-term records and from a wider geographical extent are needed in order to fully understand the various processes that can influence emission, deposition, and accumulation of mercury in continental ecosystems. In particular, the role of climate, both through controls on deposition and accumulation, is one of the major issues that needs to be addressed. In this study we analysed two peat cores from Rano Aroi, a small mire located in Easter Island, which covers the last ~71 kyrs BP. Compositional (C, N, C/N, Ti) isotopic (δ13C) organic matter data and other geochemical information summarized by Principal Component Analysis (PC1 and PC2) from the previous studies were complemented with the determination of mercury content in peat samples and in selected vegetation samples of present-day flora located in the watershed. The objectives of the present research were (i) to reconstruct the Pleistocene mercury fluxes variability at millennial to centennial resolution and (ii) to determine the 116 main factors that controlled peat mercury concentrations over long time scales. 2. Material and Methods 2.1. Regional setting Easter Island (27º07’ S, 109º22’ W) is a Chilean volcanic island Miocene in age located in the southern Pacific Ocean. The nearest continental point from the South American continent lies 3512 km away. The Island has a triangular shape and the topography is characterized by more than 70 volcano craters and the rolling surfaces of lava flows (Baker et al., 1974; González-Ferrán et al., 2004; Sáez et al., 2009). The climate is humid subtropical, with average monthly temperatures between 18 °C (August) and 24 °C (February) (Junk and Claussen, 2011) and highly variable annual rainfall ranging from 500 to 1800 mm (mean of 1130 mm). Rano Aroi is a small mire located in an ancient Pleistocene volcano crater close to most elevated area of the island (27°S, 108°W, 430 m elevation, 0.13 km2; Figure 1). The crater slopes form a small catchment area (15.82 km2). Watershed lithology mainly consists on very porphyritic olivinic tholeite, basalt and hawaiite lava flows, rich in iron (González-Ferrán et al., 2004). The mire is covered by Scirpus californicus, Polygonum acuminatum, Asplenium polyodon var. squamulosum, Vittaria elongata and Cyclosorus interruptus (Zizka, 1991), while the catchment area is covered by grassland and planted eucalyptus. There is a small ravine acting as a natural outflow that infiltrates before reaching the Figure 1. (left) Map showing the location of Rano Aroi on the island. (right) Photo of the Rano Aroi mire, indicating the location of the studied cores (ARO 06 01, center of the mire and ARO 08 02, margins of the mire). Manuscript I 117 coast. An artificial outlet built in the 1960s controls maximum water levels. 2.2. Sampling Two sediment cores (ARO 06 01, 13.9 m long and ARO 08 02, 4 m long) were retrieved from Rano Aroi on March 2006 and October 2008 respectively. The core ARO 06 01 was sampled with a UWITEC corer from the central part of the mire and the The uppermost two meters were rejected to avoid potential anthropic remobilization, as described in previous studies (Margalef et al., 2013, 2014). Meanwhile ARO 08 02 was recover with a Russian corer from the eastern boundary since the mire´s margins was potentially not affected by recent human activities (Flenley and King, 1984; Flenley et al., 1991). None of the cores retrieved in either campaign reached the bedrock. The cores were sealed, transported to a core repository and stored in a cold room at +4 °C until sampling. Both cores were lithologically described and sampled for smear slides every 5 cm (Margalef et al., 2013). 2.3. Age model The chronology for both cores were fully described Margalef et al., (2013). A total of 27 AMS 14C dates were obtained from pollen concentrates of the ARO 06 01 and ARO 02 08 cores, in the Poznan Radiocarbon Laboratory (Poland). Only 18 dates were used for the age-depth model. The AMS ages were calibrated using CALIB 6.02 software, and the INTCAL 98 curve (Reimer et al., 2004) and CalPal (Danzeglocke et al., 2008) for samples older than 20,000 radiocarbon yr BP. The age model was built by simple linear interpolation. See all details in Margalef et al., (2013). 2.4. Geochemical analysis Both cores were sampled every 5 cm (286 samples) for total carbon and nitrogen (TC and TN) and stable isotope (δ13C) analyses. Samples were dried at 60 °C over 48 h, frozen with liquid nitrogen and ground in a ring mill. Analyses were performed using a Finnigan delta Plus EA-CF-IRMS spectrometer, located at the Centres Tècnics i Tecnològics of the Universitat de Barcelona (CCiTUB) (Margalef et al., 2013). 118 Titanium concentration was analysed by Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES) in core ARO 06 01. The samples were analysed for elemental concentrations using a Varian Vista AX ICP-AES at the Department of Geological Sciences, Stockholm University, Sweden. See more information for geochemical analysis in previous works (Margalef et al., 2013, 2014). Mercury concentrations were determined in 150 samples of the two cores collected in Rano Aroi and in selected vegetation samples of present-day flora located in the island. Dried, milled samples were analysed for mercury using a Milestone DMA-80 analyser hosted at the Departamento de Bioloxía Funcional (University of Santiago de Compostela). The analysis of one in every five samples was duplicated as a control. A standard reference material of the moss Pleurozium schreberi (Steinnes et al., 1997) M3, was run with each set of samples. The quantification limit was 1.34 ng g-¹ and mean recovery was 90%. 3. Results 3.1. Peat Geochemistry 3.1.1. Organic fraction: Total Carbon, Total Nitrogen, C/N and δ13C The Rano Aroi deposits are mostly organic, with TC concentrations between 40% and 70% (Margalef et al., 2013). Total N levels of core ARO 06 01 vary between 0.4% and 2%, with C/N ratios ranging from 40 to 110 (Figure 2). ARO 08 02 shows similar values with respect to TC (50–60%), TN (0.8–2%, peaking to 2.5% at 140 and 80 cm) and C/N (20–80, with a marked decrease from 2 m upward) (Figure 3). In ARO 06 01, δ13C shows a constant value around −14‰ from depths of 14 m to 9 m, whereas from 9 m to 6 m, δ13C values gradually shift from −14 to −26‰ (Figure 2). In the upper five meters, δ13C values vary around −26‰. In ARO 08 02, a gentle shift of δ13C from −19‰ to −23‰ occurs from depths of 4 m to 3.5 m (Figure 3). δ13C curves show high-frequency changes (dips) within the long-term trend (Margalef et al., 2013). 3.1.2. Inorganic fraction: Titanium Titanium concentrations were only measured in ARO 06 01 (Figure 2). They show a baseline between 54 and 304 mg kg-1 interrupted by abrupt increases higher than 1000 mg kg-1, and reaching values of 5000 mg kg-1 at 340 cm. Manuscript I 119 3.1.3. Mercury Mercury concentrations in Rano Aroi samples varied in the same range in both cores (Figure 2 and Figure 3). Core ARO 06 08 has values that range between 35 and 333 ng g-1 (147±49, mean ± standard deviation) albeit they can reach up to 1023 ng g-1 (sample at 370 cm of core depth). The Hg minimum is located at 580 cm (Figure 2). Core ARO 08 02 displays Hg values that range between 37 and 329 ng g-1 (104±56 ng g-1), but two maxima (1088 and 910 ng g-1) are located at 195 and 285 cm of core depth. According to Hg concentrations, this profile shows three distinct sections: i) from the base of the core to 290 cm values increase with depth but with small oscillations (122±26 ng g-1); ii) from 280 to 202 cm, values are low and almost constant (73±14 ng g-1); and iii) the upper 190 cm have intermediate values (101±56 ng g-1), with two relative maxima of 177 and 192 ng g-1 at 130 and 75 cm respectively (Figure 3). Figure 2. Depth profiles of the main paleoenvironmental proxies and Hg concentration (in ng g-1) analysed in ARO 06 01 core. Lithology and radiocarbon age samples are indicated in the column. Geochemical proxies: TN (in percentage) C/N ratios and δ13C (‰), are indicative of the origin of organic matter; and Ti (in ppm) come from terrigenous inputs and together with δ13C (‰) and Facies C (Margalef et al., 2013) are wet event indicators. Periods of flood events and drought conditions are marked. 120 Mercury in vegetation samples has a large concentration range (Table 1). The lowest value (341 ng g-1) corresponds to the rhizome of Scirpus californicus and the maximum value (11 175 ng g-1) with an unidentified rush sample. Table 1. Mercury concentrations determined in vegetation samples collected in Easter Island. Local/English name Scientific name Hg (ng g-1) Kikuyu/Kikuyu grass 1 945 Nga’atu/Totora (duster) Scirpus californicus (duster) 1 804 Nga’atu/Totora (rhizome) Scirpus californicus (rhizome) 341 /Rush - no identified11 175 Maku Piro/ molassesgrass Melinis minutiflora 1 470 Figure 3 Depth profiles of the main paleoenvironmental proxies and Hg concentration (in ng g-1) analysed in ARO 08 02 core versus depth. Lithology and radiocarbon age samples are indicated in the column. Geochemical proxies: TN (in percentage) C/N ratios and δ13C (‰), are indicative of the origin of organic matter; Facies C (Margalef et al., 2013) is wet event indicator. Periods of high rainfall events and drought conditions are marked. Manuscript I 121 4. Discussion 4.1. Mercury concentration in Rano Aroi records The two Rano Aroi records (ARO 08 02 and ARO 06 01) cover different temporal ranges (38.7 kyr cal BPPresent day and 71 kyrs BP - 8.5 kyr cal BP, respectively) and were retrieved in different parts of the peatland (Figure 1), although less than 50 meters apart. Previous studies, using multiple short cores from the same peatland showed that spatial variability can be a problem for an accurate Hg study, since vegetation and micro-topography changes can affect the interception and retention of atmospheric deposition, and over relatively short timescales (Bindler et al., 2004; Martínez Cortizas et al., 2012). However, the short distance between the cores, the small size of the mire (0.13 km2) and the age-depth model based on both cores (Margalef et al., 2013) seem to have largely solved this problem. Therefore, we discuss the processes that controlled mercury concentrations in Rano Aroi using both cores together; the differences between them are illustrated based on other proxies (as TC, TN and δ13C). To the best of our knowledge, there is only one comparable record – Pleistocene peat record from a tropical-sub tropical area – to Rano Aroi in the literature that is Pinheiros mire, located in Minas Gerais state, Brazil. The range of mercury concentrations in Rano Aroi records (~ 35 and 333 ng g-1) is in general comparable with that of Pinheiros mire (~36 – 370 ng g-1) (Pérez-Rodríguez et al., 2015), although the minima found in the Brazilian mire are considerably lower than those in Rano Aroi (<2 vs. 35 ng g-1), probably due to the large quartz content of the corresponding peat sections in the former (Pérez-Rodríguez et al., 2015). On the other hand, ARO cores show two extraordinary mercury peaks of ~1000 ng g-1 at ~20.0 kyr cal BP and ~5.0 kyr cal BP (Figure 4 and 5) that are more than 2 fold the maximum found in the Brazilian record. Maxima Hg values in the Pleistocene and Holocene sections of Rano Aroi are comparatively high with respect to other peatlands worldwide, even in sections affected by anthropogenic emissions. For example, for the Industrial Period maxima of ~800 ng g-1 were reported for the Czech Republic (Zuna et al., 2012), >400 ng g-1 for Scotland (Yang et al., 2001; Farmer et al., 2009), ca. 600 ng g-1 in Northeast China (Tang et al., 2012), ca. 130 ng g-1 for Canada (Outridge et al., 2011), ca. 300 ng g-1 in Spain (Martínez Cortizas et al., 2012). Our two maxima of natural Hg deposition are, however, similar to those (up to 1100 ng g-1) found in 128 Fe content and high values of PC2 (large positive loading of Fe) during enhanced rainfall periods. At least part of the Fe in (andic) volcanic soils may be as Fehumus compounds (García-Rodeja et al., 2004) to which the mercury could be bound (Nóvoa-Muñoz et al., 2008; Peña-Rodríguez et al., 2012).These compounds may be transported to the mire by superficial runoff (Figure 6). The comparison of Hg concentration in ARO 06 01 with the PC1 and PC2 components (the long-term background fluxes of inorganic particulate material and the delivery of large amounts of terrigenous particles, respectively) only shows agreement during the shorter wet events and not for the long-term trends (see Figure 4). This suggests that matter fluxes do not affect the concentration of mercury, but precipitation may be the common driver. It is worth mentioning the presence of a short rise of Hg and PC2 in ARO 06 01 at ~37 kyr cal BP that, according to the δ13C record, does not correspond to a wet period (Figure 4) pointing to other processes as an increase of Hg concentration due to terrigenous Hg-rich input. However the existence of a chronological uncertainty in this section (see Margalef et al., (2014) invites to caution in the interpretation. The data support the idea that wet deposition - directly on the mire or through the catchment circuits –was the dominant process controlling Hg concentration in Rano Aroi peatland since ~71.0 kyrs, essentially during high-rainfall events. 4.3.3. Other processes: Mire vegetation, peat decomposition and volcanic activity The extremely high Hg concentration (1023 and 910 ng g-1, ARO 06 01 and 08 02 respectively) at ~20 kyr cal BP in both cores seems to have been caused by a phase of strong rainfall, since they coincide with an increase of PC2 and a decline in δ13C (ARO 06 01 and ARO 08 02 respectively). For these peaks Hg is 3 and 4.5 fold (ARO 06 01 and 08 02 respectively) that of the next high concentration (~300 ng g-1), which could indicate other processes may have enhanced mercury accumulation in the mire or an external factor may have increased atmospheric mercury concentration. The ~20 cal kry BP maximum corresponds to a cold phase occurred at the end of the Last Glacial Maximum, when humid conditions in Eastern Island prevailed (Margalef et al., 2014). Both, colder and humid conditions would have favored Hg accumulation in Rano Aroi since Hg deposition is controlled by temperature and Manuscript I 129 humidity variations (Martínez Cortizas et al., 1999; Corella et al., 2017). It has been shown that vegetation type and degree of decomposition can affect the mercury content in peat (Rydberg et al., 2010b). Table 1 reveals that different vegetation present in the island may involve different mercury accumulation in plant remains. According to the data presented here (Table 1) an increase in the abundance of rush, which showed large (~11,000 ng g-1) mercury concentrations, coupled to more humid conditions by ~20 kyr cal BP might help to explain these increases in Hg (Figure 6). However, we have no certainty of the presence of rush in the mire catchment at present and the available pollen and macroremains data (Margalef et al., 2013, 2014; Margalef, 2014) do not allow to support this speculation. The effect of organic matter decomposition in Hg accumulation is still a subject of debate. For short-term periods, it has been proposed that increased decomposition leads to an increase in Hg concentration (Biester et al., 2003; Martínez Cortizas et al., 2007). This could be the situation during the extreme oxidation event caused by the drought period occurred at ~40.0 kyr cal BP (Figure 4), when mercury concentration slightly rises (Figure 4, Figure 6). On the other hand, the available data for longer, Pleistocene-Holocene records, as that of the Pinheiros mire (Pérez-Rodríguez et al., 2015), suggest that the effect of peat decomposition on Hg concentrations exponentially decreases to be almost negligible in peats with ages older than 10-11 kyr (Pérez-Rodríguez et al., 2015). There is no general decomposition trend in Rano Aroi and the short-term changes (reflected by changes in C/N ratio) seem to respond to enhanced rainfall periods and associated changes in mire vegetation (Margalef et al., 2013), that may not have affected the Hg concentrations. In Rano Aroi cores there is no evidence of recent volcanic activity (i.e. tephra layers), despite there are volcanic lava flows from Terevaka whose appearance suggests that they are not older than two or three thousand years (Baker, 1967). The absence of recorded volcanic eruptions in the island does not rule out the possible of Hg degassing by active fumaroles, especially on an island containing more than 70 volcano craters (Baker et al., 1974; González-Ferrán et al., 2004). Data (tephra layers) from South Patagonia indicated volcanic eruptions at ~4.2 kyr cal BP and ~7.7 kyr cal BP identified as Mt Burney (52ºS 72ºW) and Hudson volcanic eruption (45ºS 72ºW), respectively (McCulloch and Davies, 2001; Stern, 2008). 130 We have no data to support the effect of these volcanic events for the Hg peak at ~4.7 kyr cal BP (ARO 08 02) and/or at ~8.5 kyr cal BP (ARO 06 01) in Easter Island, but they cannot be discarded as a possible Hg source. Moreover, although there is some mismatch between the Hg peak and Hudson volcanic eruption, it has been shown that this specific eruption (at ~7.7 kyr cal BP) modified dust atmospheric deposition dust during a period of ~700 years (Vanneste et al., 2016). Similarly, we cannot exclude the effect of volcanic activity or fires regimes at other periods, through increasing atmospheric Hg concentration, then precipitated by rainfall. For example during the LGM, charcoal data indicates an increased in fire activity at tropical latitudes of South America (Power et al., 2008), what might have been an extra Hg emission to the atmosphere. A higher resolution would be needed in this section of the cores as well as more cores to confirm the hypothesis of the effect of volcanism and differences in intensity (i.e. differences in Hg concentration) between both events. 5. Author Contributions MP and AM designed the scientific questions to be solved. OM, SP and SG designed fieldwork and took samples; OM processed samples and performed the geochemical analysis; MP selected the samples and performed the Hg analysis; MP, PC and AM reviewed the Hg data and interpreted the results; all the coauthors contributed to wrote and reviewed the final version of this manuscript. 6. Acknowledgments This research was funded by the Spanish Ministry of Science and Education through the projects LAVOLTER (CGL2004-00683/BTE), GEOBILA (CGL200760932/BTE), R2014/001 and GPC GI-1553 (Dirección Xeral I+D, Xunta de Galicia). We would like to thank CONAF (Chile) and the Riroroko family for the facilities provided on Easter Island and Jesús R. Aboal (Universidade de Santiago de Compostela) for laboratory facilities. 7. References Allan M., Le Roux G., Sonke J.E., Piotrowska N., Streel M., and Fagel N., 2013. Reconstructing historical atmospheric mercury deposition in Western Europe using: Misten peat bog cores, Belgium. Science of The Total Environment 442, 290–301 Amos H.M., Sonke J.E., Obrist D., Robins N., Hagan N., Horowitz H.M., Mason R.P., Manuscript I 131 Witt M., Hedgecock I.M., Corbitt E.S., et al., 2015. Observational and modeling constraints on global anthropogenic enrichment of mercury. Environmental Science and Technology 49, 4036–4047 Baker P.E., 1967. Preliminary account of recent geological investigations on Easter Island. Geological Magazine 104 (2), 116–122 Baker P.E., Buckley F., and Holland J.G., 1974. Petrology and geochemistry of Easter Island. Contributions to Mineralogy and Petrology 44, 85–100 Biester H., Martinez-Cortizas A., Birkenstock S., and Kilian R., 2003. Effect of Peat Decomposition and Mass Loss on Historic Mercury Records in Peat Bogs from Patagonia. Environmental Science and Technology 37, 32–39 Biester H., Keppler F., Putschew A., Martinez-Cortizas A., and Petri M., 2004. Halogen Retention, Organohalogens, and the Role of Organic Matter Decomposition on Halogen Enrichment in Two Chilean Peat Bogs. Environmental Science and Technology 38, 1984–1991 Biester H., Bindler R., Martinez-Cortizas A., and Engstrom D.R., 2007. Modeling the Past Atmospheric Deposition of Mercury Using Natural Archives. Environmental Science and Technology 41, 4851–4860 Bindler R., Klarqvist M., Klaminder J., and Förster J., 2004. Does within-bog spatial variability of mercury and lead constrain reconstructions of absolute deposition rates from single peat records? The example of Store Mosse, Sweden. Global Biogeochemical Cycles 18, 1–12 Carlson R.W., 2005. The Mantle and Core: Treatise on Geochemistry (Elsevier) Cloy J.M., Farmer J.G., Graham M.C., MacKenzie A.B., and Cook G.T., 2008. Historical records of atmospheric Pb deposition in four Scottish ombrotrophic peat bogs: An isotopic comparison with other records from western Europe and Greenland. Global Biogeochemical Cycles 22 Corella J.P., Wang F., and Cuevas C.A., 2017. 700 years reconstruction of mercury and lead atmospheric deposition in the Pyrenees (NE Spain). Atmospheric Environment Daga R., Ribeiro Guevara S., Pavlin M., Rizzo A., Lojen S., Vreča P., Horvat M., and Arribére M., 2016. Historical records of mercury in southern latitudes over 1600years: Lake Futalaufquen, Northern Patagonia. Science of The Total Environment 553, 541–550 Danzeglocke U., Jöris O., and Weninger B., 2008. CalPal-2007 Engstrom D.R., Fitzgerald W.F., Cooke C.A., Lamborg C.H., Drevnick P.E., Swain E.B., Balogh S.J., and Balcom P.H., 2014. Atmospheric Hg emissions from preindustrial gold and silver extraction in the Americas: A reevaluation from lake-sediment archives. Environmental Science and Technology 48, 6533–6543 Enrico M., Roux G. Le, Marusczak N., Heimbürger L.-E., Claustres A., Fu X., Sun R., and Sonke J.E., 2016. Atmospheric Mercury Transfer to Peat Bogs Dominated by Gaseous Elemental Mercury Dry Deposition. Environmental Science and Technology acs.est.5b06058 Farmer J.G., Anderson P., Cloy J.M., Graham M.C., MacKenzie A.B., and Cook G.T., 2009. Historical accumulation rates of mercury in four Scottish ombrotrophic peat bogs over the past 2000 years. Science of the Total Environment 407, 5578– 5588 132 Fitzgerald W.F. and Mason R.P., 1996. The global mercury cycle: oceanic and anthropogenic aspects. In Global and Regional Mercury Cycles: Sources, Fluxes and Mass Balances, (Springer), pp. 85–108 Flanagan F.J., Moore R., and Aruscavage P.J., 1982. Mercury in Geologic Reference Samples. Geostandards Newsletter 6, 25–46 Flenley J.R. and King S.M., 1984. Late Quaternary pollen records from Easter Island. Nature 307, 47–50 Flenley J.R., King A.S.M., Jackson J., Chew C., Teller J.T., and Prentice M.E., 1991. The Late Quaternary vegetational and climatic history of Easter Island. Journal of Quaternary Science 6, 85–115 Franzen C., Kilian R., and Biester H., 2004. Natural mercury enrichment in a minerogenic fen--evaluation of sources and processes. Journal of Environmental Monitoring : JEM 6, 466–472 García-Rodeja E., Nóvoa J.C., Pontevedra X., Martínez-Cortizas A., and Buurman P., 2004. Aluminium fractionation of European volcanic soils by selective dissolution techniques. Soils of Volcanic Regions in Europe 56, 325–351 González-Ferrán O., Mazzuoli R., and Lahsen A., 2004. Geología del complejo volcánico Isla de Pascua-Rapa Nui, Chile. V Región-Valparaiso. Hall B., 1995. The gas phase oxidation of elemental mercury by ozone. In Mercury as a Global Pollutant, (Springer), pp. 301–315 Hermanns Y. and Biester H., 2013. Anthropogenic mercury signals in lake sediments from southernmost Patagonia, Chile. Science of the Total Environment 445–446, 126–135 Hermanns Y.-M., Cortizas A.M., Arz H., Stein R., and Biester H., 2012. Untangling the influence of in-lake productivity and terrestrial organic matter flux on 4,250 years of mercury accumulation in Lake Hambre, Southern Chile. Journal of Paleolimnology 49, 563–573 Horowitz H.M., Jacob D.J., Amos H.M., Streets D.G., and Sunderland E.M., 2014. Historical mercury releases from commercial products: Global environmental implications. Environmental Science and Technology 48, 10242–10250 Hylander L.D. and Meili M., 2005. The Rise and Fall of Mercury: Converting a Resource to Refuse After 500 Years of Mining and Pollution. Critical Reviews in Environmental Science and Technology 35, 1–36 Jitaru P., Gabrielli P., Marteel A., Plane J.M.C., Planchon F.A.M., Gauchard P.-A., Ferrari C.P., Boutron C.F., Adams F.C., Hong S., et al., 2009. Atmospheric depletion of mercury over Antarctica during glacial periods. Nature Geoscience 2, 505– 508 Junk C. and Claussen M., 2011. Simulated climate variability in the region of Rapa Nui during the last millennium. Climate of the Past 7, 579–586 Kirk J.L., Muir D.C.M., Antoniades D., Douglas M.S. V, Evans M.S., Jackson T.A., Kling H., Amoureux S., Lim D.S.S., Pienitz R., et al., 2011. Climate change and mercury accumulation in canadian high and subarctic lakes. Environmental Science and Technology 45, 964–970 Lamborg C.H., Fitzgerald W.F., Damman A.W.H., Benoit J.M., Balcom P.H., and Engstrom D.R., 2002. Modern and historic atmospheric mercury fluxes in both hemispheres: Global and regional mercury cycling implications. Global Biogeochemical Cycles Manuscript I 133 16, 1104 Margalef O., 2014. The last 70 ky of Rano Aroi (Easter Island, 27 ° S) peat record: New insights for the Central Pacific paleoclimatology. Universitat de Barcelona Margalef O., Cañellas-Boltà N., Pla-rabes S., Giralt S., Pueyo J.J., Joosten H., Rull V., Buchaca T., Hernández A., Valero-Garcés B.L., et al., 2013. A 70,000 year multiproxy record of climatic and environmental change from Rano Aroi peatland (Easter Island). Global and Planetary Change 108, 72–84 Margalef O., Martínez Cortizas A., Kylander M., Pla-Rabes S., Cañellas-Boltà N., Pueyo J.J., Sáez A., Valero-Garcés B.L., and Giralt S., 2014. Environmental processes in Rano Aroi (Easter Island) peat geochemistry forced by climate variability during the last 70kyr. Palaeogeography, Palaeoclimatology, Palaeoecology 414, 438– 450 Martínez Cortizas A., Pontevedra-Pombal X., García-Rodeja E., Nóvoa-Muñoz J.C., and Shotyk W., 1999. Mercury in a Spanish Peat Bog: Archive of Climate Change and Atmospheric Metal Deposition. Science 284, 939–942 Martínez Cortizas A., Biester H., Mighall T., and Bindler R., 2007. Climate-driven enrichment of pollutants in peatlands. Biogeosciences 4, 905–911 Martínez Cortizas A., Peiteado Varela E., Bindler R., Biester H., and Cheburkin A., 2012. Reconstructing historical Pb and Hg pollution in NW Spain using multiple cores from the Chao de Lamoso bog (Xistral Mountains). Geochimica et Cosmochimica Acta 82, 68–78 McCulloch R.D. and Davies S.J., 2001. Late-glacial and Holocene palaeoenvironmental change in the central Strait of Magellan, southern Patagonia. Palaeogeography, Palaeoclimatology, Palaeoecology 173, 143–173 Nóvoa-Muñoz J.C., Pontevedra-Pombal X., Martínez-Cortizas A., and García-Rodeja Gayoso E., 2008. Mercury accumulation in upland acid forest ecosystems nearby a coal-fired power-plant in Southwest Europe (Galicia, NW Spain). Science of The Total Environment 394, 303–312 Outridge P.M., Sanei H., Stern G. A., Hamilton P.B., and Goodarzi F., 2007. Evidence for control of mercury accumulation rates in Canadian High Arctic Lake sediments by variations of aquatic primary productivity. Environmental Science and Technology 41, 5259–5265 Outridge P.M., Rausch N., Percival J.B., Shotyk W., and McNeely R., 2011. Comparison of mercury and zinc profiles in peat and lake sediment archives with historical changes in emissions from the Flin Flon metal smelter, Manitoba, Canada. The Science of the Total Environment 409, 548–563 Peña-Rodríguez S., Pontevedra-Pombal X., Fernández-Calviño D., Taboada T., AriasEstévez M., Martínez-Cortizas A., Nóvoa-Muñoz J.C., and García-Rodeja E., 2012. Mercury content in volcanic soils across Europe and its relationship with soil properties. Journal of Soils and Sediments 12, 542–555 Pérez-Rodríguez M., Horák-Terra I., Rodríguez-Lado L., Aboal J.R., and Martínez Cortizas A., 2015. Long-Term (~57 ka) controls on mercury accumulation in the souther hemisphere reconstructed using a peat record from pinheiro mire (minas gerais, Brazil). Environmental Science and Technology 49, 1356–1364 Pérez-Rodríguez M., Horák-Terra I., Rodríguez-Lado L., and Martínez Cortizas A., 2016. Modelling mercury accumulation in minerogenic peat combining FTIRATR spectroscopy and partial least squares (PLS). Spectrochimica Acta Part A: 134 Molecular and Biomolecular Spectroscopy 168, 65–72 Power M.J., Marlon J., Ortiz N., Bartlein P.J., Harrison S.P., Mayle F.E., Ballouche A., Bradshaw R.H.W., Carcaillet C., Cordova C., et al., 2008. Changes in fire regimes since the last glacial maximum: An assessment based on a global synthesis and analysis of charcoal data. Climate Dynamics 30, 887–907 Reimer P.J., Baillie M.G.L., Bard E., Bayliss A., Beck J.W., Bertrand C.J.H., Blackwell P.G., Buck C.E., Burr G.S., and Cutler K.B., 2004. IntCal04 terrestrial radiocarbon age calibration, 0–26 cal kyr BP. Ribeiro Guevara S., Meili M., Rizzo A., Daga R., and Arribére M., 2010. Sediment records of highly variable mercury inputs to mountain lakes in Patagonia during the past millennium. Atmospheric Chemistry and Physics 10, 3443–3453 Roos-Barraclough F., Martinez-Cortizas A., Garcıa-Rodeja E., Shotyk W., García-Rodeja E., and Shotyk W., 2002. A 14 500 year record of the accumulation of atmospheric mercury in peat: volcanic signals, anthropogenic influences and a correlation to bromine accumulation. Earth and Planetary Science Letters 202, 435–451 Roulet M., Lucotte M., Saint-Aubin A., Tran S., Rheault I., Farella N., Dezencourt J., Passos C.-J.S., Soares G.S., and Guimaraes J.-R., 1998. The geochemistry of mercury in central Amazonian soils developed on the Alter-do-Chao formation of the lower Tapajos River Valley, Para state, Brazil. Science of the Total Environment 223, 1–24 Rull V., Cañellas-Boltà N., Margalef O., Sáez A., Pla-Rabes S., and Giralt S., 2015. Late Holocene vegetation dynamics and deforestation in Rano Aroi: Implications for Easter Island’s ecological and cultural history. Quaternary Science Reviews 126, 219–226 Rydberg J., Klaminder J., Rosén P., and Bindler R., 2010a. Climate driven release of carbon and mercury from permafrost mires increases mercury loading to sub-arctic lakes. Science of The Total Environment 408, 4778–4783 Rydberg J., Karlsson J., Nyman R., Wanhatalo I., Näthe K., and Bindler R., 2010b. Importance of vegetation type for mercury sequestration in the northern Swedish mire, Rödmossamyran. Geochimica et Cosmochimica Acta 74, 7116–7126 Sáez A., Valero-Garcés B.L., Giralt S., Moreno A., Bao R., Pueyo J.J., Hernández A., and Casas D., 2009. Glacial to Holocene climate changes in the SE Pacific. The Raraku Lake sedimentary record (Easter Island, 27oS). Quaternary Science Reviews 28, 2743–2759 Schroeder W.H. and Munthe J., 1998. Atmospheric mercury—An overview. Atmospheric Environment 32, 809–822 Schuster P.F., Krabbenhoft D.P., Naftz D.L., Cecil L.D., Olson M.L., Dewild J.F., Susong D.D., Green J.R., and Abbott M.L., 2002. Atmospheric Mercury Deposition during the Last 270 Years: A Glacial Ice Core Record of Natural and Anthropogenic Sources. Environmental Science and Technology 36, 2303–2310 Steinnes E., Rühling Å., Lippo H., and Mäkinen A., 1997. Reference materials for large-scale metal deposition surveys. Accreditation and Quality Assurance 2, 243– 249 Stern C.R., 2008. Holocene tephrochronology record of large explosive eruptions in the southernmost Patagonian Andes. Bulletin of Volcanology 70, 435–454 Tang S., Huang Z., Liu J., Yang Z., and Lin Q., 2012. Atmospheric mercury deposition Manuscript I 135 recorded in an ombrotrophic peat core from Xiaoxing’an Mountain, Northeast China. Environmental Research 118, 145–148 Vanneste H., De Vleeschouwer F., Bertrand S., Martínez-Cortizas A., Vanderstraeten A., Mattielli N., Coronato A., Piotrowska N., Jeandel C., and Roux G. Le, 2016. Elevated dust deposition in Tierra del Fuego (Chile) resulting from Neoglacial Darwin Cordillera glacier fluctuations. Journal of Quaternary Science 31, 713–722 Wang F., Saiz-Lopez A., Mahajan A.S., Gómez Martín J.C., Armstrong D., Lemes M., Hay T., and Prados-Roman C., 2014. Enhanced production of oxidised mercury over the tropical Pacific Ocean: A key missing oxidation pathway Yang H., Rose N.., Boyle J.., and Battarbee R.., 2001. Storage and distribution of trace metals and spheroidal carbonaceous particles (SCPs) from atmospheric deposition in the catchment peats of Lochnagar, Scotland. Environmental Pollution 115, 231– 238 Zizka G., 1991. Flowering plants of Easter Island. Palmarum Hortus Francofurtensis; 3 Zuna M., Ettler V., Š O., Mihaljevi M., Šebek O., and Mihaljevič M., 2012. Mercury accumulation in peatbogs at Czech sites with contrasting pollution histories. Science of The Total Environment 424, 322–330 MANUSCRIPT II Long-term (~57 ka) controls on mercury accumulation in the Southern Hemisphere reconstructed using a peat record from Pinheiro Mire (Minas Gerais, Brazil) Environmental Science and Technology, 49, 1356-1364. JCR IF (2015): 5.393, D1 in Engineering, Environmental and Environmental Sciences Manuscript V 195 Total solar insolation and algae productivity increase mercury accumulation in a pristine Southern Hemisphere lake Harald Biester 1*, Marta Pérez-Rodríguez 2, Benjamin-Silas Gilfedder 3, Antonio Martínez Cortizas 2, Yvonne-Marie Hermanns 1 * corresponding author: H. Biester; [email protected] 1 Institud für Geoökologie, AG Umweltgeochemie, Technische Universität Braunschweig, 38106 Braunschweig, Germany 2 Departamento de Edafoloxía e Química Agrícola, Facultade de Bioloxía, Universidade de Santiago de Compostela, Campus Sur, Santiago de Compostela 15782, Spain 3 Lehrstuhl für Hydrologie, Universität Bayreuth, Universitätsstr. 30, 95440 Bayreuth, Germany Abstract Mercury (Hg) is know to highly accumulate in aquatic biota. Uptake by algae has been found to be its entry-point into the aquatic foodchain however the contribution of Hg uptake by algae under changing solar irradiance and climatic conditions is currently unknown for pre-industrial times. We have analyzed the link between Hg accumulation, cyclic changes in total solar insolation (TSI), related changes in lake productivity and climate during the past 4.5 kyrs in sediments of a small highly productive lake located in Southern Patagonia (53°S). The analyses encompass proxies for TSI (based on 10Be), sediment geochemical composition and lake productivity (FTIR spectra, and hydrogen-index (HI)). The sediment record shows high concentrations of organic matter (median 70 %) and strong vari¬ations in Hg accumulation which correspond to changes in TSI and aquatic productivity. Accumulation of Hg was highest during drier periods when insolation and lake productivity was high and erosion fluxes from the catchment were low. During these periods, accumulation was up to fourfold higher compared to those of lower TSI, lower productivity and wetter climatic conditions. This indicates that sediment Hg accumulation and potential Hg methylation in this highly productive lake were to a large extent controlled by insolation and related algae production and to a lesser extent by Hg fluxes supplied by erosion from catchment soils. We suggest that the high Hg uptake by algae is due to water-column methylation in settling particles in anaerobic micro-niches, at times of high productivity and eutrophication. Our findings implies that sediment Hg accumulation, related to the mentioned conditions, can surpass Hg fluxes from the catchment in productive lakes and does potentially increase formation and enrichment of methyl-Hg in lake foodchain. 196 1. Introduction Mercury is an element of concern as anthropogenic emissions have largely surpassed those from natural sources and the metal is known for its biomagnification in aquatic food chains. Today, nearly all freshwater sediments of the past 200 years are enriched in Hg by a factor of 3-5 or more compared to preindustrial times (Biester et al., 2007). Due to its high affinity to bind to organic components, the biogeochemical cycle of Hg is largely coupled to that of organic matter. Transport by dissolved organic matter (DOM) is the major pathway of Hg from catchments into aquatic systems (Driscoll, 1995). In freshwater systems and estuaries, algae have been shown to accumulate Hg from the water phase and influence Hg cycling by affecting its concentrations, speciation and transport to the sediments (Fitzgerald et al., 2007, Lavoie et al., 2013, Dranguet et al., 2014; Le Faucheur et al., 2014). The mechanism of Hg uptake by phytoplankton is not well understood, with both facilitated transport and diffusion involved (Le Faucheur et al., 2014, Dranguet et al., 2014). Oxidized dissolved Hg (HgII-species such as HgCl2) and methylmercury (CH3HgX (MeHg)) are the dominant species taken up and lipophilic CH3HgCl has a higher bioaccumulation factor than other Hg2+- species, whereas DOM-bound Hg appears to decrease uptake by algae (Dranguet et al., 2014; Le Faucheur et al., 2014; Schartup et al., 2015a). Direct uptake of gaseous or dissolved elemental Hg (GEM) by phytoplankton seems to be of minor importance as no detectable accumulation of Hg could be observed when GEM was used in bioaccumulation experiments. However, recent studies have emphasized the importance of DOM-bound-MeHg as well as the quality of DOM (terrestrial versus marine) for Hg enrichment in plankton (Jonsson et al., 2014; Schartup et al., 2015a) and the literature therein). Oxic water-column Hg methylation has been found to be a major process increasing Hg concentrations in settling particles (Balcom et al., 2015; Gascón Díez et al., 2016). Moreover, eutrophication has been shown to increase phytoplankton MeHg concentrations in marine systems and the export of organically bound Hg to the sediment, simultaneously decreasing the water phase Hg reservoir (Soerensen et al., 2016). Scavenging of Hg by algae has been suggested as a major process for Hg accumulation in lake sediments. From the simultaneous increase of algae derived organic matter and Hg concentrations in Arctic lakes, some studies concluded algal production represents one of the main drivers for the increase in Hg accumulation in sediments during the industrial age (Outridge et al., 2007; Stern Manuscript V 197 et al., 2009). However, other work has found that the correlation between both proxies is not consistent throughout a larger number of Arctic lakes (Kirk et al., 2011). In contrast, studies based on mass balance calculations have hypothesized that increased algae production may lead to bio-dilution of Hg concentra¬tions in algae. Investigations of lake sediment records on a millennial scale, aiming to disentangle the past climatic control underlying the relationship between algal productivity and increased Hg accumulation -independent from anthropogenic pollutionare limited. For example, Cooke et al., (2012) could not find a relationship between chlorophyll a and Hg concentrations during the Holocene in an Arctic lake. In a previous study we analyzed sediments from a small lake at the Strait of Magellan (53 °S), and found a significant correlation between proxies of algal production (Rock Eval hydrogen index (HI), C/N ratios) and the accumulation of Hg in the sediments during the past 4.5 kyr (Hermanns et al., 2013). However, it is largely unknown how and to which extent Hg uptake and enrichment by algae controls the accumulation of Hg alone, or if other factors such as insolation, temperature and/or precipitation, which control both productivity and Hg fluxes to the lake, are important. In addition, the Hg mass balance resulting from increased Hg uptake by algae is mostly unknown. A recent model study on the role of eutrophication on Hg accumulation in Baltic Sea sediments suggests that increased uptake of Hg by algae during eutrophication will decrease Hg evasion from the lake e.g. by biotic or photochemical reduction (Soerensen et al., 2016). Algal productivity in lakes is controlled by insolation, temperature and nutrient availability. Total solar irradiance (TSI), as a function of solar output, has varied throughout the Holocene and is known to show periodicities of varying frequency. Periodical changes in insolation may cause cyclic climatic effects on Earth such as the Bond Events (Bond et al., 1997), which are related to quasi-periodical (~1500 years) cooling of the North Atlantic. It has been demonstrated that cyclic variations in algal productivity of Arctic lakes are coherent with time series of cosmogenic nuclides (14C and 10Be), which are directly related to changes in solar output (Hu et al., 2003). Despite its importance to understand the natural variability of Hg accumulation in lakes the relationship between TSI, lake productivity and the sedimentation of Hg has not yet been investigated. One major problem here is the sensitivity of productivity proxies and the extraction of the Hg signal which is directly related to changes in productivity. This is especially true where sediment 198 Hg accumulation is dominated by surface run-off of DOM-Hg which tends to obscure Hg signals related to productivity. In this study, we supplemented data from a previous investigation on the importance of algal scavenging and catchment processes (terrestrial organic matter fluxes) on Hg accumulation in Lago Hambre (LH) (53 °S) (Hermanns et al., 2013) by the spectroscopic (FTIR) character-isation of the lake’s sediments and compared it to the 10Be record (a proxy for TSI). Pérez-Rodríguez et al., (2016) showed that the application of principal component analysis (PCA) to the LH FTIR data allowed sensitive detection of in-lake productivity signals which showed a strong dependency on TSI changes throughout the Holocene. Based on this preliminary result, we present a re-evaluation of the 4.0 kyrs Hg sediment record of this small, pristine lake in Patagonia (Fig.1). We use a combination of geochemical and spectroscopic proxies discussed in previous studies (Hermanns and Biester, 2013; Hermanns et al., 2013; Pérez-Rodríguez et al., 2016) to determine the influence of solar-induced lake productivity changes on mercury accumulation. 2. Methods and Materials 2.1. Location Lago Hambre is located 50 km south of Punta Arenas, near the Strait of Magellan, in southernmost Patagonia, Chile (53°360 13.1900 S, 70°570 8.7700 W). It is situated about 80 m a.s.l. and is a small lake with a surface area of 13,700 m2 and a maximum depth of 17 m. The catchment-to-lake ratio is about eight. The lake is situated in the zone of sub-Antarctic deciduous forest dominated by Nothofagus pumilio and Nothofagus antarctica. 2.2. Sediment sampling A 5 m long sediment core was recovered from the deepest part of the lake in 2008 using a piston corer. Sediment cores were stored under dark and cool (4 °C) conditions until analyses. Sub-samples were taken at 1 cm intervals in the upper 470 cm of the piston core, and at 2-cm intervals below 470 cm. 2.3. Element and organic matter analyses Mercury concentrations were determined by means of CVAAS after combustion Manuscript V 199 of the sample and pre-concentration of Hg by amalgamation on a gold trap using a DMA 80 mercury analyzer (MLS). Selected samples (n=24) were analysed in duplicates. Precision was always better than 6% RSD. Certified reference materials (CRM: NIST 1515, 44 ± 4 µg kg-1; CCRMP LKSD-4: 190 µg kg-1) were used for quality control. Results for the NIST 1515 reference material were always within the standard deviation of the certified value and recovery for LKSD-4 was always between 95-100%. Concentrations of carbon (C) and nitrogen (N) in carbonate-free samples were determined by GC-TCD after thermal combustion in an elemental analyzer (Euro EA3000, Eurovector, Germany). Samples were analyzed for copper (Cu) and zirconium (Zr) using an energy-dispersive XRF mini-probe multi-element analyzer (EMMA (Cheburkin and Shotyk, 1996)). The hydrogen index (HI), i.e. the quantity of pyrolyzable hydrocarbons (S2) per gram TOC (mg HC gC−1), was determined by means of a Rock–Eval-II-plus-S3-unit at the Alfred Wegener Institut für Polarund Meeresforschung in Bremerhaven, Germany using sample aliquots of 30–50 mg. Figure 1. Map of southern South America and location of Lake Hambre. 200 2.4. Infrared spectroscopy of sediments A total of 430 samples were analysed by Infrared Spectroscopy (IR). FTIR spectra of freeze-dried and ground samples of lake sediment were obtained using a Vector 22 FTIR spectrometer (BrukerOptik, Ettlingen, Germany) in absorption mode, with subsequent baseline substraction on KBr pellets (200 mg dried KBr and 2 mg sample). Measurements were recorded from 4,500 to 300 cm-1 using a resolution of 2 cm-1. Thirty-two scans were taken per sample and averaged to obtain the final spectra. In order to improve the comparison between samples, each FTIR spectrum was re-scaled by normalising each absorption value to the integrated spectrum area. 2.5. Statistical modelling Further discussion of the FTIR data will be confined to the use of one (Cp9) of the PCA (principal component analysis) components extracted from the FTIR data from a previous study in Lago Hambre (see Pérez-Rodríguez et al., 2016 for full details). In brief, PCA was performed on the FTIR data to identify the main geochemical signals of the lake core and relate them to the underlying environmental processes in the lake and its catchment. We performed the PCA on a transposed matrix (samples as columns and absorbance bands as rows) and a varimax-rotation was applied to maximize the loadings. Each extracted component is represented by a spectrum of scores, which permits us to identify the most relevant signals characterizing the composition of the samples. Sample loadings (or their squared values) indicate the weight of the identified compounds on the samples’ spectroscopic signal (related to its abundance in the samples). In the present work we use only the information relative to the last 4.0 kyr. PCA and Pearson correlation coefficient analyses were performed using the statistic environment R 3.1.1 (Team, 2014). 2.6. Spectral analysis The spectral analysis (REDFIT, (Schulz and Mudelsee, 2002) providing periodicities of PCA-Cp9 (an algaenan signal), HI, Hg and Cu concentrations and accu-mulation were performed using the Past software (Hammer et al., 2001). PCA-Cp9 and HI spectral analysis are from Pérez-Rodríguez et al. (2016). Manuscript V 201 3. Results and Discussion 3.1 Geochemical and climatic background Lago Hambre is a small (surface area = 13,700 m2), 17 m deep lake, located 50 km south of Punta Arenas (53°36’ 13.19’’S, 70°57’ 8.77’’ W), Patagonia (Chile) (Fig. 1). The radiocarbon dated ~5 m long sediment core dates back to ~4500 yrs (Hermanns et al., 2013). The sediment record is characterized by generally high organic matter content (carbon concentration, median 38 wt. %, ~70 wt. % organic matter) interrupted by more mineral matter-rich layers related to periodic changes in wetter and drier conditions (Hermanns et al 2013, Pérez-Rodríguez et al., 2016) and a tephra layer at 4250 cal. B.P. (Mt. Burney). At the time of sampling the lake was stratified and anoxic below a water depth of ~7 m. The variability in C/N ratios is generally high (range 12-29, median 17) pointing to frequent changes in the dominant source of organic matter (OM). Terrestrial OM (higher C/N-ratios) mainly derived from the Nothofagus forest soils, that were dominant (median C/ N=19) between 3500 and 1700 B.P. (McCulloch and Davies, 2001), followed by a shift towards more algal OM (lower C/N ratios) in the past ~1700 years (median C/N=16), when the climate became drier and colder (Fig. 2) (Caniupán et al., 2014; Pérez-Rodríguez et al., 2016). One of the PCA components (PCA-Cp9) extracted from the FTIR-spectra is characterised by high loadings of absorptions typical for aliphatic compounds associated with green algae (Pérez-Rodríguez et al., 2016). This specific spectrum is likely to be associated with algaenan, an insoluble, non-hydrolysable, and highly aliphatic macromolecule that is a structural component of the cell wall of freshwater green algae (Blokker et al., 1998). The record of PCA1-Cp9 scores shows a close correlation (r=0.74, n=226) with the hydrogen index (HI) (Fig. 2) where high HI values are typical for algae OM (Outridge et al., 2007). Both proxies of algal production (PCA1-Cp9 and HI) appear closely related to changes in TSI, showing a periodicity of ~ 180-220 years (Fig. 2 and 3). This indicates that productivity in LH followed these short term changes in insolation (PérezRodríguez et al., 2016). Similar periodicity in past climatic changes in Patagonia has been observed in other studies (Turney et al., 2016) Besides the effect on aquatic productivity, cyclic variations in sediment composition also point to changes in climate conditions at the Strait of Magellan. We used zirconium (Zr), a typical conservative lithogenic element, as a proxy of 202 erosion of mineral matter from the catchment soils (Fig. 2). Periods of high TSI and lake productivity (Cp9, HI, C/N) mostly correspond to phases of relatively low mineral matter fluxes and relatively dry conditions. Moreover, it has been shown that productivity proxies also seem to be in pace with the longer periodicity (~1500 years) of Bond Events during the past 5 kyrs indicating a link to the southernmost SH climate system with solar-forced cooling of the NH (PérezRodríguez et al., 2016). Drier and colder periods in southern Patagonia are most likely caused by weakening or shifting of the Southern Hemisphere westerly wind belt (Lamy et al., 2010; Varma et al., 2011; Kilian and Lamy, 2012; Caniupán et al., 2014). However, the relationship between drier conditions and higher productivity was not consistent throughout the entire record because local climatic variation may have influenced both productivity and mineral matter fluxes to the lake. Accordingly, some sections show an increase in both Zr accumulation and productivity proxies (e.g. at ~1600 BP, ~3600 BP), which indicate higher erosion (Zr) and productivity during wetter and probably warmer conditions. 3.2 Mercury Accumulation The LH sediments show strong variations in concentrations and accumulation rates of mercury throughout the past ~4000 years (Fig. 2 and S1 for concentrations). Both vary by a factor of ~4.2, between 87 and 362 µg kg-1 (median 187 µg kg-1) and 14 to 53 µg m-2 yr-1 (median 29 µg m-2 yr-1), respectively. Mercury accumulation rates in LH are clearly higher than background atmospheric Hg deposition rates reported from other remote lakes (1-9 µg m-2 yr-1) (Biester et al., 2007). Hermanns et al. (2013) did not find a significant covariation between mercury and the organic matter content or the fluxes of inorganic matter from the catchment at LH. Median Hg concentration in LH sediments (187 µg kg-1) is about 3.5-fold higher than the average Hg concentrations (53 µg kg-1) in the organic matter rich top soils and 8-fold higher than the mercury content (23 µg kg-1) in the rocks of the lake’s catchment (Hermanns et al. 2013). Sediment focusing is assumed to be of minor importance here as LH is too small and too deep to cause significant grain size separation during sedimentation. Even if carbon losses of ~25 % during organic matter diagenesis in the sediment are assumed (Rydberg et al., 2008), there needs to be an additional factor to explain the large enrichment in Hg in the lake sediments. In addition, the record of Hg accumulation rates appears inversely related (although p>0.05) to that of the Zr accumulation (Fig. 2) indicating that Manuscript V 203 Figure 2. Proxy records of aquatic productivity (PCA1–Cp9 and HI), Hg and Cu accumulation, Zr accumulation for mineral matter sedimentation and 10Be-based reconstruction of total solar irradiance (ΔTSI, after (Steinhilber et al., 2009) and carbon/nitrogen ratios (C/N). Grey bars indicate drier events at LH derived from Zr accumulation (Pérez-Rodriguez et al., 2016). Streaked bar shows Mt Burney tephra layer ΔTSI (Wm−2) −0.6 −0.2 0.2 0.6 01234 Calendar age (ka BP) Cp9 −2 −1 0 1 2 HI (mgHC gToC−1) 250 300 350 400 C/N 10 15 20 25 Hg (μ g m−2 yr−1) 20 30 40 50 60 70 Cu (mg m−2 yr−1) 6 8 10 12 14 16 18 Zr (mgm−2yr−1) 5 10 15 20 + Erosion Mt Burney Tephra 01234 Calendar age (ka BP)