BBC Focus Magazine

[Flakkarin]

I had good fun researching and writing this piece for class (it's the background to my research) so I thought maybe I'd share it. Enjoy.

Figs here: http://s702.photobucket.com/albums/ww27/Flakkarin/Glacio-volcanic/
_______
The anthropogenic release of excess CO2 into the atmosphere is often termed the ‘human volcano’. But the volcanic release of CO2 is sufficiently slow to only cause climate warming on millennial time scales or longer. On shorter (decadal-centennial) time scales, large volcanic eruptions are responsible for cooling the atmosphere, through injection of sulphur aerosols into the troposphere and stratosphere. These scatter the incoming short wave radiation, effectively cooling the lower atmosphere (and thus decreasing Earth surface temperatures), while volcanic dust absorbs outgoing long wave radiation warming the stratosphere (Baldwin et al., 1976) (Fig 1). Both of these results were observed in large explosive eruptions as early as 1963 (Lamb, 1970). Dust particles tend to settle out of the atmosphere relatively quickly, and this, coupled with the stronger influence of scattering by sulphur aerosols, means volcanic eruptions tend to have a ‘reverse greenhouse effect’, where the dominant result is cooling. However, the scattering of radiation is not uniform at all latitudes since it is partly dependant on the angle of solar incidence: i.e. how much of the thicker aerosol-laden atmosphere the incoming solar radiation has to travel through. This means that the higher latitudes experience greater depletion of the incoming solar radiation, and are cooled more strongly following eruptions (Budyko, 1968).
This provides volcanic eruptions with the potential to seriously affect the sensitive climate of high polar latitudes. However, single large eruptions only produce cooling trends that last a couple of years. After the eruption of Mt Pinatubo in the Philippines in 1991, global mean air temperatures were reduced by up to 0.5°C at the surface, but the effects were measureable only until 1994 (Parker et al., 1996). Even the largest volcanic eruption in recorded human history, that of Tambora, Indonesia, in 1815, the atmosphere cooled by only ~1⁰C for just a couple of years (Self, 2006). For volcanic eruptions to cause larger changes, and affect glaciations, several large eruptions need to occur closely spaced in time, to propagate the cooling signal.
The intensity of volcanic activity is known to have waxed and waned throughout Earth’s history. An exceptional record of activity from a Greenland ice core had provided a continuous record of activity for the past 110,000 years (Fig 2), (Zielinski et al., 1996). As a first indicator of the potential of volcanic activity to significantly cool the atmosphere, historical records of eruptions were compared to instrumental surface temperature records for the past 250 years (Self et al., 1981). Correlations were found between cooler temperatures and volcanic eruptions, but the paper concluded that series of eruptions had no prolonging effect on the climate signal. However, this may overlook several issues, such as the source area for each eruption. High-latitude eruptions will only permit the transportation of aerosols within one hemisphere, whereas mid-latitude or equatorial eruptions can disperse their products across both hemispheres, and cause true global cooling (Robock, 2000). The limited time frame of the study may also present a bias in the location and eruption styles. Extending correlation studies further back in time required the development of proxy records for temperature and volcanism. For example, Briffa et al. (1998), provide a correlation for the past 600 years between the summer temperature proxy from northern boreal forest tree rings to volcanic activity from historical and ice core records (Fig 3).

The link between volcanism and glaciation was first proposed for the onset of the Pleistocene Northern Hemisphere glaciation (Stewart, 1975), although a possible mechanism wasn’t provided until Bray (1975) suggested how volcanic eruptions could trigger the necessary conditions modelled by Flohn (1974). The rapid glaciation onset model relied on the sudden build up of permanent snow cover across the uplands of the sub-Arctic, which survives through the summer, creating a set of feedbacks that eventually lead to permanent snowfields and ice fields. This is often termed the ‘ice-albedo feedback’, where the build up of snow or ice increases the Earth’s albedo, reflecting more sunlight back into space, and causing further cooling, allowing more snow and ice to build up. The proposed most likely scenario that would allow a snowpack to survive through the summer included greater than normal winter snowfall, followed by a cold, cloudy summer with decreased storms and solar insolation (Flohn, 1974). Volcanic eruptions could replicate all of these phenomena.
Since this initial mechanism was proposed, intense volcanic episodes have been suggested as the cause of glaciations over a range of magnitudes. One such study aims to link volcanism to one of the largest glaciations ever experienced on Earth, through the so-called ‘volcanic winter to snowball Earth’ hypothesis (Stern, 2008). Since individual eruptions cannot be discerned so far back in time, volcanic rocks are quantified as a proxy, and a value for magmatic flux is derived for the Neoproterozoic (1000-542 Ma). The study finds correlation between episodes of high volcanic activity and glaciations, and one period with a correlative lack of both. Mechanisms are also proposed for each episode of high activity, such as the breakup of the Rodinian continent. Marinoan (∼635 Ma) glaciation in particular corresponds to a peak time of subduction-related igneous activity in the Arabian-Nubian Shield and the East African uplift.
This appears to be promising mechanism; however, a link cannot be proven. There does appear to be a correlation between increases in volcanic activity and glacial phases, but correlation does not prove causality, and indeed in one of the first papers published on the link in the Pleistocene, it was suggested that glaciations had caused the volcanic activity increase, and not the other way round (Kennett and Thunell, 1975). Glacio-eustatic changes in sea level may have caused stress release in the mantle, enhancing volcanism. This left researchers with a ‘chicken-and-egg’ challenge of which came first: the glaciers or the volcanism. It may be that they occur as part of a positive feedback system, where glaciations cause large volcanic eruptions, which enhance the cooling and consequently accelerate the growth of ice sheets, and so on. This has been proposed as a mechanism that caused the Toba super-eruption 73,000 years ago, which in turn accelerated the shift towards more glacial conditions (Rampino and Self, 1992). However, when the climate is not already well within a shift or a glaciation, the chicken-and-egg issue returns, a problem which is particularly difficult to address the further back in time we go. This is because the temporal resolution of records of volcanism and glaciation become coarser, hampering precise timing of events. However, for the past ~100,000 years, several archives record both volcanism and climate indicators at sufficient resolution to begin to discern the relative timing of volcanic pulses and the onset of glaciations.

Returning to the Pleistocene, the recent North Pacific Ocean leg 45 of the Ocean Drilling Project has produced three cores with records of both volcanism, in the form of visible ash from the North Pacific Rim; and glaciation, in the form of dropstones or ice rafted debris. The relative positions and concentrations of these two indicators have been compared in order to try to prove that volcanism came first (Prueher and Rea, 1998, 2001). The sudden increase of continentally-derived material at around 2.65 Ma represents a shift from non-glacial to glacial conditions, and the core chronology suggests this change happened in as little as 1000-2000 years. This proves that another mechanism is required to push the climate over the threshold into glaciation, since previously proposed scenarios, such as tectonic uplift (Raymo et al., 1988) and changes in orbital forcing regimes (Maslin et al., 1998), are too slow to account for this rapid shift. The authors suggest crustal unloading caused by sea level changes probably did not cause volcanism to increase after the start of the glaciation in this area, since crustal loading by ice sheets likely balanced this change. This is backed up by the appearance of volcanic ash layers immediately preceding the sudden increase in ice rafted debris in two of the cores, although in the third core the events appear synchronous. Marine cores are advantageous as they record proxies for glacial conditions and actual volcanic ash, have reasonably well-constrained age models, and simple stratigraphy to help distinguish events. However, their age constraints become less well defined further down the core (and thus further back in time), and at present, good quality long cores are sparse. Ice core records help to solve some of these issues.

The detection of volcanic signals in ice cores depends on the acidity of eruptions, formed by the sulphur aerosols. A first attempt at quantifying the acidity of ice cores, and relating peaks to volcanic activity, was carried out on a core from Crete, Greenland, which reached back to 553 AD (Hammer et al., 1981). Acidity was measured using conductivity meters, which have the advantage of not being influenced by carbon dioxide or sea salt components of ice. Since the record includes mostly historical volcanic records, peaks could be attributed to certain eruptions, and periods of low activity could be used to determine background acidity, to aid in volcanic signal selection. Ice cores also record several proxies for temperature (e.g. oxygen isotopes) and climatic conditions (e.g. dust concentrations), allowing meaningful comparisons between volcanic and climatic histories. This method was a success, and acidity measurements have been used in most major ice cores ever since. However, when the records from the Arctic and Antarctic ice cores are compared, major discrepancies in recorded events occur (Fig 4). This highlights the issue of preservation potential for volcanic signals, in terms of magnitude and plume direction, but perhaps most importantly, in terms of eruption location. A large acidity peak at one pole and not the other logically represents a high-latitude eruption in one hemisphere. Acidity peaks at both poles most likely represent a large low-latitude or equatorial eruption that spread its products over both hemispheres. However, it may also represent simultaneous high-latitude eruptions in both hemispheres (Schneider et al., 2009). This may help to explain why some seemingly very large events, such as the 1258 event, estimated as having released 5x1014 -2x1015 kg of material, has yet no single source region (Oppenheimer, 2003). Information on the source location of volcanic eruptions is crucial in assessing their role in influencing climate, since tropical eruptions are more likely to affect global temperatures.
If volcanoes are sufficiently close to an ice sheet, ash from eruptions will also be trapped in ice cores, as well as acidity peaks. Where ash can be recovered, its chemistry can be determined, and since each volcanic centre has a unique geochemistry, the ash can be matched to a particular source location (Zielinksi et al., 1995). In some cases, a volcano will produce a chemically distinct ash every eruption, and for some well-studied centres, geochemistry can be matched to certain events. Even without this, the well-constrained ice core chronologies and source location are enough for a primary assessment of relative contribution to climatic forcing. Finding identical chemical layers in different cores can also help with time correlation between records, if the ice chronology is not precise, as often happens nearer the base of cores. This method has previously relied on visible ash layers, but recent advances in ‘cryptotephra’ (ash invisible to the naked eye) detection have allowed recognition of many more direct volcanic records in ice cores and other media (Turney et al., 1997; Kuehn et al., 2009). Although cryptotephra extraction and analysis is at present time-consuming and complex, the contribution to our understanding of volcanic forcing of climate will be invaluable.
The volcanic geochemistry of certain ash layers can also be correlated between different records of palaeoclimate, better constraining the timing of abrupt climate changes. However, even as timing and records improve, providing a causal link still requires proof that volcanic forcing is sufficient to cause the cooling signals observed in the data when the two occur together. Ever-evolving climate models allow volcanic input to be considered in the climate system. One of the earliest attempts to model the influence of volcanism on a known cooling event was for the Little Ice Age (Robock, 1979). The Little Ice Age was a period roughly between 1430-1850 when temperatures were colder than before or after it. Most data was originally collected in Europe however, and since then the period has been reconsidered as at best a northern hemisphere event, and did not cool globally. However, the temperature anomaly is still significant to require explanation, and Robock compared model results to hemispheric temperature reconstructions of the past 400 years. He tested the forcing of volcanic dust, variations in the solar constant (dependant on sunspot number cycles), variations in CO2, and natural variability. The volcanic dust experimental run produced the best fit to the climate data and reconstructions for the past 400 years.
The volcanic influence has been tested most recently in the National Center for Atmospheric Research’s Community Climate System Model, Version 3 (Schneider et al., 2009). The model has fully coupled atmosphere, ocean, sea ice and land components, and was used to test forcing within the temperature reconstructions of the past 2000 years. The focus was a comparison between the affects of high-latitude versus low-latitude (tropical) eruptions. The conclusions were that in both scenarios the Arctic was the most impacted latitude band in the winter, due to ice-albedo feedbacks and changes in atmospheric and oceanic circulation. This includes the large contribution of increased sea ice growth following an eruption to the persistence of the cooling signal. When comparing the scenarios, tropical eruptions were shown to have much greater impact, and that high-latitude eruption impacts were relatively short-term, even in the Arctic. The model also predicts the temperature and climate signatures that would be created if the 1258 event were a large tropical eruption. If such climate signals can be found in proxy records, this provides another way to determine the source area of the event, should the specific volcano not be identified via geochemical ash fingerprinting.

The aerosol products of large, explosive volcanic eruptions are known to cause a cooling of the lower atmosphere, creating lower surface temperatures, typically on the order of 0.5-1°C for up to 5 years. The possibility that sustained groupings of large eruptions could prolong and intensify the cooling signal permits the idea that they could be a major contributor to global temperature trends on decadal-centennial time scales. The increased impact of any eruption on the high-latitudes of the Arctic also lends strength to the idea that eruptions can cause the snow and ice balance of the Arctic to change, and increase towards a runaway ice-albedo feedback that leads to glaciation. Correlations between high volcanic activity and colder climates have been tested in a number of ways. In ancient Earth history, volcanic rocks are used as a proxy for volcanic flux, and compared to glacial geology. In more recent history, direct climate data has been compared to volcanic written records, and extending beyond this requires the use of proxy records. Volcanic activity can be recognised in records by visible or micro ash layers, or by peaks in acidity in ice cores. Past climate can be indicated in any number of ways, but most useful are proxies for temperature (e.g. oxygen isotopes and tree ring densities) and glacial conditions (e.g. ice rafted debris in marine cores). The use of well-dated records has allowed a reasonable consideration of whether volcanicity does increase before major glaciation episodes, but a correlation does not prove a cause. To test whether eruptions could realistically produce the strength of cooling needed to initiate glaciation, climate models are used to test the reproducibility of cooling signals with volcanic forcing. This would appear to be an initial success; the current most sophisticated climate model has shown a good agreement between the forcing power of eruptions (particularly originating from the tropics) and cooling signatures for the past 2000 years.
_______