Clim. Past, 18, 485–506, 2022
© Author(s) 2022.
This work is distributed under
the Creative Commons Attribution 4.0 License.
Large volcanic eruptions occurring in the last glacial period can be detected by their accompanying sulfuric acid deposition in continuous ice cores. Here we employ continuous sulfate and sulfur records from three Greenland and three Antarctic ice cores to estimate the emission strength, the frequency and the climatic forcing of large volcanic eruptions that occurred during the second half of the last glacial period and the early Holocene, 60–9 kyr before 2000 CE (b2k). Over most of the investigated interval the ice cores are synchronized, making it possible to distinguish large eruptions with a global sulfate distribution from eruptions detectable in one hemisphere only. Due to limited data resolution and large variability in the sulfate background signal, particularly in the Greenland glacial climate, we only list Greenland sulfate depositions larger than 20 kg km−2 and Antarctic sulfate depositions larger than 10 kg km−2. With those restrictions, we identify 1113 volcanic eruptions in Greenland and 737 eruptions in Antarctica within the 51 kyr period – for which the sulfate deposition of 85 eruptions is found at both poles (bipolar eruptions). Based on the ratio of Greenland and Antarctic sulfate deposition, we estimate the latitudinal band of the bipolar eruptions and assess their approximate climatic forcing based on established methods. A total of 25 of the identified bipolar eruptions are larger than any volcanic eruption occurring in the last 2500 years, and 69 eruptions are estimated to have larger sulfur emission strengths than the Tambora, Indonesia, eruption (1815 CE). Throughout the investigated period, the frequency of volcanic eruptions is rather constant and comparable to that of recent times. During the deglacial period (16–9 ka b2k), however, there is a notable increase in the frequency of volcanic events recorded in Greenland and an obvious increase in the fraction of very large eruptions. For Antarctica, the deglacial period cannot be distinguished from other periods. This confirms the suggestion that the isostatic unloading of the Northern Hemisphere (NH) ice sheets may be related to the enhanced NH volcanic activity. Our ice-core-based volcanic sulfate records provide the atmospheric sulfate burden and estimates of climate forcing for further research on climate impact and understanding the mechanism of the Earth system.How to cite. Lin, J., Svensson, A., Hvidberg, C. S., Lohmann, J., Kristiansen, S., Dahl-Jensen, D., Steffensen, J. P., Rasmussen, S. O., Cook, E., Kjær, H. A., Vinther, B. M., Fischer, H., Stocker, T., Sigl, M., Bigler, M., Severi, M., Traversi, R., and Mulvaney, R.: Magnitude, frequency and climate forcing of global volcanism during the last glacial period as seen in Greenland and Antarctic ice cores (60–9 ka), Clim. Past, 18, 485–506, https://doi.org/10.5194/cp-18-485-2022, 2022.1 Introduction
The dispersal of gas, aerosols and ash particles by volcanic eruptions plays a major role in the climate system (Gao et al., 2007; Robock, 2000). Large volcanic eruptions injecting sulfuric gases into the stratosphere and forming sulfate aerosols have a global or hemispheric cooling effect of several degrees lasting for several years after the eruption (Sigl et al., 2015; Sinnl et al., 2021).
Estimations of volcanic stratospheric sulfur injections and of the timing and frequency of large volcanic eruptions are essential for the ability to understand and model past and future global climate conditions (Timmreck et al., 2016). For the last 1200 to 2500 years, the ice-core-based volcanic forcing records derived from Greenland and Antarctica (Crowley and Unterman, 2013; Gao et al., 2008; Toohey and Sigl, 2017) provide an essential forcing record for climate model simulations (Jungclaus et al., 2017), supporting detection and attribution studies (Schurer et al., 2014), including those applied in the IPCC. However, so far the global ice-core-based volcanic record of the last glacial period has been poorly documented.
1.1 Ice-core records of volcanic sulfate deposition
Several studies have reconstructed the volcanic sulfate deposition for part or all of the Holocene in Greenland (Cole-Dai et al., 2009; Gao et al., 2008; Sigl et al., 2013) or in Antarctica (Kurbatov et al., 2006; Castellano et al., 2004; Plummer et al., 2012; Nardin et al., 2020; Cole-Dai et al., 2021). Sigl et al. (2015) applied accurately dated ice cores synchronized between the two hemispheres to reconstruct global volcanism over the last 2500 years. This so-called bipolar synchronization allows distinguishing large global eruptions from those of hemispheric or more regional impact. During the last 2500 years, they identified 50 global (bipolar) volcanic eruptions, 5 of which had a sulfur emission strength larger than or similar to the Tambora eruption occurring in Indonesia in 1815 CE. Prior to the last glacial maximum no bipolar volcanic sulfate deposition record is currently available from ice cores.
One conclusion drawn from historical eruptions is that there is a significant variability of the same volcanic event in the sulfate deposition records derived from different ice cores on both a regional and a local scale (Sigl et al., 2014; Gao et al., 2007). Some of this regional variability can be explained by the difference in sulfate deposition fluxes at different locations. For example, in Antarctica where geographical distances are large, the sulfate deposition at a specific site will be strongly dependent on factors such as the location of the eruption, governing wind patterns and seasonality. Another reason for the lateral sulfate deposition variability is the amount and patchiness of snowfall, which may locally enhance the sulfate deposition for high snowfall areas compared to low snowfall areas for a volcanic event. Moreover, there may be more absent sulfate deposition events caused by post-depositional processes on the snow surface, such as wind erosion (Gautier et al., 2016). The spatial variability of sulfate deposition in Antarctica was studied at 19 sites covering the past 2000 years by Sigl et al. (2014), and here both accumulation and post-deposition effects were found to be important factors. In particular, on the East Antarctic Plateau where snow accumulation is very low, the sulfate deposition is lower than at more coastal and higher-accumulation sites in Antarctica. The snow accumulation effect is also observed in Greenland (Gao et al., 2007), although the effect is much less pronounced here because the accumulation rates are less variable in central Greenland than in different parts of Antarctica. In order to reduce the accumulation bias, Gao et al. (2008) selected five large low-latitude volcanic events from 54 Arctic and Antarctic ice cores and calculated the mean ratio of deposition in individual ice cores; they then applied the deposition ratio between different cores to correct the sulfate deposition for all events in all cores to obtain the Arctic and Antarctic mean sulfate depositions. In general, it is clear that more robust volcanic deposition patterns can be obtained when larger sets of ice cores are included, and preferably ice cores from high-accumulation sites should be applied (Gao et al., 2007; Sigl et al., 2014).
One complication related to the derivation of volcanic sulfate deposition in ice cores is the thinning of the ice layers with increasing depth and age. Due to glacier flow, the annual layers and thus the volcano-derived sulfate deposition become thinned with depth, an effect that is most pronounced at high-accumulation sites and close to bedrock. In central Greenland, typical thinning rates of annual layers in the 60–10 ka range are 50 %–90 % depending on age and local flow conditions (Johnsen et al., 2001). To calculate the sulfate deposition of a specific eruption from a measured ice concentration a correction for the thinning at the corresponding depth is needed to obtain the past accumulation rate at the time of snowfall. Thinning functions are obtained from ice-flow modeling, and thus there is a site-specific dependency on accurate flow modeling associated with the sulfate deposition determination.
1.2 Studies of the frequency of volcanic eruptions
The volcanic sulfate record of the Greenland GISP2 ice core has been investigated by Zielinski et al. (1997), who found that there was increased volcanic activity during the deglacial period (22–8 ka b2k) compared to the average activity of the last glacial cycle. This is interpreted as being related to the tectonic isostatic response to the melting of the large ice sheets during that period. Based on the global volcanic databases (Siebert and Simkin, 2002; Bryson et al., 2006), Huybers and Langmuir (2009) found that volcanism increased 2 to 6 times during the deglacial period of 12–7 ka b2k compared to the average level of eruptions during the 40–0 ka b2k interval.
In Antarctica, Castellano et al. (2004) determined the frequency of volcanic eruptions over the last 45 kyr based on the EPICA Dome C (EDC) ice core. They found a rather constant level of volcanic activity throughout that period except for the most recent millennia, when the activity shows an increase. Kurbatov et al. (2006) detected volcanic signals during the last 12 kyr in the Siple Dome A ice core from West Antarctica. They found that the number of volcanic sulfate signals is decreasing with age, possibly related to the relatively low sampling resolution in the deeper part of that core. Recently, Cole-Dai et al. (2021) used the high-accumulation WAIS Divide ice core to determine a fairly constant Holocene eruption frequency with larger-than-Tambora (1815 CE) events occurring approximately once per millennium. Note, however, that all these reconstructions differ in their volcanic signal detection method, which may lead to different trends in peak frequencies.
1.3 Volcanic events identified in ice cores with tephra and sulfate peak synchronization
The ice-core volcanic source identification is important as it helps to constrain the magnitude – interpreted here as sulfur emission strength rather than the mass of material erupted (Pyle, 2015) – and the climate forcing of the eruption. Furthermore, it allows for a more detailed comparison to modeling studies. In historical times, the volcanic origin of an ice-core acidity spike may be corroborated by a precise dating of the ice core (Sigl et al., 2015). Further back in time, as the uncertainty of both the ice-core dating and the identification of the erupting volcanoes increases, the origin of a volcanic ice-core layer can only be determined if it is associated with a volcanic ash (tephra) deposition in the ice (Gronvold et al., 1995). However, tephra layers do not always coincide with sulfate peaks (Davies et al., 2010), and most volcanic sulfate signals have no tephra associated with them.
In the last glacial period, many Greenland tephra deposits have been associated with Icelandic eruptions, while around a dozen of identified tephra layers originate in North America and eastern Asia (Abbott and Davies, 2012; Bourne et al., 2015; Davies et al., 2014). In Antarctica, tephra layers have been identified and associated with eruptions occurring within Antarctica and in the Southern Hemisphere (Narcisi et al., 2005, 2010, 2012). Recently, Mcconnell et al. (2017) identified tephra from the long-lasting and halogen-rich Antarctic Mount Takahe eruption that occurred around 17.80 ka. Tephra of the Oruanui eruption from the Taupo volcano in present-day New Zealand has been identified and dated to 25.32 ka before 1950 CE (BP) in the West Antarctic Ice Sheet Divide ice core (WDC) (Dunbar et al., 2017).
Volcanic eruptions generally do not deposit tephra in both Greenland and Antarctica, so the bipolar synchronization of sulfur spikes in the ice cores is dependent on an alternative matching technique. Svensson et al. (2020) applied annual layer counting in both Greenland and Antarctic ice cores to match patterns of volcanic eruptions leading to the identification of some 80 bipolar eruptions in the 60–12 ka interval. For the Holocene, a bipolar synchronization of volcanic eruptions was released with the AICC2012 timescale (Veres et al., 2013). Using sulfur isotopes, it has recently become possible to test if sulfate has indeed reached the stratosphere, which is a prerequisite for being globally distributed, as the sulfate undergoes characteristic isotope fractionation in the stratosphere (Burke et al., 2019; Gautier et al., 2018; Crick et al., 2021; Baroni et al., 2008), but these analyses are still scarce for the last glacial period.
1.4 Extending the ice-core volcanic record into the last glacial period
Here we extend the ice-core record of sulfate deposition in Greenland and Antarctica by employing sulfate records from three Greenland and three Antarctic ice cores in the interval 60–9 ka (in one core we use elemental sulfur measurements, but for the sake of brevity we will refer to sulfate records). We investigate the sulfur emission strengths (i.e., defining the climate impact potential) and the frequency of volcanic eruptions detected in either Greenland or Antarctica. For eruptions identified in both hemispheres, we estimate the climate forcing using modern analogs and determine the occurrence of very large eruptions. Unless otherwise stated, all ages provided in this work are relative to the year 2000 CE.
UNIVERSITY OF COPENHAGEN – FACULTY OF SCIENCE
Ice cores drilled in Antarctica and Greenland have revealed gigantic volcanic eruptions during the last ice age. Sixty-nine of these were larger than any eruption in modern history. According to the University of Copenhagen physicists behind the research, these eruptions can teach us about our planet’s sensitivity to climate change.
For many people, the mention of a volcanic eruption conjures up doomsday scenarios that include deafening explosions, dark ash billowing into the stratosphere and gloopy lava burying everything in its path as panicked humans run for their lives. While such an eruption could theoretically happen tomorrow, we have had to make do with disaster films and books when it comes to truly massive volcanic eruptions in the modern era.
“We haven’t experienced any of history’s largest volcanic eruptions. We can see that now. Eyjafjellajökull, which paralysed European air traffic in 2010, pales in comparison to the eruptions we identified further back in time. Many of these were larger than any eruption over the last 2,500 years,” says Associate Professor Anders Svensson of the University of Copenhagen’s Niels Bohr Institute.
By comparing ice cores drilled in Antarctica and Greenland, he and his fellow researchers managed to estimate the quantity and intensity of volcanic eruptions over the last 60,000 years. Estimates of volcanic eruptions more than 2,500 years ago have been associated with great uncertainty and a lack of precision, until now.
Sixty-nine eruptions larger than Mount Tambora
Eighty-five of the volcanic eruptions identified by the researchers were large global eruptions. Sixty-nine of these are estimated to be larger than the 1815 eruption of Mount Tambora in Indonesia – the largest volcanic eruption in recorded human history. So much sulfuric acid was ejected into the stratosphere by the Tambora eruption that it blocked sunlight and caused global cooling in the years that followed. The eruption also caused tsunamis, drought, famine and at least 80,000 deaths.
“To reconstruct ancient volcanic eruptions, ice cores offer a few advantages over other methods. Whenever a really large eruption occurs, sulfuric acid is ejected into the upper atmosphere, which is then distributed globally – including onto Greenland and Antarctica. We can estimate the size of an eruption by looking at the amount of sulfuric acid that has fallen,” explains Anders Svensson.
In a previous study, the researchers managed to synchronize ice cores from Antarctica and Greenland – i.e., to date the respective core layers on the same time scale. By doing so, they were able to compare sulphur residues in ice and deduce when sulfuric acid spread to both poles after globally significant eruptions.
When will it happen again?
“The new 60,000-year timeline of volcanic eruptions supplies us with better statistics than ever before. Now we can see that many more of these great eruptions occurred during the prehistoric Ice Age than in modern times. Because large eruptions are relatively rare, a long timeline is needed to know when they occur. That is what we now have,” says Anders Svensson.
One may be left wondering when the next of these massive eruptions will occur. But Svensson isn’t ready to make any concrete predictions:
“Three eruptions of the largest known category occurred during the entire period we studied, so-called VEI-8 eruptions (see fact box). So, we can expect more at some point, but we just don’t know if that will be in a hundred or a few thousand years. Tambora sized eruptions appears to erupt once or twice every thousand years, so the wait for that may be shorter.”
How was climate affected?
When powerful enough, volcanic eruptions can affect global climate, where there is typically a 5-10- year period of cooling. As such, there is great interest in mapping the major eruptions of the past – as they can help us look into the future.
“Ice cores contain information about temperatures before and after the eruptions, which allows us to calculate the effect on climate. As large eruptions tell us a lot about how sensitive our planet is to changes in the climate system, they can be useful for climate predictions,” explains Anders Svensson.
Determining Earth’s climate sensitivity is an Achilles heel of current climate models. Svensson concludes:
“The current IPCC models do not have a firm grasp of climate sensitivity – i.e., what the effect of a doubling of CO2 in the atmosphere will be. Vulcanism can supply us with answers as to how much temperature changes when Earths atmospheric radiation budget changes, whether due to CO2 or a blanket of sulphur particles. So, when we have estimated the effects of large volcanic eruptions on climate, we will be able to use the result to improve climate models.”
Climate of the Past
Magnitude, frequency and climate forcing of global volcanism during the last glacial period as seen in Greenland and Antarctic ice cores (60–9 ka)
ARTICLE PUBLICATION DATE