Feedback between Volcanism and GlacialInterglacial CO 2 Peter

Feedback between Volcanism and Glacial/Interglacial CO 2 Peter Huybers and Charlie Langmuir Harvard University October 2009 (photo from MODIS)

Ice-volume and CO 2 covary in near lock-step during the Pleistocene… Puzzles: 1. Why does CO 2 change? 2. What links CO 2 to ice-volume? 3. Why is the record saw-toothed?

Of course it is the ocean… …but could there be a volcanic contribution?

(Glazner et al. , GRL 1999)

(Jellinik, Manga, and Saar, JGR 2004)

Episodic Quaternary volcanism in France and Germany (Laacher See, 12. 90 Ka) (Nowell, Jones, and Pyle, JQS 2006)

~65% of eruptive volume occurs within a few thousand years of deglaciations (Bacon and Lanphere, GSA 2006)

A physical basis for deglaciation causing mantle melting Unloading 1 km of ice depressurizes the underlying mantle column equivalent to 1/3 km of mantle ascent. Because melt generation rates are ~0. 3% per km, 1 km of ice unloading melts ~0. 1% of the material in the melt zone. A melt zone 100 km thick then yields 100 m of melt.

Overview 1. Deglaciation is accompanied by increased subaerial volcanism on a global scale. 2. Using estimates of modern volcanic CO 2 fluxes, we suggest the deglacial volcanic pulse contributes ~40 ppm of the deglacial rise in atmospheric CO 2. 3. Volcanic CO 2 would then constitute a positive feedback upon deglaciation and reglaciation --- assuming an increased ice load also suppresses volcanism. (Cotopaxi Volcano, Ecuador)

Temporal distribution of observed subaerial volcanic events Data is from Volcanoes of the World dataset of Siebert and Simkin (2002) and a compilation by Bryson, Bryson and Reuter (2006).

Eruption frequency (events/ky) during the glacial (40 -20 ka)

Eruption frequency during the deglaciation (18 -7 ka)

Eruption frequency during the late Holocene (5 -0��ka)

Eruption frequency ratio: deglaciation/glaciation

Eruption frequency ratio: deglaciation/late Holocene

Ice mass balance from the NCEP reanalysis

A strong relationship exists between the activity ratios and a proxy for changes in glaciation

Rates of observed eruptions glaciated volcanoes unglaciated volcanoes But how to correct for the temporal reporting bias?

An approximate method to correct for the observational bias 1. Model of observational bias 2. Estimate of glaciated activity 3. Estimate of global activity 4. Fractional deviations

Volcanic activity over the last 40 Ka.

Comparison with the GISP 2 volcanic sulphate record (Zielinski, QSR 2000)

Timing of deglaciation in volcanic regions ~12 Ka (Yu et al. , 2008) ~12 Ka (Maclennan et al. , 2002) ~12 Ka (Dyke, 2004) 17 to 11 Ka (Mc. Culloch et al. , 2000) 1� 8 to 6 Ka (Boulton et al. , 2001)

How much CO 2 do subaerial volcanoes emit? • Modern magmatic production at arcs is ~3 km 3/yr (Crisp, 1984; Carmichael, 2002; Dimilanta, 2002), times 1% CO 2 in primary magmas puts emissions at 0. 1 Gt CO 2 per year (Marty and Tolstikshin, 1998; Fischer et al. , 1998). Subaerial emissions at non-converging margins may contribute another 0. 05 Gt CO 2 per year (Hilton et al. , 2002). ~0. 15 Gt CO 2 / yr • The CO 2/Nb ratio of mantle derived rocks is ~500. 3 ppm Nb in arc magmas leads to 0. 15% mantle CO 2. p. CO 2/3 He suggests 80 -90% of CO 2 is not mantle derived, hence 1% CO 2 in arc magmas, or ~0. 15 Gt CO 2 / yr. • Direct measurements of CO 2 indicate a range of 0. 05 to 0. 15 Gt CO 2 per year (Goff et al. , 2001; Varekamp et al. , 1992; Sano and Williams, 1996; Gorman et al. , 2006). Mt Etna alone was observed to emit ~0. 04 Gt CO 2 per year (Williams, 1992). 0. 05 to 0. 15 Gt CO 2 / yr • Scaling our estimate of global volcanic rates by 0. 1 to 0. 15 Gt CO 2/yr for modern rates and using mass-balance thresholds of between -6 m/yr and -9 m/yr leads to between 1000 to 5000 Gt CO 2 emitted above a baseline scenario of current emissions during the last deglaciation.

CO 2 distribution between the atmosphere and ocean: a simple, two-box representation. Atmosphere (a) Vt--Wo Ft Ocean (b) q is the airborne fraction (0. 1 to 0. 15) is a timescale (300 to 2000 yrs)

Volcanogenic contribution to atmospheric CO 2

And what about CO 2 emissions at mid-ocean ridges? • Modern magmatic production at mid-ocean is ~20 km 3/yr (e. g. Chen, 1996), while primary CO 2 content is only ~0. 1%, putting emissions at ~0. 1 Gt CO 2 per year (Fischer et al. , 1998; Saal et al. , 2002; Marty and Tolstikshin, 1998; Cartigny et al. , 2007), similar to arc emissions. • The ~130 m deglacial rise in sea level is equivalent to suppressing 45 m of mantle ascent. Given an ascent rate of ~3 cm/yr, the sea-level rise then suppresses ~1500 yrs of emissions at mid-ocean ridges, or 150 Gt CO 2. Some further questions: • What is the relationship between eruptions, magma production, and CO 2 emissions? • What is the role of volcanic emissions in controlling atmospheric CO 2, relative to (and in combination with) changes in the storage of carbon in the oceans and biosphere coming out the last glacial?

Conclusions 1. Global volcanism apparently increased by two to five fold between 12 to 6 Ka, roughly consistent with deglacial forcing. 2. Such an increase in volcanism is consistent with an additional 1000 to 5000 Gt of CO 2 emissions. The timing of these emissions is consistent with the secondary rise in atmospheric CO 2 during the last deglaciation. 3. By similar logic, a deficit of volcanic emission relative to weathering during reglaciation would tend draw down CO 2. 4. Thus, volcanism may forge a link between glacial variability and atmospheric CO 2 concentrations.

A test of the glacio-volcano-CO 2 hypothesis

Amplitude of deglaciation versus interglacial trends in CO 2