Sea level forcing of mid-ocean ridge magmatism on Milankovitch timescales

It is well-documented that Iceland experienced a pulse of elevated volcanism immediately following the last deglaciation (Maclennan et al., 2002). Modeling results suggest ice sheet retreat depressurized the mantle thus enhancing melt production and the supply of magma to the surface (Jull and McKenzie, 1996). Here we take a similar approach, but instead model the effect of glacial-interglacial changes in sea level on mantle melting at mid-ocean ridges. Loading rates reaching ±2 cm/year of water are comparable to the tectonic unloading rate of ~2 cm/year of mantle rock that drives magmatic activity at a slow-spreading ridge. Although the magnitude of sea level forcing is smaller than subglacial forcing, the sea level effect is globally distributed and could have significant consequences for ocean crust architecture and geothermal heat delivery to the deep ocean. We use a model of melt production based on analytical corner flow velocities coupled to the pMELTS model (Ghiorso et al. 2002; Asimow et al. 2004) of melting of the Workman and Hart (2006) depleted upper mantle source composition. For simplicity we assume that the hydrostatic pressure signal from sea-level variation is felt instantaneously by the entire melting regime, and that melts migrate from source to ridge axis at a constant rate. We neglect crustal magmatic and hydrothermal processes that might damp or delay the signal. We examined mid-ocean ridge systems with half-spreading rates from 30 mm/yr to 100 mm/yr and melt migration rates from 2.5 to 50 m/yr. For the case of 30 mm/yr half-spreading rate and 10 m/yr melt migration, we find that the rate of melt delivery to the crust varies ±30% relative to steady state conditions when the model is driven by a record of sea-level variability for the last 140 kyr. Notably, we simulate that melt delivery increased by ~30% beginning at 75 kyr BP, coincident with a rapid decrease in sea level of approximately 60 m. We also estimate a ~30% increase in melt delivery from 30 kyr BP to 20 kyr BP, driven by a 50 m decrease in sea level. Melt delivery decreased more than 50% from the Last Glacial Maximum to 5 kyr BP, driven by the ~120 m rise in sea level during the deglaciation. Our results suggest that modest changes in hydrostatic pressure driven by ice sheet growth and decay yield substantial alterations in magma flux at mid-ocean ridges relative to steady state conditions. These simulations raise the possibility that mantle melting may act as a negative feedback on ice sheet size by modulating deep ocean temperature. We estimate that enhanced melt production during sea level low-stands increases deep ocean temperature by order 0.1 °C. We speculate this modest warming may contribute to deglaciations by reducing sea-ice extent in the Southern Ocean, which may in turn promote ventilation of the abyssal ocean and the release of sequestered carbon dioxide to the atmosphere. References: Asimow, P.D., et al., 2004. G-cubed 5(1): 10.1029/2003GC000568. Ghiorso, M.S., et al., 2002. G-cubed 3(5): 10.1029/2001GC000217. Jull, M., and McKenzie, D., 1996. JGR 101(B10): 21815-21828. Maclennan, J., et al., 2002. G-cubed 3(11): 10.1029/2001GC000282 Workman, R.K. and Hart, S.R., 2005. EPSL 231(1-2): 53-72.