Constraining the Source Distribution of Meltwater Pulse 1A Using Near- and Far-Field Sea-level Data

Abstract

Meltwater pulse 1A (MWP-1A) is the largest land ice melt event of the last deglaciation. In a period of no more than 340 years, between 14.65 and 14.31 ka (Dechamps et al, 2012), ~10% of the total deglacial sea-level rise occurred (Hanebuth et al, 2000; Peltier and Fairbanks, 2006; Deschamps et al, 2012), resulting in the highest reported rate of global mean sea-level rise in the geological record, which may have exceeded 4 m per century (Deschamps et al, 2012). Yet, the implications of MWP-1A for constraining the rates of the underlying processes and its role in the sequence of climate events during Termination 1 remain unclear due to the lack of information on its melt source distribution. While glacial isostatic adjustment (GIA) modelling experiments (Clark et al, 2002; Bassett et al, 2005; Deschamps et al, 2012) and recent assessments of ice-sheet histories (Carlson and Clark, 2012) suggest that at least 50% of the event may have come from Antarctica, other interpretations of Antarctic ice-extent and sea-level records suggest a substantially smaller (including zero) Antarctic contribution (Ackert et al, 2007; Mackintosh et al, 2011; Whitehouse et al, 2012). In this study, we show that after reassessments of local MWP-1A amplitudes at Barbados and Sunda Shelf based on the well-constrained timing derived from the Tahiti sea-level record (Deschamps et al, 2012), the sea-level data from Barbados, Sunda Shelf, and Tahiti do not provide as tight of a constraint on the Antarctic contribution as previously thought. We find that between 1 to 10 m sea-level equivalent (sle) could have melted from the Antarctic, compared to 7 to 15 m sle from previous analyses (Clark et al, 2002; Bassett et al, 2005; Deschamps et al, 2012). To better constrain the source of MWP-1A, we also consider sea-level data from Scotland (Shennan et al, 2000), which have, until now, been excluded from MWP-1A fingerprinting experiments because they are strongly influenced by local ice unloading. To overcome this, we isolate the elastic MWP-1A amplitude (i.e. fingerprint signal) at this location using a suite of models that provide optimal fits to the Scottish data, and thereby remove near-field contamination. Preliminary results show that the inclusion of these data leads to an improved MWP-1A source distribution constraint compared to that obtained using the far- and intermediate-field data alone.