Climate changes across the last 24,000 years provide key insights into Earth system responses to external forcing. Climate model simulations 1, 2 and proxy data 3-8 have independently allowed for study of this crucial interval; however, they have at times yielded disparate conclusions. Here, we leverage both types of information using paleoclimate data assimilation 9, 10 to produce the first observationally constrained, full-field reanalysis of surface temperature change spanning the Last Glacial Maximum to present. We demonstrate that temperature variability across the last 24 kyr was linked to two modes: radiative forcing from ice sheets and greenhouse gases; and a superposition of changes in thermohaline circulation and seasonal insolation. In contrast with previous proxy-based reconstructions 6, 7 our reanalysis results show that global mean temperatures warmed between the early and middle Holocene and were stable thereafter. When compared with recent temperature changes 11 , our reanalysis indicates that both the rate and magnitude of modern observed warming are unprecedented relative to the changes of the last 24 kyr.

The Greenland ice sheet (GrIS) is a growing contributor to global sea-level rise 1 , with recent ice mass loss dominated by surface meltwater runoff 2,3 . Satellite observations reveal positive trends in GrIS surface melt extent 4 , but melt variability, intensity and runoff remain uncertain before the satellite era. Here we present the first continuous, multi-century and observationally constrained record of GrIS surface melt intensity and runoff, revealing that the magnitude of recent GrIS melting is exceptional over at least the last 350 years. We develop this record through stratigraphic analysis of central west Greenland ice cores, and demonstrate that measurements of refrozen melt layers in percolation zone ice cores can be used to quantifiably, and reproducibly, reconstruct past melt rates. We show significant (P < 0.01) and spatially extensive correlations between these ice-core-derived melt records and modelled melt rates 5,6 and satellite-derived melt duration 4 across Greenland more broadly, enabling the reconstruction of past ice-sheet-scale surface melt intensity and runoff. We find that the initiation of increases in GrIS melting closely follow the onset of industrial-era Arctic warming in the mid-1800s, but that the magnitude of GrIS melting has only recently emerged beyond the range of natural variability. Owing to a nonlinear response of surface melting to increasing summer air temperatures, continued atmospheric warming will lead to rapid increases in GrIS runoff and sea-level contributions.Melting across higher elevations of the GrIS results in liquid water infiltration, percolation, and either refreezing or storage within the porous firn layer. Such processes reduce ice-sheet surface albedo, increase firn temperatures 7 , and may generate impermeable ice layers that exacerbate ice-sheet runoff 8 , the proportion of surface melt leaving the ice sheet. Runoff across the margins of the GrIS is presently the leading source of mass loss from the ice sheet 2,3,6 , and has been implicated in the centennial-scale slowdown of the overturning circulation in the North Atlantic Ocean 9 .GrIS surface melting in 2012 was more expansive than at any time over the 40 years since we started measuring melt using satellites (in 1978) 4 . Proposed mechanisms driving melt in 2012 include contributions from anomalous radiative 10-12 and non-radiative 13 energy fluxes, and anticyclonic atmospheric circulation favouring advection of warm, dry air and clear sky conditions 14 . An ice core record from Summit Station (Fig. 1) demonstrated the exceptional nature of 2012 melt at high elevation (about 3,200 m), revealing 15 that it last occurred at this site in 1889. At lower elevations, where melt occurs more frequently and at greater rates, there exist only limited ice-core-derived reconstructions of melt variability [16][17][18] and no quantifiable ice-core-based reconstructions of melt intensity or runoff. Furthermore, large discrepancies among reanalysis products over Greenland before the midtwentieth century 6,19 limi...

No abstract

Climate changes across the last 24,000 years provide key insights into Earth system responses to external forcing. Climate model simulations and proxy data have independently allowed for study of this crucial interval; however, they have at times yielded disparate conclusions. Here, we leverage both types of information using paleoclimate data assimilation to produce the first observationally constrained, full-field reanalysis of surface temperature change spanning the Last Glacial Maximum to present. We demonstrate that temperature variability across the last 24 kyr was linked to two modes: radiative forcing from ice sheets and greenhouse gases; and a superposition of changes in thermohaline circulation and seasonal insolation. In contrast with previous proxy-based reconstructions our reanalysis results show that global mean temperatures warmed between the early and middle Holocene and were stable thereafter. When compared with recent temperature changes, our reanalysis indicates that both the rate and magnitude of modern observed warming are unprecedented relative to the changes of the last 24 kyr.

atellite observations 1 , ice core records 2 and climate models 3 have revealed accelerating mass loss of the Greenland Ice Sheet (GrIS) during recent decades, as well as widespread thinning, receding and speeding up of Greenland's marine-terminating outlet glaciers 4 . These glaciological changes directly contribute to global sea level rise 5 , impacting ocean overturning 6 and marine ecosystems downstream 7,8 . Yet whereas such observations highlight the sensitivity of the GrIS to industrial-era Arctic warming 9,10 , especially across daily to interannual timescales 3,4 , little is known of the long-term (multidecadal to centennial) response of Greenland's marginal environments and its peripheral glaciers and ice caps (GICs) to climatic forcing. Given recent findings that GICs accounted for upwards of 20% of Greenlandic ice losses during the early 21st century (despite encompassing less than 5% of the GrIS area 11,12 ), it is important to reconcile such uncertainties by placing contemporary GIC observations into a longer-term perspective.Naturally derived climate proxies offer the potential to extend our understanding of past GIC-climate coupling well beyond the satellite era. However, existing records are limited in scope and, in many regions, provide conflicting information. Across the climatically sensitive coastal west Greenland (CWG) and northeastern Canadian Arctic regions (Fig. 1a), for example, recent studies using proglacial sediments [13][14][15] and mosses 15,16 have provided critical age constraints on the timing of GIC expansion during the previous two millennia. These studies reveal intervals of glacier advancement during the Medieval Warm Period (MWP), a period of widespread relative warmth, as well as during the colder Little Ice Age (LIA). By assuming that GIC growth primarily coincides with declining summertime temperatures, such findings have given rise to the notion that 'paradoxical' (that is, relatively cool) climate conditions existed across CWG and northeastern Canada during the MWP [13][14][15] . However, evi-