Background: Greenland's role in the climate system
Sea level rise
More than one quarter of the global sea level rise is attributed to the mass loss of the Greenland Ice Sheet [1,2,3] (Figure 1). Between 1992 and 2011, the contribution of Greenland's melting glaciers to global sea level rise was about 7.5 mm, with a 4-fold increase in their mass loss between the 1990s and the 2000s . This process has even been accelerating its pace in the recent decade [5,6,7] (Figure 2).
Further impacts of glacial melting
In addition to the increasing contribution to sea level rise, the freshwater input of the melting ice sheet into the oceans also has strong implications for ocean currents even on a basin-wide scale [8,9,10,11]. Due to microbial activity at the interface between glacier and the bedrock, the glacial meltwater also transports a significant amount of biologically relevant substances into the ocean, such as micro-nutrients  and dissolved nitrogen . The implications for biological production and abundance of higher organisms are still unknown.
So what is known about the causes of this accelerating mass loss of the Greenland Ice Sheet? The 5th assessment report of the International Panel on Climate Change, IPCC AR5, points out that the increasing contribution of the large ice caps to global sea level rise since the 1990s is linked to the warming of the surrounding oceans . In fact, the largest increase of glacier mass loss is observed close to the ice-ocean interface, at those glaciers that are increasingly retreating, whose flow velocity is increasing rapidly, and those which are substantially thinning [15,16,17,18,19] (Figure 3).
The ice-ocean interface
Despite its small area of only about 5% of the entire Greenland Ice Sheet, the large contribution of peripheral glaciers to the total mass loss (amounting to about a quarter between 2000 and 2011 [20,21]) highlights the importance of processes at the boundaries between ice, ocean and atmosphere.
Recent observations of individual glaciers show that their accelerated flow was accompanied by a warming of the surrounding fjord water [20,21], caused by an increased inflow of warmer ocean water. The presence of such warm water at the base of the floating ice tongues of a marine terminating glaciers leads to its thinning. This process first leads to an acceleration of the ice flow velocity, and finally to a complete disintegration of the glacier tongue. Model studies also suggest that the ocean is the most likely driver of glacier changes , but the underlying mechanisms are still poorly understood due to lack of observational data and a proper process understanding.
To complicate investigations even further, subglacial meltwater entering the ocean from close to the grounding line and partly from the glacier surface may modify the basal melting in the cavity beneath the glacier tongue [25,26,27]. At present, it is not possible to quantify the mass loss due to supraglacial melting or dynamic ice mass loss due to glacier destabilization . The supraglacial contribution also includes the atmosphere as the main driver that influences the glacier surface via complex multiscale processes .
According to the IPCC, the dynamics of marine-terminating glaciers play a highly underestimated role for the mass loss of the Greenland Ice Sheet . In April 2016, during its 43rd meeting, the IPCC decided to draft a special report on "The ocean and the cryosphere in climate change". The simultaneous occurrence of mass loss along the edges of the Greenland ice sheet, the acceleration and retreat of the glaciers on the one hand, and the warming of the North Atlantic on the other, suggest a strong interaction at the glacier-ocean interface. The relevant processes still remain less studied than they should be, and as a consequence, new interdisciplinary research fields have been developing [18,29] (see also GRISOnetwork: http://web.whoi.edu/griso/).
A complex problem
The interactions between an ice sheet and the oceans they extend into are far from being sufficiently understood to provide reliable predictions for future projections of the contribution of ice sheets to sea level changes. The interactions between the Greenland Ice Sheet and the relativeky warm North Atlantic and Arctic Oceans mostly occur through a large number of small-scale peripheral and outlet glaciers. Recent investigations for example in southern Greenland suggest that the warming North Atlantic plays a crucial role in destabilizing the regional ice sheet. A linkage between oceanic warming and increased calving of icebergs is also supported by paleo studies on sediment cores [30,31].
So far, the main loss of ice has been observed on Greenland's southeastern and western outlet glaciers and the peripheral glaciers [22,23,32,33,34], where warm Atlantic water from the subpolar North Atlantic circulates in associated fjords. The retreat of the glaciers in that region coincided with the accumulation and warming of Atlantic water . But in recent years, the warming of the North Atlantic has also reached the Nordic Seas and even the Arctic Ocean [36,37]. This raises the question as to what extent this warming has already reached the marine-terminating glaciers on the northeast coast of Greenland, which were still considered stable until just a few years ago. In this region, the only major ice stream in Greenland, the "Northeast Greenland Icestream" (NEGIS, see also Fig. 5), is also located.
The 79°N-Glacier and Zachariæ Isstrøm
The thickest and most massive glacier that is still in the early stages of retreat is the 79°N glacier (Fig. 4a). For some years, this glacier has been losing mass after an observed acceleration of the glacier flow velocity . The neighboring glacier, the Zachariæ Isstrøm, has lost almost its entire ice tongue in recent years. The increased acceleration and the continous retreat of the grounding line resulted in the glacier becoming a Tidewater glacier today. Both glaciers together drain about 16% of the total Greenland Ice Sheet, so increasing their mass loss significantly changes Greenland's total contribution to sea-level rise. Mouginot et al., (2015) also showed that the surface mass balance, i.e. snowfall and surface ablation is not driving these changes, also showing no negative trend in the context of the decadal and interannual variability. However, due to the observed acceleration, the glacier outflow has increased for both glaciers . It is plausible to hypothesize that this change is caused by warming of both ocean and atmosphere. Due to the lack of process understanding, this hypothesis still remains to be be proven.