For the 19th NSIDC showed a similar melt pattern to the 18th, which only a small uptick on the graph, still around 10% of the sheet showed melt.
DMI seem to be having a few teething problems with the new site and model at the moment (doesn't any new site) so I'll write an update later.
Friv, the DMI Surface Mass Balance is a model based on current weather. Its not going to be right all the time (just as PIOMAS isn't). What they will have been doing is using the data they have got (Weather Stations to measure precipitation, River Flow Measurements, Glacier Flow Measurements, GRACE measurements) to validate the model and tweak it so it approximates the actual value over the year.
The new model counts less of the melt as runoff because it is still on or in the icesheet. It may not have refrozen yet, but there is the potential for it to do so.
Even in 2012 the Surface Mass Balance of Greenland was positive, not negative as you imply (i.e. more mass was deposited on the surface than lost through melt). This needs to be the case unless a sheet is in freefall as this is the only input to the system (glaciers cannot add mass only remove it). SMB was 300Gt less than the average though for the year, and that is a lot of ice not gained.
Recently there is more melt than previously even on more normal years (2013). For SMB at least, Greenland seems to be following the Arctic in having normal freezing seasons followed by exceptional melting seasons.
Measured sea level rise is still only rising at 3mm a year on average, and the majority of that is coming from thermal expansion, not the ice sheets (although the totals attributed to them are rising), the problem comes if this flow keeps on increasing exponentially, or there is a catastrophic collapse (WAIS).