The True Start of the Greenland Melt Season

Ruth Mottram and Peter Langen, DMI
May 13th 2016

The Greenland melt season has finally begun and then some.

Today we can announce that the Greenland melt season started on the 10th May 2016, tied joint 5th with 1999 in the rankings of early starts to the melt season.

The top 5 earliest official starts are thus:


  1. 1996 (29th April)
  2. 2010 (2nd May)
  3. 1990 (6th May)
  4. 2006 (7th May)
  5. 2016/1999 (10th May)

The latest start was last year, 12th June, 2015. 

Three days in a row with more than 5% of the area of the ice sheet melting but the Surface mass balance remains around 0 thanks to the refreezing of meltwater inside the snow layers.

As we have written about before, there is no formal definition for the start of either melt season or ablation season (when the loss of ice by melting consistently exceeds the gain of ice from snowfall) so we developed our own thresholds:


  • Melting season onset: The first day of a period of at least three consecutive days where more than 5% of the ice sheet experiences melting. In our model, we define melting to occur in a place when the melt rate is greater than 1 mm/day.
  • Ablation season onset: The first day of a period of at least three consecutive days where surface mass balance (SMB) is negative and below –1 Gt/day (1 Gt is one billion tons and corresponds to 1 cubic kilometer of water).

In part these thresholds were designed to specifically exclude the early extreme melting events like the one we observed in early April this year, when we came close to three consecutive days over 5% but just missed out. Since early April, a large dump of snow over the ice sheet in early May temporarily halted melting, which resumed in earnest last week.

Our analysis of our model results shows that early melt onset does not predict a high rate of ice loss in the summer ablation season. However it may act to assist in the loss of ice if the right weather conditions coincide. This happens in two ways.

Firstly, in the areas where there has been a high melt there may be enhanced loss of ice early in the season. Most of the meltwater so far has refrozen in the snow pack forming a thick layer of ice that further melt water cannot trickle through, which explains why the surface mass balance is so far close to 0 in spite of the large area melting. The processes that control how quickly meltwater runs off the ice sheet or refreezes in the surface layers are currently the focus of intensive research right now in the Danish funded RETAIN project led by scientists at GEUS, DMI and DTU.

Secondly, bare glacier ice and melting snow are much darker than fresh snow which means that heat from the sun is absorbed at the surface rather than being reflected, leading to more melt known as an albedo feedback. The albedo plot on the polar portal shows that the reflectiveness of the surface is now very low especially around the edges, when compared to the average in the period 2000-2009.

The warm conditions over Greenland as identified in ECMWF forecasts and the anomalously low surface reflectiveness of the ice sheet based on the MODIS satellite data processed by Prof. Jason Box, GEUS.

Ice melt this week has been driven by warm air moving over the ice sheet as in April. On the ice sheet itself for example, the KAN_U PROMICE weather station at 1840m above sea level on the ice sheet has again been measuring temperatures above freezing.

In the North of Greenland however, where Danish colleagues from the EGRIP project are busy preparing to take a new ice core, the cold remains, they tweeted earlier: