In December 2014, I posted the essay Will Greenland Begin Accumulating Ice in 2015 and Beyond. Based on 1) my understanding of Greenland’s icecap dynamics, 2) changes in the North Atlantic Oscillation and its effect on heat transport, and 3) NOAA’s 2014 Arctic Report Card indicating a rapid deceleration in loss of ice mass (an insignificant 6 gigatons Gt), I argued that Greenland could begin accumulating ice starting as early 2015. Ice mass loss can only happen if summer melt exceeds annual accumulation, or if greater basal melting of glaciers and iceberg calving increases glacier discharge and offsets accumulation. However the inflow of warm Atlantic water has been declining, and suggested basal melting and glacier discharge should be decreasing. Furthermore as the high pressure system that governs Greenland’s weather shifted, snow accumulation was once again exceeding summer melt, suggesting a reversion to the mean of 200 GT of net snow accumulation each year.

My prediction stood in stark contrast CO2 global warming projections of accelerated melting of Greenland ice. So I realized my prediction was risky, but based on natural climate dynamics, my prediction seemed very likely; if not in 2015, then in the very near future. I have always believed that making predictions (aka projections) are the best way to test the correctness of our understanding. Besides we often learn more by being wrong, because predictions can be “right” for the wrong reasons. On December 15th NOAA published their 2015 Arctic Report Card reported an increased “ice mass loss of 186 Gt over the entire ice sheet between April 2014 and April 2015” Although that loss was “22% below the average mass loss of 238 Gt for the 2002-2015 period” it was 6.4 times higher than the 29 Gt loss of the preceding 2013-2014 season.” Although my 2015 prediction failed, all the climate dynamics show a decelerating rate of loss and a high sensitivity to natural variability. Those trends still suggest an increase Greenland ice mass in the near future is more likely than ever.

Despite videos and alarmist suggestions from the NY Times that “Greenland is Melting Away”, Greenland’s Surface Mass Balance (SMB) has been increasing recently as seen in Figure 1 from the Danish Meteorological Society (DMI). Typically Greenland steadily accumulates mass from September to June until the summer solstice triggers 2+ months of melting. Despite that period of intensive melting, on average (dark black line) Greenland steadily accumulates a surface mass of 300 Gt of snow each year. That is why the “lost squadron” of WW2 bombers, forced to land in Greenland in 1942, were recently located beneath 260 feet of snow. Global warming hypotheses have suggested increasing warmth would increase the melt season, ultimately offsetting any accumulation as observed in 2011-2012. However as climate experts and NOAA’s report card suggest, it is the North Atlantic Oscillation that controls the melt season by affecting the position of a blocking high-pressure system around Greenland.

In 2011-2012 Greenland experienced an extreme heat wave that melted away that year’s accumulated snow (red curve in Fig. 1). As discussed in Hanna 2013, a blocking high-pressure system caused that heat wave (as is typical for most heat waves elsewhere). First dry descending air currents within a high-pressure system promote clear skies that increase insolation while suppressing convection that normally carries heat away. Second the location of a high-pressure system determines where and how much warm air is transported poleward and how much cold air is transported equatorward.

Unlike low-pressure storm systems that pass over a region relatively quickly, high-pressure systems resist movement and can remain stationary for weeks and months. During the periods of greater than average melting, the blocking high was centered over Greenland and pulled warm air masses from the south into western Greenland while blocking any cold air incursions from the north. Hanna 2013 concluded summer 2012 was “not really representative of the future climates projected by the CMIP5” climate models but “more likely resulted from natural variability” and the negative North Atlantic Oscillation. The recent events in Greenland should have provided the media an opportunity to inform the public about how extreme weather effects are driven by natural oscillations. Instead weather was sadly politicized to promote climate fear.

The same negative North Atlantic Oscillation also reduces the westerly winds that warm Europe during the winter. Accordingly Greenland’s heat waves were followed by extremely cold European winters. These rapid changes in extreme weather in Greenland and Europe have no direct connection to changes in radiative forcing from the sun or from CO2 (although studies suggest a solar connection). Greenland’s climate is far more sensitive to natural climate oscillations as suggested by NOAA’s reported changes in the melt season and illustrated in DMI’s Figure 2 below.

NOAA reported that the location of the high-pressure system shifted in 2015 relative to the summers of 2007-2012 during which time it was “persistently centered over the ice sheet” so that a southerly air flow brought warm air across the ice sheet. In the summer of 2015 the high-pressure was centered “over the north-central ice sheet in July” which again promoted warm, southerly airflow and enhanced melting. “On the other hand, in June and August 2015 the anomaly was centered over the Labrador Sea southwest of Greenland, which promoted the advection of cold air from the Arctic Ocean and reduced melting.”

In Figure 2 below the blue curve represents changes in the percentage of Greenland area undergoing melt and the black curve represents the average. When the High was centered over the Labrador Sea from May to mid June and from the end of July to September the extensiveness of the melt season was far below average. Only when the high-pressure system shifted to the center of Greenland in late June and early July was there any extensive melting. It would be silly to blame extensive melting in June on rising levels of CO2, when below average melting during all other months suggested no sensitivity to global warming.

NOAA’s Moveable Goal Posts

In NOAA’s 2014 report card they reported an insignificant 6 Gt loss of total mass between June 2013 and June 2014. This year they reported an increase to 189 Gt of total loss between April 2014 and April 2015. Because different start months experience different rates of snow accumulation and can not be compared directly, NOAA restated the 2014 ice loss now using an April start date and calculated a 5 fold higher ice loss of 29 Gt for 2014 than originally reported. Clearly the choice of monthly start dates significantly affects the estimates of annual ice mass that is presented to the public.

As seen in Figure 1 above, on average Greenland continues to gain ice from September to June, and then undergoes a melt season that ends about the first of September. Above average accumulation was observed from April to June 2015, and that would offset at least some of the calculated “annual” loss. Currently there has been above average accumulation since September 2015 (Fig. 1 blue curve). Logically the first of September would be the most objective choice to compare annual changes in net accumulation. So why bounce between April and June? A suspicious mind might suspect NOAA of cherry-picking the month in order to emphasize ice loss. I do not pretend to know their reasons for choosing April for the 2015 calculations. However if they were being objective and consistent then an April start date should have been consistently chosen in previous reports. Unfortunately NOAA’s annual comparisons have used various start dates that differ by 6 months. In NOAA’s 2013 Report Card they used the month of September, which maximized that year’s ice loss stating, “The rate of mass loss has accelerated during the period of observation, the mass loss of 367 Gt/y between September 2008 and September 2012 being almost twice that for the period June 2002-July 2006 (193 Gt/y).” I suggest NOAA publish the ice mass for every month so independent reviewers can develop consistent annual comparisons.

When discussing Surface Mass Balance NOAA used September start dates reporting, “The surface mass balance for September 2014 through September 2015 measured along the southwestern portion of the ice sheet at the K-transect was the third least negative since the beginning of the record in 1990; not since the 1991-1992 (when melting was low due to the Mount Pinatubo eruption) and 1995-1996 balance years has so little ice been lost. And that observation raises serious questions. How could NOAA calculate a 6-fold increase in total ice loss when 2015 experienced the least amount of ice lost since 1996? And how did using an April start and finish date affect their results. If the third least amount of ice was lost in the year 2014-15, and 2013-2014 registered a “negligible” 6 Gt loss in total mass, how did lost ice mass increase in 2015?

The only other possible explanations are there must have been greater loss of mass elsewhere. But lost surface mass seems unlikely as the K-transects has been deemed a good proxy for the continent and according to the DMI, overall surface mass balance increased by 200 Gt. That is slightly below the 1990-2013 average but most definitely there was no net loss due to increased melting.

Marine Terminating Glaciers

The only other possible cause for NOAA’s increased loss of mass in 2015 would be a greater rate of basal melt, iceberg calving and glacial retreat. However the data does not readily support that possibility either. In past decades the fraction of retreating glaciers had steadily increased from 49% in the 1970s to 89% by 2010 (Howat 2011). As previously discussed here, the shift to the negative phase of the North Atlantic Oscillation should result in an initial acceleration of glacier retreat due to a greater flow of warm Atlantic water directed towards Greenland. However eventually that inflow declines and the trend reverses. Indeed of the glaciers reported by NOAA only 45% were in retreat this the past year. That is slightly lower than the 1970s percentage. NOAA stated, “Between the end of the 2014 melt season and the end of the 2015 melt season, 22 of the 45 glaciers had retreated, but the advance of 9 relatively wide glaciers resulted in a low 1-year net area loss of -16.5 km2. This is the lowest annual net area loss in the 16-year period of observations (1999-2015) and 7.7 times lower than the annual average area change trend of -127 km2”[Emphasis mine]. This year’s loss is only 5% of the peak area retreat in 2011-2012. Of course an increase in glacial area may not translate into increased mass if glaciers thinned as their area increased. But it stills seems unlikely that there was an increased mass loss. Hopefully NOAA can clarify these inconsistencies.

In conclusion, whatever the final determination of the true total mass change for 2015 may be, it is clear that that there is no longer an acceleration of Greenland mass loss. All evidence supports a deceleration. The 189 Gt estimate of total ice loss was still “22% below the average mass loss of 238 Gt for the 2002- 2015.” The rapid decelerating rate of loss for glacier area was the lowest in the past 16 years. Transect K experienced the lowest loss of surface mass since the 1990s. And changes in both the melt season and glacier basal melting appear to be primarily under the control of the North Atlantic Oscillation and its affect on atmospheric and ocean circulation. If the current reduction in solar irradiance remains low and reduces the poleward flow of ocean heat in the Gulf Stream similar to the reduction estimated during low solar irradiance of the Little Ice, Greenland should begin accumulating mass in the very near future.