The Laboratory » Team » Dr. Xavier FETTWEIS » The 2016 melt season over Greenland as simulated by MARv3.5.2

The 2016 melt season over Greenland

 The 2015-2016 Greenland ice sheet

season as simulated by MARv3.5.2

 

1. Summary

According to the regional climate model MARv3.5.2 (run at a resolution of 20km) forced by the reanalysis NCEP-NCARv1, the total Greenland Ice Sheet (GrIS) Surface Mass Balance (SMB) of the 2015-2016 hydrological years (from Sep 2015 to Aug 2016) is 144 Gt/yr below the 1981-2010 average as a result of a significant runoff anomaly (+188 Gyt/yr) partly compensated by snowfall higher than average (+30 Gt/yr) (See Fig. 1). The run-off anomaly in 2015-2016 is the 3rd highest postive anomaly after 2009-2010 and 2011-2012. The normalized anomalies (i.e. in respect to the 1981-2010 interannual variability) are listed in the table below for the 3 years with the highest run-off:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

The 2015-2016 SMB is in average significantly lower than average in the ablation zone and higher than average in the accumulation zone (see Fig 2a). The highest snowfall anomalies occur in the south of Greenland where there is more (respectively less) snowfall in 2016 than average along the south-eastern (respectively south-western) coast (Fig. 2b) while the run-off is significantly higher than average everywhere (Fig. 2c).  


The 2015-2016 melt extent is not exceptional except near the ELA along the western coast (Fig. 3a). Along this same coast, the anomalies of bare ice extent are more significant (Fig. 3b) as well as in the north-eastern ablation zone where the highest anomalies of near-surface temperature are found (Fig. 4). Finally, the anomalies of albedo mainly follow the anomalies of the bare ice extent (Fig. 5).   


As a result of dominant negative NAO conditions in the 2016 summer, abnormal anticyclonic conditions dominate the 2016 summer with a maximum of Z500 anomaly centred on Baffin (Fig. 6a). However, these anomalies do not change significantly the mean south-west to north-east 7500 pattern of the general circulation over Greenland (Fig. 6b). The general circulation is just weakened at the south and enhanced at the north. 


To conclude, the abnormal high run-off rate simulated by MARv3.5.2 during the summer 2016 is not due to anomalies in the general circulation while the JJA NAO index was significantly negative (as the summers 2007, 2010 and 2012). The highest anomalies occurred 

- at the south-west as a result of an earlier appearance of bare ice resulting from lower 2015-2016 winter accumulation than normal
- at the north-east as a result of significant positive near-surface temperature anomalies combined with abnormal exposure of bare ice areas.  
In the other areas, the anomalies of near-surface temperature (and run-off) are not exceptional but positive everywhere suggesting that these anomalies are more a consequence of the background warming (Global Warming) enhanced by an exceptional exposure of bare ice areas than other thing.


At the daily time scale (see Fig. 8), we can not see exceptional melt events (except one on Mid-July) but a run-off higher than average through the whole melt season confirming well that the role of the general circulation has been weak this summer. We can also see that the bare ice extent is higher than normal through the whole summer while the melt extent is included most of the time in the interannual variability showing the role of negative anomalies in the winter accumulation as trigger of the melt season.

 

2. Figures

Fig. 1 Evolution of the anomalies of the total Greenland Ice Sheet (GrIS) Surface Mass Balance (SMB), Snowfall and Run-Off of the hydrological years (1 Sep => 31 Aug) simulated by MARv3.5.2 forced by NCEP-NCARv1 in respect to 1981-2010. Units are GT/yr. The anomalies for the 2015-2016 hydrological year are listed.   

 

Fig. 2a: Left) Cumulated SMB (in mmWE/yr) from the 1 Sep 2015 to 31 Aug 2016. Middle) Same as Left) but in respect to the 1981-2010 average from the 1 Sep to 31 Aug. The anomalies lower than the 1981-2010 interannual variability (1 standard deviation) are hatched. Right) Normalized anomalies by the 1981-2010 interannual variability.

 

Fig. 2b: Same as Fig 3a but for snowfall.

 

 Fig. 2c: Same as Fig 3a but for run-off.

 

 
Fig. 3a: Left) Number of melt days ( (i.e. when the daily meltwater production > 5 mmWE/day) from the 1 Jun 2016 to 31 Aug 2016. Middle) Same as Left) but in respect to the 1981-2010 average from 1 Jun to 31 Aug. Right) Normalized anomaly in respect to the 1981-2010 interannual variability (1 standard deviation).
 
 
 
Fig. 3b: Same as Fig 3a but for the number of day where bare ice (surface density > 900 kg/m³) appear in surface.

 

 
Fig. 4: Left) Anomaly of the JJA mean near surface temperature (~3m) in respect to the 1981-2010 average from 1 Jun to 31 Aug. Righ) Normalized anomaly in respect to the 1981-2010 interannual variability (1 standard deviation).

 
Fig. 5: Left) Anomaly of the JJA mean surface albedo in respect to the 1981-2010 average from 1 Jun to 31 Aug. Righ) Normalized anomaly in respect to the 1981-2010 interannual variability (1 standard deviation).
 
 
Fig. 6a: In colored background, the mean 2016 JJA temperature anomaly at 700hPa (in °C) from NCEP-NCARv1 in respect to 1981-2010. The isohypses of the mean 1981-2010 JJA geopotentiel height at 500hPa Z500 (resp. the 2016 Z500 anomalies) are also plotted in dash black (resp. bleu). 
 
 
Fig. 6b: In background, the anomaly of the JJA wind speed at 500 hPa (in m/s). The isohypses of the 2016 JJA geopotentiel height at 500hPa (not anomaly here) are plotted in bleu. 
 
 

3. Daily variability (not discussed in the summary)

 

Fig 7: a) Time series of the cumulated GrIS SMB in GT simulated by the regional climate model MAR (version 3.5.2) forced by the NCEP-NCARv1 reanalysis since 1 Sep 2011 (in green), and 1 Sep 2015 (in red). The 1981-2010 mean simulated by MARv3.5.2 forced by NCEP-NCARv1 is also plotted in black. b) Same as a) but for the daily SMB in GT/day.  The absolute maximum/minimum SMB rate of each day is plotted in blue. c) Daily  mean GrIS near-surface temperature (TAS) simulated by MAR. The absolute maximum temperature of each day is plotted in blue. d) Time series of the North Atlantic Oscillation (NAO) index from Climate Prediction Center (CPC).

 

Fig 8: Time series of a) the daily mean GrIS Run-off (in GT/yr), b) production of meltwater (in GT/day), c) daily mean GrIS incoming long/shortwave radiation (in W/m²)), d) bare ice extent (i.e. area where the surface density is > 900 kg/m^3 in % of the ice sheet surface) and e) daily mean GrIS surface albedo simulated by MARv3.5.2 forced by NCEP-NCARv1 (1981-2016).

 
 
 
 
 
 
 
 










 
Fig 9: Left) Melt extent as derived from satellite data (Credit: NSIDC/Thomas Mote). Right) Same as a) but as simulated by MARv3.5.2 forced by NCEP-NCARv1. Daily meltwater production > 5 mmWE/day is used as melt threshold in MARv3.5.2.

 
 
Fig 10: Greenland bloking index (GBI) from NCEP-NCARv1 in red and from the Global Forecast System (GFS) based forecasting in dashed red. According to Hanna et al. (2013), the GBI is defined as the 500hPa geopotential height area averaged over 60-80°N, 280-340°E.  
 
 
 
Fig 11: Idem as Fig. 6 but for 700hPa temperature averaged over 60-80°N, 280-340°E.  
 
 
 

4. Interresting links

  
5. Raw daily MARv3.5.2 outputs forced by NCEP1. The 2016 files is automatically updated every day. 

(c) Xavier Fettweis, University of Liège (ULg), Belgium