Tuesday, April 11, 2017

Preliminary results from 2016 traverse


During the 2016 field season, we measured broadband albedo (from 350-1800 nm) at 35 locations totaling 373 measurements. Likewise, we measured the snow's optical grain size and collected snow samples to determine impurity concentrations from both mineral dust and black carbon.

We compared our albedo results with the MODerate resolution Imaging Spectoradiometer (MODIS) satellite and Modele Atmospherique Regional (MAR) climate model. In the map below you can see the differences between our measurments and the satellite/climate model outputs.

Our albedo measurements agree with MODIS within their reported error, although differences from the MAR climate model are a bit larger than expected. 

RMS difference from measurements 
   0.045 ± 0.039
0.029 ± 0.029
Reported error
   ± 0.02
± 0.067

We see a negative correlation between optical grain size and albedo (R2 = 0.845, p = 0.005), as would be expected, but no significant correlation between total impurity mass and albedo (R2 = 0.0003, p = 0.96). These results agree with previous studies, which conclude that snow grain size can be 5-10 times more important in albedo reduction than black carbon content or density (Adolph et al., 2016; Tedesco et al., 2016).

Firn Cores

In the 2016 traverse we successfully collected seven firn cores, each ranging 21 - 31 m in length. Each firn core was analyzed for density and the distribution of refrozen melt. Portions of all 7 cores were melted for major ion, trace element, and water isotope samples. We are currently measuring and analyzing these samples to determine annual accumulation from each core. Core 2 provides an example of these results:

The graph below shows annual accumulation measured in Core 2 and the annual accumulation measured in PARCA core 6745. While the two cores have a similar mean annual accumulation, the year to year variability does not agree well between the two cores.

The Core 2 annual accumulation does agree well with several regional climate models. Pearson correlation (R) values are: RMAR = 0.57, RRACMO2 = 0.64 , RPolarMM5 = 0.57 and RBOX13 = 0.51

Melt Refreeze

We documented the location of melt refreeze features, such as ice layers and pipes, within each firn core using a backlit light table. We then analyzed the distribution of whole ice layers vs depth. We plan to analyze the distribution of ice layers with time once we measure the annual accumulation for all seven cores.

We found a significant increase in the total amount of ice in ice layers with decreasing depth for cores 1 to 5 (p-values < 0.02). Core 6 shows an insignificant increase in the total amount of ice layers in the firn and core 7 shows no trend. This suggests that, for the lower elevation sites, the changing summer climate over this region and increasing surface melting are resulting in greater amounts of refreezing within the firn, which has important implications for estimates of Greenland surface mass balance. 


We collected a total of 1630 km of ground penetrating radar during the 2016 campaign from Raven/Dye-2 to Summit, 800 km were along the main traverse path and 830 km were along the E-W spurs.

Preliminary analysis from the 400 MHz antenna along the main traverse path show excellent agreement with climate models. We see much more spatial variability from our radar accumulation results than the climate models indicate. It will be interesting to see how well our results compare with the climate models for the E-W spurs and for the 2017 traverse.

No comments:

Post a Comment