Elevation change of glacier facies across Devon Ice Cap, Nunavut from 2003 to 2011 Tyler de Jong ([email protected])1, Luke Copland1, David Burgess2 , Wesley Van Wychen1 1 – University of Ottawa, Ottawa, ON 2 – Geological Survey of Canada, Ottawa, ON

Firn line at 1648m Automated 6 2000 weather ENVISAT Wide Swath 5 1800 station Mode extracted σ0 1600 4 along GPR transect Firn line 1400 3 1200 2 1000 1 Sigma nought 800

600 (m) Elevation 0

Sigma nought (dB) nought Sigma Ice Cap 400 -1 elevation (m) 200 -2 0 0 1000 2000 3000 4000

Trace number Figure 6 Extracted σ0 values from Envisat ASAR Wide Figure 4 May 2012 (2011 post freeze-up) GPR radargram Swath Mode from Sept. 28, 2011 (2011 post freeze-up)

4 2000 1800

Firn line 3 1600

at 1609m 2 1400 Firn line 1200 1 1000 800 0 600 (m) Elevation Sigma nought (dB) nought Sigma -1 400 200 *240cm -2 0 Figure 1 Devon Ice Cap and location of GPR transect 0 1000 2000 3000 4000 Trace number and Sverdrup Glacier Automated Weather Station Figure 7 Extracted σ0 values from Envisat ASAR Wide Swath Mode from Sept. 29, 2010 (2010 post freeze-up)

Figure 5 May 2011 (2010 post freeze-up) GPR radargram

Introduction In recent years, Devon Ice Cap (Figure 1), along with the other glaciers and ice caps in the Canadian Arctic Archipelago (CAA) have seen an increased rate of ice mass wastage due to anomalously high summer temperatures (Sharp et al., 2011; Gardner et al., 2011). Snow accumulation and surface melt patterns form up to five distinct facies zones on the surface of glaciers (glacier ice, superimposed ice, saturation, percolation, and dry snow zones; Figure 1) (Muller 1962; Koerner 2005). The progression of these facies zones to higher elevations can therefore be indicative of an increase in surface melt and warmer summer temperatures.

Recently, remote sensing has been widely used to discriminate glacier facies (Partington 1998; Wolken et al 2009; Casey and Kelly 2010). The strength of the backscattered signal (sigma nought, σo) from SAR data can be collected to determine melt extent and characterize facies. σo is determined by the individual scatterers on the glacier surface and is expressed in Decibels (dB), which signifies the intensity of the returned signal. Percolation and saturation zone facies (Figure 2) have a higher backscatter due to the presence of effective scatterers such as ice lenses and pipes. Comparatively, superimposed ice and glacier ice facies at lower elevations have a lower backscatter due to the lack effective scatterers. The division between the superimposed ice zone and saturation zone is where the largest contrast in backscatter is seen and is termed the firn line (Figure 2).

Figure 2 Illustration from Muller (1962) of the glacier facies zonation

Methods ENVISAT ASAR Wide Swath imagery (150 m spatial resolution) was used from 2003 to 2011 during post freeze-up (autumn) periods to detect glacier facies and the progression of the firn line between years (Figure 3). Using PCI Geomatics OrthoEngine, the ASAR images were orthorectified using a rigorous math model and a digital elevation model (DEM). A radiometric correction option is selected in the orthorectification process which returns the units in calibrated σo. This process outputs a radiometric terrain corrected image in Decibel (dB) values. The images were then imported into ArcMap where the σo values were extracted along the GPR transect (Figure 1) and Figure 3 ENVISAT ASAR imagery with extracted sigma nought values along GPR transect showing the location of the graphed with the corresponding ice cap elevation (Figure 3). firn line

A 450MHz Ground Penetrating Radar (GPR) was towed behind a snowmobile along a northwest transect of DIC (Figure 1) in May 2011 (Figure 5) and repeated in May 2012 (Figure 4). The GPR was used to map the near surface (depths of ~12m) of the northwest sector of Discussion and Conclusions DIC and to provide validation of the ENVISAT ASAR signal (e.g. Figures 4, 5, 6, 7). Data from the Sverdup Glacier Automated Weather High summer temperatures since 2005 relate to an increase in elevation of the firn line, which signifies an increase in Station (Figure 1) was also used to compare changes in glacier facies elevation with annual positive degree days (PDD) (Figure 8). the glacier facies zones to higher elevations over this period. In recent years, Wolken et al (2009) showed that the dry snow zone is an intermittent phenomenon in the Arctic, as melt is ubiquitous at all elevations during warm years. The Results ground-based GPR measurements show a distinct firn line, whereas the ENVISAT ASAR signal can be less obvious The location of the firn line identified in the GPR radargrams (Figure 4,5) relates to the area of highest contrast in the extracted σo in some years. In addition, the relative ASAR σo values are seen to have generally consistent patterns from year to ENVISAT graphs (Figure 3). The elevation of the firn line shows a clear trend towards increasingly higher elevations from 2003 to 2011 year. (Figure 9). The firn line ranged from 1520 m in 2006 to 1673 m in 2009. When comparing the firn line elevations to PDDs there is a correlation of 0.6 from 2003 to 2011. By using the GPR results to calibrate and validate the ENVISAT ASAR signal it is possible to identify the approximate location of the firn line since 2003. If warming continues it is expected that the firn line and glacier 1700 facies zones will continue increasing in elevation due to an increase in surface melt. In future work, estimates of melt o

300 rates could be completed by relating in situ summer balance measurements at a single point with its corresponding σ 1650 250 value, which could then be extracted across the entire ice cap. 200 1600 Acknowledgements

150 Northern Scientific Training Program (NSTP), Members of the Laboratory for Cryospheric Research (Samantha Darling, Adrienne White, Alexandra Elevation (m) Elevation 100 1550 Waechter, Nicole Shaffer, Miriam Richer-McCallum and Alex Bevington), Alec Casey (University of Alberta) 50

Positive Degree Days Days Degree Positive 1500 0 References 2003 2005 2007 2009 2011 • Casey, JA. and Kelly, R. (2010). Estimating the equilibrium line of Devon Ice Cap, Nunavut, from RADARSAT-1 ScanSAR wide imagery. Canadian Journal of Remote 2003 2005 2007 2009 2011 Year Sensing, 36 (1) S41-S55. Year Figure 9 Firn line elevations derived from ENVISAT ASAR • Gardner, A., Moholdt G., Wouters, B., Wolken, G., Burgess, D., Sharp, M., Cogley, G., Braun, C. and Labine, C. (2011). Sharply increased mass loss from glaciers and ice Figure 8 Annual Positive Degree Days at Sverdrup Glacier (Figure 1), 2003 caps in the Canadian Arctic Archipelago. Nature, 473 (7347), 357-360. imagery. Firn line determined by the point of highest • Koerner, RM. (2005). Mass balance of glaciers in the Queen Elizabeth Islands, Nunavut,Canada. Annals of Glaciology, 42, 417-423. to 2011 sigma nought contrast (seen in the Figure 3 graphs) • Muller, F. (1962). Zonation in the accumulation area of the glaciers of Axel Heiberg Island, N.W.T., Canada. Journal of Glaciology, 4 (33), 302-311. • Partington, KC. (1998). Discrimmination of glacier facies using multi-temporal SAR data. Journal of Glaciology, 44 (146), 42-53. • Sharp, M., Burgess, D., Cogley, G., Ecclestone, M., Labine, C. and Wolken, G. (2011) Extreme melt on Canada’s Arctic ice caps in the 21st century. Geophysical Research Letters, 38, 1- 5. • Wolken, GJ., Shapr, M. and Wang, L. (2009). Snow and ice facies variability and ice layer formation on Canadian Arctic ice caps, 1999-2005. Journal of Geophysical Research, 114, 1-14.