This animation shows the temporal changes in the glacier cover types during the ablation period 2015 (from June 22 to September 14) as detected by the proposed EURAC approach, which exploits the Sentinel-1 C-band SAR instrument.
The area of interest includes all the glaciers of the Ortles-Cevedale massif, the largest glaciered group of the Italian Alps. It is located in the Eastern Italian Alps and covers an area of 1638 km². The highest peak of this mountain group is Mt. Ortles (3905 m).
The multi-temporal maps, which are computed at the time frequency of Sentinel-1 acquisitions, are made by three overlapping layers:
- The glacier cover type: in blue the firn, in white the snow and in cyan the ice
- The hill-shade of the investigated area as seen by the SAR sensor in gray scale
- The Randolph Glacier Inventory (RGI) delimiting the glacier borders in red in the condition of the late summer of 2003.
A classification method to detect snow, firn, bare ice and bare rock
In detail, in order to detect the snow, the firn, the bare ice and the bare rock we developed a supervised classification method that makes use of both:
- machine learning techniques applied to single date observation
- the multitemporal nature of the data of the entire ablation seasons in order to correct itself from wrong detection done at the leave of single data observation.
The method allows also the estimation of the snowline altitude for each investigated glacier. As an example, in the graph in the lower part of the animation is reported how the SLA of the Careser glacier is behaving.
Analysing the multi-temporal map
From the animation is possible to notice that at the end of June most of the glaciers are still fully covered by snow. While advancing during the season the snow melts and the firn and/or ice underneath the snow cover appear and therefore the SLA rises.
During the first half of August, a snowfall occurred at high altitude in the Ortles-Cevedale massif. Some fresh snow cover is therefore reported in the map of August 21st and, consequently, the retrieved SLA on the Careser glacier decreases.
From the temporal behavior is possible to identify that the highest SLA for the Careser glacier is retrieved in the map of September 2nd. The highest value of SLA during the ablation period can be interpreted as an approximation of the equilibrium line altitude (ELA), which is the line that separates the accumulation from the ablation areas and corresponds to the theoretical line where the net mass balance equals zero within a particular year. Given its high correlation with the annual mass balance, the ELA is an important proxy variable for glacier mass balance estimation. The high temporal frequency of the Sentinel-1 acquisitions and the insensitiveness to clouds of its C-band SAR sensor allow a better estimation of the SLA peak and therefore of the ELA in comparison to glacier mapping with optical sensors, which can introduce a perceptible underestimation of the ELA when only cloud-free midsummer optical images are available.
Finally, it is interesting to note that by comparing the RGI of 2003 with the glacier cover maps generated with the Sentinel-1 data in 2015 a general decrease of glacier extension is visible.
For any further questions, please contact Mattia Callegari.