Melting Record Climbs Up to High Elevation in Greenland in 2007
2 Oct, 2007 11:15 am
The year 2007 broke the melting record in Greenland for areas above 2000 meters. Places above 2000 m experienced, indeed, the greatest ?melting index?, computed as the melting area times melting duration, for the period 1988-2007, which was 153 % greater than the average.
The reasons why we do care about melting are multiple. The first, coming to mind, is that melting generates liquid water which might flow to the sea, potentially affecting sea level rise. Another reason is that liquid water might percolate at the ice sheet bottom, enabling glacier acceleration. Another less obvious reason is related to the albedo change connected to melting processes. Melted and refrozen snow have a lower albedo than fresh or unthawed snow because of the aggregation of grains as a consequence of the appearance of liquid water, which acts as a glue. A lower albedo means more energy absorbed, hence leading to more snowmelt.
Microwave remote sensing data used in my study span between 1987 and 2007 and have been collected by the Special Sensor Microwave Imager mounted on the Defense Meteorological Satellite Program platforms. Previous to 1987, data were collected by different instruments back to 1973, and accurate inter-calibration of the different instruments is necessary. The use of microwaves (19.35 GHz) allows the collection of data in all weather conditions and is not dependent from sun illumination, as in the case of visible data, when clouds also represent an obstacle. Besides, microwave data are sensitive to sub-surface melting, allowing us to detect melting even when this is not happening on the surface.
It is unlikely that snow melting at high elevation might reach the sea or the bottom of the ice sheet and accelerate glacier sliding. Also, the albedo can be re-set to high values (e.g., less energy absorbed) by new snow covering the metamorphosed one. Areas in the south part of Greenland experiencing the record in 2007 are also areas of precipitation during winter. The amount of new snow for changing the albedo depends by many factors, but it is in the order of a few tens of centimeters. Nevertheless, these factors need to be accounted for solving the puzzle depicting the changes occurring on the Greenland ice sheet.
Even though the effects of melting at high elevation might not be as direct as those occurring at low elevations and along the coast, monitoring these areas is very important because they affect, in both known and unknown ways, the ice sheet body. Melting is related to surface temperature, and the knowledge of melting areas can provide scientists a mean for studying the spatial distribution of surface temperature over the inhospitable and vast island. Results for 2007 were derived, daily, using a near-to-real time framework in which satellite data were processed and elaborated within 24 hours from their acquisition. This system was strongly supported by two colleagues of mine, Mary Brodzik and Richard Armstrong, of the National Snow and Ice Data Center in Boulder, Colorado, who set up the processing framework. Though Greenland is not going to disappear from one day to another (as a reviewer asked me about a submitted paper showing the results herein reported), monitoring melting over Greenland in real time might still provide useful information about daily surface temperature, and the data can be used to update climate models in a real time framework. Combining remote sensing tools and climate models is a fascinating approach that I am currently exploring with European colleagues for studying the trends of the whole ice sheet, not only on its surface but also in its mass.
Reference:
Tedesco, M., Eos Trans. AGU, September 25
A few oddities:
"The first, coming to mind, is that melting generates liquid water which might flow to the sea, potentially affecting sea level rise" - very unlikely for melting above 2000m
"Another reason is that liquid water might percolate at the ice sheet bottom, enabling glacier acceleration" possible, but later contradicted by "It is unlikely that snow melting at high elevation might reach the sea or the bottom of the ice sheet and accelerate glacier sliding".
Albedo also seems to get partially re-said lower. And "Besides, microwave data are sensitive to sub-surface melting, allowing us to detect melting even when this is not happening on the surface." - if that is so, it won't affect the surface albedo at all. So it would be nice to know how much of the melting was at the sfc.
[Response] Thanks for pointing out this. 1) I do not see a contraddiction. In the second section the reader is introduced to why, in general, monitoring melting is important. Then, in the 4th section I explain that some of the potential implications described above might not apply to the areas where I observed the record. However, as I am also presenting results for areas below 2000 m, I believe it is useful to give to the redaer a tool for what potentially this means for these areas. If the reviewer agrees, I would be glad to receive suggestion on how to rearrange the text, but still keeping the basic structure. 2) The distinction between surface and sub-surface melting is not possible using just microwaves at the frequencies in object. Visible or infra-red data can be useful to this aim but they are strongly limited by the sun illumination and clouds. Several approaches for differentiating among different melting conditions are under evaluation. Thanks, Marco Tedesco
This is a great article, and yet another indication of climate change. One thing that I would like to note though: Although the passive microwave record goes back to the early 1970s, how reliable are the measurements pre-1979 for Greenland? For sea ice melt onset, we are only confident in the 1979 - record.
[Response] Sheldon, you are right. The sensors changed since the 1970's up to 1987 , when SSM/I can be consistently used. I did not look at early 1970's data so I cannot tell but I know of people who are looking there. I believe there are many issues to be resolved and that's why I am not publishing results.