Extreme Greenland ice sheet melt

By Julia Figa, Mark Higgins and Sancha Lancaster (EUMETSAT)

Between 10 and 12 July 2012 the ASCAT instrument on Metop-A detected a very fast surface snow melt event, covering almost the entire Greenland ice sheet. This backs up reports of extreme melting conditions in the area this summer (see e.g. the report on NASA Earth Observatory or NPP VIIRS Blog, source: C. Seaman).

The primary function of the ASCAT (Advanced Scatterometer) instrument on the Metop satellite is to provide ocean surface wind vector measurements over the world's oceans. But it is also used over land and ice to obtain information about soil moisture, sea ice and snow cover. It transmits an electromagnetic pulse and measures the energy scattered back from the surface (known as backscatter). The fact that the ASCAT operates in the microwave energy range makes this backscatter independent of the cloud cover, which is transparent at these electromagnetic frequencies. This all-weather capability makes it extremely valuable, especially in high latitudes.

Over ocean and lakes, the backscatter is related to the small-scale water surface roughness caused by the surface wind. Over land and ice, the microwave signal penetrates deeper, and the backscatter depends on how the different land cover and soil properties scatter the electromagnetic signal, and how deep it penetrates. Microwave backscattering over snow and ice depends on their physical characteristics, such as temperature and density and the surface roughness.

Monitoring backscatter changes over land ice sheets helps us detect changes in the snow surface conditions, which can be due to surface temperature changes, accumulation, densification, etc. These events are likely to occur more dramatically during the melting and freezing seasons at the height of the summer and winter periods. During other periods, the backscatter over great ice sheets is so stable that some areas in Greenland and Antarctica are considered as very good natural targets for monitoring the calibration stability of many satellite earth observation instruments.

Extreme melting

Between 8 and 11 July the backscatter brightness over Greenland changed dramatically. This can be seen in the images and animation of ASCAT enhanced resolution radar images (see images below): between 8 and 11 July the backscatter image over Greenland got much darker. After 12 July, the bright backscatter map builds up again. This event is likely to have been due to extreme air surface temperatures during those days (see synoptic situation on 11 July at 18:00 UTC as shown by the AVHRR VIS0.6 image with ECMWF 300hPa height field and SYNOP observations, source: EUMeTrain). The fact that the backscatter goes back to expected levels indicates that the change in the surface is temporary.

However, according to Dr David Long from Brigham Young University, the melt and re-freeze is likely to have created a layer of ice that will take some years to be buried deep in the snow again. This means that it may also take years for the area to recover the old, stable target backscatter.

The images shown in the animation have been obtained from NOAA/NESDIS and are part of the Brigham Young University Scatterometer Climate Record Pathfinder. More information on the Scatterometer Image Reconstruction techniques followed is on the Brigham Young University site.

Scatterometers monitoring snow from space

ASCAT's signal is in the C-band range of the microwave frequencies and it transmits and receives vertically polarised energy. Other scatterometers, such as QuikSCAT (in the past) or the current OSCAT on ISRO's Oceansat-2, operate in Ku-band and transmit and receive, in addition, horizontally polarised energy.

The backscatter from any surface depends on the frequency and polarisation of the signal. It is for this reason that a false colour RGB (Red Green Blue) composite of daily pictures from different radars can give interesting insights into snow properties. The image below shows two daily pictures over Greenland, taken around the time of the melt event in July. The areas affected by normal summer melt are shown in dark green, the areas suffering from the fast surface melt are shown in red. This is because the different frequency/polarisation combinations have different sensitivities to specific snow physical state and the depth of penetration of the radar signal slightly different. Monitoring snow ice sheet properties is an emerging scatterometer application requiring multi-frequency, multi-polarimetric measurements.

Figure 2 is the false colour RGB images from a single day of Ku-band data (Oceansat-2) and C-band data (ASCAT). Land is grey. Summer melt is the dark green. Fast surface melt is red. Refrozen melt is bright white. Unmelted firn dark grey/blue.