Melt detection from SMOS enhanced resolution brightness temperatures on the Antarctic and Greenland ice sheets

The Soil Moisture and Ocean Salinity (SMOS) mission has been providing surface brightness temperatures (TB) at L-band since 2010. These measurements feed geophysical applications over the ice sheets including melt detection. The synergy between SMOS and other sensors operating at higher frequencies such as AMSR2 is furthermore promising to give insights on the percolation of meltwater in the snowpack and the presence of firn aquifers. However, most of the algorithms currently use the SMOS Level 3 (L3) gridded data which suffer from a coarse spatial resolution. To overcome this issue, we developed a new SMOS enhanced resolution TB dataset over Antarctica and Greenland which is further used to detect melt. The resolution enhancement process is based on the radiometer version of the Scatterometer Image Reconstruction (rSIR) algorithm which was previously successfully applied to SSM/I, SSMIS, AMSR and SMAP. Our methodology also takes advantage of the multi-incidence nature of SMOS measurements to use information with the best native resolution, i.e. low-incidence measurements (15-40°). We evaluated the effective spatial resolution to be ~30 km for the new SMOS enhanced TB maps. It is twice finer than the spatial resolution of the conventional SMOS L3 dataset (~60 km). Finally, a state-of-the-art melt detection algorithm was applied to both the enhanced resolution and the conventional L3 datasets. The new product unravels many localized melt patterns that were not detected using the SMOS L3, especially near the grounding lines of Antarctic ice shelves, due to smoothing of the TB between melting and non melting area. We also identify a clear aquifer signature over the ~25 km wide northern George 6 ice shelf in the Antarctic Peninsula thanks to the gain in resolution. The new SMOS enhanced resolution TB and melt products are posted on the same 12.5 km polar stereographic grid and are comparable to AMSR-E and 2 in terms of spatial resolution to facilitate further synergetic use of multi-frequency passive microwave datasets.