I am mostly happy with the revised version but suggest a few additional changes.

1) specify up front what you mean by thin ice (including in the abstract)
2) Page 3, first paragraph feels disjointed from the second paragraph. You mention L-band in the first paragraph but go into more detail in the second paragraph, but it's strange to say for the first time SMOS has allowed for L-band measurements. I think some better flow can be done between these two paragraphs.
3) Page 8, line 26, typo. Should be note. Also, the thickness means the mean grid cell thickness even if part is open water, or the thickness of the ice that is actually there? It's always good to specify since PIOMAS for example includes open water fraction in the mean grid cell thickness, whereas CPOM CS2 data does not. I think it's important to also point out how this compares with how climate models define it so that users of SMOS can correctly compare with thin ice in GCMs for example.
4) I would always suspect SMOS thickness to be lower than ORAS5 near the ice edge because of open water fraction. But wouldn't the ORAS5 also capture the open water? Have both the SMOS and ORAS5 been gridded to the same grid? Also I don't fully understand why the SMOS data is on a 12.5 km grid because the actual spatial resolution is coarser than this.