There has been some speculation about this since yesterday, but Paolo has confirmed that the first puddles of melted snow can be seen on images of NOAA's Arctic StarDot NetCam #1 that is located at the North Pole:
NOAA webcam 2 is also showing some dark patches:
Since 2002, the multi-agency USA-Japan joint project entitled "North Pole Environmental Observatory" (NPEO) has deployed Web Cameras along with instruments that monitor air, ice and ocean conditions. The Web Cams provide an otherwise unobtainable view of sea ice conditions throughout the Arctic summer.
The summers observed by the North Pole Web Cams were very different. (see table below). The onset of melting is typically in early June [I think that should be 'late June', N.], but occurred in late July in 2002, and late June in 2003 and 2004. The Web Cam images show very limited melt pond coverage in 2002, but widespread melt pond coverage in 2003 and 2004. In 2003, the melt ponds were widespread by July 4, but diminished in late July, and then reformed in mid-August. Freezeup began in late August in 2002, Sept 7 in 2003, and in mid-August in 2004. Summer sea ice transition information was not available for 2005 or 2007.
In 2008, the snow became waterlogged June 25, and extensive meltponds formed quickly, only five days later (June 30). Meltponds are still observed August 25 in standard web cam images. After that time, the only clear images are from another webcam with a fisheye lens observing sky conditions, but also revealing snow conditions. Meltponds are visible on the fisheye image from September 14. Although further images are at least partially obscured by water drops or snow on the lens, there appears to be snow cover in an image from Sepember 22. In summary, in 2008, the onset of melt progressed more quickly than in previous years, and melt ponds persisted longer and later than usual.
In 2009, the snow became soft and meltponds started to form July 8, but never became very widespread, with the maximum meltpond extent observed around July 14-16. Meltponds were closing over by Aug 11, but small slits of open water were still visible Sept 8. After that date, only one image was received from the web cam, on Sept 25, at which time the melt ponds were all snowcovered. In summary, in 2009, onset of melt was later than most years (later only in 2002). Meltpond coverage was less widespread than any year observed by the web cams other than 2002, but the snow remained soft later than most years (at least through Sept 8).In 2010, the snow became soft Jun 25, similar to many other years, but widespread meltponds formed by Jun 27, earlier than in any of the other years observed by the webcams. Freezeup began Aug 12 and meltponds were covered by Aug 23.