I just saw this video on the
NBC news website about the DMI investing 1 million USD to create ASIP (automated sea ice products) using artificial intelligence. I've always thought that something like this would be great for IMS/MASIE, to replace the human analysts, whilst looking at various sources and not just passive microwave data. Especially now with all the Sentinel stuff. IMS/MASIE is the most precise of all daily sea ice extent products, but due to the human factor and changes in sources over time long-time series are unreliable/inconsistent.
Would it be possible, once the AI is really good at performing the job, to go back and look and past satellite data from as many sources as possible and construct a new continuous dataset? That could perhaps be the solution to the
problems with the DMSP satellites and a possible break in the longest running dataset of Arctic sea ice cover we have.
I've attached the video below, but I'm not sure it'll play (18.8 MB is a bit much perhaps). If it doesn't play, click the link above.