Wednesday, July 20, 2011

Making your own sea ice estimates

The Sea Ice Outlook does accept estimates from outside the professional community.  Maybe not everybody involved is thrilled by this, but I do think it's a good idea from my distant vantage.  And Jim Overland, one of the people behind the SIO, is strongly in favor of it.  (I had a chance to talk with him about the outlook and other ice matters a few weeks ago.)

In the most recent report, there are 3 submissions from 'outsiders' -- Chris Randles (you've seen him comment here as crandles) and Larry Hamilton, both at Neven's Arctic Sea Ice Blog, and one from Wattsupwiththat.

Watts' entry was a poll of readers.  While perfectly legitimate as an entry, it's also perfectly useless scientifically.  One goal of science is to gain understanding of the system in order to spread the knowledge around.  Polling can't be spread.

Much more interesting are Chris's and Larry's methods.  Both are obviously methods of great brilliance, as they currently have the same estimate as I do from my statistical method -- 4.4 million km^2 for this September.  Beyond that, you can read their method descriptions in the Sea Ice Outlook report and start constructing your own method by not making the mistake they and I have made.  Whatever those turn out to be.  Larry Hamilton's write ups (one for ice extent, one for volume) are:

And you can also examine model output from the PIPS replacement model ACNFS, PIOMAS, and CFS as a basis for making your own estimates.   (And please do cite others that you know of.)

All are welcome to post your methods here in addition to (or instead of) at the SIO.


Oale said...

My albedo-oriented and weather discarding mess of an estimation protocol begun with a way too low an estimate: - I'm not going to submit it anywhere else, a) since I've not tested it with other years
b) many others do this the same way and better
c) incorporating further corrective constants (changing with the insolation) will take it too mathematical to my poor taste. :-).
d) it's not based on winter maximum data though the calculations start there

further complications would undoubtly rise if I would try to model the winter phase of ice extent the same way, so...
the current number is about 4.26Mkm2, just so you know, predicting is easy afterwards :-).

L Hamilton said...

Thanks for the mention, Robert. I guess the "great brilliance" of all our predictions will be evident (or not) pretty soon.

One caveat on outsider-ness. I'm definitely an amateur with respect to ice science, and my statistical model contains no physics. My outsider status might be slightly compromised, however, by affiliation with the science steering committee of SEARCH (as a human-dimensions guy) since its inception in the late 1990s.

In that role, I've lately been enthusiastic about the new window the SEARCH Sea Ice Outlook has opened on science as a process. Your point about its participatory potential is very important as well.

DanG said...

For fun, I'm running my own prediction league among those of us overseeing climate change adaptation activity in water companies in the UK. The prize will be a iceberg lettuce and a bottle of eis wein if I can find one. I doubt any of us have done much more than eyeball the JAXA graph as a guide, so nothing to write home about scientifically. But it gets people to follow the phenomenon. The median guestimate is 4.67m km2 - which is close to the Canadian Ice Service in the SEARCH group and, indeed, Tamino.

L Hamilton said...

Dan, I trust you'll put up a web page to share results from your prediction league? Sounds like fun.

Robert Grumbine said...

Dan: I'll echo Hamilton's interest in you posting your results. If you don't have space yourself, I'll be happy to post it here. If you do, no need to be bashful about providing the link here. (See my link policy)