Comments on: Sea Ice Utilities

By: Seb

Seb — Mon, 12 Jan 2009 15:10:08 +0000

NSIDC has a document comparing the F11 and F13 sensors and show that the differences are very small (perhaps negligible in most cases), but can be large in some regions. Nothing is said about the near real time data, which incorporates the F15 sensor, so these have both F13 and F15 data. Looking at a few subsets of both sensors, the differences do seem small again, but missing data pixels are not the same. I wonder whether it would be sensible to fill these missing pixels with data from whichever sensor does have data to produce a grid with fewer missing data…

By: Maurizio Morabito

Maurizio Morabito — Wed, 10 Sep 2008 22:18:30 +0000

I hope somebody is still reading this...has anybody else attempted to digitize Cryosphere Today's graphs for each individual Arctic sea? I just did and the results are not exactly identical to the values reported for the whole Northern Hemisphere. Am I missing a sea or what?

By: Ross Berteig

Ross Berteig — Tue, 08 Jul 2008 02:03:27 +0000

Pete, like Google earth?

How hard do you suppose it is to make R spit out KML files anyway?

By: Pete

Pete — Mon, 07 Jul 2008 22:47:14 +0000

Earth would be an oblate spheroid type of ellipsoid. (i just went to wikipedia for a refresh)

By: Pete

Pete — Mon, 07 Jul 2008 22:28:56 +0000

Why don’t they do away with the 2-D maps and just put the data on a 3D ellipsoid of the earth that you can manipulate with your mouse? The only projections would then be the projection of whatever view you have on the screen.

(I know 3D and ellipsoid are redundant but it sounds better)

By: jeez

jeez — Mon, 07 Jul 2008 22:11:12 +0000

Re: 13

You may want to consider changing your name to Steve Middleware.

By: Steve McIntyre

Steve McIntyre — Mon, 07 Jul 2008 22:08:50 +0000

When these values are plotted up as a contour, they make sense. Larger areas at the pole.

By: Steve McIntyre

Steve McIntyre — Mon, 07 Jul 2008 22:05:04 +0000

They don’t make things easy. The data is little-endian, while the areas are big-endian. Arrgh. Anyway, we’ve got the areas now and can push the peanut along a little.

By: Scott

Scott — Mon, 07 Jul 2008 21:57:57 +0000

Sorry, I could/should have included that information. In IDL world, you’d read it in as a ‘lonarr(304,448)’ and then divide by 1000. And you need to swap the endians if you’re using Unix. So, indeed, the first value is 382.068 as Dishman indicated.

By: Dishman

Dishman — Mon, 07 Jul 2008 21:42:00 +0000

I’m looking at it… looks like… integer, size =4, little endian
readBin(handle, ‘int’, 448*304, size=4, signed = FALSE)

It’s fixed point, so it’s a floating point number multiplied by some scaling and stored in an integer.. in this case, 1000.

The first value is 382.068