Comments on: When Graphs Attack!

By: Sea ice extent recovering quickly « Watts Up With That?

Sea ice extent recovering quickly « Watts Up With That? — Thu, 09 Oct 2008 22:38:32 +0000

[...] graph from the National Snow and Ice Data Center, I published with annotations on July 14th 2008, which was oft cited back in early June with the phrase “if this trend [...]

By: Gary Gulrud

Gary Gulrud — Fri, 18 Jul 2008 16:16:11 +0000

“27 months the same as cycle 23 min.” D’oh, 50% more than, rather.

By: Leon Brozyna

Leon Brozyna — Fri, 18 Jul 2008 11:32:03 +0000

The National Snow and Ice Data Center has a mid-month update where they explain why the ice is not melting as fast as expected:

http://nsidc.org/arcticseaicenews/

They’re also qualifying their statement with suggestions that the ice is so thin that significant melting could still occur. I doubt their objectivity in the way they suggest much melting could still happen since they think there is such a long melt period yet to go. With roughly six weeks to go to peak melt, I don’t see this as a very long time to go to reach peak melt, especially since we’re already 4½ months into the melt season and the melt rate seems to slow throughout August.

By: Gary Gulrud

Gary Gulrud — Thu, 17 Jul 2008 06:07:55 +0000

Leif: Ok, I eye-ball 27 months the same as cycle 23 min. You are correct not terribly predictive of solar sunspot max. No doubt “strongly correlated” is off the mark.

But then, I at least implied I was trying to stay out of that game.

My original point was that any hope of an average to large cycle is now lost. I was not deriding any particular model, rather the current obsession with sun spots. Clearly I am not above the tendency but that should reinforce the issue.

By: Leif Svalgaard

Leif Svalgaard — Thu, 17 Jul 2008 03:50:28 +0000

Gary: clearly I meant: cycle 20 had decreased to R20/n and cycle 21 had increased to R21/n . One could use other methods as well, with not much difference.

By: Leif Svalgaard

Leif Svalgaard — Thu, 17 Jul 2008 03:28:04 +0000

Gary: cycle 20 was a long cycle and the minimum 20-21 lasted quite some time making that ‘minimum long’. Of course, all depends on how you eye-ball the duration. A rather objective way would be to measure the time between when cycle 20 had decreased to R20/n and cycle 21 to R21/n [both smoothed values], where R20 and R21 are the maximum values and ‘n’ is a parameter to be chosen. Setting n = 3 or 4 seems good choices. So by several measures 20-21 was rather ‘wide’ compared to e.g. 21-22 [clearly visible on the butterfly diagram], yet R21 and R22 were about the same, showing that the ‘width’ is a poor predictor.

By: Leif Svalgaard

Leif Svalgaard — Thu, 17 Jul 2008 02:08:13 +0000

Vincent: the July 2, 1954 value is likely to be ‘glitch’. There was one too on June 27th, 1954. In any event, these have nothing really to do with cycle 19.

By: Gary Gulrud

Gary Gulrud — Thu, 17 Jul 2008 01:16:58 +0000

Leif: We are agreed on a low cycle 24.

The ‘plateau’ shape is a little too constricting. Establishing SDs from the maximum clearly won’t work for regime changes like 23 to 24.

I don’t get the 20-21 example, the min was over 10 leading into a max of 160. The coming max will be more like 50.

By: Vincent Guerrini

Vincent Guerrini — Thu, 17 Jul 2008 00:57:23 +0000

Yet at Solar 24: By the sites owner: “Today the solar flux has dipped to a new low of 64.2. Just so you do not worry too much, on July 2, 1954 a value of 64.4 was observed. What followed was one of the strongest Cycles ever recorded (Solar Cycle 19)”. Was that event exceptional?

By: Leif Svalgaard

Leif Svalgaard — Wed, 16 Jul 2008 23:26:46 +0000

Gary: so you do mean a kind of size of the 'through' around minimum. Nothing wrong with that. The problem comes in, when you say that the length of that flat plateau is strongly correlated with the size of the following maximum. It is not, David A, notwithstanding. The minimum between cycles 20 and 21 is a good counter example. But, hey, what are we quarreling about? We do gree, I think, that we are in for a low, low cycle 24.

By: Gary Gulrud

Gary Gulrud — Wed, 16 Jul 2008 22:16:41 +0000

Leif:
Your dear friend Dave Archibald is more conversant on that subject than I, so what follows is just reasoning by a rank amateur, meaning the duration of low solar activity appropriate to a minimum.

Remember 23′s minimum, the 13 month smoothed minimum came in May and the following Sept. the longest spotless period.

So we could call those 4 or 5 months the absolute trough and perhaps the minimum (9503 – 9709) 18 months in length. I don’t have numeric criteria to delimiting a minimum to recommend, but looking at the daily data for ’96 we can see the sun was at its local ebb in radio flux, sun spots, planetary indicies and flaring during the trough, with significant terminus in spotting at either end of the 18 month period.

Well this time around the decline seems to have been well begun by 3/07 and is yet to bottom. So the minimum, in this non-scientific, woolly thinking sort of way, should extend through all of 2009, i.e., end in being twice as long.

The current low-level is adequate for minimum as was the activity of a year ago. So lets just say the duration of solar activity of sufficient depression for official minimum.

By: Leif Svalgaard

Leif Svalgaard — Wed, 16 Jul 2008 20:47:01 +0000

Gary: you were talking about ‘the length of the minimum’. What is that?

By: Gary Gulrud

Gary Gulrud — Wed, 16 Jul 2008 18:38:37 +0000

Leif, the minimum of a cycle is taken to precede its maximum. I know you know that, can I demonstrate it? No, just eye-balling as usual. Can one take eye-balling to the bank? I should hope not!

By: woodfortrees (Paul Clark)

woodfortrees (Paul Clark) — Wed, 16 Jul 2008 16:53:01 +0000

… and those all important (but sometimes misleading) trend-lines for each:

http://www.woodfortrees.org/plot/nsidc-seaice-n/mean:12/plot/nsidc-seaice-n/trend
http://www.woodfortrees.org/plot/nsidc-seaice-s/mean:12/plot/nsidc-seaice-s/trend

Note the scales are not the same!

By: woodfortrees (Paul Clark)

woodfortrees (Paul Clark) — Wed, 16 Jul 2008 16:50:22 +0000

Returning to topic, I’ve now finally found a way to get NSIDC monthly data (don’t ask; it involves merging 12 month-specific files and interpolating for missing values…).

Full monthly data (dominated by annual signal), plus annual rolling mean:

http://www.woodfortrees.org/plot/nsidc-seaice-n/mean:12/plot/nsidc-seaice-n
http://www.woodfortrees.org/plot/nsidc-seaice-s/mean:12/plot/nsidc-seaice-s