Comments on: Ocean Heat Content Stumbles

By: Andrew

Andrew — Mon, 19 Oct 2009 21:46:39 +0000

I might try and get to that, but I also expect that the data is too sparse to get any good (that is, statistically “robust” in the sense that they are fairly firm-not “good” in the “I like them” kind of way) answers.

By: davids99us

davids99us — Mon, 19 Oct 2009 20:32:06 +0000

IMHO these analyses need the CI's to show the probability the rate exceeds the expected rate. Not sure if there is enough data though.

By: Andrew

Andrew — Sun, 18 Oct 2009 20:56:40 +0000

Here's my initial reaction:

http://devoidofnulls.wordpress.com/2009/10/18/a...

My impression is that the Ocean Heat Content data unsurprisingly shows that same spurious acceleration effect I discussed with you earlier (namely, the curve changes sign of rate of change suddenly, without much change in the rate before or after.) Needless to say, I don't regard the claims of accelerating OHC rise as very impressive.

I would also appreciate it if anyone can find any errors I made in my analysis. For instance, can anyone else verify the very large slopes I seem to be getting?

By: Nick Stokes

Nick Stokes — Fri, 16 Oct 2009 06:17:11 +0000

No worries. I'm generally a worrier about OHC data. I wish it was better.

By: Bob Tisdale

Bob Tisdale — Thu, 15 Oct 2009 21:21:12 +0000

Nick Stokes: My apologies. Your suggestion of caution turned out to be correct. The NODC updated their OHC data today.
http://bobtisdale.blogspot.com/2009/10/nodcs-co...

By: cohenite

cohenite — Wed, 14 Oct 2009 10:45:12 +0000

Increasingly it is apparent that the OHC is not an indice of AGW;

http://wattsupwiththat.com/2009/08/11/ocean-hea...

One of the outstanding issues of the NOAA/Levitus record is this;

http://wattsupwiththat.com/2009/06/02/anomalous...

That transition spike, if excluded, in combination with the recent further decrease in OHC has taken OHC almost back to where it started in the 50's.

By: davids99us

davids99us — Wed, 14 Oct 2009 10:14:26 +0000

Aslak, thanks for the references, much appreciated. I was actually
wondering if in fact the 'acceleration' is significant, given the short
period and high autocorrelation. I haven't had time to do the analysis yet.

By: Aslak Grinsted

Aslak Grinsted — Wed, 14 Oct 2009 09:25:26 +0000

The OHC in the graph you show has accelerated: It starts out with a slightly negative trend, followed by a positive trend. I dont see any reason to look specifically at “since 1998″.

Some other OHC estimates can be seen here:
http://www.realclimate.org/index.php/archives/2...
Ishii and Levitus have also since revised their estimates to account for the problem described in the RC post. I personally like their estimates better, because they are independent of the topex/poseidon altimetry data. Afaik Domingues uses “optimal” interpolation, with the EOFs determined from the altimetry data set.

By: Bob Tisdale

Bob Tisdale — Mon, 12 Oct 2009 02:11:27 +0000

Nick Stokes: You wrote, "Of course there is posted data - that's where I noted all the 999's. I'm not familiar with the format, and and didn't make a detailed analysis. But I think there is likely to be a reason why NODC has not posted plots yet, and so I suggest caution."

Read the Instructions, Nick. Here's a link:
http://www.nodc.noaa.gov/OC5/WOA05/readwoa5.html

Scroll down and you'll find the following illustration, which is one of the NODC’s Coordinate Systems for Statistical Fields:
http://i38.tinypic.com/103s7s8.png

The illustration looks pretty straight forward. So if you’re looking at 1-degree data, the first Field would have the global coordinates of 90S, 1E. Or if you’re looking at the 5-degree data, the first Field would have the global coordinates of 90S-85S, 0-5E. Those should be right in the center of Antarctica and they are represented by -999.999.

Now, with a quick check (less than 1 minute) you could open a data file in Word or Notepad and determine that there a 64,800 Fields in the HC_0-700_A9A0406.dat File and of those 64,800 Fields, 23,380 or 36% of them are represented by -999.999. Do the same thing for the first 1955 data file and you’ll find the same number of Fields (64,800) and approximately the same number of fields with -999.999. (It wouldn’t be much of an Ocean Heat Content dataset if they were missing 36% of the data in approximately the same areas for the entire period.) Land mass represents approximately 31% of the global surface area. Do we assume the other 5% might be sea ice?

Feel free to continue to use caution, Nick, if you still think it’s necessary. I’ll continue to plot the data and post about it.

Have a nice day.

By: Nick Stokes

Nick Stokes — Sun, 11 Oct 2009 18:54:52 +0000

Bob,
It isn't incorrect. I carefully said “posted NODC plots“, and there are no new ones. Of course there is posted data – that's where I noted all the 999's. I'm not familiar with the format, and and didn't make a detailed analysis. But I think there is likely to be a reason why NODC has not posted plots yet, and so I suggest caution.