Guest Post by Willis Eschenbach
Following on my previous post, “Jason and the Argo Notes”, just a couple of graphs in passing:
Figure 1. Argo surface temperatures, northern hemisphere. Colors show the latitude of the floats, from red at the Equator to blue in the north. Click on image for full size version. UPDATE: several readers pointed out that the x-axis is years, not “Latitude” as it is incorrectly labeled.
Figure 2. Argo surface temperatures, southern hemisphere. Colors show the latitude of the floats, from red at the Equator to blue in the south. Click on image for full size version.
Lots of interesting stuff there. I note that in the northern hemisphere, you can see how the limit on the ocean temperature affects the areas a bit further away from the equator (yellow and green). Those it only clips in the summer, while at the equator it affects temperatures for much of the year.
Also, the shape of the annual swings is interesting. The high temperature in the summer comes to a sharp peak and drops quickly, while the winters drop and change slowly, leading to a rounded shape.
Always more to learn,
w.
An article about temperatures and no mention of climate change. I find that alarming. You need to at least say more research is needed. /sarc
Very nice graphs. Is there any way to plot the standard deviation of the temperatures for each… 1 year by 20 degrees latitude blocks for example? Maybe 10 degrees would be enough. It’d be interesting to see the variability of temperature across the years for the latitudes. Is there a way to do this without the unavailability of data skewing the results too much?
Thank you Willis. It is very interesting to see this data plotted. The hard top on the prior posted graph creates some interesting hypothesis regarding a temperature safety relief valve.
Thanks Willis,
Very good graphic views of the data; Lots to learn!
I am an engineer by profession and all I see is two separate oscillating parameters ( the hemispheric temperatures) pertaining to two, quite asymmetric bodies of water that are obviously connected, yinging and yanging around two asymmetric sets of land masses all of which are located on a rotating planet in a rotating solar system featuring various bodies rotating and beating at a range of periods and with a sun that exhibits oscillating solar activity at a range of periods on the one hand and a scientific “confectus” (= a confected consensus ) that it is overwhelmingly down to a steady increase in CO2. These cretins would find a trend line in a pure sine wave.
This data set looks like a good candidate for a 3-D treatment.
http://en.wikipedia.org/wiki/VisIt
This is worth a look
http://www.climate.noaa.gov/cpo_pa/cdep/GODAS_ArgoS.pdf
I don’t suppose that I could prevail on you to look at the salinity data and see if Tmax is coupled to salinity?
Talking about a non existing problem (global warming/climate change) is propaganda.
Pretty colors. Might want to look at relabelling the abscissa.
it looks a lot less noisy than the surface data – do you have any thoughts on that? Also, if the heat is being sequestered in the deep ocean, how likely is it that you would detect it through Argo?
The Reference Frame has some remarks on Willis’ first post on this subject.
DocMartyn says:
February 10, 2012 at 3:21 pm (Edit)
Interesting stuff, Doc, thanks. Salinity will be high wherever evaporation is high, because fresh water evaporates and salt is left behind. So there is a connection, although perhaps not the one you are looking for.
w.
Mike McMillan says:
February 10, 2012 at 3:38 pm
I tried and tried to do that, but no luck. R is using a “date” object to date the data. This is a text object like “2002-01-07”. I haven’t been able to figure out what it is using internally, because using the usual “xaxp = c(1996,2012,17)” doesn’t do a dang thing. Any help gratefully accepted, I’m still learning R … and will be for the next half century, I hope …
w.
Some interesting charts coming out of the Argo data.
I’d be interested in the date of the peak high and low temperatures. Specifically, how long is the lag in days behind the solar solstices and is there is any difference in that by latitude.
It might be worth plotting each latitude band separately. See which ones have more climate signal, and drill those down to even smaller latitude bands.
Cool charts Willis. I had to see what a blink would look like, it took some horizontal scrunching to get the dates to align, but it looks pretty close. It’s 2.3Mb.
http://i42.tinypic.com/zx9ute.jpg
Willis,
These graphs are very interesting. Is horizontal axis (year) and not (latitude)?
Steve
Shape of the Antarctic winter curves are truncated / flattened while the curves near the Arctic are rounded. Wonder if that is due to the buoys getting locked up in the ice pack around Antarctica for a long time while the ice near the Arctic tends to be more mobile. Also looks like it gets colder down south and stays colder longer and that there is better coverage in the southern oceans than in the northern one. Intentional? Artifact? If nothing else, good for some head scratching. Cheers –
Cargonauts
Willis says: “…The high temperature in the summer comes to a sharp peak and drops quickly, while the winters drop and change slowly, leading to a rounded shape….”
Isn’t this just a result of an item (packet of water) losing heat faster when it’s hot than when it is rather less hot? I think that would work with both conduction and radiation of heat, whichever predominates.
Wills, I suspect that Tmax is a function of the evaporation rate.
If it is windy, lower Tmax at the same light flux.
Higher salt mean less evaporation, more sensible heat, hence higher Tmax.
My guess is that Tmax should plot nicely against salinity.
Ok, so the horizontal is time?
Or latitude?
What is the false-color?
Units?
It would be interesting to put depth on the z axis.
@ Jerry
Took me a bit too, to get to the “ah ha!” part of the graph.
Yes, horizontal is time, vertical is temperature. The color is latitude. Seems a bit odd at first, but quite a nice way to get that third variable to show up.
It appears to be oscillating regularly. More time will help us with the frequency.
A couple of years ago, I plotted the monthly SST’s at 45N and 45S in the N/S Atlantic, N/S Pacific, and the Indian Ocean against a sine curve. The actual curve always had a steeper peak and a broader valley than the sine model.
https://docs.google.com/spreadsheet/ccc?key=0AiP3g3LokjjZdDg3UDNPR21QRk9TUmVZZWlORU5uU3c&authkey=CL-ow8wM#gid=10
diogenes:
“…if the heat is being sequestered in the deep ocean, how likely is it that you would detect it through Argo?”
I think that if the earth is heating up, one could expect an increase in convection, much like a pot of water on a stove. I’m guessing that would mean an increase in the upwelling in the tropics with the warm tropical waters sinking further poleward. Ergo, I would expect to see the greatest warming trends in the mid-latitudes. If the heat is being sequestered in the deep, then I would expect to see it being diffusely returned in the deep tropical upwelling. If the convective overturning is increasing, then I’d also expect to see an increase in the salinity in the deep tropics and a decrease in the mid-latitudes. Much like these plots I made from the ARGO data:
https://sites.google.com/site/climateadj/argo-analysis
This was just using the 2005-2010 data in which NINO 3.4 had a negative trend. So I can’t argue that it’s a persistent pattern.
AJ:
“I’d also expect to see an increase in the salinity in the deep tropics and a decrease in the mid-latitudes”
I hate it when I have to call myself out. I’d expect to see an increase in salinity in the tropics, but with a long term persistent upward trend in convection, shouldn’t I also see an increase in salinity in the mid-latitudes as well?
Surely under cloudless skies the trade winds will evaporate strongly, increasing salinity. However under the ITCZ Cu-nim belt, Willis’s thermal governor, the evaporated water returns in a narrowish band, so rainfall reduces salinity.
A salinity index between the cu-nim band and the higher salinities either side would indicate how strongly the cu-nim thermal governor had been working.
Willis if those graphs you show are a true representation of the argo data in easy to see form they show a decline in the oceans temperature over the entire planet. Your graphs have a slope down in all latitudes, so much for the hidden heat it must be hiding on the dark side of the moon.
Willis; I think this may be of interest for You:
Look at http://polar.ncep.noaa.gov/ofs/viewer.shtml?-natl-temp-0-small-rundate=latest
See temperature and depth down to >1000 m box by box.
West of north Africa are the ocean well mixed down to 1000 meter. Nearly same temperature as the surface. Not as cold as if cold water is upwelling as west of south Africa.
http://en.wikipedia.org/wiki/Upwelling
It looks to me as the water is mixed by high salinity surface water sinking.
Why could otherwise the temperature be so high 1000 meter down in a otherwise cold ocean?
To me is this a salinity driven deepwater formation which is not mentioned at all in classic Thermohaline circulation theory. As if dry Sahara air increase ocean salinity and make the surface water to sink.
http://en.wikipedia.org/wiki/Thermohaline_circulation
And by the way, is this the missing heat? http://peswiki.com/index.php/Directory:Mixing_Sea_and_River_Water
😉
If this is a dumb question I apologize. But would it not relate better to the world we are used to by turning it upside down so north (blue) is up and the equator (red) is down?
Why has Arctic been behaving badly?
http://www.vukcevic.talktalk.net/EAA.htm
Likely cause may be volcanic ash deposits on the Arctic ice.
I reckon wayne Job is right. I’ll have to go read Part 1 to see if you already pulled out the trends in lat bands/bins.
>
Also, the shape of the annual swings is interesting. The high temperature in the summer comes to a sharp peak and drops quickly, while the winters drop and change slowly, leading to a rounded shape.
>
Notable feature (along with general trend of any isotherm to migrate towards the equator).
This could looked at in terms of feedback.
The primary driver (annual variation in insolation) is sinusoidal. 12m at high lat. 6m at equator. Two off-set, superimposed 12m cycles are visible in SH plot.
The asymmetry you note is more marked in NH where there is less thermal inertia (more land).
The narrowing of the warm side of the cycle could be the effect of a negative feedback to increased temperature. A feedback that is not symmetrical in lower than average temp.
Evaporation could be a part of this. The asymmetry is probably because warm water floats and cold water sinks.
This asymmetry provides a mechanism whereby El Nino/Nina cycles can capture solar energy into the climate system.
mid lat swings in NH seem c. 20K , twice that of SH.
Since everyone is offering suggestions: I would love to see the data presented by Ocean – eg, North Atlantic, South Atlantic, North Pacific, Indian, etc…
Eyeballing/rulering: NH getting slightly warmer, SH getting slightly cooler (max and mins).
Fuzz is is the +/-, visible, not 1000th of a degree. Can the fuzz be eliminated statistically, or just quantified, and if so, what is it?
Does not the inexactitude of originating data provide the top end error of subsequent manipulations and “corrections”? What sort of corrections or adjustments are appropriate or defensible?
Your split by hemisphere only needs proportionality of globe and comparison with GISTemp for same latitude band to determine required land-only portion for GISTemp to be consistent with ARGO data. Wish I could do this, but I can’t.
The cross-correlation of data-parts with GISTemp merged data strikes me as the easiest way to show that adjustments and “corrections” in the GISTemp computers are excessive and distorted. If you can’t take the pieces and combine them in a whole that is identical to the alleged whole, then you know you have a problem of generalization.
In Superfreakonomics, the authors demonstrate that statistical interpretation of the human species can be bizarre, by saying that the average human being has one breast and one testicle. The average Earth temperature, I suspect, is similarly non-real. There are places that are warming and those that are cooling, but it is neither global nor meaningful as a linear CO2 result.
Steve from Rockwood says:
February 10, 2012 at 4:43 pm
Yes, years.
w.
AFPhys says:
February 11, 2012 at 10:27 am (Edit)
Dang … so would I … but then there’s my day job, and my monkey mind is always going “Ooooh, shiny!” about something new and interesting …
Right now I’m looking at the ~ 30°C limit to the ocean temperatures, and how it plays out. I’ll post more as they come off of the presses …
w.
Willis, if I remember correctly, once you asked (Climate Etc) what causes annual atmospheric CO2 variation. I think it’s the annual SST variation. Maybe you find something in the ARGO data.
How many latitude bands, and what are they? Looks like 6: 0-10, 10-20, 50-60.
My compliments on the graphs. much information, informatively displayed.
Again I have to say: Absolutely brilliant illustrations. Willis I really understand how thrilling it must be to generate these graphic. Not without irony that the far best graphical illustrations of Argo data is shown on WUWT, and carried out by scientist not hired by the Argo project.
K.R.Frank
moderator, thanks.
Please disregard my earlier post.
Long-term, all the salts on the continents are headed for the oceans, so all ocean salinities are increasing. Every once in a while the Med gets cut off and dries out, and lays down a layer of salt which is subtracted from the total, tho’. 😉