Global Oceanic Climate Update for August 2009
Dr. Roy Spencer September 1st, 2009
This is the first of what might turn into a series of monthly updates of some maritime climate parameters monitored by the AMSR-E instrument on NASA’s Aqua satellite. All monthly statistics have been computed by me from daily global gridpoint data produced and archived by Remote Sensing Systems (RSS) under the direction of Frank Wentz, a member of our U.S. AMSR-E Science Team. Since Aqua was launched in 2002, the data are available only since June, 2002. A description of how these products were derived, and where they reside, is provided here.
There are 5 “ocean products”: sea surface temperature [SST]; near-surface wind speed; vertically-integrated water vapor; vertically integrated cloud water; and rain rate. I will present time series of monthly anomalies averaged over the global, ice-free oceans (56 deg. N to 56 deg. S latitude), and separately for the deep tropics (20 deg. N to 20 deg. S latitude). ‘Anomalies’ are departures from the average seasonal cycles in those parameters, which will be recomputed as each new month of data is added.
GLOBAL OCEANS
In the first figure below are plotted the 5 ocean products for the global ice free-oceans (56N to 56S). As can be seen in the top panel, SSTs in August cooled slightly from the unusually warm conditions experienced in July.
I have added linear trend lines to each time series, which you are free to misinterpret as you wish.
Since the AMSR-E period of record is only 7.25 years long, a calculated trend won’t have much meaning…although it will be interesting to see how long it takes before the climate system obeys the UN’s command to warm, and the SST trend line begins to go uphill again.
How these different variables change relative to each other is illustrated in the following lag-correlation plot of SST versus the other variables. “PDO” is the Pacific Decadal Oscillation Index, while “SOI” is the Southern Oscillation Index (negative for El Nino, positive for La Nina). A discussion of these curves is provided later, below.
TROPICAL OCEANS
The next figure shows the ocean product anomalies for just the deep tropics, 20N to 20S latitude….
…and the lag correlation plot for the deep tropics is next:
DISCUSSION
Using the 20N-20S lag correlation plot as an example, you can see that total integrated water vapor is highly correlated with SST, which in turn is highly correlated with El Nino conditions (negative SOI values).
Also note that sea surface temperature tends to peak after months of anomalously low wind conditions, then falls as wind speeds increase.
Cloud water and rain rates increase as SST increases, reaching a maximum 1 to 3 months after the SST peaked.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.





So, what heats the Oceans?
Sunshine.
Many thanks Dr Spencer. Excellent data.
If we could add OHC (Ocean Heat Content), Solar Irradiance and LWR to these time series, we would get a comprehensive picture of the heat flows in the Earth’s climate system.
Who knows. Maybe it would show a GHG signature.
I also do not like the term “anomaly”.
That’s fine, but it’s widely used. Widely. See also: gravity anomaly.
You may ask yourself: what is unexpected or abnormal about gravity anomalies? Technically, nothing. Gravity anomalies have been fairly extensively mapped, so it’s not like we don’t know they’re there.
There’s not much of an analogy between gravity anomalies and things like SST, because gravity anomalies are more or less fixed over a span of years, while SSTs aren’t. Think of it as spacial vs. temporal anomaly.
Somewhat on topic: OI.v2 Monthly SST anomaly update:
http://bobtisdale.blogspot.com/2009/09/august-2009-sst-anomaly-update.html
As you can see, the Global SST anomalies are nowhere close to the record levels in the OI.v2 data…
http://i31.tinypic.com/2wbui5v.png
…but the ERSST.v3b data is again in record monthly territory. Raw data here:
ftp://eclipse.ncdc.noaa.gov/pub/ersstv3b/pdo/aravg.mon.ocean.90S.90N.asc
The Hadley Centre’s HADSST2 as of last month also was not at record highs (first column after date):
http://hadobs.metoffice.com/hadsst2/diagnostics/global/nh+sh/monthly
As soon as the Hadley Centre releases their August HADSST2 update, I’ll put together a post comparing the official NOAA ERSST.v3b data with the other two datasets. HADSST2 is not satellite based, so that excuse is not applicable.
Nogw (11:13:03) :
“It is relieving that there is no “end of the world” ahead. All are real nano variabilities.
All are not anomalous ( “inconsistent with or deviating from what is usual, normal, or expected “). Everything is OK. No global warming, no climate change ahead.
Good to know everything is usual, normal and as expected.”
I agree, but who would have thought that Labor Day would bring snow to Mt. Bachelor in Oregon?
http://www.katu.com/news/local/57656322.html
Very interesting plots on the lag times and sequence.
If we accept at face value that, the sequence might be that at some low SST the winds essentially stop, or drop to a low enough threshold so that evaporative heat loss does not keep up with solar isolation heat gain. Then that changing heat balance, starts a gradual warming of the SST. That as the SST rises, water vapor and rainfall rise almost in step with it, with about a 0.5 – 1 month delay. Then all the others follow.
Now for a back test! Did that pattern and sequence occur during the 1998 SST peak?
Unfortunately, as mentioned in the original post, this data set only goes back to 2002, but can the necessary data be assembled from sources that were around during that time, perhaps satellite, and or buoy based data sets that report the three primary variables, SST, water vapor, and wind speed for the 1998 super El Nino area.
For example, was there a wind gyre formed over that area which had very low wind speeds, or maintained high enough humidity, without rain out, to effectively cut off evaporative cooling of the Sea Surface?
If the concept is correct, you would need to have a combination of two variables, both wind speed and prevailing humidity determine evaporative cooling of the surface.
Is there a wind pattern, barometric pressure system configuration or vertical lapse rate profile that would trap high humidity air near the surface in this area and prevent cloud formation and rain out?
Is that pattern stable (ie persists for weeks or months)?
Becalmed sailors stories lead me to believe the answer is yes.
Does that pattern result in low wind speeds or a captive cell of high humidity at the surface that cannot cool the ocean surface by evaporation?
If so, even at moderate to low wind conditions little if any evaporative cooling would happen if you had some sort of a cap on convection and mixing.
The effect is obvious for anyone who has experienced a hot muggy night where even though there is a breeze, it is so humid that it does not cool you off.
Larry
I too do not like the term “anomaly”. When i was working as an applied mathematician doing orbit determination and trajectory analysis we used the word residual which to me has a different meaning to me then anomaly.
We also throw away the outliers as bad data while the climate change gurus seem to think they are the truth incarnate.
Dear Dr. Spencer
First of all thank you for all your efforts on this blog. It is really nice to see important people spending their time in the open space!
I take that opportunity to ask what are certainly very basic questions, but up to now I couldn’t find any clear information about it, certainly due to my lack of physics knowledge.
This is about remote measurements: what is in fact measuring the sensors in the satellites?
For SST I understand this is about microwave, but what about other measurements (wind, water vapor, rain) ?
And for the microwave measurement:
– why only the sea surface? and not land?
– how is it known it is actually the sea surface temperature (I saw the calibration, but hey, proxy reconstruction for past temperature uses calibration as well :))
– is it temperature of the first 10 cm of sea? 1 meter? temperature of the air on top of the surface?
I still have a box full of other questions of the same type, but that is certainly enough for now.
So if you could provide me with some hints, papers or books where I could find those explanation, I would really appreciate.
Of course anyone who has an answer is welcomed to answer as well 🙂
thank you in advance
Ron de Haan (14:34:45) :
No doubt, as expected and predicted by all barycentrists.
So if the sun heats the water, and wind cools it, what creates the wind? If I remeber elementry school correctly, it is water evaporation wich is trigered by warm water (wich we now now is due to lack of wind). This all produces a nice feedback loop to keep everything in check. These are the kinds of things that will slowly add up and kill the warmies. Yah!
My understanding of an anomaly is that it is the difference from some base period average. So if TBj is the average temperature for month j in the base period, then the anomaly Aj,y is simply Tj,y-TBj where y denotes the year and j the month and Tj,y is the observed temperature for that year and month. If I am correct in this, then one cannot say, as Dr Spencer has, that August has cooled compared to July. It could be that the observed value of .12 for August is the largest ever recorded for August. It would then be a stretch to say that August has cooled.
Stringing together anomalies in a time series seems problematic to me.
Seems like anomaly is quite unpopular; which suggests many of us here aren’t formally trained in “climatology”; which includes me; nor do I have formal training in Astrology.
I’d prefer to see data in SI units of course; but I’m perfectly ok with letting the climate science guys assign their special non-colloquial meaning to “anomaly”, and simply ignore my layman understanding of that word.
Likewise in Photometry; we don’t like to use the word “brightness” because it has a common everyday connotation that is unspecific enough to be useless in Photometry, so we use “Luminance” instead, as the Photometric analog of “Radiance”; which of course is measured in Watts per Steradian, per square metre. Lay persons are likely to use “Brightness” instead of Luminance, Luminous Intensity, Luminous emittance, Illuminance; all of which are quite different Photmetric quantities; they have neither the same dimensions nor units., so to lazily use brightness is to thoroughly confuse the reader.
So i’m happy to discard, my lay dictionary understanding of “anomaly” as meaning off base, or improper; but I would still rather see this brought into the mainstream of ordinary physics metrology.
It can get a lot worse; some physicists like to use a system of units which says that e = c = h (or maybe it’s “h bar” =h/2pi) = 1; works for them I guess.
Climate scientists could help their own cause, if they simply added to their graphs the SI value of the zero point of their anomaly data.
It doesn’t matter to me whether the base line is the 1960-2000 average value or whether it is the average from 1066 to 1776 ; just tell me the SI unit value.
As for all the curves that Roy Spencer (Dr) has posted here; notice that no cause and effect information is included; they are as Al Gore likes to say, clearly correlated, but that alone doesn’t tell us who did what to whom.
But I believe that these sorts of data graphs, and Frank Wentz et al’s paper “How Much More rain will Global Warming Bring ?” SCIENCE July 7/2007 make a powerful argument (along with the cloud graphs, which weren’t in that paper), that the oceanic water /cloud cycle is what is in robust feedback control of the earth’s temperature range; and it all doesn’t have much to do with CO2; it is entirely dictated by the physical and other properties of H2O and that weird 104 degree angle of the molecule.
George; waiting to celebrate the new ice season.
The Cloud Water measurement appears to be increasing over time (2002 -> 2009) relative to all the other measurements (rain rate, wind speed, sst, water vapor) .
Is it possible to confirm this increase?
If it exists could it be related to the solar cycle?
The only thing missing from the charts is how solar insolation varies with water vapor, rain, and/or cloud water. That might complete the cycle. As those three decrease, insolation increases and heats the sea surface.
Call it what you will, my understanding is, departure of a point from the mean (mathematical average).
To my question … Dr Spencer (or Bob Tisdale) ?
Inspired by E.M.Smith and his delving into GISS I had a look at some of the long record stations and plotted my own anomaly graphs.
All I did was take a station with a long continuous record (from GHCN mean V2) and set a base line for a particular month that was the average (say x100 years) of that month. With the base line I then plotted the anomaly for that month for each year of record for that particular station. (no smearing/in fill from other stations)
(now that is out of the way) What I think I see is that the anomaly in the area I focused on (Iceland) in July is remarkably (stunningly) stable over 100+ years.
(the questions) Am I right in seeing this stable range of anomalies (for one station)? Does satellite data show the same stability at any point on the globe?
Looking at the graphs, it occurs to me that a simple global scale climate model might have good predictive ability over a 6 to 12 month time scale and who knows, perhaps over longer timescales.
A climate model with good predictive ability 12 months out would create quite a stir.
re: curiousgeorge’s SOI link and values, it seems that the Australian Bureau of Meteorology’s understanding of the current state of the index is different to the one quoted. They reckon it’s -1.4 at the moment. Their data is available on this page:
http://www.bom.gov.au/climate/enso/soi.dt3
It’s used to generate the graph on this page:
http://www.bom.gov.au/climate/enso/
-1.4 is still in El Nino territory. It has been for a while.
Maybe we’re comparing different values?
wkkruse (15:08:39) : My understanding of an anomaly is that it is the difference from some base period average. So if TBj is the average temperature for month j in the base period, then the anomaly Aj,y is simply Tj,y-TBj where y denotes the year and j the month and Tj,y is the observed temperature for that year and month. If I am correct in this, then one cannot say, as Dr Spencer has, that August has cooled compared to July. It could be that the observed value of .12 for August is the largest ever recorded for August. It would then be a stretch to say that August has cooled.
The “anomaly” in a single graph must be the difference from some single absolute value (average) or it would make no sense. I dont think you are correct in saying that each month is the difference from some average of that month, or the averages keep changing.
Hence August will have cooled compared to July
Chickens and Eggs me suspects.
Richard, See
http://discover.itsc.uah.edu/amsutemps/execute.csh?amsutemps. Note that the global temperature varies substantially over the course of a year. If the anomaly were based on the single yearly average, the January anomaly would always be negative and the July anomaly always positive. Each month has to be compared against its own average or the anomaly doesn’t make sense.
My point is essentially that the anomalies for each month have to be normalized in order for them to be compared to one another.
For the record, I don’t like “normal” and “anomaly” either, but they have specific meanings (average, and departure from average are probably better terms) that the field is stuck with.
Microwave SSTs probably represents the upper few millimeters of water, but I’ve seen claims of even a cm or more.
Land surface temperature retrievals are much more difficult because the emissivity of land varies so widely. Ocean surface emissivity variations are much better behaved…Frank Wentz is probably the leading expert in this area.
Wind speed is from the polarization-dependent increase in brightness temperature due to both foam and capillary waves on the surface. It’s empirically calibrated to a 10 m anemometer height. Accurate retrieval of SST requires a good simultaneous accounting for the wind effect, since both act to increase brightness temperatures, but by different proportions in H versus V polarizations.
Water vapor retrieval is from the frequency dependent microwave absorption by water vapor near the 22.235 GHz absorption line. It is probably one of the most accurate of the microwave retrievals, although the fact it is vertically integrated reduces its utility.
Rain retrievals are from the frequency dependent absorption by microwaves by raindrops…even though its calibrated in terms of a surface rain rate, it is more directly related to the total vertically integrated water content, with large drops (precipitation-size) having a greater influence that small (cloud water) drops. This (rain retrieval) is the field I was originally an ‘expert’ in, before global warming took over my life.
The separation between cloud water and rain water retrievals is pretty fuzzy, as one inevitably affects the other.
wkkruse (16:52:38) : ..the global temperature varies substantially over the course of a year. If the anomaly were based on the single yearly average, the January anomaly would always be negative and the July anomaly always positive. </I
Correct and if you examine that graph thats how it seems to be
Each month has to be compared against its own average or the anomaly doesn’t make sense. My point is essentially that the anomalies for each month have to be normalized in order for them to be compared to one another.
No a single graph has to be compared against a single average (a range of years not one) or it makes no sense
LarryOldtimer (10:13:56) :
…
But then, science has become a Humpty Dumpty world:
When I use a word,’ Humpty Dumpty said in rather a scornful tone, `it means just what I choose it to mean — neither more nor less.’
`The question is,’ said Alice, `whether you can make words mean so many different things.’
`The question is,’ said Humpty Dumpty, `which is to be master – – that’s all.’
Who would have thought that Lewis Carrol would turn out to be a futurist, foreseeing the craziness that “science” has become? We have indeed “gone through the looking glass”.
Lets hope that we don’t see an emergence of a “Queen of Hearts” figure on the political landscape, declaring “Off with their Heads!”
I may have read the above wrong, but the first sentence was OK and the second seems to contradict it.
For an individual station we have records (January in this example). They are 100 years worth of average January temperatures. To find the mean “baseline” we might sum them all and /100. Then for each year in the series we compare each individual “Jan” to our calculated baseline for jan and come up with a difference (the anomaly).
Comparing different months is pretty meaningless. Anomalies are only valid compared to themselves. Even then the absolute anomaly has little value in the sense that if the anomaly for September (08) was +10 and the September (09) anomaly is -10 then the overall trend is zero.
Earlier in the comments I asked about the narrow band that July anomalies in Iceland sit – stunning I thought. You can see the peak of the trend in the 30’s/40’s and the trough in the 60’s/70’s but overall (100+ years) it seemed to be remarkably stable year on year. Three sounds like a lot of anomaly until you see 100 years of -3 to +3.
Reading anything into a spot anomaly of 0.18 is just anal as far as I am concerned. Going back to the Icelandic stations a particular January (very wild month in Iceland) anomaly might be -6.00 but it means little, the following year may be +4.00. If that continued over the 100 years we would be “loosing” 2.00 per year and would all be living in ice cathedrals by now. It doesn’t, which leaves us arguing over 150 year trends of (imaginary) 0.6/Cent. and how best to flagellate ourselves – far more anal IYWMO
PS those are sea temperaures they vary differently to air temperatures you pointed me to