Are you sure about that claim?

Absolutely.

Read The Code.

Published papers are fine to get an idea what someone claims or maybe even what they intended but the only thing that actually does anything is the code.

And the code says “ALL LAND ALL THE TIME THROUGH STEP3″. (Which step quite definitely creates fictional sea temperatures in boxes of ocean near islands. See the links I gave and you can see the log file output showing the particular ocean boxes created from particular LAND stations on islands.)
It then says “Optionally, you MAY add SST data via the HadSBBX data from Hadley”.

I have no idea what goes into the links you gave, I do know exactly what GIStemp does. It is also possible that the sea temps you are looking at are via the Hadley file merger in STEP4_5, but that just puts us back at asking what Hadley’s dog did with their homework…

The rest of your argument appears to amount to “Such a coincidence seems to unlikely to be true”. You might enjoy an anecdote from Richard Feynman:

“Feynman walks into a class late, and announces to the class that he has encountered something astounding. While walking through a parking lot, he saw a car with the plate number 186CSC. What, he asks the class, do they think the odds are of seeing that precise number?”

More seriously, again unless you account for the fact that the response to volcanic eruptions is DEFINITELY amplified in the LT relative to the surface, OF COURSE you will get a higher satellite trend starting from right after one which diverges less from the surface record.

Now, mind telling me why the response to volcanoes, El Ninos, etc. is clearly amplified and the long term trend isn’t? I’ll tell you it seems clear to me that the best explanation of this difference is that the long term trends in one or more datasets is WRONG.

John Finn (02:48:58) : That trend line is 1. starting right in the middle of Pinatubo, and the amplification factor for volcanoes appears to be even greater than 1.2 (see the plot I showed above) and 2. Right around the time of RSS’s step change from the transition between NOAA 12 and 11 and 10.

I’m not sure I understand your points. I used 1992 because off the discrepancy between RSS and UAH before then. What ever the reason for the discrepancy it no longer seems to exist.

Can we agree that since 1992 RSS and UAH now agree:

http://www.woodfortrees.org/plot/rss/from:1992/to:2009/trend/offset:-0.05/plot/rss/from:1992/to:2009/offset:-0.05/plot/uah/from:1992/to:2009/trend/plot/uah/from:1992

It is also the case that, since 1992, Hadley and GISS are very much in agreement with the 2 satellite records.

Perhaps I’m doing you a disservice but your argument seems to hinge on the tropospheric amplification factor (1.2) being exactly the same as the UHI trend in the surface records.

1880-11/1981: Hadley HadISST1 (Rayner 2000),
12/1981-present: Reynolds-Rayner-Smith (2001)

Are you sure about that claim?

Look at these maps:

http://data.giss.nasa.gov/cgi-bin/gistemp/do_nmap.py?year_last=2009&month_last=08&sat=-1&sst=1&type=trends&mean_gen=0112&year1=1979&year2=2008&base1=1951&base2=1980&radius=1200&pol=reg

http://data.giss.nasa.gov/cgi-bin/gistemp/do_nmap.py?year_last=2009&month_last=08&sat=4&sst=0&type=trends&mean_gen=0112&year1=1979&year2=2008&base1=1951&base2=1980&radius=1200&pol=reg

http://data.giss.nasa.gov/cgi-bin/gistemp/do_nmap.py?year_last=2009&month_last=08&sat=4&sst=1&type=trends&mean_gen=0112&year1=1979&year2=2008&base1=1951&base2=1980&radius=1200&pol=reg

Does it look like the combined land and sea surface looks more in the over lap areas in the land only map or the sea surface only? Note the strong “land” warming in the South Atlantic, which isn’t present in the sea surface only…and isn’t present in the combined either. I suspect that the land surface may be an issue closer to shore, but I don’t see the issue you imply showing up.

As pointed out in the above links, the “Sea” temperatures in GIStemp are UHI adjusted LAND temperatures that are smeared out over 1200 km radius boxes of OCEAN surface. Then, and only then, may the Hadley SST anomaly map be merged in to fill in any missing bits.

IMHO, the application of that UHI error bias correlation in this analysis is sound in direct proportion to the degree to which GIStemp (and one presumes Hadley, if only they would tell us what their method might be…) does in fact spread UHI land data over the sea.

Basically, you have no sea temperatures to work with only “anomaly boxes” based on land temps from STEP3 with a fudge applied from some simulated interpolated spliced corrected anomaly maps (in optional STEP4_5) for any boxes not already fabricated from the UHI land data in STEP3.

And that, boys and girls, is why “reasoning from what ought to be reasonable” is useless with this AGW game of “hide the data”. You MUST look at the code and see just what shenanigans were really being pulled.

And that is not pretty.

http://chiefio.wordpress.com/2009/03/05/illudium/

Here is a short quote:

http://www.emc.ncep.noaa.gov/research/cmb/sst_analysis/

Oh, and the satellite data are adjusted based on an optimal interpolation method. We’re getting even further away from “data” and into the land of processed data food product…

So your “Ocean Temperatures” are not quite temperatures and didn’t quite come from thermometers. These are then BLENDED in with all those boxes and grids made up in STEP3 of GIStemp that are largely based on AIRPORTS on islands. There are dozens of Islands in the Sun, each with an airport warming an area of ocean about 1/2 the size of the continental USA. To the extent a ship wanders into that box, it might contribute a bit to the “average of averages”, but good luck figuring out how much. At this point both anomaly maps are way past the average-the-temperature steps and all you have is to anomaly maps to merge. One data point each per box.

GIStemp is all about smearing airport tarmac into the countryside and oceans.

http://chiefio.wordpress.com/2009/08/26/agw-gistemp-measure-jet-age-airport-growth/

And it’s UHI method is broken too:

http://chiefio.wordpress.com/2009/08/30/gistemp-a-slice-of-pisa/

No…I believe that most of the ocean temperature data are from sea surface temperatures measured by ships.

Nope. You have a simplistic and wrong understanding of the process.

GIStemp uses only LAND data up through STEP3 (all the anomaly, boxing, gridding, etc. steps). Only in the optional STEP4_5 does it allow you to bring in a Hadley SST anomaly map (not actual temperatures) and that is based on a variety of things including some completely made up re-imagined data. But since Hadley raw data have become “The Dogs Lunch”, it is rather unclear what value it might add.

http://chiefio.wordpress.com/2009/09/08/gistemp-islands-in-the-sun/

FWIW, STEP4 is pretty ugly:

http://chiefio.wordpress.com/2009/03/07/gistemp-step4-the-process/

Luckily, the STEP3 script makes it clear that it is optional. From:

http://chiefio.wordpress.com/2009/03/07/gistemp-step3-the-process/

echo ; echo “If you don’t want to use ocean data, you may stop at this point”
echo “You may use the utilities provided on our web site to create maps etc”
echo “using to_next_step/SBBX1880.Ts.${label}.$rad as input file” ; echo

echo “In order to combine this with ocean data, proceed as follows:”
echo “move SBBX1880.Ts.${label}.$rad from STEP3/to_next_step to STEP4_5/input_files/.”
echo “create/update the SST-file SBBX.HadR2 and move it to STEP4_5/input_files/.”
echo “You may use do_comb_step4.sh to update an existing SBBX.HadR2 file”
echo “You may use do_comb_step5.sh to create the temperature anomaly tables”
echo “that are based on land and ocean data”

And what sea data is used is hideously mangled by the time it gets to GIStemp as an “anomaly data food product”:

http://chiefio.wordpress.com/2009/02/28/hansen-global-surface-air-temps-1995/

Inclusion of Marine Temperatures

Yup, next paragraph. They talk about “Empirical Orthogonal Functions” used to fill in some South Pacific data… but it uses “Optimal Interpolations” which sure sounds like they are just cooking each datapoint independently… From here on out when they use EOF data they are talking about this synthetic data. It also looks like they use 1982-1993 base years to create the offsets that are used to cook the data for 1950-81. Wonder if any major ocean patterns were different in those two time periods, and just what surface (ship / bouy) readings were used to make the Sea Surface Temp reconstructions? They do say “The SST field reconstructed from these spatial and temporal modes is confined to 59 deg. N – 45 deg S because of limited in situ data at higher latitudes.” OK, got it. You are making up data based on what you hope are decent guesses. But in GIStemp “nearby” can be 1000km away with no consideration for climate differences, so I’m concerned that the same quality of care is being given here.

In short, no real sea data need apply to the GIStemp temperature series. Only a fabricated interpolated filled in sparse modified composite splice anomaly map is used.

Yeah, it’s that good…

This is not significantly different from what GIStemp does. It UHI adjusts the data (badly, IMHO) and then uses these UHI data to fabricate “boxes and grids” of temperatures over the ocean up to 1200 km from the source thermometer.

That means, for example, that the Military Base at Diego Garcia warms an area the size of 1/2 the United States, and it is almost 100% ocean.

So goose, meet gander.

This is not significantly different from what GIStemp does. It UHI adjusts the data (badly, IMHO) and then uses these UHI data to fabricate “boxes and grids” of temperatures over the ocean up to 1200 km from the source thermometer.

That means, for example, that the Military Base at Diego Garcia warms an area the size of 1/2 the United States, and it is almost 100% ocean.

So goose, meet gander.

John Finn (02:48:58) : That trend line is 1. starting right in the middle of Pinatubo, and the amplification factor for volcanoes appears to be even greater than 1.2 (see the plot I showed above) and 2. Right around the time of RSS’s step change from the transition between NOAA 12 and 11 and 10.

John Finn (12:15:56) : Except that if you go over to RC, they will tell you-beat it into your head in fact-that warming simply could not occur without being amplified aloft, whatever the cause .

I’m not here to defend RC, but my interpretation of their article is that they do expect amlification to occur. They do not say that is happening. In any case you appear to be saying there are 2 wrongs here, i.e. temperature discrepancy and tropospheric hot spot. If the ‘hot spot’ theory is wrong – then there is no discrepancy and vice versa.

Incidentally most of the disagreement between UAH and the others appears to originate in the pre-1992 period. Roy Spencer has commented on the RSS/UAH disagreement on his blog. The trends for all 4 records since ~1992 are remarkably similar. Here’s Joel’s plot using 1992 as a start point (and a slightly different offset)

http://www.woodfortrees.org/plot/gistemp/from:1992/to:2009/trend/offset:-0.28/plot/gistemp/from:1992/to:2009/offset:-0.28/plot/rss/from:1992/to:2009/trend/offset:-0.05/plot/rss/from:1992/to:2009/offset:-0.05/plot/uah/from:1992/to:2009/trend/plot/uah/from:1992

The very slightly higher warming trend in the GISS record is probably due to their arctic extrapolation. The arctic has been warmer over the past few years.

UAH lists their errorbars as +/- 0.05 C/decade (see http://vortex.nsstc.uah.edu/data/msu/t2lt/readme.18Jul2009 , under “Update 7 Aug 2005 *”), at least as of 2003. I don’t know what RSS and HADCRUT list their errorbars as. But, I just wonder if you are taking the data beyond where it can be trusted.

And, by the way, despite John Christy’s papers (and one paper that appears to be written by some independent folks), I don’t think the final word has yet been written on which satellite analysis trends are the most accurate.

http://www.realclimate.org/index.php/archives/2007/12/tropical-troposphere-trends/

And not only is the 1.2 ratio what is generally found in models, it is also found in the residuals of the detrended observations ~the same:

http://www.climateaudit.org/phpBB3/viewtopic.php?f=3&t=740

Because it isn’t (ACCORDING TO RC) something the models have wrong which causes this, it’s the nature of the way the system is supposed to behave.

Joel Shore (13:34:16) : Ditto, and RSS is not shown, UMD is not shown, UW is not shown, because those analyses are erroneous (the latter two almost certainly so, but the case against RSS is pretty good to.).

Christy, J. R., and W. B. Norris (2006), Satellite and VIZ-radiosonde intercomparisons for diagnosis of non-climatic influences, J. Atmos. Oceanic Tech., 23, 1181-1194.

Christy, J. R. and W. B. Norris (2009), Discontinuity issues with radiosonde and satellite temperatures in the Australian region: 1979-2006, J. Atmos. Oc. Tech., 25, doi: 10.1175/2008JTECHA1126.1.

Christy, J. R., W. B. Norris, R. W. Spencer, and J. J. Hnilo (2007), Tropospheric temperature change since 1979 from tropical radiosonde and satellite measurements, J. Geophys. Res., 112, D06102. doi: 10.1029/2005JD006881.

Randall, R. M., and B. M. Herman (2008), Using limited time period trends as a means to determine attribution of discrepancies in microwave sounding unit-derived tropospheric temperature time series, J. Geophys. Res., 113, D05105, doi:10.1029/2007JD008864.

But even if you compare surface temperatures to RSS it is still clear that the surface measures have a warm bias:

http://www.woodfortrees.org/plot/hadcrut3gl/from:1979/offset:-.15/mean:12/scale:1.2/plot/uah/mean:12/plot/hadcrut3gl/from:1979/offset:-.15/mean:12/scale:1.2/trend/plot/uah/mean:12/trend/plot/rss/mean:12/plot/rss/mean:12/trend

Again, the evidence that the surface data has a warm bias is extensive. And yet the resistance…the ghost of Kuhn is slowly nodding his head saying “Yup”.

I grow weary of having to repeat this again and again, but what you and now Shore are doing is totally inappropriate. This is the correct comparison:

http://www.woodfortrees.org/plot/hadcrut3gl/from:1979/offset:-.15/mean:12/scale:1.2/plot/uah/mean:12/plot/hadcrut3gl/from:1979/offset:-.15/mean:12/scale:1.2/trend/plot/uah/mean:12/trend

You are all growing increasingly tedious.

It is not clear to me how resilient that expected 1.2 amplification factor is. And, of course, there are some reasonably substantial errorbars on the trends, especially for the satellite data. Furthermore, I am not sure why you have chosen to plot just the UAH satellite analysis and not RSS also. In fact, I think there are now 4 different analyses for temperatures on the global scale (the third and fourth being from U of Md and U of Washington if I remember correctly) and I believe it is UAH which has the lowest trend.