I'm sure you understand that a very slight change in prevailing winds greatly affects large areas of ocean surface temperatures. Smoothing is essential along with even sampling and satellite is the only way to do that. A large part of the surface record for ocean areas is taken on islands which almost invariably have UHI (mostly airports).That's not what I've read. ERSST is based on COADS, and COADS is based on data from ships of opportunity augmented by buoys in later years when oceanographic buoy deployment became more common. I've NEVER heard of SST being measured at an island airport.
ICOADS: The International Comprehensive Ocean-Atmosphere Data Set Project
http://www.ncdc.noaa.gov/oa/climate/research/Smith-Reynolds-dataset-2004.pdf
Improvements to NOAA’s Historical Merged Land–Ocean Surface Temperature Analysis (1880–2006)
If you go to the link in the previous post ("Global surface temperature anomalies") you'll see that they say they put satellite data in, but this caused problems, but they took it back out. I'm not sure how well satellite data works with SST data, because satellite must only look at the exact top of the ocean radiatively, while even a buoy or a ship is measuring some volume of the surface.
The reason that you sense no resolution to the adjustments argument is that GG has moved on. The statistical methods used for adjustments have satisfied him since there are even numbers of positive and negative trends in a gaussian distribution. That should be expected from a homogenization adjustment (not an adjustment for UHI or other physical error).
Oh well.
As Eschenbach showed (and verified by others, e.g. http://diggingintheclay.blogspot.com/2009/12/reproducing-willis-eschenbachs-wuwt.html, the effect of homogenization is to introduce artificial trends, positive in the case of Darwin (but also negative).
Well, Darwin's one station. What's it do to the rest of Australia?
While the bottom line effect is "only" about a 0.15C to 0.2C rise per century overall (depending on methods), there are drops in 1940-1970 and much faster rises in 1975-2000. The essence of what GG is missing (or hiding) is that the adjustments vary over time to create recent warming compared to the raw data. It is impossible to see that on GG's graph that adds up all adjustments over all time periods. GG repeats that several times in his replies to critics. To quote one "If you adjust down first, then up you do NOT create a total warming trend: you create two trends one cooling and one warming that sum with each other. Why would you consider only the latter? It seems many people got stucked on this. I am going to explain it better in my post in a little while." He never explained it better.
Unfortunate.
This still doesn't go back to my original and probably more important question. What's wrong with the graph in post #73? Because if there's nothing significantly wrong with it, then the bet stands as stated -- and it should be obvious why. (Or more succinctly; if LT and surface trends match so well, then what's wrong with the adjustments?)
Or maybe this one is more clear:
From Before and after
What's wrong with the graph in post #73 Because the data for the chart in post 73 is wrong. Look at the difference between satellite LT there and satellite LT in post 64. Take the 80's and 90's (until 97) from Spencer's graph in post 64 and see that it fluctuates around zero for the period going as low as -0.5 and a bit above 0.3 In Tamino's graph in post 73 he has picked points (not averaged or smoothed) that lie between -0.3 and a bit under 0.2. In other words he compressed the y axis over the early years.
Then in 1998 all of a sudden he accurately captures the peak at 0.6 But after that he compresses the y axis again from about 0.25 to about 0.38 after 2001 or so. The actual values for that time period are from -0.1 to about 0.6 Then Tamino seems to truncate the x axis to avoid displaying the -0.2 in 2008.
Somehow the chart in post 73 contains y-axis data compressed by some unknown process. As for the chart in your latest post, it is from a different Tamino thread http://tamino.wordpress.com/2008/03/02/whats-up-with-that/ which is a junky post in response to Anthony Watts' even more junky post. Tamino uses graphic spaghetti to merge satellite with GISS by using the same shades of red and adding Watts' two other irrelevant data sets rather them separating them and comparing by pairs.
I've NEVER heard of SST being measured at an island airport.
The surface record uses lots of island airports to fill the grids in ocean areas whereas the satellite record simply measures within the grids whether there are islands there or not and thus more faithfully reflects SST. The result is a gradual warm bias in the surface record but with suppressed SST effects since the airport doesn't feel the full effect of the El Nino SST increase mainly due to island induced rain clouds. The result is exactly what we see, a surface temperature record that relentlessly rises but suppresses the El Nino peaks, compared to a satellite record as shown in post 64.
Well, Darwin's one station. What's it do to the rest of Australia?
Darwin was apparently adjusted using data from neighboring islands (not Australia). The other northern Australian stations don't show warming (see http://joannenova.com.au/2009/12/smoking-guns-across-australia-wheres-the-warming/#more-5172 for example), but GHCN does show warming overall. In this case adjustments via data homogenization have a warm bias over a large area, not just that one site although it does appear to be a worst case (Willis says he chose it because it was the first one on the list).
The bottom line in all of this is that we are trying to compare apples to apples, or 98 El Nino measured most accurately by satellite to the upcoming El Nino. Not to data that is both adjusted with a warm bias and polluted with both UHIE and cooling effects.