New Temperature Versions

UAH has a new version 6.0(beta) which includes some adjustments of course.  It is being gone over with a fine tooth comb by the usual suspects since Spencer and Christy are notorious "skeptics".  "Surface" temperatures are mandatory because everything climate is based on "surface" temperature change.  "Surface" temperature will always have some issues because there is no real surface.;  Since there is an elevation consideration over land and a lapse rate that is variable, there would need to be considerable altitude and specific heat capacity adjustments, latent heat is really "hidden" as far as temperature goes and the ocean readings are mainly sub-surface rather than surface readings in many cases.  Satellites measuring the lower troposphere have to estimate a specific altitude which appears to be around 2000 meters based on the RSS version on Climate Explorer that is available in degrees Kelvin.  So they are measuring a different "surface" with different latent heat considerations.

Since satellites tend to have issues at the poles as do "surface" stations, I tend to prefer looking at the tropics, specifically tropical oceans since they represent the lion's share of total energy.  The chart above compares the UAH beta version with the newest version of ERSST.  What I see is a pretty fair comparison considering all the issues involved with both products.  There is a very small difference in the two trends and as usual the lower specific heat troposphere has more variation than the high thermal mass ocean surface/sub-surface.  I am of course no body, so I will leave it to the 'spurts to really screw this up.

 For the "global" oceans there is a little bit bigger trend difference which could be due to any number of real and calibration issues.  UAH is trying a different averaging method with the intent of improving regional temperatures.  ERSSTv4 I believe is more focused on a "global" average which would mean longer range interpolation.  ERSSTv4 no longer uses the Reynolds oiv2 satellite temperature data for its interpolation, because is caused some "significant" cooling which would most likely improve the correlation between these two data sets.  The difference really doesn't amount to a hill of beans, but the hyper-precision junkies will find some flaws that they think are "significant".

Since land "surface" temperature is an average of Tmax and Tmin and there are rumors the Tmin is suspect due to nocturnal atmospheric boundary layer variation, I am staying out of that mess.  What will likely the case though is the longer the interpolation range the greater the discrepancy between UAH and whatever land "surface" data set.

Nick Stokes has his critique on his blog and Roy Spencer has his pretty detailed explanation of the changes on his blog.