Wednesday, October 28, 2015

Does Senator Smith Know NOAA's Adjustments *Lower* the Trend?

I wonder if Senator Lamar Smith, who wants all the data, emails, and the birthdays of the children of the NOAA scientists who contributed to Karl et al, Science (2015), knows that the adustments made to the raw data lower the long-term warming trend?


That would be fun to watch in Congressional testimony..... And that the raw data is easily available, as Karl et al write in their supplementary material:
"The data sets are the publically available ERSST v4 data set of sea surface temperature anomalies (13) ftp://ftp.ncdc.noaa.gov/pub/data/cmb/ersst/v4, and the ISTI Databank v1.0.0 of land surface air temperature (14) http://www.surfacetemperatures.org/databank." 

12 comments:

David in Cal said...

Data adjustments are hard for me to evaluate. E.g., one study looked a the qualilty of the weather station sites:

The new improved assessment, for the years 1979 to 2008, yields a trend of +0.155C per decade from the high quality sites, a +0.248 C per decade trend for poorly sited locations, and a trend of +0.309 C per decade after NOAA adjusts the data.

I don't have enough expertise to compare this study with NOAA's adjusted trend. However, it's worth noting that the trend based on troposphere studies 1979-2015 is pretty close to the 0.155C/decade trend from high quality sites. In any event, it seems to me appropriate that NOAA make public their adjustment methods.

Cheers
David in Cal

Thomas said...

Unknown, the "study" you refer to never got further than a blog post on WUWT. You can find a critique here:
http://www.skepticalscience.com/watts_new_paper_critique.html

David in Cal said...

Thomas, the critic you cite correctly points out that the change in time of observation needs adjustment. However, I do think it's better to include only well-sited station. Sadly, this is such a political football that AFAIK nobody makes both sets of adjustments. Warmists adjust for time of observation, pushing the trend up; skeptics adjust for station siting, pushing the trend down.

Because of the uncertainty in all these adjustments, I personally think the Troposphere temperatures are a more reliable basis for estimating the warming trend. As I said, the trend on this basis for the full period available, 1979-2015, is somewhere around 1.5 degrees C per century. I have read that physicists say that the impact of rising CO2 would cause temperatures to rise at roughly this rate, if there were no positive or negative feedbacks. For these reasons, my personal opinion FWIW is that this figure is the best guess we can make for temperature rises going forward.

Cheers
David in Cal

David Appell said...

David in Cal: Why do you accept adjustments to the tropospheric model, but not to the surface model?

It seems only because you like their results.

David Appell said...

PS: To believe there are no feedbacks, you'd have to believe that

1) Ice doesn't melt at higher temperatures, and
2) Evaporation doesn't increase at higher temperatures.

Toby said...

David in Cal,

Tropospheric temperatures from satellites go through far more adjustment than any ground data - there are issues with the atmospheric layers the microwaves (not infra-red, mind) must pass through, with satellites orbital decay, with calibration of sensors that have been years in space, and with the difficulties of an elliptical orbit. And satellites are not even measuring surface temperature, anyway, and if anything the last two El Ninos (1998 & 2010) led to an over-estimate of temperature by satellite. Carl Mears of RSS is on the record to say that ground station data is more reliable than satellite data.

JoeT said...

David in Cal,

There are few to none legitimate scientists who would argue that you can possibly get a temperature increase of 1.5C/century without feedbacks.

Let's do a simple calculation. First of all, I'll give you your 1.5C/century. NASA GISS would be more like 0.16C/decade from 1970 to now. UAH in the old version (before the latest "update") agreed very well with that trend.

But let's use your value. In 1970 CO2 concentration was 326 ppm and in 2015 it will be around 400 or so. That 45 year span then gives you a temperature rise of 0.675 C.

We can then calculate a simple transient climate sensitivity from

S = delta T/log2(C2/C1) where delta T = 0.675, C2 = 400 and C1 = 326.

That gives us a transient climate sensitivity of 2.3 C per doubling of CO2. That number is in the ballpark of more sophisticated analysis. The equilibrium climate sensitivity would be even higher. CO2 by itself only gives 1.1 per doubling of CO2.

This alone shows that feedbacks are important.

David Appell said...

Unknown 10/29 wrote:
"I don't have enough expertise to compare this study with NOAA's adjusted trend."

Personally, it's far easier for me to understand adustments to the surface stations, than adjustments to the satellite data.

David, can you even list all the adjustments done to the satellite data?

I can't.

David Appell said...

David in Cal wrote:
"In any event, it seems to me appropriate that NOAA make public their adjustment methods."

David, can you show me where RSS and UAH make their adjustments to the satellite data known?

In fact, can you point me to the paper where UAH explains their version 6? (No, you can't. They put the monthly numbers out there, but haven't published a paper on their methodlogy.)

David Appell said...

JoeT wrote:
"That gives us a transient climate sensitivity of 2.3 C per doubling of CO2."

And this is probably a lower bound, since anthropogenic aerosols are a cooling factor, holding down GHG warming.

Gerkmonster1 said...

David Appell writes : David, can you even list all the adjustments done to the satellite data?


So David, Can you list all the adjustments done to the surface data... How many times has the 1880 data, at any given site been adjusted? Do you have a list of revisions dating back to the original recording?

David Appell said...

Gerk: The adjustments are done anew every month (and the methods vary by research group). So a list of revisions wouldn't be very meaningful.