29 November 2015

Teasing the data to get the trend you want

"The cherries are picked mainly on the plains..."

Feldman et al 2015 (link) [1, 2] is supposed to be one of these 'smoking guns' of CO2-induced doom Team Consensus likes to refer to [3, 4]. Here's a graph from that paper.

At first glance a few things seemed odd:

28 November 2015

Harries et al 2001 corrected graphs

The labels were switched on graph 1a. So I switched them back to correct them. You can verify that the graph is now correct considering the difference graph figure 1b:


27 November 2015

The broken pea shooter

I've been looking into so-called smoking gun of greenhouse gas absorption: Harries et al 2001, and came across a rather hard to obtain, obscure 2003 follow-up paper by Harries co-authored with Brindley that basically nullifies the 2001 finding.

The paper is called Observations of the Infrared Outgoing Spectrum of the Earth from Space: The Effects of Temporal and Spatial Sampling by Brindley and Harries (hereafter B & H 2003).

Turns out sampling limitations (etc) in the later 1997 IMG instrument gave rise to errors that effectively nullify the following graph (as depicted on the SkS website):

17 November 2015

RSS Satellite Temperature Lower Troposphere Australia

UPDATE 10 March 2017: I got the longitude wrong by 180 degrees. Explanation here:


Graph should look like this:

It was still close enough though. There is no warming in Australia during satellite period.


Now that I can open NetCDF files in Matlab I can look at various climate-related data files and reveal data by longitude and latitude, kind of like Steven Goddard does only not as good.

You can get the data for RSS TLT here:



Using the file:
"uat4_tb_v03r03_anom_chtlt_197812_201509.nc3.nc" I can dial in any latitude and longitude or area.

First, for comparison sake here's what I get averaging the whole earth: