Comments on Spencer’s strange statements on CRUTEM3

I was reading this article from Open Mind the other day and I even read the comment section. There a commenter named “Robert” quoted Roy Spencer saying something about CRUTEM3 surface temperature set (the land part of the HadCRUT3 land & ocean surface temperature analysis) that really seemed strange to me. Below I’ll explain why.

The quote in question is from this Roy Spencer blog article. There Spencer first explains how he is in the process of comparing the CRUTEM3 and NOAA ISH datasets (link is to a summary page, Spencer uses more complete data set which doesn’t seem to be available for laypeople). He then shows a figure showing both datasets from 1973 to 2009 and notes:

But while the monthly variations are very similar, the warming trend in the Jones dataset is about 20% greater than the warming trend in my ISH data analysis.

The comment I found strange comes next:

This is a little curious since I have made no adjustments for increasing urban heat island (UHI) effects over time, which likely are causing a spurious warming effect, and yet the Jones dataset which IS (I believe) adjusted for UHI effects actually has somewhat greater warming than the ISH data.

So, what is there so strange about this quote?

First, Roy Spencer is a scientist so he should know how to look for information on these things and yet he just seems to proceed by guessing (as he only “believes” CRUTEM3 is adjusted for UHI). Now, how should we check if “Jones dataset” (CRUTEM3) is adjusted for UHI effects? This is the website Spencer got the data from. Would it be a good idea to look what they say about their dataset? Well, they don’t say much but they do give references to scientific research articles on their dataset. Roy Spencer as a scientist most likely has access to all of the given papers but gladly they offer the latest paper for free there, here’s the link (Brohan et al. 2006).

Second, CRUTEM3 is not adjusted for UHI effects. The UHI effect is included to the uncertainty values which doesn’t show if one only uses nominal values like Spencer does (shouldn’t a scientist consider uncertainty too when determining a difference between two datasets?). You can see the effect of UHI in Brohan et al. Fig. 10 where it is as one factor to the blue uncertainty band. (IPCC AR4 WGI also discusses this.)

There you have it: regardless of what is causing the difference in the datasets Spencer is comparing (and I see Spencer pondering much about CRUTEM3 and UHI but I don’t see him considering the ISH dataset at all – it is as if Spencer would have already decided before the analysis that there is a problem in the CRUTEM3 or in the “Jones dataset” as he calls it for some reason), my message here is that the approach presented in Spencer’s blog entry doesn’t appear very scientific and the strange part is that the Author of the blog entry is a scientist. But when I read the final conclusion (“It is increasingly apparent that we do not even know how much the world has warmed in recent decades, let alone the reason(s) why. It seems to me we are back to square one.”), the whole thing takes even more bizarre turn. Here we have what seems to be half-baked comparison of two datasets followed by what seems to be quite standard nonsense conclusions you usually see in denialist blogs. From a scientist I would expect far better than this. I cannot help wondering if there are two persons who call themselves Dr. Roy Spencer and write about climate issues.

P.S. There has been lot of research on urban heat island effects, by the way.