Christy et al (J Clim 2009), Surface Temperature Variations in East Africa and Possible Causes, is a really excellent article that will interest many readers interested in surface temperature data sets. It’s interesting on a number of counts, not all of which I have time to summarize today.
It is a detailed study of station records from Kenya and Tanzania in East Africa an area which more or less covers nine 5-degree gridcells from 10S to 5N and from 30E to 45E. They collected and digitized original British East Africa and German colonial station data, as well as GHCN, GISS and their sources (but not CRU data), resulting in a substantial expansion of available data. Christy obviously has an excellent record of placing data online and I hope that this extends to the newly collected station data (which is not online at the moment.)
Although surface data is the backbone of temperature history, detailed analyses of station data are rare and detailed analyses of non-US non-European data are even rarer. The absence of such analyses is an indictment of the authors of the major temperature indices (CRU, GISS, NOAA). They are funded to publish temperature indices and this sort of technical study should be part and parcel of their obligations.
Christy et al approach the calculation of gridcell temperatures a little differently than GISS (CRU still not providing an operational description of methodology). First, they try to identify different versions of the same station and to obtain a station history from these versions. This is along the lines of GISS’ calculation: GISS collecting various versions in dset0 and combining them in dset1. The approach of Christy et al looks more sensible than the GISS approach, though I doubt that the difference “matters” a lot to the final answer, other than being more logical.
Next Christy et al apply a breakpoints algorithm to distinguish unreported step changes in the station history citing the radiosonde method of Haimberger et al 2007. Breakpoint detection, as I read the article, is done through internal properties of the series, rather than through neighbor comparison (a la USHCN), neighbors being a lot sparser than in the US. They set the sensitivity of the changepoint parameter at three different settings and report on its impact on the trend (it’s non-trivial.) Breakpoint detection is a complicated statistical procedure and not one that I’ve studied enough to have an independent opinion on the merits of the various approaches (both the Christy et al code and USHCN code are unarchived in any event). However, Christy et al describe their test statistic (p 3345) and their methodology seems clearly preferable to the weird GISS two-legged coercion (CRU methodology needless to say is unknown).
Then they average (“merge”) the station anomaly series. I can’t tell in a first reading when they converted the data into anomalies – their Figure 1 shows anomaly series.
Then they compare their trends with trends calculated from HadCRUT3v, CRUTEM3v and GISS, reporting that they were unable to replicate the high trend from the major indices. It seems that the CRU series is dominated by the Nairobi airport:
The recent trends of TMean calculated from global datasets do not agree with our results for this cell. As shown in Table 2, the 1979–2004 TMean trend of the central cell as produced by HadCRUT3v, CRUTEM3v, and GISS (0.31, 0.47, and 0.35 deg C/decade, respectively) are markedly inconsistent with all of the time series for that cell constructed in this study. Evidently, the main signal used by HadCRUT3v for this cell since 1979 is derived from the single Nairobi, Kenya, station at Jomo Kenyatta Airport (P. Jones 2004, personal communication). Our unadjusted time series for this site does indeed show significant warming since 1979 (0.25 deg C/decade), but the higher trend is not corroborated by the many nearby stations used in our analysis. Such differences were also found in central California (Christy et al. 2006) and northern Alabama (Christy 2002), where our more comprehensive reconstructions were on average about 0.1 deg C/decade more negative in the cells covering those areas versus values for the cell from global databases.
Christy et al continue with an interesting discussion of Tmin versus Tmax, arguing that Tmax samples a bigger volume of air than Tmin and is a more reliable index of large-scale changes (citing Pielke et al 2007).
The idea that the Jones CRUTEM series is dominated by Nairobi airport will come as little surprise to CA readers – recently we saw that the CRUTEM Hawaii series is little more than an alter ego for Honolulu airport.
I think that it is time to recognize that calling CRUTEM an index of “Land” temperature is a misnomer as, in addition to “Ocean” and “Land”, there is a third important surface in major temperature indices: airport tarmac. In appreciation for Phil Jones’ efforts at measuring airport temperatures around the world, perhaps it would be appropriate that his series be rebranded CRUTAR.
Climate Audit 
35 Comments
Steve, another great post. I love the CRUTAR idea.
Tarmac and UHI appear to a lot more important than Mr. Jones has previously insisted. No wonder he doesn’t want his methods audited.
The Surface Stations project has shown that even in the U.S. where travel and communication is easy, those in charge of the records are doing poorly at examining and managing their data.
“Surface Temperature Variations in East Africa and Possible Causes” DOI: 10.1175/2008JCLI2726.1
CRUTAR
ROTFL
How about some CRUTON?
Tasty!
CRUTAR… pretty good, but what I really get out of this posting is that some people are working really hard to make sense out of the data they have and are applying tests I never would have dreamed of. They deserve credit for that. Keeps me humble … OK, that doesn’t take much…..
Some 2 yrs ago Jonathan Lowe blogged about the Tmin issue in his Gust of Hot Air blog. He was urged at the time to publish more formally by more than one observer. He used good Aust BOM stats for Mt Gambier in South Aust. I think its the airport but its a quiet country town with a stable population. The net was that Tmin in Mount Gambier is some time after first light. Tav cannot be got from (Tmax-Tmin)/2. JL also observed that trends in Tmin Tmax vs trends in T3am T6am T9am etc make temp vs co2, aerosols, clouds etc a very complicated analysis. I wonder if JL had done a paper and got it accepted would the discussions be different today? Congrats to Christy et al.
Maybe one can paraphrase GIGO: CRUData in, CRUD out…
Has anyone ever asked Phil Jones for his code? There’s no mention of it in the file of Warwick Hughes/Eschenbach/Keenan correspondence.
I’ve not read the article, but i’m interested by the approach of demonstrating the sensitivity to various adjustment parameters. Clearly there is an issue of un-reported station moves, it looks like this study gives some insight into the uncertainty that this introduces into the record
These tarmac and other urban effects are probably omnipresent and we often view their large effects on the “mainstream” surface records as amusing.
But I think that this discussion also has its flip side: if we were behaving rationally, we would actually care about the temperature in the big enough cities/towns (and airports) because that’s where most people either live or work.
This city index would show a pretty clear warming – one that is almost completely caused by very different drivers than CO2. We might say that it is interesting that these effects have not been attacked by the green movement.
We may also define the “anthropocentric global mean temperature” which would be the average temperature measured in the vicinity of all 6.7 billion humans – i.e. it would be weighted by the population density. Such an anthropocentric global mean temperature would be much more directly relevant for human lives and it would arguably show a dramatically different behavior than either HadCRUT or UAH-like indices.
Re: Luboš Motl (#13),
The ultimate “anthropocentric global mean temperature” should be around 36.6C, no matter how it’s measured!
More seriously, Africa presents the greatest challenge of all continents in terms of historical records. Not only are long, unbroken records very rare, but their reliability in many cases is questionable. E.g., Kinshasa and Brazaville lie opposite one another on the Congo River, yet the few years of mutually available record show the kind of correlation that might be expected from stations many hundreds of miles apart and with intervening topographic barriers. On the other hand, in many cases urban heating is relatively modest compared to the developed world, where automobiles on paved streets are ubiquitous and colder winter temperatures require buildings to be heated.
Christy et al. should be commended for tackling the region of Africa whose historical temperatures are most in doubt. GHCN seems quite content to exploit that doubt, by updating urban records quite selectively, which GISS and HADCRUT then project regionally. Thus Nairobi is current, whereas Zanzibar, Harare, Antanarive etc. are not. The results of such selectivity are all too predictable.
Re: John S. (#18), Hence the need for a physical meteorological understanding over there as elsewhere. I cannot insist enough on the imperious need to read Marcel Leroux’s books:
“Meteorology and climate of Tropical Africa” Springer-Praxis that includes all synoptic charts to support the author understanding. This is a fascinating read and the basis for his general circulation work exposed in later works.
Re: Antonio San (#21),
Physical understanding is always imperative in interpreting data. Some 70 yrs before Leroux published his fine monograph, Hamilton and Archibold provided such insight for West Africa along the Guinea Coast in an extensive article in J. Royal Society. It stands up surprisingly well even today.
Re: Luboš Motl (#13),
That is of course interesting, and what Roger Pielke, Sr. has been saying for some time now, that land use and other human induced changes have a significant impact on temperatures on the regional and local level, and that enough regional impacts can produce a near global effect. You are right, since that’s where the people live, that is an area that should generate significant interest, and is something you might actually be able to impact (even the Romans knew how to use green areas and fountains to help control city temps). Focusing on CO2 is focusing on the wrong problem to the detriment of trying to improve things in cities and towns, which as you observe is where actually live, most of us.
I won’t make comment as to why CO2 regulation is preferable, that tends towards politics, but I think the rational is apparent to most reading this blog.
It is interesting indeed that Phil Jones et al last fall published a paper showing that in fact about half the warming trend in China was due to UHI (I posted the ref here somewhere recently) but it has NOT caused him to redo his methods. Too much work to get the right answer?
It has been pointed out that if warming was so detrimental to human health, why do people keep moving to cities with a pronounced UHI and also to the South (in the USA)?
Re: Craig Loehle (#14), When my family and I moved from Minnesota to Texas, the average temperature increased from 45F to 65F, an increase in “local” warming of 20F (~10C). I don’t know what catastrophe is going to befall if we stay us but I’ll tell this, I’d have to get a fresh wife if I moved back.
I was going to ask you, Steve M, on a different thread whether you had changed your position on defining breakpoints in a time series, but it appears that, although you have produced some breakpoint analyses, you continue to have reservations.
I judge that using breakpoints when comparing the differences between two series that should produce the same results, e.g. RSS and UAH in the tropics is more readily related to a cause than doing so when determining breakpoints for a single series.
As you noted, USHCN in Version 2 does breakpoint analyses using differences of the station in question with its neighboring ones. This procedure I presume assumes that a stand alone station will not have temperature changes significantly different than its neighbors. I am not sure what the result of all this is once a few iterations of this process adjust the neighboring stations based on their neighboring stations.
Any way, when I look at those USHCN station adjustments, primarily from the breakpoint discovery process, with largest magnitude changes and compare them with the stations with minimal or no adjustments it does not immediately strike me why or how the adjustments were made. I do not have the where with all to do the extensive iterations that the USHCN method apparently involves, but I would assume that the single stand-alone station series should have breakpoints that could be established internally and the point of doing breakpoints on differences is to insure, or at least give good confidence, that the breakpoints were not caused by a true climate effect and one that effected the stations in the neighborhood.
Re: Kenneth Fritsch (#15),
Kenneth, if a series has breakpoints, then its properties are obviously different than if it doesn’t. I’m not conversant with all the issues in this topic. My problem with USHCNv2 was that it seemed to sort of average out all the series, adjusting the “best” series to bad series. It looks to me like biases can creep in and that the USHCN proponents haven’t necessarily canvassed all the properties of their method. Since their code and details have not been made public, it’s impossible to evaluate their method right now.
I have thought about this proposition often and find it telling on how we measure and adjust temperatures currently. Urban areas have probably from a strict temperature standpoint seen much greater effects from AGW (primarily from causes other than GHGs) than is predicted in the worst scenario for future temperatures as currently defined in most series.
Of course, when one considers the underlying assumptions that go into our current temperature series it becomes clearer why we ignore the urban effects. We assume that climate change is affected by the temperatures as measured, excluding the urban effects. It is that temperature index that will be sensitive to GHGs. It is those detrimental climate effects that are predicted (without mention of any beneficial effects) that will depend on that temperature index (that excludes the urban effects).
Where can the GISS “adjustment” methodology be found? I assume it will include computer code?
Steve: Hansen was forced to make the code public in 2007 after the Y2K schmozzle and is at the GISS website. The adjustments are poorly described in the published articles. They’ve been discussed here in various threads – see Categories in left frame.
Re: Jim (#20), if you are really interested:
http://clearclimatecode.org/
Mike
Re: thefordprefect (#30),
Almost nothing since Sep ’08. Seems to be hibernating.
A don’t really know where to post this comment, but here is a recently published paper in GRL which concludes that the Atlantic Multidecadal Oscillation was the major driver of recent temperature changes in the Arctic region.
Chylek, P., C. K. Folland, G. Lesins, M. K. Dubey, and M. Wang (2009): Arctic air temperature change amplification and the Atlantic Multidecadal Oscillation (published 16 July 2009)
From the abstract:
C. K. Folland is affiliated with the UK Met Office Hadley Centre for Climate Change and he was an author of the IPCC Third Assessment Report. Despite this he just have been co-authored an article which tries to demonstrate that most of the recently observed warming in the Arctic region is due to natural causes. Itt seems to me a ‘great leap backward’ from previous statements.
Re: Adam Soereg (#22),
Adam,
While this is an encouraging sign, I wouldn’t place too much hope for a sea change (that’s a pun by the way – see later links) on it!
Folland along with Parker are the authors of a key IPCC (Folland and Parker 1995) reference which justifies adjustments to the historically recorded SST temperatures which make a very significant contribution to the derived mean global surface temperature (MGST)anomaly. Steve has written at length on the topic of the ‘bucket adjustments’ on his blog. Just do a search for Folland and Parker on this blog and you’ll find lots of links. Here are just two
Changing Adjustments to 19th Century SST
Out of Africa
In the second link at about #34/#35 Steve explains to Michael Hansen that ‘Folland and Parker used their “bucket” adjustments to change SST measurements to better match the land records. I’ve discussed this: their Pearl Harbor hypothesis is ludicrous and contradicted by available meta-data, but continues to be used by the Team’ The link to the thread in which Steve discusses the ‘Pearl Harbor hypothesis’ is here
The ‘take home’ message for me from Steve’s auditing of the SST adjustments is that there are considerable problems arising in the derived MGST anomaly as a result of these ‘bucket adjustments’ since they directly affect the derived pre-WWII ‘warming’ trend followed by the post-WWII cooling trend and subsequent post-1970s warming trend that we should all be very concerned about as this is the ‘smoking gun’ for the ‘unprecedented warming’ that our continued use of burning fossil fuels is having on our climate.
KevinUK
I’ve always felt UHI is much more important than CO2 induced warming (if ever determined to be significant, of which I have doubts). Personally, I live in the Sierra Nevada foothills, in the conifers, and I love living where everything is 95% dirt and 5% concrete, rather than the other way around (just doesn’t seem natural…I wouldn’t have it any other way)! Seems to me we need to be discussing strategies on how to create cities while maintaining vegetation, streams, natural open spaces, old/large trees (really minimize the concrete). Live within our environment, rather than completely remaking it. I always hear how urban sprawl is so bad for the environment, but I think it’s certainly better than condensed concrete cities! At least for most inhabitants of the planet. We need the room for vegetation…
Anyway, hope some day they’ll pay me to maintain my slice of the forest (and my perennial garden)!
Ed
Steve or anyone else knowledgable in the matter:
Do you have any good references or starters to breakpoint analysis?
thanks
Re: Jonathan Duff (#26),
Strucchange reference manual ( http://cran.r-project.org/web/packages/strucchange/strucchange.pdf )
has these references:
Bai J. (1994), Least Squares Estimation of a Shift in Linear Processes, Journal of Time Series
Analysis, 15, 453-472.
Bai J. (1997a), Estimating Multiple Breaks One at a Time, Econometric Theory, 13, 315-352.
Bai J. (1997b), Estimation of a Change Point in Multiple Regression Models, Review of Economics
and Statistics, 79, 551-563.
Bai J., Perron P. (1998), Estimating and Testing Linear Models With Multiple Structural Changes,
Econometrica, 66, 47-78.
Bai J., Perron P. (2003), Computation and Analysis of Multiple Structural Change Models, Journal
of Applied Econometrics, 18, 1-22.
Zeileis A., Kleiber C., Krämer W., Hornik K. (2003), Testing and Dating of Structural Changes in
Practice, Computational Statistics and Data Analysis, 44, 109-123. doi:10.1016/S0167-9473(03)00030-
6.
I guess these are related to regime shifts, as in Wilson 07 ( http://www.climateaudit.org/?p=1285 ) fig 7, I had some reservations with the implicit iid assumption for the data.
Re: Jonathan Duff (#26)
UC provided good resources for statistical analysis. If you are also looking for references on climate shifts for analysis on a physical level, here are a few:
Arctic air temperature change amplification and the Atlantic Multidecadal Oscillation mentioned by Adam Soereg above. I have a great deal of respect for Petr Chylek.
Has the climate recently shifted? is a draft of a paper. Note that it is the author’s opinion “there is no comfort to be gained” from greater than expected (from IPCC’s POV) natural climate variability, a conclusion that seems entirely counter-intuitive to me since the 20th century had two warming regimes lasting 30 years or so and only one cooling regime. The 21st century is expected to have two cooling regimes and only one warming regime. The authors’ write “warming over the 21st century may well be larger than that predicted by the current generation of models, given the propensity of those models to underestimate climate internal variability.” It seems to me that underestimating internal climate variability during the 20th century means the climate is less sensitive to rising atmospheric CO2. Go figure.
The pacemaker of major climate shifts which was the subject of a news story titled UW-Milwaukee Study Could Realign Climate Change Theory – Scientists Claim Earth Is Undergoing Natural Climate Shift blogged about recently by Roger Pielke.
Tropical Pacific decadal variability and global warming published in 2002 and predicting a change to a cooling climate regime in “about four years.”
I am certain more papers can be added to this list and hope others will add them as I am very interested in this subject.
UC and Ron Cram
Thanks very much, most appreciated and will start reading.
re 30. nice.
A lot more happening here.
As someone who grew up on a farm, I have to add one point: It is true that cities (if you include suburbs) are where most of the people live in industrialized countries, but it is even more true that almost all our food is produced outside those cities. (As I recall, more than 50 percent of the US population now lives in suburbs.)
So we want to look at the effects of possible climate changes in both cities and rural areas. And I suspect we will often want to treat central cities differently from their suburbs.
HI for more information go to this link http://www.pegasyssoft.com
2 Trackbacks
[…] posted on Watts Up With That, ICECAP and Climate Audit, there is an excellent new paper on the issue of surface temperature trends. The paper […]
[…] posted on Watts Up With That, ICECAP and Climate Audit, there is an excellent new paper on the issue of surface temperature trends. The paper […]