Contaminated Temperature Data
Ross McKitrick, Financial Post
December 05, 2007
http://www.nationalpost.com/related_links/Story.html?id=145245
QUOTE: "In other words, we have confirmed, on new and stronger grounds,
that the IPCC's global surface-temperature data is exaggerated, with a
large warming bias. Claims about the amount of surface warming since
1980, and its attribution to anthropogenic greenhouse-gas emissions,
should be re*****sed using uncontaminated data. And governments that
rely on the IPCC for advice should begin asking why it was allowed to
suppress earlier evidence of this problem."
Below is the famous graph of "global average surface temperature," or
"global temperature" for short. The data come from thermometers around
the world, but between the thermometer readings and the final, famous,
warming ramp, a lot of statistical modelling aims at removing known
sources of exaggeration in the warming trend. In a new article just
published in the Journal of Geophysical Research -- Atmospheres, a
co-author and I have concluded that the manipulations for the steep
post-1980 period are inadequate, and the above graph is an exaggeration.
Along the way, I have also found that the United Nations agency
promoting the global temperature graph has made false claims about the
quality of its data.
The graph at right comes from data collected in weather stations around
the world. Other graphs come from weather satellites and from networks
of weather balloons that monitor layers of the atmosphere. These other
graphs don't show as much warming as the weather-station data, even
though they measure at heights where there is supposed to be even more
greenhouse-gas-induced warming than at the surface. The discrepancy is
especially clear in the tropics.
The surface-measured data has many well-known problems. Over the
post-war era, equipment has changed, station sites have been moved, and
the time of day at which the data is collected has changed.
Many long-term weather records come from in or near cities, which have
gotten warmer as they grow. Many poor countries have sparse
weather-station records and few resources to ensure data quality. Fewer
than one-third of the weather stations operating in the 1970s remain in
operation.
Scientists readily acknowledged that temperature measurements are
contaminated for the purpose of measuring climate change. But they argue
that adjustments fix the problem. To deal with a false warming generated
by urbanization, they have the "Urbanization Adjustment." To deal with
biases due to changing the time of day when temperatures are observed,
they have the "Time of Observation Bias Adjustment." And so forth.
How do we know these adjustments are correct? In most studies, the
question is simply not asked. A few studies argue that the adjustments
must be adequate since adjacent rural and urban samples give similar
results. But closer inspection shows some of these papers don't actually
give similar results at all, or when they do they define "rural" so
broadly that it includes partly urbanized places. Other studies say the
adjustments must be adequate because trends on windy nights look the
same as trends on calm nights. But the long list of data problems
includes issues just as serious under both windy and calm conditions.
The papers describing the adjustments aim to construct data showing what
the temperature would be in a region if nobody had ever lived there. If
the adjustments are right, the final output should not be correlated
with the extent of industrial development and variations in
socioeconomic conditions. But in a 2004 study with climatologist Patrick
Michaels, we found that the adjustment models were not removing the
contamination patterns as claimed. If the contamination were removed, we
estimated the average measured warming rate over land would decline by
about half. Dutch meteorologists using different data and a different
testing methodology had come to the same conclusions.
In response to criticisms of our paper, I began assembling a more
complete database, covering all available land areas and a more
extensive set of climatological and economic indicators. Meantime, in
2005, I was asked to serve as an external reviewer for the IPCC re****t,
which was released earlier this year. I accepted, in part to address the
data-contamination problem.
Scientists who attribute warming to greenhouse gases argue that their
climate models cannot reproduce the surface trends from natural
variability alone. They then attribute it to greenhouse gases, since
(they assume) all other human influences have been removed from the data
by the adjustment models. If that has not happened, however, they cannot
claim to be able to identify the role of greenhouse gases. Despite the
vast number of studies involved, and the large number of contributors to
the IPCC re****ts, the core message of the IPCC hinges on the assumption
that their main surface climate data set is uncontaminated. And by the
time they began writing the recent Fourth *****sment Re****t, they had
before them a set of papers proving the data are contaminated.
How did they handle this issue? In the first draft of the IPCC re****t,
they simply claimed that, while city data are distorted by urban
warming, this does not affect the global averages. They cited two
familiar studies to sup****t their position and ignored the new
counter-evidence. I submitted lengthy comments criticizing this section.
In the second draft there was still no discussion, so again I put in
lengthy comments. This time the IPCC authors wrote a response. They
conceded the evidence of contamination, but in a stunning admission,
said: "The locations of socioeconomic development happen to have
coincided with maximum warming, not for the reason given by McKitrick
and Mihaels [sic] (2004), but because of the strengthening of the Arctic
Oscillation and the greater sensitivity of land than ocean to greenhouse
forcing, owing to the smaller thermal capacity of land." Note the irony:
Confronted with published evidence of an anthropogenic (but
non-greenhouse) explanation for warming, they dismissed it with an
unproven conjecture of natural causes. Who's the "denialist" now?
Furthermore, the claim is preposterous. The comparison of land and ocean
is irrelevant since we were only talking about land areas. The Arctic
Oscillation is a wind-circulation pattern that affects long-term weather
trends in the Arctic. It certainly plays a role in explaining Arctic
warming over the past few decades. But for IPCC lead authors to invoke
it to explain a worldwide correlation between industrialization and
warming patterns is nonsense.
The final version of the re****t, published in May, 2007, included the
following paragraph (Chapter 3, page 244):
McKitrick and Michaels (2004) and [Dutch meteorologists] de Laat and
Maurellis (2006) attempted to demonstrate that geographical patterns of
warming trends over land are strongly correlated with geographical
patterns of industrial and socioeconomic development, implying that
urbanization and related land surface changes have caused much of the
observed warming. However, the locations of greatest socioeconomic
development are also those that have been most warmed by atmospheric
circulation changes (Sections 3.2.2.7 and 3.6.4), which exhibit
large-scale coherence. Hence, the correlation of warming with industrial
and socioeconomic development ceases to be statistically significant. In
addition, observed warming has been, and transient greenhouse-induced
warming is expected to be, greater over land than over the oceans
(Chapter 10), owing to the smaller thermal capacity of the land.
In the first sentence, the phrase "attempted to demonstrate" should be
replace with "showed." This kind of slanted wording arises when
organizations like the IPCC fail to control the biases of their lead
authors.
The above paragraph acknowledges the correlation between warming trends
and socioeconomic development. But it claims it is a mere coincidence,
due to unspecified "atmospheric circulation changes." The two cited
sections discuss some natural circulation patterns, but do not show they
overlap with the pattern of industrialization -- the topic simply does
not come up. And the de Laat and Maurellis paper refuted that
explanation, anyway.
The IPCC authors also claimed that, in view of the natural circulation
changes, "the correlation of warming with industrial and socioeconomic
development ceases to be statistically significant." Statistical
significance is a precise scientific term, and a claim that results are
insignificant requires specific numerical evidence. The effects in the
underlying papers were all statistically significant. The IPCC's claim
to the contrary is false.
So there are two points to note here. First, the IPCC concedes the
existence of a correlation pattern that shows its main data set is
contaminated, and it has no coherent counterargument. Its claim that it
is due to natural circulation changes contradicts its later (and
prominently advertised) claims that recent warming patterns cannot be
attributed to natural atmospheric circulation changes. Second, the claim
that our evidence is statistically insignificant is, in my opinion, a
plain fabrication. The IPCC offered no sup****ting evidence. Confronted
with two lines of independent evidence that the data set on which it
bases its fundamental conclusions is contaminated, it conceded the
point, but then dismissed it on the basis of non-existent
counter-evidence.
This is no mere tiff among duelling experts. The IPCC has a monopoly on
scientific advising to governments concerning climate change.
Governments who never think to conduct due diligence on IPCC re****ts
send delegates to plenary meetings at which they formally "accept" the
conclusions of IPCC re****ts. Thereafter they are unable -- legally and
politically -- to dissent from its conclusions. In the years ahead,
people around the world, including here in Canada, could bear costs of
climate policies running to hundreds of billions of dollars, based on
these conclusions. And the conclusions are based on data that the IPCC
lead authors concede exhibits a contamination pattern that undermines
their interpretation of it, a problem they concealed with untrue claims.
Our new paper presents a new, larger data set with a more complete set
of socioeconomic indicators. We showed that the spatial pattern of
warming trends is so tightly correlated with indicators of economic
activity that the probability they are unrelated is less than one in 14
trillion. We applied a string of statistical tests to show that the
correlation is not a fluke or the result of biased or inconsistent
statistical modelling. We showed that the contamination patterns are
largest in regions experiencing real economic growth. And we showed that
the contamination patterns account for about half the surface warming
measured over land since 1980.
In other words, we have confirmed, on new and stronger grounds, that the
IPCC's global surface-temperature data is exaggerated, with a large
warming bias. Claims about the amount of surface warming since 1980, and
its attribution to anthropogenic greenhouse-gas emissions, should be
re*****sed using uncontaminated data. And governments that rely on the
IPCC for advice should begin asking why it was allowed to suppress
earlier evidence of this problem.
-- Ross McKitrick is associate professor and director of graduate
studies, Department of Economics, University of Guelph.
--
Warmest Regards
Bonzo
"The question scientists should now be asking is not how much it will
warm over the next 50 to 100 years, but why has it warmed so little
during the major carbon dioxide buildup?" Patrick J. Michaels,
Environmental Scientist , University of Virginia
|
