Transcript Document

The influence of anthropogenic
surface processes and
inhomogeneities on gridded
global climate data
Ross McKitrick
Department of Economics
University of Guelph
Guelph ON Canada
Presentation to the
American Chemical Society
Denver CO via Webinar
August 28 2011
Surface Climate Data

The “global temperature”
ross.mckitrick.weebly.com
2
Summary


Climate data is the output of a model

Raw data: daily T-Min and T-Max readings from inhabited places

This isn’t what the climate analyst is interested in: it must be
converted into “climate data” using a statistical adjustment model.
How do we know the adjustment model “works”?

Many papers merely describe the adjustment steps in
enthusiastic detail

I have focused on devising statistical tests of the results
ross.mckitrick.weebly.com
3
Conclusions

Based on analysis of multiple data sets, and after addressing a
long list of statistical rebuttals, I find the evidence convincing that:

The adjustment models are inadequate

The resulting climate record over land is contaminated with patterns
of socioeconomic development

This adds a net warming bias to the global trend and may lead to
misattribution of spatial patterns to greenhouse gases

A valid empirical model of the spatial pattern of observed warming
must include anthropogenic surface processes
ross.mckitrick.weebly.com
4
Papers

McKitrick, Ross and Patrick J. Michaels (2004). “A Test of Corrections for
Extraneous Signals in Gridded Surface Temperature Data” Climate Research 26
pp. 159-173.

McKitrick, Ross R. and Patrick J. Michaels. (2007) “Quantifying the influence of
anthropogenic surface processes and inhomogeneities on gridded surface
climate data.” Journal of Geophysical Research-Atmospheres 112, D24S09,
doi:10.1029/2007JD008465.

McKitrick, Ross R. and Nicolas Nierenberg (2010) “Socioeconomic Patterns in
Climate Data.” Journal of Economic and Social Measurement, 35(3,4) pp. 149175. DOI 10.3233/JEM-2010-0336.

McKitrick, Ross R. (2010) “Atmospheric Oscillations do not Explain the
Temperature-Industrialization Correlation.” Statistics, Politics and Policy, Vol 1
No. 1, July 2010.

rossmckitrick.weebly.com
ross.mckitrick.weebly.com
5
Core Methodology

There is a spatial pattern of warming and
cooling trends since 1980

Climate models predict the pattern as a response to
GHG’s, solar changes, etc.

The predicted pattern is uncorrelated with
spatial pattern of socioeconomic development

But raw weather data is known to be correlated with socioeconomic
development

The adjustment models are supposed to remove these effects.

Therefore: If the adjustments are adequate, the climate data should be
uncorrelated with socioeconomic patterns
ross.mckitrick.weebly.com
6
Core Methodology

There is a spatial pattern of warming and
cooling trends since 1980

Climate models predict the pattern as a response to
GHG’s, solar changes, etc.

The predicted pattern is uncorrelated with
spatial pattern of socioeconomic development

But raw weather data is known to be correlated with socioeconomic
development

The adjustment models are supposed to remove these effects.

Therefore: If the adjustments are adequate, the climate data should be
uncorrelated with socioeconomic patterns
ross.mckitrick.weebly.com
7
Core Methodology

There is a spatial pattern of warming and
cooling trends since 1980

Climate models predict the pattern as a response to
GHG’s, solar changes, etc.

The predicted pattern is uncorrelated with
spatial pattern of socioeconomic development

But raw weather data is known to be correlated with socioeconomic
development

The adjustment models are supposed to remove these effects.

Therefore: If the adjustments are adequate, the climate data should be
uncorrelated with socioeconomic patterns
ross.mckitrick.weebly.com
8
Core Methodology

There is a spatial pattern of warming and
cooling trends since 1980

Climate models predict the pattern as a response to
GHG’s, solar changes, etc.

The predicted pattern is uncorrelated with
spatial pattern of socioeconomic development

But raw weather data is known to be correlated with socioeconomic
development

The adjustment models are supposed to remove these effects.

Therefore: If the adjustments are adequate, the climate data should be
uncorrelated with socioeconomic patterns
ross.mckitrick.weebly.com
9
Core Methodology

Hypothesis:


{spatial pattern of trends in surface climate data}
is uncorrelated with
{spatial pattern of socioeconomic development}
In a series of papers I have shown that this
hypothesis is strongly rejected
ross.mckitrick.weebly.com
10
Sources of climate data

CRU, NOAA, NASA all produce “global climate data”
products

All rely on same underlying archive


Global Historical Climatology Network (run by NOAA)
The 3 data products are very similar since they all
use the same input data and similar, though not
identical, averaging methods
ross.mckitrick.weebly.com
11
Sources of observational error:







Changing sample size
Changing sample locations
Build up of surrounding landscape
Equipment changes
Poor quality control
Local air pollution
Waste heat from buildings and traffic, etc.
ross.mckitrick.weebly.com
12
GHCN sample 1885

Locations of weather stations
ross.mckitrick.weebly.com
13
GHCN sample 1925

Locations of weather stations
ross.mckitrick.weebly.com
14
GHCN sample 1945
ross.mckitrick.weebly.com
15
GHCN sample 1965

Locations of weather stations
ross.mckitrick.weebly.com
16
GHCN sample 1985

Locations of weather stations
ross.mckitrick.weebly.com
17
GHCN sample 2005

Locations of weather stations
ross.mckitrick.weebly.com
18
GHCN
sample
size over
time
ross.mckitrick.weebly.com
19
GHCN fraction of sample from
urban airports
ross.mckitrick.weebly.com
20
“Climate” data: the record as if the
land surface was never modified
and equipment never varied
Temp data from cities
“True” record
adjustment algorithm
+
=
ross.mckitrick.weebly.com
21
Structure of data set



Cross-sectional
Observational unit is a 5ox5o grid cell
Dependent variable is 1979-2002 trend
ross.mckitrick.weebly.com
22
Measurement Model
q i  Ti  f (S i )  g ( I i )
Where
qi = observed climatic trend oC/decade
Ti = “true” trend
f (Si) = surface processes like urbanization and agriculture
g (Ii) = data inhomogeneities
ross.mckitrick.weebly.com
23
For gridcell i

Ti (ideal temperature trend) represented by
Ti   0  1TROPi   2 PRESSi  3 DRYi   4 DSLPi  5WATERi
  6 ABSLATi

TROPi = trend in troposphere over same
gridcell as measured by satellites
ross.mckitrick.weebly.com
24
For gridcell i

Surface processes f (Si) measured by
pi = % growth in population density
mi = % growth in real average income
yi = % growth in real national GDP
ci = % growth in national coal consumption
ross.mckitrick.weebly.com
25
For gridcell i

Inhomogeneities g (Ii) measured by
gi = GDP density (GDP per square km)
ei = availability of educated workers (sum of literacy +
postsecondary education)
xi = rate of missing observations (# missing months in cell)
ross.mckitrick.weebly.com
26
Regression equation
qi   0  1TROPi   2 PRESSi   3 DRYi   4 DSLPi   5WATERi   6 ABSLATi
  7 pi   8 mi   9 yi  10 ci  11ei  12 g i  13 xi  ui
Surface proc.

Inhom.
GLS with clustering-robust std error matrix
ross.mckitrick.weebly.com
27
First pair of studies:

McKitrick and Michaels (2004)
 Tested 218 raw series and corresponding CRU gridded data
 Both exhibited significant imprint of socioeconomic data with v.
similar coefficients
 ‘Adjustment’ hypothesis rejected at high confidence level

McKitrick and Michaels (2007)
 Complete sample of (available) surface grid cells
 ‘Independence’ hypothesis again rejected at high confidence
level

Both studies: nonclimatic signals likely add up to a net warming
bias in global average
ross.mckitrick.weebly.com
28
Variable
trop
slp
2007 Results
dry
dslp
Probability that effects are zero:
water
abslat
g

Joint P = 0.0000
(7x10-14)
e
x
p
Effect on Trend of Doubling Level
0.80
m
0.60
deg C
0.40
y
0.383
0.407
c
0.20
_cons
0.006
0.00
Population Density
Real Average Income
Real National GDP
National Coal Use
-0.20
-0.303
-0.40
-0.60
ross.mckitrick.weebly.com
N
R2
ll
P(I)
P(S)
P(all)
SURF
0.8625
(8.61)
0.0043
(0.99)
0.4901
(0.09)
-0.0004
(-0.07)
-0.0284
(-1.34)
0.0006
(0.49)
0.0434
(3.38)
-0.0027
(-5.11)
0.0041
(1.66)
0.3831
(2.70)
0.4075
(2.37)
-0.3032
(-2.19)
0.0059
(3.25)
-4.0368
(-0.92)
440
0.53
139.01
0.0000
0.0005
0.0000
29
Specification tests





Bootstrap resampling
Remove outliers, re-estimate
RESET test
Cross-validation tests
Hausman endogeneity test (P = 0.9962)
ross.mckitrick.weebly.com
30
Generating ‘clean’ trends



Set GDP density and education to US levels
Set all other surface and inhomogeneity effects to 0
Use model coeff’s to generate adjusted predicted values
Observed average surface trend:
MSU average:
Adjusted average surface trend:
0.30 oC/decade
0.23
0.17
ross.mckitrick.weebly.com
31
IPCC Report

How did the IPCC deal with this?

IPCC AR4 page 244:


McKitrick and Michaels (2004) and 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 urbanisation 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.
No supporting citation given
ross.mckitrick.weebly.com
32
IPCC Report

I obtained correlation fields between gridded temperatures and
AO, ENSO and PDO
ross.mckitrick.weebly.com
33
IPCC Report
I augmented data sets for M&M 2004 and M&M 2007 with circulation
terms

2004 Model:



Circulation index effects are insignificant
Including them anyway does not remove the significance of the conclusions
2007 Model


Circulation index effects are jointly barely significant
Including them increases size and significance of socioecononomic terms
Conclusion: IPCC claim is false.
(McKitrick 2010, Statistics Politics and Policy July 2010)
ross.mckitrick.weebly.com
34
IPCC Report
I augmented data sets for M&M 2004 and M&M 2007 with circulation
terms

2004 Model:



Circulation index effects are insignificant
Including them anyway does not remove the significance of the conclusions
2007 Model


Circulation index effects are jointly barely significant
Including them increases size and significance of socioecononomic terms
Conclusion: IPCC claim is false.
(McKitrick 2010, Statistics Politics and Policy July 2010)
ross.mckitrick.weebly.com
35
IPCC Report
I augmented data sets for M&M 2004 and M&M 2007 with circulation
terms

2004 Model:



Circulation index effects are insignificant
Including them anyway does not remove the significance of the conclusions
2007 Model


Circulation index effects are jointly barely significant
Including them increases size and significance of socioecononomic terms
Conclusion: IPCC claim is false.
(McKitrick 2010, Statistics Politics and Policy July 2010)
ross.mckitrick.weebly.com
36
IPCC Report
I augmented data sets for M&M 2004 and M&M 2007 with circulation
terms

2004 Model:



Circulation index effects are insignificant
Including them anyway does not remove the significance of the conclusions
2007 Model


Circulation index effects are jointly barely significant
Including them increases size and significance of socioecononomic terms
Conclusion: IPCC claim is false.
(McKitrick 2010, Statistics Politics and Policy July 2010)
ross.mckitrick.weebly.com
37
Schmidt (2009) “Spurious correlation between recent
warming and indices of local economic activity.”
International Journal of Climatology 10.1002/joc.1831

3 arguments against our findings

surface temperature field exhibits spatial autocorrelation (SAC) so
results are insignificant

Use of RSS satellite series rather than UAH series removes
significance of results

Data generated by climate model yields apparent correlations with
socioeconomic data, yet is uncontaminated by construction, so
effects must be a fluke
ross.mckitrick.weebly.com
38
Schmidt (2009) “Spurious correlation between recent
warming and indices of local economic activity.”
International Journal of Climatology 10.1002/joc.1831

3 arguments against our findings

surface temperature field exhibits spatial autocorrelation (SAC) so
results are insignificant

Use of RSS satellite series rather than UAH series removes
significance of results

Data generated by climate model yields apparent correlations with
socioeconomic data, yet is uncontaminated by construction, so
effects must be a fluke
ross.mckitrick.weebly.com
39
Schmidt (2009) “Spurious correlation between recent
warming and indices of local economic activity.”
International Journal of Climatology 10.1002/joc.1831

3 arguments against our findings

surface temperature field exhibits spatial autocorrelation (SAC) so
results are insignificant

Use of RSS satellite series rather than UAH series removes
significance of results

Data generated by climate model looks correlated with
socioeconomic data, yet is uncontaminated by construction, so
effects must be a fluke
ross.mckitrick.weebly.com
40
McKitrick & Nierenberg
“Socioeconomic patterns in climate data”
J Econ Soc Measurement 2010

Responses

Schmidt did not actually test SAC. We do, and show that while
depvar is AC’d, regression residuals are not, as long as socioecon
variables are included in model.

Use of RSS data diminishes individual significance but effect due to
a small number of outliers. Once these removed, RSS yields
strongest results of all data sets

Model-based data cannot replicate observed patterns; predicts
opposite signs
ross.mckitrick.weebly.com
41
McKitrick & Nierenberg
“Socioeconomic patterns in climate data”
J Econ Soc Measurement 2010

Responses

Schmidt did not actually test SAC. We do, and show that while
depvar is AC’d, regression residuals are not, as long as socioecon
variables are included in model.

Use of RSS data diminishes individual significance but effect due to
a small number of outliers. Once these removed, RSS yields
strongest results of all data sets

Model-based data cannot replicate observed patterns; predicts
opposite signs
ross.mckitrick.weebly.com
42
McKitrick & Nierenberg
“Socioeconomic patterns in climate data”
J Econ Soc Measurement 2010

Responses

Schmidt did not actually test SAC. We do, and show that while
depvar is AC’d, regression residuals are not, as long as socioecon
variables are included in model.

Use of RSS data diminishes individual significance but effect due to
a small number of outliers. Once these removed, RSS yields
strongest results of all data sets

Model-based data cannot replicate observed patterns; predicts
opposite signs
ross.mckitrick.weebly.com
43
Data variations

Surface



Observed: CRU, CRU2v, CRU3v
Modeled: GISS-E; GCM average
Troposphere


Observed: UAH, RSS
Modeled: GISS-E; GCM average
ross.mckitrick.weebly.com
44
Spatial Autocorrelation Tests
OBSERVED: SAC DISAPPEARS
MODELS: SAC REMAINS
ross.mckitrick.weebly.com
45
Estimation with SAC model
ross.mckitrick.weebly.com
46
Estimation with SAC model
OBSERVATIONS:
MODELS:
SIGNIFICANT
INSIGNIFICANT
ross.mckitrick.weebly.com
47
GCM Counterfactual

Schmidt 2009, p.2:
There is a relatively easy way to assess whether there is
any true significance to these correlations. We can take
fully consistent model simulations for the same period and
calculate the distribution of the analogous correlations.
Those simulations contain no unaccounted-for processes
(by definition!) but plenty of internal variability, locally
important forcings and spatial correlation. If the distribution
encompasses the observed correlations, then the null
hypothesis (that there is no contamination) cannot be
rejected.
ross.mckitrick.weebly.com
48
Results
ross.mckitrick.weebly.com
49
1 = climate model
reproduces observed
effect,
Results
0 = failure to do so
ross.mckitrick.weebly.com
50
Filtering results on surface
data



Set GDP density and education to US levels
Set all other surface and inhomogeneity effects to 0
Use model coeff’s to generate adjusted predicted values
ross.mckitrick.weebly.com
51
Filtering results on surface
data



Set GDP density and education to US levels
Set all other surface and inhomogeneity effects to 0
Use model coeff’s to generate adjusted predicted values
ross.mckitrick.weebly.com
52
Filtering results on surface
data



Set GDP density and education to US levels
Set all other surface and inhomogeneity effects to 0
Use model coeff’s to generate adjusted predicted values
This method should not reduce mean trend in GISS data
ross.mckitrick.weebly.com
53
Conclusions

In general, I reject the null hypothesis that
adjustment models yield “climate” data



socioeconomic patterns are highly significant across wide
variety of specifications and data combinations
socioeconomic data are necessary for well-specified error
term
This suggests a causal interpretation of the regression
results
ross.mckitrick.weebly.com
54
Responses to critiques:

IPCC claim that the results were statistically
insignificant & due to natural circulation patterns was
a fabrication


The claim was both unsubstantiated and untrue
Various critiques have not held up



SAC is not a source of bias
Results hold up across numerous data sets
Climate models cannot reproduce results
ross.mckitrick.weebly.com
55
Thank you

[email protected]

rossmckitrick.weebly.com
ross.mckitrick.weebly.com
56