Four Things I Think I Know About Climate

John R. ChristyUniversity of Alabama in Huntsville

Alabama State Climatologist

We should always begin our scientific assessments with the following statement:

“At our present level of ignorance, we think we know …”

Paraphrase of Richard Mallory Hoover High School

Fresno CA Physics Teacher 1968

Testing Hypothesis (assertions) about ClimateUAHuntsville builds datasets from scratch

1. Popular surface temperature datasets are poor metrics for checking on the greenhouse effect - and they are poorly measured as well

2. Warming is occurring but at a rate and in a manner that is inconsistent with model projections of enhanced greenhouse warming

3. Sensitivity research suggests the climate system is less sensitive to CO2 increases as depicted in models due to unaccounted-for negative cloud feedbacks

4. Impacts on global emissions of current legislative actions are minuscule and will have no discernable impact on whatever the climate is going to do

Testing Hypotheses on Global Warming

1. Testing Assertions based on Popular Surface Temperature

Datasets

Popular surface datasets tend to:

(a) overstate the warming, and (b) serve as a poor greenhouse

metric

"Global" Surface TemperatureTmean is Average of Day and Night

HadCRUT3 (2010 Jan-Jun only)

-0.8-0.6-0.4-0.20.00.20.40.6

1850 1870 1890 1910 1930 1950 1970 1990 2010

CO2 up 38% at current rate of 0.6% per year

Day vs. Night Surface Temp

Nighttime - disconnectedshallow layer/inversion. Temperature affected by land-use changes, buildings, farming, etc.

Daytime - deep layer mixing, connected with levels impacted by enhanced greenhouse effect

Warm air above inversion

Cold air near surface

Greenhouse signal

Night Surface Temp

Nighttime - disconnectedshallow layer/inversion. But this situation can be sensitive to small changes such as roughness or heat sources.

Buildings, heat releasing surfaces, aerosols, greenhouse gases, etc. can disrupt the delicate inversion, mixing warm air downward - affecting TMin.

Warm air above inversion

Cold air near surface

Warm air

MODIS21 Jul 2002

Jacques DescloitresMODIS Land Rapid Response Team NASA GSFC

CA Valley and Sierra (Jun-Nov) 1910-2003

8

10

12

14

16

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

°C

Valley TMinSierra TMin

Nighttime temperatures rising but not because of greenhouse gas warming, but nighttime readings are included in popular datasets

Daytime temperatures tell more accurate story

Christy 2002, Christy et al. 2006, 2007, 2009, Pielke et al 2008, Walters et al. 2007

CA Valley and Sierra Annual Avg TMax 1910-2003 .

18

20

22

24

26

28

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

°C

Valley TMaxSierra TMax

CA Valley and Sierra (Jun-Nov) 1910-2003

8

10

12

14

16

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

°C

Valley TMinSierra TMin

Nighttime temperatures rising but not because of greenhouse gas warming, but nighttime readings are included in popular datasets

Daytime temperatures tell more accurate story

Christy 2002, Christy et al. 2006, 2007, 2009, Pielke et al 2008, Walters et al. 2007

1. (c) Some surface data sources are

simply poor

East Africa TMax (Christy et al. 2009).

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

1900 1920 1940 1960 1980 2000

Obs: (HadCRUT3) +0.14 °C/decade

Kilimanjaro

East Africa TMax (Christy et al. 2009).

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

1900 1920 1940 1960 1980 2000

Obs: (UAH) +0.02 °C/decade

Obs: (HadCRUT3) +0.14 °C/decade

Kilimanjaro

Testing Hypothesis (assertions) about ClimateUAHuntsville builds datasets from scratch

1.1. Popular surface temperature datasets tend to be poor Popular surface temperature datasets tend to be poor metrics for checking on the greenhouse effect - and metrics for checking on the greenhouse effect - and they are often poorly measured as wellthey are often poorly measured as well

2. Warming is occurring but at a rate and in a manner that is inconsistent with model projections of enhanced greenhouse warming

3. Sensitivity research indicates the climate system is less sensitive to CO2 increases as depicted in models due to unaccounted-for negative cloud feedbacks

4. Impacts on global emissions of current legislative actions are minuscule and will have no discernable impact on whatever the climate is going to do

Testing Hypotheses on Global Warming

2. (a) Testing Assertions based on Climate Models for

global trend magnitude

Climate models tend to overstate or misrepresent

the warming

History Lesson 1988

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

1960 1970 1980 1990 2000 2010 2020

GISS-A(88)GISS-B(88)GISS-C(88)

Predictions

History Lesson 1988

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

1960 1970 1980 1990 2000 2010 2020

GISS-A(88)GISS-B(88)GISS-C(88)UAH-LT (SfcAdj)RSS-LT (SfcAdj)

2010 to July Only .Observations

Predictions

G. Schmidt NASA/GISS

12

13

14

15

16

17

1979 1983 1987 1991 1995 1999 2003 2007

BCCRCCMaT47_avgCCMaT63GFDL_CM2.0GFDL_CM2.1HadGEM1CSIROMk3.0CSIROMk3.5GISS EH (3)GISS ER (4)ECHAM5MRI232a (5)CCSM3.0 (7)PCM1 (2)MIROC3.2HiRes (1)CNRM CM3 (1)INGV ECHAM4 (1)INMCM3.0 (1)IPSL CM4 (1)MIROC3.2 MidRes (3)ECHO G (3)ESSENCEUKMO HadCM3 (1)QUMP HadCM3 (17)Median

Individual Model Surface Trends1979-2010

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

1979 1985 1990 1995 2000

HadCRUT3vUAHLTsfcRSSltsfcBCCRCCMaT47_avgCCMaT63GFDL_CM2.0GFDL_CM2.1HadGEM1CSIROMk3.0CSIROMk3.5GISS EH (3)GISS ER (4)ECHAM5MRI232a (5)CCSM3.0 (7)PCM1 (2)MIROC3.2HiRes (1)CNRM CM3 (1)INGV ECHAM4 (1)INMCM3.0 (1)IPSL CM4 (1)MIROC3.2 MidRes (3)ECHO G (3)ESSENCEUKMO HadCM3 (1)QUMP HadCM3 (17)Medianmean w/oMRI,MIROC HiRes

Trends ending in 2008 with various start yearsIPCC AR4 Model Runs (22 models) vs. Obs.

Start Year

0.00

0.05

0.10

0.15

0.20

0.25

0.30

1979 1984 1989 1994 1999

Mean of Model trends(w/o Max,Min)Median of ModelTrendsHadCRUT3v

UAHLTsfc

RSSLTsfc

Trends ending in 2008 with various start yearsIPCC AR4 Model Runs (22 models) vs. Obs.

Start Year

0.00

0.05

0.10

0.15

0.20

0.25

0.30

1979 1985 1990 1995 2000

Mean of Model trends(w/o Max,Min)Median of ModelTrendsHadCRUT3v

UAHLTsfc

RSSLTsfc

Trends ending in 2010 (Jun) with various start yearsIPCC AR4 Model Runs (22 models) vs. Obs.

Start Year

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1978 1982 1986 1990 1994 1998 2002 2006 2010

Global Bulk Atmospheric TemperaturesUAH Satellite Data

Warming rate 50% of model projectionsChristy et al. 2007, 2009

Testing Hypotheses on Global Warming

Testing Assertions based on Climate Models - Sierra

Nevada loses 80% of snow by 2100

Observations contradict this

Snyder et al. 2002

Sierras warm faster than Valley in model simulations

Southern Sierra Nevada Snowfall1915-16 to 2009-10

0

200

400

600

800

1000

1200

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010

CM

So. Sierra Snowfall

Christy and Hnilo 2010 .

Trend +1.1 cm/decade

Testing Hypotheses on Global Warming

2. (b) Testing Assertions based on Climate Models concerning

Tropical Upper Air Temperature trends

Climate models tend to misrepresent the observed

relationship

A Climate Model Simulation is a Hypothesis

How does one define a falsifiable test for a model hypothesis?

Douglass, Christy, Pearson and Singer 2007

Select a prominent metric dependent on the main perturbation in forcing - a large signal - test against observations

One such signal is the vertical structure of the tropical tropospheric temperature trend - i.e. how surface and upper air trends compare

Vertical Temperature Change due to Greenhouse Forcing in

Models

Model Simulations of Tropical Troposphere Warming:About 2X surfaceLee et al. 2007

100

200

300

400

500

600

700

800

900

1000

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6

Models: Mean and Standard Error

Model -2SEModel +2SEModel Best Guess

Best Estimate of Models - given surface trend close to observed

°C/decade

Observations: Four "corrected" Datasets100

200

300

400

500

600

700

800

900

1000

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6

HadATIGRARATPACRICH

Upper air trends of four observed datasets are significantly cooler in this apples to apples comparison

°C/decade

100

200

300

400

500

600

700

800

900

1000

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6

Models vs. Obs

Model -2SEModel +2SEModel Best GuessHadATIGRARATPACRICH

Upper air trends of four observed datasets are significantly cooler in this apples to apples comparison (Douglass et al. 2007).

°C/decade

Model Sfc-to-Upper Air Trends Normalized to Observed Sfc100

200

300

400

500

600

700

800

900

1000

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6

Model Best GuessHadATIGRARATPACRICH

Cool ExtWarm Ext

Putting the hot and cold extremes (thick red) of model trends tied to actual surface trend, models are still too hot which in this case all models have been tied to the actual observed surface trend.

°C/decade

A different test asks whether the models and observational upper air trends agree if NO restriction is placed on the model surface trends (i.e. apples to oranges.)

Extremely weak hypothesis to test and does NOT address the sfc-to-upper air relationship (a key signature of greenhouse warming in models.)

Model Sfc-to-Upper Air Trends Normalized to Observed Sfc100

200

300

400

500

600

700

800

900

1000

-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6

Model Best GuessHadATIGRARATPACRICHS08-2SES08+2SE

Putting the hot and cold extremes (thick red) of model trends tied to actual surface trend, models are still too hot which in this case all models have been tied to the actual observed surface trend.

°C/decade

0.00.20.40.60.81.01.21.41.61.82.0

Model ARObs AR

Christy et al. 2007, 2010; Christy and Norris 2006, 2009; Randall and Herman 2008; Klotzbach et al. 2009, 2010; McKitrick et al. 2010

Ratio of Lower Tropospheric Trend to Surface Trend(Amplification Ratio)

ModelMedian

Land Surface Temperatures overstate warming Model Expectation (Surface less than upper air)

Observations (Surface much more than upper air)

McKitrick et al. 2010Over the interval 1979-2009, model-projected temperature trends are two to four times larger than observed trends in both the lower and mid-troposphere and the differences are statistically significant at the 99% level. [Note: recalculated Santer et al. 2008 method, and even with surface trend variation found Santer et al.’s result is not verified.]

Klotzbach et al. 2010

[Our] result is inconsistent with model projections which show that significant amplification of the modeled surface trends occurs in the modeled tropospheric trends.

Christy et al. 2010

Table 2 displays the new per decade linear trend calculations [of difference between global surface and troposphere using model amplification factor] … over land and ocean. All trends are significant at the 95% level.

Testing Hypothesis (assertions) about ClimateUAHuntsville builds datasets from scratch

1. Popular surface temperature datasets tend to be poor metrics for checking on the greenhouse effect - and they are often poorly measured as well

2.2. Warming is occurring but at a rate and in a manner that Warming is occurring but at a rate and in a manner that is inconsistent with model projections of enhanced is inconsistent with model projections of enhanced greenhouse warminggreenhouse warming

1. Sensitivity research suggests the climate system is less sensitive to CO2 increases (as depicted in models) due to unaccounted-for negative cloud feedbacks

4. Impacts on global emissions of current legislative actions are minuscule and will have no discernable impact on whatever the climate is going to do

Testing Hypotheses on Global Warming

3. Testing Assertions on the sensitivity of the climate to increases in greenhouse gas

forcing

Climate models tend to overstate the sensitivity due to

missing negative cloud feedbacks

Response of Clouds and Water Vapor (shortwave and longwave) to Increasing CO2

0

20

40

60

80

100

Relative Temperature Effect

Water Vapor and CloudsCO2Other

Negative Feedback?(mitigates CO2 impact)

Positive Feedback?(enhances CO2 impact - models)

A Climate Model Simulation is a Hypothesis

How does one define a falsifiable test for a model hypothesis?

Select a prominent metric dependent on the main perturbation in forcing - a large signal - test against observations

One such test is to calculate the climate feedback parameter during monthly and annual scale variations

Spencer and Braswell

Global Warming in Models isGreatly Magnified by Positive Feedbacks

Warming from CO2 only

Warming in modelsamplified by clouds

& vapor(~3 deg. C by 2100)

Satellite data suggests cloudsreduce warming(~0.5°C by 2100)

Spencer and Braswell .

Testing Hypothesis (assertions) about ClimateUAHuntsville builds datasets from scratch

1. Popular surface temperature datasets tend to be poor metrics for checking on the greenhouse effect - and they are often poorly measured as well

2. Warming is occurring but at a rate and in a manner that is inconsistent with model projections of enhanced greenhouse warming

3.3. Sensitivity research suggests the climate is less Sensitivity research suggests the climate is less sensitive to CO2 increases than depicted in models due sensitive to CO2 increases than depicted in models due to unaccounted-for negative cloud feedbacksto unaccounted-for negative cloud feedbacks

4. Impacts on global emissions of current legislative actions are minuscule and will have no discernable impact on whatever the climate is going to do

Testing Hypotheses on Global Warming

4. Testing Assertions the impact of regulations on

climate

Regulations will have a minuscule impact on whatever

the climate is going to do

What did California do?• Force a limit on emissions of Light

Duty Vehicles• California AB 1493 seeks to reduce

tail-pipe emissions of CO2 by 26% by 2016

• 11 NE States adopted AB 1493• Trial in Federal Court (Burlington

VT) to address the engineering, legal and climate issues of AB 1493, April-May 2007

What did California do?• Force a limit on emissions of Light

Duty Vehicles• California AB 1493 seeks to reduce

tail-pipe emissions of CO2 by 26% by 2016

• 11 NE States adopted AB 1493• Trial in Federal Court (Burlington

VT) to address the engineering, legal and climate issues of AB 1493, April-May 2007

What did California do?• Force a limit on emissions of Light

Duty Vehicles• California AB 1493 seeks to reduce

tail-pipe emissions of CO2 by 26% by 2016

• 11 NE States adopted AB 1493• Trial in Federal Court (Burlington

VT) to address the engineering, legal and climate issues of AB 1493, April-May 2007

What did California do?• Force a limit on emissions of Light

Duty Vehicles• California AB 1493 seeks to reduce

tail-pipe emissions of CO2 by 26% by 2016

• 11 NE States adopted AB 1493• Trial in Federal Court (Burlington

VT) to address the engineering, legal and climate issues of AB 1493, April-May 2007

IPCC “Best Estimate”

0.0

0.5

1.0

1.5

2.0

2.5

3.0

A1B

California AB 149326% CO2 reduction LDV 2016

0.0

0.5

1.0

1.5

2.0

2.5

3.0

A1BCANo. EastUSA

Net Impact if all US 0.01°C 2100

The temperature impact on global temperatures if the entire world adopted AB 1493 is an undetectable 0.03°C.

Latest sensitivity results suggest the impact is even smaller.

Pg 46

“Plaintiffs’ expert Dr. Christy estimated that implementing the regulations across the entire United States would reduce global temperature by about 1/100th (.01) of a degree by 2100. Hansen did not contradict that testimony.”

Judge William Sessions III Ruling 12 Sept 2007

AB 1493 is legal

Questions• What could make a “dent” in

forecasted global temperatures?

• What would be the impact of building 1000 nuclear power plants and putting them on-line by 2020?– (average 1.4 gigawatt output

each)

Questions• What could make a “dent” in

forecasted global temperatures?

• What would be the impact of building 1000 nuclear power plants and putting them on-line by 2020?– (average 1.4 gigawatt output

each)

IPCC “Best Estimate”

0.0

0.5

1.0

1.5

2.0

2.5

3.0

A1B

Net Effect of 10% CO2 emission reduction to A1B Scenario

(~1000 Nuclear Plants by 2020)

0.0

0.5

1.0

1.5

2.0

2.5

3.0A1B Emissions10% Reduction A1B

Net Impact 0.07°C 2050

Testing Hypothesis (assertions) about ClimateUAHuntsville builds datasets from scratch

1. Popular surface temperature datasets tend to be poor metrics for checking on the greenhouse effect - and they are often poorly measured as well

2. Warming is occurring but at a rate and in a manner that is inconsistent with model projections of enhanced greenhouse warming

3. Sensitivity research suggests the climate is less sensitive to CO2 increases than depicted in models due to unaccounted-for negative cloud feedbacks

4. Impacts on global emissions of current legislative Impacts on global emissions of current legislative actions are minuscule and will have no discernable actions are minuscule and will have no discernable impact on whatever the climate is going to doimpact on whatever the climate is going to do

Kenya, East Africa

Energy Transmission

Energy System

Energy Use

Energy Source

Consensus is not Science Michael Crichton

Consensus is not Science

William Thomson (Lord Kelvin)

All Science is numbers

Michael Crichton