Climate Sensitivity: Uncertainties & Learning Workshop on GHG - - PowerPoint PPT Presentation

Climate Sensitivity: Uncertainties & Learning

Workshop on GHG Stabilization Scenarios Tsukuba, Japan, 23 January 2004 Michael Schlesinger and Natasha Andronova

Climate Research Group Department of Atmospheric Sciences University of Illinois at Urbana-Champaign

Introduction

• Climate sensitivity, ∆T2x: The change in

global-average near-surface temperature resulting from a doubling of the preindustrial carbon dioxide concentration.

• If ∆T2x is small, then the problem of

human-induced climate change may not be

• acute. If ∆T2x is large, then human-induced

climate change may be one of the most severe problems of the 21st century.

Outline

• Primer on Climate Sensitivity, ∆T2x
• Estimates of Climate Sensitivity, ∆T2x
• Uncertainty in ∆T2x due to uncertainty in

the radiative forcing

• Causes of temperature changes from

1856 to present

• Learning ∆T2x over time

dH t

( )

dt = − ΔT t

( )

λ + F t

( )

• F(t): Radiative Forcing

– The change in the net downward radiative flux at some level in the atmosphere, usually the tropopause, caused by some “external” factor, such as changed solar insolation or GHGs.

• Instantaneous Radiative Forcing

– Radiative forcing before any temperature changes.

• Adjusted Radiative Forcing

– Radiative forcing after temperatures above the tropopause change, with tropospheric temperatures held constant.

• dH(t)/dt – The change in heat storage of the

climate system; on earth, essentially the heat taken up or lost by the ocean.

• λ – Climate Sensitivity
• Equilibrium Climate Sensitivity , λeq

– F(t) = constant and sufficient time elapsed for dH/dt = 0. ∆T = ∆Teq. – λ = λeq = ∆Teq/F ; e.g., λeq = ∆T2x/F2x. F2x = 3.71 W/m2. ∆T2x taken as a synonym for λeq.

dH t

( )

dt = − ΔT t

( )

λ + F t

( )

∆T2x Simulated By The UIUC Stratosphere/Troposphere General Circulation Model

• 0.5

0.5 1 1.5 2 2.5 3 3.5

• 0.5

0.5 1 1.5 2 2.5 3 3.5 10 20 30 40 50 60 2xCO2 – 1xCO2 surface air temperature (°C) Time (years)

y = m1*(1-exp(-0.5*M0/m1)) Error Value 0.011291 2.3625 m1 NA 23.68 Chisq NA 0.76598 R2 y = m1*(1-exp(-0.5*M0/m1)) Error Value 0.01223 2.2227 m1 NA 14.79 Chisq NA 0.82407 R2

WO PCHEM W PCHEM

Outline

• Primer on Climate Sensitivity, ∆T2x
• Estimates of Climate Sensitivity, ∆T2x
• Uncertainty in ∆T2x due to uncertainty in

the radiative forcing

• Causes of temperature changes from

1856 to present

• Learning ∆T2x over time

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

Range

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

Range

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

Range RCM

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

RCM GCM Range

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

RCM GCM Paleo Range

• U.S. National Research Council study

chaired by Jule Charney wrote: "We estimate the most probable global warming for a doubling of CO2 to be near 3°C with a probable error of ±1.5°C".

• The Intergovernmental Panel on Climate

Change interpreted the findings of the Charney report to mean that 1.5°C ≤ ∆T2x ≤ 4.5°C.

• We assert that this interpretation is incorrect

and that the correct interpretation is that there is only a 50% likelihood that ∆T2x lies within 1.5° to 4.5°C.

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

RCM GCM Paleo NRC 97

Range

79

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

RCM GCM Paleo NRC 97 IPCC

Range

79

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

NRC 97 IPCC "Expert"

• pinion

Range

79

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

NRC 97 IPCC "Expert"

• pinion

Tol & de Vos Expert Forest et al Expert Wigley & Raper

Range

79

Uniform

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10 Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

IPCC

NRC 79

Andronova & Schlesinger Knutti et al Gregory et al Forest et al

Range

Uniform Forest et al Gregory et al Knutti et al

• 0.1

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

• 3
• 2
• 1

1 2 3 4 5 6 7 8 9 10

Arrhenius (1896) EBM RCM GCM Paleo IPCC Expert Uniform Forrest et al. Expert Forrest et al. Gregory et al. Andronova & Schlesinger Expert Wigley & Raper Knutti et al. Tol & de Vos NRC 79

Cumulative distribution function Climate sensitivity, ∆T2x (°C)

Arrhenius

EBM

IPCC

Range

Outline

• Primer on Climate Sensitivity, ∆T2x
• Estimates of Climate Sensitivity, ∆T2x
• Uncertainty in ∆T2x due to uncertainty in

the radiative forcing

• Causes of temperature changes from

1856 to present

• Learning ∆T2x over time

Simple climate/ocean model

NH Atmosphere

• ver Ocean

NH Mixed Layer SH Mixed Layer SH Atmosphere

• ver Ocean

NH Polar Ocean SH Polar Ocean κ W κ W βa βo NH Interior Ocean SH Interior Ocean κ W W κ βo W W κ κ NH Bottom Ocean SH Bottom Ocean βo σo, N σo, S λ λ ΔF oN ΔF oS Atmosphere

• ver Ocean

Mixed Layer Atmosphere

• ver Land

Interior Ocean Land Bottom Ocean σo , i σ

L , i

λ λ ΔF oi F Δ

Li

λ Na,o λa ,L

k S βa βo

λ Sa,o λ Na,o λ Sa,o

k N

Radiative Forcing

• 1.5
• 1
• 0.5

0.5 1 1.5 2 2.5 1750 1800 1850 1900 1950 2000

GHG Sulphate aerosol Tropospheric ozone Total anthropogenic

(A) 1363 1364 1365 1366 1367 1368 1369 1750 1800 1850 1900 1950 2000

LN HS

Year (C)

• 6
• 5
• 4
• 3
• 2
• 1

1 1750 1800 1850 1900 1950 2000 (B)

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1850 1880 1910 1940 1970 2000 Temperature Departure,

• C

(A) (NH+SH)/2 Obs

• 0.4
• 0.2

0.2 0.4 0.6 1850 1880 1910 1940 1970 2000 (B) (NH–SH) Obs Temperature Departure,

• C

Observed Surface Air Temperature Changes

Simulated Vs. Observed Surface Air Temperature Change

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1850 1880 1910 1940 1970 2000 Temperature Departure,

• C

(A) (NH+SH)/2 Obs Sim(GTAS)

• 0.4
• 0.2

0.2 0.4 0.6 1850 1880 1910 1940 1970 2000 (B) (NH–SH) Obs Sim(GTAS) Temperature Departure,

• C
• 0.3
• 0.2
• 0.1

0.1 0.2 0.3 0.4 1850 1880 1910 1940 1970 2000 Obs – sim temperature (

• C )

(C) (NH+SH)/2 Res = Obs – Sim(GTAS)

• 0.4
• 0.3
• 0.2
• 0.1

0.1 0.2 0.3 1850 1880 1910 1940 1970 2000 Obs – sim temperature (

• C )

(D) NH–SH Res = Obs – Sim(GTAS)

Schlesinger & Ramankutty (1994)

1 10 100 5 10 15 20 25 30 35 40 A Eigenvalues λk/ Total Variance x 100 (%) IPCC 92 (1858-1992) N = 135, M = 40 Mode

• 0.08
• 0.06
• 0.04
• 0.02

0.02 0.04 0.06 0.08 1850 1870 1890 1910 1930 1950 1970 1990

∧

X 1

∧

X 2 Mode Time Series (°C) C Year

• 0.3
• 0.2
• 0.1

0.0 0.1 0.2 0.3 5 10 15 20 25 30 35 40

ρ1 ρ2

Eigenvectors (EOFs,

• C)

B i (years)

• 0.4
• 0.3
• 0.2
• 0.1

0.1 0.2 0.3 0.4 1850 1870 1890 1910 1930 1950 1970 1990

∧

X 1 +

∧

X 2 D IPCC Obs Year Detrended Temperature Anomaly (°C)

Schlesinger & Ramankutty (1994)

180 120W 60W 0 60E 120E 180 90N 60N 30N 30S 60S 90S 90N 60N 30N 30S 60S 90S

2 1 3 5 9 11 6 7 9 5 8 4

180 120W 60W 0 60E 120E 180

10

∆T2x Versus Radiative Forcing

1 2 3 4 5 GT GTA GTAV1 GTAS GTASV1 GTASV2 ∆T2x (°C) Radiative forcing model IPCC Range

Conclusion 1

• To reduce the uncertainty in

climate sensitivity requires reducing the uncertainty in the radiative forcing, not only by aerosols, but also by the Sun and volcanoes.

Outline

• Primer on Climate Sensitivity, ∆T2x
• Estimates of Climate Sensitivity, ∆T2x
• Uncertainty in ∆T2x due to uncertainty in

the radiative forcing

• Causes of temperature changes from

1856 to present

• Learning ∆T2x over time

Contributions to the Observed Temperature Changes

• 50

50 100 150 Anthro Volcano Sun Residual Contribution to ∆T (%) (A) 1856–1990

• 50

50 100 150 Anthro Volcano Sun Residual Contribution to ∆T (%) (B) 1904–1944

• 50

50 100 150 Anthro Volcano Sun Residual anthro anthro+HS anthro+LN anthro+volc anthro+volc+HS anthro+volc+LN Contribution to ∆T (%) (C) 1944–1976

• 40
• 20

20 40 60 80 Anthro Volcano Sun Residual Contribution to ∆T (%) (D) 1976–1990

Conclusions 2

• The observed warming during

1904-1944 and cooling during 1944-1976 were not human induced.

• The observed warming during

1976-1990 was equally due to humans and the residual.

Conclusions 2

• The observed warming during

1856-1990 was mostly human induced.

Outline

• Primer on Climate Sensitivity, ∆T2x
• Estimates of Climate Sensitivity, ∆T2x
• Uncertainty in ∆T2x due to uncertainty in

the radiative forcing

• Causes of temperature changes from

1856 to present

• Learning ∆T2x over time

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1860 1880 1900 1920 1940 1960 1980 2000

Temperature departure, (oC)

5 10 15 20 25 30 1940 1950 1960 1970 1980 1990 2000

Climate sensitivity, (

• C)

each is based on 40 realisations

• f the residuals

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1860 1880 1900 1920 1940 1960 1980 2000

Temperature departure, (oC)

5 10 15 20 25 30 1940 1950 1960 1970 1980 1990 2000

Climate sensitivity, (

• C)

each is based on 40 realisations

• f the residuals

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1860 1880 1900 1920 1940 1960 1980 2000

Temperature departure, (oC)

5 10 15 20 25 30 1940 1950 1960 1970 1980 1990 2000

Climate sensitivity, (

• C)

each is based on 40 realisations

• f the residuals

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1860 1880 1900 1920 1940 1960 1980 2000

Temperature departure, (oC)

5 10 15 20 25 30 1940 1950 1960 1970 1980 1990 2000

Climate sensitivity, (

• C)

each is based on 40 realisations

• f the residuals

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1860 1880 1900 1920 1940 1960 1980 2000

Temperature departure, (oC)

5 10 15 20 25 30 1940 1950 1960 1970 1980 1990 2000

Climate sensitivity, (

• C)

each is based on 40 realisations

• f the residuals

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1860 1880 1900 1920 1940 1960 1980 2000

Temperature departure, (oC)

5 10 15 20 25 30 1940 1950 1960 1970 1980 1990 2000

Climate sensitivity, (

• C)

each is based on 40 realisations

• f the residuals

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1860 1880 1900 1920 1940 1960 1980 2000

Temperature departure, (oC)

5 10 15 20 25 30 1940 1950 1960 1970 1980 1990 2000

Climate sensitivity, (

• C)

each is based on 40 realisations

• f the residuals

Conclusion 3

• The uncertainty in climate

sensitivity due to climate noise can be reduced by learning over time, that is, by performing future estimations using longer

• bservational records.

Conclusions 4

• It is quite likely that the

formulation and negotiation of policies to abate human-induced climate change will, for the foreseeable future, continue to be made against a backdrop of deep uncertainty.

Recommendation

• Focus not only on where we end,

but also on how we can begin.

• Thus, need to develop near-term

hedging strategy(ies) that will get buy-in by the US and Australia.

• Effective Climate Sensitivity, λe(t)

– F(t) need not be constant and/or equilibrium achieved. – λ(t) = λe(t) = ∆T(t)/[F(t) – dH(t)/dt] . – If F = constant, then λe(t) –> λeq as t –>infinity; for F > 0, from above, λe(t) ≥ λeq .

• Transient Climate Response

– Change in global-mean surface air temperature for a transient climate forcing at the time of doubling of the CO2 concentration.

dH t

( )

dt = − ΔT t

( )

λ + F t

( )

∆T2x Simulated By Some General Circulation Models

Transient Climate Response

Outline

• Primer on Climate Sensitivity, ∆T2x
• Estimates of Climate Sensitivity, ∆T2x
• Uncertainty in ∆T2x due to uncertainty in

the radiative forcing

• Causes of temperature changes from

1856 to present

• Uncertainty in ∆T2x due to uncertainty in

the temperature data

• Learning ∆T2x over time

Estimates of ∆T2x and ∆FSO4 from three datasets for GA

Dataset/Quantity ΔT

2x °C

( )

ΔF

SO4 (Wm−2)

Jones et al. [1999] 4.8564 –0.2183 Jones and Moberg [2003] 4.2437 –0.2068 Folland et al. [2001a] 1.2771 +0.0095

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1850 1870 1890 1910 1930 1950 1970 1990 ∆T (

• C)

Northern Hemisphere FEA JAM

• 0.6
• 0.4
• 0.2

0.2 0.4 0.6 1850 1870 1890 1910 1930 1950 1970 1990 ∆T (

• C)

Southern Hemisphere FEA JAM

Conclusion 3

• The uncertainty in the southern

hemisphere temperatures must be reduced.

Conclusions 5

• Such policy formulation can be

aided by the tool of robust adaptive decision analysis. But, that is a topic for a future lecture.