upton, new york · global energy balance global and annual average energy fluxes in watts per...

CONSIDER A SPHERICAL EARTH

Stephen E. Schwartz

Upton, New York

Ithaca, New York19 November 2007

http://www.ecd.bnl.gov/steve

GLOBAL ENERGY BALANCEGlobal and annual average energy fluxes in watts per square meter

Schwartz, 1996, modified from Ramanathan, 1987

RADIATIVE FORCING

A change in a radiative flux term in Earth’s radiationbudget, ∆F, W m-2.

Working hypothesis:On a global basis radiative forcings are additive andfungible.

• This hypothesis is fundamental to the radiativeforcing concept.

• This hypothesis underlies much of the assessment ofclimate change over the industrial period.

CLIMATE RESPONSEThe change in global and annual mean temperature,∆T, K, resulting from a given radiative forcing.

Working hypothesis:The change in global mean temperature isproportional to the forcing, but independent of itsnature and spatial distribution.

∆T = λ-1∆F

CLIMATE SENSITIVITYThe change in global and annual mean temperature perunit forcing, λ, K/(W m-2),

λ-1 = ∆T/∆F.

Climate sensitivity is not known and is the objective ofmuch current research on climate change.

Climate sensitivity is often expressed as thetemperature for doubled CO2 concentration ∆T2×.

∆T2× = λ-1∆F2×

∆F2× ≈ 3.7 W m-2

CLIMATE SENSITIVITY ESTIMATESTHROUGH THE AGES

Estimates of central value and uncertainty range from majornational and international assessments

6

5

4

3

2

1

0

∆T2×

, Sen

sitiv

ity to

2 ×

CO

2, °C

190018901880

Arrhenius

Stefan-Boltzmann

2010200019901980

CharneyNRC – – – – IPCC – – – –

1 sigma> 66%

"Likely"

1.5

1.0

0.5

0.0

Sensitivity, K

/(W m

-2)

Despite extensive research, climate sensitivity remains highly uncertain.

COMMITTED WARMING IN CLIMATE MODEL RUNSAtmospheric composition held constant at 2000 value (IPCC, 2007)

models models

} Committedwarming

Temperature continues to increase for composition held constant.

Projected incremental 21st century is 50% beyond warming alreadyrealized.

“COMMITTED WARMING,” “THERMAL INERTIA,”“WARMING IN THE PIPELINE”

“ Additional global warming of ... 0.6˚C is “in the pipeline” and willoccur in the future even if atmospheric composition and other climateforcings remain fixed at today’s values.

Hansen et al, Science, 2005

“ Even if the concentrations of greenhouse gases in the atmosphere hadbeen stabilized in the year 2000, we are already committed to furtherglobal warming of about another half degree.

Meehl, Washington, et al., Science, 2005

“ Even if atmospheric composition were fixed today, global-meantemperature ... rise would continue due to oceanic thermal inertia. Thewarming commitment could exceed 1˚C.

Wigley, Science, 2005

“COMMITTED WARMING,” “THERMAL INERTIA,”“WARMING IN THE PIPELINE” (cont’d)

“ Because of the long time scale required for removal of CO2 from theatmosphere as well as the time delays characteristic of physicalresponses of the climate system, global mean temperatures are expectedto increase by several tenths of a degree for at least the next 20 yearseven if CO2 emissions were immediately cut to zero; that is, there is acommitment to additional CO2-induced warming even in the absenceof emissions.

Friedlingstein and Solomon, PNAS, 2005

WHAT CAN WE LEARN FROMENERGY BALANCE MODELS?

EMPIRICAL DETERMINATIONOF EARTH’S CLIMATE

SENSITIVITY

JUST PUBLISHED IN JGR

STOVE-TOP MODEL OFEARTH’S CLIMATE SYSTEM

STOVE-TOP MODEL OF EARTH’S CLIMATE SYSTEM

dH

dtC

dT

dtQ k T T= = − −( )amb

H = heat content T = temperature

C = system heat capacity

Q = heating rate from stove

Tamb = ambient temperature

Steady State T: T TQ

k∞ = +amb

let Q Q F→ + : ∆TF

k∞ =

Sensitivity: λ− ∞≡ =1 1∆T

F k

Time constant: τ λ= −C 1

Tem

pera

ture

Time

T = T0 + (Tf - T0)(1-e-t/τ )

0 τ

T0

Tf

τ is the time constant of the system response to a perturbation.

DEPENDENCE OF RESPONSE ON SYSTEM HEAT CAPACITY

Tem

pera

ture

Time 0 τ

T0

Tf

T

empe

ratu

reTime

0 τ

T0

Tf

For constant k, ∆T∞ and λ−1 are independent of system heat capacity C.

Time constant τ varies linearly with heat capacity: τ λ= −C 1

Sensitivity can be inferred from τ and C as λ τ− =1 / C.

BILLIARD BALL MODEL OFEARTH’S CLIMATE SYSTEM

BILLIARD BALL TEMPERATURESENSITIVITY AND TIME

CONSTANTEvaluated according to the

Stefan-Boltzmann radiation law

Global energy balance: dH

dtQ E Q T= − = − σ 4

Initially Q T0 04= σ

Temperature sensitivity: ∆ ∆T Qss = −λ 1 ; ∆ ∆T t Q e t( ) = −− −λ τ1 1( )/

For Stefan-Boltzmann planet sensitivity is λS-B-1 =

T

Q4

Relaxation time constant is τ λS-B S-B= = −T C

QC0

00

1

4 , or λ τS-B

S-B,0

1− =C

BILLIARD BALLTEMPERATURE SENSITIVITY

Evaluated according to theStefan-Boltzmann radiation law

For Q S0 0 4= γ / where S0 is the solar constant = 1370 W m-2

and γ is global mean co-albedo = 0.69

Climate sensitivity is λS-B-1 = 0.27 K/(W m-2)

For 2 × CO2 forcing F2× = 3.71 W m-2, ∆T2× = 1.0 K

ENERGY BALANCE MODEL OFEARTH’S CLIMATE SYSTEM

ENERGY BALANCE MODEL OFEARTH’S CLIMATE SYSTEM

Global energy balance: CdT

dt

dH

dtQ E J Ts

s4= = − = −γ εσ

C is heat capacity coupled to climate system on relevant time scale

Ts is global mean surface temperature H is global heat content

Q is absorbed solar energy E is emitted longwave flux

J is 14

solar constant γ is planetary co-albedo

σ is Stefan-Boltzmann constant ε is effective emissivity

ENERGY BALANCE MODEL OFEARTH’S CLIMATE SYSTEM

Apply step-function forcing:

At “equilibrium”

F Q E= −∆( )

∆T Fs( )∞ = −λ 1

Tem

pera

ture

Time 0 τ

T0

T(∞)

λ−1 is equilibrium climate sensitivity λγ

−1 04

= 0 S

fT

J K / (W m )-2

f is feedback factor fd

d T

d

d T= − +

−1

14

140 0

1lnln

lnln

γ ε

Time-dependence: ∆T t F e ts( ) ( )/= −− −λ τ1 1

τ is climate system time constant τ λ= −C 1 or λ τ− =1 / C

TEMPERATURE RESPONSE TOLINEARLY INCREASING

FORCINGβ = dforcing/dtime

Energy balance:

Time-dependence:

CdT

dtt J Ts

S s4= + −β γ εσ

∆T t t e ts( ) [( ) ]/= − +− −βλ τ τ τ1

Tem

pera

ture

Time2τ 3τ 4τ τ 0 . .

∆ T = βλ-1τ Teq

T

λ−1 and τ are the same as before: λ τ− =1 / C

For t /τ >∼ 3, ∆T t ts( ) ( )= −−βλ τ1

Temperature lags equilibrium response by: ∆Tlag = −βλ τ1

APPROACH

Empirically determine heat capacity C andtime constant τ of Earth’s climate systemfrom observations over the instrumentalperiod.

Evaluate sensitivity as λ-1 = τ/C.

DETERMINING EARTH’SHEAT CAPACITY

BY OCEAN CALORIMETRY

HEAT CAPACITY OF EARTH’SCLIMATE SYSTEM FROM GLOBAL

MEAN HEAT CONTENT ANDSURFACE TEMPERATURE TRENDS

CdH dt

dT dt

dH

dT= =/

/s s

C: Global heat capacity

H: Global ocean heat content

Ts: Global mean surface temperature

ZONAL AVERAGE HEAT CONTENT TREND (1955-2003)1018 J (100 m)-1 (1˚ latitude)-1 yr-1

• Heating is greatest in upper ocean, with downwelling plumes.

HEAT CONTENT OF WORLD OCEANS, 1022 J

Levitus et al., 2005

HEAT ABSORPTION BY COMPONENTSOF EARTH’S CLIMATE SYSTEM

The world ocean is responsible for ~84% of the increase in globalheat content. Levitus et al., 2005

GLOBAL TEMPERATURE TREND OVER THE INDUSTRIAL PERIOD

0.8

0.6

0.4

0.2

0.0

-0.2

-0.4

-0.6

Tem

pera

ture

Ano

mal

y, K

2000

2000

1990

1990

1980

1980

1970

1970

1960

1960

1950

1950

1940

1940

1930

1930

1920

1920

1910

1910

1900

1900

1890

1890

1880

1880

1870

1870

1860

1860

GISS, Goddard Institute for Space Studies, NASA, USA CRU, Climatic Research Unit, East Anglia, UK

EMPIRICAL DETERMINATION OF OCEAN HEAT CAPACITY

CdH dtdT dt

= //s

Surfacetemperature Ts:

GISS, CRU

Ocean heatcontent H:

Levitus et al., 2005 -10

-8

-6

-4

-2

0

2

4

6

Oce

an H

eat C

onte

nt A

nom

aly

H, W

yr

m-2

20001990198019701960

1.5

1.0

0.5

0.0

-0.5

Tem

perature anomaly ∆

T, K

Heat Content← L300 L700

L3000

Temperature→

GISS CRU

0.200.150.100.050.00

Fraction equilibrated

3500

3000

2500

2000

1500

1000

500

0

Dep

th, m

150100500

Equlibrated depth, m

3500

3000

2500

2000

1500

1000

500

0

Dep

th, m

20151050

Heat capacity C, W yr m-2 K-1

3500

3000

2500

2000

1500

1000

500

0

Dep

th, m

• ~50% of heat capacity is between surface and 300 m.

• Other heat sinks raise global heat capacity to 17 ± 7 W yr m-2 K-1.

CHARACTERISTIC TIME OFEARTH’S CLIMATE SYSTEM

FROM TIME SERIES ANALYSIS

DETERMINATION OF TIME CONSTANT OF EARTH’S CLIMATESYSTEM FROM AUTOCORRELATION OF TIME SERIES

Recipe (GISS annual global mean surface temperature anomaly Ts)

0.4

0.0

-0.4T

emp.

Ano

mal

y, K

2000198019601940192019001880

1. Remove long term trend; plot the residuals:210

-1-2

Nor

m. R

esid

ual

2000198019601940192019001880

2. Calculate autocorrelogram (& standard deviations; Bartlett, 1948):

-1.0-0.50.00.51.0

Aut

ocor

rela

tion

r

20151050Lag time ∆t, yr

Recipe for determining climate system time constant, continued

3. Examine the lag-1 autocorrelation:

2

1

0

-1

-2

Lag1

Nor

m. R

esid

ual

210-1-2

Normalized Residual

4. Remove the trend; plot the residuals:210

-1-2

Nor

m. R

esid

ual

120100806040200Time, yr

5. Examine for any remaining autocorrelation:

-1.0-0.50.00.51.0

Aut

ocor

rela

tion

2520151050Lag time ∆t, yr

Recipe for determining climate system time constant, continued

6. If no residual autocorrelation (Markov process) calculate timeconstant τ for relaxation of system to perturbation:

r t e t( ) /∆ ∆= − τ or τ( ) / ln ( )∆ ∆ ∆T T r T= − (Leith, 1973)

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

• Time constant τ increases with increasing lag time.

• Implies coupling of Ts to a system of longer time constant.

• On decadal scale time constant asymptotes to 5 ± 1 yr.

• This is the e-folding time constant for relaxation of global meansurface temperature to perturbations on the decadal scale.

THIS RESULT IS ROBUST1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

GISS 1880-2004

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

CRU 1880-2004

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

GISS 1880-1942

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

CRU 1880-1942

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

GISS 1943-2004

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

CRU 1943-2004

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

GISS NH 1880-2005

1086420

Tim

e C

onst

. τ, y

r20151050

Lag time ∆t, yr

CRU NH 1880-2005

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

GISS SH 1880-2005

1086420

Tim

e C

onst

. τ, y

r

20151050Lag time ∆t, yr

CRU SH 1880-2005

SUMMARY RESULTSQuantity Unit Value 1 σ

Effective global heat capacity C W yr m-2 K-1 17 7

Effective climate system time constant τ yr 5 1

Equilibrium climate sensitivity λ τ− =1 / C K/(W m-2) 0.30 0.14

Equilibrium temperature increase for 2 × CO2,∆T2×

K 1.1 0.5

Total forcing over the 20th century,F T20 20

1= −∆ / λW m-2 1.9 0.9

Forcing in 20th century other than GHGs(mainly aerosols), F F F20 20 20

other ghg= −W m-2 -0.3 1.0

Lag in temperature change, ∆Tlag K 0.03

COMPARISON WITH PREVIOUS RESULTS

6

5

4

3

2

1

0

∆T2×

, Sen

sitiv

ity to

2 ×

CO

2, °C

190018901880

Arrhenius

Stefan-Boltzmann

2010200019901980

CharneyNRC – – – – IPCC – – – –

1 sigma

> 66%

"Likely"

This study

1.5

1.0

0.5

0.0

Sensitivity, K

/(W m

-2)

Sensitivity obtained in this study is much lower than that fromclimate models and paleo studies.

WHAT MIGHT BE WRONG WITH THISANALYSIS?

• Ocean heat capacity too great, resulting in low sensitivity.

Erroneous or nonrepresentative data.

Obtaining heat capacity from measurements.

• Time constant too short, resulting in low sensitivity.

Time series too short to give true time constant.

Detrending emphasizes the rapid fluctuations.

• Earth’s climate system is much more complex than can berepresented by a single-compartment model.

Multiple time constants, multiple heat capacities.

CLIMATE SENSITIVITY AND INFERRED20th CENTURY TOTAL AND AEROSOL FORCING

Inverse calculation of forcing as function of climate system time constant τ

λ τ− =1 / C F T C T20 201

20= =−∆ ∆/ /λ τ F F Faer GHG= −20

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Clim

ate

sens

itivi

ty λ

-1, K

/(W

m-2)

16141210864

Climate system time constant τ, yr

5

4

3

2

1

0 ∆

T2×

, K

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Aerosol forcing over 20

th century, W m

-2

161284

Climate system time constant τ, yr

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Tot

al fo

rcin

g ov

er 2

0th c

entu

ry, W

m-2

Time constant from autocorrelation is τ = 5 ± 1 yr.

Submitted comment suggests τ too small because of length of data record.

Climate sensitivity and inferred forcing depend strongly on time constant.

SUMMARY• Despite intense research Earth’s climate sensitivity

remains uncertain to at least a factor of 2.

• Energy balance considerations and empirical observationsmay usefully refine sensitivity estimates.

• Climate sensitivity can be determined as time constantupon heat capacity.

• The effective heat capacity of Earth’s climate system is17 ± 7 W yr m-2 K-1 ≈ 150 m of the world ocean.

The time constant of Earth’s climate system is 5 ± 1 years.

• Climate system response to greenhouse forcing is innear steady state, with little further warming (due topresent GH gases) “in the pipeline.”

• The equilibrium sensitivity of Earth’s climate system is0.30 ± 0.14 K / (W m-2); ∆T2× = 1.1 ± 0.5 K.

CONCLUDING OBSERVATION

• The time constant, heat capacity and sensitivity of Earth’sclimate system are important integral properties that mightinstructively be examined in model calculations as well asin observations.