1 00/XXXX © Crown copyright Apportioning climate change indicators between regional emitters Jason Lowe and Geoff Jenkins Hadley Centre for Climate Prediction.

1 00/XXXX © Crown copyright

Apportioning climate change indicators between regional

emitters

Jason Lowe and Geoff Jenkins

Hadley Centre for Climate Prediction and Research

25th September 2002


What this talk is not about

This talk is not about the HadCM3 validationdata. Choice of this validation was arbitrary andother datasets are available.

What this talk is aboutThis talk is about the Hadley Centre contributionto this simple modelling exercise.

Building our capacity in this area


Contents

Introduction and models Results of phase 1 Results of phase 2 Conclusions


Estimating regional share

CONCENTRATIONS FROM EACH REGION

EMISSIONS FROM EACH REGION

RADIATIVE FORCING FOR EACH REGION

TEMPERATURE CHANGE FOR EACH REGION

SHARE


Choice of model units (1)Input data:-Linearly interpolated between values

Carbon cycle model:-Impulse response function fitted to Bern model

Default case uses the SAR standard parameters

CH4 and N2O:-Single fixed lifetime for each gas, taken from TAR (page 244)


Choice of model units (2)“Climate model”:-Impulse response function fitted to Hadley Centre 4xCO2stabilisation experiment.

The forcing caused by a doubling of CO2 quoted in the IPCC TAR (page 358) is 3.71 Wm-2.

Forcing expressed as a multiple of the 4xCO2 forcing.


Extended model

In order to achieve a better fit to the A2 CO2 and temperature

predicted by more complex models the forcing and emissions were modified by temperature dependent functions.

The form of these functions was chosen arbitrarily. An iterative

calculation was used to calculate the CO2 and temperatures.

Carbon cycle function

=0.46(1+0.7(T/Tmax)2)


Choices and uncertainties

Start year End year for emissions End year for calculation Emissions scenario Attribution method Choice of species Size of regional groupings

Gas cycle parameters Climate model Feedback Emissions scenario Attribution method Choice of species Aerosols and other forcing Choice of historical

emissions Size of regional groupings

ScientificPolicy options


Contents



CDIAC (CO2) – Basic model


CDIAC – Extended model (feedback)


Can we simulate B1 CO2 concentrations using a simple

model? Input to HadCM3 is used as a comparison

B1 CO2 Concentrations

0

200

400

600

800

1000

1890 1950 2000 2050 2100

HadCM3

nofeedback

feedback

HadCM3 CO2 concentrations derived from Bern carbon cycle model.

Pre-1990 values agree well with observations


Can we simulate A1FI CO2 concentrations using a simple

model?Input to HadCM3 is used as a comparison

A1fi CO2 Concentrations

0

200

400

600

800

1000

1890 1950 2000 2050 2100

HadCM3

nofeedback

feedback


Can we simulate temperature rise using a simple model?

HadCM3 simulation is used as a comparison

A2

Had

CM

3

A2

nofe

edba

ck

A2

feed

back

A1f

i Had

CM

3

A1f

i nof

eedb

ack

A1f

i fee

dbac

k

B2

Had

CM

3

B2

nofe

edba

ck

B2

feed

back

B1

Had

CM

3

B1

nofe

edba

ck

B1

feed

back

00.5

11.5

22.5

33.5

44.5

5

Temperature rise from 1890

1990 2100


Contents



Attribution methods1. “All minus one” - Marginal2. Differential

Base case Simple linear version of model Edgar Hyde historic emissions + A2

future 1890 is start year for emissions


Global temperature rise from regional emissions


Regional share of temperature rise

0

0.1

0.2

0.3

0.4

0.5

0.6

OECD REF ASIA ALM

2000

2100


Sensitivity Studies Choice of indicator Effect of different emissions start years Effect of different emissions scenarios Effect of different climate and carbon cycle

parameters Effect of including a temperature feedback Effect of different attribution methods


Regional share for various indicators


Choice of indicator?

0

0.1

0.2

0.3

0.4

0.5

0.6

OECD REF ASIA ALM

Emit

Total Emit

Conc

Force

Temp

Share estimated at year 2000


Regional share of temperature rise for different emission start

dates


Are the results different for other scenarios?

A2

A1FI B1


Does the amount of carbon cycle fertilization affect the

result?

Bern high caseBern low case


Does a slower climate response (only long time constant) affect the result?

Slow climate model response


Does using the extended model affect the apportionment

calculation?

Basic model Extended model


Does using the extended model affect the apportionment

calculation?

0

0.1

0.2

0.3

0.4

0.5

0.6

OECD REF ASIA ALM

No feedback

Feedback

At year 2000

0

0.1

0.2

0.3

0.4

0.5

0.6

OECD REF ASIA ALM

No feedback

Feedback

At year 2100


Comparing attribution methods

All minus oneDifferential


Conclusions The apportionment calculation has been carried out with a number of

greenhouse gases and for a range of future emissions scenarios.

Using a more elaborate model (which includes temperature feedback) improves the simulation of gas concentrations and temperature. There is also an effect on the apportionment calculation.

If the share is not evaluated until the end of the period (2100), the results vary with emissions scenario. If the share is evaluated earlier the difference between scenarios is smaller.

Not including emissions before 1950 or 1990 tends to reduce the share of earlier emitters (e.g. OECD).

A shorter atmospheric carbon lifetime or a slower climate response can both modify the attribution results.

1 00/XXXX © Crown copyright Apportioning climate change indicators between regional emitters Jason Lowe and Geoff Jenkins Hadley Centre for Climate Prediction.

Documents

xxxx crown copyright

basic model slide

regional emissions slide

comparison slide

dates slide

onedifferential slide

area slide

a2 a1fib1 slide