Ingeborg Levin, Tobias Naegler, Renate Heinz, Daniel Osusko ...

Atmospheric observation-based global SF 6
emissions: comparison of top-down and
bottom-up estimates
I. Levin 1 , T. Naegler 1 , R. Heinz 1 , D. Osusko 1 , E. Cuevas 2 ,
A. Engel 3 , J. Ilmberger 1 , R. L. Langenfelds 4 ,
B. Neininger 5 , C. v. Rohden 1 , L. P. Steele 4 ,
A. Varlagin 6 , R. Weller 7 , D. E. Worthy 8 , S. A. Zimov 9
1
IUP, Univ. Heidelberg, 2 INM Izaña Obs., 3 IAU, Univ. Frankfurt, 4 CMAR,
Aspendale, 5 MetAir AG, Hausen, 6 IPEE, Moscow, 7 AWI,
Bremerhaven, 8 Env. Canada, Toronto, 9 Cherskii Obs.

Sulphur Hexafluoride: SF 6
… is a very stable man-made Greenhouse Gas
Mean atmospheric lifetime: ≈ 3 000 years
Global Warming Potential: ≈ 23 000 x CO 2 (100 yr time horizon)
→ Kyoto - reported
Atmospheric mixing ratio today: ≈ 7 ppt (10 -12 mol/mol)
Sources of SF 6 :
- ca. 75% from electrical applications
- Magnesium industry
- adiabatic applications
Sinks of SF 6 : only in the Mesosphere > 60 km
- UV Absorption (λ < 130 nm)
- electron reactions

Heidelberg co-operative network of
tropospheric SF 6 observations
& stratospheric profiles from Kiruna, Aire sur l‘Adour, Teresina

Global long-term trend of SF 6 in the
troposphere
SF 6
[ppt]
7
6
5
4
3
Alert 82°N (cylinders)
Alert 82°N (flasks)
Syktyvkar 61°N (flasks >2500m)
Izaña 28°N (cylinders)
Cape Grim 41°S (air archive)
Cape Grim 41°S (flasks)
Neumayer 71°S (cylinders)
Neumayer 71°S (flasks)
northern hemisphere
2
1
0
southern hemisphere
1980 1985 1990 1995 2000 2005 2010
Year

Observed tropospheric SF 6 growth rates
SF 6
growth rate [ppt yr -1 ]
0.35
0.30
0.25
0.20
0.15
0.10
Alert (cylinders & flasks)
Cherskii (flasks)
Syktyvkar (flasks)
Schauinsland (flasks)
Izana (cylinders)
Cape Grim (archive & flasks)
Neumayer (cylinders & flasks)
UNFCCC adopted
± 0.014 ppt yr -1
(ca. 6 %)
0.05
1980 1985 1990 1995 2000 2005 2010
Year
Practically no SF 6 sinks:
global mean growth rate ≅ global mean SF 6 emissions

Atmospheric observation-inferred global
SF 6 emissions
SF 6
emission [Gg/a]
10
8
6
4
2
0
total this study (±1σ)
1980 1985 1990 1995 2000 2005

How well do the 2009 bottom-up EDGAR
estimates compare to our
atmospheric observation-based
top-down emissions

Atmospheric observation-inferred global
SF 6 emissions
SF 6
emission [Gg/a]
10
8
6
4
2
0
total this study (±1σ)
1980 1985 1990 1995 2000 2005

Comparison with new global (bottom-up)
EDGAR-estimated SF 6 emissions
SF 6
emission [Gg/a]
10
8
6
4
2
0
total this study (±1σ)
total (EDGAR)
1980 1985 1990 1995 2000 2005

How do UNFCCC-reported SF 6 emissions
compare to our top-down estimate
Problem:
Only industrialised countries (Annex I) are required to report their
GHG emissions to UNFCCC, these are
Western Europe, Canada, U.S.A., Japan, Australia, New Zealand,
Eastern Europe, Russia & Turkey
Therefore, we separate here into
Annex I and non-Annex I
which are newly industrialised countries, i.e. China, India, Brazil,
others

UNFCCC-based and EDGAR estimated
Annex I SF 6 emissions
SF 6
emission [Gg/a]
10
8
6
4
2
0
total this study (±1σ)
total (EDGAR)
Annex I (EDGAR)
Annex I (UNFCCC-based)
1980 1985 1990 1995 2000 2005

UNFCCC-based and EDGAR Annex I
& non-Annex I SF 6 emissions
SF 6
emission [Gg/a]
10
8
6
4
2
0
6
4
total this study (±1σ)
total (EDGAR)
Annex I (EDGAR)
Annex I (UNFCCC-based)
Non-Annex I (EDGAR)
Non-Annex I (UNFCCC-based)
total
Annex I
non-Annex I
2
0
1980 1985 1990 1995 2000 2005

Are non-Annex I
countries really
responsible for the
major part of global
SF 6 emissions today
SF 6
emission [Gg/a]
10
8
6
4
2
0
6
total this study (±1σ)
total (EDGAR)
Annex I (EDGAR)
Annex I (UNFCCC-based)
Non-Annex I (EDGAR)
Non-Annex I (UNFCCC-based)
4
2
0
1980 1985 1990 1995 2000 2005

Are non-Annex I
countries really
responsible for the
major part of global
SF 6 emissions today
SF 6
emission [Gg/a]
10
8
6
4
2
0
6
total this study (±1σ)
total (EDGAR)
Annex I (EDGAR)
Annex I (UNFCCC-based)
Non-Annex I (EDGAR)
Non-Annex I (UNFCCC-based)
4
2
This would require
much larger SF 6
emissions per electrical
power production in
non-Annex I than in
Annex I countries
Gg SF 6
/a per GWh
0
1.5
1.0
0.5
0.0
total (EDGAR)
Annex I (EDGAR)
non-Annex I (EDGAR)
Annex I (UNFCCC-based)
inferred non Annex I (UNFCCC-based)
1980 1985 1990 1995 2000 2005

The SF 6 north-south gradient principally also
provides information on the distribution of emissions
SF 6
[ppt]
7
6
5
4
3
Alert 82°N (cylinders)
Alert 82°N (flasks)
Syktyvkar 61°N (flasks >2500m)
Izaña 28°N (cylinders)
Cape Grim 41°S (air archive)
Cape Grim 41°S (flasks)
Neumayer 71°S (cylinders)
Neumayer 71°S (flasks)
northern hemisphere
2
1
0
southern hemisphere
1980 1985 1990 1995 2000 2005 2010
Year

Observed difference between Alert (82°N)
and Neumayer (71°S)
ΔSF 6
[ppt]
0.6
0.4
0.2
Alert - Neumayer:
observed
GRACE: Edgar adjusted
GRACE: UNFCCC-based
0.0
1980 1985 1990 1995 2000 2005

Observed and simulated difference between
Alert (82°N) and Neumayer (71°S)
ΔSF 6
[ppt]
0.6
0.4
0.2
Alert - Neumayer:
observed
GRACE: Edgar adjusted
GRACE: UNFCCC-based
0.0
1980 1985 1990 1995 2000 2005
Simulation with the coarseresolution
atmospheric box
model GRACE based on
UNFCCC emissions

Observed and simulated difference between
Alert (82°N) and Neumayer (71°S)
ΔSF 6
[ppt]
0.6
0.4
0.2
Alert - Neumayer:
observed
GRACE: Edgar adjusted
GRACE: UNFCCC-based
0.0
1980 1985 1990 1995 2000 2005
Simulations with
EDGAR distribution

Observed and simulated difference between
Alert (82°N) and Neumayer (71°S)
ΔSF 6
[ppt]
0.6
0.4
0.2
Alert - Neumayer:
observed
GRACE: Edgar adjusted
GRACE: UNFCCC-based
0.0
1980 1985 1990 1995 2000 2005
→ Model transport uncertainties limit constraints on the
north-south distribution of emissions, this would also be a
concern for high-resolution models !

Summary
Global atmospheric SF 6 mixing ratio has increased from almost
zero in the 1970s to almost 7 ppt today
After a decrease of annual global emissions in 1996-1998, SF 6
sources increase again since 1998
Bottom-up estimates by EDGAR compare well with our inferred
emissions, however, for some periods, they are significantly
different
Annex I reported emissions are surprisingly low and leave a
large gap of non-reported emissions
… but model transport uncertainties and the number of
observational sites in our network limit emission
apportionment to Annex I or non-Annex I countries

Thank you !

The coarse-resolution GRACE model

Simulating tropospheric SF 6 with EDGARestimated
SF 6 emissions
SF 6
[ppt]
7
6
5
4
3
6.0
5.5
5.0
Alert 82°N (tanks & flasks)
Neumayer 71°S (tanks & flasks)
2003 2004 2005
2
1
Cape Grim 41°S (archive & flasks)
GRACE: EDGAR original
0
1980 1985 1990 1995 2000 2005

Simulating tropospheric SF 6 with
observation-inferred emissions
SF 6
[ppt]
7
6
5
4
3
6.0
5.5
5.0
Alert 82°N (tanks & flasks)
Neumayer 71°S (tanks & flasks)
2003 2004 2005
2
1
0
Cape Grim 41°S (archive & flasks)
GRACE: EDGAR original
GRACE: EDGAR adjusted
1980 1985 1990 1995 2000 2005