This presentation summarizes greenhouse gas emissions data from the VERIFY H2020 project. It finds that top-down estimates show higher variability than bottom-up inventories. For CO2, land use models agree with reported inventories in direction of the sink or source, though variability is high. For CH4 and N2O, differences between reported and modeled estimates are largely due to use of different methodologies and tiers. The presentation identifies remaining challenges around harmonizing definitions, reducing model uncertainties, and better understanding sectoral contributions at regional levels.

1. This presentation contains unpublished data and figures.
It should only be used as information purposes.
The data belongs to the VERIFY H2020 project and the use of it its
forbidden without asking permission from the author(s).
The data behind the figures is subject to regular updates and the
version included in this presentation does not represent the latest
version.
For more information please contact: a.m.r.petrescu@vu.nl
Thank you !
DISCLAIMER
https://verify.lsce.ipsl.fr/

2. This project has received funding from the European Union’s Horizon 2020
research and innovation programme under grant agreement No 776810
ICOS Science Conference Online
Plenary, 17 September, 2020
A synthesis of European greenhouse gas emissions
and their uncertainties
A.M. Roxana Petrescu, et al.
Vrije Universiteit Amsterdam, The Netherlands
visiting at: JRC European Commission, Ispra, Italy

3. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
Contents
Main achievements since the last ICOS Science Conference
in Prague (publications and findings)
Current synthesis results – preliminary key messages for the
three main GHGs
Challenges
3

4. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
THE VERIFY SYNTHESIS
4

5. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
ACHIEVEMENT
Andrew, R. M.: A comparison of estimates of global carbon dioxide
emissions from fossil carbon sources, Earth Syst. Sci. Data, 12,
1437–1465
https://essd.copernicus.org/articles/12/1437/2020/
Petrescu et al., synthesis on European anthropogenic AFOLU
bottom-up GHG emissions
https://essd.copernicus.org/articles/12/961/2020/
• It highlights the importance of reliable quantification of GHG
emissions to support action under the Paris Agreement
• It provides an overview of existing BU data sets for AFOLU sector
• It identifies uncertainties related to the calculations of emissions,
and their sources
5

6. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
MAIN FINDINGS
For CH4 and N2O emissions the main differences between
NGHGI reports and models are the use of tiers and
methodologies (for both emissions and uncertainty
calculation).
One detected similarity between all sources is the use of
EFs, as almost all sources make use of the IPCC defaults.
AD is shared, often data sources rely on the same basic
activity data (FAOSTAT or MS) but there is some complexity
to it (see Fig. 4 in Petrescu et al., 2020)
For CO2 and LULUCF sector, there is the need to reduce the
gap between inventories and models by defining common
definitions in land use reporting
6

7. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
DATA AVAILABILITY
7
CO2, CH4 and N2O – freely available
http://webportals.ipsl.jussieu.fr/VERIFY/FactSheets/
We thank Patrick Brockmann, Matthew McGrath and Philippe Peylin for the design and help with the web portal
We thank Matthew McGrath, Chunjing Qiu and Robbie Andrew for the work on the plots

8. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
SUMMARY OF THE SYNTHESIS PAPERS – CO2
8
sink
source

9. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
CO2 KEY MESSAGES
9
• TD results show a much higher variability (min and max)
as well in the extremes of the min/max.
• Regional EUROCOM ensemble mean seems to be the
closest to the NGHGI but it shows high variability:
does an ensemble of inversion systems performs better then
one model? Should we develop inversion systems for
diversity purposes?
• fossil CO2 emissions from EU27+UK, split by major
source categories show good agreement between all
data sources and UNFCCC NGHGI (2019) as well as the
inversion mean.
• land CO2: net land use and HWP: despite high variability
of CO2 estimates from DGVMs and ORCHIDEE models,
we see some consensus in the time series mean with the
NGHGI total LULUCF estimate.
• Bookkeeping models and FAOSTAT are using country
statistics AD, same as NGHGI, therefore a good
agreement is seen
Sources of uncertainty:
- allocation, aggregation (fossil CO2)
- input data (e.g. area) and structural/parametric
uncertainty of models
- definitions (e.g. land use)
- simulation setup (TD, process based, inventory based)

10. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
SUMMARY OF THE SYNTHESIS PAPERS - CH4
10

11. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
ANTHROPOGENIC ESTIMATE FOR CH4
11
From both regional and global
inversions we subtracted the peatland
FMI natural emissions, Etiope
geological fluxes (scaled by Hmiel et
al) and inland water emissions
The EU27+UK JSBACH-HIMMELI (FMI)
CH4 peatland estimates are similar to
the natural wetlands estimates
provided by GCP (1.43 vs 1.44 Tg
CH4 yr−1, averaged over 2005-2017)

12. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
CH4 KEY MESSAGES
12
• Regarding TD results, we advocate the need
of better quantification and possible
inclusion of natural estimates in national
reporting which, at both global and regional
level, might be the key for explaining the
differences between anthropogenic BU and
total TD estimates
• The UNFCCC 2019 Refinement advices the
MS to actively try to include total TD
estimates in their country reporting
BU estimates are, in general within the uncertainty range
of the UNFCCC NGHGI data. The main differences are
caused by the application of different tiers and methods
used in calculating emissions, especially for agriculture
(e.g. the use of same AD and EFs) as discussed in Petrescu
et al., 2020 AFOLU publication

13. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
SUMMARY OF THE SYNTHESIS PAPERS – N2O
13

14. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
ANTHROPOGENIC ESTIMATE FOR N2O
14
• From both regional and global inversions we subtracted the natural inland water emissions (9.6-15.2
kton N2O yr-1). We are aware that for Europe a large (~80%) is anthropogenic (leaching of N-fertilizers)
and we will apply the correction.
• The N2O emission from naturals soils are small in the EU, as most of the land is considered to be
managed

15. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
N2O KEY MESSAGES
15
• Overall, in EU27+UK the highest uncertainty in the
UNFCCC NGHGI reporting (2018) comes from the
waste sector (626 %), followed by agriculture
(107%).
• The main sources for these uncertainties are related
to the wastewater treatment and discharge (913 %)
and the direct and indirect emissions from
agricultural soils (121 %)
• For TD it is impossible to separate the N2O natural
from anthropogenic sources (uncertainty introduced
by definitions)
• Further improvement of inverse methods for N2O is
needed to determine the total level of emissions
and, most importantly, the trends.
• NGHGI N2O uncertainties are very large
• For all IPCC sectors, the BU anthropogenic estimates
show consistent trends and values with the NGHGI
(agriculture, IPPU)

16. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
FUTURE CHALLENGES
We need to identify the reasons for differences between BU and TD not only for EU as a whole but
regional/country level: e.g. Northern Europe is dominated by natural emissions (wetlands), Western
Europe dominated by agriculture.... might be useful in defining actions related to key sector emitters and
define mitigation strategies
16

17. VERIFY GA meeting | July 7th -9th , 2020 |Teleconference
QUESTIONS (CHALLENGES) TO YOU ALL
Can we learn from this current synthesis and make progress next year?
(e.g. include biomass burning emissions from GFED, GFAS, UNFCCC)
Can BU modelers be encouraged to use common definitions, terminology?
(e.g. land use reporting, input parameters, ground observation data)
Can we find common grounds for the model selection?
Can more products report their uncertainties?
Can TD models reduce their uncertainty (e.g. CSR)?
Can we identify the reasons for differences between BU and TD not only
for EU but regional/country level? e.g. Northern Europe being dominated by natural
emissions (wetlands) Western Europe dominated by agriculture.... might be useful in defining
actions related to key sector emitters and define mitigation strategies
For consistent comparison with UNFCCC NGHGI, can we better understand
the prior information (sectoral partitions) used by inverse models to
reduce the uncertainty in calculating the anthropogenic component from
inversions?
17

18. This project has received funding from the European Union’s Horizon 2020
research and innovation programme under grant agreement No 776810
Thank you for your attention
For questions/comments please send an email to:
a.m.r.petrescu@vu.nl