EMEPTFMM Workshop on the review of the MSCE models on HMs and POPs 1314 October 2005, Moscow

1
EMEP/TFMM Workshop on the review of the MSC-E
models on HMs and POPs 13-14 October 2005, Moscow
An evaluation of emission data for POPs and HMs
with emphasis on model applicability
Knut Breivik1, Vigdis Vestreng2, Olga M.
Rozovskaya3, Jozef M. Pacyna1 1 EMEP CCC
Norwegian Institute for Air Research (NILU) 2
EMEP MSC-W Norwegian Meteorological Institute
(met.no) 3 EMEP MSC-E
2
An evaluation of available emission data with
emphasis on model applicability
Motivation and Conclusion
If we want to understand and control POPs and
HMs, knowledge about the sources is essential!
3
An evaluation of available emission data with
emphasis on model applicability
Outline

• Introduction
• Sources of HMs and POPs
• Features used for categorisation of emission data
• Scientific features
• Policy-oriented features
• European emission inventories
• Overview
• Intercomparison for 1990
• Relative changes in key sources (1990 vs 2003)
• Discussion

4
An evaluation of available emission data with
emphasis on model applicability
Simplified source categories (POPs and HMs)
5
An evaluation of available emission data with
emphasis on model applicability
Emission inventories (features and aims)

• Complete coverage of sources (control strategies)
• Complete coverage of Parties (political domain)

• Complete spatial coverage (geographical domain)
• Complete temporal coverage (POPs persistence,
  old vs new)
• Complete compound coverage (speciation, phys-chem
  state)
• Complete media coverage (POPs multimedia)

Common to both

• As accurate as possible (absolute numbers)
• Information on uncertainty requested

Key question Is this information available (on
a Europan level)?
6
An evaluation of available emission data with
emphasis on model applicability
Emission inventory features a reflection of aims
5 Prevedouros et al 2004 EST
1 webdab.emep.int
4 e.g. Chemos. 2000
3 Atmos Environ 2003
2 Berdowski et al. 1997 TNO-report
9 Bailey Chemos 2001
7 Y.-F. Li, numerous publications
8 Breivik et al STOTEN 2002a,b
6 Denier van der Gon et al. 2005 TNO-report
7
An evaluation of available emission data with
emphasis on model applicability
EMEP data Number of reporting parties by
pollutant
CLRTAP has 49 parties, the POPs protocol has been
ratified by 24 parties (in force from 2003)
The numbers in parentheses are the number of
parties submitting a numerical value (gt0). Data
for 1980-1989 indicate the annual maximum
reporting during the period.
8
An evaluation of available emission data with
emphasis on model applicability
Emissions of PAHs in 1990 (EMEP vs TNO)Sorted
from EMEP high to EMEP low
PAHs (and PCDD/Fs) extensively reported
(stack-derived approach) Discrepancies can in
part be caused be differences in speciation (TNO
Borneff 6 EMEP 4) TNO ca 3 times EMEP More than
an order of magnitude differences for AT, BG, DE,
DK, EE, FR, IS, RU
9
An evaluation of available emission data with
emphasis on model applicability
Emissions of PCBs in 1990 (EMEP vs TNO)
Only 11 Parties have reported data Differences in
speciation TNO SPCB209/SPCB7 EMEP ? TNO about
2 EMEP ES and NL are reporting zeros.
Plausible?
10
An evaluation of available emission data with
emphasis on model applicability
Emissions of HCB in 1990 (EMEP vs TNO vs NILU)
For 12 parties there are 3 estimates
available Large differences and in many cases
more than an order of magnitude Some EMEP and TNO
data include zeros Sum 12 parties NILU gt EMEP gt
TNO
11
An evaluation of available emission data with
emphasis on model applicability
Emissions of HCB in 1990 (EMEP vs TNO vs NILU)
EMEP gt NILU gt TNO Fair agreement in many cases
(e.g. CH) EMEP reports O for DK. Impossible!
12
An evaluation of available emission data with
emphasis on model applicability
EMEP Key source analysis for 1990 and 2003 (in )
PCDD/Fs and PAHs AT, BE, CA, DK, FR, IS, NL, NO,
ES, SE, GB PCBs FR, SE, GB HCB AT, BE, CA, FR,
ES, SE, GB
13
An evaluation of available emission data with
emphasis on model applicability
EMEP data Number of reporting parties by
pollutant
CLRTAP has 49 parties, the HMs protocol has been
ratified by 27 parties (in force from 2003)
The numbers in parentheses are the number of
parties submitting a numerical value (gt0). Data
for 1980-1989 indicate the annual maximum
reporting during the period.
14
An evaluation of available emission data with
emphasis on model applicability
Emissions of Pb in 1990 (EMEP vs TNO)
For 20 countries out of 29 countries, the results
are within a factor of 2 Differences may in
extreme cases exceed several orders of magnitude
(PT, LV, CY) PT Unit error?
15
An evaluation of available emission data with
emphasis on model applicability
Emissions of Cd in 1990 (EMEP vs TNO)
All estimates are within an order of magnitude,
For 18 out of 27 countries, the estimates are
within a factor of 2 Differences found factor
of gt5 (IT), factor of gt2 (RU, ES, CZ, EE, HR,
AT) The sum of EMEP and TNO estimates deviate by
70 (TNO higher)
16
An evaluation of available emission data with
emphasis on model applicability
Emissions of Cd in 1990 (EMEP vs TNO)
All estimates with the exception of BY are within
an order of magnitude. For 17out of 26 countries,
the results are within a factor of 2.
Differences found factor of gt5 (IT), factor of
gt2 (RU, ES, CZ, EE, HR, AT) The sum of EMEP and
TNO estimates deviate by 78 (TNO higher)
17
An evaluation of available emission data with
emphasis on model applicability
EMEP Key source analysis for HMs, 1990 and 2003
(in )
18
An evaluation of available emission data with
emphasis on model applicability
Discussion (I) official emission data

• Official emission data are of limited value in
  terms of model applicability as there is
  insufficient emphasis and information on the
  relevant features (temporal and spatial
  resolution, speciation, media coverage).
• It seems safe to claim that most emission data
  for HMs and POPs (policy and research-driven
  estimates) still suffer from significant
  uncertainties and poor accuracy.
• Further improvement of official emission data
  with respect to research-driven features seems to
  be the most sensible way to proceed (more
  research seems required before this can be
  accomplished as the current empirical basis is
  weak)

19
An evaluation of available emission data with
emphasis on model applicability
Discussion (II) implications for MSC/E modelling

• Knowledge on the sources and emissions are likely
  to remain the least understood feature with
  respect to the overall behaviour and fate of many
  of these compounds. This needs to be taken into
  account when evaluating the MSC/E models.
• Great care needs to be taken when selecting /
  adjusting / combining emission inventories for
  the purpose of model use or data interpretation.
  Additional assumptions leads to additional
  uncertainties and lack of transparency.
• Certain features of emission inventories may
  still be fair (spatial patterns, temporal
  patterns, congeneric patterns, absolute levels
  etc) and model evaluation exercises may take
  advantage of these features