The POP Model Intercomparison Study: Steps Towards a Mechanistic Interpretation of the Results Marti

1
The POP Model Intercomparison Study Steps
Towards a Mechanistic Interpretation of the
ResultsMartin Scheringer, Fabio WegmannSwiss
Federal Institute of Technology Zürich EMEP
Task Force on Measurements and ModellingMoscow14
October 2005
2
Overview

• Where do we stand?
• Stage II overview and trends
• Next steps

3
Overview

• Where do we stand?
• Stage II overview and trends
• Next steps

4
POP Model Intercomparison Study

• 10 highly different models
• Models have different purposes and endpoints
• Study planned in three stages, start March 2002
  (TFMM meeting Geneva)
• Stage I individual phase transfer processes
• Stage II mass balances and concentration and
  deposition fields sensitivity analysis
• Stage III persistence and long-range transport
  potential
• Three expert meetings in Moscow (20022005)
• Current status stage I finished, stage II
  partly finished, stage III started

5
Participating Models

• ADEPT/LOTOS (Netherlands)
• ADOM-POP (Germany)
• CAM/POPs (Canada)
• CliMoChem (Switzerland)
• DEHM-POP (Denmark)

• EVN-BETR and
• UK-MODEL (UK)
• G-CIEMS (Japan)
• HYSPLIT 4 (USA)
• MSCE-POP (MSC-E)
• SimpleBox (Netherlands)

6
Results from Stages I and II

• Reports Shatalov et al. 2004 and 2005

7
Results from Stages I and II

• Stage I
• Model descriptions
• Selection and treatment of chemical properties
• Process descriptions
• Stage II
• Masses in environmental compartments
• Mass fluxes for degradation, transport, and phase
  exchange
• Concentrations at interfaces

8
Overview

• Where do we stand?
• Stage II overview and trends
• Next steps

9
Computational Experiments for Stage II

• EMEP region as model domain (35N70N,
  10W30E)
• Calculations for PCBs 153 (and 28, 180) in 2000,
  four scenarios
• Reference chemical properties (Li et al.)
• Own chemical properties (various sources)
• Zero initial concentrations
• Initial concentrations from historical emissions
  (Breivik et al.)

sensitivity study
10
Results from Stage II What Do They Show?

• Large body of data
• General conclusion most models agree in most
  respects (Shatalov et al. 2005, p. 265268)
• However
• Agreement not always necessary we can learn
  from discrepancies!
• Many discrepancies at more detailed levels
• Why?

11
Results from Stage II What Do They Show?

• Large body of data
• General conclusion most models agree in most
  respects (Shatalov et al. 2005, p. 265268)
• However
• Agreement not always necessary we can learn
  from discrepancies!
• Many discrepancies at more detailed levels
• Why?

12
Discrepancies in Stage II (I)

• Example G-CIEMS and CliMoChem

mass in troposphere (10 km), reference properties
mass in troposphere (10 km), own properties
G-CIEMS
CliMoChem
large discrepancy (factor of 10)
large discrepancy (factor of 5 to 6)
13
Discrepancies in Stage II (II)

• Example G-CIEMS and CliMoChem

mass in soil (10 cm), own properties
mass in water (200 m), own properties
G-CIEMS
CliMoChem
no strong discrepancy (factor of 2 to 3)
no strong discrepancy (factor of 1.5)
14
Discrepancies in Stage II (III)

• Example G-CIEMS and CliMoChem

mass degraded in troposphere (10 km), reference
properties
mass degraded in troposphere (10 km), own
properties
G-CIEMS
CliMoChem
both models with highest mass losses, but
different by a factor of 2 to 3
CliMoChem similar to other models G-CIEMS with
different shape
15
Discrepancies in Stage II (IV)

• Example G-CIEMS and CliMoChem

outflow out of model domain in troposphere (10
km), reference properties
outflow out of model domain in water (200
m), reference properties
CliMoChem
G-CIEMS
CliMoChem with lowest values, G-CIEMS with high
values (factor of 5 to 15)
CliMoChem (and MSCE-POP) low G-CIEMS higher by
factor of 10
16
Outflow out of Model Domain

• Example G-CIEMS and CliMoChem

G-CIEMS with one box for the model
domain outflow in all directions
17
Outflow out of Model Domain

• Example G-CIEMS and CliMoChem

CliMoChem with latitudinal zones outflow only in
N and S directions
18
Discrepancies in Stage II (V)

• Example G-CIEMS and CliMoChem

net deposition from troposphere to
soil, reference properties
net deposition from troposphere to
water, reference properties
G-CIEMS
CliMoChem
CliMoChem low, shape similar to other
models G-CIEMS with different shape
both models with (very) low deposition G-CIEMS
with different shape
19
Results from Stage II What Do They Show?

• Large body of data
• General conclusion most models agree in most
  respects (Shatalov et al. 2005, p. 265268)
• However
• Agreement not always necessary we can learn
  from discrepancies!
• Many discrepancies at more detailed levels
• Why?

20
Next Steps (I)

• Reports on stages I and II provide valuable basis
  for developing research questions
• two models in comparison
• individual models in comparison to all other
  models
• Concept for a systematic analysis of similarities
  and discrepancies
• priority of questions
• focal points of interest
• development of hypotheses about model behavior

21
Next Steps (II)

• Answers to research questions
• perform additional calculations as experiments
  to test hypotheses
• use consistent chemical properties
• use simple and most consistent release pattern
• exclude influence of different model geometry
• This will constitute part 2 of stage II of the
  POPs Model Intercomparison Study (next modelers
  meeting Saturday, October 15).

22
Next Steps (III)

• Stage III
• Use reference chemicals from OECD model
  comparison study
• aldrin, atrazine, biphenyl, p-cresol, CCl4,
  a-HCH, PCB-28, PCB-153, PCB-180, HCB
• and also
• BDE-47, BDE-99, BaP, hexachlorobutadiene
• Rank reference chemicals according to Pov and
  LRTP in all models
• Again analyze reasons for differences

23
(No Transcript)