Performance of the MATCH model in EURODELTA CITYDELTA, EURODELTA workshop, Matera, 20031029 Joakim L

1
Performance of the MATCH model in EURODELTA
CITY-DELTA, EURODELTA workshop, Matera,
2003-10-29Joakim Langner, Robert Bergström and
ValentinFoltescuSwedish Meteorological and
Hydrological Institute joakim.langner_at_smhi.se,
www.smhi.se

• The MATCH CTM
• Global statistics for 1999
• Interpretation of model results

2
Status of delivery of results

• EURODELTA - regional model inter-comparison
• Base case (1999) delivered
• Scenarios (7) produced but not delivered yet
• All concentrations are diagnosed to 3 m
• TNO-EMEP - Regional model inter-comparison for
  three different years
• Base case (1999) delivered
• 1997and 2001 could be run if additional models do
  this

3
Required output matrices (1999)
Variable Output Frequency Hourly
Daily Monthly O3, NO2,HNO3, H2O2,NO,HCHO,CO
X PPM2.5, PPM10, PM2.5,PM10,AMMO,SULF,NITR, SO
2,NH3,HNO3,NA,CL,NMVOC,C2H2,C5H8,HCHO
X Black Carbon, ORG (if possible) X Wet
depos. for SOx, NOx, NHx X Dry depos.
for SOx, NOx, NHx (total grid area) X Dry
depos. for SOx, NOx, NHx (forest area)
X Total depos. (drywet) for SOx, NOx, NHx
X
4
MATCH (Multi-scale Atmospheric Transport and
Chemistry model)Robertson et al. (1999) JAM

• 3D, Eulerian, limited area, off-line, h- or
  s-coordinates in the vertical
• 84x106 points, 44x44 km, rotated lat-long grid
  used in EURODELTA
• 14 levels reaching 6 km in EURODELTA
• meteorology, HIRLAM analysis, 3and 6h forecast
  (clouds and precip)
• ECMWF, ALADIN
• Transport
• Bott-type advection scheme, different orders can
  be applied
• fifth order scheme horizontally and second order
  vertically
• implicit vertical diffusion
• optional time interpolation of input met. data
• optional adjustment to get consistent 3D wind
  fields
• optional diagnostic calculation of boundary layer
  parameters
• optional subgrid scale vertical transport in
  convective clouds (not used in EURODELTA)
• Emissions (point and area sources), coupled to
  dynamic emission data bases
• submodel for point sources (not used in
  EURODELTA)
• EURODELTA time factors, vertical distribution
  using sector emissions
• biogenic (C5H8) E94, Simpson et al.

5
MATCH (Multi-scale Atmospheric Transport and
Chemistry model)Robertson et al. (1999) JAM

• Dry deposition
• resistance approach, optional number of land use
  classes (4 in EURODELTA)
• size dependent deposition for particles (Zhang)
• coupling to soil moisture (ozone in EURODELTA)
• Wet deposition
• simple - based on scavenging coefficients and
  2D precipitation
• standard - based on 3D precipitaion and
  cloudiness - including aqeous phase chemistry
• size dependent scavenging of particles
• Chemistry
• parameterized 10 component chemistry for sulfur,
  oxidized and reduced nitrogen
• photochemistry with 60 components (based on
  EMEP)EMEP acid deposition chemistry
• KPP used to generate solver for chemistry
• Particulates
• size-resolved sea salt (not reported in
  EURODELTA)
• primary anthropogenic PM (4 classes)
• bulk scheme for other inorganics, coupled scheme
  under development
• Fully parallellised using RSL, FLIC and MPI. Runs
  on Linux clusters, SGI-Origin, IBM, True64, HP
  etc.

6
Global statistics for 1999, EURODELTA emissions
7
Ozone 1999
8
Ozone 1999
9
Ozone 1999
10
Ozone 1999
11
Global statistics for 1999, EURODELTA emissions
12
Global frequency distributions SO2, nssSO4 1999
13
Global statistics for 1999, EURODELTA emissions
14
Global frequency distributions NO2,
HNO3NO3-part. 1999
15
Global statistics for 1999, EURODELTA emissions
16
Global frequency distributions NH3NH4 part,
NH4-part. 1999
17
Global statistics for 1999, EURODELTA emissions
18
Global frequency distributions wet deposition 1999
19
MATCH-PMPM10 sourcecontributionsannual avg.
1999
20
MATCH-PMNumber concentrationsannual avg. 1999
21
Global statistics for 1999, EURODELTA emissions
22
Global frequency distributions PM10, SIA 1999
23
Summary

• Ozone
• Positive bias
• Suggested improvements dry deposition, improved
  landuse improved simulation of
  NOx surface diagnosis? reduce correlation!
• NOx (NO2, HNO3, NO3-part, wd-NOx)
• Negative bias
• Suggested improvements dry deposition improve
  d gas-aerosol scheme
• SOx (SO2, nssSO4, wd-SOx)
• Slight positive bias
• Suggested improvements dry deposition of
  SO2 cloud chemistry (mixing)
• NHx
• Negative bias (NH3NH4-part, wd-NHx)
• Suggested improvements seasonal variation of
  emissions
• PM10
• negative bias
• Suggested improvements add more
  sources better understanding of chemical
  composition better emission data

24
Important observations

• Interpolation of emission data may have a
  negative impact on the results
• Outliers in the official EMEP data affect the
  statistics. One case, 1999-08-04, with extremely
  high observed values at DK03 reduce the global
  correlation for HNO3NO3 from 0.58 to 0.26. Other
  parameters for the same day and station are
  probably in error.

25
MATCH-PMseasaltannual avg. 1999
26
MATCH 3D transport/chemistry/deposition model
3D meteorological data, HIRLAM, ECMWF, MM5,
Aladain...
Emission data e.g. EMEP-2000
3D transport/chemistry model, MATCH
Chemical mechanism, KPP syntax
Land use - forest
27
Applications

• Operational emergency response (nuclear
  accidents) SMHI, EU-RODOS-system
• Operational national and Nordic mapping of
  deposition (S, N, base cations)
• Large scale chemical modelling, deposition - long
  term as well as operational forecasts
• Climate impact on air pollution budgets
• Aerosol modelling (building up)
• Air quality, urban scale
• Atmospheric chemistry projects in developing
  countries (Asia, South America)
• List of publications at www.smhi.se. look under
  research and development/meteorology/amospheric
  chemistry