Scale Dependency Analysis of Atmospheric Pollutants with EMEP - - PowerPoint PPT Presentation

Scale Dependency Analysis of Atmospheric Pollutants with EMEP MSC-W Model

Semeena Valiyaveetil Shamsudheen

Component NO3 in air ug/m3 HNO3 in air ugN/m3 Ammonia in air ugN/m3 Nitrate wd mgN/m2 Amm wd mgN/m2 Nitrate in pptn mgN/l Amm in pptn mgN/l

• No. of

Stns Obs Bias% 56 Bias% 28 Bias% 14 Bias% 7 Corrn 56 Corrn 28 Corrn 14 Corrn 7 22 1.92

• 11

1.72

• 12

1.70

• 13

1.68

• 15

1.67 .85 0.87 0.87 0.87 13 0.17

• 27

0.12

• 28

0.12

• 21

0.13

• 22

0.13 0.49 0.54 0.52 0.54 43 1.62

• 30

1.14

• 28

1.17

• 30

1.13

• 24

1.17 0.69 0.66 0.72 0.72 63 14690

• 7

13621

• 5

13937

• 6

13818

• 6

13798 0.80 0.77 0.78 0.77 62 17015

• 2

16624 2 17379 2 17416 4 17773 0.66 0.62 0.63 0.63 63 0.28

• 7

0.27

• 8

0.26

• 8

0.26

• 9

0.26 0.65 0.67 0.66 0.66 62 0.34

• 3

0.33

• 2

0.33

• 1

0.34 0.34 0.54 0.50 0.51 0.50

http://emep.int/publ/reports/2013/MSCW_technical_1_2013.pdf

Bias% = Model -Obs Obs ´100

Model Validation

Time series of Oxidised Nitrogen

Computer requirements

HP BL 460c Gen 8 cluster, 518 nodes, 32GB memory each --> ~3.30 hr for 50 x 50, 1 yr SGI Altix 8600, 1440 nodes, 32GB memory each

• -> ~3 hours for 50X50km, 1 yr.

EMEP Monitoring network and Data Uncertainties

EMEP Monitoring Network

Denmark - 12 Germany - 30 Poland - 5 Latvia - 3 Lithuania - 2 Estonia - 3 Russia - 7 Finland - 18 Sweden - 18

Data quality and uncertainities

• In the data quality objective (DQO) of EMEP measurement program,

there is a requirement that the measurements should be within 15-25% uncertainity for the combined sampling and chemical analysis. http://www.nilu.no/projects/ccc/qa/

• We have annual laboratory intercomparisons which check the

performance in the laboratory and these are generally very good, within 5-10% uncertainity for S and N components. However the main uncertainity is in the field, whether it is used bulk or wet only, collection efficiency (should be checked with parallel rainguages, but few sites report this), sampling intervals (higher uncertainity for longer periods, especially longer than weekly) etc. http://www.nilu.no/projects/ccc/intercomparison/info.html

Data quality and uncertainities

• To asess the real uncertainity in the field measurements, you need

parallel sampling, i.e., field intercomparison. CCC had conducted some of these in the past, at NO01 and PL05 the uncertainities in the wet deposition of SO4 and NO3 is better than 10% http://www.nilu.no/projects/ccc/qa/

• Finally it is upto the parties to report the data within the requirements

set by the programme. CCC do check the data quality by annual laboratory intercomparison and field intercomparison. Further a time series and ion balance plots were prepared to get a more subjective feeling of the data quality. http://www.nilu.no/projects/ccc/ionbal/index.html

• Before CCC used to write annual report on the data quality where also

the Parties reported on the uncertainties of their measurements. The last one was in 2006 http://www.nilu.no/projects/ccc/reports/cccr3-2008.pdf Wenche.aas@nilu.no