[PPT] - Development and application of a reactive plume-in-grid model: PowerPoint Presentation

SLIDE 1

Development and application of a reactive plume-in-grid model: evaluation over greater Paris

13th Harmo Conference

Irène Korsakissok1,2, Vivien Mallet1,3

1 CEREA, joint laboratory ENPC/EDF R&D, Paris-Est university, France 2 IRSN, Fontenay-aux-roses, France 3 INRIA, Paris-Rocquencourt research center, France

1-4 June 2010 - Paris

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 1 / 18

SLIDE 2

Subgrid-scale modeling of emissions

Outline

1

Subgrid-scale modeling of emissions Why use a subgrid model ? Model coupling Non-linear chemistry

2

Application over Greater Paris

3

Conclusions

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 2 / 18

SLIDE 3

Subgrid-scale modeling of emissions Why use a subgrid model ?

A wide range of scales

From µm (particles) to km (meteo) Gridded representation : usually from 1 to 50 km... Subgrid-scale phenomena : emissions, chemistry, clouds, land use, turbulence...

global scale continental scale regional scale local scale

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 3 / 18

SLIDE 4

Subgrid-scale modeling of emissions Model coupling

Model coupling within Polyphemus platform

Using Polyphemus modeling platform : modularity, easy coupling Plume-in-grid method : coupling an Eulerian model (Polair3D) and a Gaussian puff model to model point source emissions Puffs are “injected” into the Eulerian model after a given time (“injection time”)

Eulerian Model Puff Model

Plume-in- grid interface

Puff location, puff size Meteorological data in puff cell Meteorological data (wind, stability) Puff transfer

Saved concentrations: Eulerian + Gaussian

Eulerian concentrations Gaussian concentrations

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 4 / 18

SLIDE 5

Subgrid-scale modeling of emissions Model coupling

Model coupling within Polyphemus platform

real plume Gaussian plume model Gaussian puff model Wind u uΔtpuff

2σz 2σx

2σy

tinj: injection in the Eulerian model

Puffs size given by standard deviations σx, σy, σz (similarity theory, Briggs) ∆tpuff time step between two puffs’ emissions tinj injection time (puff “lifetime”)

tinj ∆tpuff : total number of puffs handled

by the model for one continous source Reference : Korsakissok, I. et Mallet, V. (2010). Subgrid-scale treatment for major point sources in an Eulerian model : A sensitivity study on the European Tracer Experiment (ETEX) and Chernobyl cases. Journal of Geophysical Research. 115 :D03303.

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 4 / 18

SLIDE 6

Subgrid-scale modeling of emissions Non-linear chemistry

Reactive plume-in-grid model

Phase 1: photostationary state NO/NO2/O3 Phase 2: acid formation through OH and NO3 Phase 3: full chemistry acid and ozone production

Advantages of subgrid model

Better representation of local-scale diffusion Source height and plume rise Near-source chemistry

Chemistry within puffs

The species in one puff α react with each other The species in two overlapping puffs α and β react with each other The species in one puff react with the background species (from the Eulerian model)

Vα Vαβ Vβ

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 5 / 18

SLIDE 7

Subgrid-scale modeling of emissions Non-linear chemistry

Chemistry between puffs and background species

A + B

k

− → P cα

A , cα B puff

cb

A, cb B

background d(cα

A + cb A)

dt = −k(cα

A cα B puff

+ cb

A cb B background

+ cα

A cb B + cα B cb A

interaction

) dcb

A

dt = −kcb

A cb B

background chemistry (Eulerian) dcα

A

dt = d

cα

A + cb A

dt

− dcb

A

dt puff=background perturbation

zone titration

O3 + NO

k

− → NO2 + O2 plume of NOx (NO+NO2) uniform background of O3 → Decrease of in-plume O3 concentration

10 20 30 40 50 60 70 Time after emission (minutes)

6
4
2

2 4 6 8

plume mass (micrograms)

x1e+10

NO2 mass NO mass O3 mass

Evolution of in-plume mass for several species (µg) in a continuous plume of NOx emitted within a background of O3.

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 6 / 18

SLIDE 8

Application over Greater Paris

Outline

1

Subgrid-scale modeling of emissions

2

Application over Greater Paris Spatial impact of subgrid-scale modeling Results on measurement stations Sensitivity study

3

Conclusions

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 7 / 18

SLIDE 9

Application over Greater Paris

Issues

1

What is the impact of a subgrid-scale modeling of point emissions on regional photochemistry ?

2

Impact on primary vs secondary species ?

3

Impact on results over six months vs particular days ?

4

Sensitivity to local-scale modeling ? Reference : Korsakissok, I. et Mallet, V. (2010). Development and application

f a reactive plume-in-grid model : Evaluation over Greater Paris.

Atmospheric Chemistry and Physics Discussions 10, 5091-5134

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 8 / 18

SLIDE 10

Application over Greater Paris

s23 s32 s25 s24 s23 s25 s24 s15

Point sources ( ) and measurement stations (rural and urban ). Left : SO2, right : NO. The circle diameters are proportional to the sources emission rates.

Ile-de-France (Paris region), summer 2001, six months Meteorological fields from ECMWF (0.36◦ resolution) Full gaseous chemistry (RACM mechanism) Polair3D (0.05◦ resolution) with/without subgrid modeling (similarity theory, tinj = 20 min, ∆tpuff = 100 s) 89 point sources : Qs > 106 µg s−1 for NOx (20% of total emissions) or SO2 (55% of total emissions)

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 9 / 18

SLIDE 11

Application over Greater Paris Spatial impact of subgrid-scale modeling

Spatial impact of subgrid-scale modeling

1.5 2.0 2.5 3.0 3.5 48.2 48.4 48.6 48.8 49.0 49.2 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

SO2

1.5 2.0 2.5 3.0 3.5 48.2 48.4 48.6 48.8 49.0 49.2 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

NO

1.5 2.0 2.5 3.0 3.5 48.2 48.4 48.6 48.8 49.0 49.2 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

NO2

1.5 2.0 2.5 3.0 3.5 48.2 48.4 48.6 48.8 49.0 49.2

4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5

O3

Differences in mean ground concentrations : Polair3D - plume-in-grid. Concentrations averaged over six months (µg m−3).

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 10 / 18

SLIDE 12

Application over Greater Paris Spatial impact of subgrid-scale modeling

Spatial impact of subgrid-scale modeling during a low-dispersion day (sulfur dioxide)

20.1
17.6
15.1
12.6
10.1-7.6 -5.1 -2.6 -0.1 2.4

4.9 7.5 10.0 12.5 15.0 17.520.0 22.5 25.0

Differences in hourly-averaged SO2 ground concentrations : Polair3D - plume-in-grid (µg m−3), for day 2001-08-24 between 03 and 15h (local hour).

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 11 / 18

SLIDE 13

Application over Greater Paris Spatial impact of subgrid-scale modeling

Spatial impact of subgrid-scale modeling during a low-dispersion day (ozone)

12.9
12.0
11.1
10.1-9.2 -8.2 -7.3 -6.4 -5.4 -4.5 -3.5 -2.6 -1.6 -0.7 0.2

1.2 2.1 3.1 4.0

Differences in hourly-averaged O3 ground concentrations : Polair3D - plume-in-grid (µg m−3), for day 2001-08-20 between 03 and 15h (local hour).

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 12 / 18

SLIDE 14

Application over Greater Paris Results on measurement stations

Results on stations for SO2 and NO

RMSE = v u u t 1 n

n

X

i=1

(xi − yi )2, with xi simulated values, yi observed values.

Polair3D, plume-in-grid Black % : urban stations Green % : periurban and rural stations.

5 10 15 20 25

10.3%
15.0%
6.2%
8.1%
12.5%
11.8%
8.9%
13.1%
16.2%
9.4%
11.6%

6.2%

8.4%
17.4%
13.6%
14.2%
0.8%
12.2%
10.7%

SO2 – Mean observed values : 6.2 µg m−3 10 20 30 40 50

4.2%
3.3%
5.9%
4.4%
4.1%
6.9%
3.6%
4.9%
2.8%
4.6%
4.5%
4.5%
4.1%
3.2%
5.0%
5.2%
1.3%
6.1%
6.7%
3.0%
2.2%
0.3%
5.0%

NO – Mean observed values : 10.42 µg m−3

Comparison to observations on measurement stations, over six months. Mean and RMSE in µg m−3.

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 13 / 18

SLIDE 15

Application over Greater Paris Results on measurement stations

Results on stations for SO2 and NO

RMSE = v u u t 1 n

n

X

i=1

(xi − yi )2, with xi simulated values, yi observed values.

Polair3D, plume-in-grid Black % : urban stations Green % : periurban and rural stations.

5 10 15 20 25

10.3%
15.0%
6.2%
8.1%
12.5%
11.8%
8.9%
13.1%
16.2%
9.4%
11.6%

6.2%

8.4%
17.4%
13.6%
14.2%
0.8%
12.2%
10.7%

SO2 – Point sources : 55% of total emissions 10 20 30 40 50

4.2%
3.3%
5.9%
4.4%
4.1%
6.9%
3.6%
4.9%
2.8%
4.6%
4.5%
4.5%
4.1%
3.2%
5.0%
5.2%
1.3%
6.1%
6.7%
3.0%
2.2%
0.3%
5.0%

NO – Point sources : 20% of total emissions

SO2 more impacted (more point sources) than NO, urban/rural stations equally impacted

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 13 / 18

SLIDE 16

Application over Greater Paris Results on measurement stations

Results on stations for NO2 and O3

RMSE = v u u t 1 n

n

X

i=1

(xi − yi )2, with xi simulated values, yi observed values.

Polair3D, plume-in-grid Black % : urban stations Green % : periurban and rural stations.

10 15 20 25 30 35

0.2% 2.0% 3.8% 0.7% 2.8%

0.5%
0.1%

1.1% 1.3% 1.9% 1.8%

0.1%

0.7% 0.5% 6.2% 1.6% 0.5%

0.2%

2.6%

1.2%
0.4%
0.4%
0.8%

0.2%

0.9%
0.5%

NO2 – Mean observed values : 34.64 µg m−3 20 25 30 35 40 45

1.8%
0.9%
2.2%
2.4%
1.4%
2.8%
2.2%
2.3%
2.0%
1.0%
1.7%
1.3%
0.2%
0.5%
1.2%
0.2%
1.1%
0.2%
0.7%
0.4%
0.7%

O3 – Mean observed values : 56.87 µg m−3

Comparison to observations on measurement stations, over six months. Mean and RMSE in µg m−3.

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 14 / 18

SLIDE 17

Application over Greater Paris Results on measurement stations

Results on stations for NO2 and O3

RMSE = v u u t 1 n

n

X

i=1

(xi − yi )2, with xi simulated values, yi observed values.

Polair3D, plume-in-grid Black % : urban stations Green % : periurban and rural stations.

10 15 20 25 30 35

0.2% 2.0% 3.8% 0.7% 2.8%

0.5%
0.1%

1.1% 1.3% 1.9% 1.8%

0.1%

0.7% 0.5% 6.2% 1.6% 0.5%

0.2%

2.6%

1.2%
0.4%
0.4%
0.8%

0.2%

0.9%
0.5%

NO2 – Mean observed values : 34.64 µg m−3 20 25 30 35 40 45

1.8%
0.9%
2.2%
2.4%
1.4%
2.8%
2.2%
2.3%
2.0%
1.0%
1.7%
1.3%
0.2%
0.5%
1.2%
0.2%
1.1%
0.2%
0.7%
0.4%
0.7%

O3 – Mean observed values : 56.87 µg m−3

Primary species (NO, SO2) more impacted than secondary species (NO2, O3)

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 14 / 18

SLIDE 18

Application over Greater Paris Sensitivity study

Sensitivity study

SO2

b a s e b r i g g s d t p u f f = 6 s t i n j = 4 m i n s 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 RMSE difference

NOx

b a s e b r i g g s d t p u f f = 6 s t i n j = 4 m i n s 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 RMSE difference

Base case : similarity theory, tinj = 20 min and ∆tpuff = 100 s

Alternative cases

sigma parameterization : Briggs ∆tpuff = 600 s tinj = 40 min O3

b a s e b r i g g s d t p u f f = 6 s t i n j = 4 m i n s 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 RMSE difference

Differences (Polair3D - plume-in-grid) in RMSE (µg m−3), computed on all stations and six months for SO2, NOx and O3.

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 15 / 18

SLIDE 19

Conclusions

Outline

1

Subgrid-scale modeling of emissions

2

Application over Greater Paris

3

Conclusions

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 16 / 18

SLIDE 20

Conclusions

Summary and conclusions

1

Full non-linear gaseous chemistry implemented in plume-in-grid model

2

Spatial impact, especially during low-dispersion days

3

Impact of plume-in-grid model on averaged statistics is limited by :

◮ limited amount of emissions from point sources compared to traffic

(except for SO2)

◮ averaging effect (smoothing spatial variability) ◮ stations locations (background stations) 4

However, significant improvement is shown for primary species

5

O3 sensitive to time step between two puffs, primary/less-reactive species (SO2, NOx) sensitive to injection time

Future work

1

Handling chemistry for particulate matter

2

Extension to line sources and application to road emissions

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 17 / 18

SLIDE 21

Conclusions

Thank you for your attention

I. Korsakissok (CEREA/IRSN)

1-4 June 2010 18 / 18