Air Quality Prediction in Beijing - Home - Molina Center...

Air Quality Prediction in Beijing

Liu WeidongBeijing Meteorological Bureau

Phone:86-10-88512940E-mail: [email protected]

Content

• 1. Introduction• 2. The Prediction Model of Air quality• 3. Application of the Model in Beijing• 4. Conclusion

1. Introduction

• 1.1 Main Pollutant in Beijing• 1.2 Air Quality Measurement Network

in Beijing

1.1 Main Pollutant

• PM10, the most serious pollutant • SO2, (Winter)• No2 (Winter)• O3 (Summer)• CO(Winter)

1.2 Air Quality Measurement Network• There are ten Measurement Sites in Beijing

2. Study and Prediction Model of Air quality

• Atmospheric boundary layer and air quality evaluation;

• Relationship between atmospheric pollution , meteorological condition and urban layout and development.

• Air pollution forecast model and system

2.1 The Method of Air Pollution Forecast

• Potential Forecast• Statistical Forecast• Numerical Forecast• Experiential correction

Air Pollution Potential Forecast

• Relation between pollution concentration and the surface’s field of pressure

• Analysis of correlation between pollutant and each weather condition.(surface meteorological element,weather phenomenon, upper meteorological element)

The statistical Method • Select the factor for the forecast. (850hPa

Temperature, relative humidity, surface weather situation,surface wind,upper wind,inversion,mixture height etc.)

• Get a statistical equations (stepwise regression analysis) in different seasons.

• Forecast main Pollutant’s Concentration. (SO2、CO、NOX、PM10、O3)

• City Air Pollution Prediction System (CAPPS)• Based on the result of MM5 numerical weather

forecast products.• Rely on the previous real pollutant data.• Output the air pollution index of SO2、NOX、

CO、PM10 for the next day.

Numerical Forecast for Air Pollution

Flow chart of numerical forecasting technology for urban air pollution

Mesoscale Numerical Weather Forecast Model

Meteorological elements within boundary

Box model of air pollutants diffusion(CAPPS)

Initial Pollutants’Concentration

Air pollutants’concentration

Air pollution index

Single city CAPPS system scheme

Pollutant concentation at start time

Numerical Forecast physical field

PPI index API index

Meso-Scale Meteorological Mode System MM5

Air Pollutant Advection

diffusion box mode

Visulization

Multi-cities CAPPS Scheme

City air pollution prediction numerical Model

CAPPS

National meteorological Center meso-scale numerical

prediction product

Pollutant concentration of

47 cities

Prediction Pollution index of

47 cities

Whole country air pollutant potential

Index forecast

Results verified system

Plot and output

Transmitted pollution index prediction in situ at different place

City area coefficient

Introduction of CAPPS• Pollution Index and Potential Prediction (by

CAPPS model)Establishment of Advective Diffusion Equation and Grid-cell Prediction Model

• The advective diffusion equation of airborne pollutants without considering chemical reactions can be written as:

ckvcvcrqcVtc

wdii ∇⋅⋅∇+⋅∇−⋅∇−∑=∇⋅+∂∂ rrrrrr

)()()(δ

After averaging over each cell, ∂∂ τ

τ

τδ τ

ττ

ττ

τ

τ τ τ

( ) ( + )

ct

V c d

q r d cv cv d k c di i d w

+ ⋅ ∇

= − ∇ ⋅ + ∇ ⋅ ⋅ ∇

∫∫∫

∑∫∫∫ ∫∫∫ ∫∫∫

1

1 1 1

v

v v v vv

and using the virtual turbulent transport speed (theconcept from model “PIC”) vt = ( u u c c v v c c w w c ct t t= ′ ′ = ′ ′ = ′ ′/ , / , / ) the prediction equation of mean concentration can beobtained as

τ∂∂

( ) ct

Q c V V v v d stS d w= − + + + ⋅∫∫v v v

where

V c c V V v v d sC tS d w= + + + ⋅∫∫1 ( )

v v v

and

TT

c

eceVQc

δτ

δτ

V

0

V CC

)1(−−

+−=

b

c

g

f

a

dh

ekj+i

i+1 j ki j k

k+1i+1

i+1 k

ji k+1 i+1 j k+1

i j+1 k+1 j+1

j+11

b

c

h

f

1a

d

k+1

d

j+1

kj

i

δx

δy

δz e

g

i+1

The grid-cell in CAPPS model

Source strength Q

v v v vcV vd vw Vt

Initial measurments0

Metorological prediction model

Vv vd

v vwv v

Vt

t = (n

+1)

δT

c-di

stri

buti

on

At

C

By Inversion

c

Fig.4 The flowchart of the grid-cell model

At t he t i me t =nδT

At t he t i me t =( n+1) δT

TT

c

eceVQc

δτ

δτ

V

0

V CC

)1(−−

+−=

Air Pollution Potential Index and Trial Prediction of the Pollution Index

In order to assess the performance of the above calculation, theday to day predictions of daily mean SO2, TPS and NOx concentrations were made at Shanghai on August-December 1998. The number of days with both predicted and measuredconcentrations is 118 and the comparison and correlation analysis were madebetween PSI predictions and the corresponding measurements. PSI index is defined as:

PSI=( ionconcentrat standard national ionconcentratpollutant measuredor predicted

)× 100

0 50 100 150 2000

50

100

150

200

Fig5. The comparis sion o f predition and measr ements of so2 pollution indexPrediction

Mea

surm

ents

0 20 40 60 80 100 12010

20

30

40

50

60

70

80

Mea

surm

ents

Prediction

Fig6. The comparission of pred ition and measrements of TSP pol lution index

FI G. 8 SO2 pr edi ct i on( sol i d bl ue) and measur ment s( vi r t ual r ed)

0

20

40

60

80

100

120

t i me days

PSI

0

50

100

150

200

TI ME Days

PSI

FI G. 9 TSP pr edi ct i on( sol i d bl ue) and measur ment s( vi r t ual r ed)

0

100

200

300

400PSI

FIG.8 NOx prediction(solid blue)and measurments(virtual green)

time days

Application of the Model in Beijing

Beijing Institute of meteorology sciences

SO2相关系数：0.89 准确率：94%

0

20

40

60

80

100

120

140

160

180

12-13 12-18 12-23 12-28 1-03 1-08 1-13

实况

预报 SO2相关 y = 1.2672x - 30.17

R2 = 0.7985

0

50

100

150

200

0 50 100 150 200

Fig 9 The correlation coefficient between measurements and prediction of SO2(1999/12/13-2000/1/13)

Winter

NO2相关系数：0.82 准确率：68%

0

20

4060

80

100

120

140160

180

200

12-13 12-18 12-23 12-28 1-03 1-08 1-13

实况

预报

NO2相关 y = 1.2295x - 27.056

R2 = 0.6786

0

50

100

150

200

0 50 100 150

Fig 10 The correlation coefficient between measurements and prediction of NO2(1999/12/13-2000/1/13)

Winter

RSP相关系数：0.90 准确率：65%

0

50

100

150

200

250

12-13 12-18 12-23 12-28 1-03 1-08 1-13

实况

预报

rsp相关 y = 0.8392x - 10.979

R2 = 0.8047

0

50

100

150

200

250

0 50 100 150 200 250

Fig 11 The correlation coefficient between measurements and prediction of PM10 (1999/12/13-2000/1/13)

Winter

CO相关系数：0.90 准确率：78%

0

20

40

60

80

100

120

140

160

180

12-13 12-18 12-23 12-28 1-03 1-08 1-13

实况

预报

CO相关 y = 1.0596x - 5.2822

R2 = 0.8105

0

50

100

150

200

0 20 40 60 80 100 120 140

Fig 12 The correlation coefficient between measurements and prediction of CO (1999/12/13-2000/1/13)

Winter

二氧化硫

准确率96.6% 相关系数0.94

0

10

20

30

40

50

60

6-11 6-16 6-21 6-26 7-1 7-6 7-11 7-16 7-21 7-26 7-31 8-5 8-10

SO2实测

S预300

二氧化硫 R=0.94 R2 = 0.8834

0

10

20

30

40

50

60

0 10 20 30 40 50 60

Fig13 The correlation coefficient between measurements and prediction of SO2

(2000/06/11-2000/8/10)

Summer

二氧化氮准确率75.9% 相关系数0.88

0

20

40

60

80

100

6-11

6-13

6-15

6-17

6-19

6-21

6-23

6-25

6-27

6-29

7-1

7-3

7-5

7-7

7-9

7-11

7-13

7-15

7-17

7-19

7-21

7-23

7-25

7-27

7-29

7-31

8-2

8-4

8-6

8-8

8-10

8-12

NO2实测

N预300

二氧化氮 R=0.88 R2 = 0.7719

0

20

40

60

80

100

0 20 40 60 80 100

Fig14 The correlation coefficient between

measurements and prediction of NO2 (2000/06/11-

2000/8/10)

Summer

可吸入颗粒物

准确率46.6% 相关系数0.96

050

100150200250300350

6-11

6-13

6-15

6-17

6-19

6-21

6-23

6-25

6-27

6-29

7-1

7-3

7-5

7-7

7-9

7-11

7-13

7-15

7-17

7-19

7-21

7-23

7-25

7-27

7-29

7-31

8-2

8-4

8-6

8-8

8-10

8-12

RSP实测

R预300

可吸入颗粒物 R=0.96 R2 = 0.9123

0

50

100

150

200

250

300

0 100 200 300 400


measurements and prediction of PM10 (2000/06/11-

2000/8/10)

Summer

一氧化碳

准确率100% 相关系数0.87

0

10

20

30

40

50

6-11

6-13

6-15

6-17

6-19

6-21

6-23

6-25

6-27

6-29

7-1

7-3

7-5

7-7

7-9

7-11

7-13

7-15

7-17

7-19

7-21

7-23

7-25

7-27

7-29

7-31

8-2

8-4

8-6

8-8

8-10

8-12

O3实测

C预300

一氧化碳 R=0.87 R2 = 0.7636

0

10

20

30

40

50

0 10 20 30 40 50


measurements and prediction of CO (2000/06/11-

2000/8/10)

Summer

二氧化硫

准确率：87% 相关系数：0.95

0

50

100

150

0922

0925

0928

1002

1005

1008

1011

1014

1017

1020

1031

1103

1106

1109

1112

1117

1122

1125

日期

指数

SO实况

SO预报

本图为北京市气象科学研究所（污染预报业务单位）提供

The correlation coefficient between measurements and prediction of SO2

(2000/09/22-2000/11/25)Fall

The correlation coefficient between measurements and prediction of NO2

(2000/09/22-2000/11/25)

二氧化氮准确率：79% 相关系数：0.76

0

50

100

150

200

0922

0925

0928

1002

1005

1008

1011

1014

1017

1020

1031

1103

1106

1109

1112

1117

1122

1125

日期

指数

NO实况

NO预报

123


Fall

The correlation coefficient between measurements and prediction of PM10

(2000/09/22-2000/11/25)

可吸入颗粒物

准确率：72% 相关系数：0.93

0

50

100

150

200

250

300

092209

2509

2810

0210

0510

0810

1110

1410

1710

2010

3111

0311

0611

0911

1211

1711

2211

25

日期

指数

PM10实况

PM10预报


Fall

2001/Oct/ The comparison between the prediction by statistic approach and Capps

4. Conclusion remark

For prediction of urban air quality, thesimple grid-cell Prediction Modelwithout the requirement of emissioninventories is helpful.

Study in the future• The mechanism and regulating principle

about air,water and soil pollution in Beijing city and it’s neighbor

• The operational system study of air pollution forecast (especially the heavy air pollution pre-warning system)

• Photochemical model• Need high resolution atmospheric -

chemistry model ---emission inventory

Air Quality Prediction in Beijing - Home - Molina Center...

Documents

Transcript of Air Quality Prediction in Beijing - Home - Molina Center...