Urban air quality and regional haze weather forecast in Yangtze...

Urban air quality and regional haze weather forecast in Yangtze River

Delta of China

Wang Tijian Jiang Fei Deng Junjun Shen YiSchool of Atmospheric Science，Nanjing University

Fu Qinyan Wang QianShanghai Environmental Monitoring Center

Zhang Danning Fu Yin Xu JinahuaNanjing Environmental Monitoring Center

Dec.12,2012, Geneva

Outline

Air pollution in Yangtze River Delta

Urban air quality and regional haze weather forecast model

Air quality forecast in urban

Haze weather forecast in YRD

Summary

Yangtze River Delta(YRD)

Located on the western coast of the Pacific Ocean and covers an area of about 99,600 km2, with a population of 75 million.

Regional haze

Photochemical smog

Acid rain

Dust storm

PM10

0

0.1

0.2

0.3

0.4

0.5

0.6

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85

PM10

Heavy haze case 2008.10.27-29

Dust storm occurred in Mar.20,2010

Continuous air pollution for more than 10 days

chemical weather forecasting (CWF) modelsin the word

On line model: SKIRON/Dust, WRF-Chem, ENVIROHIRLAM, TAPM

Off line model:ALADINCAMx, CAMx-AMWFG, EURAD-RIU, FARM, LOTOSEUROS, MATCH, MM5-CAMx, MM5-CHIMERE,MM5/WRF-CMAQ, MOCAGE,NAME, OPANA,RCG, SILAM, THORRegAEMS, NAQPS, PATH,TAQM

Urban air quality forecast system

Display and Release to Website

WRF Preprocessing System

WRF-Chem Model

Products Output

Emission Inventory

Meteorological initial and Boundary conditions

Chemical initial and Boundary conditions

Previous forecast results3-D grid chemical concentrations

Automatic Download GFS Data

Emission Preprocessing Model

Weather processes simulation

Transport and diffusion

Dry and wet deposition

Chemical reactions and aerosol process

Air Pollution Index

Hourly surface concentrations distribution, SO2, NO2, O3, PM10

Hourly concentrations at selected sites

WRF-CHEM

Regional haze weather forecast system

RegAEMS

Different aerosol schemes in RegAEMS

SO42-, NO3

-:ISORROPIA( Nenes,1998 )EC,OCSOA: SORGAM(Schell,2001)Sea salt: Monahan（1986）Dust: Gillette(1988)

ext scat abs ag sp apb b b b b b= + + + +

/ ext

Visibility scheme

K bν =

( )( )4 4 4 32

2

3 ( )

4 100.6 0.175 10

extb f RH NH SO NH NO

Organics Soot SoilCoarsemass NO

= • + +

+ + ++ +

(K is 3.912)

Haze level classfication

PM2.5>75μg/m3

RH<0.8

Visibility<10km

Super heavy haze：Visibility<2km

Heavy haze：2km≤Visibility<4km

Medium haze：4km≤Visibility<7km

Light haze：7km≤Visibility<10km

Air quality and haze weather forecast step

Today Tomorrow The day after tomorrow

00:00 20:00 12:00 12:00

Spin up Day one forecast ……

……

……

16 24 24

48 hours forecast with 16 hours spin up

Model settings for Shanghai forecastHorizontal grid: 4 nested domains

Domain 1: 88*75, 81kmDomain 2: 85*70, 27kmDomain 3: 76*67, 9 kmDomain 4: 88*73, 3 km

Vertical level: 24 sigma level

Model top: 100hpa

Gas phase mechanism: RACM

Aerosol: MADE/SORGAM

Anthro. Emi.: INTEX-B + Shanghai emission inventory

Natrual Emi.: calculated online by Guenther scheme

Point source in Shanghai

CO NOx SO2

VOC PM2.5 PM10

Area source in Shanghai

CO NOx SO2

VOC PM2.5 PM10

Line source in Shanghai

CO NOx SO2

VOC PM2.5 PM10

SO2 hourly concentration

各月份预报的平均情况预报平均 相关性 平均偏差

时段 观测平均24-hour 48-hour 24-hour 48-hour 24-hour 48-hour

5月 34.20 41.17 43.22 0.16 0.21 7.00 9.97 6月 31.20 34.86 32.31 0.34 0.32 4.29 1.70 7月 24.71 40.31 36.71 0.21 0.08 15.40 11.45 8月 21.01 35.25 41.45 0.16 0.19 14.46 20.30 9月 18.15 55.67 49.41 0.09 0.05 37.57 31.51 10月 28.11 60.25 61.94 0.19 0.17 32.04 33.63 11月 39.78 48.47 45.12 0.18 0.24 12.99 12.21 12月 60.43 51.32 60.06 0.27 0.20 -6.17 0.46 1月 48.99 45.29 48.64 0.19 0.21 -3.55 -0.35 2月 28.22 36.01 39.16 0.20 0.18 7.70 10.96 3月 41.93 33.52 33.12 0.17 0.24 -8.37 -8.71 平均 34.25 43.83 44.65 0.20 0.19 10.31 11.20

0.00 

0.05 

0.10 

0.15 

0.20 

0.25 

0.30 

0.35 

0.40 

‐10.0 

‐5.0 

0.0 

5.0 

10.0 

15.0 

20.0 

25.0 

30.0 

35.0 

40.0 

5月 6月 7月 8月 9月 10月 11月 12月 1月 2月 3月

相关

性

平均

偏差

偏差

相关性

Before  emi updated

After emi updated

NO2 hourly concentration

各月份预报的平均情况预报平均 相关性 平均偏差

时段 观测平均24-hour 48-hour 24-hour 48-hour 24-hour 48-hour

5月 54.39 49.09 52.37 0.33 0.34 -3.86 -1.26 6月 48.48 44.92 48.20 0.41 0.33 -3.08 3.44 7月 40.47 51.12 52.61 0.33 0.21 10.14 11.59 8月 32.28 45.28 50.51 0.22 0.42 14.41 19.25 9月 39.21 47.72 48.86 0.21 0.00 7.92 9.28 10月 56.85 50.59 54.13 0.39 0.42 -6.20 -2.67 11月 61.00 49.52 49.80 0.31 0.25 -12.32 -10.52 12月 67.15 52.45 56.69 0.24 0.23 -12.49 -8.57 1月 57.89 47.64 51.58 0.23 0.25 -10.17 -6.23 2月 43.23 44.98 46.40 0.28 0.21 1.74 3.22 3月 51.78 44.49 46.14 0.29 0.30 -7.15 -5.39 平均 50.25 47.98 50.66 0.29 0.27 -1.91 1.10

0.00 

0.05 

0.10 

0.15 

0.20 

0.25 

0.30 

0.35 

0.40 

‐10.0 

‐5.0 

0.0 

5.0 

10.0 

15.0 

20.0 

25.0 

30.0 

35.0 

40.0 

5月 6月 7月 8月 9月 10月 11月 12月 1月 2月 3月

相关

性

平均

偏差

偏差

相关性

PM10 hourly concentration

各月份预报的平均情况预报平均 相关性 平均偏差

时段 观测平均24-hour 48-hour 24-hour 48-hour 24-hour 48-hour

5月 81.91 60.69 65.73 0.11 0.14 -25.44 -22.56

6月 90.35 52.37 52.45 0.35 0.33 -35.92 -34.47

7月 56.85 66.68 75.29 0.17 0.10 9.35 18.06

8月 56.44 63.88 83.46 0.16 0.20 8.66 26.96

9月 51.35 68.93 68.42 0.06 -0.01 17.53 18.32

10月 92.98 81.70 83.90 0.25 0.26 -10.97 -8.83

11月 84.00 91.05 88.75 0.30 0.33 5.03 9.38

12月 119.61 108.57 128.66 0.27 0.28 -7.56 10.53

1月 90.67 95.93 107.80 0.31 0.37 5.46 17.23

2月 56.54 82.43 91.77 0.09 0.09 25.71 35.07

3月 87.16 81.76 86.09 0.10 0.13 -5.23 -0.73

平均 78.90 77.63 84.76 0.20 0.20 -1.22 6.27

0.00 

0.05 

0.10 

0.15 

0.20 

0.25 

0.30 

0.35 

0.40 

‐35.0 

‐25.0 

‐15.0 

‐5.0 

5.0 

15.0 

25.0 

35.0 

5月 6月 7月 8月 9月 10月 11月 12月 1月 2月 3月 相关

性

平均

偏差

偏差

相关性

Statistics on API prediction

‐15 

‐10 

‐5 

0 

5 

10 

15 

20 

0.00 

0.10 

0.20 

0.30 

0.40 

0.50 

0.60 

0.70 

0.80 

0.90 

May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr

Deviation

Accuracy & Correlatio

n

Month

Accuracy

Correlation

Deviation

Air quality forecast for 2nd Youngth Olympic game in Nanjing in 2014

DOMAIN 1: 88×75,    81 kmDOMAIN 2: 85×70,    27 kmDOMAIN 3: 70×64,       9 kmDOMAIN 4: 55×61,       3 km

Area source Point source

NOxNOx

PM10PM10

SO2SO2

Meteorological variables in Nanjing during Oct., 2007

Observed and predicted concentrations of PM10, SO2 and NO2 in Nanjing during Oct., 2007

Paved road dustConstruction dust

Biomass burning dustSoil dust

Haze weather forecast in YRD using RegAEMS

Regional Center:(32.06°N, 119.86°E)

Vertical level:Vertical: 10 levels

Horizontal Grid Size:DOMAIN 1: 88×75, 81 kmDOMAIN 2: 85×70, 27 kmDOMAIN 3: 70×64, 9 km

Prediction Period:2009.1.1—2009.12.31

Observed vs. predicted meteorological variables

Daily concentrations of PM10 and PM2.5 in YRD during 2009

Hourly visibility from observation and prediction in YRD during 2009

Daily RH, PM2.5, Visibilityand Haze level forecast

Heavy haze case prediction

Backward trajectory, fire spot and biomass emission

117E 118E 119E 120E 121E 122E

30N

31N

32N

33N

34N

Nanjing

Hefei

Shanghai

Hangzhou

Biomass burning

Fire spot

RegAEMS

10月28日18:00~20:00

Data Assimilation in dust prediction

Directly analyze 3D aerosol mass concentration with a one-step procedure of variational minimization within the GSI

Do NOT apply any assumption about vertical shape and relative weight of individual species.

14 WRF/Chem-GOCART 3D aerosol mass concentration as analysis variables

need background error covariance statistics for each aerosol species

Use CRTM as the AOD observation operator, including both forwardand Jacobian models

In short, no much difference from 3DVAR DA for meteorological obs.

38

J(x) =12

(x − xb )T B−1(x − xb ) +12

[y − H(x)]T R−1[y − H(x)]

OMB/OMA of MODIS AOD

40

41

SummaryWRF-CHEM and RegAEMS are useful tools for urban quality and regional haze forecast.

To improve the spatial resolution of emission inventory, especially emissions from fugitive dust, transportation and biomass burning in Yangtze River Delta region, to improve the performance of model prediction.

To include urban canopy model in the meteorological model to improve the performance of meteorology prediction.

To improve the model performance of forecast on heavy air pollution case using data assimilation and satellite spot, such as dust storm, and biomass burning.