1 Surface O 3 over Beijing: Constraints from New Surface Observations Yuxuan Wang, Mike B. McElroy, J. William Munger School of Engineering and Applied.

1

Surface O3 over Beijing: Constraints from New Surface Observations

Yuxuan Wang, Mike B. McElroy, J. William MungerYuxuan Wang, Mike B. McElroy, J. William MungerSchool of Engineering and Applied Sciences, Harvard UniversitySchool of Engineering and Applied Sciences, Harvard University

Jiming Hao, Hong Ma, Yaosheng ChenJiming Hao, Hong Ma, Yaosheng Chen Tsinghua University, Beijing, ChinaTsinghua University, Beijing, China

GC 3rd Users Meeting, April 15, 2007

2http://www.fallingrain.com/world/CH/22/Miyun.html

Beijing

Miyun Station

3

Overview of Miyun Station

1.1. Established by the China Project; started Established by the China Project; started in Nov 2004, continuously to datein Nov 2004, continuously to date

2.2. Miyun (Miyun (40.3N, 116.8E40.3N, 116.8E): Location chosen to ): Location chosen to observe both clean continental air and observe both clean continental air and pollution in the Beijng plumepollution in the Beijng plume

3.3. Directly observe trace-gas covariances Directly observe trace-gas covariances near the source; near the source;

4.4. Species measured: CO, OSpecies measured: CO, O33, CO, CO2 2 ((NONOxx, , NONOyy, SO, SO22))

5.5. Data used here: 30-min averages Data used here: 30-min averages prepared by J.W. Mungerprepared by J.W. Munger

4

Sometimes It’s Clear

Courtesy of Bill Munger

5

Mean Flow Patterns at the surface

East Asian Monsoon

Jan July

6

Seasonal variations of other factors

Qiu et al., Atmos. Environ., 2000

Beijing AOD (1984-1994)

AOD

Model sulfate AOD peaks in July (0.5)

Precipitation

obs

model

7

O3 observations: monthly statistics

• distributions of O3 skewed toward high values in summer months

• O3 peak in early summer and early fall

• Summertime ozone pollution

daytimedaytime O3

80ppbv

Legend

variations in summer

8

Seasonality of Ozone at Lin An

Wang et al., 2002

obs: daytime mean (1999-2000)

model

9

CO observations : monthly statistics

• small seasonality in CO• distributions of CO close to bimodal: clean continental air clean continental air and local pollutionand local pollution

Legenddaytime CO

variations in summer

>3ppm CO!

10

GEOS-Chem Model Set-up

GEOS-4 global simulation; 2GEOS-4 global simulation; 2o o x 2.5x 2.5oo resolution; version 7-04-09resolution; version 7-04-09

Updated Chinese emissions of CO and Updated Chinese emissions of CO and NONOxx with monthly variability with monthly variability

Simulation period: Nov 2004 – Dec Simulation period: Nov 2004 – Dec 2006. Focus on 20062006. Focus on 2006

1-hr model outputs sampled by 1-hr model outputs sampled by observations. observations.

11

Chinese emissions of CO and NOx

NOx (2004) CO (2001)

Zhang, Streets, et al., 2007

China total: 17.8 TgNO2 (up from 11.3 for TRACE-P)

Streets, et al., 2006

China total: 146 Tg/yr (up from 100)

12

Monthly variations in Chinese emissions

0.040

0.060

0.080

0.100

0.120

0.140

0.160

0.180

0.200

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

All SourcesPowerIndustryHeating & Residential

NOx

(1.3: 1)

Zhang, Streets, et al., 2007

CO (1.5:1)

heating and residential

Seasonality not yet in the standard model

13

Soil NOx emissions

Wang et al., 2007

a priori soil

a posteriori soil

• derived from 3-year GOME observations over east China • a posteriori 0.85 TgN/yr vs a priori 0.24 TgN/yr• peak in summer, significant in spring and fall

14

Simulated Seasonal Variations (O3)

• model captures the overall seasonal patterns

• model fails to simulate the decrease of O3 in July

• model cannot reproduce very high ozone levels

Afternoon

model

model high bias in

summer

15

Simulated Seasonal Variations (CO)

• model tends to underestimate mean CO concentrations• model captures the overall seasonal patterns

model

16

CO-O3 relationship

Parrish et al., 1998

Miyun

• O3/CO ratio is much lower at Miyun

• seasonal variability are generally consistent between the two sites

Sable Island

small correlation in July

17

Average diurnal pattern of summertime ozone

obs June

obs July

Model July

very low ozone in the afternoon!

(< 20 ppb!)

not captured by model!

18

Low Ozone Events in July

obs model

9-10

18 24-2621 28

61

rain or thunderstorm!

19

Flow patterns between June and July

JulyJune

20

frequency distribution of CO

O3 levels as a function of CO

R.H. as a function of CO

June

July

June

July

June

July

0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 4000

0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 4000

0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 4000

O3

RH

<200ppb

21

July CO-O3 Relationship

model

Sensitivity: increase cloud optical depth

Impact of precipitation and Impact of precipitation and cloudiness on local ozone cloudiness on local ozone production!production!

22

Summary Observations reveal seasonal variations in OObservations reveal seasonal variations in O3, 3, CO, CO,

and their correlations in north China regionand their correlations in north China region

Model capture the general characteristics in Model capture the general characteristics in observations of CO and Oobservations of CO and O33

Major discrepancies in the modelMajor discrepancies in the model: : Underestimate CO concentrationsUnderestimate CO concentrations Not able to represent small-scale dynamics (in Not able to represent small-scale dynamics (in

transition seasons)transition seasons) Not able to capture low-ozone events in July, likely Not able to capture low-ozone events in July, likely

due to underestimate of cloudinessdue to underestimate of cloudiness still work in progress….still work in progress….

Need for a finer-resolution model -- nested-grid Need for a finer-resolution model -- nested-grid capability in GEOS-5capability in GEOS-5