Sensitivity of Ozone FormationTo Photons

Sasha MadronichNational Center for Atmospheric ResearchBoulder Colorado USA

Mexico City, 14 August 2009

Tropospheric Ozone Formation:

Urban ozone is generated when air containing hydrocarbons and nitrogen oxides is exposed to ultraviolet radiation

- Haagen-Smit (1950s)

3

Mexico City’s O3 Production is VOC-limited, NOx-inhibited

WRF-Chem --- sensitivity studies ● observations CAMx --- sensitivity studies

Tie et al., 2007 Lei et al., 2007

NOx-VOC Regimes NOx-limited

Very low NOx:O3 ~ J0.5 [NOx]

VOC-limited

NOx-inhibited

Very high NOx:O3 ~ J [VOC] / [NOx]

Sensitivity (%/%) of O3 in Mexico City

NCAR Master Mechanism box model

Madronich, unpubl..

O3 production is always PHOTON-LIMITED

Radiative transfer modeling ok for ideal conditions: cloud-free, pollution-free

Large uncertainties for realistic conditions, not well modeled

Affected by long term trends in aerosols, absorbing gases, clouds

Few comprehensive studies on photon-limitation

Quantifying Photolysis Processes

Photolysis reaction: AB + hn A + B

Photolysis frequency (s-1) J = l F(l) s(l) f(l) dl

(other names: photo-dissociation rate coefficient, J-value)

Photolysis rates:

CALCULATION OF PHOTOLYSIS COEFFICIENTS

J (s-1) = l F(l) s(l) f(l) dl

F(l) = spectral actinic flux, quanta cm-2 s-1 nm-1

probability of photon near molecule.

s(l) = absorption cross section, cm2 molec-1

probability that photon is absorbed.

f(l) = photodissociation quantum yield, molec quanta-1

probability that absorbed photon causes dissociation.

Solar Spectrum

UNEP, 2002

O2 and O3 absorball UV-C (l<280 nm)before it reaches the troposphere

Spectral Region ForTropospheric Photochemistry

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

280 300 320 340 360 380 400 420

Wavelength, nm

dJ/d

l (r

el) O3->O2+O1D

NO2->NO+OH2O2->2OHHONO->HO+NOCH2O->H+HCO

surface, overhead sunMadronich, unpubl..

Typical Vertical Optical Depths, t

0.01

0.1

1

10

100

280 300 320 340 360 380 400 420

Wavelength, nm

Opt

ical

dep

th O3 (300 DU)RayleighAerosol (25 km)Cloud (32)

Direct transmission = exp(-t)Diffuse transmission can be much larger

Madronich, unpubl..

Effect of Pollutants on UV Irradiance

-30

-20

-10

0300 320 340 360

Wavelength, nmU

V re

duct

ion,

% SO2 30 ppbhaze 25kmNO2 50 ppbO3 120ppb

Model calculations for 21 June, 35 N, noon, pollutants distributed over a 1 km boundary layer

Madronich, unpubl..

UV Actinic Flux Reduction Slower Photochemistry

300 320 340 360 380 4000

100000000000000

200000000000000

300000000000000

obs

tuv-clean

tuv-polluted

Wavelength, nm

Qua

nta

cm-2

s-1

nm

-118 March 16:55 LTsolar zenith angle = 65o

Madronich, Shetter, Halls, Lefer, AGU’07

Mexico City (T1)

JNO2 Observed/Model_cleanMarch 2006 T1 supersite

thin curves = individual daysthick blue curve = average

Madronich, unpubl..

Aerosol Impacts on Photochemistry

Outside Mexico City (Tres Marias) 15 April 94

6 9 12 15 18

Local time, hrs.

Mexico City 11 Feb 94

0.E+00

2.E-03

4.E-03

6.E-03

8.E-03

1.E-02

6 9 12 15 18

Local time, hrs

J NO

2, s-1

JNO2_expcleanwo=0.95wo=0.80

Castro et al. 2001

NO2 + hn NO + O (at surface)

O3 Suppression from Aerosol (Mexico City)

Castro et al. 2001

Vertical Structure of Aerosol EffectsNO2 Photolysis Frequency

19N, April, noon, AOD = 1 at 380 nm

0

0.5

1

1.5

2

5.0E-03 1.0E-02 1.5E-02

JNO2, s-1

z, k

m

clean

purelyscattering

moderately absorbing(wo=0.8)

Castro et al. 2001

Aerosol Single Scattering AlbedoMexico City

UV-MFRSR (T1)AERONET (T1)Barnard et al. (CENICA)

Corr et al., 2009

DIURNAL CYCLE OF AEROSOL OPTICS550 nm

Paredes-Miranda et al., 2008

Clouds

UNIFORM CLOUD LAYER

• Above cloud: - high radiation because of reflection

• Below cloud: - lower radiation because of attenuation by cloud

• Inside cloud: - complicated behavior– Top half: very high values (for high sun)– Bottom half: lower values

EFFECT OF UNIFORM CLOUDS ON ACTINIC FLUX

340 nm, sza = 0 deg., cloud between 4 and 6 km

02468

10

0.E+00 4.E+14 8.E+14

Actinic flux, quanta cm-2 s-1

Altit

ude,

km od = 100

od = 10od = 0

Madronich, 1987

SPECTRAL EFFECTS OF PARTIAL CLOUD COVER

Crafword et al., 2003

PARTIAL CLOUD COVERBiomodal distributions

Crafword et al., 2003

WRF-Chem Regional O3 Prediction

Observed daily 1-h maximum O3 for all EPA AIRNOW surface stations in the model domain, 21 July - 4 August 2002.

G.A. Grell et al. / Atmospheric Environment 39 (2005) 6957–6975

Correlation coefficient - R

G.A. Grell et al. / Atmospheric Environment 39 (2005) 6957–6975

scatter mostly from clouds not modeled correctly !?

Mean Bias

G.A. Grell et al. / Atmospheric Environment 39 (2005) 6957–6975

bias mostly from aerosols not modeled correctly !?

Problems and Opportunities

O3 productions is– sometimes NOx limited– sometimes VOC limited– always photon limited

Pollution affects photon availability (10-30% reductions are not uncommon).

Aerosols and clouds change the vertical gradient of photochemistry– usually brighter above, dimmer below (but not always)

UV properties of aerosols are poorly known– Composition– Size distributions– Morphologies, mixing states– Vertical distribution

VOC-NOx-photon interactions: Photon availability may change NOx-limited transition point:

NOx

O3High J

Low J

Delay of reactivity: slower urban photochemistry allows more export of precursors for regional oxidants.

Regional photochemistry may be accelerated by scattering aerosols (Dickerson et al., 1997)

Clouds: need improved cloud statistics for parameterizing optical properties. Also, how to deal with model vs. real clouds?

Need evaluation of model J-values with in situ measurements under realistic conditions. Need to demonstrate closure through the vertical extent.