Photochemistry of
Atmospheric Species
Photolysis Reactions
Molecular Spectra and Quantum Yields
Solar Radiation
Fundamentals of Radiative Transfer
Aerosols
Clouds
Radiative Forcing of Climate
Biological Effects of UV Radiation
Sasha Madronich
National Center for Atmospheric Research
Boulder Colorado USA
[email protected]
Boulder, 8 February 2011
Energetics of Oxygen in the Atmosphere
DHf (298K)
kcal mol-1
Excited atoms
O*(1D)
104.9
Ground state atoms
O (3P)
59.6
Ozone
O3
34.1
Normal molecules
O2
0
Increasing
stability
2
Atmospheric Oxygen
Thermodynamic vs. Actual
1
1E-10
Concentration, atm.
1E-20
1E-30
O3
1E-40
O2 (=0.21)
thermodyn. O3
thermodyn. O
O
1E-50
thermodyn. O*
1E-60
1E-70
observed O3
inferred O
1E-80
inferred O*
1E-90
O*
1E-100
1E-110
200
220
240
260
280
300
Temperature, K
3
Photochemistry
 Thermodynamics alone cannot explain
atmospheric amounts of O3, O, O*
 Need
– energy input, e.g.
O2 + hn  O + O
(l < 240 nm)
– chemical reactions, e.g.
O + O2 (+ M)  O3 (+ M)
= Photochemistry
4
Some Important Photolysis Reactions
O2 + hn (l < 240 nm)  O + O
source of O3 in stratosphere
O3 + hn (l < 340 nm)  O2 + O(1D)
source of OH in troposphere
NO2 + hn (l < 420 nm)  NO + O(3P)
source of O3 in troposphere
CH2O + hn (l < 330 nm)  H + HCO
source of HOx, everywhere
H2O2 + hn (l < 360 nm)  OH + OH
source of OH in remote atm.
HONO + hn (l < 400 nm)  OH + NO
source of radicals in urban atm.
5
Quantifying Photolysis Processes
Photolysis reaction:
AB + hn  A + B
Photolysis rates:
Photolysis frequency (s-1)
J=
l F(l) s(l) f(l) dl
(other names: photo-dissociation rate coefficient, J-value)
6
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.
Measurement of Absorption Cross Section s(l)
Pressure
gage
Gas
supply
Pump for
clean out
Gas
in
Prism
Absorption cell
Light
detector
Lamp
L
Transmittance = I / Io = exp(-s n L)
s = -1/(nL) ln( I/Io )
Easy: measure pressure (n = P/RT), and relative change in light: I/Io
8
O2 Absorption Cross Section
9
O3 Absorption Cross Section
10
H2O2 Absorption Cross Section
11
NO2 Absorption Cross Section
12
CH2O Absorption Cross Section
13
Measurement of Quantum Yields f(l)
Pressure
gage
Gas
supply
Pump for
clean out
Gas
in
Prism
Absorption cell
Light
detector
Lamp
L
Transmittance = I / Io = exp(-s n L)
Quantum Yield = number of breaks per photon absorbed
f = Dn / DI
Difficult: must measure absolute change in n (products) and I (photons absorbed)
14
Measured Quantum Yields
15
Compilations of Cross Sections & Quantum Yields
http://www.atmosphere.mpg.de/enid/2295
http://jpldataeval.jpl.nasa.gov/
16
Spectral Range For
Tropospheric Photochemistry
1
0.9
0.8
dJ/d l (rel)
0.7
0.6
O3->O2+O1D
NO2->NO+O
0.5
H2O2->2OH
HONO->HO+NO
0.4
CH2O->H+HCO
0.3
0.2
0.1
0
280
300
320
340
360
380
400
420
Wavelength, nm
surface, overhead sun
17
Solar Spectrum
O2 and O3 absorb
all UV-C (l<280 nm)
before it reaches the
troposphere
18
UNEP, 2002
Typical Vertical Optical Depths, t
100
Optical depth
10
O3 (300 DU)
Rayleigh
Aerosol (25 km)
Cloud (32)
1
0.1
0.01
280
300
320
340
360
380
Wavelength, nm
400
420
Direct transmission = exp(-t)
Diffuse transmission can be much larger
19
DIFFUSE LIGHT - CLEAN SKIES, SEA LEVEL
Irradiance
Actinic flux
Diffuse fraction
1.0
0.8
320 nm
0.6
0.4
550 nm
0.2
0.0
0
20
40
60
Solar zenith angle, deg.
80
100
RADIATIVE TRANSFER
CONCEPTS
21
RADIATIVE TRANSFER THEORY
…is hard
22
Spectral Radiance, I
I(l,q,f) = N(hc/l) / (dt dA dw dl)
units: J s-1 m-2 sr-1 nm-1
(old name = spectral Intensity)
23
INTEGRALS OVER ANGULAR INCIDENCE

E=
2 2
  I (q ,  ) cos q sin q d q d 
0 0
Watts m-2
 2
F =   I (q ,  ) sin q d  d q
0 0
Watts m-2 or quanta s-1 cm-2
24
Absorption and Scattering
• Absorption – inelastic, loss of radiant energy:
• Scattering – elastic, radiant energy is conserved,
direction changes:
25
Beer-Lambert Law
Volume of box = V = Area x Thickness = A dz
Molecules (or particles):
each has cross sectional area s
total number = N = n V
total area of molecules = s N = s n A dz
z
I(z)
z+dz
I(z+dz)
Fraction of light lost = fraction of area shaded
by molecules = s n A dz / A = s n dz
Fraction of light surviving box:
I(z+dz)/I(z) = 1 - s n dz
or
[I(z+dz) - I(z)] / dz = - s n I(z)
or (lim dz  0)
(differential form)
(integral form)
dI/dz = - s n I
I(z2) = I(z1) exp [- s n (z2-z1)]
26
Definition of Optical Depth
Beer-Lambert Law: I(z2) = I(z1) exp [- s n (z2-z1)]
Optical depth: t = s n (z2-z1)
If s and/or n depend on z, then
z2
t =  s ( z ) n( z ) d z
z1
In monochromatic radiative transfer,
t is often used as the spatial coordinate
(instead of z)
27
SCATTERING PHASE FUNCTIONS
P(q, f; q’, f’)
28
The Radiative Transfer Equation
Propagation derivative
cos q
dI (t, q, f )
dt
Scattering from
direct solar beam
Beer-Lambert
attenuation
=  I (t , q, f) 

wo
4
2
wo
4
F e
 t/ cos q o
P( q, f; q o , f o ) 
1
0 1 I( t, q' , f' )P (q, f; q' , f' )d cos q' df'
Scattering from diffuse light
(multiple scattering)
29
NUMERICAL SOLUTIONS TO RADIATIVE
TRANSFER EQUATION
• Discrete ordinates
n-streams (n = even), angular distribution
exact as n but speed  1/n2
• Two-stream family
delta-Eddington, many others
very fast but not exact
• Monte Carlo
slow, but ideal for 3D problems
• Others
matrix operator, Feautrier, adding-doubling,
successive orders, etc.
30
Multiple Atmospheric Layers
Each Assumed to be Homogeneous
Must specify three optical properties:
Optical depth, Dt
Single scattering albedo, wo = scatt./(scatt.+abs.)
Asymmetry factor, g: forward fraction, f ~ (1+g)/2
31
For each layer, must specify Dt, wo, g:
1. Vertical optical depth, Dt(l, z) = s(l, z) n(z) Dz
for molecules: Dt(l, z) ~ 0 - 30
Rayleigh scatt. ~ 0.1 - 1.0 ~ l-4
O3 absorption ~ 0 - 30
for aerosols: 0.01 - 5.0
Dt(l, z) ~ l-a
for clouds: 1-1000
a~0
cirrus ~ 1-5
cumulonimbus ~ > 100
32
For each layer, must specify Dt, wo, g:
2. Single scattering albedo, wo(l, z) = scatt./(scatt.+abs.)
range 0 - 1
limits:
pure scattering = 1.0
pure absorption = 0.0
for molecules, strongly l-dependent, depending on
absorber amount, esp. O3
for aerosols:
sulfate ~ 0.99
soot, organics ~ 0.8 or less,
not well known but probably higher
at shorter l, esp. in UV
for clouds: typically 0.9999 or larger (vis and UV)
33
For each layer, must specify Dt, wo, g:
3. Asymmetry factor, g(l, z) = first moment of phase function
1
range -1 to + 1
pure back-scattering = -1
isotropic or Rayleigh = 0
pure forward scattering = +1
1
g =  P () cos  d(cos )
2 1
strongly dependent on particle size
for aerosols:, typically 0.5-0.7
for clouds, typically 0.7-0.9
Mie theory for spherical particles: can compute Dt, wo, g
from knowledge of l, particle radius and complex index of refraction
34
SIMPLE
2-STREAM
METHOD:
3 Equations
for each layer
AEROSOLS
36
Many different types of aerosols
• Size distributions
• Composition (size-dependent)
Need to determine aerosol optical properties:
t(l) = optical depth
wo = single scattering albedo
P() = phase function or g = asymmetry factor
37
Mie Scattering Theory
For spherical particles, given:
Complex index of refraction: n = m + ik
Size parameter:
a = 2r / l
Can compute:
Extinction efficiency
Qe(a, n)
x
r2
Scattering efficiency
Qs(a, n)
x
r2
Phase function
or asymmetry factor
P(, a, n)
g(a, n)
38
Extinction Efficiency, Qext
5
n = 1.55 + 0.01i
4
300 nm
3
Qext
400 nm
2
1
0
10
100
1000
10000
Radius, nm
39
Phase function or Asymmetry factor, g
Asymmetry factor, g
n = 1.5 + 0.01 i
1
0.8
100 nm
0.6
g
300 nm
0.4
1000 nm
0.2
0
200
300
400
500
600
700
Wavelength, nm
40
Single Scattering Albedo = Qscatt/Qext
Single Scattering Albedo, w o
n = 1.5 + 0.01 i
1.2
1
0.8
100 nm
w o 0.6
300 nm
1000 nm
0.4
0.2
0
200
300
400
500
600
700
Wavelength, nm
41
Aerosol size distributions
42
Optical properties of aerosol ensembles

Ke (l ) =  r 2Qe (r , l )n(r ) d r
Total extinction coefficient =
0

K s (l ) =  r 2Qs (r , l )n(r ) d r
Total scattering coefficient =
0
Average single scattering albedo =
 o (l ) = K s (l ) / Ke (l )

Average asymmetry factor =
g (l ) =
2
g
(
r
,
l
)

r
Qs (r , l )n(r ) d r

0

2

r
 Qs (r , l )n(r ) d r
0
43
UV Actinic Flux Reduction  Slower Photochemistry
Mexico City
3E+14
18 March 16:55 LT
solar zenith angle = 65o
Quanta cm-2 s-1 nm-1
obs
2E+14
tuv-clean
tuv-polluted
1E+14
0
300
320
Madronich, Shetter, Halls, Lefer, AGU’07
340
360
Wavelength, nm
380
400
44
Aerosol Effects
NO2 Photolysis Frequency
19N, April, noon, AOD = 1 at 380 nm
2
clean
z, km
1.5
1
0.5
moderately
absorbing
(wo=0.8)
purely
scattering
0
5.0E-03
1.0E-02
JNO2, s
1.5E-02
-1
Castro et al. 2001
CLOUDS
46
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
47
EFFECT OF UNIFORM CLOUDS ON ACTINIC FLUX
Altitude, km
340 nm, sza = 0 deg.,
cloud between 4 and 6 km
10
8
6
4
2
0
0.E+00
od = 100
od = 10
od = 0
4.E+14
8.E+14
Actinic flux, quanta cm-2 s-1
48
INSIDE CLOUDS:
Photon Path Enhancements
Cumulonimbus, od=400
Mayer et al., 1998
49
Broken Clouds
50
SPECTRAL EFFECTS OF PARTIAL CLOUD COVER
Crafword et al., 2003
51
PARTIAL CLOUD COVER
Biomodal distributions
Crawford et al., 2003
Independent Pixel Approximation
 Cloud free:
• So = direct sun
• Do = diffuse light from sky
• Go = total = So + Do
 Completely covered by clouds:
• S1 = direct sun (probably very small)
• D1 = diffuse light from base of cloud
• G1 = total = S1 + D1
 Mix: Clouds cover a fraction c of the sky
• If sun is not blocked:
GNB = So + cD1 + (1-c)Do
• If sun is blocked:
GB = S1 + cD1 + (1-c)Do
53
Photochemistry Inside Liquid Particles
aerosol
clouds
Mayer and Madronich, 2004
54
Radiative Forcing
of
Climate
55
Shortwave and Longwave Radiation
56
Energy Balance
57
Infrared Absorbers
58
Very Structured Spectra
59
Radiative Forcing
Incoming solar ~340 W m-2
1827 – Fourier recognizes atmospheric heat
trapping
1860 – Tyndall measures infrared spectra
1896 – Arrhenius estimates doubling of CO2
would increase global temperatures by 5-6 oC
Changes since 1750:
long-lived gases ~ 3 W m-2
ozone ~ 0.4 W m-2
aerosols and clouds ~ -1 W m-2
IPCC, 2007
Some Climate Model Problems
11 models compared
All predict 20th century climate
well, but for different reasons
Trade off between
- aerosol forcing
- climate sensitivity
Climate sensitivity = DT for doubling of CO2
Kiehl, 2007
61
Biological Effects
of
Ultraviolet Radiation
62
Non-melanoma Skin Cancers
63
Moles and Melanoma
64
Melanoma Warning Signs
Asymmetry, Border, Color, Diameter (ABCD)
65
Damage to Eyes
66
ACTION SPECTRUM DETERMINATION
67
Ocular damage action spectra (normalized at 300 nm)
100
cataract pig (Oriowo)
cataract rat (Merriam)
cataract rabbit (Pitts)
conjunctiva human (Pitts)
cornea rabbit (Pitts)
cornea primate (Pitts)
cornea human (Pitts)
uvitis rabbit (Pitts)
ave-Pitts only
ave-all ocular
erythema human (CIE)
Sensitivity
10
1
0.1
0.01
0.001
0.0001
280
300
320
340
360
380
Wavelength, nm
68
SPECTRALLY INTEGRATED RADIATION
 Radiometry
Signal (W m-2) = l E(l) R(l) dl
 Biological effects
Dose rate (W m-2) = l E(l) B(l) dl
 Photo-dissociation of atmospheric chemicals
Photolysis frequency (s-1) = l F(l) s(l) f(l) dl
69
Changes in surface irradiance, for 1% O3 change
Near 300 DU
Solid: Absolute
(mW m-2 nm-1)
Dotted: Relative
(percent)
Heavy: sza = 0o
Light: sza = 70o
70
Observed dependence
of UV on O3 changes
(clear sky)
For erythemal
(sunburning) radiation
Madronich et al. 1998
Global Climatology of UVR from Satellites
 Total Ozone Mapping Spectrometer (TOMS):
– Nimbus 7 (1 Nov 1978 – 31 Dec 1992)
– Meteor 3 (22 Aug 1991 – 11 Dec 1994)
– Earth Probe (2 Jul 1996 – 30 Jun 2000)
 Data (Level 3/Version 7):
– O3 by UV-B wavelength pairs
– Cloud reflectivity at 380 nm
– 1.25o lon x 1.00o lat
 Model:
– TUV look up tables for erythemal irradiance, as functions of
sza, O3, surface elevation
– Calculate for each location, 15 min of each half day, each
month, each year
72
UVERY climatology, 1979 to 2000 - Clear skies
UVERY climatology, 1979 to 2000 - All skies
UVERY change, 1980’s to 1990’s
Clear sky (ozone changes only)
Cloud factor % changes, 1980’s to 1990’s
UVERY % change, 1980’s to 1990’s
All sky conditions (ozone and cloud changes)