Biofuels & Atmospheric Chemistry:
What can a global model tell us
about our future decisions?
Rachel Catherine Pike
Emmanuel College
This dissertation is submitted for the degree of Doctor of Philosophy
Declaration
This dissertation is the result of my own work and includes nothing which is the
outcome of work done in collaboration except where specifically indicated in the
text.
This dissertation has not been submitted, in whole or in part, for any other degree,
diploma or other qualification at any other university, and is less than 60,000 words
in length.
Rachel Pike
December 2009
Abstract
Biofuels & Atmospheric Chemistry:
What can a global model tell us about our future decisions?
Rachel Pike
Emmanuel College, University of Cambridge
Biomass energy is the oldest form of energy harnessed by humans. Currently,
processed biofuels, which are derived from biomass, are being pursued as a possi-
ble route to decarbonize the transport sector—a particularly difficult task for both
technological and sociological reasons. In this thesis I explore the impacts that
large scale biofuel use could have on atmospheric chemistry.
I review the current state of biofuels politically and technologically, focusing
on ethanol and biodiesel. I discuss the salient features of tropospheric chemistry
and in particular the oxidation of isoprene, an important biogenic volatile organic
compound. I examine the impact that including isoprene oxidation has in a new
chemistry-climate computer model, UKCA; the response of ozone turns out to
depend on local chemical conditions. To evaluate the global model, I use data
from the OP3 field campaign in Malaysia and compare it with output from the
model chemical mechanism. The mechanism is able to reproduce NOx and ozone
measurements well, though is more sensitive to representations of physical rather
than chemical processes. I also perform a simple sensitivity study which examines
crop changes in the region of Southeast Asia.
In the final two chapters, I turn to two distinct phases of the biofuel life cycle. I
characterize a potential future atmosphere through an ozone attribution study, then
examine the impact of future cropland expansion (phase I of a biofuel life cycle)
on tropospheric chemistry. I find that land use change has a large impact on ozone,
and that it is more acute than another perturbation (CO2 suppression) to isoprene
emissions. I then move to phase III of the life cycle—combustion—and look at the
sensitivity of the model chemistry to surface transport emissions from biofuels as
a replacement for conventional fuels. I find that biodiesel reduces surface ozone,
while ethanol increases it, and that the response has both a linear and nonlinear
component.
Acknowledgments
Writing a PhD is an intense and solitary process. Without the support of colleagues,
friends, and family, I’m not sure how anyone finishes one of these things.
I have first to thank Professor John Pyle for his guidance and advice over the
last three years; I couldn’t have asked for a better supervisor. Thank you to Paul
Young, for humoring (humouring?) countless questions; he actually taught me
how a global model works, what isoprene is, and how to design a set of exper-
iments. To G6 members, past and present, thank you for debugging advice and
office olympics. I owe a particular note of gratitude to Luke Abraham for his help
with UKCA. Thank you to Juliette Lathie`re for collaborating on the isoprene crop
experiments, and for being enthusiastic throughout the process. I need to acknowl-
edge James Lee for his patience with the box model work, and to say that the entire
OP3 team are fantastic. We all miss Kate Furneaux; may she be in peace.
Thank you to the Gates Trust for funding.
I had a number of moments of panic while writing, and I am desperately grate-
ful to my friends in Cambridge for rescuing me from myself. Thank you to my
friends at Emma, fellow Gates Scholars, the residents of Earl Street, and a cer-
tain northeastern chemist. To my friends further away: some of those phone calls
were the best distraction I could have asked for—thank you. Finally, to my fam-
ily: Mom, Dad, and Max. This PhD would never have been submitted without
occasional respite in California and many phone calls around the world. I love
you.
Rachel Pike
December 2009
for granny
Contents
Prologue 1
1 Biofuels: A Solution? 3
1.1 A brief history of climate agreements . . . . . . . . . . . . . . . 4
1.2 The transport sector . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 An introduction to biofuels . . . . . . . . . . . . . . . . . . . . . 10
1.3.1 Biomass potential . . . . . . . . . . . . . . . . . . . . . . 11
1.3.2 Ethanol . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3.3 Biodiesel . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4 Life cycle of a biofuel . . . . . . . . . . . . . . . . . . . . . . . . 18
1.4.1 Phase I: Cultivation . . . . . . . . . . . . . . . . . . . . . 19
1.4.2 Phase II: Transformation . . . . . . . . . . . . . . . . . . 20
1.4.3 Phase III: Combustion . . . . . . . . . . . . . . . . . . . 21
1.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2 Biofuels: Potential Atmospheric Impacts 23
2.1 Tropospheric chemistry . . . . . . . . . . . . . . . . . . . . . . . 23
2.1.1 Initiation . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1.2 Propagation . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.1.3 Chain termination . . . . . . . . . . . . . . . . . . . . . 30
2.1.4 PAN chemistry . . . . . . . . . . . . . . . . . . . . . . . 33
2.1.5 Nighttime chemistry . . . . . . . . . . . . . . . . . . . . 33
iii
2.2 Isoprene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.2.1 Isoprene chemistry . . . . . . . . . . . . . . . . . . . . . 34
2.2.2 Observations of isoprene emission . . . . . . . . . . . . . 37
2.2.3 Estimations of isoprene emissions . . . . . . . . . . . . . 40
2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3 The Impact of Isoprene in UKCA 46
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.2 UKCA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.1 Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.2 Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2.3 Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.2.4 Isoprene emissions . . . . . . . . . . . . . . . . . . . . . 56
3.3 Model evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.3.1 Stratosphere-troposphere exchange . . . . . . . . . . . . 58
3.3.2 Comparison to measurements . . . . . . . . . . . . . . . 60
3.4 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.5 Impact of isoprene in UKCA . . . . . . . . . . . . . . . . . . . . 62
3.5.1 Ozone . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.5.2 Methane . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4 Case Study: South East Asia 79
4.1 OP3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.1.2 The measurements . . . . . . . . . . . . . . . . . . . . . 81
4.1.3 Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.2 Regional sensitivity study . . . . . . . . . . . . . . . . . . . . . . 102
4.2.1 Experimental setup . . . . . . . . . . . . . . . . . . . . . 102
4.2.2 Impact on ozone . . . . . . . . . . . . . . . . . . . . . . 103
iv
4.2.3 Uncertainties . . . . . . . . . . . . . . . . . . . . . . . . 107
4.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5 Future Ozone Attribution and Phase I: Cultivation 110
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.1.1 How will ozone change in the future? . . . . . . . . . . . 110
5.1.2 The impacts of future land use change . . . . . . . . . . . 112
5.2 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.3 Future isoprene emissions . . . . . . . . . . . . . . . . . . . . . . 118
5.4 Ozone attribution in a potential future climate . . . . . . . . . . . 124
5.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 124
5.4.2 The impact of climate change . . . . . . . . . . . . . . . 125
5.4.3 The impact of biogenic isoprene emissions . . . . . . . . 129
5.4.4 The impact of anthropogenic emissions . . . . . . . . . . 130
5.4.5 The combined impacts . . . . . . . . . . . . . . . . . . . 132
5.4.6 The future tropospheric ozone budget . . . . . . . . . . . 134
5.5 Future land use change . . . . . . . . . . . . . . . . . . . . . . . 135
5.5.1 Ozone . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.5.2 Crops: exposure to ozone . . . . . . . . . . . . . . . . . . 138
5.6 Land use change versus CO2 suppression . . . . . . . . . . . . . 141
5.6.1 Land use and CO2 suppression: ozone . . . . . . . . . . . 141
5.6.2 Land use and CO2 suppression: HOx . . . . . . . . . . . 145
5.6.3 Land use and CO2 suppression: budgets . . . . . . . . . . 146
5.6.4 Land use and CO2 suppression: reaction fluxes . . . . . . 148
5.6.5 Uncertainties . . . . . . . . . . . . . . . . . . . . . . . . 149
5.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6 Phase III: Combustion 151
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.1.1 Impact on the atmosphere . . . . . . . . . . . . . . . . . 152
v
6.1.2 Biofuel emissions from the tailpipe . . . . . . . . . . . . 153
6.2 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.3 The BASE case . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.4 What is the impact of biofuels? . . . . . . . . . . . . . . . . . . . 161
6.4.1 Ozone . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.4.2 NOx-VOC sensitivity . . . . . . . . . . . . . . . . . . . . 164
6.4.3 Seasonality . . . . . . . . . . . . . . . . . . . . . . . . . 166
6.5 Uncertainties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
7 Conclusions 170
7.1 Summary of novel work . . . . . . . . . . . . . . . . . . . . . . 170
7.2 The impact of the full life cycle . . . . . . . . . . . . . . . . . . . 173
7.3 Potential further studies . . . . . . . . . . . . . . . . . . . . . . . 174
7.4 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Bibliography 176
vi
List of Figures
1.1 End use oil consumption by sector . . . . . . . . . . . . . . . . . 7
1.2 Stabilisation wedges from Pacala and Socolow [2004] . . . . . . 9
1.3 Schematic of biofuel production . . . . . . . . . . . . . . . . . . 10
1.4 Chemical structure of ethanol and biodiesel . . . . . . . . . . . . 11
1.5 Schematic of the three stages of a biofuel life cycle . . . . . . . . 18
2.1 Schematic of the tropospheric ozone budget . . . . . . . . . . . . 24
2.2 A visual summary of tropospheric chemistry . . . . . . . . . . . . 26
2.3 VOC-NOx topography from Sillman and He [2002] . . . . . . . . 32
2.4 Relationships between CO2 and isoprene emission . . . . . . . . 39
3.1 Present day isoprene emissions for January and July . . . . . . . . 56
3.2 Regional emissions of isoprene . . . . . . . . . . . . . . . . . . . 57
3.3 Stratosphere-troposphere exchange in UKCA . . . . . . . . . . . 59
3.4 Comparison of model output to climatological ozone . . . . . . . 60
3.5 Change in surface ozone with and without isoprene . . . . . . . . 63
3.6 NO2 emissions in UKCA . . . . . . . . . . . . . . . . . . . . . . 65
3.7 Chemical production of ozone between BASE and NOISOP . . . 67
3.8 Key reaction fluxes between BASE and NOISOP . . . . . . . . . 68
3.9 Relative increase in PAN between BASE and NOSIOP . . . . . . 69
3.10 Change in ozone versus change in isoprene . . . . . . . . . . . . 71
3.11 Change in ozone versus NOx emission . . . . . . . . . . . . . . . 72
3.12 Difference in surface OH with and without isoprene . . . . . . . . 74
vii
3.13 Seasonal variation of methane lifetime with and without isoprene . 76
4.1 Time series of measured NO, NO2, and O3 during OP3 . . . . . . 83
4.2 Diurnal average NO, NO2, and O3 measurements during OP3 . . 85
4.3 The base case box model . . . . . . . . . . . . . . . . . . . . . . 89
4.4 Chemical sensitivities compared to NO, NO2, and O3 data . . . . 93
4.5 Model comparison with and without ozone venting . . . . . . . . 96
4.6 Cost function analysis for model fitting NO and NO2 . . . . . . . 98
4.7 Cost function analysis for model fitting O3 . . . . . . . . . . . . . 99
4.8 ‘Best fit’ model-measurement comparison . . . . . . . . . . . . . 101
4.9 Area of emission perturbation in two sensitivity studies . . . . . . 102
4.10 Difference in surface ozone between OILPALM and BASE . . . . 103
4.11 Difference in zonal mean ozone between OILPALM and BASE . . 104
4.12 Difference in PAN between OILPALM and BASE . . . . . . . . . 105
4.13 Difference in ozone between SUGARCANE and BASE . . . . . . 106
4.14 Linearity in maximum ozone response . . . . . . . . . . . . . . . 107
5.1 A summary of emissions and climate sensitivity experiments . . . 115
5.2 Matrix of land use change and CO2 suppression experiments . . . 116
5.3 Change in isoprene emissions from climate change, CO2 suppres-
sion, and cropland expansion . . . . . . . . . . . . . . . . . . . . 120
5.4 Change in isoprene emissions between BOTH and BASE . . . . . 121
5.5 Crop mask used to estimate isoprene emissions used in FutCrops . 122
5.6 Change in isoprene emissions between BOTH and FutCROPS . . 123
5.7 Difference in annual ozone for from emissions and climate . . . . 126
5.8 Impact of climate change on temperature, ozone production, and
NOx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.9 Impact of climate change on ozone in the U.S. and Europe . . . . 129
5.10 Future change in surface NO2 emissions . . . . . . . . . . . . . . 131
5.11 Change in ozone between BOTH and BASE . . . . . . . . . . . . 133
viii
5.12 Change in ozone between BOTH and FutCROPS . . . . . . . . . 136
5.13 Difference in ozone production and NOz between FutCROPS and
BOTH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.14 Increase exposure to surface high ozone due to land use change . . 140
5.15 Annual average change in ozone for crops and CO2 suppression . 143
5.16 Average monthly change in the HO2:OH ratio for crops and CO2
suppression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.1 NO2 emissions from surface transport . . . . . . . . . . . . . . . 155
6.2 CO emissions from surface transport . . . . . . . . . . . . . . . . 156
6.3 Seasonal variation in ratio of formation of H2O2 and HNO3 . . . 160
6.4 Average change in surface ozone between BD100 and E85 and
BASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
6.5 Maximum change in ozone between BD100, BD20, and E85 and
BASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
6.6 Percentage change in ratio of formation of H2O2 and HNO3 . . . 165
6.7 Seasonal change in ozone between the E85 and BASE runs . . . . 167
ix
List of Tables
1.1 Estimates of future energy supply from biomass . . . . . . . . . . 12
1.2 Energy balance of ethanol fuel-feedstock systems . . . . . . . . . 14
1.3 Energy balance of biodiesel fuel-feedstock systems . . . . . . . . 17
2.1 Tropospheric lifetimes of various hydrocarbon species . . . . . . 35
2.2 Isoprene flux rates for various taxa . . . . . . . . . . . . . . . . . 38
2.3 Literature estimates of total global isoprene flux . . . . . . . . . . 42
3.1 MIM species . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.2 Present day emissions in UKCA . . . . . . . . . . . . . . . . . . 53
3.3 Present day NMVOC emissions in UKCA . . . . . . . . . . . . . 55
3.4 Ozone budgets with and without isoprene . . . . . . . . . . . . . 66
4.1 Initial concentrations of six species used in the box model. . . . . 88
4.2 Summary of chemical sensitivity tests . . . . . . . . . . . . . . . 92
5.1 Future B2+CLE emissions . . . . . . . . . . . . . . . . . . . . . 118
5.2 Total global isoprene flux for eight studies . . . . . . . . . . . . . 119
5.3 Ozone budgets in emissions and climate change attribution experi-
ments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.4 Ozone budgets with CO2 suppression and land use change . . . . 147
5.5 Annual average reaction fluxes from land use change andCO2 sup-
pression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
x
6.1 Emission scaling factors for traffic experiments . . . . . . . . . . 157
6.2 Emission scaling factors for E85 VOCs . . . . . . . . . . . . . . 157
Acronyms, Abbreviations &
Symbols
BD20 A fuel blend of 20% biodiesel and 80% conventional diesel
BD100 100% biodiesel fuel
BL Boundary layer
BVOC Biogenic volatile organic compound
CCM Climate composition model
DGVM Dynamic global vegetation model
DM8H Daily 8-hour ozone maximum
E85 A fuel blend of 85% ethanol and 15% petrol
ETS Emissions trading scheme
GCM General circulation model
GHG Greenhouse gas
HDV Heavy duty vehicle
IPCC Intergovernmental Panel on Climate Change
LCA Life cycle assessment
LDV Light duty vehicle
MIM Mainz isoprene mechanism [Po¨schl et al., 2000]
NOx Nitrogen oxides = NO + NO2
OP3 Oxidation and Particle Photochemical Processes; a NERC-funded field campaign
PAN Peroxyacyl nitrate
PAR Photosynthetically active radiation
PFT Plant functional type
SOA Secondary organic aerosol
SST Sea surface temperature
STE Stratosphere-troposphere exchange
Tg Teragram (= 1 x 1012 g)
UKCA United Kingdom Chemistry & Aerosol Model (a CCM)
UNFCCC United Nations Framework Convention on Climate Change
VOC Volatile organic compound
h Planck’s constant
λ Wavelength
ν Frequency
xii
Prologue
Biofuels could form an important part of our future energy supply. Although the
impact of biofuels on the carbon budget and greenhouse gas emissions has been
previously studied, the air quality implications of large scale biofuel use have not
yet been well characterized. The goal of this thesis is to explore some of the po-
tential feedbacks between biofuels and tropospheric chemistry, and to address this
question, I use the concept of a life cycle assessment and a new global climate
composition computer model. A model is a good place to test the sensitivities of
the system before any sweeping decisions are made.
In Chapter 1, I describe the political agreements which relate to decarboniza-
tion and fuel standards, the Earth’s potential to grow biomass, and state of technol-
ogy in biofuel growth and production. I also introduce the concept of a life cycle
assessment, and posit some types of emissions that could impact the atmosphere.
Due to its climatic, health, and chemical importance, I examine ozone frequently
throughout the thesis, and in Chapter 2 the chemistry of ozone production and de-
struction is introduced. I also spend parts of further chapters examining the impact
of isoprene (an important biogenic volatile organic compound) emissions on tropo-
spheric chemistry, so the background to isoprene emission, chemistry, and global
flux estimation is also included.
Chapter 3 is the first full description of the tropospheric version the United
Kingdom Chemistry and Aerosol model (UKCA) which is used at the University
of Cambridge. In addition to a brief background to the dynamical core of the
model, I describe the chemistry and emissions in detail. The second half of the
1
chapter focuses on how including isoprene in the model impacts the model results,
particularly for ozone. In Chapter 4, I use a number of measurements from the
OP3 field campaign in Borneo, Malaysia, to assess the ability of the model chem-
ical mechanism. Can this mechanism represent data, and therefore, is it useful?
I also perform two sensitivity studies which examine the impact of changing iso-
prene emissions in the region; these changes in emissions could be due to growth
of biofuel feedstock crops.
Chapter 5 begins with an ozone attribution study of a potential future atmo-
sphere. I also ask: how could phase I of the biofuel life cycle impact tropospheric
chemistry through changing isoprene emissions? I compare the impact of land use
change to the influence of CO2 suppression of isoprene emission on ozone, HOx,
and certain reaction fluxes. In Chapter 6, I move to the third phase of the life
cycle, combustion. How would air quality change if all the cars in the world ran
on ethanol (biodiesel)? I examine the impacts of these fuels on surface ozone in
particular.
The thesis closes in Chapter 7 with a summary of the novel work and con-
clusions gained from my research. I then present some further work (short term)
which could be directly related to the experiments I conducted. Finally, I briefly
describe the outlook (long term) for this type of biofuel-atmosphere research.
2
Chapter 1
Biofuels: A Solution?
Past, present, and projected changes in the Earth System have been unequivocally
linked to anthropogenic activity [Solomon et al., 2007b]. The majority of posi-
tive radiative forcing resulting from human activity has occurred due to increasing
atmospheric concentrations of carbon dioxide [Solomon et al., 2007b], a product
of using petroleum-derived hydrocarbon fuels to power the world’s economies. A
number of other radiatively active greenhouse gases have also increased due to
anthropogenic activity: notably, methane, nitrous oxide, halocarbons, and tropo-
spheric ozone. In order to reduce the probability of large changes in the climate
system, atmospheric concentrations of greenhouse gases, and therefore emissions
of these gases, will have to be reduced. In this light, a number of technologies
which reduce carbon emissions have been proposed.
In the beginning part of this chapter, I explain why biofuels could form part
of this renewable energy mix. In particular, I review a brief history of climate
agreements (section 1.1) that have codified emissions reductions, and I explain
why transportation is a particularly difficult sector to decarbonize (section 1.2).
Section 1.3 is concerned with an overview of biofuels, including the various types
of fuels and the differences between them. In section 1.4, I describe the life cycle
assessment of two biofuels, ethanol and biodiesel, which are currently the most
likely contenders to replace conventional fuels in the short term. This chapter
3
closes with a description of the potential atmospheric impacts of each phase of the
biofuel life cycle.
1.1 A brief history of climate agreements
The first international conference on climate change was sponsored by the World
Meteorological Organization and held in Geneva in 1979 [World Meteorological
Organization, Geneva, February 1979]. Nearly a decade later, the Intergovern-
mental Panel on Climate Change (IPCC) was set up by the United Nations; the
Panel released its first assessment report in 1990 [Houghton et al., 1990]. In 1992,
the United Nations Framework Convention on Climate Change (UNFCCC) was
completed, which committed signatory parties to voluntarily reductions in emis-
sions. The stated aim was that in the year 2000, emissions levels were to equal
1990 levels [United Nations Framework Convention on Climate Change, 1992]. In
1997, the Kyoto Protocol was adopted [United Nations Framework Convention on
Climate Change, 1997], which shifted the reductions from voluntary to mandatory.
It wasn’t until Russia ratified the Protocol in 2004, however, that it was able to
come into force [Hopkin, 30 September, 2004]. The next Conference of the Par-
ties (COP-15) to the UNFCCC is to be held in Copenhagen in December, 2009, at
which a new emissions reduction scheme may be introduced.
Although international agreements exist at the macro scale, regional and na-
tional commitments have been ratified as well. The European Union has set a
goal of reducing emissions by 20% below 1990 levels by the year 2020 [Com-
munication from the commission to the European Parliamant, 23 January 2008],
which is amendable to 30% if other nations commit to similar reductions [Euro-
pean Commission, 23April 2009]. The main tool through which they are encourag-
ing emissions control is the carbon emissions trading scheme (ETS). The EU ETS
is the largest cap-and-trade system in the world, in which total carbon emissions
are ‘capped’, and emissions permits are ‘traded’; the goal is to create both a price
4
for carbon and a functioning carbon market. Individual European countries have
set additional targets, separate from the regulation of carbon emissions required
under the EU ETS. For example, the UK has set reduction targets of 26 per cent by
2020 and at least 60 per cent by 2050, compared to 1990 levels [United Kingdom
Parliament, 2007-2008 Session].
Another national level agreement is the energy bill currently under review in
the United States Congress. The targets are set to 3% by 2012, and 85% by 2050,
using 2005 as the base year. The renewable energy requirement would begin at
6% in 2012 and gradually rise to 25% in 2025, and a national lower carbon fuel
standard would also come into effect. In addition to federal legislation, individual
states have recently been allowed to regulate their greenhouse gas (GHG) emis-
sions under the Clean Air Act. In 2006, California legislated that greenhouse gas
emissions must be reduced to 1990 levels by the year 2020 [California Congress,
2005-2006 Session]. Obtaining energy from renewable sources has become a leg-
islative priority in individual states—twenty four states have renewable portfolio
standards, in which a certain percentage of electricity is required to be sourced
renewably [U.S. Department of Energy: Energy efficiency and renewable energy,
2009]. The obligations range from 10% (Michigan, North Dakota, South Dakota,
Wisconsin and Vermont) to 33% (California).
Within the larger framework of decarbonizing the energy sector, specific tar-
gets and regulations have recently been passed concerning transportation, some of
which affect biofuels. In the EU, a biofuels directive was passed in 2003, which
mandated 5.75% of fuel be replaced by biofuels by 2010, and 10% by 2020 [Eu-
ropean Commission, 8 May 2003]. Recently, these volumetric requirements were
repealed concurrent with the implementation of ‘sustainability criteria’ for biofu-
els. A life cycle assessment (LCA) mechanism has been adopted to assess the full
greenhouse gas benefits of particular biofuels and pathways, and using the formula
described in European Commission [23 April 2009], greenhouse gas savings for
biofuels must be at least 35% in comparison with conventional fuels. This value
5
will scale to 50% savings by 2017 [European Commission, 23 April 2009].
In the United States, both volumetric and GHG requirements have been adopted.
The total renewable fuel requirement in 2008 was 9.0 billion gallons [United States
Environmental Protection Agency, 2009]. Over the course of the regulation, which
falls under the Energy Independence and Security Act [One Hundred Tenth Congress
of the United States of America, 4 January, 2007], the total renewable requirement
increases to 36 billion gallons by 2022. The ‘advanced’ biofuel requirement (com-
prised of cellulosic ethanol and biodiesel) is 21 billion gallons in the same year. All
renewable fuels are required to reduce GHG emissions by 20% compared to their
conventional fuel counterparts. Advanced biofuels are required to have a 50% sav-
ings and cellulosic biofuels are required to have even greater savings of 60%. Indi-
vidual states have also enacted and regulated fuel carbon standards. For example,
the Governor of California issued an Executive Order which mandated a low car-
bon fuel standard [Office of the Governor of California, 18 January 2007]. The low
carbon standard requires a reduction in carbon intensity of fuels by 10% by 2020
[California Environmental Protection Agency: Air Resources Board, Accessed 15
September 2009]. The fuels will be assessed using a life cycle assessment model,
GREET [Wang, 2009].
In the UK, the Renewable Transport Fuels Obligation (RTFO), which requires
a small percentage of fuel to be renewable, was passed in 2007 [United Kingdom
Statutory Instrument No. 3072, 26 October, 2007]. The RTFO requires renewable
fuel to comprise 2.5% of supply in 2008-2009, which has been met, growing to
3.5% in 2010-2011. For these fuels, reductions in GHG emissions, which are out-
put using a ‘carbon calculator’, are supposed to be 40% below conventional fuels
in 2008-2009 and 50% below in 2010-2011 [Renewable Fuels Agency, Accessed
21 October, 2009].
6
1.2 The transport sector
Reaching political agreements for emissions reductions is only one challenge in
the decarbonization process. Each sector requires its own technological solutions,
and not all sectors consume energy, or even types of energy, equally. The transport
sector proves a particular challenge because its growth rate has been high during
the last 35 years. Figure 1.1 shows the global end use consumption of oil by sector.
In terms of oil consumption, the transport sector has grown by the largest amount
of any sector in recent decades [International Energy Agency, 2008]. In terms of
total energy consumption (not just oil), during the same period, growth globally
(206%) and in OECD countries (182%) has also been high.
1976 1981 1986 1991 1996 2001 20061971
Non-energy useOther sectors TransportIndustry
0
500
1000
1500
2000
2500
3000
3500
by sector (Mtoe)
[M
to
e]
Figure 1.1. Global end use oil consumption [Mt oil equivalent] by sector from Inter-
national Energy Agency [2008]. Other sectors includes agriculture, public services,
and residential use.
A high growth rate is one of four factors which make the transport sector par-
ticularly difficult to decarbonize. A second contributing feature is the complicated
infrastructure which is required to support a vehicle fleet. If future fuels differ
widely from their current analogues, both the fleet itself (the cars and trucks) may
need to be replaced, and the distribution network (the pipelines, tankers, storage
systems, and refueling stations) may also need to be changed. Both would require
7
time, capital investment, and energy. These costs—temporal, financial, and ener-
getic—are the second reason that the transport sector is hard to change.
The sociological barriers for decarbonizing transport are especially challeng-
ing. A centralized sector such as power generation is rather easier to transition;
as long as electricity continues to flow with minimal change in cost, the end user
is generally satisfied. In contrast, the requisites for maintaining status quo in the
transport sector are more stringent, and include rapid refueling, certain power and
speed expectations, and maintenance of a 300-500 mile driving range. The power
and driving range both rely on energy density; that is, a fuel system must supply
enough energy in a small enough volume to be carried on board individual ve-
hicles. These phenomena contribute to the technological barriers in creating the
future transport systems.
Large growth, the cost of infrastructure change, and sociological expectations
are three parts of the challenge that faces the transport sector. In addition, a so-
lution must be technologically feasible. This question of technical capability and
feasibility was addressed in a study by Pacala and Socolow [2004], in which the
authors proposed that stabilisation of atmospheric CO2 concentrations would re-
quire a suite of carbon-reducing technologies. Figure 1.2 shows their ‘stabilization
triangle’. Each of the ‘wedges’ shown in the bottom of Figure 1.2 represent a
type of technology that could, by 2054, reduce carbon emissions by one Gt per
annum. The authors proposed 15 wedges, of which seven are required in order to
stabilise CO2 concentrations at 500±50 ppmv. Only two of these proposals relate
to the transportation sector: hydrogen fuel derived from wind power used in fuel
cell vehicles, and biofuels. With both hydrogen technology [Solomon and Baner-
jee, 2006] and fuel cells [Romm, 2006] still not ready for large scale production
or use, not to mention infrastructure development costs, biofuels are one of the
only potential options for reducing carbon emissions from the transport sector in
the short to medium term. Biofuels can be carbon neutral, but only if the carbon
absorbed during cultivation equals the carbon emitted during combustion [Tilman
8
Figure 1.2. Stabilisation wedges from Pacala and Socolow [2004]. A) The top curve
represents a business as usual path, and the lower curve represents an emissions sta-
bilisation path at 500 ppmv CO2. The green area represents the area of avoided car-
bon required in order to stabilise. B) The stabilisation triangle is divided into seven
‘wedges’, each of which reaches 1 Gt C yr−1 by 2054.
et al., 2009; Sustainable biofuels: prospects and challenges, 2008]. Pacala and
Socolow [2004] also described another possible wedge for future carbon savings:
increased afforestation and decreased deforestation. Because the carbon storage
capacity of agricultural land is lower than forests [e.g. Foley et al., 2005; Jain and
Yang, 2005], if we cut down the forest to plant biofuel crops, the carbon savings
are lost.
9
Biochemical 
Conversion
Ethanol
Biodiesel
Thermochemical 
Conversion
Pryolysis
Gasification
Liquefication
Syn-Oil
Syn-Gas
Bio-chemicals
BIOFUELS
Hydroysis,
Fermentation
Rening
Rening
Synthetic
 diesel Synthetic
 petrol
Transesterication
Figure 1.3. Schematic of biofuel production through biological and thermochemical
processing, after Demirbas [2007a].
1.3 An introduction to biofuels
Biomass energy is the oldest form of energy harnessed by humans. Primary biomass,
such as charcoal or wood, is still used throughout a large portion of the develop-
ing world for energy. This type of primary biomass energy contributed more than
10% to global energy supply in 2005 [Anderson et al., 2004]. In Sub-Saharan
Africa, combustion of primary biomass and waste comprises 61.5% of energy sup-
ply [Anderson et al., 2004]. Combustion of this type of primary biomass energy
has a strong effect on the atmosphere [e.g. Crutzen and Andreae, 1990; van der
Werf et al., 2004; Jain et al., 2006; Aghedo et al., 2007; Ziemke et al., 2009, and
references therein]. These types of fuels, however, cannot fuel transportation.
On the other hand, secondary, or processed, biofuels are viable liquid fuels
for transport. The potential efficacy and uses of a particular biofuel depend both
on its feedstock and its conversion method. Biomass converts approximately 1%
10
OH O
O
R
(a) (b)
Figure 1.4. Chemical structure of a) ethanol and b) biodiesel.
of incoming solar radiation into usable energy [Schiermeier et al., 2008]. After
biomass cultivation, two major branches of conversion methods exist [Demirbas,
2007a]: biochemical and thermochemical. Figure 1.3 summarises the branching
and families of biofuel production. Biological processes usually deliver a fairly
pure product, although the process may be slow; alternatively, thermochemical
processes are more rapid but yield heterogenous products that can be difficult to
distill or refine [Sustainable biofuels: prospects and challenges, 2008]. Thermo-
chemical products are molecularly similar to their conventional fuel analogues,
while biological products are different. Because currently the most common bio-
fuels for the transportation sector are produced biologically, they are the focus of
this thesis. The chemical structure of the two main biologically produced biofuels,
ethanol and biodiesel, and shown in Figure 1.4. In biodiesel, the R group is a chain
of 14-24 carbons, depending on feedstock, and has between 0 and 3 double bonds
[Ma and Hanna, 1999; Pinzi et al., 2009].
1.3.1 Biomass potential
The feasibility of large scale biofuel deployment depends not only on the vehicle
technologies and distribution pathways, but also on the amount of biomass that
could be produced. Berndes et al. [2003] reviewed a number of studies of global
biomass potential; Table 1.1 shows a summary of biomass potential estimates for
a range of years from Berndes et al. [2003] and Moreira [2006]. The range of po-
11
Table 1.1. Estimates of future energy supply [EJ (1018 joules) yr−1] from biomass
summarized from Berndes et al. [2003] and Moreira [2006].
Year 1990 2020-30 2050 2100
Hall et al. [1993] 31
Johansson et al. [1993] 62 78
Battjes [1994] 47
Williams [1995] 30 38 46
Fujino et al. [1999] 88
Houghton et al. [2001] 441
Fischer and Schrattenholzer [2001] 350-450
Yamamoto et al. [2001] 84 265
tential biomass estimates is large, from 31 to 88 EJ (1018 joules) yr−1 in 1990, and
from 38 to 450 EJ yr−1 in 2050. The large majority of the variability in projections
is due to differences in estimates for global land cover and the potential yield or
production capacity from a unit of land [Berndes et al., 2003]. In addition, as-
sumptions about food productivity and availability, degraded land availability, and
competing land use types contribute to variability in potential biomass estimates.
The total global energy demand in 2004 was 308.8 EJ yr−1 , and the global trans-
portation demand was 82.4 EJ yr−1 [Sims and Schock, 2007]. From the estimates
listed in Table 1.1, it is clear that biomass could supply a significant part of the
transportation energy demand; by some estimates, biomass could cover the energy
required by the entire global fleet.
I should note that the energy harnessed from biomass does not have to take the
form of liquid fuel for transport. A number of studies have suggested that using
biomass to generate combined heat and power (CHP) would be more efficient than
producing liquid fuels for transport. The use of this type of energy by the transport
sector would require development of electric vehicles and therefore advanced bat-
tery technology. Estimates of the increase in efficiency for the CHP-electric fuel
system compared to liquid fuel production range from 33% [Sims et al., 2006], to
100% [Ohlrogge et al., 2009] and 112% [Campbell et al., 2009].
12
1.3.2 Ethanol
1.3.2.1 Introduction
The use of ethanol as a liquid fuel for transport is as old as the motor vehicle; Henry
Ford outfitted the Model T to be a flex fuel vehicle that could run on alcohol-based
fuels [Solomon et al., 2007a]. For most of the twentieth century, however, mineral
fuels proved to be cheaper and easier to produce, and alcohol-based fuels never
took a significant part of the market. The first country to revive the ethanol market
was Brazil, which developed their ‘Proalcool’ program in 1975 in the wake of the
1973 OPEC oil embargo [Solomon et al., 2007a]. Today, all liquid transport fuel
in Brazil contains at least 24% ethanol [Moreira, 2000; Solomon et al., 2007a]
and ethanol comprises approximately 40% of all transport fuel used in the country
[Solomon et al., 2007a].
Ethanol production in the United States, the other major world producer, has
grown somewhat more gradually, aided by a series of tax exemptions beginning in
1978 [Solomon et al., 2007a]. Ethanol forms a relatively small percent of the liquid
fuel for transport used in the United States—E85 forms 1% of the alternative fuel
market, and 0.02% of the entire market. E85 is a mixture of 85% ethanol and
15% conventional gasoline. Ethanol in ‘gasohol’ (a mixture of 10% ethanol and
90% gasoline by volume, or E10) forms a slightly larger section of the total end
use market for transport fuel (2%), but it is still relatively small. Despite this, the
United States is the world’s largest producer of fuel ethanol, producing 9 billion
gallons in 2008 [Renewable Fuels Association, Accessed 15 September, 2009]. In
the same year, Brazil produced 6.5 billion gallons of ethanol for transport.
1.3.2.2 Feedstocks
Sugar- and starch-producing crops comprise the main feedstocks of most cur-
rent ethanol production. The main sugar crops are sugarcane, sugar beet, and
sweet sorghum [Demirbas, 2007a; Sustainable biofuels: prospects and challenges,
13
Table 1.2. Energy balance of ethanol fuel-feedstock systems, as in Goldemberg
[2007]. Energy balance is the renewable output energy divided by the fossil fuel
input energy, such as transport, fertilizer, and industrial processing.
Feedstock Energy Balance
Sugarcane 10.2
Sugar beet 2.1
Corn 1.4
Cellulose 10.0
2008]. Starchy crops include corn, sorghum, and wheat, of which corn is the most
prevalent in present day production [Demirbas, 2007a].
Cellulosic ethanol is derived from the sugars within the cell walls of the en-
tire biomass of a plant (see section 1.3.2.3). Instead of optimizing growth and
crop choice for maximum sugar output, cellulosic ethanol crops are chosen for
their total biomass output. Perennial grasses (either single species or polycultures)
and fast growing trees (short rotation coppice) are examples of dedicated energy
crops that could be used to produce cellulosic ethanol [Tilman et al., 2006; Sustain-
able biofuels: prospects and challenges, 2008]. Cellulosic ethanol could also be
extracted from waste or co-products, such as the extra biomass from cereal produc-
tion, forestry waste, or municipal biomass waste [Sustainable biofuels: prospects
and challenges, 2008]. Because development of the sugar extraction techniques
from cell walls is still underway, cellulosic ethanol is sometimes referred to as
‘second generation’ ethanol.
Each of these feedstocks differs in its energy balance, or total output of renew-
able energy per unit of input energy. Table 1.2 lists a summary of common ethanol
feedstocks and their energy balances. Ethanol derived from sugarcane, for exam-
ple, is very favorable in its energy balance and produces over 10 times the amount
of energy required as inputs. One reason for this is that a high proportion of the
input production is powered from the waste products within the sugarcane produc-
tion cycle, called bagasse [Moreira, 2000; Wang et al., 2008]. Bagasse, because
14
it is also grown during the cultivation period (and pulls down atmospheric carbon)
does not count as an energy input. Although sugar beet and corn have positive en-
ergy balances, they are not as positive as sugarcane; one reason for this is the starch
and sugar can only be extracted from the fruit of the plant. Cellulosic ethanol also
shows a high return on energy inputs, partly because the entire biomass of the plant
is used.
1.3.2.3 Production
Currently, the most established method of ethanol production is biological fer-
mentation. For sugar-producing crops, the production process involves four or
five steps: milling, pressing, fermentation, distillation, and possibly denaturing
[Solomon et al., 2007a]. A starchy crop, such as corn, requires seven steps: milling,
liquefaction, saccharification, fermentation, distillation, dehydration, and denatur-
ing [Solomon et al., 2007a]. The increased number of steps, particularly the need
for a hydrolysis reaction, decreases the energy balance significantly for starchy
crops versus those which produce sugar directly [Lin and Tanaka, 2006]. Saccha-
rification, the hydrolysis of long starch polymers, is necessary because the yeast
which produces ethanol can only digest sugars. Distillation is also a very energy
intensive step, which would be improved by increasing the concentration of ethanol
produced during the fermentation process. The obstacle to this is the alcohol tol-
erance of the bacteria used in fermentation, which is currently approximately 18%
[Sustainable biofuels: prospects and challenges, 2008].
A crop is only a small part of the entire biomass of a plant. Sugar polymers exist
throughout the entire plant; cellulose is a polymer of six-carbon glucose sugars
that constructs the majority of cell walls. It is intertwined both with hemicellulose,
which is a branched, shorter polymer of five-carbon sugars, and with lignin, a stiff
polymer which crosslinks the former two [Service, 2007]. In order to use the full
biomass of a plant as a feedstock for ethanol production, lignin must be removed
before fermentation can begin. This difficult deconvolution is currently the subject
15
of much research [Service, 2007; Weng et al., 2008].
1.3.3 Biodiesel
1.3.3.1 Introduction
In an early public display of the compression ignition engine, Robert Diesel used
peanut oil as fuel [Shay, 1993]. Despite early testing of a variety of vegetable oils in
diesel engines [Shay, 1993; AltIn et al., 2001], concentrated interest in generating
fuel from vegetable oils did not occur until the 1970s, in response to the OPEC oil
embargo [Shay, 1993]. Reducing carbon emissions has sparked renewed interest
in diesel derived from biomass. In the United States, biodiesel constitutes 5.3%
of alternative fuels, and 0.14% of total transport fuel. Biodiesel production in the
United States sums to 1.8 billion litres in 2008 [Sims et al., 2008]. In the same
year, the EU produced 5.8 billion litres [Sims et al., 2008].
1.3.3.2 Feedstocks
Biodiesel can be derived from a number of oil-producing crops. The most com-
monly used feedstocks are soybean, palm, sunflower, safflower, cottonseed, rape-
seed (canola), and peanut [Demirbas, 2007b]. Because of many centuries of use in
the food industry, the oil produced per hectare from these crops has been optimized
to maximize output. A more recently cultivated plant, Jatropha, which produces
the physic nut, is also a potential feedstock for biodiesel production [Sims et al.,
2008]. Jatropha growth has not been fully optimized for maximum production, and
its successful use in biodiesel production is therefore considered to be ‘second gen-
eration’, similar to cellulosic ethanol. Microalgal production of oils is early in the
development process, and is also considered to be a ‘second generation’ biofuel.
There are no commercial algal biodiesel production facilities [Lardon et al., 2009],
although the energy company ExxonMobile has recently funded Craig Venter at
16
Table 1.3. Energy balance of biodiesel fuel-feedstock systems. Energy balance is the
renewable output energy divided by the fossil fuel input energy.
Feedstock Energy Balance Source
Soybean 1.93 Hill et al. [2009]
Palm 4.8 Ya´n˜ez Angarita et al. [2009]
Microalgae 0.98-3.5 Lardon et al. [2009]
Rapeseed 1.93 Patterson et al. [2008]
Synthetic Genomics to build one [Service, 2009]1. The most common feedstock
for biodiesel production in the United States is soybean [Hill et al., 2006], while in
Europe it is rapeseed [Sustainable biofuels: prospects and challenges, 2008].
As with ethanol, each biodiesel feedstock and each pathway has its own en-
ergy balance. Table 1.3 shows some example energy balances for soybean, palm,
and microalgae derived biodiesel. The most positive energy balance is for palm
derived biodiesel, followed by microalgae. The range for microalgae estimates is
large primarily because of energy consumption levels for the different oil extrac-
tion techniques [Lardon et al., 2009]. The total net energy balance for biodiesel, no
matter what the input feedstock, is never as high as for ethanol. One of the reasons
for this is that only the fruit of a plant can be used to extract oils, while the entire
biomass of a plant can contain sugars (e.g. sugarcane).
1.3.3.3 Production
There are four main stages in biodiesel production: extraction, transesterifica-
tion, phase separation, and distillation [Demirbas, 2007b; Sims et al., 2008]. Oil-
producing crops produce triglycerides. The main transformation step for biodiesel
is the removal of the glycerin backbone and replacement with a methyl group. For
one molecule with three different fatty acid chains, this will produce three different
fatty acid methyl esters (see Figure 1.4). The removal of glycerin is the bulk of the
1Confirmed by personal communication with Juan Enriquez, co-founder of Synthetic Genomics;
July, 2009 in Oxford, UK.
17
Cultivation
CO2
CO2
Transformation
Combustion
Inputs: 
fossil fuels
 minerals
 fertilizers
 equipment
 land
Outputs:  
heat
GHG emissions
other emissions
Figure 1.5. Schematic of the three stages of a biofuel life cycle including inputs and
outputs, adapted from Sustainable biofuels: prospects and challenges [2008].
phase separation step [Ma and Hanna, 1999]; glycerol is also a valuable coproduct
used in household products such as cosmetics and soap.
1.4 Life cycle of a biofuel
The life cycle of a biofuel can be broken down into three distinct stages: cultiva-
tion, transformation, and combustion. Figure 1.5 shows a schematic of the inputs
and outputs from a biofuel life cycle, each phase of which will have a unique en-
vironmental and atmospheric footprint. A typical life cycle assessment attempts
to quantify the energy and environmental inputs and outputs over the entire life-
time of a product. For fuel-feedstock systems, LCA is also sometimes described as
18
‘well-to-wheels’ analysis.
The full analysis of a fuel-system life cycle is complex, and depends both on the
fuel pathway and the vehicle technology (sometimes also known as a ‘propulsion
system’). A number of LCA studies were reviewed by MacLean and Lave [2003],
who reported ranges of energy balance and GHG estimates from the literature. The
authors found that for fuel pathways and vehicle propulsion systems that have been
around for a long period of time, both the energy and greenhouse balance are well
constrained to a fairly small range. Newer and future fuels and systems, however,
have a much broader range; for example, ethanol derived from cellulose is reported
to have a range of greenhouse gas emissions of -85 to +14 g CO2 equivalent per
MJ of fuel energy [MacLean and Lave, 2003].
1.4.1 Phase I: Cultivation
There are two main types of feedstock for biofuels: dedicated energy crops and
waste products [Sustainable biofuels: prospects and challenges, 2008]. The tech-
nology to use waste products (especially sugar extraction from cellulose) efficiently
is not yet available, so I will focus on dedicated energy crops.
The growth and cultivation of crops requires both energy and equipment, and
releases a particular set of emissions into the atmosphere. Life cycle analyses have
considered the energy and GHG cost of fertilizer production, machinery manu-
facturing, and direct energy needed to cultivate and store the feedstock crop [Wu
et al., 2006]. For an oily crop, the cultivation stage can also include crude oil ex-
traction on site, and yield from this extraction process [Huo et al., 2008]. Trace gas
emissions associated with the cultivation stages include soil NOx and volatile or-
ganic compounds (VOC) emitted by the plant species themselves. Biogenic VOC
emissions from the cultivation step are a focus in chapters 4 and 5.
Although emissions of reactive trace gases are not usually considered in phase
I of the biofuel life cycle, agricultural emissions of the greenhouse gas N2O emis-
sions often are [MacLean and Lave, 2003]. Crutzen et al. [2008] considered the
19
impact of increasing N2O emissions resulting from large scale biofuel use on
the atmosphere by using an fertilizer application/atmospheric release rate of 3-
5%, which was higher than previous estimates. The authors found that the in-
creased emission of N2O, which is a greenhouse gas (and also destroys strato-
spheric ozone), outweighed any carbon savings gained by converting to biofuels.
Soybean biodiesel and corn ethanol were the fuels considered.
While energy, greenhouse gas emissions, and some trace gas emissions have
been considered in LCAs to date, these factors do not necessarily include all of the
potential environmental impacts of biofuel cultivation. For example, cultivating
energy crops would require water [Schilling et al., 2008]. On the other hand, pos-
itive effects could include improved soil health and increased biodiversity [Rowe
et al., 2009].
1.4.2 Phase II: Transformation
The transformation step involves the conversion of a feedstock (in the case of
ethanol) or crude mixture (in the case of biodiesel) into a fuel that is usable in
vehicle engines. Calculating the contribution of the second stage of a biofuel life
cycle to the entire LCA requires the inclusion of a number of smaller steps. These
include the energy to make and transport needed solvents and chemicals, wet or
dry milling (for ethanol), feedstock transportation, fuel transportation, and the en-
ergy to carry out industrial chemical reactions [Wu et al., 2006; Huo et al., 2008].
Wang et al. [2007] examined the impact of the processing plant used to transform
corn to ethanol, and found that the energy balance and GHG emission were af-
fected significantly. The total life cycle can also be influenced by the production
of coproducts at the transformation stage. For ethanol, a common coproduct is
animal feed [Wang et al., 2007; Sims et al., 2008], while for biodiesel, a common
coproduct is glycerol [Huo et al., 2008]. Phase II of a life cycle assessment also
usually includes emissions of criteria air pollutants, whose calculation usually de-
pends on reported emissions factors [e.g. Wu et al., 2006]. These trace gases may
20
include a number of typical anthropogenic industrial emissions such as CO, NOx,
anthropogenic VOCs, and toxic aromatic compounds.
1.4.3 Phase III: Combustion
The end use, or Phase III, of the biofuel life cycle is combustion in a vehicle engine.
This stage usually encompasses transport from a distribution centre to a refueling
station, vehicle operation, and sometimes vehicle service. This stage also includes
the evaporative emissions that occur due to leaks and errors in the distribution
process [Jacobson, 2007; Huo et al., 2009]. Trace gas emissions linked to the
combustion stage may include CO, NOx, and vehicle tailpipe VOCs. Because of
the widespread nature of mobile vehicle emissions, the footprint of combustion
is the largest of the three phases. A more in depth exploration of the reason for
the different emissions profile of ethanol and biodiesel versus their conventional
counterparts can be found in section 6.1.2.
As I describe further in the next chapter, the presence of NOx is often a deter-
mining factor in whether a local chemical regime produces or destroys ozone [e.g.
Liu et al., 1987, and section 2.1.2]. The emission of NOx during the biofuel life-
cycle is therefore an important aspect of their potential impact on the atmosphere.
Variations in fuel quality and combustion characteristics lead to changes in tailpipe
NOx emissions. NO formation in a combustion engine normally proceeds through
thermal formation by the Zeldovich mechanism [Zeldovich, 1946]. The reactions
proceed at very high temperatures (2000+ ◦C) [Hanson and Saliman, 1984], and
are initiated by reaction of atomic oxygen with molecular nitrogen:
O+N2 → N+NO (1.1)
which is the rate limiting step. This leads to reaction of atomic nitrogen with
molecular oxygen:
N+O2 → O+NO, (1.2)
21
or reaction with OH:
N+OH→ H+NO (1.3)
both of which form a second molecule of NO. High engine temperatures are highly
correlated to changes in NO formation and emission [Hanson and Saliman, 1984].
A much smaller fraction of tailpipe NOx emission can also arise from ‘prompt
NOx’ (a mechanism which proceeds through the formation of hydrogen cyanide)
or from ‘fuel NOx’ (from combustion of nitrogen within the fuel content) reactions
[Fernando et al., 2006].
1.5 Summary
In this chapter I have explored the potential for biofuels to form part of a future
renewable energy mix. This has been recognized in a series of political agree-
ments, which I documented in the first section of this chapter. I then described the
particular difficulties of solving the transportation energy problem: high growth
rates in the transport sector, inertia in the infrastructure of the system, sociological
requirements, and technological feasibility.
After introducing biofuels, I described the concept of biomass potential for
which there is currently a high level of uncertainty in the literature. I also went
through the two major types of biofuels, ethanol and biodiesel, and documented
their major feedstocks and production pathways. I finished this chapter with a
discussion of the three phases of a biofuel life cycle—cultivation, transformation,
and combustion—and the potential trace gas emissions profile of each.
In the next chapter, I describe the atmospheric chemistry that these emissions
have the potential to perturb.
22
Chapter 2
Biofuels: Potential Atmospheric
Impacts
Emissions of NOx, CO, and VOC from the biofuel life cycle could have impli-
cations for air quality and tropospheric chemistry. In section 2.1, I explore the
initiation, propagation, and termination steps of radical chemistry in both a pol-
luted and unpolluted environment, focusing on the formation and destruction of
ozone. I then describing the nonlinearities of the chemical system, and the differ-
ences between a VOC-sensitive and a NOx-sensitive chemical regime. In section
2.2, I turn to isoprene, the focus of Chapters 3 and 5. The second half of this
chapter examines the chemistry and emission of isoprene into the atmosphere.
2.1 Tropospheric chemistry
The troposphere is the lowermost part of the atmosphere which extends from the
surface to between eight and eighteen kilometres, depending on latitude. It is char-
acterized by rapid meteorological processes which create a well-mixed environ-
ment, significant presence of water vapour, and high temperatures. The principal
components of the troposphere are nitrogen (78%) and oxygen (21%), but a num-
ber of trace gas constituents play important chemical roles.
23
stratosphere-
troposphere
exchange
OH
HO2, RO2
NO NO2emissions
CO, CH4, VOC
deposition
O3photon
photon
OH
CO, O3
O3
HO2
chemical
production
chemical
destruction
Figure 2.1. Schematic of the tropospheric ozone budget, showing the two production
processes (stratosphere-troposphere exchange and chemical production) and the two
loss processes (deposition and chemical destruction). Based on Fowler et al. [2008]
and Young [2007].
Ozone is of particular importance because of its role in initiating the radical
oxidation chemistry that dominates in the troposphere. Besides its chemical role
(described below in sections 2.1.1 and 2.1.3), ozone is also a radiatively active gas
[e.g. Gauss et al., 2003; Shindell et al., 2006a], and can be detrimental to both hu-
man [e.g. Wilkins et al., 2001; Jerrett et al., 2009] and crop health [e.g. Long et al.,
2005; Hayes et al., 2007; Mills et al., 2007; van Dingenen et al., 2009]. Figure 2.1
shows a schematic representation of the ozone budget in the troposphere. Ozone
production is dominated by the chemical term, the model mean of which was be-
tween 3448 and 5110 Tg yr−1 in a number of recent model intercomparisons [see
Wild, 2007, and references therein]. Ozone production depends on relative concen-
trations of NOx and VOC, which I describe further in section 2.1.3. An additional
source of ozone is the flux from the stratosphere, which has been modelled be-
24
tween 552 and 765 Tg yr−1 [Wild, 2007]. Destruction of ozone is also dominated
by chemical processes, and has been estimated to be between 3435 and 4668 Tg
yr−1 [Wild, 2007]. The final term in the ozone budget is deposition, or physical
loss of ozone to the surface of the earth. Deposition has been modelled at between
818 and 1003 Tg yr−1 [Wild, 2007].
The chemistry of ozone and other tropospheric constituent gases is complex
and nonlinear; Figure 2.2 presents a visual summary of this chemistry. What fol-
lows is a descriptive summary of the key reactions that govern the chemistry and
composition of trace gases the troposphere.
2.1.1 Initiation
The troposphere constitutes a sort of low temperature oxidation chamber which
proceeds by radical chemistry. The generation of radicals arises from the photo-
chemical lability of a number of trace gases. The major source of radicals in the
troposphere is the photolysis of ozone, which proceeds through two channels:
O3 + hν(λ ≤ 1150nm)→ O(3P) + O2, (2.1)
O3 + hν(λ ≤ 310nm)→ O(1D) + O2. (2.2)
The ground state O(3P) generated from the longer wavelength photolysis channel
reforms O3 by:
O(3P) + O2 +M→ O3 +M. (2.3)
A large portion of theO(1D) formed through the short wavelength photolysis chan-
nel is collisionally quenched:
O(1D) +M→ O(3P) +M (2.4)
25
OH HO2
RORO2
O3
HCHO
H2O2
RH
ROOH
HNO3
NO2
NO
R'CHO
O3
NO2
NO
O2
O2 O2
H2O
HO2
HO2
NO2
photon
photon
photon
photon
O3
CO, O2
(carbonyl)
Figure 2.2. A visual summary of tropospheric chemistry as described in the text,
adapted from Jenkin and Clemitshaw [2000] and Young [2007]. Radicals are pro-
duced through two main initiation steps (green), and cycle through the main oxidation
radical propagation steps (black). In the absence of NOx, ozone is consumed (pur-
ple), whereas in the presence of NOx, ozone is produced (red). The cycle ends with
radical-radical chain termination reactions (blue).
26
after which the O(3P) reforms ozone as in reaction 2.3. The small fraction of
O(1D) remaining can react with water vapour to generate the hydroxyl radical:
O(1D) + H2O→ 2OH. (2.5)
The formation of OH is therefore dependent on the concentration of water vapour,
(which is linked to temperature and therefore climate), the concentration of ozone,
and the photolysis frequency.
A more minor channel of radical production results from the photochemical
lability of carbonyl species. The main radical-producing channel of photolysis
of formaldehyde, the most abundant carbonyl in the atmosphere [Levy II, 1971;
Carlier et al., 1986], is:
HCHO+ hν(λ ≤ 338nm)→ H+HCO (2.6)
after which the radicals H and HCO go on to form HO2 by:
H+O2 +M→ HO2 +M (2.7)
HCO+O2 → CO+HO2 (2.8)
HO2 and OH are rapidly interconverted; the two together comprise the HOx fam-
ily. Once generated, these radical species can participate in a host of reaction cycles
in the troposphere.
2.1.2 Propagation
Once formed, the hydroxyl radical can react with numerous trace gases. Princi-
pally, OH is consumed by the oxidation of CO:
CO+OH→ CO2 +H (2.9)
27
which accounts for 70% of OH reactivity in the troposphere [e.g. Wayne, 2000].
The H atom then goes on to form HO2 by:
H+O2 +M→ HO2 +M (2.7)
A second major fate for the hydroxyl radical is reaction with a volatile organic
compound (VOC). Though there are numerous VOCs in the atmosphere, each with
a separate oxidation mechanism, a number of salient features are common to all.
For this reason, I take the most abundant and simple VOC, methane, as a repre-
sentative example. The oxidation of methane by hydroxyl begins with hydrogen
abstraction:
CH4 +OH→ CH3 +H2O, (2.10)
after which oxygen rapids adds to the methyl radical to form methylperoxy:
CH3 +O2 +M→ CH3O2 +M. (2.11)
The formation of methylperoxy from methane oxidation can lead to the oxidation
of NO, if it is present:
CH3O2 +NO→ CH3O+NO2. (2.12)
The resulting methoxy radical can react with O2 to regenerate HOx by:
CH3O+O2 → HCHO+HO2 (2.13)
where the production of formaldehyde can lead to further radical production through
photolysis. Methylperoxy is a representative radical product of VOC, from which
NO oxidation is common. Using R to represent a generic hydrocarbon, this process
can be written as:
RO2 +NO→ RO+NO2. (2.14)
28
An analogous reaction exists for HO2 (formed during CO oxidation, reaction 2.9),
from which OH is recycled:
HO2 +NO→ OH+NO2; (2.15)
though in the absence of NO, HO2 can react with O3:
HO2 +O3 → OH+ 2O2. (2.16)
The two reactions above (2.14 and 2.15), in which NO2 is generated from the
oxidation of NO, form the basis for ozone formation in the troposphere, as NO2 is
rapidly photolysed:
NO2 + hν(λ ≤ 400nm)→ NO+O(3P) (2.17)
from which O(3P) goes on to form ozone via reaction 2.3.
The oxides of nitrogen have a particularly important role in the atmosphere be-
cause of their role in reducing peroxy radical species. During this process, NO2 is
formed (reactions 2.14 and 2.15), which leads to the formation of ozone (reactions
2.17 and 2.3). In the absence of these conversion processes, NO will react with
ozone by:
NO+O3 → NO2 +O2. (2.18)
The sum of reactions 2.3, 2.18, and 2.17 forms the photostationary state, which is
a null cycle. The presence of radical species, such as RO2 and HO2, perturb this
cycle and lead to the production of ozone. This ozone producing regime proceeds
as:
CO+OH→ CO2 +H (2.9)
H + O2 +M→ HO2 +M (2.7)
HO2 +NO→ OH+NO2 (2.15)
29
NO2 + hν → NO+O(3P) (2.17)
O(3P) + O2 +M→ O3 +M (2.3)
with the net reaction:
CO+ 2O2 → CO2 +O3. (2.19)
In contrast, in a cleaner environment with lower concentrations of NOx, ozone
destruction may occur by:
CO+OH→ CO2 +H (2.9)
H + O2 +M→ HO2 +M (2.7)
HO2 +O3 → OH+ 2O2 (2.16)
with the net reaction:
CO+O3 → CO2 +O2 (2.20)
The determining factor between these two regimes is the presence or absence of
the oxides of nitrogen [e.g. Liu et al., 1987; Chameides et al., 1992].
2.1.3 Chain termination
The radical chemistry that prevails in the troposphere is terminated through two
main routes. These chain terminating (radical-radical) steps either form hydroper-
oxides:
HO2 +HO2 → H2O2, (2.21)
and other peroxides:
HO2 +RO2 → ROOH, (2.22)
or nitric acid:
OH+NO2 +M→ HNO3 +M. (2.23)
30
A more minor channel of nitric acid formation was recently reported by Butkovskaya
et al. [2005, 2007], in which HO2 reacts with NO:
HO2 +NO→ HNO3 (2.24)
which competes with the recycling of NO2 in reaction 2.15, but is more impor-
tant in the upper troposphere [Cariolle et al., 2008]. Both hydroperoxides [Zhang
et al., 2003b] and nitric acid are extremely soluble in water and are rapidly and
permanently removed by wet or dry deposition [e.g. Wesely and Hicks, 2000, and
references therein]. Nitric acid is also subject to attack by OH, a more minor chan-
nel, and photolysis, which is more important in the upper troposphere.
Ozone production depends on the concentrations of both NOx and VOC in
the local environment. Figure 2.3 shows an ozone production map for a range of
NOx and VOC emission rates.1 This ‘topography’ of ozone production shows that
as both NOx and VOC increase, ozone production also increases (along the dark
blue line, the maximum ‘ridge’ of ozone production). When emissions of either
NOx or VOC are particularly high, increasing the emissions of the same species
decreases ozone production. This occurs largely through the chain termination
processes, formation of nitric acid or peroxides, described above. The dominating
chain termination step occurring in a particular location can be indicative of the
type of chemical regime [Sillman, 1995].
In a ‘NOx-sensitive’ environment, the addition of NOx to the chemical system
increases ozone production; this is depicted on the bottom half of Figure 2.3. In
these environments, concentrations of peroxy radicals are high, and the formation
of peroxides (reactions 2.21 and 2.22) competes with recycling of NO by reac-
tions 2.14 and 2.15. High fluxes through the reactions which form peroxides are
therefore indicative of a NOx-sensitive regime.
In contrast, in a ‘VOC-sensitive’ chemical regime, the addition of VOC in-
1Although a version of this graphic appears in Sillman and He [2002], the slightly edited ver-
sion shown here was accessed from Sanford Sillman’s website: http://www-personal.umich.edu/ sill-
man/ozone.htm.
31
Figure 2.3. Topography of ozone based on NOx and VOC emission rates, from Sill-
man and He [2002]. Ozone isopleths are shown in light blue [ppbv], and the maxi-
mum ‘ridge’ of ozone production is shown in the dark blue dashed line.
32
creases ozone production, and this is depicted on the top half of Figure 2.3. In
these areas, high concentrations of NOx lead to HOx loss via the formation of
nitric acid (reactions 2.23 and 2.24). It is the formation of nitric acid which is in-
dicative of a VOC-sensitive chemical regime. It is noteworthy that in the ozone
production topography shown in Figure 2.3, the way to increase ozone production
most rapidly is to add VOC to a VOC-sensitive environment.
2.1.4 PAN chemistry
In addition to these loss processes, the formation of reservoir species such as PAN
can also play a role in the temporary loss of HOx and NOx [Sillman et al., 1990;
Sillman and He, 2002]. The most abundant of these is peroxyacetyl nitrate (PAN)
[Crutzen, 1979], which is formed by the addition of NO2 to the acetyl radical. PAN
is thermally labile and is in equilibrium with acetyl and NO2:
CH3C(O)O2 +NO2 
 CH3C(O)OONO2. (2.25)
The thermal dependence of the lifetime of PAN is in the range of typical tropo-
spheric temperatures. PAN becomes particularly stable higher up in the tropo-
sphere [Singh and Hanst, 1981; Talukdar et al., 1995; Moxim et al., 1996], where
thermal decomposition does not occur but photolysis may. The increased lifetime
of PAN at these high altitudes allows for the transport of NOx around the globe; as
PAN descends into the warmer regions of the troposphere, NO2 is released as the
equilibrium shifts towards the left side of reaction 2.25.
2.1.5 Nighttime chemistry
Once the sun sets, the formation of OH by photolysis of ozone ceases. At night,
NO3 can abstract a hydrogen, usually from alkenes, producing nitric acid, and
providing an initiation step for radical chemistry. NO3 is photolytically labile, and
as such does not exist during the day. At night, it can be formed from the reaction
33
between ozone and NO2:
O3 +NO2 → NO3 +O2. (2.26)
Once formed, NO3 and NO2 are in equilibrium with dinitrogen pentoxide:
NO3 +NO2 → N2O5 (2.27)
which can be hydrolysed on aerosol surfaces to form nitric acid; this provides a
nocturnal sink for NOx. Ozonolysis of VOCs is an additional initiation step that
can occur at night.
2.2 Isoprene
Isoprene, or 2-methyl-1,3-butadiene, is particularly important for tropospheric chem-
istry because of its high levels of emission and fast reactivity. In this section I
describe the atmospheric chemistry of isoprene, a number of observations of its
emission, and three types of studies that have attempted to estimate its global flux
into the atmosphere.
2.2.1 Isoprene chemistry
Isoprene is a very reactive trace gas. Its reactivity arises from double unsaturation
as well as methyl substitution, which stabilizes the formation of radical products.
Table 2.1 lists the lifetimes of several common VOCs found in the troposphere, and
shows that propane, ethene, and formaldehyde have lifetimes on the order of days.
The isoprene lifetime for reaction with OH is the lowest of any of these VOCs,
with a value of 1.4 hours only.
Initial reaction between one of isoprene’s two pi bonds andOH, plus addition of
molecular oxygen yields a hydroxyperoxy radical (RO2). In the presence of NO,
this can be reduced to a hydroxyalkoxy radical (RO). The formation of HO2 by
34
Table 2.1. Tropospheric lifetimes of various hydrocarbon species for reaction with
the hydroxyl and nitrate radicals. Data are from Atkinson and Arey [2003].
VOC Lifetime OH Lifetime NO3
Propane 10 day 7 yr
Ethene 1.4 day 225 day
Formaldehyde 1.2 day 80 day
α-pinene 2.6 h 5 min
Isoprene 1.4 h 50 min
hydrogen abstraction yields formaldehyde and a first generation carbonyl product;
the highest molar yields of this carbonyl are for the formation of methacrolein
(MACR) and methyl vinyl ketone (MVK) [e.g. Ruppert and Heinz-Becker, 2000,
and references therein]. These first generation carbonyl products are then subject to
attack by OH, particularly because of the presence of a second double bond, which
begins the cycle again. The relative ratio of MVK to MACR can be indicative of
the surrounding chemistry, owing to differences in the rate of removal of each by
reaction with OH and ozone [see e.g. Kuhn et al., 2007]. In addition, the ratio of the
sum of MVK+MACR to isoprene has been shown to correlate to NOx [Biesenthal
et al., 1998] and OH concentrations [Kuhn et al., 2007].
A second and much smaller reaction pathway for isoprene is direct reaction
with ozone through ozonolysis. Ozone can add to either of isoprene’s double bonds
to form an ozonide, which rapidly decomposes to a Criegee intermediate. This
intermediate can then form formaldehyde, methacrolein, and methyl vinyl ketone
[Grosjean et al., 1993; Aplincourt and Anglada, 2003].
2.2.1.1 Isoprene nitrates
Another possible fate for a hydroxyperoxy radical along the isoprene oxidation
chain is the formation of chain terminating species. Most important for ozone
chemistry is the formation of isoprene nitrates, which occurs by collisional sta-
bilization of the peroxy-NO intermediate [Zhang et al., 2003a]. The first step is
35
formation of the excited intermediate:
RO2 +NO→ ROONO∗ (2.28)
which either forms NO2 and the hydroxyalkoxy radical,:
ROONO∗ → RO+NO2, (2.29)
or is collisionally stabilized to form the nitrate:
ROONO ∗+M→ RONO2 +M. (2.30)
These nitrates are formed in yields between 4% and 12% [see Horowitz et al.,
2007, and references therein]. Isoprene nitrates act as a reservoir species for NOx,
which can be reversed by chemical reaction with hydroxyl, or which can form a
permanent loss if dry or wet deposition occurs [Horowitz et al., 2007]. These rates
and are still subject to further research, as isoprene nitrate measurements in the
field are few. The role of isoprene nitrates has also been identified as a weakness
in global tropospheric chemistry modelling [e.g. Wu et al., 2007].
2.2.1.2 Isoprene and aerosol formation
The existence of aerosol derived from natural organic carbon-containing com-
pounds was first noted by Went [1960]. Carlton et al. [2009] reviewed laboratory
findings and model experiments for isoprene secondary organic aerosol (SOA), and
reported that estimates of isoprene SOA remain highly uncertain, with a range of 6
to 30 Tg yr−1. Global SOA has been found to be 27% [Hoyle et al., 2007] to 78%
[Heald et al., 2008] resultant from isoprene oxidation. Paulot et al. [2009] recently
proposed that reaction between the hydroxyhydroperoxide (a product of the first
cycle of isoprene oxidation) and OH can form an epoxide whilst also recycling
OH. The epoxides measured by Paulot et al. [2009] demonstrated rapid uptake
36
onto aerosol seeds. More remains to be done in order to reduce the uncertainty
in isoprene derived SOA formation, but this is an important impact of isoprene on
the atmosphere, as aerosol has consequences for both climate [e.g. Bellouin et al.,
2005; Kanakidou et al., 2005] and human health [Englert, 2004].
2.2.2 Observations of isoprene emission
Discoveries of biogenic isoprene emission occurred in both the United States [Ras-
mussen and Went, 1965] and the former Soviet Union [Sanadze, 1957]. Early flux
measurements confirmed that isoprene is emitted from specific taxa common to
forest ecosystems; early on, these were documented in North America [e.g. Ras-
mussen, 1970] and the tropics [e.g. Rasmussen and Khalil, 1988]. In conjunction
with these early flux measurements, inventories of VOC and isoprene emission
were calculated for the United States [Zimmerman, 1979; Lamb et al., 1987] and
the globe [Mu¨ller, 1992].
Thousands of measurements of isoprene emission have been made for various
plant species. A summary of relevant measurements of species-specific isoprene
fluxes can be found Table 2.2 as reviewed by Kesselmeier and Staudt [1999]. The
data presented are for plants whose emission rates are notably high; also listed are
common crop species that are of potential interest for biofuel production. More
recently, an electronic database of biogenic VOC measurements has also been es-
tablished [Wiedinmyer et al., 2004]2.
The biosynthetic pathway of many of the biogenic VOCs emitted into the tropo-
sphere are similar, and most monoterpenes and isoprene share the same precursor
compound dimethylallyl pyrophosphate (DMAPP) [Fuentes et al., 2000]. Isoprene
synthesis is carried out by an enzyme, isoprene synthase, whose activity is corre-
lated with isoprene emission [Kuzma and Fall, 1993]. Isoprene escapes from a
plant through its stomata [Sharkey, 1996], though the physiological reason for iso-
prene emission still remains unknown. It has been noted that many less evolved
2Enclosure measurements are available online: http://bai.acd.ucar.edu/Data/BVOC/index.shtml
37
Table 2.2. Isoprene flux rates for high-emitting and biofuel relevant species. Emis-
sion factors are standardized to 1000 µmol−1 s−1 incoming photosynthetically active
radiation (PAR) and 30 ◦C and are given in units of µg LDW−1 h−1, or µg isoprene
emitted per gram of leaf dry weight per hour. Data are summarized from Kesselmeier
and Staudt [1999].
Species Familiar Name Flux Rate
µg · LDW−1 h−1
Mucuna Pruriens Velvet Bean 317
Eleais guineesis Palm Oil Tree 172.9
Acacia Nigrescens African Acacia 110
Salix Babylonica Weeping Willow 115
Populus Balsamifera Southern Poplar 100±46
Quercus Alba White Oak 7.8-120
Querus Coccinea Scarlet Oak 40-130
Querus Velutina Black Oak 60-100
Liquidambar styraciflua Sweetgum 63-99
Securinega Virosa Securinega 81
Eucalyptus Globulus Blue Gum Eucalyptus 15-49
Arundo Donax Giant Reed 38.6
Oryza Sativa Rice 0
Secale Cereale Rye 0
Sorghum Bicolor Sorghum 0
Triticum Aestivum Wheat 0
Glycine Max Soy Bean 0
plants emit isoprene [Fuentes et al., 2000, and references therein]. Emission for
thermal protection has been proposed [Sharkey and Singsaas, 1995], as well as
protection from ozone [e.g. Loreto et al., 2001].
2.2.2.1 Isoprene emission, light, and temperature
Although the physiological reason for isoprene emission remains unknown, the
relationship between isoprene emission and light and temperature has been docu-
mented for over thirty years.
The relationship between isoprene emissions and temperature was recorded
by both Rasmussen and Jones [1973] and Tingey et al. [1979]. That emission is
correlated to the activity of isoprene synthase rather than stomatal conductance was
first reported by Jones and Rasmussen [1975]. The activity of isoprene synthase
38
is temperature dependent [Fall and Wildermuth, 1998; Logan et al., 2000], and
isoprene emission generally increases strongly with increased temperature. This
relationship holds until a threshold temperature is reached (approximately 40 ◦C),
when isoprene synthase begins to denature [Fall and Copley, 2000].
The correlation of isoprene emission with radiation was first recognized by
Sanadze [1964], and was clarified by Rasmussen and Jones [1973]. Isoprene emis-
sion requires light to be emitted [Lerdau and Gray, 2003], and increases with in-
creased incident radiation, though the physiological reasons behind this depen-
dence remain unknown. In addition to the correlation between laboratory mea-
surements of emissions and light and temperature, correlation of in situ flux mea-
surements and light and temperature have been reported [e.g. Rinne et al., 2002].
2.2.2.2 Isoprene suppression by CO2
C i (μmol mol–1)
200 400 600 800 1000 1200
Re
lat
ive
 e
m
iss
ion
 ra
te
0.5
1.0
1.5
2.5
3.0
C
Normalised isoprene 
emission rate
Figure 2.4. Empirically deduced relationships between CO2 and isoprene emission
from Possell et al. [2005] (left) and Wilkinson et al. [2009] (right).
That CO2 concentration levels impact emission of isoprene was reported as
early as 1973 by Rasmussen and Jones. The suppression of emission at high CO2
was measured in leaf cuvettes by both Monson and Fall [1989] and Loreto and
Sharkey [1990]. More recently, isoprene emission has been found to be suppressed
when CO2 concentrations were artificially raised in an otherwise open environ-
39
ment [Rosenstiel et al., 2003]. The study of Possell et al. [2005] compiled the
reported values in the literature with new measurements to deduce a simple empir-
ical relationship between CO2 concentration and leaf-scale isoprene emission; this
relationship is shown in Figure 2.4. The inhibition of emission from present day
values (300-400 ppbv) of CO2 to potential future values (600-800 ppbv) was also
reported by Wilkinson et al. [2009], whose emission relationship is also shown in
Figure 2.4. The stimulation of emission at lower ambient concentrations was not
supported by Wilkinson et al. [2009], who linked the relationship between CO2
and isoprene emission to competing processes occurring at the cellular level for
pyruvate, an important molecule in isoprene production. Arneth et al. [2007a]
conducted a review of the potential explanations for CO2 inhibition of isoprene
emission at the plant physiological level.
2.2.3 Estimations of isoprene emissions
Three types of modelling studies have attempted to estimate the global flux of
isoprene into the atmosphere. In this section, I review the empirical, process based,
and satellite studies and the resulting estimations of global isoprene emission.
2.2.3.1 Empirical
The first empirical attempt to estimate global VOC emission was undertaken by
Rasmussen and Went [1965]. The authors reported a value of 438 Tg VOC yr−1
using a simple methodology; they multiplied four factors together which accounted
for above-forest concentration of VOC, the height of the boundary layer, the total
land area covered by VOC-emitting vegetation, and the season of emission. Much
lower estimates for the contiguous United States were reported by Zimmerman
[1979] (65 Tg yr−1) and Lamb et al. [1987] (31 Tg VOC yr−1). In the latter, the
estimate accounted for land use, the temperature dependence of emission, and light
and seasonal dependence. The parameterizations were simple: isoprene emissions
were not allowed after the first frost and before spring, and emission was also not
40
allowed at night.
Global isoprene emission estimates increased in their complexity throughout
the following decade. Rasmussen and Khalil [1988] considered land area, leaf sur-
face area, percent of emitters in an ecosystem type, rate of emission, number of
emission days, and the length of days in their calculation of global isoprene emis-
sion. The authors estimated global isoprene emission to be 397 Tg C yr−1 (450
Tg isoprene yr−1). Turner et al. [1991] used foliar biomass instead of percentage
of emitters, and the exponential temperature dependence of Tingey et al. [1981] to
calculate a value of 251 Tg C yr−1 (285 Tg isoprene yr−1). Using the future dis-
tribution of vegetation from a climate-vegetation model, Turner et al. [1991] also
calculated a future isoprene emission of 318 Tg C yr−1 (360 Tg isoprene yr−1). A
similar estimate was calculated by Mu¨ller [1992], who used vegetation type, the
exponential temperature dependence of Lamb et al. [1987], and leaf biomass den-
sity to estimate total global isoprene emission of 221 Tg C yr−1 (250 Tg isoprene
yr−1). The authors included a calculation of net primary productivity which in-
clude parameterizations for precipitation and temperature, which was then taken to
be proportional to leaf biomass density.
The algorithm of Guenther et al. [1995] improved on these existing method-
ologies. The authors included more accurate emission factors, more nuanced pa-
rameterizations of the emission dependence on light and temperature, and a greater
horizontal (0.5◦×0.5◦) and temporal (monthly) resolution of input parameters than
previous estimates. Their flux estimations were calculated according to:
F = D ·  · CT · CL (2.31)
where F is the flux, D is the foliar density (kg dry matter m−2),  is an emission
factor dependent on the ecosystem type, and CT and CL are empirically derived
activity adjustment factors which account for the emission dependence on tempera-
ture and light, respectively. The foliar density, D, was calculated using net primary
productivity, which was dependent on temperature and precipitation.  was nor-
41
Table 2.3. Literature estimates of total present day global isoprene flux [Tg C yr−1]
from a number of models.
Source Emissions estimate [Tg C yr−1]
Rasmussen and Khalil [1988] 397
Turner et al. [1991] 251
Mu¨ller [1992] 221
Guenther et al. [1995] 503
Wang and Shallcross [2000] 530
Adams et al. [2001] 495
Potter et al. [2001] 559
Levis et al. [2003] 507
Sanderson et al. [2003] 484
Naik et al. [2004] 454
Lathie`re et al. [2005] 402
Shim et al. [2005] 566
Tao and Jain [2005] 601
Lathie`re et al. [2006] 460
Guenther et al. [2006] 529
Arneth et al. [2007b] 410
Lathie`re et al. [2009] 413
malized to incoming photosynthetically active radiation (PAR) of 1000 µmol m−2
s−1 and a temperature of 303.15 K. Five possible  values were assigned according
to the ecosystem type; these emission factors were derived from analysis of twelve
separate measurement studies. The algorithm developed by Guenther et al. [1995]
is the basis for most of the global isoprene flux estimates listed in Table 2.3.
Although the original empirical equation (2.31) remains the basis for the es-
timates listed in Table 2.3, a number of additional factors have been added over
subsequent years. Examples include plant growth stage [Potter et al., 2001], leaf
age [Lathie`re et al., 2005; Tao and Jain, 2005], and the percentage of an emis-
sion flux which escapes from the canopy [Tao and Jain, 2005]. Some amount of
the variability between estimates is due to difference input datasets. The distribu-
tion of ecosystem types has also been estimated from dynamical global vegetation
models (DGVMs) [e.g. Sanderson et al., 2003; Naik et al., 2004; Lathie`re et al.,
42
2005].
The emission model of Lathie`re et al. [2009] is based on the Model of Emis-
sions of Gases and Aerosols from Nature (MEGAN), which was described by
Guenther et al. [2006]. The MEGAN scheme made improvements on the work
of Guenther et al. [1995] by including the impacts of soil moisture, leaf age, and
canopy loss processes on isoprene fluxes. Although Guenther et al. [2006] im-
plemented emission factors () which vary spatially, Lathie`re et al. [2009] use
mean emission factors due to long integration periods. Six ecosystems, or plant
functional types (PFTs) were used. The impact of CO2 according to Possell et al.
[2005] was included in the model calculations. The isoprene emission distributions
used throughout this thesis are from the model of Lathie`re et al. [2009].
2.2.3.2 Process based
A second method for bottom-up estimations of isoprene emission is process based.
These models use the cellular processes which control isoprene emission to calcu-
late their estimates rather than ecosystem level empirical data. So far, three process
based models have been developed to estimate isoprene emission fluxes [Niinemets
et al., 1999; Martin et al., 2000; Zimmer et al., 2003], and their differences are
reviewed in Arneth et al. [2007a]. Arneth et al. [2007b] used the algorithm of Ni-
inemets et al. [1999] to calculate global isoprene emission of 410 Tg C yr−1 using
10 ecosystem types.
Keenan et al. [2009] used the model of Guenther et al. [1995] as well as the
processed based models of Niinemets et al. [1999] and Martin et al. [2000] to
estimate isoprene fluxes from European forests. The authors found that for the
present day calculation, the choice of emissions model did not significantly affect
the results; they reported 1.03 Tg C yr−1 isoprene emission from European forests.
For the future, however, the models diverged in their projections of total isoprene
emissions; the model of Martin et al. [2000] calculated an isoprene flux (2.36 Tg
C yr−1) almost 50% higher than the model of Niinemets et al. [1999] (1.58 Tg C
43
yr−1).
2.2.3.3 Satellite derived
Finally, a top-down method for isoprene flux calculations has been achieved using
satellite data. Formaldehyde is a product of isoprene oxidation and is detectable
by remote sensors. Using the yield of formaldehyde, which is calculated using a
model chemical mechanism, and the formaldehyde data as measured by the Global
Ozone Monitoring Experiment (GOME) satellite, Palmer et al. [2003] calculated
summertime isoprene fluxes over North America. Their results (5.7 Tg C in July)
were lower than an analogous calculation from the Guenther et al. [1995] method
(7.1 Tg C in July). Shim et al. [2005] extended this method to calculate a global
isoprene emission flux of 566 Tg C yr−1.
A second top-down estimation, using the satellite data retrieved by the Ozone
Monitoring Instrument (OMI), was undertaken by Millet et al. [2008]. In the
method which is most directly comparable to the results described in Palmer et al.
[2003], Millet et al. [2008] reported 8.2 Tg C yr−1 emitted by North America in the
period June-August. This compared with 10.6 Tg C yr−1 predicted by Guenther
et al. [2006].
2.2.3.4 Magnitude & range of estimates
The total emission of isoprene is roughly equal to that of methane (582 Tg yr−1
in Solomon et al. [2007b]). Generally, biogenic VOC comprises a large fraction
of total VOC emissions; in my control run (BASE, in chapters 3, 5, and 6) an-
thropogenic VOC was only 22% of the total global flux. Since the development of
the original empirical model by Guenther et al. [1995], the range of estimates has
been relatively small, between 400 and 600 Tg C yr−1. This appears to be well
constrained, though Arneth et al. [2008a] argued that this was due to proliferation
of a few algorithm types rather than low levels of scientific uncertainty. The range
that does occur can be explained by variations in input data; Pfister et al. [2008a]
44
noted that a 28% change in their emissions calculation arose from changing the
input land cover characteristics.
2.3 Summary
In this chapter, I introduced the salient features of tropospheric chemistry. I de-
scribed the initiation of the radical oxidation processes that occur in the atmo-
sphere, followed by the propagation of these reactions. This included the separa-
tion between ozone-producing and ozone-destroying chemical regimes. I described
the chain termination reactions which can be used to describe a NOx-sensitive or
VOC-sensitive chemical environment.
After explaining the basics of PAN chemistry and nighttime chemistry, I turned
to isoprene. The lifetime of isoprene in the atmosphere was discussed, as well as a
series of observations of isoprene emission and its correlation to light, temperature,
and CO2 concentration. I devoted the last section of this chapter to a discussion of
global isoprene flux estimations, describing empirical, process based, and satellite
based methods.
In the next chapter, I describe the atmospheric model that is used in this thesis,
and the impact that including isoprene has on model chemistry.
45
Chapter 3
The Impact of Isoprene in UKCA
In this chapter, I introduce the global climate composition model UKCA (the
United Kingdom Chemistry and Aerosol model). In section 3.2, I describe the
basic model dynamics, chemistry, and emissions used in my experiments. After
a short model evaluation (section 3.3) and experimental description (section 3.4),
I explore the impact of including isoprene oxidation in UKCA in a present day
atmosphere (section 3.5).
3.1 Introduction
Numerical approaches to weather prediction and meteorology were first proposed
by Cleveland Abbe in the early twentieth century [Abbe, 1901]. His suggestions
were realized in part by Vilhelm Bjerknes, the first scientist to propose the diagnostic-
prognostic system for weather prediction [see the review of Lynch, 2008, and refer-
ences therein]. Bjerknes also put forward the first set of independent mathematical
equations to describe the present and future states of the atmosphere from a set of
initial conditions [Bjerknes, 1904]. Solving these equations was emboldened by
the vision of Lewis Fry Richardson, the advent of the modern computer, and the
leadership of John von Neumann [Richardson, 1922; Tribbia and Anthes, 1987;
Lynch, 2008]. The first computational predictions of weather occurred in the early
46
1950’s [Platzman, 1979; Tribbia and Anthes, 1987].
Simultaneously, the importance of atmospheric chemistry for air quality was
being established; in 1952, Haagen-Smit described the effects of photochemi-
cal smog. The nature of tropospheric photochemistry has been elucidated and
characterized in more depth since [e.g. Crutzen, 1970; Liu et al., 1987; Atkinson,
2000], through the use of a number of tools, one of which is computer models.
Global modelling has been approached in two distinct ways: first, with the use of
chemistry-transport models (CTMs) [see e.g. Logan et al., 1981; Brasseur et al.,
1998; Wild and Prather, 2000; Bey et al., 2001; Park et al., 2004; Savage et al.,
2004; Teysse`dre et al., 2007] which are externally forced, and second with gen-
eral circulation models (GCMs) [see e.g. Roelofs and Lelieveld, 1995; Johnson
et al., 1999; Horowitz et al., 2003; Zeng and Pyle, 2003; Hauglustaine et al., 2004;
Schmidt et al., 2006], which generate their own meteorological fields. Tropospheric
chemistry modelling has historically been approached through the lens of air qual-
ity.
The inclusion of isoprene oxidation, which is particularly important for tropo-
spheric chemistry, is a more recent addition to global models [e.g. Fehsenfeld et al.,
1992; Pfister et al., 2008b]. Isoprene chemistry is complex, and involves hundreds
of intermediates; for this reason, its chemistry is represented in models in a number
of condensed schemes. The particular mechanism which represents isoprene oxi-
dation has been shown to influence model results [e.g. Archibald et al., in press].
von Kuhlmann et al. [2004] found a large differences (up to 60%) in ozone when
the various lumped mechanisms were used in their model. The production of PAN
is also cited as a reason for the differences in isoprene modelling, with global PAN
concentrations varying by over 50% in several sensivity studies in von Kuhlmann
et al. [2004]. Differences in model representations of isoprene degradation has also
been identified as one of the key areas of uncertainty in model intercomparisons.
Wu et al. [2007] found that the yield of organic nitrates from isoprene oxidation
plays a large role in the differences between model ozone budgets.
47
Pfister et al. [2008a] described the contribution of representing isoprene oxi-
dation in a CTM chemical budget by implementing a tracer tagging scheme. Over
hot-spot emitting regions in the southern hemisphere, column PAN was found to be
up to 70% resultant from isoprene degradation. Global ozone burden decreased by
3% when isoprene was removed from the model calculations, and regional changes
in ozone of up to 4 Dobson units occurred over high NOx-emitting regions. Over
the Amazon, a reduction in ozone was calculated when isoprene was included in
their model, which the authors attributed to ozonolysis. 9-16% of the CO burden
was derived from isoprene oxidation [Pfister et al., 2008a], similar to the estimate
of Granier et al. [2000], in which isoprene contributed 10% to the tropospheric CO
burden.
The photooxidation of isoprene is important for air quality on both global [e.g.
Fehsenfeld et al., 1992; Houweling et al., 1998; Folberth et al., 2006] and regional
[e.g. Pierce et al., 1998; Tao et al., 2003] scales, though precise results seem to
be sensitive to the configuration of the base model. The inclusion of isoprene in
one model led to a 250% increase in ozone in the northern hemisphere [Houweling
et al., 1998]. Wang and Shallcross [2000] found that isoprene chemistry increased
ozone by 4 ppbv over oceans and 8-12 ppbv over land and across northern midlat-
itudes in August. Folberth et al. [2006] reported somewhat higher values between
8-25 during northern hemisphere summer due to isoprene. Isoprene is also impor-
tant for the ozone budget, as reported by Wild [2007], in which isoprene emissions
were found to be one of the leading causes for budget variability. In the work of
Roelofs and Lelieveld [2000] net chemical production increased by 110 g yr −1 due
to isoprene, while the annually averaged burden increased by 23 Tg.
One of the most important ways in which isoprene impacts ozone concentra-
tions is through interaction with NOx. There is some uncertainty in the chemistry
of PAN and other nitrates that are formed during the degradation of isoprene (e.g.
isoprene nitrates, higher peroxy nitrates, or nitroxy-aldehydes). The yield of iso-
prene nitrates from reaction between isoprene peroxy radicals and NO remains
48
uncertain, and has been reported at 4.4% [Chen et al., 1998], 8% [Carter and
Atkinson, 1996], and even 8-12% [Sprengnether et al., 2002]. The fate of these
nitrates is also uncertain, including the rate of potential recycling of NOx through
reaction with OH [see Horowitz et al., 2007, and references therein] and rates of
deposition [Shepson et al., 1996]. The sensitivity of models to these parameters
has been tested in Horowitz et al. [1998], Fiore et al. [2005], and von Kuhlmann
et al. [2004]. The maximum relative changes in ozone burden in von Kuhlmann
et al. [2004] were found to arise from variations in deposition velocity as opposed
to the chemical mechanism, total isoprene emission, or treatment of nitrates. Fiore
et al. [2005] did not test the deposition sensivity, but did report that a change in the
yield and recycling of NOx from isoprene nitrates altered ozone maxima by over
10 ppbv.
In this chapter, my aim is describe UCKA and to quantify the impact that iso-
prene has on model results.
3.2 UKCA
UKCA has both a stratospheric version, which has been described in Morgen-
stern et al. [2009], and a tropospheric version, which has been described briefly
in O’Connor et al. [2009]. The stratospheric version of the model runs with a
different chemistry and is more often used with interactive chemistry-climate feed-
backs [e.g. Morgenstern et al., 2008]. What follows is the first description of the
tropospheric version of the model which is run in Cambridge.
3.2.1 Dynamics
UKCA is a chemistry module embedded in the unified model (UM), which is de-
veloped by the UK Met Office for both research and numerical weather predic-
tion. UM6.1 (the version used here) is a non-hydrostatic model which uses a semi-
Lagrangian tracer transport scheme [Priestley, 1993]. The horizontal resolution is
49
3.75 by 2.5 degrees. The model has semi-implicit timestepping, and is inherently
mass conserving. The dynamical core of the model is described in Davies et al.
[2005]. Gravity wave drag has both an orographic [Webster et al., 2003] and a
spectral component [Scaife et al., 2002]. The radiation scheme of Edwards and
Slingo [1996] is used, and for the BASE run, carbon dioxide was set to 345 ppmv
and was methane set to 1655 ppbv. In addition, the mixing ratios of N2O, CFC-11,
and CFC-12 were set to 207 ppbv, 230 pptv, and 382 pptv, respectively. Ozone
can interact on-line with the radiation scheme [e.g. Morgenstern et al., 2008], but
in these experiments values are set by the climatology of Li and Shine [1995] to
reduce variability between simulations. Cloud microphysics are represented as in
Wilson and Ballard [1999] and cloud top height for convection is parameterized
using Gregory and Rowntree [1990].
3.2.2 Chemistry
UKCA uses the ASAD chemistry package [Carver et al., 1997], which has been
used in the CTM TOMCAT [Law et al., 1998; Cook et al., 2007; Hamilton et al.,
2008] and the GCM UM CAM [Zeng and Pyle, 2005, 2003; Zeng et al., 2008]. The
chemistry is solved by a Newton-Raphson solver derived from Wild and Prather
[2000]. It is a medium-sized chemistry, simulating the Ox, HOx, and NOx chem-
ical cycles and the oxidation of CO, methane, ethane, propane, and isoprene. The
mechanism without isoprene chemistry has been described in Arnold et al. [2005].
The model has 55 chemical tracers, 114 bimolecular reactions, 15 termolecular re-
actions, and 35 species are photolysed. Isoprene oxidation is parameterized using
the condensed Mainz isoprene mechanism (MIM).
The MIM is described in Po¨schl et al. [2000] and its implementation into the
ASAD chemistry package is discussed in Young [2007] and Young et al. [2009].
The MIM performed favourably in comparison with a much more complex mech-
anism (the Master Chemical Mechanism, or MCM, [Jenkin et al., 1997; Saunders
et al., 1997]). It varied from the MCM by less than 10% in most cases for compar-
50
Table 3.1. Tracers involved in the MIM representation of isoprene oxidation, after
Po¨schl et al. [2000].
Tracer Description
C5 compounds
C5H8 isoprene
ISO2 peroxy radicals from C5H8 + OH
ISOOH β-hydroxyhydroperoxides from ISO2 + HO2
ISON β-hydroxy alkylnitrates from ISO2 + NO and
alkyl nitrates from C5H8 + NO3
C4 compounds
MACR methacrolein, methyl vinyl ketone, and other C4
carbonyls
MACRO2 peroxy raicals from MACR + OH
MACROOH hydroperoxides form MACRO2 + HO2
MPAN peroxymethacrylic nitric anhydride and other higher
peroxyacyl nitrates
C3 compounds
HACET hydroxyacetone and other C3 ketones
MGLY methylglyoxal and other C3 aldehydes
C2 and C1 compounds
CH3C(O)O2 peroxyacetyl radical
PAN peroxyacetyl nitrate
CH3C(O)OOH peroxy acetic acid
CH3C(O)OH acetic acid
NALD nitrooxy acetaldehyde
HCOOH formic acid
51
isons with O3, OH, CO, and PAN among other species [Po¨schl et al., 2000]. Table
3.1 shows a list of the 16 tracers involved in the representation of isoprene chem-
istry, which go through 44 chemical reactions. The main oxidation pathway of
isoprene is via reaction with OH, from which ISO2 is produced in the MIM. This
can either form a peroxide (ISOOH), resulting from reaction with eitherHO2 or an-
other ISO2 molecule, or can react with NO. The ISO2 + NO reaction forms either
a nitrate species (ISON), or a representative C4 species (MACR), and formalde-
hyde; from the latter reaction, other tracers are formed in much smaller quantities.
The MACR tracer can produce MPAN or hydroxyacetone, which can be oxidized
by OH to form methylglyoxal; oxidation of methylglyoxal by OH yields the per-
oxyacetyl radical. The peroxyacetyl radical can form PAN (see section 2.1.4) or
formaldehyde and methylperoxy, followed by final oxidation to carbon dioxide and
water.
Photoysis is calculated using the off-line scheme of Law and Pyle [1993a],
which relies on tabulated values for photolysis as calculated by the Cambridge 2D
model [Harwood and Pyle, 1975]. Photolysis rates are adjusted for solar zenith
angle and have a diurnal variation. 32 species are dry deposited, and dry deposi-
tion is calculated from tabulated velocities [Ganzeveld and Lelieveld, 1995; Zhang
et al., 2003b] and is adjusted for season, time of day, and surface roughness as
described by Giannakopoulos et al. [1999]. Five surface types are accounted for:
snow/ice, water, forest, desert, and savannah/grassland. The deposition velocity is
then extrapolated into the middle of the lowermost model level using the method
of Sorteberg and Hov [1996]. Wet deposition is described as a first order loss us-
ing rainfall rates as calculated by the dynamics of the UM [Giannakopoulos et al.,
1999] and is calculated for 24 species.
Concentrations of ozone, NOy, and CH4 are overwritten in the stratosphere
above approximately 30 hPa with zonal mean concentrations generated by the
Cambridge 2D model [Harwood and Pyle, 1975; Law and Pyle, 1993b,a]. This al-
lows for representative seasonal changes in tracer concentrations without running a
52
Table 3.2. Present day emissions of NOx, CO, and non-methane VOCs in UKCA.
Species Amount
NO2, Tg N yr−1
Domestic 1.5
Industrial 10.3
Traffic 16.1
Soil 7.0
Biomass Burning 10.2
Total 45.0
CO, Tg CO yr−1
Domestic 238.3
Industrial 37.6
Traffic 194.6
Biogenic 89.0
Ocean 11.0
Biomass Burning 507.5
Total 1078.0
VOC, Tg VOC yr−1
Domestic 28.8
Industrial 39.3
Traffic 47.7
Natural (acetone) 48.3
Natural (isoprene) 468.1
Biomass Burning 31.3
Total 663.5
full stratospheric chemistry. Carbon dioxide (345 ppmv) and methane (1655 ppbv)
are set at a fixed value in the chemistry to avoid excessively long spin-up periods.
3.2.3 Emissions
Eight chemical species are emitted in UKCA: nitrogen dioxide (NO2), carbon
monoxide (CO), formaldehyde (HCHO), ethane (C2H6), propane (C3H8), ac-
etaldehyde (CH3CHO), acetone (CH3C(O)CH3), and isoprene (C5H8). Tracers
are emitted into the lowermost model level, and emissions are updated every five
model days. Isoprene is the only species emitted with diurnal variation. Geo-
graphic distributions of the emissions of the first seven species were compiled by
53
Dentener et al. [2005]. These emissions were used by the model intercompar-
isons of Stevenson et al. [2006] and Shindell et al. [2006b], as well as in the fourth
assessment of the Intergovernmental Panel on Climate Change (IPCC) [Solomon
et al., 2007b]. Emissions of NOx, CO, and total non-methane VOCs are summa-
rized in Table 3.2 and are broken down by source sector. Anthropogenic emissions
are based on the Edgar3.2 dataset [Olivier and Berdowski, 2001], which includes
ship emissions estimated for 1995 and assumes a 1.5% growth rate per annum.
Biomass burning emissions are distributed according to van der Werf et al. [2003],
with totals derived from the 1997-2002 average. Ecosystem-specific biomass burn-
ing emission factors were used from Andreae and Merlet [2001].
Soil NOx emissions are distributed according to the GEIA emissions inventory
[Global Emissions Inventory Activity, Accessed 21 October, 2009] based on the
model of Yienger and Levy II [1995] and are scaled to 7 Tg N yr−1 as recommended
by Stevenson et al. [2006]. Aircraft emissions are from AERO2K [Eyers et al.,
2004]. Lightning NOx emission is based on the parameterization of [Price and
Rind, 1992] and is scaled to 5 Tg N yr−1 based on an idealized five year model run.
Natural CO emissions (biogenic and oceanic) are distributed as recommended by
Stevenson et al. [2006], according to the POET project [POET Inventory, Accessed
21 October, 2009], based on Mu¨ller and Brasseur [1995] and are scaled to 100 Tg
CO yr−1. Table 3.3 shows the breakdown of VOC emissions. Non-methane non-
isoprene hydrocarbons are scaled to 116 Tg VOC yr−1 for anthropogenic sources,
as suggested by the fourth assessment report of the IPCC [Solomon et al., 2007b].
Speciation is provided by the third assessment of the IPCC Prather et al. [2001].
Not all the species listed in Prather et al. [2001] are initialized in UKCA, so after
speciation the total is re-scaled to 100%. Biogenic acetone emissions of 30.1 Tg
C yr−1 are added to this using the spatial and temporal distribution of the GEIA
database.
54
Table 3.3. Present day emissions of ethane, propane, formaldehyde, acetaldehyde,
acetone, and isoprene in UKCA [Tg VOC yr−1].
Species Amount
Ethane
Domestic 9.5
Industrial 13.0
Traffic 15.8
Biomass burning 12.8
Total 51.1
Propane
Domestic 11.2
Industrial 15.2
Traffic 18.5
Biomass burning 3.7
Total 48.5
Formaldehyde
Domestic 1.0
Industrial 1.4
Traffic 1.7
Biomass burning 2.2
Total 6.3
Acetaldehyde
Domestic 3.2
Industrial 4.4
Traffic 5.4
Biomass burning 11.1
Total 24.2
Acetone
Domestic 3.9
Industrial 5.3
Traffic 6.4
Biomass burning 1.5
Natural 48.3
Total 65.3
Isoprene
Total 468.1
55
January
July
Figure 3.1. January (top) and July (bottom) isoprene emissions [µg m2 s−1] modelled
by the SDVGM and a BVOC emission model.
3.2.4 Isoprene emissions
Isoprene emissions were derived using a BVOC emissions model1. The emissions
model is based on the empirical parameterizations of Guenther et al. [1995], which
were updated in Guenther et al. [2006] and have been described in Chapter 2. In
the version used here, parameterizations were adjusted from their original form
1Isoprene emissions were calculated by Juliette Lathie`re at the University of Sheffield. Our col-
laboration is part of the QUEST project.
56
and did not include geographic variation in emissions factors from various plant
functional types (PFTs). The distribution of PFTs was calculated by the Sheffield
Dynamic Global Vegetation Model (SDGVM) [Woodward et al., 1995; Beerling
et al., 1997]. The vegetation model was forced with meteorological fields output
from the UM, which was run without chemistry. The present-day global vegetation
map is based on Loveland et al. [2000] and accounts for crops by using Ramankutty
and Foley [1999] and anthropogenic grasses from Klein Goldewijk [2001]1.
0 20 40 60 80 100 120 140 160 180 
North America 
Europe 
Central Africa 
Southeast Asia 
Amazon 
Figure 3.2. Regional isoprene emissions [Tg C yr−1].
Figure 3.1 shows the geographical location and magnitude of present day iso-
prene emissions for January and July, when emissions are at their peak in the south-
ern and northern hemisphere, respectively. A meridional seasonality appears in the
emissions that follows the path of the sun, as emissions of isoprene are closely cor-
related to radiation and temperature (see section 2.2.2.1). Tropical and sub-tropical
forests are notably high emitting, and key emission regions include the Amazon,
central Africa, Southeast Asia and in July, the southeastern United States. The total
1The crop map was compiled by Nathalie du Noblet-DuCoudre´ and Jean-Yves Peterschmitt in
collaboration with Juliette Lathie`re [Lathie`re et al., 2009].
57
global emissions of isoprene are 413 Tg C yr−1 (468 Tg isoprene yr−1), which is
78% of total flux of VOC into the atmosphere in the BASE experiment.
Figure 3.2 shows isoprene emissions for high emitting regions. The Amazon
emits the most of any region, at 174 Tg C yr−1. Central Africa and Southeast Asia
emit 80 and 91 Tg C yr−1, respectively. Annual isoprene emissions in North Amer-
ica are lower than the other hot-spot areas, but the numbers are slightly misleading
due to the seasonality of emissions. In January, North American emissions are as
low as 1.9 Tg C month−1, and in July, this value rises to 9.35 Tg C month−1. If this
high level of emission was maintained, the annual emission would come to 112.2
Tg C yr−1, higher than Southeast Asia and central Africa.
3.3 Model evaluation
3.3.1 Stratosphere-troposphere exchange
Initially, a 38-level version of the model was developed. This version is still em-
ployed at the UK Met Office, and is described briefly in O’Connor et al. [2009]. I
found ozone concentrations that were too high in remote marine areas when using
this version of the model, especially in the northern hemisphere during March-
April and the southern hemisphere during September-October. This seasonal cycle
is suggestive of too strong stratosphere-troposphere exchange (STE), which is a
result of the Brewer-Dobson circulation [Dobson et al., 1946; Brewer, 1949]. The
seasonality of the circulation leads to a strengthening of the downward motion
in the winter hemisphere [Holton et al., 1995]. The Brewer-Dobson circulation is
stronger in the northern hemisphere, due to to increased wave activity in the region,
and this may contribute to observed maxima in spring concentrations of ozone in
the region [e.g. Monks, 2000].
The introduction of an on-line STE diagnostic1 reinforced our hypothesis. Sea-
sonally varying annual STE [Tg O3 yr−1] is shown in Figure 3.3 for two model
1The diagnostic was implemented by Dr. N. Luke Abraham.
58
ST
E
(
Tg
/y
ea
r)
D J F M A M J J A S O N
1600
1400
1200
1000
800
600
400
Figure 3.3. Stratosphere-troposphere exchange [Tg O3 yr−1] for two model runs.
The 38 level version is shown in pink, and the 60 level version is shown in green (the
‘BASE’ run).
runs. The 38-level version of the model simulates the atmosphere up to 56 km,
while the 60-level version captures the whole stratosphere and simulates up to 84
km. This extension into the mesosphere meant that the Brewer-Dobson overturn-
ing could be fully and more accurately modelled. Indeed, the 60-level version has
an average annual STE is 424 Tg yr−1, which lies within the first standard devi-
ation of the model average from the recent comprehensive model assessment by
Stevenson et al. [2006] of 556 ± 154 Tg yr−1. This was a marked improvement
over the 38 level version, which had an average annual STE of 1324 Tg yr−1. The
estimate from the 60-level version of the model is similar to that reported in Mur-
phy and Fahey [1994], who used observed ratios to estimate the flux of ozone from
the stratosphere to be 450 Tg yr−1. Gettelman et al. [1997] estimated STE to be
450-590 Tg yr−1 from satellite data.
59
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
30S-EQ 250hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 15
xdeea
EQ-30N 250hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 8
EQ-30N 250hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
90S-30S 500hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 5
30S-EQ 500hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 15
EQ-30N 500hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 8
30N-90N 500hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 18
90S-30S 750hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 5
30S-EQ 750hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 15
EQ-30N 750hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 8
30N-90N 750hPa
J F M A M J J A S O N D
month
20
40
60
80
100
O
3 /
 p
p
b
v
n = 18
Figure 3.4. Comparison of climatological ozone (black diamonds) as described in
[Stevenson et al., 2006] with the BASE model run (green line). Lines show the stan-
dard deviation of the measurements and the lighter green shows the standard devia-
tions for the model. Both model and measurements are sampled for a series of latitude
bands and three pressure levels identically. Data are from Logan [1999] and Thomp-
son et al. [2003].
3.3.2 Comparison to measurements
Ozone is of particular chemical importance for tropospheric chemistry (see section
2.1.1), and also has an impact on the radiative balance of the atmosphere, human
health, and crop health. Throughout this thesis, I use ozone as a sort of exemplar
trace constituent in the troposphere, and examine the sensitivity of ozone to vari-
ous perturbations. Though a more exhaustive analysis of the model’s performance
can be found in a forthcoming paper by Fiona O’Connor (to be submitted as Part
II of Morgenstern et al. [2009]), here I briefly describe the model’s ability to sim-
ulate ozone. Chapter 4 elaborates on the model chemical mechanism’s ability to
reproduce measurements.
60
Figure 3.4 shows the performance of UKCA against the climatological ozone
fields described in Logan [1999] and Thompson et al. [2003] as shown in Steven-
son et al. [2006]. Data are broken down into four latitude bands—the southern
high latitudes, the southern tropics, the northern tropics, and the northern high
latitudes—and three pressure levels—750 hPa, 500 hPa, and 250 hPa. Near the
surface, UKCA compares well with measurements, capturing both the seasonal
variation and the magnitude of the climatologies. One key weakness is apparent:
the northern tropics and northern high latitudes are slightly overestimated in mag-
nitude (by approximately 8 and 6 ppbv, respectively). The overprediction is also
evident in the multi-model assessment of Stevenson et al. [2006], which uses the
same emissions, and could be an artefact thereof.
In the middle of the troposphere at 500 hPa, the seasonal variation in the clima-
tology continues to be reproduced by the model, though the tropical measurements
are slightly underestimated in magnitude. The opposite occurs in the northern high
latitudes, where the model slightly overpredicts ozone concentrations; this is visi-
ble in the multimodel comparison, and again could be an artefact of the emissions
[Stevenson et al., 2006]. The comparison is poorest in the the upper tropical tropo-
sphere, where the underprediction is exacerbated and reaches up to 15 ppbv. This
is not the case for the multi-model average of Stevenson et al. [2006], where upper
troposphere ozone is overpredicted by approximately 7 ppbv in the northern hemi-
sphere during winter. The negative bias is only partially (up to -2 ppbv) explained
by the positive bias in the model humidity in the same region [O’Connor et al.,
2009]. It could also be linked to lightning NOx, for which the uncertainty has re-
cently been estimated at 60% [Schumann and Huntrieser, 2007]; perturbations to
lightning NOx led to up to 10% changes in the ozone burden in the work of Wild
[2007]. Another contributing factor could be the underestimation of HOx from the
photolysis of formaldehyde, for which the absorption cross section has recently
been found to be greater than previously thought [Carbajo et al., 2008].
61
3.4 Experimental setup
I ran two experiments: BASE, in which isoprene was included, and NOISOP, in
which isoprene emissions were set to zero. Although in older work [e.g. Zeng and
Pyle, 2003] the lack of biogenic VOC emissions was often compensated for by in-
creasing CO emissions, here I kept CO emissions constant in both experiments to
minimize the variable differences between the two runs. I employed the 60 level
version of UKCA. Model meteorology in the atmosphere-only version is forced us-
ing sea surface temperatures (SSTs) and sea ice distribution. For these two present
day runs, SSTs were constructed using a two stage interpolation of observations
[Rayner et al., 2003]. Both simulations were run for five years and four months
between 1999 and 2005, with the first sixteen months taken as spin up.
3.5 Impact of isoprene in UKCA
3.5.1 Ozone
3.5.1.1 The global impact
Figure 3.5 shows the impact of including isoprene on near-surface monthly mean
ozone in January and July. Isoprene reduces ozone by 5-10 ppbv in the Ama-
zon, Southern Africa, and Southeast Asia in January. These three regions also
displayed decreases in ozone when a similar sensitivity study was carried out by
Young [2007], although the absolute magnitude of the change was higher in that
study (-20 to -40 ppbv), which is likely a result of differing distributions in emis-
sions. In January, only one region shows a large increase in ozone concentrations
due to the inclusion of isoprene in model runs, which is central Africa. Central
Africa has strong NOx emissions from biomass burning in January, enough so that
the local chemical environment is VOC-sensitive in terms of ozone production.
Thus the inclusion of isoprene increases ozone production by providing VOC to
the system (see section 2.1.3).
62
January
July
Figure 3.5. Change [ppbv] in five year average near-surface (600 m and below) ozone
January (left) and July (right) between the BASE and NOISOP simulations.
In July, the reduction in ozone is limited to a smaller area of the Amazon, cen-
tral Africa, and Southeast Asia to a lesser degree. The more noteworthy feature in
boreal summer is the dramatic increase in ozone production throughout the north-
ern hemisphere, which is particularly strong in the eastern United States, Europe,
and eastern Asia. The strength of the signal over terrestrial areas is between 10-30
ppbv, which is higher than the values of 8-12 ppbv reported in Wang and Shallcross
63
[2000], though similar values are described in the work of Folberth et al. [2006]
(8-25 ppbv) and Young [2007] (5-20+ ppbv). Over the oceans, smaller changes
were noted in all three studies, and have been modelled here as well. Increases
in the summer ozone concentrations with the inclusion of ‘higher’ (non-methane)
hydrocarbons are also reported in Roelofs and Lelieveld [2000] and Houweling
et al. [1998]. Variations in the exact magnitude of the changes in ozone resulting
from representing isoprene could be due to variations in emissions used in each
model, and to variations in the number and reactivity of hydrocarbons in the base
chemistry.
Ozone concentrations also increase in the southern part of South America and
in southern Africa in July, though the two hemispheric signals have separate and
distinct causes. The increase in ozone in the northern hemisphere is due to the
seasonality of isoprene emissions (see Figure 3.1). In the NOx-rich environment
of the more polluted northern hemisphere, the addition of VOC increases ozone
production by altering the balance between NOx and VOC. In the southern hemi-
sphere, the signal is more closely linked to seasonal increases in NOx (as opposed
to increases in VOC), as shown in Figure 3.6. The southern hemisphere elevations
in NOx emissions are due to biomass burning emissions.
Table 3.4 shows the impact of isoprene on the global tropospheric ozone bud-
get. The greatest absolute impact on the ozone budget is the increase in production
of 129 Tg yr−1 (4.2%) when isoprene is included, although the greatest relative
change is seen for net chemical production. A similar increase in total ozone pro-
duction (93 Tg yr−1) was reported in Wu et al. [2007]. Loss also increases as
ozonolysis (O3 + isoprene) is introduced and as elevated HO2 concentrations in-
crease the rate of the HO2 + O3 reaction (section 2.1.2). The ozone lifetime de-
creases by a day and a half when isoprene is included in the model (-4.5%), which
is the same as the relative difference described in Pfister et al. [2008a], although
the absolute values are very slightly higher here. The lifetime was also reduced in
the study of Wild [2007] by 1.1 days. The burden is nearly unchanged, similar to
64
a   January
b   July
Figure 3.6. NO2 [g m−2 month−1] emissions in the BASE simulation for UKCA.
the result (-0.4%) in an earlier version of the model [Young, 2007]. Low changes
in the ozone burden resulting from isoprene oxidation have also been reported by
Wu et al. [2007] and Pfister et al. [2008a].
Once isoprene is included in the model, the total chemical production of ozone
is 3215 Tg yr−1, which is lower than the mean of 3984 reported in Wild [2007]
(which includes the ACCENT study of Stevenson et al. [2006]). The net chemical
production in UKCA of 402 Tg O3 yr−1 is closer to the multi-model mean in Wild
65
Table 3.4. Annual ozone budgets with and without isoprene included in UKCA. Each
term is given in Tg O3 yr−1 except the burden [Tg] and the lifetime [days].
Experiment NOISOP BASE Difference Relative change
Production 3086 3215 129 4.2%
Loss 2748 2813 65 2.4%
Net Chem. Production 338 402 64 18.9%
Deposition 762 795 33 4.3%
Inferred STE 424 397 27 -6.4%
O3 τ 33.1 31.6 1.5 -4.5%
O3 Burden 323 317 6 -1.9%
[2007] of 245 Tg yr−1. Both terms have a significant range in their estimations
(761 and 346 Tg yr−1, respectively). Deposition of 795 Tg yr−1 and STE of 397 Tg
yr−1 are also lower in UKCA than in the multi-model mean, which were 902 and
636 Tg yr−1. Despite these underestimations, the atmospheric burden of ozone in
UKCA (317 Tg yr−1) is close to the mean estimate reported in Wild [2007] of 307
Tg yr−1. These values are lower than those reported by a previous implementation
of the model chemistry, for which production was found to be 4025 Tg yr−1 and
loss was 3225 Tg yr−1 [Young, 2007]. A value of 3620 Tg yr−1 for chemical
production was also reported using a similar chemical mechanism with an older
version of the dynamics [Zeng et al., 2008]. The values in UKCA could be partially
resultant from the lumped isoprene oxidation mechanism, which has been shown
to underpredict ozone in both high and low isoprene conditions [Archibald et al.,
in press].
3.5.1.2 The chemical impact
Figure 3.7 shows the change in annual average chemical production of ozone in
the troposphere resulting from the addition of isoprene oxidation to UKCA. Three
of the world’s isoprene hot-spots show a reduction in chemical production: the
Amazon, central Africa, and Southeast Asia/the maritime continent. Terrestrially,
the notable exception to this is the southeastern United States, where the introduc-
66
Figure 3.7. Annual average tropospheric chemical production of ozone (8500 m and
below) between BASE and NOISOP runs [kg O3 gridbox month−1]
tion of isoprene instead leads to an increase in ozone production. This is a region
where NOx emissions are high, and the same is true for Europe and eastern Asia,
areas that also show an increase in ozone production when isoprene is included in
the model. In addition, ozone production increases over the oceans downwind of
regions of high emission (the equatorial Pacific near South America, the tropical
Atlantic between central Africa and the Amazon, the Indian ocean adjacent to the
maritime continent, and the Pacific east of Southeast Asia.) The formation of ni-
trates, notably PAN, allows NOx to be transported to these remote areas. There is
also a widespread increase in ozone production over all of the tropical oceans. The
production of ozone in southern South America and central Africa due to biomass
burning emissions, described earlier, are not evident in this annual average pic-
ture. A very similar distribution in the change of ozone production has been found
previously, using the same chemical mechanism in a predecessor model [Young,
67
2007].
January                                             RO2 + NO July                                                     RO2 + NO
January                                             O3 + isoprene July                                                     O3 + isoprene
Figure 3.8. Key reaction fluxes [moles gridbox−1 second−1] between the BASE and
NOISOP simulations.
One of the ways in which isoprene impacts the ozone budget is through the
increased regeneration of NO2 by reaction of NO with RO2. The top panels in
Figure 3.8 show the flux through the reaction RO2 + NO in January and July.
This leads to the formation of ozone via photolysis of NO2 (see section 2.1), but
depends both on RO2 radical concentrations (to which isoprene contributes), and
NO concentrations. The strongest signal is modelled over the southeastern United
States, where NOx emissions are high and background concentrations of NO are
elevated. Other strong signals appear in areas of high NOx emission such as central
Africa in January, and the central Amazon in July. In contrast, isoprene also leads
directly to ozone loss via ozonolysis. The flux through this reaction is shown in the
bottom panels of Figure 3.8.
Another way in which isoprene chemistry influences ozone is through the for-
68
H
yb
rid
 H
t [
m
]
Figure 3.9. Annual average relative increase [%] in tropospheric zonal mean PAN
between BASE and NOISOP runs.
mation of PAN, a reservoir species for NOx. The lifetime of PAN has a strong
dependence on temperature, so that when PAN is lofted into the higher, colder
altitudes of the troposphere, it can exist for up to one month in winter, and ap-
proximately 10 days in summer [Crutzen, 1979; Moxim et al., 1996]. In warmer,
surface regions it dissociates with a lifetime of hours, releasing NOx, which can
then participate in further radical reactions. The formation of nitrates during iso-
prene oxidation is stabilized by a feedback loop involving OH. Reaction with OH
can destroy nitrates, though this is slow in comparison with thermal degradation for
PAN. In regions of high isoprene emission, where nitrate formation can be high,
isoprene reacts with OH, reducing the potential sink of nitrates by chemical de-
struction. This small feedback loop is an additional reason why nitrate formation
is marked when isoprene is added to UKCA. The stabilization of nitrate species is
particularly important for ISON, a lumped isoprene nitrate species (formed both
by reaction of first generation hydroxyalkoxy RO with NO2, and by reaction of
69
isoprene with the NO3 radical).
The relative difference in zonal mean PAN between the BASE and NOISOP
simulations is shown in Figure 3.9. A strong hemispheric gradient appears, arising
from the the fact that the majority of isoprene is emitted in the southern hemisphere
and tropics (up to 10◦N, see Figure 3.1). The seasonal emissions of the southeast-
ern United States appear as an area of 50-70% increase in PAN between 30◦N
and 40◦N. Because the PAN lifetime is strongly dependent on temperature, the in-
crease (140%) in PAN at high altitudes is due to low temperatures in this region.
In contrast, the rise nearer to the surface is due to the very high concentrations of
isoprene, where PAN is produced in high quantities despite the shorter lifetime.
Young [2007] also found a high change in PAN burden due to isoprene oxida-
tion, with increases of up to 900%. The largest relative changes were also over
the southern tropics, similar to what is modelled by UKCA. Large changes in PAN
were also reported by Wang and Shallcross [2000].
3.5.1.3 Relationships between emission and ozone changes
In order to understand the ozone changes shown in Figure 3.5, I attempted to delin-
eate different chemical regimes by correlation to levels of isoprene or NOx emis-
sion. The two regimes of ‘VOC-sensitive’ and ‘VOC-rich’ become apparent when
the change in ozone is plotted as a scatter function against isoprene emission, as
shown in Figure 3.10. The VOC-sensitive regime appears as a dense, positive cor-
relation on the left side of the plot and is highlighted by the orange box. These
points represent gridboxes in which the addition of isoprene has increased ozone
concentrations. A number of gridboxes show increases in ozone of up to 20 ppbv,
with some showing an even higher response. The longer, less dense correlation
that corresponds to a VOC-rich environment extends through to the right side of
the plot, and is highlighted in the blue box. These are points for which the inclusion
of isoprene oxidation has decreased ozone concentrations. The addition of VOC to
a VOC-sensitive regime results in a larger change in ozone concentrations (0-20+
70
Figure 3.10. Scatter plot of the change in ozone [ppbv] versus the change in isoprene
[ppbv] for all surface gridpoints in the BASE and NOISOP model runs. The VOC-
sensitive regime is highlighted in orange, and the VOC-rich regime is highlighted in
blue.
ppbv), while changes in the VOC-rich regions are relatively smaller in magnitude
(-10-0 ppbv). In VOC-sensitive locations, the sinks ofHOx are lower than the sinks
of NOx, and the radical concentrations are not sufficient to oxidize all the NO in a
gridbox [Sillman and He, 2002]. This creates the sharp gradient in the upper right
of Figure 2.3, and the greater response in VOC-sensitive regions (orange box).
The addition of VOC to a VOC-sensitive chemical environment can stimulate
ozone production by promoting NO2 regeneration through reaction of NO with
RO2 (see the ozone production topography in section 2.1.3). It can also promote
the formation of nitrate species, which remove NOx from the chemical system
and reduce ozone production. These two competing effects are evident in Figure
3.11. In gridboxes with low NOx emission (shown in the orange box), which can
be either NOx- or VOC-sensitive, isoprene does not have a definitive effect, in
71
Figure 3.11. Scatter plot of the change in ozone [ppbv] versus NOx emission [mg m2
month−1] for all surface gridpoints in the BASE and NOISOP model runs. The NOx-
sensitive regime is highlighted in orange, and the NOx-rich regime is highlighted in
blue.
some areas reducing ozone, and in some areas increasing it. The response even
shows a sort of Gaussian distribution around zero. In a remote forest location, for
example, low NOx emissions in one box could be sequestered as PAN or ISON
(reducing ozone production), but downwind upon the release of NOx, ozone pro-
duction could increase. In areas of higher NOx emission (shown in the blue box),
however, the inclusion of isoprene nearly always results in elevated ozone levels.
In most cases increases of up to 15 ppbv ozone are modelled, with exceptional val-
ues above 20 ppbv shown for a fewer number of gridboxes. The correlation shows
a wide spread, though a similar relationship and spread has been described in Sill-
man et al. [1990]. A positive NOx-O3 correlation in response to isoprene was also
modelled in relation to the PROPHET field campaign [Barket Jr. et al., 2004].
72
3.5.2 Methane
3.5.2.1 The impact on OH
Another way in which isoprene alters the chemistry of the troposphere is through
its chemical relationship with HOx. The direct impacts of isoprene on surface
OH can be seen in Figure 3.12. Surface OH decreases everywhere in the model
when isoprene oxidation is included, though the high-emitting regions (Figure 3.2)
exhibit the strongest reductions in OH concentrations. The seasonality of emis-
sions appears in the fluctuation of OH reductions from south to north. The largest
changes are modelled over South America and Asia, with values of up to -4.5× 106
molec cm−3. Slightly lower changes between 2-4× 106 molec cm−3 are modelled
over other hot-spot regions. These changes in OH in UKCA are similar to work
in a predecessor version of the model [Young, 2007], and have also been shown
in a number of other global models [Houweling et al., 1998; Jo¨ckel et al., 2006;
Folberth et al., 2006]. Away from areas of very high emission, OH concentrations
change by a smaller magnitude of 0-1 × 106 molec cm−3. Reduced OH arises
from an increase in the formation of peroxides (section 2.1.3) which can form a
permanent sink for HOx. Secondly, the recycling of OH from HO2 (a product of
VOC oxidation) by ozone, which proceeds as:
HO2 +O3 → OH+ 2O2 (2.16)
is less efficient than recycling by NO:
HO2 +NO→ OH+NO2 (2.15).
Isoprene oxidation therefore reduces the OH/HO2 ratio [Young, 2007] in a low
NOx environment.
The chemical relationship between isoprene and OH is complex. In high NOx
environments, the oxidation of VOC leads to the formation of HO2 which reacts
73
-4.5 -4.0 -3.6 -3.1 -2.7 -2.2 -1.8 -1.3 -0.9 -0.4 0.0
January
July
Figure 3.12. Difference in surface OH [106 molec cm−3] between BASE and
NOISOP runs for January (top) and July (bottom).
with NO to form OH. This generally maintains a high concentration of OH. In low
NOx environments, the chemistry is more complex, though depletion of OH has
been reported by a number of global models [Houweling et al., 1998; Jo¨ckel et al.,
2006; Folberth et al., 2006]. Recent work has suggested that high measured con-
centrations of OH in low NOx environments [e.g. Ren et al., 2008; Hofzumahaus
et al., 2009] indicate a missing OH source that models fail to consider [Lelieveld
et al., 2008; Butler et al., 2008]. OH is also a byproduct of ozonolysis reactions be-
tween ozone and terpenes, including isoprene [Atkinson et al., 1992]. The inclusion
of the O3 + isoprene reaction was not enough to account for measured concentra-
74
tions of OH in the work of Biesenthal et al. [1998] or Carslaw et al. [2001]. Future
discoveries of the ‘missing OH source’ may shift our understanding of the VOC-
HOx chemical relationship. Future work will perhaps model decreases in OH (and
the resultant increase in methane lifetime) of a reduced magnitude in response to
evolving scientific understanding of isoprene oxidation in low NOx environments
[see e.g. Karl et al., 2009].
3.5.2.2 Methane lifetime
Although changes in OH are especially important for the oxidizing capacity of the
troposphere, they are also important because reaction with OH forms the main sink
for methane, a radiatively active greenhouse gas [Solomon et al., 2007b]. Figure
3.13 shows the average annual seasonal cycle of methane lifetime in the atmo-
sphere for simulations with and without isoprene emissions. Two things are no-
table: first is the seasonal cycle of methane lifetime, arising from the fact that OH
concentrations are elevated in the northern hemisphere summer due to the pres-
ence of anthropogenic pollution [e.g. Spivakovsky et al., 2000]. Second, there is a
large difference between the runs with and without isoprene. The average annual
methane lifetime for these runs is 10.2 years (without isoprene) and 12.5 years
(with). These values are higher than those reported in Stevenson et al. [2006] of
8.7 ± 1.3 years. The methane lifetime decresed 4.2% when the negative bias in
humidity in the tropics in UKCA was removed [O’Connor et al., 2009], where OH
concentrations are high, and by inference, so is the methane sink, [e.g. Thompson
et al., 1989; Lelieveld et al., 1998; Bousquet et al., 2006]. Despite this, the methane
lifetime in UKCA remained high (9.61 years). Wild [2007] found that humidity,
temperature, and treatment of wet deposition were the largest reasons for methane
lifetime variability in global model budgets.
Through increasing the sink for OH, isoprene chemistry accounts for a 22.5%
increase in the methane lifetime. Young [2007] found that isoprene accounted for a
28% rise in the methane lifetime of UKCA’s predecessor model, UM CAM. Simi-
75
lar results have been reported for a halving of isoprene emissions, where resulting
OH concentrations increased by 7.2% and methane lifetime decreased by 7 months
[Young et al., 2009]. Consistent with this picture, a 7.3% decrease in global OH
due to isoprene was also reported in Pfister et al. [2008a].
8 
9 
10 
11 
12 
13 
14 
Ja
nu
ar
y 
Fe
br
ua
ry
 
M
ar
ch
 
Ap
ril
 
M
ay
 
Ju
ne
 
Ju
ly
 
Au
gu
st
 
Se
pt
em
be
r 
O
ct
ob
er
 
N
ov
em
be
r 
D
ec
em
be
r 
Figure 3.13. Seasonal variation of tropospheric methane lifetime with respect to OH
[years] for simulations with (blue) and without (red) isoprene emissions.
The magnitude of the seasonal cycle in methane lifetime also increases (from
2.3 years to 3.1 years, a change of 32%) when isoprene is included in model calcu-
lations. Two competing factors are at play: first, the direct interaction between OH
and isoprene, and second, the indirect effect of isoprene in the formation of ozone,
which in turn leads to elevated concentrations of OH. As isoprene emissions are
concentrated in the southern hemisphere (see Figures 3.1 and 3.2), the removal of
OH due to chemical reaction is more marked than in the northern hemisphere, thus
the first effect dominates during December, January, and February. In the northern
hemisphere summer, (June, July, and August) ozone increases strongly with the
inclusion of isoprene. Increased ozone, along with increasing radiation due to the
season leads to higher values of OH. The seasonal difference in the effects of these
76
two signals leads the larger magnitude in the methane lifetime seasonal cycle. This
change was also reported in Young [2007], in which isoprene led to a 36% increase
in the seasonal cycle of the methane lifetime.
3.6 Summary
In this chapter, I described the CCM UKCA in a tropospheric configuration. I be-
gan with a description of the dynamical core of the model (the Unified Model),
and the representation of chemistry used (the ASAD chemistry package), and the
the emissions used in the BASE and NOISOP present-day model simulations. The
model’s vertical top was considered in the discussion of STE, which showed that
a full representation of the stratosphere was necessary in order to model STE ac-
curately. A short model comparison was done against ozone climatologies similar
to the work of Stevenson et al. [2006], which showed that the model performs well
at the surface but underestimates ozone in the tropical upper troposphere. After
describing the emissions modelling, geographical distribution, and seasonality of
isoprene emissions worldwide, I examined the impact that isoprene has model re-
sults.
As expected, the change in surface ozone resulting from isoprene oxidation
was closely correlated to the local chemical environment into which isoprene was
emitted. The largest increases in ozone were modelled throughout the northern
hemisphere in summer, while decreases were modelled over most of the high emit-
ting regions in the tropics, though this was dependent on season. I showed NOx
emissions to help identify the environment into which isoprene emission occurred,
and I was able to separate the chemical regime of the ozone change through cor-
relations with both isoprene and NOx emissions. Finally, I described the impact
that isoprene has on methane lifetime and OH in UKCA, which was to increase
methane lifetime by 2.3 years and to increase its seasonal cycle. Consistent with
other models, OH concentrations were reduced substantially over high emitting
77
regions.
In the next chapter, I describe a set of regional and local experiments in South-
east Asia, an area of high isoprene emission.
78
Chapter 4
Case Study: South East Asia
In this chapter, Southeast Asia is used as a case study for two types of model ex-
periments. In section 4.1, I perform experiments in connection with data from the
OP3 field campaign. I use a box model to assess the ability of the UKCA chemical
mechanism to reproduce NO, NO2, and O3 measurements from a remote tropical
rainforest field site in Malaysian Borneo. Section 4.2 is comprised of two simple
sensitivity studies performed at the regional scale; in the first of these isoprene
emissions are doubled in Southeast Asia, and in the second they are reduced by
twenty five percent.
4.1 OP3
4.1.1 Introduction
A four month field campaign, part of the NERC-funded ‘Oxidant and Particle Pho-
tochemical Processes’ (OP3) [OP3, Accessed 11 November, 2009], was conducted
in the Malaysian state of Sabah, on the island of Borneo (the location is shown in
Figure 4.9), between April and July of 2008 [Hewitt et al., 2009a]. There were
two intensive periods of observation, the first between 8 April and 3 May, and the
second between 25 June and 23 July. A key goal of the project is to assess our
understanding of photochemical processes above a rainforest and their impacts on
79
various scales; to this end, the campaign utilised simultaneous ground, airborne,
and satellite measurements [a full list of instrumentation can be found in Hewitt
et al., 2009a]. A further aim is to understand the scale relationships of these mea-
surements as they are used by and contribute to mesoscale, regional, and global
models.
Atmospheric oxidation above a tropical rainforest is complex [e.g. Kuhn et al.,
2007; Lelieveld et al., 2008], and it is therefore beyond current computational re-
sources to represent it explicitly in a global model. Furthermore, the horizontal res-
olution of the current generation of global models is 2-5◦ [Stevenson et al., 2006]
(approximately equivalent to 220 and 550 km at the equator), which limits their
ability to model sub-grid scale processes such as emissions variability. At the same
time, these models attempt to simulate the production and destruction of ozone,
which is dependent on local chemical conditions [Crutzen, 1973; Sillman et al.,
1990; Jenkin and Clemitshaw, 2000]. Our understanding of the future impacts of
ozone often relies on the output of global models [Forster et al., 1996; Fuglestvedt
et al., 1999; Stevenson et al., 2006, e.g.] and it is therefore essential to understand
how global chemical mechanisms perform in relation to the local measurements
which help to constrain them.
NO and NO2 act as catalysts in many oxidation cycles in the atmosphere due
to their rapid interconversion; the availability of NOx largely determines whether
ozone production or destruction dominates in a specific region of the tropical bound-
ary layer [Liu et al., 1987]. Nitrogen oxides are emitted both by natural pro-
cesses and human activities. Of the biogenic sources, emission from soils [Yienger
and Levy II, 1995; Delon et al., 2008] and formation during lightning storms
[Franzblau and Popp, 1989; Schumann and Huntrieser, 2007] are the major con-
tributors. Fossil fuel combustion, biomass burning and aircraft emissions are the
major anthropogenic sources [Kasibhatla, 1993; Levy II et al., 1999; Toenges-
Schuller et al., 2006]. Though fluxes from tropical areas are not yet well quan-
tified, the work of Bakwin et al. [1990] suggested significant NOx emissions from
80
tropical forested areas. Jaegle´ et al. [2004] reported that soil emissions of NOx
could be as large as biomass burning emissions in Africa. In these remote tropical
areas the potential for NOx species to influence local chemistry is significant, due
to low background NOx and high concentrations of both the hydroxyl radical and
biogenic VOC [Hewitt et al., 2009b; Steinkamp et al., 2009]. An increase in the
frequency and spatial distribution of tropical NOx measurements will help quantify
local tropical fluxes and sources. But global models will largely play the role of
quantifying the impact of these fluxes on a regional and global scale. For this rea-
son, it is important to understand how global models relate to local measurements.
To this end, I use the chemical mechanism incorporated in the global CCM
UKCA, and study what alterations are required to match measured NO, NO2, and
O3. In this section, my aim is to assess the ability of UKCA’s chemical mechanism.
4.1.2 The measurements
4.1.2.1 Methods
Nitrogen oxides and ozone measurements were taken at the Bukit Atur Global At-
mospheric Watch (GAW) station (04◦58’53”, 117◦50’37”, and elevation 426 m)
the location of which is shown in Figure 4.91. NO measurements were made by
chemiluminescence using an Ecophysics CLD 780 TR nitric oxide analyser, with
an Ecophysics PLC 762 NO2 photolytic converter connected to allow conversion
of NO2 to NO. NO and NO2 concentrations were measured from an inlet situ-
ated at 5 m above ground level. Each measurement cycle lasted for 1 minute and
consisted of 12 seconds of NO measurement, 12 seconds of NO2 measurement
and 24 seconds of interference determination. The remaining 12 seconds allowed
for switching between the different modes and purging of the reaction cell. The
1σ limit of detection for 10 minute frequency data was approximately 2.8 pptv for
NO and 7 pptv for NO2. Ozone concentrations were measured using a Thermo
1Measurements were made by James Lee and Sarah Moller from the University of York. Our
collaboration in this measurement-model comparison is part of the NERC OP3 project.
81
Environmental Instruments (TEI) 49i UV absorption ozone analyser. The data was
internally averaged to one minute frequency and the detection limit was 0.6 ppbv.
On board the FAAM BAe 146 aircraft, NO and NO2 were measured using the
the University of East Anglia (UEA) NOxy instrument, which employed the same
technique as the ground based instrument described above2. Zeros were carried out
at the beginning of level runs during the flights and calibrations took place during
transit to and from the airport. Detection limits of the UEA NOxy are of the order
of 3 pptv for NO and 15 pptv for NO2 for 10s data, with estimated accuracies of
10% for NO at 1 ppbv and 10% for NO2 at 1 ppbv. The instrument is described in
detail by Brough et al. [2003]. Ozone was measured on board the aircraft using a
TEI 49C UV absorption analyser.
Isoprene fluxes, used in the box modelling experiments, were measured using a
PTR-MS instrument at the Bukit Atur site3. Its response was optimized to achieve
the best compromise between the optimal detection limit for VOCs and the min-
imization of the impact of high relative humidity. The operational details of the
instrument have been presented elsewhere [e.g. Lindinger et al., 1998; de Gouw
et al., 2003; Blake et al., 2009].
4.1.2.2 Discussion
Time series of NO, NO2, and O3 data for the first OP3 intensive observation period
are shown in Figure 4.1. Although the frequency of data collection is 1 minute (sec-
tion 4.1.2.1), it is shown here with a running average of 10 minutes for smoothing
purposes. NO levels were typically below 0.1 ppbv, although there were regular
spikes above this level which reached up to 0.4 ppbv. NO2 levels were higher, gen-
erally below 0.4 ppbv but reaching 0.8 ppbv. Ozone concentrations ranged from
near zero up to 30 ppbv, but were only consistently above 20 ppbv on three days
2Aircraft measurements were carried out by Dave Stewart and Claire Reeves of the University of
East Anglia.
3Isoprene flux measurements were carried out by Ban Langford from Lancaster University and
Pawel Misztal from the Centre for Ecology and Hydrology and the University of Edinburgh.
82
(11-13th April).
Figure 4.1. Time series of measured NO, NO2, and O3 at the Bukit Atur GAW
ground site, plotted versus local time (GMT+8) during the first intensive observation
period of OP3.
Figure 4.2 shows the median diurnal profiles for the entire April measurement
period for all three species. The 25-75 quartile limit is shown in shaded regions
around each of the profiles. The ozone diurnal cycle shows a minimum of approx-
imately 6 ppbv around 7:00 h followed by a rise through the morning. Ozone con-
centrations of approximately 11 ppbv remain until the evening, when concentra-
tions slowly fall to their minimum in the morning. NO2 concentrations exhibit the
most marked diurnal cycle, which peaks at midnight around 240 pptv and reaches
a low of 80 pptv in mid-afternoon. The slow loss of NO2 between midnight and
midday occurs less rapidly than the buildup between late afternoon and evening.
An NO peak of around 70 pptv is observed at 8:00 h and quickly recovers to a
fairly constant level between 30 and 40 pptv. This persists until 18:00 h when a
83
further drop to 20 pptv occurs. Non-zero NO concentrations between 15-20 pptv
persist throughout the night.
In July, an aircraft joined the campaign in order to make dedicated measure-
ments above the site and over the surrounding areas. In the lower panel of Figure
4.2, the diurnal cycles of NO, NO2, and O3 are shown for this second observation
period. These diurnal cycles are sampled only for the four days for which equiva-
lent aircraft data is also available. The average measurements made in profile flight
patterns directly over the site are plotted as whiskered points and show values for
both boundary layer and free troposphere.
O3 shows little vertical structure compared to ground measurements. The
ground based measurements vary little in this time period, remaining around 9
ppbv. Morning aircraft measurements are slightly higher (10-12 ppbv) than the
ground based. Aircraft measurements of ozone levels rise slightly to approximately
13 ppbv in the late afternoon, though boundary layer and free troposphere values
remain identical (within uncertainty) to each other. A diurnal pattern in the ground
based O3 observations is not clear.
84
-20 
-15 
-10 
-5 
0 
5 
10 
15 
20 
0.0 
0.1 
0.2 
0.3 
0.4 
0.5 
0.6 
6:00:00 AM 9:00:00 AM 12:00:00 PM 3:00:00 PM 6:00:00 PM 
O
3  [p
p
b
v] 
N
O
x
 [p
p
b
v]
 
NO NO2 
NO (BL) NO2 (BL) 
NO (FT) NO2 (FT) 
O3 
O3 (BL) 
O3 (FT) 
-20 
-15 
-10 
-5 
0 
5 
10 
15 
20 
0.0 
0.1 
0.2 
0.3 
0.4 
0.5 
0.6 
0.7 
0.8 
12:00:00 AM 6:00:00 AM 12:00:00 PM 6:00:00 PM 12:00:00 AM 
O
3  [p
p
b
v] 
N
O
x
 [p
p
b
v]
 
Figure 4.2. Top: Median diurnal cycle of ground-based measured NO (dark red),
NO2 (dark green), and O3 (dark blue) in April. The corresponding 25-75 quartile
interval is shown with each measurement: NO in pink, NO2 in light green, and O3
in light blue. Bottom: median diurnal measurements in July, shown only for the
days when corresponding flight data is available between 6:00 and 18:00 h; diurnal
profiles are the same color as above. Average flight data for morning and afternoon
profiles above the site are shown as whiskered points and are separated by height. NO
measurements are show in red (boundary layer) and yellow (free troposphere). NO2
is shown in light green (boundary layer) and brown (free troposphere). O3 is shown
in light blue (boundary layer) and purple (free troposphere)
85
Boundary layer NO2 matches the ground based measurements very closely,
which remain in the range of 100-200 pptv for most of the day. NO2 measurements
show a similar structure (rise until midnight and subsequent decrease afterwards)
to the first campaign, but because only four days are sampled here the full diurnal
cycle is not shown. NO2 measurements of around 20 pptv in the free troposphere
are much lower than those in the boundary layer and at the surface, demonstrating
that NO2 has a strong vertical structure. NO displays a similar pattern to NO2 with
boundary layer values of 80-200 pptv, which resemble ground-based measurements
well, and free tropospheric values that are much lower (less than 10 pptv). The
diurnal cycle of NO also bears strong resemblance to that of the first campaign,
(i.e. a rise in early morning followed by a slow tapering into the afternoon).
For comparison, the NO concentrations at the ground site in both measurement
periods are in between measurements made in the Amazon Rainforest of 20 pptv
[Lelieveld et al., 2008] and 100 pptv [Karl et al., 2009]. Ozone, on the other
hand, is lower at the Borneo site than in reported values for the Amazon for both
the boundary layer (19 ppbv) and the free troposphere (37 ppbv) [Lelieveld et al.,
2008].
4.1.3 Modelling
4.1.3.1 Model description
I use a zero-dimensional, stationary box model to simulate the diurnal cycles of
NO, NO2, and O3 observed during the first intensive observation period. The
model uses the same chemical mechanism (the ASAD chemistry package [Carver
et al., 1997]) which is implemented in UKCA and was run at 10 minute chemical
timesteps. In the box model, six species are deposited: NO, NO2, O3, peroxy
acetyl nitrate (PAN), peroxy-methacrylic nitric anhydride (MPAN), and a lumped
nitrate species representing the products from isoprene + NO3 and isoprene nitrates
(ISON, [see Po¨schl et al., 2000, and section 3.2.2]), with deposition velocities from
the literature [Zhang et al., 2003b; Ganzeveld and Lelieveld, 1995] which are set
86
to either an average daytime or an average nighttime value depending on the model
timestep. The photolysis follows the scheme of the Master Chemical Mechanism
[MCM, Saunders et al., 2003]. As those photolysis rate constants are originally for
July, at 45◦N for clear sky conditions, the rate constant for JNO2 was reduced by
50% to account for clouds and aerosol.
Only NO and isoprene are emitted into the box. NO emissions were 600 pptv
day−1 (the flux is constant) unless otherwise mentioned. Isoprene emissions into
the model are taken from ground based flux measurements. Occasionally, flux
measurements were not available; in these instances, we used linear interpolation
to connect the fluxes. In the absence of NO flux measurements at the site, we were
not able to constrain NO emissions to a diurnal pattern. NO flux measurements
were made in a nearby site underneath the canopy layer [Hewitt et al., 2009a]. The
Bukit Atur GAW station was in a clearing, and therefore canopy flux measurements
are not representative of this site, as there can be a strong difference between below-
and above-canopy fluxes of NOx [e.g. Duyzer et al., 2004]. For these reasons, we
assume that NO is emitted constantly into the model; this is also consistent with
how emissions are done in the global version of UKCA.
The box model boundary layer height is fixed to a set value during the day (6:00
to 18:00 h) and fixed to a different value at night. For the first set of sensitivities,
the daytime boundary layer height was set to 1000 m, and the nighttime height was
set to 200 m. The boundary layer height is effectively a mixing depth, and therefore
controls the range over which emissions are mixed into the model, and the rate of
sinks via deposition.
Temperature is set to 25◦C and pressure to a surface value of 1013 mb in the
box, appropriate to the conditions of the rainforest site. A 5◦C variation in tem-
perature showed negligible impact on box model output (not shown), so a constant
temperature was used. Latitude was set to 5◦N. Initial concentrations of chemical
species in the box model are set to the values shown in Table 4.1. NO, NO2, and
O3 are initialized to their midnight values from the diurnal cycle in the measure-
87
Table 4.1. Initial concentrations of six species used in the box model.
Species Concentration
CO 130 ppbv
HOOH 3 ppbv
C2H6 500 pptv
C3H8 50 pptv
HCHO 1 ppbv
CH3COCH3 50 pptv
ments. All other species are initialized to zero.
For the following model-measurement comparisons, both the model and the
data are sampled for 15 days to account for day to day variability in isoprene flux
measurements.
4.1.3.2 The base case
Figure 4.3 shows the initial results from the box model (dark blue line). The box
model overestimates NO concentrations (80 pptv instead of 60 pptv), though the
structure is decently captured. The shape of the NO2 curve is very different from
the measurements, with NO2 building up throughout the night. This occurs be-
cause any NO emitted into the model reacts with ozone to form NO2. Ozone is
also overestimated in the box model. The box model calculates values of approxi-
mately 30 ppbv, with an 8 ppbv diurnal cycle; measurements of 6-11 ppbv are far
below this level.
88
N
O
 [p
p
tv
]
N
O
2 [
p
p
tv
]
O
3
 [p
p
b
v]
0
20
40
60
80
100
120
100
200
300
400
500
0
10
20
30
40
0 4 8 12 16 20 24
0 4 8 12 16 20 24
0 4 8 12 16 20 24
Hour
Figure 4.3. Box model comparison to measured diurnal cycle of NO, NO2, and O3
(black line with shading) shown dark blue (without venting) and light blue (the ‘base
case’).
89
To obtain the ‘base case’ box model run (Figure 4.3, light blue line), one key
adjustment was made to the original box model output. A dilution parameter was
introduced to simulate mixing with the free troposphere resulting from a collapse
of the boundary layer at night. This ‘venting parameter’ removes a small frac-
tion—between 0 and 4 percent—of chemical tracers at each 10 minute timestep
between 24:00 h and 6:00 h. Doppler lidar measurements of the backscatter from
aerosol [Pearson et al., 2009] provide strong evidence for dilution of aerosol in the
Danum Valley boundary layer during this period. These measurements were made
slightly afield from the Bukit Atur GAW site used for NOx and ozone measure-
ments in an area of complex topography; the elevation of the lidar measurement
site was 198 m compared with 426 m for the Bukit Atur site. The median bound-
ary layer height dropped to approximately 200 m according to these measurements,
suggesting that on some nights the Bukit Atur site may have effectively been in the
free troposphere. The venting parameter is a simple way to simulate the mixing be-
tween the boundary layer box and the free troposphere by parameterizing dilution
of species which are concentrated in the boundary layer. As is evident in Figure
4.3, the box model was not able to capture the diurnal structure of NO2 before the
introduction of the venting parameter.
The base case (shown in Figures 4.3 and 4.4) shows both the strengths and
weaknesses of the box model with venting. A good fit to the data for NO is ob-
tained, simulating the morning rise in concentration due in part to the onset of
photolysis of NO2. During the day modelled NO follows a similar decay pattern
to the measured data, an improvement over the model without venting.
NO2 measurements and model simulations are in good agreement following
the addition of the venting parameter (the base case) and the buildup of NO2 until
midnight and subsequent reduction in concentrations is well captured. At sunrise,
when the boundary layer begins to grow, a steep drop in NO2 concentrations ap-
pears around 6:00 h in the box model due to the onset of photolysis. The largest
divergence between modelled and measured values occurs in the afternoon, be-
90
tween 12:00 and 16:00 h, though this is still relatively small. Base case modelled
and measured ozone display similar diurnal cycles. Both show minima between
7:00 and 8:00 h and maxima in the late afternoon, though the measurements have
more structure in the afternoon than the modelled ozone. However, the box model
underestimates the total concentration of ozone, simulating values of 0-8 ppbv in-
stead of 6-11 ppbv.
In summary, the ‘base case’ box model run performs reasonably well, although
ozone is too low. The inclusion of the venting parameter, which parameterizes
dilution of boundary layer air between 24:00 and 6:00 h, seems necessary in order
to simulate the structure of NO2 measurements at night.
4.1.3.3 Chemical sensitivities
I have performed a range of sensitivity experiments to explore the model perfor-
mance. First, a series of emissions sensitivities were carried out (not shown.) In
order to determine if the nighttime NO concentration could be captured if emis-
sions were altered, a sensitivity study was performed in which emissions of NO
were tripled to 1.8 ppbv day−1. This did not improve the agreement between the
modelled and measured values at night. Nonzero nighttime NO could arise from
emission taking place very near to the measurement inlet, which is not reproduced
by the box model as NO quickly reacts with O3 to form NO2 in a zero dimensional
model.
I also examined the sensitivity of the model to changing isoprene emissions.
Doubling the emission fluxes into the model reduced NO during the day by 12
pptv, and NO2 was reduced by a similar amount; ozone hardly changed. Overall,
the diurnal patterns were very similar between the two runs, and between these two
studies I determined that the regime was likely not emissions controlled. Although
it has been proposed that some chemistry in high-VOC environments might be ex-
plained by the presence of unknown reactive hydrocarbons [see Di Carlo et al.,
2004, and references therein], any VOC with similar reactivity to that of isoprene
91
Table 4.2. Summary of chemical sensitivity tests
Short name Figure Colour Description
Base Light Blue The base case
O3 phot Aqua JO3 divided by 3
Vd O3 Orange Ozone dep. velocities
reduced by 75%
Horowitz Green NOx recycling rates
as in Horowitz et al. [2007]
ISON dep Yellow ISON tracer dep. velocities
set equal to those of PAN
Reinit Red Reinitialized the model
each day at midnight
Init O3=9 Purple Reinitialise at midnight
with 9 ppbv O3
seems to be unable to explain any divergence in the model-measurement compari-
son.
A summary of the various chemical sensitivity runs is shown in Table 4.2,
with corresponding results plotted in Figure 4.4. Generally, the diurnal cycle was
modelled reasonably well for all the chemical parameters tested and was relatively
insensitive to chemical changes. The overall diurnal structure for NO is well cap-
tured, with the maxima at 8:00 h. With the inclusion of the venting parameter, the
NO2 diurnal structure is also always well simulated. All the chemical sensitivities
capture the cycle of ozone but not the magnitude.
92
0 4 8 12 16 20 24
0
20
40
60
80
100
120 Data
Base
Reinit
Horowitz
Init O3=9
O3 phot
Vd O3
ISON dep
0 4 8 12 16 20 24
50
100
150
200
250
300
350
400
N
O
2
 [p
p
tv
]
0 4 8 12 16 20 24
Hour
0
5
10
15
20
O
3
 [p
p
b
v]
N
O
 [p
pt
v]
Figure 4.4. 15 day average diurnal NO [pptv], NO2 [pptv], and O3 [ppbv] from
measurements (black) with 25-75 quartile limits shown in the shaded grey. Seven
model experiments are overlaid in various colours: the base run is shown in light
blue, reduction of ozone photolysis rate is shown in dark blue, reduction of ozone
deposition velocities is shown in orange, adjustment of NOx recycling rates is shown
in green, ISON deposition change is shown in yellow, reinitialization at midnight is
shown in red, and reinitialization with high ozone is shown in purple.
93
In the first chemical sensitivity test (Figure 4.4, aqua line), the photolysis rate
of O3 was reduced by a factor of three. The chemical mechanism shows very
little sensitivity to JO3 , barely changing from the base case run. In the second test,
ozone deposition velocities (both daytime and nighttime values) were reduced by
75% (Figure 4.4, orange line). For ozone, this simulation has the most impact of
any of the chemical sensitivity studies, but still does not increase the concentration
enough to match measured values. The change in deposition velocities also alters
the shape of the diurnal cycle, as nighttime deposition velocities approach a limit
of zero.
In an attempt to keep ozone production values high by increasing the concen-
tration of NOx in the system, an additional simulation was carried out. Recycling
of NOx from the reaction of ISON with OH was modified by increasing the ISON
+ OH rate constant from 1.3 x 10−11 cm3 s−1 [Chen et al., 1998; Po¨schl et al.,
2000] to 4.5 x 10−11 cm3 s−1 [Horowitz et al., 2007] (Figure 4.4, green line). I
also performed an experiment in which NOx concentrations would decrease; in
this sensitivity study, ISON deposition velocity was increased to match nitric acid
(Figure 4.4, yellow line). Neither experiment has an impact on the modelled values
of O3.
Two computational tests were also performed. In the first, the model species
concentrations are reinitialized each day at midnight, rather than using the values
calculated by the model the previous day (Figure 4.4, red line). This introduced a
stronger bias in NO and NO2 around 6:00 h, the first time photochemistry turns on
after reinitialization. A second study reinitialized the model at an artificially high
value of ozone, and this too displays a similar model bias at sunrise (Figure 4.4,
purple line). These two experiments give confidence that the model sensitivity to
initial conditions is eliminated by reusing the concentrations calculated from the
previous day.
The six studies discussed emphasize that the results from the UKCA chem-
ical mechanism are relatively robust to chemical, photolytic, and deposition rate
94
changes. As models with this level of complexity, particularly for isoprene ox-
idation, seem to perform relatively similarly [e.g. Emmerson and Evans, 2009;
Archibald et al., in press], it seems plausible that this is not an artefact of the
model mechanism itself. From my analysis, it appears likely that the regime is
more sensitive to physical processes and parameterizations than chemical ones. In
order to assess the impact of these chemistry factors in relation to physical param-
eters controlling the processes of emission, mixing and deposition, I conducted a
further series of experiments based on physical variables.
4.1.3.4 Physical sensitivities
As shown in Figure 4.2, the measured data show vertical structure in both NO and
NO2 with much lower concentrations in the free troposphere compared with the
boundary layer. The venting parameter simulates exchange with free tropospheric
air at night and assumes that this incoming air has lower concentrations of NO,
NO2, and ozone. However, O3 displays little to no vertical structure in the mea-
surements (Figure 4.2). For this reason, a simulation was run in which venting of
ozone was turned off while all other species continued to be diluted. Not venting
O3 is the numerical equivalent of removing O3 and introducing an equal amount
during the same amount of time, such that a collapse of the boundary layer and
influence of the free troposphere may well bring in ‘new’ ozone, but the concen-
trations will be similar to the boundary layer air it is replacing. This hypothesis is
reinforced by the difference between the species in their distribution of sources and
sinks; NOx has a source which is largely surface dominated at a remote rainforest
location (higher in the troposphere, lightning can contribute as well), while ozone
has a strong surface sink due to deposition.
95
0 4 8 12 16 20 24
0
20
40
60
80
100
120
N
O
 [p
p
tv
]
50
100
150
200
250
300
350
400
N
O
2
 [p
p
tv
]
Hour
0
5
10
15
20
O
3
 [p
p
b
v]
0 4 8 12 16 20 24
0 4 8 12 16 20 24
Figure 4.5. Model comparison to diurnal cycle of NO, NO2, and O3 (black with
shading) without (orange) and with (the base case, blue) venting ozone.
96
Figure 4.5 shows the results of the box model when ozone is not vented, which
displays much better agreement with the measurements. The amplitude of the di-
urnal cycle of ozone is reduced when O3 is no longer vented, as the nighttime sinks
are reduced. This dampened cycle more closely matches the measured diurnal pro-
file. NO2 changes little when ozone is not vented, but NO is greatly improved,
particularly in the early daytime hours. By keeping O3 in the box during the night,
the chemical sink for NO remains higher, and the elevated values of NO in the
morning are correspondingly reduced.
Three variables are used to further test the physical boundaries of the box
model: the exact quantity of material lost at night (the venting parameter), the
height of the boundary layer during the day, and the height of the boundary layer at
night. In order to obtain the best value for these three parameters, a cost function
analysis was used:
CFx =
1
t
∑
t
(|modelx −measuredx|)
measuredx
(4.1)
where for each species (denoted by x) and at each timestep (t), the difference be-
tween modelled and measured values of NO, NO2, O3 are evaluated and averaged
over a 24 hour day and 15 experiment days. The cost function gives the average
deviation of the model from the measurements expressed as a fraction, where zero
is a perfect match. The NO cost function is only evaluated between 6:00 and 18:00
h due to the mechanism’s inability to capture nighttime NO concentrations, so that
results are not skewed because of nighttime values. The results of the three cost
functions are shown in Figures 4.6 and 4.7, where a lower value of the cost function
represents better agreement between measurements and modelled concentrations.
The NO cost function shows a dependence on the venting fraction until the
value of 2% per timestep, at which point model and measured data converge to
a reasonable agreement of less than 30% difference in value. At the zero value
for venting, NO shows almost no dependence on the nighttime boundary layer
97
a b
Cost Function Analysis for NO
2
Ve
n
ti
n
g
 F
ra
ct
io
n
1%
2%
3%
4%
0%
Boundary Layer Height Night Boun
dary L
ayer H
eight 
Day
500
1000
1500
2000
500
1000
1500
2000
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
 
0.40
0.20
0.20
0.40
0.40
0.20
0.20
0.40
0.20
0.20
0.20
0.40
Cost Function Analysis for NO
Ve
n
ti
n
g
 F
ra
ct
io
n
1%
2%
3%
4%
0%
Boundary Layer Height Night Boun
dary L
ayer H
eight 
Day
500
1000
1500
2000
500
1000
1500
2000
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
 
0.40
0.60
0.20
0.40
0.20
0.20
0.20
0.20
Figure 4.6. Cost function analysis of model comparison to diurnal average NO (a),
and NO2 (b). A lower value represents closer agreement between the box model and
the measurements.
height, which reflects the fact that the cost function is only evaluated during day-
time hours. NO matches the measurements best (with a cost function value below
0.20) for high values of the daytime boundary layer height, though the gradient of
dependence on daytime boundary layer height decreases with increasing venting
fraction.
At a 0% value for the venting parameter, the NO2 cost function shows values
of 0.30 to 0.80. With venting, the levels are lower (up to 0.30), which suggests that
the best fit requires at least some material to be removed at night. Between 1% and
4%, however, NO2 displays little variation in the cost function, and the entire cost
function ‘space’ is valued under 0.40. NO2 also shows very little dependence on
the nighttime boundary layer height, demonstrating that venting is a more impor-
tant loss process than deposition. The height of the boundary layer during the day
is important only at heights less then approximately 500 m and above 1800 m.
Ozone was not vented in these experiments, so the cost function for ozone is
98
Cost Function Analysis for O
3
Ve
n
ti
n
g
 F
ra
ct
io
n
1%
2%
3%
4%
0%
Boundary Layer Height Night Boun
dary L
ayer H
eight 
Day
500
1000
1500
2000
500
1000
1500
2000
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
0.20
0.20
0.40
0.60
0.80
0.20
0.40
0.60
0.80
0.20
0.40
0.40
0.60
0.60
0.80
0.20
0.40
0.60
0.60
0.80
0.40
0.60
0.80
Figure 4.7. Cost function analysis of model comparison to diurnal average O3.
relatively stable in relation to venting parameter. Ozone shows a very high sen-
sitivity to the boundary layer height during the day (with values ranging between
0.10 and 1.0), presumably due to deposition, and little dependence at night except
below 500 m.
4.1.3.5 Best fit to measurements
Figure 4.8 shows the best fit to the measurements obtained using the box model.
The values for the venting parameter (2%), boundary layer height during the day
(1200 m) and night (750 m) were taken from the cost function analysis minima.
The results show good agreement between measured and modelled values, captur-
ing the majority of structure and diurnal variation for all three measured species.
NO matches particularly well, though the model is still not able to simulate the
residual concentrations at night. As described before, these could arise from a
99
highly stratified boundary layer, or rapid mixing times up from the soil to the mea-
surement inlet before chemical reaction. In either case, these processes are very
small scale, and beyond the capability of a global model (with a resolution of hun-
dreds of kilometres) to capture physically.
Modelled NO2 is slightly higher than the measured values but captures the
structure of the measurements effectively. In particular, the nighttime structure
of NO2 is well simulated once venting was included in the box model. After-
noon NO2 concentrations are also slightly higher in the model than measurements.
Since my analysis shows that transport and physical processes dominate the diur-
nal structure, perhaps this afternoon discrepancy arises from afternoon convection.
Ozone looks very similar to measurements, though the rapid rise in the morning is
not entirely captured. Nevertheless, the magnitude and basic form of the diurnal
cycle are simulated well.
With the success of the chemical mechanism in this region in mind, I ran a
regional sensitivity study in which phase I of the biofuel lifecycle was represented
in a simple, idealized manner.
100
020
40
60
80
100
120
N
O
 [p
p
tv
]
50
100
150
200
250
300
350
400
N
O
2 [
p
p
tv
]
0
5
10
15
20
O
3 [
p
p
b
v]
0 4 8 12 16 20 24
0 4 8 12 16 20 24
0 4 8 12 16 20 24
Hour
Figure 4.8. ‘Best fit’ model (red) compared to diurnal average NO, NO2, and O3
(black with shading) after optimized adjustments to the venting parameter and bound-
ary layer heights.
101
4.2 Regional sensitivity study
The expansion of oil palm production in Southeast Asia has been identified as a
major threat to the carbon balance [Danielsen et al., 2008], bird [Peh et al., 2006]
and insect [Turner and Foster, 2009] populations, and to biodiversity in general
[Pin Koh and Wilcove, 2007]. Oil palm is also an extremely high emitter of iso-
prene [Kesselmeier and Staudt, 1999, also shown in Table 2.2], and could therefore
influence air quality significantly. In this section, my aim is to test the hypothesis
that phase I of the biofuel life cycle can impact tropospheric chemistry through iso-
prene emissions. The first experimental sensitivity study, in which isoprene emis-
sions are doubled, could represent a transformation of forested areas to managed
oil palm plantations; the second study could represent a transition to a non-emitting
crop, such as sugar cane from oil palm or rainforest. Both types of crops can be
used in biofuel production (see section 1.3).
4.2.1 Experimental setup
Figure 4.9. The area of emission perturbation in two sensitivity studies (outlined in
red). The location of the OP3 field campaign is denoted by the blue cross.
Using UKCA, I ran two simple studies in which the emissions of isoprene
102
were altered. In the first, isoprene emissions were doubled (the OILPALM run) in
the region between -25◦S, 25◦N, 70◦E, and 160◦E. In the second study, isoprene
emissions were reduced by 25% (the SUGARCANE run) in the same region. The
area of perturbation is shown in Figure 4.9. The model was run for 16 months with
the first four months taken as spin up. The model year is 2001, and the sea surface
temperatures are the same as in chapter 3 [Rayner et al., 2003].
4.2.2 Impact on ozone
Figure 4.10. Difference in monthly mean July surface ozone [ppbv] between the
OILPALM and BASE simulations.
Changes in surface ozone resulting from the OIL PALM simulation in South-
east Asia are depicted in Figure 4.10. July values are shown, as isoprene emissions
in the region peak in July (see Figure 3.1) and thus the impact on the local envi-
103
ronment regime will be largest at that time. Over the area of perturbation, ozone
concentrations drop when isoprene emissions are doubled. The strongest changes
modelled are up to -10 ppbv, though over the oceans the difference is generally re-
stricted to values above -6 ppbv. The outflow area from Southeast Asia through to
the South Pacific is evident from an area of small change in ozone concentrations
off the western coast of South America.
hy
br
id
 h
ei
gh
t
Figure 4.11. Difference in monthly mean July ozone [ppbv] averaged between 70E
and 160E (the area of perturbation) between the OILPALM and BASE simluations.
The chemical tropopause of 150 ppbv of ozone, as defined by Stevenson et al. [2006],
is shown as a black line.
The effect of the OILPALM run is evident higher in the troposphere as well,
as shown in Figure 4.11. The magnitude of the impact in the free troposphere is
lower (6 pbbv maximum) than near the surface, though the distribution is more
widespread. The largest differences are concentrated in the colder regions of the
tropical upper troposphere: at 8000 m, changes in ozone of -1 to -4 ppbv are ev-
ident across much of southern Asia and the maritime continent, as well as most
of the Pacific (not shown). The lower temperature is important for the lifetime
104
of PAN [Moxim et al., 1996], where concentrations increase more (5-10%) in this
region than elsewhere in the free troposphere. To show the spatial distribution
in relation to ozone, the change in surface PAN concentrations resulting from the
OILPALM simulation is shown in Figure 4.12. The formation of PAN increases
in this study, removing NOx from the local chemical environment and impeding
ozone formation. The simultaneous reductions in ozone can therefore be explained
by this ‘locking up’ of NOx, preventing the conversion of NO radicals to NO2.
Figure 4.12. Difference in monthly mean July surface PAN [pptv] between OIL-
PALM and BASE.
To investigate the linearity of the relationship between isoprene emissions and
ozone production, a second sensitivity test was performed in which isoprene emis-
sions were reduced by 25% (the SUGARCANE run). The change in surface ozone
in July resulting from this second experiment is shown in Figure 4.13. Ozone con-
105
Figure 4.13. Difference in monthly mean July surface ozone [ppbv] between SUG-
ARCANE and BASE
centrations increase in the SUGARCANE run; in fact, the values shift by a greater
amount than when emissions are doubled. The perturbation to emissions does not
yield a direct change in ozone; rather, a small reduction (-25%) changes ozone
by a larger amount than a large increase (200%). The linearity of the maximum
change in ozone as a response to these simple perturbations to isoprene emissions
is shown in Figure 4.14. The response curve shown here is just one transect from
a more complete ozone production isopleth plot such as the one shown in section
2.1.3.
Young [2007] performed a similar isoprene emission reduction sensitivity study
to my SUGARCANE run using the UM CAM model in a future atmosphere with
A2 emissions. In January, when emissions shift South, an increase in ozone of up
106
M
ax
im
um
 c
ha
ng
e 
in
 o
zo
ne
 [p
pb
v]
Perturbation to isoprene [%]-15 
-10 
-5 
0 
5 
10 
15 
-40 -20 0 20 40 60 80 100 120 
Perturbation to isoprene [%]
Figure 4.14. Linearity in maximum ozone response between OILPALM, SUGAR-
CANE, and BASE. The x axis shows the amount of perturbation to isoprene emis-
sions (%), and the y axis shows the maximum response in July surface ozone [ppbv]
resulting from the perturbation.
to 5 ppbv over Australia was found, similar to the results from UKCA. Interest-
ingly, a decrease in ozone in southeast Asia and India was also reported, colocated
with strong NOx emissions in the A2 scenario which effectively convert the region
to a VOC-sensitive environment. Wiedinmyer et al. [2006] also did a simple sensi-
tivity test assuming a 50% reduction of isoprene emissions in the Amazon and the
southeastern United States. In the VOC-rich environment of the Amazon, ozone
increased by 3-5 ppbv in July, lower in value than the response given by UKCA in
the southeast Asia. In contrast, the more polluted environment of the southeastern
United States displayed a decrease in monthly mean ozone of 2-4 ppbv [Wiedin-
myer et al., 2006].
4.2.3 Uncertainties
The uncertainties associated with this type of simple sensitivity study are large.
With a large-scale shift towards oil palm production, other, non-isoprene emissions
could change simultaneously. Carbon [Houghton, 2003] and nitrogen [Williams
107
et al., 1992; Bouwman et al., 2002] emissions associated with land use change, fer-
tilizer emissions, and post-processing power plant emissions could increase. Some
terpenoid emissions from the forest or reclaimed land could decrease, depending
on the speciation of the forest; simultaneously, others could increase. Furthermore,
changes in biogenic isoprene and terpenoid emissions could also influence SOA
production (section 2.2.1.2), which could have consequences for local climate and
meteorology.
4.3 Summary
In this chapter, I began with a description of measurements of NO, NO2, and O3
taken during the OP3 field campaign. The box model was then presented, and a
number of chemical sensitivity studies were reported on; their success in match-
ing measured ozone levels was limited. A series of physical sensitivity studies
proved to have greater success in matching the model output to measured data.
I optimized a cost function based on daytime boundary layer height, nighttime
boundary layer height, and venting parameter. This yielded a ‘best fit’ box model
simulation which was quite successful at reproducing the data. My work demon-
strates that the chemical mechanism in UKCA can indeed reproduce measurement
data in a remote tropical rainforest site, but that it is sensitive to parameterizations
of physical processes.
In the second section, I performed two simple sensitivities in which isoprene
emissions were altered. Although simple, these changes could be associated with
the growth of biofuel feedstocks (either oil palm for biodiesel, or sugarcane for
ethanol) in the region. I then described the summertime ozone reductions resulting
from the OILPALM perturbation and the associated increase in PAN. Looking at
the linearity in the response, the change in ozone was stronger for the decrease
in VOC emission (SUGARCANE) than for the increase (OILPALM). Since the
response in ozone is non-negligible, it appears that phase I of the biofuel life cycle
108
can indeed influence air quality and tropospheric chemistry.
In the next chapter, I attempt to simulate a more realistic representation of
future crop changes and to characterize componenents of a potential future atmo-
sphere.
109
Chapter 5
Future Ozone Attribution and
Phase I: Cultivation
In this chapter, I examine a global representation of phase I of the biofuel life
cycle. In section 5.2 I describe the experimental setup and methodology, and in
section 5.3 I look at how isoprene emissions could evolve in the future. Section
5.4 is concerned with a characterization of a potential future atmosphere using
the concept of ozone attribution, with particular focus on emissions and physical
climate change. In section 5.5, I examine the impact of land use change (expansion
of land dedicated to growing crops) related to phase I of the biofuel life cycle. I
then compare the impact of land use change to another potential isoprene emission
perturbation, CO2 suppression (previously described in section 2.2.2.2) in section
5.6.
5.1 Introduction
5.1.1 How will ozone change in the future?
The relationship between ozone and climate is complex. On a theoretical level, it
starts with the direct relationship between temperature and chemistry. Higher tem-
110
peratures drive faster rates of chemistry [e.g. Wild, 2007], which can increase or
decrease ozone production depending on the local chemical environment (section
2.1.2). High ozone concentrations and temperature are also correlated on the ba-
sis of the decrease in PAN lifetime (which has a thermal dependence, see section
2.1.4) and the associated increase in HOx and NOx released from PAN degradation
[Sillman and Samson, 1995]. An increase in isoprene emission driven by higher
temperatures could increase ozone in a high-NOx environment, while decreasing
ozone in a low-NOx environment [Roelofs and Lelieveld, 2000; Fiore et al., 2005].
Climate and chemistry are linked on an indirect level as well. Increasing at-
mospheric humidity, a result of higher temperatures [Solomon et al., 2007b], will
increase the destruction of ozone through the reaction of O(1D) with water vapour
[Thompson, 1992; Johnson et al., 1999]. In high-NOx environments, this could
increase ozone by stimulating oxidation processes through higher concentrations
of OH, while in low-NOx environments, the scavenging could reduce ozone levels
[Liu et al., 1987; Sillman and Samson, 1995]. Other competing climate variables
are also at play, including an increase in lightning NOx resulting from increased
convection [Zeng et al., 2008], an increase in stratosphere-troposphere exchange
[Zeng and Pyle, 2003; Hegglin and Shepherd, 2009], and changes in both wet
and dry deposition [Zhang et al., 2008]. Finally, elevated temperatures could po-
tentially alter the meteorology of the atmospheric system in ways that generally
encourage ozone formation, such as increasing stagnation events, heat waves, and
frequency of droughts [e.g. Mickley et al., 2004].
The relationship between ozone episodes and temperature has been empirically
established through observation [Cox and Chu, 1996; Lin et al., 2001], and has
led to a number of modelling studies investigating the relationship between future
climate change and air quality. Some studies have examined this relationship on
relatively short time horizons (between 2020 and 2030) [e.g. Johnson et al., 1999;
Stevenson et al., 2000, 2005; Racherla and Adams, 2006; Stevenson et al., 2006;
Unger et al., 2006; Wu et al., 2008a]. Several studies have examined the impact
111
of climate change on ozone using the same time horizon (2100) used here [e.g.
Stevenson et al., 2000; Hauglustaine et al., 2005; Brasseur et al., 2006; Liao et al.,
2006; Zeng et al., 2008]. The complexity of the relationship between ozone and
climate is such that there is not a consensus prediction for how ozone will change
in the future; for example, Hauglustaine et al. [2005] and Liao et al. [2006] found
positive changes in surface ozone over land areas, while Zeng et al. [2008] reported
a more geographically heterogeneous effect.
5.1.2 The impacts of future land use change
Another potential future perturbation to the Earth system is land use change. The
human species relies on the earth system for water, fuel, food, and other natural
resources to sustain itself. The result is that human influence over the biosphere
has been in effect for centuries [Vitousek et al., 1997a], and will continue to grow
as global population augments.
Ramankutty and Foley [1999] were the first to quantify the total historical influ-
ence of humans on global land, and they found that 1140 million ha of woodland
and 670 million ha of savannas/grasslands have been lost to cropland expansion
since the year 1700. A second attempt was undertaken by Klein Goldewijk [2001],
whose showed 1206 million ha of land has been cleared between 1700 and 1990,
while pastureland has grown by 2927 million ha. Hurtt et al. [2006] used an ac-
counting model to estimate that 42-68% of global land has been influenced by
anthropogenic forces in the last three hundred years. Pongratz et al. [2008] com-
piled a land use change dataset for the last millenium; they were able to extend
their analysis into longer timescales by using population data as a proxy for land
use change. The authors reported that 500 million ha of natural vegetation was
transformed to agricultural and pastoral land between 800 and 1700.
Anthropogenic land use change has been found to influence the climate system.
As a result of changes in surface albedo and evapotranspiration, Davin et al. [2007]
reported a -0.29 W m−2 radiative forcing due to anthropogenic land cover change
112
between 1860 and 1992. Pongratz et al. [2009] examined a longer timescale of 800
years in order to assess the impact of human-driven land use and land cover change
on radiative forcing between 1100 and 1900 and found a value of -0.21 W m−2.
Besides radiative forcing, land use has also been found to impact temperature and
wind estimates [Civerolo et al., 2000; Feddema et al., 2005; Lamptey et al., 2005;
Voldoire et al., 2007], soil suitability for growing crops [Ramankutty et al., 2006],
and precipitation [Lamptey et al., 2005; Junkermann et al., 2009]. These impacts
on the climate system will carry into the future; for example, Davin et al. [2007]
projected a -0.7 W m−2 radiative forcing between 1992 and 2100 on account of
land use change.
A number of gas fluxes into the atmosphere have also been influenced by land
use change. The fluxes of NO, N2O, and NH3 have been increased by approxi-
mately 80%, 37%, and 70% respectively as a result of agricultural control of the
nitrogen cycle [Vitousek et al., 1997b]. The flux of carbon to the atmosphere as a
result of anthropogenic land use change has been estimated to be between 1.6 and
2.1 Gt C yr−1 in the 1980’s [Jain and Yang, 2005] and 1.1 to 1.3 Gt C yr−1 in the
1990’s [Shevliakova et al., 2009]. As for the total flux, Houghton [2003] estimated
a carbon flux of 156 Pg C between 1850 and 2000 as a result of human driven land
use change. These types of calculations have associated uncertainties due to the
use of proxy data to create records about the past; Quaife et al. [2008] quantified
this uncertainty to be approximately -15% to 8% using the UK as a case study.
Carbon fluxes into the atmosphere due to land use change will continue into the
future, though the extent seems to be temperature, scenario, and model dependent
[Cramer et al., 2004; Gitz and Ciais, 2004; Sitch et al., 2005; Friedlingstein et al.,
2006; Mu¨ller et al., 2007]
Besides the biogeophysical and biogeochemical variables mentioned above,
land use can impact biogenic emissions such as isoprene. Tao and Jain [2005]
modelled changes in global isoprene emissions between 1981 and 2000, and found
that 2% of isoprene variability between 1981 and 2000 arose from variations in
113
land use (climate change and CO2 fertilization had larger effects). Guenther et al.
[2006] noted that present day global isoprene emissions change by -0.2 to 3% de-
pending on the land use dataset used as input to their model. Lathie`re et al. [2006]
ran two sensitivity studies: in the first case, representing tropical deforestation,
isoprene emissions were reduced by -29% globally and -35% in the tropics; in the
second, representing European afforestation, emissions changed by +4% globally
and +126% in Europe. The impact of land cover on isoprene emissions has also
been studied at smaller scales, notably for European forests [Arneth et al., 2008b],
though replacement of natural forests by woody biomass plantations did not alter
isoprene emissions significantly in their model study.
These changes in isoprene emissions have the potential to transform local tro-
pospheric chemistry. Wiedinmyer et al. [2006] found altered ozone concentrations
of -5 ppbv (over the United States) to 5 ppbv (over the Amazon) from a sensitivity
study in which isoprene emissions were lowered to represent urbanization. In a
second sensitivity, isoprene emissions were increased to represent a transition to
high-emitting tree plantations; this resulted in a decrease of ozone concentrations
over the Amazon of up to 9 ppbv and a mixed signal in the north and western United
States between -5 and 20 ppbv [Wiedinmyer et al., 2006]. Chen et al. [2009a] found
that future land use change decreased surface ozone over the United States under
an A2 scenario by up to 5 ppbv due to changing isoprene. Avise et al. [2009] re-
ported that the impact of future land use change on biogenic isoprene emissions
in the United States increased ozone by 1.3 ppbv in July. Using a single column
model, Ganzeveld and Lelieveld [2004] found that for two deforestation scenarios
in the Amazon, isoprene concentrations decreased, ozone rose by under 2 ppbv at
the surface, and surface OH increased by 50%. A 100% increase in biogenic VOC
emissions resulted in an increase in maximum 1-hour June ozone concentrations of
30% in the Mid-Atlantic region of the United States in the work of [Bell and Ellis,
2004]. Finally, a 1-6 ppbv change in daily 8-hour ozone maximum (DM8H) under
the A1B scenario in 2050 was attributed to future land use change in the Houston
114
BASE
CConly
ISOP
ANTH
BOTH
FutCROPS
FutCROPS_CO2
BOTH_CO2
Isop Emi. Anth Emi.Climate Crops CO2Name
Figure 5.1. A summary of emissions and climate sensitivity experiments.
area [Jiang et al., 2008].
In this chapter, my goal is to test the model sensitivity to land use change
and CO2 suppression of isoprene emission within the context of a potential future
atmosphere.
5.2 Experimental setup
In order to understand how and why ozone changes in a potential future atmo-
sphere, I ran a number of experiments to test the sensitivity of UKCA to climate
change, future anthropogenic emissions, and future biogenic emissions. I used a
sequential or serial methodology, such that each experiment added one more of
each of these variables; the difference between each sequential experiment yielded
the change from a single variable. Figure 5.1 shows the simulation names and a
summary of the experimental design.
The BASE run (1999-2005) has been described in detail in section 3.2, and
so will not be described here. In addition to the chemistry and emissions used in
the BASE simulation, the CConly integration included long wave radiative forc-
ing from CO2 and CH4 of 621 ppbv and 2973 ppbv, respectively. In addition, the
radiation code includes long wave absorption of three more species, whose concen-
trations were set as follows: N2O was set to 362 ppbv, CFC-11 was set to 45 pptv,
115
Present Crops Future Crops
No CO2
Supp.
CO2
Supp.
BOTH FutCROPS
BOTH_CO2 FutCROPS_CO2
Figure 5.2. Matrix of land use change and CO2 suppression experiments.
and CFC-12 was set to 222 pptv. Future sea surface temperatures (SSTs) were gen-
erated using the Met Office Hadley Centre Earth System Model (HadGEM) [Stott
et al., 2006], in which the runs of a coupled ocean-atmosphere model followed
the A1B IPCC SRES scenario [Nakicenovic et al., 2000] with 150 years spin up
period, 30 years of which included anthropogenic forcing. SSTs are a boundary
condition in UKCA (there is no slab ocean) and vary monthly and interannually
throughout the course of each integration. Each of the UKCA integrations set in
the future were allowed 16 months to spin up and were run for a further four years
corresponding approximately to the end of the century (2096-2099).
The ISOP integration included the physical climate forcing from the additional
long wave radiation and SSTs as well as future isoprene emissions. The ISOP run
was designed to test the changes resulting from future biogenic isoprene emissions.
The ANTH simulation also included future climate forcings, but future emissions
were limited to anthropogenic emissions (isoprene emissions remained at present
day values.) The BOTH simulation included both climate and emission (anthro-
pogenic and biogenic) future forcings as a sort of ‘best guess’ future run. The
difference between ANTH and CConly, and the difference between BOTH and
ISOP, test the response resulting from change anthropogenic emissions with both
present day and future isoprene emissions.
In addition to these, the FutCROPS run included the change in isoprene emis-
sions associated with projected future cropland expansion, which could be asso-
116
ciated with growth of biofuel feedstocks. The impact of CO2 suppression of
isoprene emission was included in the experiment called FutCROPS CO2. Fi-
nally, to compare the relative impacts of cropland expansion and CO2 suppression,
the BOTH CO2 run included all future forcings and CO2 suppression without a
change in cropland expansion. The last four experiments represent a matrix, shown
in Figure 5.2, designed to test CO2 suppression with present and future crops, and
land use with and without CO2 suppression all in a potential future atmosphere.
Future anthropogenic emissions were taken from a B2+CLE scenario devel-
oped for use in Fowler et al. [2008] and are summarized in Table 5.1. The scenario
was based on an updated IPCC B2 storyline [Riahi et al., 2007], and was further
revised to include legislation (Current LEgislation = CLE) and regulation passed
between 2002 and 2006 as described in Dentener et al. [2005] and Cofala et al.
[2007]. The B2+CLE scenario is an optimistic scenario that involves ‘local solu-
tions to economic, social and environmental sustainability’, and in which global
population increases at a moderate rate [Fowler et al., 2008]. I chose this moderate
scenario because my intention is to study phase I of the biofuel life cycle, and it
seems unlikely that large scale implementation of biofuels will occur contempora-
neously with a very pessimistic storyline (such as the A2). Biofuel regulation and
emissions reductions are almost always colegislated (see section 1.1).
The B2+CLE emissions used here were for the year 2050, but CO and NO2
emissions were projected to change less than five and ten percent, respectively,
between 2050 and 2095. This is well within the range of uncertainties between
and within the economic scenarios; for example, carbon dioxide emissions in the
original B2 scenario vary between 7 and 24 Gt C −1 (240%) within the sub-family
scenarios [Nakicenovic et al., 2000]. When compared with present day, NOx emis-
sions are reduced by 9.6 Tg N yr−1 (-17%) and CO by 38.6 Tg yr−1 (-3.6%).
NMVOC increases slightly by 16.1 Tg yr−1 (2.4%). The emissions listed in Table
5.1 are due to anthropogenic sources only. Soil, biomass burning, oceanic, and
natural (acetone) emissions were all kept constant with respect to the present day
117
Table 5.1. Future B2+CLE emissions of NOx, CO, and non-methane non-isoprene
VOCs.
Emitted species Amount [Tg yr−1]
NO2, Tg N yr−1
Anthropogenic 20.2
Soil 7.0
Biomass Burning 10.2
Total 37.4
CO, Tg CO yr−1
Anthropogenic 432.0
Biogenic 89.0
Ocean 11.0
Biomass Burning 507.5
Total 1039.5
VOC, Tg VOC yr−1
Anthropogenic 99.7
Natural (acetone) 48.3
Biomass Burning 31.3
Total 179.3
(Table 3.2, section 3.2.3). The increase in NMVOC between present and future
appears quite small, but is larger when isoprene is taken into account.
5.3 Future isoprene emissions
Three models were used to generate isoprene emission fields1. The Sheffield dy-
namic global vegetation model (SDGVM) [Woodward et al., 1995; Beerling et al.,
1997] was forced with mean monthly temperature, humidity, and precipitation de-
rived from the Unified Model [Johns et al., 1997] run without chemistry. Using this
data, the SDGVM was run to vegetative equilibrium around a fixed (no seasonal
variation) map of global crop distribution, as in section 3.2.3. Future (2100) crop
distribution was modelled by the Integrated Model to Assess the Global Environ-
ment [IMAGE, Alcamo, 1999] economic and policy model and compiled for input
1Isoprene emissions fields were generated by Juliette Lathie`re, as in Chapter 3.
118
into a dynamic global vegetation model. To explore the land use changes in their
most extreme and in the context of large scale biofuel implementation, the crop
map was forced with the SRES A2 scenario [Nakicenovic et al., 2000]. The output
from the SDGVM was then used in a BVOC emission model (also described in
section 3.2.3) based on the parameterizations of Guenther et al. [2006] to calculate
global isoprene emissions. Table 5.2 summmarizes the emissions estimates.
Table 5.2. Total global isoprene flux [Tg C yr−1] for eight studies, calculated by the
Sheffield Dynamic Global Vegetation Model.
Experiment Names Climate Crops CO2 Supp. Isop. [Tg C yr−1]
BASE, CConly, ANTH Pres. Pres. No 413
ISOP, BOTH Fut. Pres. No 482
BOTH CO2 Fut. Pres. Yes 283
FutCROPS Fut. Fut. No 314
FutCROPS CO2 Fut. Fut. Yes 184
The present day estimate for isoprene emissions of 413 Tg C yr−1 compares
well with recent literature estimates between 402 and 601 Tg C yr−1 [see section
2.2.3.4, and Lathie`re et al., 2009]. The present day calculation includes some land
use change forcing from the existing distribution of agricultural land (see section
3.2.4). Without any crops at all, the model estimated present day isoprene emis-
sion to be 471 Tg C yr−1, even closer to the median literature estimate for global
isoprene flux [Lathie`re et al., 2009].
In the future, a number of factors could influence biogenic isoprene emissions.
Figure 5.3 shows the relative percentage impact of climate change, CO2 suppres-
sion, and cropland expansion on isoprene emissions (each discussed below.)
5.3.0.1 Future isoprene
In the future calculation with present day crops (i.e. crops do not vary; this emis-
sion field was used in the ISOP and BOTH integrations), isoprene emissions in-
creased by 17% to 482 Tg C yr−1 compared to the present day. This estimate is
119
-50 -40 -30 -20 -10 0 10 20 
Climate
Land use
CO2 Suppression
Figure 5.3. Change in isoprene emissions [%] resulting from climate change, CO2
suppression, and cropland expansion. For the latter two, relative percentages are av-
erage for the two sets of model runs; for example, the CO2 suppression bar is an
average of the relative difference between emissions in the BOTH and BOTH CO2
and FutCROPS and FutCROPS CO2.
slightly lower than some other estimations of the future global isoprene flux [e.g.
Heald et al., 2009; Young et al., 2009] because crops are included in these calcula-
tions.
Figure 5.4 shows the difference in isoprene emissions between the present and
future runs. As the vegetation responds dynamically to future climate forcings, the
increases in future isoprene emission are due to increases in temperature (direct
effect) and to CO2 fertilisation of the biosphere (indirect effect). Directly, higher
temperatures drive higher isoprene emissions, as described in section 2.2.2.1. Indi-
rectly, because CO2 is the main source of carbon for photosynthesis and therefore
plant growth, elevated CO2 levels stimulate the growth of the biosphere; expansion
of the high growth, high-emitting ecosystems (PFTs) such as tropical rainforests
tends to increase global isoprene emissions in tropical regions such as the Amazon
and central Africa. The few areas which show a decrease in isoprene emission (e.g.
the northwest Amazon) are most closely linked to a decrease in soil moisture, one
of the driving variables for the BVOC emission model2.
2Juliette Lathı`ere, personal communication, 10 November 2009.
120
a  January
b  July
Figure 5.4. Change in isoprene emissions [g isoprene m−2 month−1] between
ISOP/BOTH and BASE simulations.
5.3.0.2 Land use change
Figure 5.5 shows the future change in crop fraction projected by the A2 scenario
according to the IMAGE model, illustrating large scale conversion of other land
types to managed agricultural and pastoral systems. The gridbox fraction of land
dedicated to growing crops increases dramatically in South America (425%), the
western United States (123%), central Africa (209%), and southeast Asia (141%).
In the central North America, Europe, and eastern China, crop fraction decreases
slightly (by less than 10%). Globally, the land area used for crops increases by
97%.
When forests (high emitter) were replaced by agricultural land (very low emit-
ter) in the FutCROPS simulation, reduced global isoprene emissions of 314 Tg C
121
Figure 5.5. Crop mask derived from the IMAGE model and applied to the Sheffield
Dynamic Global Vegetation Model for estimates of isoprene emissions used in the
FutCROPS and FutCROPS CO2 simulation. Units are in percent of gridbox assigned
to crops.
yr−1 were calculated, as shown in Figure 5.6. In the emission model, the crop
plant functional type (PFT) was assigned an emission factor of 0.09 mg isoprene
m−2 h−1 after Guenther et al. [2006]. For reference, a high emitting PFT such as
broadleaf rainforest had an emission factor of 12.6 mg isoprene m−2 h−1 [Lathie`re
et al., 2009]. The low isoprene emission of the crop PFT could be representative
of a number of biofuel feedstocks, notably corn and sugarcane, which are com-
monly used to produce ethanol (see section 1.3.2.2). Globally isoprene emissions
are reduced by 35% on account of land use change (both with and without CO2
suppression, Figure 5.3). Emissions from current isoprene ‘hotspots’ are dramat-
ically reduced, notably over the Amazon (-57%), and central Africa (-26%), and
summertime emissions from south eastern United States (-27%), and Southeast
122
a  January
b July
Figure 5.6. Change in isoprene emissions [g isoprene m−2 month−1] between BOTH
and FutCROPS simulations.
Asia (-16%). Where the crop fraction increases (Figure 5.5) and isoprene emis-
sions remain constant (Figure 5.6), this is because the previous plant functional
type (PFT) in that region was a low emitter (e.g. the southwest United States).
5.3.0.3 CO2 suppression
The inhibition of isoprene emission at high ambient atmospheric CO2 concentra-
tions was introduced in section 2.2.2.2. Both Possell et al. [2005] and Wilkinson
et al. [2009] have empirically parameterized the isoprene emission-CO2 relation-
ship, and Wilkinson et al. [2009] supplied a plant physiological rationale behind the
reduction in emissions based on competition within the plant for pyruvate. At the
123
time this work started, the study of Wilkinson et al. [2009] was not yet available,
so the relationship of Possell et al. [2005] was employed. Within the range stud-
ied here (345-621 ppmv CO2), however, there is hardly any difference between
the two. When CO2 suppression was included and crops were left at present day
levels, isoprene emissions were reduced by 199 Tg C yr−1, or 41% (the difference
between BOTH and BOTH CO2). When crops were held at the future distribution
instead, CO2 suppression accounted for a 130 Tg C yr−1 decrease, again a reduc-
tion of 41% (the difference between the FutCROPS and FutCROPS CO2 runs.)
CO2 suppression is thus the largest relative factor in controlling global isoprene
emissions within the four matrix experiments (Figure 5.3). The relative reduction
in isoprene emissions resulting from CO2 suppression is similar to the values pre-
viously reported in Heald et al. [2009] of -31% and Young et al. [2009] of -55%.
5.4 Characterizing a potential future atmosphere through
ozone attribution
5.4.1 Introduction
Attribution is commonly used as a method to deconvolve the various causes of
changes in air pollutants or climate change. In the United States, this is closely tied
to modelling of specific air quality targets or ‘policy-relevant’ background ozone
levels (the level of ozone that would be present without anthropogenic emissions
from North America) [Wu et al., 2008a]. Depending on the model and emissions,
the greatest positive contributing factor to future ozone has been attributed to bio-
genic emissions [Hogrefe et al., 2004], anthropogenic emissions and scenario [Tao
et al., 2007], climate [Steiner et al., 2006; Tagaris et al., 2007], and long range
transport [also chemical boundary conditions, see Avise et al., 2009]. In this sec-
tion, my aim is to find out what the largest contributors are to future ozone changes
in UKCA.
124
5.4.2 The impact of climate change
The upper left of Figure 5.7 shows the effect of climate change only on future av-
erage near surface ozone in UKCA. The northeast of the United States, California,
the extended region around the Mediterranean, parts of Eurasia, and east Africa all
exhibit increases in ozone concentrations between 6 and 12 ppbv. Over the conti-
nents, there are colocated increases in OH of up to 70%. A very widespread but
lower (2-6 ppbv) increase in ozone is also modelled throughout much of the model
domain. Small decreases are modelled throughout the tropics over the oceans,
with an area near Greenland and the ITCZ region east of the maritime continent
both showing stronger negative signals. The area near Greenland may be linked
to colocated decreases in temperature of -2 ◦C, a response that was also found in
Solomon et al. [2007b].
The change in ozone resulting from climate change is slightly larger than val-
ues reported by Zeng et al. [2008] (± 10 ppbv) using a similar chemical mecha-
nism, though their results were more heterogeneous spatially. Decreases in ozone
throughout the tropics are evident in the work of Young [2007] and Zeng et al.
[2008]. The results in UKCA are closer to those reported in Hauglustaine et al.
[2005] for 2100, though in UKCA decreases in ozone over the oceans are limited
to the tropics rather than the more widespread phenomenon reported in that study.
It is worth clarifying that both Zeng et al. [2008] and Hauglustaine et al. [2005]
used future emissions in both the runs with and without climate change, so the pho-
tochemical response may not be exactly comparable (the CConly and BASE runs
both use present day emissions). In addition, the general increase in global ozone
modelled by UKCA is in contrast to the general background decrease in ozone
noted by Liao et al. [2006] using the GISS model.
125
d) BOTH-BASE
b) ISOP-CConly
c) BOTH-ISOP
a) CConly-BASE
Figure 5.7. Difference in annual average near surface (600 m and below) ozone
[ppbv] for the four year model period (2096-2099). a) the difference between CConly
and BASE runs, or the impact of climate change; b) the difference between ISOP and
CConly runs, or the impact of future biogenic isoprene emissions; c) the difference
between the BOTH and ISOP runs, or the impact of future anthropogenic emissions;
and d) the difference between BOTH and BASE runs, or the sum of the total impact
of emissions and climate.
126
hy
br
id
 h
ei
gh
t
hy
br
id
 h
ei
gh
t
hy
br
id
 h
ei
gh
t
a  Temperature  [˚C]
b  O3 Production  [kg gridbox month-1]
c  NOx  [%]
Figure 5.8. Difference in annual average zonal mean a) temperature [◦C], b) ozone
production [kg O3 gridbox month−1], and c) NOx [%] between CConly and BASE
runs.
127
Figure 5.8 shows an example of the type of chemistry-climate interactions de-
scribed in section 5.1.1. The change in annual average zonal mean temperature
is shown in the top panel, showing increases of 2-3.5 ◦C throughout much of the
troposphere, and even higher changes in the upper troposphere (4-5 ◦C), where
greenhouse gases are most effective. The distribution of the change is similar to
that reported in Solomon et al. [2007b] and Young [2007]. At the surface, much
of the increases occur over the Artic and over continental regions, the latter due to
higher heat capacities and changes in boundary layer humidity and soil moisture
[Joshi et al., 2008]. One of the key changes that happens in the free troposphere
due to climate change is the increase in ozone production (middle panel of Figure
5.8), which rises by over 1 kg O3 gridbox month−1 around 4000 m. Colocated
ozone concentrations increase by 8-14 ppbv, even higher than changes at the sur-
face. One of the contributing factors to increased ozone production in the region
is the thermal dissociation of PAN, which releases NOx into the chemical system
(Figure 5.8, bottom panel). NOx concentrations in much of the free troposphere
increase by 5-40 %. PAN concentrations throughout the troposphere decrease be-
tween 10% and 50%, and the tropospheric burden drops by 17% on account of
climate change.
The impact of climate change on ozone at a regional level has been well studied
for both the United States and Europe [e.g. Liao et al., 2006; Dawson et al., 2007;
Meleux et al., 2007; Tagaris et al., 2007; Racherla and Adams, 2008]. In Figure
5.9, the impact of climate change on four-year average July ozone is shown. The
‘ozone season’ is defined as a period in which ozone (usually DM8H) can exceed a
threshold value such as 75 ppbv [Chen et al., 2009b], and usually occurs between
April and September for polluted northern hemisphere sites. July is often taken as
a case study month for this period [e.g. Cox and Chu, 1996; Nolte et al., 2008].
In UKCA, increases of 2-16 ppbv are modelled over much of this domain; the
Mediterranean and Sahara show particularly strong increases, as does the western
United States. A decrease of 4-8 ppbv is modelled over the Pacific ocean near
128
Figure 5.9. Difference in four-year average July ozone [ppbv] between CConly and
BASE runs for the United States and Europe (the difference due to climate change).
the Baja peninsula of Mexico. These results compare favorably with the literature.
Lin et al. [2008] and Liao et al. [2006] reported increases in surface ozone over
the United States of +3-12 ppbv, and Hogrefe et al. [2004] found values of up
to 14 ppbv. Hogrefe et al. [2004] also reported small decreases in ozone over the
southeast US. Over Europe, Liao et al. [2006] reported increases of 2-8 ppbv, while
Meleux et al. [2007] reported values of +8-20 ppbv.
5.4.3 The impact of biogenic isoprene emissions
The upper right panel of Figure 5.7 shows the change in ozone concentrations when
future biogenic isoprene emissions are included in calculations. In the southern
Amazon, central Africa, and maritime continent regions, future isoprene emissions
decrease ozone by 2-4 ppbv. This is due to an increase in ozonolysis and the
removal of ozone precursor gases; these regions display similar behaviour when
isoprene was incorporated into UKCA (section 3.5.1.2). Two regions show an
increase in ozone as a result of increased future isoprene emissions, which are the
eastern United States and the northern part of South America. The differences
between regions are due to the variation in chemical regime into which isoprene
is emitted (see section 3.5.1.3) and the season. Globally, the impact is relatively
129
small: the PAN burden increases by only 6%, compared with -17% due to climate
change, and the CO burden increases by 5%.
The decreases in ozone in the Amazon, central Africa, and the maritime con-
tinent are similar to those reported in Zeng et al. [2008] of more than 3 ppbv. In
their study, ozone concentrations increased throughout much of the northern hemi-
sphere, which was not reproduced by UKCA. Two possible explanations for this
divergence are the much higher NOx emissions used in that study (A2), and per-
haps more importantly, the geographically constant scaling factor of 50% applied
to all isoprene emissions. Here, vegetation is allowed to vary with climate and
thus global isoprene emissions only increase by 17% and have more geographical
variability (Figure 5.4).
5.4.4 The impact of anthropogenic emissions
In this scenario, the major polluting regions in North America, Europe, Japan, and
the Antipodes are all projected to reduce NOx emissions dramatically—between
100 and 500 Gg N yr−1. Figure 5.10 shows future NOx emissions compared to
those for the present day. Emissions in developing countries are projected to in-
crease, particularly in Nigeria, Kenya, South Africa, southern Brazil, India, and
China.
The influence of anthropogenic emissions on ozone is evaluated from the dif-
ference between the BOTH and ISOP simulations (lower left panel of Figure 5.7).
The overwhelming impact is the reduction in ozone due to the reduction in precur-
sor emissions in the B2+CLE scenario. The greatest changes are modelled over
the southeastern United States, where ozone is reduced by up to 16 ppbv. Rel-
atively large changes (4-8 pbbv) are also modelled over the Mediterranean and
east Asia. Slight (2-4 ppbv) increases are modelled in the Pacific in the midlati-
tudes. This response was relatively robust to the level of isoprene emissions. Us-
ing present day isoprene emissions (ANTH-CConly, not shown), the distribution
of ozone reductions was very similar, although the absolute value of the change
130
Figure 5.10. Change in annual average surface NO2 emissions [Gg N yr−1] between
future (B2+CLE) and present day emissions.
was approximately 4 ppbv lower throughout the domain. This result demonstrates
that reductions of anthropogenic emissions, as projected by the B2+CLE scenario,
are an effective way to reduce future surface ozone, and that reductions are not de-
pendent on the isoprene emission scenario. In fact, future isoprene emissions only
increase the importance of NOx reductions.
Precursor emissions have been previously shown to be an important factor in
controlling future ozone concentrations. Simulations with low emissions, particu-
larly for NOx, have generally shown decreases in ozone [e.g. Szopa et al., 2006].
In contrast, high levels of emissions generally lead to higher future ozone concen-
trations [e.g. Wu et al., 2008b; Zeng et al., 2008]. The reduction in surface ozone
due to the B2+CLE emissions was also shown in the results of Fowler et al. [2008],
which used a mean of five model simulations. The results were presented for each
season, and the greatest reductions were modelled in spring (MAM) and summer
(JJA). Smaller reductions (2-4 ppbv) were apparent in fall (SON) and winter (DJF).
In winter, however, Fowler et al. [2008] reported slight increases over the United
States and Europe (between 2-8 ppbv). Similar (2-10 ppbv) increases were mod-
131
elled by UKCA. In summary, the results of UCKA match well with the reported
multi-model means in Fowler et al. [2008].
5.4.5 The combined impacts
In the lower right panel of Figure 5.7, the combined effect of anthropogenic emis-
sions, biogenic emissions, and climate change is shown. The total change in ozone
is quasi-linear when considered in comparison to the impact of each of the indi-
vidual variables. Much of the globe retains low to mid level increases (2-6 ppbv)
in ozone due to the physical climate change forcings, and the low levels and re-
ductions of ozone across the tropical oceans are also present. Over western South
America, the impact of biogenic emissions cancels out the increase due to climate
change, so no change is modelled. Even more dramatically, the reductions in ozone
due to anthropogenic emissions across North America, the Mediterranean, and
parts of Eurasia negate the impact of climate change. In the southeastern United
States this even leads to a decrease in ozone in the BOTH simulation compared
with the BASE case (-2 to -6 ppbv).
For simplicity, and to eliminate differences due to seasonal variability of emis-
sions and climate change, changes in ozone presented thus far have been tempo-
rally averaged over the study period. Figure 5.11 shows the seasonal variation for
the total future impact (all emissions and climate) compared with the BASE run.
In January, UKCA models an increase in ozone throughout much of the north-
ern hemisphere. The decrease in ozone east of the maritime continent is strong in
this season. The signal throughout the rest of tropics is heterogenous, and little
change is modelled through the southern hemisphere. It seems that climate change
forcings are the dominant cause of changes in ozone concentrations in this season.
In July, the polluted regions of North America, Europe, southeast Asia, and east
Asia show a decrease in surface ozone between the BASE and BOTH simulations.
Reductions in anthropogenic precursor emissions appear to be the strongest cause
for changes in ozone in the summer. In both seasons, ozone reductions over the
132
a   January
b   July
Figure 5.11. Change in near-surface (up to 600 m) ozone concentrations [ppbv] be-
tween BOTH (2096-2099) and BASE (2001-2005) simulations for a) January, and b)
July.
tropical oceans are evident, presumably due to climate change.
As modelled by UKCA, climate change has the most widespread impact, while
anthropogenic emissions have the largest and most acute impact on future surface
ozone. Biogenic emissions, on the other hand, play a much smaller role. In terms
of policy relevance, the results from UKCA support reductions in anthropogenic
ozone precursor emissions as a way to improve future air quality. In fact, the
133
reductions are concentrated in the summer months (the ozone season), when they
are most warranted.
5.4.6 The future tropospheric ozone budget
Table 5.3 shows the ozone budgets for the attribution experiments. The budget of
the BASE integration has been discussed in section 3.5.1.1. The budget changes
markedly in the CConly simulation, such that production increases by 16%, loss in-
creases by 20%, and deposition rises by 9%. These are very similar to the numbers
reported by Zeng and Pyle [2003] of 11%, 21%, and 10% respectively, in which a
predecessor mechanism (albeit without isoprene chemistry) was employed. Simi-
lar values have also been reported by a number of other studies [Stevenson et al.,
2000; Hauglustaine et al., 2005; Liao et al., 2006, as summarized in Wu et al.
[2008a]].
Table 5.3. Ozone budgets from emissions and climate change attribution experiments.
Production, loss, and are in [Tg O3 yr−1], ozone lifetime is in days, and ozone burden
is given in Tg. All values are for the troposphere below the chemical tropopause (150
ppbv) defined in Stevenson et al. [2006].
Experiment Prod. Loss Dep. O3 τ O3 Burden
BASE 3215 2813 795 31.6 317
CConly 3728 3367 868 29.1 343
ISOP 3717 3350 865 28.8 338
ANTH 3607 3287 842 29.4 337
BOTH 3572 3255 837 29.1 331
In UKCA, the ozone lifetime decreased by 8% while the global burden in-
creased by 8% as a result of climate change. In the survey of Wu et al. [2008a],
ozone lifetime decreased by 10-18% (summarized from four studies), similar to my
results. The direction of change in the ozone burden was not consistent between the
same four model studies, with values of between -12 and +5% reported, indicat-
ing that the complex relationship between ozone and climate can lead to changes
in both directions depending on model parameterizations [Wu et al., 2008a]. The
134
ozone burden remained relatively constant throughout the future model studies. Fu-
ture isoprene emissions negligibly reduced both ozone production (11 Tg, 0.3%)
and loss (17 Tg, 0.5%) compared to the run with only climate change, similar to the
low change in surface ozone. In contrast, future anthropogenic emissions reduced
ozone production and loss (121 Tg, 3.2% and 80 Tg and 2.4%, respectively). The
BOTH run had the lowest levels of ozone production and loss of any of the model
simulations set in the future.
5.5 Future land use change: Phase I of the biofuel life
cycle
It has been shown that cropland expansion, which could be associated with an in-
creased growth of biofuel feedstocks, could reduce future isoprene emissions (sec-
tion 5.3.0.2). It has also been shown in sections 3.5 and 5.4.3 that isoprene emis-
sions can influence tropospheric chemistry. In this section, I examine how reduced
isoprene emissions from land use change could impact atmospheric chemistry in
the future.
5.5.1 Ozone
To the best of my knowledge, the impact of anthropogenic land use change on
atmospheric chemistry at a global scale has not previously been studied.
Figure 5.12 shows that future cropland expansion (Figures 5.5 and 5.6) leads to
an increase in regional ozone of up to 50% (over 10 ppbv in hotspot regions) in the
FutCROPS simulation compared with the BOTH simulation. In January, the largest
changes are modelled over the Amazon, the tropical Atlantic, and the tropical Pa-
cific. Over the Amazon, dramatic reductions in isoprene flux due to cropland ex-
pansion generate elevated ozone concentrations by both reducing isoprene ozonol-
ysis and reducing NOx sequestration which forms organic nitrate species. As a
result of the latter, NOx concentrations rise by 150% over the Amazon. The strong
135
a   January
-14      -10       -8        -6        -4       -2        -1          1         2         4         6         8          10       14
b   July
Figure 5.12. Change in near-surface (averaged up to 600 m) ozone concentrations
[ppbv] between BOTH and FutCROPS (2096-2099) for a) January, and b) July.
link between isoprene chemistry and nitrate species is demonstrated by the signal
over the tropical oceans, where changes in PAN concentrations lead to changes in
NOx concentrations of -30% over the Pacific and +50% over the Atlantic.
In July, with the peak growing and emitting season in the northern hemisphere,
the terrestrial ozone signal shifts northwards. Large ozone changes (over 8 ppbv)
are modelled over North America, the Mediterranean, China, and the Amazon. The
shape of the impact in ozone over the Amazon correlates spatially with changes in
emission of isoprene due to crops. Over North America and China, where isoprene
emissions decrease with anthropogenic land use change, regional NOx concentra-
tions increase by 40% and 90%, respectively, leading to higher ozone concentra-
136
tions. The signal over Northern Africa is linked to elevated ozone concentrations
in the Mediterranean, again via the formation and transport of nitrate species such
as PAN.
Both Wiedinmyer et al. [2006] and Chen et al. [2009a] reported decreases in
ozone over the United States as a result of allowing land use change to reduce bio-
genic isoprene emissions. The present day emissions in Wiedinmyer et al. [2006]
and the A2 emissions in Chen et al. [2009a] both have higher NOx emissions than
the B2+CLE scenario, perhaps explaining the divergence in our results. High NOx
emissions would lead to a VOC-sensitive chemical regime, and thus a reduction
in VOC would lead to a reduction in ozone. Wiedinmyer et al. [2006] reported
increases in ozone over the Amazon (+3-5 ppbv) on account of land use, which are
also modelled by UKCA. Avise et al. [2009] found very small increases in ozone
resulting from land use change. From these studies, it seems that the change in
ozone resulting from land use change is both model and scenario specific for pol-
luted regions such as the United States, but that the response in VOC-rich regions
may be more robust. Keeping in mind that these studies are just sensitivities, it
seems most appropriate to propose that large scale land use change throughout the
high isoprene emitting regions of the tropics will be detrimental to air quality, but
that the effect on the northern hemisphere is more likely to be scenario dependent.
The change in chemical ozone production between the FutCROPS and BOTH
runs is shown in Figure 5.13. In this image it becomes apparent that ozone pro-
duction increases directly over the regions of perturbation to isoprene. Part of this
increase in ozone can be explained by a reduction in NOz which occurs in the same
regions (NOz = all nitrogen containing compounds - NOx), which is also shown
in Figure 5.13. The decrease in isoprene due to cropland expansion decreases the
formation of NOz, and ozone production rises in these areas (most of which are
NOx-sensitive).
137
ab
Figure 5.13. Difference in annual average a) chemical ozone production [kg O3
gridbox month−1] and b) column NOz [%] between the FutCROPS and BOTH simu-
lations.
5.5.2 Crops: exposure to ozone
In these simulations, the inclusion of anthropogenic land use change leads to in-
creases in ozone over the very areas where cropland is projected to expand in the
future. Ozone has been found to be damaging to crops [Hayes et al., 2007; Mills
et al., 2007; van Dingenen et al., 2009], whose productivity has been shown to
decline after prolonged exposure to ozone concentrations above 35 ppbv [Long
et al., 2005], and the geographical area exposed to this concentration increases in
the model when land use change is included. Figure 5.14 shows the change in the
138
number of months that a gridbox is exposed to monthly mean ozone concentrations
above 35 ppbv for an average model year between the FutCrops and BOTH runs.
Zero values indicate gridboxes that were either exposed to mixing ratios of ozone
greater than 35 ppbv for the same number of months, or which never attain ozone
values of that level. Three regions show a marked increase in the duration of ex-
posure: the southern and eastern United States, Southeast Asia, and the Amazon.
These are precisely the three regions in which cropland is predicted to expand and
where exposure to ozone has the greatest potential to reduce the productivity of
crops. It should be noted that the monthly mean output of a global model is not
equal to the maximum ozone or to the DM8H values; still, these results suggest a
positive ozone feedback loop.
This potential feedback loop between crop expansion, ozone elevation, and
crop health and productivity is robust to the inclusion of CO2 suppression, as
shown in the bottom panel of Figure 5.14. Even with the inclusion of CO2 sup-
pression in the model, in which the total global isoprene flux is 41% lower, the
three regions in which land use markedly reduces isoprene emissions (southeast-
ern United States, southeast Asia, and the Amazon) show up to 3 months increase
in exposure to ozone above 35 ppbv. With CO2 suppression included, over the
United States and southeast Asia the maxima of the responses are elevated, though
in the United States the spatial extent is slightly smaller. In the Amazon, though
the change is still evident, the total change in exposure is lower. Africa shows
a more widely distributed response to land use change when CO2 suppression is
accounted for versus when it is not.
139
ab
Figure 5.14. Increase in the number of months a gridbox experiences surface ozone
above 35 ppbv in an average model year as a result of land use change. a) differ-
ence between FutCROPS and BOTH simulations (without CO2 suppression) and (b)
difference between FutCROPS CO2 and BOTH CO2 (with CO2 suppression).
140
5.6 How does land use change compare to CO2
suppression of isoprene emission?
As CO2 rises in a potential future atmosphere, another possible impact on iso-
prene is CO2 suppression of its emission. Arneth et al. [2007a] and Heald et al.
[2009] both found that the suppression was enough to offset CO2 fertilization and
temperature induced increases in emission that have been predicted elsewhere (see
e.g. Lathie`re et al. [2005]; Tao and Jain [2005]; Guenther et al. [2006]). Heald
et al. [2009] found that CO2 suppression reduced isoprene emissions by 217 Tg
yr−1 when vegetation was kept fixed, while emissions dropped by 610 Tg yr−1
when vegetation was dynamic. Using the parameterizations described in Arneth
et al. [2007a], Arneth et al. [2007b] reported reductions in future (2081-2100) iso-
prene emission estimates of 930-990 Tg C yr−1 for the A2 scenario and 450-480
Tg C yr−1 for the B1 scenario for European forests. Young et al. [2009] assessed
the potential that CO2 suppression of isoprene emission could have on the oxidiz-
ing capacity of the troposphere by using the UM CAM model, a predecessor to
UKCA. They found that CO2 suppression of isoprene emission impacted surface
ozone concentrations by up to 10 ppbv over terrestrial emission regions, though it
reduced ozone in Europe, east Asia, and the Indian ocean.
The response of the Sheffield Dynamic Global Vegetation model showed a
slightly larger decrease in isoprene resulting from CO2 suppression than for land
use change (Figure 5.3). In this section I compare the two effects on atmospheric
chemistry.
5.6.1 Land use and CO2 suppression: ozone
In order to compare the impact on the oxidizing capacity of the atmosphere from
CO2 suppression and future land use change, the differences in average near-
surface ozone from the four matrix experiments are shown in Figure 5.15. The
top two panels display the difference in ozone as caused by land use change both
141
with and without CO2 suppression. Without CO2 suppression (top panel), the
change in ozone concentrations is the most marked of any of the experiments, with
particularly strong increases appearing over the Amazon of over 6 ppbv. Smaller
terrestrial differences come through over southeast Asia and central and northern
Africa, though these are on a similar level (less than 2 ppbv) to the change mod-
elled over parts of the Pacific and Atlantic oceans. The change over the maritime
continent is mixed, ranging between -2 and 4 ppbv. A similar profile occurs when
land use change is studied with CO2 suppression included (second panel of Figure
5.15). Again, the Amazon shows the highest levels of change in ozone concen-
trations (up to 6 ppbv). The differences in equatorial Africa and the southeastern
United States are slightly lower than without CO2 suppression. The change in
southeast Asia remains on a similar scale as it was before.
142
bc
d
a
Figure 5.15. Annual average change in near surface (0-600 m average) ozone [ppbv]
over the four year model period 2096-2099 for a) the difference from crops without
CO2 suppression (FutCROPS-BOTH), b) the difference from crops with CO2 sup-
pression (FutCrops CO2-BOTH CO2), c) the difference from CO2 suppression with
present day crops (BOTH CO2-BOTH), and d) the difference from CO2 suppression
with future crops (FutCrops CO2-FutCROPS).
143
In contrast, the impact of CO2 suppression when crops are kept to their present
day geographic area (Figure 5.15, third panel) shows less drastic changes in ozone
with a wider spatial extent than changes from land use. All four hot-spot emitting
regions (southeastern United States, the Amazon, central Africa, and southeast
Asia) display ozone change on the order of 1-4 ppbv as a result ofCO2 suppression.
In many ways, this is to be expected: the changes occurring from land use change
are spatially heterogeneous and will effect different areas with variable levels of
intensity. CO2 suppression, on the other hand, is applied to all emissions with a
simple empirical relationship (described in section 2.2.2.2). Finally, the change in
ozone from CO2 suppression when crops are held at future values is shown in the
bottom panel. Again, the four hot-spot regions appear with increases in average
ozone concentrations. The maritime continent reveals the strongest signal, while
the signal from the Amazon is almost negligible compared to land use change.
Since the future land use scenario predicts notable increases in cropland in the
Amazon, the number of gridboxes which emit isoprene in the region has already
been reduced significantly. Thus the potential reduction in emissions due to CO2
suppression is fractionally much smaller.
For all the studies, the change in annual average ozone is higher than the ozone
attibuted to biogenic isoprene emissions in Figure 5.7. The change in isoprene
resulting from land use change and CO2 suppression is also higher (section 5.3),
explaining some degree of the difference in magnitude. An additional contributing
factor to the acute increase in ozone modelled due to land use change would be de-
position. One weakness of my experiments is that the land surface type in UKCA
remains constant throughout the model study, so deposition is not altered while iso-
prene emissions are. Cropland expansion would likely decrease deposition, as the
UKCA deposition velocity for shrubs is 23% lower than for forests, which would
further exacerbate the acute response of surface ozone. This feedback would not
exist for CO2 suppression, for which the surface characteristics would not change.
144
5.6.2 Land use and CO2 suppression: HOx
The impact of both CO2 suppression and land use change on the oxidizing capacity
of the atmosphere is also significant for the concentration of OH and its role in
the HOx cycle. Figure 5.16 shows the change in the HO2:OH ratio, which has
been shown to be a good indicator for HOx cycling [e.g. Ren et al., 2008]. Here,
as with ozone, the difference in OH between land use change (upper panel) and
CO2 suppression (lower panel) does not change much with the variation in the
other variable. Because the change in the ratio of HO2 to OH is very similar for
both the land use experiments (with and without CO2 suppression) and the CO2
suppression experiments (with present and future crops), only one example for
each variable is shown in Figure 5.16.
The maximum impact of land use change on the ratio of HO2 to OH in mod-
elled over the Amazon, where the change is greater than -90%. This arises largely
because the sink of OH via reaction with isoprene is lower with lower isoprene
emissions. Lower isoprene emissions will also decrease the amount of HO2 in
the system, as the recycling of HO2 is more effective with NO than O3 (see sec-
tion 3.5.2.1). Isoprene is therefore altering both the numerator and denominator
of this ratio through direct (reaction with OH) and indirect (altering the O3 and
NO concentrations in the region) means. Three other regions show decreases in
the percentage HO2 to OH ratio: the southeastern United States, southeast Asia,
and central Africa. None are as large as the impact over the Amazon. In contrast,
CO2 suppression has a much more widespread impact on the HOx family and the
decreases are lower in magnitude. In general, the ratio decreases by less than 60%,
but it does so across all the four key emitting regions. The increase in OH concen-
trations as a result of CO2 suppression has been reported elsewhere [Young et al.,
2009].
145
ab
Figure 5.16. Average monthly change [%] in the HO2:OH ratio over the four year
model period 2096-2099 in the troposphere up to 5000 m for (a) FutCROPS-BOTH,
or the difference from crops without CO2 suppresion, and (b) BOTH CO2-BOTH the
difference from CO2 suppression with present day crops.
5.6.3 Land use and CO2 suppression: budgets
Although the changes in regional ozone concentrations resulting from land use
change and CO2 suppression can be marked, the global impact on the budget
seems to have a rather smaller effect. The production of ozone increases by a max-
imum of 54 Tg yr−1 (1.5%) when both crops and CO2 suppression and included
146
Table 5.4. Ozone budgets as a result of CO2 suppression and land use change. All
values are averaged from four model years (2096-2099) and averaged for the tropo-
sphere using the chemical tropopause defined in Stevenson et al. [2006]. Production,
loss, and deposition are in [Tg O3 yr−1], ozone lifetime is in days, and the burden is
given in Tg.
Experiment Prod. Loss Dep. O3 τ O3 Burden
BOTH 3572 3255 837 29.1 331
FutCROPS 3586 3255 850 30.0 338
FutCROPS CO2 3626 3281 854 30.0 344
BOTH CO2 3587 3261 845 29.7 339
(FutCROPS CO2 minus BOTH). The ozone burden is nearly identical for all four
of the runs that include either land use change or CO2 suppression. For the BOTH
(with the highest isoprene) integration the ozone lifetime is the shortest of the ma-
trix, at 29.1 days. The ozone lifetime and burden are largest for the run with the
lowest amount of isoprene emission (FutCROPS CO2), indicating that the global
impact of isoprene in these experiments is to reduce total tropospheric ozone. This
is supported by both the surface ozone distributions (Figure 5.15) and the exposure
values (Figure 5.14); in both cases, reducing isoprene emission increased ozone.
Young et al. [2009] reported that CO2 suppression accounted for a reduction in
ozone production, while here we show an increase. A likely reasons for this diver-
gence is the difference in anthropogenic emissions used; their study employed the
SRES A2 scenario, a pessimistic scenario with large increases in NOx emission
compared to present day. With over three times (331% more, 124.1 Tg N yr−1)
more NOx in the system, it is likely that ozone production is limited instead by
VOC, such that a reduction in VOC due to CO2 suppression will reduce ozone
production. In this work, much lower NOx emissions are employed, and the pro-
duction of ozone is more likely NOx-sensitive.
147
5.6.4 Land use and CO2 suppression: reaction fluxes
The inverse relationship between isoprene emission and ozone concentrations in
UKCA is driven by both direct and indirect means. The relative significance of
the two can be described by the annual flux through key reactions, as shown in
Table 5.5, which demonstrates that changes in ozonolysis (direct) are nearly equal
to changes in the formation of NO2 (indirect) from VOC oxidation. The average
impact of land use is to reduce the ozonolysis flux by 42% per year, while CO2
suppression leads to an average -53% change in the flux. The impact on NO2
regeneration is lower; land use change changes the flux by -11% per year, while
CO2 suppression accounts for a 13% reduction.
Table 5.5. Annual average fluxes for four experiments involving CO2 suppression
and land change given in [Tg O3 yr−1].
Flux BOTH FutCROPS FutCROPS CO2 BOTH CO2
RO2 + NO 342 305 264 295
Ozonolysis 93 53 25 43
The correlation of the total flux through these reactions with the change in
isoprene emission also gives some indication of the strength of their chemical re-
lationship. Using all four experiments, the slope of the correlation between total
ozonolysis flux and global isoprene emission is 0.23 (with an R2 value of .997),
while the relationship is 0.26 (R2 of .989) for the recycling of NO. These slopes
are remarkably similar. From the basis of these two reactions only, it would be diffi-
cult to predict a positive or negative change in ozone. However, besides ozonolysis,
another loss process for NOx (which reduces ozone production) is the formation
of reservoir species such as PAN or isoprene nitrates. When this is considered
in conjunction with the the fluxes above, it becomes evident that the globally aver-
aged effect of increasing isoprene emissions in UKCA is to increase ozone loss, for
which chemical recycling of NO2 does not compensate. This result is in agreement
with the budgets, and with the surface ozone concentrations.
148
5.6.5 Uncertainties
There is a range of uncertainty associated with this type of study. The extent of
future land use change depends on the model and scenario [Alo and Wang, 2008].
Auxiliary changes would inevitably accompany a large scale change in land use, in-
cluding biogenic [Yienger and Levy II, 1995] and fertilizer [Bouwman et al., 2002]
nitrogen emissions [Vitousek et al., 1997b] and surface climate and radiation im-
pacts [Georgescu et al., 2009; Pongratz et al., 2009]. The scientific understand-
ing of the impact of land use change on climate is not robust, as past land cover
change has not been shown to have a common or singular effect in models, but
rather that the response depends on the parameterizations and assumptions in the a
model physics [Pitman et al., 2009]. Future climate change will also have impact
on the terrestrial carbon cycle, though the extent and direction of the flux is also
still uncertain [Arora and Matthews, 2009]. In addition, there are multiple sce-
narios which describe different levels of population growth, economic decisions
[Nakicenovic et al., 2000; Fowler et al., 2008], and climate forcing, which impacts
the anthropogenic influence on the landscape [Alo and Wang, 2008; Lapola et al.,
2009].
5.7 Summary
In this chapter I have described a series of eight experiments designed to both
attribute ozone in the future as well as test the sensitivity of the model to CO2
suppression and land use change as they impact isoprene emissions. In the first set
of experiments, I used a serial methodology to attribute certain changes in future
ozone, as calculated by UKCA, to climate change, anthropogenic (B2+CLE), or
future biogenic emissions. The future run which included all variables was quasi-
linear in response to the sum of the variables versus the integrations in which each
perturbation was tested individually. I found that the largest and most widespread
positive changes in future surface ozone arose from physical climate change forc-
149
ings, and that the largest negative changes were linked to reductions in anthro-
pogenic emissions. Biogenic isoprene emissions did not have that much of an
impact.
In the second set of experiments, I focused on the impact of land use change and
CO2 suppression on isoprene emissions and the resulting change in tropospheric
chemistry. Land use change, which could be associated with large scale growth of
biofuel feedstocks, increased ozone production over key hot-spot emitting regions.
I found that there could be a feedback loop between land use change, reduced
isoprene emissions, and increased ozone; in these areas exposure to high levels
of ozone could reduce crop productivity, a potentially significant impact on the
agricultural cycle and the production of biofuel feedstocks. In the comparison
between land use change and CO2 suppression, I found that land use change had
a more acute effect on ozone and HOx, and that the impact from CO2 suppression
was more widespread. Despite these regional differences, the change in global
budget numbers was relatively small between the experiments.
In the next chapter, I examine the final phase of the biofuel life cycle.
150
Chapter 6
Phase III: Combustion
In this chapter, I turn to the final phase of the biofuel life cycle: combustion. I
present a set of four experiments which focus on tailpipe emissions from biofuels
and their potential impact on tropospheric chemistry. In section 6.2 I describe
the experimental methodology, and then I examine the BASE scenario using a
chemical sensitivity indicator in section 6.3. I look at the change in ozone resulting
from a shift to biofuels in section 6.4, and break the response down into both a
linear and nonlinear component.
6.1 Introduction
Emissions from transportation comprise a large fraction of total anthropogenic
emissions worldwide. Because transportation emissions are emitted from mobile
sources, the spatial distribution is wider than that of other sectors, such as power
generation or industry. It is also the largest growing sector for end use of oil, grow-
ing by 1146.47 Mtoe [Million tonnes of oil equivalent] (106%) in OECD countries
and 584.4 Mtoe (82%) globally between 1973 and 2006 [International Energy
Agency, 2008]. This growth is expected to continue at a significant rate [Inter-
national Energy Agency, 2007]. Because of its wide spatial distribution, complex
emissions, and projected growth, transportation is particularly important to study
151
in order to understand future changes in the troposphere.
6.1.1 Impact on the atmosphere
The transport sector has been shown to impact the radiative balance of Earth’s at-
mosphere. Fuglestvedt et al. [2008] found that the transport sector has contributed
15% of CO2 radiative forcing and 31% of tropospheric ozone forcing since prein-
dustrial times. For the present day, Niemeier et al. [2006] reported that 30 mW m−2
in January and 50 mW m−2 in July of radiative forcing could be attributed to ozone
caused by emissions from road transport. In the future, the radiative forcing impact
due transport could change due to changing emissions. Shindell et al. [2008] found
a 30% reduction in road transport emissions resulted in a change of -21 mW m−2
in RF resulting from CO2 and -5 mW m−2 due to O3. Unger et al. [2009] reported
that a 50% reduction of road transport emissions resulted in a -19 mW m−2 shift
in radiative forcing, but when the reduction was applied in the United States only,
it resulted in a change of only -5 mW m−2.
In addition to impacting the climate system, road transport has a profound im-
pact on air quality and tropospheric chemistry. Niemeier et al. [2006] found that
road transport accounted for up to 10-50 ppbv ozone in polluted regions in July. For
the same month, Matthes et al. [2007] reported that road traffic accounted for over
15% of surface ozone, particularly over North America and Europe. Castellanos
et al. [2009] found that instead of diurnal variation, temporally uniform emissions
from mobile sources led to a reduction in the number of days with DM8H ozone
exceeding 80 ppbv.
Changes in the emissions from transportation can also lead to changes in air
quality and atmospheric chemistry. Hoor et al. [2009] found that for a 5% re-
duction in all transport emissions, ozone increased in an average of five models.
Individual sector reductions showed that reducing emissions from road transport
(instead of from shipping or aviation) had the largest acute effect on ozone. Ship-
ping had a more widespread impact, as the global average change in ozone re-
152
sulting from a 5% decrease in ship emissions was 0.06 ppbv, while for road it
was 0.0435 ppbv, and for aircraft was 0.015 ppbv. When Niemeier et al. [2006]
scaled global road transport emissions to European levels (representing a growth
in transport in the developing world), ozone increased by 5-40 ppbv, with particu-
larly strong changes in southern and eastern Asia. When the scale was adopted for
North American road transport emissions (representing even higher growth) ozone
increased by 10-100 ppbv, again with the strongest response in southern and east-
ern Asia. Finally, Bell and Ellis [2004] found that when road transport emissions
were increased, the average increase in summertime hourly max ozone went up by
9 ppbv in urban locations and 6 ppbv in rural locations in the eastern United States.
6.1.2 Biofuel emissions from the tailpipe
The impact of the transportation sector on climate and air quality would be mod-
ified by the use of non-conventional fuels, which have a different emissions pro-
file than their conventional analogues. One potential alternative liquid fuel for
transport, biodiesel, emits higher NOx and lower CO and hydrocarbons than con-
ventional diesel. Elevated NOx emissions have been linked to both the level of
unsaturation of biodiesel fuel molecules [McCormick et al., 2001; Fernando et al.,
2006; Knothe et al., 2006; Ban-Weiss et al., 2007] and to the advance timing of
fuel injection [Monyem and H. Van Gerpen, 2001; Szybist et al., 2005, 2007]. Un-
saturated molecules have a slightly higher flame temperature, encouraging NOx
formation in the piston [Ban-Weiss et al., 2007]. The change in injection timing in
a compression-ignition engine leads to a longer time period in which the tempera-
tures are high enough to form thermal NOx (section 1.4.3) before the reactions are
quenched by expansion of the piston [Szybist et al., 2007]. The advance in timing
has been attributed to lower compressibility of biodiesel in comparison to diesel
[Lapuerta et al., 2008]. Biodiesel is also an oxygenated fuel, and therefore com-
bustion in a compression-ignition engine will be more complete, reducing both the
unburnt hydrocarbon and CO emissions [Agarwal, 2007]. This level of comple-
153
tion in the combustion chamber is a contributing factor to the higher temperatures
which are conducive to NOx formation [Agarwal, 2007].
For ethanol, the comparison to conventional gasoline emissions profile is dif-
ferent. VOC emissions are much higher for ethanol, a fuel with a higher vapour
pressure than conventional fuel [Koc et al., 2009]. Evaporative emissions therefore
rise [Huo et al., 2009]. Cracking of ethanol leads to the production of acetaldehyde
[Gaffney and Marley, 2009], the emissions of which have been shown to be much
higher for vehicles run with higher percentages of ethanol [Graham et al., 2008].
CO emissions are slightly elevated for ethanol, and are controlled by the flammabil-
ity and oxygenation of the fuel [Koc et al., 2009]. NOx emissions depend on peak
temperatures achieved in the spark-ignition engine during combustion. Because
the latent heat of vaporization is higher for ethanol than conventional gasoline,
the flame temperature is lowered, which in turn leads to reduced NOx emissions
[Koc et al., 2009]. In general, ethanol is mixed with a small fraction of gasoline
(commonly 15%) to allow for fuel ignition in the engine on cold days [Sustainable
biofuels: prospects and challenges, 2008].
These changes in tailpipe emissions resulting from the combustion of biofu-
els (either for ethanol or biodiesel) have the potential to alter emissions from the
transport sector. As described above, surface transport has already been shown to
influence both tropospheric chemistry and climate. It follows that any change in
emissions resulting from a shift in fuel mix could alter air quality. In this chapter,
my aim is to understand how surface transport emissions from biofuels could affect
tropospheric chemistry.
6.2 Experimental setup
I ran four sensitivity experiments—BASE, E85, BD20, and BD100—in order to
test the sensitivity of the model chemistry to surface transport emissions. Each
was run between 1999 and 2005 with 16 months removed for spin-up. Physical
154
Figure 6.1. High resolution (1x1◦) NO2 emissions from surface transport for the
BASE case, in 106 kg NO2 year−1.
forcing factors (radiation and SSTs) were the same as the BASE run in Chapters 3
(section 3.2) and 5 (section 5.2). Present day emissions [Dentener et al., 2005] are
also the same for the BASE case and are described in section 3.2.3. The emissions
were put together using the Regional Air Pollution Information and Simulation
(RAINS) model [Dentener et al., 2005]. For transportation emissions, NOx emis-
sions factors for Europe, North America, Russia, and Southeast Asia were specific
to national characteristics such as fleet composition, fuel quality, and maintenance
levels. In other places, default emission factors were applied. Emissions from
surface transport do not have seasonal or monthly variation.
High resolution NO2 emissions from the surface transport sector are depicted
in Figure 6.1. These emissions are degraded to the resolution of UKCA (3.75◦ by
2.5◦) when used in the model, but are shown here at higher resolution to illustrate
the heterogeneity and complexity of surface transport emissions. NO2 emissions
155
Figure 6.2. High resolution (1x1◦) CO emissions from surface transport for the BASE
case, in 106 kg CO year−1.
from road transport cover most of the exposed (i.e. without seasonal or permanent
ice) land surface of the earth, with the except of the Sahara, parts of southern
Africa, central Eurasia, and central Australia. Key shipping routes are also visible,
especially the Atlantic corridor between Europe and North America. There are
a few areas of particularly high density in the NO2 emissions field: the eastern
United States, Europe, India, Thailand, and China. In contrast, emissions of CO
(shown in Figure 6.2), which are elevated in the same key regions, are also elevated
in South Africa, Nigeria, central America, and a larger extent of eastern Europe.
This could be related to the increased average age of vehicles in these countries,
which has been shown to be linked to emissions [Pandian et al., 2009]. Surface
transport emissions comprise 36% of the total NO2 emitted in the BASE run, and
18% of CO.
These surface transport emissions fields were altered for the sensitivity studies
156
Table 6.1. Emission scaling factors used to modify existing traffic emission fields.
Scaling factors for biodiesel were reported in United States Environmental Protection
Agency [2002], and scaling factors for E85 are from Jacobson [2007].
Fuel Type NOx CO VOCs
E85 .70 1.05 See Table 6.2
BD20 1.024 .869 .821
BD100 1.134 .573 .368
Table 6.2. Emission scaling factors for E85 VOCs as reported in Jacobson [2007].
Species Scaling Factor
Ethane 1
Propane .35
Formaldehyde 1.6
Acetaldehyde 20.0
Acetone 1
by scaling factors appropriate to three fuel types: 20% biodiesel mixed with 80%
conventional diesel (BD20), 100% biodiesel (BD100), and 85% ethanol mixed
with 15% conventional gasoline (E85). The scaling factors for each study are listed
in Table 6.1. The scaling factors for biodiesel (both BD20 and BD100) were re-
ported in United States Environmental Protection Agency [2002]. The authors used
39 emission datasets for heavy duty diesel vehicle (HDDV) engines tested with
soybean derived biodiesel. The data was analyzed using statistical curve fitting in
order to obtain a relationship between the percentage of fuel comprised of biodiesel
and the measured emissions of criteria air pollutants. According to United States
Environmental Protection Agency [2002], the impact of biodiesel on HDDV emis-
sions was to increase NOx by 13.4% for 100% biodiesel, and to decrease both CO
and VOCs by 42.7% and 63.2%, respectively. BD20 scaled almost linearly with the
100% biodiesel results, in other words, the change in emissions for 20% biodiesel
is approximately equal to 20% of the change for 100% biodiesel.
The scaling factors for E85 were reported in Jacobson [2007] and are the aver-
157
age of 10 studies which examined the relative impacts of E85 versus conventional
gasoline or petrol. NOx emissions are reduced by 30% for E85 compared to con-
ventional fuel, and CO rises by 5%. Importantly, the work of Jacobson [2007]
included a speciation of emission scaling factors for individual organic species. To
capture the most variability possible (and the best information in the literature),
these individual scaling factors were used in the E85 run. The values for these
speciated VOC scaling factors are listed in Table 6.2. The largest change in VOC
emissions for E85 is the 2000% jump in acetaldehyde emissions, which is a conser-
vative estimate from the observed data in the 10 studies used by Jacobson [2007].
Propane emissions were reduced by 65%, and formaldehyde emissions jumped by
60%. Ethane and acetone emissions were left unchanged.
6.3 The BASE case
One of the major difficulties in modelling tropospheric chemistry is understanding
and identifying when a chemical regime is NOx-sensitive or VOC-sensitive with
respect to ozone production. Sillman [1995] addressed this issue by examining pos-
sible indicator species for VOC- versus NOx-sensitivity; defined as which species
(NOx or VOC) will promote an increase in ozone production. In a NOx-sensitive
environment, where many peroxy radicals are present, a significant radical termi-
nation process is the formation of hydroperoxides,
HO2 +HO2 → H2O2. (2.21)
In contrast, in a VOC-sensitive environment, where large amounts of NO2 are
present, a stronger loss process is the formation of nitric acid,
OH+NO2 +M→ HNO3 +M (2.23)
158
both of these radical-radical reactions have been described in section 2.1.3. The
dominant loss process of HOx radicals (which initiate the majority of tropospheric
oxidation cycles, see section 2.1.1) therefore can be indicative of the type of chem-
ical regime.
Sillman [1995] established that the ratio of concentrations of H2O2/HNO3
was indeed a strong indicator of the sensitivity of a chemical regime with re-
spect to ozone production. Using measurements from eight sites, the authors found
that this flux ratio was able to distinguish between the NOx-sensitivity and VOC-
sensitivity across the widest range of emissions and meteorological conditions.
Sillman [1999] reviewed VOC-NOx-ozone chemistry, revisiting the H2O2/HNO3
ratio as an appropriate and useful indicator for the sensitivity of a chemical regime
to VOC versus NOx. More recently, Sillman and West [2009] tested this sensitiv-
ity parameter against measurements in Mexico City and found that the formation
flux ratio of H2O2/HNO3 was one of the only useful parameters in that polluted
environment.
Both Sillman [1995] and Sillman and West [2009] used instantaneous tracer
concentrations in their studies. Equivalent concentrations are not available as out-
put of a global model (I have monthly mean data instead); however, the flux of
formation of the two molecules can be a diagnostic output. In theory, the flux of
formation is even more closely linked to the sensitivity of the chemistry because it
eliminates differences in chemical (e.g. via reaction with OH) and physical (e.g.
via wet deposition) sinks between the two species. This approach was used in a
model by Steiner et al. [2006] to identify chemical regimes in California. The
authors successfully differentiated between a NOx-rich region (the Bay area) and
a NOx-sensitive regime (the southern Sierra Nevada mountains) using afternoon
production fluxes for hydrogen peroxide and nitric acid. Zhang et al. [2009] also
successfully used the production fluxes to verify the sensitivity of chemical regimes
in the United States to VOC or NOx.
In Figure 6.3, the seasonality of the indicator flux ratio H2O2/HNO3 is shown
159
a  JFM b  AMJ
c  JAS d  OND
VO
C 
se
ns
iti
vi
ty
   
   
   
   
   
   
   
   
   
 N
O
x s
en
si
tiv
ity
Figure 6.3. Seasonal variation in the ratio of the flux of formation of H2O2 over
HNO3 for averaged a) January, February, and March; b) April, May, and June; c)
July, August, and September; and d) October, November, and December.
for the BASE integration. Sillman [1995] found that the VOC sensitive regime was
generally indicated by values less than 0.3, the NOx sensitive regime was indicated
by values above 0.5, and that values in between indicated a transition regime. For
their work in Mexico city, Sillman and West [2009] reported that VOC sensitive
air masses were marked by ratio values below 0.1, while NOx sensitive air was
marked by ratio values above 0.2. The H2O2/HNO3 indicator flux ratio is therefore
not an absolute measure of the chemical sensitivity, but rather an indicative one
which needs to be calibrated for a particular location. Generally, a low value of
the flux ratio will be suggestive of a VOC-sensitive regime (there is little peroxide
production in comparison with nitric acid), while high values of the ratio suggest
NOx-sensitivity (much more peroxide is produced compared with nitric acid.)
In the winter months for the northern hemisphere, a strong VOC-sensitivity
covers much of North America and Europe. Most of South America, Australia,
160
and Africa are NOx-sensitive, with the exception of a strip across the Sahel where
biomass burning NOx emissions are present in the months between November
and March. The northern parts of the Pacific and Atlantic are also NOx-sensitive
during this time period. In the northern hemisphere summer months, parts of North
America switch to NOx-sensitive regimes, concurrent and colocated with seasonal
biogenic emissions of isoprene, a result that is consistent with Zhang et al. [2009].
A ridge through South America and a swath of southern Africa display areas of
VOC-sensitivity which were not present during the winter months, both are areas
where biomass burning emissions are elevated during this period. The northern
section of Australia also transitions to VOC-sensitivity due to seasonal isoprene
emissions; a change from the stronglyNOx-sensitivity of the boreal winter (Austral
summer). Southeast Asia also transitions strongly between the two halves of the
year, albeit with opposite seasonality to Australia.
6.4 What is the impact of biofuels?
6.4.1 Ozone
Figure 6.4 shows the annual average change in surface ozone due to a shift to 100%
biodiesel (BD100) or 85% ethanol (E85). The strongest impact of BD100 is to
reduce ozone throughout the northern hemisphere. The response is on the order of
1-4 ppbv, with greater reductions in southern California (-8 ppbv) and in Japan and
Korea (-6 ppbv). Slight increases up to 2 ppbv are modelled over parts of central
Eurasia and a few scattered areas throughout the tropics. Biodiesel increases NOx
emissions and decreases VOC emissions, and as shown in Figure 6.3, the northern
hemisphere is strongly VOC-sensitive throughout most of the year.
In contrast, the effect of E85 fuel (Figure 6.4, bottom panel) is to increase
ozone strongly in UKCA. Increases are modelled over most of the northern hemi-
sphere on the order of 1-10 ppbv. Increases over Europe reach up to 15 ppbv, while
increases over Asia and North America are even higher: over Asia, values reach 20
161
a  BD100-BASE
b  E85-BASE
Figure 6.4. Average change in monthly mean surface ozone [ppbv] between a)
BD100 and b) E85 and BASE. Note the different scales.
ppbv, while over the northeastern United States values reach 25 ppbv. The largest
increases are modelled over southern California and climb to 35 ppbv. Similar
to the BD100 sensitivity study, the opposite signal is modelled throughout central
Eurasia (up to -1 ppbv) and scattered areas in the tropics (up to -2 ppbv.) Most of
the affected domain throughout the northern hemisphere was VOC-sensitive (Fig-
ure 6.3), and thus increasing VOC emissions and reducing NOx is a powerful way
162
-40 
-20 
0 
20 
40 
60 
80 
BD
10
0
BD
20
BA
SE E8
5
N
O
x P
er
tu
rb
at
io
n 
an
d
 O
zo
ne
 R
es
po
ns
e 
[%
]
Experiment
Figure 6.5. Maximum change in relative surface ozone [%] between BD100, BD20,
and E85 and BASE (blue), and the change in NOx emissions for the same runs (red).
to increase ozone production. Jacobson [2007] performed a very similar exper-
iment in which the scaling factors for E85 were applied to the present day fleet
in the United States and California (they used two model domains). They found
that ozone increased in the Los Angeles area by up to 15 ppbv in August. They
also found that in the polluted environment of Los Angeles, the change in ozone
was directly related to the relative reduction in NOx. The NOx-VOC sensitivity
indicator results above indicate that this region is VOC sensitive, consistent with
their results. An increase in acetaldehyde emissions in a 2012 future scenario with
E10 fuel (10% ethanol mixed with 90% conventional fuel) also showed high ozone
increases in Bangkok in the work of Milt et al. [2009].
In order to investigate the relative relationships between all the sensitivity stud-
ies, Figure 6.5 shows the linearity of the maximum relative response in ozone and
the change in NOx for each set of model experiments. The largest relative change
163
in ozone is modelled for the E85 sensitivity, with a value of 70%; the change in
NOx is 30% for this study. The maximum response for BD100 is -10 ppbv (with an
increase in NOx of 13.4%), while the smaller perturbation from the BD20 experi-
ment changes ozone by a maximum of only 3 ppbv (for a +2.4% change in NOx).
Matthes et al. [2007] examined the relative contribution of NOx and VOC emis-
sions from traffic, found that NOx was the largest contributor, a result which is in
agreement with the high correlation of between ozone and NOx emissions shown
in Figure 6.5. These numbers show both linear and nonlinear components. For
ethanol, as an example, the perturbation of 30% in NOx emissions leads to a rela-
tive change in ozone of 70%. In contrast, BD100 leads to a 20% change in ozone
through a 13.4% change in NOx emissions. The linearity of the response seems to
be dependent on the direction of the perturbation and the chemical regime.
6.4.2 NOx-VOC sensitivity
With the diagnostic flux indicator, it is possible to understand the nonlinear com-
ponent of the change in ozone described above. The flux ratio of H2O2 to HNO3
formation changes with the emissions depending on fuel type used in the model,
as shown in Figure 6.6. When 100% biodiesel is used, a fuel that emits more NOx
and less VOC, the sensitivity flux parameter is reduced, indicating a transition to
a more VOC-sensitive regime (in which more nitric acid and less hydroperoxide
is produced). The largest changes are modelled over the UK and northern France,
southern California, and the eastern seaboard of North America. These are all ar-
eas of very high emissions from surface transport (Figures 6.1 and 6.2). The BD20
simulation also showed decreases in the flux parameter, although to a lower extent
(not shown.)
In contrast, the choice to move to E85 fuel moves the chemical system towards
NOx-sensitivity. E85 emits more VOC (+22%) and CO (+5%) (see Table 6.1) than
conventional fuel while emitting much less NOx (-30%). The largest response is
found above polluted continental regions; every region with high surface transport
164
a  BD100-BASE
b  E85-BASE
VO
C 
se
ns
iti
vi
ty
   
   
   
   
  N
O
x s
en
si
tiv
ity
Figure 6.6. Percentage change in ratio of formation of H2O2 and HNO3 between
BD100 and E85 runs relative to the BASE case.
emissions shows a move towards NOx sensitivity. The northern oceans show a
less extreme shift in the opposite direction—towards VOC sensitivity. For both
sensitivity studies, the largest changes in the indicator flux parameter occur in the
northern hemisphere. This is colocated with the areas of highest surface transport
emissions (Figures 6.1 and 6.2), and with highest VOC-sensitivity (Figure 6.3).
165
6.4.3 Seasonality
The nonlinear component of the ozone response, because it is dependent on local
chemical conditions, is likely to be seasonally dependent. Despite the fact that in
these experiments surface transport emissions do not vary with time, other emis-
sions in the model do change. There is also some suggestion from the literature that
the response does vary with season; Niemeier et al. [2006] performed an experi-
ment in which road traffic emissions were shut off entirely, and found that ozone
decreased in Europe in January by over 25 ppbv while increasing over much of
the continental tropics. In July, however, the signal was more uniform, and ozone
increased by up to 25 ppbv over most of the industrialized areas of northern hemi-
sphere.
In Figure 6.7, the change in ozone for average January and July are shown for
the E85 simulation compared to the BASE simulation. Because the E85 perturba-
tion involves an increase in VOC emissions and a decrease in NOx, the strongest
response occurs in January. The eastern United States, where some of the the high-
est increases occur, has been shown to be VOC-sensitive during this period [Jacob
et al., 1995; Zhang et al., 2009, and Figure 6.3]. If we take this to be representative
of the northern hemisphere, then this would explain the large ozone changes that
occur in winter that are not present in summer. In July, the response is markedly
smaller in both magnitude and geographic extent, coinciding with the period when
biogenic VOC emissions are higher in the northern hemisphere and ozone produc-
tion is NOx-sensitive in some regions. The southeastern United States, a partic-
ularly strong emission hot-spot, even shows a decrease (over -5 ppbv) in ozone
during this time period due to extreme VOC saturation of the chemical system.
Smaller decreases (between 2 and 5 ppbv) are scattered across the northern hemi-
sphere during this season. Similar decreases are modelled throughout the south-
ern hemisphere during January. These small decreases (between 2 and 5 ppbv of
ozone) follow the seasonality of isoprene emissions.
166
a  January
b  July
Figure 6.7. Change in near surface (up to 600 m) average ozone [ppbv] between the
E85 and BASE runs for a) January and b) July.
167
6.5 Uncertainties
A sensitivity study of this nature could not possibly capture the complexity of the
vehicle fleet in a single country, let alone worldwide. It is intended to probe the
possible limits or boundaries in the potential atmospheric perturbation that a switch
to biofuels would entail. Notably, the whole world does not operate one type of
fuel, for which there could possibly be one type of fuel replacement. Instead,
the fuel market is split largely down country lines [International Energy Agency,
2007]. For example, the United States light duty vehicle market veers heavily to-
wards gasoline/petrol (76% in 2006) [Energy Information Administration, 2007].
In contrast, the Germany light duty market operates 54% (in 2005) on diesel [Min-
erallwirtschaftsverband, 2006]. A switch to biofuels in the United States would
therefore likely veer towards ethanol (a gasoline replacement), while a similar shift
in Germany would likely head to biodiesel (a diesel replacement). A single change
applied across the globe is therefore not an accurate portrayal of what is most likely
to occur in the future, as changes are more likely to be regional.
In addition, the scaling factors used for E85 were based on mostly light duty
vehicle (LDV) testing, while the scaling factors for biodiesel were based on emis-
sions testing from heavy duty vehicles (HDVs). Perhaps the response in emissions
testing would change for testing of different vehicles. Finally, the exact changes
in emissions are scaled up from small laboratory experiments, and may not be
representative of the entire fleet. For example, the NOx emissions from biodiesel
reported in United States Environmental Protection Agency [2002] are still under
discussion. The exact increase in NOx emissions (13%) was replicated by Monyem
and H. Van Gerpen [2001], but smaller changes were observed in Szybist et al.
[2005] (10%) and Karavalakis et al. [2009] (11%).
168
6.6 Summary
In this chapter, I used a set of four sensitivity studies to estimate the upper bounds
of the impact that combustion of biofuels could have on air quality. I character-
ized the BASE simulation through the use of flux ratio indicators, which showed
the strongest seasonality in the northern hemisphere, where VOC-sensitive condi-
tions dominate in the winter, but where the sensitivity is more heterogeneous in the
summer. I then described the impact of phase III of the life cycle (combustion) on
tropospheric chemistry. I found that biodiesel decreased surface ozone by up to 10
ppbv, and that ethanol had a much more stark effect as ozone increased by 35 ppbv;
a strongly nonlinear response. I also found that seasonality was important, espe-
cially for ethanol, and that the positive response in surface ozone was dominant in
winter.
Although there are uncertainties associated with this type of large scale study,
my results point out that our choice of future fuel will have different and profound
effects on air quality. Increases in ozone of such a scale due to ethanol could
be disastrous for human and crop health, and for positive radiative forcing of the
climate system. Although these results cannot be taken as a projection, they do
call for more research and perhaps raise a red flag for the future ethanol industry.
In contrast, the move to biodiesel seems to be a safe choice for air quality; in the
best case scenario, high NOx emissions from biodiesel decrease ozone, while in
the worst case, ozone will remain the same.
In the next chapter, I summarize the novel work performed in my thesis and
describe some of the summed impacts of phase I and phase III of the biofuel life
cycle.
169
Chapter 7
Conclusions
In this final chapter, I review the novel work that was presented throughout this
thesis (section 7.1). I then discuss the aggregated impact of the two phases of the
biofuel life cycle on air quality in section 7.2, followed by some ideas for further
studies that could build directly from my work (section 7.3). In section 7.4, I
conclude with a brief longer term outlook on the future of this type of life cycle-
atmosphere research.
7.1 Summary of novel work
1. Implemented isoprene chemistry in UKCA
I included a representation of isoprene oxidation chemistry in the tropospheric
configuration of UKCA. My analysis showed that isoprene had a particularly strong
positive impact on ozone in the northern hemisphere during summer, and that gen-
erally over hot-spot emission regions with low NOx emissions, ozone was reduced.
I was able to distinguish between both VOC-rich and VOC-sensitive regimes and
NOx-rich and NOx-sensitive regimes through examining scatter relationships. I
also found that isoprene increased the methane lifetime by 22.5% in UKCA, and
that OH was substantially reduced over hot-spot emission regions. I characterized
the impact of isoprene oxidation on key tropospheric trace gases, setting up that
170
changes in isoprene emissions resulting from the growth of energy crops could
have strong influences on air quality in regions where land use could change in the
future.
2. Used a box model to compare the chemical mechanism with
measurements from a field campaign
Using a box model, I examined the ability of UKCA’s chemical mechanism
against NO, NO2, and O3 data from the OP3 field campaign in Malaysian Borneo.
I found that the mechanism was relatively insensitive to a series of chemical, pho-
tolytic, and deposition parameters, but that representations of physical processes
were more important. In particular, the implementation of a nighttime ‘venting pa-
rameter’ was necessary in order to capture the diurnal cycle of NO2. Using a cost
function analysis of three physical variables, I produced a best fit simulation which
matched the measurements closely and demonstrated that the UKCA mechanism
could capture the chemistry of NOx and ozone during the campaign. My work
demonstrated that a global model chemical mechanism can capture the chemistry
at an individual field site, but that parameterizations of physical processes are im-
portant in order to do so successfully.
3. Performed two global sensitivity studies for Southeast Asia
Considering the region of Southeast Asia in a global model, I performed two
sensitivity studies, OILPALM and SUGARCANE, which were simple represen-
tations of how changes in isoprene could result from cultivation of biofuel feed-
stocks. When isoprene emissions were increased in the OILPALM study, surface
ozone concentrations dropped, while when isoprene emissions were decreased (for
the SUGARCANE study) ozone strongly increased. This simple study substan-
tiated the hypothesis that phase I of the biofuel life cycle could indeed impact
tropospheric chemistry through changes in isoprene emissions. The magnitude of
the response depended on the direction of the isoprene perturbation, and therefore
on the type of feedstock crop.
171
4. Characterized a potential future atmosphere through ozone
attribution
Using the concept of attribution and a serial methodology, I found that future
surface ozone increased due to climate change and decreased due to changes in
anthropogenic emissions. Biogenic isoprene emissions had a smaller effect, and
the total surface ozone in the ‘best guess’ future run was quasi-linear with respect to
each individual variable. My results suggest that reducing anthropogenic precursor
emissions is an effective way to reduce future ozone. I also found that these ozone
reductions were largest in the summer months, coinciding with the ‘ozone season’.
5. Examined changes due to land use change and CO2 suppression
of isoprene emission at the global scale
I performed the first global study which examined isoprene emissions, land use
change, and the resulting change in tropospheric chemistry. I found that ozone in-
creased as a result of land use change, and that this increase occured directly over
the regions of perturbation and was robust to the inclusion of CO2 suppression.
This could be significant for crop health and productivity, and in the worst case
scenario, may encourage more cropland expansion (as existing crops produce less
than expected). I also compared the impact of land use change with another pertur-
bation to isoprene emission, CO2 suppression, and found that the spatial variation
was different but that both increased ozone. Overall, land use had a more acute ef-
fect than CO2 suppression on both ozone and HOx, but the global budget numbers
were relatively unchanged.
6. Explored the sensitivity to biofuel emissions from surface
transport
I performed the first study of changes in tropospheric chemistry resulting from
biofuel use at the global level. I found that ethanol increased ozone markedly (+35
ppbv) and biodiesel decreased ozone, but to a smaller extent (-10 ppbv). I used
chemical reaction fluxes to identify NOx- and VOC-sensitive regimes. I reported
that most extreme ozone increases due to ethanol occured in the northern hemi-
172
spheric winter, when the regime is strongest in its VOC-sensitivity. My results
indicate that biofuels could have a strong impact on tropospheric chemistry for
phase III (combustion), and that ethanol seems to be a worse choice for air quality
than biodiesel.
7.2 The impact of the full life cycle
Although I examined them separately, the aggregate impacts of phase I and phase
III of the biofuel life cycle on air quality and tropospheric chemistry are worth
considering. The response will not necessarily be linear, but from both the ozone
attribution study (section 5.4) and the surface transport experiments (section 6.4), it
seems probable that there will be a large linear component in the ozone response. I
found that the changes in ozone resulting from the oil palm biodiesel fuel-feedstock
system were consistently negative across both phases of the life cycle. In the sim-
ple phase I sensitivity study (OILPALM, chapter 3) the increase in isoprene led
to reduced ozone (-10 ppbv) in the VOC-rich region of Southeast Asia. In the
phase III studies (BD20 and BD100, chapter 6), the change in ozone was again
negative (-10 ppbv). Phase I would reduce ozone in tropical regions (where oil
palm can be cultivated), while phase III would be more influential in the northern
midlatitudes, where more surface transport emissions are present. These are big
changes; it seems plausible that these reductions in ozone should be considered in
the discourse surrounding biodiesel policy.
In contrast, the sugarcane ethanol fuel-feedstock system led to increases in
ozone. This occured in the simple sensitivity (SUGARCANE, chapter 3 and in the
more realistic representation of cropland expansion (FutCrops and FutCROPS CO2,
chapter 5). When I considered the final phase of the life cycle, ozone again in-
creased; in some regions ozone was elevated by over 30 ppbv. These positive in-
creases in ozone concentrations resulting from two phases of the ethanol life cycle
are clearly an important consideration in the decision to produce and use ethanol
173
as a liquid fuel for transport.
While many other factors, including the carbon balance, need to be considered,
my work shows that the impact on air quality can also be very significant.
7.3 Potential further studies
A number of further studies could be carried on directly from this work. In chapter
5, I examined phase I of the life cycle by looking at cropland expansion. The crops
in my experiments were low emitting crops (representing sugarcane or corn, for
example). The experiment in chapter 4 in which I increased isoprene emissions,
however, also displayed changes in ozone resulting from emissions perturbation;
from this it follows that the cropland expansion work could be made more robust
by also studying a high isoprene emitting crop. It would be interesting to see if the
surface ozone response from the simple sensitivity remain as consistent for a high
emitting crop (negative) as they did for a low emitting crop (positive).
In chapter 6, I performed three sensitivity studies in which surface transport
emissions were scaled to biofuel tailpipe emissions. A more accurate representa-
tion of the change in transport emissions from biofuels would entail an additional
scaling factor which accounts for the portion of a vehicle fleet which uses a partic-
ular type of fuel. A set of regional sensitivity studies could also be run, in which
ethanol is used in one part of the domain and biodiesel another. These types of
studies might mitigate the response of surface ozone to a particular biofuel mix,
and could give a more realistic quantification of potential future changes in ozone
and oxidizing capacity.
Perhaps most importantly, the sensitivity of air quality and atmospheric chem-
istry to phase I and phase III of the biofuel life cycle should be tested together
as well as separately. An experiment in which both cropland expansion (chapter
5) and changes in surface transport emissions (chapter 6) are included would be
a good first step. As I described above, my hypothesis is that a large part of the
174
summed response would be linear, and thus the positive ozone responses to the sug-
arcane ethanol system would compound each other, while the negative responses
to the oil palm biodiesel system would also add up. Some portion of the change
would likely be nonlinear, however, and would be interesting to investigate.
7.4 Outlook
Although there are multiple experiments which follow directly from my simula-
tions, some more general additions to the methodology could also be made. Most
importantly, I have neglected the second phase of the biofuel lifecycle, transforma-
tion, in examining the impact of biofuel production on tropospheric chemistry. The
transformation stage, particularly the processing to form crude biofuel on feedstock
cultivation sites, could change the impact of changing VOC emissions in these ar-
eas; for example, if NOx emissions from on site crude processing plants increase,
perhaps the ozone response to increased isoprene would shift. I can imagine a se-
ries of experiments using the box model and oil palm isoprene flux data that could
probe these questions; a global study in which point source NOx emissions are
added to feedstock growing areas would also be interesting. Phase II of the life cy-
cle should be examined on its own to begin with, and then could be added to both
phase I and phase III in order to test the total response to a particular feedstock-fuel
system.
The ultimate goal of this type of biofuel research is to include the entire life
cycle in the emissions of a regional or global model. In this thesis, I have examined
two phases of the biofuel life cycle and their influence on tropospheric chemistry.
The changes in trace gas constituents and the oxidizing capacity of the atmosphere
are significant, warranting more research in this area. The more we understand
about the impact of the entire life cycle on air quality and atmospheric chemistry,
the more informed we will be in making decisions about the implementation of
biofuels at a large scale.
175
Bibliography
Abbe, C. (1901), The physical basis of long-range weather forecasts, Mon. Wea.
Rev., 29, 551–561.
Adams, J. M., J. V. H. Constable, A. B. Guenther, and P. Zimmerman (2001), An
estimate of natural volatile organic compound emissions from vegetation since
the last glacial maximum, Chemosphere, 3, 73–91.
Agarwal, A. K. (2007), Biofuels (alcohols and biodiesel) applications as fuels for
internal combustion engines, Prog. Energ. Combust., 33, 233–271.
Aghedo, A. M., M. G. Schultz, and S. Rast (2007), The influence of African air
pollution on regional and global tropospheric ozone, Atmos. Chem. Phys., 7,
1197–1212.
Alcamo, K. E. . L. R., J. (Ed.) (1999), Global Change Scenarios of the 21st Cen-
tury: Results from the IMAGE 2.1 Model, Oxford: Pergamon/Elsevier Science.
Alo, C. A., and G. Wang (2008), Potential future changes of the terrestrial ecosys-
tem based on climate projections by eight general circulation models, J. Geo-
phys. Res., 113, G01,004, doi:10.1029/2007JG000528.
AltIn, R., S. C¸etinkaya, and H. S. Yu¨cesu (2001), The potential of using vegetable
oil fuels as fuel for diesel engines, Energ. Convers. Manage., 42, 529–538.
Anderson, D., et al. (2004), World Energy Assessment: Overview, New York:
176
United Nations Development Programme, United Nations Department of Eco-
nomic and Social Affairs, and World Energy Council.
Andreae, M. O., and P. Merlet (2001), Emission of trace gases and aerosols from
biomass burning, Global Biogeochem. Cycles, 15, 955–966.
Aplincourt, P., and J. M. Anglada (2003), Theoretical studies of the isoprene
ozonolysis under tropospheric conditions. 2. Unimolecular and water assisted
decomposition of α-hydroperoxides, J. Phys. Chem. A, 107, 5812–5820.
Archibald, A., M. Jenkin, and D. Shallcross (), An isoprene mechanism inter-
comparison, Atmos. Environ., In Press, Accepted Manuscript, –, doi:10.1016/
j.atmosenv.2009.09.016.
Arneth, A., et al. (2007a), Process-based estimates of terrestrial ecosystem isoprene
emissions: incorporating the effects of a direct CO2-isoprene interaction, Atmos.
Chem. Phys., 7, 31–53.
Arneth, A., P. M. Miller, M. Scholze, T. Hickler, G. Schurgers, B. Smith, and I. C.
Prentice (2007b), CO2 inhibition of global terrestrial isoprene production: Po-
tential implications for atmospheric chemistry, Geophys. Res. Lett., 34, L18,813,
doi:10.1029/2007GL030615.
Arneth, A., R. K. Monson, G. Schurgers, U. Niinemets, and P. I. Palmer (2008a),
Why are estimates of global terrestrial isoprene emissions so similar (and why
is this not so for monoterpenes)?, Atmos. Chem. Phys., 8, 4605–4620.
Arneth, A., G. Schurgers, T. Hickler, and P. A. Miller (2008b), Effects of species
composition, land surface cover, CO2 concentration and climate on isoprene
emissions from European forests, Plant Biology, 10, 150–162.
Arnold, S. R., M. P. Chipperfield, and M. A. Blitz (2005), A three-dimensional
model study of the effect of new temperature-dependent quantum yields for ace-
tone photolysis, J. Geophys. Res., 110, D22305, doi:10.1029/2005JD005998.
177
Arora, V. K., and H. D. Matthews (2009), Characterizing uncertainty in model-
ing primary terrestrial ecosystem processes, Global Biogeochem. Cycles, 23,
GB2016, doi:10.1029/2008GB003398.
Atkinson, R. (2000), Atmospheric chemistry of VOCs and NOx, Atmos. Environ.,
34, 2063–2101.
Atkinson, R., and J. Arey (2003), Gas-phase tropospheric chemistry of biogenic
volatile organic compounds: A review, Atmos. Environ., 37, S197–S219.
Atkinson, R. A., S. M. Aschmann, J. Arey, and B. Shorees (1992), Formation
of OH radicals in the gas-phase reactions of O3 with a series of terpenes, J.
Geophys. Res., 97, 6065–6073.
Avise, J., J. Chen, B. Lamb, C. Wiedinmyer, A. Guenther, E. Salath, and C. Mass
(2009), Attribution of projected changes in summertime US ozone and PM2.5
concentrations to global changes, Atmos. Chem. Phys., 9, 1111–1124.
Bakwin, P. S., S. C. Wofsy, S. Fan, M. Keller, S. E. Trumbore, and J. M. D. Costa
(1990), Emission of nitric oxide (NO) from tropical forest soils and exchange
of NO between the forest canopy and atmospheric boundary layers, J. Geophys.
Res., 95, 16,755–16,764.
Ban-Weiss, G. A., J. Chen, B. A. Buchholz, and R. W. Dibble (2007), A numerical
investigation into the anomalous slight nox increase when burning biodiesel; a
new (old) theory, Fuel Process.Technol., 88, 659–667.
Barket Jr., D. J., et al. (2004), A study of the NOx dependence of isoprene oxida-
tion, J. Geophys. Res., 109, D11,310, doi:10.1029/2003JD003965.
Battjes, J. (1994), Global options for biofuels from plantations according to IM-
AGE simulations, The Netherlands: Interfacultaire Vakgoep Energie en Mi-
lieukunde, Rijksuniversiteit Groningen.
178
Beerling, D. J., F. I. Woodward, M. Lomas, and A. J. Jenkins (1997), Testing the
responses of a dynamic global vegetation model to environmental change: A
comparison of observations and predictions, Global Eco. Biogeogr., 6, 439–450.
Bell, M., and M. Ellis (2004), Sensitivity analysis of tropospheric ozone to modi-
fied biogenic emissions for the Mid-Atlantic region, Atmos. Environ., 38, 1897–
1889.
Bellouin, N., O. Boucher, J. Haywood, and M. Shekar Reddy (2005), Global es-
timate of aerosol direct radiative forcing from satellite measurements, Nature,
438, 1138–1141.
Berndes, G., M. Hoogwijk, and R. van den Broek (2003), The contribution of
biomass in the future global energy supply: a review of 17 studies, Biomass &
Bioenergy, 25, 1–28.
Bey, I., et al. (2001), Global modeling of tropospheric chemistry with assimilated
meteorology: Model description and evaluation, J. Geophys. Res., 106, 23,073–
23,096.
Biesenthal, T. A., J. W. Bottenheim, P. B. Shepson, S. Li, and P. C. Brickell (1998),
The chemistry of biogenic hydrocarbons at a rural site in eastern Canada, J.
Geophys. Res., 103, 25,487–25,498.
Bjerknes, V. (1904), Das problem der wettervorhersage, betrachtet vom stand-
punkte der mechanik und der physik., Meteor. Z., 21, 1–7.
Blake, R. S., P. S. Monks, and A. M. Ellis (2009), Proton-transfer reaction mass
spectrometry, Chem. Rev., 109, 861–896.
Bousquet, P., et al. (2006), Contribution of anthropogenic and natural sources to
atmospheric methane variability, Nature, 443, 439–443.
Bouwman, A. F., L. J. M. Boumans, and N. H. Batjes (2002), Emissions of N2O
179
and NO from fertilized fields: Summary of available measurement data, Global
Biogeochem. Cycles, 16, 1058.
Brasseur, G. P., D. A. Hauglustaine, S. Walters, P. J. Rasch, J. F. Mu¨ller, C. Grainer,
and X. X. Tie (1998), MOZART, a global chemical transport model for ozone
and related chemical tracers: 1. Model description, J. Geophys. Res., 103,
28,265–28,291.
Brasseur, G. P., M. Schultz, C. Granier, M. Saunois, T. Diehl, M. Botzet, E. Roeck-
ner, and S. Walters (2006), Impact of climate change on the future chemical
composition of the global troposphere, J. Climate, 19, 3932–3951.
Brewer, A. W. (1949), Evidence for a world circulation provided by the measure-
ments of helium and water vapour distribution in the stratosphere, Q. J. Roy.
Meteor. Soc., 75, 351–363.
Brough, N., et al. (2003), Intercomparison of aircraft instruments on board the C-
130 and Falcon 20 over southern Germany during export 2000, Atmos. Chem.
Phys., 3, 2127–2138.
Butkovskaya, N., A. Kukui, and G. Le Bras (2007), HNO3 forming channel of the
HO2 + NO reaction as a function of pressure and temperature in the ranges of
72–600 Torr and 223–323 K, J. Phys. Chem. A, 111, 9047–9053.
Butkovskaya, N. I., A. Kukui, N. Pouvesle, and G. Le Bras (2005), Formation of
nitric acid in the gas-phase HO2 + NO reaction: Effects of temperature and water
vapor, J. Phys. Chem. A, 109, 6509–6520.
Butler, T. M., D. Taraborrelli, C. Bru¨hl, H. Fischer, H. Harder, M. Martinez,
J. Williams, M. G. Lawrence, and J. Lelieveld (2008), Improved simulation
of isoprene oxidation chemistry with the ECHAM5/MESSy chemistry-climate
model: lessons from the GABRIEL airborne field campaign, Atmos. Chem.
Phys., 8, 4529–4546.
180
California Congress (2005-2006 Session), Global Warming Solutions Act, Resolu-
tion AB32.
California Environmental Protection Agency: Air Resources Board (Ac-
cessed 15 September 2009), Low carbon fuel standard program, Online:
http://www.arb.ca.gov/fuels/lcfs/lcfs.htm#wg.
Campbell, J. E., D. B. Lobell, and C. B. Field (2009), Greater transportation energy
and GHG offsets from bioelectricity than ethanol, Science, 324, 1055–1057.
Carbajo, P. G., S. C. Smith, A.-L. Holloway, C. A. Smith, F. D. Pope, D. E. Shall-
cross, and A. J. Orr-Ewing (2008), Ultraviolet photolysis of HCHO: Absolute
HCO quantum yields by direct detection of the HCO radical photoproduct, J.
Phys. Chem. A, 112, 12,437–12,448.
Cariolle, D., M. J. Evans, M. P. Chipperfield, N. Butkovskaya, A. Kukui, and
G. Le Bras (2008), Impact of the new hno3-forming channel of the ho2+no re-
action on tropospheric hno3, nox, hox and ozone, Atmos. Chem. Phys., 8, 4061–
4068.
Carlier, P., H. Hannachi, and G. Mouvier (1986), The chemistry of carbonyl com-
pounds in the atmosphere: A review, Atmos. Environ., 20, 2079–2099.
Carlton, A. G., C. Wiedinmyer, and J. H. Kroll (2009), A review of Secondary
Organic Aerosol (SOA) formation from isoprene, Atmos. Chem. Phys., 9, 4987–
5005.
Carslaw, N., et al. (2001), OH and HO2 radical chemistry in a forested region of
north-western Greece, Atmos. Environ., 35, 4725–4737.
Carter, W. P. L., and R. Atkinson (1996), Development and evaluation of a detailed
mechanism for the atmospheric reactions of isoprene and NOx, Int. J. Chem.
Kinet., 28, 497–530.
181
Carver, G. D., P. D. Brown, and O. Wild (1997), The ASAD atmospheric chemistry
integration package and chemical reaction database, Comput. Phys. Commun.,
105, 197–215.
Castellanos, P., J. W. Stehr, R. R. Dickerson, and S. H. Ehrman (2009), The sen-
sitivity of modeled ozone to the temporal distribution of point, area, and mobile
source emissions in the eastern united states, Atmos. Environ., 43, 4603–4611.
Chameides, W. L., et al. (1992), Ozone precursor relationships in the ambient at-
mosphere, J. Geophys. Res., 97, 6037–6055.
Chen, J., J. Avise, A. Guenther, C. Wiedinmyer, E. Salathe, R. B. Jackson, and
B. Lamb (2009a), Future land use and land cover influences on regional biogenic
emissions and air quality in the united states, Atmos. Environ., 43, 5771–5780.
Chen, J., et al. (2009b), The effects of global changes upon regional ozone pollution
in the united states, Atmos. Chem. Phys., 9, 1125–1141.
Chen, X., D. Hulbert, and P. B. Shepson (1998), Measurement of the organic nitrate
yield from OH reaction with isoprene, J. Geophys. Res., 103, 25,563–25,568.
Civerolo, K. L., G. Sistla, S. T. Rao, and D. J. Nowak (2000), The effects of land
use in meteorological modeling: Implications for assessment of future air quality
scenarios, Atmos. Environ., 34, 1615–1621.
Cofala, J., M. Amann, Z. Klimont, K. Kupiainen, and L. Ho¨glund-Isaksson (2007),
Scenarios of global anthropogenic emissions of air pollutants and methane until
2030, Atmos. Environ., 41, 8486–8499.
Communication from the commission to the European Parliamant (23 January
2008), 20 20 by 2020, Europe’s Climate Change Opportunity, Brussels.
Cook, P. A., et al. (2007), Forest fire plumes over the North Atlantic: p-
TOMCAT model simulations with aircraft and satellite measurements from
182
the ITOP/ICARTT campaign, J. Geophys. Res., 112, D10S43, doi:0.1029/
2006JD007563.
Cox, W. M., and S.-H. Chu (1996), Assessment of interannual ozone variation in
urban areas from a climatological perspective, Atmos. Environ., 30, 2615–2625.
Cramer, W., A. Bondeau, S. Schaphoff, W. Lucht, B. Smith, and S. Sitch (2004),
Tropical forests and the global carbon cycle: impacts of atmospheric carbon
dioxide, climate change and rate of deforestation, Phil. Trans. R. Soc. B, 359,
331–343.
Crutzen, P. J. (1970), The influence of nitrogen oxides on the atmospheric ozone
content, Q. J. Roy. Meteor. Soc., 96, 320–325.
Crutzen, P. J. (1973), A discussion of the chemistry of some minor constituents in
the stratosphere and troposphere, Pure Appl. Geophys., 106, 1385–1399.
Crutzen, P. J. (1979), The role of NO and NO2 in the chemistry of the troposphere
and stratosphere, Annu. Rev. Earth Pl. Sci., 7, 443–472.
Crutzen, P. J., and M. O. Andreae (1990), Biomass burning in the tropics: Impact
on atmospheric chemistry and biogeochemical cycles, Science, 250, 1669–1678.
Crutzen, P. J., A. R. Mosier, K. A. Smith, and W. Winiwarter (2008), N2O release
from agro-biofuel production negates global warming reduction by replacing
fossil fuels, Atmos. Chem. Phys., 8, 389–395.
Danielsen, F., et al. (2008), Biofuel plantations on forested lands: double jeopardy
for biodiversity and climates, Conservation Biology, 23, 348 – 358.
Davies, T., M. J. P. Cullen, A. J. Malcolm, M. H. Mawson, A. Staniforth, A. A.
White, and N. Wood (2005), A new dynamical core for the met office’s global
and regional modelling of the atmosphere, Q. J. Roy. Meteor. Soc., 131, 1759–
1782.
183
Davin, E. L., N. de Noblet-Ducondre´, and P. Friedlingstein (2007), Impact of land
cover change on surface climate: Relevance of the radiative forcing concept,
Geophys. Res. Lett., 34, L13,702, doi:10.1029/2007GL029678.
Dawson, J. P., P. J. Adams, and S. N. Pandis (2007), Sensitivity of ozone to sum-
mertime climate in the eastern USA: A modeling case study, Atmos. Environ.,
41, 1494–1511.
de Gouw, J., C. Warneke, T. Karl, G. Eerdekens, C. van der Veen, and R. Fall
(2003), Sensitivity and specificity of atmospheric trace gas detection by proton-
transfer-reaction mass spectrometry, Int. J. Mass Spectrom., 223-224, 365–382.
Delon, C., C. E. Reeves, D. J. Stewart, D. Serc¸a, R. Dupont, C. Mari, J.-P.
Chaboureau, and P. Tulet (2008), Biogenic nitrogen oxide emissions from soils
impact on NOx and ozone over West Africa during AMMA (African Monsoon
Multidisciplinary Experiment): modelling study, Atmos. Chem. Phys., 8, 2351–
2363.
Demirbas, A. (2007a), Progress and recent trends in biofuels, Prog. Energ. Com-
bust., 33, 1–18.
Demirbas, A. (2007b), Importance of biodiesel as transportation fuel, Energy Pol-
icy, 35, 4661–4670.
Dentener, F., D. Stevenson, J. Cofala, R. Mechler, M. Amann, P. Bergamaschi,
F. Raes, and R. Derwent (2005), The impact of air pollutant and methane emis-
sion controls on tropospheric ozone and radiative forcing: CTM calculations for
the period 1990–2030, Atmos. Chem. Phys., 5, 1731–1755.
Di Carlo, P., et al. (2004), Missing OH reactivity in a forest: Evidence for unknown
reactive biogenic VOCs, Science, 304, 722–724.
Dobson, G. M. B., A. W. Brewer, and B. M. Cwilong (1946), Bakerian lecture.
meteorology of the lower stratosphere, Phil. Trans. R. Soc. A, 185, 144–175.
184
Duyzer, J. H., J. R. Dorsey, M. W. Gallagher, K. Pilegaard, and S. Walton (2004),
Oxidized nitrogen and ozone interaction with forests. ii: Multi-layer process-
oriented modelling results and a sensitivity study for douglas fir, Q. J. Roy. Me-
teor. Soc., 130, 1957–1971.
Edwards, J. M., and A. Slingo (1996), Studies with a flexible new radiation code.
I: Choosing a configuration for a large-scale model, Q. J. Roy. Meteor. Soc., 122,
689–719.
Emmerson, K. M., and M. J. Evans (2009), Comparison of tropospheric chemistry
schemes for use within global models, Atmos. Chem. Phys., 9, 1831–1845.
Energy Information Administration (2007), Alternatives to traditional transporta-
tion fuels 2007, Tech. rep., Department of Energy, United States Government.
Englert, N. (2004), Fine particles and human health–a review of epidemiological
studies, Toxicology Letters, 149, 235–242.
European Commission (23 April 2009), Directive 2009/28/EC of the European
Parliament and of the Council on the promotion of the use of energy from renew-
able sources and amending and subsequently repealing Directives 2001/77/EC
and2003/30/EC, Official Journal of the European Union, p. L140/16.
European Commission (23April 2009), Directive 2009/29/EC of the European Par-
liament and of the Council of amending Directive 2003/87/EC so as to improve
and extend the greenhouse gas emission allowance trading scheme of the Com-
munity, Official Journal of the European Union, p. L140/63.
European Commission (8 May 2003), Directive 2003/30/EC of the European Par-
liament and of the Council on the promotion of the use of biofuels or other re-
newable fuels for transport, Official Journal of the European Union, p. L 123/42.
Eyers, C., et al. (2004), AERO2K global aviation emissions inventories for 2002
and 2025, Tech. Rep. QinetiQ/04/01113, QinetiQ, Farnborough.
185
Fall, R., and S. D. Copley (2000), Bacterial sources and sinks of isoprene, a reactive
atmospheric hydrocarbon, Environ. Microbiol., 2, 123–130.
Fall, R., and M. C. Wildermuth (1998), Isoprene synthase: From biochemical
mechanism to emission algorithm, J. Geophys. Res., 103, 25,599–25,609.
Feddema, J. J., K. W. Oleson, G. B. Bonan, L. O. Mearns, L. E. Buja, G. A. Meehl,
and W. M. Washington (2005), The importance of land-cover change in future
climates, Science, 310, 1674–1678.
Fehsenfeld, F., et al. (1992), Emissions of volatile organic compounds from veg-
etation and their implications for atmospheric chemistry, Global Biogeochem.
Cycles, 6, 389–430.
Fernando, S., C. Hall, and S. Jha (2006), NOx reduction from biodiesel fuels,
Energy & Fuels, 20, 376–382.
Fiore, A. M., L. W. Horowitz, D. W. Purves, H. Levy II, M. J. Evans, Y. Wang,
Q. Li, and R. M. Yantosca (2005), Evaluating the contribution of changes in
isoprene emissions to surface ozone trends over the eastern United States, J.
Geophys. Res., 110, D12,303, doi:10.1029/2004JD005485.
Fischer, G., and L. Schrattenholzer (2001), Global bioenergy potentials through
2050, Biomass Bioenerg., 20, 151–159.
Folberth, G. A., D. A. Hauglustaine, J. Lathie`re, and F. Brocheton (2006), Interac-
tive chemistry in the Laboratoire de Me´te´orologie Dynamique general circula-
tion model: Model description and impact analysis of biogenic hydrocarbons on
tropospheric chemistry, Atmos. Chem. Phys., 6, 2273–2319.
Foley, J. A., et al. (2005), Global consequences of land use, Science, 309, 570–574,
doi:10.1126/science.1111772.
186
Forster, P. M. D., C. E. Johnson, K. S. Law, J. A. Pyle, and K. P. Shine (1996), Fur-
ther estimates of radiative forcing due to tropospheric ozone changes, Geophys.
Res. Lett., 23, 3221–3324.
Fowler, D., et al. (2008), Ground-level ozone in the 21st century: future trends,
impacts and policy implications, London: The Royal Society.
Franzblau, E., and C. J. Popp (1989), Nitrogen oxides produced from lightning, J.
Geophys. Res., 94, 11,089–11,104.
Friedlingstein, P., et al. (2006), Climate-carbon cycle feedback analysis: Results
from the C4MIP model intercomparison, J. Climate, 19, 3337–3353.
Fuentes, J. D., et al. (2000), Biogenic hydrocarbons in the boundary layer: A re-
view, Bull. Amer. Meteor. Soc., 81, 1537–1575.
Fuglestvedt, J., T. Berntsen, G. Myhre, and K. Rypdal (2008), Climate forcing
from the transport sectors, Proc. Natl. Acad. Sci. USA, 105, 454–458.
Fuglestvedt, J. S., T. K. Berntsen, I. S. A. Isaksen, H. Mao, X.-L. Liang, and W.-
C. Wang (1999), Climatic forcing of nitrogen oxides through changes in tro-
pospheric ozone and methane; global 3D model studies, Atmos. Environ., 33,
961–977.
Fujino, J., K. Yamaji, and H. Yamamoto (1999), Biomass-balance table for evalu-
ating bioenergy resources, Appl. Energ., 63, 75–89.
Gaffney, J. S., and N. A. Marley (2009), The impacts of combustion emissions on
air quality and climate–from coal to biofuels and beyond, Atmos. Environ., 43,
23–36.
Ganzeveld, L., and J. Lelieveld (1995), Dry deposition parametisation in a chem-
istry general circulation model and its influence on the distribution of reactive
trace gases, J. Geophys. Res., 100, 20,999–21,012.
187
Ganzeveld, L., and J. Lelieveld (2004), Impact of Amazonian deforestation
on atmospheric chemistry, Geophys. Res. Lett., 31, L06105, doi:10.1029/
2003GL01925.
Gauss, M., et al. (2003), Radiative forcing in the 21st century due to ozone changes
in the troposphere and the lower stratosphere, J. Geophys. Res., 108, D94,292,
doi:10.1029/2002JD002624.
Georgescu, M., D. B. Lobell, and C. B. Field (2009), Potential impact of u.s. bio-
fuels on regional climate, Geophys. Res. Lett., 36, L21,806.
Gettelman, A., J. R. Holton, and K. H. Rosenlof (1997), Mass fluxes of O3, CH4,
N2O and CF2Cl2 in the lower stratosphere calculated from observational data, J.
Geophys. Res., 102, 19,149–19,159.
Giannakopoulos, C., M. P. Chipperfield, K. S. Law, and J. A. Pyle (1999), Vali-
dation and intercomparison of wet and dry deposition schemes using Pb-210 in
a global three-dimensional off-line chemical transport model, J. Geophys. Res.,
104, 23,761–23,784.
Gitz, V., and P. Ciais (2004), Future expansion of agriculture and pasture acts to
amplify atmospheric CO2 levels in response to fossil-fuel and land-use change
emissions, Climatic Change, 67, 161–184.
Global Emissions Inventory Activity (Accessed 21 October, 2009), Online:
http://www.geiacenter.org.
Goldemberg, J. (2007), Ethanol for a sustainable energy future, Science, 315, 808–
810.
Graham, L. A., S. L. Belisle, and C.-L. Baas (2008), Emissions from light duty
gasoline vehicles operating on low blend ethanol gasoline and E85, Atmos. Env-
iron., 42, 4498–4516.
188
Granier, C., G. Pe´tron, J.-F. Mu¨ller, and G. Brasseur (2000), The impact of natural
and anthropogenic hydrocarbons on the tropospheric budget of carbon monox-
ide, Atmos. Environ., 34, 5255–5270.
Gregory, D., and P. R. Rowntree (1990), A mass flux convection scheme with rep-
resentation of cloud ensemble characteristics and stability-dependent closure,
Mon. Wea. Rev., 118, 1483–1506.
Grosjean, D., E. L. Williams II, and E. Grosjean (1993), Atmospheric chemistry
of isoprene and its carbonyl products, Environ. Sci. Technol., 27, 830–840, doi:
10.1021/es00042a004.
Guenther, A., et al. (1995), A global model of natural volatile organic compound
emissions, J. Geophys. Res., 100, 8873–8892.
Guenther, A., T. Karl, P. Harley, C. Wiedinmyer, P. I. Palmer, and C. Geron (2006),
Estimates of global terrestrial isoprene emissions using MEGAN (Model of
Emissions of Gases and Aerosols from Nature), Atmos. Chem. Phys., 6, 3181–
3210.
Haagen-Smit, A. J. (1952), Chemistry and physiology of Los Angeles smog, Ind.
Eng. Chem., 44, 1342–1346.
Hall, D., R.-C. F., R. Williams, and J. Woods (1993), Renewable Energy: Sources
for Fuels and Electricity, chap. Biomass for energy: supply prospects, pp. 593–
691, Washington, DC: Island Press.
Hamilton, J. F., et al. (2008), Observations of an atmospheric chemical equator
and its implications for the tropical warm pool region, J. Geophys. Res., 113,
D20,313, doi:10.1029/2008JD009940.
Hanson, R. K., and S. Saliman (1984), Combustion Chemistry, chap. Survey of
Rate Constants in the N/H/O System, pp. 361–421, New York: Springer-Verlag.
189
Harwood, R. S., and J. A. Pyle (1975), 2-dimensional mean circulation model for
the atmosphere below 80 km, Q. J. Roy. Meteor. Soc., 101, 723–747.
Hauglustaine, D. A., F. Hourdin, L. Jourdain, M.-A. Filiberti, S. Walters, J.-F.
Lamarque, and E. A. Holland (2004), Interactive chemistry in the Laboratoire
de Me´te´orologie Dynamique general circulation model: Description and back-
ground tropospheric chemistry evaluation, J. Geophys. Res., 109, D04,314, doi:
10.1029/2003JD003957.
Hauglustaine, D. A., J. Lathie`re, S. Szopa, and G. A. Folberth (2005), Future tro-
pospheric ozone simulated with a climate-chemistry-biosphere model, Geophys.
Res. Lett., 32, L24,807, doi:10.1029/2005GL024031.
Hayes, F., M. Jones, G. Mills, and M. Ashmore (2007), Meta-analysis of the rela-
tive sensitivity of semi-natural vegetation species to ozone, Environmental Pol-
lution, 146, 754–762.
Heald, C. L., et al. (2008), Predicted change in global secondary organic aerosol
concentrations in response to future climate, emissions, and land use change, J.
Geophys. Res., 113, D05,211, doi:10.1029/2007JD009092.
Heald, C. L., M. J. Wilkinson, R. K. Monson, C. A. Alo, G. L. Wang, and A. Guen-
ther (2009), Response of isoprene emission to ambient co2 changes and impli-
cations for global budgets, Global Change Biol., 15, 1127–1140.
Hegglin, M. I., and T. G. Shepherd (2009), Large climate-induced changes in ul-
traviolet index and stratosphere-to-troposphere ozone flux, Nature Geosci, 2,
687–691.
Hewitt, C. N., et al. (2009a), Oxidant and particle photochemical processes above
a south-east asian tropical rain forest (the OP3 project): introduction, rationale,
location characteristics and tools, Atmos. Chem. Phys. Disc., 9, 18,899–18,963.
190
Hewitt, C. N., et al. (2009b), Nitrogen management is essential to prevent tropi-
cal oil palm plantations from causing ground-level ozone pollution, Proc. Natl.
Acad. Sci. USA, 106, 18,447–18,451, doi:10.1073/pnas.0907541106.
Hill, J., E. Nelson, D. Tilman, S. Polasky, and D. Tiffany (2006), From the Cover:
Environmental, economic, and energetic costs and benefits of biodiesel and
ethanol biofuels, Proc. Natl. Acad. Sci. USA, 103, 11,206–11,210.
Hill, J., S. Polasky, E. Nelson, D. Tilman, H. Huo, L. Ludwig, J. Neumann, H. C.
Zheng, and D. Bonta (2009), Climate change and health costs of air emissions
from biofuels and gasoline, Proc. Natl. Acad. Sci. USA, 106, 2077–2082.
Hofzumahaus, A., et al. (2009), Amplified trace gas removal in the troposphere,
Science, 324, 1702–1704.
Hogrefe, C., et al. (2004), Simulating changes in regional air pollution over the
eastern united states due to changes in global and regional climate and emissions,
J. Geophys. Res., 109, D22,301, doi:10.1029/2004JD004690.
Holton, J. R., P. H. Haynes, M. E. McIntyre, A. R. Douglass, R. B. Rood, and
L. Pfister (1995), Stratosphere-troposphere exchange, Rev. Geophys., 33, 403–
439.
Hoor, P., et al. (2009), The impact of traffic emissions on atmospheric ozone and
oh: results from quantify, Atmos. Chem. Phys., 9, 3113–3136.
Hopkin, M. (30 September, 2004), Russia backs Kyoto treaty, Nature, p. Published
Online, doi:10.1038/news040927-15.
Horowitz, L. W., J. Liang, G. M. Gardner, and D. J. Jacob (1998), Export of reactive
nitrogen from North America during summertime: Sensitivity to hydrocarbon
chemistry, J. Geophys. Res., 103, 13,451–13,476.
191
Horowitz, L. W., et al. (2003), A global simulation of tropospheric ozone and re-
lated tracers: Description and evaluation of MOZART, version 2, J. Geophys.
Res., 108, D24,4784, doi:10.1029/2002JD002853.
Horowitz, L. W., A. M. Fiore, G. P. Milly, R. C. Cohen, A. Perring, P. J.
Wooldridge, P. G. Hess, L. K. Emmons, and J.-F. Lamarque (2007), Observa-
tional constraints on the chemistry of isoprene nitrates over the eastern United
States, J. Geophys. Res., 112, D12S08, doi:10.1029/2006JD007747.
Houghton, J., G. Jenkins, and J. Ephraums (Eds.) (1990), Scientific Assessment
of Climate Change; Contribution of Working Group I, Cambridge: Cambridge
University Press.
Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, X. Dai,
K. Maskell, and C. A. Johnson (Eds.) (2001), Climate Change 2001: The Sci-
entific Basis, Contribution of Working Group I to the Third Assessment Report
of the Intergovernmental Panel on Climate Changes, Cambridge: Cambridge
University Press.
Houghton, R. A. (2003), Revised estimates of the annual net flux of carbon to the
atmosphere from changes in land use and land management 1850-2000, Tellus
B, 55, 378–390.
Houweling, S., F. Dentener, and J. Lelieveld (1998), The impact of non-methane
hydrocarbon compounds on tropospheric photochemistry, J. Geophys. Res., 103,
10,673–10,696.
Hoyle, C. R., T. Berntsen, G. Myhre, and I. S. A. Isaksen (2007), Secondary or-
ganic aerosol in the global aerosol—chemical transport model Oslo CTM2, At-
mos. Chem. Phys., 7, 5675–5644.
Huo, H., M. Wang, C. Bloyd, and V. Putsche (2008), Life-cycle assessment of
energy and greenhouse gas effects of soybean-derived biodiesel and renewable
192
fuels, Tech. Rep. ANL/ESD/08-2, Argonne National Laboratory, Energy Systems
Division.
Huo, H., Y. Wu, and M. Wang (2009), Total versus urban: Well-to-wheels assess-
ment of criteria pollutant emissions from various vehicle/fuel systems, Atmos.
Environ., 43, 1796–1804.
Hurtt, G. C., S. Frolking, M. G. Fearon, B. Moore, E. Shevliakova, S. Malyshev,
S. W. Pacala, and R. A. Houghton (2006), The underpinnings of land-use history:
three centuries of global gridded land-use transitions, wood-harvest activity, and
resulting secondary lands, Global Change Biol., 12, 1208–1229.
International Energy Agency (2007), World Energy Outlook 2007, Paris:
OECD/IEA.
International Energy Agency (2008), Key World Energy Statistics, Paris:
OECD/IEA.
Jacob, D. J., L. W. Horowitz, J. W. Munger, B. G. Heikes, R. R. Dicker-
son, R. S. Artz, and W. C. Keene (1995), Seasonal transition from NOx to
hydrocarbon-limited conditions for ozone production over the eastern United
States in September, J. Geophys. Res., 100, 9315–9324.
Jacobson, M. (2007), Effects of ethanol (E85) versus gasoline vehicles on cancer
and mortality in the united states, Environ. Sci. Technol., 41, 4150–4157.
Jaegle´, L., et al. (2004), Satellite mapping of rain-induced nitric oxide emissions
from soils, J. Geophys. Res., 109, D21,310, doi:10.1029/2004JD004787.
Jain, A. K., and X. Yang (2005), Modeling the effects of two different land
cover change data sets on the carbon stocks of plants and soils in concert
with co2 and climate change, Global Biogeochem. Cycles, 19, GB2015, doi:
10.1029/2004GB002349.
193
Jain, A. K., Z. Tao, X. Yang, and C. Gillespie (2006), Estimates of global biomass
burning emissions for reactive greenhouse gases (CO, NMHCs, and NOx) and
CO2, J. Geophys. Res., 111, D06,304, doi:10.1029/2005JD006237.
Jenkin, M. E., and K. C. Clemitshaw (2000), Ozone and other secondary photo-
chemical pollutants: Chemical processes governing their formation in the plan-
etary boundary layer, Atmos. Environ., 34, 2499–2527.
Jenkin, M. E., S. M. Saunders, and M. J. Pilling (1997), The tropospheric degra-
dation of volatile organic compounds: A protocol for mechanism development,
Atmos. Environ., 31, 81–104.
Jerrett, M., R. T. Burnett, I. Pope, C. Arden, K. Ito, G. Thurston, D. Krewski,
Y. Shi, E. Calle, and M. Thun (2009), Long-term ozone exposure and mortality,
N. Engl. J. Med., 360, 1085–1095.
Jiang, X., C. Wiedinmyer, F. Chen, Z. Yang, and J. C. Lo (2008), Predicted impacts
of climate and land use change on surface ozone in the Houston, Texas, areas, J.
Geophys. Res., 113, D20,312, doi:10.1029/2008JD009820.
Jo¨ckel, P., et al. (2006), The atmospheric chemistry general circulation model
ECHAM5/MESSy 1: consistent simulation of ozone from the surface to the
mesosphere, Atmos. Chem. Phys., 6, 5067–5104.
Johansson, T., H. Kelly, A. Reddy, and R. Williams (1993), Renewable Energy:
Sources for Fuels and Electricity, chap. A renewables-intensive global energy
scenario (appendix to Chapter 1), Washington, DC: Island Press.
Johns, T. C., R. E. Carnell, J. F. Crossley, J. M. Gregory, J. F. B. Mitchell, C. A.
Senior, S. F. B. Tett, and R. A. Wood (1997), The second Hadley Centre cou-
pled ocean-atmosphere GCM: Model description, spinup and validation, Cli-
mate Dyn., 13, 103–134.
194
Johnson, C. E., W. J. Collins, D. S. Stevenson, and R. G. Derwent (1999), Relative
roles of climate and emissions changes on future tropospheric oxidant concen-
trations, J. Geophys. Res., 104, 18,631–18,645.
Jones, C. A., and R. A. Rasmussen (1975), Production of isoprene by leaf tissue,
Plant Physiol., 55, 982–987.
Joshi, M., J. Gregory, M. Webb, D. Sexton, and T. Johns (2008), Mechanisms
for the land/sea warming contrast exhibited by simulations of climate change,
Climate Dynamics, 30, 455–465.
Junkermann, W., J. Hacker, T. Lyons, and U. Nair (2009), Land use change sup-
presses precipitation, Atmos. Chem. Phys., 9, 6531–6539.
Kanakidou, M., et al. (2005), Organic aerosol and global climate modelling: A
review, Atmos. Chem. Phys., 5, 1053–1123.
Karavalakis, G., S. Stournas, and E. Bakeas (2009), Effects of diesel/biodiesel
blends on regulated and unregulated pollutants from a passenger vehicle op-
erated over the european and the athens driving cycles, Atmos. Environ., 43,
1745–1752.
Karl, T., A. Guenther, A. Turnipseed, G. Tyndall, P. Artaxo, and S. Martin (2009),
Rapid formation of isoprene photo-oxidation products observed in amazonia,
Atmos. Chem. Phys., 9, 7753–7767.
Kasibhatla, P. S. (1993), NOy from sub-sonic aircraft emissions: A global three-
dimensional model study, Geophys. Res. Lett., 20, 1707–1710.
Keenan, T., . Niinemets, S. Sabate, C. Gracia, and J. Peuelas (2009), Process based
inventory of isoprenoid emissions from european forests: model comparisons,
current knowledge and uncertainties, Atmos. Chem. Phys., 9, 4053–4076.
Kesselmeier, J., and M. Staudt (1999), Biogenic volatile organic compounds
195
(VOC): An overview of emission, physiology and ecology, J. Atmos. Chem.,
33, 23–88.
Klein Goldewijk, K. (2001), Estimating global land use change over the past 300
years: the hyde database, Global Biogeochem. Cycles, 15, 417–433.
Knothe, G., C. A. Sharp, and T. W. Ryan (2006), Exhaust emissions of biodiesel,
petrodiesel, neat methyl esters, and alkanes in a new technology engine, Energ.
Fuels, 20, 403–408.
Koc, M., Y. Sekmen, T. Topgul, and H. S. Yucesu (2009), The effects of ethanol-
unleaded gasoline blends on engine performance and exhaust emissions in a
spark-ignition engine, Renew. Energ., 34, 2101–2106.
Kuhn, U., et al. (2007), Isoprene and monoterpene fluxes from Central Amazonian
rainforest inferred from tower-based and airborne measurements, and implica-
tions on the atmospheric chemistry and the local carbon budget, Atmos. Chem.
Phys., 7, 2855–2879.
Kuzma, J., and R. Fall (1993), Leaf isoprene emission rate is dependent on leaf
development and the level of isoprene synthase, Plant Physiol., 101, 435–440.
Lamb, B., A. Guenther, D. Gay, and H. Westberg (1987), A national inventory of
biogenic hydrocarbon emissions, Atmos. Environ., 21, 1695–1705.
Lamptey, B. L., E. J. Barron, and D. Pollard (2005), Simulation of the relative
impact of land cover and carbon dioxide to climate change from 1700 to 2100,
J. Geophys. Res., 110, D20103, doi:10.1029/2005JD005916.
Lapola, D. M., M. D. Oyama, and C. A. Nobre (2009), Exploring the range
of climate biome projections for tropical south america: The role of CO2
fertilization and seasonality, Global Biogeochem. Cycles, 23, GB3003, doi:
10.1029/2008GB003357.
196
Lapuerta, M., O. Armas, and J. Rodrguez-Fernndez (2008), Effect of biodiesel
fuels on diesel engine emissions, Prog. Energ. Combust., 34, 198–223.
Lardon, L., A. Helias, B. Sialve, J.-P. Steyer, and O. Bernard (2009), Life-cycle
assessment of biodiesel production from microalgae, Environ. Sci. Technol., 43,
6475–6481.
Lathie`re, J., D. A. Hauglustaine, A. D. Friends, N. De Noblet-Ducoudre´, N. Viovy,
and G. A. Folberth (2006), Impact of climate variability and land use changes
on global biogenic volatile organic compound emissions, Atmos. Chem. Phys.,
6, 2129–2146.
Lathie`re, J., C. Hewitt, and D. Beerling (2009), Sensitivity of isoprene emissions
from the terrestrial biosphere to 20th century changes in atmospheric co concen-
tration, climate and land use, Global Biogeochem. Cycles, in press.
Lathie`re, J. A., D. A. Hauglustaine, N. De Noblet-Ducoudre´, G. Krinner, and G. A.
Folberth (2005), Past and future changes in biogenic volatile organic compound
emissions simulated with a global dynamic vegetation model, Geophys. Res.
Lett., 32, L20,818, doi:10.1029/2005GL024164.
Law, K. S., and J. A. Pyle (1993a), Modeling trace gas budgets in the troposphere
1. Ozone and odd nitrogen, J. Geophys. Res., 98, 18,377–18,400.
Law, K. S., and J. A. Pyle (1993b), Modeling trace gas budgets in the troposphere
2. CH4 and CO, J. Geophys. Res., 98, 18,401–18,412.
Law, K. S., P. H. Plantevin, D. E. Shallcross, H. J. Rogers, J. A. Pyle, C. Grouhel,
V. Thouret, and A. Marenco (1998), Evaluation of modeled O3 using Measure-
ment of Ozone by Airbus In-Service Aircraft (MOZAIC) data, J. Geophys. Res.,
103, 25,721–25,737.
Lelieveld, J., P. J. Crutzen, and F. J. Dentener (1998), Changing concentration,
lifetime and climate forcing of atmospheric methane, Tellus B, 50, 128–150.
197
Lelieveld, J., et al. (2008), Atmospheric oxidation capacity sustained by a tropical
forest, Nature, 452, 737–740.
Lerdau, M., and D. Gray (2003), Ecology and evolution of light-dependent and
light-independent phytogenic volatile organic carbon, New Phytol., 157, 199–
211.
Levis, S., C. Wiedinmyer, G. B. Bonan, and A. Guenther (2003), Simulating bio-
genic volatile organic compound emissions in the Community Climate System
Model, J. Geophys. Res., 108, D21,4659, doi:10.1029/2002JD003203.
Levy II, H. (1971), Normal atmosphere: Large radical and formaldehyde concen-
trations predicted, Science, 173, 141–143.
Levy II, H., W. J. Moxim, A. A. Klonecki, and P. S. Kasibhatla (1999), Simulated
tropospheric NOx : Its evaluation, global distribution and individual source con-
tributions, J. Geophys. Res., 104, 26,279–26,306.
Li, D., and K. P. Shine (1995), A 4-dimensional ozone climatology for UGAMP
models, UGAMP Internal Report No. 35, United Kingdom University Global
Atmospheric Modelling Programme, Reading, UK.
Liao, H., W.-T. Chen, and J. H. Seinfeld (2006), Role of climate change in global
predictions of future tropospheric ozone and aerosols, J. Geophys. Res., 111,
D12304, doi:10.1029/2005JD006852.
Lin, C. Y. C., D. J. Jacob, and A. M. Fiore (2001), Trends in exceedances of the
ozone air quality standard in the continental united states, 1980-1998, Atmos.
Environ., 35, 3217–3228.
Lin, J.-T., K. O. Patten, K. Hayhoe, X.-Z. Liang, and D. J. Wuebbles (2008), Ef-
fects of future climate and biogenic emissions changes on surface ozone over
the United States and China, J. Appl. Meteor. Climatol., 47, 1888–1909.
198
Lin, Y., and S. Tanaka (2006), Ethanol fermentation from biomass resources: cur-
rent state and prospects, Appl. Microbio. Biot., 69, 627–642.
Lindinger, W., A. Hansel, and A. Jordan (1998), On-line monitoring of volatile
organic compounds at pptv levels by means of proton-transfer-reaction mass
spectrometry (PTR-MS) medical applications, food control and environmental
research, Int. J. Mass Spectrom. Ion Processes, 173, 191–241.
Liu, S. C., M. Trainer, F. C. Fehsenfeld, D. D. Parrish, E. J. Williams, D. W. Fahey,
G. Hu¨bler, and P. C. Murphy (1987), Ozone production in the rural troposphere
and the implications for regional and global ozone distributions, J. Geophys.
Res., 92, 4191–4207.
Logan, B. A., R. K. Monson, and M. J. Potosnak (2000), Biochemistry and physi-
ology of foliar isoprene production, Trends Plant Sci., 5, 477–481.
Logan, J. A. (1999), An analysis of ozonesonde data for the troposphere: Recom-
mendations for testing 3-D models and development of a gridded climatology
for tropospheric ozone, J. Geophys. Res., 104, 16,115–16,150.
Logan, J. A., M. J. Prather, S. C. Wolfsy, and M. B. McElroy (1981), Tropospheric
chemistry: A global perspective, J. Geophys. Res., 86, 7210–7254.
Long, S. P., E. A. Ainsworth, A. D. Leakey, and P. B. Morgan (2005), Global
food insecurity. treatment of major food crops with elevated carbon dioxide or
ozone under large-scale fully open-air conditions suggests recent models may
have overestimated future yields, Phil. Trans. R. Soc. B, 360, 2011–2020.
Loreto, F., and T. D. Sharkey (1990), A gas-exchange study of photosynthesis and
isoprene emission in quercus-rubra l, Planta, 182, 523–531.
Loreto, F., M. Mannozzi, C. Maris, P. Nascetti, F. Ferranti, and S. Pasqualini
(2001), Ozone quenching properties of isoprene and its antioxidant role in
leaves, Plant Physiol., 126, 993–1000.
199
Loveland, T. R., B. C. Reed, J. F. Brown, D. O. Ohlen, Z. Zhu, L. Yang, and J. W.
Merchant (2000), Development of a global land cover characteristics database
and IGBP DISCover from 1 km AVHRR data, Int. J. Remote Sens., 21, 1303–
1330.
Lynch, P. (2008), The origins of computer weather prediction and climate model-
ing, Journal of Computational Physics, 227, 3431–3444.
Ma, F., and M. A. Hanna (1999), Biodiesel production: a review, Bioresource
Technol., 70, 1–15.
MacLean, H. L., and L. B. Lave (2003), Evaluating automobile fuel/propulsion
system technologies, Prog. Energ. Combust., 29, 1–69.
Martin, M. J., C. M. Stirling, S. W. Humphries, and S. P. Long (2000), A process-
based model to predict the effects of climatic change on leaf isoprene emission
rates, Ecol. Model., 131, 161–174.
Matthes, S., V. Grewe, R. Sausen, and G.-J. Roelofs (2007), Global impact of road
traffic emissions on tropospheric ozone, Atmos. Chem. Phys., 7, 1707–1718.
McCormick, R. L., M. S. Graboski, T. L. Alleman, A. M. Herring, and K. S. Tyson
(2001), Impact of biodiesel source material and chemical structure on emissions
of criteria pollutants from a heavy-duty engine, Environ. Sci. Technol., 35, 1742–
1747.
Meleux, F., F. Solmon, and F. Giorgi (2007), Increase in summer european ozone
amounts due to climate change, Atmos. Environ., 41, 7577–7587.
Mickley, L. J., D. J. Jacob, B. D. Field, and D. Rind (2004), Effects of future
climate change on regional air pollution episodes in the United States, Geophys.
Res. Lett., 31, L24,103, doi:10.1029/2004GL021216.
Millet, D. B., D. J. Jacob, K. F. Boersma, T.-M. Fu, T. P. Kurosu, K. Chance, C. L.
Heald, and A. Guenther (2008), Spatial distribution of isoprene emissions from
200
North America derived from formaldehyde column measurements by the OMI
satellite sensor, J. Geophys. Res., 113, D02,307, doi:10.1029/2007JD008950.
Mills, G., A. Buse, B. Gimeno, V. Bermejo, M. Holland, L. Emberson, and H. Plei-
jel (2007), A synthesis of aot40-based response functions and critical levels of
ozone for agricultural and horticultural crops, Atmos. Environ., 41, 2630–2643.
Milt, A., A. Milano, S. Garivait, and R. Kamens (2009), Effects of 10substitution
on ground level ozone formation in bangkok, thailand, Atmospheric Environ-
ment, 43, 5962–5970.
Minerallwirtschaftsverband (2006), Mwv-prognose 2025 fr die bundesrepublik
deutschland, Tech. rep., Minerallwirtschaftsverband.
Monks, P. S. (2000), A review of the observations and origins of the spring ozone
maximum, Atmos. Environ., 34, 3545–3561.
Monson, R. K., and R. Fall (1989), Isoprene emission from aspen leaves: Influence
of the environment and relation to photosynthesis and phytorespiration, Plant
Physiol., 90, 267–274.
Monyem, A., and J. H. Van Gerpen (2001), The effect of biodiesel oxidation on
engine performance and emissions, Biomass and Bioenerg., 20, 317–325.
Moreira, J. R. (2000), Sugarcane for energy—recent results and progress in brazil,
Energy for Sustainable Development, 4, 43–54.
Moreira, R. (2006), Global biomass energy potential, Mitigation and Adaptation
Strategies for Global Change, 11, 313–333.
Morgenstern, O., P. Braesicke, M. M. Hurwitz, F. M. O’Connor, A. C. Bushell,
C. E. Johnson, and J. A. Pyle (2008), The World Avoided by the Montreal Pro-
tocol, Geophys. Res. Lett., 35, L16,811, doi:10.1029/2008GL034590.
201
Morgenstern, O., P. Braesicke, F. M. O’Connor, A. C. Bushell, C. E. Johnson,
S. M. Osprey, and J. A. Pyle (2009), Evaluation of the new UKCA climate-
composition model Part I: The stratosphere, Geosci. Model Dev., 2, 43–57.
Moxim, W. J., H. Levy II, and P. S. Kasibhatla (1996), Simulated global tropo-
spheric PAN: Its transport and impact on NOx, J. Geophys. Res., 101, 12,621–
12,638.
Mu¨ller, C., B. Eickhout, S. Zaehle, A. Bondeau, W. Cramer, and W. Lucht (2007),
Effects of changes in CO2, climate, and land use on the carbon balance of the
land biosphere during the 21st century, J. Geophys. Res., 112, G02,032, doi:
10.1029/2006JG000388.
Mu¨ller, J.-F. (1992), Geographical distribution and seasonal variation of surface
emissions and deposition velocities of atmospheric trace gases, J. Geophys. Res.,
97, 3787–3804.
Mu¨ller, J.-F., and G. Brasseur (1995), IMAGES: A three-dimensional chemi-
cal transport model of the global troposphere, J. Geophys. Res., 100, 16,445–
16,490.
Murphy, D. M., and D. W. Fahey (1994), An estimate of the flux of stratospheric
reactive nitrogen and ozone into the troposphere, J. Geophys. Res., 99, 5325–
5332.
Naik, V., C. Delire, and D. J. Wuebbles (2004), Sensitivity of global biogenic iso-
prenoid emissions to climate variability and atmospheric CO2, J. Geophys. Res.,
109, D06301, doi:10.1029/2003JD004236.
Nakicenovic, N., et al. (2000), IPCC Special Report on Emissions Scenarios, Cam-
bridge: Cambridge University Press.
Niemeier, U., C. Granier, L. Kornblueh, S. Walters, and G. Brasseur (2006), Global
202
impact of road traffic on atmospheric chemical composition and on ozone cli-
mate forcing, J. Geophys. Res., 111, D09301, doi:10.1029/2005JD006407.
Niinemets, U., J. D. Tenhunen, P. C. Harley, and R. Steinbrecher (1999), A model
of isoprene emission based on energetic requirements for isoprene synthesis and
leaf photosynthetic properties for Liquidambar and Quercus, Plant Cell Envi-
ron., 22, 1319–1335.
Nolte, C. G., A. B. Gilliland, C. Hogrefe, and L. J. Mickley (2008), Linking global
to regional models to assess future climate impacts on surface ozone levels in
the United States, J. Geophys. Res., 113, D14,307, doi:10.1029/2007JD008497.
O’Connor, F. M., C. E. Johnson, O. Morgenstern, and W. J. Collins (2009), In-
teractions between tropospheric chemistry and climate model temperature and
humidity biases, Geophys. Res. Lett., 36, L16,801, doi:10.1029/2009GL039152.
Office of the Governor of California (18 January 2007), Governor’s Executive Or-
der S-01-07.
Ohlrogge, J., D. Allen, B. Berguson, D. DellaPenna, Y. Shachar-Hill, and
S. Stymne (2009), Driving on biomass, Science, 324, 1019–1020.
Olivier, J., and J. Berdowski (2001), The Climate System, chap. Global emissions
sources and sinks, pp. 33–78, Leiden, The Netherlands: A.A. Balkema Publish-
ers/Swets & Zeitlinger Publishers.
One Hundred Tenth Congress of the United States of America (4 January, 2007),
HR6: Energy Independence and Security Act.
OP3 (Accessed 11 November, 2009), Online: http://www.es.lancs.ac.uk/op3/.
Pacala, S., and R. Socolow (2004), Stabilization Wedges: Solving the Climate
Problem for the Next 50 Years with Current Technologies, Science, 305, 968–
972.
203
Palmer, P. I., D. J. Jacob, A. M. Fiore, R. V. Martin, K. Chance, and T. P. Kurosu
(2003), Mapping isoprene emissions over North America using formaldehyde
column observations from space, J. Geophys. Res., 108, D64,180.
Pandian, S., S. Gokhale, and A. K. Ghoshal (2009), Evaluating effects of traf-
fic and vehicle characteristics on vehicular emissions near traffic intersections,
Transport. Res. D–Tr. E., 14, 180–196.
Park, R. J., K. E. Pickering, D. J. Allen, G. L. Stenchikov, and M. S. Fox-
Rabinovitz (2004), Global simulation of tropospheric ozone using the University
of Maryland Chemical Transport Model (UMD-CTM): 1 Model description and
evaluation, J. Geophys. Res., 109, D09301, doi:10.1029/2003JD004266.
Patterson, T., R. Dinsdale, and S. Esteves (2008), Review of energy balances and
emissions associated with biomass-based transport fuels relevant to the united
kingdom context, Energy & Fuels, 22, 3506–3512.
Paulot, F., J. D. Crounse, H. G. Kjaergaard, A. Kurten, J. M. St. Clair, J. H. Sein-
feld, and P. O. Wennberg (2009), Unexpected epoxide formation in the gas-phase
photooxidation of isoprene, Science, 325, 730–733.
Pearson, G., F. Davies, and C. Collier (2009), Remote sensing of the tropical rain
forest boundary layer using pulsed doppler lidar, Atmos. Chem. Phys., in prep.
Peh, K. S.-H., N. S. Sodhi, J. de Jong, C. H. Sekercioglu, C. A.-M. Yap, and
S. L.-H. Lim (2006), Conservation value of degraded habitats for forest birds in
southern peninsular Malaysia, Divers. Distrib., 12, 572–581.
Pfister, G. G., L. K. Emmons, P. G. Hess, J.-F. Lamarque, J. J. Orlando, S. Walters,
A. Guenther, P. I. Palmer, and P. J. Lawrence (2008a), Contribution of isoprene
to chemical budgets: A model tracer study with the NCAR CTM MOZART-4,
J. Geophys. Res., 113, D05,308, doi:10.1029/2007JD008948.
204
Pfister, G. G., L. K. Emmons, P. G. Hess, J.-F. Lamarque, A. M. Thompson,
and J. E. Yorks (2008b), Analysis of the summer 2004 ozone budget over the
united states using intercontinental transport experiment ozonesonde network
study (ions) observations and model of ozone and related tracers (MOZART-4)
simulations, J. Geophys. Res., 113, D23,306, doi:10.1029/2008JD010190.
Pierce, T., C. Geron, L. Bender, R. Dennis, G. Tonnesen, and A. Guenther (1998),
Influence of increased isoprene emissions on regional ozone modeling, J. Geo-
phys. Res., 103, 25,611–25,630.
Pin Koh, L., and D. S. Wilcove (2007), Cashing in palm oil for conservation (com-
mentary), Nature, 448, 993–994.
Pinzi, S., I. L. Garcia, F. J. Lopez-Gimenez, M. D. Luque de Castro, G. Dorado,
and M. P. Dorado (2009), The ideal vegetable oil-based biodiesel composition:
A review of social, economical and technical implications, Energy & Fuels, 23,
2325–2341.
Pitman, A. J., et al. (2009), Uncertainties in climate responses to past land cover
change: First results from the LUCID intercomparison study, Geophys. Res.
Lett., 36, L14,814, doi:10.1029/2009GL039076.
Platzman, G. W. (1979), The ENIAC computations of 1950–gateway to numerical
weather prediction, Bull. Am. Meteor. Soc., 60, 302–312.
POET Inventory (Accessed 21 October, 2009), Online:
http://www.aero.jussieu.fr/projet/ACCENT/POET.php.
Pongratz, J., C. Reick, T. Raddatz, , and M. Claussen (2008), A reconstruction
of global agricultural areas and land cover for the last millennium, Global Bio-
geochem. Cycles, 22, GB3018, doi:10.1029/2007GB003153,2008.
Pongratz, J., T. Raddatz, C. H. Reick, M. Esch, and M. Claussen (2009), Radiative
205
forcing from anthropogenic land cover change since a.d. 800, Geophys. Res.
Lett., 36, L02,709, doi:10.1029/2008GL036394.
Po¨schl, U., R. von Kulhmann, N. Poisson, and P. J. Crutzen (2000), Development
and intercomparison of condensed isoprene oxidation mechanisms for global
atmospheric modelling, J. Atmos. Chem., 37, 29–52.
Possell, M., C. N. Hewitt, and D. J. Beerling (2005), The effects of glacial at-
mospheric CO2 concentrations and climate on isoprene emissions by vascular
plants, Global Change Biol., 11, 60–69.
Potter, C. S., S. E. Alexander, J. C. Coughlan, and S. A. Klooster (2001), Modeling
biogenic emissions of isoprene: Exploration of model drivers, climate control
algorithms and use of global satellite observations, Atmos. Environ., 35, 6151–
6165.
Prather, M., et al. (2001), Climate Change 2001: The Scientific Basis, Contribution
of Working Group I to the Third Assessment Report of the Intergovernmental
Panel on Climate Change, chap. Atmospheric chemistry and greenhouse gases,
pp. 239–287, Cambridge: Cambridge University Press, Cambridge, UK.
Price, C., and D. Rind (1992), A simple lightning parameterization for calculating
global lightning distributions, J. Geophys. Res., 97, 9919–9933.
Priestley, A. (1993), A quasi-conservative version of the semi-lagrangian advection
scheme, Mon. Weather Rev., 121, 621–629.
Quaife, T., S. Quegan, M. Disney, P. Lewis, M. Lomas, and F. I. Woodward (2008),
Impact of land cover uncertainties on estimates of biospheric carbon fluxes,
Global Biogeochem. Cycles, 22, GB4016, doi:10.1029/2007GB003097.
Racherla, P. N., and P. J. Adams (2006), Sensitivity of global tropospheric ozone
and fine particulate matter concentrations to climate change, J. Geophys. Res.,
111, D24,103, doi:10.1029/2005JD006939.
206
Racherla, P. N., and P. J. Adams (2008), The response of surface ozone to climate
change over the eastern United States, Atmos. Chem. Phys., 8, 871–885.
Ramankutty, N., and J. A. Foley (1999), Estimating historical changes in global
land cover: Croplands from 1700 to 1992, Global Biogeochem. Cycles, 13, 997–
1027.
Ramankutty, N., C. Delire, and P. Snyder (2006), Feedbacks between agriculture
and climate: An illustration of the potential unintended consequences of human
land use activities, Global Planet. Change, 54, 79–93.
Rasmussen, R., and C. Jones (1973), Emission isoprene from leaf discs of
hamamelis, Phytochemistry, 12, 15–19.
Rasmussen, R. A. (1970), Isoprene: Identified as a forest-type emission to the
atmosphere, Environ. Sci. Technol., 4, 667–671.
Rasmussen, R. A., and M. A. Khalil (1988), Isoprene over the Amazon basin, J.
Geophys. Res., 93, 1417–1421.
Rasmussen, R. A., and F. W. Went (1965), Volatile organic carbon of plant origin
in the atmosphere, Proc. Natl. Acad. Sci. USA, 53, 215–220.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, E. C.
Rowell, D. P.and Kent, and J. Kaplan (2003), A global analyses of sea surface
temperature, sea ice, and night marine air temperature since the late nineteenth
century, J. Geophys. Res., 108, D14,4407, doi:10.1029/2002JD002670.
Ren, X., et al. (2008), HOx chemistry during INTEX-A 2004: Observation,
model calculation, and comparison with previous studies, J. Geophys. Res., 113,
D05,310, doi:10.1029/2007JD009166.
Renewable Fuels Agency (Accessed 21 October, 2009), RTFO targets, Online:
http://www.renewablefuelsagency.org/index.cfm.
207
Renewable Fuels Association (Accessed 15 September, 2009), Ethanol industry
statistics, Online: http://www.ethanolrfa.org/industry/statistics/.
Riahi, K., A. Grbler, and N. Nakicenovic (2007), Scenarios of long-term socio-
economic and environmental development under climate stabilization, Technol.
Forecast. Soc., 74, 887–935.
Richardson, L. (1922), Weather Prediction by Numerical Process, Cambridge:
Cambridge University Press.
Rinne, H. J. I., A. B. Guenther, J. P. Greenberg, and P. C. Harley (2002), Isoprene
and monoterpene fluxes measured above amazonian rainforest and their depen-
dence on light and temperature, Atmos. Environ., 36, 2421–2426.
Roelofs, G.-J., and J. Lelieveld (1995), Distribution and budget of O3 in the tropo-
sphere calculated with a chemical general circulation model, J. Geophys. Res.,
100, 20,983–20,998.
Roelofs, G.-J., and J. Lelieveld (2000), Tropospheric ozone simulation with a
chemistry-general circulation model: Influence of higher hydrocarbon chem-
istry, J. Geophys. Res., 105, 22,697–22,712.
Romm, J. (2006), The car and fuel of the future, Energy Policy, 34, 2609–2614.
Rosenstiel, T. N., M. J. Potosnak, K. L. Griffin, R. Fall, and R. K. Monson (2003),
Increased CO2 uncouples growth from isoprene emission in an agriforest ecosys-
tem, Nature, 421, 256–259.
Rowe, R. L., N. R. Street, and G. Taylor (2009), Identifying potential environmen-
tal impacts of large-scale deployment of dedicated bioenergy crops in the UK,
Renew. Sust. Energ. Rev., 13, 271–290.
Ruppert, L., and K. Heinz-Becker (2000), A product study of the OH radical-
initiated oxidation of isoprene: Formation of C5-unsaturated diols, Atmos. Env-
iron., 34, 1529–1542.
208
Sanadze, G. (1964), Light-dependent excretion of isoprene by plants, Photosynthe-
sis Research, 2, 701–707.
Sanadze, G. A. (1957), The nature of gaseous substances emitted by leaves of
Robina pseudoacacia, Soobshch. Akad. Nauk Gruz. SSR, 19, 83–86.
Sanderson, M. G., C. D. Jones, W. J. Collins, C. E. Johnson, and R. G. Derwent
(2003), Effect of climate change on isoprene emissions and surface ozone levels,
Geophys. Res. Lett., 30, 18,1936, doi:10.1029/2003GL017642.
Saunders, S. M., M. E. Jenkin, R. G. Derwent, and M. J. Pilling (1997), World
wide web site of a master chemical mechanism (MCM) for use in tropospheric
chemistry models, Atmos. Environ., 31, 1249–1249.
Saunders, S. M., M. E. Jenkin, R. G. Derwent, and M. J. Pilling (2003), Protocol
for the development of the Master Chemical Mechanism, MCM v3 (Part a):
Tropospheric degradation of non-aromatic volatile organic compounds, Atmos.
Chem. Phys., 3, 161–180.
Savage, N. H., K. S. Law, J. A. Pyle, A. Richter, H. Nuss, and J. P. Burrows (2004),
Using GOME NO2 satellite data to examine regional differences in TOMCAT
model performance, Atmos. Chem. Phys., 4, 1895–1912.
Scaife, A. A., N. Butchart, C. D. Warner, and R. Swinbank (2002), Impact of
a spectral gravity wave parameterization on the stratosphere in the met office
unified model, J. Atmos. Sci., 59, 1473–1489.
Schiermeier, Q., J. Tollefson, T. Scully, A. Witze, and O. Morton (2008), Energy
alternatives: Electricity without carbon, Nature, 454, 816–823.
Schilling, K. E., M. K. Jha, Zhang, You-Kuan, P. W. Gassman, and C. F. Wolter
(2008), Impact of land use and land cover change on the water balance of a large
agricultural watershed: Historical effects and future directions, Water Resour.
Res., 44, W00A09, doi:10.1029/2007WR006644.
209
Schmidt, G. A., et al. (2006), Present-day atmospheric simulations using GISS
ModelE: Comparison to in situ, satellite, and reanalysis data, J. Climate, 19,
153–192.
Schumann, U., and H. Huntrieser (2007), The global lightning-induced nitrogen
oxides source, Atmos. Chem. Phys., 7, 3823–3907.
Service, R. (2007), Cellulosic ethanol: Biofuel researchers prepare to reap a new
harvest, Science, 315, 1488 – 1491.
Service, R. F. (2009), ExxonMobil fuels Venter’s efforts to run vehicles on algae-
based oil, Science, 325, 379.
Sharkey, T. D. (1996), Isoprene synthesis by plants and animals, Endeavour, 20,
74–78.
Sharkey, T. D., and E. L. Singsaas (1995), Why plants emit isoprene, Nature, 374,
769–769.
Shay, E. G. (1993), Diesel fuel from vegetable oils: Status and opportunities,
Biomass and Bioenergy, 4, 227–242.
Shepson, P. B., E. Mackay, and K. Muthuramu (1996), Henry’s Law constants and
removal processes for several atmospheric β-hydroxy alkyl nitrates, Environ.
Sci. Technol., 30, 3618–3623.
Shevliakova, E., et al. (2009), Carbon cycling under 300 years of land use change:
Importance of the secondary vegetation sink, Global Biogeochem. Cycles, 23,
GB2022, doi:10.1029/2007GB003176.
Shim, C., Y. Wang, Y. Choi, P. I. Palmer, D. S. Abbot, and K. Chance (2005), Con-
straining global isoprene emissions with Global Ozone Monitoring Experiment
(GOME) formaldehyde column measurements, J. Geophys. Res., 110, D24301,
doi:10.1029/2004JD005629.
210
Shindell, D., G. Faluvegi, A. Lacis, J. Hansen, R. Ruedy, and E. Aguilar (2006a),
Role of tropospheric ozone increases in 20th-century climate change, J. Geo-
phys. Res., 111, D08302, doi:10.1029/2005JD006348.
Shindell, D., J.-F. Lamarque, N. Unger, D. Koch, G. Faluvegi, S. Bauer, M. Am-
mann, J. Cofala, and H. Teich (2008), Climate forcing and air quality change
due to regional emissions reductions by economic sector, Atmos. Chem. Phys.,
8, 7101–7113.
Shindell, D. T., et al. (2006b), Multimodel simulations of carbon monoxide: Com-
parison with observations and projected near-future changes, J. Geophys. Res.,
111, D19,306, doi:10.1029/2006JD007100.
Sillman, S. (1995), The use of NOy, H2O2, and HNO3 as indicators for ozone-
NOx-hydrocarbon sensitivity in urban locations, J. Geophys. Res., 100, 14,175–
14,188.
Sillman, S. (1999), The relation between ozone, NOx and hydrocarbons in urban
and polluted rural environments, Atmos. Environ., 33, 1821–1845.
Sillman, S., and D. He (2002), Some theoretical results concerning O3-NOx-VOC
chemistry and NOx-VOC indicators, J. Geophys. Res., 107, D224,659, doi:10.
1029/2001JD001123.
Sillman, S., and P. J. Samson (1995), Impact of temperature on oxidant photo-
chemistry in urban, polluted rural and remote environments, J. Geophys. Res.,
100, 11,497–11,508.
Sillman, S., and J. J. West (2009), Reactive nitrogen in mexico city and its rela-
tion to ozone-precursor sensitivity: results from photochemical models, Atmos.
Chem. Phys., 9, 3477–3489.
Sillman, S., J. A. Logan, and S. C. Wofsy (1990), The sensitivity of ozone to
211
nitrogen oxides and hydrocarbons in regional ozone episodes, J. Geophys. Res.,
95, 1837–1851.
Sims, R., A. Hastings, B. Schlamadinger, G. Taylor, and P. Smith (2006), Energy
crops: current status and future prospects, Global Change Biol., 12, 2054–2076.
Sims, R., M. Taylor, J. Saddler, and W. Mabee (2008), From 1st- to 2nd-generation
biofuel technologies: An overview of current industry and RD&D activities,
Tech. rep., International Energy Agency.
Sims, R. E., and R. N. Schock (Eds.) (2007), Mitigation of Climate Change, Con-
tribution of Working Group III to the Fourth Assessment Report of the Inter-
governmental Panel on Climate Change, chap. Energy Supply, pp. 251–322,
Cambridge: Cambridge University Press.
Singh, H. B., and P. L. Hanst (1981), Peroxyacetyl nitrate (PAN) in the unpolluted
atmosphere: An important reservoir for nitrogen oxides, Geophys. Res. Lett., 8,
941–944.
Sitch, S., V. Brovkin, W. von Bloh, D. van Vuuren, B. Eickhout, and A. Ganopolski
(2005), Impacts of future land cover changes on atmospheric CO2 and climate,
Global Biogeochem. Cycles, 19, GB2013, doi:10.1029/2004GB002311.
Solomon, B. D., and A. Banerjee (2006), A global survey of hydrogen energy
research, development and policy, Energy Policy, 34, 781–792.
Solomon, B. D., J. R. Barnes, and K. E. Halvorsen (2007a), Grain and cellulosic
ethanol: History, economics, and energy policy, Biomass and Bioenerg., 31,
416–425.
Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. Averyt, M. Tignor,
and H. Miller (Eds.) (2007b), Climate Change 2007: The Physical Science Ba-
sis, Contribution of Working Group I to the Fourth Assessment Report of the
212
Intergovernmental Panel on Climate Change, Cambridge: Cambridge Univer-
sity Press.
Sorteberg, A., and O. Hov (1996), Two parametrizations of the dry deposition ex-
change for SO2 and NH3 in a numerical model, Atmos. Environ., 30, 1823–1840.
Spivakovsky, C. M., et al. (2000), Three-dimensional climatological distribution of
tropospheric OH: Update and evaluation, J. Geophys. Res., 105, 8931–8980.
Sprengnether, M., K. L. Demerjian, N. M. Donahue, and J. G. Anderson (2002),
Product analysis of the OH oxidation of isoprene and 1,3-butadiene in the pres-
ence of NO, J. Geophys. Res., 107, D15,4268, doi:10.1029/2001JD000716.
Steiner, A. L., S. Tonse, R. C. Cohen, A. H. Goldstein, and R. A. Harley (2006),
Influence of future climate and emissions on regional air quality in California, J.
Geophys. Res., 111, D18,303, doi:10.1029/2005JD006935.
Steinkamp, J., L. N. Ganzeveld, W. Wilcke, and M. G. Lawrence (2009), Influence
of modelled soil biogenic NO emissions on related trace gases and the atmo-
spheric oxidizing efficiency, Atmos. Chem. Phys., 9, 2663–2677.
Stevenson, D., R. Doherty, M. Sanderson, C. Johnson, B. Collins, and D. Derwent
(2005), Impacts of climate change and variability on tropospheric ozone and its
precursors, Faraday Discuss., 130, 41–57.
Stevenson, D. S., C. E. Johnson, W. J. Collins, R. G. Derwent, and J. M. Edwards
(2000), Future estimates of tropospheric ozone radiative forcing and methane
turnoverthe impact of climate change, Geophys. Res. Lett., 27, 2073–2076.
Stevenson, D. S., et al. (2006), Multimodel ensemble simulations of present-
day and near-future tropospheric ozone, J. Geophys. Res., 111, D08301, doi:
10.1029/2005JD006338.
Stott, P. A., G. S. Jones, J. A. Lowe, P. Thorne, C. Durman, T. C. Johns, and J.-C.
213
Thelen (2006), Transient climate simulations with the hadgem1 climate model:
Causes of past warming and future climate change, J. Climate, 19, 2763–2782.
Sustainable biofuels: prospects and challenges (2008), Policy Document ISBN 978
0 85403 662 2, The Royal Society.
Szopa, S., D. A. Hauglustaine, R. Vautard, and L. Menut (2006), Future global
tropospheric ozone changes and impact on European air quality, Geophys. Res.
Lett., 33, L14,805, doi:10.1029/2006GL025860.
Szybist, J. P., A. L. Boehman, J. D. Taylor, and R. L. McCormick (2005), Evalu-
ation of formulation strategies to eliminate the biodiesel NOx effect, Fuel Pro-
cess. Technol., 86, 1109–1126.
Szybist, J. P., J. Song, M. Alam, and A. L. Boehman (2007), Biodiesel combustion,
emissions and emission control, Fuel Processing Technology, 88, 679–691.
Tagaris, E., K. Manomaiphiboon, K.-J. Liao, L. R. Leung, J.-H. Woo, S. He,
P. Amar, and A. G. Russell (2007), Impacts of global climate change and
emissions on regional ozone and fine particulate matter concentrations over the
United States, J. Geophys. Res., 112, D14,312,, doi:10.1029/2006JD008262.
Talukdar, R. K., J. B. Burkholder, A.-M. Schmoltner, J. M. Roberts, R. R. Wil-
son, and A. R. Ravishankara (1995), Investigation of the loss processes for per-
oxyacetyl nitrate in the atmosphere: UV photolysis and reaction with OH, J.
Geophys. Res., 100, 14,163–14,174.
Tao, Z., and A. K. Jain (2005), Modeling of global biogenic emissions for key
indirect greenhouse gases and their response to atmospheric CO2 increases and
changes in land cover and climate, J. Geophys. Res., 110, D21309, doi:10.1029/
2005JD005874.
Tao, Z., S. M. Larson, D. J. Wuebbles, A. Williams, and M. Caugh (2003),
A summer simulation of biogenic contributions to ground-level ozone over
214
the continental United States, J. Geophys. Res., 108, D14,4404, doi:10.1029/
2002JD002945.
Tao, Z., A. Williams, H.-C. Huang, M. Caughey, and X.-Z. Liang (2007), Sensi-
tivity of U. S. surface ozone to future emissions and climate changes, Geophys.
Res. Lett., 34, L08,811, doi:10.1029/2007GL029455.
Teysse`dre, H., et al. (2007), A new chemistry-climate tropospheric and strato-
spheric model MOCAGE-Climat: evaluation of the present-day climatology and
sensitivity to surface processes, Atmos. Chem. Phys., 7, 5815–5860.
Thompson, A. M. (1992), The oxidizing capacity of the earth’s atmosphere: Prob-
able past and future changes, Science, 257, 1157–1165.
Thompson, A. M., R. W. Stewart, M. A. Owens, and J. A. Herwehe (1989), Sen-
sitivity of tropospheric oxidants to global chemical and climate change, Atmos.
Environ., 23, 519–532.
Thompson, A. M., et al. (2003), Southern Hemisphere Additional Ozoneson-
des (SHADOZ) 1998-2000 tropical ozone climatology 2. Tropospheric vari-
ability and the zonal wave-one, J. Geophys. Res., 108, D2,8241, doi:10.1029/
2002JD002241.
Tilman, D., J. Hill, and C. Lehman (2006), Carbon-Negative Biofuels from Low-
Input High-Diversity Grassland Biomass, Science, 314, 1598–1600.
Tilman, D., et al. (2009), Beneficial biofuels–the food, energy, and environment
trilemma, Science, 325, 270–271.
Tingey, D. T., M. Manning, L. C. Grothaus, and W. F. Burns (1979), The influ-
ence of light and temperature on isoprene emission rates from live oak, Physiol.
Plantarum, 47, 112–118.
Tingey, D. T., R. Evans, and M. Gumpertz (1981), Effects of environmental condi-
tions on isoprene emission from live oak, Planta, 152, 565–570.
215
Toenges-Schuller, N., et al. (2006), Global distribution pattern of anthro-
pogenic nitrogen oxide emissions: Correlation analysis of satellite measure-
ments and model calculations, J. Geophys. Res., 111, D05,312, doi:10.1029/
2005JD006068.
Tribbia, J. J., and R. A. Anthes (1987), Scientific basis of modern weather predic-
tion, Science, 237, 493–499.
Turner, D. P., J. V. Baglio, A. G. Wones, D. Pross, R. Vong, B. D. McVeety, and
D. L. Phillips (1991), Climate change and isoprene emissions from vegetation,
Chemosphere, 23, 37–56.
Turner, E. C., and W. A. Foster (2009), The impact of forest conversion to oil palm
on arthropod abundance and biomass in Sabah, Malaysia, J. Trop. Ecol., 25,
23–30.
Unger, N., D. T. Shindell, D. M. Koch, M. Amann, J. Cofala, and D. G. Streets
(2006), Influences of man-made emissions and climate changes on tropospheric
ozone, methane, and sulfate at 2030 from a broad range of possible futures, J.
Geophys. Res., 111, D12313, doi:10.1029/2005JD006518.
Unger, N., D. T. Shindell, and J. S. Wang (2009), Climate forcing by the on-road
transportation and power generation sectors, Atmos. Environ., 43, 3077–3085.
United Kingdom Parliament (2007-2008 Session), Climate change bill, Available
Online: http://services.parliament.uk/bills/2007-08/climatechangehl.html.
United Kingdom Statutory Instrument No. 3072 (26 October, 2007), Transport en-
ergy, sustainable and renewable fuels: The renewable transport fuel obligations
order, London: The Stationery Office Limited.
United Nations Framework Convention on Climate Change (1992), Online:
http://unfccc.int/essential background/convention/background/items/1349.php.
216
United Nations Framework Convention on Cli-
mate Change (1997), Kyoto Protocol, Online:
http://unfccc.int/essential background/kyoto protocol/items/1678.php.
United States Environmental Protection Agency (2002), A comprehensive analysis
of biodiesel impacts on exhaust emissions, Tech. Rep. EPA420-P-02-001.
United States Environmental Protection Agency (2009), EPA proposes new regu-
lations for the national renewable fuel standard program for 2010 and beyond,
Regulatory Annoucement EPA-420-F-09-023, Office of Transportation and Air
Quality.
U.S. Department of Energy: Energy efficiency and renewable en-
ergy (2009), States with renewable portfolio standards, Online:
http://apps1.eere.energy.gov/states/maps/renewable portfolio states.cfm.
van der Werf, G., J. T. Randerson, G. J. Collatz, and L. Giglio (2003), Carbon emis-
sions from fires in tropical and subtropical ecosystems, Global Change Biol., 9,
547–562.
van der Werf, G. R., J. T. Randerson, G. J. Collatz, L. Giglio, P. S. Kasibhatla,
A. F. Arellano Jr., S. C. Olsen, and E. S. Kasischke (2004), Continental-scale
partitioning of fire emissions during the 1997 to 2001 El Nin˜o/La Nin˜a period,
Science, 303, 73–76.
van Dingenen, R., F. J. Dentener, F. Raes, M. C. Krol, L. Emberson, and J. Cofala
(2009), The global impact of ozone on agricultural crop yields under current and
future air quality legislation, Atmos. Environ., 43, 604–618.
Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo (1997a), Human
domination of earth’s ecosystems, Science, 277, 494–499.
Vitousek, P. M., J. D. Aber, R. W. Howarth, G. E. Likens, P. A. Matson, D. W.
217
Schindler, W. H. Schlesinger, and G. D. Tilman (1997b), Human alteration of
the global nitrogen cycle: Sources and consequences, Ecol. Appl., 7, 737–750.
Voldoire, A., B. Eickhout, M. Schaeffer, J. F. Royer, and F. Chauvin (2007), Cli-
mate simulation of the twenty-first century with interactive land-use changes,
Clim. Dynam., 29, 177–193.
von Kuhlmann, R., M. G. Lawrence, U. Po¨schl, and P. J. Crutzen (2004), Sensitiv-
ities in the global scale modelling of isoprene, Atmos. Chem. Phys., 4, 1–17.
Wang, K.-Y., and D. E. Shallcross (2000), Modelling terrestrial biogenic isoprene
fluxes and their potential impact on global chemical species using a coupled
LSM-CTM model, Atmos. Environ., 34, 2909–2925.
Wang, M. (2009), The Greenhouse Gases, Regulated Emissions,
and Energy Use in Transportation (GREET) Model, Online:
http://www.transportation.anl.gov/modeling simulation/GREET/.
Wang, M., M. W. Hong, and H. Huo (2007), Life-cycle energy and greenhouse
gas emission impacts of different corn ethanol plant types, Environ. Res. Let., 2,
024,001.
Wang, M., M. Wu, H. Huo, and J. H. Liu (2008), Life-cycle energy use and green-
house gas emission implications of Brazilian sugarcane ethanol simulated with
the GREET model, Int. Sugar J., 110, 527.
Wayne, R. P. (2000), Chemistry of Atmospheres, third ed., Oxford: Oxford Univer-
sity Press, New York.
Webster, S., A. R. Brown, D. R. Cameron, and C. P. Jones (2003), Improvements
to the representation of orography in the Met Office Unified Model, Q. J. Roy.
Meteor. Soc., 129, 1989–2010, doi:10.1256/qj.02.133.
Weng, J.-K., X. Li, N. D. Bonawitz, and C. Chapple (2008), Emerging strategies of
218
lignin engineering and degradation for cellulosic biofuel production, Curr. Opin.
Biotech., 19, 166–172.
Went, F. W. (1960), Organic matter in the atmosphere, and its possible relation to
petroleum formation, Proc. Natl. Acad. Sci. USA, 46, 212–221.
Wesely, M. L., and B. B. Hicks (2000), A review of the current status of knowledge
on dry deposition, Atmos. Environ., 34, 2261–2282.
Wiedinmyer, C., A. Guenther, P. Harley, C. Hewitt, C. Geron, P. Artaxo, R. Stein-
brecher, and R. Rasmussen (2004), Emissions of Atmospheric Trace Com-
pounds, chap. Global organic emissions from vegetation, pp. 115 –170, Dor-
drecht, The Netherlands: Kluwer Academic Publishers.
Wiedinmyer, C., X. Tie, A. Guenther, R. Neilson, and C. Granier (2006), Fu-
ture changes in biogenic isoprene emissions: How might they affect regional
and global atmospheric chemistry?, Earth Interactions, 10, Paper No. 3, doi:
10.1175/EI174.1.
Wild, O. (2007), Modelling the global tropospheric ozone budget: exploring the
variability in current models, Atmos. Chem. Phys., 7, 2643–2660.
Wild, O., and M. J. Prather (2000), Excitation of the primary tropospheric chem-
ical mode in a global three-dimensional model, J. Geophys. Res., 105, 24,647–
24,660.
Wilkins, C. K., P. A. Clausen, P. Wolkoff, S. Larsen, M. Hammer, K. Larsen,
V. Hansen, and G. D. Nielsen (2001), Formation of strong airway irritants
in mixtures of isoprene/ozone and isoprene/ozone/nitrogen dioxide, Environ.
Health Persp., 109, 937–941.
Wilkinson, M. J., R. K. Monson, N. Trahan, S. Lee, E. Brown, R. B. Jackson, H. W.
Polley, P. A. Fay, and R. Fall (2009), Leaf isoprene emission rate as a function
of atmospheric co2 concentration, Global Change Biol., 15, 1189–1200.
219
Williams, E. J., G. L. Hutchinson, and F. C. Fehsenseld (1992), NOx and N2O
emissions from soil, Global Biogeochem. Cycles, 6, 351–388.
Williams, R. (1995), Variants of low CO2-emitting supply system (LESS) for the
world: Prepared for the IPCC Second Assessment Report Working Group IIa,
Energy Supply Mitigation Options, Tech. Rep. PNL-IO851, Pacific Northwest
Laboratories.
Wilson, D., and S. Ballard (1999), A microphysically based precipitation scheme
for the UK Meteorological Office Unified Model, Q. J. Roy. Meteor. Soc., 125,
1607–1636, doi:10.1256/smsqj.55706.
Woodward, F. I., T. M. Smith, and W. R. Emanuel (1995), A global land primary
productivity and phytogeography model, Global Biogeochem. Cycles, 9, 471–
490.
World Meteorological Organization (Geneva, February 1979), The first world cli-
mate conference, 12-23 February 1979, Geneva, Switzerland, in Confrence mon-
diale sur le climat: Confrence d’experts sur le climat et l’homme.
Wu, M., M. Wang, and H. Huo (2006), Fuel-cycle assessment of selected
bioethanol production pathways in the united states, Tech. Rep. ANL/ESD/06-
7, Argonne National Laboratory, Energy Systems Division.
Wu, S., L. J. Mickley, D. J. Jacob, J. A. Logan, R. M. Yantosca, and D. Rind
(2007), Why are there large differences between models in global budgets of tro-
pospheric ozone?, J. Geophys. Res., 112, D05,302, doi:10.1029/2006JD007801.
Wu, S., L. J. Mickley, D. J. Jacob, D. Rind, and D. G. Streets (2008a), Effects of
2000–2050 changes in climate and emissions on global tropospheric ozone and
the policy-relevant background surface ozone in the United States, J. Geophys.
Res., 113, D18,312, doi:10.1029/2007JD009639.
220
Wu, S., L. J. Mickley, E. M. Leibensperger, D. J. Jacob, D. Rind, and D. G. Streets
(2008b), Effects of 2000–2050 global change on ozone air quality in the United
States, J. Geophys. Res., 113, D06,302, doi:10.1029/2007JD008917.
Ya´n˜ez Angarita, E. E., E. E. Silva Lora, R. E. da Costa, and E. A. Torres (2009),
The energy balance in the palm oil-derived methyl ester (pme) life cycle for the
cases in brazil and colombia, Renewable Energy, 34, 2905–2913.
Yamamoto, H., J. Fujino, and K. Yamaji (2001), Evaluation of bioenergy potential
with a multi-regional global-land-use-and-energy model, Biomass & Bioenergy,
21, 185–203.
Yienger, J. J., and H. Levy II (1995), Empirical model of global soil-biogenic NOx
emissions, J. Geophys. Res., 100, 11,477–11,464.
Young, P., A. Arneth, G. Schurgers, G. Zeng, and J. A. Pyle (2009), The CO2
inhibition of terrestrial isoprene emission significantly affects future ozone pro-
jections, Atmos. Chem. Phys., 9, 2793–2803.
Young, P. J. (2007), The influence of biogenic isoprene emissions of atmospheric
chemistry: A model study for present and future atmospheres, Ph.D. thesis, Uni-
versity of Cambridge.
Zeldovich, Y. (1946), The oxidation of nitrogen in combustion and explosions,
Acta. Physiochim., 21, 577–628.
Zeng, G., and J. A. Pyle (2003), Changes in tropospheric ozone between 2000
and 2100 modeled in a chemistry-climate model, Geophys. Res. Lett., 30, 1392,
doi:10.1029/2002GL016708.
Zeng, G., and J. A. Pyle (2005), Influence of El Nin˜o Southern Oscillation on
stratosphere/troposphere exchange and the global tropospheric ozone budget,
Geophys. Res. Lett., 32, L01814, doi:10.1029/2004GL021353.
221
Zeng, G., J. A. Pyle, and P. J. Young (2008), Impact of climate change on tropo-
spheric ozone and its global budgets, Atmos. Chem. Phys., 8, 369–387.
Zhang, D., R. Zhang, and S. W. North (2003a), Experimental study of NO reaction
with isoprene hydroxyalkyl peroxy radicals, J. Phys. Chem. A, 107, 11,013–
11,019.
Zhang, L., J. R. Brook, and R. Vet (2003b), A revised parametisation for gaseous
dry deposition in air-quality models, Atmos. Chem. Phys., 3, 2067–2082.
Zhang, Y., X.-M. Hu, L. R. Leung, and W. I. Gustafson Jr. (2008), Impacts of
regional climate change on biogenic emissions and air quality, J. Geophys. Res.,
113, D18,310, doi:10.1029/2008JD009965.
Zhang, Y., X.-Y. Wen, K. Wang, K. Vijayaraghavan, and M. Z. Jacobson (2009),
Probing into regional o3 and particulate matter pollution in the united states: 2.
an examination of formation mechanisms through a process analysis technique
and sensitivity study, J. Geophys. Res., 114, D22,305.
Ziemke, J. R., S. Chandra, B. N. Duncan, M. R. Schoeberl, O. Torres, M. R.
Damon, and P. K. Bhartia (2009), Recent biomass burning in the tropics and
related changes in tropospheric ozone, Geophys. Res. Lett., 36, L15,819, doi:
10.1029/2009GL039303.
Zimmer, W., R. Steinbrecher, C. Ko¨rner, and J.-P. Schnitzler (2003), The process-
based SIM-BIM model: Towards more realistic prediction of isoprene emission
from adult Quercus petraea forest trees, Atmos. Environ., 37, 1665–1671.
Zimmerman, P. R. (1979), Determination of emission rates of hydrocarbons from
indigenous species of vegetation in the Tampa/St. Petersburg, Florida area. Ap-
pendix C. Tampa Bay area photochemical oxidant study, Tech. rep., U.S. Envi-
ronmental Protection Agency, Region IV, Atlanta, Georgia.
222