ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-16-10621-2016Impacts of current and projected oil palm plantation expansion on air
quality over Southeast AsiaSilvaSam J.samsilva@mit.eduhttps://orcid.org/0000-0001-6343-8382HealdColette L.https://orcid.org/0000-0003-2894-5738GeddesJeffrey A.AustinKemen G.KasibhatlaPrasad S.https://orcid.org/0000-0003-3562-3737MarlierMiriam E.https://orcid.org/0000-0001-9333-8411Department of Civil and Environmental Engineering, Massachusetts
Institute of Technology, Cambridge, MA, USADepartment of Physics and Atmospheric Science, Dalhousie University,
Halifax, Nova Scotia, CanadaNicholas School of the Environment, Duke University, Durham, NC, USADepartment of Ecology, Evolution and Environmental Biology, Columbia
University, New York, NY, USASam J. Silva (samsilva@mit.edu)26August20161616106211063519February201626February201616July201620July2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://acp.copernicus.org/articles/16/10621/2016/acp-16-10621-2016.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/16/10621/2016/acp-16-10621-2016.pdf
Over recent decades oil palm plantations have rapidly expanded across
Southeast Asia (SEA). According to the United Nations, oil palm production in
SEA increased by a factor of 3 from 1995 to 2010. We investigate the impacts
of current (2010) and near-term future (2020) projected oil palm expansion in
SEA on surface–atmosphere exchange and the resulting air quality in the
region. For this purpose, we use satellite data, high-resolution land maps,
and the chemical transport model GEOS-Chem. Relative to a no oil palm
plantation scenario (∼ 1990), overall simulated isoprene emissions in
the region increased by 13 % due to oil palm plantations in 2010 and a
further 11 % in the near-term future. In addition, the expansion of palm
plantations leads to local increases in ozone deposition velocities of up to
20 %. The net result of these changes is that oil palm expansion in SEA
increases surface O3 by up to 3.5 ppbv over dense urban regions, and in
the near-term future could rise more than 4.5 ppbv above baseline levels.
Biogenic secondary organic aerosol loadings also increase by up to
1 µg m-3 due to oil palm expansion, and could increase by a
further 2.5 µg m-3 in the near-term future. Our analysis
indicates that while the impact of recent oil palm expansion on air quality
in the region has been significant, the retrieval error and sensitivity of
the current constellation of satellite measurements limit our ability to
observe these impacts from space. Oil palm expansion is likely to continue to
degrade air quality in the region in the coming decade and hinder efforts to
achieve air quality regulations in major urban areas such as Kuala Lumpur and
Singapore.
Introduction
Palm oil is currently the most popular source of food oil in the world, and
is rapidly gaining importance as a source of biofuel (Corley, 2009). Over
80 % of global palm oil production takes place in Southeast Asia (SEA),
with more than 10 million hectares of land farmed in 2005 (Fitzherbert et
al., 2008). According to the Food and Agriculture Organization of the United
Nations (FAO, 2015), palm oil production in SEA has grown by a factor of 5
over the past 20 years. By 2020, the total area occupied by oil palm
plantations in SEA is expected to increase even further (Austin et al., 2015;
Marlier et al., 2015b). This represents a significant and rapid change in
land use over a relatively small region. The palm expansion in SEA is just
one example of land use change that may drive changes in atmospheric
composition; however, to date there has been limited exploration of these
impacts (Heald and Spracklen, 2015).
The impact of oil palm expansion on global carbon stocks and biodiversity has
been studied extensively. Austin et al. (2015) suggest that up to
47 MtC yr-1 will be emitted due to oil palm land use change in
Kalimantan alone from 2010 to 2020. Carlson et al. (2013) estimate that the
plantations could generate over 25 % of Indonesia's projected carbon
emissions by 2020. Furthermore, replacing native forests with oil palm has
and will continue to threaten biodiversity across the region. Fitzherbert et
al. (2008) found that only 15 % of all species present in the natural
forest were also found in oil palm plantations.
This rapid land use change across SEA has also altered surface–atmosphere
exchange of trace gases over the region, with potential implications for
regional air quality. Airborne observations over northeastern Borneo during
the 2008 OP3 campaign indicated that oil palm plantations emit 7 times more
isoprene than the nearby rainforest (Hewitt et al., 2010). Isoprene is the
most abundant volatile organic compound (VOC) emitted by plants, is a
precursor of secondary organic aerosol (SOA), and can play an important role
in surface ozone formation, depending on the local chemical environment.
Despite this dramatic increase in isoprene emissions, no appreciable
difference in surface O3 over palm vs. native forest was observed
during OP3 (Hewitt et al., 2009). Growth of oil palm plantations also impacts
the deposition of gases and particles from the atmosphere due to changes in
the total amount of plant surface area available for deposition. The OP3
observations show that deposition velocities of ozone over oil palm are half
the magnitude of deposition velocities over the natural forest (Fowler et
al., 2011). Fowler et al. (2011) suggest that this is due to smaller
non-stomatal exchange of ozone from the oil palm canopy. Expansion of oil
palm may also impact soil NOx emissions due to fertilizer application,
changing land types, changes in the amount of NO that escapes through the
canopy, and changes in the amount of sunlight that reaches the ground (Hewitt
et al., 2009). The links between these rapid land use changes and air quality
become even more important in light of the tremendous population growth and
significant air quality problems already present throughout SEA. The region
contains more than 570 million people, an increase of more than 10 %
since 2000 (United Nations Population Division, 2011). Forouzanfar et
al. (2015) found that indoor and outdoor atmospheric pollution are both
within the top 10 leading risk factors for premature mortality throughout
SEA.
The impact of oil palm plantations on air quality has been investigated in
previous modeling studies. Warwick et al. (2013) simulate the influence of
modern-day oil palm distributions over Borneo, constraining both isoprene and
NO2 fluxes to values measured during the OP3 campaign, and converting
the entire island of Borneo to oil palm plantations. They find the potential
for up to 70 % (30–45 ppbv) increases in regional O3
concentrations. Ashworth et al. (2012) use the HadGEM2 model to assess the
impact of biofuel feedstocks on air quality. Part of their study included oil
palm plantations in SEA, where they scaled forest emissions to meet the
observed plantation emissions. They found that the oil palm-related increase
in isoprene emissions by 2020 could lead to local increases in surface
O3 and SOA of up to 3 ppbv and 0.4 µg m-3,
respectively.
Agricultural burning across SEA (partially related to palm oil production) is
an additional source of air pollution in the region. Marlier et al. (2015b)
assessed different land use scenarios across Sumatra to determine their
overall influence on current and future fire emissions. They found that the
scenario with the highest amount of oil palm had the largest associated fire
emissions (100 Tg DM yr-1); these fires contribute up to 60 % of
the total smoke concentrations across equatorial SEA.
In this study, we use the GEOS-Chem model to simulate the potential impacts
of oil palm on air quality broadly across SEA for current and future oil
palm expansion scenarios. We go beyond previous studies by explicitly
simulating the concurrent perturbations to biogenic emissions, soil NOx
emissions, and dry deposition, and exploring the net impacts on air quality
in the region. We compare these results with available satellite
observations to investigate whether the current constellation of satellite
instruments can detect the changes in air quality driven by rapid land use
change.
Model descriptionThe GEOS-Chem model
We use the global chemical transport model GEOS-Chem v9-02
(www.geos-chem.org) to investigate the changes in air quality
associated with oil palm plantations in SEA. The model is driven by
assimilated meteorology from the Goddard Earth Observing System (GEOS). We
use GEOS-5 meteorology for the year 2006 in all of our simulations, due in
part to the availability of updated anthropogenic emission inventories for
this year. For this analysis, we perform a series of nested simulations of
GEOS-Chem over the Asian domain (70–150∘ E,
10∘ S–55∘ N) at 0.5∘× 0.667∘
horizontal resolution with 47 vertical layers. Boundary conditions are
produced using the same version of the global model at
2∘× 2.5∘ horizontal resolution. The model was
initialized with a 1-year simulation at 2∘× 2.5∘
horizontal resolution, and then an additional 6-month simulation at
0.5∘× 0.667∘. A map of the Asian domain, and the
region of particular interest to this study is shown in Fig. 1.
The nested GEOS-Chem Asian domain simulated at 0.5 × 0.67
(top) and the particular region of interest explored in this work (bottom).
The location of major regions discussed in the text are also shown.
The GEOS-Chem oxidant-aerosol simulation includes
H2SO4-HNO3-NH3 aerosol thermodynamics (Park et al.,
2006; Pye et al., 2009) coupled to a detailed
HOx-NOx-VOC-O3-BrOx chemical mechanism (Bey et al.,
2001; Mao et al., 2013). Secondary organic aerosol (SOA) is produced from
the oxidation of biogenic hydrocarbons (isoprene, monoterpenes, and
sesquiterpenes), aromatics, and IVOCs and represented with a volatility
basis set approach (Pye et al., 2010; Pye and Seinfeld, 2010).
The global simulations are driven by anthropogenic emissions from the
Emissions Database for Global Atmospheric Research version 3 (EDGARv3),
including emissions from ship exhaust (Olivier and Berdowski, 2001). Over the
Asian region, a more recent anthropogenic emission inventory from the year
2006 is used (Streets et al., 2003, 2006). Geddes et al. (2015) show that the
long-term (1996–2012) trend in satellite-derived estimates of ground-level
NO2 concentrations over SEA is relatively small, indicating that
regional changes in anthropogenic emissions are not large. Global emissions
from aviation are based on the AEIC inventory (Stettler et al., 2011; Simone
et al., 2012). Biomass burning emissions for 2006 follow the GFED3 inventory
(van der Werf et al., 2010). We note that the GFED3 emissions indicate that
2006 was a higher than average fire year in the region. Additional
simulations using emissions and meteorology from 2007 and 2008 indicate that
the choice of model year does not substantially influence the results of this
work, spatial patterns are all consistent and concentration changes are all
less than 10 %.
The GEOS-Chem land use module developed by Geddes et al. (2016) is used to
drive surface–atmosphere exchange processes in the model. Land use is
described using 16 plant functional types (PFTs), with an associated monthly
leaf area index (LAI) per PFT. The baseline global PFTs and associated LAI
are from the year 2000 inputs to the Community Land Model v.4
(http://www.cgd.ucar.edu/tss/clm), which are based on satellite
observations (Lawrence et al., 2011). The dominant vegetation types in SEA
are plotted in Fig. 2. Biogenic emissions of VOCs (BVOCs) are calculated
online for each PFT following MEGANv2.1 (Guenther et al., 2012). These BVOC
emissions respond to temperature, available sunlight, leaf area index, leaf
age, and soil moisture. These responses are quantified as activity factors
which are applied to the basal emission factor to calculate emissions that
vary with meteorology and phenology. Dry deposition is calculated following
Wesely (1989). The depositional velocity is a function of the aerodynamic,
boundary layer, and canopy resistances, added in series. The aerodynamic
resistance depends on atmospheric stability and surface roughness height, the
boundary layer resistance is a function of the chemical species and
meteorology, and the canopy resistance varies with the chemical properties of
the deposited species and the land type. To calculate the canopy resistance,
land types are mapped from the 16 PFTs to the 11 depositional surfaces
described in Wesely (1989). For the calculation of the aerodynamic and canopy
resistances, we take into account the influence of LAI on the land
properties. We account for all depositional surface types within a grid cell
by preserving the fractional land cover of each grid box. Soil NOx
emissions are a function of soil moisture, temperature, available nitrogen,
and land use type and are calculated following Hudman et al. (2012). The land
types for these emissions are mapped from the PFTs to 24 biomes described in
Steinkamp and Lawrence (2011). Additionally, a canopy reduction factor, to
account for the loss of NOx within the canopy, is calculated as a function
of LAI and meteorological parameters.
Percentage of vegetated area occupied by dominant vegetation classes
in the No Palm scenario at the native 0.23∘× 0.31∘
resolution.
GEOS-Chem has previously been used to study air quality in SEA.
Trivitayanurak et al. (2012) use the nested model to better understand the
distribution and sources of atmospheric trace constituents over Asia. They
find that the model captures the vertical and spatial variability of trace
constituents such as CO, isoprene, and sulfate to within 30 % of
observations from several aircraft campaigns. They also note that the model
under predicts AOD as measured by MODIS, which they attribute to an
underestimate of local biogenic SOA. Fu et al. (2007) use the model in
combination with satellite measurements of formaldehyde to constrain
non-methane VOC emissions broadly across Asia from 1996 to 2001. They show that
observed and modeled HCHO are highly correlated in the region. Jiang et
al. (2015) compare the model and satellite estimates of CO, O3, and
NO2 in Asia, and conclude from these comparisons that the GEOS-Chem
simulation of tropospheric O3 is reliable within the Asian domain.
Description of SEA land use
To account for oil palm plantations, we add a new plant functional type to
the land module to calculate palm specific biogenic emissions, soil NOx
emissions, and dry deposition. The basal isoprene emission factor of oil palm
is set to the basal rate of 7.8 mg m-2 h-1 measured during OP3
(Misztal et al., 2011); this basal emission factor is modulated online by
local meteorology and phenology to estimate emissions. Previous studies
scaled modeled emissions of broadleaf evergreen trees (Ashworth et al.,
2012), or kept these emissions fixed to the measured rate during OP3 (Warwick
et al., 2013). The basal isoprene emission factor of the native forests
(considered to be broadleaf evergreen tropical trees) is reduced to
1.6 mg m-2 h-1 to match the OP3 measurements (Langford et al.,
2010), which is a factor of 4 lower than the emission factors for broadleaf
evergreen tropical trees within MEGANv2.1, and consistent with the
rainforests of Southeast Asia emitting less isoprene than South American and
African rainforests. This difference is likely due to a previous dearth of
measurements across the rainforests of SEA to constrain MEGANv2.1 (Guenther
et al., 2006, 2012). Observations suggest that oil palm is
not a significant emitter of monoterpenes and sesquiterpenes, but the
mechanisms behind this are not clear (Misztal et al., 2011). The observations
of Misztal et al. (2011) do indicate that oil palm plantations have higher
emissions of estragole and toluene. However their impact on atmospheric
chemistry and composition is not well constrained, and Guenther et al. (2012)
indicate that emissions of these species are relatively small, particularly
as compared to isoprene. In light of this, and the large uncertainties on the
estimates of basal emissions for these other biogenic compounds from natural
forests in the region, we modify only the emission factor of isoprene for oil
palm relative to background forests.
During the OP3 campaign Fowler et al. (2011) observed that relatively few of
the oil palm plantations were fertilized. When extrapolated across the whole
expanse of oil palm plantations, it was found that the average soil NOx
emission of plantations is similar to that of the background forest. In
addition, we have no information on the relative impacts of palm plantations
and natural tropical forests on nitrogen cycling in the soil. Therefore, the
biome type-specific soil NOx emissions parameters for oil palm are
identical to the tropical forest conditions in our simulations.
The leaf area index (LAI) of an oil palm plantation changes as the trees age.
Oil palms themselves live for more than a decade, and their leaves live for
600–700 days; this contributes to LAI values that can vary from 2 to 8 in
mature plants (Van Kraalingen et al., 1989). Oil palm LAI is set to 4.5 for
these simulations; this value is selected as an approximation of the average
over the lifespan of the plant (Van Kraalingen et al., 1989), and therefore a
likely average for the plantation as a whole. We explore the sensitivity of
our results to this specified LAI value in Sect. 3. For the Wesely
depositional scheme land types, oil palm is assumed to be most similar to the
native forest, as opposed to the “crop” land type. This assumption is made
simply because oil palm plantations are physically much closer to a tall tree
forest than a wheat field or cornfield. We do not attempt to model the
detailed growth, harvesting, and senescence history of the plantation (Fan et
al., 2015). Previous studies (Ashworth et al., 2012, and Warwick et al.,
2013) did not explicitly simulate the concomitant changes to deposition due
to oil palm expansion. They instead performed short sensitivity studies
wherein Warwick et al. (2013) doubled the modeled deposition velocities, and
Ashworth et al. (2012) scaled modeled deposition parameters (roughness
length and LAI) by observed canopy height and biomass density.
Several land use maps are used to describe the modern and future distribution
of oil palm over SEA (Fig. 3). For the modern-day scenario, we use a land use
map developed by Miettinen et al. (2012), which describes land use across
insular SEA (-10 to 10∘ N, 95 to 140∘ E) in 2010 on a
250 m grid, and includes a land classification for palm plantations. The
future expansion scenario maps are adapted from Marlier et al. (2015b) and
Austin et al. (2015). Marlier et al. (2015b) developed a variety of scenarios
at 1 km resolution to understand changes in fire emissions associated with
land use change in Sumatra. We use the “High oil palm” scenario to
represent a realistic upper limit on the 2020 distribution of Sumatran oil
palm. This map was reported originally as the probability that a given grid
box will contain oil palm in 2020. The probability for each grid box was
treated as the percent area covered by palm, and converted to fractional PFT
coverage at the 0.23∘× 0.31∘ resolution that is
input into the GEOS-Chem land module. This leads to an increase of 112 %
in total palm coverage in Sumatra from 2010 to 2020. Austin et al. (2015)
mapped the future oil palm distribution in Kalimantan (Indonesian Borneo)
using a logistic regression model at 250 m resolution to understand the
impact that oil palm could have on emissions of CO2. We constrained that
logistic regression model to a total of 3.6 Mha of oil palm expansion to
create a map for the 2020 distribution of oil palm in Kalimantan, following
the totals given by Austin et al. (2015). This leads to an increase of
108 % in total palm coverage over Kalimantan from 2010 to 2020. Note that
these two data sets cover only Sumatra and Kalimantan, which together
represent 54 % of palm production in the region in 2010, but much of the
available land for future expansion. From 2010 to 2020 oil palm production in
these two regions increase by 10.8 Mha. Our projections do not consider the
potential oil palm expansion in the rest of the SEA region. For both
scenarios, the oil palm plantations are added as a land type and fractional
coverage of all pre-existing vegetation classes from the base land map are
reduced accordingly. This differs from previous studies where Warwick et
al. (2013) assumed the entire island of Borneo was covered in oil palm
plantations, and Ashworth et al. (2012) did not consider any changes to the
underlying vegetation characteristics, instead they scaled isoprene emissions
from selected tropical broad-leaved trees. It is important to note that the
2020 distribution used here is the best estimation of a near-term future
wherein large increases in oil palm plantations continue to occur. The
distribution may not be an accurate prediction for the specific year 2020. It
is meant to represent a realistic near-term scenario, and for this reason we
refer to it from here on as the “future” distribution.
Percentage of vegetated area occupied by oil palm plantations for
Modern (2010) and Future scenarios. Note that estimates for palm plantation
increases from 2010 to the near-term future are only available for Sumatra
and Kalimantan; palm plantation coverage in other regions are assumed
constant.
Impacts of palm expansion on air quality in Southeast Asia
To understand the overall air quality impact of oil palm plantations in SEA,
we explore three major plantation scenarios. The first uses a land map with
no palm to establish a baseline. This simulation is referred to as “No
Palm” and is representative of conditions that pre-date the major palm
expansion in SEA (∼ 1990). The next is a simulation we call “Modern
Palm”, using the 2010 distribution from Miettinen et al. (2012). The third
simulation is “Future Palm,” and uses the merged land maps from Marlier et
al. (2015b) and Austin et al. (2015) for Sumatra and Kalimantan. In addition,
we perform sensitivity scenarios for the Modern Palm distribution to
disaggregate land use change-driven impacts on BVOC emissions from dry
deposition and soil NOx emissions; this is referred to as “BVOC-only”.
Results are shown here for annual means; seasonal differences are minor.
Changes in surface–atmosphere exchange
The direct influence of oil palm plantation expansion is on
surface–atmosphere exchange, most significantly BVOC emissions and dry
deposition.
The changes in BVOC emissions are as anticipated: where palm is added,
surface fluxes of BVOCs increase. By far the largest increase is that of
isoprene. There are increases in other BVOCs due to replacement of unforested
regions (wetlands, pastureland, etc.) with oil palm, but they are at least a
full order of magnitude smaller than the concomitant changes in isoprene.
Figure 4 shows the increase in isoprene emissions associated with the
addition of oil palm plantations in both the Modern and Future scenarios,
compared to the baseline No Palm scenario. The largest increases in isoprene
emissions in the Modern Palm scenario occur in northern Borneo, Sumatra, and
the southern Malay Peninsula. This addition of palm increases isoprene
emissions by a factor of three over northeastern Borneo, where the OP3
campaign took place (see Sect. 4). This growth corresponds to an absolute
increase on the order of 14 µmol m-2 h-1. The largest
relative changes in isoprene emissions occurred over the northern half of
Sumatra and the southern Malay Peninsula, with up to a 4.5 fold increase over
the simulation without oil palm. Similar to northeastern Borneo, the absolute
change in emissions is on the order of 10 µmol m-2 h-1.
Sumatra and the Malay Peninsula are of particular importance due to their
proximity to the large urban centers of Kuala Lumpur and Singapore. Oil palm
plantations in 2010 result in an additional 1.26 TgC yr-1 of isoprene
emission from SEA, a 13 % increase from the No Palm scenario.
Annual mean simulated isoprene emissions over SEA (top) and the
change due to Modern (middle) and Future (bottom) oil palm expansion.
The Future Palm emission scenario changes are limited to Sumatra and
Kalimantan (Indonesian Borneo) as prescribed in the land use change scenarios
considered (Sect. 2.2). Isoprene emissions in Kalimantan increase by a factor
of 2 (∼ 10.6 µmol m-2 h-1) from the No Palm
simulation. These are mostly regions that are as yet undeveloped and are good
candidates for future palm agriculture. There are many regions in Sumatra where
the changes in isoprene emissions are greater than a factor of 3–4, with an
absolute difference in excess of 14 µmol m-2 h-1.
Similar to the changes in the Modern Palm scenario, these isoprene emissions
increases are very near large urban regions. In this scenario, the total
change in isoprene emissions from 2010 to the near-term future scenario in
SEA is 1.1 TgC yr-1 across SEA, a 10 % increase in isoprene
emissions. This emission increase is nearly twice as high as previous work
(Ashworth et al., 2012), due to differences in the assumed future
distribution of oil palm.
The impacts of oil palm expansion on dry deposition velocities are more
complex than the BVOC emissions changes. As stated in Sect. 2.2, LAI
influences the resistance terms in the calculation of depositional
velocities, and oil palm plantations tend to have a higher LAI (prescribed
here to a fixed value of 4.5) than both the natural forest (ranges from 0 to
6.77, with a median value of 4 for SEA), and grasslands or previously cleared
agricultural lands. The LAI for the various scenarios is shown in Fig. 5. The
addition of oil palm plantations increases the LAI for much of SEA. For a
highly reactive species such as O3, an increase in LAI directly leads to
an increase in the depositional velocity over that surface, as seen in
Fig. 6. It should be noted that there are minor perturbations to the
deposition of other gas phase species as well. However, O3 deposition is
most sensitive to this land use change due to its high reactivity and strong
LAI dependence. The impact of oil palm expansion on particle deposition is
negligible.
Annual average leaf area index (LAI) over SEA (top) and the change
due to Modern (middle) and Future (bottom) palm expansion.
Modern Palm distributions increase
O3 dry deposition velocities most significantly in the Malay Peninsula
and North Sumatra. In both regions, the largest changes are increases of
0.05 cm s-1, or nearly 15 %. This change is the opposite sign and
smaller in magnitude than the measured difference in O3 deposition
velocity across the forest to palm transition reported by Fowler et
al. (2011); however our values are not directly comparable to those
measurements given the heterogeneity of land types within each grid cell, the
resolution of the model, and the fact that the simulated depositional
velocity changes are an aggregate of many land types transitioning to oil
palm plantations (not purely forest to palm). Figure 5 shows that the Malay
Peninsula and North Sumatra exhibit the largest changes in deposition
partially due to the dense palm plantations in the region, and also due to
the large changes in LAI due to those plantations replacing cleared land, low
LAI forests, and other varied land types. Across SEA, the overall impact of
palm plantations on ozone deposition is small, with a net 0.5 % increase
in O3 depositional velocity.
The future distribution of oil palm produces the largest changes in ozone dry
deposition velocities over Kalimantan, and again Sumatra (Fig. 6). In both
these regions, increases in deposition velocities are as large as
0.06 cm s-1, or 20 % relative to the No Palm scenario. As with the
Modern Palm scenario, the most significant changes occur over areas where oil
palm plantations replace cleared lands. The average change in ozone dry
deposition fluxes across the region from No Palm to the future scenario is an increase of 1.0 %.
Annual average ozone dry deposition velocity over SEA (top) and the
change due to Modern (middle) and Future (bottom)
Palm expansion.
In this work, the basal emission of soil NOx in oil palm plantations is
identical to that in natural forests. However the loss of NOx to the
canopy is impacted by differences in LAI. We find that there are negligible
changes in the net soil NOx emission over regions where the land
transitioned from forest to palm. This is largely consistent with the OP3
field campaign observations (Fowler et al., 2011), which indicated that few
of the palm plantations are fertilized. The soil NOx changes are more
significant where other land type changes occurred; for instance, moving
from pastureland to oil palm. However there remains substantial uncertainty
surrounding fertilization practices in SEA palm plantations. Therefore, this
simulated perturbation in soil NOx emissions is likely a lower limit.
Additional small changes in the re-emission of nitrogen occur as a feedback
related to changes in the total amount of nitrogen deposited from the
atmosphere.
LAI sensitivity tests indicate that assigning oil palm plantation LAI from 3
to 6, as opposed to the 4.5 used in our simulations, has modest impacts on
the changes in surface–atmosphere exchange over this region. Lower LAI
reduces isoprene emissions and dry deposition, and increases soil NOx
emissions broadly across oil palm regions by ∼ 5 % compared to the
Modern Palm simulation. The inverse is true of higher LAI values. The maximum
changes in these processes are of the order 10 % relative to the Modern Palm
simulation.
Changes in air quality
Atmospheric composition over SEA is impacted by oil palm plantation expansion
via the perturbations in surface–atmosphere exchange discussed in Sect. 3.1.
We focus here on how these changes connect to surface air quality in the
region.
Formaldehyde (HCHO) is an oxidation product of isoprene, a toxic pollutant,
and an O3 precursor. Figure 7 shows the sensitivity of simulated HCHO to
changes in the oil palm distribution. The largest increases in surface HCHO
due to Modern Palm are in regions where surface fluxes of isoprene change the
most, and are locally isolated due to the short atmospheric lifetime of HCHO
(∼ hours). The largest relative increases in HCHO (up to 70 %) are
seen over northeastern Borneo, while concentrations near the urban centers on
the Malay Peninsula show increases greater than 50 %. In terms of
absolute values, the largest changes occur over Sumatra, with surface values
increasing by as much as 2 ppbv. The increase over northeastern Borneo is
lower, at around 1.4 ppbv. Across SEA, mean surface HCHO increases by
1.6 %. HCHO sensitivities to Future Palm distributions share similar
spatial characteristics to the Modern Palm scenario, with more pronounced
changes over Kalimantan. Fractionally, the largest changes are in Sumatra,
where surface concentrations of HCHO increase by up to a factor of 1.8
compared to the No Palm baseline. Mean surface HCHO concentrations over SEA
increase by 2.8 % compared to the No Palm baseline. Absolute changes in
surface HCHO are still quite high over Sumatra, above 2.5 ppbv in many
regions.
Annual average surface formaldehyde (HCHO) concentrations over
SEA (top) and the change due to Modern (middle) and Future (bottom)
Palm expansion.
Figure 8 shows that the simulated response of surface NOx to the oil
palm expansion is very small. In principle, this response is influenced by
changes to deposition, soil NOx emissions, and isoprene fluxes. Given
the modest difference in deposition and soil NOx emissions, the dominant
impact is the elevated concentrations of isoprene. Additional isoprene leads
to more conversion of NO to NO2, and therefore increases the formation
of HNO3, leading to a net loss of NOx. This effect is only apparent
across the southern Malay Peninsula, a region with high surface NOx
concentrations, due in large part to significant anthropogenic activity.
These changes are typically less than 0.1 ppbv, on the order of 5 %
decreases. The Future oil palm simulation shows similar decreases in the
surface NOx response. These decreases are as large as 1 ppbv over
Sumatra, a 5 % drop. There is a decrease in NOx across Kalimantan on
the order of ∼ 0.5 ppbv related to the same chemistry. In reality,
these changes may be dwarfed by the impact of anthropogenic emissions of
NOx associated with production and processing facilities as well as oil
palm fertilization; these changes are highly uncertain, and have not been
included here.
Annual average surface nitrogen oxides (NOx) concentrations
over SEA (top) and the change due to Modern (middle) and Future (bottom)
Palm expansion.
The introduction of oil palm and the resulting increase in concentrations of
isoprene can lead to changes in concentrations of ozone through VOC-NOx
chemistry. At the same time, the increase in the deposition velocity of ozone
leads to a shorter average lifetime, which decreases concentrations. In our
modeled responses we see both of these signatures across SEA. Figure 9 shows
that the surface ozone response to Modern Palm is most prominent over the
southern Malay Peninsula (up to 4 ppbv), with changes over northeastern
Borneo and Sumatra as well. Over the Malay Peninsula and Sumatra, a region
not sampled during OP3, surface ozone concentrations increase by up to
26 % (3–4 ppbv) due to palm expansion. Ozone formation is enhanced in
these regions, where additional isoprene emissions combine with NOx rich
air near the major urban centers. Surface ozone increases in northeastern
Borneo are on the order of 2 ppbv, located in the near vicinity of the oil
palm plantations. These results differ spatially from Warwick et al. (2013),
likely due to the substantially different land maps used for oil palm
emissions of VOCs over Borneo. Hewitt et al. (2009) did not observe a change
in surface O3 concentrations due to oil palm at all over northeastern
Borneo. Much of the discrepancy between our results and the Hewitt et
al. (2009) observations can likely be explained by sampling and the different
spatial resolution of the measurements and the model. The
0.5∘× 0.666∘ grid box resolution used in this study
is on the order of the entire study region for OP3.
Annual average surface ozone concentrations over SEA (top) and the
change due to Modern (middle) and Future (bottom)
Palm expansion.
Adding oil palm plantations usually increases the LAI (Fig. 5), leading to an
increased depositional velocity (Fig. 6), which ultimately results in an
increased sink of O3. However, this is generally counteracted by the
large increase in isoprene emissions. Our BVOC-only sensitivity simulation
indicates that the changes in biogenic emissions are the dominant factor
controlling the changes in O3. The ratio of the changes in the Modern
Palm simulation to those in the BVOC-only simulation is shown in Fig. 10.
Across most of the region, this ratio is nearly 1, indicating that the
changes in soil NOx and deposition have little impact on O3. That
said, over regions that undergo large forest to palm transitions and have
relatively low background O3 concentrations, such as Indonesian Borneo,
up to 50 % of the O3 change is related to dry deposition and soil
NOx. A small decrease in the annual average O3 concentrations
(∼ 10 pptv) is simulated over southwestern Borneo, where under
low-NOx conditions, the additional isoprene from palm consumes O3.
The ratio of the changes in ozone concentrations in the Modern Palm
scenario compared to the changes in ozone in the BVOC-only simulation.
Regions less than 1 show where increasing deposition velocities over palm
plantations counteract some of the ozone increases driven by increasing
isoprene emissions.
The changes in surface ozone are exacerbated in the Future Palm scenario
(Fig. 9). The relative near-term future surface concentrations over the Malay
Peninsula increase by up to 4.5 ppbv (25 %) over the No Palm scenario.
This is larger than the impact (< 1 ppbv) estimated by Ashworth et
al. (2012), likely due to their more modest future estimate of palm
expansion. O3 concentrations in southern Kalimantan increase by up to
1 ppbv (∼ 5 %).
These changes have substantial impacts on local urban air quality in Kuala
Lumpur and Singapore, which aim to adhere to the World Health Organization
(WHO) (http://www.who.int/) guidelines for the daily maximum 8 h
average ozone not to exceed 50 ppbv. The number of days per year which
exceed this standard in our simulation are shown in Fig. 11 for both
Singapore and Kuala Lumpur. The No Palm simulation suggests that surface
O3 exceeds the WHO standard for 35 days in Singapore and 23 days in
Kuala Lumpur. We show how additional isoprene emissions associated with palm
expansion increases O3 concentrations in these urban regions,
exacerbating air quality issues. In particular, over Kuala Lumpur, Modern
Palm is associated with 33 additional days of O3 exceedance, increasing
to 62 total days in the Future Palm scenario. The impacts over Singapore are
more modest; nevertheless palm expansion is associated with 8 more days above
the WHO guideline levels in the future scenario. From Fig. 9, we observe that
ozone air quality in Jakarta, another large urban center in the region, is
relatively unaffected by oil palm expansion, due to the local transport
patterns and the spatial distribution of the plantations.
The number of days with simulated daily maximum 8 h average surface
ozone concentrations exceeding 50 ppbv over Singapore and Kuala Lumpur.
The changes in biogenic secondary organic aerosol (SOA) all track very
closely with HCHO and isoprene, due to the rapid formation of SOA from
biogenic precursors. These changes are shown in Fig. 12. Relative increases
in surface biogenic SOA concentrations due to Modern Palm expansion are
highest in northeastern Borneo and the southern Malay Peninsula, with
increases larger than 60 %, approximately 1 µg m-3.
Future Palm expansion may lead to
further substantial enhancements of SOA in the region, as high as
3.5 µg m-3 and generally at least 1.5 µg m-3
over regions with high oil palm density. Again, these values are larger than
those in Ashworth et al. (2012), who employ a more modest oil palm expansion
scenario. Palm expansion in the coming decade could lead to an average
5 % increase in surface SOA across the region, degrading visibility and
enhancing air pollution exposure. Though most of these changes are local,
they do stretch into protected nature preserves and dense urban regions.
Though these change to biogenic SOA are large, it is important to note that
biogenic SOA is not the dominant source of particulate matter pollution
across SEA. Marlier et al. (2013) show that regional fires can contribute to
annual average particulate matter concentrations of more than
100 µg m-3, several orders of magnitude higher than the
changes in SOA related to oil palm plantations.
Annual average surface biogenic secondary organic aerosol (SOA)
concentrations over SEA (top) and the change due to Modern (middle) and
Future (bottom) Palm expansion.
The changes in air quality are not as sensitive to the choice of palm LAI
(between 3–6), as compared to the changes in surface–atmosphere exchange
processes. The magnitude of the changes in HCHO, O3, and NOx are
all on the order of 1 %, and not more than 5 % relative to the Modern
Palm scenario. The relative changes in SOA are also generally small, but are
slightly higher (∼±8 %) over the southern Malay Peninsula,
near Singapore.
Limitations of observing systems for detecting the impacts of land use
change on air quality
There are no long-term surface measurements of air quality in SEA to assess
and validate our simulated impacts of oil palm expansion, beyond the
snapshots provided by the OP3 campaign discussed above. However, a suite of
space-based instruments has been making global measurements during the peak
of the palm expansion (2004–Present). The rapid and extensive expansion of
oil palm in SEA is arguably the most dramatic example of local land use
change during the satellite era for atmospheric composition. We investigate
whether the anticipated changes in air quality have been detectable from
space over the last decade. We focus on the record of HCHO, O3,
NO2, and aerosol optical depth (AOD) observations, the suite of
observable species that may have been impacted by oil palm development, and
we use our model simulation to direct this analysis.
All of the measurements analyzed here are part of the A-Train constellation
of polar-orbiting satellites. As such they provide daily coverage in a sun
synchronous orbit, with local overpasses ∼ 10:00 and ∼ 13:00. We
use HCHO, NO2, and O3 measurements from the OMI instrument on board
the NASA/Aura satellite, which has been operating since late 2004. The
satellite has a 14 × 24 km footprint size. HCHO observations are
from the NASA OMI HCHOv3 (OMHCHO) retrieval; filtering and quality control
screening is described in González Abad et al. (2015). These data have been
used previously to successfully analyze large urban source regions and assess
biogenic isoprene emissions (e.g. Zhu et al., 2014). We use the NO2
retrievals from the DOMINOv2 retrieval product, as described in Boersma et
al. (2011). These data have been used previously to assess emissions of
NO2 across Asia (Vinken et al., 2014). We use the NASA OMO3PR product
(Bak et al., 2015) that retrieves a vertical profile of ozone concentrations
in 18 layers extending from the surface up to 0.3 hPa. There are alternative
satellite measurements of all the above chemical species, but we focus on
this suite of measurements that provide a consistent record during the peak
palm expansion, aboard the same observing platform. We also explore aerosol
optical depth (AOD) measurements from two MODIS instruments on board the NASA
Terra and Aqua satellites. Our analysis uses the MODIS collection 6 product
(Sayer et al., 2014).
All satellite data are filtered spatially to best capture the signal of palm
expansion against the significant background of other sources. These include
a large urban signature from the Kuala Lumpur, Singapore, and Jakarta
megacities and substantial fire activity in both Sumatra and Indonesian
Borneo. To address this, we focused our analysis on the remote northwest
corner of Borneo, shown in the map of Fig. 13. This contains the region where
the airborne measurements were made during the OP3 campaign (Hewitt et al.,
2010). Our model simulations suggest that this region has exhibited
significant changes in air quality due to oil palm expansion, particularly in
HCHO. Using Miettinen et al. (2010) as a base map, satellite measurements
over potential palm plantation land types are identified within the “Palm”
region, non-urban forest land-types in the “Forest” region, and ocean in
the “Ocean” region. This classification allows for various time series
analyses to be performed over what should be three distinct emission and
depositional land types, and thus best isolate the impact of oil palm on
local air quality.
The three regions in northeastern Borneo used for the spatial
filtering of the satellite data.
Of all the atmospheric constituents observed, changes in HCHO concentrations
should represent the strongest air quality perturbation due to oil palm
expansion according to our model simulation (Fig. 7). This is due to the
strong local source of HCHO through oxidation of isoprene from oil palm,
combined with the small urban sources in the region and the relatively
constant background HCHO from methane oxidation. Background HCHO
concentrations in our model simulation are on the order of
1016 molecules cm-2. The background concentrations observed by
the OMI instrument are also of the order 1016 molecules cm-2.
From our model simulations, the expected change in column concentration of
HCHO due to Modern Palm expansion is of order
1015 molecules cm-2. OMI is most sensitive to regions with very
high HCHO concentrations (>2×1016 molecules cm-2).
Even at peak sensitivity, the instrument has an error of 30 % per
retrieval. This error compounds to more than 100 % over areas with lower
signals. The anticipated changes in HCHO due to oil palm expansion are
therefore near the detection limits of the OMI instrument. Furthermore, the
OMI sensor has been slowly degrading with time, causing a significant drop in
data density since mid 2007 (González Abad et al., 2015) due to an issue
known as the “row anomaly”. Instrument degradation is apparent in plots of
annual average OMI HCHO, shown in Fig. 14. Figure 14 suggests that HCHO
concentrations have increased over SEA from 2005 through 2014, but that this
is consistent with an overall increase in background HCHO that is not limited
to SEA. This issue with instrument sensitivity, combined with a large
decrease in the number of available observations makes it difficult to
identify a significant trend in the HCHO satellite record. Figure 15 shows
the monthly mean HCHO columns from OMI across all three selected land
regions, as well as a LOWESS fit of the data with the shaded regions
representing the 95 % confidence interval obtained through a LOWESS block
bootstrapping scheme (Cleveland, 1979). The measured HCHO over the forest and
palm regions are both generally higher than over the ocean. The LOWESS
analysis shows that HCHO concentrations are highest over the palm region,
with a mean difference of ∼0.6×1015 molecules cm-2,
slightly lower than but of similar magnitude to the difference expected from
our model simulations. This supports the results of our simulations, but also
demonstrates how challenging it is to identify this signal from satellite
observations. While the LOWESS analysis also suggests that the OMI HCHO
column concentrations across the region increased over this time period, much
of this may be driven by the row anomaly and we do not see evidence for
significant palm plantation growth within our limited palm region selected in
Fig. 13.
Annual average HCHO columns measured by the OMI instrument in 2005
and 2014.
There is large uncertainty with regard to the distribution and application of
fertilizer on oil palm plantations (Kusin et al., 2015; Fowler et al., 2011)
and industrial emissions of NOx associated with palm processing
facilities (Hewitt et al., 2009). OMI NO2 tropospheric retrievals are
shown in Fig. 16. Similar to the HCHO retrievals, both monthly mean columns
and the LOWESS fit with 95 % confidence intervals are shown. The NO2
columns above the palm region are generally higher relative to the forest
region, but again there is no increasing difference over time. This is
consistent with a lack of significant growth in fertilizer application or
industrial emissions over the oil palm region selected in Fig. 13. The
constant palm forest difference of 1014 molecules cm-2
(∼ 25 %) agrees with our modeled analysis, which shows surface
concentrations that differ by ∼ 29 % in this region, related to
elevated anthropogenic NOx emissions. This suggests that we are not
missing major palm-related sources of NOx emissions (fertilizer or
industrial processing) in our simulation. The constant increase in retrieved
NO2 concentrations over all three regions is consistent with Geddes et
al. (2015), who show a broad increasing trend in NO2 across all of
northern Borneo, possibly due to warmer surface temperatures, and transport
from urban regions.
The satellite-derived signal in tropospheric ozone from oil palm development
near Kuala Lumpur is not apparent against the background of urban
development. Even though there are significant changes in the local ozone
concentrations, too many confounding sources exist to identify the oil palm
signal. Fires in SEA dominate the measured AOD, with an additional
contribution from urban sources (Cohen and Lecoeur, 2015). Since the AOD
measurement is a net observation of extinction from all aerosols at all
altitudes, detecting changes in surface-level SOA is not straightforward. It
is therefore challenging to identify the impact of oil palm expansion on air
quality in SEA with the current constellation of polar-orbiting satellites.
In light of this result, it is important to consider the monitoring
capabilities of future observing systems, such as the North American
geostationary mission TEMPO (Chance et al., 2013). Geostationary observations
offer more frequent sampling of the diurnal cycle, which may enable a
separation of source signatures (e.g. urban, fire, biogenic, etc.) and a
better identification of perturbations associated with land use change. In
terms of HCHO measurements, TEMPO has a similar precision to that of the OMI
sensor (1016 molecules cm-2). However, TEMPO will have much
smaller spatial footprint (8×4.5 km) as compared to OMI (14×24 km), and will sample HCHO three times daily, as opposed to the once
daily measurements from OMI. The combination of these factors will likely
make changes in the emissions of HCHO and its precursors more detectable, if
they occur within the geostationary viewing field. Our results indicate that
the instrument precision and footprint size are the most important limiting
factor in detection of the perturbation associated with oil palm plantations
with the current suite of satellite observations.
A monthly mean and LOWESS time series of HCHO from OMI in the three
separate regions of northern Borneo (see Fig. 13). The LOWESS shaded regions
represent the 95 % confidence interval.
Conclusions
In this study, we simulate the impact of recent and near-term projected oil
palm expansion across SEA on air quality. We go beyond previous work by
consistently treating the impact of land use change on a suite of
land–atmosphere exchange processes relevant to atmospheric chemistry. Our
simulations suggest that oil palm plantation expansion in the region has had
a significant impact on air quality. As oil palm expansion continues, the
potential impact on surface O3 concentrations is significant. The
predicted ozone changes are largely due to increasing isoprene emissions.
Locally however, increases in depositional velocities counteract these
elevated emissions. If the oil palm crop expansion continues unabated,
near-term future ozone concentrations in urban regions could be up to
30 % higher (compared to the No Palm scenario) due to the plantations alone. Exposure to ozone is a
significant cause of premature mortality, responsible for more than 200 000
deaths globally in 2013 (Forouzanfar et al., 2015). The increase in ozone
attributable to oil palm plantations has the potential to bring many regions
in SEA (including the dense urban areas of Singapore and Kuala Lumpur)
further above WHO recommended threshold concentrations. The increase in
particulate matter due to biogenic secondary organic aerosol formation
presents an additional health concern. Over Singapore, our results indicate
that the addition of oil palm plantations contributes 9 % of the target
WHO recommended ozone concentrations, and 4 % of the recommended 24 h
particulate matter concentrations. This is on the same order as the seasonal
average predicted contribution of fires to the same particulate matter air
quality targets (∼ 8 %) (Marlier et al., 2015a). This work
illustrates that in trying to reach local air quality objectives, it is
important to consider the impacts of local land use change.
A monthly mean and LOWESS time series of NO2 from OMI in the
three separate regions of northern Borneo (see Fig. 13). The LOWESS shaded
regions represent the 95 % confidence interval.
In this study we do not include the potential air quality impacts associated
with local oil palm processing plants and prescribed burning of the fields.
These are likely to lead to additional impacts on regional air quality, some
of which have been described by Austin et al. (2015), Marlier et al. (2015a),
and Hewitt et al. (2009). We also have limited constraints on how
depositional fluxes are altered by land use transitions, and the
fertilization practices for oil palm. However our simulation of NOx and
ozone deposition in the region appears to be broadly consistent within the
range measurements from the OP3 campaign.
Though the rapid oil palm expansion in SEA has led to substantial changes in
the concentration of many atmospheric species, including HCHO, O3, and
aerosols, many of these changes occur in areas with high fire and urban
activity. Because of this, the signal of oil palm impacts on air quality is
difficult to disentangle from the satellite record. This issue of strong
confounding sources is compounded with issues of instrument sensitivity and
degradation over time. This study suggests that future observing systems will
require better sensitivities and stability to capture the impacts of land use
change on air quality.
This work is an example of the impact of significant land use conversion on
the regional scale. The growing pressures on the global food supply are
likely to lead to further land conversions to support agricultural activity
in the coming decades. Future atmospheric composition will respond to these
changes, with implications for air quality and climate, and will remain
important to monitor and understand.
Data availability
The model data used in this study are archived at MIT and are available from
the authors upon request (samsilva@mit.edu). The FAO agricultural statistics
are available at http://faostat3.fao.org/home/E (FAO, 2015). The OMI
HCHO (González Abad et al., 2015), OMI NO2 (Boersma et al., 2011),
OMI Ozone Profile (Bak et al., 2015), and the MODIS AOD (Sayer et al., 2014)
satellite observations used in this study are made available by NASA at
http://disc.sci.gsfc.nasa.gov/.
Acknowledgements
This work was supported by NSF (AGS-1238109).
Edited by: A. B. Guenther Reviewed by: three anonymous
referees
ReferencesAshworth, K., Folberth, G., Hewitt, C. N., and Wild, O.: Impacts of
near-future cultivation of biofuel feedstocks on atmospheric composition and
local air quality, Atmos. Chem. Phys., 12, 919–939,
10.5194/acp-12-919-2012, 2012.Austin, K. G., Kasibhatla, P. S., Urban, D. L., Stolle, F., and Vincent, J.:
Reconciling Oil Palm Expansion and Climate Change Mitigation in Kalimantan,
Indonesia, PLoS ONE, 10, e0127963, 10.1371/journal.pone.0127963, 2015.Bey, I., Jacob, D. J., Yantosca, R. M., Logan, J. A., Field, B. D., Fiore, A.
M., Li, Q., Liu, H. Y., Mickley, L. J., and Schultz, M. G.: Global Modeling
of Tropospheric Chemistry with Assimilated Meteorology: Model Description and
Evaluation, J. Geophys. Res.-Atmos., 106, 23073–23095,
10.1029/2001JD000807, 2001.Bak, J., Liu, X., Kim, J. H., Chance, K., and Haffner, D. P.: Validation of
OMI total ozone retrievals from the SAO ozone profile algorithm and three
operational algorithms with Brewer measurements, Atmos. Chem. Phys., 15,
667–683, 10.5194/acp-15-667-2015, 2015.Boersma, K. F., Eskes, H. J., Dirksen, R. J., van der A, R. J., Veefkind, J.
P., Stammes, P., Huijnen, V., Kleipool, Q. L., Sneep, M., Claas, J.,
Leitão, J., Richter, A., Zhou, Y., and Brunner, D.: An improved
tropospheric NO2 column retrieval algorithm for the Ozone Monitoring
Instrument, Atmos. Meas. Tech., 4, 1905–1928, 10.5194/amt-4-1905-2011,
2011.Carlson, K. M., Curran, L. M., Asner, G. P., McDonald Pittman, A., Trigg, S.
N., and Adeney, J. M.: Carbon Emissions from Forest Conversion by Kalimantan
Oil Palm Plantations, Nature Climate Change, 3, 283–287,
10.1038/nclimate1702, 2013.Chance, K., Xiong, L., Suleiman, R. M., Flittner, D. E., Al-Saadi, J., and
Janz, S. J.: Tropospheric emissions: Monitoring of pollution (TEMPO),
Proceedings of SPIE, Vol. 8866, Earth Observing Systems XVIII, 88660D
(23 September 2013), San Diego, CA, USA, available at:
https://www.cfa.harvard.edu/atmosphere/publications/TEMPO-SPIE-2013-24aug2013.pdf
(last access: 16 May 2016), 2013.
Cleveland, W. S.: Robust Locally Weighted Regression and Smoothing
Scatterplots, J. Am. Stat. Assoc., 74, 829–36, 1979.Cohen, J. B. and Lecoeur, E.: Decadal-scale relationship between measurements
of aerosols, land-use change, and fire over Southeast Asia, Atmos. Chem.
Phys. Discuss., 15, 26895–26957, 10.5194/acpd-15-26895-2015, 2015.Corley, R. H. V.: How Much Palm Oil Do We Need?, Environ. Sci. Policy, 12,
134–139, 10.1016/j.envsci.2008.10.011, 2009.Fan, Y., Roupsard, O., Bernoux, M., Le Maire, G., Panferov, O., Kotowska, M.
M., and Knohl, A.: A sub-canopy structure for simulating oil palm in the
Community Land Model (CLM-Palm): phenology, allocation and yield, Geosci.
Model Dev., 8, 3785–3800, 10.5194/gmd-8-3785-2015, 2015.FAO: Livestock, Crop and Fertilizer data, Food and Agricultural Organization
of the United Nations: Statistics Division, available at:
http://faostat3.fao.org/home/E, last access: October 2015.Fitzherbert, E. B., Struebig, M. J., Morel, A., Danielsen, F., Brühl, C.
A., Donald, P. F., and Phalan, B.: How Will Oil Palm Expansion Affect
Biodiversity?, Trend. Ecol. Evol., 23, 538–545,
10.1016/j.tree.2008.06.012, 2008.Forouzanfar, M. H., Alexander, L., Anderson, H. R., Bachman, V. F., Biryukov,
S., Brauer, M., Burnett, R., et al.: Global, Regional, and National
Comparative Risk Assessment of 79 Behavioural, Environmental and
Occupational, and Metabolic Risks or Clusters of Risks in 188 Countries,
1990–2013: A Systematic Analysis for the Global Burden of Disease Study
2013, The Lancet, 386, 2287–2323, 10.1016/S0140-6736(15)00128-2, 2015.Fowler, D., Nemitz, E., Misztal, P., Di Marco, C., Skiba, U., Ryder, J.,
Helfter, C., Cape, J. N., Owen, S., Dorsey, J., Gallagher, M. W., Coyle, M.,
Phillips, G., Davison, B., Langford, B., MacKenzie, R., Muller, J., Siong,
J., Dari-Salisburgo, C., Di Carlo, P., Aruffo, E., Giammaria, F., Pyle, J.
A., and Hewitt, C. N.: Effects of Land Use on Surface-Atmosphere Exchanges of
Trace Gases and Energy in Borneo: Comparing Fluxes over Oil Palm Plantations
and a Rainforest, Philos. T. R. Soc. B, 366, 3196–3209,
10.1098/rstb.2011.0055, 2011.Fu, T.-M., Jacob, D. J., Palmer, P. I., Chance, K., Wang, Y. X., Barletta,
B., Blake, D. R., Stanton, J. C., and Pilling, M. J.: Space-Based
Formaldehyde Measurements as Constraints on Volatile Organic Compound
Emissions in East and South Asia and Implications for Ozone, J. Geophys.
Res.-Atmos., 112, D06312, 10.1029/2006JD007853, 2007.Geddes, J. A., Martin, R. V., Boys, B. L., and van Donkelaar, A.: Long term
trends worldwide in ambient NO2 concentrations inferred from satellite
observations, Environ. Health Persp., 124, 281–289, 10.1289/ehp.1409567,
2015.Geddes, J. A., Heald, C. L., Silva, S. J., and Martin, R. V.: Land cover
change impacts on atmospheric chemistry: simulating projected large-scale
tree mortality in the United States, Atmos. Chem. Phys., 16, 2323–2340,
10.5194/acp-16-2323-2016, 2016.González Abad, G., Liu, X., Chance, K., Wang, H., Kurosu, T. P., and
Suleiman, R.: Updated Smithsonian Astrophysical Observatory Ozone Monitoring
Instrument (SAO OMI) formaldehyde retrieval, Atmos. Meas. Tech., 8, 19–32,
10.5194/amt-8-19-2015, 2015.Guenther, A., Karl, T., Harley, P., Wiedinmyer, C., Palmer, P. I., and Geron,
C.: Estimates of global terrestrial isoprene emissions using MEGAN (Model of
Emissions of Gases and Aerosols from Nature), Atmos. Chem. Phys., 6,
3181–3210, 10.5194/acp-6-3181-2006, 2006.Guenther, A. B., Jiang, X., Heald, C. L., Sakulyanontvittaya, T., Duhl, T.,
Emmons, L. K., and Wang, X.: The Model of Emissions of Gases and Aerosols
from Nature version 2.1 (MEGAN2.1): an extended and updated framework for
modeling biogenic emissions, Geosci. Model Dev., 5, 1471–1492,
10.5194/gmd-5-1471-2012, 2012.Heald, C. L. and Spracklen, D. V.: Land Use Change Impacts on Air Quality and
Climate, Chem. Rev. 115, 4476–4496. 10.1021/cr500446g, 2015.
Hewitt, C. N., MacKenzie, A. R., Di Carlo, P., Di Marco, C. F., Dorsey, J.
R., Evans, M., Fowler, D., Gallagher, M. W., Hopkins, J. R., Jones, C. E.,
Langford, B., Lee, J. D., Lewis, A. C., Lim, S. F., McQuaid, J., Misztal, P.,
Moller, S. J., Monks, P. S., Nemitz, E., Oram, D. E., Owen, S. M., Phillips,
G. J., Pugh, T. A. M., Pyle, J. A., Reeves, C. E., Ryder, J., Siong, J.,
Skiba, U., and Stewart, D. J.: Nitrogen Management Is Essential to Prevent
Tropical Oil Palm Plantations from Causing Ground-Level Ozone Pollution, P.
Natl. Acad. Sci., 106, 18447–18451, 2009.Hewitt, C. N., Lee, J. D., MacKenzie, A. R., Barkley, M. P., Carslaw, N.,
Carver, G. D., Chappell, N. A., Coe, H., Collier, C., Commane, R., Davies,
F., Davison, B., DiCarlo, P., Di Marco, C. F., Dorsey, J. R., Edwards, P. M.,
Evans, M. J., Fowler, D., Furneaux, K. L., Gallagher, M., Guenther, A.,
Heard, D. E., Helfter, C., Hopkins, J., Ingham, T., Irwin, M., Jones, C.,
Karunaharan, A., Langford, B., Lewis, A. C., Lim, S. F., MacDonald, S. M.,
Mahajan, A. S., Malpass, S., McFiggans, G., Mills, G., Misztal, P., Moller,
S., Monks, P. S., Nemitz, E., Nicolas-Perea, V., Oetjen, H., Oram, D. E.,
Palmer, P. I., Phillips, G. J., Pike, R., Plane, J. M. C., Pugh, T., Pyle, J.
A., Reeves, C. E., Robinson, N. H., Stewart, D., Stone, D., Whalley, L. K.,
and Yin, X.: Overview: oxidant and particle photochemical processes above a
south-east Asian tropical rainforest (the OP3 project): introduction,
rationale, location characteristics and tools, Atmos. Chem. Phys., 10,
169–199, 10.5194/acp-10-169-2010, 2010.Hudman, R. C., Moore, N. E., Mebust, A. K., Martin, R. V., Russell, A. R.,
Valin, L. C., and Cohen, R. C.: Steps towards a mechanistic model of global
soil nitric oxide emissions: implementation and space based-constraints,
Atmos. Chem. Phys., 12, 7779–7795, 10.5194/acp-12-7779-2012, 2012.Jiang, Z., Worden, J. R., Jones, D. B. A., Lin, J.-T., Verstraeten, W. W.,
and Henze, D. K.: Constraints on Asian ozone using Aura TES, OMI and Terra
MOPITT, Atmos. Chem. Phys., 15, 99–112, 10.5194/acp-15-99-2015, 2015.
Kusin, F. M., Akhir, N. I. M., Mohamat-Yusuff, F., and Awang, M.: The Impact
of Nitrogen Fertilizer Use on Greenhouse Gas Emissions in an Oil Palm
Plantation Associated with Land Use Change, Atmósfera, 28, 243–250,
2015.Langford, B., Misztal, P. K., Nemitz, E., Davison, B., Helfter, C., Pugh, T.
A. M., MacKenzie, A. R., Lim, S. F., and Hewitt, C. N.: Fluxes and
concentrations of volatile organic compounds from a South-East Asian tropical
rainforest, Atmos. Chem. Phys., 10, 8391–8412, 10.5194/acp-10-8391-2010,
2010.Lawrence, D. M., Oleson, K. W., Flanner, M. G., Thornton, P. E., Swenson, S.
C., Lawrence, P. J., Zeng, X., Yang, Z.-L., Levis, S., Sakaguchi, K., Bonan,
G. B., and Slater, A. G.: Parameterization Improvements and Functional and
Structural Advances in Version 4 of the Community Land Model, J. Adv. Model.
Earth Syst., 3, M03001, 10.1029/2011MS000045, 2011.Mao, J., Paulot, F., Jacob, D. J., Cohen, R. C., Crounse, J. D., Wennberg, P.
O., Keller, C. A., Hudman, R. C., Barkley, M. P., and Horowitz, L. W.: Ozone
and Organic Nitrates over the Eastern United States: Sensitivity to Isoprene
Chemistry, J. Geophys. Res.-Atmos., 118, 11256–11268,
10.1002/jgrd.50817, 2013.Marlier, M. E., DeFries, R. S., Voulgarakis, A., Kinney, P. L., Randerson, J.
T., Shindell, D. T., Chen, Y., and Faluvegi, G.: El Nino and Health Risks
from Landscape Fire Emissions in Southeast Asia, Nature Climate Change, 3,
131–136, 10.1038/nclimate1658, 2013.Marlier, M. E., DeFries, R. S., Kim, P. S., Gaveau, D. L. A., Koplitz, S. N.,
Jacob, D. J., Mickley, L. J., Margono, B. A., and Myers, S. S.: Regional Air
Quality Impacts of Future Fire Emissions in Sumatra and Kalimantan, Environ.
Res. Lett., 10, 054010, 10.1088/1748-9326/10/5/054010, 2015a.Marlier, M. E., DeFries, R., Pennington, D., Nelson, E., Ordway, E. M.,
Lewis, J., Koplitz, S. N., and Mickley, L. J.: Future Fire Emissions
Associated with Projected Land Use Change in Sumatra, Glob. Change Biol., 21,
345–362, 10.1111/gcb.12691, 2015b.Miettinen, J., Shi, C., Tan, W. J., and Liew, S. C.: 2010 Land Cover Map of
Insular Southeast Asia in 250-M Spatial Resolution, Remote Sens. Lett., 3,
11–20, 10.1080/01431161.2010.526971, 2012.Misztal, P. K., Nemitz, E., Langford, B., Di Marco, C. F., Phillips, G. J.,
Hewitt, C. N., MacKenzie, A. R., Owen, S. M., Fowler, D., Heal, M. R., and
Cape, J. N.: Direct ecosystem fluxes of volatile organic compounds from oil
palms in South-East Asia, Atmos. Chem. Phys., 11, 8995–9017,
10.5194/acp-11-8995-2011, 2011.
Olivier, J. G. J. and Berdowski, J. J. M.: Global emissions sources and
sinks, in: The Climate System, edited by: Berdowski, J., Guicherit, R., and
Heij, B. J., 33–78, A. A. Balkema Publishers/Swets & Zeitlinger
Publishers, Lisse, the Netherlands, 2001.Park, R. J., Jacob, D. J., Kumar, N., and Yantosca, R. M.: Regional
Visibility Statistics in the United States: Natural and Transboundary
Pollution Influences, and Implications for the Regional Haze Rule, Atmos.
Environ., 40, 5405–5423, 10.1016/j.atmosenv.2006.04.059, 2006.Pye, H. O. T. and Seinfeld, J. H.: A global perspective on aerosol from
low-volatility organic compounds, Atmos. Chem. Phys., 10, 4377–4401,
10.5194/acp-10-4377-2010, 2010.Pye, H. O. T., Liao, H., Wu, S., Mickley, L. J., Jacob, D. J., Henze, D. K.,
and Seinfeld, J. H.: Effect of changes in climate and emissions on future
sulfate-nitrate-ammonium aerosol levels in the United States, J. Geophys.
Res.-Atmos., 114, D01205, 10.1029/2008JD010701, 2009.Pye, H. O. T., Chan, A. W. H., Barkley, M. P., and Seinfeld, J. H.: Global
modeling of organic aerosol: the importance of reactive nitrogen (NOx and
NO3), Atmos. Chem. Phys., 10, 11261–11276, 10.5194/acp-10-11261-2010,
2010.Sayer, A. M., Munchak, L. A., Hsu, N. C., Levy, R. C., Bettenhausen, C., and
Jeong, M.-J.: MODIS Collection 6 Aerosol Products: Comparison between Aqua's
E-Deep Blue, Dark Target, and “merged” Data Sets, and Usage
Recommendations, J. Geophys. Res.-Atmos., 119, 13965–13989,
10.1002/2014JD022453, 2014.Steinkamp, J. and Lawrence, M. G.: Improvement and evaluation of simulated
global biogenic soil NO emissions in an AC-GCM, Atmos. Chem. Phys., 11,
6063–6082, 10.5194/acp-11-6063-2011, 2011.
Simone, N. W., Stettler, M. E. J., and Barrett, S. R. H.: Rapid estimation of
global civil aviation emissions with uncertainty quantification, Transport.
Res. D-Tr. E., 25, 33–41, 2012.
Stettler, M. E. J., Eastham, S., Barrett, S. R. H.: Air quality and public
health impacts of UK airports. Part I: Emissions, Atmos. Environ., 45,
5415–5424, 2011.Streets, D. G., Bond, T. C., Carmichael, G. R., Fernandes, S. D., Fu, Q.,
Klimont, Z., Nelson, S. M., Tsai, N. Y., Wang, M. Q., Woo, J.-H., and Yarber,
K. F.: An inventory of gaseous and primary aerosol emissions in Asia in the
year 2000, J. Geophys. Res., 108, 8809, 10.1029/2002JD003093, 2003.Streets, D. G., Zhang, Q., Wang, L., He, K., Hao, J., Wu, Y., Tang, Y., and
Carmichael, G. C.: Revisiting China's CO emissions after the Transport and
Chemical Evolution over the Pacific (TRACE-P) mission: Synthesis of
inventories, atmospheric modeling, and observations, J. Geophys. Res., 111,
D14306, 10.1029/2006JD007118, 2006.Trivitayanurak, W., Palmer, P. I., Barkley, M. P., Robinson, N. H., Coe, H.,
and Oram, D. E.: The composition and variability of atmospheric aerosol over
Southeast Asia during 2008, Atmos. Chem. Phys., 12, 1083–1100,
10.5194/acp-12-1083-2012, 2012.United Nations, Department of Economic and Social Affairs, Population
Division: World Population Prospects: The 2010 Revision, Volume I:
Comprehensive Tables, ST/ESA/SER.A/313, 2011.
van der Werf, G. R., Randerson, J. T., Giglio, L., Collatz, G. J., Mu, M.,
Kasibhatla, P. S., Morton, D. C., DeFries, R. S., Jin, Y., and van Leeuwen,
T. T.: Global fire emissions and the contribution of deforestation, savanna,
forest, agricultural, and peat fires (1997–2009), Atmos. Chem. Phys., 10,
11707–11735, 10.5194/acp-10-11707-2010, 2010.Van Kraalingen, D. W. G., Breure, C. J., and Spitters, C. J. T.: Simulation
of Oil Palm Growth and Yield, Agr. Forest Meteorol., 46, 227–244,
10.1016/0168-1923(89)90066-X, 1989.Vinken, G. C. M., Boersma, K. F., Maasakkers, J. D., Adon, M., and Martin, R.
V.: Worldwide biogenic soil NOx emissions inferred from OMI NO2
observations, Atmos. Chem. Phys., 14, 10363–10381,
10.5194/acp-14-10363-2014, 2014.Warwick, N. J., Archibald, A. T., Ashworth, K., Dorsey, J., Edwards, P. M.,
Heard, D. E., Langford, B., Lee, J., Misztal, P. K., Whalley, L. K., and
Pyle, J. A.: A global model study of the impact of land-use change in Borneo
on atmospheric composition, Atmos. Chem. Phys., 13, 9183–9194,
10.5194/acp-13-9183-2013, 2013.Wesely, M. L.: Parameterization of surface resistances to gaseous dry
deposition in regional-scale numerical models, Atmos. Environ., 23,
1293–1304, 10.1016/0004-6981(89)90153-4, 1989.Zhu, L., Jacob, D. J., Mickley, L. J., Marais, E. A., Cohan, D. S., Yoshida,
Y., Duncan, B. N., González Abad, G., and Chance, K. V.: Anthropogenic
Emissions of Highly Reactive Volatile Organic Compounds in Eastern Texas
Inferred from Oversampling of Satellite (OMI) Measurements of HCHO Columns,
Environ. Res. Lett., 9, 114004, 10.1088/1748-9326/9/11/114004, 2014.