ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-16-10111-2016Atmospheric CH4 and CO2 enhancements and biomass burning emission ratios derived from satellite observations of the 2015 Indonesian fire plumesParkerRobert J.https://orcid.org/0000-0002-0801-0831BoeschHartmutWoosterMartin J.MooreDavid P.WebbAlex J.GaveauDavidMurdiyarsoDanielEarth Observation Science, Department of Physics and Astronomy, University of Leicester, Leicester, UKKing's College London, Department of Geography, London, UKNERC National Centre for Earth Observation, UKCenter for International Forestry Research, P.O. Box 0113 BOCBD, Bogor, IndonesiaDepartment of Geophysics and Meteorology, Bogor Agricultural University, Bogor, IndonesiaRobert Parker (rjp23@le.ac.uk)11August20161615101111013117March201621April201623June201610July2016This 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/16/10111/2016/acp-16-10111-2016.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/16/10111/2016/acp-16-10111-2016.pdf
The 2015–2016 strong El Niño event has had a dramatic impact on the amount
of Indonesian biomass burning, with the El Niño-driven drought further
desiccating the already-drier-than-normal landscapes that are the result of
decades of peatland draining, widespread deforestation, anthropogenically
driven forest degradation and previous large fire events. It is expected that
the 2015–2016 Indonesian fires will have emitted globally significant
quantities of greenhouse gases (GHGs) to the atmosphere, as did previous El
Niño-driven fires in the region. The form which the carbon released from
the combustion of the vegetation and peat soils takes has a strong bearing on
its atmospheric chemistry and climatological impacts. Typically, burning in
tropical forests and especially in peatlands is expected to involve a much
higher proportion of smouldering combustion than the more
flaming-characterised fires that occur in fine-fuel-dominated environments
such as grasslands, consequently producing significantly more CH4
(and CO) per unit of fuel burned. However, currently there have been no
aircraft campaigns sampling Indonesian fire plumes, and very few ground-based
field campaigns (none during El Niño), so our understanding of the
large-scale chemical composition of these extremely significant fire plumes
is surprisingly poor compared to, for example, those of southern Africa or
the Amazon.
Here, for the first time, we use satellite observations of CH4 and
CO2 from the Greenhouse gases Observing SATellite (GOSAT) made in
large-scale plumes from the 2015 El Niño-driven Indonesian fires to probe
aspects of their chemical composition. We demonstrate significant
modifications in the concentration of these species in the regional
atmosphere around Indonesia, due to the fire emissions.
Using CO and fire radiative power (FRP) data from the Copernicus Atmosphere
Service, we identify fire-affected GOSAT soundings and show that peaks in
fire activity are followed by subsequent large increases in regional
greenhouse gas concentrations. CH4 is particularly enhanced, due to
the dominance of smouldering combustion in peatland fires, with CH4
total column values typically exceeding 35 ppb above those of background
“clean air” soundings. By examining the CH4 and CO2 excess
concentrations in the fire-affected GOSAT observations, we determine the
CH4 to CO2 (CH4/CO2) fire emission
ratio for the entire 2-month period of the most extreme burning
(September–October 2015), and also for individual shorter periods where the
fire activity temporarily peaks. We demonstrate that the overall CH4
to CO2 emission ratio (ER) for fires occurring in Indonesia over this
time is 6.2 ppb ppm-1. This is higher than that found over both the
Amazon (5.1 ppb ppm-1) and southern Africa (4.4 ppb ppm-1),
consistent with the Indonesian fires being characterised by an increased
amount of smouldering combustion due to the large amount of organic soil
(peat) burning involved. We find the range of our satellite-derived
Indonesian ERs (6.18–13.6 ppb ppm-1) to be relatively closely matched
to that of a series of close-to-source, ground-based sampling measurements
made on Kalimantan at the height of the fire event
(7.53–19.67 ppb ppm-1), although typically the satellite-derived
quantities are slightly lower on average. This seems likely because our field
sampling mostly intersected smaller-scale peat-burning plumes, whereas the
large-scale plumes intersected by the GOSAT Thermal And Near infrared Sensor
for carbon Observation – Fourier Transform Spectrometer (TANSO-FTS)
footprints would very likely come from burning that was occurring in a
mixture of fuels that included peat, tropical forest and already-cleared
areas of forest characterised by more fire-prone vegetation types than the
natural rainforest biome (e.g. post-fire areas of ferns and scrubland, along
with agricultural vegetation).
The ability to determine large-scale ERs from satellite data
allows the combustion behaviour of very large regions of burning to be
characterised and understood in a way not possible with ground-based studies,
and which can be logistically difficult and very costly to consider using
aircraft observations. We therefore believe the method demonstrated here
provides a further important tool for characterising biomass burning
emissions, and that the GHG ERs derived for the first time for
these large-scale Indonesian fire plumes during an El Niño event point to
more routinely assessing spatiotemporal variations in biomass
burning ERs using future satellite missions. These will have more
complete spatial sampling than GOSAT and will enable the contributions
of these fires to the regional atmospheric chemistry and climate to be better
understood.
Introduction
The 2015–2016 strong El Niño event, which is ongoing in the tropical
Pacific at the time of writing, has had a dramatic impact on the amount of
landscape burning occurring across large parts of Indonesia. Landscape fires
are commonly used in this environment to clear forest and help manage land
for agriculture, but an El Niño-driven drought has further dried out the
already human-modified landscapes of Central Kalimantan and south Sumatra.
These regions are already more flammable than their natural state due to
decades of peatland draining and deforestation, anthropogenically driven
forest degradation, as well as the legacy of previous large fire events
. Even short, localised fire events in these environments
can lead to significant greenhouse gas (GHG) emissions, as demonstrated by
, who report that a 1-week fire event in Riau province
(Sumatra) was responsible for emitting 172 ± 59 Tg CO2eq.,
approximately 5–10 % of Indonesia's average annual GHG emissions. The 2015
El Niño-driven fire season in Indonesia is already known to have been far
more extensive than in normal years , and during the last
very strong El Niño (1997–1998; the strongest yet on record) massive
increases in Indonesian fire activity were similarly recorded
. Indeed all previous El Niño events, back
to the next-strongest event after 1997–1998 (i.e. that of 1982–1983),
appear to have produced significant increases in burning over Indonesia, as
detailed in . Whilst the degree of fire activity increase
associated with El Niño is possible to gauge using, for example,
satellite-derived active fire counts, forest cover change or burned area maps
(e.g. ), what
is equally valuable is information on the emissions to the atmosphere
resulting from these burns, so that their atmospheric impacts can be more
fully determined.
Using satellite-derived estimates of burned area along with assumptions on
peatland depths of burn, estimated that the 1997 El
Niño-driven Indonesian fires released an amount of carbon (0.81–2.57 Pg)
equivalent to between 13 and 40 % of that year's annual global carbon
emissions from fossil fuels, contributing to the largest annual increase in
atmospheric CO2 concentration detected since records began in the
1950s . More recently, have reported
similarly anomalous estimates for that year's Indonesian burning, based on
related methodologies but different data sets. Anomalies in both inter-annual
variability and the atmospheric growth rates of CO2 and CH4
continue to be attributed to biomass burning events, including El
Niño-driven Indonesian fires . It is possible that the 2015–2016 El Niño-driven
Indonesian fires, which, at the time of writing, have largely ceased due to
heavy rains (but which may well return in 2016), may ultimately be of a
similarly anomalous magnitude to those driven by prior El Niño events.
Therefore, there exists a strong interest in both quantifying the amount of
fire activity occurring and calculating the overall carbon emissions to the
atmosphere that result. Furthermore, the types of biomes being affected are
important, because whilst post-fire vegetation regrowth in fire-affected
areas does subsequently take up some of the released carbon, areas of burned
tropical forest are often replaced by plants holding far less carbon per unit
area, and the burning of peat represents an effectively permanent transfer of
carbon from the land to the atmosphere . The form in which
the carbon is emitted into the atmosphere also has a strong bearing on the
emission impacts, with most carbon being released as either the long-lived
GHG carbon dioxide, the shorter-lived but much stronger GHG methane, or the
air pollutant carbon monoxide . Typically, burning in
tropical forests and especially in peatlands is expected to involve a much
higher proportion of smouldering combustion than the more
flaming-characterised combustion that occurs in fine-fuel-dominated
environments such as grasslands . Hence,
fires in peatlands and tropical forests are expected to produce more CO and
CH4 per unit of fuel burned, with a consequent reduction in the
amount of CO2. However, currently the only information on emission
makeup in fires in Indonesian biomes come from relatively few lab-based
studies where samples of fuels have been burned in combustion chambers (i.e.
). At present there are no
known field-based studies of emission makeup, certainly none conducted during
El Niño years where the dry conditions may promote different combustion
behaviour than occurs under more normal meteorological and fuel moisture
conditions, and none where the constituents of the large-scale plumes that
most likely contain the bulk of the emitted gases (and aerosols) are
assessed. The latter point is important because whilst ground-based sampling
can measure emission makeup close to the source, including in the field under
real landscape combustion conditions, such an approach is, by necessity,
limited to capturing smoke from individual fire locations, usually from
smaller fires. These measurements may not fully represent the emission
characteristics of the type of large-scale plumes that may actually be
responsible for holding most of the combustion products. Aircraft sampling
can provide a means to capture the latter's characteristics (e.g.
), but such campaigns are costly, infrequent and
logistically challenging. An alternative approach to characterising the
emission makeup of large-scale fire plumes is via satellite-based sounding of
wildfire plume chemistry, which has so far been demonstrated only a few
times, by using IASI onboard MetOp and by
using the TANSO-FTS instrument onboard the Greenhouse gases
Observing SATellite (GOSAT). Here, we build on the latter work to exploit
GOSAT's observations of CH4 and CO2 over the 2015 El
Niño-driven Indonesian fires, using these to demonstrate the increase in
atmospheric concentrations of CH4 and CO2 associated with the
large-scale biomass burning plumes, and deriving from these observations the
CH4 to CO2 emission ratios (ERs) for these El Niño-driven
Indonesian fires for the first time. We compare these Indonesian fire
emission ratios to those derived from GOSAT in alternative tropical biomes
having different combustion characteristics (southern African savannah and
the Amazon basin). In combination with fire radiative power (FRP) and
atmospheric carbon monoxide data taken from the new Copernicus Atmosphere
Service Global Fire Assimilation System (), we demonstrate
the Indonesian fires are occurring in peatland-dominated landscapes that
explain certain characteristics of the noted ERs, which are themselves
important in determining the so-called emission factors representative of the
combustion processes occurring in these very large-scale landscape fires.
Emission factors (EFs) are necessary when converting estimates of the amount
of fuel burned (obtained from burned area or FRP-based methods) into a
quantity of each trace gas . These EFs are
themselves often calculated through the use of ERs which are
determined from the ratio of the excess concentrations emitted from wildfires
. Whilst the emission factors are only one aspect of
calculating the overall emitted amounts, due to the fact that satellite
observations of burned area and FRP have significantly improved in recent
years, the accuracy of the emission factors is becoming more crucial to the
overall accuracy of the emissions . The capability to
measure the CH4 to CO2 (CH4/CO2) ERs from a variety of wildfires in
different biomes across the globe, consistently using a single
instrument and/or approach and from very large-scale plumes that represent some of
the largest individual fires emission sources, is therefore a significant
advancement. We first demonstrated this capability in , and
here we focus on extending this determination of satellite-derived
CH4/CO2 ERs to Indonesia during the anomalously large
El Niño-driven fire season of September–October 2015. The emission ratios
themselves are of clear interest in helping to determine the relative amounts
of these two key GHGs released by the fires, but also the relative amounts of
CH4 and CO2 being released in a smoke plume are known to vary
with the dominance of smouldering and flaming combustion of the causal fire,
as do the more commonly used CO2 and CO measures (e.g.
). Furthermore, knowledge of
the relative amounts of these two phases of combustion are known to exert
strong controls on the relative emissions of many other compounds (e.g.
), and thus, if we can better
understand the relative CO2 and CH4 emission makeup of the
large-scale plumes emanating from these fires, it may provide useful
information to better estimate the type of combustion occurring and thus
potentially the overall emission characteristics beyond the two species
observed.
El Niño and Indonesian fire activity
El Niño describes a large-scale climate anomaly that typically occurs once
or twice per decade, with one of the key characteristics being significantly
warmer than normal sea surface temperatures (SSTs) in the equatorial eastern
Pacific Ocean . The many other effects associated with
an El Niño event are complex and not always consistent between different El
Niños, but most events are accompanied by warmer temperatures across much
of South America, Africa, Southeast Asia and western Europe, decreased
precipitation over central/southern Africa, central America, Southeast Asia
and increased precipitation over the southern United States and western
Europe . Indonesia is located in the equatorial region
and can be particularly affected by El Niño events, for example, usually
experiencing warmer temperatures and significant reductions in rainfall that
exacerbate certain aspects of a landscape already heavily modified by human
actions. In particular, much of the low-lying land on the Indonesian islands
of Sumatra and Kalimantan that was originally covered by moist, forested
peatlands has been cleared and drained for agriculture, and this has led to
much drier landscape conditions. Fire is commonly used to manage the land,
and during the droughts associated with El Niño, the already heavily
disturbed peatland landscapes can become so dry that they can be ignited from
the vegetation fires that are widespread even in normal years
. Such fires can burn down into the carbon-rich peat for
weeks, whilst also spreading across the landscape to ignite new areas –
including spreading into areas of remaining tropical forest that normally are
not prone to fire. During El Niño these peat and forest fires can thus
affect areas that are very much larger than those burned during normal years,
particularly during the strongest El Niño events when fire activity can be
more than an order of magnitude higher .
As described in Sect. , during the 1997–1998 El Niño, fires
in Indonesia are estimated to have released huge amounts of carbon into the
atmosphere and because of the smouldering nature of peat (and to some extent,
tropical forest as well), a greater proportion of these emissions is likely
to be in the form of non-CO2 gases, primarily the air pollutant CO
and the strong greenhouse gas CH4, than is the case for flaming fires
. This contrasts with the burning of the
El Niño-dried finer fuels, which will typically burn primarily via flaming
combustion and thus release a lower proportion of CO and CH4 and a
higher proportion of CO2, whose global warming potential is
significantly lower than that of methane . It is estimated
that approximately three-fourths of the fire activity over this time period
was due to peatland burning .
Magnitude of El Niño events and the associated fire activity
There are many different ways to quantify the magnitude of an El Niño event
but one of the most widely accepted is the Multivariate ENSO Index (MEI)
. This is based on observations of a variety of
meteorological parameters over the tropical Pacific Ocean. By this metric
, the current El Niño event that we are experiencing
(2015–2016) is already the third strongest event on record (behind 1997–1998
and 1982–1983), with the potential to be classified even higher before it is
complete.
To investigate the magnitude of the increased fire activity over Indonesia
that has been associated with the current El Niño we examined the fire
radiative power being released from the identified combustion zones. FRP is a
measure of a fire's release rate of thermal radiation, and is strongly related
to the rate of fuel consumption, and trace gas and aerosol emission
. FRP is therefore both an indicator for
the presence of fire, and an estimator for the amount of material being
emitted to the atmosphere from that fire. Global satellite observations of
FRP are made from the MODIS instruments onboard the NASA Terra and Aqua
satellites at a nadir spatial resolution of 1 km, and these are incorporated
into the Copernicus Atmosphere Monitoring Services (CAMS) Global Fire
Assimilation System (GFAS), set up under the Monitoring Atmospheric
Composition and Climate (MACC) series of research projects
(). Using FRP data converted to FRP density by dividing
by the grid cell area (0.1∘× 0.1∘) and adjusting
for the impact of unseen parts of the land surface due to gaps in satellite
coverage and variations in cloud cover , GFAS produces
estimates of trace gas emissions from the mapped fire-affected areas, which
CAMS then uses in its atmospheric chemistry transport model to identify
atmospheric abundances of the released chemical species.
Time series of the monthly total fire radiative power density
(W m-2) recorded over the Indonesian region (defined as
5∘ N–10∘ S, 90–150∘ E) between 2009 and 2015,
calculated using data from the CAMS Global Fire Assimilation System . September and
October 2015 are clearly anomalous compared to the previous years shown,
highlighting the effect of this year's El Niño on the region's fire
activity.
Figure shows the monthly total FRP density (in
W m-2) over the Indonesian region (defined as
5∘ N–10∘ S, 90–150∘ E) for the last 7 years,
calculated from the GFAS data, including adjustments for observation
frequency and cloud cover . Whilst significant landscape
burning takes place every year between July and October in this Indonesian
region, the fires that took place in the latter part of 2015 (particularly
September and October 2015) were clearly of an extreme magnitude, with the
cumulative FRP density for October 2015 exceeding 7500 W m-2, compared
to the second-highest value of just over 2000 W m-2 (October 2014).
Whilst FRP gives an indication of the intensity of fires and their associated
emissions to the atmosphere, the number of fires is also a useful indicator
of fire activity, especially in regions which may see many small fires as
opposed to fewer, but larger, events .
For this reason, the original MODIS MOD14/MYD14 fire counts were also
examined . The number of fires observed by MODIS across
Indonesia during September–October 2015 is shown in
Fig. . Overlain onto this in green are the
locations of known peatlands in Sumatra, Kalimantan and Papua
. It is clear that the majority of the most fire-affected
regions of Indonesia during the September and October extreme fire event,
i.e. Central Kalimantan and the southeastern region of Sumatra, are located
in areas dominated by peatlands.
Fire emissions and combustion regimes
As already stated, in contrast to the flaming combustion involved in the
burning of wood and/or grass, peatland fires are typically dominated by deeper
smouldering combustion. As smouldering combustion is less efficient than
flaming combustion, there is a higher proportion of CO, CH4 and other
non-methane hydrocarbons (NHMCs) released compared to CO2.
A literature review of previous ground- and aircraft-based measurements of the
CH4/CO2 ER indicates a wide range of values,
demonstrating the variability that can be dependent on not only the fuel type
but also on additional factors, such as fuel moisture content, the ratio of
living to dead matter and how recently the area last burned
. To take just one example,
present CH4/CO2 ER values for flaming fires of
2.6 ppb ppm-1 from sugar cane fields, increasing to
10.3 ppb ppm-1 over fires dominated by smouldering combustion in
forest and shrubland. Fires with intermediate values were reported to
represent a mixture of smouldering and flaming combustion. Similarly,
report mean ER values of 2.1 ± 1.5 ppb ppm-1
for flaming combustion, 5.3 ± 2.0 ppb ppm-1 for mixed combustion
and 10.1 ± 3.9 ppb ppm-1 for smouldering combustion. A further
study presents values of 3.2–4.6 ppb ppm-1 for
flaming combustion, increasing to 7.8 ppb ppm-1 for smouldering
combustion in savannah or forest regions. The wide range of CH4 to
CO2 ERs reported by these different studies demonstrates
that, even when measured close to the source, as all these were, there is a high
degree of variability intrinsic to the CH4/CO2 ER but
the relative behaviour remains consistent, namely that flaming processes
produce smoke with a lower CH4 to CO2 ratio than smouldering
processes, and thus it may be possible to distinguish between these two types
of combustion using measurements of this ratio.
MODIS fire counts for September–October 2015 over the Indonesia,
gridded into 0.5∘× 0.5∘ boxes. Also overlaid are
the locations of known peatlands in Sumatra (left), Kalimantan (centre) and
Papua (right).
The objectives of this work are to first determine whether the expected high
concentrations of CH4 emitted by the extreme peatland burning in
September–October 2015 over Indonesia are observable from satellite data
and, if that is the case, to then determine the CH4/CO2
ER of the resulting large-scale smoke plumes and compare this to
measurements made in situ. The capability to examine the large-scale ERs
of a region such as this is important because if GOSAT can measure
CH4/CO2 ERs, such observations contain
information related to the mix of combustion types occurring and can thus
help discriminate predominantly smouldering from predominantly flaming
regions. Not only is this of direct interest for the CH4 and
CO2 emissions themselves but is also useful when considering the many
other species contained within the smoke, because the relative abundance of
most of these is in part dependent on the amount of flaming and smouldering
combustion occurring.
This work is presented as follows. Section introduces the
GOSAT satellite data used in this work, providing details on the retrieval
method and how the CH4/CO2 data have been used to
determine fire ERs. Section describes the
methodology used for determining whether a GOSAT sounding is affected by fire
and provides statistics on the number of fire-affected soundings that we
observe over the Indonesian fire region. Section goes on to
examine the enhancement in CH4 as observed from the fire-affected
data while Sect. then uses these data to determine
CH4/CO2 fire ERs, comparing them to in situ
observations of the same El Niño-driven fire event. Finally, we summarise
our findings and comment on the outlook for further study in this area of
research.
GOSAT Proxy XCH4 data
GOSAT was the first dedicated greenhouse gas measurement mission based on an
Earth observation satellite approach, and was launched by the Japanese Space
Agency (JAXA) on 23 January 2009 . GOSAT is equipped with
two instruments. The first is the Thermal And Near infrared Sensor for carbon
Observation – Fourier Transform Spectrometer (TANSO-FTS), which provides
point-based measurements of total column CO2 and CH4 with
near-surface sensitivity because of its use of a shortwave infrared (SWIR) as
well as a thermal infrared (TIR) band sensitive to the mid-troposphere. The
second is the Cloud and Aerosol Imager (TANSO-CAI), which provides
multispectral imagery at 0.5 km resolution with bands at 0.38, 0.67, 0.87
and 1.6 µm. This allows additional cloud–aerosol information about
the region of interest within which the TANSO-FTS measurement footprints
fall.
The TANSO-FTS measurement pattern originally consisted of five (later changed to
three) across-track points with a footprint of ∼ 10.5 km, each separated
by approximately 100 km on the ground. GOSAT also has capabilities for
agile pointing, allowing both target mode and observations of the glint spot
over the ocean. Near-surface sensitivity to the target gases is achieved by
the TANSO-FTS instrument utilising three SWIR spectral bands at 0.76, 1.6 and
2.0 µm, with mid-tropospheric sensitivity available from a fourth
band operating between 5.5 and 14.3 µm in the TIR.
provide extensive details of the performance and operation
of the TANSO-FTS instrument over the past 6 years. In short, although GOSAT
has experienced three major anomalies over its lifetime (a solar paddle
failure in May 2014, a pointing system issue in January 2015 and a
cryocooler restart in September 2015), it continues to operate well,
providing high-quality atmospheric radiance measurements from which we are
able to retrieve dry-air, column-averaged fractions of CO2 and
CH4 (denoted as XCO2 and XCH4, respectively).
Details of the University of Leicester Proxy XCH4 GOSAT retrieval,
including recent updates and uncertainty characterisation, can be found in
. In brief, the retrieval utilises the
original Orbiting Carbon Observatory (OCO) so-called full-physics retrieval
algorithm (as the radiative transfer attempts to explicitly model the
physical behaviour of the aerosol-scattered light) developed to obtain XCO2 from a simultaneous
fit of NIR/SWIR O2 and CO2 bands, and subsequently modified
to operate on GOSAT spectral data to retrieve XCH4 using the
light-path Proxy approach. Developed by for use on
SCIAMACHY data, this Proxy method utilises the fact that the majority of the
influence of atmospheric scattering on the retrieved XCH4 can be
negated through the co-retrieval of the spectrally close 1.6 µm
CO2 band, since the signal related to both species undergoes the same
light-path enhancement through scattering. The resulting
XCH4/XCO2 ratio is therefore robust to the effects of
aerosol. Generally the final Proxy XCH4 is obtained via the
application of XCO2 model fields to this ratio. Typically, due to the
fact that there is significantly less influence from aerosol on the final
product than with the typical full-physics retrieval approach
, high-quality retrievals are possible even under
cloud–aerosol conditions where the typical full-physics retrieval struggles.
Not only does this result in many more successful soundings globally but it
also allows studies over cloudy or smoke-affected regions where no data at
all may be available from the typical full-physics retrieval approach.
Time series showing the monthly 95th percentile values over
Indonesia for the GOSAT Proxy XCH4/XCO2 (top) as well
as the individual XCH4 (middle) and XCO2 (bottom) components
of the Proxy data for the entire GOSAT data record (2009–present).
XCH4 data obtained using the Proxy approach described above have been
used in many inversion studies to estimate both global and regional
emissions of XCH4. Normally the main disadvantage of the Proxy
XCH4 retrieval is that it requires an accurate and unbiased
XCO2 model to convert the ratio back into XCH4. However, in our current study of the atmospheric
impacts and ERs of the El Niño-driven fires in Indonesia, we use only the
individual retrieved XCH4 and XCO2 components of the Proxy
retrieval and hence, we have no dependence on any CO2 model. For the
purposes of this study, the standard GOSAT Proxy data record (typically
generated as part of the ESA GHG-CCI project , 4–6
months behind real time due to the use of ECMWF ERA-Interim data in the
processing chain) has been extended with the use of ECMWF Analysis data in
order to produce results more quickly than possible with the normal route. In
this way, the Proxy XCH4 time series has been extended from June to
November 2015 and includes retrievals both over land and also over ocean when
GOSAT measures in a sun-glint geometry.
In Sect. , it was shown that September–October 2015 exhibited
significantly higher FRP over Indonesia than previous years. Before exploring
the GOSAT data over Indonesia in more detail, it is first useful to put the
GHG observations for September and October 2015 into the context of the
longer GOSAT time series. Figure shows the
95th percentile values for the monthly GOSAT data over Indonesia for the
entire data record from April 2009 to November 2015. The upper panel shows
the XCH4/XCO2 ratio, with the central and lower panels
showing the individual XCH4 and XCO2, respectively.
In order to quantify the extreme nature of the October 2015 observations and
to account for the annual growth rate, we define the magnitude of the
enhancement as the October–July difference for each year, with July typically
signifying the start of the fire season in this region. For CO2, we
observe a magnitude of 4.35 ppm for October 2015 compared to a mean value of
1.05 ± 1.42 ppm for the previous years (2009–2014). In the case of
XCH4, the enhancement value for October 2015 is found to be
45.65 ppb compared to an average for previous years of
11.93 ± 3.60 ppb. The enhancement of both the XCO2 and, in
particular, XCH4 in October 2015 is therefore significantly higher
than that observed over the region in previous years, corresponding to the
extreme in fire activity observed in Fig. .
Identifying fire-affected GOSAT soundings
Section established that a significant increase is observed in
the monthly maximum values for the XCH4, XCO2 and the
XCH4/XCO2 during September–October 2015 (calculated as
the 95th percentile values) recorded over Indonesia by GOSAT. To further
investigate the atmospheric GHG anomalies identified over Indonesia by GOSAT
in Fig. , it is first necessary to identify which
GOSAT soundings are directly affected by fire emissions, and which can be
considered background (clear) cases.
We use the CAMS CO fields to determine if a particular GOSAT sounding is
likely to be fire affected. In addition to emissions from CO sources and
their atmospheric transport, the CAMS CO fields incorporate carbon monoxide
total column measurements from the IASI and MOPITT instruments
. We sampled the CO fields at the time and location of each
GOSAT sounding, and based on the CO distribution and data from the GOSAT
CAI, values in excess of 0.003 kg m-2 were
determined as being likely affected by the local fire emissions. Conversely,
if the CO value was less than 0.00075 kg m-2 then the sounding was
classed as clear (i.e. unaffected by local fire emissions). GOSAT
soundings corresponding to locations and times having CO values between these
thresholds were not able to be confidently classed as either fire
affected or clear. Out of 3946 GOSAT soundings over Indonesia during
September–October 2015, the CAMS CO identified 341 (8.6 %) of these as
being affected by fire and 1272 (32.3 %) as clear (i.e. unaffected by
fire), with the remainder lying between these thresholds.
False-colour image (RGB indicates CAI band 3, 2, 1) derived from
data taken by the GOSAT CAI, collected when the GOSAT satellite passed over
the island of Borneo on 21 October 2015 (around 13:00 LT, 05:00 UTC), a
period when extreme fires were burning across much of Central Kalimantan.
GOSAT TANSO-FTS sounding locations are identified by the numbered large red
circles, with the MODIS active fire detections identified by the small purple
circles.
Figure shows a GOSAT CAI false-colour image covering much of
Kalimantan on 21 October 2015, a time when a massive pall of smoke enveloped
Central Kalimantan and parts of the surrounding regions. The active fire
detections for this day made from MODIS are also shown (small purple
circles), along with the numbered locations of the individual GOSAT TANSO-FTS
soundings (red circles). All GOSAT soundings made coincident with this CAI
image were in locations where the simultaneous CAMS CO field indicated the
corresponding TANSO-FTS measurement was fire affected.
Observations of enhanced methane concentrations
Once we had identified a set of GOSAT soundings that were able to be clearly
classed as fire affected or clear, it became possible to examine the
XCH4, XCO2 and XCH4/XCO2 distributions
in order to determine the changes in the column amount and trace gas ratio
characteristics related to the extreme levels of fire activity.
Figure shows histograms of the
XCH4/XCO2 ratio, as well as the individual XCH4
and XCO2 components, for all the clear (blue) and fire-affected (red)
soundings, as well as for the entire data set (green).
Histograms showing the distributions over Indonesia in
September–October 2015 of the XCH4/XCO2 ratio (left),
the retrieved XCO2 (centre) and the retrieved XCH4 (right)
for all data (green), data determined to be definitely affected by fire
emissions (red) and those classed as clear (blue). Also included are the
corresponding mean and standard deviation values for each distribution.
Table showing the mean and standard deviation over Indonesia in
September–October 2015 of the XCH4/XCO2 ratio (left),
the retrieved XCO2 (centre) and the retrieved XCH4 (right)
for all data, data determined to be unaffected by fire and data determined to
be affected by fire.
As Table shows, for the XCH4/XCO2
ratio, the mean ratio calculated from all the data is 4.54 ppb ppm-1,
with a standard deviation of 0.033 ppb ppm-1. The histograms for the
clear and fire-affected data show two clearly separated distributions, with
means of 4.52 and 4.59 ppb ppm-1, respectively. When examining just the
XCO2 distributions, there appears to be less of a distinct
separation, with means of 399.9 and 401.1 ppm, respectively, for the clear and
fire-affected cases. This corresponds to a XCO2 increase of 0.3 %
percent over the background XCO2 concentrations, whereas the
XCH4 distribution for the fire-affected scenes shows a much larger
mean enhancement of 1.9 % percent over the background (1840.1 ppb vs.
1805.5 ppb).
Indonesia trace gas distributions for September–October 2015
showing (top to bottom): the GOSAT-retrieved XCH4, XCO2 and
XCH4/XCO2 ratio, along with the CAMS carbon monoxide
(CO) total column and the measured IASI CO total column. The left column
shows all data gridded at 2∘× 2∘, the central
column shows only those points determined to be clear using the criteria
of Sect. and the right column shows the data determined to be
fire affected based on the same criteria.
In order to examine the spatial distribution of the atmospheric GHGs and
XCH4/XCO2 ratio enhancements,
Fig. shows (top to bottom) maps of the
GOSAT-retrieved XCH4, XCO2, XCH4/XCO2
ratio, along with the CAMS total column CO and IASI total column CO for all
TANSO-FTS sounding locations (left), clear locations (centre) and
fire-affected locations (right). These data show that the spatial extent
of the enhancements in XCH4, XCO2 and in the resulting
XCH4/XCO2 ratio, as well as in the CAMS and IASI CO,
are related to the enhanced fire activity seen over parts of Sumatra and
Kalimantan (shown in Fig. ), whose emissions are
being transported primarily westwards over the ocean (last column of
Fig. ).
This finding confirms that the anomalously large amount of fire activity seen
occurring in September and October 2015 during the El Niño
(Fig. ) and which included fires in the extensive
peatlands of Central Kalimantan and south Sumatra
(Fig. ) resulted in a significant increase in
atmospheric column amounts of CH4 and CO2 downwind of the
fires. These enhancements are observable from GOSAT satellite observations,
and in the following section we examine the CH4/CO2 ER
of this smoke to better understand the combustion characteristics.
Determination of CH4/CO2 ERs
As discussed in Sect. , the capability to determine large-scale
regional ERs during intense fire activity is important, as it allows
information to be gained not only on the emissions of these gases themselves
but also potentially on the relative dominance of flaming vs. smouldering
combustion. Our previous work, , demonstrated for the first
time an ability to determine CH4/CO2 fire ERs from
satellite data, in that case using GOSAT to study ERs of boreal forest
(Canada and Russia), tropical forest (Brazil) and savannah (southern Africa)
fires. The satellite-derived ERs obtained appeared to be in good agreement
with those derived during ground and aircraft sampling studies in the same
biomes, albeit these in situ data themselves show relatively large
variations. Such variability is likely a function of differences in fuel
type, fuel moisture and fire behaviour that occurred between different
measurement campaigns, fire locations and time of year or day
. Here, we apply the technique of to
our current GOSAT Proxy retrievals of XCO2 and XCH4 made
during the September–October 2015 Indonesian fires, in order to determine
the ERs characterising the very large-scale plumes seen during this
anomalously large climate-related fire event.
As a first step in this process, it is useful to calculate the excess (or
Δ) XCH4 and XCO2 values prior to any subsequent
processing, since, for example, the fire emissions can be superimposed into a
background atmosphere that itself contains spatially and/or temporally
varying amounts of XCH4 and XCO2. Calculating such excess
amounts removes the impact of potentially varying background concentrations.
However, since we utilise the XCH4 and XCO2 components of the
GOSAT Proxy XCH4 retrieval, which themselves do not account
explicitly for aerosol scattering (but instead rely on these effects to ratio
out when computing the final Proxy XCH4 values; see
Sect. ), this does provide some potential for error to be
introduced in any subsequently calculated CH4/CO2
emission ratio. Such errors are related to the fact that the degree of
scattering may be different between the fire-affected (i.e. smoke-laden) and
matching background (i.e clear) TANSO-FTS soundings from which the excess
amounts are calculated. We analysed the magnitude of this effect using a
simple model, included in Appendix , and the results indicate that
it is possible to underestimate the CH4/CO2 ER by
∼ 10 % if appropriate care is not taken during selection of the clear
soundings whose column amounts are to be subtracted from those of the
fire-affected soundings in order to calculate the excess column amounts. In a
region such as Indonesia during El Niño, where large-scale fire activity is
clearly greatly affecting the aerosol composition of the local atmosphere,
this aspect becomes even more challenging. To deal with this, we only used
fire-affected TANSO-FTS soundings made over land, so as to minimise the
effect of mixing and/or dilution as smoke-laden air was transported longer
distances over the ocean. For each fire-affected sounding, a matching
background measurement was selected from the group of clear soundings located
over the same island and as close as possible to the fire-affected
measurement (e.g. the background for the Sumatra soundings were selected from
clear soundings between 90–108∘ E and
5∘ N–10∘ S) in order to minimise impacts stemming from use
of non-uniform background measurements as detailed in .
Out of 131 fire-affected soundings, a suitable background sounding was
identified for 105 (80 %) of the soundings. Each background XCH4
and XCO2 value was then subtracted from the concentration derived
from its corresponding fire-affected sounding in order to produce the
ΔXCH4 and ΔXCO2 values, from which the ERs
could then be calculated.
Scatterplot of GOSAT-derived ΔXCH4 vs.
ΔXCO2 values for large-scale fire plumes seen over Indonesian
region (of the type seen in Fig. ) from 1 September to
31 October 2015, calculated as the total column difference between the fire-affected
and corresponding clear background TANSO-FTS soundings. The
CH4/CO2 ER, (ppb ppm-1) is
calculated from the gradient of a linear best fit, shown as the dashed line.
Also shown are the correlation coefficient R and the number of soundings
N.
Daily fire radiative power density (red line) taken from the Global
Fire Assimilation System (GFAS) , operated as part of CAMS.
Data are shown from 1 September to 31 October 2015, for the entire Indonesian landmass (red) and separately
for the regions of Sumatra and Kalimantan. Two specific time periods are
highlighted (referred to as Period 1 and Period 2), Period 1 covering
9–15 September (Sumatra) and 8–17 September (Kalimantan) and Period 2
covering 19–27 October (Sumatra) and 14–25 October (Kalimantan).
Figure shows the ΔXCH4 values plotted
against the simultaneously derived ΔXCO2 values for all of the
fire-affected soundings measured over Indonesia for the
September–October 2015 period. The CH4/CO2 ER derived
from the linear best fit to these data is 6.2 ppb ppm-1 (correlation
coefficient of 0.937). Whilst this calculated CH4/CO2
ER is significantly above that of the ambient background
(∼ 4.52 ppb ppm-1), the many fire plumes sampled by GOSAT
soundings across the September–October 2015 period mean that potential
variations in the emission ratios over time (and space) can also be explored.
Figure once again shows the daily CAMS FRP
density data, but this time for September–October 2015 only, and as a daily
average for the entire Indonesian region as well as for Sumatra and
Kalimantan individually. There is very significant variability seen in the
fire activity across these 2 months, and we identify several distinct time
periods to examine in more detail for both Sumatra and Kalimantan. The period
9–15 September over Sumatra is characterised by a steady increase in FRP
density, peaking on 12 September at over 200 W m-2 before decreasing
again and reducing to below 50 W m-2 by 15 September. We take this as
Period 1 for Sumatra. By contrast, over Kalimantan at around the same time
(specifically between the 8 and 17 September) there is a peak in FRP on
8 September of nearly 300 W m-2, followed by a lull around the middle
period before a second increase to almost 150 W m-2 on
13–14 September. We take this as Period 1 for Kalimantan. In contrast to the
differing behaviours during Period 1, both Sumatra and Kalimantan exhibit
somewhat more similar trends in fire activity during Period 2, starting with
a high peak (over 300 W m-2) on 19 and 14 October for Sumatra and
Kalimantan, respectively, however, while over Kalimantan the fire activity
then immediately reduced to a lower level (around 100 W m-2). Over
Sumatra, high FRP density values in excess of 300 W m-2 are maintained
over several days before slowly decreasing. This suggests a significantly
larger fire event over Sumatra than over Kalimantan at this time, a finding
consistent with the CAMS total column CO fields (e.g. as seen in
Fig. ). Although Fig.
suggests that an additional period centred around 22 September should be of
interest, there are insufficient GOSAT soundings during this time from which
to determine an ER, demonstrating that the somewhat limited GOSAT sampling
strategy can lead to a sparseness of observations in certain situations.
Scatterplots of ΔXCH4 vs. ΔXCO2
derived for Sumatran large-scale fire plumes via analysis of TANSO-FTS data
for the time periods detailed in Fig. :
September–October 2015 (top), Period 1: 9–15 September (middle) and Period
2: 19–27 October (bottom). The CH4/CO2 ER
is calculated as the gradient of a linear fit to the data (dashed line). The
correlation coefficient R and the number of soundings N are also shown.
Figure shows the ΔXCH4 vs.
ΔXCO2 measurements recorded over Sumatra, for the entire
2-month period (September–October) (top) and for Period 1 (middle) and
Period 2 (bottom) only. Over the 2 months, a total of 66 fire-affected
TANSO-FTS measurements are identified that have a suitable matching
background available from which to calculate ΔXCH4 and
ΔXCO2. The linear best fit to these data give a
CH4/CO2 ER of 6.64 ppb ppm-1 (R=0.893) for these Sumatran fires. When examining the Periods 1 and 2 only,
which Fig. shows correspond to times of
increased fire activity over the island, higher ERs of 8.1 and
8.8 ppb ppm-1 are derived (R=0.91 and 0.92, respectively). These
higher CH4/CO2 ERs are consistent with the
region being characterised by a larger proportion of smouldering combustion,
most likely of peatland, given the preponderance of that land cover in the
fire-affected area (Fig. ), resulting in enhanced
CH4 concentrations as already observed in Sect. .
Scatterplots of ΔXCH4 vs. ΔXCO2
derived for Kalimantan large-scale fire plumes via analysis of TANSO-FTS data
for the time periods detailed in Fig. :
September–October 2015 (top), Period 1: 8–17 September (middle) and Period
2: 14–25 October (bottom). The CH4/CO2 ER
is calculated as the gradient of a linear fit to the data (dashed line). The
correlation coefficient R and the number of soundings N are also shown.
Similar to Fig. ,
Fig. shows the ΔXCH4 and
ΔXCO2 retrievals for Kalimantan, plotted on a scatterplot from
which the CH4/CO2 ER can be derived. Over
the 2 months of September and October 2015,
Fig. shows that Kalimantan appears characterised
by typically lower amounts of fire activity than Sumatra, interspersed with
relatively short but intense episodes, such as those on 8 September and
14 October. The CH4/CO2 ER calculated for
the Kalimantan data across the entire 2-month period is found to be
6.2 ppb ppm-1, calculated from 39 separate ΔXCH4 and
ΔXCO2 observations (correlation coefficient of 0.974).
However, when examining Period 1 only (8–17 September), although derived
from only nine data points (R=0.92) an extremely high ER is found
(13.6 ppb ppm-1, R=0.92). By contrast, during Period 2
(14–25 October) the ER is found once again to be lower, at
6.2 ppb ppm-1 (R=0.97). This lower value may be affected by the
fact that throughout Period 2 extensive smoke aerosol covers much of
Kalimantan (as seen in Fig. ) and the selection of clear
TANSO-FTS that appropriately represent the clean background of the
fire-affected measurements is significantly more difficult. This is further
compounded by the fact that the wind vectors (not shown) for this period
indicate that the background air is likely to be coming from further south,
potentially having a different CH4 and CO2 concentration.
Ground-based ERs from El Niño-enhanced peat and vegetation fires
In addition to space-based observations described above, during October 2015,
at the height of the fire activity on Kalimantan
(Fig. ), a short field campaign was conducted to
derive CH4 to CO2 ERs for comparison to the GOSAT-derived
values. During this campaign, smoke was sampled and ERs derived for
individual fire plumes stemming from the El Niño-enhanced landscape fires.
Trace gas mixing ratio measures of CO2 and CH4 were made at
1 Hz frequency in plumes from fires at four different locations within
∼ 30 km of Palangkaraya (2.21∘ S, 113.92∘ E), the
capital of Central Kalimantan and one of the most fire-affected regions
during the 2015 El Niño related drought . We use a
ground-based, more portable version of the cavity-enhanced laser absorption
spectrometer described in . The precision (Allan variance,
1σ at 1 Hz) of the mixing ratios derived via the laser spectroscopy
was 1.71 ppb for CH4 and 2.63 ppm for CO2, with a total
absolute uncertainty of around 1 % of the measured concentrations. Fires at
the four different locations were sampled between 12 and 16 October 2015,
with each site located on peat but with plumes encompassing both pure peat
burning and also times when both peat and some overlying vegetation were
being consumed. The CH4/CO2 ERs determined from these
close-to-source measurements varied between 7.53 and 19.7 ppb ppm-1
(mean ± sd = 12.9 ± 3.9 ppb ppm-1), a range relatively
consistent with that determined from the GOSAT-derived, space-based
observations (6.18–13.6 ppb ppm-1). However, the majority of the
ground-based ERs were derived from locations dominated by almost pure peat
burning sampled close to the source, whereas the space-based observations
from GOSAT are derived from measurements of the smoke filling a 10.5 km
diameter TANSO-FTS footprint and thus representative of much larger areas of
combustion, very likely comprising a mix of peat and vegetation burning in
the majority of cases.
Despite this potential for the GOSAT-derived CH4 to CO2 ER to
be somewhat less characteristic of pure peat burning than are some of the
close-to-source measurements, and the potential for the measurements to be
influenced by cleaner air (such as that transported from the south), it is
still expected that the emissions over Indonesia will be largely dominated by
smouldering combustion, resulting in a typically higher
CH4/CO2 ER than that observed from flaming combustion
as is generally characteristic of most African savannah burning
. To confirm that this is the case, the same GOSAT-based
analysis performed here for the 2015 Indonesian fires was also performed for
southern Africa (defined as 0∘ N–40∘ S,
30∘ W–60∘ E) and the Amazon (defined as
0∘ N–40∘ S, 30–75∘ W), both of which underwent
significant fire activity during this same time period. The calculated
CH4/CO2 ER for southern Africa was found to be
4.35 ppb ppm-1 (see Appendix B, Fig. ), consistent
with observations of flaming-dominated combustion in savannah regions.
, using a ground-based, open-path FTIR system, reported
CH4 to CO2 ERs for different phases of southern African
savannah burns conducted on 7 ha plots in Kruger National Park, South
Africa. Backfires (spreading against the wind) typically produced emissions
with very complete combustion characteristics, with CH4 to
CO2 ERs of 1.9–2.2 ppb ppm-1, apart from one case where a
value of 6.0 ppb ppm-1 was recorded. Residual areas of smouldering
combustion present after the fire front had passed were recorded as having
CH4 to CO2 ERs of 3.1–14.1 ppb ppm-1, although it was
possible that the lowest ERs reported were significantly influenced by
remaining pockets of flaming activity. The headfire emissions, which combine
the smoke from the most intense flaming part of the burn with those from the
smouldering zone immediately behind, were found to have CH4 to
CO2 ERs of 2.4–5.4 ppb ppm-1. The overall fire-averaged
CH4 to CO2 ER, calculated from the ERs of the individual
phases and using airborne measures of FRP to estimate the amount the fuel
consumption in each for the purposes of the weighting calculation, was
4.3 ± 1.7 ppb ppm-1, very close to the 4.35 ppb ppm-1
derived from GOSAT's observations of large-scale southern African savannah
plumes. This provides further evidence for the representative nature of our
GOSAT approach, which is currently the only method able to assess the ERs of
the largest plumes emanating from landscape fires, albeit only at the
relatively sparse sampling locations targeted by GOSAT. The
CH4/CO2 ER for the Amazon is, as perhaps expected,
somewhat in between that of African savannah and Indonesian peatlands and
forests, being 5.1 ppb ppm-1 (see Fig. ).
and references therein report the presence of significant
smouldering combustion in Amazon fires occurring in forested regions, much
more than typically seen in African savannah and primarily stemming from the
coarse woody fuels that represent a significant component of the fuel in this
biome. However, smouldering in the peat-dominated fuels of the Indonesian
fires would still be expected to be more prevalent ,
and so the CH4 to CO2 ER would be expected to be higher
there, as we have indeed found.
Summary and outlook
The objective of this study was to utilise XCH4 and XCO2
observations made by the GOSAT satellite when passing over Indonesia to probe
the composition of large-scale plumes from the 2015 Indonesian fires for the
first time, with these extreme fires being driven by the ongoing strong El
Niño, the largest seen since 1997–1998. We wished to both identify the
atmospheric greenhouse gas impacts of the very significant increase in fire
activity and use any such measurements to determine the biomass burning
ERs of these two important GHGs using the technique we pioneered
in . This would enable the characterisation of certain
aspects of the chemical makeup of these large-scale El Niño-driven fires
for the first time, which, in 1997–1998, were responsible for the largest
release of fire-emitted GHGs seen worldwide, and indeed which are believed to
be of a magnitude not seen since that period anywhere on Earth
.
Our analysis of GOSAT data confirms a significant enhancement of both
XCH4 and XCO2 in the fire-affected GOSAT soundings, with the
greatest change seen in the XCH4 mixing ratios where we see an
average value of 1840.1 ppb compared to an average value in the clear
(non-fire-affected) cases of 1805.5 ppb. For these fire-affected soundings,
the CH4/CO2 ER was estimated from the
gradient of the linear best fit to the excess XCH4 and XCO2
values. We find an overall ER for the entire Indonesian fire-affected region
during the September–October 2015 fire peak of 6.2 ppb ppm-1, with
Sumatra showing slightly higher mean ERs (6.6 ppb ppm-1) than
Kalimantan (6.2 ppb ppm-1). When examining shorter periods of time to
focus on specific fire episodes on each island, we find ERs as low as 6.1 and
as high as 13.6 ppb ppm-1. This range is consistent with that seen in
field-sampled GHG data taken in October 2015 on Kalimantan, close to the fire
sources, but we believe the large-scale sampling provided by the GOSAT
TANSO-FTS 10.5 km diameter footprints enables sampling of a much more
representative amount of smoke than does the relatively limited, small-scale
sampling possible on the ground. We therefore believe that our GOSAT-derived
ERs are well suited for use in studies attempting to understand
the impact of these extreme El Niño-driven fires on the larger-scale,
regional atmosphere.
Our GOSAT-derived ERs for Indonesia indicate plumes that appear
more dominated by the products of smouldering combustion than the plumes
sampled by GOSAT in southern Africa and in the Amazon during the same period,
consistent with prior expectation and previous ground-based and airborne
sampling campaigns that suggest less smouldering-dominated combustion in
these latter biomes (especially in the savannah case). GOSAT's capability to
determine not only the enhancement in greenhouse gas abundance stemming from
such large fire events but also to provide the data necessary to calculate
the GHG emission ratios and help identify the relative balance of
smouldering and flaming activity ongoing in very large regions is an
extremely valuable aid to understanding the composition of the plumes and
their impact on regional atmospheric composition and climate. Some challenges
remain, mainly relating to obtaining an accurate representation of the
background non-fire-affected XCH4 and XCO2 amounts (see
Appendix ). However, the technique that we present here and in
should be easily applicable to future satellite missions
focused on atmospheric composition. Several of these will have increased
spatial and temporal resolutions that will greatly aid in obtaining the most
suitable background observations. One such mission, Sentinel-5 Precursor, is
planned for launch in 2016 and is capable of measuring both CH4 and
CO at a high spatial resolution, providing an ability that GOSAT currently
lacks. We believe, therefore, that this work will prove valuable in eventually
facilitating the routine determination of regional biomass burning ERs from
space, and their spatiotemporal variations whose importance is described in,
e.g. . Such a capability might ultimately allow the
characterisation of such burning events under different climatic and biome
conditions.
Data availability
The University of Leicester Proxy XCH4 data used in this work is available as part of the ESA GHG-CCI Project, freely available from http://www.esa-ghg-cci.org. Please contact Dr. Robert Parker for more details.
As discussed in Sect. , there exists the potential to introduce
errors into the GOSAT-derived ERs if the fire-affected sounding
and the background sounding each contain sufficiently different aerosol
scattering characteristics. As the fire-affected sounding will, by
definition, usually contain a non-trivial amount of smoke aerosols, whilst the background
sounding is, in theory, supposed to be free of smoke, some quantification of this
affect is needed. In this section, we derive and use a simple mathematical
representation to determine the magnitude of such effects.
Let the observed excess concentration be the difference between the observed fire and background concentrations:
ΔXCH4=XCH4fire-XCH4bgd.
Both soundings will have an error due to scattering associated with them,
which typically lengthens the light path and hence reduces the inferred gas
mixing ratio. This error factor, here termed A, will be different for the
fire and background cases.
Implementation of Eq. () with ΔXCH4
varied between 5 and 50 ppb for ERs ranging from 0.003 to 0.009 for
different ranges of Aratio. The true ERs and the ERs derived from
the observed correlation are shown in each panel. The top left figure shows a
fixed value of Aratio=1 while the remaining panels (clockwise)
show the value of Aratio ranging from 0.999 to 1, 0.99 to 1 and
0.995 to 1.
Therefore,
ΔXCH4=AfireXCH4fire-AbgdXCH4bgd.
Similarly, for XCO2 we have
ΔXCO2=AfireXCO2fire-AbgdXCO2bgd.
The ratio of the excess concentrations due to the fire emissions, from which
we calculate the CH4 to CO2 ER, is then given by
ΔXCH4ΔXCO2=AfireXCH4fire-AbgdXCH4bgdAfireXCO2fire-AbgdXCO2bgd.
Now, we set the observed CH4 concentration in the fire sounding to the
background concentration, plus the true excess concentration related to the
fire:
XCH4fire=XCH4bgd+ΔXCH4true.
Doing the same for XCO2 now gives
ΔXCH4ΔXCO2=Afire(XCH4bgd+ΔXCH4true)-AbgdXCH4bgdAfire(XCO2bgd+ΔXCO2true)-AbgdXCO2bgd.
This can then be expanded to
ΔXCH4ΔXCO2=AfireXCH4bgd+AfireΔXCH4true-AbgdXCH4bgdAfireXCO2bgd+AfireΔXCO2true-AbgdXCO2bgd,
and then rearranged to
ΔXCH4ΔXCO2=(Afire-Abgd)XCH4bgd+AfireΔXCH4true(Afire-Abgd)XCO2bgd+AfireΔXCO2true.
Now, let the ratio of the two error terms be
Aratio=Afire/Abgd,
which then rearranged gives
Afire=AratioAbgd.
Substituting this in now gives
ΔXCH4ΔXCO2=(Abgd(Aratio-1))XCH4bgd+AratioAbgdΔXCH4true(Abgd(Aratio-1))XCO2bgd+AratioAbgdΔXCO2true.
The Abgd terms cancel, giving
ΔXCH4ΔXCO2=(Aratio-1)XCH4bgd+AratioΔXCH4true(Aratio-1)XCO2bgd+AratioΔXCO2true.
This equation therefore relates the observed excess concentrations
(ΔXCH4 and ΔXCO2) calculated from the
difference in GOSAT's fire-affected and background soundings to the
true background concentrations (XCH4bgd and XCO2bgd),
the true excess concentrations (ΔXCH4true and
ΔXCO2true) and the ratio between the error terms,
Aratio. Furthermore, we can use this simple relationship to explore
the likely error in the calculated ER for a given value of
Aratio.
Figure shows the implementation of Eq. ()
for various scenarios. The background XCH4 and XCO2
concentrations are fixed at 1850 ppb and 400 ppm, respectively,
representing the normal fire-free atmosphere. The true methane enhancement
(ΔXCH4true) is varied between 0 and 50 ppb in 5 ppb
increments and the true ER between CH4 and CO2 is varied
between 0.003 and 0.012. The different panels then show the behaviour for
various ranges of Aratio. The top left panel has Aratio
set at a constant value of 1 (i.e. the error in the background is exactly the
same as the error in the fire cases), which is the ideal situation, and the
true ERs are reproduced exactly. The top right panel allows Aratio
to vary between 0.999 and 1.0. The effect of this is a slight spreading of
the lines and the difference between the true and observed ER is minimal. The
bottom left panel shows the increased range of Aratio from 0.995 to
1.0, which causes the observed ERs to differ more from the truth, with a true
ER of 0.008 only appearing as an observed ratio of 0.00755, an error of
5.6 %. Finally, in the bottom right panel, the value of Aratio is
allowed to vary between 0.99 and 1.0. This relatively large variation
decreases the observed ER further, with a true ER of 0.008 appearing as an
observed ratio of 0.00686, for example.
Whilst there are many unknowns that impact the value of Aratio, and
so it is not possible to know its exact value for a particular pair of GOSAT
fire-affected and background observations used to derive an emission ratio,
it is possible to determine its expected range. By comparing the scatter of
the fire-affected and background XCH4 values to the Proxy
XCH4 data (which are much less affected by aerosol and in this case
used as the truth) it is possible to estimate the likely range of values of
Aratio. The standard deviation of the ratio between the
XCH4/ Proxy XCH4 for the background and fire cases is
found to be 0.00494, suggesting that values of Aratio are likely in
the 0.995–1.0 range (i.e. up to a 0.5 % reduction in A). This means that
whilst we are likely to tend to underestimate the true CH4 to
CO2 emission ratios with GOSAT, for the majority of cases
(CH4 to CO2 ERs in the range 0.005–0.008) the effect can be
considered small, with typical biases of 0.4–5.6 %. Even in extreme cases
with high ERs (e.g. 0.012), we expect an error of less than 15 %.
This section contains ΔXCH4 vs. ΔXCO2
correlation plots for southern Africa (Fig. ) and the
Amazon (Fig. ), as discussed in the main text in
Sect. .
Scatterplot showing the ΔXCH4 vs.
ΔXCO2 values, calculated as the difference between the values
in the fire-affected soundings to those in the background cases over the
entire southern African region from 1 September to 31 October 2015. The
CH4/CO2 ER is calculated from the gradient of the
linear best fit, which is shown along with the correlation coefficient R
and the number of sounding pairs N. This GOSAT-derived ER is very similar
to the fire-averaged CH4 to CO2 ER of
4.3 ± 1.7 ppb ppm-1 derived by using
open-path FTIR spectroscopy measurements close to the source on these types
of savannah fire events.
Scatterplot showing the ΔXCH4 vs.
ΔXCO2 values, calculated as the difference between the values
in the fire-affected soundings to those in the background cases over the
entire Amazonian region from 1 September to 31 October 2015. The
CH4/CO2 ER is calculated from the gradient of a linear
fit to the data. This line of best fit is also shown, along with the
correlation coefficient R and the number of soundings N.
Acknowledgements
R. J. Parker is funded via an ESA Living Planet Fellowship with additional
funding from the UK National Centre for Earth Observation (NCEO) and the ESA
Greenhouse Gas Climate Change Initiative (GHG-CCI). H. Boesch, D. Moore and
M. Wooster are also supported by NERC NCEO. A. J. Webb is funded by a NERC
PhD. Field measurements in Indonesia were part supported by NERC grant
NE/J010502/1 (NERC SAMBBA). Field measurements were also part supported by a
DFID grant to CIFOR (project no. 203034). Bruce Main at King's College London
and Agus Salim at CIFOR are thanked for their technical contributions to the
field measurement campaign. D. Murdiyarso acknowledges the support provided
by USAID and USFS. This work also relates to NERC grant NE/N01555X/1.
We thank the Japanese Aerospace Exploration Agency, National Institute for
Environmental Studies, and the Ministry of Environment for the GOSAT data and
their continuous support as part of the Joint Research Agreement. We also
thank CAMS for provision of the data from GFAS. We also acknowledge the use
of MODIS Active Fire Detections extracted from MCD14ML distributed by NASA
FIRMS (available on-line
at https://earthdata.nasa.gov/active-fire-data). We thank EUMETSAT for
the IASI CO Level 2 data. IASI is a joint mission of EUMETSAT and the Centre
National détudes Spatiales (CNES, France).
Finally, this research used the ALICE High Performance Computing Facility at
the University of Leicester.Edited by: M.
Chipperfield Reviewed by: two anonymous referees
ReferencesAlexe, M., Bergamaschi, P., Segers, A., Detmers, R., Butz, A., Hasekamp, O.,
Guerlet, S., Parker, R., Boesch, H., Frankenberg, C., Scheepmaker, R. A.,
Dlugokencky, E., Sweeney, C., Wofsy, S. C., and Kort, E. A.: Inverse
modelling of CH4 emissions for 2010–2011 using different satellite
retrieval products from GOSAT and SCIAMACHY, Atmos. Chem. Phys., 15,
113–133, 10.5194/acp-15-113-2015, 2015.Andreae, M. O. and Merlet, P.: Emission of trace gases and aerosols from
biomass burning, Global Biogeochem. Cy., 15, 955–966,
10.1029/2000GB001382, 2001.Bertschi, I. T., Yokelson, R. J., Ward, D. E., Christian, T. J., and Hao,
W. M.: Trace gas emissions from the production and use of domestic biofuels
in Zambia measured by open-path Fourier transform infrared spectroscopy,
J. Geophys. Res.-Atmos., 108, D13,
10.1029/2002JD002158, 2003.Boesch, H., Baker, D., Connor, B., Crisp, D., and Miller, C.: Global
Characterization of CO2 Column Retrievals from Shortwave-Infrared Satellite
Observations of the Orbiting Carbon Observatory-2 Mission, Remote Sensing, 3,
270–304, 10.3390/rs3020270, 2011.Bonsang, B., Boissard, C., Le Cloarec, M., Rudolph, J., and Lacaux, J.:
Methane, carbon monoxide and light non-methane hydrocarbon emissions from
African savanna burnings during the FOS/DECAFE experiment, J.
Atmos. Chem., 22, 149–162, 10.1007/BF00708186, 1995.Buchwitz, M., Reuter, M., Schneising, O., Boesch, H., Guerlet, S., Dils, B.,
Aben, I., Armante, R., Bergamaschi, P., Blumenstock, T., Bovensmann, H.,
Brunner, D., Buchmann, B., Burrows, J., Butz, A., Chadin, A., Chevallier, F.,
Crevoisier, C., Deutscher, N., Frankenberg, C., Hase, F., Hasekamp, O.,
Heymann, J., Kaminski, T., Laeng, A., Lichtenberg, G., Maziere, M. D., Noel,
S., Notholt, J., Orphal, J., Popp, C., Parker, R., Scholze, M., Sussmann, R.,
Stiller, G., Warneke, T., Zehner, C., Bril, A., Crisp, D., Griffith, D.,
Kuze, A., O'Dell, C., Oshchepkov, S., Sherlock, V., Suto, H., Wennberg, P.,
Wunch, D., Yokota, T., and Yoshida, Y.: The Greenhouse Gas Climate Change
Initiative (GHG-CCI): Comparison and quality assessment of
near-surface-sensitive satellite-derived {CO2} and {CH4} global data
sets, Remote Sens. Environ., 162, 344–362,
10.1016/j.rse.2013.04.024, 2015.Butz, A., Hasekamp, O. P., Frankenberg, C., Vidot, J., and Aben, I.: CH4
retrievals from space-based solar backscatter measurements: Performance
evaluation against simulated aerosol and cirrus loaded scenes, J.
Geophys. Res.-Atmos., 115, D21, 10.1029/2010JD014514, 2010.CAMS-GFAS, E.: CAMS Global Fire Assimilation System,
available at: http://apps.ecmwf.int/datasets/data/cams-gfas/ (last access: 1 August 2016),
2016.Christian, T. J., Kleiss, B., Yokelson, R. J., Holzinger, R., Crutzen, P. J.,
Hao, W. M., Saharjo, B. H., and Ward, D. E.: Comprehensive laboratory
measurements of biomass-burning emissions: 1. Emissions from Indonesian,
African, and other fuels, J. Geophys. Res.-Atmos., 108, 4719,
10.1029/2003JD003704, 2003.Christian, T. J., Yokelson, R. J., Carvalho, J. A., Griffith, D. W. T.,
Alvarado, E. C., Santos, J. C., Neto, T. G. S., Veras, C. A. G., and Hao,
W. M.: The tropical forest and fire emissions experiment: Trace gases emitted
by smoldering logs and dung from deforestation and pasture fires in Brazil,
J. Geophys. Res.-Atmos., 112, d18308,
10.1029/2006JD008147, 2007.Cofer, W. R., Winstead, E. L., Stocks, B. J., Goldammer, J. G., and Cahoon,
D. R.: Crown fire emissions of CO2, CO, H2, CH4, and TNMHC from a dense Jack
pine boreal forest fire, Geophys. Res. Let., 25, 3919–3922,
10.1029/1998GL900042, 1998.Cogan, A., Boesch, H., Parker, R., Feng, L., Palmer, P., Blavier, J.-F.,
Deutscher, N. M., Macatangay, R., Notholt, J., Roehl, C., Warneke, T., and
Wunch, D.: Atmospheric carbon dioxide retrieved from the Greenhouse gases
Observing SATellite (GOSAT): Comparison with ground-based TCCON observations
and GEOS-Chem model calculations, J. Geophys. Res.-Atmos., 117, D21, 10.1029/2012JD018087,
2012.Coheur, P.-F., Clarisse, L., Turquety, S., Hurtmans, D., and Clerbaux, C.: IASI measurements of
reactive trace species in biomass burning plumes, Atmos. Chem. Phys., 9, 5655–5667, 10.5194/acp-9-5655-2009, 2009.Cressot, C., Chevallier, F., Bousquet, P., Crevoisier, C., Dlugokencky, E. J., Fortems-Cheiney, A., Frankenberg, C., Parker, R.,
Pison, I., Scheepmaker, R. A., Montzka, S. A., Krummel, P. B., Steele, L. P., and Langenfelds, R. L.: On the consistency between global and regional methane emissions
inferred from SCIAMACHY, TANSO-FTS, IASI and surface measurements, Atmos. Chem. Phys., 14, 577–592, 10.5194/acp-14-577-2014, 2014.Drake, N.: Indonesia blazes threaten endangered orangutans, Nature, 527, 18,
10.1038/527018a, 2015.Frankenberg, C., Meirink, J. F., Bergamaschi, P., Goede, A. P. H., Heimann,
M., Körner, S., Platt, U., van Weele, M., and Wagner, T.: Satellite
chartography of atmospheric methane from SCIAMACHY on board ENVISAT: Analysis
of the years 2003 and 2004, J. Geophys. Res.-Atmos.,
111, d07303, 10.1029/2005JD006235, 2006.Fraser, A., Palmer, P. I., Feng, L., Boesch, H., Cogan, A., Parker, R.,
Dlugokencky, E. J., Fraser, P. J., Krummel, P. B., Langenfelds, R. L.,
O'Doherty, S., Prinn, R. G., Steele, L. P., van der Schoot, M., and Weiss, R.
F.: Estimating regional methane surface fluxes: the relative importance of
surface and GOSAT mole fraction measurements, Atmos. Chem. Phys., 13,
5697–5713, 10.5194/acp-13-5697-2013, 2013.Fraser, A., Palmer, P. I., Feng, L., Bösch, H., Parker, R., Dlugokencky, E.
J., Krummel, P. B., and Langenfelds, R. L.: Estimating regional fluxes of
CO2 and CH4 using space-borne observations of XCH4: XCO2, Atmos.
Chem. Phys., 14, 12883–12895, 10.5194/acp-14-12883-2014, 2014.Freeborn, P. H., Wooster, M. J., Hao, W. M., Ryan, C. A., Nordgren, B. L.,
Baker, S. P., and Ichoku, C.: Relationships between energy release, fuel mass
loss, and trace gas and aerosol emissions during laboratory biomass fires,
J. Geophys. Res.-Atmos., 113, D1,
10.1029/2007JD008679, 2008.Gaveau, D. L., Salim, M. A., Hergoualc'h, K., Locatelli, B., Sloan, S.,
Wooster, M., Marlier, M. E., Molidena, E., Yaen, H., DeFries, R., Verchot, L., Murdiyarso, D.,Nasi, R., Holmgren, P., and Sheil, D.:
Major atmospheric emissions from peat fires in Southeast Asia during
non-drought years: evidence from the 2013 Sumatran fires, Scientific reports,
4, 6112, 10.1038/srep06112, 2014.Giglio, L., Descloitres, J., Justice, C. O., and Kaufman, Y. J.: An enhanced
contextual fire detection algorithm for MODIS, Remote Sens. Environ.,
87, 273–282, 10.1016/S0034-4257(03)00184-6, 2003.Guild, L. S., Kauffman, J. B., Cohen, W. B., Hlavka, C. A., and Ward, D. E.:
Modeling biomass burning emissions for amazon forest and pastures in
Rondônia, Brazil, Ecol. Appl., 14, 232–246,
10.1890/01-6009, 2004.Hartmann, D., Klein Tank, A., Rusticucci, M., Alexander, L., Brönnimann,
S., Charabi, Y., Dentener, F., Dlugokencky, E., Easterling, D., Kaplan, A.,
Soden, B., Thorne, P., Wild, M., and Zhai, P.: Observations: Atmosphere and
Surface, book section 2, Cambridge University Press, Cambridge,
United Kingdom and New York, NY, USA, 159–254, 10.1017/CBO9781107415324.008,
2013.Huijnen, V., Wooster, M. J., Kaiser, J. W., Gaveau, D. L. A., Flemming, J.,
Parrington, M., Inness, A., Murdiyarso, D., Main, B., and van Weele, M.: Fire
carbon emissions over maritime southeast Asia in 2015 largest since 1997,
Scientific Reports, 6, 26886, 10.1038/srep26886, 2016.Hurst, D. F., Griffith, D. W., and Cook, G. D.: Trace gas emissions from
biomass burning in tropical Australian savannas, J. Geophys.
Res.-Atmos., 99, 16441–16456, 10.1029/94JD00670, 1994.Inness, A., Blechschmidt, A.-M., Bouarar, I., Chabrillat, S., Crepulja, M.,
Engelen, R. J., Eskes, H., Flemming, J., Gaudel, A., Hendrick, F., Huijnen,
V., Jones, L., Kapsomenakis, J., Katragkou, E., Keppens, A., Langerock, B.,
de Mazi`ere, M., Melas, D., Parrington, M., Peuch, V. H., Razinger, M.,
Richter, A., Schultz, M. G., Suttie, M., Thouret, V., Vrekoussis, M., Wagner,
A., and Zerefos, C.: Data assimilation of satellite-retrieved ozone, carbon
monoxide and nitrogen dioxide with ECMWF's Composition-IFS, Atmos. Chem.
Phys., 15, 5275–5303, 10.5194/acp-15-5275-2015, 2015.Kaiser, J. W., Heil, A., Andreae, M. O., Benedetti, A., Chubarova, N., Jones,
L., Morcrette, J.-J., Razinger, M., Schultz, M. G., Suttie, M., and van der
Werf, G. R.: Biomass burning emissions estimated with a global fire
assimilation system based on observed fire radiative power, Biogeosciences,
9, 527–554, 10.5194/bg-9-527-2012, 2012.Kasischke, E. S. and Bruhwiler, L. P.: Emissions of carbon dioxide, carbon
monoxide, and methane from boreal forest fires in 1998, J.
Geophys. Res.-Atmos., 107, 8146,
10.1029/2001JD000461, 2002.Koppmann, R., Khedim, A., Rudolph, J., Poppe, D., Andreae, M. O., Helas, G.,
Welling, M., and Zenker, T.: Emissions of organic trace gases from savanna
fires in southern Africa during the 1992 Southern African Fire Atmosphere
Research Initiative and their impact on the formation of tropospheric ozone,
J. Geophys. Res.-Atmos., 102, 18879–18888,
10.1029/97JD00845, 1997.Koppmann, R., von Czapiewski, K., and Reid, J. S.: A review of biomass burning emissions, part I: gaseous emissions of carbon monoxide, methane, volatile
organic compounds, and nitrogen containing compounds, Atmos. Chem. Phys. Discuss., 5, 10455–10516, 10.5194/acpd-5-10455-2005, 2005.Korontzi, S., Ward, D. E., Susott, R. A., Yokelson, R. J., Justice, C. O.,
Hobbs, P. V., Smithwick, E. A. H., and Hao, W. M.: Seasonal variation and
ecosystem dependence of emission factors for selected trace gases and
PM2.5
for southern African savanna fires, J. Geophys. Res.-Atmos., 108, 4758, 10.1029/2003JD003730, 2003.Kuze, A., Suto, H., Nakajima, M., and Hamazaki, T.: Thermal and near infrared
sensor for carbon observation Fourier-transform spectrometer on the
Greenhouse Gases Observing Satellite for greenhouse gases monitoring, Appl.
Opt., 48, 6716–6733, 10.1364/AO.48.006716, 2009.Kuze, A., Suto, H., Shiomi, K., Kawakami, S., Tanaka, M., Ueda, Y., Deguchi, A.,
Yoshida, J., Yamamoto, Y., Kataoka, F., Taylor, T. E., and Buijs, H. L.: Update on GOSAT TANSO-FTS performance,
operations, and data products after more than 6 years in space, Atmos. Meas. Tech., 9, 2445–2461, 10.5194/amt-9-2445-2016, 2016.Langner, A. and Siegert, F.: Spatiotemporal fire occurrence in Borneo over a
period of 10 years, Glob. Change Biol., 15, 48–62,
10.1111/j.1365-2486.2008.01828.x, 2009.Langner, A., Miettinen, J., and Siegert, F.: Land cover change 2002–2005 in
Borneo and the role of fire derived from MODIS imagery, Glob. Change
Biol., 13, 2329–2340, 10.1111/j.1365-2486.2007.01442.x, 2007.Lee, T., Sullivan, A. P., Mack, L., Jimenez, J. L., Kreidenweis, S. M.,
Onasch,
T. B., Worsnop, D. R., Malm, W., Wold, C. E., Hao, W. M., and Collett Jr., J.
L. :
Chemical Smoke Marker Emissions During Flaming and Smoldering Phases of
Laboratory Open Burning of Wildland Fuels, Aerosol Sci. Tech.,
44, i–v, 10.1080/02786826.2010.499884, 2010.Liu, D., Allan, J. D., Young, D. E., Coe, H., Beddows, D., Fleming, Z. L.,
Flynn, M. J., Gallagher, M. W., Harrison, R. M., Lee, J., Prevot, A. S. H.,
Taylor, J. W., Yin, J., Williams, P. I., and Zotter, P.: Size distribution,
mixing state and source apportionment of black carbon aerosol in London
during wintertime, Atmos. Chem. Phys., 14, 10061–10084,
10.5194/acp-14-10061-2014, 2014.
Lobert, J. M. and Warnatz, J.: Emissions from the combustion process in
vegetation, in: Fire in the Environment: The Ecological, Atmospheric and
Climatic Importance of Vegetation Fires,, edited by: Crutzen, P. J. and
Goldammer, J. G., John Wiley, 15–37, 1993.Myhre, G., Shindell, D., Bréon, F.-M., Collins, W., Fuglestvedt, J.,
Huang,
J., Koch, D., Lamarque, J.-F., Lee, D., Mendoza, B., Nakajima, T., Robock,
A., Stephens, G., Takemura, T., and Zhang, H.: Anthropogenic and Natural
Radiative Forcing, in: Climate Change 2013: The Physical Science Basis.
Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change, edited by: Stocker, T., Qin, D.,
Plattner, G.-K., Tignor, M., Allen, S., Boschung, J., Nauels, A., Xia, Y.,
Bex, V., and Midgley, P., Cambridge University Press,
chapt. 8, 659–740, 10.1017/CBO9781107415324.018, 2013.O'Dell, C. W., Connor, B., Bösch, H., O'Brien, D., Frankenberg, C.,
Castano, R., Christi, M., Eldering, D., Fisher, B., Gunson, M., McDuffie, J.,
Miller, C. E., Natraj, V., Oyafuso, F., Polonsky, I., Smyth, M., Taylor, T.,
Toon, G. C., Wennberg, P. O., and Wunch, D.: The ACOS CO2 retrieval
algorithm – Part 1: Description and validation against synthetic
observations, Atmos. Meas. Tech., 5, 99–121, 10.5194/amt-5-99-2012,
2012.O'Shea, S. J., Bauguitte, S. J.-B., Gallagher, M. W., Lowry, D., and Percival, C. J.: Development of a
cavity-enhanced absorption spectrometer for airborne measurements of CH4 and CO2, Atmos. Meas. Tech., 6, 1095–1109, 10.5194/amt-6-1095-2013, 2013.Othman, M. and Latif, M.: Dust and gas emissions from small-scale peat
combustion, Aerosol Air Qual. Res., 13, 1045–1059,
10.4209/aaqr.2012.08.0214, 2013.Page, S. E., Siegert, F., Rieley, J. O., Boehm, H.-D. V., Jaya, A., and
Limin,
S.: The amount of carbon released from peat and forest fires in Indonesia
during 1997, Nature, 420, 61–65, 10.1038/nature01131, 2002.Parker, R., Boesch, H., Cogan, A., Fraser, A., Feng, L., Palmer, P. I.,
Messerschmidt, J., Deutscher, N., Griffith, D. W., Notholt, J., Wennberg, P. O., and Wunch, D.:
Methane observations from the Greenhouse Gases Observing SATellite:
Comparison to ground-based TCCON data and model calculations, Geophys.
Res. Lett., 38, 15, 10.1029/2011GL047871, 2011.Parker, R. J., Boesch, H., Byckling, K., Webb, A. J., Palmer, P. I., Feng, L., Bergamaschi, P., Chevallier, F.,
Notholt, J., Deutscher, N., Warneke, T., Hase, F., Sussmann, R., Kawakami, S., Kivi, R., Griffith, D. W. T., and Velazco, V.:
Assessing 5 years of GOSAT Proxy XCH4 data and associated uncertainties, Atmos. Meas. Tech., 8, 4785–4801, 10.5194/amt-8-4785-2015, 2015.Reid, J. S., Koppmann, R., Eck, T. F., and Eleuterio, D. P.: A review of
biomass burning emissions part II: intensive physical properties of biomass
burning particles, Atmos. Chem. Phys., 5, 799–825,
10.5194/acp-5-799-2005, 2005.Ritung, S., Wahyunto, and Nugroho, K.: Peatland map of Indonesia, Office of
Research and Development of Land Resources, Ministry of Agriculture,
available at: http://www.litbang.pertanian.go.id/unker/one/600/ (last access: 1 August 2016),
2016.Ross, A. N., Wooster, M. J., Boesch, H., and Parker, R.: First satellite
measurements of carbon dioxide and methane emission ratios in wildfire
plumes, Geophys. Res. Lett., 40, 4098–4102,
10.1002/grl.50733, 2013.Schepers, D., Guerlet, S., Butz, A., Landgraf, J., Frankenberg, C., Hasekamp,
O., Blavier, J.-F., Deutscher, N. M., Griffith, D. W. T., Hase, F., Kyro, E.,
Morino, I., Sherlock, V., Sussmann, R., and Aben, I.: Methane retrievals from
Greenhouse Gases Observing Satellite (GOSAT) shortwave infrared measurements:
Performance comparison of proxy and physics retrieval algorithms, J.
Geophys. Res.-Atmos., 117, D10,
10.1029/2012JD017549, 2012.Schroeder, W., Oliva, P., Giglio, L., and Csiszar, I. A.: The New {VIIRS}
375
m active fire detection data product: Algorithm description and initial
assessment, Remote Sens. Environ., 143, 85–96,
10.1016/j.rse.2013.12.008, 2014.Simmonds, P., Manning, A., Derwent, R., Ciais, P., Ramonet, M., Kazan, V.,
and
Ryall, D.: A burning question. Can recent growth rate anomalies in the
greenhouse gases be attributed to large-scale biomass burning events?,
Atmos. Environ., 39, 2513–2517,
10.1016/j.atmosenv.2005.02.018, 2005.Stockwell, C. E., Yokelson, R. J., Kreidenweis, S. M., Robinson, A. L.,
DeMott, P. J., Sullivan, R. C., Reardon, J., Ryan, K. C., Griffith, D. W. T.,
and Stevens, L.: Trace gas emissions from combustion of peat, crop residue,
domestic biofuels, grasses, and other fuels: configuration and Fourier
transform infrared (FTIR) component of the fourth Fire Lab at Missoula
Experiment (FLAME-4), Atmos. Chem. Phys., 14, 9727–9754,
10.5194/acp-14-9727-2014, 2014.Trenberth, K. E.: The Definition of El Niño, B. Am.
Meteorol. Soc., 78, 2771–2777,
10.1175/1520-0477(1997)078<2771:TDOENO>2.0.CO;2, 1997.Trigg, S. N., Curran, L. M., and McDonald, A. K.: Utility of Landsat 7
satellite data for continued monitoring of forest cover change in protected
areas in Southeast Asia, Singapore J. Trop. Geo., 27, 49–66,
10.1111/j.1467-9493.2006.00239.x, 2006.Turner, A. J., Jacob, D. J., Wecht, K. J., Maasakkers, J. D., Lundgren, E.,
Andrews, A. E., Biraud, S. C., Boesch, H., Bowman, K. W., Deutscher, N. M.,
Dubey, M. K., Griffith, D. W. T., Hase, F., Kuze, A., Notholt, J., Ohyama,
H., Parker, R., Payne, V. H., Sussmann, R., Sweeney, C., Velazco, V. A.,
Warneke, T., Wennberg, P. O., and Wunch, D.: Estimating global and North
American methane emissions with high spatial resolution using GOSAT satellite
data, Atmos. Chem. Phys., 15, 7049–7069, 10.5194/acp-15-7049-2015, 2015.van der Werf, G. R., Randerson, J. T., Collatz, G. J., Giglio, L.,
Kasibhatla,
P. S., Arellano, A. F., Olsen, S. C., and Kasischke, E. S.: Continental-Scale
Partitioning of Fire Emissions During the 1997 to 2001 El Niño/La
Niña Period, Science, 303, 73–76, 10.1126/science.1090753, 2004.van der Werf, G. R., Dempewolf, J., Trigg, S. N., Randerson, J. T.,
Kasibhatla,
P. S., Giglio, L., Murdiyarso, D., Peters, W., Morton, D. C., Collatz, G. J.,
Dolman, A. J., and DeFries, R. S.: Climate regulation of fire emissions and
deforestation in equatorial Asia, Proc. Natl. Acad.
Sci., 105, 20350–20355, 10.1073/pnas.0803375105, 2008.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 Leeuwen, T. T. and van der Werf, G. R.: Spatial and temporal variability in the ratio of trace gases emitted from biomass burning, Atmos. Chem. Phys., 11, 3611–3629, 10.5194/acp-11-3611-2011, 2011.Voiland, A.: Seeing Through the Smoky Pall: Observations from a Grim
Indonesian
Fire Season, available at:
http://earthobservatory.nasa.gov/Features/IndonesianFires (last access: 1 August 2016),
2016.Wang, W., Ciais, P., Nemani, R. R., Canadell, J. G., Piao, S., Sitch, S.,
White, M. A., Hashimoto, H., Milesi, C., and Myneni, R. B.: Variations in
atmospheric CO(2) growth rates coupled with tropical temperature, Proc.
Natl. Acad. Sci. USA, 110,
13061–13066, 10.1073/pnas.1219683110, 2013.Wecht, K. J., Jacob, D. J., Sulprizio, M. P., Santoni, G. W., Wofsy, S. C.,
Parker, R., Bösch, H., and Worden, J.: Spatially resolving methane
emissions in California: constraints from the CalNex aircraft campaign and
from present (GOSAT, TES) and future (TROPOMI, geostationary) satellite
observations, Atmos. Chem. Phys., 14, 8173–8184,
10.5194/acp-14-8173-2014, 2014.Wolter, K.: Multivariate ENSO Index (MEI), available at:
http://www.esrl.noaa.gov/psd/enso/mei/, last access: 1 August 2016.Wolter, K. and Timlin, M. S.: Measuring the strength of ENSO events: How does
1997/98 rank?, Weather, 53, 315–324,
10.1002/j.1477-8696.1998.tb06408.x, 1998.Wooster, M. and Zhang, Y.: Boreal forest fires burn less intensely in Russia
than in North America, Geophys. Res. Lett., 31, 20,
10.1029/2004GL020805, 2004.Wooster, M. J., Freeborn, P. H., Archibald, S., Oppenheimer, C., Roberts, G.
J., Smith, T. E. L., Govender, N., Burton, M., and Palumbo, I.: Field
determination of biomass burning emission ratios and factors via open-path
FTIR spectroscopy and fire radiative power assessment: headfire, backfire and
residual smouldering combustion in African savannahs, Atmos. Chem. Phys., 11,
11591–11615, 10.5194/acp-11-11591-2011, 2011.
Wooster, M. J., Ceccato, P., and Flasse, S. P.: Cover – Indonesian fires
observed using AVHRR, Int. J. Remote Sens., 19, 383–386,
10.1080/014311698216035, 1998.Wooster, M. J., Roberts, G., Perry, G. L. W., and Kaufman, Y. J.: Retrieval
of
biomass combustion rates and totals from fire radiative power observations:
FRP derivation and calibration relationships between biomass consumption and
fire radiative energy release, J. Geophys. Res.-Atmos.,
110, d24311, 10.1029/2005JD006318, 2005.Wooster, M. J., Perry, G. L. W., and Zoumas, A.: Fire, drought and El Niño relationships on Borneo (Southeast Asia)
in the pre-MODIS era (1980–2000), Biogeosciences, 9, 317–340, 10.5194/bg-9-317-2012, 2012.Yokelson, R. J., Griffith, D. W. T., and Ward, D. E.: Open-path Fourier
transform infrared studies of large-scale laboratory biomass fires, J.
Geophys. Res.-Atmos., 101, 21067–21080,
10.1029/96JD01800, 1996.Yokelson, R. J., Goode, J. G., Ward, D. E., Susott, R. A., Babbitt, R. E.,
Wade, D. D., Bertschi, I., Griffith, D. W., and Hao, W. M.: Emissions of
formaldehyde, acetic acid, methanol, and other trace gases from biomass fires
in North Carolina measured by airborne Fourier transform infrared
spectroscopy, J. Geophys. Res.-Atmos., 104,
30109–30125, 10.1029/1999JD900817, 1999.Yokelson, R. J., Christian, T. J., Karl, T. G., and Guenther, A.: The
tropical forest and fire emissions experiment: laboratory fire measurements
and synthesis of campaign data, Atmos. Chem. Phys., 8, 3509–3527,
10.5194/acp-8-3509-2008, 2008.Yokelson, R. J., Andreae, M. O., and Akagi, S. K.: Pitfalls with the use of
enhancement ratios or normalized excess mixing ratios measured in plumes to
characterize pollution sources and aging, Atmos. Meas. Tech., 6, 2155–2158,
10.5194/amt-6-2155-2013, 2013.