ESDEarth System DynamicsESDEarth Syst. Dynam.2190-4987Copernicus PublicationsGöttingen, Germany10.5194/esd-9-663-2018The biomass burning contribution to climate–carbon-cycle feedbackThe biomass burning contributionHarrisonSandy P.s.p.harrison@reading.ac.ukhttps://orcid.org/0000-0001-5687-1903BartleinPatrick J.BrovkinVictorhttps://orcid.org/0000-0001-6420-3198HouwelingSanderhttps://orcid.org/0000-0002-6189-1009KlosterSilviaPrenticeI. ColinDepartment of Geography and Environmental Science, University of
Reading, Whiteknights, Reading, RG6 6AB, UKDepartment of Geography, University of Oregon, Eugene, Oregon
97403–1251, USAMax Planck Institute for Meteorology, Bundesstraße 53, 20146
Hamburg, GermanyDepartment of Earth Sciences, Vrije Universiteit Amsterdam, De
Boelelaan 1085, 1081 HV Amsterdam, The NetherlandsAXA Chair of Biosphere and Climate Impacts, Department of Life
Sciences, Imperial College London, Ascot, SL5 7PY, UKSandy P. Harrison (s.p.harrison@reading.ac.uk)28May20189266367718February201826February201828April2018This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/This article is available from https://esd.copernicus.org/articles/9/663/2018/esd-9-663-2018.htmlThe full text article is available as a PDF file from https://esd.copernicus.org/articles/9/663/2018/esd-9-663-2018.pdf
Temperature exerts strong controls on the incidence and severity of fire. All
else equal, warming is expected to increase fire-related carbon emissions,
and thereby atmospheric CO2. But the magnitude of this feedback is very
poorly known. We use a single-box model of the land biosphere to quantify
this positive feedback from satellite-based estimates of biomass burning
emissions for 2000–2014 CE and from sedimentary charcoal records for the
millennium before the industrial period. We derive an estimate of the
centennial-scale feedback strength of 6.5 ± 3.4 ppm CO2 per
degree of land temperature increase, based on the satellite data. However,
this estimate is poorly constrained, and is largely driven by the
well-documented dependence of tropical deforestation and peat fires
(primarily anthropogenic) on climate variability patterns linked to the El
Niño–Southern Oscillation. Palaeo-data from pre-industrial times provide
the opportunity to assess the fire-related climate–carbon-cycle feedback over
a longer period, with less pervasive human impacts. Past biomass burning can
be quantified based on variations in either the concentration and isotopic
composition of methane in ice cores (with assumptions about the isotopic
signatures of different methane sources) or the abundances of charcoal
preserved in sediments, which reflect landscape-scale changes in burnt
biomass. These two data sources are shown here to be coherent with one
another. The more numerous data from sedimentary charcoal, expressed as
normalized anomalies (fractional deviations from the long-term mean), are
then used – together with an estimate of mean biomass burning derived from
methane isotope data – to infer a feedback strength of
5.6 ± 3.2 ppm CO2 per degree of land temperature and (for a
climate sensitivity of 2.8 K) a gain of 0.09 ± 0.05. This finding
indicates that the positive carbon cycle feedback from increased fire
provides a substantial contribution to the overall climate–carbon-cycle
feedback on centennial timescales. Although the feedback estimates from
palaeo- and satellite-era data are in agreement, this is likely fortuitous
because of the pervasive influence of human activities on fire regimes during
recent decades.
Introduction
Fire is a natural, recurring event in most terrestrial ecosystems. About
4 % of the global land area is burnt every year (Giglio et al., 2013),
resulting in global CO2 emissions of around 2 PgC per year (van der
Werf et al., 2010), substantial contributions to the budgets of other direct
or indirect greenhouse gases (including CH4, CO, N2O, and ozone
precursors), and further contributions to the atmospheric aerosol loading
(black carbon, organic compounds). Climate-induced inter-annual variability
in biomass burning, particularly variability associated with the El
Niño–Southern Oscillation (ENSO), is an important component of the
inter-annual variability of the atmospheric CO2 growth rate (van der
Werf et al., 2010). However, changes in biomass burning also occur in
response to longer-term climate variability and trends (Marlon et al., 2008,
2013; Power et al., 2008; Mooney et al., 2011; Daniau et al., 2012). Changes in biomass burning therefore need to be taken into account in
estimating the “climate–carbon-cycle feedback”, i.e. the longer-term
positive feedback by which global warming leads to a reduction in land
carbon storage, a consequent reduction in the net uptake of CO2 so that
more CO2 remains in the atmosphere, and thus an amplification of the
initial warming (Arora et al., 2013; Cox et al., 2013; Wenzel et al., 2014).
The dominant terms in the terrestrial carbon balance are gross primary
production and total ecosystem respiration. The climate–carbon-cycle
feedback is generally attributed to the temperature-dependent balance of
these two large annual fluxes (Keenan et al., 2016; Ballantyne et al., 2017;
Jung et al., 2017); but this neglects the potential contribution of biomass
burning, which we consider here.
Although there have been attempts to quantify the contribution of
deforestation fires (Bowman et al., 2009) or the aerosol-related component
of biomass burning (Arneth et al., 2010), the global-scale contribution of
biomass burning to the climate–carbon-cycle feedback has been quantified
only once (Ward et al., 2012). That study reported a variety of feedbacks
based on simulations using a single Earth system model (ESM). Ward et al. (2012) found that the simulated total climate feedback due to fire was
negative, but their conclusion rested mainly on a large (and highly
uncertain: Boucher et al., 2013; Carslaw et al., 2013; Lee et al., 2016)
indirect aerosol effect that exceeded the simulated fire feedback through
the carbon cycle. In contrast, Arneth et al. (2010) estimated the aerosol
feedback from biomass burning to be small and of uncertain sign.
Remotely sensed observations of biomass burning offer a uniquely detailed
global perspective on fire regimes. However, they cover only a limited
period, and our ability to use these records to derive an empirical estimate
of the biomass burning contribution to the carbon cycle feedback is further
compromised by the complexity of the controls on fire. Climate influences
the occurrence and magnitude of fires on daily to seasonal timescales; both
climate and fire affect vegetation productivity and hence the availability
of fuel on yearly to decadal timescales; and human activities increase
ignitions but more importantly decrease fuel availability and fire spread
(Bistinas et al., 2014: Knorr et al., 2014; Andela et al., 2017). Analyses
of the independent influence of individual controls, when other factors are
held constant, show that burnt area and biomass burning is extremely
sensitive to, and positively correlated with, spatial and temporal
variations in global temperature (Krawchuk et al.. 2009; Daniau et al.,
2012; Bistinas et al., 2014). Regional analyses also show positive
relationships between biomass burning and temperature, although the strength
of this relationship relative to other controls on fire varies between
regions (see, e.g., Marlon et al., 2013). The overwhelming importance of
temperature for fire regimes is unsurprising given that temperature changes
influence atmospheric circulation patterns and are closely tied to changes
in precipitation and surface climates (Held and Soden, 2006; Li et al.,
2013). The positive relationship between temperature and fire at global and
regional scales suggests that the contribution of fire to the climate–carbon-cycle feedback is likely to be positive. Yet burnt area has declined over
the last decade. This decline has been attributed to the effects of fire
suppression and landscape fragmentation outweighing the influence of
climate-induced changes in biomass burning (Andela et al., 2017).
The use of palaeo-climate records obviates the problem of limited record
length and avoids those various human influences that have been so large as
to dominate the fire record over at least the past 150 years (Marlon et al.,
2008). Ice cores provide direct evidence for past changes in atmospheric
composition, and the concentration and stable carbon-isotope composition of
methane (CH4) in ice cores have been used together to reconstruct
changes in biomass burning during the last millennium: see Rubino et al. (2016) for a review. CH4 is released during the smouldering phase of
fires, roughly in proportion to total CO2 emission (Andreae and Merlet,
2001). Although this process is a relatively minor contributor to total
atmospheric CH4, it disproportionately influences the 13C content
of CH4 because pyrogenic CH4 carries the isotopic signature of
photosynthesis. This is much less negative than that of the dominant
(microbial) sources of CH4 (Barker and Fritz, 1981). But measurements
of the 13C content of CH4 in ice cores are currently
available with limited temporal resolution and are subject to large
uncertainties in the isotopic fractionation factors for different CH4
sources. The abundance of sedimentary charcoal provides an alternative and
more direct measure of relative changes in biomass burning (Power et al.,
2008; Harrison et al., 2010) and has been shown to mirror changes in
biomass burning CH4 (Wang et al., 2010). Sedimentary charcoal data are
far more numerous than ice-core isotopic records for the last millennium. If
it is possible to establish a quantitative relationship between charcoal
abundance and biomass burning CH4, it should then be worthwhile to
exploit the greater number and temporal resolution of these records to
quantify the fire contribution to the carbon-climate feedback. This is the
approach we adopt in this paper. We use a single-box model of the land
biosphere to derive an estimate of the contemporary biomass burning
contribution to the climate–carbon-cycle feedback using remote-sensing-based
estimates of biomass burning carbon emissions for the interval 2000–2014
CE. We then demonstrate that the charcoal and methane records of biomass
burning during the pre-industrial Common Era (1–1700 CE) are in good
agreement. Finally, we exploit a good correlation of normalized anomalies of
global charcoal abundance with global land temperatures during the last
millennium to derive an alternative estimate of the strength of the
climate–carbon-cycle feedback.
Schematic of the analysis of global fire–temperature relationships
for the (a) satellite and (b) pre-industrial eras. Fb: biomass burning
flux; Tland: global mean land temperature; T: global mean
temperature; W: global land carbon storage; NPP: global net primary production;
W/ NPP: residence time of land carbon; AF: airborne fraction; 2.12: conversion
factor from ppm to Pg C; S: climate sensitivity; C: atmospheric CO2 mole
fraction; Nt: normalized anomalies of charcoal data; ΔFb:
biomass burning flux anomalies; μ(Fb): mean biomass burning flux.
See text for sources of temperature data.
Data and methods
We use a single-box model of the land biosphere to quantify the fire
feedback, making separate calculations of the feedback strength and gain for
the satellite-era and the pre-industrial period (Fig. 1). Feedback strength
is measured as the increase in atmospheric CO2 concentration (ppm) per
degree increase in temperature (K). In the satellite era, we use the
relationship between the satellite-derived fire emissions and land
temperature to estimate feedback strength, with a correction for the fact
that land temperature variations are stronger than global mean temperature
variations. We then convert feedback strength to gain, assuming well-founded
values for the total biomass, airborne fraction, climate sensitivity, and
atmospheric CO2 concentration (Fig. 1a). We follow the same approach
for the pre-industrial era (Fig. 1b), but using sedimentary charcoal data to
estimate variations in fire emissions. Use of the sedimentary charcoal data
in this way is predicated on our demonstration here that there is a strong
statistical relationship, conforming to an expected mathematical form,
between the charcoal data and the ice-core record of atmospheric methane and
its stable carbon-isotope composition.
Remotely sensed burned area and carbon emissions
Burnt area and carbon emissions for 2000 to 2014 were derived from the GFED4s
database (Randerson et al., 2015,
http://www.geo.vu.nl/~gwerf/GFED/GFED4/, last access: 16 February 2018). GFED4s provides monthly burnt area estimates on a 0.5∘
spatial grid from 1997 through 2014, but prior to August 2000 these
observations were derived by calibrating the Along Track Scanning Radiometers (ATSR) and the Visible and Infrared Scanner (VIRS) active fire data
against MODIS-derived burnt area (van der Werf et al., 2017). We therefore
exclude the pre-MODIS period 1997 to 1999 because of the large uncertainties
in the burnt-area and emission estimates (Giglio et al., 2013). We also test
whether the retention of the mixed-source estimates for 2000 (with only 5
months from MODIS) has an impact on the results (Supplement, Sect. 8). Carbon
emissions in GFED4s are divided into source sectors: savannah, grassland, and
shrubland fires; boreal forest fires; temperate forest fires; deforestation
fires; peatland fires; agricultural fires. The estimates we use for total
fire emissions include all of these sectors except agricultural fires. We
exclude agricultural fires on the assumption that these are only started by
people and therefore that the incidence, timing, and size of these fires are
unrelated to climate or other environmental factors. We also use an estimate
of the total emissions from natural sources, that is, also excluding
deforestation and peatland fires. Global mean land temperatures for this
period, for comparison with the fire data, were taken from the NOAA Merged
Land Ocean Global Surface Temperature Analysis (NOAAGlobalTemp v4.0.1,
10.7289/V5FN144H; Vose et al. 2012):
https://www.ncdc.noaa.gov/data-access/marineocean-data/noaa-global-surface-temperature-noaaglobaltemp (last access: 16 February 2018),
with specific data found at
http://www1.ncdc.noaa.gov/pub/data/noaaglobaltemp/operational/ (last access: 16 February 2018).
Charcoal data
The sedimentary charcoal data were obtained from version GCDv3 of the Global
Charcoal Database (Marlon et al., 2016). This new version of the database
contains considerably more individual site records than previous versions and provides better
spatial coverage. Charcoal data were read directly from
the database file GCDv03_Marlon_et_al_2015.mdb. The data were processed using
the protocol described in Power et al. (2010) and Blarquez et al. (2014)
except that the transformed charcoal influx values (or their equivalents)
were expressed as normalized anomalies (normans, Nt at time t) or
Nt=(ct∗-c‾∗)/c‾∗,
where the ct∗ is the optimally Box–Cox-transformed influx
values from a particular record at time t, and c‾∗ is the mean
transformed influx for that record over the interval 1–1700 CE (the
transformation and normalization base period). A 10-year interval was used for
pre-binning the 633 records used for the creation of the composite curve.
Methane concentration and stable carbon-isotope data
Methane concentration data were taken from the composite Law Dome records
(Etheridge et al., 2010). We used a composite data set of δ13C
of CH4 from Ferretti et al. (2005), Mischler et al. (2009), and Sapart
et al. (2012). We used the published age models for each record. We then
applied the 0.51 ‰ correction described by Sapart et al. (2012)
to the Northern Hemisphere data, in order to create the global
composite.
Global palaeo-temperature data
We calculated annual area-weighted averages of mean annual temperature
anomalies for land grid points, using the 5∘ gridded data
set of Mann et al. (2009), which covers the interval from 500 through 2006 CE.
We used a base period of 1961–1990 CE to calculate anomalies. We did
not use the global average of the PAGES 2k Consortium (2013) because this
reconstruction is dominated by records from the Arctic and Antarctic, where
there are few or no fires, prior to 800 CE. Although there are many
last-millennium temperature reconstructions for the Northern Hemisphere,
global data sets are few and the rest cover shorter time intervals than Mann
et al. (2009).
Composite curves of charcoal, δ13C of CH4,
CH4, and palaeo-temperature data
The individual charcoal records have a median sampling interval of 16.75 years
over the interval 1–100 CE (with 250 sites contributing data), and
16.90 years over the interval 1601–1700 CE (350 sites), for a typical sample
density of over 1000 per century. The δ13C of CH4 and
CH4 records average 2.5 and 3.0 samples per century over the interval
1–500 CE, increasing to 10 per century over the interval 1601–1700 CE. The
temperature data have annual resolution. Consequently, for the regression
analyses we developed composite (across sites, in the case of charcoal) or
smoothed curves (for the other variables) with a common sampling interval,
and an appropriate smoothing window for each series. We used the R package
locfit (R Core Team, 2016; Loader, 2013) to fit these curves.
Data smoothing can induce spurious cross-correlations between series (Loader,
1999; Granger and Newbold, 1986), while using an overly high-resolution
sampling interval can create temporal pseudoreplication, whereby sequential
observations do not provide independent information (Hurlbert, 1984). Both
could inflate the apparent significance of relationships among series. We
chose the sampling interval and smoothing window by examining diagnostic
checks of the regression analyses of charcoal (as the response variable) with
temperature, or δ13C of CH4 and CH4 (as predictors),
attempting to minimize the autocorrelation of the residuals as a guard
against pseudoreplication. This process led to the selection of a 50-year
time step for the evaluation of the smoothed curves. For the charcoal and
temperature data, we selected a 50-year (half-width) fixed smoothing window,
which suppresses inter-annual to decadal-scale variability in those series,
while preserving longer-term variations. The δ13C of CH4 and
CH4 data are too sparse in the first part of the record to use a
fixed-width smoothing window, and so we used the variable window width or
“span” approach with the span parameter equal to 0.1. This strategy led to
some interpolation in the sparser parts of these records. We obtained
bootstrap confidence intervals for the smoothed curves. For charcoal, we used
the “bootstrap-by-site” approach described by Blarquez et al. (2014), which
allows the impact of the variations in the spatial distribution of the
charcoal records to be assessed, and the standard approach for the other
series. The R code used to produce the composite/smoothed curves is included
in the Supplement (Sects. 2–5).
Comparison of charcoal and methane records
The isotopic composition of atmospheric CH4 depends on the magnitudes
and isotopic discrimination factors of different contributors to the global
CH4 budget. Thus, although variations in biomass burning emission of
CH4 are expected to influence its isotopic composition, there is no direct correspondence between isotopic composition and the biomass burning
flux. The isotopic composition of CH4 can also be influenced by changes
in the magnitude and/or isotopic discrimination of other methane fluxes, of
which the microbial source (methanogenesis in wetlands and wet soils and in
other anoxic environments including ruminant stomachs) dominates. Moreover,
isotopic discrimination by methanogenesis shows large geographic variations
and cannot be assumed to be the same now (with widespread agricultural
grazing and draining of natural wetlands) as it was in pre-industrial
times. We therefore chose to compare the CH4 isotopic record with the
charcoal record by treating the isotopic discrimination factors as unknown
and using a regression approach (Fig. 1), respecting the isotopic mass
balance, to test whether the two types of record are systematically related
to one another. After 1700 CE, the relationships between charcoal and
temperature and between charcoal and δ13C [CH4] and
[CH4] become significantly distorted. Regressions were therefore fitted
using composite/smoothed curve data only up to and including 1700 CE.
The mass balance equation for the principal (non-fossil fuel) annual
CH4 fluxes to the atmosphere is
F=Fm+Fg+Fb,
where F is the total flux, Fm is the microbial flux, Fg is the
geological flux (natural seepage from underground gas reservoirs), and
Fb is the biomass burning flux. The isotopic mass balance equation is
δ=δm(Fm/F)+δg(Fg/F)+δb(Fb/F)-ε,
where δ is the isotopic signature (δ13C) of global
atmospheric CH4, δm, δg, and δb
are the isotopic signatures of the microbial, geological, and biomass burning
sources respectively and ε is the isotopic discrimination of
CH4 oxidation in the atmosphere and soils. Rearrangement of Eqs. (2) and (3) yields
Fb=F(δ-δm+ε)/(δb-δm)-Fg(δg-δm)/(δb-δm).
The total flux F is related to the global CH4 concentration M in steady
state by F=fM/τ, where f is the conversion factor between atmospheric
concentration and mass and τ is the atmospheric lifetime of CH4,
which we assume to be constant. The geological flux can also be assumed to be constant, although its magnitude is disputed (Schwietzke et al., 2016;
Petrenko et al., 2017). The steady-state assumption is appropriate because
we are considering variations over periods longer than the atmospheric
lifetime of CH4, approximately 9 years (Schwietzke et al., 2016).
Equation (4) can then be resolved into the sum of three components: a constant
intercept, a component proportional to M, and a component proportional to the
product δM. Equation (4) also holds, with appropriate adjustment of
units, if the Fb is expressed in normans (normalized anomalies); then all of the fluxes are
relative to the mean value of Fb. We used ordinary linear regression of
charcoal normans with M and δM as predictors to quantify the
relationship between the charcoal data and CH4 isotopic composition.
The inclusion of CH4 concentration in this analysis is essential,
because variations in δ could be brought about irrespective of
biomass burning by variations in Fm, which is generally much larger than
Fb.
Co-evolution of temperature and fire-related emissions over the
period between 2000 and 2014. The temperature data are from the NOAA data
set (NOAAGlobalTemp v4.0.1; doi10.7289/V5FN144H; Vose et al., 2012) and the
emissions data are from GFED4 (Randerson et al., 2015,
www.globalfiredata.org, last access: 16 February 2018). The top panel
(a) shows global temperature (in blue) and emissions (in red) after excluding agricultural areas;
the bottom panel (b) shows temperature (in blue) and emissions (in red) from areas of natural vegetation only,
excluding both deforestation fires and peatland fires.
Calculation of feedback strengths and gain
The global relationship between biomass burning CO2 emissions and
temperature provides an estimate of the strength of the feedback. We define
feedback strength as the equilibrium sensitivity of atmospheric CO2 to
global land temperature in ppm K-1. This can be further converted to
gain (Lashof et al., 1997). Following the convention established by Hansen
et al. (1984), gain (g) is the product of the feedback strength and the
climate sensitivity (i.e. the global mean surface temperature change for a
doubling of CO2 concentration) expressed in K ppm-1. Then the
temperature amplification ΔT/ΔT0, where ΔT is the
actual temperature change and ΔT0 is the reference temperature
change without the feedback, is
ΔT/ΔT0=1/(1-g).
Note that this convention (Hansen et al., 1984) is widely applied in the
literature on terrestrial biogeochemical feedbacks. However, an alternative
convention exists in which the quantity defined in Eq. (5) is called
the gain, while the quantity we call gain is called the feedback factor (see, e.g., Roe, 2009).
The equilibrium sensitivity of atmospheric CO2 concentration to a
change in the biomass burning flux was estimated using a box model, with
parameters derived from either present-day or palaeo-relationships. The
principle is that an increased rate of removal of land carbon due to fire
results in a reduced steady-state carbon storage and a correspondingly
increased atmospheric CO2 content. The change in atmospheric CO2
concentration is given to a good approximation by
ΔC≈(W/NPP)ΔFbAF/2.12,
where ΔC is the change in atmospheric CO2 concentration (ppm),
W is total land ecosystem carbon storage (Pg C), NPP is total land net
primary production (Pg C a-1), ΔFb is the change in biomass
burning carbon flux (Pg C a-1), AF is the airborne fraction (the
fraction of emitted CO2 remaining in the atmosphere), and the factor
2.12 converts Pg C to ppm (http://cdiac.ornl.gov/pns/convert.html, last access: 16 February 2018; Ciais et al., 2014). (The full
derivation of Eq. (6) is given in the Appendix.) For the satellite era,
we related ΔFb (Pg C a-1) statistically to temperature
data. For the pre-industrial era, we related normalized charcoal anomalies
(dimensionless) statistically to temperature data and multiplied them by an
estimate of the long-term mean Fb for the period up to 1600 CE
(3.87 Pg C a-1). This estimate was based on the calibration of the methane
isotope record by Sapart et al. (2012), as follows: we multiplied the
contemporary flux of 2.02 Pg C a-1 (the average of five satellite-based
estimates from Shi et al., 2015) by the ratio of the global biomass burning
CH4 flux inferred for 1–1600 CE (27.4 Tg CH4 a-1) to the
same flux inferred from GFED4s (14.3 Tg CH4 a-1). Since feedback
strength is related to timescale (Roe, 2009), we assumed an AF appropriate
to the centennial timescale (Joos et al., 2013), and standard values for
global net primary production and total carbon storage in vegetation, litter, and non-permafrost soils. The derivation of Eq. (6), and details of
calculations including the uncertainty propagation, are provided in the
Appendix.
Relationship between global fire-related emissions and temperature
over the period between 2000 and 2014. Panel (a) shows the
relationship between global temperature and emissions after excluding
agricultural areas; panel (b) shows the relationship between
temperature and emissions from areas of natural vegetation only, excluding
both deforestation fires and peatland fires.
Indices of pre-industrial global biomass burning trends, 0–1750 CE: (a) normalized
charcoal anomalies, (b)δ13C of CH4
(‰) based on a composite of the data from Ferretti et
al. (2005), Mischler et al. (2009), and Sapart et al. (2012), and (c) CH4
concentration (ppb) from Etheridge et al. (2010). Panel (d) shows
global average temperature anomalies over land (∘C) from Mann et al. (2009).
The plots show the 50-year smoothed record for each indicator,
with 95 % bootstrap confidence intervals; the individual data points for
δ13C, CH4, and land temperature are shown by grey points.
There are too many individual charcoal points to be shown.
ResultsRelationship between biomass burning flux and global average land
temperature during the satellite era
The sensitivity of the MODIS-era biomass burning flux to temperature (Fig. 2)
was obtained by regression of GFED4s annual fluxes against global (annual
average) land temperature data, yielding a slope of 0.71 Pg C K-1 with
a standard error of ±0.34 Pg C K-1 (Fig. 3). Although approaching
statistical significance, this relationship was weak (R2= 0.25; p=0.058). The slope of the relationship, however, was shown to be insensitive to
individual extreme years (see Supplement, Sect. 8).
Estimation of feedback strength during the satellite era
The fitted relationship of annual biomass burning flux to temperature
provides an estimate of the feedback strength of 6.5 ± 3.4 ppm K-1
with respect to global land temperature. We took account of the
greater variability of land versus global mean temperatures by means of a
regression of land versus global mean temperature anomalies for 2000–2014
(Fig. 3a), yielding a slope of 1.364 ± 0.098 K K-1. Correcting
the estimated land-based feedback strength with this slope yielded a
corrected feedback strength of 8.9 ± 4.7 ppm K-1. Assuming a
value of S=2.8 K, the central value for climate sensitivity recently
obtained by a novel emergent-constraint method (Cox et al., 2018), led to
∂T/∂C=S/ (Cln2) = 0.0106 K ppm-1 (evaluated at
C=380 ppm) and an estimated gain of 0.09 ± 0.05. (The uncertainty
of the gain estimate does not include the uncertainty in S, which affects all
estimates of gain but does not affect comparisons of gain made with the same
value of S.)
However, if deforestation and peat fires (which account for 18–28 % of
emissions) were excluded from the calculations (Fig. 3b), no significant
relationship of biomass burning emissions to temperature remained (p=0.476).
Inter-annual variability in tropical deforestation and peatland fires
is well known to be correlated with ENSO (van der Werf et al., 2010),
whereas ENSO-related changes in temperature and precipitation are both
positive and negative across extratropical regions – resulting in
compensatory impacts on total non-anthropogenic fire emissions, which show
no clear general relationship to temperature during the satellite era
(Prentice et al., 2011).
Relationship between methane and charcoal records of biomass
burning
The fitted regression equation relating charcoal normans (dimensionless) to
the concentration of CH4 (Mt at time t, ppb) and the product of the
δ13C of CH4 (δt at time t, ‰) with Mt
(δtMt, ‰ ppb) is
Nt=0.0659+0.00118Mt+0.00004679δtMt
(R2=0.771; F=54.04 with 1 and 32 df; p<0.0001). The
standard errors of the fitted regression coefficients in Eq. (7) are as
follows: ±0.0147 for the intercept, ±0.000 70 ppb-1 for the
coefficient of Mt, and ±0.000 012 37 ‰-1 ppb-1
for the coefficient of δtMt (see Supplement, Sect. 7 for more
details). The Ljung–Box statistic (Ljung and Box, 1978) is 16.9 with 12 df
and p=0.15, i.e. not significant, indicating that pseudoreplication and
the possibility of spurious correlation are absent.
This analysis shows, for the first time, that the charcoal and methane data
sources (Fig. 4) are in good agreement (Fig. 5b). It is therefore
appropriate to use charcoal normans (based on a global compilation, albeit
with some unnevenness is sampling) as an indicator for normalized anomalies
of global biomass burnt.
Relationship between normalized charcoal anomalies and global land
temperature. The data points refer to 50-year binned data. The top panel (a) shows
observed charcoal normans; estimated values based on the linear
regression of charcoal normans against the δ13C of CH4 and
the product of this δ13C value with the concentration of
CH4 are plotted in (b); and estimated values based on the linear
regression of charcoal normans against temperature are plotted in (c). Note
that the slope and intercept of the relationship shown in panel (b) are
necessarily 1.0 and 0.0 respectively – the key point is the goodness of
fit shown between the two data sources after the charcoal data have been
calibrated against the CH4 and CH4 isotopic records.
The ratio r of the coefficient of Mt to the coefficient of δtMt could in principle provide an independent estimate of the
microbial discrimination factor, as δm=ε-r by rearrangement of Eq. (4). However, in practice this calculation
does not provide a strong constraint on δm. Assuming
ε=-6.3 ‰ (Schwietzke et al., 2016) and
with r=25.2± 16.4 ‰ from Eq. (7), δm is estimated as -31.5 ± 16.4 ‰. The
central estimate is small in magnitude compared to typical values around
-60 % (e.g. Sapart et al., 2012), but its standard error is large.
Relationship between charcoal records and global average land
temperature
The fractional sensitivity of the millennium-scale biomass burning flux to
temperature was obtained by regression of charcoal normans against global
land temperature. The fitted regression equation relating anomalies of
charcoal normans and temperature (Fig. 5c) is
Nt=-0.0205+0.158Tt,
where the Nt is charcoal normans (dimensionless) and Tt is the
area-weighted average temperatures (∘C; R2=0.646; F=41.98 with 1 and 23 df; p<0.0001). The standard errors of the fitted
regression coefficients in Eq. (8) are ±0.005 for the intercept,
and ±0.024 K-1 for the coefficient of Tt. The Ljung–Box
statistic is 16.2 with 12 df and p=0.184, i.e. non-significant (see
Supplement, Sect. 6).
Relationship between normalized charcoal anomalies and land
temperature for the (a) northern extratropics, (b) northern tropics, (c) southern tropics, and
(d) southern extratropics. The data points refer to
50-year binned data.
Regional analyses show that the observed strongly positive global-scale
relationship between temperature and normalized charcoal anomalies is
mirrored in the northern extratropics, northern tropics, and southern tropics
(Fig. 6), but not in the southern extratropics. However the Mann et al. (2009)
data set contains relatively few observations from the southern
extratropics and shows an anomalously large temperature decline from 500 to
1500 CE compared to other reconstructions (e.g. Neukom et al., 2014; Gergis
et al., 2016; Supplement, Sect. 11). We reserve judgment as
to whether this regional difference in the relationship is meaningful. In
any case, the land area represented by the southern extratropics is small.
Estimation of feedback strength during the pre-industrial era
Applying an estimated long-term mean value Fb=3.87± 1.94 Pg C a-1
yielded ΔFb= 0.61 ± 0.32 Pg C a-1 K-1. The resulting
estimate of feedback strength is 5.6 ± 3.2 ppm K-1 with respect to
land temperature. A regression of land
versus global mean temperatures based on the 500–1700 CE data in Mann et al. (2009) yielded a slope of 1.146 ± 0.0018 K K-1
(Supplement Sect. 9). Correcting the estimated land-based feedback strength with this slope, and
assuming S=2.8 K as before, led to ∂T/∂C=S/ (Cln2) = 0.0144 K ppm-1
(evaluated at C=280 ppm) and an estimated gain of
0.09 ± 0.05. The uncertainty in this value is dominated by the large
uncertainty assigned to the mean pre-industrial biomass burning flux.
Discussion and conclusions
Our analyses of data from the pre-industrial era yielded an estimate of the
feedback strength of 5.6 ± 3.2 ppm K-1 for land
temperature and a gain of 0.09 ± 0.05. Our analyses for the satellite
era yielded 6.5 ± 3.4 ppm K-1 for land temperature and also a
gain of 0.09 ± 0.05. The agreement between the two gain estimates is
fortuitous, however. The pre-industrial estimate is founded on a strong
relationship between charcoal data and reconstructed temperatures. Its
uncertainty is largely due to uncertainty about the absolute magnitude of
average biomass burning emissions in pre-industrial times. In contrast, the
uncertainty of the satellite-era estimate is largely due to the weakness of
the relationship between emissions and observed temperatures. Moreover, this
relationship is dominated by the well-known correlation of anthropogenic
burning in the tropics with the ENSO cycle. The period for which reliable
satellite-based estimates of biomass burning emissions are available is too
short to have allowed the effects of longer-term climate variability to
emerge, especially given the uncertainties associated with the large
differences between different satellite products (Hantson et al., 2016).
It is unclear whether the magnitude of the fire feedback estimated on the
basis of inter-annual variability should be different from the estimate
obtained based on decadal to centennial variability. The palaeo-record does
not provide a test of this because there are too few annually resolved
charcoal records, while the satellite-era records cover too short a period
to be able to examine longer-term sensitivity. However, even if the
satellite-era data provided a strong constraint on fire feedback, the
estimate of gain based on pre-industrial, centennial-scale climate
variability would likely still be more relevant to long-term climate
projections.
Many of the influences on fire have changed dramatically between
pre-industrial and recent times. The geographic pattern of fire frequency
shows an unambiguous decline with human population density, a relationship
that holds across more than 4 orders of magnitude of population density
(Bistinas et al., 2014; Knorr et al., 2014). Moreover, global biomass
burning has declined precipitously since its peak in the mid-nineteenth
century, as shown by both charcoal data (Marlon et al., 2008, 2016) and carbon
monoxide isotopes in ice and contemporary air (Wang et al.,
2010). On the other hand, tropical deforestation and burning of peat
substrates yield intense, localized pyrogenic sources of CO2 that
closely co-vary with inter-annual variation in the duration and intensity of
the dry season (van der Werf et al., 2010). Our estimate of gain based on
pre-industrial, centennial-scale climate variability is likely more relevant
to long-term climate projections, but any realistic estimation of future
fire risks and feedbacks must consider the pervasive effects of human
settlement and land use (Knorr et al., 2014). It is also possible that the
influence of temperature variability on inter-annual timescales might
generally differ from its influence on decadal-to-millennial timescales, but
we cannot establish this from currently available palaeo-data because there
is too little annually resolved information, while the interval for which we
have satellite data is too short even to resolve decadal variability.
Charcoal abundances have generally been interpreted as a measure of “fire
activity” or relative changes in the quantity of burned biomass (e.g. Power
et al., 2008; Harrison et al., 2010; Daniau et al., 2012; Marlon et al.,
2016). There have been some attempts to quantify the relationship between
charcoal abundance and burnt area or total biomass consumed at a local scale
(see, e.g., Peters and Higuera, 2007; Duffin et al., 2016; Leys et al., 2017).
These analyses, however, show a strong dependency on vegetation type and
fire regime and the need to apply calibrations accounting for charcoal
source area in the same way as for the interpretation of pollen abundances
(Prentice, 1985: Sugita, 1994). Such calibrations have been made for Europe
(Adolf et al., 2018) but not for other regions. Our analyses establish for
the first time that there is a good relationship (R2=0.77) between
global charcoal abundance, expressed as normalized anomalies, and the
methane and methane-isotopic record from ice cores. Since emissions reflect
the amount of biomass consumed by fire, which in turn is influenced by area
burnt and fire intensity, these analyses support the idea that the
sedimentary charcoal record – when synthesized at continental to global
scales – can provide quantitative evidence for changes in the global
biomass burning carbon flux. Establishing the quantitative relationship
between charcoal abundance and fire emissions is key to being able to exploit
the continued expansion of the spatial and temporal coverage of charcoal
records (Marlon et al., 2016) to examine regional changes in fire regimes on
multiple timescales.
The strength of the global land climate–carbon-cycle feedback has been
assessed by Arora et al. (2013) on the basis of nine CMIP5 Earth system
models. Five models that do not explicitly represent fire yield feedback
strengths (after converting Pg C to ppm and multiplying them by the airborne
fraction) in the range 6.8 to 19.9 ppm K-1 with a median of 17.5 ppm K-1.
Of four models that do represent fire, two yield values in the
same range; the other two (sharing the same land model) yield lower values
but have been shown to greatly underestimate the feedback based on the
observed relationship between tropical land temperatures and the rate of
increase in atmospheric CO2 concentration (Wenzel et al., 2014). Our
global estimate of the biomass burning contribution as 5.6 ± 3.2 ppm K-1,
based on the pre-industrial period, suggests that the contribution
of fire emissions to the climate–carbon-cycle feedback is substantial. Our
estimate may even be conservative. Sapart et al. (2012) estimated the
intertemporal coefficient of variation in the biomass burning CH4 flux
to be 7.3 % for the period 1–1600 CE, compared to only 2.9 % in the
charcoal anomalies.
Although some of the models in the assessment by Arora et al. (2013)
included fire as an interactive process, none considered deforestation or
peat fires. A substantial component of the total contemporary land
climate–carbon-cycle feedback appears to be attributable to anthropogenic
fires in the tropics and their spatially coherent association with ENSO
variability. This is in contrast with extratropical fire regimes, which show
regionally asynchronous responses to climate variability (Prentice et al.,
2011), and the response of net ecosystem exchange to warming, which is
asymmetrical between low and high latitudes (Wenzel et al., 2014). The
importance of deforestation and peatland fires in driving fire feedback in
the recent decades suggests that measures to protect tropical forests and
peatlands could appreciably reduce the magnitude of the climate–carbon-cycle
feedback.
The climate–carbon-cycle feedback is an important benchmark for ESMs.
Despite growing interest in the environmental and human drivers and impacts
of fire (Bowman et al., 2009, 2011; Harrison et al., 2010;
Fischer et al., 2016), the global-scale contribution of biomass burning to
the climate–carbon-cycle feedback has been poorly quantified. Our analyses
provide an independent estimate of this feedback, illustrating the use of
the palaeo-record to estimate Earth system quantities that may be difficult
or impossible to derive from contemporary observations.
All the data used in the analyses are public and
available from the sites given in the text or references. Our analyses are
fully documented in the Supplement.
The box model, parameter estimates, and their uncertainties
In steady state, carbon inputs to biomass and subsequently (via litter
production) to soil organic matter, corresponding to net primary production
(NPP), must be balanced by outputs: heterotrophic respiration, RH, and
biomass burning, Fb. Here we designate rates of carbon transfer by
heterotrophic respiration and biomass burning respectively as kr and
kb, such that kb=Fb/W; kb*=Fb*/W* (where the
asterisk denotes new steady-state values after a change in the burning
rate); then kr=kr*=RH/W= (NPP -Fb)/W= (NPP -Fb*) /W*,
assuming that the impact of an altered fire frequency on NPP is
small compared to its effect on W (Martin Calvo and Prentice, 2015). Hence,
W*/W= (NPP -Fb*) / (NPP -Fb) and upon rearrangement
ΔW=-W.ΔFb/(NPP-Fb),
where ΔW=W*-W and ΔFb=Fb*-Fb or, to a
close approximation (given Fb≪ NPP),
ΔW≈-W.ΔFb/NPP.
This calculation is insensitive to CO2 effects on NPP, as an increase
in NPP in steady state implies a proportionate increase in W.
Global terrestrial biosphere C is given by Ciais et al. (2014) as the sum of
450–650 Pg C (vegetation C) and 1500–2400 (soil C),
i.e. 550 ± 100 Pg C and 1950 ± 450 Pg C respectively – yielding a combined
uncertainty of ± 461 Pg C (18.4 %) For global NPP, the two bottom–up
estimates given by Prentice et al. (2001) are 59.9 and 62.6 Pg C a-1,
yielding a mean of 61.25 and a standard error (n=2) of ± 1.35 Pg C a-1 (2.2 %).
We therefore assigned values of W=2500± 461 Pg C and NPP = 61.25 ± 1.35 Pg C a-1.
For contemporary biomass burning C emissions (Shi et al., 2015; Table 3), five
satellite-derived estimates together provide a global mean of
7391.7 Tg CO2 a-1 (2.02 Pg C a-1) with a standard
deviation (n=5) of ±1291.2 Tg CO2 a-1, corresponding to
a standard error of ±0.157 Pg C a-1 (7.8 %). We therefore
assigned Fb= 2.02 ± 0.157 Pg C a-1 for the
satellite era. For the pre-industrial era, we estimated the long-term mean
biomass burning C flux as the product of the contemporary flux of
2.02 Pg C a-1 (Shi et al., 2015) with the ratio of the global biomass
burning CH4 flux inferred from methane isotope data for the period
1–1600 CE (27.4 Tg CH4 a-1) to the same flux inferred from
GFED4s (14.3 Tg CH4 a-1) by Sapart et al. (2012), yielding
Fb= 3.87 Pg C a-1. However, while Sapart et al. (2012)
assigned an uncertainty of only ± 2.8 Tg CH4 a-1
(10 %) to their estimate of global biomass burning CH4 flux, we
inflated the uncertainty of our estimate of Fb to
±1.94 Pg C a-1 (50 %) in order to include additional
potential sources of error, which include variability of the isotopic
fractionation factors and of the emission factor for CH4 with respect to
CO2.
For the centennial-scale airborne fraction (AF in Eq. 6) we adopted the
estimate of 0.476 ± 0.057 (12.0 %) obtained by Joos et al. (2013).
This estimate was derived from multiple models performing identical
pulse-response experiments. The mean value here is the multi-model mean
(converted from units of years to fractions by dividing by the timescale),
and the uncertainties are 1 standard deviation of the variation among
models. The mean value is close to the empirical estimate of 0.44 given by
Ciais et al. (2014).
Conversion of the feedback strength (∂C/∂T) into a gain
requires a further assumption about the climate sensitivity (S), defined as
the equilibrium change in global mean temperature for a doubling of
atmospheric CO2. We have used S=2.8 K, the central estimate provided
by Cox et al. (2018).
The Supplement related to this article is available online at https://doi.org/10.5194/esd-9-663-2018-supplement.
SPH, ICP, PJB, and SK designed and performed the
analyses. SPH and ICP wrote the first draft of the manuscript and all
authors contributed to the final version.
The authors declare that they have no conflict of interest.
Acknowledgements
This work was inspired by discussions at workshops of the
Global Palaeofire Working Group in 2012 and 2013 and partially funded by
the University of Reading. Sandy P. Harrison acknowledges the support from the ERC-funded
project GC2.0 (Global Change 2.0: Unlocking the past for a clearer future,
grant number 694481). This research is a contribution to the AXA Chair
Programme in Biosphere and Climate Impacts and the Imperial College
initiative on Grand Challenges in Ecosystems and the Environment (ICP). Patrick J. Bartlein acknowledges funding from the US National Science Foundation Geography and
Spatial Science Program (1435744). We thank Sam Rabin for his review of the
manuscript.
Edited by: Somnath Baidya Roy
Reviewed by: Sam Rabin and four anonymous referees
References
Adolf, C., Wunderle, S., Colombaroli, D., Weber, H., Gobet, E., Heiri, O.,
van Leeuwen, J. F. N., Bigler, C., Connor, S. E., Gałka, M., La Mantia,
T., Makhortykh, S., Svitavska-Svobodova, H., Vannière, B., and Tinner, W.:
The sedimentary and remote-sensing reflection of biomass burning in Europe,
Global Ecol. Biogeog., 27, 199–212, https://doi.org/10.1111/geb.12682, 2018.Andela, N., Morton, D. C., Giglio, L., Chen, Y., van der Werf, G. R.,
Kasibhatla, P. S., DeFries, R. S., Collatz, G. J., Hantson, S., Kloster, S.,
Bachelet, D., Forrest, M., Lasslop, G., Li, F., Mangeon, S., Melton, J. R.,
Yue, C., and Randerson, J. T.: A human-driven decline in global burned area,
Science, 356, 1356–1362, 10.1126/science.aal4108, 2017.Andreae, M. O. and Merlet, P.: Emission of trace gases and aerosols from
biomass burning, Glob. Biogeochem. Cycl., 15, 955–966,
10.1029/2000GB001382, 2001.Arneth, A. Harrison, S. P., Zaehle, S., Tsigaridis, K.,Menon, S.,
Bartlein, P. J., Feichter, H., Korhola, A., Kulmala, M., O'Donnell, D.,
Schurgers, S., Sorvari, S., and Vesala, T.: Terrestrial biogeochemical
feedbacks in the climate system, Nat. Geosci., 3, 525–532, 10.1038/ngeo905, 2010.Arora, V. K, Boer, G. J., Friedlingstein, P., Eby, M., Jones, C. D.,
Christian, J. R., Bonan, G., Bopp, L., Brovkin, V., Cadule, P., Hajima, T.,
Ilyina, T., Lindsay, K., Tjiputra, J. F., and Wu, T.: Carbon–concentration
and carbon–climate feedbacks in CMIP5 Earth System Models, J. Climate, 26,
5289–5314, 10.1175/JCLI-D-12-00494.1, 2013.Ballantyne, A., Smith, W., Anderegg, W., Kauppi, P., Sarmiento, J., Tans,
P., Shevliakova, E., Pan, Y., Poulter, B., Anav, A., Friedlingstein, P.,
Houghton, R., and Running, S.: Accelerating net terrestrial carbon uptake
during the warming hiatus due to reduced respiration, Nat. Clim. Change,
7, 148–152, 10.1038/nclimate3204, 2017.Barker, J. F. and Fritz, P.: Carbon isotope fractionation during microbial
methane oxidation, Nature, 293, 289–291, 10.1038/293289a0, 1981.Bistinas, I., Harrison, S. P., Prentice, I. C., and Pereira, J. M. C.: Causal
relationships versus emergent patterns in the global controls of fire frequency,
Biogeosciences, 11, 5087–5101, 10.5194/bg-11-5087-2014, 2014.Blarquez, O., Vanniere, B. , Marlon, J. R., Daniau, A-L., Power, M. J.,
Brewer, S., and Bartlein, P. J.: paleofire: An R package to analyse
sedimentary charcoal records from the Global Charcoal Database to
reconstruct past biomass burning, Comp. Geosci., 72, 255–261, 10.1016/j.cageo.2014.07.020, 2014.
Boucher, O., Randall, D., Artaxo, P., Bretherton, C., Feingold, G., Forster,
P., Kerminen, V.-M., Kondo, Y., Liao, H., Lohmann, U., Rasch, P., Satheesh,
S. K., Sherwood, S., Stevens B., and Zhang, X. Y.: Climate and aerosols. 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. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.
K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M.,
Cambridge University Press, Cambridge, UK and New York, NY, USA, 2013.Bowman, D. M. J. S., Balch, J. K., Artaxo, P., Bond, W. J., Cochrane, M. A.,
D'Antonia, C. M., DeFries, R. S., Doyle, J. C., Harrison, S. P.,
Johnston, F. H., Keeley, J. E., Krawchuk, M. A.,
Kull, C. A., Marston, J. B., Moritz, M. A., Prentice, I. C., Roos, C.
I., Scott, A. C., Swetnam, T. W., van der Werf, G. R., and Pyne, S. J.,:
Fire in the earth system, Science, 324, 481–484, 10.1126/science.1163886, 2009.Bowman D. M. J. S., Balch, J., Artaxo, P., Bond, W. J., Cochrane, M. A.,
D'Antonio, C. M, DeFries, R., Johnston, F. H., Keeley, J. E., Krawchuk, M.
A. Kull, C. A., Mack, M., Moritz, M. A., Pyne, S., Roos, C. I., Scott, A.
C., Sodhi, N. S., Swetnam, T. W., and Whittaker, R.: The human dimension of
fire regimes on Earth, J. Biogeog., 38, 2223–2236,
10.1111/j.1365-2699.2011.02595.x, 2011.Carslaw, K. S., Lee, L. A., Reddington, C. L., Pringle, K. J., Rap, A.,
Forster, P. M., Mann, G. W., Spracklen, D. V., Woodhouse, M. T., Regayre, L.
A., and Pierce, J. R.: Large contribution of natural aerosols to uncertainty
in indirect forcing, Nature, 503, 67–74,
10.1038/nature12674, 2013.
Ciais, P., Sabine, C., Bala, G., Bopp, L., Brovkin, V., Canadell, J.,
Chhabra, A., DeFries, R., Galloway, J., Heimann, M., Jones, C., Le
Quéré, C., Myneni, R. B., Piao, S., and Thornton, P.: Carbon and
other biogeochemical cycles. 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. F., Qin, D., Plattner, G.-K., Tignor,
M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P.
M., Cambridge University Press, Cambridge, UK and New York, USA, 2013.Cox, P. M., Huntingford, C., and Williamson, M. S.: Emergent constraint on
equilibrium climate sensitivity from global temperature variability, Nature,
553, 319–322, 10.1038/nature25450, 2018.Cox, P. M., Pearson, D., Booth, B. B., Friedlingstein, P., Huntingford, C.,
Jones, C. D., and Luke, C. M.: Sensitivity of tropical carbon to climate
change constrained by carbon dioxide variability, Nature, 494, 341–344,
10.1038/nature11882, 2013.Daniau, A.-L., Bartlein, P. J., Harrison, S. P., Prentice, I. C., Brewer,
S., Friedlingstein, P., Harrison-Prentice, T. I., Inoue, J., Marlon, J. R.,
Mooney, S., Power, M. J., Stevenson, J., Tinner, W., Andrič, M.,
Atanassova, J., Behling, H., Black, M., Blarquez, O., Brown, K. J.,
Carcaillet, C., Colhoun, E., Colombaroli, D., Davis, B. A. S., D'Costa, D.,
Dodson, J., Dupont, L., Eshetu, Z., Gavin, D. G., Genries, A., Gebru, T.,
Haberle, S., Hallett, D. J., Horn, S., Hope, G., Katamura, F., Kennedy, L.,
Kershaw, P., Krivonogov, S., Long, C., Magri, D., Marinova, E., McKenzie, G.
M., Moreno, P. I., Moss, P., Neumann, F. H., Norström, E., Paitre, C.,
Rius, D., Roberts, N., Robinson, G., Sasaki, N., Scott, L., Takahara, H.,
Terwilliger, V., Thevenon, F., Turner, R. B., Valsecchi, V. G.,
Vannière, B., Walsh, M., Williams, N., and Zhang, Y.: Predictability of
biomass burning in response to climate changes, Glob. Biogeochem. Cyc., 26,
10.1029/2011GB004249, 2012.Duffin, K. I., Gillson, L., and Willis, K. J.: Testing the sensitivity of
charcoal as an indicator of fire events in savanna environments:
quantitative predictions of fire proximity, area and intensity, The
Holocene, 18, 279–291, 10.1177/0959683607086766, 2008.Etheridge, D. M. et al.: Law Dome Ice Core 2000-Year CO2, CH4, and
N2O Data,IGBP PAGES/World Data Center for Paleoclimatology, Data Contribution
Series # 2010-070, NOAA/NCDC Paleoclimatology Program, Boulder
CO, USA, available at:
ftp://ftp.ncdc.noaa.gov/pub/data/paleo/icecore/antarctica/law/law2006.xls),
(last access: 16 February 2018), 2010.Ferretti, D. F., Miller, J. B., White, J. W. C., Etheridge, D. M., Lassey,
K. R., Lowe, D. C., MacFarling Meure, C. M., Dreier, M. F., Trudinger, C.
M., van Ommen, T. D., and Langenfelds, R. L.: Unexpected changes to the
global methane budget over the past 2000 years, Science, 309, 1714–1717,
10.1126/science.1115193, 2005.Fischer, A. P., Spies, T. A., Steelman, T. A., Moseley, C., Johnson, B. R.,
Bailey, J. D., Ager, A. A., Bourgeron, P., Charnley, S., Collins, B. M.,
Kline, J. D., Leahy, J. E., Littell, J. S., Millington, J. D. A.,
Nielsen-Pincus, M., Olsen, C. S., Paveglio, T. B., Roos, C. I., Steen-Adams,
M. M., Stevens, F. R., Vukomanovic, J., White, E. M., and Bowman, D. M.:
Wildfire risk as a socioecological pathology, Front. Ecol. Environ., 14,
276–284, 10.1002/fee.1283, 2016.Gergis, J., Neukom, R., Gallant, A. J., and Karoly, D. J.: Australasian
temperature reconstructions spanning the Last Millennium, J. Climate, 29,
5365–5392, 10.1175/JCLI-D-13-00781.1, 2016.Giglio, L., Randerson, J. T., and van der Werf, G. R.: Analysis of daily,
monthly, and annual burned area using the fourth-generation global fire
emissions database (GFED4), J. Geophys. Res-Biogeo., 118, 317–328,
10.1002/jgrg.20042, 2013.
Granger, C. W. J. and Newbold, P.: Forecasting Economic Time Series,
Academic Press, 2nd edition, 1986.Hansen, J., Lacis, A., Rind, D., Russell, G., Stone, P., Fung, I., Ruedy,
R., and Lerner, J.: Climate sensitivity: Analysis of feedback mechanisms,
in: Climate Processes and Climate Sensitivity, edited by: Hansen, J. E. and Takahashi,
T., AGU Geophysical Monograph 29, Maurice Ewing, Vol. 5.
American Geophysical Union, 130–163, 1984.Hantson, S., Arneth, A., Harrison, S. P., Kelley, D. I., Prentice, I. C.,
Rabin, S. S., Archibald, S., Mouillot, F., Arnold, S. R., Artaxo, P., Bachelet, D.,
Ciais, P., Forrest, M., Friedlingstein, P., Hickler, T., Kaplan, J. O., Kloster, S.,
Knorr, W., Lasslop, G., Li, F., Mangeon, S., Melton, J. R., Meyn, A., Sitch, S.,
Spessa, A., van der Werf, G. R., Voulgarakis, A., and Yue, C.: The status and
challenge of global fire modelling, Biogeosciences, 13, 3359–3375, 10.5194/bg-13-3359-2016, 2016.Harrison, S. P., Marlon, J., and Bartlein, P. J.: Fire in the Earth System,
edited by: Dodson, J., Changing Climates, Earth Systems and Society,
Springer-Verlag, 21–48, 10.1007/978-90-481-8716-4_3, 2010.Held, I. M. and Soden, B. J.: Robust responses of the hydrological cycle to
global warming, J. Climate, 19, 5686–5699, 10.1175/JCLI3990.1, 2006.Hurlbert, S. H.: Pseudoreplication and the design of ecological field
experiments, Ecological Monographs, 54, 187–211, 10.2307/1942661, 1984.Joos, F., Roth, R., Fuglestvedt, J. S., Peters, G. P., Enting, I. G., von
Bloh, W., Brovkin, V., Burke, E. J., Eby, M., Edwards, N. R., Friedrich, T.,
Frölicher, T. L., Halloran, P. R., Holden, P. B., Jones, C., Kleinen, T.,
Mackenzie, F. T., Matsumoto, K., Meinshausen, M., Plattner, G.-K., Reisinger,
A., Segschneider, J., Shaffer, G., Steinacher, M., Strassmann, K., Tanaka,
K., Timmermann, A., and Weaver, A. J.: Carbon dioxide and climate impulse
response functions for the computation of greenhouse gas metrics: a
multi-model analysis, Atmos. Chem. Phys., 13, 2793–2825,
10.5194/acp-13-2793-2013, 2013.Jung, M., Reichstein, M., Schwalm, C. R., Huntingford, C., Sitch, S.
Ahlström, A., Arneth, A., Cams-Valls, G., Ciais, P., Friedlingstein, P.,
Gans, F., Ichii, K., Jain, A. K., Kato, E., Papale, D., Poulter, B., Raduly,
B., Rödenbeck, C., Tramontana, G., Viovy, N., Wang, Y.-P., Weber, U.,
Zaehle, S., and Zeng, N.: Compensatory water effects link yearly global land
CO2 sink changes to temperature, Nature, 541, 516–520, 10.1038/nature20780, 2017.Keenan, T. F., Prentice, I. C., Canadell, J. G., Williams, C. A., Wang, H.,
Raupach, M., and Collatz, G. J.: Recent pause in the growth rate of
atmospheric CO2 due to enhanced terrestrial uptake, Nature Comms., 13428, 10.1038/ncomms13428, 2016.Knorr, W., Kaminski, T., Arneth, A., and Weber, U.: Impact of human population density
on fire frequency at the global scale, Biogeosciences, 11, 1085–1102,
10.5194/bg-11-1085-2014, 2014.Krawchuk, M., Moritz, M., Parisien, M.-A., Van Dorn, J., and Hayhoe, K.:
Global pyrogeography: The current and future distribution of wildfire, PLoS
ONE, 4, e5102, 10.1371/journal.pone.0005102, 2009.Lashof, D. A., DeAngelo, B. J., Saleska, S. R., and Harte, J.: Terrestrial
ecosystem feedbacks to global climate change, Annu. Rev. Energy
Environ, 22, 75–118, 10.1146/annurev.energy.22.1.75, 1997.Lee, L. A., Reddington, C. L., and Carslaw, K. S.: On the relationship
between aerosol model uncertainty and radiative forcing uncertainty, PNAS,
113, 5820–5827, 10.1073/pnas.1507050113, 2016.Leys, B. A., Commerford, J. L., and McLauchlan, K. K.: Reconstructing
grassland fire history using sedimentary charcoal: Considering count, size
and shape, PLOS ONE, 12, e0176445, 10.1371/journal.pone.0176445, 2017.Li, G., Harrison, S. P., Bartlein, P. J., Izumi, K., and Prentice, I. C.,
Precipitation scaling with temperature in warm and cold climates: an
analysis of CMIP5 simulations, Geophys. Res. Lett., 40, 4018–4024, 10.1002/grl.50730, 2013.Ljung, G. and Box, G. E. P.: On a measure of lack of fit in time series
models, Biometrics, 65, 297–303, 10.1093/biomet/65.2.297,
1978.
Loader, C.: Local Regression and Likelihood, Springer, 1999.Loader, C.: locfit: Local Regression, Likelihood and Density Estimation, R
package version 1.5-9.1, available at: https://CRAN.R-project.org/packagelocfit, (last access: 16 February 2018), 2013.Mann, M. E., Zhang, Z., Rutherford, S., Bradley, R. S., Hughes, M. K.,
Shindell, D., Ammann, C., Faluvegi, G., and Ni, F.: Global signatures and
dynamical origins of the Little Ice Age and Medieval Climate Anomaly,
Science, 326, 1256–1260, 10.1126/science.1177303, 2009.Marlon, J., Bartlein, P. J., Carcaillet, C., Gavin, D. G., Harrison, S. P.,
Higuera, P. E., Joos, F., Power, M., and Prentice, I. C.: Climate and
human influences on global biomass burning over the past two millennia,
Nat. Geosci., 1, 697–702, 10.1038/ngeo313, 2008.Marlon, J. R., Bartlein, P. J., Walsh, M. K., Harrison, S. P., Brown, K. J.,
Edwards, M. E., Higuera, P. E., Power, M. J., Anderson, R. S., Briles, C.,
Brunelle, A., Carcaillet, C., Daniels, M., Hu, F. S., Lavoie, M., Long, C.,
Minckley, T., Richard, P. J. H., Scott, A. C., Shafer, D. S., Tinner, W.,
Umbanhower, C. E. Jr., and Whitlock, C.: Wildfire responses to abrupt
climate change in North America, PNAS, 10.1073/pnas.0808212106, 2009.
Marlon, J. R., Bartlein, P. J. Daniau, A-L., Harrison, S. P., Maezumi, S.
Y., Power, M. J., Tinner, W., and Vannière, B.: Global biomass burning:
a synthesis and review of Holocene paleofire records and their controls,
Quat. Sci. Rev., 65, 5–25, 2013.Marlon, J. R., Kelly, R., Daniau, A.-L., Vannière, B., Power, M. J., Bartlein, P.,
Higuera, P., Blarquez, O., Brewer, S., Brücher, T., Feurdean, A., Romera, G. G.,
Iglesias, V., Maezumi, S. Y., Magi, B., Courtney Mustaphi, C. J., and Zhihai, T.:
Reconstructions of biomass burning from sediment-charcoal records to improve
data-model comparisons, Biogeosciences, 13, 3225–3244, 10.5194/bg-13-3225-2016, 2016.Martin Calvo, M. and Prentice, I. C.: Effects of fire and CO2 on
biogeography and primary production in glacial and modern climates, New
Phytol., 208, 987–994, 10.1111/nph.13485, 2015.Mischler, J. A., Sowers, T. A., Alley, R. B., Battle, M., McConnell, J. R.,
Mitchell, L., Popp, T., Sofen, E., and Spencer, M. K.: Carbon and hydrogen
isotopic composition of methane over the last 1000 years, Glob. Biogeochem.
Cy., 23, GB4024, 10.1029/2009GB003460, 2009.
Mooney, S., Harrison, S. P., Bartlein, P. J., Daniau A.-L., Stevenson,
J., Brownlie, K., Buckman, S., Cupper, M., Luly, J., Black, M.,
Colhoun, E., D'Costa, D., Dodson, J., Haberle, S., Hope, G. S., Kershaw, P.,
Kenyon, C., McKenzie., M., and Williams, N.: Late Quaternary fire regimes of
Australasia, Quat. Sci. Rev., 30, 28–46, https://doi.org/10.1016/j.quascirev.2010.10.010,
2011.Neukom, R., Gergis, J., Karoly, D. J., Wanner, H., Curran, M., Elbert, J.,
González-Rouco, F., Linsley, B. K. Moy, A. D., Mundo, I., Raible, C. C.,
Steig, E. J., van Ommen, T., Vance, T., Villalba, R., Zinke, J., and Frank,
D.: Inter-hemispheric temperature variability over the last millennium, Nat.
Clim. Change, 4, 362–367, 10.1038/nclimate2174, 2014.PAGES 2k Consortium: Continental-scale temperature variability during the
past two millennia, Nat. Geosci., 6, 339–346, 10.1038.ngeo1797, 2013.Peters, M. E. and Higuera, P. E.: Quantifying the source area of
macroscopic charcoal with a particle dispersal model, Quat. Res., 67,
304–10, 10.1016/j.yqres.2006.10.004, 2007.
Petrenko, V. V., Smith, A. M., Schaefer, H., Riedel, K., Brook, E.,
Baggenstos, D., Harth, C., Hua, Q., Buizert, C., Schilt, A., Fain, X.,
Mitchell, L., Bauska, T., Orsi, A., Weiss, R. F., and Severinghaus, J. P.:
Minimal geological methane emissions during the Younger Dryas–Preboreal
abrupt warming event, Nature, 548, 443–446, 2017.Power, M.J., Ortiz, N., Marlon, J., Bartlein, P. J., Harrison, S. P., Mayle,
F., Ballouche, A., Bradshaw, R., Carcaillet, C., Cordova, C., Mooney, S.,
Moreno, P., Prentice, I. C., Thonicke, K., Tinner, W., Whitlock, C., Zhang,
Y., Zhao, Y., Anderson, R. S., Beer, R., Behling, H., Briles, C., Brown, K,
Brunelle A., Bush, M., Clark, J., Colombaroli, D., Chu, C. Q., Daniels, M.,
Dodson, J., Edwards, M. E., Fisinger, W., Gavin, D. G., Gobet, E., Hallett,
D. J., Higuera, P., Horn, S., Inoue, J., Kaltenrieder, P., Kennedy, L.,
Kong, Z. C., Long, C., Lynch, J., Lynch, B., McGlone, M., Meeks, S., Meyer,
G., Minckley, T., Mohr, J., Noti, R., Pierce, J., Richard, P., Shuman, B.
J., Takahara, H., Toney, J., Turney, C., Umbanhower, C., Vandergoes, M.,
Vanniere, B., Vescovi, E., Walsh, M., Wang, X., Williams, N., Wilmshurst,
J., and Zhang, J. H.: Changes in fire activity since the LGM: an assessment
based on a global synthesis and analysis of charcoal data, Clim. Dynam., 30, 887–907, 10.1007/s00382.00.0334x, 2008.Power, M. J., Marlon, J. R., Bartlein, P. J., and Harrison, S. P.: Fire
history and the Global Charcoal Database: A new tool for hypothesis testing
and data exploration, Palaeoclim., Palaeogeog. Palaeoecol., 291, 52–59, 10.1016/j.palaeo.2009.09.014, 2010.Prentice, I. C.: Pollen representation, source area, and basin size: toward
a unified theory of pollen analysis, Quat. Res., 23, 76–86, 10.1016/0033-5894(85)90073-0, 1985.
Prentice, I. C., Farquhar, G. D., Fasham, M. J. R., Goulden, M. L., Heimann,
M., Jaramillo, V. J., Kheshgi, H. S., Le Quéré, C., Scholes, R. J.,
and Wallace, D. W. R.: The carbon cycle and atmospheric carbon dioxide. In:
Climate Change 2001: The Scientific Basis, edited by: Houghton, J. T., Ding, Y., Griggs, D.
J., Noguer, M., van der Linden, P. J., Dai, X., Maskell, K., and Johnson, C.
A., Cambridge University Press, Cambridge, 183–237, 2001.Prentice, I. C., Kelley, D. I., Foster, P. N. Friedlingstein, P.,
Harrison, S. P., and Bartlein, P. J.: Modeling fire and the
terrestrial carbon balance, Glob. Biogeochem. Cy., 25, GB3005,
10.1029/2010GB003906, 2011.R Core Team. R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria,
https://www.r-project.org/, (last access: 16 February 2018), 2016.Randerson, J. T., van der Werf, G. R., Giglio, L., Collatz, G. J., and
Kasibhatla, P. S.: Global Fire Emissions Database, Version 4, (GFEDv4), ORNL
DAAC, Oak Ridge, Tennessee, USA, 10.3334/ORNLDAAC/1293, 2015.Roe, G.: Feedbacks, timescales, and seeing red, Annu. Rev. Earth Planet.
Sci., 37, 93–115, 10.1146/annurev.earth.061008.134734, 2009.Rubino, M., D'Onofrio, A., Seki, O., and Bendle, J. A.: Ice-core records of
biomass burning, Anthrop. Rev., 3, 140–162,
10.1177/2053019615605117, 2016.Sapart, C. J., Monteil, G., Prokopiou, M., van der Wal, R. S. W., Kaplan, J.
O., Sperlich, P., Krumhardt, K. M., van der Veen, C., Houweling, S., Krol,
M. C., Bluneir, T., Sowers, T., Martinerie, P., Witrant, E., Dahl-Jensen,
D., and Röckmann, T.: Natural and anthropogenic variations in methane
sources during the past two millennia, Nature, 490, 85–88, 10.1038/nature11461, 2012.Schwietzke, S. Sherwood, O. A., Bruhwiler, L. M. P., Miller, J. B., Etiope,
G., Dlugokencky, E. J., Michel, S. E., Arling, V. A., Vaughn, B. H., White,
J. W. C., and Tans, P. P.: Upward revision of global fossil fuel methane
emissions based on isotope database, Nature, 538, 88–91, 10.1038/nature19797, 2016.Shi, Y., Matsunaga, T., Saito, M., Yamaguchi, Y., and Chen, X.: Comparison
of global inventories of CO2 emissions from biomass burning during
2002–2011 derived from multiple satellite products, Environ. Poll., 206,
479–487, 10.1016/j.envpol.2015.08.009, 2015.Sugita, S.: Pollen representation of vegetation in Quaternary sediments:
Theory and method in patchy vegetation, J. Ecol., 82, 881–897, 10.2307/2261452, 1994.
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 der Werf, G. R., Peters, W., van Leeuwen, T. T., and Giglio, L.: What could have
caused pre-industrial biomass burning emissions to exceed current rates?, Clim. Past, 9,
289–306, 10.5194/cp-9-289-2013, 2013.van der Werf, G. R., Randerson, J. T., Giglio, L., van Leeuwen, T. T., Chen, Y.,
Rogers, B. M., Mu, M., van Marle, M. J. E., Morton, D. C., Collatz, G. J., Yokelson,
R. J., and Kasibhatla, P. S.: Global fire emissions estimates during 1997–2016, Earth
Syst. Sci. Data, 9, 697–720, 10.5194/essd-9-697-2017, 2017.Vose, R. S., Arndt, D., Banzon, V. F., Easterling, D. R., Gleason, B.,
Huang, B., Kearns, E., Lawrimore, J. H., Menne, M. J., Peterson, T. C.,
Reynolds, R. W., Smith, T. M. Williams, C. N., Jr., and Wuertz, D. L.:
NOAA's merged land-ocean surface temperature analysis, B. Am. Meterol.
Soc., 93, 1677–1685, 10.1175/BAMS-D-11-00241.1, 2012.Wang, Z., Chappellaz, J., Park, K., and Mak, J. E.: Large variations in
southern hemisphere biomass burning during the last 650 years, Science, 330,
1663–1666, 10.1126/science.1197257, 2010.Ward, D. S., Kloster, S., Mahowald, N. M., Rogers, B. M., Randerson, J. T., and Hess,
P. G.: The changing radiative forcing of fires: global model estimates for past,
present and future, Atmos. Chem. Phys., 12, 10857–10886, 10.5194/acp-12-10857-2012, 2012.Wenzel, S., Cox, P. M., Eyring, V., and Friedlingstein, P.: Emergent
constraints on climate-carbon cycle feedbacks in the CMIP5 Earth system
models, J. Geophys. Res.-Biogeosci., 119, 794–807,
10.1002/2013JG002591, 2014.