The Agriculture, Forestry and Other Land Use (AFOLU) sector contributes with
ca. 20–25 % of global anthropogenic emissions (2010), making it a key
component of any climate change mitigation strategy. AFOLU estimates,
however, remain highly uncertain, jeopardizing the mitigation effectiveness
of this sector. Comparisons of global AFOLU emissions have shown divergences
of up to 25 %, urging for improved understanding of the reasons behind
these differences. Here we compare a variety of AFOLU emission datasets and
estimates given in the Fifth Assessment Report for the tropics (2000–2005)
to identify plausible explanations for the differences in (i) aggregated
gross AFOLU emissions, and (ii) disaggregated emissions by sources and gases
(CO
Modelling studies suggest that, to keep the global mean temperature increase
to less than 2
The Agriculture, Forestry, and other Land Use (AFOLU) sector roughly
contributes a quarter (10–12 Pg CO
Currently, data on AFOLU emissions are available through national greenhouse
gas inventories, which are submitted to the United Nations Framework
Convention on Climate Change (UNFCCC), but these national estimates cannot be
objectively compared due to differences in definitions, methods, and data
completeness (Houghton et al., 2012; Abad-Viñas et al., 2015). More
comparable AFOLU data are offered in global emission databases such as EDGAR
(Emissions Database for Global Atmospheric Research) or FAOSTAT (Food and
Agriculture Organization Corporate Statistical Database)(Smith et al., 2014;
Tubiello et al., 2015), or more sectorial datasets such as the Houghton
Forestry and other Land Use (FOLU) data (Houghton et al., 2012), and the US
Environmental Protection Agency non-CO
Understanding the inconsistencies among AFOLU datasets is an urgent task since they preclude our accurate understanding of land–atmosphere interactions, GHG effects on climate forcing and, consequently, the utility of modelling exercises and policies to mitigate climate change (Houghton et al., 2012; Grace et al., 2014; Smith et al., 2014; Sitch et al., 2015; Tian et al., 2016). The land use sector plays a prominent role in the Paris Agreement (Article 5), with many countries including it in their mitigation targets for their Nationally Determined Contributions (NDCs; Grassi and Dentener, 2015; Richards et al., 2015; Streck, 2015). It is then urgent to understand how much and why different AFOLU datasets differ in their emission estimates, so that we can better navigate countries' land-based mitigation efforts, and help to validate their proposed claims under the UNFCCC.
Here we compare gross AFOLU emissions estimates for the tropics, for
2000–2005, from six datasets: FAOSTAT, EDGAR, Houghton, Baccini, the US
Environmental Protection Agency data (EPA), and a recently produced,
spatially explicit AFOLU dataset, that we will hereafter call Hotspots
(Roman-Cuesta et al,. 2016). We aim to identify differences and plausible
explanations behind (i) aggregated AFOLU, FOLU, and agricultural gross
emissions, (ii) disaggregated contributions of the emission sources for the
different datasets, (iii) disaggregated contribution of the different gases
(CO
Our study area covers the tropics and the subtropics, including the more
temperate regions of South America (33
This is a multi-gas (CO
FAOSTAT covers agriculture, forestry, and other land uses and their
associated emissions of CO
The Emissions Database for Global Atmospheric Research (EDGAR) provides
global GHG emissions from multiple gases (CO
Differences and similarities of the assessed AFOLU datasets.
Houghton's bookkeeping model calculates the net and gross fluxes of carbon
(CO
These are gross FOLU tropical emissions published by Baccini et al. (2012). Data are gross emissions for the period 2000–2010 disaggregated
into deforestation (4.18 PgCO
The EPA dataset contains global non-CO
The AR5 is a synthesis report, not a repository of global data. However, new AFOLU data are produced by merging peer-reviewed data such as Figs. 11.2, 11.4, 11.5, and 11.8 in chapter 11 of the AR5 (Smith et al., 2014). We will compare our six datasets with the data from these newly produced figures.
Table 1 shows a summary of key similarities and differences of the assessed AFOLU datasets and the data from the AR5. The exact variables used for each database are described in Table S1 in the Supplement.
We focus on human-induced gross emissions only, excluding fluxes from
unmanaged land (i.e. natural wetlands). We focus on direct emissions
excluding indirect emissions whenever possible (i.e. nitrate leaching and
surface run-off from croplands). Delayed fluxes (legacies) are important
(i.e. underestimations of up to 62 % of the total emissions when recent
legacy fluxes are excluded; Houghton et al., 2012) but are frequently omitted
in GHG assessments that are derived from remote sensing, such as the deforestation
emissions used in the Hotspots database, which relies on Harris et al.,
2012). Wood-harvesting emissions also excluded legacy fluxes. We assumed
instantaneous emissions of all carbon that is lost from the land after human
action (Tier 1, IPCC, 2006; i.e. deforested and harvested wood), with no
transboundary considerations (i.e. the emissions are assigned wherever the
disturbance takes place, particularly important for Harvested Wood Products).
Life cycle substitution effects were neither considered for harvested wood
(Peters et al., 2012). Some exceptions were allowed when data were already
aggregated (i.e. for the Houghton and EPA datasets we could not exclude
indirect emissions linked to forest decay and agriculture respectively), or
because their legacy (past decay) estimates corresponded to an important
source (i.e. EDGAR post-burned decay and decomposition emissions represent
deforestation; Tubiello et al., 2015). Databases include a diversity of
emission sources and gases under AFOLU, not always following IPCC
requirements (some exclude peatland emissions, some include energy into the
AFOLU emissions, some exclude non-CO2 emissions, etc.). However, to compare
the AFOLU emission estimates between databases, we choose exactly the same
sources: deforestation, wood harvesting, fire, livestock (enteric
fermentation and manure management), cropland soil emissions, rice emissions,
emissions from drained histosols, and the same gases CO
Tubiello et al. (2015) identified four main differences that resulted in
larger estimates for the EDGAR data than for FAOSTAT, under the AFOLU
estimates of the AR5 (Smith et al., 2014): (1) the inclusion of energy
emissions under the agriculture budget, (2) the inclusion of savanna burning,
(3) higher rice emissions due to the use of the IPCC 1996 guidelines instead
of the IPCC 2006 guidance, and (4) FOLU's unresolved differences due to
unclear metadata on EDGAR's proxy for deforestation (post-burned decay and
decomposition). We have corrected for the first two in our data comparison.
No energy or CO
Summary of
To characterize the emission variability between countries we estimated the standard deviations for the different emission sectors: (i) forest (deforestation, fire, and wood harvesting), (ii) agriculture (cropland soils and paddy rice), (iii) livestock, and the aggregated AFOLU emissions, for the three most complete datasets (Hotspots, FAOSTAT, EDGAR), per country. We grouped the standard deviations into four percentiles to aggregate countries into levels of emission variability: high agreement (corresponds to low variability, low standard deviations, < 25th percentile), moderate agreement (25th–50th percentiles), low agreement (25th–50th percentiles), and very low agreement (equals very high variability, very high standard deviations, > 75th percentile). See Supplement for a further discussion on issues regarding emission variability.
AFOLU (Agriculture, Forestry, and Other Land Use) emissions
estimates (PgCO
We found good agreement among datasets for the aggregated tropical scales
with AFOLU values of 8.0 (5.5–12.2; 5th–95th percentiles), 8.4 and
8.0 Pg CO
Tropical gross annual emissions (2000–2005) comparisons for the
leading emission sources in the AFOLU sector, for the Hotspots, FAOSTAT,
EDGAR, Baccini, EPA, and Houghton datasets. Bars indicate uncertainty
estimates (1
The IPCC AR5 offers a FOLU gross value for the tropics of ca.
8.4 PgCO
Emissions in the agricultural sector are mostly net, since sink effects in
the soils are small and frequently temporal (USEPA, 2013; Smith et al.,
2014). Comparisons with global agricultural emissions show that for the
year 2000, global estimates more than doubled the Hotspots values (i.e. 5
and 5.5 Pg CO
While the gross aggregated estimates suggested a good level of agreement
among datasets (Fig. 1), differences occur when comparing the emissions
sources leading to the AFOLU budgets (Fig. 2). The FOLU sector showed the
largest differences, mainly due to the estimates of forest degradation, and
particularly fire (FAOSTAT and EDGAR showed the lowest and highest values).
The forest sector is the most uncertain term in the AFOLU emissions due to
both uncertainties in areas affected by land use changes and other
disturbances, and by uncertain forest carbon densities (Houghton et al.,
2012; Grace et al., 2014; Smith et al., 2014). Agricultural sources were
more homogeneous (ca. 2 Pg CO
Characteristics of the emission sources used in this comparative
assessment disaggregated by greenhouse gases for the period 2000–2005, for
the Hotspots, FAOSTAT, EDGAR, EPA, Houghton, and Baccini datasets (based on
gross emissions from Baccini et al., 2012). Superindices specify differences
between datasets and/or indicate the exact data included in our database
comparisons. EPA offers only non-CO
Deforestation emissions were 2.9 (1.0–10.1), 3.7, 2.5, and
4.2 PgCO
Forest degradation can be defined in many ways (Simula, 2009), but no single operational definition has been agreed upon by the international community (Herold et al., 2011a). It typically refers to a sustained human-induced loss of carbon stocks within forest that remains forest. In this study, similarly to Federici et al. (2015), we consider degradation to be any annual removal of carbon stocks that does not account for deforestation, without temporal-scale considerations (i.e. time needed for disturbance recovery or time to guarantee a sustained reduction of the biomass). We assessed two major degradation sources: wood harvesting and fire. Soil degradation is poorly captured in many datasets, and mainly focuses on fire in equatorial Asian peatland forests and drained peatlands (Hooijer et al., 2010). A better understanding of the processes and emissions behind forest degradation is key for climate mitigation efforts, not only because forest degradation is a widespread phenomenon (i.e. affects much larger areas than deforestation; Herold et al., 2011b), but also because the lack of knowledge of net carbon effects frequently results in assumptions of carbon neutrality of the affected standing forests, particularly for fire (Houghton et al., 2012; Le Quéré et al., 2014), which likely leads to an underestimation of forest and AFOLU emissions (Brando et al., 2014; Turetsky et al., 2015; Roman-Cuesta et al., 2016).
Gross emissions from forest degradation were larger than deforestation for the Hotspots, EDGAR, and Baccini datasets, with degradation-to-deforestation ratios of 108, 120, and 128 % respectively. FAOSTAT had degradation emissions of 60 % of the deforestation, partly due to its anomalously low fire contribution (see next section). Houghton et al. (2012) pointed out that global FOLU net fluxes were led by deforestation with a smaller fraction attributable to forest degradation, while the opposite was true for gross emissions (degradation being 267 % of deforestation emissions). This large ratio relates to their inclusion of shifting cultivation under degradation. This is a definition issue, which would not fit the definition of degradation chosen in this study, where a complete forest cover loss would represent deforestation and not degradation.
Continental disaggregated emissions for the individual emission sources (PgCO
Fire led the gross forest degradation emissions in the tropics in 2000–2005
(Fig. 2): 2 (1.1–2.7), 0.2, 3.4, 2.9 Pg CO
In spite of the importance of fire as a degradation source, this variable is
frequently incompletely included, either through unaccounted gases (i.e.
CH
Fire suffers, moreover, from a series of assumptions that do not apply so easily to other types of degradation: (1) assuming a non-human nature of the fires (deforestation fire vs. wildfires), which in tropical areas contrasts with multiple citations referring to the 90 % human causality of fires (Cochrane et al., 1999; Roman-Cuesta et al., 2003; Alencar et al., 2006; Van der Werf et al., 2010); (2) assuming force majeure conditions that lead to non-controllable fires due to extreme climate conditions, which frequently result in incomplete assessment and reporting of emissions. This assumption contrasts with research on how human activities have seriously increased fire risk and spread in the tropics (Uhl and Kauffman, 1990; Laurance and Williamson, 2001; Roman-Cuesta et al., 2003; Hooijer et al., 2010), and clearly expose how most of the fires in the humid tropics would not occur in the absence of human influences over the landscape (Roman-Cuesta et al., 2003). (3) We assuming carbon neutrality and full biomass recovery after fire in standing forests. This is a generous assumption that contrasts with numerous studies on tropical forest die-back following fire events in non-fire adapted humid tropical forests (Cochrane et al., 1999; Barlow and Peres, 2008; Roman-Cuesta et al., 2011; Brando et al., 2012; Oliveras et al., 2013; Balch et al., 2015). All these phenomena cast doubts on the robustness of these assumptions and call for a much more comprehensive inclusion of fire emissions into forest degradation budgets.
There is not a unique way to estimate wood harvesting emissions as exposed
in the guidelines for harvested wood products of the IPCC (IPCC, 2006).
Assumptions regarding the final use of the wood products, decay times,
substitution effects, international destination of the products, and time
needed for forests to recover their lost wood can fully change the emission
budgets. In our study, wood-harvesting emissions were 1.2 (0.7–1.6), 2.0,
1.7 PgCO
Livestock emissions were the most homogeneous among the emissions sources
(Fig. 2) with estimates of 1.2 (0.8–1.5), 1.1, 1.2,
1.1 Pg CO
The estimates of cropland emissions reached values of 0.18 (0.16–0.19),
0.56, 0.6, and 0.64 Pg CO
Identification of the least reliable emission source (x) for each dataset considering disagreements among emission estimates due to biased/divergent/incomplete definitions and methods.
Estimates of drained peatlands (mainly for agricultural purposes) suggest
large omissions in the Hotspots database with emissions 1 order of
magnitude lower (28 TgCO
Disaggregation of cropland soil emissions from drained peatlands for datasets with available data: FAOSTAT and Hotspots. Organic soils are excluded in EPA cropland emissions.
When paddy fields are flooded, decomposition of organic material gradually
depletes the oxygen present in the soil and floodwater, causing anaerobic
conditions in the soil that favour methanogenic bacteria that produce
CH
Based on the explanations above, Table 4 points out the likely least reliable emission sources for each dataset considering disagreements among emission estimates due to biased/divergent/incomplete definitions and methods. Houghton's sinks are suggested as least reliable since they suffer from compatibility issued with IPCC guidance and exclude sinks from non-disturbed areas and forests undergoing disturbances other than wood harvesting or recovery from shifting cultivation (Grassi and Dentener, 2015; Federici et al., 2016).
Contribution of the different AFOLU greenhouse gases (CO
GHG emission contribution (CO
GHG emissions (CO
Non-CO
The importance of multi-gas assessments relates to their role in climate
change mitigation due to their radiative forcing (RF), understood as a
measure of the warming strength of different agents (gases and not gases) in
causing global warming (W m
Country comparisons showed poor agreement among datasets for all the emission sectors, particularly for the largest emitters (i.e. Brazil, Argentina, India, Indonesia; Figs. 7, 8). Forests led the AFOLU disagreements (as observed by the similarity of Fig. 7a, b). From a continental perspective, Central and South America had more countries with high levels of disagreement, suggesting a need for further data research.
Country-level agreement for
Different datasets were developed for different purposes that have influenced the methods and approaches chosen to estimate their land use GHGs. Thus, EDGAR was created with an air pollution focus making its land emissions weaker. In contrast, FAOSTAT carries FAO's focus on land, particularly agriculture (data available since the 60s), with forest data added later through the FRA assessments (1990, 2005, 2010, 2015). The “Hotspot” database was created to identify the areas with the largest land use emissions in the tropics (emissions hotspots), while Houghton's accent is on historical LULUCF emission trends (since 1850). EPA concentrates on industrial, energy, and agricultural emissions. Forests are excluded with an interest on human health and mitigation. Moreover, due to its long existence, several datasets rely on FAOSTAT long-term agricultural data, which is probably the reason behind the higher homogeneity of agricultural emission estimates (i.e. crops, rice, and livestock among datasets). FAOSTAT forest emissions use FRA data, which get updated every 5 years. Different FRA versions strongly influence forest emission estimates which makes it important to acknowledge the FRA version used when contrasting FAOSTAT emissions and when comparing estimates (i.e. differences up to 22 % between the forest sink estimates using FRA2015 and FRA2010 have been reported by Federici et al., 2015). Similarly, official updates of Houghton's bookkeeping TRENDS data, as well as researchers' self-tuned versions of his model, result in emission differences that are difficult to track.
Country-level agreement for
Under the UNFCCC, countries are requested to use the latest IPCC AFOLU guidelines to estimate their GHG emissions (i.e. IPCC, 2006, 2003 for developed and developing countries respectively). The use of different guidelines, tiers, and approaches influences the final emission estimates. Compliance with IPCC has two main consequences: (1) the total area selected to report emissions and (2) the choice of land use over land cover. In the first case, under IPCC guidance, the total area selected to report emissions would include all the land under human influence (the managed land concept, which includes areas under active and non-active management). Houghton's bookkeeping model and the carbon modelling community in general do not comply well with the managed land concept, resulting in different net emissions from forest land uses and land use changes (LULUCF) than IPCC compliant country emissions (Grassi and Dentener, 2015; Federici et al., 2016). In the second case, the selection of land uses instead of land covers has partly been behind the recent controversy between FAO and the Global Forest Watch's reported estimates on deforestation trends (Holmgren, 2016). Estimates of deforestation that rely on land cover are higher than those using land use, since forest losses under forest land uses that remain forest land use are not considered deforestation (i.e. logged areas will regrow). In our analysis, FAO and Houghton rely on land use for deforestation, while Hotspots and EDGAR rely on land cover. FAOSTAT and Hotspots rely on the 2006 IPCC guidelines for National Greenhouse Gas Inventories (IPCC, 2006). FAOSTAT uses Tier 1 and standard emission factors, while Hotspots uses a combination of tiers (Tier 3 for all emissions except wood harvesting and cropland emissions over histosols that rely on Tier 1). EDGAR reports the use of 2006 IPCC guidelines for the selection of the emission factors but some of their methodological approaches are not always consistent with IPCC guidelines (i.e. deforestation expressed as the decay of burned forests, wood harvesting is part of the energy sector, agricultural energy balances are included in the AFOLU budget). EPA methods are reported to be consistent with IPCC guidelines and guidance, with Tier 1 methodologies used to fill in missing or unavailable data (USEPA, 2013).
The Paris Agreement (COP21) counts on the Intended Nationally Determined
Contributions (INDCs) as the core of its negotiations to fight climate
change. As of March 2016, 188 countries had submitted their INDCs under the
UNFCC (FAO, 2016) with agriculture (crops, livestock, fishery, and
aquaculture) and forests as prominent features in meeting the countries'
mitigation and adaptation goals (86 % percent of the countries include
AFOLU measures in their INDCs, placing it second after the energy sector;
FAO, 2016). However, there exists large variability in the way countries
present their mitigation goals, and quantified sector-specific targets are
rare (FAO, 2016). Variability relates not only to the lack of a standardized
way of reporting mitigation commitments under the INDCs, but also to
uncertainties and gaps in the AFOLU data. The Paris Agreement relies on a
5-year cycle stock-taking process to enhance mitigation ambition, and to keep
close to the 2 Data aggregation offers more homogeneous emission estimates than
disaggregated data (i.e. continental level, gas level, emission source
level). Forest emissions are the most uncertain of the AFOLU sector, with
deforestation having the highest uncertainties. Agricultural emissions, particularly livestock, are the most homogeneous of
the AFOLU emissions. Forest degradation, both fire and wood harvesting, show the largest
variabilities among databases. CO Among the non-CO Emissions from histosols/peatlands remain incomplete or fully omitted in
most datasets. Large emitters show the highest levels of data disagreement in the tropics,
enhancing the need for data improvement to guarantee effective mitigation
action. Forest lead the emission disagreement in the total AFOLU emissions. Central and South America showed the largest continental disagreements on
emission data for all the land sectors.
For the country and continental scales, we found the following.
Research run by the carbon community is pivotal for AFOLU assessments and,
while these two research communities overlap, they do not focus on exactly
the same topics. The carbon community works with CO
The quality of the reported AFOLU emissions can be assessed through the
UNFCCC principles: completeness, comparability, consistency, accuracy, and
transparency, which can help navigate the improvements of national
monitoring systems. From these principles, the reviewed datasets performed
well in consistency (they applied similar methods and assumptions over time, with the
exception of Hotspots that did not include temporal data). Transparency was excellent
for FAOSTAT with well elaborated and publicly available metadata linked to
their offered data, while EDGAR performed poorly due to insufficient
metadata. Improving transparency requires an urgent call for future action.
Improving accuracy and uncertainty also requires urgent action. Thus, in spite of their importance in fully
understanding the emission trends and dynamics, only Houghton and the
Hotspots provided uncertainties. FAO offered uncertainties as a percent
value for each emission source. Completeness and omissions are also urgent tasks because all datasets
are incomplete, i.e. missing pools, missing gases (Table 1), and omissions
affect all datasets. Complete emission reporting should consider the
importance of the following:
forest soil CO emissions from CH all forest fire types (i.e. temperate conifers and woodlands; understory
fires over humid closed canopy forests (Alencar et al., 2006; Morton et al.,
2013; i.e. 85 500 km CO CO CO
Further suggestions on improving data gaps and knowledge for the AFOLU
sector have been reported by Smith et al. (2014), Houghton et al. (2012),
USEPA (2013), and Sist et al. (2015), with a focus on soil data and crop production systems, as
well as an improved understanding of the mitigation potentials, costs, and
consequences of land use mitigation options.
Data will be available at my website:
Rosa Maria Roman-Cuesta, Mariana C. Rufino, and Martin Herold designed the study. Stephen Ogle and Benjamin Poulter provided data and ran quality controls of the data. Rosa Maria Roman-Cuesta, Mariana C. Rufino, Martin Herold, Klaus Butterbach-Bahl, Todd S. Rosenstock, Louis Verchot, Christopher Martius, Simone Rossi, Richard A. Houghton, Stephen Ogle, Benjamin Poulter, and Sytze de Bruin discussed the results and contributed to writing. Sytze de Bruin advised on statistical choices.
This research was generously funded by the Standard Assessment of Mitigation Potential and Livelihoods in Smallholder Systems (SAMPLES) project as part of the CGIAR Research Program Climate Change, Agriculture, and Food Security (CCAFS). Funding also came from two European Union FP7 projects: GEOCarbon (283080) and Independent Monitoring of GHG Emissions-No. CLIMA.A.2/ETU/2014/0008. Partial funds came through CIFOR from the governments of Australia (Grant Agreement #46167) and Norway (Grant Agreement #QZA-10/0468). In memory of Changsheng Li. Edited by: A. Ito Reviewed by: two anonymous referees