Aligning climate scenarios to emissions inventories shifts global benchmarks

Abstract

Taking stock of global progress towards achieving the Paris Agreement requires consistently measuring aggregate national actions and pledges against modelled mitigation pathways¹. However, national greenhouse gas inventories (NGHGIs) and scientific assessments of anthropogenic emissions follow different accounting conventions for land-based carbon fluxes resulting in a large difference in the present emission estimates2,3, a gap that will evolve over time. Using state-of-the-art methodologies⁴ and a land carbon-cycle emulator⁵, we align the Intergovernmental Panel on Climate Change (IPCC)-assessed mitigation pathways with the NGHGIs to make a comparison. We find that the key global mitigation benchmarks become harder to achieve when calculated using the NGHGI conventions, requiring both earlier net-zero CO2 timing and lower cumulative emissions. Furthermore, weakening natural carbon removal processes such as carbon fertilization can mask anthropogenic land-based removal efforts, with the result that land-based carbon fluxes in NGHGIs may ultimately become sources of emissions by 2100. Our results are important for the Global Stocktake⁶, suggesting that nations will need to increase the collective ambition of their climate targets to remain consistent with the global temperature goals.

Full text

102 | Nature | Vol 624 | 7 December 2023
Article
Aligning climate scenarios to emissions
inventories shifts global benchmarks
Matthew J. Gidden1,2,10 ✉, Thomas Gasser1,1 0, Giacomo Grassi3, Nicklas Forsell1, Iris Janssens1,4,
William F. Lamb5,6, Jan Minx5,6, Zebedee Nicholls1, 7,8 , Jan Steinhauser1,9 & Keywan Riahi1
Taking stock of global progress towards achieving the Paris Agreement requires
consistently measuring aggregate national actions and pledges against modelled
mitigation pathways1. However, national greenhouse gas inventories (NGHGIs) and
scientic assessments of anthropogenic emissions follow dierent accounting
conventions for land-based carbon uxes resulting in a large dierence in the present
emission estimates2,3, a gap that will evolve over time. Using state-of-the-art
methodologies4 and a land carbon-cycle emulator5, we align the Intergovernmental
Panel on Climate Change (IPCC)-assessed mitigation pathways with the NGHGIs to
make a comparison. We nd that the key global mitigation benchmarks become harder
to achieve when calculated using the NGHGI conventions, requiring both earlier
net-zero CO2 timing and lower cumulative emissions. Furthermore, weakening natural
carbon removal processes such as carbon fertilization can mask anthropogenic
land-based removal eorts, with the result that land-based carbon uxes in NGHGIs
may ultimately become sources of emissions by 2100. Our results are important for the
Global Stocktake6, suggesting that nations will need to increase the collective ambition
of their climate targets to remain consistent with the global temperature goals.
The 2021 UN Climate Change Conference (COP26) marked a shift in
the focus of climate policy from pledge-making to implementation
towards the long-term temperature goal of the Paris Agreement, the
collective progress towards which is assessed through periodic Global
Stocktakes (GSTs). In spring 2022, the first GST was launched
7
and con-
tinues through the 2023 United Nations Climate Change Conference
(COP28) to establish evaluation mechanisms among parties. Compar-
ing present emission trends from the NGHGIs and future targets in a
collective benchmarking effort rooted in the best available science
will be key for a rigorous, precedent-setting first GST and the overall
success of the Paris Agreement6.
Countries have gradually increased the ambition of their national
targets in response to the latest IPCC report findings
1,8
. Notably, sev-
eral nations made long-term net-zero emission commitments in the
run-up to COP26 (ref. 9), which brought the long-term temperature
goal of the Paris Agreement within striking distance, although much
of the assessed temperature reductions arose from long-term and
non-binding promises rather than immediate climate action10–12. Global
climate scenarios show that bothdeep reductions of near-term emis-
sions as well as enhancement of anthropogenic land-based carbon sinks
are needed to achieve net-zero emissions and limit global warming to
achieve the temperature goal of the Paris Agreement13,14.
A key discrepancy exists, however, in how model-based scientific
studies and NGHGIs account for the role of anthropogenic land-based
carbon fluxes
4,15,16
, with national inventories incorporating a broader
scope of removals2, resulting in lower emission estimates in NGHGIs.
Previous studies
2–4
have quantified the magnitude of this difference
to be approximately 5.5–6.7 Gt CO
2
yr
−1
. This conceptual difference
hinders the comparability of the aggregate targets set by countries
and future mitigation benchmarks. Although this problem has been
acknowledged in the most recent IPCC assessment
17
and raised by par-
ties during the GST18, the impact of this discrepancy on national and
global mitigation benchmarks is still not well understood. Aligning
mitigation pathways assessed by the IPCC with NGHGI conventions
is therefore needed to support the science-based formulation of
nationally determined contributions (NDCs) and to measure collec-
tive global action against emission levels necessary to achieve the Paris
Agreement goal.
Aligning climate pathways and inventories
The IPCC-assessed mitigation pathways are typically generated by inte-
grated assessment models (IAMs) that capture transitions in anthropo-
genic energy and land-use systems consistent with stated global climate
policy objectives. The reporting conventions for land-use, land-use
change and forestry (LULUCF) carbon fluxes of these models follow
that of detailed global carbon-cycle models (that is, ‘bookkeeping’
models). These models simulate and account for direct anthropogenic
fluxes (due to human-induced land-use changes, forest harvest and
regrowth) separately from indirect fluxes that are the natural response
https://doi.org/10.1038/s41586-023-06724-y
Received: 13 February 2023
Accepted: 6 October 2023
Published online: 22 November 2023
Open access
Check for updates
1International Institute for Applied Systems Analysis, Laxenburg, Austria. 2Climate Analytics, Berlin, Germany. 3Joint Research Centre, European Commission, Ispra, Italy. 4Department of
Computer Science, imec, University of Antwerp, Antwerp, Belgium. 5Mercator Research Institute on Global Commons and Climate Change, Berlin, Germany. 6Priestley International Centre of
Climate, School of Earth and Environment, University of Leeds, Leeds, UK. 7Melbourne Climate Future’s Doctoral Academy, School of Geography, Earth and Atmospheric Sciences, University
of Melbourne, Parkville, Victoria, Australia. 8Climate Resource, Northcote, Victoria, Australia. 9Potsdam Institute for Climate Impact Research, Potsdam, Germany. 10These authors contributed
equally: Matthew J. Gidden, Thomas Gasser. ✉e-mail: gidden@iiasa.ac.at
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature | Vol 624 | 7 December 2023 | 103
of land to environmental changes (for example, CO
2
fertilization or
response to climate change)
5,15,19,20
and define anthropogenic emissions
as those owing to the direct effect. Because it is practically not possible
to separate direct and indirect fluxes through observations, the NGHGIs
submitted by parties to the United Nations Framework Convention on
Climate Change (UNFCCC) follow reporting conventions
21
that define
anthropogenic fluxes using an area-based approach
22
in which all fluxes
occurring on managed land are considered anthropogenic, with few
exceptions to isolate natural disturbences
16,23,24
. The NGHGIs include
a wider definition of managed land compared with models, which
includes any forested area that ‘perform[s] production, ecological,
or social functions’
25
(Fig.1). As a result, present-day LULUCF fluxes
estimated with scientific modelling conventions indicate that the land
sector is a net source of emissions3, whereas the NGHGIs collectively
report it as a net sink
26
, resulting in fundamentally different present
and future perspectives of the role of land-based fluxes.
To estimate the direct and indirect components of land-based carbon
fluxes necessary to align mitigation pathways with conventions used
in the NGHGIs, we use a reduced-complexity model with an explicit
treatment of the land-use sector, OSCAR5, one of the models used for
the annual Global Carbon Budget
3
, applied in a probabilistic setup
and at a resolution of five global regions used in the IPCC assessments
(Methods). We calculate a difference of 4.4 ± 1.0 Gt CO
2
yr
−1
in LULUCF
emissions globally averaged over 2000–2020 between model-based
(higher) and NGHGI-based (lower) accounting conventions, which is
in line with the existing estimates
2,5
. We then assess the pathways with
OSCAR to quantify how the difference between conventions evolves
over time. A total of 914 of the 1,202 IPCC-assessed pathways provided
sufficient land-use change data to enable this alignment (Extended
Data Table1; data are available at https://data.ece.iiasa.ac.at/genie/).
Across both the 1.5 °C and 2.0 °C scenarios (Fig.2a,b; see defini-
tions in the Methods), LULUCF emissions estimated using the NGHGI
conventions show a strong increase in the total land sink until around
mid-century. However, the NGHGI alignment factor (that is, the dif-
ference between the two accounting conventions; Fig.2c) decreases
over this period, nearing zero in the 2050s to 2060s for the 1.5 °C
scenarios and the 2070s to 2080s for 2.0 °C scenarios. This convergence
is primarily a result of the simulated stabilization and then decrease
of the CO2-fertilization effect as well as background climate warming
reducing the overall effectiveness of the land sink
27,28
, which in turn
reduces the indirect removals included in NGHGIs. These dynamics
lead to land-based emissions reversing their downward trend in most
NGHGI-aligned scenarios by mid-century and result in the LULUCF
sector becoming a net source of emissions by 2100 in about 25% of
both the 1.5 °C and 2.0 °C scenarios.
Global and regional ambition implications
More ambitious mitigation action is required to meet the global
emission benchmarks derived from scenarios when assessed using
the NGHGI conventions compared with model-based conventions
(Extended Data Table2 and Extended Data Fig.1). The NGHGI-aligned
pathways result in earlier net-zero CO2 emissions by 1–5 years for
the 1.5 °C and −1 to 7 years for the 2.0 °C scenarios (Fig.3a). Emission
+
+
+=
Enabling like-for-like comparison between the two conventions
Scientic models (red) do not currently match NGHGIs (green) resulting in different emissions estimates.
To align them, indirect uxes (blue) that occur on all land considered managed in NGHGIs, simulated with vegetation models,
need to be added to direct uxes (red) calculated with bookkeeping models.
Alignment factor
Misalignment between NGHGIs and scientific models
Differences stem from denitions of managed land and the carbon uxes that are included
Land with extensive
human activity
Land with limited or
no human activity
Other land serving production,
ecological and social functions
ManagedManaged
UnmanagedManaged
Direct uxes Indirect uxes
ClimateHuman
Bookkeeping models
Track carbon uxes
caused directly by
human activities
Vegetation models
Simulate uxes caused by
environmental changes
(for example, carbon fertilization)
Scientific
modelling convention
Based on models
of the carbon cycle
NGHGI
convention
Based on
observations
Not included
in NGHGIs
Unmanaged
Unmanaged
IAM Not included in
alignment factor
Fig. 1 | Dif ference in pr esent est imates of LULUC F carbon f luxes under
NGHGI and model-based accounting conventions. Schematic showin g the
difference in accounting conventions between NGHGIs (green) and scientific
models (b ookkeeping m odels in red and ve getation mod els in blue). Models
such as IA Ms are based on b ookkeeping a pproaches and co nsider direc t fluxes
due to land use (for exam ple, wood harve st) and land-cover cha nges. Additio nal
indirect f luxes due to evolv ing environme ntal conditi ons can be est imated by
process ed-based vege tation model s. NGHGIs co nsider a wider m anaged land
area and are gen erally based on p hysical obser vations, and th us include both
direct and i ndirect f luxes. We use the te rm ‘unmanaged ’ to describe l and not
considered managed by NGHGIs to be consistent with previous literature, but
recogni ze that this includ es land that has b een managed by ind igenous and
traditional communities for centuries to millienia38,39. In this stud y, we estimate
the alignm ent factor to tra nslate betwe en both conventi ons (the indirect f lux
considere d in NGHGIs but no t in models, blue). This f actor will chan ge over
time base d on future land-us e decisions an d overall mitigatio n efforts be cause
of, for example, changing atmospheric CO2 levels.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

104 | Nature | Vol 624 | 7 December 2023
Article
reductions in 2030 relative to 2020 are between 3 and 6 percentage
points greater for both pathway categories (Fig.3b). The assessed
cumulative net CO
2
emissions to global net-zero CO
2
also decreases
by 15–18% for both the 1.5 °C and 2.0 °C scenarios (Fig.3c) because of
extra land-based carbon removal when using the NGHGI conventions.
Although the NGHGI-aligned benchmarks strengthen, they are still
consistent with the climate assessment of the IPCC. All land-use fluxes
(direct and indirect) are included in the physical climate models used
by the IPCC—that is, the temperature outcomes of each pathway are
the same even if flux components are accounted for differently by
models and inventories. When considering the additional land sink
following the conventions of the NGHGIs, however, multiple dynamics
interact that contribute to the revisions of the benchmarks, including
the change in historical emission baseline, the enhanced anthropogenic
land sink compared with what was reported by IAMs and declining
sequestration from that additional sink in the future.
Parties to the UNFCCC use the net land CO2 flux reported in the
NGHGIs as a basis to assess compliance with their NDCs and track
the progress of their long-term emission reduction strategies under
the Paris Agreement
2,29,30
as with previous climate pacts
31
. Historically,
NDCs have been compared with scenario-based estimates of needed
emission reductions by either aligning the IPCC-assessed pathways to
NGHGIs with constant offsetting methods
1
or excluding LULUCF emis-
sions entirely
9,29
. Comparing our results with one of the most recent
aggregate NDC estimates1 (Methods and Extended Data Fig.2), we
find that the gap between unconditional NDCs and a median 2.0 °C
outcome is approximately 18% larger, whereas our assessment of the
gap between unconditional NDCs and a median 1.5 °C outcome is
around 4% smaller (Extended Data Fig.3). It is therefore important to
incorporate a dynamic estimation of indirect fluxes when assessing
national climate targets because their changing role in achieving these
targets depends on domestic land-management decisions as well as
the overall strength of global mitigation (Fig.4).
Aligning pathways to inventory-based LULUCF accounting con-
ventions can additionally affect how equitable mitigation action is
understood, as around 60% of the historical NGHGI adjustment falls
in Non-Annex I countries
26
. Assessed regionally, 1.5 °C-consistent emis
-
sion reductions are higher for developed countries, whereas they are
slightly lower in most developing regions when assessing scenario
outcomes using the NGHGI-based conventions (Extended Data Fig.4).
In the 2.0 °C pathways, the NGHGI alignment results in more stringent
2020–2030 emission reductions globally compared with the unad-
justed pathways, as the strength of the indirect flux continues to grow
with increasing atmospheric carbon concentrations. This strength-
ening most directly affects regions with large forested areas such as
Latin America and Russia, whereas others such as the Organisation for
Economic Co-operation and Development (OECD) countriesand Asia,
on average, see a decrease in emission reductions. Our results span
both positive and negative values across many regions, showcasing
the diversity of future responses to land-sink changes and complexities
when setting both equitable and ambitious climate targets based on
national inventories. They also highlight the risk of over-dependence
on land sinks to measure mitigation progress using national inventory
conventions against ambitious climate targets.
Considering carbon removal
In most 1.5 °C and 2.0 °C pathways, hundreds of gigatonnes of CO
2
are
removed from the atmosphere over the course of this century, with
ultimate levels dependent on the strength of near-term mitigation
action
17,32,33
. Because our assessment relies on a bookkeeping model
that explicitly tracks land carbon reservoirs, we are able to isolate
LULUCF emissions—model reporting conventions
LULUCF emissions—NGHGI reporting conventions
Ref. 3 model 2002–2011
Ref. 3 model 2012–2021
Ref. 3 NGHGI 2002–2011
Ref. 3 NGHGI 2012–2021
Ref. 2 NGHGI 2000–2020
2000 2020 2040 2060 2080 2100
−8
−6
−4
−2
0
2
4
6
8
a
b
c
2000 2020 2040 2060 2080 2100
−8
−6
−4
–2
0
2
4
6
8
Gt CO2 yr–1
Carbon-cycle
uncertainty
Scenario
uncertainty
Model convention
estimates higher
LULUCF uxes
NGHGI conventio
n
estimates higher
LULUCF uxes
2000–2020
−4
−2
0
2
4
6
8
−4
−2
0
2
4
6
8
Gt CO2 yr–1
2025 2050 2075 2100
1.5 ºC
Ref. 2
OSCAR (this study)
Ref. 3
2.0 ºC
+=
Gt CO2 yr–1
Fig. 2 | Land -use emiss ions in re-an alysed IPCC pa thways with model-
based and NGHGI-based accounting conventions. a,b, Land-use emissions
pathways before a nd after alignm ent to match NGH GIs for 1.5 °C (a) and 2 ,0 °C
(b) pathways. Hist orical esti mates2,3 a re shown with car bon-cycle unce rtainty
(1σ), and the medi an of scenario pat hways are shown with th e scenario
interquar tile range in shade d plumes. Pathways c onsistent wi th model-
based convention are shown in red, whereas the NGHGI convention is shown
in green. c, C omparing the t wo conventions resu lts in a differenc e between
re-analysed and NGHGI-adjusted pathways—that is, an alignment factor,
which evolves as a f unction of the s trength of land -based climate m itigation.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature | Vol 624 | 7 December 2023 | 105
those LULUCF fluxes that effectively constitute carbon removals from
carbon emissions (for example, deforestation), thereby quantifying
total land-based carbon dioxide removal (CDR) consistently across
scenarios and filling a gap in the IPCC Sixth Assessment Report
(see footnote 53 in ref. 17) as well as underlying scenario database that
constitutes a widely used resource in the climate change community
34
.
Scenarios see a marked increase by 2030 in CDR from the LULUCF
sector, resulting in around 60% higher removals of CO2 by 2030
compared with the 2020 levels in the 1.5 °C pathways and 10%
higher removals in the 2.0 °C pathways (Fig.5a). Taken together
with engineered (non-LULUCF) CDR options, models deploy 2.6
[1.4–3.2] Gt CO2 yr−1 (interquartile range) and 0.7 [0.3–2.5] Gt CO2 yr−1
additional CDR between 2020 and 2030 in the 1.5 °C and 2.0 °C path-
ways, respectively. Land-based sinks account for nearly 100% of
current CDR. By 2030, in the 1.5 °C pathways, 95% [88–98%] of total CDR
is delivered by land-based sinks (Fig.5b). By 2100, CDR from LULUCF
accounts for about 30% [21–42%] of the annual total.
Although deep mitigation scenarios assessed by the IPCC show a
notable and continued dependence on land-based removals over the
whole century, LULUCF removals of the same pathways aligned to
NGHGIs would peak by mid-century and decline thereafter. Over time,
the reduced effectiveness of indirect removals counterbalances the
gains from direct removals35 (Extended Data Fig.5), maintaining
the overall direct and indirect removals at around 6–7 Gt CO2 yr−1 by
mid-century. The 1.5 °C pathways cumulatively sequester around 20%
more carbon from direct removals but 20% less carbon from indi-
rect removals compared with the 2.0 °C pathways over that period
(Extended Data Fig.6). Considering the changing dynamics of indi-
rect carbon removals included in NGHGIs can dramatically change
the estimated carbon removals on land over time. Although the 1.5 °C
scenarios show growth in total assessed net land removal by 2030
(Fig.5c), thescenariosaligned with current policies approximately
double removals compared with the 1.5 °C and 2.0 °C scenarios by
mid-century, because of the increasing strength of indirect removals
(notably through strong CO2 fertilization) (Fig.5d).
Thus, although the addition of a larger ‘managed land’ sink in NGHGIs
may reduce the reported levels of present-day national emissions in
some cases, maintaining the strength of this carbon sink on these land
areas may pose a fundamental challenge in the long term. Not only do
estimates of needed progress in anthropogenic emission reductions
risk being masked by natural sink enhancement in the near term, but
even the maintenance of existing natural sinks requires additional
efforts to remove carbon, for example, through the expansion of forest
areas, from the NGHGI accounting perspective. In other words, the
future effort needed to achieve or maintain net-zero economy-wide
emissions would be underestimated using NGHGI accounting conven-
tions as the indirect contribution to land sinks loses efficacy and may
eventually become a net source of emissions in low-warming scenarios.
Balancing practicality and policy advice
We provide here an estimation of the LULUCF emissions consistent
with NGHGIaccounting conventions for all IPCC-assessed scenarios
that provide sufficient land-use cover information using probabilistic
and constrained estimates from a single established model, OSCAR.
Repeating this work with additional models would increase robustness
by averaging model biases and structural uncertainties, although this
would require land-use scenario information at a much finer resolution
than the five regions.
Because the pathways are aligned with the NGHGI conventions by
re-allocating indirect carbon fluxes caused by environmental changes
to anthropogenic fluxes, our results do not change any climate out-
come or mitigation benchmark produced by the IPCC, but provide
a translational lens to view those outcomes from the perspective of
national emissions reporting frameworks as deployed by the UNFCCC
parties. For example, the fact that we find net-zero timings for the 1.5 °C
pathways advance by up to 5 years compared with the IPCC-assessed
benchmarks does not imply that 5 years have been lost in the race to
net-zero, but that following the reporting conventions for natural sinks
used by parties to the UNFCCC results in net-zero needing to be reached
5 years earlier to match the modelled benchmarks. Our results reinforce
the importance of a rapid decline in fossil fuel and industry emissions
in this decade while limiting reliance on nature-based solutions that
can weaken over time to keep global temperature rise within the limit
prescribed in the Paris Agreement.
0 0.02 0.04 0.06 0.08 0.10
Density (a.u.)
−5.0
−2.5
0
2.5
5.0
7.5
10.0
12.5
15.0
Δ
years
1.5 ºC
1.5 ºC-OS
2.0 ºC
0 0.05 0.10 0.15 0.20 0.25
Density (a.u.)
−8
−6
−4
−2
0
0 0.002 0.004 0.006 0.008
Density (a.u.)
0
50
100
150
200
250
a
b
c
Δ
%
Δ
Gt CO
2
Fig. 3 | Chan ges in global m itigation b enchmark s across as sessed s cenarios .
a–c, Scenari o-wise distr ibutions of th e estimated ch ange in the net-zero CO2
year (a), 2020–2030 CO2 emission reductions (b) and cumulative emissi ons
until net-zero CO2 (c) betwee n the re-analysed m odel-based an d the NGHGI
LULUCF accou nting conventio ns are shown for 1. 5 °C (blue, IPC C category C1),
1.5 °C- OS (green, I PCC category C 2) and 2.0 °C (purp le, IPCC categor y C3)
scenari os. A positive valu e indicates that t he benchmark co mes later (for
net-zero years) or is higher (for cumul ative emission s) in the model-base d
framework co mpared with the N GHGI-base d framework, wh ereas a negati ve
value indica tes that the ben chmark is highe r in the NGHGI-ba sed framework
(for emission red uctions). Across all b enchmarks, N GHGI-base d accounting
tends to res ult in more stringe nt outcomes (earli er net-zero years, higher
emission re ductions a nd lower cumulative em issions to net-zero CO2 emission).
A comparis on with the ori ginal AR6 bench marks is shown in E xtended Dat a
Fig.1. a.u., ar bitrary unit s.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

106 | Nature | Vol 624 | 7 December 2023
Article
Importantly, this ‘new’ net-zero year is conceptually consistent with
the meaning of balancing of sources and sinks of greenhouse gases
(GHGs) as stipulated in Article 4 of the Agreement (although in the
context of CO2). Yet it occurs before the climatological milestone
that results in halting further warming, as is the case of the net-zero
year under the scientific modelling convention. Understanding and
addressing how these different frameworks can be mutually interpreted
is a fundamental challenge for evaluating progress towards the Paris
Agreement, given the reality that carbon removals from anthropogenic
and natural land-based processes cannot be estimated separately by the
NGHGIs, which are typically based on direct observations. The outcomes
presented here highlight that the conventions by which land-based
carbon removals are considered have important implications for
NDC assessment and transparency, operationalization of removals
under carbon markets as laid out in Article 6.4 of the Paris Agreement
and monitoring, reporting and verification of these removals.
The policy and scientific communities can take steps to meet this
challenge by reconciling terms, definitions and estimates of land-based
CO
2
fluxes in four concrete ways. First, climate targets can be formu-
lated explicitly for areas of critical mitigation action, including gross
CO
2
emission reductions without LULUCF, net land-based removals,
engineered carbon removals and non-CO2 GHG emissionreductions,
allowing for parties to define their expected contributions and to meas-
ure progress in each domain separately. Second, parties can clarify
the nature of their deforestation pledges, because direct and indirect
carbon fluxes vary greatly in different forest types
36
. Third, scientific
and practitioner communities can convene discussions on how to
enhance monitoring, reporting and verification of LULUCF fluxes to
better align estimates from both groups. Fourth, IAM teams can provide
their individual assumptions and estimates for direct LULUCF emis-
sions and removals, including the indirect flux component consistent
with the NGHGIs
37
and their assumptions about the land-use contribu-
tion of NDCs and long-term strategies. Future IPCC assessments could
either use such scenario data if available or use an approach such as
that presented here to provide a holistic scenario assessment aligned
with the NGHGIs and better inform necessary collective action to meet
global climate goals.
Although science and policy processes continue to co-evolve, inform-
ing one another, a full reconciliation of the conceptual discrepancies
outlined here will take time. However, the first iteration of the GST will
be completed by the end of 2023 and new NDCs will be formulated soon
thereafter, necessitating earlier compatibility between national targets
and benchmarks estimated by global models. Our results provide esti-
mates and a line of evidence that can be directly used by parties and
Carbon sinks
CO2 removals
Carbon source
CO2 emissions
Net
emissions
Carbon sinks
CO2 removals
Carbon source
CO2 emissions
Net
emissions
Direct
Direct
IndirectDirect
Impact of indirect fluxes on ability to achieve national climate targets
Future national land carbon balance under NGHGI convention shift with global CO2 levels
Low global mitigation
CO
2 levels in line with
current climate policies
National land sector
Global emissions
Paris Agreement
alignment
National land sector
Global emissions
Paris Agreement
alignment
National land sector
Global emissions
Paris Agreement
alignment
National land sector
Global emissions
Paris Agreement
alignment
Direct
NGHG
IN
GHGI
NGHGI
IndirectDirect IndirectDirect
++=++=
Direct
NGHGIIndirectDirect ++=
++=
Increased land-based mitigation
Strong mitigation effort pursued in
the land sector of a country
Unchanged land-based mitigation
No additional mitigation effort in
the land sector of a country
Stylized uxes
Care is needed when national climate targets rely on indirect fluxes
Direct uxes are consistent with the degree of mitigation in the land sector. Indirect uxes depend on how environmental
conditions (for example, CO2, climate) change over time, which is in turn dependent on global mitigation efforts. Under the NGHGI convention,
a Paris-consistent world could lead to weaker indirect removals, masking increased direct ones.
High global mitigation
CO
2 levels in line with the
Paris Agreement
CO2
CO2
CO2
CO2
CO2
Fig. 4 | The f uture role of ind irect f luxes in nation al climate tar gets. In a
future wit h strong mitiga tion action i n line with the goal s of the Paris Agr eement
(bottom r ow), stabilizing or even de creasing atm ospheric CO2 wil l result in a
weakening of th e indirect sink ( blue arrows), whereas a futu re with weak
mitigati on action will s ee the indirec t sink increas e(as long as CO2 fertilization
dominate s over climate feedba cks, top row). The direc t component o f LULUCF
fluxe s (red arrows) is dueentirely to lan d-use manageme nt decisions (colu mns).
Future estima tes of net LULUCF e missions (g reen arrows) will dif fer between
conventions depending on how much overall mitigation occurs and how much
land-base d mitigation o ccurs, which ca n have unexpecte d consequen ces on
national climate target achievement.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature | Vol 624 | 7 December 2023 | 107
the UNFCCC to meaningfully compare aggregated national targets and
mitigation benchmarks. No matter what the reporting conventions
are, the near-term action that is needed to meet the Paris Agreement
is clear: emissions must peak as soon as possible and reduce markedly
this decade. This message must not be lost in the translation between
different concepts of anthropogenic land carbon fluxes.
Online content
Any methods, additional references, Nature Portfolio reporting summa-
ries, source data, extended data, supplementary information, acknowl-
edgements, peer review information; details of author contributions
and competing interests; and statements of data and code availability
are available at https://doi.org/10.1038/s41586-023-06724-y.
1. den Elzen, M. G. J. etal. Updated nationally determined contributions collectively raise
ambition levels but need strengthening further to keep Paris goals within reach. Mitig.
Adapt. Strateg. Glob. Change 27, 33 (2022).
2. Grassi, G. etal. Harmonising the land-use lux estimates of global models and national
inventories for 2000–2020. Earth Syst. Sci. Data 15, 1093–1114 (2023).
3. Friedlingstein, P. etal. Global Carbon Budget 2022. Earth Syst. Sci. Data 14, 4811–4900
(2022).
4. Grassi, G. etal. Critical adjustment of land mitigation pathways for assessing countries’
climate progress. Nat. Clim. Change 11, 425–434 (2021).
5. Gasser, T. etal. Historical CO2 emissions from land use and land cover change and their
uncertainty. Biogeosciences 17, 4075–4101 (2020).
6. Ogle, S. M. & Kurz, W. A. Land-based emissions. Nat. Clim. Change 11, 382–383 (2021).
7. UNFCCC. Synthesis Report for the Technical Assessment Component of the First Global
Stocktake (UNFCCC, 2023).
8. van Beek, L., Oomen, J., Hajer, M., Pelzer, P. & van Vuuren, D. Navigating the political: an
analysis of political calibration of integrated assessment modelling in light of the 1.5 °C
goal. Environ. Sci. Policy 133, 193–202 (2022).
9. UNFCCC. Nationally Determined Contributions under the Paris Agreement. Revised Note
by the Secretariat (UNFCCC, 2021).
2020 2030
1.5 ºC
2030
2.0 ºC
2050
1.5 ºC
2050
2.0 ºC
2100
1.5 ºC
2100
2.0 ºC
0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
Gt CO2 yr–1
Total CDR
Land CDR
Non-land CDR
2020 2030 2040 2050 2060 2070 2080 2090 2100
0
0.2
0.4
0.6
0.8
1.0
Fraction of total CDR
1.5 ºC
2.0 ºC
1.5 ºC 2.0 ºC Current policies
–3
–2
–1
0
1
2
3
4
Gt CO2 yr–1
1.5 ºC 2.0 ºC Current policies
−3
−2
−1
0
1
2
3
4
Indirect removals
Land CDR (direct)
Total land removals
Scenario uncertainty
Carbon-cycle uncertainty
Gt CO2 yr–1
a b
cd
Fig. 5 | CDR characteristics in mitigation and current-policy pathways.
a, Net land-us e carbon removal le vels from direct f luxes (g reen bars) are
compared w ith non-land CDR ( brown bars) and total leve ls (summing land-us e
and CDR, grey b ars) with whiskers de noting the inte rquartile range o f each
estimate a cross 1.5 ° C and 2.0 °C sc enarios. Her e, non-land CDR com prises
technolo gies include d in the IAM pathw ays assesse d in AR6other than thos e
due solely to la nd-use change, su ch as bio-ene rgy with car bon capture and
storage, dire ct air capture of CO2 w ith storage and enhanced mineral
weathering. b, The share of land- based CDR redu ces over time acros s both
1.5 °C a nd 2.0 °C pathways w ith the median (so lid line) and interquar tile range
(shaded area) show n for the populatio n of scenarios a ssessed . The direct
component of land-based removal flux, which constitutes land-based CDR, and
the indirec t component o f the removal flu x evolve differently a cross pathways.
c, In the near te rm, until 2030, th e 1.5 °C pathw ays see a strong enh ancement
of addition al removals (pink bar), wherea s the 2.0 °C pathway s see a similar
addition of t otal removals as cur rent-policy pathways. d, By mid-century,
additiona l removals in current-polic y pathways out-pace both th e 1.5 °C and
2.0 °C pat hways, because o f the continued e nhancement o f indirect remova ls
compared w ith an overall weakeni ng of this flu x in mitigati on pathways. Sc enario
uncert ainty in c,d is est imated by the interq uartile range of s cenario-bas ed
estimate s, whereas t he carbon-c ycle uncert ainty is est imated by the inte rquartile
range of the me dian ensembl e of climate runs (Me thods).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

108 | Nature | Vol 624 | 7 December 2023
Article
10. Höhne, N. etal. Wave of net zero emission targets opens window to meeting the Paris
Agreement. Nat. Clim. Change 11, 820–822 (2021).
11. Meinshausen, M. etal. Realization of Paris Agreement pledges may limit warming just
below 2 °C. Nature 604, 304–309 (2022).
12. Ou, Y. etal. C an updated climate pledges limit warming well below 2 °C? Science 374,
693–695 (2021).
13. Rogelj, J. etal. Scenarios towards limiting global mean temperature increase below 1.5 °C.
Nat. Clim. Change 8, 325 (2018).
14. Roe, S. etal. Contribution of the land sector to a 1.5 °C world. Nat. Clim. Change 9, 817–828
(2019).
15. Intergovernmental Panel on Climate Change. Climate Change and Land: An IPCC
Special Report on Climate Change, Desertiication, Land Degradation, Sustainable Land
Management, Food Security, and Greenhouse Gas Fluxes in Terrestrial Ecosystems
(Cambridge Univ. Press, 2022).
16. Grassi, G. etal. Reconciling global-model estimates and country reporting of anthropogenic
forest CO2 sinks. Nat. Clim. Change 8, 914–920 (2018).
17. IPCC. in Climate Change 2022: Mitigation of Climate Change. Working Group III Contribution
to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change
(eds Shukla, P. R. etal.) 3–48 (Cambridge Univ. Press, 2022).
18. UNFCCC. Summary Report on the First Meeting of the Technical Dialogue of the First Global
Stocktake under the Paris Agreement (UNFCCC, 2022).
19. Houghton, R. A. & Nassikas, A. A. Global and regional luxes of carbon from land use and
land cover change 1850–2015. Global Biogeochem. Cycles 31, 456–472 (2017).
20. Hansis, E., Davis, S. J. & Pongratz, J. Relevance of methodological choices for accounting
of land use change carbon luxes. Global Biogeochem. Cycles 29, 1230–1246 (2015).
21. Buendia, E., Guendehou, S., Limmeechokchai, B. & Pipatti, R. 2019 Reinement to the
2006 IPCC Guidelines for National Greenhouse Gas Inventories (UNFCCC, 2019).
22. Task Force on National Greenhouse Gas Inventories. Revisiting the Use of Managed Land as
a Proxy for Estimating National Anthropogenic Emissions and Removals (UNFCCC, 2009).
23. Ogle, S. M. etal. Delineating managed land for reporting national greenhouse gas
emissions and removals to the United Nations framework convention on climate change.
Carbon Balance Manage. 13, 9 (2018).
24. Canadell, J. G. etal. Factoring out natural and indirect human effects on terrestrial carbon
sources and sinks. Environ. Sci. Policy 10, 370–384 (2007).
25. IPCC. 2006 IPCC Guidelines for National Greenhouse Gas Inventories (UNFCCC, 2008).
26. Grassi, G. etal. Carbon luxes from land 2000–2020: bringing clarity to countries’ reporting.
Earth Syst. Sci. Data 14, 4643–4666 (2022).
27. Mitchard, E. T. A. The tropical forest carbon cycle and climate change. Nature 559,
527–534 (2018).
28. Lee, J.-Y. etal. in Climate Change 2021: The Physical Science Basis. Contribution of
Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on
Climate Change (eds Masson-Delmotte, V. etal.) 553–672 (Cambridge Univ. Press, 2021).
29. Fyson, C. L. & Jeffery, M. L. Ambiguity in the land use component of mitigation contributions
toward the Paris Agreement goals. Earths Future 7, 873–891 (2019).
30. Forsell, N. etal. Assessing the INDCs’ land use, land use change, and forest emission
projections. Carbon Balance Manag. 11, 26 (2016).
31. Schlamadinger, B. etal. A synopsis of land use, land-use change and forestry (LULUCF)
under the Kyoto Protocol and Marrakech Accords. Environ. Sci. Policy 10, 271–282
(2007).
32. Prütz, R., Streler, J., Rogelj, J. & Fuss, S. Understanding the carbon dioxide removal range
in 1.5 °C compatible and high overshoot pathways. Environ. Res. Commun. 5, 041005
(2023).
33. Smith, S. M. etal. The State of Carbon Dioxide Removal 1st edn (The State of Carbon Dioxide
Removal, 2023).
34. Byers, E. etal. AR6 scenarios database. Zenodo https://doi.org/10.5281/zenodo.5886912
(2022).
35. Jiang, M. etal. The fate of carbon in a mature forest under carbon dioxide enrichment.
Nature 580, 227–231 (2020).
36. Gasser, T., Ciais, P. & Lewis, S. L. How the Glasgow Declaration on Forests can help keep
alive the 1.5 °C target. Proc. Natl Acad. Sci. USA 119, e2200519119 (2022).
37. Gusti, M., Augustynczik, A. L. D., Di Fulvio, F., Lauri, P. & Forsell, N. in The 2nd International
Electronic Conference on Forests—Sustainable Forests: Ecology, Management, Products
and Trade 23 (eds de Dios, V. R. & Dimopoulos, P.) (MDPI, 2021).
38. Fletcher, M.-S., Hamilton, R., Dressler, W. & Palmer, L . Indigenous knowledge and the
shackles of wilderness. Proc. Natl Acad. Sci. USA 118, e2022218118 (2021).
39. Ellis, E. C. etal. People have shaped most of terrestrial nature for at least 12,000 years.
Proc. Natl Acad. Sci. USA 118, e2023483118 (2021).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional afiliations.
Open Access This article is licensed under a Creative Commons Attribution
4.0 International License, which permits use, sharing, adaptation, distribution
and reproduction in any medium or format, as long as you give appropriate
credit to the original author(s) and the source, provide a link to the Creative Commons licence,
and indicate if changes were made. The images or other third party material in this article are
included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your
intended use is not permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a copy of this licence,
visit http://creativecommons.org/licenses/by/4.0/.
© The Author(s) 2023
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Methods
Selection of AR6 scenarios
As part of its Sixth Assessment Report, IPCC Working Group III authors
analysed more than 2,200 scenarios for potential inclusion in its miti-
gation pathway assessment40. Of those, 1,202 were eventually vetted:
deemed to have provided enough detail to allow a climate analysis using
the climate assessment architecture of the IPCC
41
. Those scenarios were
then divided into different scenario categories based on their peak and
end-of-century temperature probabilities34.
In this study, we focus on three scenarios: C1, C2 and C3 as defined
in AR6 of the IPCC (ref. 40). C1 scenarios are as likely as not to limit
warming to 1.5 °C and have been interpreted as consistent with the
1.5 °C long-term temperature goal of the Paris Agreement as outlined
in Article 2 (ref. 42), although arguments have been made that fur-
ther delineation should be made into scenarios that do and do not
achieve net-zero CO
2
emissions to better reflect its Article 4 (ref. 43). We
assess outcomes from the 2.0 °C C3 scenarios given their historic policy
relevance, their capability to show progress towards 1.5 °C and their
use in examining climate impacts beyond what is envisioned by
the Paris Agreement. We also highlight mitigation outcomes of C2
scenarios, also called high overshoot scenarios, which are as likely
as not to limit warming to 1.5 °C in 2100 but are likely to exceed
1.5 °C in the interim period. Such pathways are nominally similar in
mitigation and impact assessment with C3 scenarios until at least
mid-century43.
For this analysis, we require that scenarios have been vetted by
the IPCC climate analysis framework and provide a minimum set of
land-cover variables such as Land Cover|Cropland, Land Cover|Forestry
and Land Cover|Pasture. We analyse the presence of each of these vari-
ables and their combination in Extended Data Table3 at the global, IPCC
5-region (R5) and IPCC 10-region (R10) levels. Balancing concerns of
greater regional detail and greater scenario coverage, we perform our
analysis based on the R5 regions (Extended Data Table4) given that
nearly all models with full global variable coverage also provide detail
at the R5 regional level for the C1–C3 scenarios.
To understand how well our scenario subset containing R5 land-cover
variables corresponds statistically to the full database sample of the
C1–C3 scenarios, we perform a Kolmogorov–Smirnov test over key
mitigation variables of interest including GHG and CO
2
2030 emis-
sion reductions, median peak warming, median warming in 2100, year
of median warming, cumulative net CO
2
emissions throughout the
century, cumulative net CO2 until net-zero and cumulative net nega-
tive CO2 after net-zero (Extended Data Fig.7). For all variables, the
Kolmogorov–Smirnov test is not able to determine whether the R5
subset comes from a different distribution than the full database sam-
ple, whereas it is able to determine the non-R5 subset is different for
peak warming and cumulative net CO
2
emissions, both of which are
shown in Extended Data Fig.8. These results indicate that the sub-
set of about 75–80% of all the C1–C3 scenarios we chose to perform
our analysis will result in sufficiently similar macro-mitigation out-
comes to represent such outcomes from the original distribution
of scenarios.
Reanalysis with OSCAR
We use OSCAR v.3.2: a version structurally similar to the one used for
the 2021 Global Carbon Budget (GCB)
44
, albeit used here with a regional
aggregation that matches the R5 IPCC regions. We first run a historical
simulation (starting in 1750 and ending in 2020) using the same experi-
mental setup as for the 2021 GCB
5,44
, with the updated input data used in
ref. 36. This historical simulation is used not only to initialize the model
in 2014 for the scenario simulations but also to constrain the Monte
Carlo ensemble (n = 1,200) using two values (instead of one in the GCB):
the cumulative land carbon sink in the absence of land-cover change
over 1960–2020 and the NGHGI-compatible emissions averaged over
2000–2020. The former is a constraint of 135 ± 25 Gt CO
2
yr
−1
(ref. 44).
The latter is a constraint of −0.45 ± 0.77 Gt CO
2
yr
−1
, using ref. 2 as a
central estimate and combining uncertainties in ELUC and SLAND from
the GCB. All physical uncertainties in this section are 1 standard devia-
tion (1σ). All values reported in the main text and figures are obtained
using the weighted average and standard deviation of the Monte Carlo
ensemble, using these two constraints for the weighting5.
To run the final scenario simulations over 2014–2100, OSCAR needs
two types of input data: (1) CO2 and local climate projections and (2) land
use and land-cover change projections. The former mostly affects the
land carbon sink (that is, the indirect effect), whereas the latter mostly
affects the bookkeeping emissions (that is, the direct effect). OSCAR
follows a theoretical framework
45
that enables a clear separation of both
direct and indirect effects. Only the direct effect is reported annually in
the GCB. Note that we do not re-evaluate the land-cover change albedo
effect because this was already included in the original AR6 database
climate projections.
Atmospheric CO2 time series is taken directly from the database, as
the median outcome estimated by the Model for the Assessment of
Greenhouse Gas Induced Climate Change(MAGICC). However, local
climate temperature and precipitation changes are not directly avail-
able. These are, therefore, computed using the internal equations
of OSCAR46, and the time series of global temperature change and
species-based effective radiative forcing (ERF) from the database (same
source). The missing components of the global ERF were treated as fol-
lows. Black carbon on snow and stratospheric H2O start at a historical
level in 2014 (ref. 47) and follow the same relative annual change as the
ERF of the scenario from black carbon and CH
4
, respectively. Contrails
are assumed constant after 2014. Solar forcing is assumed to follow the
same pathway common to all Shared Socioeconomic Pathways(SSPs).
Volcanic aerosols are assumed to be constant and equal to the average
of the historical period (that is, to have a zero ERF). Finally, we apply a
linear transition over 2014–2020 between the observed and projected
CO
2
and climate, so that these variables are 100% observed in 2014
and 100% projected in 2020. We note that the observed and projected
CO2 are virtually indistinguishable over that period but the observed
and projected regional climate changes do differ by up to a few tenths
of a degree. We further note that, because only median atmospheric
CO2, ERF and global temperature are used as input, we do not sample
and report the full physical uncertainty of the Earth system, but only
the biogeochemical uncertainty from the terrestrial carbon cycle in
response to these median outcomes.
Land use and land-cover change input data for OSCAR have three
variables: the land cover change perse, wood harvest data (expressed
in carbon amount taken from woody areas without changing the land
cover) and shifting cultivation (a traditional activity consisting of cycles
of cutting forest for agriculture, abandoning to recover soil fertility
and then returning). Wood harvest and shifting cultivation informa-
tion are not provided in the database; so we use proxy variables to
extrapolate the historical 2014 values. Wood harvest is scaled using
the Forestry Production|Roundwood variable, and shifting cultiva-
tion is scaled using Primary Energy|Biomass|Traditional as a proxy
of the development level of a region. When scenarios did not report
these proxy variables, we assumed a constant wood harvest or shifting
cultivation in the future, because these are second-order effects on the
global bookkeeping emissions.
Land-cover change is split between gains and losses that are deduced
directly as the year-to-year difference (gain if positive, loss if nega-
tive) using the following land-cover variables of the database: Land
Cover|Forest, Land Cover|Cropland, Land Cover|Pasture and Land
Cover|Built-up Area (built-up area is assumed to be constant if not
available). Land-cover change in the remaining biome of OSCAR
(non-forested natural land) is deduced afterwards to maintain a con-
stant land area. To build the transitions matrix required as input by
OSCAR, it is then assumed that the area increase of a given biome
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Article
occurs at the expense of all the biomes that see an area decrease
(within the same region and at the same time step), in proportion
to the share of total area decrease of the biomes. By construction,
this approach provides only net land-cover transitions because it is
impossible to have gain and loss in the same year, in a given region.
Therefore, and because our historical data account for gross tran-
sitions but scenarios do not, we add to this net transitions matrix a
constant amount of reciprocal transitions equal to their average
historical value over 2008–2020 to obtain a gross transitions matrix.
Finally, the three land use and land-cover change input variables
follow the same linear transition over 2014–2020 as the CO2 and
climate forcings.
We extract two key variables (and their subcomponents) from these
scenario simulations: the bookkeeping emissions (ELUC in the GCB) and
the land carbon sink (SLAND in the GCB). Following the approach in ref. 4,
the adjustment flux (that is, the indirect flux included in the NGHGIs
but not included by the IAMs, also called the factor in the main text)
required to move from bookkeeping emissions to NGHGI-compatible
emissions is calculated as the part of the land carbon sink that occurs in
forests that are managed. Therefore, we obtain the adjustment flux by
multiplying the value of SLAND simulated for forests by the fraction of
(officially) managed forests. We set this fraction to the one estimated by
ref. 4 for 2015, which also allows us to deduce the area of managed and
unmanaged (that is, intact) forest in our base year. We then estimate
how the area of intact forest evolves in each scenario, assuming that
forest gains are always managed forest (that is, they do not change
intact forest area) and that half of the forest losses are losses of intact
forest with the other half being losses of the managed forest. This
fraction is deduced from ref. 48 that estimated that around 92 Mha of
intact forest disappeared between 2000 and 2013, whereas the FAO
Global Forest Resources Assessment 2020 reports about 170 Mha of
gross deforestation over the same period. We acknowledge, however,
that applying a global and constant value for this fraction is a coarse
approximation that should be refined in future work, possibly using
information from the scenario database itself. This assumption also
implies that, as long as there is a background gross deforestation
(as is the case here, given the added reciprocal transitions), countries
will report more and more managed forest area. This is not neces-
sarily inconsistent with the Glasgow Declaration on Forest made at
COP26, as its implications in terms of pristine forest conservation
are unclear36. The subcomponents of the bookkeeping emissions are
extracted following the land categories defined in ref. 2, and we con-
sider that the net flux happening in the forest land category, excluding
shifting cultivation, is the direct contribution to land CDR. The indi-
rect contribution to land CDR would be exactly the adjustment flux
described above.
The re-analysed bookkeeping net emissions (that is, direct effect)
show an average deviation of −87 Gt CO
2
for C1 scenarios and −63 Gt
CO2 for C3 scenarios from the reported emissions in the database,
accumulated over the course of the century. Using the best-guess
transient-climate response to cumulative emissions estimated by the
IPCC (ref. 49), this implies that the global temperature outcomes of
these scenarios would differ by about −0.04 °C and −0.03 °C, respec-
tively, from what was reported in the IPCC report, if our estimates of
bookkeeping emissions were used instead of those reported by the
IAM teams.
Furthermore, after re-allocating the indirect effect in managed
forest (to align with the NGHGIs), we observe a 4.4 ± 1.0 Gt CO2 yr−1
difference between the aligned and unaligned historical LULUCF emis-
sions over 2000–2020. This number is at the lower end of the latest
6.4 ± 1.2 Gt CO2 yr−1 provided in the 2022 GCB3. Compared with the
6.7 ± 2.5 Gt CO
2
yr
−1
difference reported in ref. 2, and correcting for
the absence of organic soils emissions in our simulations with OSCAR
(about 0.8 Gt CO2 yr−1), OSCAR can explain about 75% of the observed
difference. Although OSCAR typically produces fairly central estimates
of the direct effect
3
, its estimates of the indirect effect show a biased
high CO2 fertilization50.
Comparing adjusted pathways with current policy and NDC
estimates
We use the latest available estimate of aggregate NDCs from ref. 1 to
compare with the NGHGI-adjusted global pathways. The 1.5 °C and
2.0 °C pathways we use are the same as previously discussed: the IPCC
C1 and C3 pathways with sufficient land cover detail at the R5 region.
We additionally re-analyse the current-policy pathways from the IPCC
AR6 database. These correspond to pathways consistent with the
current policies as assessed by the IPCC, or the P1b pathways as per the
AR6 database metadata indicator Policy_category_name.
We incorporate an endogenous estimation of the indirect effect with
OSCAR, which varies over time based on land-cover pattern changes
and changes to carbon-cycle dynamics and carbon fertilization. As
such, we compare our central estimate of global GHG emissions in 2015,
approximately 49.4 Gt CO
2
-equiv to that in ref. 1, 51.2 Gt CO
2
-equiv,
resulting in a difference of 1.8 Gt CO2-equiv. We then apply this offset
value (1.8 Gt) to all estimations of 2030 emission levels in ref. 1 to pro-
vide comparable levels with our pathways. This ensures that the NDC
targets calculated based on national inventories become comparable
with the NGHGI-adjusted modelled pathways.
Data availability
All data generated and analysed here are available at GENIE Scenario
Explorer (https://data.ece.iiasa.ac.at/genie).
Code availability
OSCAR is an open-source model available at GitHub (https://github.
com/tgasser/OSCAR). Source code for all analysis files is available at
GitHub (https://github.com/iiasa/gidden_ar6_reanalysis).
40. Riahi, K. etal. in IP CC, 2022: Climate Change 2022: Mitigation of Climate Change.
Contribution of Working Group III to the Sixth Assessment Report of the Intergovernmental
Panel on Climate Change (eds Shukla, P. R. etal.) (Cambridge Univ. Press, 2022).
41. Kikstra, J. S. etal. The IPCC Sixth Assessment Report WGIII climate assessment of
mitigation pathways: from emissions to global temperatures. Geosci. Model Dev. 15,
9075–9109 (2022).
42. Mace, M. J. Mitigation commitments under the Paris Agreement and the way forward.
Climate Law 6, 21–39 (2016).
43. Schleussner, C.-F., Ganti, G., Rogelj, J. & Gidden, M. J. An emission pathway classiication
relecting the Paris Agreement climate objectives. Commun. Earth Environ. 3, 135
(2022).
44. Friedlingstein, P. etal. Global Carbon Budget 2021. Earth Syst. Sci. Data 14, 1917–2005
(2022).
45. Gasser, T. & Ciais, P. A theoretical framework for the net land-to-atmosphere CO2 lux and
its implications in the deinition of “emissions from land-use change”. Earth Syst. Dyn. 4,
171–186 (2013).
46. Gasser, T. etal. The compact Earth system model OSCAR v2.2: description and irst results.
Geosci. Model Dev. 10, 271–319 (2017).
47. Smith, C. etal. in Climate Change 2021: The Physical Science Basis. Contribution of Working
Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change
(eds Masson-Delmotte, V. etal.) 923–1054 (Cambridge Univ. Press, 2021).
48. Potapov, P. etal. The last frontiers of wilderness: Tracking loss of intact forest landscapes
from 2000 to 2013. Sci. Adv. 3, e1600821 (2017).
49. Canadell, J. G. etal. in Climate Change 2021: The Physical Science Basis. Contribution
of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on
Climate Change (eds Masson-Delmotte, V. etal.) 673–816 (Cambridge Univ. Press,
2021).
50. Quilcaille, Y., Gasser, T., Ciais, P. & Boucher, O. CMIP6 simulations with the compact Earth
system model OSCAR v3.1. Geosci. Model Dev. 16, 1129–1161 (2023).
Acknowledgements We thank M. Sanz and C.-F. Schleussner for the comments on an initial
draft of the paper, M. Beer and F. Spagopoulou from the Designers for Climate for expert
support on visualizations, and the reviewers whose comments improved the quality of the
paper. We acknowledge the funds received from the Horizon Europe research of the European
Union and the innovation programme RESCUE, grant agreement no. 101056939 (M.J.G. and T.G.);
the Horizon 2020 research of the European Union and the innovation programme ESM2025—
Earth System Models for the Future, grant no. 101003536 (T.G. and Z.N.); and the ERC-2020-SyG
Content courtesy of Springer Nature, terms of use apply. Rights reserved

GENIE grant of the European Union, grant no. 951542 (M.J.G., W.F.L., J.M. and K.R.). The views
expressed are those of the writers and may not under any circumstances be regarded as
stating an oficial position of the European Commission.
Author contributions M.J.G., T.G. and K.R. contributed to the conceptualization; M.J.G.,
T.G., G.G., I.J. and Z.N. devised the methodology; M.J.G. and T.G. helped in the investigation;
T.G. provided the software support; M.J.G. helped with visualization; M.J.G. wrote the
original draft; and M.J.G., T.G., G.G., N.F., I.J., W.F.L., J.M., Z.N., J.S. and K.R. reviewed and
edited the paper.
Competing interests The authors declare no competing interests.
Additional information
Supplementary information The online version contains supplementary material available at
https://doi.org/10.1038/s41586-023-06724-y.
Correspondence and requests for materials should be addressed to Matthew J. Gidden.
Peer review information Nature thanks Jennifer Burney, H. Damon Matthews and Chris Jones
for their contribution to the peer review of this work. Peer reviewer reports are available.
Reprints and permissions information is available at http://www.nature.com/reprints.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Article
Extende d Data Fig. 1 | Sce nario-wi se mitigatio n benchmark s hift. Th e
change bet ween estimat es of mitigati on benchmarks for 1 .5 C (blue, IP CC
categor y C1), 1.5C-OS ( green, IPCC ca tegory C2), and 2 C (purp le, IPCC
category C3) scenar ios. Origin al values from the A R6 database (which foll ows
IAM repo rting conventi ons) are shown as circle s whereas values d erived from
reanalyze d scenarios i n this study (in line w ith NGHGI rep orting convent ions)
are shown as tr iangles. T he estimate s of the year of globa l net-zero CO2 (pane l a),
emission s reductions b etween 2020 a nd 2030 (panel b), and cumula tive
CO2 emission s (panel c) are shown. E ach pair of dots and t riangles rep resents
results f rom a single scena rio, with scen arios ordered alo ng the y-axis base d
on the values i n the original A R6 dataset.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Extende d Data Fig. 2 | NGH GI-adjuste d global GHG pat hways compared with N DCs and curr ent policie s. The interqu artile range show n and median
highlight ed is plotted to gether with cu rrent estima tes of 2030 ag gregated nat ional climate t arget levels and curr ent policy est imates from de n Elzen etal. (2022).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Article
Extende d Data Fig. 3 | The 2 030 emissi ons gap betwe en current
policie s and pledge s. 1.5 C an d 2 C as assess ed in this study an d by den Elzen
(2022) is compared against levels of current policies, conditional NDCs, and
uncondit ional NDCs as rep orted in den El zen (2022). Median e stimates of all
values are us ed to compute the re spective e mission gaps .
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Extende d Data Fig. 4 | Th e change in the e stimated 2 030 emissi on gap
betwee n due to alignm ent to NGHGI co nventions. T he total magn itude, left,
relative value , right. Each b ar represent s the median value w ith the interqu artile
range of the es timate across s cenarios. T hese change s occur differe ntly across
different re gions bet ween pathways follow ing model-bas ed conventions an d
adjusted p athways following NGH GI-based conve ntions. A posi tive value
means tha t the gap is larger wh en consider ing both (i.e. wh en aligned to N GHGIs),
and a negat ive value means the g ap is smaller. Regions l abels conform to I PCC
5-region la bels for Asia, La tin America , Middle East and A frica, the O ECD and
EU, and the Reform ed Economies , respective ly (Extended D ata Table4).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Article
Extende d Data Fig. 5 | Gro ss carbon re moval levels. Gros s carbon removal leve ls from LULUCF (reanaly zed with OSC AR) by direct ef fects (g reen) and indirec t
effect s (purple) across 1 .5 C and 2 C pathways . Interquart ile ranges of eac h estimate are show n by error bars.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Extende d Data Fig. 6 | Cum ulative carbo n sequeste red on land st arting
from 2020. Gross cumulative c arbon removal levels s tarting fr om 2020 from
LULUCF (reanaly zed with OSCA R) by direct ef fects (gr een) and indirect e ffects
(purple) acros s 1.5 C and 2 C pat hways. Removals in bot h categorie s increase by
midcentur y, but at different level s. Both pathway ca tegories se e similar total
cumulative re moval levels by the end of th e century wi th varying st rength of
indirect removals.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Article
Extende d Data Fig. 7 | K-S test of s cenario di stributi ons. Kolmogorov-
Smirnov (K-S) test result s for key mitigation i ndicators for the f ull set of C1-C3
scenari os, those scen arios having all l and-cover variabl es define d at the R5
region level, a nd those not havi ng all land-cover vari ables defi ned at the R5
region level . The null hypoth esis of the K-S test is that t wo dataset va lues are
from the sam e distributi on. For all indicato rs derived from s cenarios incl uding
land-cover var iables data at th e R5 region level, we c an not reject th e null
hypothe sis (p > 0.05). Some indic ators of the sce nario set wit hout land-cover
data (not use d in this analysis) do reje ct the null hyp othesis.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Extende d Data Fig. 8 | Mit igation met rics from sce nario subs ets. Key
mitigation metrics where scenarios without R5 region coverage (in red)
cannot repl icate the full dat abase outcom e. The blue bar pre sents the
outcome for th e full database , scenarios w ith global values of l and-cover
variables a nd R5 values are sh own in yellow, and scenar ios with land-cover
variables a t the R5 regio n are shown in green . The red bar shows how t he
distribution changes when considering the population of scenarios without
full variabl e coverage (‘No R5 all ’).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Article
Extended Data Table 1 | Indirect LULUCF lux estimates aligned with NGHGIs
Median values and interquartile ranges of the indirect lux in Gt CO2yr−1 estimated by OSCAR per R5 IPCC region (see Extended Data Table4 for region deinitions). This value is computed for
every scenario with suficient land-use data (seeMethods) in each model region for every point in time. This value constitutes the ‘NGHGI Adjustment Factor’ and is computed and added to
each scenarios’ estimated direct LULUCF lux values to quantify emissions pathways from global models aligned with NGHGI LULUCF reporting conventions.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Extended Data Table 2 | Updated mitigation benchmarks
Net mitigation outcomes from scenarios: (a) as assessed by the IPCC in AR6, (b) with direct effects of LULUCF reanalyzed by OSCAR, and (c) including both direct and indirect effects of LULUCF
(i.e. aligned to NGHGIs). All values provided as medians with 5th−95th percentile ranges in parentheses.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Article
Extended Data Table 3 | Variable coverage of scenarios
Fraction of AR6 database scenarios with land-use variables of interest, per scenario category.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Extended Data Table 4 | Regional deinitions
Deinitions of IPCC 5-region macro regions as listed in the IPCC AR6 database.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com