Page 1
1
Soil organic carbon stocks potentially at risk of decline in organically farmed croplands
1
2
Ulysse Gaudaré
1
, Matthias Kuhnert
2
, Pete Smith
2
, Manuel Martin
3
, Pietro Barbieri
4
, Sylvain
3
Pellerin
1
, Thomas Nesme
4
.
4
5
1
INRAE, UMR 1391 ISPA, CS 20032, Villenave d’Ornon, France.
6
2
Institute of Biological and Environmental Sciences, University of Aberdeen, 23 St. Machar Drive,
7
Aberdeen AB24 3UU, UK
8
3
INRAE, Info&Sol, Orléans, France
9
4
Bordeaux Sciences Agro, University of Bordeaux, UMR 1391 ISPA, CS 40201, Gradignan,
10
France.
11
12
13
Increasing soil organic carbon (SOC) stocks in agricultural lands is key to mitigate climate
14
change and organic farming shows promising results. Evidence of higher SOC stocks in
15
organic farms compared to conventional farms reflects current situations where organic
16
farming occupies small fractions of agricultural areas, with access to ample amounts of
17
resources for organic fertilisation. Using a modelling approach, we estimated global SOC
18
stocks following a 100% conversion to organic farming of global croplands under a normative
19
and an optimal organic scenario. We found that global soil carbon inputs would be reduced by
20
39% and 29% for both scenarios respectively, leading to a 9% and a conservation of global
21
SOC stocks reduction (with spatial variations) after 20 years in the normative and optimal
22
organic scenario, respectively. These results suggest that an expansion of organic farming
23
might reduce its potential to mitigate climate change through soil carbon sequestration unless
24
appropriate practices – such as widespread cover cropping and enhanced residue recycling –
25
are implemented.
26
27
28
29
The agricultural sector is responsible for 23% of global anthropogenic greenhouse gas (GHG)
30
emissions worldwide
1,2
, but there is an opportunity for mitigation of climate change through carbon
31
sequestration in agricultural soils. While arable lands have lost up to half of their organic carbon
32
stocks since the industrial revolution
5
, agricultural practices could help increase soil organic carbon
33
stocks, by increasing carbon inputs to soils or by reducing soil carbon mineralisation
6
.
34
35
Organic farming is proposed as a way to increase soil organic carbon (SOC) stocks
7
.
Meta-analyses
36
of field experiments have shown that organically managed cropland soils have, on average, higher
37
SOC stocks (+3.5 tC.ha
-1
) and soil carbon sequestration rate (+0.45 tC.ha
-1
.yr
-1
) than conventional
38
(i.e. non-organic) ones
8,9
. These results are largely explained by higher soil carbon inputs in organic
39
systems through both enhanced manure application rates and the use of more complex crop
40
rotations with higher frequency of temporary pastures and cover crops
10
, resulting in higher organic
41
carbon inputs to soils. However, concerns have been raised that these positive effects of organic
42
farming may result from carbon transfers from other ecosystems through manure and compost
43
inputs, so that there may be no net change in carbon stocks over the whole land area
11
. Accounting
44
for these lateral carbon transfers and capturing their effects is therefore essential for obtaining
45
accurate estimates of the potential of organic farming to sustain global SOC stocks.
46
47
Organic farming occupies less than 2% of the global utilized agricultural area (UAA)
12
. Evidence
48
provided by meta-analyses reflect situations where organic materials, such as animal manure or
49
compost are readily available for fertilisation of organically managed soils
13
. In contrast, the
50
Page 2
2
expansion of organic farming might trigger competition for fertilising resources, possibly resulting
51
in a reduction of potential for soil carbon inputs and soil carbon sequestration. A recent study has
52
shown that organic farming upscaling to 100% of the UAA would lead to a 56% crop yield
53
reduction due to severe nitrogen (N) limitation
14
– a significant drop compared to the 20-30% yield
54
reduction previously reported by field-based meta-analyses
15,16
. This drop is mostly due to the ban
55
of synthetic N fertilizers in organic guidelines that reduces both the range of N fertilization
56
resources (e.g., crop residues, livestock manure) and their global availability, with large
57
consequences for soil fertilisation – a result confirmed by recent studies highlighting N fertilisation
58
limitation when organic farming is upscaling
17–19
. Expansion of organic farming is thus likely to
59
have major consequences for soil carbon inputs from crop residues and fertilising materials,
60
potentially resulting in large changes in SOC stocks.
61
62
Capturing these systemic feedbacks is key to accurately estimating soil carbon inputs in scenarios of
63
large-scale organic farming. We addressed these knowledge gaps by combining (i) GOANIM, a
64
spatially explicit model simulating cropland N cycle, crop productivity and livestock populations
65
under scenarios of large organic farming expansion
14
with (ii) RothC, a dynamic, first order kinetic
66
model simulating carbon dynamics in soils
20,21
. We used GOANIM outputs about livestock manure
67
and crop residue production to estimate carbon fluxes between croplands, grasslands and livestock,
68
and to estimate soil carbon inputs (SCI) in scenarios of large organic farming expansion for
69
croplands. We then used the estimated SCI as an input to RothC to simulate the changes in SOC
70
stocks under different time horizons. We assessed different scenarios combining (i) variations in
71
organic farming practices (e.g., cover cropping, use of conventional manure on organic croplands,
72
residue recycling) and (ii) variations in the level of organic farming expansion globally, each
73
compared with a baseline scenario of no changes in current agricultural practices.
74
75
Although all organic regulations are gathered under the ban of synthetic fertilisers
22
, organic
76
farming encompasses a diverse set of farming practices, depending on regional regulations, farming
77
contexts and markets
16,23,24
. In particular, organic farmers may adopt cropping practices that are
78
known to improve soil carbon sequestration (e.g. cover cropping, extensive crop residues recycling,
79
diversified crop rotations including pasture). We captured this variability in cropping practices by
80
considering both (i) a normative organic scenario in which organic farming is restricted to the ban
81
of synthetic fertilizers, some differences in crop rotations, no cover-crops and a redistribution of
82
livestock population compared to conventional farming and (ii) an optimal organic scenario that
83
may favour carbon inputs to cropland soils mostly through extensive cover-cropping and enhanced
84
residue recycling. Note that the assumptions related to the normative scenario were well aligned
85
with those of a previous study about organic farming expansion that resulted in drastic reduction of
86
global cropland production and livestock population reduction in a fully organically managed
87
world, with a large shift towards ruminant animal species
14
. In contrast, the optimal scenario was
88
well aligned with observational data that show that covering soils by catch and cover-crops is a
89
common practice that many organic farmers implement
10,11
. We hypothesized that, in the normative
90
organic scenario, both soil carbon inputs and SOC stocks would be negatively affected by a global
91
transition to organic farming whereas those negative effects can be partly ameliorated when
92
additional cropping practices are considered, as in the optimal organic scenario. Hereafter, we first
93
focus on results from a hypothetical 100% conversion of cropland areas to organic farming before
94
analysing scenarios with an intermediate level of organic farming expansion. The scenarios are not
95
intended to be prescriptive; rather they are exploratory, offering a framework for analysis. Thus, the
96
primary goal of our modelling exercise is not to assess if organic farming will change SOC stocks,
97
but rather to explore if, how and where SOC stocks could be at risk of decline under organic
98
farming expansion.
99
100
Page 3
3
101
Results
102
103
Carbon flows and stocks in an organic world
104
105
SCI reduction in an organic world
106
107
Table 1.
Global soil carbon inputs (PgC.yr
-1
) for croplands under both 100% organic
108
scenarios and the baseline.
109
Plant-based
residues
Manure
Total
Baseline
2.50
0.22
2.72
100% organic
scenario
Normative
1.51
0.11
1.62
Optimal
1.77
0.11
1.87
Ratio organic /
baseline
Normative
0.61
0.48
0.60
Optimal
0.71
0.48
0.69
110
Globally, we found a 40 and 31% reduction in the total SCI to croplands for the normative and
111
optimal organic scenarios, respectively (
Table 1
). Such massive drop of SCI is primarily due to (i)
112
39% and 29% reduction in plant-based residues returned to the soil (-1 PgC.yr
-1
and -0.7 PgC.yr
-1
),
113
followed by (ii) a 68% reduction in farmyard manure application rate (-0.11 PgC.yr
-1
) in both 100%
114
organic scenarios compared to the baseline. In the normative organic scenario, the reduction in
115
plant-based residues returns is mainly due to a 51% reduction of annual crop dry matter production,
116
partially attenuated by increased frequency of temporary rotational pastures, resulting in an overall
117
47% reduction of cropland biomass production
(Supplementary Table 1)
. The reduction in manure
118
application rate is mainly due to a 66% reduction in the global livestock population, as well as
119
changes in animal types and in the regional distribution of livestock populations. In the optimal
120
organic scenario, the additional 0.25 PgC.yr
-1
carbon inputs compared to the normative organic
121
scenario is explained by 83% of additional SCI from the use of cover crops on 50% of organically
122
managed croplands (+0.21 PgC.yr
-1
, +0.07 tC.ha
-1
.yr
-1
on average).
123
124
These global changes in soil carbon inputs mask large variations among world regions
(Figure 1)
.
125
In some specific regions – such as Central Africa or Russia – soil carbon inputs are increased in the
126
normative 100% organic scenario compared to the baseline. This is explained by higher inputs as
127
plant-based residues
(Supplementary Figure 1)
due to (i) high manure application rates that help
128
to sustain high crop yields in organic farming
(Supplementary Figure 1)
and (ii) high share of
129
carbon fixing crops – such as temporary pastures – in organic rotations
10,25
. Note, that in other
130
regions – such as Northern Brazil – the increase in plant-based residues resulting from more
131
frequent carbon fixing crops in organic rotations is offset by a drop in farmyard manure application,
132
resulting in reduced soil carbon inputs to cropland soils. In the optimal 100% organic scenario, we
133
found that regions such as Central Canada, Eastern Europe or Southern Russia have a higher soil
134
carbon inputs compared to the baseline (
Figure 1b
). In those regions, additional soil carbon inputs
135
from cover crops are sufficient to compensate the reduction of soil carbon inputs due to drop in crop
136
production resulting from the ban of synthetic fertilizers (
Supplementary Figure 1
).
137
138
Page 4
4
139
Figure 1.
Maps of annual
organic-to-baseline ratios of soil total carbon inputs for the normative (left) and
140
optimal (right) 100% organic scenario.
141
142
143
SOC stocks
144
145
In the normative scenario, the transition to 100% organic farming would result in a 9, 13 and 18%
146
SOC stock reduction in croplands after 20, 50 and 100 years, respectively, compared to the baseline
147
(
Table 2
). This reduction would represent an overall loss of -6.8 PgC from croplands in the first 20
148
years after that transition and a mean loss of 0.23 tC.ha
-1
.yr
-1
. However, a transition to 100%
149
organic farming in the optimal scenario would result in the conservation or slight increase in
150
croplands SOC stock. In particular, cropland SOC stocks would slightly increase, by 0.3 PgC 20
151
years after the transition to organic farming, leading to an average storage of 0.01 tC.ha
-1
.yr
-1
.
152
153
Table 2. Global changes in SOC stocks (PgC) in croplands after 20, 50, and 100 years following conversion to
154
organic farming.
Ratios and differences between the organic and the baseline are indicated.
155
Global soil organic carbon stocks [PgC]
20 years
50 years
100 years
Baseline
75.7
100% organic scenario
Normative
68.9
65.5
62.3
Optimal
76.1
77.1
78.5
Ratio org / baseline
Normative
0.91
0.87
0.82
Optimal
1.00
1.02
1.04
Difference org -
baseline [tC.ha
-1
.yr
-1
]
Normative
-0.23
-0.23
-0.18
Optimal
0.01
0.03
0.04
156
Again, these global results mask spatial variations among world regions (
Figure 2
). In the
157
normative scenario, cropland SOC stocks increase in some regions (such as central Africa) while
158
others they decrease in others (such as India and Mexico) (
Figure 2b
) – a result largely explained
159
by regional variations in soil carbon inputs (
Figure 1a
). In the optimal scenario, some of those latter
160
regions (such as India) would experience an increase in cropland SOC stocks. Those regions are
161
marked by high potential of additional SOC stocks per hectare due to cover cropping (
Figure 3
).
162
This positive effect of cover crops in the optimal scenario is due to (i) an additional soil carbon
163
input of +0.07 tC.ha
-1
.yr
-1
on average on global cropland soils and (ii) a ground covering effect that
164
reduces soil carbon mineralisation. Both effects result in an additional global mean increase in
165
cropland SOC of +0.47 tC.ha
-1
.yr
-1
over the 20 first years following conversion to organic farming.
166
Page 5
5
167
Figure 2.
Global changes in soil organic carbon (SOC) stocks (PgC) in croplands over time, and maps of the SOC
168
stock ratios between the 100% organic scenarios (either normative or optimal) and the baseline at 20 years.
169
Changes in global SOC stocks in croplands and spatial distribution are reported for the normative (red line) and optimal
170
(blue line) 100% organic scenarios. The black dashed line represents the global SOC stocks for croplands in the
171
baseline.
172
173
174
Figure 3
.
Additional SOC stocks per ha [tC.ha
-1
.yr
-1
] due to cover cropping in the optimal organic scenario
175
compared to the normative organic scenario.
176
177
In the normative scenario, SOC stocks reduced drastically in the first 20 years after transitioning to
178
organic farming (-0.5 % per ha and per
year on average), whereas the SOC reduction would slow
179
down thereafter (-0.2 % per ha and per year on average)
(Supplementary Figure 2)
. This rapid
180
Page 6
6
decline in the first 20 years followed by slower loss after 20 years is frequently observed in field
181
studies
26,27
.
182
183
184
Intermediate scenarios of organic farming expansion
185
186
Because converting the entire agricultural area to organic farming is a drastic thought experiment,
187
we also explored more realistic scenarios of intermediate conversion to organic farming. In those
188
intermediate scenarios, manure surplus from conventional farming systems – i.e. conventional
189
manure that is in excess compared with conventional cropland N requirements – may be applied on
190
organically farmed lands. Therefore, we introduced two variants of our normative and optimal
191
organic scenarios by considering (i) the application or (ii) the ban of conventional manure surplus
192
on organically managed lands.
193
194
We found that, in situations without conventional manure application, changes in global SOC stocks
195
in croplands was linearly correlated with increasing share of the UAA under organic farming. This
196
linear relationship would be strongly negative in the normative organic scenarios, reflecting that
197
expanding normative organic systems would put SOC stocks in global croplands at risk. In contrast,
198
the slightly positive relationship between global SOC stocks and share of UAA under organic
199
farming in the optimal organic scenarios suggests that sustaining expansion of diversified organic
200
systems would help to protect SOC stocks
(Figure 4a)
.
201
202
Using conventional manure surplus as an additional, external source of organic fertilising material
203
on organically managed croplands – a practice often implemented by organic farmers
13,28
– would
204
make SOC stocks non-linearly correlated with the share of the global UAA under organic farming
205
(Figure 4a)
. In both the normative and optimal organic scenarios, applying conventional manure
206
would help to increase global SOC stocks as well as SOC sequestration rates
(Figure 4a and b)
.
207
For instance, in the normative scenario, when 20% of the global UAA is converted to organic
208
farming, agricultural SOC stocks would be close to those reported in the baseline 20 years after the
209
conversion. Transferring animal manure from conventional to organic systems increases SOC
210
stocks in organically managed lands through both direct effects (through the application of
211
additional soil carbon input to organic soils) and indirect effects (by alleviating at least partly their
212
often reported N deficiency
14–16
thereby boosting organic crop yields with positive feedback on crop
213
residues returns to soils).
214
Some regions – such as the UK, Northern India and Northern China – would see their cropland
215
SOC stocks increasing compared to the baseline in both the normative and optimal scenarios
216
(
Figure 4c
). In those same regions, SOC stock would decrease in a scenario with 20% of the UAA
217
under organic farming without conventional manure application compared to the baseline. This
218
regional effect is explained by the uneven geographic distribution of conventional manure surpluses
219
at the global scale
(Supplementary Figure 3)
, with major consequences for soil carbon inputs.
220
Interestingly, our results also show that SOC stocks in conventionally managed lands would remain
221
constant with or without the use of conventional manure surplus on organically managed lands
222
(
Supplementary table 2
). This absence of an effect of transferring carbon from conventionally to
223
organically managed lands is explained by the small share (less than 1%) that conventional manure
224
surplus represents over the total soil carbon inputs in conventionally managed lands.
225
226
Page 7
7
227
Figure 4.
Evolution of global SOC stocks (PgC) at 20 years (a) and mean difference (organic minus baseline) of
228
SOC sequestration rate (tC.ha
-1
.yr
-1
) over the first 20 years (b) with maps of SOC stock ratio at 20 years and
229
with 20% of the global UAA under organic farming (c).
In both upper panels, the red lines represent the normative
230
organic scenario and the blue line the optimal organic scenario. The dashed lines represent situations where
231
conventional manure surplus are applied on organically managed croplands whereas the solid lines represent situations
232
without conventional manure application.
233
234
Achieving 20% of the global UAA under organic farming – although being far above the current
235
1.5% share of organic farming – is the most realistic of the situations we simulated. In this situation,
236
we found that global SOC stocks would decrease by -2% to -1% in the normative organic scenario
237
(without and with conventional manure, respectively) whereas they would increase by +0.1% to
238
+1% in the optimal organic scenario (without and with conventional manure, respectively). This
239
would translate into a -0.118 tC.ha
-1
.yr
-1
difference in SOC sequestration rate between organic and
240
conventional farming (with conventional manure) in the normative organic scenario, whereas this
241
difference would increase to +0.124 tC.ha
-1
.yr
-1
in the optimal organic scenario (
Figure 4b
,
242
Supplementary table 2
).
243
244
245
Page 8
8
Discussion and conclusion
246
247
Contrary to what is sometimes claimed
29,30
, our results suggest that global SOC stocks may be at
248
risk of decline if organic farming expands, especially if the expansion occurs through normative
249
organic farming systems. This would result from a drastic reduction in global soil carbon inputs
250
(SCI), mostly as crop residues and animal manure due to large N deficiency, resulting in severe
251
decline in crop production, as well as a reduction in livestock populations
14
. In addition, our results
252
show that SOC stocks could be conserved under the optimal organic scenarios, thanks to extensive
253
cover-cropping and enhanced residue recycling. Our findings are in contrast to previous studies
254
reporting strong carbon sequestration potential of organic farming based on field observations at the
255
local scale
8
. These results highlight that soil carbon impacts of organic farming cannot be assessed
256
simply from extrapolation of local field observations without considering whole-system effects. The
257
assessment of the impacts of expansion of organic farming systems needs to consider the systemic
258
feedbacks that go along with organic farming expansion itself
31
, in particular the availability of
259
fertilising resources and related effects on crop production
14,32
.
260
261
Our results are, however, fairly well aligned with local reports on organic farming expansion. For
262
instance, the N deficiency – and its resulting effects on crop biomass production –simulated by the
263
GOANIM model here is consistent with local observations that N fertilising resources may become
264
scarce if organic farming expands widely, as recently highlighted in France
33
, India
34
or Bhutan
35
. In
265
addition, our results on limited SOC benefits from organic farming are consistent with findings
266
from a recent meta-analysis that organic farming may not increase SOC stocks compared to
267
conventional farming if there is no lateral carbon transfer from other agroecosystems
11
. Finally, our
268
global estimates of 0.124 tC.ha
-1
.yr
-1
SOC sequestration rates in the optimal organic scenario and
269
under 20% of the global UAA under organic farming are close to the 0.07-0.14 tC.ha
-1
.yr
-1
values
270
reported from an extensive meta-analysis on SOC sequestration potential of organic farming when
271
lateral carbon transfers are controlled
8
.
272
273
Besides those global estimates, our results also show that a range of additional cropping practices
274
could sustain or increase SOC stocks in organically managed croplands. In particular, we found that
275
the extensive use of cover crops is key to increase SOC stocks through both increasing SCI and
276
reducing SOC mineralisation
36–39
. Estimating the real benefits that extensive use of cover-crops
277
could bring for SOC stocks in organic farming at the global scale is subject to many uncertainties
278
given the lack of precise information on (i) potential areas available for cover cropping, (ii) spatially
279
explicit species composition of the cover crops and (iii) cover crops biomass potential production.
280
However, the potential additional SOC stocks offered by cover crops that we found in our study
281
(0.29 tC.ha
-1
.yr
-1
) is very similar to the 0.32 tC.ha
-1
.yr
-1
value reported in a recent meta-analysis
40
.
282
283
Other practices – such as agroforestry
41,42
, enhanced circularity
38,43
and increased frequency of
284
temporary N-fixing leys or cover-crops in organic rotations
15,16
– may have positive impacts on N
285
resource conservation (by avoiding nitrate leaching
45
), N supply to plants
38,44
and SOC stocks.
286
External fertilising organic materials – such as urban compost, green wastes, food industry by-
287
products or eventually sewage sludge – could also provide N to soils as well as providing additional
288
soil carbon inputs
46
. Modelling the benefits brought by this extensive set of additional cropping
289
practices was beyond the scope of this study but our results suggest that making organic farming
290
more climate beneficial will require some of these additional practices.
291
292
Modelling variations in soil organic carbon stocks in different farming scenarios at the global scale
293
has some limitations. In particular, SOC stocks were modelled using RothC, a model that has
294
proved its potential to accurately simulate SOC changes at the local
47
and large
48
scales, but that
295
Page 9
9
requires some specific modelling assumptions. Among them, we had to assume that carbon stocks
296
in the baseline are at the equilibrium
48
. It is likely that this assumption does not always reflect the
297
reality
49,50
which may have implications for our findings. However, we found evidence that the
298
error brought by this assumption was negligible with only 1% reduction of global croplands SOC
299
stocks after 100 years compared to the initial situation when SOC stocks were not considered at the
300
equilibrium in the baseline (see Supplementary Information). Another limitation may be related to
301
the fact that the soil organic carbon mineralisation tracks nitrogen mineralisation
51
which may
302
sustain plant growth, a factor we did not consider in our study. This may lead to a slight over-
303
estimation of SOC stock reduction due to over-estimating the reduction in soil carbon inputs
304
compared to the baseline, an effect that should be addressed in further analyses.
305
306
The estimates of global changes in SOC stocks in croplands provided by this study should be
307
complemented by similar estimates for grasslands. Indeed, carbon transfers between grasslands and
308
croplands through livestock grazing and manure collection and disposal on croplands – although
309
probably minimal at the global scale – may affect local SOC stocks under grasslands, especially
310
when livestock species and spatial distribution are modified in organic farming. However, we found
311
that converting global agriculture to organic farming would result in small changes in grassland
312
SOC stocks (see Supplementary Information). Additionally, the region with the biggest effects is
313
India, where information on grasslands management is highly uncertain
52–55
, calling for caution in
314
interpreting the estimates of grassland SOC stocks.
315
316
Simulations were performed considering recent past climate. However, ongoing climate change is
317
likely to affect (i) crop yields and livestock farming, with major consequences on soil carbon inputs
318
to agricultural soils and (ii) SOC mineralisation through a series of processes that are soil
319
temperature and moisture dependent. Accounting for those climate change effects would make
320
sense to allow mitigation and adaptation the be explored together. However, modelling climate
321
change effects on SOC stocks in organic farming would require a series of additional and disputable
322
assumptions (about climate change effects on crop yields, cropping area spatial distribution,
323
livestock farming and animal production
57
), and would likely result in increased uncertainties. .
324
More importantly, the literature critically lacks of data about how climate change effects would
325
differ in organic vs. conventional farming
9
. Addressing these issues is necessary to derive accurate
326
estimates of SOC stocks in organic farming under future climate.
327
328
This study provides important information to estimate the potential of organic farming to reduce
329
GHG emissions from agriculture. Our results provide an alternative estimate of changes in SOC
330
stocks following conversion to organic farming, to those which upscale SOC stock differences
331
based on field observations
17,59
. Because organic farming expansion is also likely to affect CH
4
and
332
N
2
O emissions through a series of processes related to rice cultivation, animal husbandry, manure
333
management, and N fertilisation, deriving accurate estimates for those emissions is much needed in
334
order to complement our SOC stock change estimates provided in this study.
335
336
Page 10
10
Methods
337
338
The objective of this study was to estimate the potential impact of global organic farming expansion
339
on soil organic carbon (SOC) stocks. To do so, we used a modelling approach to estimate the SOC
340
stock changes in scenarios of global organic farming expansion compared to the currently observed
341
SOC stocks. Currently, organic farming occupies less than 2% of the global agricultural lands.
342
Therefore, we consider that the currently observed SOC stocks are those observed under
343
conventional farming, hereafter called the baseline. The modelling approach was based on two
344
separate steps, as explained below.
345
346
First, we estimated the soil carbon inputs (SCI) in scenarios of large organic farming expansion and
347
in the baseline for croplands in a spatially explicit way (5 arc-min resolution, i.e. ~10x10km at the
348
equator). In both the organic scenarios and the baseline, we estimated the SCI as a sum of (i) the
349
amount of carbon that is returned to agricultural lands as plant residues (crop-based and grass-based
350
residues) and (ii) the amount of carbon excreted by animals as farmyard manure (FYM) applied to
351
lands after accounting for C losses during manure storage. The SCI estimates for organic farming
352
scenarios were computed using outputs from the GOANIM model
14
. GOANIM is a spatially
353
explicit (5 arc-min resolution) linear optimisation model that simulates nitrogen flows to and from
354
croplands and grasslands under scenarios of organic farming upscaling. GOANIM calculates
355
cropland N budget and its effects on crop yield for 61 crop species. The optimising module of
356
GOANIM is designed to maximise food availability at the global scale (from both crop-based and
357
animal-based products) by spatially optimising the global livestock population and the N allocation
358
from animal manure to the different considered crops. We used the latest version of GOANIM,
359
accounting for (i) differences in feed rations and feed use efficiency between organic farming and
360
conventional farming
60
, (ii) the 2019 refinement of the IPCC guidelines values on manure
361
management and nitrogen losses (as direct N
2
O emissions, nitrate leaching and ammonia
362
volatilisation) and (iii) representation of non-productive, young animals. Further details about the
363
GOANIM model can be found in Barbieri et al. 2021
14
, especially about the case of Sub-Saharan
364
Africa where drops in yields following the conversion to organic farming due to factors other than
365
N limitation (e.g., poor pest and weed control) were negligible. In addition, two organic farming
366
scenarios were considered in this study: (i) a normative organic scenario in which organic farming
367
is restricted to the ban of synthetic fertilizers, differences in the type of crop grown in crop rotations
368
as reported by Barbieri et al. 2019
25
, no cover-crops and redesign of the global livestock population
369
as reported by Barbieri et al. 2021
14
, and (ii) an optimal organic scenario that draw upon the
370
normative scenario but with cover cropping implemented on 50% of the bare soil periods between
371
two cash crops (in organically managed lands), increased root-shoot ratio and enhanced plant-based
372
residues recycling on croplands (see below for additional details on this optimal scenario).
373
374
Second, we used the estimated SCI from both organic scenarios as inputs to the RothC
20,21
model to
375
estimate changes in SOC stocks over the 0-30 cm soil depth, in context of large organic farming
376
upscaling, considering only annual crops (which represents 45 of the 61 crops in GOANIM, thereby
377
assuming no changes in carbon inputs to soils for perennial crops). RothC is a model that estimates
378
soil organic carbon turnover in both croplands and grasslands according to SCI, soil covering,
379
climate and soil properties. RothC considers four active soil organic carbon compartments: the
380
resistant plant pool (RPM), the decomposable plant pool (DPM), the microbial pool (BIO) and the
381
humic pool (HUM). An additional inert organic matter (IOM) pool is considered but the latter is
382
supposed to be constant over time in RothC; it is thus assumed unchanged in the organic scenarios
383
vs. in the baseline, and is not included in the equations below. RothC estimates the carbon flows
384
among the four active compartments as well as the amount of carbon mineralised from each
385
Page 11
11
compartment, with a monthly time step and through first order kinetic equations. In this study, we
386
used the continuous formulation of RothC
21
summarized in equation (1).
387
388
(1):
𝑆𝑂𝐶′(𝑡) = ⍴(𝑡) ∗ 𝐴 ∗ 𝑆𝑂𝐶(𝑡) + 𝐵(𝑡)
389
390
Where
SOC’(t)
represent the derivative of
SOC
with respect of time,
SOC(t)
represent the SOC
391
stocks at time
t
.
A
is a 4x4 matrix representing the mineralisation and carbon flows among the four
392
active soil organic carbon pools.
⍴
(t)
is the decomposition rate modifier and depends on the
393
climatic, edaphic and soil covering conditions. Note that soil covering affects SOC dynamics by
394
reducing its mineralisation rate in RothC. We assumed similar rates of soil organic carbon
395
stabilisation and mineralisation in both the organic scenarios and the baseline – a rather
396
conservative estimate due to lack of consistent data, despite preliminary evidence of more active
397
carbon cycling in organically managed soils
61
. Spatially explicit climatic data were retrieved from
398
the AgMERRA dataset
62
combined with the Penman equation to estimate potential
399
evapotranspiration. Spatially explicit data on soil clay content were retrieved from the harmonized
400
world soil database
63
. Finally, spatially explicit soil covering data for all crops considered where
401
extracted from Sacks et al. 2010
64
.
B(t)
represents the soil carbon inputs at time
t
and was estimated
402
using equation (2):
403
404
(2):
𝐵(𝑡) = [(𝑎
𝑑𝑝𝑚
𝑎
𝑟𝑝𝑚
𝑎
𝑏𝑖𝑜
𝑎
ℎ𝑢𝑚
)
𝑐𝑟𝑜𝑝𝑟𝑒𝑠𝑖𝑑𝑢𝑒𝑠
𝑇
∗ (1 − %𝐹𝑌𝑀) +
405
(𝑎
𝑑𝑝𝑚
𝑎
𝑟𝑝𝑚
𝑎
𝑏𝑖𝑜
𝑎
ℎ𝑢𝑚
)
𝑓𝑎𝑟𝑚𝑦𝑎𝑟𝑑𝑚𝑎𝑛𝑢𝑟𝑒
𝑇
∗ %𝐹𝑌𝑀] ∗ 𝑏
𝑡
406
407
Where
a
dpm
,
a
rpm
,
a
bio
and
a
hum
are four coefficients that define the proportions of the carbon inputs
408
to soils attached to the four active soil organic carbon pools for both crop residues and farmyard
409
manure. Here,
a
dpm
,
a
rpm
,
a
bio
and
a
hum
were parametrised as follows: (0.6,0.4,0,0) for crop-based
410
residues, (0.4,0.6,0,0) for grass residues and (0.49,0.49,0,0.02) for farmyard manure.
%FYM
411
represents the share of farmyard manure in total soil carbon inputs and
b
t
represents the total soil
412
carbon inputs at time
t
(in t C.ha
-1
).
413
414
415
Soil carbon input (SCI) estimation
416
417
For both the organic scenarios and the baseline, we estimated the annual SCI using equation (3):
418
419
(3):
𝑆𝐶𝐼 = 𝐴𝑔𝐶 ∗ %𝑅𝑒𝑐𝑦𝑐𝑙𝑒𝑑 + 𝐵𝑔𝐶 + 𝐹𝑌𝑀
𝑎𝑝𝑝𝑙𝑖𝑒𝑑
420
421
Where SCI represents the inputs of organic carbon to either cropland or grassland soils (in t C.ha
-
422
1
.yr
-1
).
AgC
and
BgC
(in t C.ha
-1
.yr
-1
) are respectively the above and belowground plant carbon
423
biomass (the latter being estimated over the 0-30 cm soil depth).
%Recycled
(in %) represents the
424
percentage of the
AgC
that remains on field. In croplands the
%Recycled
data were extracted from
425
the GOANIM model
14
. In grasslands,
%Recycled
represents the non-grazed carbon share of the
426
entire grassland biomass production. Finally, FYM
applied
(in t C.ha
-1
) is the carbon from farmyard
427
manure applied to the cropland or grassland soils. We assumed that biomass quality and its related
428
carbon stabilisation and mineralisation properties were similar in both the organic scenarios and the
429
baseline due to inconsistent data in the literature
65
. We estimated AgC and BgC using equation (4)
430
and (5):
431
432
(4):
𝐴𝑔𝐶 = 𝑌𝑖𝑒𝑙𝑑 ∗ 0.5 𝐻𝐼
⁄
433
(5):
𝐵𝑔𝐶 = 𝐴𝑔𝐶 ∗ 𝑅𝑆
434
435
Page 12
12
Where
HI
and
RS
represent the crop-specific harvest index (unit-less) and the root-shoot ratio (unit-
436
less), respectively, for each of the considered 45 crop species. Both
HI
and
RS
values were retrieved
437
from Monfreda et al. 2008
66
and Smil et al. 1999
67
.
Yield
refers to the crop yields (in tons DM.ha
-1
)
438
as retrieved from Monfreda et al. 2008
66
(for the baseline) or from the GOANIM model (for the
439
organic scenarios)
10
. To convert the estimated dry matter production in C, we used a 0.5 coefficient
440
value (in t C.t DM
-1
).
441
442
FYM
applied
was estimated using equation (6) and (7)
443
444
(6):
𝐹𝑌𝑀
𝑎𝑝𝑝𝑙𝑖𝑒𝑑
=
𝐶
𝑒𝑥
∗(1−𝛽)
𝐻𝐴
445
(7):
𝐶
𝑒𝑥
= ∑ 𝑉𝑆
𝑎
∗ 𝑃𝑜𝑝
𝑎
𝑎
446
447
Where
C
ex
(in tC.yr
-1
) is the total amount of carbon excreted by the livestock population as
448
farmyard manure and
HA
is the total harvested area (ha).
β
represents the share of
C
ex
that is not
449
applied to the agricultural lands. In croplands,
β
represents the share of
C
ex
that is left on pasture
450
during animal grazing, used for non-agricultural purposes (e.g., as fuel) and is lost during the
451
manure management process. In grasslands,
β
the share of
C
ex
that is not left on pasture during
452
animal grazing.
β
was estimated following the 2019 IPCC guidelines refinement
68
. The amount of
453
carbon lost in the manure management process was estimated according to Bareha et al. 2021
69
. In
454
equation (7),
Pop
a
is the livestock population (in heads) for each of the nine considered animal
455
species
a
.
VS
(in tC.head
-1
.yr
-1
) is the amount of volatile solid carbon excreted per animal and per
456
year and was estimated using equation 10.24 of the 2019 refinement of IPCC guidelines represented
457
in equation (8).
458
459
Equation 8:
𝑉𝑆 = [𝐺𝑟𝐸 ∗ (1 −
𝐷𝐸
100
) + (𝑈𝐸 ∗ 𝐺𝑟𝐸)] ∗ [(
1−𝐴𝑆𝐻
18.45
)]
460
461
Where,
GrE
is the gross energy intake (MJ.day
-1
),
DE
is the feed digestibility (%),
UE
is the urinary
462
energy (% of
GrE
) and ASH is the ash content of the feed (% of DM).
UE
had a value of 0.02 for
463
pigs and 0.04 for all other animals. In the organic scenario, the estimation of
GrE
,
DE
and
ASH
464
where made using the feed nutritional composition from feedipedia (feedipedia.org). In the
465
baseline, we used data from Herrero et al. 2013
70
to estimate
DE
and
ASH
and used equation (9)
71
to
466
estimate
GrE
.
467
468
Equation 9:
𝐺𝑟𝐸 = 𝐶𝑃 ∗ 0.056 + 𝐹𝑎𝑡 ∗ 0.096 + (100 − 𝐶𝑃 − 𝐹𝑎𝑡 − 𝐴𝑆𝐻) ∗ 0.042
469
470
Where,
CP
is the crude protein content of the ration (%),
Fat
is the fat content of the ration (%) and
471
ASH
is the mean ash content of the ration (%).
CP
,
Fat
and
Ash
were retrieved from Herrero et al.
472
2013
70
.
473
We made sure that the
VS
excretion would remain in a range of 10 to 50% of the total C ingested by
474
livestock animals
72
. This helped to close the carbon cycle within both the organic scenarios and the
475
baseline, thereby avoiding any overestimation of soil carbon inputs.
476
477
478
SCI for the optimal organic scenario
479
480
We designed the optimal organic scenario to estimate the benefits brought by a more carbon-
481
oriented farming and to capture the potential effect of additional cropping practices on SOC stocks.
482
Based on a preliminary sensitivity analysis of SCI and SOC stocks to various cropping parameters
483
(see
Supplementary table 3
), we built the optimal organic scenario on the assumption that the
484
Page 13
13
fraction of crop residues recycled on croplands (
%Recycled
) and
RS
would be increased. More
485
precisely, we used equation (3) using modified
%Recycled
, AgC and BgC (hereafter called AgC
opt
486
and BgC
opt
) values, with
%Recycled
being increased by 10% and
AgC
opt
and
BgC
opt
being estimated
487
using equations (10), (11) and (12).
488
489
(10):
𝑇𝑜𝑡𝑎𝑙 = 𝑌𝑖𝑒𝑙𝑑 ∗ 0.5 ∗ (1 + 𝑅𝑆) 𝐻𝐼
⁄
490
(11):
𝐴𝑔𝐶
𝑜𝑝𝑡
=
𝑇𝑜𝑡𝑎𝑙
(1+𝑅𝑆′)
491
(12):
𝐵𝑔𝐶
𝑜𝑝𝑡
= 𝑇𝑜𝑡𝑎𝑙 − 𝐴𝑔𝐶
𝑜𝑝𝑡
492
493
Where
Total
is the total carbon biomass produced.
AgC
opt
and
BgC
opt
are the total carbon in the
494
above-ground and below-ground biomass in the optimal organic scenarios, respectively. Evidences
495
show that
RS
is up to twice higher for crops in conditions of low N availability compared to
496
conditions of high N availability
73
. We estimated a modified
RS’
root-shoot ratio for situations of N
497
availability in the optimal organic croplands using equation (13):
498
499
(13):
{
𝑖𝑓𝑌𝑖𝑒𝑙𝑑 < 𝑌𝑖𝑒𝑙𝑑
𝑚𝑎𝑥
𝑡ℎ𝑒𝑛𝑅𝑆
′
= (2 −
𝑌𝑖𝑒𝑙𝑑
𝑌𝑖𝑒𝑙𝑑
𝑚𝑎𝑥
) ∗ 𝑅𝑆
𝑖𝑓𝑌𝑖𝑒𝑙𝑑 = 𝑌𝑖𝑒𝑙𝑑
𝑚𝑎𝑥
𝑡ℎ𝑒𝑛𝑅𝑆
′
= 𝑅𝑆
500
501
Where Yield
max
is the crop specific maximum attainable yield for organic farming (in tons C.ha
-1
) as
502
defined in the GOANIM model
14
.
503
504
In addition, we also simulated extensive use of cover-crops in the optimal organic scenario based on
505
the observed higher share of cover-crops in organic crop rotations compared to conventional ones
10
.
506
The use of cover crops is limited by agronomic and pedo-climatic conditions. Based on a previous
507
meta-analysis on the extent of cover-crops, we considered that cover cropping could be potentially
508
applied on 50% of global croplands
40
where bare-soil periods exist between main cash crops. We
509
estimated the additional SCI from cover crops using equation (14). Meanwhile, we assumed that
510
there were no cover crops in the baseline.
511
512
(14):
𝑆𝐶𝐼
𝑐𝑐,𝑖,𝑚𝑜𝑛𝑡ℎ
=
1.87
𝐺𝑀𝐵𝑆𝑃
∗𝑌𝑖𝑒𝑙𝑑
𝑝𝑙𝑎𝑛𝑡,𝑖
𝑌𝑖𝑒𝑙𝑑
𝑝𝑙𝑎𝑛𝑡,𝑤𝑜𝑟𝑙𝑑
513
514
Where
SCI
cc,i,month
(in t C.ha
-1
.
.
month
-1
) is the soil carbon input from cover crops in country
i
per
515
month of cover cropping. The 1.87 value (in t C.ha
-1
.
yr
-1
) is the global annual mean of soil carbon
516
input from cover crops estimated by Poeplau et al. 2015
40
. We divided this 1.87 value by the
517
estimated global mean duration of the bare soil period in the baseline (
GMBSP
, expressed in
518
month). To account for the variability of cover cropping productivity among countries – that is
519
driven by climatic and farming factors
74
– we multiplied this global mean cover-cropping biomass
520
production by the ratio of the country specific mean yield (
Yield
plant,i
) to the global mean yield
521
(
Yield
plant,world
) for the most productive crop species between wheat and maize in the country.
522
Finally, for each of the considered grid-cells, this monthly
SCI
cc,i,month
was multiplied by the average
523
bare-soil period (in months) between main cash crops, based on sowing and harvesting dates
524
retrieved from Sacks et al. 2010
64
.
525
526
Note that sharp differences in SCI for this optimal scenario may appear among countries in Figure
527
1, such as between Spain and France. Those differences are likely due to differences in climate.
528
Because crop productivity is significantly lower in Spain compared to France due to its more arid
529
conditions, even small additional carbon inputs to soils from cover crops are likely to raise the SCI
530
ratio above 1 in Spain. On contrast, because of higher crop productivity in France, much higher
531
Page 14
14
carbon provisioning is needed from cover-crops to raise the SCI ratio above 1 in that country. The
532
same holds true for several Sub-Saharan African countries. Another explanations lie in the data and
533
model parametrisation we used in our simulations. Several parameters – such as the biomass
534
productivity of cover crops – were in fact defined by country or climatic region. These effects are in
535
fact quite common in global databases, and they are in most cases an artefact from the interpolation
536
of climate data.
537
538
539
RothC parametrisation
540
541
We used RothC assuming carbon pools to be at steady state in the baseline. This necessary
542
assumption translates into a steady state assumption for climatic conditions and soil carbon inputs
543
over the years for both the organic farming scenarios and the baseline. Although partly unrealistic,
544
this assumption is consistent with the thought experiment of large organic farming expansion that
545
we report in this study. To remain in line with this steady state assumption in the baseline, we first
546
estimated the SCI that are required to keep baseline SOC stocks at their current level (
SCI
0
) by
547
using the method developed by Martin et al. 2007
21
and summarized in equation (15).
548
549
(15):
𝑆𝐶𝐼
0
= (𝐼
4
− 𝐹) ∗ 𝑆𝑂𝐶
*
550
551
Where
SCI
0
is the carbon inputs (in t C.ha
-1
.yr
-1
) required to maintain SOC stocks at their current
552
level.
F
is a 4x4 matrix representing the mineralisation and carbon flows among the four active soil
553
organic carbon pools.
F
values depend on the climatic, edaphic and soil covering conditions.
SOC
*
554
is the current active (i.e. not comprised in the IOM pool) SOC stocks that is assumed to be at the
555
equilibrium (in either croplands or grasslands). Total SOC stocks were retrieved from the AEZEF
556
dataset
75
that provides estimates of soil organic carbon stocks for croplands on the first 30 cm of
557
topsoils per country and for 18 agroecological zones.
SOC*
was estimated after subtracting the
558
IOM content which was estimated using the Falloon’s et al. (1998) equation
47
.
559
560
To estimate the SCI in the organic farming scenarios (SCI
1
), we corrected SCI
0
by the ratio of
561
SCI
org
to SCI
baseline
(
RCI
) as detailed in equation (16).
562
563
(16):
𝑆𝐶𝐼
1
= 𝑆𝐶𝐼
0
∗ 𝑅𝐶𝐼 =
𝑆𝐶𝐼
0
∗𝑆𝐶𝐼
𝑜𝑟𝑔
𝑆𝐶𝐼
𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒
564
565
Where
SCI
org
and
SCI
baseline
are the soil carbon inputs for the organic farming scenarios and the
566
baseline, respectively, estimated using the methods presented in the previous sections. We used
SCI
1
567
as input in the RothC model to estimate the changes in SOC stocks in the organic farming scenarios
568
– 20, 50 and 100 years after a global conversion to this farming system – using equation (1). We
569
assumed constant climate data over the simulation periods. This assumption is disputable given
570
current and future climate change, but it remains consistent with our thought experiment that
571
consists in exploring situations of drastic expansion of organic farming. Further studies that are
572
beyond the scope of this article would be needed to account for future climate scenarios. The
573
estimated SCI
1
is expressed in tC.ha
-1
.yr
-1
, though RothC requires monthly data. We assumed that
574
the annual soil carbon inputs were equally distributed between the twelve months of the year.
575
576
In order to account for the observed differences in crop rotations between organic and conventional
577
farming
10
, we ran RothC in the organic farming scenarios for each of the 45 considered crop species
578
separately, and then, estimated a weighted mean of SOC stocks according to crop species harvested
579
areas, as detailed in equation (17).
580
581
Page 15
15
(17):
𝑆𝑂𝐶
𝑡,𝑚𝑒𝑎𝑛
=
∑ 𝑆𝑂𝐶
𝑡,𝑖
∗𝐻𝐴
𝑖
𝑖
𝐻𝐴
𝑡𝑜𝑡𝑎𝑙
582
583
Where
SOC
t,mean
is the weighted mean of SOC stocks at time
t
and
SOC
t,i
is the SOC stock
584
estimated by the run of RothC for each specific crop
i,
HA
i
represents the harvested area of crop
i
in
585
the organic farming scenarios and
HA
total
is the total harvested area (all crop considered).
HA
i
and
586
HA
total
were retrieved from Barbieri et al. 2019
25
.
587
588
589
Limitations and uncertainties
590
Although the modelling foundations of our work are solid, its global extent requires a large set of
591
input data that may come with some limitations. In particular, both the baseline and the organic
592
scenarios required detailed, spatially explicit distribution of cropland areas, types of crops grown
593
and crop yields. These data were derived from Ref
66
and Earthstat, and were centred circa year
594
2000. Many changes have occurred in agriculture during these last 20 years (including about
595
expanding irrigation and changes in varieties) that may affect our simulations. However, to the best
596
of our knowledge, these databases remain the most appropriate given their global extent, higher
597
number of crop species considered, and data quality and cross-validation. Note that uncertainties
598
and possibly caveats may remain in those databases, e.g. about cropland areas in the island of
599
Guinea or about grassland areas in India, as already mentioned.
600
601
Finally, several of our input data may be affected by some uncertainties. The complexity of the
602
GOANIM and RothC models and limited knowledge about several aspects of input data makes the
603
quantification of these uncertainties very difficult. However, the SOC stocks we estimated were
604
determined over long periods (20, 50 and 100 years). Long term averages show reduced errors on
605
estimated variables due to reduced aggregation effects by the input data – especially the climate
606
data
58
. In addition, this study is based on the comparison of organic farming to a baseline, that are
607
both affected by the same errors and uncertainties. Therefore, concentrating the analysis on the
608
ratios (or differences) of organic to conventional estimation helps to reduce errors and uncertainties.
609
610
611
Data treatment & code availability
612
613
All analyses were made using R x64 3.5.3. GOANIM was used in its most recent version deposited
614
in a public repository
(
https://github.com/Pie90/GOANIM_public
)
. For RothC we used the
615
cin_month
and
runExplicitSol
functions from the RothC package
76
to respectively estimate SCI
0
,
616
and SOC stock evolution across time.
617
618
619
Conflicts of interest
620
621
The authors declare no competing interests.
622
623
Further correspondence
624
625
Any correspondence and requests for materials should be addressed to Ulysse Gaudaré.
626
627
628
Acknowledgments
629
630
Page 16
16
We would like to thank Romain Girault and Younes Behara for their help regarding carbon losses in
631
manure management process. We would also thank Denis Angers, Eric Ceschia and Christopher
632
Poeplau for their inputs on how to consider cover crops. This work was funded by ADEME,
633
Bordeaux Sciences Agro (Univ. Bordeaux), INRAE’s intern committee on organic farming (MP
634
Métabio) and Aberdeen’s university. Matthias Kuhnert and Pete Smith acknowledge support from
635
the CIRCASA project which received funding from the European Union's Horizon 2020 Research
636
and Innovation Programme under grant agreement no 774378.
637
638
639
Authors’ contributions
640
641
U.G., M.K., S.P. and T.N. designed the study; U.G. performed the modelling work, with the help of
642
P.B. for the GOANIM model and M.K, and M.M. for the RothC model. All authors were involved
643
in the interpretation of results and contributed actively to writing and revising the manuscript.
644
645
646
References
647
648
1. Foley, J. A.
et al.
Solutions for a cultivated planet.
Nature
478
, 337–342 (2011).
2. IPCC.
AR6 Full Report
. (2021).
3. Fuss, S.
et al.
Negative emissions—Part 2: Costs, potentials and side effects.
Environ. Res. Lett.
13
, 063002 (2018).
4. Amelung, W.
et al.
Towards a global-scale soil climate mitigation strategy.
Nat. Commun.
11
,
1–10 (2020).
5. Lal, R. Agricultural activities and the global carbon cycle.
Nutr. Cycl. Agroecosystems
70
, 103–
116 (2004).
6. Paustian, K.
et al.
Climate-smart soils.
Nature
532
, 49–57 (2016).
7. Clark, M. & Tilman, D. Comparative analysis of environmental impacts of agricultural
production systems, agricultural input efficiency, and food choice.
Environ. Res. Lett.
12
,
064016 (2017).
8. Gattinger, A.
et al.
Enhanced top soil carbon stocks under organic farming.
Proc. Natl. Acad.
Sci.
109
, 18226–18231 (2012).
9. Seufert, V. & Ramankutty, N. Many shades of gray—the context-dependent performance of
organic agriculture.
Sci. Adv.
3
, e1602638 (2017).
10. Barbieri, P., Pellerin, S. & Nesme, T. Comparing crop rotations between organic and
conventional farming.
Sci. Rep.
7
, 1–10 (2017).
Page 17
17
11. Alvarez, Roberto. Organic farming does not increase soil organic carbon compared to
conventional farming if there is no carbon transfer from other agroecosystems. A meta-analysis.
Soil Res
60
, 211–223 (2021).
12. IFOAM.
Annual Report 2020.pdf
. (2020).
13. Nowak, B., Nesme, T., David, C. & Pellerin, S. Disentangling the drivers of fertilising material
inflows in organic farming.
Nutr. Cycl. Agroecosystems
96
, 79–91 (2013).
14. Barbieri, P.
et al.
Global option space for organic agriculture is delimited by nitrogen
availability.
Nat. Food
2
, 363–372 (2021).
15. Seufert, V., Ramankutty, N. & Foley, J. A. Comparing the yields of organic and conventional
agriculture.
Nature
485
, 229–232 (2012).
16. Ponisio, L. C.
et al.
Diversification practices reduce organic to conventional yield gap.
Proc. R.
Soc. B Biol. Sci.
282
, 1–7 (2015).
17. Muller, A.
et al.
Strategies for feeding the world more sustainably with organic agriculture.
Nat.
Commun.
8
, 1–13 (2017).
18. Erb, K. H.
et al.
Exploring the biophysical option space for feeding the world without
deforestation.
Nat. Commun.
48
, 829–834 (2016).
19. Billen, G.
et al.
Reshaping the European agro-food system and closing its nitrogen cycle: The
potential of combining dietary change, agroecology, and circularity.
One Earth
4
, 839–850
(2021).
20. Coleman, K.
et al.
Simulating trends in soil organic carbon in long term experiments using
RothC-26.3.
Geoderma
81
, 29–44 (1997).
21. Martin, M. P., Cordier, S., Balesdent, J. & Arrouays, D. Periodic solutions for soil carbon
dynamics equilibriums with time-varying forcing variables.
Ecol. Model.
204
, 523–530 (2007).
22. Seufert, V., Ramankutty, N. & Mayerhofer, T. What is this thing called organic? – How organic
farming is codified in regulations.
Food Policy
68
, 10–20 (2017).
23. Puech, C., Baudry, J., Joannon, A., Poggi, S. & Aviron, S. Organic vs. conventional farming
dichotomy: Does it make sense for natural enemies?
Agric. Ecosyst. Environ.
194
, 48–57
(2014).
Page 18
18
24. Kirchmann, H., Kätterer, T., Bergström, L., Börjesson, G. & Bolinder, M. A. Flaws and criteria
for design and evaluation of comparative organic and conventional cropping systems.
Field
Crops Res.
186
, 99–106 (2016).
25. Barbieri, P., Pellerin, S., Seufert, V. & Nesme, T. Changes in crop rotations would impact food
production in an organically farmed world.
Nat. Sustain.
2
, 378–385 (2019).
26. Dignac, M.-F.
et al.
Increasing soil carbon storage: mechanisms, effects of agricultural practices
and proxies. A review.
Agron. Sustain. Dev.
37
, 14 (2017).
27. Roberta_et_al_GCB_EnsembleModellingUncertainty_AAMCombined.pdf.
28. Oelofse, M., Jensen, L. S. & Magid, J. The implications of phasing out conventional nutrient
supply in organic agriculture: Denmark as a case.
Org. Agric.
3
, 41–55 (2013).
29. Leifeld, J.
et al.
Organic farming gives no climate change benefit through soil carbon
sequestration.
Proc. Natl. Acad. Sci.
110
, E984–E984 (2013).
30. IFOAM EU Group. Organic farming, climate change mitigation and beyond. (2016).
31. Nesme, T., Barbieri, P., Gaudaré, U., Pellerin, S. & Angers, D. Sound methods are needed to
assess GHG mitigation potential of organic farming deployment. A comment on.
Agric. Syst.
187
, 102994 (2021).
32. Smith, L. G., Kirk, G. J. D., Jones, P. J. & Pearce, B. D. The greenhouse gas impacts of
converting food production in England and Wales to organic methods.
Nat. Commun.
10
, 1–10
(2019).
33. Nowak, B., Nesme, T., David, C. & Pellerin, S. Nutrient recycling in organic farming is related
to diversity in farm types at the local level.
Agric. Ecosyst. Environ.
10 (2015).
34. Dorin, B. Theory, Practice and Challenges of Agroecology in India.
Int. J. Agric. Sustain.
20
,
153–167 (2022).
35. Feuerbacher, A., Luckmann, J., Boysen, O., Zikeli, S. & Grethe, H. Is Bhutan destined for
100% organic? Assessing the economy-wide effects of a large-scale conversion policy.
PLOS
ONE
13
, e0199025 (2018).
36. Lal, R. Soil carbon sequestration and aggregation by cover cropping.
J. Soil Water Conserv.
70
,
329–339 (2015).
Page 19
19
37. Mazzoncini, M., Sapkota, T. B., Bàrberi, P., Antichi, D. & Risaliti, R. Long-term effect of
tillage, nitrogen fertilization and cover crops on soil organic carbon and total nitrogen content.
Soil Tillage Res.
114
, 165–174 (2011).
38. Baethgen, W. E., Parton, W. J., Rubio, V., Kelly, R. H. & M. Lutz, S. Ecosystem dynamics of
crop–pasture rotations in a fifty‐year field experiment in southern South America: Century
model and field results.
Soil Sci. Soc. Am. J.
85
, 423–437 (2021).
39. Crystal-Ornelas, R., Thapa, R. & Tully, K. L. Soil organic carbon is affected by organic
amendments, conservation tillage, and cover cropping in organic farming systems: A meta-
analysis.
Agric. Ecosyst. Environ.
312
, 107356 (2021).
40. Poeplau, C. & Don, A. Carbon sequestration in agricultural soils via cultivation of cover crops -
A meta-analysis.
Agric. Ecosyst. Environ.
200
, 33–41 (2015).
41. Riggers, C., Poeplau, C., Don, A., Frühauf, C. & Dechow, R. How much carbon input is
required to preserve or increase projected soil organic carbon stocks in German croplands under
climate change?
Plant Soil
460
, 417–433 (2021).
42. Ledo, A.
et al.
Changes in soil organic carbon under perennial crops.
Glob. Change Biol.
26
,
4158–4168 (2020).
43. Morais, T. G.
et al.
Agroecological measures and circular economy strategies to ensure
sufficient nitrogen for sustainable farming.
Glob. Environ. Change
69
, 102313 (2021).
44. Smith
et al.
Greenhouse gas mitigation in agriculture.
Philos. Trans. R. Soc. B Biol. Sci.
363
,
789–813 (2008).
45. Abdalla, M.
et al.
A critical review of the impacts of cover crops on nitrogen leaching, net
greenhouse gas balance and crop productivity.
Glob. Change Biol.
25
, 2530–2543 (2019).
46. Ryals, R., Hartman, M. D., Parton, W. J., DeLonge, M. S. & Silver, W. L. Long-term climate
change mitigation potential with organic matter management on grasslands.
Ecol. Appl.
25
,
531–545 (2015).
47. Falloon, P., Smith, P., Coleman, K. & Marshall, S. How important is inert organic matter for
predictive soil carbon modelling using the Rothamsted carbon model?
Soil Biol. Biochem.
32
,
433–436 (2000).
Page 20
20
48. Martin, M. P.
et al.
Feasibility of the 4 per 1000 aspirational target for soil carbon: A case study
for France.
Glob. Change Biol.
27
, 2458–2477 (2021).
49. Meersmans, J.
et al.
Spatial analysis of soil organic carbon evolution in Belgian croplands and
grasslands, 1960-2006.
Glob. Change Biol.
17
, 466–479 (2011).
50. Chen, L., Smith, P. & Yang, Y. How has soil carbon stock changed over recent decades?
Glob.
Change Biol.
21
, 3197–3199 (2015).
51. Cai, A.
et al.
Carbon and Nitrogen Mineralization in Relation to Soil Particle-Size Fractions
after 32 Years of Chemical and Manure Application in a Continuous Maize Cropping System.
PLOS ONE
11
, e0152521 (2016).
52. Bouwman, A. F., Van der Hoek, K. W., Eickhout, B. & Soenario, I. Exploring changes in world
ruminant production systems.
Agric. Syst.
84
, 121–153 (2005).
53. Chang, J.
et al.
Combining livestock production information in a process-based vegetation
model to reconstruct the history of grassland management.
Biogeosciences
13
, 3757–3776
(2016).
54. Fetzel, T.
et al.
Quantification of uncertainties in global grazing systems assessment.
Glob.
Biogeochem. Cycles
31
, 1089–1102 (2017).
55. Petz, K.
et al.
Mapping and modelling trade-offs and synergies between grazing intensity and
ecosystem services in rangelands using global-scale datasets and models.
Glob. Environ.
Change
29
, 223–234 (2014).
56. Osborne, T., Rose, G. & Wheeler, T. Variation in the global-scale impacts of climate change on
crop productivity due to climate model uncertainty and adaptation.
Agric. For. Meteorol.
170
,
183–194 (2013).
57. Rojas-Downing, M. M., Nejadhashemi, A. P., Harrigan, T. & Woznicki, S. A. Climate change
and livestock: Impacts, adaptation, and mitigation.
Clim. Risk Manag.
16
, 145–163 (2017).
58. Kuhnert, M.
et al.
Impact analysis of climate data aggregation at different spatial scales on
simulated net primary productivity for croplands.
Eur. J. Agron.
88
, 41–52 (2017).
59. Tuomisto, H. L., Hodge, I. D., Riordan, P. & Macdonald, D. W. Does organic farming reduce
environmental impacts? - A meta-analysis of European research.
J. Environ. Manage.
112
, 309–
320 (2012).
Page 21
21
60. Gaudaré, U.
et al.
Comparing productivity and feed-use efficiency between organic and
conventional livestock animals.
Environ. Res. Lett.
16
, 024012 (2021).
61. Lori, M., Symnaczik, S., Mäder, P., De Deyn, G. & Gattinger, A. Organic farming enhances soil
microbial abundance and activity—A meta-analysis and meta-Regression.
PLoS ONE
12
,
e0180442 (2017).
62. Ruane, A. C., Goldberg, R. & Chryssanthacopoulos, J. Climate forcing datasets for agricultural
modeling: Merged products for gap-filling and historical climate series estimation.
Agric. For.
Meteorol.
200
, 233–248 (2015).
63. Nachtergaele, F., van Velthuizen, H. & Verelst, L.
Harmonized World Soil Database - Version
1.1
. 43 (2009).
64. Sacks, W. J., Deryng, D., Foley, J. A. & Ramankutty, N. Crop planting dates: an analysis of
global patterns: Global crop planting dates.
Glob. Ecol. Biogeogr.
19
, 607–620 (2010).
65. García-Palacios, P.
et al.
Crop traits drive soil carbon sequestration under organic farming.
J.
Appl. Ecol.
55
, 2496–2505 (2018).
66. Monfreda, C., Ramankutty, N. & Foley, J. A. Farming the planet: 2. Geographic distribution of
crop areas, yields, physiological types, and net primary production in the year 2000.
Glob.
Biogeochem. Cycles
22
, 1–19 (2008).
67. Smil, V. Nitrogen in crop production: An account of global flows.
Glob. Biogeochem. Cycles
13
, 647–662 (1999).
68. IPCC.
CHAPTER 10 EMISSIONS FROM LIVESTOCK AND MANURE MANAGEMENT
.
(2019).
69. Bareha, Y., Affes, R., Moinard, V., Buffet, J. & Girault, R. A simple mass balance tool to predict
carbon and nitrogen fluxes in anaerobic digestion systems.
Waste Manag.
135
, 47–59 (2021).
70. Herrero, M.
et al.
Biomass use, production, feed efficiencies, and greenhouse gas emissions
from global livestock systems.
Proc. Natl. Acad. Sci. U. S. A.
110
, 20888–20893 (2013).
71. Weiss, W. P. & Tebbe, A. W. Estimating digestible energy values of feeds and diets and
integrating those values into net energy systems.
Transl. Anim. Sci.
3
, 953–961 (2019).
Page 22
22
72. Soussana, J.-F. & Lemaire, G. Coupling carbon and nitrogen cycles for environmentally
sustainable intensification of grasslands and crop-livestock systems.
Agric. Ecosyst. Environ.
190
, 9–17 (2014).
73. Marschner, H. & Marschner, P.
Marschner’s mineral nutrition of higher plants
.
(Elsevier/Academic Press, 2012).
74. Justes, E.
Réduire les fuites de nitrate au moyen de cultures intermédiaires
. (INRAE, 2012).
75. Gibbs, H., Yui, S. & Plevin, R. New Estimates of Soil and Biomass Carbon Stocks for Global
Economic Models.
GTAP Tech. Pap.
33
, 1–43 (2014).
76. Martin, M. P. RothC, a R package, INRAE (https://forgemia.inra.fr/manuel.martin/rothc).
(2018).
649