Con inen al-scale e idence o a m
1
managemen impac s on soil ca bon
2
3
Julian Hel ens ein1, Nick an Dijk1, Anna Edlinge 2, Gab iel Y.K. Moine 3, Sophie an Rijssel1,
4
Alexand e M.J.-C. Wadoux4, Rachel C eame 3, Ca men Vazquez3, Ve a L. Mulde 1
5
1 Soil Geog aphy and Landscape G oup, Wageningen Uni e si y & Resea ch, Wageningen, he
6
Ne he lands
7
2 Wageningen En i onmen al Resea ch, Wageningen Uni e si y & Resea ch, Wageningen, he
8
Ne he lands
9
3 Soil Biology G oup, Wageningen Uni e si y & Resea ch, Wageningen, he Ne he lands
10
4 College o Science and Enginee ing, James Cook Uni e si y, Cai ns, Aus alia
11
12
Keywo ds: ca bon a ming, clima e-sma a ming, ag icul u al managemen , soil ca bon
13
dynamics, soil o ganic ca bon s ocks
14
15
16
ABSTRACT
17
The e a e high expec a ions ha ag icul u al p ac ices can mi iga e clima e change and imp o e
18
soil heal h by inc easing soil o ganic ca bon (SOC). Howe e , exis ing la ge scale SOC moni o ing
19
ea s ag icul u al managemen as a black box, meaning ha obse ed pa e ns and ends canno
20
in o m on sus ainable p ac ices. He e, we o he i s ime combine managemen da a om
21
sys ema ic a m su eys (n =81,688 a ms) and ep esen a i e soil moni o ing da a (n = 8,834
22
loca ions) o quan i y he impac o ag icul u al p ac ices on h ee SOC me ics ac oss Eu ope:
23
s ocks, s ocks ela i e o pedoclima ic benchma ks, and yea ly change in SOC concen a ion. Ou
24
indings show ha managemen in ensi y is a signi ican con ibu o o SOC loss ac oss Eu ope,
25
wi h a ying impac by soil and clima e egion. In u n, se e al p ac ices (e.g. high sha e o
26
manu e, o ganic managemen , and a high p opo ion o leys in c op o a ion) demons a ed
27
po en ial o inc easing SOC in di e en ag icul u al sys ems. This esea ch opens new a enues
28
o unde s anding le e age poin s o enhance soil heal h, and p o ides unp eceden ed e idence o
29
in o ming policies p omo ing sus ainable a ming p ac ices.
30
31
MAIN
32
Main aining soil o ganic ca bon (SOC) s ocks in ag icul u al sys ems is c i ical o sus aining soil
33
heal h, and a oiding CO₂ emissions om deg ada ion1–4. P o ec ing exis ing SOC helps p ese e
34
nu ien cycling, soil s uc u e, biodi e si y, and wa e egula ion, all o which unde pin p oduc i e
35
and sus ainable a ming sys ems5,6. Inc easing SOC s ocks, de ined as soil ca bon seques a ion7,
36
can also emo e CO₂ om he a mosphe e, bu global seques a ion po en ial is modes and ime-
37
limi ed due o s eady-s a es, e e sibili y, and socio-economic cons ain s8. E en op imis ic
38
scena ios sugges i could o se only a small sha e o he emission educ ions needed o clima e
39
a ge s4. None heless, op imal SOC managemen is a cen al componen o sus ainable soil
40
managemen and si s on he igh side o he clima e equa ion. A signi ican amoun o esea ch
41
has shown ha he adop ion o sus ainable managemen p ac ices such as di e se c op o a ions,
42
o ganic amendmen s, o pe ennial c opping - among o he s - can slow o e e se declines in SOC
43
s ock1,9.
44
Howe e , cu en unde s anding o managemen e ec s on SOC s ocks mos ly s ems om
45
con olled ield expe imen s9–11, which e eal ha bo h managemen e ec s on SOC and he
46
in luence o SOC on o he ecosys em unc ions a e highly con ex dependen 8,12. Me a-analyses o
47
dozens o con olled expe imen s allow o isola e he in luence o pedoclima ic a iabili y, and
48
ha e shown ha biocha applica ions, o ganic e ilize s, pe ennial c ops, and ag o o es y ha e
49
posi i e e ec s on SOC in a la ge majo i y o cases, while illage in ensi y, c op o a ion, and
50
mine al e ilize applica ion ha e smalle o mo e ambi alen e ec s1,13,14. While hese p o ide
51
impo an scien i ic insigh s, me a-analyses o con olled expe imen s do no cap u e he complex
52
in e ac ions and a iabili y o a ming p ac ices as obse ed in eal a ms ac oss di e en pedo-
53
clima ic condi ions12. This because managemen p ac ices on con olled expe imen s di e om
54
on- a m eali y, as single p ac ices a e emphasized o es ing, igno ing ha hese p ac ices o en
55
come in bundles15. A ecen syn hesis o managemen p ac ices ( illage, c op o a ion and
56
e iliza ion) iden i ied mo e han 285,000 possible managemen bundles16, he applica ion o hese
57
bundles is hen u he mul iplied by he di e ing pedo-clima ic condi ions when conside ing he
58
ou comes in e ms o SOC s ock and s ock change17.
59
Oppo uni ies o la ge-scale assessmen s o SOC s ocks ha e a isen h ough he inc ease in
60
na ional and con inen al soil moni o ing18–20. Fo example, analysis o changes in SOC s ocks
61
be ween 2009 and 2018 om he Eu opean soil moni o ing ne wo k LUCAS Soil, showed small
62
inc eases in SOC s ocks, especially in g assland soils21. Howe e , in his and o he la ge-scale
63
SOC assessmen s22–24, ag icul u al managemen was no conside ed, he e o e while we can
64
moni o he change in SOC s ocks, we do no unde s and he ole o ag icul u al p ac ices. Clea
65
unde s anding o how managemen impac s SOC s ocks, and how ha a ies wi h pedoclima ic
66
con ex , is essen ial o science-in o med policy, e.g. de ining sus ainable managemen p ac ices25.
67
Thus he e is an u gen need o empi ical e idence on he impac o eal-wo ld managemen
68
p ac ices on SOC s ocks o in o m policy and implemen a ion e o s26.
69
Assessmen s o managemen impac s on SOC unde eal a m condi ions emain a e, p ima ily
70
due o wo majo challenges: limi ed a ailabili y o a m managemen da a and he high complexi y
71
o co a ying ac o s (pedo-clima ic condi ions, social in luence, economic lock-ins e c.). The i s
72
challenge can be add essed by combining on- a m soil sampling wi h a me in e iews o
73
documen managemen p ac ices27–29. The second challenge can be mi iga ed ei he by limi ing
74
geog aphic scope o cons ain a iabili y, some imes including a pai ed-si e app oach whe e
75
con ols a e adjacen o he s udied p ac ice5,12,30,31, o by ocusing on speci ic c opping sys ems
76
ac oss b oad pedoclima ic g adien s17. Howe e , bo h o hese challenges demand labo in ensi e
77
app oaches which es ic sample size and ep esen a i eness o ea lie s udies.
78
To add ess his knowledge gap, we combined s anda dized Eu opean a m su eys (Fa m
79
Accoun ancy Da a Ne wo k (FADN), n = 81,688) wi h la ge scale soil moni o ing da a o
80
disen angle he e ec o a m managemen p ac ices on SOC ac oss all 27 Eu opean Union
81
membe s a es plus he Uni ed Kingdom (Supplemen a y Figu e 1). We i s used FADN o p edic
82
likely managemen p ac ices a all ag icul u al sampling loca ions o he LUCAS Soil moni o ing
83
p og am (n = 8,834) 32. We hen analyzed ela ionships be ween managemen and h ee SOC
84
me ics: 1) SOC s ocks, 2) de ia ions om benchma ked SOC s ocks, and 3) he change in SOC
85
concen a ion om 2009 o 2018—while aking a iabili y in soil ypology and clima e in o
86
accoun . We hypo hesized ha :
87
1. Pe manen c ops, g asslands, and ex ensi e c ops (e.g., empo a y g asslands, o age c ops)
88
a e associa ed wi h mo e posi i e SOC ou comes han in ensi e c ops (e.g., po a oes, suga
89
bee o ce eals).
90
2. Highe managemen in ensi y nega i ely co ela es wi h SOC ou comes.
91
3. The sha e o manu e in he e ilize egime, c op o a ional di e si y, sha e o leys in he
92
c op o a ion, and o ganic cul i a ion a e posi i ely associa ed wi h SOC ou comes, while
93
illage in ensi y is nega i ely associa ed.
94
4. Pedo-clima ic condi ions signi ican ly in luence he ela ionship be ween managemen and
95
SOC s ocks.
96
Resul s
97
Va iabili y in ag icul u al managemen ac oss Eu ope
98
Analysis o a m managemen da a om 81,688 a ms e ealed ha ag icul u al managemen
99
p ac ices exhibi ed signi ican a iabili y ac oss Eu ope, bo h be ween and wi hin egions. Fo
100
e ilize egimes, o al N and K inpu s showed a simila spa ial pa e n, wi h highes a e age
101
applica ion a es in NW Eu ope, while o al P inpu s we e highes in no he n I aly, NW and NE
102
Eu ope (Supplemen a y Fig 2a-c). The sha e o N, P, and K om manu e ollowed a simila spa ial
103
pa e n o all h ee nu ien s, wi h highes a e age alues in alpine egions and coas al a eas o
104
no he n Spain (Supplemen a y Fig 2d- ). Howe e , he p opo ion o K de i ed om manu e was
105
consis en ly highe han ha o N o P. O ganic a ming was mos p e alen in ma ginal egions,
106
i.e. wi h clima ic cons ain s such as in he e y no h and sou h o in moun ainous egions
107
(Supplemen a y Fig. 2g). C op o a ional di e si y was highes in eas e n Ge many and Czech
108
Republic (Supplemen a y Fig. 2h), a eas known o la ge a m sizes. The sha e o ley and odde
109
c ops in he c op o a ion was highes in Sweden, SE F ance, Po ugal, and I aly (Supplemen a y
110
Fig. 2i). Tillage in ensi y was gene ally high, indica ing he p e alence o con en ional illage,
111
wi h lowes alues in eas e n hal o Ge many, pa s o he UK and F ance, and sou he n Po ugal
112
(Supplemen a y Fig. 2j).
113
While spa ial pa e ns in a m managemen ac oss Eu ope ha e been desc ibed p e iously33–36, ou
114
s udy is no el in ha he use o indi idual a m da a enabled us o desc ibe c op- and al i ude-
115
speci ic managemen p ac ices o each egion. Andalusia, he sou he nmos egion o Spain,
116
p o ides an example o illus a e c op- and al i ude‑speci ic di e ences in managemen p ac ices
117
iden i ied om he a m su eys. While he Andalusian a e age N inpu le el is 73.5 kg N ha-1 y-
118
1, ou app oach unco e ed ha epo ed applica ion a es ange om an a e age o 0 o nu s o
119
220 kg N ha-1 o s awbe ies, wi h also di e ences wi hin a c op depending on al i ude
120
(Supplemen a y Fig. 3). Simila in a- egional a iabili y was obse ed o all o he managemen
121
a iables, wi h o ganic a ming p o iding ano he s iking example (Supplemen a y Fig. 3). In
122
Andalusia, only 6% o a ms g owing whea below 300 m al i ude (n = 191) we e o ganic, whe eas
123
67% o hose abo e 600 m (n = 89) p ac iced o ganic a ming. These c op- egion-, and al i ude-
124
speci ic managemen p ac ices we e used o p edic he mos likely managemen a each LUCAS
125
soil sampling loca ion.
126
O he 257 egions analyzed, ou app oach iden i ied signi ican di e ences in ag icul u al
127
managemen in mos egions, wi h 93% o egions showing signi ican c op and al i ude speci ic
128
di e ences in N inpu , 92% in P inpu , and 95% in K inpu . Signi ican di e ences we e u he
129
obse ed in 98%, 97% and 97%, espec i ely o egions o he sha e o N, P, and K de i ed om
130
manu e; 88% o o ganic a ming, 97% o c op o a ional di e si y, 95% o he sha e o ley, 80%
131
o he sha e o o age, and 89% o illage in ensi y.
132
133
Rela ionship be ween land co e and soil o ganic ca bon
134
SOC s ocks (χ² = 1608, p < 10-15), benchma ked SOC s ocks (χ² = 184, p < 10-15), and he change
135
in SOC concen a ion be ween 2009 and 2018 (χ² = 143, p < 10-15) we e all signi ican ly dependen
136
on land co e . SOC s ocks we e highes in g asslands, ollowed by a able c ops wi h high soil
137
co e such as empo a y g asslands and odde c ops (Fig. 1a). SOC s ocks we e calcula ed using
138
a mass-co ec ed ixed dep h app oach (see me hod) aking in o accoun a ia ions in concen a ion
139
o SOC and bulk densi y. While he concen a ion o SOC was conside ably highe in g asslands
140
han any a able c ops, hey also p esen ed lowe bulk densi y han a able c ops (Supplemen a y
141
Fig. 4). Gene ally, soils unde a able cul i a ion, including empo a y g asslands, had signi ican ly
142
highe bulk densi y han g assland o pe manen c ops (Supplemen a y Fig. 4). Some pe manen
143
c ops showed low obse ed SOC s ocks (Fig. 1a and b), as a combina ion o low bulk densi y and
144
a e age o mode a e SOC concen a ion (Supplemen a y Fig. 4). To accoun o SOC accumula ion
145
po en ial, obse ed SOC s ocks we e benchma ked by di iding by “ ypical” SOC s ock alues o
146
ha pedo-clima ic uni 37. These benchma ked SOC s ocks we e highes in pe manen c ops, se
147
aside land and low in ensi y a able c ops such as empo a y g asslands and luce ne (Fig. 1b).
148
Lowes benchma ked SOC s ocks we e ound in in ensi e a able c ops such as po a o and suga
149
bee , which a e associa ed wi h in ensi e soil dis u bance and minimal c op esidue e u n.
150
A majo i y o si es (55%) showed an inc ease in SOC concen a ion om 2009 o 2018, con i ming
151
De Rosa e al. 21. Addi ionally, we ound signi ican di e ences be ween indi idual c ops, wi h
152
mos pe manen c ops and g assland ypes ending o ha e lowe isk o SOC loss (Fig. 1c). The
153
c ops mos likely o ha e SOC losses we e po a o, ye, oa s, and i icale.
154
155
156
Figu e 1. Rela ionship be ween soil o ganic C (SOC) me ics and land co e . SOC s ocks (a),
157
χ² = 1545, p < 10-15), he benchma ked SOC s ocks (b), χ² = 163, p < 10-15), and he change in
158
soil o ganic C be ween 2009 and 2018 (c), χ² = 143, p < 10-15) we e all signi ican ly dependen
159
on land co e . SOC and land co e da a om LUCAS Soil. Only land co e s wi h n > 10 a e
160
shown. To al n = 6795, 6795, and 6362 espec i ely.
161
Managemen e ec on soil o ganic ca bon
162
Mixed linea models wi h pedoclima ic zone as a andom e ec showed ha managemen in ensi y
163
signi ican ly a ec ed all h ee SOC me ics (Supplemen a y Table 1). Managemen in ensi y was
164
calcula ed om e ilize inpu , sha e o manu e, p obabili y o o ganic cul i a ion, illage in ensi y
165
(a able only), and sha e o leys and o age in he c op o a ion (a able only). SOC me ics in a able
166
si es showed he s onges esponse o managemen in ensi y (Fig. 2), wi h a e y clea nega i e
167
e ec o in ensi y on benchma ked SOC. A able si es wi h low managemen in ensi ies showed a
168
benchma ked SOC o 30% abo e he pedoclima ic ypical alues, while he mos in ensi e si es
169
we e 12% below (Fig. 2b). While managemen in ensi y o pe manen c ops had a sligh ly nega i e
170
ma ginal e ec , managemen in ensi y o g asslands was posi i ely co ela ed wi h bo h SOC
171
s ocks (Fig. 2a) and benchma ked SOC s ock (Fig. 2b). The a e o change in SOC showed he
172
same nega i e co ela ion wi h in ensi y index o all h ee land co e classes (Fig. 3c).
173
174
Figu e 2. Rela ionship be ween SOC me ics and managemen in ensi y. Ma ginal e ec s
175
calcula ed om linea mixed models wi h pedoclima ic zones 37 as andom e ec and in ensi y and
176
land co e class as ixed e ec . The model χ² a e 124 (p < 10-15) o SOC s ocks (a, n = 6,776),
177
113 (p < 10-15) o benchma ked SOC s ocks (b, n = 6,776), and 7.3 (p = 0.007) o he yea ly
178
change in SOC concen a ion (c, n = 6,346). See supplemen a y able 1 o u he model de ails.
179
O ange = a able, blue = pe manen c ops, g een = g assland, and pu ple = all si es (no signi ican
180
in e ac ion e ec ). Shaded a eas show he 95% con idence in e als. The black ho izon al line in
181
b) indica es he benchma ked alue.
182
Rega ding he e ec o indi idual managemen p ac ices, N inpu had a nega i e ma ginal e ec
183
on SOC me ics o a able soils, bu a posi i e e ec up o 300 kg N ha-1 o g assland soils (excep
184
o he change in SOC concen a ion which inc eased linea ly wi h inc easing N inpu s) (Fig. 3).
185
This obse ed posi i e ma ginal e ec o N inpu on SOC me ics o g asslands explains he
186
posi i e e ec o managemen in ensi y o e all (Fig. 2). The sha e o manu e in he e ilize
187
egime had a consis en ly posi i e ma ginal e ec on SOC s ocks and benchma ked SOC s ocks
188
(Fig. 3b, g). The o ganic a ming had consis en ly posi i e ma ginal e ec on SOC s ocks and
189
benchma ked SOC s ocks (Fig. 3c, h), bu no signi ican e ec on he change in SOC concen a ion
190
o a able si es (Fig. 3l).
191
The sha e o ley and odde c ops in he c op o a ion had he s onges posi i e ma ginal e ec on
192
obse ed and benchma ked SOC s ocks in a able soils (Fig. 3e, j). A sligh ly nega i e e ec o
193
c op o a ional di e si y on SOC s ocks (Fig. 3d, i) was likely due o o a ions wi h a high sha e o
194
ley and odde c ops,—which end o ha e lowe di e si y. Tillage in ensi y was no ound o ha e
195
a signi ican e ec on any SOC me ic (Supplemen a y able 3).
196
197
Figu e 3. Rela ionship be ween SOC me ics and indi idual managemen a iables. Ma ginal
198
e ec s calcula ed om mixed linea models wi h pedoclima ic zones 37 as andom e ec and
199
managemen a iables as ixed e ec s. Panels a) o e) show ma ginal e ec s on SOC s ocks, panels
200
) o j) on benchma ked SOC s ocks, and panels k) and l) on he change in SOC concen a ion om
201
2009 o 2018. Only managemen a iables ha had a signi ican e ec a e shown, hus some plo s
202
a e oid. O ange = a able, blue = pe manen c ops, g een = g assland, and pu ple = all si es (no
203
in e ac ion e ec ). Shaded a eas show he 95% con idence in e als. The black ho izon al line in
204
b) indica es he benchma ked alue. C op o a ional di e si y and he sha e o ley and odde c ops
205
in he c op o a ion only applies o a able c ops, hus only o ange lines. See supplemen a y able
206
2 and 3 o model de ails.
207
Soil- and clima e-speci ic managemen e ec s
208
To es how managemen e ec s compa ed o di e en pedoclima ic zones, we calcula ed he
209
di e ence in obse ed SOC s ocks be ween he 10% mos op imally managed and 10% leas
210
op imally managed ields pe soil pedoclima ic zone. Op imal managemen was de ined based on
211
esul s om he mixed linea models (see Fig. 4 and me hods o de ails), e.g. si es wi h high alues
212
o sha e o manu e in e ilize mix, o ganic a ming, and sha e o leys in c op o a ion. In a able
213
si es, op imal managemen consis en ly had a posi i e e ec on obse ed SOC s ocks compa ed o
214
subop imal managemen , excep o A lan ic a able sandy soils, which may be due o he low
215
sample size (n = 118) o his soil ype (Fig. 4a). In e es ingly, e en cen al a able sandy soils, o
216
which we obse ed li le managemen a iabili y, had a signi ican managemen e ec , sugges ing
217
ha hese soils a e pa icula ly sensi i e o managemen . The s onges e ec was obse ed on
218
Alpine and bo eal soils, whe e he di e ence be ween 10% mos op imally managed si es and 10%
219
leas op imally managed si es was 41.2 ± 10.2 Mg C ha-1.
220
221
Figu e 4. Pedoclima ic zone speci ic managemen e ec on SOC s ock as a unc ion o managemen
222
a iabili y. A) a able, b) g assland, c) pe manen c op. Managemen e ec is de ined as he
223
di e ence be ween SOC s ock obse ed on he 10% mos op imally managed and 10% leas
224
op imally managed ields pe soil pedoclima ic zone. Op imal managemen is de ined as high
225
p obabili y o o ganic cul i a ion, high sha e o manu e, low o a ional di e si y and high sha e
226
o leys in he o a ion o a able; high p obabili y o o ganic cul i a ion, high sha e o manu e,
227
and in e media e le els o N inpu o g assland and pe manen c ops (see Fig. 4). Managemen
228
a iabili y cap u es how a iable managemen is wi hin he pedoclima ic zone and is calcula ed
229
as he mean no malized s anda d de ia ion pe managemen indica o . E o ba s show he 95%
230
con idence in e al. See37 o de ini ions and dis ibu ions o pedoclima ic zones.
231
Fo g assland si es, ou ou o se en pedoclima ic zones had signi ican managemen e ec s, wi h
232
s onges e ec s obse ed o cold clima e semi-na u al loamy and clayey soils (29.3 ± 4.8 Mg C
233
ha-1), Medi e anean a able loamy and clayey soils (23.2 ± 8.3 Mg C ha-1), and con inen al semi-
234
na u al loamy and clayey soils (23.1 ± 3.6 Mg C ha-1) (Fig. 4b). Fo he o he pedoclima ic zones,
235
he combina ion o p ac ices de ined as op imal managemen o g asslands in his s udy we e no
236
ound o co ela e wi h highe obse ed SOC s ocks. Fo pe manen c op si es, a signi ican
237
managemen e ec o 16.8 ± 7.7 and 15.0 ± 3.3 Mg C ha-1 was obse ed o A lan ic and Cen al
238
a able loamy and clayey soils and Medi e anean loamy and clayey soils, espec i ely (Fig. 5c),
239
wi h no consis en e ec s o he o he pedoclima ic zone.
240
Discussion
241
Managemen impac s on SOC
242
While cu en unde s anding o how di e en ag icul u al p ac ices impac SOC is la gely based
243
on analyses o con olled ield expe imen s1,9,10, ou s udy is he i s o le e age la ge-scale, eal-
244
a m da a o accoun o he complexi y o a ming p ac ices and hei in e ac ions wi h di e se
245
soils and clima es. O he s udies using on- a m da a a smalle spa ial scales ha e also ound
246
nega i e co ela ions be ween managemen in ensi y and SOC in a able soils, e.g. in he
247
Ne he lands5 o Sou he n Sweden30, which we show a he con inen al scale. The nega i e e ec
248
o managemen in ensi y highligh s he combined e ec o a ious managemen p ac ices ha may
249
de e io a e SOC s ocks.
250
A key ad an age o ou app oach is ha bo h he soil and a m managemen da ase s used he e was
251
designed o be ep esen a i e18,32, hus cap u ing he ac ual a iabili y o soils and managemen
252
p ac ices a con inen al scale—an aspec ha has been la gely absen om p e ious s udies5,17. The
253
ou comes o ou esea ch hus p o ide c ucial empi ical e idence o guide e ec i e policy design
254
and p ac ical implemen a ion. Fo example, ou analysis shows ha he posi i e co ela ion o SOC
255
wi h o ganic a ming14, sha e o manu e in he e ilize mix1,10, high c op co e such as h ough
256
empo a y g asslands and o he pe ennial c ops9,13, and no e ec o illage in ensi y1,9 p e iously
257
shown o con olled ield expe imen s also holds o eal- a ms ac oss Eu ope. While con olled
258
expe imen s ha e shown ha c op o a ion posi i ely impac s soil ca bon s o age ela i e o
259
monocul u e1,38,39, ou s udy shows ha , in p ac ice, he composi ion o he c op o a ion (e.g. high
260
sha e o leys o o age c ops) is mo e impo an han di e si y o c ops pe se.
261
E en less is cu en ly known abou managemen impac s on SOC in g asslands and pe manen
262
c ops9. Ou esul s show ha pe manen c ops we e associa ed wi h highe benchma ked SOC
263
s ocks compa ed o a able sys ems and we e mo e likely o ha e posi i e ΔSOC o e ime (Fig.
264
2), suppo ing he po en ial con ibu ion o ee c op sys ems o inc easing SOC s ocks1,40. In
265
e ms o sus ainable managemen o g asslands, he posi i e co ela ion o e ilize amoun wi h
266
SOC (Fig. 3) adds impo an nuance o sus ainabili y ade-o s and con ex -dependency o
267
e ilize use, con i ming simila pa e ns obse ed a he egional scale on- a m su eys in
268
England41.
269
While ΔSOC is po en ially mos in e es ing om a moni o ing pe spec i e, he noise o signal a io
270
o ΔSOC is pa icula ly high, which is a challenge aced by all SOC moni o ing s udies. This
271
7. Don, A. e al. Ca bon seques a ion in soils and clima e change mi iga ion—De ini ions and
499
pi alls. Global Change Biology 30, e16983 (2024).
500
8. Moine , G. Y. K., Hijbeek, R., an Vuu en, D. P. & Gille , K. E. Ca bon o soils, no soils o
501
ca bon. Global Change Biology 29, 2384–2398 (2023).
502
9. Wiesmeie , M. e al. Soil o ganic ca bon s o age as a key unc ion o soils - A e iew o d i e s
503
and indica o s a a ious scales. Geode ma 333, 149–162 (2019).
504
10. López i Losada, R. e al. Syne gis ic e ec s o mul iple “good ag icul u al p ac ices” o
505
p omo ing o ganic ca bon in soils: A sys ema ic e iew o long- e m expe imen s. Ambio
506
(2025) doi:10.1007/s13280-025-02188-8.
507
11. Tang, S. e al. Soil ca bon seques a ion enhanced by long- e m ni ogen and phospho us
508
e iliza ion. Na . Geosci. 1–9 (2025) doi:10.1038/s41561-025-01789-y.
509
12. Schepens, J. A. B. e al. F om global ends o local eali ies: SOC acc ual o imp o ed
510
managemen p ac ices in no hwes e n Eu ope. Geode ma Regional 42, e00983 (2025).
511
13. Poeplau, C. & Don, A. Ca bon seques a ion in ag icul u al soils ia cul i a ion o co e c ops
512
– A me a-analysis. Ag icul u e, Ecosys ems & En i onmen 200, 33–41 (2015).
513
14. Bai, Z. e al. E ec s o ag icul u al managemen p ac ices on soil quali y: A e iew o long-
514
e m expe imen s o Eu ope and China. Ag icul u e, Ecosys ems and En i onmen 265, 1–7
515
(2018).
516
15. Debe na dini, M., Candel, J. & Schul e, R. P. O. F om he g ound up: explo ing Eu opean
517
ca bon a ming h ough social p ac ice heo y. Jou nal o Ru al S udies 120, 103850 (2025).
518
16. Rillig, M. C. & Lehmann, A. Explo ing he ag icul u al pa ame e space o c op yield and
519
sus ainabili y. The New Phy ologis 223, 517–519 (2019).
520
17. Edlinge , A. e al. The impac o ag icul u al managemen on soil agg ega ion and ca bon
521
s o age is egula ed by clima ic h esholds ac oss a 3000 km Eu opean g adien . Global
522
Change Biology 29, 3177–3192 (2023).
523
18. O giazzi, A., Ballabio, C., Panagos, P., Jones, A. & Fe nández-Ugalde, O. LUCAS Soil, he
524
la ges expandable soil da ase o Eu ope: a e iew. Eu opean Jou nal o Soil Science 69,
525
140–153 (2018).
526
19. Pan, G., Xu, X., Smi h, P., Pan, W. & Lal, R. An inc ease in opsoil SOC s ock o China’s
527
c oplands be ween 1985 and 2006 e ealed by soil moni o ing. Ag icul u e, Ecosys ems &
528
En i onmen 136, 133–138 (2010).
529
20. Román Doba co, M. e al. Mapping soil o ganic ca bon ac ions o Aus alia, hei s ocks,
530
and unce ain y. Biogeosciences 20, 1559–1586 (2023).
531
21. De Rosa, D. e al. Soil o ganic ca bon s ocks in Eu opean c oplands and g asslands: How much
532
ha e we los in he pas decade? Global Change Biology 30, e16992 (2024).
533
22. Le y, P. e al. The e ec s o land use on soil ca bon s ocks in he UK. Biogeosciences 21,
534
4301–4315 (2024).
535
23. Wadoux, A. M. J.-C. e al. Baseline high- esolu ion maps o o ganic ca bon con en in
536
Aus alian soils. Sci Da a 10, 181 (2023).
537
24. Xie, Z. e al. Soil o ganic ca bon s ocks in China and changes om 1980s o 2000s. Global
538
Change Biology 13, 1989–2007 (2007).
539
25. Alblas, E. e al. B idging Law and Soil Science o P omo e Soil Heal h. Eu opean Jou nal o
540
Soil Science 76, e70127 (2025).
541
26. Lacos e, M. e al. On-Fa m Expe imen a ion o ans o m global ag icul u e. Na u e Food 3,
542
11–18 (2021).
543
27. Edlinge , A. e al. Compos Applica ion Enhances Soil Heal h and Main ains C op Yield:
544
Insigh s F om 56 Fa me -Managed A able Fields. Jou nal o Sus ainable Ag icul u e and
545
En i onmen 4, e70041 (2025).
546
28. Ga land, G. e al. C op co e is mo e impo an han o a ional di e si y o soil
547
mul i unc ionali y and ce eal yields in Eu opean c opping sys ems. Na Food 2, 28–37 (2021).
548
29. Vah e , T. e al. Landscapes, managemen p ac ices and hei in e ac ions shape soil ungal
549
di e si y in a able ields – E idence om a na ionwide a me s’ ne wo k. Soil Biology and
550
Biochemis y 168, 108652 (2022).
551
30. Williams, H., Colombi, T. & Kelle , T. The in luence o soil managemen on soil heal h: An
552
on- a m s udy in sou he n Sweden. Geode ma 360, 114010 (2020).
553
31. Mudge, P. L. e al. I iga ing g azed pas u e dec eases soil ca bon and ni ogen s ocks. Global
554
Change Biology 23, 945–954 (2017).
555
32. Eu opean Commission. Fa m accoun ancy da a ne wo k. (2022).
556
33. Ballo , R., Guilpa , N. & Jeu oy, M.-H. The i s map o c op sequence ypes in Eu ope o e
557
2012–2018. Ea h Sys em Science Da a 15, 5651–5666 (2023).
558
34. Debonne, N. e al. The geog aphy o mega ends a ec ing Eu opean ag icul u e. Global
559
En i onmen al Change 75, (2022).
560
35. Panagos, P. e al. Imp o ing he phospho us budge o Eu opean ag icul u al soils. Science o
561
The To al En i onmen 853, 158706 (2022).
562
36. Hel ens ein, J. e al. Di e gen ag icul u al de elopmen pa hways ac oss a m and landscape
563
scales in Eu ope: Implica ions o sus ainabili y and a me sa is ac ion. Global En i onmen al
564
Change 86, 102855 (2024).
565
37. Feeney, C. J. e al. Benchma king soil o ganic ca bon (SOC) concen a ion p o ides mo e
566
obus soil heal h assessmen han he SOC/clay a io a Eu opean scale. Science o The To al
567
En i onmen 951, 175642 (2024).
568
38. McDaniel, M. D., Tiemann, L. K. & G andy, A. S. Does ag icul u al c op di e si y enhance
569
soil mic obial biomass and o ganic ma e dynamics? A me a-analysis. Ecological
570
Applica ions 24, 560–570 (2014).
571
39. Zheng, F. e al. Posi i e e ec s o c op o a ion on soil agg ega ion and associa ed o ganic
572
ca bon a e mainly con olled by clima e and ini ial soil ca bon con en : A me a-analysis.
573
Ag icul u e, Ecosys ems & En i onmen 355, 108600 (2023).
574
40. Kay, S. e al. Ag o o es y c ea es ca bon sinks whils enhancing he en i onmen in
575
ag icul u al landscapes in Eu ope. Land Use Policy 83, 581–593 (2019).
576
41. Wa d, S. E. e al. Legacy e ec s o g assland managemen on soil ca bon o dep h. Global
577
Change Biology 22, 2929–2938 (2016).
578
42. an Leeuwen, C. C. E., Mulde , V. L., Ba jes, N. H. & Heu elink, G. B. M. S a is ical
579
modelling o measu emen e o in we chemis y soil da a. Eu opean Jou nal o Soil Science
580
73, e13137 (2022).
581
43. B eu e, T. S. e al. Re isi ing he soil ca bon sa u a ion concep o in o m a isk index in
582
Eu opean ag icul u al soils. Na Commun 16, 2538 (2025).
583
44. The S a e Council o he People’s Republic o China. China’s hi d na ional soil su ey makes
584
signi ican p og ess. (2024).
585
45. Reddy, A. A. The Soil Heal h Ca d Scheme in India: Lessons Lea ned and Challenges o
586
Replica ion in O he De eloping Coun ies. Jou nal o Na u al Resou ces Policy Resea ch 9,
587
124–156 (2019).
588
46. Home | Soils 4 A ica - Ho izon 2020 p og amme o he Eu opean Union.
589
h ps://www.soils4a ica-h2020.eu/ (2025).
590
47. Aus alian Go e nmen - DAFF. Na ional Soil Moni o ing P og am.
591
h ps://www.ag icul u e.go .au/ag icul u e-land/ a m- ood-d ough /na u al-
592
esou ces/soils/na ional-soil-moni o ing-p og am (2025).
593
48. Wadoux, A. M. J.-C. e al. On soil dis ic s. Geode ma 452, 117065 (2024).
594
49. Geo giou, K. e al. Soil Ca bon Sa u a ion: Wha Do We Really Know? Global Change
595
Biology 31, e70197 (2025).
596
50. Eu opean Commission. Glossa y:Fa m ypology. h ps://ec.eu opa.eu/eu os a /s a is ics-
597
explained/index.php? i le=Glossa y:Fa m_ ypology (2025).
598
51. Wes bu y, D. B., Pa k, J. R., Mauchline, A. L., C ane, R. T. & Mo ime , S. R. Assessing he
599
en i onmen al pe o mance o English a able and li es ock holdings using da a om he Fa m
600
Accoun ancy Da a Ne wo k (FADN). Jou nal o En i onmen al Managemen 92, 902–909
601
(2011).
602
52. Ca dillo, C., Di Fonzo, A. & Libe a i, C. The Fa m’s O ien a ion owa ds Sus ainabili y: An
603
Assessmen Using FADN Da a in I aly. Land 12, 301 (2023).
604
53. Dabkiene, V., Balezen is, T. & S eimikiene, D. De elopmen o ag i-en i onmen al oo p in
605
indica o using he FADN da a: T acking de elopmen o sus ainable ag icul u al de elopmen
606
in Eas e n Eu ope. Sus ainable P oduc ion and Consump ion 27, 2121–2133 (2021).
607
54. S a is ics Ne he lands. S anda dised Calcula ion Me hods o Animal Manu e and Nu ien s:
608
S anda d Da a 1990-2008. (2012).
609
55. Holland, J. M. The en i onmen al consequences o adop ing conse a ion illage in Eu ope:
610
e iewing he e idence. Ag icul u e, Ecosys ems & En i onmen 103, 1–25 (2004).
611
56. Lüdecke, D. sjs a s: S a is ical Func ions o Reg ession Models. [objec Objec ]
612
h ps://doi.o g/10.5281/ZENODO.1284472 (2022).
613
57. Hollande , M. & Wol e, D. A. Nonpa ame ic S a is ical Me hods. (John Wiley & Sons, L d,
614
New Yo k, 1973).
615
58. Fowle , A. F., Basso, B., Milla , N. & B in on, W. F. A simple soil mass co ec ion o a mo e
616
accu a e de e mina ion o soil ca bon s ock changes. Sci Rep 13, 2242 (2023).
617
59. W igh , M. N. & Ziegle , A. ange : A Fas Implemen a ion o Random Fo es s o High
618
Dimensional Da a in C++ and R. Jou nal o S a is ical So wa e 77, 1–17 (2017).
619
60. Kuzne so a, A., B ockho , P. B. & Ch is ensen, R. H. B. lme Tes Package: Tes s in Linea
620
Mixed E ec s Models. Jou nal o S a is ical So wa e 82, 1–26 (2017).
621
61. Lüdecke, D. gge ec s: Tidy Da a F ames o Ma ginal E ec s om Reg ession Models.
622
Jou nal o Open Sou ce So wa e 3, 772 (2018).
623
62. Hel ens ein, J. e al. Fa me su eys in Eu ope sugges ha specialized, in ensi e a ms we e
624
mo e likely o pe cei e nega i e impac s om COVID-19. Ag onomy o Sus ainable
625
De elopmen 42, 84 (2022).
626
627
628
Supplemen a y in o o :
629
Con inen al-scale e idence o a m managemen impac s on
630
soil ca bon
631
632
Julian Hel ens ein1, Nick an Dijk1, Anna Edlinge 2, Gab iel Y.K. Moine 3, Sophie an Rijssel1,
633
Alexand e M.J.-C. Wadoux4, Rachel C eame 3, Ca men Vazquez3, Ve a L. Mulde 1
634
1 Soil Geog aphy and Landscape G oup, Wageningen Uni e si y & Resea ch, Wageningen, he
635
Ne he lands
636
2 Wageningen En i onmen al Resea ch, Wageningen Uni e si y & Resea ch, Wageningen, he
637
Ne he lands
638
3 Soil Biology G oup, Wageningen Uni e si y & Resea ch, Wageningen, he Ne he lands
639
4 College o Science and Enginee ing, James Cook Uni e si y, Cai ns, Aus alia
640
641
642
Supplemen a y igu es
643
644
Figu e 5. Spa ial co e age o a m managemen and soil p ope y da ase s. Da a om he Fa m
645
Accoun ancy Da a Ne wo k (FADN, n = 81,688 o he yea 2018) was used o p edic a m
646
managemen and (a) shows he numbe o a m eco ds pe NUTS2 adminis a i e egion o 2018.
647
The Eu opean soil moni o ing ne wo k (LUCAS Soil) co e s a ound 20,000 sample loca ions, o
648
which 8,834 we e on ag icul u al land and me ou inclusion c i e ia (b).
649
650
Figu e 6. Spa ial a iabili y in a m managemen based on 81,688 ep esen a i e indi idual a m
651
obse a ions om 2018. Maps show he a ea-weigh ed mean alue o a a m managemen
652
indica o pe NUTS2 adminis a i e egion. a) – c) o al N, P, K inpu in kg ha-1, including bo h
653
mine al e ilize and manu e; d) – ) he sha e o N, P, K espec i ely de i ed om manu e; g) is
654
he p e alence o o ganic a ming (% o a ea); h) o a ional di e si y measu ed as he Gini-
655
Simpson index –sco e close o one means high di e si y; i) sha e o ley and odde legumes in he
656
c op mix; j) illage in ensi y, see me hods o de ails. See Supplemen a y Fig. 1 o he numbe o
657
a ms (sample size) in each NUTS2 egion.
658
659
660
