Received: 26 November 2017 Revised: 8 August 2018 Accepted: 11 August 2018

DOI: 10.1002/ldr.3136

RESEARCH ARTICLE

Agroforestry systems: Meta‐analysis of soil carbon stocks,

sequestration processes, and future potentials
Lingling Shi1,2,3 | Wenting Feng4 | Jianchu Xu1,2 | Yakov Kuzyakov3,5,6

1
Key Laboratory for Plant Diversity and
Biogeography of East Asia, Kunming Institute Abstract
of Botany, Chinese Academy of Sciences, Agroforestry (AF) has the potential to restore degraded lands, provide a broader range
Kunming 650201, Yunnan, PR China
2
of ecosystem goods and services such as carbon (C) sequestration and high biodiver-
East and Central Asia, World Agroforestry
Centre, Kunming 650201, Yunnan, PR China sity, and increase soil fertility and ecosystem stability through additional C input from
3
Department of Soil Science of Temperate trees, erosion prevention, and microclimate improvement. Advantages and processes
Ecosystems, Department of Agricultural Soil
Science, University of Göttingen, Göttingen,
for global C sequestration in AF are unknown. We used a meta‐analysis of 427 soil C
37707, Germany stock data pairs grouped into four main AF systems—alley cropping, windbreaks,
4
National Engineering Laboratory for silvopastures, and homegardens—and evaluated changes in AF and adjacent control
Improving Quality of Arable Land, Institute of
Agricultural Resources and Regional Planning, cropland or pasture. Mean soil C stocks in AF (1‐m depth) were 126 Mg C·ha−1, which
Chinese Academy of Agricultural Sciences, is 19% more than that in cropland or pasture. The highest C stocks in soil were in sub-
Beijing, 100081, PR China
5 tropical homegardens, AF with younger trees, and topsoil (0–20 cm). Increased soil C
Institute of Environmental Sciences, Kazan
Federal University, Kazan, 420049, Russia stocks in AF were lower than aboveground C stocks in most AF systems, except alley
6
Soil Science Consulting, Göttingen, 37077, cropping. Homegardens stored the highest C in both aboveground and belowground,
Germany
especially in the subsoil (20–100 cm). Advantages of AF ecosystem services focusing
Correspondence
Yakov Kuzyakov, Department of Soil Science on mechanisms of belowground C sequestration were analyzed. AF could store
of Temperate Ecosystems, Department of
5.3 × 109 Mg additional C in soil on 944 Mha globally, with most in the tropics and
Agricultural Soil Science, University of
Göttingen, Göttingen 37707, Germany. subtropics. AF systems could greatly contribute to global soil C sequestration if used
Email: kuzyakov@gwdg.de
in larger areas. Future investigations of AF should include (a) mechanistic‐ and pro-
Funding information
Sino‐German (CSC‐DAAD) Postdoc Scholar-
cess‐based studies (instead of common monitoring and inventories), (b) models linking
ship Program; Key Project from the Ministry of forest and crop growth with soil water and C and nutrient cycling, and (c) accurate
Sciences and Technology of China, Grant/
Award Number: 2017YFC0505100
assessments of the AF area worldwide based on the remote sensing approaches.

KEY W ORDS

agroforestry management, carbon sequestration, ecosystem services, homegardens, meta‐analysis,

sustainable land use

1 | I N T RO D U CT I O N Kyoto Protocol since 2007 (Baah‐Acheamfour et al., 2017; Stavi &

Agroforestry (AF) is an intermediate land use practice that combines
trees or shrubs with crops or pastures and thus merges the ecosystem
characteristics of forests and cropland or pastures (Figure 1). AF is an
integrated approach to sustainable land use because of its socioeco-
nomic advantages and environmental benefits, including prevention
of soil erosion and microclimate improvement, as well as higher crop
productivity (Nair et al., 2011, Nair, 2013; Torralba, 2016). AF has also
been adopted as a greenhouse gas mitigation strategy under the
Lal, 2013). The current review has summarized and documented the
lower C loss through greenhouse gas emissions in AF systems com-
pared with croplands (Kim, Kirschbaum, & Beedy, 2016). Thus, there
is increasing research interest in the C sequestration potential, pro-
cesses, and mechanisms in AF systems (Nair, Kumar, & Nair, 2009a;
Lorenz & Lal, 2014; Zomer et al., 2016).
AF has diverse systems based on differences in tree cover per-
centage, plantation arrangement, labor intensity, and livestock density
(Nair, 2011; Figure 1). Many terms are used for the AF systems, such

3886 © 2018 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/ldr Land Degrad Dev. 2018;29:3886–3897.
SHI ET AL.
FIGURE 1 Schema of four dominating agroforestry (AF) systems (modified from Nair, 1993): Alley cropping (AC), homegardens (HG),
silvopastures (SP), and windbreaks (WB). Top: Depending on land use intensity, labor intensity, percentage of cropland and trees, and livestock.
Bottom: AF systems according to increasing land use intensity and anthropogenic impacts on crops and trees [Colour figure can be viewed at
wileyonlinelibrary.com]
as ‘agrisilviculture,’ ‘tree‐based intercropping,’ and ‘hedgerow
intercropping.’ Within these terms and the very broad range of
AF systems, four major groups of AF systems have been commonly
adopted worldwide (Figure 1; Nair, 1991): (a) alley cropping (e.g.,
tree‐based intercropping or hedgerow intercropping) consists of
rows of trees with wide interrow spacing, with annual crops grown
in the alleyways; (b) windbreaks (shelterbelts) consist of single or
multiple rows of trees, shrubs, or perennial grasses with high stems
that protect crops, livestock, wildlife, or people from intensive
winds; and (c) silvopastures (tree‐pasture or silvia forest) combine
trees with forage and livestock production. Trees are managed for
high‐value sawlogs and provide shade and shelter for livestock
and forage; (d) homegardens (multistrata systems) are the deliber-
ate management of multipurpose trees and shrubs in close associ-
ation with annual or perennial agricultural crops and livestock by
individual households. The total global area currently under AF
has been estimated to be approximately 1,023 Mha (Verchot et al.,
2007). Global areas under tree intercropping, multistrata systems,
protective systems, silvopastures, and tree woodlots are
estimated at 700, 100, 300, 450, and 50 Mha, respectively (Lorenz
& Lal, 2014).
The intercropped trees have deeper and broader root systems,
which could increase C allocation in soil and also improve soil aggre-
gate formation for C stabilization (Cardinael et al., 2015; Le
Bissonnais et al., 2017; Upson & Burgess, 2013). The lower intensity
of human management in AF, such as fertilization and tillage,
reduces the C loss by soil disturbances (Nair & Graetz, 2004). Fur-
thermore, established trees around or inside the croplands suppress
3887
wind and water erosion, which reduces aboveground and below-
ground C losses (Béliveau et al., 2017; Lacoste, Viaud, Michot, &
Walter, 2015). For example, in tropical areas, intercropping trees
protect soil against raindrop effects, reduce runoff velocity by
increasing surface roughness and water infiltration, and provide a lit-
ter layer and tree roots to decrease runoff (Béliveau et al., 2017).
Therefore, multiple mechanisms and processes are responsible for
the increased C stocks in AF systems compared with pure agricul-
tural lands, and a synthesis is necessary.
The C sequestration potential of AF has been estimated in vari-
ous individual or review studies over the past three decades, but
the estimates at the global scale vary widely (Cardinael et al., 2017;
Kim et al., 2016; Oelbermann, Voroney, & Gordon, 2004). The
aboveground and belowground vegetation biomass C stocks could
be up to 2.2 Pg C (1 Pg = 1015 g) in AF systems over 50 years, but
the estimates of global AF area are highly uncertain (Lorenz & Lal,
2014). The estimates of C accumulation in AF ranged from
0.29 to 15.2 Mg C·ha−1·yr−1 aboveground and from 30 to
300 Mg C·ha−1·yr−1 for soils down to 1‐m depth (Nair, Nair, et al.,
2009b). In other forest‐related land use types, soil organic C (SOC)
stocks decrease, as follows: old forest (288 Mg C·ha−1) > mixed species
reforestation (181 Mg C·ha−1) > plantations (142 Mg C·ha−1) > grass-
slands (110 Mg C·ha−1) > AF (99 Mg C·ha−1) > cropland (40 Mg C·ha
−1
; Albrecht & Kandji, 2003; Pandey, 2002). Carbon stocks in AF vary
with geographical locations and climatic zones. For example, in
Southeast Asia, agrosilviculture systems have the capacity for storing
12–228 Mg C·ha−1 in humid tropical lands and 68–81 Mg C·ha−1 in
dry lowlands (Basu, 2014). In particular, the differences between C
3888
stocks and C sequestration potentials among AF systems are yet to
be determined.
Although there are various estimates of C stocks in AF, the results
are highly divergent, and some studies do not have control plots for
comparison, making it difficult to accurately evaluate C sequestration
potential in AF. Furthermore, C sequestration potentials, processes,
and consequences of AF and their specifics have not yet been gener-
alized, and studies that estimate both the aboveground and below-
ground C are very rare. We collected literature data on soil C stocks
in AF and grouped all systems as alley cropping, windbreaks,
silvopastures, or homegardens. We aimed (a) to quantify C stocks in
soils in four AF systems compared with adjacent agricultural control
plots under various soil and climatic conditions; (b) to evaluate the
effects of AF characteristics, such as tree age, soil properties, and
depth on soil C stocks; and (c) to identify knowledge gaps regarding
the main processes and mechanisms in AF for short‐, medium‐, and
long‐term C accumulation in AF.
2 | MATERIALS AND METHODS
2.1 | Data collection from literature studies
Data sources on the effects of AF on C stocks were assembled from
the studies published from 1950 to 2018. To avoid the diverse terms
used in AF, we limited our search on four major AF systems. The stud-
ies were identified using the database Web of Science. We limited the
search parameters to papers whose title, abstract, or keywords
referred to ‘agroforestry,’ ‘alley cropping,’ ‘windbreaks/shelterbelts,’
‘silvopasture,’ or ‘homegardens’ in combination with ‘soil,’ ‘carbon,’ or
‘organic matter.’ The search yielded 1,440 literature sources.
Studies were screened using the following criteria: Studies
included field data, and each study site had an AF system and adjacent
cropland or pasture control plots. Modeling studies and reviews were
excluded. With the use of these criteria, 350 studies were selected
FIGURE 2 Worldwide distribution of the studies evaluated in the meta‐analysis: 76 papers with 427 pair comparisons. The numbers before slash
reflect the number of studies, and after the slash, the number of data pairs [Colour figure can be viewed at wileyonlinelibrary.com]
SHI ET AL.
and used for data collection from the full‐published text. The final car-
bon‐AF database contained 76 papers with 427 pair comparisons (AF
vs. cropland or pasture; Supporting Information). From these studies,
we coded information according to AF systems/location/tree age/
aboveground and soil C stocks/soil depths. If necessary, we used dig-
itizer software to extract data from original figures.
The database covered six continents and 16 countries, ranging
from temperate to tropical climatic zones (Figure 2). Most of the stud-
ies were in Asia and in tropical areas. The paired‐site design (AF with
cropland or pasture control) was the main experimental approach to
analyze AF effects. A meta‐analysis was conducted for soil C stocks,
considering five parameters: climate, geographical location, AF sys-
tems, tree age, and soil depths. Finally, the effects of environmental
factors (e.g., mean annual temperature) and AF properties (e.g., tree
percentage) were examined.
2.2 | Carbon stocks data collection, standardizations,
and calculations
Soil C stocks were present in original studies in various forms, includ-
ing root biomass, microbial biomass, and SOC. In this meta‐analysis,
we focused on SOC stocks, which present a long‐term sequestered
C pool and are the main driver for various ecosystem functions and
services. If a study reported C contents as percentages of soil mass,
we converted the values to metric tons per hectare (Mg C·ha−1) by
the following equations. The results were extrapolated to generate
SOC stocks on a hectare basis (Gelaw, Singh, & Lal, 2015):
C density per soil layer ¼ C content × BD × Layer thickness;
where BD is the bulk density, the units of all the variables are trans-
formed to SI units, and all results for C stocks from individual studies
were recalculated and presented here in units of Mg C·ha−1. For
papers that did not provide BD, we estimated BD from soil texture
or using the following equation (Post & Kwon, 2000):
SHI ET AL.
100
BD ¼ ;
OM% 100 − OM%
− systems
0:244 MBD
where OM% is the percentage of soil organic matter, assuming that
organic matter equals the percentage of soil C divided by 0.58 (Mann,
1986). MBD is mineral bulk density, and a typical value of 1.64 was
used (Mann, 1986).
To compare the changes in C both aboveground and belowground
(soil C), we collected 36 pairs of aboveground biomass C and soil C
data from 12 studies of the SOC meta‐analysis references database.
In these studies, the changes of aboveground biomass C compared
with controls (aboveground ΔC stock) were calculated as additional
tree biomass C. Tree biomass stocks were calculated by using the
Dixon method (Dixon, Winjum, Andrasko, Lee, & Schroeder, 1994),
on the basis of the volume of stem wood multiplied by species‐specific
wood densities. The obtained stem wood mass was then multiplied by
1.6 to estimate whole tree biomass, and C content was assumed to be
50% of the estimated whole tree biomass.
2.3 | Meta‐analysis
We extracted the means and standard errors (SEs) of each comparison,
and the standard deviations (SDs) were calculated as
pffiffiffi
SD ¼ SE n;
where n is the sample size. If data were given with a mean and a con-
fidence interval (CI), the SD was calculated as
pffiffiffi
SD ¼ ðCIu − CIl Þ n=2Z α=2 ;
where CIu and CIl are the upper and lower limits of CI and Zα/2 is the Z
score for a given level of significance, equal to 1.96 where α = 0.05
and 1.645 where α = 0.10. In studies without reported SE, SDs, or
CI, we assigned SDs of 1/10 of the means.
The means for each AF system and corresponding cropland (or
pasture) were used to compute the response ratio (RR) by
     
RR ¼ ln X1 =X 2 ¼ ln X 1 − ln X 2 :
RR gives the log of the proportional difference between the
groups. For each comparison, we calculated the unbiased standardized
mean difference (Hedge's d) between the mean C stocks in AF and
adjacent reference systems (cropland or pasture), where X1 is the mean
value of the C stock in AF (the average value of combined forest and
crop components in AF systems), and X2 is the mean C stock under
the reference systems (Arora et al., 2014). We calculated effect sizes
and conducted formal meta‐analyses in R with the ‘ROM’ function in
the ‘metaphor’ package (Viechtbauer, 2010), assuming that effect sizes
varied randomly among comparisons (Gurevitch et al., 2001). We used
the R package “metaphor” to calculate overall effect sizes and 95% CI.
Changes in C stocks were statistically significant if the calculated 95%
CI did not overlap with zero. The log RR was back‐transformed and
expressed as percentage C stock change ([RR − 1] × 100) for presenta-
tion in figures and text (ΔC stock, % change).
3889
2.4 | Analysis of factors affecting C stocks in AF
The following explanatory variables, selected a priori from literature,
were included in the range of mixed effects meta‐regression models
to test whether they explained any of the variations in RR: (a) soil
properties: soil N and P contents, soil pH, and soil clay, silt, and sand
contents; (b) vegetation properties: tree density and tree cover per-
centage; and (c) climatic factors: mean lowest and highest monthly
temperature, mean lowest and highest monthly precipitation, and
mean annual temperature and precipitation. Where appropriate, multi-
ple meta‐regressions were used to compare the effects of continuous
variables among categorical groups. Uncertainty in the regression
coefficients was quantified, using 95% CIs. For each meta‐analysis
model, residual heterogeneity (between studies variation) was esti-
mated by restricted maximum likelihood (Viechtbauer, 2010). We used
these estimates to calculate the I2 value. The I2 statistic was commonly
used to quantify the dispersion of effect sizes in a meta‐analysis. A
low I2 value indicated that variability among effect sizes was mainly
caused by sampling error within studies, whereas a high I2 value indi-
cated that variability was caused by true heterogeneity among studies.
We calculated the regression in R with the “rmd” functions in the met-
aphor package (Viechtbauer, 2010).
3 | RESULTS AND DISCUSSION
3.1 | Soil C stocks under AF systems
Soil C stocks in AF were similar across climatic zones and continents
(Figure 3). The absence of significant effects was connected with high
variation among the studies. To avoid the influence of background
variation, instead of absolute values, we compared soil C stocks in
AF with those in nearby control plots (cropland/pasture). At the conti-
nent scale, the meta‐analysis resulted in 40% higher soil C stocks in AF
of Asia compared with controls (Figure 4). In total, the soil C stocks in
AF were 18% higher than in nearby control lands. The variability
between studies explained 40% (I2) of the increase, indicating that
studies were highly affected by local conditions, such as microclimate,
soil, and vegetation characteristics (Figure 4). The number of studies
was weighted in the meta‐analysis. For example, the highest study
numbers in Asia and Africa, compared with cropland, showed the larg-
est soil C stock increase (Figure 4).
The mean absolute C stocks in AF systems were similar among AF
systems (35–40 Mg C·ha−1), with slightly lower stocks in homegardens
(Figure 3). Homegardens had 50% more soil C stocks than had con-
trols, which was twice than those of other AF systems (25% on aver-
age; Figure 4). The highest increase of soil C stocks in homegardens
was due to its highest tree density and litter production than in other
AF systems, which positively correlate with changes of soil C (Islam,
Dey, & Rahman, 2015). Silvopasture has similar or even higher tree
density than have alley cropping and windbreaks, but it had the lowest
increase or even had no changes in soil C stocks (Figure 4). C stocks in
topsoil under silvopasture, compared with nearby pastures, are
affected by grass inputs instead of tree inputs; thus, soil C increase
3890
FIGURE 3 Soil C stocks in agroforestry systems across continents, climatic zones, agroforestry systems, and tree ages. The black solid and red
dashed lines, lower and upper edges, bars, and black circles represent the median and mean values, the 25th and 75th percentiles, the 5th and
95th percentiles, and the maximum and minimum values. Letters indicate significant differences among the groups (p < 0.05). The numbers above
the boxes indicate the numbers of data pairs included [Colour figure can be viewed at wileyonlinelibrary.com]
FIGURE 4 Changes (Δ) of soil C stocks in agroforestry (AF)
compared with cropland (or pasture) on six continents, in three
climate zones, across four AF systems and depending on tree age.
Squares with error bars denote the overall mean response ratio ± 95%
confidence interval (CI). The 95% CIs that do not cross the zero line
indicate significant differences of C stocks between AF and cropland
(or pasture). The numbers indicate the numbers of data pairs included.
I2 is the measure of heterogeneity: 0–40% means little or no
heterogeneity, 30–60% means moderate heterogeneity, 50–90%
indicates substantial heterogeneity, and 75–100% indicates high
heterogeneity [Colour figure can be viewed at wileyonlinelibrary.com]
was not significant as in the other AF systems (Figure 1; Upson, Bur-
gess, & Morison, 2016).
At the regional scale, soil C stocks and sequestration potentials in
AF also changed with tree age. With the tree growth up to 40 years,
SHI ET AL.
the mean soil C stocks slightly increased from 30 to 40 Mg C·ha−1,
but AF with older trees had no clear trends because of a limited num-
ber of studies (Figure 3). Across all age classes, increased soil C stocks
in AF were reduced with tree age, indicating the highest changes in
the younger trees (younger than 10 years; Figure 4). This result was
in opposition to the general opinion that tree‐mediated soil C seques-
tration is lower at early tree age because most of the photosynthesis C
was used to support tree growth (Takimoto, Nair, & Nair, 2008).
Cardinael et al. (2017) suggested that young plantations stored addi-
tional SOC but mainly under the tree rows, possibly as a result of
the herbaceous vegetation. The increased C input might be the result
of fast‐growing tree species chosen for most AF systems, especially in
alley cropping and homegardens, which can quickly increase litter
inputs by tree pruning and accumulate SOC especially at the early
age (Singh & Gill, 2014).
Soil depth was another important factor for the differences in C
stocks. AF increased in deep soil C accumulation tree roots. The soil
C stocks of AF systems were greater than those of cropland/pasture
for depths down to 100 cm (Figure 5). The meta‐analysis indicated
that soil C stocks in the uppermost 60–80 cm were significantly higher
in AF than in cropland, but the differences decreased with soil depths
(Figures 4 and 5). The results might be biased because a lower number
of studies (162 vs. 265) reported C stocks in deep soil (>30 cm). Soil C
stocks in AF were 126 Mg C·ha−1, which is 19% more compared than
those in cropland or pasture (Figure 5).
The changes in soil C stocks in topsoil were sensitive to litter
inputs by tree pruning and herbaceous vegetation root inputs.
Although the increases in C stocks were lower in deeper soil than in
topsoil, they still suggest that tree plantations add C transport into
deeper layers through rhizodeposition and fine root turnover (Thakur,
SHI ET AL.
−1
FIGURE 5 Soil C stocks (Mg C·ha ) in 20‐cm layers under
agroforestry (averaged over all systems; green squares) and controls
(cropland/pasture; red triangles) depending on depths. Error bars
indicate ±standard errors. Asterisks indicate significant (p < 0.05)
differences between agroforestry and cropland/pasture. The values
close to the legend show the total C stocks in 1‐m depth under
agroforestry and cropland. The C stock surplus in 1‐m depth is equal
to 19% of that in the controls [Colour figure can be viewed at
wileyonlinelibrary.com]
Kumar, & Kunhamu, 2015). This is especially important for C seques-
tration considering the long mean residence time of C in deep soil
(Schneckenberger & Kuzyakov, 2007; Zang et al., 2018).
The increases of soil C stocks in AF compared with controls were
influenced by soil properties (texture, pH, and P and N content; Figure
S1), climate (temperature and precipitation), and vegetation (tree den-
sity and cover percentage). These factors were correlated with soil C
accumulation in AF relative to cropland or pasture but explained a
small amount of variation in the RR (Table S1; Figure S1). The low
explaining power of soil properties, climate, and vegetation was con-
nected with complex interactions among these factors because AF
combines the characteristics of forest and agriculture.
3.2 | Soil C stocks versus aboveground biomass C
Previous AF studies have focused on the survey of tree biomass C
accumulated aboveground, because tree stem biomass C can be stable
for a long time. Recently, the belowground C stocks are receiving
more and more attraction, but a comparison of the differences
between these two C pools is rare. We collected 12 studies (36 data
pairs that contained both the aboveground and belowground C stocks)
to explore whether high aboveground C accumulation in AF is accom-
panied by larger soil C stocks.
3891
Generally, AF systems have much higher increases in above-
ground C stocks than in soil C, especially in silvopastures (Figure 6a).
Homegardens and windbreaks stored high aboveground C and more
soil C stocks than did cropland (Figure 6a). Alley cropping had similar
increases in aboveground and soil C stocks, but the sum of both C
stocks was lower than that of other AF systems (Figure 6a). Above-
ground C was correlated with tree density and growth rate, which
was higher in AF systems such as homegardens, windbreaks, and
silvopastures. However, tree management in the alley cropping sys-
tem, such as pruning, might limit the biomass accumulation in stems
(Gamble, Johnson, Sheaffer, Current, & Wyse, 2014). Homegardens
have the highest increase of soil C, which was higher in the subsoil
(20–100 cm) than in topsoil (0–20 cm; Figure 6b). This difference
was due to the intensive root C turnover and litter inputs below-
ground (Becker, Pabst, Mnyonga, & Kuzyakov, 2015). Additionally,
homegardens are mainly located in the tropics with higher rainfall
and leaching potential, which could result in higher C transport into
deep soil (Kumar & Nair, 2004).
We compared aboveground and belowground C stocks in AF sys-
tems at the continent scale. Most of the AF systems have similar
aboveground and belowground C stocks, but the stocks depend on
study locations (Figure 6c). For example, alley cropping had higher
total C stocks (aboveground and belowground) in Asia, but lower total
C stocks in alley cropping were found in South America (Figure 6c).
Such a difference is explained by tree species used in each continent.
Windbreaks have more total C stocks in North America than in Europe
(Figure 6c). Windbreaks function by reducing the C loss by wind (and
partly by water) erosion, and this situation is more serious in North
America in several deserts and drylands than in Europe (Brandle,
Hodges, & Zhou, 2004). Changes in total C stocks in AF systems also
depended on climate (Figure 6d). Alley cropping in the tropics had sig-
nificantly higher soil C stocks than aboveground biomass C, which
might be attributed to large amounts of fast‐growing trees and N‐fix-
ing trees in AF plantations that play an important role in soil C seques-
tration (Tsonkova, Böhm, Quinkenstein, Freese, & Jose, 2012).
Silvopasture had higher tree biomass in the tropics than in temperate
climate because of higher tree density and growth rate in the tropics
(Figure 6d). For better understanding and management, the processes
and mechanisms of C sequestration in AF systems still need to be
determined.
3.3 | Mechanisms of C sequestration in AF soils and
knowledge gaps
AF systems improve soil C sequestration through increased plant C
inputs, improved microclimate, reduced soil erosion, and closed nutri-
ent and water cycling. One of the major AF characteristics is to
increase plant diversity by additional tree planting in cropland or pas-
tures. Tree density and diversity are higher in homegardens, followed
by alley cropping and silvopastures, and lower in windbreaks.
Increases in aboveground plant biomass are varied by plant and crop
species, growth rate, and tree age (Zomer et al., 2016). The C stocks
under AF increase mainly through root biomass, rhizodeposition, and
dissolved organic C (DOC) leached from litter and SOC (Table 1).
Besides root C inputs, in alley cropping, trees also introduce additional
3892 SHI ET AL.

FIGURE 6 (a) Changes of C stocks (aboveground and soil) in four agroforestry (AF) systems compared to cropland/pasture. AC, alley cropping;
HG, homegardens; WB, windbreaks; SP, silvopastures. (b) Changes of soil C stocks in topsoil (0–20 cm) and subsoil (20–100 cm) of four AF
systems compared with cropland. (c) Changes of C stocks aboveground and belowground in AF systems located in six continents. (d) Changes of C
stocks between aboveground and belowground in AF systems depending on climate zones. Error bars indicate ±SE. Positive and negative values
correspond respectively to increase and decrease of C stocks compared with that in cropland [Colour figure can be viewed at wileyonlinelibrary.
com]

TABLE 1 Mechanisms of direct and indirect effects on C inputs and soil C stabilization in agroforestry

C input and stabilization Direct effects Indirect effects

C input Aboveground ↓ Reduction/stop of water and wind erosion (+) (WB,
↑ Tree and crop stem and leaves litter biomass (+) (HG, AC, AC)
SP, WB) ↑ Stem leaching (+) (AC, HG)
↑ Dust deposition (+) (WB) ↓ Light by shading (−) (AC)
Belowground ↓ Pathogenic fungi (+) (AC, HG)
↑ Tree and crop root biomass (+) (HG, AC, SP, WB)
↑ Mycorrhizal fungal biomass (+/−) (HG, AC)
↑ Rhizodeposition of tree and crop roots (+) (HG, AC)
↑ Downward DOM transport (+) (SP)
C stabilization (solely ↑ Formation of mineral‐associated OM (+) (HG, AC) ↑ Soil structure (aggregation) (+) (HG, AC)
belowground) ↑ Other GHG such as CH4 (+) (HG, AC) ↓ Soil temperature (+) (WB, SP)
↑ SOM decomposition (−) (HG, AC) ↑ Burrowing animals (+) (HG)
↑ Soil and rhizosphere CO2 (−) (HG, AC) ↓ Agriculture management such as tillage and fertilizer
↑ Litter decomposition (+/−) (HG, AC, SP, WB) (+) (AC)
↑ Fungi/bacteria ratio (+/−) (HG, AC) ↑ Aeration and O2 availability (−) (HG, AC)
↑ AMF (tropical/subtropical trees, crops, shrubs) (+/−) (HG, ↑ Soil moisture in topsoil and subsoil (−) (SP)
AC)
↑ ECM (boreal/temperate trees) (+/−) (HG, AC)
↑ ↓ SA fungi (+/−) (HG, AC)

Note. ↑ shows increase or improvement. ↓ means decrease or degradation. (+) or (−) means positive or negative effects on C sequestration, respectively. (+/
−) means unclear effects or the effect depends on other environmental factors. AMF, ECM, SA, AC, HG, SP, and WB show especially strong effects under
the agroforestry systems. AC: alley cropping; AMF: arbuscular mycorrhizal fungi; ECM: ectomycorrhizal fungi; GHG: greenhouse gas; HG: homegardens; SA:
saprotrophic fungi; SP: silvopastures; WB: windbreaks.
SHI ET AL.
microbial groups, such as mycorrhizal fungi and N‐fixing and P‐solubi-
lizing bacteria. Mycorrhizal fungal hyphae are an important C pool in
soil. Mycorrhizal fungi and fine roots transfer large proportions of pri-
mary production into soils, but the estimates vary substantially across
studies (Janssens et al., 2010; Pausch & Kuzyakov, 2018; Sommer,
Dippold, Flessa, & Kuzyakov, 2016). Additionally, the biological N fix-
ation and nutrient mobilization from deep soils have increased soil fer-
tility, prompting C sequestration (Araujo, Iwata, Jr, & Xavier, 2012;
Balakrishna, Lakshmipathy, Bagyaraj, & Ashwin, 2017; Vallejo, Roldan,
& Dick, 2010). Trees in alley cropping and windbreaks modify microcli-
mate, increase soil moisture, and lower temperate, resulting in slow C
decomposition and increasing crop production at the same time.
Silvopasture has a similar or even higher tree density than has
alley cropping. Conversely, the tree species in silvopasture commonly
have deeper root systems to avoid the competition with grasses in the
surface layer (Sharrow, Buck, Lassoie, & Fernandes, 1999). These deep
and broad root systems have great effects on C input and SOC stabi-
lization (Schmitt‐Harsh, Evans, Castellanos, & Randolph, 2012; Zang,
Wang, & Kuzyakov, 2016; Zang et al., 2018). Rhizodeposition and
DOC sorption onto minerals are the controlling steps in the formation
of mineral‐associated organic matter (Boffa, Taonda, Dickey, & Boffa,
2000; Kaiser & Kalbitz, 2012; Kukal & Bawa, 2014; Lorenz & Lal,
2014). The quantity of rhizodeposition from trees and DOC derived
from through‐fall, stem‐fall, and recently deposited litter is virtually
unknown (Luo, Wang, Feng, Luo, & Baldock, 2017; Mganga &
Kuzyakov, 2014; Pausch & Kuzyakov, 2018). Tree roots grow deep
and deliver rhizodeposition deep into soils that are important for long
SOC preservation (Schneckenberger & Kuzyakov, 2007; Zang et al.,
2018). SOC in subsoil has three to 10 times longer mean residence
time than in topsoil (Kaiser et al., 2010; Zang et al., 2018). Additionally,
tree roots increase soil C stabilization by facilitating the formation of
aggregates and mineral‐associated organic matter (Gelaw et al.,
2015; Guimaraes, Mendona, Passos, & Andrade, 2014; Gunina &
Kuzyakov, 2014). Compared with cropland/pasture, trees have larger
root biomass, rhizodeposition, and especially the hyphae of
ectomycorrhizae and glycoproteins of arbuscular mycorrhizae, all of
which bind mineral for the aggregate formation (Lenka, Choudhury,
Sudhishri, Dass, & Patnaik, 2012; Rillig, 2004).
Windbreaks play an important role in preventing wind damage
and erosion (Brandle et al., 2004), although it has lower tree density
than has the other AF systems. It slows down wind flow and leads
to the deposition of soil particles removed from the surface of crop-
land and the accumulation under tree lines (Wiesmeier et al., 2018).
This is also a reason for higher C stocks under the windbreak stripes.
The windbreaks reduce soil loss runoff in the rainy season (Atangana,
Khasa, Chang, et al., 2014) and improve microclimate conditions.
Lower temperature achieved in hot seasons, and higher moisture
increases crop and tree production, thus increasing crop yields and
plant C inputs (Zheng, Zhu, & Xing, 2016). Tree species used in wind-
breaks generally have stronger root systems than those in
silvopasture, helping them stand against wind and use deep soil mois-
ture during drought (Brandle et al., 2004).
Homegardens are getting more and more attention in AF
research, because of their high efficiency in crop production, espe-
cially in mountain areas where land used for crop production is
3893
strongly limited (Kumar & Nair, 2004). Having the highest similarity
with forest, homegardens harbored very high aboveground plant den-
sity and richness, which correlated well with soil C accumulation
(Mattsson, Ostwald, Nissanka, & Marambe, 2013; Mattsson, Ostwald,
Nissanka, & Pushpakumara, 2015; Mganga, Razavi, & Kuzyakov, 2016;
Saha, Nair, Nair, & Kumar, 2009). In mountain areas, homegardens are
used in sloping areas (C. Hemp & Hemp, 2008). The multitiered
homegarden canopy and root architecture, besides the litter layer,
act as a multilayer defense against the effects of the falling raindrops
and thus reduce soil erosion rates (Kumar & Nair, 2004). Furthermore,
the runaway C and nutrients are taken up by plants and microflora
(Kumar, 2011; Mganga et al., 2016; Pandey & Srivastava, 2009). The
AF systems also indirectly reduce the use of fertilizers, tillage, and
other disturbing activities, improving biodiversity (Hemp, 2005,
2006), nutrient retention, and soil C stability (Pabst, Gerschlauer,
Kiese, & Kuzyakov, 2016).
3.4 | Global up‐scaling: C sequestered in AF soil and
potential sequestration
To estimate total C stocks at the global scale, we need not only the C
stocks per unit land but also the total area of each AF system. In our
meta‐analysis, we compared soil C stocks in AF systems on the basis
of available data of the area for each AF type from the literature. Of
the four AF systems, alley cropping covers a much larger area than
do homegardens but sequesters less C in soil (Cardinael et al., 2015).
The AF area in a temperate climate is similar to that in the tropics
(470 Mha). However, total soil C stocks in temperate AF were approx-
imately 10 times less than those in the tropics and subtropics
(Table 2). We estimated that the global area of AF is at least
960 Mha, which accumulated 3.6 × 109 Mg more soil C stocks than
did cropland/pasture (Table 2).
The soil C stocks in AF are still an underestimation because little
data are available from other AF systems, such as forest farming.
The potential for C sequestration in AF at the global scale is remark-
ably high, because degraded forests, low‐productivity cropland, and
bare land after stem harvest could be converted to AF (Ilany, Ashton,
Montagnini, & Martinez, 2010). There is 1.9 billion ha of land degraded
because of erosion, which could be sustainably improved by AF
(Brown, 2006). Assuming the average surplus of 36 Mg C·ha−1 stored
in AF soils compared with cropland, these additional AF areas would
sequester 6.9 × 1010 Mg C globally (=69 Pg C). This survey also
pointed out the high C sequestration potential of homegardens.
Homegardens are common traditional farming practices in Asia and
some parts of Africa and have attracted growing attention because
of their economic and ecological benefits (Hemp & Hemp, 2008).
The area under homegardens was assessed to be 8.0 Mha, much less
than that of other AF systems (e.g., 400 Mha under silvopasture;
Takimoto, Nair, & Nair, 2009). Therefore, increasing the area of
homegardens could strongly contribute to global C sequestration.
Despite its global importance, it is hard to find data on the actual
extent of AF around the world. The working group of World Agrofor-
estry Centre (ICRAF) used the percentage of tree cover as a criterion
to calculate AF area for each category on the basis of remote sensing
data (Zomer, Trabucco, Coe, & Place, 2009). This method provides
3894 SHI ET AL.

TABLE 2 Global areas and soil carbon (C) stocks under four main agroforestry (AF) systems in the tropical, subtropical, and temperate climatic
zones

Region AF systems Number of references Area (million ha) Soil C in AF (Mg C·ha−1) Δ Soil C (Mg C·ha−1)
a
Tropics and subtropics Alley cropping 35 540 38.6 5.3
Homegardens 5 100b 40.8 8.7/23.4c
Silvopastures 6 n.d. 28.1 −1.8
Windbreaks 2 n.d. 30.1 6.3
Sum 48 >640 Δ4.5 × 109 Mg C
Temperate Alley cropping 13 164d 17.7 2.2
Homegardens 2 n.d. 30.2 10.0
Silvopastures 4 70b 36.8 0.70
Windbreaks 8 80d 66.5 0.90
Sum 27 >304 Δ8.0 × 108 Mg C
Sum of global 75 >944 Δ5.3 × 109 Mg C

Note. n.d.: no data; Δ Soil C represents additional C sequestered in soils under AF compared to the local cropland (or pasture).
a

b
Data extracted from Nair and Nair (2003).
c
Δ Soil C of homegardens highly varied between tropical and subtropical climates (8.7 vs. 23.4 Mg C·ha−1). To estimate C sequestration potentials, we used
the largest area estimated by the references in the calculation. The estimate of alley cropping in tropical and subtropical areas is based on the sum of tree‐
base intercropping AF area from Nair (2012) (700 Mha) and subtracted from the temperate alley cropping area (million ha).
d
Data from Udawatta and Jose (2011). Temperate alley cropping data are mainly from the United States.

accurate estimates of AF area sizes and aboveground biomass C
stocks than do other methods (Zomer et al., 2016). We have found a
weak correlation between tree density and soil C stocks. Therefore,
further studies could include other factors in the regressions between
soil C and tree characteristics (e.g., density, coverage, and richness),
then characterize each AF systems using tree characteristics, and
finally use remote sensing data to accurately estimate the distribution
of various AF systems.
4 | CONCLUSIONS AND OPEN RESEARCH
QUESTIONS
AF promotes C sequestration relative to cropland or pasture. This
review indicated that all four main AF systems—alley cropping, wind-
breaks, silvopastures, and homegardens—sequestered significantly
more C than did cropland (or pasture). Of the four main AF systems,
silvopasture sequestrated the most C aboveground, whereas
homegardens were best at storing C in soil per area unit.
Homegardens, however, currently cover only a small area and thus
have great potentials for future C sequestration by increasing the area.
Carbon sequestered in AF aboveground is slightly more than in the
upper 20‐cm soil but much smaller than C stocks in soil up to 1‐m
depth. Mean soil C stocks in AF were 126 Mg C·ha−1, which is 19%
more than those in cropland or pasture. AF systems in Asia seques-
tered much more C both aboveground and belowground than in other
continents. Intercropping with younger trees (<20 years) increased C
in soil more than with older trees. The following important knowledge
gaps regarding C sequestration in AF were identified: (a) quantitative
assessment of C inputs and stocks in various parts of AF systems, with
special focus on deep soil C and its dependence on tree species and
their age; (b) optimization of the area allocated to trees and crops
within each AF system to reach maximal C sequestration, increase
yield, maximize ecosystem services, and improve environmental condi-
tions; and (c) development of new remote sensing approaches to
distinguish AF from the background of forests, plantations, and gar-
dens and to quantify the AF area.
ACKNOWLEDGMENTS
We greatly acknowledge the work carried out by the researchers,
whose published data were used for this meta‐analysis. We are very
thankful to three anonymous reviewers who strongly helped to
improve the manuscript. This research was financially supported by
the Sino‐German (CSC‐DAAD) Postdoc Scholarship Program for Dr.
Lingling Shi and Key Project from the Ministry of Sciences and
Technology of China (Grant No: 2017YFC0505100). The work is
performed according to the Russian Government Program of
Competitive Growth of Kazan Federal University. We thank Soil
Science Consulting (https://soilscicon.wordpress.com) for help with
the paper preparation.
ORCID
Lingling Shi http://orcid.org/0000-0001-8213-689X
Wenting Feng http://orcid.org/0000-0002-3189-3687
Yakov Kuzyakov http://orcid.org/0000-0002-9863-8461
RE FE RE NC ES
Albrecht, A., & Kandji, S. T. (2003). Carbon sequestration in tropical agro-
forestry systems. Agriculture Ecosystems and Environment, 99, 15–27.
https://doi.org/10.1016/S0167‐8809(03)00138‐5
Araujo, A. S. F., Iwata, B. D. F., de Andrade Lira, M. Jr., & Xavier, G. R.
(2012). Microbiological process in agroforestry systems. A review.
Agronomy for Sustainable Development, 32, 215–226. https://doi.org/
10.1007/s13593‐011‐0026‐0
Arora, G., Chaturvedi, S., Kaushal, R., Nain, A., Tewari, S., Alam, N. M., &
Chaturvedi, O. P. (2014). Growth, biomass, carbon stocks, and seques-
tration in age series of Populus deltoides plantations in Tarai region of
central Himalaya. Turkish Journal of Agriculture and Forestry, 38,
550–560. https://doi.org/10.3906/tar‐1307‐94
Atangana A, Khasa D, Chang S, Ann Degrande 2014. Agroforestry for soil
conservation. Tropical agroforestry. Dordrecht: Springer. 2014: 203–216
SHI ET AL.
Baah‐Acheamfour, M., Chang, S. X., Bork, E. W., & Carlyle, C. N. (2017).
The potential of agroforestry to reduce atmospheric greenhouse gases
in Canada: Insight from pairwise comparisons with traditional agricul-
ture, data gaps and future research. Forestry Chronicle, 93, 180–189.
https://doi.org/10.5558/tfc2017‐024
Balakrishna, A. N., Lakshmipathy, R., Bagyaraj, D. J., & Ashwin, R. (2017).
Influence of alley copping system on am fungi, microbial biomass C
and yield of finger millet, peanut and pigeon pea. Agroforestry Systems,
91, 487–493. https://doi.org/10.1007/s10457‐016‐9949‐4
Basu, J. P. (2014). Agroforestry, climate change mitigation and livelihood
security in India. New Zealand Journal of Forestry Science, 44, 1–10.
https://doi.org/10.1186/1179‐5395‐44‐S1‐S11
Becker, J., Pabst, H., Mnyonga, J., & Kuzyakov, Y. (2015). Annual litterfall
dynamics and nutrient deposition depending on elevation and land
use at Mt. Kilimanjaro. Biogeosciences, 12, 5635–5646. https://doi.
org/10.5194/bg‐12‐5635‐2015
Béliveau, A., Lucotte, M., Davidson, R., Paquet, S., Mertens, F., Passos, C. J.,
& Romana, C. A. (2017). Reduction of soil erosion and mercury losses in
agroforestry systems compared to forests and cultivated fields in the
Brazilian Amazon. Journal of Environmental Management, 203,
522–532. https://doi.org/10.1016/j.jenvman.2017.07.037
Boffa, J. M., Taonda, S. J. B., Dickey, J. B., & Boffa, J. M. (2000). Field‐scale
influence of karite (Vitellaria paradoxa) on sorghum production in the
Sudan zone of Burkina Faso. Agroforestry Systems, 49, 153–175.
https://doi.org/10.1023/a:1006389828259
Brandle, J. R., Hodges, L., & Zhou, X. H. (2004). Windbreaks in North Amer-
ican agricultural systems. New Vistas in Agroforestry. Springer
Netherlands.. https://doi.org/10.1007/978‐94‐017‐2424‐15
Brown, D. R. (2006). Personal preferences and intensification of land use:
Their impact on southern Cameroonian slash‐and‐burn agroforestry
systems. Agroforestry Systems, 68, 53–67. https://doi.org/10.1007/
s10457‐006‐0003‐9
Cardinael, R., Chevallier, T., Cambou, A., Béral, C., Barthès, B. G., Dupraz,
C., … Kouakoua, E. C. C. (2017). Increased soil organic carbon stocks
under agroforestry: A survey of six different sites in France. Agriculture,
Ecosystems and Environment, 236, 243–255. https://doi.org/10.1016/J.
AGEE.2016.12.011
Cardinael, R., Mao, Z., Prieto, I., Stokes, A., Dupraz, C., Kim, J. H., & Jour-
dan, C. (2015). Competition with winter crops induces deeper rooting
of walnut trees in a Mediterranean alley cropping agroforestry system.
Plant and Soil, 391, 219–235. https://doi.org/10.1007/s11104‐015‐
2422‐8
Dixon, R. K., Winjum, J. K., Andrasko, K. J., Lee, J. J., & Schroeder, P. E.
(1994). Integrated land‐use systems: Assessment of promising
agroforest and alternative land‐use practices to enhance carbon con-
servation and sequestration. Climatic Change, 27, 71–92. https://doi.
org/10.1007/bf01098474
Gamble, J. D., Johnson, G., Sheaffer, C. C., Current, D. A., & Wyse, D. L.
(2014). Establishment and early productivity of perennial biomass alley
cropping systems in Minnesota, USA. Agroforestry Systems, 88, 75–85.
https://doi.org/10.1007/s10457‐013‐9657‐2
Gelaw, A. M., Singh, B. R., & Lal, R. (2015). Organic carbon and nitrogen
associated with soil aggregates and particle sizes under different land
uses in tigray, northern Ethiopia. Land Degradation & Development, 26,
690–700. https://doi.org/10.1002/ldr.2261
Guimaraes, G. P., Mendona, E. D., Passos, R. R., & Andrade, F. V. (2014).
Soil aggregation and organic carbon of Oxisols under coffee in agrofor-
estry systems. Revista Brasileira De Ciencia Do Solo, 38, 278–287.
https://doi.org/10.1590/S0100‐06832014000100028
Gunina, A., & Kuzyakov, Y. (2014). Pathways of litter C by formation of
aggregates and SOM density fractions: Implications from 13C natural
abundance. Soil Biology and Biochemistry, 71, 95–104. https://doi.org/
10.1590/S0100‐06832014000100028
Gurevitch, J., Curtis, P. S., & Jones, M. H. (2001). Meta‐analysis in ecology.
Advances in Ecological Research, 32, 199–247. https://doi.org/10.1016/
S0065‐2504(01)32013‐5
3895
Hemp, A. (2005). The Chagga home gardens—Relict areas for endemic Saltatoria
species (Insecta: Orthoptera) on Mount Kilimanjaro. Biological Conservation,
125, 203–209. https://doi.org/10.1016/j.biocon.2005.03.018
Hemp, A. (2006). Vegetation of Kilimanjaro: Hidden endemics and missing
bamboo. African Journal of Ecology, 44, 305–328. https://doi.org/
10.1111/j.1365‐2028.2006.00679.x
Hemp C, Hemp A. (2008). The chagga homegardens on Kilimanjaro. Maga-
zine of the International Human Dimensions Programme on Global
Environmental Change. Mountainous Regions: Laboratories for Adapta-
tion 12–17. https://doi.org/10.1016/j.biocon.2005.03.018
Ilany, T., Ashton, M. S., Montagnini, F., & Martinez, C. (2010). Using agro-
forestry to improve soil fertility: effects of intercropping on Ilex
paraguariensis (yerba mate) plantations with Araucaria angustifolia.
Agroforestry Systems, 80, 399–409. https://doi.org/10.1007/s10457‐
010‐9317‐8
Islam, M., Dey, A., & Rahman, M. (2015). Effect of tree diversity on soil
organic carbon content in the homegarden agroforestry system of
north‐eastern Bangladesh. Small‐Scale Forestry, 14, 91–101. https://
doi.org/10.1007/s11842‐014‐9275‐5
Janssens, I. A., Dieleman, W., Luyssaert, S., Subke, J.‐A., Reichstein, M.,
Ceulemans, R., … Matteucci, G. (2010). Reduction of forest soil respira-
tion in response to nitrogen deposition. Nature Geoscience, 3, 315–322.
https://doi.org/10.1038/ngeo844
Kaiser, C., Koranda, M., Kitzler, B., Fuchslueger, L., Schnecker, J.,
Schweiger, P., … Richter, A. (2010). Belowground carbon allocation by
trees drives seasonal patterns of extracellular enzyme activities by
altering microbial community composition in a beech forest soil. New
Phytologist, 187, 843–858. https://doi.org/10.1111/j.1469‐
8137.2010.03321.x
Kaiser, K., & Kalbitz, K. (2012). Cycling downwards—Dissolved organic
matter in soils. Soil Biology and Biochemistry, 52, 29–32. https://doi.
org/10.1016/j.soilbio.2012.04.002
Kim, D. G., Kirschbaum, M. U. F., & Beedy, T. L. (2016). Carbon sequestra-
tion and net emissions of CH4 and N2O under agroforestry:
Synthesizing available data and suggestions for future studies. Agricul-
ture Ecosystems and Environment, 226, 65–78. https://doi.org/
10.1016/j.agee.2016.04.011
Kukal, S. S., & Bawa, S. S. (2014). Soil organic carbon stock and fractions in
relation to land use and soil depth in the degraded Shiwaliks hills of
lower Himalayas. Land Degradation & Development, 25, 407–416.
https://doi.org/10.1002/ldr.2151
Kumar, B. M. (2011). Species richness and aboveground carbon stocks in
the homegardens of central Kerala, India. Agriculture, Ecosystems and
Environment, 140, 430–440. https://doi.org/10.1016/j.
agee.2011.01.006
Kumar, B. M., & Nair, P. K. R. (2004). The enigma of tropical homegardens.
Agroforestry Systems, 62, 135–152. https://doi.org/10.1023/B:
AGFO.0000028995.13227.ca
Lacoste, M., Viaud, V., Michot, D., & Walter, C. (2015). Landscape‐scale
modelling of erosion processes and soil carbon dynamics under land‐
use and climate change in agroecosystems. European Journal of Soil Sci-
ence, 66, 780–791. https://doi.org/10.1111/ejss.12267
Le Bissonnais, Y., Prieto, I., Roumet, C., Nespoulous, J., Metayer, J., Huon,
S., … Stokes, A. (2017). Soil aggregate stability in Mediterranean and
tropical agro‐ecosystems: Effect of plant roots and soil characteristics.
Plant and Soil, 424, 303–317. https://doi.org/10.1007/s11104‐017‐
3423‐6
Lenka, N. K., Choudhury, P. R., Sudhishri, S., Dass, A., & Patnaik, U. S.
(2012). Soil aggregation, carbon build up and root zone soil moisture
in degraded sloping lands under selected agroforestry based rehabilita-
tion systems in eastern India. Agriculture Ecosystems and Environment,
150, 54–62. https://doi.org/10.1016/j.agee.2012.01.003
Lorenz, K., & Lal, R. (2014). Soil organic carbon sequestration in agrofor-
estry systems. A review. Agronomy for Sustainable Development, 34,
443–454. https://doi.org/10.1007/s13593‐014‐0212‐y
3896
Luo, Z., Wang, E., Feng, W., Luo, Y., & Baldock, J. (2017). The importance
and requirement of belowground carbon inputs for robust estimation
of soil organic carbon dynamics: Reply to Keel et al. (2017). Global
Change Biology, 2017, 1–2. https://doi.org/10.1111/gcb.13949
Mann, L. K. (1986). Changes in soil carbon storage after cultivation. Soil Sci-
ence, 142, 279–288. https://doi.org/10.1097/00010694‐198611000‐
00006
Mattsson, E., Ostwald, M., Nissanka, S. P., & Marambe, B. (2013).
Homegardens as a multi‐functional land‐use strategy in Sri Lanka with
focus on carbon sequestration. Ambio, 42, 892–902. https://doi.org/
10.1007/s13280‐013‐0390‐x
Mattsson, E., Ostwald, M., Nissanka, S. P., & Pushpakumara, D. K. N. G.
(2015). Quantification of carbon stock and tree diversity of
homegardens in a dry zone area of Moneragala District, Sri Lanka.
Agroforestry Systems, 89, 435–445. https://doi.org/10.1007/s10457‐
014‐9780‐8
Mganga, K. Z., & Kuzyakov, Y. (2014). Glucose decomposition and its incor-
poration into soil microbial biomass depending on land use in Mt.
Kilimanjaro ecosystems. European Journal of Soil Biology, 62, 74–82.
https://doi.org/10.1016/j.ejsobi.2014.02.015
Mganga, K. Z., Razavi, B. S., & Kuzyakov, Y. (2016). Land use affects soil
biochemical properties in Mann LK. 1986. Changes in soil carbon stor-
age after cultivation. Soil Science, 142, 279–288. https://doi.org/
10.1097/00010694‐198611000‐00006
Nair, P. K. R. (1991). State‐of‐the‐art of agroforestry systems. Forest Ecol-
ogy and Management, 45, 5–29. https://doi.org/10.1016/0378‐
1127(91)90203‐8
Nair, P. K. (1993). Ramachandran. An introduction to agroforestry. Springer
Science & Business Media.
Nair, P. K. R. (2011). Agroforestry systems and environmental quality:
Introduction. Journal of Environmental Quality, 40, 784–790. https://
doi.org/10.2134/jeq2011.0076
Nair, P. K. R. (2012). Carbon sequestration studies in agroforestry systems:
A reality‐check. Agroforestry Systems, 86, 243–253. https://doi.org/
10.1007/s10457‐011‐9434‐z
Nair, P. K. R. (2013). Agroforestry: Trees in support of sustainable agricul-
ture. In S. A. Elias, S. Marshall, M. Goldstein, et al. (Eds.), Reference
module in Earth systems and environmental sciences. 10.1016/B978‐0‐
12‐409548‐9.05088‐0
Nair, P. K. R., Kumar, B. M., & Nair, V. D. (2009a). Agroforestry as a strategy
for carbon sequestration. Journal of Plant Nutrition and Soil Science,
172, 10–23. https://doi.org/10.1002/jpln.200800030
Nair, P. K. R., Nair, V. D., Kumar, B. M., Haile, S. G., Dilly, O., & Pannell, D.
(2009b). Soil carbon sequestration in tropical agroforestry systems: A
feasibility appraisal. Environmental Science and Policy, 12, 1099–1111.
https://doi.org/10.1016/j.envsci.2009.01.010
Nair, P. K. R., Saha, S. K., Nair, V. D., & Haile, S. G. (2011). Potential for
greenhouse gas emissions from soil carbon stock following biofuel cul-
tivation on degraded lands. Land Degradation & Development, 22,
395–409. https://doi.org/10.1002/ldr.1016
Nair, P. R., & Nair, V. D. (2003). Carbon storage in North American agrofor-
estry systems. In M. J. Kimble, R. Lal, R. Birdsey, & L. S. Heath (Eds.),
2003The potential of US forest soils to sequester carbon and mitigate
the greenhouse effect (pp. 333–346). New York: CRC Press LLC.
Nair, V. D., & Graetz, D. A. (2004). Agroforestry as an approach to minimiz-
ing nutrient loss from heavily fertilized soils: The Florida experience.
Agroforestry Systems, 61, 269–279. https://doi.org/10.1023/B:
AGFO.0000029004.03475.1d
Oelbermann, M., Voroney, R. P., & Gordon, A. M. (2004). Carbon seques-
tration in tropical and temperate agroforestry systems: A review with
examples from Costa Rica and southern Canada. Agriculture Ecosystems
and Environment, 104, 359–377. https://doi.org/10.1016/j.
agee.2004.04.001
Pabst, H., Gerschlauer, F., Kiese, R., & Kuzyakov, Y. (2016). Land use and
precipitation affect organic and microbial carbon stocks and the spe-
cific metabolic quotient in soils of eleven ecosystems of Mt.
SHI ET AL.
Kilimanjaro, Tanzania. Land Degradation & Development, 27, 592–602.
https://doi.org/10.1002/ldr.2406
Pandey, C. B., & Srivastava, R. C. (2009). Plant available phosphorus in
homegarden and native forest soils under high rainfall in an equatorial
humid tropics. Plant and Soil, 316, 71–80. https://doi.org/10.1007/
s11104‐008‐9759‐1
Pandey, D. N. (2002). Carbon sequestration in agroforestry systems. Cli-
mate Policy, 2, 367–377. https://doi.org/10.3763/cpol.2002.0240
Pausch, J., & Kuzyakov, Y. (2018). Carbon input by roots into the soil:
Quantification of rhizodeposition from root to ecosystem scale. Global
Change Biology, 24, 1–12. https://doi.org/10.1111/gcb.13850
Post, W. M., & Kwon, K. C. (2000). Soil carbon sequestration and land‐use
change: Processes and potential. Global Change Biology, 6, 317–327.
https://doi.org/10.1046/j.1365‐2486.2000.00308.x
Rillig, M. C. (2004). Arbuscular mycorrhizae, glomalin, and soil aggregation.
Canadian Journal of Soil Science, 84, 355–363. https://doi.org/
10.4141/S04‐003
Saha, K. S., Nair, P. K. R., Nair, D. V., & Kumar, M. B. (2009). Soil carbon
stock in relation to plant diversity of homegardens in Kerala, India.
Agroforestry Systems, 76, 53–65. https://doi.org/10.1007/s10457‐
009‐9228‐8
Schmitt‐Harsh, M., Evans, T. P., Castellanos, E., & Randolph, J. C. (2012).
Carbon stocks in coffee agroforests and mixed dry tropical forests in
the western highlands of Guatemala. Agroforestry Systems, 86,
141–157. https://doi.org/10.1007/s10457‐012‐9549‐x
Schneckenberger, K., & Kuzyakov, Y. (2007). Carbon sequestration under
Miscanthus in sandy and loamy soils estimated by natural 13C abun-
dance. Journal of Plant Nutrition and Soil Science, 170, 538–542.
https://doi.org/10.1002/jpln.200625111
Sharrow, S. H. (1999). Silvopastoralism: competition and facilitation
between trees, livestock and improved grass‐clover pastures on tem-
perate rainfed lands. In: L. E. Buck, J. P. Lassoie, and E. C. M.
Fernandes, (eds.) Agroforestry in Sustainable Agricultural Systems. USA:
CRC Press, Boca Raton, FL.
Singh, B., & Gill, R. I. S. (2014). Carbon sequestration and nutrient removal
by some tree species in an agrisilviculture system in Punjab, India.
Range Management and Agroforestry, 35, 107–114.
Sommer, J., Dippold, M. A., Flessa, H., & Kuzyakov, Y. (2016). Allocation
and dynamics of C and N within plant–soil system of ash and beech.
Journal of Plant Nutrition and Soil Science, 179, 376–387. https://doi.
org/10.1002/jpln.201500384
Stavi, I., & Lal, R. (2013). Agroforestry and biochar to offset climate change:
A review. Agronomy for Sustainable Development, 33, 81–96. https://
doi.org/10.1007/s13593‐012‐0081‐1
Takimoto, A., Nair, P. K. R., & Nair, V. D. (2008). Carbon stock and seques-
tration potential of traditional and improved agroforestry systems in
the West African Sahel. Agriculture Ecosystems and Environment, 125,
159–166. https://doi.org/10.1016/j.agee.2007.12.010
Takimoto, A., Nair, V. D., & Nair, P. K. R. (2009). Contribution of trees to
soil carbon sequestration under agroforestry systems in the West Afri-
can Sahel. Agroforestry Systems, 76, 11–25. https://doi.org/10.1007/
s10457‐008‐9179‐5
Thakur, S., Kumar, B. M., & Kunhamu, T. K. (2015). Coarse root biomass,
carbon, and nutrient stock dynamics of different stem and crown clas-
ses of silver oak (Grevillea robusta, A. Cunn. ex R. Br.) plantation in
central Kerala, India. Agroforestry Systems, 89, 869–883. https://doi.
org/10.1007/s10457‐015‐9821‐y
Torralba, M. (2016). Do European agroforestry systems enhance biodiver-
sity and ecosystem services? A meta‐analysis. Agriculture Ecosystems
and Environment, 230, 150–161. https://doi.org/10.1016/j.
agee.2016.06.002
Tsonkova, P., Böhm, C., Quinkenstein, A., Freese, D., & Jose, S. (2012). Eco-
logical benefits provided by alley cropping systems for production of
woody biomass in the temperate region: a review. Agroforestry Systems,
85, 133–152. https://doi.org/10.1007/s10457‐012‐9494‐8
SHI ET AL.
Udawatta, R. P., & Jose, S. (2011). Carbon sequestration potential of agro-
forestry practices in temperate North America. In Carbon sequestration
potential of agroforestry systems (pp. 17–42). Dordrecht: Springer.
Upson, M. A., & Burgess, P. J. (2013). Soil organic carbon and root distribu-
tion in a temperate arable agroforestry system. Plant and Soil, 373,
43–58. https://doi.org/10.1007/s11104‐013‐1733‐x
Upson, M. A., Burgess, P. J., & Morison, J. I. L. (2016). Soil carbon changes after
establishing woodland and agroforestry trees in a grazed pasture.
Geoderma, 283, 10–20. https://doi.org/10.1016/j.geoderma.2016.07.002
Vallejo, V. E., Roldan, F., & Dick, R. P. (2010). Soil enzymatic activities and
microbial biomass in an integrated agroforestry chronosequence com-
pared to monoculture and a native forest of Colombia. Biology and
Fertility of Soils, 46, 577–587. https://doi.org/10.1007/s00374‐010‐
0466‐8
Verchot, N. L., Noordwijk, V. M., Kandji, S., Tomich, T., Ong, C., Albrecht,
A., et al. (2007). Climate change: Linking adaptation and mitigation
through agroforestry. Mitigation and Adaptation Strategies for Global
Change, 12, 901–918. https://doi.org/10.1007/s11027‐007‐9105‐6
Viechtbauer, W. (2010). Conducting meta‐analyses in R with the metafor
package. Journal of Statistical Software, 36, 1–48. https://doi.org/
10.18637/jss.v036.i03
Wiesmeier, M., Lungu, M., Cerbari, V., Boincean, B., Hübner, R., & Kögel‐
Knabner, I. (2018). Rebuilding soil carbon in degraded steppe soils of
Eastern Europe: The importance of windbreaks and improved cropland
management. Land Degradation & Development, 29, 875–883. https://
doi.org/10.1002/ldr.2902
Zang, H., Blagodatskaya, E., Wen, Y., Xu, X., Dyckmans, J., & Kuzyakov, Y.
(2018). Carbon sequestration and turnover in soil under the energy
crop Miscanthus: Repeated 13C natural abundance approach and
3897
literature synthesis. Global Change Biology. Bioenergy, 10, 262–271.
https://doi.org/10.1111/gcbb.12485
Zang, H., Wang, J., & Kuzyakov, Y. (2016). N fertilization decreases soil
organic matter decomposition in the rhizosphere. Applied Soil Ecology,
108, 47–53. https://doi.org/10.1016/j.apsoil.2016.07.021
Zheng, X., Zhu, J., & Xing, Z. (2016). Assessment of the effects of shelter-
belts on crop yields at the regional scale in northeast China. Agricultural
Systems, 143, 49–60. https://doi.org/10.1016/j.agsy.2015.12.008
Zomer, R. J., Neufeldt, H., Xu, J. C., Ahrends, A., Bossio, D., Trabucco, A., …
Wang, M. C. (2016). Global tree cover and biomass carbon on agricul-
tural land: The contribution of agroforestry to global and national
carbon budgets. Scientific Reports, 6(29987). https://doi.org/10.1038/
srep29987
Zomer RJ, Trabucco A, Coe R, Place F. (2009). Trees on farm: Analysis of
global extent and geographical patterns of agroforestry. ICRAF Work-
ing Paper no 89. Nairobi, Kenya: World Agroforestry Centre.
SUPPORTI NG INFORMATION
Additional supporting information may be found online in the
Supporting Information section at the end of the article.
How to cite this article: Shi L, Feng W, Xu J, Kuzyakov Y.
Agroforestry systems: Meta‐analysis of soil carbon stocks,
sequestration processes, and future potentials. Land Degrad
Dev. 2018;29:3886–3897. https://doi.org/10.1002/ldr.3136