ArticlePDF Available

Which agroforestry options give the greatest soil and above ground carbon benefits in different world regions?

Authors:
  • Freelance Scientist

Abstract and Figures

Climate change mitigation and food security are two of the main challenges of human society. Agroforestry systems, defined as the presence of trees on external and internal boundaries, cropland, or on any other available niche of farmland, can provide both climate change mitigation and food. There are several types of agroforestry systems with different rates of above ground and soil carbon (C) sequestration. The amount of carbon sequestered can depend on the type of system, climate, time since land use change and previous land use. Data was collected from a total of 86 published and peer reviewed studies on soil and above ground carbon sequestration for different agroforestry systems, climates and regions in the world. The objective was to understand which agroforestry systems provide the greatest benefits, and what are the main factors influencing, soil and above ground carbon sequestration. The results show that, on average, more soil carbon sequestration occurs in agroforestry systems classified as silvopastoral (4.38 tC ha⁻¹ yr⁻¹), and more above ground carbon sequestration occurs in improved fallows (11.29 tC ha⁻¹ yr⁻¹). On average, carbon benefits are greater in agroforestry systems Tropical climates when compared to agroforestry systems located in other climates, both in terms of soil (2.23 tC ha⁻¹ yr⁻¹) and above ground (4.85 tC ha⁻¹ yr⁻¹). In terms of land use change, the greatest above ground carbon sequestration (12.8 tC ha⁻¹ yr⁻¹) occurs when degraded land is replaced by improved fallow and the greatest soil carbon sequestration (4.38 tC ha⁻¹ yr⁻¹) results from the transition of a grassland system to a silvopastoral system. Time since the change is implemented was the main factor influencing above ground carbon sequestration, while climate mainly influences soil carbon sequestration most. The results of the analysis may be used to inform practitioners and policy makers on the most effective agroforestry system for carbon sequestration. The lack of data on carbon stocks before the implementation land use change and the lack of reporting on soil sampling design and variances were the main limitations in the data. The need to report this data should be considered in future studies if agroforestry systems are expected to play an important role as a climate change mitigation strategy.
Content may be subject to copyright.
1
Which agroforestry options give the greatest soil and above ground carbon benefits in 1
different world regions? 2
3
Diana Feliciano1, Alicia Ledo, Jon Hillier, Dali Rani Nayak 4
1Corresponding author. E-mail: diana.feliciano@abdn.ac.uk; 5
Institute of Biological and Environmental Sciences, School of Biological Sciences, University of Aberdeen, 23 6
St Machar Drive, Aberdeen, AB24 3UU, Scotland, UK 7
8
Abstract 9
Climate change mitigation and food security are two of the main challenges of human 10
society. Agroforestry systems, defined as the presence of trees on external and internal 11
boundaries, cropland or on any other available niche of farmland, can provide both climate 12
change mitigation and food. There are several types of agroforestry systems with different 13
rates of above ground and soil carbon (C) sequestration. The amount of carbon sequestered 14
can depend on the type of system, climate, time since land use change and land use type. 15
Data was collected from a total of 86 published and peer reviewed studies on soil and above 16
ground carbon sequestration for different agroforestry systems, climates and regions in the 17
world. The objective was to understand which agroforestry systems provide the greatest 18
benefits, and what are the main factors influencing, soil and above ground carbon 19
sequestration. The results show that, on average, greater soil carbon sequestration occurs in 20
agroforestry systems classified as silvopastoral (4.38tCha-1yr-1) and that greater above ground 21
carbon sequestration occurs improved fallows (11.29tCha-1yr-1). On average, carbon benefits 22
are greater, both in terms of soil (2.23tCha-1yr-1) and above ground (4.85tCha-1yr-1) carbon 23
sequestration. In addition, greater benefits occur when the land use change is from degraded 24
to improved fallow in the case of above ground carbon sequestration (12.8 tCha-1yr-1), and 25
from grassland to silvopastoral in the case of soil carbon sequestration (4.38 tCha-1yr-1). Time 26
since the change is implemented was the main factor influencing above ground carbon 27
sequestration while climate influences soil carbon sequestration, regardless of time. The 28
results of the analysis can inform practitioners and policy makers about the most effective 29
agroforestry system for carbon sequestration. The lack of data on carbon stocks before land 30
use change and lack of reporting on soil sampling design and variances were the main 31
limitations in the data collected from the literature. The need to report this data should be 32
considered in future studies if agroforestry systems are expected to play an important role as 33
a climate change mitigation strategy. 34
2
35
Keywords: climate change, meta-analysis, mitigation, trees, agriculture 36
37
38 39 1. Introduction 40 41 The relationship between land management and climate change has previously been 42
identified across some of the key global agricultural systems (FAO, 2011a). The rural land 43
use sector (forest, moorland, peatland, agriculture) has the unique capacity of delivering zero 44
and negative carbon emissions since it can act as a sink and reservoir for carbon dioxide 45
(Feliciano et al., 2013). Mitigation of climate change through increased carbon sequestration 46
in the soil can be particularly useful when addressed in combination with other challenges 47
that affect people’s livelihoods, such as reverting land degradation and ensuring food security 48
(Batjes, 2003). Potential increases in carbon sequestration may occur in agricultural and 49
forest lands via improved land use management, conversion to land use with higher carbon 50
stocks, or increased carbon storage in harvested products (IPCC, 2000). Agroforestry systems 51
are an option to mitigate climate change while promoting an increase in crop yields as well as 52
other positive environmental outcomes (Tubiello et al., 2008, Smith et al., 2013, Mbow et al., 53
2014). In these systems, woody perennials (trees, shrubs, palms, bamboos, etc.) are cultivated 54
in the same land-management unit with crops and/or animals, in some form of a spatial 55
arrangement or a temporal sequence (Nair, 1993, Montero et al. 1998, Joffre et al. 1999). The 56
diversification of the farm system into an agroforestry system can increase agricultural 57
productivity, improve soil fertility, control erosion, conserve biodiversity, and diversify 58
income for households and communities (Bishaw et al., 2013). According to Sanchez (1997), 59
about 20% of the world’s population depend directly on agroforestry products and services in 60
rural and urban areas of developing countries. Agroforestry systems are more common in the 61
temperate, sub-tropical and tropical zones, and include a wide range of land uses and 62
practices (Torquebiau, 2000; Nair, 1985). In the tropics agroforestry systems are especially 63
practised by smallholder farmers (Lorenz and Lal, 2014). Dixon (1995) estimated that in 64
tropical latitudes one hectare (ha) of sustainable agroforestry can provide goods and services 65
which potentially offset 5–20 ha of deforestation. According to FAO (2011a), there are five 66
main forms of agroforestry, namely alley cropping, forest farming, silvopastoralism, riparian 67
forest buffers and windbreaks. These integrate technologies such as contour farming, 68
multistorey cropping, intercropping, multiple cropping, bush and tree fallows, parkland, or 69
3
homegardens. Other authors (e.g. Schoeneberger, 2009; Kandji et al., 2006) have also 70
considered other nomenclatures, including agrisilvicultural systems, woodlots, boundary 71
planting, lives fences or multistrata agroforests. Agroforestry systems have received 72
increased attention because of their capacity to sequester carbon dioxide from the atmosphere 73
in above ground biomass, i.e., stems, branches and foliage, and in below ground biomass, i.e. 74
roots, and in the soil (Oke and Olatiilu, 2011; Nair 2012; Lorenz and Lal, 2014). This fact can 75
represent an economic opportunity for subsistence farmers in developing countries to sell 76
carbon sequestered through agroforestry activities to industrialised countries under the Kyoto 77
Protocol Clean Development Mechanism (Nair et al. 2009). Currently, payments for carbon 78
sequestration are limited to voluntary carbon markets, but it is expected that emerging 79
domestic legislation in several developed countries may soon increase the demand for 80
emission reductions from land management activities (Lipper et al., 2010). According to Nair 81
et al. (2009) there is currently an area of 1,023 Mha under agroforestry worldwide, with a 82
carbon sequestration potential of 1.9 Pg C over 50 years. There is, therefore, enormous 83
potential to further expand the area of agroforestry systems and consequently, to sequester 84
more carbon (FAO, 2011b). As agroforestry options become more popular, it is crucial to 85
quantify the potential of these systems for carbon sequestration and climate change 86
mitigation. Although the IPCC Guidelines for National Greenhouse Gas Inventories (IPCC, 87
2006) provide carbon, they only cover potential carbon storage for agrosilvicultural and 88
silvopastoral systems in dry lowlands and humid tropical regions (Dixon et al., 1995). There 89
is a well-established body of literature on the potential of agroforestry systems for carbon 90
sequestration, indicating a good understanding of the potential of different agroforestry 91
systems to sequester carbon. However, this literature is very heterogeneous in purpose and 92
approaches and the potential benefits have not been systematically compared, so to date there 93
is only a fragmented view of carbon sequestration potentials around the world. Coarse 94
estimates of the potential for above ground and soil carbon sequestration in agroforestry 95
systems lack the refinement to enable the differential effects of important practice-related, or 96
other contextual variables, on the sequestration potential to be predicted. Jose and Bardhan 97
(2012) consider that if agroforestry is to be used in carbon sequestration schemes such as the 98
clean development mechanisms, better information is required about above and below ground 99
carbon stocks and soil carbon in areas under agroforestry systems. There is, thus, a need for 100
estimates which are sensitive to specific region, climate and practice factors for agroforestry 101
systems. Farmers, project planners and project managers must be able to assess the likely 102
4
benefit, including carbon sequestration potential, of agroforestry systems according to site 103
circumstances. 104
105
The aim of this study is to provide an understanding of soil and above ground carbon 106
sequestration in agroforestry systems with respect to region, climate, agroforestry type and 107
time of implementation. This study also aims at understanding the main factors influencing 108
soil and above ground carbon sequestration. This will enable identification of the most 109
effective agroforestry systems in terms of carbon sequestration for different regions and 110
climates and support policy makers and project planners in the design of policy measures and 111
carbon offset schemes that consider agroforestry systems as climate change mitigation option. 112
More specifically, the objectives of this study are: 113
114
1) Identifying agroforestry systems implemented in different regions of the globe; 115
2) Understanding the capacity of different agroforestry systems to sequester carbon in soils 116
and above ground biomass and; 117
3) Providing a better understanding of the factors influencing carbon sequestration in 118
agroforestry systems. 119
120
121
2. Carbon sequestration in agroforestry systems: review of studies 122
Agroforestry systems can increase carbon stocks in soil, and below ground and above ground 123
tree components, contributing to greenhouse gas (GHG) emission mitigation (Mutuo et al., 124
2005). The research on the potential of agroforestry systems to sequester carbon dioxide 125
focus mainly on country or region studies (e.g. Parrota, 1999, Swamy and Puri, 2005; 126
Oelbermann et al., 2006; Kirby and Potvin, 2007; Takimoto et al., 2008; Roshetko et al, 127
2002; Gama-Rodrigues et al., 2011, Schroth et al., 2013). There are also studies that use 128
models to estimate carbon sequestration potentials (e.g. Grogan & Matthews, 2002; Lal, 129
2005; Negash & Kanninen, 2015; Luedeling & Neufeldt, 2012) and a number of existing 130
reviews (e.g. Dixon et al., 1994; Albrecht & Kandji, 2003; Jarecki and Lal, 2003; Mutuo et 131
al., 2009; Jose, 2009; Nair et al., 2009; Somarriba et al., 2013; Lorenz and Lal, 2014). In 132
addition to this, Kumar and Nair (2011) compiled the state-of-the art of research results and 133
evaluations relating to different aspects of agroforestry. Some of the studies focus on soil 134
carbon sequestration (e.g. Oelbermann et al, 2006; Mutuo et al., 2005) while others cover 135
5
above ground and soil carbon sequestration (e.g. Lal, 2002; Kirby and Potvin, 2007). The 136
review of this literature revealed a wide nomenclature for agroforestry systems, namely 137
rotational woodlot systems, parklands with trees, cropland and trees, homegardens, fallows 138
with indigenous trees or fruit trees, species-rich systems, intercropping of corn and tree 139
species, grassland and trees, cocoa and coffee shadow systems, multilayer cocoa, homesteads, 140
fast growing trees, live fences, strip planting, improved fallows, shelterbelts, fodderbank and 141
more. Each of these systems may differ in terms of plant characteristics (tree species, age, 142
crops, biodiversity and tree density), system characteristics (structure, function, stability) and 143
management factors (tillage, fertilization, residues, holding size, harvesting regime). These 144
characteristics together with the agroecological conditions (altitude, climate, wind) and soil 145
characteristics (structure, texture, fertility and physical, chemical and biological conditions) 146
where the different agroforestry systems are implemented, influence soil and above ground 147
carbon sequestration (Nair et al., 2009; Jose and Bardhan, 2012). For example, the potential 148
of agroforestry systems to sequester carbon in the above ground components is lower in arid, 149
semiarid and temperate regions than in tropical regions (Nair et al., 2009). In degraded sites, 150
carbon sequestration potential is also lower than in fertile humid sites (FAO, 2001). Several 151
individual field experiments have been conducted to test the effectiveness of agroforestry 152
practices in above ground and soil carbon sequestration in different world regions. For 153
example, Sharrow and Ismail (2004) studied soil organic carbon storage in agroforests of 154
Douglas-fir and grassland in western Oregon, North America. Kimaro et al. (2007, 2008) 155
reported above ground carbon storage and accumulation in wood of different tree species 156
under rotational woodlot systems in Tanzania, Africa. Brakas & Aune (2011) assessed carbon 157
accumulation in the smallholder farming systems under nineteen different land use types in 158
the Philippines, Asia. Somarriba et al. (2013) presented carbon stored in cocoa-based 159
agroforestry systems in several countries of Central America. Howlett et al. (2011) reported 160
soil organic carbon storage in silvopastural systems with birch in Nortwestern Spain 161
(Europe). More recently, Kim et al. (2016) compiled 56 individual studies on net rates of 162
change of biomass and soil carbon stocks in agroforestry systems and found a high variability 163
in net carbon sequestration rates in both biomass and soils depending on the type of 164
agroforestry. These authors reported carbon increments ranging from 0.3 to 7.7 t C ha-1 yr-1 in 165
biomass and 1.0 to 7.4 t C ha-1 yr-1 in soils. Although many individual studies have been 166
conducted and a few have synthesised the available data, none have been published about the 167
differences in carbon sequestration according to the type of agroforestry system, different 168
world regions, climates, previous land uses and age of the agroforestry systems, and about the 169
6
influence these factors have on carbon sequestration. There is, therefore, scope for an in-170
depth analysis of the literature in order to provide a comprehensive assessment of the 171
contribution of agroforestry systems to soil carbon and above carbon sequestration. 172
173
3. Methods 174
175
3.1 Data collection 176
The systematic search identified 40 peer reviewed studies that reported above ground carbon 177
sequestration in agroforestry systems (date range 1984-2015) and 46 peer-reviewed studies 178
that reported soil carbon sequestration in agroforestry systems (date range 1995-2015) 179
(Supplementary material). This provided two independent databases, one for above ground 180
carbon sequestration and another one for soil carbon sequestration. Peer reviewed studies 181
were selected through a search in the ISI-Web of Knowledge, Google Scholar and Scopus. 182
The searches were performed using several words related to agroforestry systems and carbon 183
sequestration, more specifically agroforestry*OR land management practices*OR carbon 184
sequestration*OR soil carbon sequestration*OR climate change*OR mitigation*OR above 185
ground carbon sequestration (or the same terms in Spanish or Portuguese). The terms were 186
used separately or in combination with each other. Both review articles and original studies 187
were considered. The reference lists of the published reviews on the topic were also searched 188
for eligible studies, i.e., snowballing. The articles with relevant titles were retrieved and the 189
abstracts read and the studies selected for further reading and analysis were those that 190
reported on: 191
192
1) Above ground carbon sequestration per year (MgCha-1yr-1) or total carbon storage per 193
hectare (MgCha-1) before the agroforestry system being implemented and after the 194
agroforestry system being implemented. Above ground carbon sequestration covers 195
carbon sequestered in tree above ground biomass; 196
2) Soil carbon sequestration per year (MgC.ha-1.yr-1) or total carbon storage per hectare 197
(MgC.ha-1) before and after implementation of the agroforestry system. Soil carbon 198
sequestration covers soil carbon only and not tree roots; 199
3) Land use before and after the implementation of the agroforestry system; 200
4) Time since implementation of the agroforestry system (age of the agroforestry system 201
in number of years); 202
7
5) Climate; 203
6) Country. 204
205
3.2 Data treatment 206
Depending on how data was reported, some adjustments had to be made. Therefore, if only 207
total carbon storage per hectare (MgCha-1) in soil or above ground was reported before and 208
after implementation of the agroforestry system, both values were divided by the number of 209
years since implementation in order to estimate carbon sequestration rates in MgCha-1yr-1. 210
Information on land-use before the implementation of the agroforestry system was often 211
reported. For few cases, land use before was inferred from careful reading of the section 212
about the characteristics of the study site. Whenever this information was not reported or 213
could not be inferred from the study, the information was recorded in the database (for above 214
ground and soil carbon) as “Not reported”. In relation to climate information, the Met Office 215
climate guide was used to reclassify the climate region reported by the studies into Arid, 216
Mediterranean, Polar, Semiarid, Temperate and Tropical classes. 217
218
For soil carbon data, an extra step was required because this was usually reported for 219
different soil depths, and often the upper and lower positions of the depth intervals did not 220
match across studies. In order to allow a standardised analysis compatible with the IPCC 221
guidelines the quadratic density function (Equation 1) based on Smith et al. (2000) was used 222
to derive a scaling cumulative distribution functions (c.d.f.) for soil density as a function of 223
soil depth in (metres) up to 1 m as follows: 224
225
cdf(d)=((22.1-(33.3d2)/2+(14.9d3)/3))⁄10.41667 (1) 226
227
And equation 2 allowed soil carbon at a given depth d to be scaled to the equivalent values at 228
0.30 m as follows: 229
230
SOC(0.3m)=SOC(d)×(cdf(0.3))/(cdf(d)) (2) 231
232
The 16 agroforestry systems reported by the literature were reclassified in 7 agroforestry 233
systems, namely agrisilvicultural, boundary planting, homegarden, improved fallow, shadow 234
systems, silvopastoral, woodlots (Table 1). 235
236
8
[Table 1 here] 237
238
Other factors such plant characteristics (type, intensity), soil characteristics (type, texture, 239
classification, pH), agro-ecological characteristics (temperature, rainfall, latitude), system 240
characteristics (e.g. tree density) and management factors (e.g. fertiliser application) were 241
also extracted since these influence soil carbon and above ground carbon. 242
243
Studies only reported on tree density very occasionally, and therefore, this variable was not 244
included in the soil or above ground carbon databases and consequently in the analysis. Soil 245
(pH, type, classification) and management characteristics (fertiliser application, fertiliser 246
type) were infrequently reported and, similarly, excluded from the analysis. Other variables 247
such as the authorship of the study and study year were also included in the databases (soil 248
and above ground) since there were several observations from the same study, and the effect 249
this has on soil or above carbon can be controlled, as recommended by Nelson and Kennedy 250
(2009). 251
252
Data compiled in the database therefore included information on the following variables: 253
world region, climate, country, agroforestry system, land-use before implementation of the 254
agroforestry system, time since implementation (in years), tree density, and above ground and 255
soil carbon sequestration (Supplementary material). Table 2 include a description of each of 256
the variables included in the database and how they were coded for the data analysis. The 257
number of observations (n) is presented in brackets. 258
259
[Table 2 here] 260
261
Absolute above ground and soil carbon sequestration (AGCS and SCS, respectively) effects 262
were estimated by calculating the difference between carbon sequestration in the current 263
agroforestry system (AFcurrent) and carbon sequestration in the land use before (LUBefore) 264
the agroforestry system being implemented (Equations 3 and 4). 265
266
AGCS (tC.ha-1.yr-1)=AGCS_AFcurrentAGCS_LUBefore (3) 267
268
SCS (tC.ha-1.yr-1) = SCS_AFcurrentSCS_LUBefore (4) 269
270
9
3.3 Data analysis 271
272
3.3.1 Descriptive meta-analysis 273
274
The mean above ground and soil carbon sequestration was calculated for each agroforestry 275
system in the six World regions identified, namely, Africa, Asia, Australia, Europe, Latin 276
America, North America. In addition, the mean, minimum and maximum above ground and 277
soil carbon sequestration were calculated for each agroforestry system, climate and also for 278
agroforestry systems in tropical climates. A more in depth analysis was undertaken for 279
agroforestry systems implemented in tropical climates because it was for this climate that soil 280
and above ground carbon was mostly reported. The calculations were undertaken with IBM 281
SPSS Statistics 24. 282
283
In order to better understand how soil and above ground carbon sequestration changed 284
according to the type of land use change, two forest plots were created in using Microsoft 285
Excel 2010. Similarly, in order to better understand how soil and above ground carbon 286
changed with respect to time since implementation, two graphs were created using the R 287
statistical platform (R-Development-Core-Team 2008). 288
289
3.3.2 Statistical modelling to quantify the drivers of above ground and soil carbon 290
sequestration in in agroforestry systems 291
292
In order to identify the factors acting on AGCS and SCS, a set of General Linear Mixed 293
Models (GLMM) were fitted. The potential covariates tested were: the effect of time after 294
conversion, climate, continent, land use before change, current agroforestry system and land 295
use change (type of change). Details can be found in Table 2. In these analyses, fallow”, 296
degraded landand underutilised landwere merged in to a single category in the “land use 297
beforesection. A first set of GLMMs was fitted for AGCS and a second set was fitted for 298
SCS, including different combinations of the available covariates. A stepwise modelling 299
procedure was followed, i.e., a series of models which included each individual variable 300
separately were first fitted. Then, several models including systematically different 301
combinations of those covariates were fitted. The models which included correlated 302
covariates were discarded. The current agroforestry system was initially included as a fixed 303
effect in the models, but it was later included as a random effect in order to account for the 304
10
intrinsic differences of the different agroforestry systems. Therefore, the fitted models had the 305
following structure: 306
307
Carbon sequestration =+
 + + (5) 308
309
where was the intercept, were the coefficients of the explanatory variables (fixed effects, 310
F),  was the random intercept that varied with the covariate included as a random effect, 311
and was the error term, assumed to follow a normal distribution (0, ). The restricted 312
maximum likelihood (REML) was used to fit the models to the data. For model diagnosis, 313
absence of patterns and deviance from a central tendency in the residuals were checked via 314
visual exploration of the residual versus fitted values plot. In addition, the normality of the 315
residuals was tested by visual inspection of the Q-Q plot and the efficiency of the GLMM 316
fitted models was checked and compared with nested models using the Akaike Information 317
Criterion (AIC). The best model identified was the one including all the covariates that were 318
significant at the 97.5 % level and had the highest R2. The best models for AGCS and SCS 319
were named AGCS model and SCS model, respectively. In order to quantify the effect of the 320
selected covariates on the AGCS and the SCS responses, the variance of the dataset explained 321
by the AGCS and the SCS models was calculated. In addition, to assess the importance of 322
one covariate over the others, the variance explained by each of the fixed and random effects 323
was also calculated. 324
325
To this end, the conditional and marginal R2GLMM values were calculated using the method 326
proposed by Nakagawa and Schielzeth (2013). The marginal R2 (
) provides the variance 327
explained by the fixed factors and the conditional R2 (
) the variance explained by the entire 328
model, including both fixed and random effects (Vonesh et al. 1996). 329
330
=



and
=




(6) 331
332
where
is the variance of the fixed effects components,
is the variance of the random 333
effect (the current agroforestry system),
is the additive dispersion and
the is the 334
distribution specific variance (Nakagawa and Schielzeth 2013). In order to evaluate the 335
11
contribution of each fixed effect to the explained variance, the marginal
values including 336
sequentially the variance of each fixed component was calculated: 337
=



(7) 338
where
is the variance of each fixed effect. 339
340
The lmer function from the “lme4” package (Bates et al. 2015) in the R statistical platform 341
(R-Development-Core-Team 2008) was used to fit the models and to obtain the values of the 342
explained variance. The R code is provided in the supplementary material. 343
344
4. Results 345
346
4.1 Above ground and soil carbon sequestration in different agroforestry systems and 347
regions 348
349
When only studies that report above ground carbon sequestration are considered Africa, Asia 350
and Latin America are the regions with the greatest diversity of agroforestry systems (Table 351
3). The most common systems in Africa are improved fallows (n=17) and shadow systems 352
(n=18), in Asia, the most common systems are homegardens (n=27) and woodlots (n=10), 353
and in Latin America the most common systems are shadow systems (n=22). In Africa, the 354
highest mean above ground carbon sequestration rate occurs in improved systems (mean= 355
12.95). In both Asia and Latin America the highest mean above ground carbon sequestration 356
rate occurs in woodlots (mean= 6.28 and 12.63, respectively). 357
358
[Table 3 here] 359
360
When only studies that report soil carbon sequestration are considered, Africa and Latin 361
America are the regions with the greatest diversity of agroforestry systems (Table 4). The 362
most common systems in Africa are agrisilvicultural systems (n=25) and improved fallows 363
(n=17) and in Latin America the most common systems are shadow systems (n=22), followed 364
by agrisilvicultural systems (n=13). In Africa, the highest mean soil carbon sequestration rate 365
occurs in improved fallow and shadow systems (mean= 1.91) and in Latin America the 366
highest mean soil carbon sequestration rate occurs in silvopastoral systems (mean= 6.54). 367
368
12
[Table 4 here] 369
370
4.2 Mean above ground and soil carbon sequestration according to agroforestry system 371
and climate 372
Figure 1A shows that the most common agroforestry systems found in the literature on above 373
ground carbon sequestration were shadow systems (n=40) and woodlots (n=34). Figure 1B 374
shows that the most common agroforestry systems found in the literature on soil carbon 375
sequestration were agrisilvicultural systems (n=52) and shadow systems (n=24). For above 376
ground carbon sequestration, the systems with the higher mean rates of above ground carbon 377
sequestration in tCha-1yr-1 are improved fallows, boundary planting and woodlots (Figure 1A) 378
and for soil carbon sequestration the systems with the higher mean rates of soil carbon 379
sequestration in tC.ha-1.yr-1 are homegardens and silvopastural systems (Figure 1B). Most 380
studies report on agroforestry systems implemented in Tropical climates (n=133 for above 381
ground carbon and n=88 for soil carbon, Figures 2A and 2B). It is also in Tropical climates 382
where above ground and soil carbon sequestration rates are higher, on average. A more 383
detailed analysis of agroforestry systems in Tropical climates indicates that mean above 384
ground carbon sequestration is higher in improved fallows and boundary planting (Figure 3A) 385
and that mean soil carbon sequestration is higher in silvopastoral systems and homegardens 386
(Figure 3B). 387
388
[Figures 1A and 1B here] 389
[Figures 2A and 2B here] 390
[Figures 3A and 3B here] 391
13
392
4.3 Above ground and soil carbon sequestration according to land use change and time 393
since the agroforestry system has been implemented 394
395
Mean above ground carbon sequestration is higher when the land use changes from degraded 396
land to improved fallows (12.8 tC.ha-1.yr-1, n=14), cropland to improved fallows (9.4 tC.ha-
397
1.yr-1, n=5) and grassland to woodlots (8.3 tC.ha-1.yr-1, n=17) (Figure 4). In all these cases, the 398
standard deviations are 5.41, 5.19, 5.23, 5.48 tC.ha-1.yr-1, respectively, indicating that the data 399
points are spread out over a wider range of values. In general, the land use change from a 400
non-agroforestry system to an agroforestry system shows a positive sequestration potential. 401
402
[Figure 4 here] 403
404
Absolute change in soil carbon sequestration is higher when the land use changes from 405
grassland to silvopastoral (4.48 tC.ha-1.yr-1, n=9), underutilised land to homegarden (3.83 406
tC.ha-1.yr-1, n=19) and cropland to improved fallow (1.9 t C.ha-1.yr-1, n=17) (Figure 5). In 407
these cases, the standard deviation is 0.86, 1.54, 1.85 t C.ha-1.yr-1, respectively. There are also 408
negative absolute changes in soil organic carbon under transition from cropland to woodlot (-409
0.5 tC.ha-1.yr-1, n=10), and grassland to agrisilvicultural (-0.8 t C.ha-1.yr-1, n=5), with 410
standard deviations of 2.57 and 0.98 t C.ha-1.yr-1, respectively. This might be due to initial 411
soil disturbance caused by the plantation of trees in non-tree systems such as cropland and 412
grassland. 413
414
[Figure 5 here] 415
416
After transition to an agroforestry system, above ground carbon sequestration varies with 417
time since implementation, this coinciding with the tree growing cycle (establishment phase, 418
initial phase, full vigour phase) (Figure 6). In the first years there is high carbon 419
sequestration, probably due to fast tree growth during the establishment phase (>5 years after 420
planting) and the initial phase (between 5 and 10 years) (Figure 6). During the full vigour 421
phase (>10 years), tree growth slows down and the carbon sequestration rate is sustained for 422
a number of years. At the end of the full vigour phase carbon sequestration starts to decline 423
probably due to tree aging, respiration and death and because the trees approach maturity 424
(when carbon sequestration and carbon losses due to respiration are equal). The duration and 425
14
magnitude of tree phases (establishment, initial phase and full-vigour phase) is determined to 426
a large extent by the combination of tree species, site characteristics (e.g. nutrient 427
availability), climatic conditions and management. 428
429
[Figure 6 here] 430
431
Soil carbon decreases immediately after the transition to an agroforestry system, then 432
increases after this initial period (5-10 years), and then it decreases again (after ~15years) 433
(Figure 7). 434
435
[Figure 7 here] 436
437 438
4.4 Drivers of above ground and soil carbon sequestration in agroforestry systems 439
440
In the AGCS model the significant fixed covariates were the time since implementation of the 441
agroforestry system (Time) and the land use before (LUBefore), with the current agroforestry 442
system (AFcurrent) included as a random effect in the model (Equation 9). 443
444
 = 7.31 0.089  +  + || (9) 445
446
The coefficients () for the different land uses before the implementation of the 447
agroforestry system () are (Table 5): 448
[Table 5 here] 449
This model explained 46% of the variance of the data (R2GLMM-C values) and it can be used to 450
predict changes in AGCS as a function of these variables, and using the factors in Table 5. 451
Details of the model fitting and validation are presented in supplementary material. The time 452
since the agroforestry system has been implemented was the main variable influencing 453
AGCS, followed by the current agroforestry systems (Figure 8). 454
[Figure 8 here] 455
15
These results show that the rate of carbon sequestration decreases with time, i.e., agroforestry 456
systems sequester more carbon immediately after land use change. This also agrees with 457
Figure 6 above. In addition, a change to an agroforestry system always results in an increase 458
carbon in above ground biomass (Equation 9). This increment is larger if the previous land 459
use is “degraded” or grassland, not so large but still positive if the previous land use was 460
“cropland”. Values of AGCS will afterward depend on the current agroforestry system 461
(AFcurrent). 462
In the SCS model the significant fixed covariates were the climate (Climate) and the land use 463
before (LUBefore), with the current agroforestry system included as a random effect in the 464
model (Equation 10): 465
466
 = 1.31 +
 +
 + || (10) 467
468
The coefficients (Coef) for climate (Climate) and for the different land uses before the 469
implementation of the agroforestry system (LUBefore) are presented in Table 6: 470
471
[Table 6 here] 472
473 This model explained 44% of the variance of the data (R2GLMM-C values) and it can be used to 474
predict changes in SCS as a function of these variables, and using the factors in Table 6. 475
Details of model fitting and validation are presented in supplementary material. The main 476
factors influencing SCS was the land use before (LUBefore) the agroforestry system has been 477
implemented, followed by climate (Climate). The current agroforestry system (AFcurrent) 478
was less important in the SCS model (Figure 9). 479
[Figure 9 here] 480
481
In this case, not all land use changes resulted in positive carbon sequestration as the 482
conversion of grassland to an agroforestry system was negative. In relation to the effect of 483
climate, polar systems result in less SCS. 484
485
5. Discussion 486
487
16
5.1 Potential of agroforestry systems to capture and store carbon 488
489
This study provides a review of literature about carbon sequestration in agroforestry systems 490
in different world regions. It is observed that a change from a non-agroforestry to an 491
agroforestry system usually leads to an increment of carbon in the system. This change is 492
very clear for above ground carbon, with a notable increment of carbon after implementation 493
followed by a constant carbon accumulation through the time. Results for soil carbon 494
sequestration were not so consistent, even though a positive increment in carbon was 495
observed in most cases (Figure 5). This might be due to the fact that the variables climate and 496
current agroforestry systems had a higher effect for soil carbon sequestration than for above 497
ground carbon sequestration. Therefore, the results for soil carbon sequestration values in 498
different agroforestry systems should be used with some caution. Large differences in soil 499
carbon sequestration (SCS) values among the land-use systems can happen due to biophysical 500
and socio-economic characteristic of the system and/or methodological issues (Stockmann et 501
al., 2013). Section 5.3 discusses more in depth the limitations of the study related to 502
methodological problems. 503
504
The results for soil carbon sequestration according to time (Figure 7) differ from Smith et al. 505
(1997a, b) who observed greatest increases in soil carbon soon after the implementation of a 506
land use or land management change. Kim et al. (2016) pointed out that soil carbon may 507
increase during the tree-growing phase, but while crops are cultivated after tree harvesting or 508
burning, soil carbon stocks are likely to decrease again. This justification coincides with the 509
reduced carbon sequestration after 15 years observed in Figure 7. Soil carbon sequestration 510
has limited potential to sequester carbon (sink saturation) when the annual benefits reach a 511
new quasi-equilibrium (Freibauer et al., 2004; Stockmann et al., 2013). In order to account 512
for this, the IPCC recommends at least a 20 year period for soil carbon sequestration 513
accounting in national greenhouse gas inventories (IPCC, 1997). 514
515
The results showed that for above ground carbon sequestration, a change from a non-516
agroforestry to an agroforestry system was always positive (Figure 4). The greatest benefits 517
were achieved when changing from degraded and cropland to improved fallows. For soil 518
carbon sequestration, the greatest benefit resulted from changing grassland to silvopastoral 519
(Figure 5). In order to predict above ground carbon and soil carbon sequestration potential of 520
different agroforestry systems in different world climates general linear mixed models 521
17
(GLMM) were developed and two equations were derived (for soil and above ground 522
carbon). These equations can be used both globally and regionally, with some confidence, 523
whenever data on carbon sequestration is sparse. Therefore, this study provides a screening 524
tool that can be used as the first method to investigate carbon sequestration benefits resulting 525
from the implementation of agroforestry systems. 526
527
5.2 Agroforestry systems as climate change mitigation strategy? 528
529
Apart from soil and above ground carbon sequestration, which mostly depend upon the type 530
of agroforestry system implemented, there are other opportunities to mitigate GHG 531
emissions. Table 10 presents some of these practices and their influence on GHG emissions. 532
In the cases of cocoa or coffee production, for example, the implementation of an 533
agroforestry system might mean the loss of a primary forest which would entails a drastic 534
reduction of forest carbon. In Indonesia, the transformation of primary forests into cocoa 535
agroforestry has decreased forest carbon (Stephan-Dewenter et al., 2007; Smiley and 536
Kroschel, 2008). It seems likely that it is better to have cocoa with shade trees when there is 537
only cocoa in the system, but that it is worse to have cocoa and trees when a primary forest is 538
felled to implement the cocoa agroforestry system. In addition, in most tropical forest 539
systems a few large trees contain over 80% of the carbon stored in the landscape. If these 540
trees are selectively kept in the system, most of the carbon can be preserved while 541
introducing a managed system. 542
543
An important objective of this meta-analysis was to determine whether agroforestry systems 544
elicited quantitatively different responses in soil carbon stocks. The analysis, although to 545
some extent providing empirical support for the 20 year accounting period for soil carbon 546
contained in the 2006 IPCC guidelines, offers some refinement. The IPCC factors do not 547
allow comparison between non-agroforestry and agroforestry systems. Albrecht & Kandji 548
(2003), who provide the factors used by the IPCC, collected the potential carbon 549
sequestration for agroforestry systems in different ecoregions of the world but did not 550
differentiate soil and above ground carbon when an agroforestry system is implemented 551
(Table 2 in Albrecht & Kandji, 2003). It is important to report carbon stocks, but to report 552
carbon sequestration changes can be even more relevant if agroforestry is to be considered as 553
a climate change mitigation option (e.g. in the Clean Development Mechanism). This study 554
provides a review of literature about carbon sequestration in agroforestry systems in the 555
18
world, and a framework for deriving the mitigation potential of different agroforestry systems 556
in a systematic way to use in situations when data about carbon stock changes are not 557
available. This information would help practitioners to choose the best agroforestry system to 558
implement for carbon sequestration purposes, and where to implement it. 559
560
However, the analysis provides information only on absolute change in soil and above ground 561
carbon sequestration, and not on other GHG emissions resulting from the implementation of 562
these systems (e.g. nitrous oxide, methane). According to Tsuruta et al. (2000), in Sumatra, a 563
jungle rubber system had lower nitrous oxide (N2O) emissions than a primary forest, but also 564
lower (CH4) uptake. And a multi-storey coffee with a leguminous tree shade canopy had N2O 565
emissions five times higher than open-grown coffee and about half the CH4 uptake. Practices 566
such as shifting cultivation, pasture maintenance by burning, nitrogen fertilisation and animal 567
production usually increase GHG emissions (Dixon, 1995; Mbow et al., 2014). Kim et al. 568
(2016) analysed carbon sequestration and net emissions of CH4 and N2O under agroforestry 569
and estimated a mitigation potential of 27±14t CO2 equivalents ha-1 y-1 for the first 14 years 570
after establishment. 571
572
Therefore, agroforestry systems are important sinks and reservoirs of carbon but a full GHG 573
emission accounting should be undertaken before agroforestry is to be considered a climate 574
mitigation option. But carbon sequestered in agroforestry systems could be sold in carbon 575
credit markets where such opportunities exist (Jose, 2009). In addition to the potential of 576
agroforestry system to accumulate and sequester carbon, these systems could evolve into a 577
technological alternative for reducing deforestation rates in tropical zones (Murthy, et al., 578
2013). Agroforestry systems might bring also important benefits in terms of climate change 579
adaptation. The incorporation of trees and crops are usually able to biologically fix nitrogen 580
in tropical climates and these systems can enhance the physical, chemical and biological 581
properties of the soil by adding significant organic matter, contributing to the release and 582
recycle of nutrients (Jose, 2009; Pandey, 2007). Agroforestry increases the resilience of 583
farming systems by safeguarding farmers against biophysical risks by improving hydraulic 584
lift and soil fertility and financial risks through income and livelihood diversification 585
(Verchot et al., 2007). Barriers for implementation of agroforestry systems should be 586
assessed and analysed before advising on the implementation of these systems. It should also 587
be recognised that optimum configuration of agroforestry systems for maximum carbon 588
sequestration might not be achieved since these are complex systems that provide a multitude 589
19
of products and services that communities depend upon. Therefore, a rigorous and 590
coordinated research effort with a holistic view of the interrelationships between soil and 591
above ground carbon sequestration and other ecosystems services provided is needed (Gama-592
Rodrigues et al., 2011. Particularly in developing countries, social, economic, governance, 593
and even cultural factors might have to be prioritised when agroforestry systems are 594
implemented. 595
596
5.3 Limitations of the study 597
598
The lack of standardised procedures for studying and reporting soil carbon sequestration in 599
different agroforestry systems is currently an impediment to synthesis and hence realisation 600
of the potential benefits. Only few studies directly reported the land use before the 601
agroforestry system implementation, and this often had to be inferred from the information 602
provided in the case study description. Other variables that influence above ground and 603
carbon sequestration were also not systematically reported. For example, tree density - which 604
was found to be positively correlated with soil carbon in homegardens in Bangladesh (Islam 605
et al, 2015) - was hardly reported. Other variables only occasionally reported were soil type, 606
soil pH and land management practices (e.g. fertiliser application, soil organic matter 607
additions, tillage practices, harvesting). Minasny et al. (2011) showed an increase in soil 608
organic carbon (SOC) in cropland potentially due to fertiliser application. Results from 609
several meta-analyses compiled by Stockmann et al. (2013) show an increase in SOC in soil 610
surface when no-tillage is adopted in comparison to conventional tillage, with this increase 611
improving with time. 612
613
Statistical information such as the variance or the inventory design or sample size was also 614
omitted in the great majority of studies. The effect sizes from individual studies in a meta-615
analysis should be combined using weighted statistical models whose weights are based on 616
the studies’ sampling variances (Rosenberg et al., 2000). As many of the studies did not 617
report any measure of variance for the response variables collected for the study, a weighted 618
meta-analysis could not be undertaken. For future studies of this type, it is recommended the 619
reporting of previous land uses (real or hypothetical), soil sampling methods, soil type, soil 620
pH, management practices and soil and above ground carbon sequestration values before and 621
after the implementation of the agroforestry system. It is also recommended that future 622
studies present better consistency in experimental design and reporting, and that further 623
20
research on C sequestration changes due to implementation of agroforestry systems in 624
different world regions and climates is undertaken. Resources should be put in place to 625
collect more secondary data from published work on carbon sequestration in above ground 626
and soils under agroforestry systems in order to undertake future analyses. Cheng and Kimble 627
(2001) considered that inadequate methods to quantify and measure carbon are a major 628
problem faced by policy makers and designers of carbon offset schemes. Albrecht and Kandji 629
(2003) argued that the scarcity of quantitative data to support the inclusion of agroforestry in 630
climate change discussions is still a problem. 631
632
Another factor that can affect quality of the analysis is the homogenisation of soil depths to 0-633
30 cm, using equations 1 and 2. Although the IPCC tier 1 level calculation considers SOC 634
stock changes to a reference depth of 30 cm (IPCC, 2006), the latest considerations on top 635
soil carbon have questioned whether this is the right depth to sample the soil or not. For 636
example, Powlson et al, 2014 conclude that the previously widely accepted advantages of no-637
till on SOC may have been overstated partly due to an artefact of preferentially observing the 638
top 30 cms of soil, which ignores the redistribution of topsoil carbon resulting from inversion 639
ploughing. According to Batjes (2010) it is useful to have estimates for SOC stocks to a 640
greater depth because the potential impact of a change in land use and management of SOC30 641
stocks may vary markedly according to IPCC climate zone and soil class, depending on the 642
actual depth, type and intensity of soil disturbance. 643
644
A total of 86 the studies reporting above ground and soil carbon in agroforestry systems were 645
compiled but more studies might have been available other languages. A large number of 646
studies reported on agroforestry systems located in Latin American countries and Asia 647
(Tables 2 and 3) which suggest that there are more studies available in Spanish, Portuguese or 648
Chinese (e.g. Scientific Electronic Library Online - SciELO). Some examples of Spanish 649
language literature in this topic are Alvarado et al. (2013), Ibrahim et al. (2007), and Ortiz et 650
al. (2008). A review of these databases was beyond the scope of this study. 651
652
6. Conclusion 653
Agroforestry systems can have an important role in climate change mitigation, especially if 654
this is linked with climate change adaptation strategies. Based on previous studies, there are 655
benefits in terms of carbon sequestration from the implementation of agroforestry systems, 656
21
and these are especially relevant in tropical climates. Some agroforestry systems are more 657
effective than others in sequestering soil and above ground carbon but a holistic assessment, 658
where other greenhouse gases are accounted and barriers to implementation as well as the 659
goods and services provided by the different agroforestry systems are taken into account. 660
Despite the potential of agroforestry systems to sequester and store carbon, inconsistent 661
methodologies and lack of previous quantitative reviews have held back the implementation 662
of reward schemes for farmers and communities. This study addresses this problem by 663
providing more information about carbon sequestration potential in different agroforestry 664
systems, regions and climates. Further, it provides two global empirical models than can be 665
used to predict carbon sequestration potential for different agroforestry systems in different 666
regions and climates, which may be used, particularly, where primary data is not available. 667
This can inform the upcoming IPCC guidelines for greenhouse gas emissions and carbon 668
sequestration inventory. It is recommended that studies systematically report on previous land 669
uses, soil characteristics and management as well as on data sampling and statistical 670
information so that estimates can be improved in the future. 671
672
Acknowledgements 673
This work was implemented as part of the CGIAR Research Program on Climate Change, 674
Agriculture and Food Security (CCAFS), which is carried out with support from CGIAR 675
Fund Donors (RG12839-10) and through bilateral funding agreements. For details please visit 676
https://ccafs.cgiar.org/donors. The views expressed in this document cannot be taken to 677
reflect the official opinions of these organisations. This work has also been partially funded 678
by the UK Natural Environment Research Council (NERC). 679
680
Appendix A. Supplementary material 681
Above ground carbon data 682
Soil carbon data 683
List of studies analysed 684
Estimates above ground 685
Estimates Soil 686
R Code 687
GLMM 688
689
690
22
References 691
Albrecht, A., Kandji, S. T. (2003). Carbon sequestration in tropical agroforestry systems. 692 Agriculture, Ecosystems and Environment 99:15–27 693 694 Alvarado, J., Andrade, H., (2013). Storage of soil organic carbon in coffee (Coffea arabica 695 L.) production systems in the municipality of Libano, Tolima, Colombia. Colombia Forestal 696 16:21-31/enero-junio. 697 698 Bates, D., Mächler, M., Bolker, B., Walker, S. (2015). Fitting linear mixed-effects models 699 using lme4. Journal of Statistical Software 67:1-48 700 701 Batjes, N.H. (2003) Estimation of soil carbon gains upon improved management within 702 croplands and grasslands of Africa. Environment, Development and Sustainability 6:133-143. 703 704 Bekele-Tesemma, A. (2007) Useful Trees and Shrubs for Ethiopia: Identification, 705 Propagation and Management for 17 Agroclimatic Zones. Tengäs, B., Kelbessa, E., 706 Demissew, S., Maundu, P (Eds). Nairobi: RELMA in ICRAF Project 707 708 Bishaw, B., Neufeldt, H., Mowo, J., Abdelkadir, A., Muriuki, J., Dalle, G., Assefa, T., 709 Guillozet, K., Kassa, H., Dawson, I., Luedeling, E., Mbow, C. (2013). Farmers’ Strategies 710 for Adapting to and Mitigating Climate Variability and Change through Agroforestry in 711 Ethiopia and Kenya. Davis, C.M., Bernart, B., and Dmitriev, A. (Eds). Forestry 712 Communications Group, Oregon State University, Corvallis, Oregon. 713
Brakas, S.G., Aune, J.B. (2011). Biomass and carbon accumulation in land use systems of 714 Claveria, The Philippines. In: Kumar, B.M. & Nair, R.P.K (Eds). Carbon Sequestration 715 Potential of Agroforestry Systems: Opportunities and Challenges. Vol 8: Advances in 716 Agroforestry. New York: Springer. 717 718 Cheng and Kimble (2001) Cheng, H.H., Kimble, J., 2001. Characterization of soil organic 719 carbon pools. In: Lal, R., Kimble, J., Follett, R.F., Stewart, B.A. (Eds.). Assessment Methods 720 for Soil Carbon. Advances in Soil Science. Lewis Publishers, Boca Raton, FL, 117–129. 721 722 Dixon R.K., (1995). Agroforestry systems: sources or sinks of greenhouse gases? 723 Agroforestry Systems 31:99-116. 724 725 Dixon, R., Winjum, J.K., Andrasko, K. J., Lee, J. J., Schroeder, P. E. (1994). Integrated land-726 use systems: Assessment promising agroforest and alternative land-use practices to enhance 727 carbon conservation and sequestration. Climatic Change 27:71-92. 728 729 FAO (2001) Soil Carbon Sequestration for Improved Land Management. Food and 730 Agriculture Organization of the United Nations, Rome. Available at http://www.fao.org/3/a-731 bl001e.pdf 732 733 FAO (2011a). Climate change, water and food security. FAO Water Reports 36. Rome, 734 FAO. Available at: http://www.fao.org/nr/water/jsp/publications/search.htm 735 736 FAO (2011b). The state of the world’s land and water resources for food and agriculture 737 (SOLAW) Managing systems at risk. Food and Agriculture Organization of the United 738 Nations, Rome and Earthscan, London. 739
23
740 Feliciano 2013 Feliciano, D., Slee, B., Hunter, C. & Smith, P. (2013). Estimating the 741 contribution of rural land uses to greenhouse gas emissions: A case study of North East 742 Scotland. Environmental Science & Policy 25:36-49. 743 744 Gama-Rodrigues, E.F., Gama-Rodrigues, A.C., and Nair, P.K.R. (2011) Soil Carbon 745 Sequestration in Cacao Agroforestry Systems: A Case Study from Bahia, Brazil. In: Kumar, 746 B.M. and Nair, P.K.R. Carbon Sequestration Potential of Agroforestry Systems: 747 Opportunities and Challenges. Vol 8: Advances in Agroforestry. New York: Springer, 85-99. 748 749 Howlett, D.S., Mosquera-Losada, M.R., Nair, P.K.R., Nair, V.D., Rigueiro-Rodríguez, A. 750 (2011) Soil C storage in silvopastoral systems and a treeless pasture in northwestern Spain. J 751 Environ Qual 40:784–790. 752 753 Huai, H., Hamilton, A. (2009) Characteristics and functions of traditional homegardens: a 754 review. Frontiers of Biology in China 4:151-157. 755 756 Ibrahim et al. (2007) Almacenamiento de carbono en el suelo y la biomasa arbórea en 757 sistemas de usos de la tierra en paisajes ganaderos de Colombia, Costa Rica y Nicaragua 758 Agroforestería en las Américas Nº 452007. 759 760 IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by 761 the National Greenhouse Gas Inventories Programme. Eggleston H.S., Buendia L., Miwa K., 762 Ngara T. and Tanabe K. (eds). Institute for Global Environmental Strategies, Tokyo, Japan. 763 Available at: http://www.ipcc-nggip.iges.or.jp/public/2006gl/index.html [Accessed 13 764 September 2011]. 765 766 IPCC (2000). Summary for Policymakers, Special Report on Land Use, Land Use Change, 767 and Forestry, Cambridge, UK: Cambridge University Press. 768 769 IPCC (1997) The Regional Impacts of Climate Change: An Assessment of Vulnerability. 770 Watson, R.T., Zinyowera, M.C., Moss, R.H. (Eds). Cambridge, UK: Cambridge University 771 Press. Available at: http://www.ipcc.ch/ipccreports/sres/regional/index.php?idp=0 772 773 Islam, M., Dey, A., Rahman, M. (2015) Effect of tree diversity on soil organic carbon content 774 in the homegarden agroforestry system of North-Eastern Bangladesh. Small-Scale Forestry 775 14:91-101. 776 777 Jarecki. M.K., Lal, R. (2003) Crop management for soil carbon sequestration. Crit Rev Plant 778 Sci 22:471-502 779 780 Joffre, R., Rambal, S., Ratte, J.-P. (1999). The dehesa system of southern Spain and Portugal 781 as a natural ecosystem mimic. Agroforestry Systems 45:57–79. 782 783 Jose, S. (2009). Agroforestry for ecosystem services and environmental benefits: An 784 overview. Agroforestry Systems 76:1-10 785 786 Jose, S., Bardhan, S. (2012) Agroforestry for biomass production and carbon sequestration: 787 An overview. Agroforestry Systems 86:105-111. 788 789
24
Kandji, S.T., Verchot, L.V., Mackensen, J., Boye, A., Van Noordwijk, M., Tomich, C.K., 790 Ong, C.K., Albrecht, A., Palm, C.A. (2006). Opportunities for linking climate change 791 adaptation and mitigation through agroforestry systems. In: World Agroforestry into the 792 Future. Garrity, D.P., Okono, A., Grayson, M., Parrott, S., (Eds). Nairobi: World 793 Agroforestry Centre, 113-121 794 795 Kaur, B., Gupta, S.R., Sahu, K.P., Paikra, P.R. (2002) Carbon storage and nitrogen cycling in 796 silvopastoral systems on a sodic soil in northwestern India. Agroforestry Systems 54:21-29. 797 798 Kim, D-G., Thomas, A.D., Pelster, D., Rosenstock, T.S., Sanz-Cobena, A. (2016) 799 Greenhouse gas emissions from natural ecosystems and agricultural lands in sub-Saharan 800 Africa: synthesis of available data and suggestions for further research. Biogeosciences 801 13:4789-4809. 802 803 Kimaro, A.A., Isaac, M.E., Chamshama, S.A.O. (2011) Carbon Pools in Tree Biomass and 804 Soils Under Rotational Woodlot Systems in Eastern Tanzania. In: Kumar. B.M., Nair, P.K.R. 805 (Eds). Carbon Sequestration Potential of Agroforestry Systems: Opportunities and 806 Challenges. Vol 8: Advances in Agroforestry. New York: Springer, 129-143. 807 808 Kimaro, A. A., Timmer, V. R., Mugasha, A. G., Chamshama, S. A. O., Kimaro, D. A. (2007) 809 Nutrient use efficiency and biomass production of tree species for rotational woodlot systems 810 in semi-arid Morogoro, Tanzania. Agroforest Systems (2007) 71:175–184. 811 812 Kirby, K. R., Potvin, C. (2007) Variation in carbon storage among tree species: Implications 813 for the management of a small-scale carbon sink project. Soil Use and Management 20:350–814 356. 815 816 Kumar, B.M., Nair, P.K.R. (Eds) Carbon Sequestration Potential of Agroforestry Systems: 817 Opportunities and Challenges. Vol 8: Advances in Agroforestry. New York: Springer. 818 819 Lal, R. (2005). Forest soils and carbon sequestration. Forest Ecology and Management 820 220:242–258. 821 822 Lal, R. (2002) The potential of soils of the tropics to sequester carbon and mitigate the 823 greenhouse effect. Adv. Agron. 76:1-30. 824 825 Lipper, L., Dutilly-Diane, C., and McCarthy, N. (2010). Supplying carbon sequestration from 826 West African rangelands: opportunities and barriers. Rangeland Ecology and Management 827 63:155–166. 828 829 Lorenz, K., Lal, R. (2014) Soil organic carbon sequestration in agroforestry systems. A 830 review. Agron Sustain Dev 34:443–454. 831 832 Luedeling, E., Neufeldt, H. (2012) Carbon sequestration potential of parkland agroforestry in 833 the Sahel. Climatic Change 115:443-461. 834 835 Mbow, C., Smith, P., Skole, D., Duguma, L., Bustamante, M. (2014). Achieving mitigation 836 and adaptation to climate change through sustainable agroforestry practices in Africa. Current 837 Opinion in Environmental Sustainability 6:8–14. 838 839
25
Minasny et al. (2011) Minasny, B., Sulaeman, Y., McBratney, A.B., 2011. Is soil carbon 840 disappearing? The dynamics of soil organic carbon in Java. Global Change Biol. 17:1917–841 1924. 842 843 Montero, G., San Miguel, A., Cañellas, I. (1998). Systems of Mediterranean silviculture ‘La 844 Dehesa’. In: Jiménez Díaz RM, Lamo de Espinos J (eds.) Agricultura Sostenible. Mundi 845 Prensa, Madrid 519-554. 846 847 Murthy IK, Gupta M, Tomar S, Munsi M, Tiwari R, et al. (2013) Carbon Sequestration 848 Potential of Agroforestry Systems in India. J Earth Sci Climate Change 4:131. 849 850 Mutuo, P.K., Cadisch, G., Albrecht, A., Palm, C.A., Verchot, L. (2009) Potential of 851 agroforestry for carbon sequestration and mitigation of greenhouse gas emissions from soils 852 in the tropics. Nutrient Cycling in Agroecosystems 71:43-54. 853 854 Mutuo, P.K., Cadisch, G., Albrecht, A., Palm, C.A. and Verchot, L. (2005). Potential of 855 agroforestry for carbon sequestration and mitigation of greenhouse gas emissions from soils 856 in the tropics. Nutrient Cycling in Agroecosystems 71:45-54. 857 858 Nair, P.K.R. (2012) Carbon sequestration studies in agroforestry systems: a reality-check. 859 Agroforestry Systems 86:243-253. 860 861 Nair, P.K.R. (1993) An Introduction to Agroforestry. Dordrecht, The Netherlands: Kluwer 862 Academic Publishers. 863 864 Nair, P.K.R. (1985) Classification of agroforestry systems. Agroforestry Systems 3:97-128. 865 866 Nair, P. K. R., Nair, V. D., Kumar, M. B., Haile, S. G. Soil carbon sequestration in tropical 867 agroforestry systems: a feasibility appraisal. Environmental Science & Policy 12:1099-1111. 868 869 Nakagawa, S., Schielzeth, H. (2013). A general and simple method for obtaining R2 from 870 generalized linear mixed-effects models. Methods in Ecology and Evolution 4:133–142. 871 872 Negash, M., Kanninen, M. (2015) Modelling biomass and soil carbon sequestration of 873 indigenous agroforestry systems using CO2FIX approach. Agriculture, Ecosystems and 874 Environment 203:147-155. 875 876 Nelson, J.P., Kennedy, P.E., 2009. The use (and abuse) of meta-analysis in environmental 877 and natural resource economics: an assessment. Environ. Resour. Econ. 42:345–377. 878 879 Nyadzi, G.I., Otsyina, R.M., Banzi, F.M., Bakengesa, S.S., Gama, B.M., Mbwambo, L. 880 Asenga, D. (2003). Rotational woodlot technology in northwestern Tanzania: Tree species 881 and crop performance. Agroforestry Systems 59:253–263. 882 883 Oelbermann, M., Voroney, R.P., Thevathasan, N.V., Gordon, A.M., Kass, D.C.L., 884 Schlonvoigt, A.M. (2006). Soil carbon dynamics and residue stabilization in a Costa Rican 885 and southern Canadian alley cropping system. Agroforestry Systems 2006 68:27–36. 886 887
26
Oke, D., Olatiilu, A. (2011) Carbon storage in Agroecosystems: A Case study of the cocoa 888 based Agroforestry in Ogbese Forest Reserve, Ekiti state, Nigeria. Journal of Environmental 889 Protection 2:1069-1075. 890 891 Ortiz, A., Riascos, L., Somarriba, E. (2008) Almacenamiento y tasas de fijación de biomasa y 892 carbono en sistemas agroforestales de cacao (Theobroma cacao) y laurel (Cordia alliodora). 893 Agroforestería en las Américas Nº 46. 894 895 Pandy, D.N. (2007) Multifunctional agroforestry systems in India. Current Science 92:455-896 463. 897 898 Parrotta, J.A. (1999). Productivity, nutrient cycling, and succession in single- and mixed-899 species plantations of Casuarina equisetifolia, Eucalyptus robusta, and Leucaena 900 leucocephala in Puerto Rico. Forest Ecology and Management 124:45-77. 901 902 Powlson, D.S., Stirling, C.M., Jat, M.L., Gerard, B.G., Palm, C.A., Sanchez, P.A., Cassman, 903 K.G. (2014) Limited potential of no-till agriculture for climate change mitigation. Nature 904 Climate Change 4 678-683. 905 906 Rosenberg, M.S., Adams, D. C., Gurevitch, J. (2000). Metawin: Statistical Software for 907 Meta-Analysis: Version 2.0. Ecology 78:1277-1283. 908 909 Roshetko, J.M., Delaney, M., Hairiah, K., Purnomosidhi, P. (2002) Carbon stocks in 910 Indonesian homegarden systems: Can smallholder systems be targeted for increased carbon 911 storage. American Journal of Alternative Agriculture 17:138-148. 912 913 Sanchez, P.A., Buresh, R.J. and Leakey, R.R.B.: (1997), Trees, soils and food security. 914 Philosophical transactions of the Royal Society of London, Series B 353:949–961. 915 916 Schoeneberger, M.M. (2009) Agroforestry: working trees for sequestering carbon on 917 agricultural lands. Agroforestry systems 75:27-37. 918 919 Schroth, G., Bede, L.C., Paiva, A.O., Cassano, C.R., Amorim, A.M., Faria, D., Mariano-920 Neto, E., Martini, A.M.Z., Sambuichi, R.H.R., bo, R.N. (2013) Contribution of agroforests 921 to landscape carbon storage. Mitig Adapt Strateg Glob Change 20:1175-1190. 922 923 Sharrow, S.H., Ismail, S. (2004). Carbon and nitrogen storage in agroforests, tree plantations, 924 and pastures in western Oregon, USA. Agroforestry Systems 60:123–130. 925 926 Smiley, G.L., Kroschel, J., (2008). Temporal change in carbon stocks of cocoa – Gliricidia 927 agroforests in Central Sulawesi, Indonesia. Agrofor. Syst. 73:219–231. 928 929 Smith, P., Haberl, H., Popp, A., Erb, K.-h., Lauk, C., Harper, R., Tubiello, F. N., de Siqueira 930 Pinto, A., Jafari, M., Sohi, S., Masera, O., Böttcher, H., Berndes, G., Bustamante, M., 931 Ahammad, H., Clark, H., Dong, H., Elsiddig, E. A., Mbow, C., Ravindranath, N. H., Rice, C. 932 W., Robledo Abad, C., Romanovskaya, A., Sperling, F., Herrero, M., House, J. I. and Rose, 933 S. (2013) How much land-based greenhouse gas mitigation can be achieved without 934 compromising food security and environmental goals? Global Change Biology 19:2285–935 2302. 936 937
27
Smith, P., Powlson, D.S., Glendining, M.J., Smith, J.U., 1997a. Using long-term experiments 938 to estimate the potential for carbon sequestration at the regional level: an examination of five 939 European scenarios. Agroke´m. Talajt 46:25–38. 940 941 Smith, P., Powlson, D.S., Glendining, M.J., Smith, J.U., 1997b. Potential for carbon 942 sequestration in European soils: preliminary estimates for five scenarios using results from 943 long-term experiments. Global Change Biology 3:67–79. 944 945 Somarriba, E., Cerda, R., Orozco, L., Cifuentes, M., Dávila, H., Espin, T., Mavisoy, H., 946 Ávila, G., Alvarado, E., Poveda, V., Astorga, C., Say, E., Deheuvels, O. (2013) Carbon 947 stocks and cocoa yields in agroforestry systems of Central America. Agriculture, Ecosystems 948 and Environment 173:46–57. 949 950 Stephan-Dewenter, I., Kesslerc, M., Barkmannc, J., Bosa, M., Buchorig, D., Erasmih, S., 951 Fausth, H., Geroldh, G., Glenke, K., Gradsteind, S.R., Guhardjai, E., Harteveldd, M., Herteld, 952 D., Höhna, P., Kappash, M., Köhlerh, S., Leuschnerd, C., Maertensj, M., Marggrafe, R., 953 Migge-Kleiank, S., Mogeai, J., Pitopangl, R., Schaeferk, M., Schwarzem, S., Spornd, S.G., 954 Steingrebek, A., Tjitrosoedirdjoi, S.S., Tjitrosoemitoi, S., Tweleh, A., Weberh, R., 955 Woltmannk, L., Zellerm, M.N., Tscharntkea, T., (2007) Tradeoffs between income, 956 biodiversity, and ecosystem functioning during tropical rainforest conversion and 957 agroforestry intensification. PNAS 104:4973–4978. 958 959 Stockmann, U., Adams, M.A., Crawford, J.W., Field, D.J., Henakaarchchi, N., Jenkins, M., 960 Minasny, B., McBratney, A.B., de Remy de Courcelles, V., Singh, K., Wheeler, I., Abbott, 961 L., Angers, D.A., Baldock, J., Bird, M., Brookes, P.C., Chenu, C., Jastrow, J.D., Lal, R., 962 Lehmann, J., O’Donnell, A.G., Parton, W.J., Whitehead, D., Zimmermann, M. (2013) The 963 knowns, known unknowns and unknowns of sequestration of soil organic carbon. 964 Agriculture, Ecosystems and Environment 164:80-99. 965 966 Swamy, S.L., Puri, S. (2005) Biomass production and C-sequestration of Gmelina arborea in 967 plantation and agroforestry system in India. Agroforestry Systems 64:181-195. 968 969 Takimoto, A., Nair, P.K.R., Nair, V., D. (2008). Carbon stock and sequestration potential of 970 traditional and improved agroforestry systems in the West African Sahel. Agriculture, 971 Ecosystems and Environment 125:159–166. 972 973 Torquebiau, E.F. (2000) A renewed perspective on agroforestry concepts and classification. 974 C.R. Acad. Sci. - Sciences de la Vie 323:1009-1017. 975 976 Tsuruta, H., Ishizuka, S., Ueda, S. and Murdiyarso, D. (2000). Seasonal and spatial 977 variations of CO2, CH4, and N2O fluxes from the surface soils in different forms of land-978 use/cover in Jambi, Sumatra. In: Murdiyarso, D. and Tsuruta, H. (Eds.) The Impacts of 979 Land-Use/Cover Change on Greenhouse Gas Emissions in Tropical Asia. IC-SEA, Bogor, 980 Indonesia and NIAES, Tsukuba, Japan, 7-30. 981
Tubiello, F., Schmidhuber, J., Howden, M., Neofotis, P.G., Park, S., Fernandes, E., Thapa, D. 982 (2008) Climate Change Response Strategies for Agriculture: Challenges and Opportunities 983 for the 21st Century. Agriculture and rural development discussion paper, no. 42, 984 Washington, DC:World Bank. 985 986
28
Verchot, L.V., Noordwijk, M.V., Kandji, S., Tomich, T., Ong, C., Albrecht, A., Mackensen, 987 J., Bantilan, C., Anupama, K.V., Palm, C. Climate change: linking adaptation and mitigation 988 through agroforestry. Miti Adapt Stat Glob Change 12:901-918. 989 990 991
Table 1 Reclassification of agroforestry systems reported by the studies
Agroforestry
system
reclassified
Definition
Agroforestry system
as reported in the
study
Agrisilvicultural This system involves simultaneously growing crops and trees on the same
piece of land (Brakas and Aune, 2011).
Agrisilvicultural,
parklands,
intercropping (e.g.
corn and banana),
taungya
Silvopastoral
Silvopastoral
Boundary
planting
between roads and farms, while providing timber, poles, fruits, fuelwood and
agroforestry systems, the common species used in boundary planting systems
are Grevillea Robusta, Markhamia lutea, Eucalyptus spp. and Alnus
Boundary planting,
shelterbelt, live fence
Improved fallows
some of the fertility lost through continuous cropping with limited or no
fertilizer application. Improved fallow consists of planting trees, mainly
legume tree species, in order to enrich the soil within a shorter time period,
compared with natural fallow (Bekele-Tesemma 2007; Brakas and Aune,
Improved fallows
Shadow systems Agroforestry systems that combine coffee, tea or cocoa shrubs with multi-
purpose shade species (Somarriba et al., 2013).
Cocoa-based
agroforestry, tea-based
agroforestry, coffee-
based agroforestry
Homegardens
parcels of land surrounding homesteads. It comprises numerous woody
species in close, multi-storied association with herbs, annual and perennial
crops, and livestockall managed in the same piece of land (Roshetko et al.,
Homegardens,
multistrata agroforests
Woodlots
between properties, such as housing subdivisions, industrial forests, or public
agroforestry option that attempts to simulate the traditional fallow system in
shifting cultivation, where trees
combine the principles of crop production and forest management to provide
multiple products. The technology involves growing trees and crops in three
interrelated phases: (i) initial tree establishment, where trees are intercropped
with crops; (ii) tree fallow; and (iii) cropping after tree harvests (Nyadzi et al.
Woodlots, fast
growing trees
Table 2 Variables collected and analysed
Variables Above ground carbon
sequestration Total n Soil carbon sequestration Total n
Independent
variable
Above ground carbon sequestration
t CO2 ha
-1
yr
-1
162 Soil carbon sequestration t CO2 ha-1 yr-1 142
Study ID
Identification number for each
study reporting above ground
carbon sequestration
162 Identification number for each study
reporting soil carbon sequestration 142
Study year
Year of publication of the study
reporting above ground carbon
sequestration
162 Year of publication of the study reporting
soil carbon sequestration 142
Continent
Africa (n=60)
Asia (n=50)
Australia (n=1)
Europe (n=2)
Latin America (n=45)
North America (n=4)
162
Africa (n=59)
Asia (n=20)
Australia (n=0)
Europe (n=3)
Latin America (n=44)
North America (n=15)
141
Climate
Arid (n=5)
Mediterranean (n=1)
Polar (n=3)
Semiarid (n=17)
Temperate (n=3)
Tropical (n=133)
162
Arid (n=24)
Mediterranean (n=0)
Polar (n=1)
Semiarid (n=11)
Temperate (n=18)
Tropical (n=88)
142
Agroforestry
system
Agrisilvicultural (n=15)
Boundary Planting (n=6)
Homegarden (n=33)
Improved fallows (n=21)
Shadow systems (n=40)
Silvopastoral (n=13)
Woodlots (n=34)
162
Agrisilvicultural (n=52)
Boundary Planting (n=2)
Homegarden (n=21)
Improved fallows (n=17)
Shadow systems (n=24)
Silvopastoral (n=9)
Woodlots (n=17)
142
Land use
before
Cropland (n=40)
Degraded land (n=16)
Forest (n=3)
Grassland (n=26)
Not reported (n=77)
162
Cropland (n=74)
Degraded land (n=4)
Fallow (n=5)
Grassland (n=20)
Unshaded (n=20)
Underutilised (n=19)
Not reported (n=0)
142
Time
Time since the implementation of
the agroforestry system (years)
162
Time since the implementation of the
agroforestry system (years)
142
Land use
change
Cropland to Agrisilvicultural
(n=11)
Cropland to Boundary Planting
(n=3)
Cropland to Homegarden (n=8)
Cropland to Improved Fallows
(n=5)
Cropland to Shadow Systems (n=3)
Cropland to Woodlots (n=12)
Degraded to Improved Fallows
(n=14)
Degraded to Woodlots (n=2)
Forest to Shadow systems (n=2)
Grassland to Agrisilvicultural (n=2)
Grassland to Boundary planting
(n=1)
Grassland to Homegarden (n=1)
Grassland to Silvopastoral (n=5)
Grassland to Woodlots (n=17)
Not reported (n=76)
162
Cropland to Agrisilvicultural (n=46)
Cropland to Homegarden (n=1)
Cropland to Improved Fallows (n=17)
Cropland to Woodlots (n=10)
Degraded Land to Agrisilvicultural (n=1)
Degraded Land to Boundary Planting
(n=1)
Degraded Land to Woodlots (n=2)
Fallow Woodlots (n=5)
Grassland to Agrisilvicultural (n=5)
Grassland to Boundary Planting (n=1)
Grassland to Homegarden (n=1)
Grassland to Shadow Systems (n=4)
Grassland to Silvopastoral (n=9)
Unshaded Systems to Shadow Systems
(n=20)
Underutilised Land to Homegardens
(n=19)
142
Table 3 Mean above ground carbon sequestration by agroforestry system type and
continent (tC ha-1 yr-1)
Continent Agroforestry
system type Mean Variance Number of
observations
Africa (n=60)
Agrisilvicultural 0.88 0.14 5
Homegarden 0.52 0.07 5
Improved fallows 12.95 20.12 17
Shadow systems 2.27 2.36 18
Silvopastoral 0.15 - 1
Woodlots
3.36
1.85
14
Asia (n=50)
Agrisilvicultural 1.13 2.52 4
Homegarden 2.77 5.8 27
Improved fallows 2.9 0.08 2
Silvopastoral 2.65 4.35 7
Woodlots
6.28
26.57
10
Australia (n=1) Silvopastoral 0.79 - 1
Europe (n=2) Woodlots 5.74 1.86 2
Latin America
(n=45)
Agrisilvicultural 2.94 5.56 6
Boundary
Planting
9.14 54.72 5
Homegarden 3.25 - 1
Improved fallows 5.55 5.45 2
Shadow systems 2.87 2.79 22
Silvopastoral 2.29 0.29 3
Woodlots
12.63
8.57
6
North America
(n=4)
Boundary
Planting
1.12 - 1
Silvopastoral 0.52 - 1
Woodlots
3.73
11.23
2
Note: Total number of observations (n) per continent in brackets
Table 4 Mean soil carbon sequestration by agroforestry system type and continent (tC ha-1yr-
1)
Continent Agroforestry
system type Mean Variance Number of
observations
Africa
(n=59)
Agrisilvicultural 0.32 2.42 25
Boundary
Planting -0.98 0.37 2
Homegarden 0.19 - 1
Improved
fallow 1.91 3.42 17
Shadow
systems 1.91 13.01 3
Woodlots 0.85 0.67 11
Asia
(n=20)
Agrisilvicultural 0.27 - 1
Homegarden 3.83 2.36 19
Europe
(n=3)
Agrisilvicultural 6.7 - 1
Woodlots 4.2 28.96 2
Latin
America
(44)
Agrisilvicultural 1.73 3.07 13
Homegarden -1.17 - 1
Continent Agroforestry
system type Mean Variance Number of
observations
Africa
(n=59)
Agrisilvicultural 0.32 2.42 25
Boundary Planting -0.98 0.37 2
Homegarden 0.19 - 1
Improved fallow 1.91 3.42 17
Shadow systems 1.91 13.01 3
Woodlots
0.85
0.67
11
Asia
(n=20)
Agrisilvicultural 0.27 - 1
Homegarden 3.83 2.36 19
Europe
(n=3) Agrisilvicultural 6.7 - 1
Woodlots
4.2
28.96
2
Latin
America
(44)
Agrisilvicultural 1.73 3.07 13
Homegarden -1.17 - 1
Shadow systems 1.05 5.75 21
Silvopastoral 6.54 2.99 6
Woodlots
-4.04
0.37
3
North
America
(n=15)
Agrisilvicultural 0.39 10.99 12
Silvopastoral 0.06 0.38 3
Shadow
systems 1.05 5.75 21
Silvopastoral 6.54 2.99 6
Woodlots -4.04 0.37 3
North
America
(n=15)
Agrisilvicultural 0.39 10.99 12
Silvopastoral 0.06 0.38 3
Note: Total number of observations (n) per continent in brackets
Table 5: Coefficients for the different land uses before the agroforestry system being
implemented
LUBefore
Coef
Cropland
-1.23
Grassland
0.90
Degraded land
0
Note: The agroforestry systems fallow and underutilised were included as degraded land
Table 6: Coefficients (A and B) for the different climates and land uses before the
agroforestry system being implemented
Climate
CoefA
Polar
-1.53
Temperate
1.08
Semiarid
2.40
Tropical
2.39
LUBefore
CoefB
Cropland
-1.86
Grassland
-2.18
Unshaded
-1.62
Degraded land
0
Note: The agroforestry systems fallow and underutilised and degraded land were included as degraded land
1
1
Figure 1 Mean, maximum and minimum above ground (A) and soil (B) carbon sequestration in different agroforestry systems. Number of 2
observations (n) is presented in brackets. 3
4
-5
0
5
10
15
20
25
A- Above ground carbon (tCha-1yr-1)
Mean
-10
-8
-6
-4
-2
0
2
4
6
8
10
12
B- Soil carbon (tCha-1yr-1)
Mean
2
Figure 2 Mean, maximum and minimum above ground (A) and soil (B) carbon sequestration in different climates. Number of observations (n) is 5
presented in brackets. 6
7
8
9
10
11
12
13
14
-5
0
5
10
15
20
25
A- Above ground carbon (tCha-1 yr-1)
Mean
-10
-8
-6
-4
-2
0
2
4
6
8
10
12
Tropical (88) Semiarid (11) Temperate (18) Polar (1) Arid (24)
B- Soil carbon (tCha-1 yr-1)
Mean
3
15
Figure 3 Mean, maximum and minimum above ground (A) and soil (B) carbon sequestration in agroforestry systems implemented in Tropical 16
climates. Number of observations (n) is presented in brackets. 17
18
19
-5
0
5
10
15
20
25
A - Above ground carbon (tCha-1.yr-1)
Mean
-8
-6
-4
-2
0
2
4
6
8
10
12
B- Soil carbon (tCha-1.yr-1)
Mean
4
20
21
Figure 4 Mean absolute change in above ground carbon sequestration resulting from the 22
implementation of an agroforestry system in tC.ha-1.yr-1. 23
Note: Mean and 95% Lower and Upper Confidence Limits (CL) for the mean. Number of observations is presented in brackets next to the 24
type of land use change. Land use changes with fewer than five observations were not included in the graph: cropland to boundary planting 25
(n=3), open ground to shadow systems (n=3), degraded to woodlots (n=2), forest to shadow systems (n=2), grassland to agrisilvicultural 26
(n=2), grassland to boundary planting (n=1), grassland to homegarden (n=1) 27
28
29
Grasslands to Woodlots
Mean=8.27, n=17
Degraded to Improved
fallow
Mean=12.8, n=14
Cropland to Woodlots
Mean=3.9, n=12
Cropland to
Agrisilvicultural
Mean=1.2, n=11 Cropland to
Homegarden Mean=3,
n=8
Cropland to Improved
fallow
Mean=9.4, n=5
Grassland to
Silvopastoral
Mean=1.65, n=5
-5 -4 -3 -2 -1 012345678910 11 12 13 14 15 16 17 18 19 20
Mean tC ha-1.yr-1
5
30
Figure 5 Mean absolute change in soil carbon sequestration resulting from the 31
implementation of an agroforestry system 32
Note: Mean and 95% Lower and Upper Confidence Limits (CL) for the mean, and number of observations presented in brackets. Land use 33
changes with fewer than five observations were not included in the graph: cropland to boundary planting (n=3), open ground to shadow 34
systems (n=3), degraded to woodlots (n=2), forest to shadow systems (n=2), grassland to agrisilvicultural (n=2), grassland to boundary 35
planting (n=1), grassland to homegarden (n=1). 36
37
38
Cropland to
Agrisilvicultural
Mean=1.0, n=46
Unshaded to Shadow
systems
Mean=1.3, N=20 Underutilised to
Homegarden
Mean 3.83, N=19
Cropland to Improved
Fallow
Mean=1.9, N=17
Cropland to Woodlots
Mean=-0.5, N=10
Grassland to
Silvopastoral
Mean=4.38, N=9
Fallow to Woodlots
Mean=0.5, N=5
Grassland to
Agrisilvicultural
Mean=-0.8, N=5
-5 -4 -3 -2 -1 012345678910
Mean tC.ha-1.yr-1
6
39
Figure 6 Above ground carbon sequestration over time since implementation of the 40
agroforestry system. Note: Each point corresponds to a value in the dataset. The dark blue line is the mean value of carbon 41
sequestration for that particular year, and the blue shade shows the standard error of the mean. 42
43
44
45 Figure 7 Soil carbon sequestration over time since implementation of the agroforestry 46
system. Note: Each point corresponds with a value in the dataset. The dark blue line is the mean value of carbon sequestration at that 47
particular year, and the blue shade shows the standard error of the mean. 48
49
7
50
Figure 8: Percentage of variance explained by each of the significant factors in the GLMM 51
AGCS model. 52
53
Figure 9: Percentage of variance explained by each of the significant factors in the GLMM 54
SCS model. 55
56
... Among the mechanisms contributing to SOC stability, soil aggregation stands out as a protective barrier against microbial decomposition, enhancing nutrient retention and soil physical properties (Keller and Phillips 2019; Augusto and Boca 2022). Studies suggest that soil texture signi cantly affects SOC accumulation, with ner-textured soils, such as clays, exhibiting a greater capacity for SOC storage due to their ability to stabilize organic matter through physical and chemical interactions (Castellano et Agroforestry systems have emerged as a viable solution for improving SOC storage and stability (Cardinael et al. 2015;Feliciano et al. 2018). By integrating trees with agricultural crops or pastures, these systems offer a multifunctional approach that enhances carbon sequestration, soil fertility, and microbial activity while fostering resilient ecosystems (Pardon et al. 2017). ...
Preprint
Full-text available
Agroforestry systems play a critical role in enhancing soil organic carbon (SOC) stability and mitigating climate change by integrating trees and crops to improve soil fertility and carbon sequestration. This study investigates the SOC stability, aggregate dynamics, and temperature sensitivity of SOC mineralization across four agroforestry systems ( Michelia oblonga, Parkia roxburghii, Alnus nepalensis , and Pinus kesiya ). Tree traits, soil properties, and aggregate characteristics were analyzed alongside a 60-day incubation experiment under three temperature regimes (25°C, 30°C, and 35°C). The results revealed the SOC mineralization significantly varied amongst the agroforestry systems with highest value in M. oblonga (25.59 mg CO 2 g − 1 ) and lowest in A. nepalensis (20.39 mg CO 2 g − 1 ). Macroaggregates consistently showed higher SOC concentrations and biochemical indicators, such as polysaccharides and total glomalin-related soil proteins (TG-RSP), compared to microaggregates and bulk soil. The temperature and aggregate sizes statistically influenced the SOC mineralization rates, with noticeable interaction effect. SOC mineralization rates increased with temperature, but Alnus nepalensis exhibited the highest temperature sensitivity (Q 10 = 0.955 and activation energy = 24.25 kJ mol − 1 ), highlighting its resilience to thermal stress. Strong positive correlations were observed between soil aggregate stability and soil biochemical indicators such as SOC, polysaccharides and TG-RSP of bulk soil and aggregates. Temporal trends indicated that carbon mineralization peaked at 30 days before stabilizing, reflecting the decomposition of labile carbon pools. These findings highlight the critical role of tree traits, soil aggregates, and thermal stability in driving SOC retention in agroforestry systems.
... Los sistemas agroforestales se definen como la integración de árboles en los bordes externos e internos de tierras de cultivo o en cualquier otro espacio disponible, lo que permite obtener beneficios tanto en la mitigación del cambio climático como en la producción de alimentos. Existen diversos tipos de SAF, cada uno con diferentes tasas de secuestro de carbono, tanto en el suelo como en la biomasa (7). ...
Article
Full-text available
El análisis estructural del sistema agroforestal en la finca "Magallanes", ubicada en Caña Brava, Paján, Manabí, Ecuador, se llevó a cabo con el objetivo de evaluar la diversidad y la estructura de las especies arbóreas presentes. Se establecieron siete parcelas de 20 x 25 metros en un área total de 7 hectáreas donde se identificaron 17 especies pertenecientes a 11 familias botánicas, con un total de 43 individuos, la investigación reveló que la mayoría de los árboles eran jóvenes con predominancia en la clase III de altura y la clase I en diámetro. Las especies más relevantes desde un punto de vista ecológico fueron Inga spectabilis, Cordia alliodora y Pseudosamanea guachapele, mientras que especies como Annona muricata, Citrus sinensis Bixa orellana mostraron menor presencia. Los índices de diversidad calculados, como Shannon-Weaver, Simpson, y Margalef, indicaron una diversidad media a alta en el sitio estudiado. La metodología incluyó un inventario forestal que consideró el diámetro a la altura del pecho (DAP) y la altura total de cada especie. Se utilizó un diseño aleatorio para las parcelas, georreferenciadas con GPS, y se aplicaron fórmulas para calcular la abundancia, dominancia y frecuencia relativa de las especies. El Índice de Valor de Importancia Ecológica (IVIE) fue fundamental para evaluar el papel ecológico de cada especie en el ecosistema. Este estudio subraya la importancia de los sistemas agroforestales no solo para la producción agrícola sino también para la conservación del suelo y la biodiversidad destacando la necesidad de prácticas sostenibles que integren tanto especies forestales como frutales para garantizar la viabilidad económica y ecológica a largo plazo.
... Agroforestry plays a crucial role in conserving biodiversity and ecosystem services within the agricultural landscape (De Beenhouwer et al. 2013). Agroforestry is instrumental in mitigating and adapting to climate change by sequestering atmospheric carbon in the soil, reducing environmental risks, and increasing resilience (Abbas et al. 2017;Feliciano et al. 2018;Patel and Moore 2019;Teeple and McMichael 2020). ...
Article
For efficient utilization of land, it is important to study the impact of locally suited agroforestry tree species on soil properties and their implications for crop yield and yield components. The aim of this study was to investigate the effects of Cordial africana Lam (C africana) on selected soil properties, yield, and yield components of maize (Zea mays L.). Soil data were collected from five radial transects at distances of 2 m (middle of the tree canopy), 4 m (edge of the tree canopy), 6 m (1-fold distance from the edge of the tree canopy), 8 m (2-fold distance from the edge of the tree canopy), and 20 m (control point). Three similar C. africana trees were purposefully selected from cultivated land. Maize was cultivated underneath the C. africana tree using a randomized complete block design (RCBD), and data on yield and yield components were collected from five radial transects. Soil bulk density (P = 0.003) significantly increased with increasing distance from the tree trunk. In contrast, soil organic carbon (P = 0.001), total nitrogen (P < 0.0001), available phosphorous (P = 0.001), cation exchange capacity (P = 0.02), and exchangeable calcium (P = 0.04) significantly decreased with increase in distance from C. africana tree trunk and were lowest at 20 m control plots. Flowering and maturity dates exhibited a substantial variation with distance from the tree trunk. Grain yield (P = 0.003), aboveground biomass (P = 0.004), and hundred-seed weight (P = 0.003) were significantly different with respect to the distance from the tree trunk. The 100 seed weight of maize was significantly higher (34.7 g) 6 m from the tree trunk. Maize grain yield (4.98 ton ha-1) and aboveground biomass (12.92 ton ha-1) were significantly higher at 6 m distance from the C. africana tree trunk showing similar trend with that of 100 seed weight. These findings suggest that incorporating C. africana tree species into agricultural land can enhance soil fertility and maize productivity, and help mitigate climate change.
... Furthermore, to adopt sustainable agriculture in place of conventional farming, numerous scientific studies have proposed the adoption of agroforestry, specifically the introduction of trees into crops. A meta-analysis of the most effective agroforestry options in terms of carbon storage in different regions reveals that agroforestry systems located in tropical areas exhibit high carbon storage values, estimated at 4.85 tCha⁻ 1 and 2.23 tCha⁻ 1 for aboveground carbon and underground carbon, respectively (Feliciano et al. 2018). Moreover, according to Vroh and Akoi (2024), agroforestry in cocoa cultivation offers many benefits, as it provides a wide variety of goods and services, such as non-timber forest products, firewood, construction wood for housing, fruits, other food items, and medicinal materials. ...
Article
Full-text available
With about 46% of global production, Côte d’Ivoire is the world’s leading producer of cocoa beans. However, this production contributes to deforestation, exacerbating the effects of climate change. In response to this observation, this study aims to deepen knowledge on the contribution of agroforestry systems in cocoa production areas in Côte d’Ivoire to atmospheric carbon storage. These main areas are the Centre-West, South-West, and West. In these areas, floristic richness was determined in 115 plots. Carbon stocks in living biomass, dead matter, and soil were evaluated. The dynamics of carbon stocks with age were also determined. The results revealed that the West area contains the most diversified cocoa agroforests, with 161 species compared to 71 and 119 in the Centre-West and South-West, respectively. Entandrophragma angolense, Nesogordonia papaverifera, and Sterculia oblonga, common to these areas, are on the IUCN Red List. Carbon stock varies by area, its history, the practices present, and especially the associated species. Thus, in the former cocoa production zone (Centre-West) and the current main production zone (South-West), Elaeis guineensis is the main carbon reservoir, with 25.576 tC.ha⁻¹ in the Centre-West and 36.862 tC.ha⁻¹ in the South-West. In the West, local trees form the main carbon reservoir with 11.701 tC.ha⁻¹. The dynamics of total carbon stocks show heterogeneous changes in production areas according to the different stages of development of agroforestry systems. This is evidence of the complexity of carbon flow and the dynamics of cocoa systems, which are strongly influenced by the sociology of the producers.
Chapter
Carbon farming is a new approach for land management that allows carbon sequestration by reducing greenhouse gas (GHG) emissions. Agriculture, forestry, and other land-use practices lead to 24% of global GHG emissions. Silvopastoral system and agroforestry generate more amount of biomass and are more efficient in the sequestration of large amounts of carbon. Carbon farming increases the carbon sinks in soil because of improvement in soil aeration from the addition of organic carbon, which reduces the process of denitrification and increases sink capacity for methane. Carbon stocking helps to increase water-holding capacity and nutrient availability of the soil which ultimately improves soil health. Also, in limited moisture conditions, biochar can efficiently improve soil moisture storage as it has a good amount of recalcitrant carbon. Again, in hot and harsh climate, Jatropha curcas is a well-adapted plant that has the ability to capture 17–25 t of carbon dioxide per hectare per year from the atmosphere (over a 20-year period). A bill to amend the Energy Conservation Act of 2001 was adopted by the Central Government of India in 2022 with the motive to establish voluntary carbon credit trading scheme and market within 2023. Carbon credit is a permit that allows the holding company to release a certain amount of carbon dioxide or any GHG. Conversely, government or any other authorized agency has issued carbon credit certificates to an organization/institution that has helped to remove GHGs from the atmosphere. Gujarat created the first carbon credit trading market in India in 2022. So, one carbon credit certifies that one metric ton of carbon dioxide has been removed from the atmosphere. Despite so many advantages of carbon farming, well-informed advisory services are necessary to motivate farmers to adopt carbon farming practices that will help improve production and soil health.
Article
Coffee agroforestry systems (AFS) have been shown to enhance soil, which results in a positive impact on above and belowground organic inputs. However, the specific temporal impact of coffee AFS to soil carbon and nitrogen pools remains uncertain. Thus, we aim to answer the following questions: how do soil total carbon (TC) and total nitrogen (TN) stocks respond to different coffee cultivation systems? And to what extent can a coffee AFS contribute to replacing the TC?. We evaluated three coffee cultivation systems (agroforestry system with grevillea – AFS; consortium with banana – CBC; and coffee monoculture – CM). We used two reference systems in order to compare these systems: a native forest (NF) and a pasture (PA). Soil samples were collected at six different depths (0–10, 10–20, 20–40, 40–60, 60–80, 80–100 cm). The samples were then analyzed using an elemental analyzer to determine total carbon (TC) and total nitrogen (TN) levels. An isotope ratio mass spectrometer was used to further assess soil composition by measuring the natural abundance of δ13C and δ15N. The two mixed coffee systems had similar TC stocks in the topsoil to those of NF and PA, while CM had the lowest stock. However, only CBC maintained a similar TC stock to NF and PA at 100 cm depth (on average 140.5 Mg ha− 1). TN stocks followed a similar pattern to TC. It was found that over 70% of soil carbon under AFS and PA was derived from C3 plants. In the upper soil layer, AFS and CBC maintain TC and TN stocks in relation to NF. However, when considering the total stock at 100 cm depth, only CBC (compared to AFS) is able to maintain similar levels to NF. Despite this, AFS has great capacity to replace soil organic carbon, replacing more than 50% of C4 in PA.
Research
Full-text available
Final report on the StartClim project “Agroforestry - How trees in the field can contribute to solving the biodiversity and climate crisis”
Article
Agroforestry as a planting strategy for restoration, conservation, and climate change mitigation, can alter plant carbon (C) input and microbial-mediated C output, which in turn affects soil organic carbon (SOC) accumulation. However, a quantitative analysis of the C sequestration potential of agroforestry at the global scale is lacking. Here, by collecting 561 pairs of observations worldwide, we conducted a meta-analysis to quantify the impact of agroforestry on SOC sequestration as well as soil properties and microorganisms. On average, agroforestry increased SOC by 10.7 %, dissolved organic carbon by 10.2 % and CO2 emissions by 10.2 % compared to other land uses (cropland, forest and uncultivated land). Across all climate zones, SOC of agroforestry increased the most in arid areas (18.7 %). Compared with monoculture, agroforestry has more advantages in terms of microbial biomass, diversity and soil nutrient content. Our findings highlight the response of agroforestry SOC sequestration to different land managements and climate zones.
Article
Full-text available
To achieve global food security, we need to approximately double food production over the coming decades. Conventional agriculture is the mainstream approach to achieving this target but has also caused extensive environmental and social harms. The consensus is that we now need an agriculture that can “multi-functionally” increase food production while simultaneously enhancing social and environmental goals, as committed to in the sustainable development goals (SDGs). Farming also needs to become more resilient to multiple insecurities including climate change, soil degradation, and market unpredictability, all of which reduce sustainability and are likely to exacerbate hunger. Here, we illustrate how agroforestry systems can increase yield while also advancing multiple SDGs, especially for the small developing-world agriculturalists central to the SDG framework. Agroforestry also increases resilience of crops and farm livelihoods, especially among the most vulnerable food producers. However, conventional yield-enhancement strategies have naturally dominated the debate on food production, hindering implementation of more multifunctional alternatives. Governments and institutions now have the opportunity to rebalance agricultural policy and investment toward such multigoal approaches. In doing so, they could achieve important improvements on multiple international commitments around the interlinked themes of food security, climate change, biodiversity conservation, and social well-being.
Article
Full-text available
Canadian agriculture is a source of greenhouse gases (GHG) and agroforestry has the potential to sequester carbon (C), and mitigate agricultural GHG emissions. Agroforestry systems are common features in Canada's agricultural landscape; however, there are limited empirical data to support implementation of agroforestry practices for GHG mitigation. This shortfall of data may be a contributing factor to the lack of policy that supports the use of agroforestry for GHG mitigation in the Canadian agricultural landscape. We reviewed published studies that compared C stocks in vegetation and soils, and/or GHG emissions in agroforestry systems to traditional agriculture across Canada, with the aims of assessing the benefit of adopting agroforestry for GHG reduction. We then identified data gaps and obstacles that could direct future research. We found that most studies reported increases in vegetation and soil organic C storage in areas with woody species compared to herbaceous crops. Agroforestry systems also reduced the emission of CH4 and N2O, and increased CO2 respiration from soil, but few studies have examined these gases. The small set of studies we reviewed demonstrated the potential of agroforestry to store terrestrial C and mitigate GHG emissions. However, additional research is required to verify this pattern across geographic regions, determine the regional potential for development of agroforestry systems, and assess the potential atmospheric GHG reduction at regional and national scales.
Article
Intercropping and agroforestry are mixed plant species cultivation systems that can potentially reduce pressure on land and water resources by generating higher crop yields and by increasing resource use efficiencies through exploitation of complementarities between species. While it is frequently claimed that mixed cultivation systems remain popular in Chinese agriculture, there are also reports that it is in decline. Little quantitative evidence is available on the prevalence of intercropping and agroforestry. Therefore, we conducted a village-level survey in 68 villages across six provinces in China in 2014 to document the prevalence of these mixed species cultivation systems in 2009 and 2014. We found that intercropping was practiced on approximately three percent of the arable land in the surveyed villages, while agroforestry was practiced on approximately one percent of the arable land and one percent of the area of plantation plus forest land. The use of both systems did not significantly change between 2009 and 2014. An explorative village-level analysis of factors associated with mixed species cultivation practices reveals a significant positive association with labour availability, and a smaller, but mostly significant, negative association with agricultural machinery power. These survey results provide the first systematically collected quantitative evidence on the current use and trend of mixed cultivation systems in China. The collected evidence indicates that these systems are currently used on a small proportion of the land, but have not declined in recent years. Intercropping continues to provide pathways for ecological intensification of agricultural food production.
Article
Agroforestry is increasingly regarded as an important adaptation and mitigation strategy against climate change. In particular, the use of fertilizer trees has been promoted as a practice that contributes to improved soil fertility through nitrogen fixation, by increasing supply of nutrients for crop production. While a lot of the evidence on the impact of fertilizer trees relies on on-farm experiments and correlational analysis, there is a paucity of rigorous evidence under actual smallholder farming conditions. This paper analyzes the impacts of adopting fertilizer trees such as Gliricidia sepium and Faidherbia albida on household food security. We draw on survey data of 338 farmers in Malawi and use an endogenous switching regression to rigorously analyze adoption impacts. Econometric results show that use of fertilizer tree adoption increases the value of food crops by 35%. Disaggregation of the impacts through stratification by land ownership further reveal that farmers with smaller farms of up to 2 acres realize the highest gains. Furthermore, fertilizer tree use in conjunction with improved maize seed also significantly increased value of food crops. This study offers preliminary insights that contribute to an emerging field of research on quantitative assessment of agricultural interventions such as agroforestry practices using novel analytical approaches. We provide some policy insights and recommend the need for future research to be designed around development initiatives that consider fine-scale variation in social, economic and ecological context of farmers to improve uptake and adaptation to realize the full potential of agroforestry in improving soil fertility and household food security.
Article
Rubber-based agroforestry (Hevea brasiliensis) systems are considered the best way to improve soil properties and the overall environmental quality of rubber monoculture, but few reports have examined soil aggregate stability in such systems. The objective of this study was to examine the management and landscape effects on water stable soil aggregates, soil aggregate-associated carbon, nitrogen content and soil carbon, and nitrogen accumulation in Xishuangbanna, southwestern China. Treatments were rubber monoculture (Rm) and four rubber-based agroforestry systems: H. brasiliensis–C. arabica (CAAs), H. brasiliensis–T. cacao (TCAs), H. brasiliensis– F. macrophylla (FMAs) and H. brasiliensis–D. cochinchinensis (DCAs). The results showed that, with the exception of CAAs, the rubber-based agroforestry treatments significantly increased total soil organic carbon (SOC) and N contents and enhanced the formation of macroaggregates compared to the rubber monoculture treatment. SOC and N contents in all water-stable aggregate fractions were significantly higher in rubber-based agroforestry systems (except CAAs) compared to rubber monoculture. The macroaggregate fractions contained more organic carbon and nitrogen than the microaggregate fractions. The proportions of C and N loss from slaking and sieving were shown to have significantly negative correlations with the mean weight diameter and the SOC and N concentrations in bulk soil. The results suggest that soil surface cover with constant leaf litter fall and extensive root systems in the rubber-based agroforestry systems increased soil organic carbon and nitrogen, helped improve soil aggregation, reduced soil erosion, decreased carbon and nitrogen loss, and ultimately improved the carbon and nitrogen accumulation rates. Given that the soil physical-chemical properties improvement and the patterns of the intercropping system played key roles in managing artificial forests, we recommend that local governments and farmers should prefer T. cacao, F. macrophylla and D. cochinchinensis and not C. arabica as the alternative interplanted tree species within rubber plantations.