Regional Environmental Change (2021) 21: 44

https://doi.org/10.1007/s10113-021-01776-0

ORIGINAL ARTICLE

Carbon and nutrient cycling in tree plantations vs. natural forests:

implication for an efficient cocoa agroforestry system in West Africa
Michel K. Yao 1 & Armand W. Koné 2 & Abigael N. Otinga 3 & Emmanuel K. Kassin 4 & Yao Tano 5

Received: 22 April 2020 / Accepted: 10 April 2021 / Published online: 30 April 2021
# The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021

Abstract
Nutrient cycling has been widely studied in tree plantations. However, studies on the observed negative changes relative to natural
vegetations and how these could be capitalized toward setting up ecofriendly agroecosystems are rare. This study was conducted in
Central-West Côte d’Ivoire to establish the changes in carbon and macronutrient cycling occurring in full-sun cocoa (Theobroma
cacao L.) and Teak (Tectona grandis) relative to primary forest and subsequently make cocoa agroforestry-based recommendations.
Leaf litterfall and associated carbon and macronutrient inputs, rates of leaf litter decomposition, and macronutrient release were
assessed. In addition, soil (0–10 cm depth) chemical and microbial parameters were evaluated. Litterfall yields were 10.6 ± 2.0, 9.3 ±
0.8, and 10.1 ± 0.4 Mg dry mass ha−1 year−1 in forest, cocoa, and teak, respectively. Compared to the forest, the cocoa plantation
supplied lower inputs of C (−736 kg ha−1 year−1 or −15.5%) and N (−75 kg ha−1 year−1 or −27%), similar P, but greater K. Similar
quantities of C and N were recorded in the teak plantations and the forest. However, the teak plantation supplied lower K but higher
P inputs than the forest. Cocoa leaf litters decomposed at the same rate as those of the forest (k = 0.3 month−1) but faster than the
teak’s, the initial leaf litter N:P ratio being the most influencing factor. Except for P, the macronutrient release from cocoa and forest
litters exhibited similar patterns and rates, which were significantly different from those observed in the teak leaf litter. Soil C
mineralization rate and mineral N concentration drastically declined in both tree plantations, the greatest gap occurring in cocoa
(Cmin: −40%, mineral N: –54.2%) due to lower litter Ca input and soil acidity. Teak appears to be a good candidate for shade as it
may compensate for the deficit in litter C (and N, to a lesser extent) supply exhibited by cocoa relative to the forest. The trends in
microbial activity underscore the need to grow cocoa in association with trees that provide quality litter materials, for improved
cocoa nutrient availability and faster C storage in soil. In line with this, some suggestions were made and discussed. This study can
be used in support of developing an efficient cocoa agroforestry system in West Africa.

Keywords Agroecological transition . Cocoa agroforestry system . Complementary plant traits . Litter decomposition . Soil
acidity . Soil–litter nutrient translocation

Communicated by Anne Bousquet-Melou

* Armand W. Koné 1
UFR Sciences et Gestion de l’Environnement, UR Gestion Durable
konearmand@yahoo.fr; konearm.sn@univ-na.ci des Sols, Université Nangui Abrogoua, 02 BP 801, Abidjan 02, Côte
d’Ivoire
Michel K. Yao 2
UFR Sciences de la Nature, UR Gestion Durable des Sols, Université
michel.yaok@gmail.com Nangui Abrogoua, 02 BP 801, Abidjan 02, Côte d’Ivoire
3
Department of Soil Science, School of Agriculture and
Abigael N. Otinga
Biotechnology, University of Eldoret, P.O. Box 1125-30100,
abigaelno@uoeld.ac.ke
Eldoret, Kenya
4
Emmanuel K. Kassin Centre National de Recherche Agronomique, Programme Cacao,
emmanuel.kassin@mdlz.com Station de Divo, BP 808, Divo, Côte d’Ivoire
5
UFR Biosciences, Laboratoire de Zoologie, Université Félix
Yao Tano Houphouët-Boigny de Cocody, 22 BP 582, Abidjan 22, Côte
tanoy03@yahoo.fr d’Ivoire
44 Page 2 of 14 Reg Environ Change (2021) 21: 44

Introduction (Sauvadet et al. 2020; Koné and Yao 2021) while improving
the cocoa physiology (Vanhove et al. 2016).
Natural forest conversion to agricultural land has been shown Litter production is a major ecological process in transfer-
to disrupt the carbon and nutrient cycles, with consequences ring carbon and nutrients from above ground plant parts to
on soil quality and climate (van Straaten et al. 2015; Han and the soil through decomposition (Fontes et al. 2014; Naik et al.
Zhu 2020). Greater drawbacks are associated with 2018; Asigbaase et al. 2021b). Litter decomposition is con-
monospecies with respect to multispecies tree plantations trolled by three main factors - environmental conditions, de-
worldwide; however, proper management of carbon in con- composer community, and substrate quality - the latter being
verted lands can reverse many of the effects of deforestation a major determinant at small scales, i.e., within the sites
(Liu et al. 2018; Veldkamp et al. 2020). One promising way is (Fontes et al. 2014). Any study focusing on carbon and nu-
the implementation of agroforestry systems (AFS) adapted to trient cycling should not be restricted to leaf litter because
local contexts (Jagoret et al. 2017; Bayala et al. 2020). To significant elements are embedded in other parts of the plant
design better AFS interventions, diagnostic field works focus- (stems, roots, etc.). However, these may be immobilized in
ing on the changes in C and nutrient cycling in converted the coarse parts or excluded from cycling in the short term
lands relative to the natural ecosystems are necessary, since (Fontes et al. 2014). Thus, soil fertility and functioning are
this would highlight potential negative trends to be curbed maintained mostly owing to litters, which is crucial in the low
(Blaser et al. 2017). fertility-tropical soils, particularly under low-fertilization
Cocoa—Theobroma cacao L. (Sterculiaceae)—is a major small-holder cocoa farms (Hartemink 2005; Ofori-
cash crop in many tropical countries, particularly in Africa. In Frimpong et al. 2007; Naik et al. 2018). Importantly, the litter
West Africa, some 6 million ha of forest zones are planted layer serves as a habitat and source of energy to soil fauna
with cocoa (Wessel and Quist-Wessel 2015); this contributes thereby promoting biodiversity (Bisseleua et al. 2009) and
to about 70% of the total world production (ICCO 2018). Côte activity of soil microorganisms which underpins plant pro-
d’Ivoire is the world’s first cocoa producing country with ductivity (Barrios et al. 2018). In West Africa, research
about 2,000,000 t in 2017 or up to 40% of the global cocoa works that have focused on cocoa leaf litterfall and decom-
market (ICCO 2018). The total area planted with cocoa in that position are related only to shaded-cocoa plantations where
country was about 2.5 million ha in 2013 (FOAStat 2015). In litterfall consists of mixed leaf litters, with shade tree’s leaf
the past, cocoa was grown in association with selected native litter sometimes being greater than that of cocoa (Hartemink
shade trees (Ofori-Frimpong et al. 2007) and/or with fruit 2005). The share of the shade tree leaf litter may amount to
plants (Koko et al. 2013). Today, almost monocropped, 60–70% of the total (Ofori-Frimpong et al. 2007; Saj et al.
“full-sun” cocoa (FSC) and cocoa farms with very few shade 2021). Furthermore, in these studies, only leaf litter break-
trees are dominant in West Africa (Ruf 2011; Andres et al. down is examined; the greater segment of element recycling
2016). In Côte d’Ivoire, FSC was reported to cover 35% of the that occurs at finer level through mineralization is in general
total cocoa area at the end of the twentieth century (Freud and overlooked, both in soil surface litter and soil (Asigbaase
Richard 2000). This cocoa farming system is now adopted by et al. 2021b). Yet, this microbial-mediated process is respon-
the large majority of farmers in the Central- and South- sible for most nutrient transformations in soil, regenerating
Western parts of the country, the main cocoa production zones minerals that may limit plant productivity (Delgado-
with 40% of the national production (Wessel and Quist- Baquerizo et al. 2016). Because cocoa leaf litterfall may hold
Wessel 2015). This is a consequence of earlier research results 10 to 45% of the total macronutrient stock (Hartemink 2005)
that show FSC as being more productive than the shaded and up to 17% of the total carbon stock (N’Gbala et al. 2017)
cocoa (Ahenkorah et al. 1987). However, this resulted to ris- in the cocoa vegetation, addressing the mineralization of that
ing ecological and agronomic concerns among them faster soil biomass pool is essential to the agroecosystem functioning.
degradation (Adeniyi et al. 2017; Asigbaase et al. 2021a), To date, unfortunately, very few studies have investigated
early degeneration and death of cocoa trees from about 20 the impacts of cocoa on the soil microbial activities in
years apparent in Côte d’Ivoire and Ghana (Wessel and Africa (Sauvadet et al. 2020). The few ones conducted in
Quist-Wessel 2015). Today, faced with the limitations of Latin America were restricted to the microbial carbon, mi-
FSC cultivation, an agroecological transition to sustainable crobial community structure, and few enzyme activities
cocoa farming through cocoa agroforestry systems (cAFS) (Alfaro-Flores et al. 2015; Buyer et al. 2017). Although thou-
in West Africa is essential particularly when the world choc- sands of hectares of timber plantations have been established
olate demand is expected to at least double by 2050 (Bisseleua in West Africa, including teak—Tectona grandis
et al. 2009). Indeed, a well-fitted integration of trees into co- (Lamiaceae), there is scarcity of literature on teak’s leaf litter
coa farms has the potential to counteract, even reverse the soil decomposition (Attignon et al. 2004; Swarnalatha and
organic matter (SOM) and nutrient depletion through input of Reddy 2011) and shortcomings in terms of soil functioning
complementary-trait litters from the associated shade species (Limtong et al. 1998; de Gannes et al. 2016).
Reg Environ Change (2021) 21: 44 Page 3 of 14 44

Although nutrient cycling has been widely studied in tree buffer zone (ESM 1). For the purpose of this study, five pri-
plantations, studies on the observed negative changes relative mary forest, five teak and five cocoa stands were randomly
to natural vegetations that could be capitalized toward setting demarcated in the three aforementioned areas, respectively.
up ecofriendly agroecosystems are rare. This study explores The characteristics of the land-use systems (LUS) are as
how C and nutrient inputs and cycling in soil are modified follows:
when natural forests are converted into FSC and timber tree
plantations in Central-West Côte d’Ivoire, to make proper – the primary forest stands harbored diverse types and spe-
agroforestry-based recommendations to mitigate potential cies of plants, the most abundant being Cedrela odorata
negative trends. For this purpose, annual C and nutrient inputs 23.5% (Meliaceae), Triplochiton scleroxylon 7.5%
via leaf litterfall over 1 year, as well as patterns of litter de- (Sterculiaceae), Prieuriana pentandra 7.4%
composition and nutrient release, were examined. Further, the (Meliaceae), Mansonia altissima 5.9% (Sterculiaceae),
microbial biomass carbon and the rate of carbon mineraliza- and Ceiba pentandra 5.9% (Bombacaceae), in addition
tion in the topsoil (0–10 cm depth) were assessed. The main to leguminous and herbaceous species. The basal area,
hypotheses are as follows: (1) carbon and macronutrient in- biomass and density of the forest stands were 22 m2
puts would decrease in FSC plantations, calling for the inclu- ha−1, 166 Mg ha−1 and 214 trees ha−1, respectively
sion of shade trees to meet the deficit of carbon supply relative (N’Drin 2018).
to the natural forest and to compensate for annual nutrient – the teak stands were 12-year old, with a density of 850
exports through cocoa beans; (2) carbon flow in the teak stand trees ha−1. The trees were planted following clearance of
would mimic that in the natural forest (Blaser et al. 2017; degraded forest areas.
N’Gbala et al. 2017), making this tree a good candidate for – cocoa fields were mature, 10–20 years, the main produc-
shade and carbon sequestration in cAFS; and (3) soil micro- tive stage. The trees were planted at a density of 1320
bial biomass C and mineralization activities would decline in plants ha−1 (3 m × 2.5 m) after forest clearance. The
the monocultures due to the recalcitrant low-quality leaf litter fields, 2-ha size in average, essentially belong to small-
they shed, calling for the inclusion of shade trees that have holders and were not fertilized.
potential of providing quality litter in cocoa fields (Gnahoua
et al. 2013; Faboya et al. 2015). In each replicate stand, a representative frame of 100 m ×
50-m size (0.5 ha) was demarcated, where measurements were
done on a 100-m transect superimposed on one diagonal.
Material and methods
Leaf litterfall and standing leaf litter stock
Study area determination

The study was conducted around the village of Goulikao Litterfalls were collected monthly over a 12-month period
(6°30′N, 5°31′W, 200 m asl) situated at the outskirts of Téné (August 2006 to July 2007) using 1m × 1m × 0.2m littertraps
National Forest, in the Departement of Oumé, mid-West Côte raised at 0.8 m above the floor of the selected stands. In each
d’Ivoire (ESM 1). The natural vegetation is the Guinean stand, three pseudoreplicate-traps were set at three distinct
semideciduous forest, stretching from the republic of Guinea collecting points, separated by 50-m intervals along the des-
to Cameroon. The climate falls under the subequatorial type, ignated transect. In total, 15 littertraps were used in each LUS.
with average temperature and hygrometry of 26°C and 85%, Collected leaves were then oven-dried at 50°C to constant
respectively. The region experiences a bimodal rainfall pattern weight. Annual litterfall for a given plot was obtained by
(ESM 2) with two dry and two wet seasons, the mean annual summing monthly masses. Concommitantly, soil surface leaf
rainfall being 1275 mm. Soils for this area are Oxisols accord- litter was sampled monthly using a 1-m2 frame demarcated
ing to the American classification system (Soil Survey Staff near the littertrap. Dry surface leaf litter mass determination
1999) with geological material consisting of crystalline rocks, was performed as described above.
essentially granite, and few schist and basic intruding rock
formations. In general, soils are clayey (30–60%) and acidic Litter decomposition monitoring
(pH 4–6.5), with 20–30 g OC kg−1 soil (Kassin and Yoro
2009). Because cocoa and teak plantations were above 10 years, tree
canopies overlapped, thereby limiting the growth of
Stand selection and description understorey plant species. The decomposition trials were thus
restricted to the respective dominant leaf litter species. For
The study took place within a 400 ha-grid extending from the this, two complementary methods were used: (1) the direct
Téné Protected Forest to the agricultural landscape, via a decomposition on the soil surface, which allows for the
44 Page 4 of 14 Reg Environ Change (2021) 21: 44

determination of the exact in-situ rate of leaf litter dry mass surface. Measurements were restricted to the 0–10-cm depth
loss and (2) the decomposition in litterbags that additionally because this is in direct contact with the litter layer and harbors
allows tracking the nutrient dynamics in decomposing litters. the bulk of the biological processes in tropical soils (Sauvadet
The decomposition experiments spanned two seasons, rainy et al. 2020; Ochoa-Hueso et al. 2021). Then, the ten
and dry. pseudoreplicate samples were thoroughly mixed into a single
Direct decomposition, expressed as monthly percent dry composite sample of which one part was kept fresh at 4°C in a
mass loss on each LUS floor was monitored over 9 months. refrigerator for microbial biomass-C (MBC), carbon mineral-
The dry mass loss was calculated using both litterfall masses ization rate (Cmin), and mineral N concentration, and the other
collected from littertraps and standing litter stock, as follows part was air-dried at room temperature, crushed and sieved (2-
(Loranger et al. 2002): mm) for chemical analyses.
Dð%Þ ¼ 100ðL1 þ C−L2Þ=ðL1 þ CÞ
Laboratory analyses of soil and leaf litter samples
where L1 and L2 are soil surface litter masses at the times T1
and T2, and C the mass of the leaves fallen between the two For both the fresh and decomposing leaf litter materials,
times (1-month interval). phosphorus was determined according to the procedure
Decomposition in litterbags was monitored over 8 months described by Murphy and Riley (1962) while the major cat-
using nylon litterbags of 20 cm × 20 cm size and 2-mm mesh, ions were extracted using ammonium acetate (buffered at pH
which is small enough to limit losses of leaves while large 7) before their concentration was established by means of
enough to enable access of micro- and macroorganisms to atomic absorption spectrophotometry techniques (VARIAN
the packed litter (Loranger et al. 2002). For each replicate- SPECTRAA 220 SF model) (Anderson and Ingram 1993).
stand, 24 litterbags were loosely filled with 10 g of dry leaf Total C and nitrogen were determined on both leaf litter
litter, of which eight were then installed on the floor around (4 mg dried, ground and ˂ 0.1 mm sieved) and soil (25 mg
each littertrap. In total, there were 120 litterbags per LUS. The dried, ground and 0.2 mm sieved) samples by dry combustion
litterbags were fixed with a short wire to prevent movement. procedures using a CHN autoanalyzer (EA1112 Thermo
Three litterbags were subsequently retrieved monthly from Finnigan Series, France). It is worth mentioning that soil total
each replicate-stand (one from each litterbag point). After re- C corresponds to organic C (SOC) in the soils of this study.
trieval, the litterbags were placed in separate paper bags and Soil pHwater (1:2.5 w/v) and texture were determined using a
transported to the laboratory. Adhering soil particles and or- glass electrode and the Robinson pipette method, respectively.
ganic debris were removed gently using tweezers, and then Soil bulk density was determined on core samples collected
litter detritus were oven-dried at 50°C for 48 h and weighed, using the cylinder method, at the 10 soil sampling points. The
ground, sieved and stored in plastic bags for chemical analy- fresh and dry weights of the soil cores were taken, the latter
ses (Loranger et al. 2002). The rates of decomposition in the after oven drying at 105°C for 48 h, to calculate both moisture
litterbags were obtained by fitting mass loss to the single ex- and bulk density. All these analyses were done according to
ponential model (Olson 1963): the methods described by Anderson and Ingram (1993).
Soil MBC was determined using the chloroform
Wt ¼ W0 e−kt fumigation–extraction method (Vance et al. 1987). For each
replicate site, a 20-g soil sample was fumigated with an
where k is the decomposition coefficient, an estimate of the ethanol-free chloroform solution while a 10-g unfumigated
decomposition rate; W0 is the initial leaf litter mass; and Wt is sample was kept in a desiccator and placed in a dark room.
the remaining litter mass after a given time span t (in months). After a 24-h incubation period, carbon from both fumigated
The time required for 50% (half-life time) and 95% decompo- and control samples was extracted with 0.5M K2SO4 solution
sition of leaf litters were computed as t50 = 0.693/k and t95 = after shaking, centrifugation and filtration of the soil suspen-
3/k, respectively (Olson 1963). sion. Then the dissolved organic carbon (DOC) of the soil
solutions was determined by measuring chemical oxygen de-
Soil sampling mand (COD) with a spectro-colorimeter DR/700 (HachTM
method). The MBC was determined as the difference between
Soil sampling was done at the end of the wet season according DOC from fumigated and unfumigated soil samples. The
to the procedures described in the Tropical Soil Biology and MBC/SOC ratio, which is a strong driver of SOM build-up
Fertility (TSBF) manual (Anderson and Ingram 1993). This is (Kallenbach et al. 2016) was also determined.
a period where soil biological activity is maximum. Soil sam- Soil Cmin was measured for each plot in triplicate according
ples were collected from the top 0–10-cm depth at ten points to the method decribed by Nacro et al. (1996). Each soil rep-
separated by 10-m intervals along the designated transect, licate (15 g equivalent dry mass) was rehumidified at 80%
after plant residues have been carefully removed from the soil water-holding capacity and incubated in a 130-ml jar at 28 ±
Reg Environ Change (2021) 21: 44 Page 5 of 14 44

0.5 °C for 3 days in the dark. Air samples from incubated soil
samples were analyzed for CO2 efflux through direct injection
into a Chromatograph (Chrompack CP-2002 P Micro GC)
equipped with a thermal conductivity detector
(nanocatharometer). The results were expressed as μg CO2-
C g−1 soil. The metabolic quotient (qCO2), also termed as
"specific respiration", was determined as cumulative 3-day
respiration per unit MBC. Mineral N (Nmin) was obtained by
determining N-NH4+ and N-NO3- + N-NO2- concentrations in
the incubated soil samples (Franzluebbers et al. 1996).

Data analyses

Carbon and nutrient inputs through litterfall (kg ha−1 year−1)

were obtained by multiplying the concentration of the element
(g kg−1 dry litter) by the annual litterfall (kg ha−1 year−1). The
remaining nutrient stocks in decomposing litters were obtain-
ed by multiplying the nutrient concentration (% dry litter) by
the remaining litter mass (g). Mean differences in the litter and
soil parameters were assessed using one-way ANOVA after
variances were tested for homogeneity using Levene’s test.
Where differences were significant, means were separated
using the least significant difference (LSD) test. Pearson’s
correlation was used to explore relationships between soil mi-
crobial variables (MBC, Cmin, MBC/SOC and qCO2) and the
chemical characteristics of (1) leaf litter and (2) soil. Results
were considered significant when p values < 0.05. These sta-
tistical analyses were performed using the R ver. 4.0.3 soft- Fig. 1 Annual leaf litterfall mass (a) and mean monthly standing leaf
ware (http://www.r-project.org/). litter stock (b) across the land-use systems. Horizontal bold lines of the
boxes indicate the median; the lower and upper bounds of the box repre-
sent the 25th and 75th percentiles, respectively. The vertical dotted bars
include all values. Different letters indicate significant difference between
Results land-use systems (least significant difference (LSD) test) at the 005 level

Annual yields of leaf litterfall and standing leaf litter

stocks LUS followed the same trend as C concentration (F = 7.5; p =
0.007). Remarkably, P concentration was twice as high in the
Average annual litterfall yields did not differ significantly (p ˃ teak litter (1.3 ± 0.2 g kg−1) compared to the cocoa (0.6 ± 0.0 g
0.05) across the LUS (Fig. 1a), contrary to the monthly stand- kg−1) and natural forest litters (0.7 ± 0.1 g kg−1). K concen-
ing leaf litter mass (F = 5.6; p = 0.02), with the teak plantations tration was significantly greater (F = 15.8; p ˂ 0.001) in the
ranking highest and the forest and cocoa plantations remain- cocoa litter (5.8 ± 0.5 g kg−1) compared to the forest and teak
ing close to each other (Fig. 1b). litters (4.1 ± 0.3 and 3.1 ± 0.2 g kg−1, respectively).
Leaf litter C:N ratio was greater (F = 12.4; p = 0.001) in the
Carbon and nutrient inputs through leaf litterfalls teak (32.2 ± 0.7) followed by cocoa plantation (19.6 ± 0.7)
relative to the natural forest (16.8 ± 0.9). Conversely, the leaf
Carbon concentration in the teak litter (526.2 ± 3.1 g kg−1) litter N:P ratios in the teak (20.3 ± 2.6) and cocoa (34.5 ± 2.0)
was significantly higher (F = 5.7; p = 0.02) compared to the plantations were respectively lower (F = 7.9; p = 0.006) than
concentration in both the cocoa and natural forest litters (432.0 that in natural forest (41.4 ± 5.8).
± 6.0 and 453.0 ± 5.5 g kg−1, respectively). However, the litter Annual leaf litterfall C and N inputs significantly declined
C concentration in the latter two did not differ significantly. in cocoa (−736 kg C ha−1 or −15.5% C; −75 kg N ha−1 or
The concentrations of N in the litter of both cocoa and teak −27% N) compared to the natural forest whereas inputs via
plantations (23.6 ± 0.7 and 22.1 ± 0.7 g kg−1) were signifi- teak litterfall was similar to that of the natural forest (Table 1).
cantly lower (F = 7.4; p = 0.008) than that of the natural forest Annual leaf litter P input in cocoa did not significantly differ
(27.1 ± 1.3 g kg−1). The concentration of P in the different from that in the natural forest whereas teak supplied greater
44 Page 6 of 14 Reg Environ Change (2021) 21: 44

Table 1 Annual carbon and

selected nutrient inputs (mean ± Nutrient inputs Land-use systems Anova
standard error) through leaf
litterfall, and related Anova (kg ha−1 year−1) Cocoa Teak Primary forest F value p value
results
C 4022.9 ± 366.7 b 5287.2 ± 204.4 a 4759.0 ± 476.9 ab 3.9 0.05
N 206.5 ± 21.5 b 239.2 ± 10.7 ab 281.5 ± 16.6 a 5.0 0.03
P 5.9 ± 0.3 b 12.7 ±1.8 a 7.3 ± 1.1 b 8.5 0.005
K 55.1 ± 9.4 a 31.7 ± 2.9 a 44.5 ± 7.1 a 2.7 0.1
Ca 198.0 ± 20.7 b 215.1 ± 20.3 b 360.2 ± 54.4 a 6.3 0.01
Mg 56.7 ±3.4 a 27.5 ± 2.2 c 41.7 ± 5.9 b 12.5 0.001

In a given row, means with the same letters are not significantly different at α = 5%

amount (two-fold) of P compared to forest. With regard to K The litter mass remaining at the EDP was negatively cor-
and Mg, cocoa supplied greater quantities than forest, contrary related to the initial leaf litter N:P ratio (r = 0.6; p = 0.03)
to teak. Both tree plantations supplied lower Ca relative to whereas the k-value positively correlated with initial leaf litter
forest. N (r = 0.5; p = 0.04).

Leaf litter decomposition rate Macronutrient dynamics in decomposing leaf litters

Direct decomposition rates on the LUS floors were 27.8 ± 4.6, The litter N exhibited similar dynamics of release in cocoa and
26.3 ± 5.5 and 23.4 ± 4.2% month−1 for forest, cocoa, and teak forest (Fig. 3a); the stock of N decreased abruptly during the
plantations, respectively. These rates, however, did not differ first month, then slowed down. In the teak litter however, the
significantly among the three LUS. release of N was delayed by 2 months and then steadily de-
Decomposition in the litterbags showed two distinct creased until the EDP. At the end, the greatest proportion of
patterns (Fig. 2). The first, observed with forest and released N occurred in cocoa litter, followed by forest litter,
cocoa litters exhibiting identical dynamics, consisted of and the least in teak litters; the difference being significant (F
two phases: an abrupt mass loss followed by a slower = 4.0; p = 0.04).
one. Teak litter mass loss was rather linear and signif- The dynamics of P release in decomposing teak leaf litter
icantly slower. At the end of the 8-month decomposi- was identical to that in forest: P accumulated during the first
tion period (EDP), the remaining litter masses were all month and then abruptly decreased in the next 2 months (Fig.
lower than 50% of the initial mass. Remarkably, the 3b). A distinct pattern was observed in cocoa leaf litter, where
remaining mass of cocoa litter was similar to that of P showed a steady decrease from the first month. However, at
forest litter and half of the remaining mass of teak litter the EDP, the proportions of released P were close to each
(F = 6.1; p = 0.01). Furtheremore, forest and cocoa other.
litters showed identical k-value, which was greater than K release in the decomposing cocoa litter followed the
teak’s k-value (Table 2). Accordingly, the half-life time same pattern as that in the forest leaf litter consisting of a
(t50) and t95 of cocoa and forest litters were significantly gradual decline over time (Fig. 3c). In teak litter however, a
shorter than those of teak. contrasting pattern was observed: a sharp decrease of K was

Table 2 Leaf litter decay

parameters (mean ± standard Decay Land-use systems Anova
error) across the land-use sys- parameters
tems, and related Anova results Cocoa Teak Primary forest F p
value value

Leaflitter Wt = W0 ℮−0.3t Wt = W0 ℮−0.2t Wt = W0 ℮−0.3t – –

decay model
k-value 0.3 ± 0.0 a 0.2 ± 0.0 b 0.3 ± 0.0 a 4.2 0.04
(month−1)
t50 (month−1) 2.8 ± 0.3 b 4.1 ± 0.5 a 2.6 ± 0.1 b 6.0 0.01
t95 (month−1) 12.1 ± 1.2 b 17.6 ± 2.0 a 11.4 ± 0.4 b 6.0 0.01

W0 is the initial leaf litter mass; Wt is the remaining litter mass after a given time span t (in months)
In a given row, means with the same superscript letters are not significantly different at α = 5%
Reg Environ Change (2021) 21: 44 Page 7 of 14 44

12
Litter Ca exhibited almost similar dynamics of release in
Forest cocoa and forest (Fig. 3d). The stock of Ca decreased abruptly
10 Cocoa during the first 2 months, stabilized for one month and then
Remaining litter mass (g)

Teak consistently decreased until the EDP. In teak litter, the Ca

8
decreased slightly during the first month, accumulated over
6 the next 2 months and then, it decreased until the EDP. The
proportion of Ca released from forest and cocoa leaflitters
4 were close to each other and twice as high as that from teak
litter (F = 24.9; p ˂ 0.001).
2 Mg release from decomposing litters followed the same
trends as K both over time and across the LUS (Fig. 3e).
0
I Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. The proportions of Mg released varied across litters (F =
18.1; p ˂ 0.001), the greatest observed with cocoa and forest
Fig. 2 Monthly leaf litter mass loss in the litterbags across the land-use
systems. Bars represent standard errors. I, time of litterbag installation and the least with teak.

Soil chemical and microbial attributes

observed in the first month followed by a sudden increase
which stabilized in the next 2 months and then decreased Lower concentrations of SOC (F = 5.7; p = 0.02) and total N
abruptly untill the EDP. The proportions of K released varied (F = 4.2; p = 0.04), and higher C:N ratios (F = 10.3; p = 0.002)
significantly across litters (F = 6.2; p = 0.01), the greatest were observed in both the tree plantations relative to the forest
occurring in cocoa and forest litters and the least in the teak. (Fig. 4a, b, and c). Mineral N declined in the tree plantations

Fig. 3 Nutrient dynamics in 125 150

decomposing leaf litters. Bars (a) (b)
125
Remaining N stock (%)

Remaining P stock (%)

represent standard errors. I, time 100
of litterbag installation 100
75
75
50
50
Forest
25 Cocoa 25
Teak
0 0

125 150
Remaining K stock (%)

100 (c) 125 (d)

Remaining Ca stock (%)

100
75
75
50
50
25 25

0 0
I Sept. Oct. Nov. Dec. Jan.

150
(e)
Remaining Mg stock (%)

125

100

75

50

25

0
I Sept. Oct. Nov. Dec. Jan.
44 Page 8 of 14 Reg Environ Change (2021) 21: 44

Fig. 4 Soil physical and chemical parameters across the land-use types. respectively. Vertical dotted bars include all values. Different letters in-
Horizontal bold lines of the boxes indicate the median; the lower and dicate significant difference between land-use systems (least significant
upper bounds of the boxes represent the 25th and 75th percentiles, difference (LSD) test) at the 005 level

(F = 10.9; p = 0.002) relative to the primary forest (Fig. 4d). b). This was however not the case for Cmin (F = 4.2; p = 0.04)
Further, soils under the cocoa plantations were acidic (F = and qCO2 (F = 5.2; p = 0.02), which significantly declined in
13.4; p = 0.001) whereas those under the forest and teak plan- the tree plantations, especially cocoa (Fig. 5c and d).
tations showed slightly alkaline conditions (Fig. 4g). The rest Of the leaf litter and soil physico-chemical parameters con-
of the soil parameters (Fig. 4e, f, and h) showed similar levels sidered, only the leaf litter N:P ratio positively correlated with
across the LUS. soil MBC (Table 3). Further, the MBC:SOC positively corre-
Soil MBC (F = 0.7; p = 0.5) and MBC:SOC ratio (F = 1.6; lated with leaf litter cation concentrations (K and Mg), and
p = 0.2) did not vary significantly across the LUS (Fig. 5a and negatively with SOC, total soil N, soil pH, and the clay+fine

Fig. 5 Soil microbial attributes

across the land-use types.
Horizontal bold lines of the boxes
indicate the median; the lower and
upper bounds of the boxes repre-
sent the 25th and 75th percentiles,
respectively. Vertical dotted bars
include all values. Different let-
ters indicate significant difference
between land-use systems (least
significant difference (LSD) test)
at the 005 level
Reg Environ Change (2021) 21: 44 Page 9 of 14 44

Table 3 Pearson’s correlation (r)

between chemical and physical Soil–litter system Sol microbial parameters
characteristics of the soil–litter
system and soil microbial MBC (mg kg−1) Cmin (mg CO2 kg−1) MBC/SOC qCO2
attributes
Litter chemical parameters
C (g kg−1) −0.29 ns −0.06 ns −0.35 ns 0.20 ns
N (g kg−1) 0.12 ns 0.24 ns −0.34 ns 0.120 ns
P (g kg−1) −0.33 ns −0.21 ns −0.36 ns −0.13 ns
K (g kg−1) 0.18 ns −0.30 ns 0.56 * −0.36 ns
Ca (g kg−1) 0.24 ns 0.68 * −0.10 ns 0.43 ns
Mg (g kg−1) 0.13 ns −0.26 ns 0.53 * −0.40 ns
N:P 0.05 * 0.26 ns 0.36 ns −0.00 ns
C:N −0.26 ns −0.17 ns 0.01 ns −0.08 ns
Soil chemical parameters
SOC (g kg−1) 0.24 ns 0.5 * −0.57 * 0.28 ns
Total N (g kg−1) 0.21 ns 0.5 * −0.58 * 0.32 ns
C:N −0.15 ns −0.6 * 0.49 ns −0.42 ns
pH 0.03 ns 0.6 * −0.58 * 0.58 *
Soil physical parameters
Bulk density (g cm−3) 0.08 ns −0.4 ns 0.47 ns 0.52 *
Clay+fine silt (%) 0.02 ns 0.1 ns −0.60 * 0.15 ns
Moisture (%) −0.29 ns −0.1 ns −0.37 ns 0.04 ns

MBC, microbial biomass carbon; Cmin, carbon mineralization rate; SOC, soil organic carbon; qCO2, metabolic
quotient (or specific respiration)
Asterisks denote significant correlation at α = 5%

silt fraction size. Cmin positively correlated with leaf litter Ca, through litterfall. A greater immobilization of Ca was ob-
SOC, total soil N and soil pH but negatively with the soil C:N served in teak, also relative to forest, in addition to K and
ratio. Finally, the qCO2 positively correlated with soil pH but Mg. Teak rather fostered the recycling of P. However, the
negatively with the soil bulk density. observed difference in nutrient recycling intensity could have
been either mitigated or amplified, had fine roots of the trees
been considered (Kähkölä et al. 2011).
Discussion
Leaf litter decomposition and soil chemistry
Litterfall, leaf litter carbon, and macronutrient budget
Among soil decomposer organisms, microbes are those in-
The annual litterfall yield recorded in cocoa plantations falls volved in all stages of decomposition, making some of the
within the 5 to 21 t ha−1 year−1 range reported for West influential abiotic factor determinant across the successive de-
African cocoa systems (Hartemink 2005). The value recorded composition stages. In this study, the initial litter C:N and N:P
in the present study is similar to that recently found by Saj ratios proved determinant in litter decay and macronutrient
et al. (2021) in Cameroon (9.4 Mg ha−1 year−1). According to release. The slower rate of teak litter decomposition as op-
Carrera et al. (2008), the amount of leaf litterfall is related to posed to cocoa and forest could be ascribed to the lower initial
the canopy development; yet in the present study, cocoa trees litter N:P ratio, in addition to its recalcitrance. The influence of
that had least canopy magnitude yielded litterfall that was the N:P ratio on cocoa leaf litter decomposition was reported
close to those by forest and teak trees. It is likely that litterfall previously by Hartemink (2005). Cocoa N:P values reported
yield was plant species-specific and was thus driven by intrin- by this author for other countries varied from 9 (Cameroon) to
sic plant traits (Zhang et al. 2014). 21 (Malaysia), which are far lower than the 34.5 recorded in
The lower N and Ca inputs by cocoa compared to the the present study. The recalcitrance of teak leaf litter is trans-
primary forest may suggest that these elements were more lated by their specific weight of 16.5 mg cm−2, which is twice
immobilized, reflecting a greater biochemical use (Fontes as high as that (8.2 mg cm−2) of forest species encountered in
et al. 2014). In turn, cacao trees fostered the recycling of K our study area, such as Afzelia africana (Fabaceae) and Ceiba
and Mg, as these elements were supplied in greater amount pentandra (Bombacaceae) (Attignon et al. 2004). One striking
44 Page 10 of 14 Reg Environ Change (2021) 21: 44

result is that cocoa and forest litters showed similar break- to feed on litters may constitute stress conditions for the soil
down and macronutrient release patterns and rates, despite microorganisms (Wardle and Ghani 1995), the qCO2 would
cocoa leaf litter showing greater C:N and lower N:P ratios. normally have been greater in teak and cocoa than in forest,
This suggests that extrinsic factors, e.g., microhabitat, litter but this was not the case. It is likely that the recalcitrant leaf
consumption by termites and earthworms, possibly accounted litter environment together with the reduced nitrogen avail-
for this similarity (Asigbaase et al. 2021b). However, this ability under the monocultures fostered the development of
observation differed from that reported by Dawoe et al. k-strategist microorganisms, including fungi (Tate 2000),
(2010) and Saj et al. (2021) who found cocoa leaf litter to which are associated with lower Cmin rate compared to the r-
decompose more slowly than semideciduous forest litter in strategist microorganisms (Fierer et al. 2007).
Ghana and Cameroon, respectively.
Although the leaf litter decay and nutrient release dynamics Implication for sustainable management of cocoa
were determined by the initial litter chemistry, these two dy- plantations
namics may not occur simultaneously for some nutrients be-
cause of the influence of extrinsic factors. The N immobiliza- Cocoa is native to forest, and as such, it is susceptible to
tion in the teak leaf litter over the first 2 months could be due cultivation in forest-like conditions. However, conflicting im-
to the relatively high C:N ratio. Indeed, microbial develop- pacts of shade trees on cocoa growth and yields, and soil
ment, when C is available, requires a corresponding nitrogen carbon stocks are reported. In Ghana, Ahenkorah et al.
supply (Bonanomi et al. 2016). In general, the other macronu- (1987) found cocoa bean yields to be lower under shade con-
trients were continuously released from forest and cocoa litters dition compared to FSC while Blaser et al. (2017) found no
over the decomposition period. However, this was not the case impact. The latter trend was also reported by Rajab et al.
in teak litters where the stocks of K, Mg, and, to a lesser (2016) in Indonesia. Distinctively, more recently, Asare
extent, Ca increased for a while after an initial decrease, and et al. (2018) in Ghana found that cocoa bean yields increase
then declined. This rise could be a result of an active coloni- significantly with increased crown cover of shade trees, dou-
zation of the litters during that period associated with the bling when going from zero to approximately 30% crown
translocation of macronutrients from the soil to the litter sys- cover. Shaded-cocoa fields are reported to remain productive
tem by filamentous decomposer fungi as commonly occurring for up to 60 years in Cameroon (Jagoret et al. 2011) and
at the soil–litter interface (Frey et al. 2003). This additionally Nigeria (Adeniyi et al. 2017) owing to an environment that
presumably denotes a greater contribution of fungi to the teak improves the cocoa physiology (Vanhove et al. 2016; Rajab
leaf litter decomposition compared to the cocoa and forest leaf et al. 2016). Blaser et al. (2017) found no significant improve-
litters, which seemed to be proportional to the C:N ratio ment of SOC stocks under shade conditions in Ghana.
(Malik et al. 2016). Although macronutient contents in litters of mature and old
The initial litter characteristics remained influential on the cocoa fields are enough to offset the nitrogen exported
soil microbial parameters, particularly cations, in addition to through cocoa beans (Snoeck et al. 2016; Pérez-Flores et al.
some soil physical and chemical attributes. In the soils of this 2018), well-managed shaded cocoa appears far more sustain-
study where carbonates, Al3+ and Fe2+ are low (Kassin and able than (fertilized) full-sun cocoa since, at minimum, it is
Yoro 2009), one could presume that the increase in soil acidity associated with increased carbon storage in the total biomass
in the cocoa stands is due to a greater exsudation of acidic and longevity. This invariably slows down destruction of
compounds by cocoa roots as reported for other plant species existing forests and reduces CO2 concentration in the atmo-
(Dickson et al. 2017). Such an increase in soil acidity under sphere without causing significant cocoa yield loss (Rajab
cocoa following forest conversion was recently reported by et al. 2016; Asare et al. 2018). Importantly, in the current
Adeniyi et al. (2017) in Nigeria. This acidity, in conjunction context of climate change, the use of adaptation strategies like
with the cocoa leaf litter recalcitrance mostly explain the drop cAFS will be necessary to reduce the vulnerability of cocoa to
in the Cmin rate. Indeed, acidic soil and poor-quality litter excessive temperatures (Schroth et al. 2016). Based on these
conditions are known to promote fungal growth and microbial observations, an interest in promoting cAFS among cocoa
C build-up (increased MBC and MBC/SOC ratio under co- farmers arose in West Africa (Mbow et al. 2014; Jagoret
coa) while inhibiting the microbial activities (Rousk et al. et al. 2017). In Côte d’Ivoire, this is occurring under the lead-
2009; Yao et al. 2010; Malik et al. 2016). In teak plantation, ership of the World Agroforestry Centre (ICRAF), the Ivorian
however, the recalcitrant low-quality leaf litter most likely rural development agency (ANADER), cocoa certification
explains the drop in soil Cmin given that soils in that LUS were bodies including the Rainforest Alliance and UTZ (Gyau
slightly alkaline. The ability of teak to maintain or increase et al. 2014) and the German development agency – GIZ
soil pH was also reported in India by Swarnalatha and Reddy (Milz and Kouamé 2013; Koné and N’Guessan 2015). In this
(2011) who recorded a value of 6.26 against 5.46 in a control framework, a set of shade tree species were listed for the study
degraded forest. Considering that soil acidity and the difficulty region, based on the farmers’ interests (medicine, nutrition,
Reg Environ Change (2021) 21: 44 Page 11 of 14 44

firewood, timber) and their willingness to adopt the cAFS, to be species that are usually maintained on cocoa fields by
according to Gyau et al. (2014). farmers or grown in the region. These species were also re-
However, for proper cocoa agroforestry management, ported to yield significant quantity of low-relacitrance leaf
shade trees should also be selected taking into account their litter (Norgrove and Hauser 2000). As N-rich leaf litter pro-
impact on soil quality, as well as on the environment. To this vider trees, legumes such as Acacia auriculiformis and
end, tree species that have potential of providing quality litter A. mangium (Fabaceae)—up to 30 m tall—were suggested
and increasing carbon and nutrient inputs and cycling rate by Gnahoua et al. (2013). Another exotic tree that is well
should be targeted (Kähkölä et al. 2011; Gnahoua et al. adapted to the region and sheds abundant low-recalcitrance
2013). As another key criteria, shade trees should be at least leaf litter is Gmelina arborea (Verbenaceae), which is up to
8 m tall, so that mature cocoa trees have an overhead shade 30 m tall. This species is increasingly reported to foster soil
which is crucial for their physiology and longevity, and ade- biological activity and SOM build up better than legumes such
quate pod production beyond 20 years (Milz and Kouamé as Gliricidia sepium (Fabaceae) and Acacia auriculiformis
2013). The present study showed that leaf litter C and N inputs (Faboya et al. 2015; Rojas et al. 2017). However, the total
and mineralization in the topsoil decline in FSC stands relative density of shade trees should be kept at an optimal level.
to the primary forest. In turn, Tectona grandis which is a While the Sustainable Agriculture Network (SAN) recom-
native species, up to 40 m tall, supplied similar amount of mends 70 emergent shade tree species per ha at the global
litterfall and C and N as the natural forest while holding high scale, which must include a minimum of 12 native species,
economic value. This makes the species a relevant candidate estimated to provide 40% shade cover; 12–18 emergent shade
for carbon sequestration and shade in cocoa plantations. It is trees per ha amounting to 30–40% shade cover are recom-
noted that T. grandis may be capable of meeting to a certain mended for cocoa in the West African agroecological zone
extent the deficit of C (and probably N) supply exhibited in the (Asare and Anders 2016).
cocoa plantations. Moreover, teak supplied twice as much P as
cocoa, which support the inclusion of this tree in cocoa fields.
Indeed, phosphorus is key to cocoa nutrition, and as such, is Conclusions
invariably part of the fertilizers prescribed for cocoa (Snoeck
et al. 2016; Asigbaase et al. 2021b). As the soil tends to acidify This study revealed the contrasting impacts of cocoa and teak
under cocoa, the soil pH increase under teak is another reason plantations on C and macronutrient inputs and cycling inten-
to grow cocoa in combination with teak. sity. Specifically, C and N inputs through leaf litterfalls de-
Furtheremore, cocoa and teak leaf litter exhibited contrast- clined in cocoa stands relative to the primary forest whereas
ing decomposition and macronutrient release rates, the pat- teak stands showed similar C and N inputs as the primary
terns observed with cocoa being closer to that with forest leaf forest. Teak may thus be recommended for shade and a spe-
litter. This predicts easier decomposition of the teak leaf litter, cies that may, through its litterfall, compensate for part of the
if mixed with cocoa (Cuchietti et al. 2014; Koutika et al. 2014; deficit in carbon and nitrogen supply exhibited by cocoa. In
Faboya et al. 2015). Higher C mineralization rate is not nec- the condition of this study, cocoa leaf litter decomposed at the
essarily advantageous as it may induce C and N loss through same rate as forest leaf litter; teak’s leaf litter decomposed at a
volatilization (Cotrufo et al. 2015). However, it should be far lower rate probably with a greater contribution of fungi.
maintained at a certain level that allows sufficient availability However, the soil C mineralization rate and mineral N drasti-
of nutrients to crops and consistent build-up of SOM. In the cally declined in both the cocoa and teak stands compared to
present study, SOM mineralization rate in both tree planta- the levels in the primary forest, which should be overcome.
tions dropped to half of that observed in the natural forest. These results suggest that for sustainability, cocoa should be
Therefore, to cope with this limitation and ensure nutrient grown in association with tree species that shed abundant
availability to cocoa, the inclusion of teak in cocoa fields quality litter (N-rich and/or low-recalcitrance) materials while
should be accompanied with tree species that have the poten- preventing soil acidification and the spread of cocoa pests and
tial of providing N-richer or lower recalcitrance leaf litter diseases. However, further studies focusing on the actual con-
(Faboya et al. 2015; Sauvadet et al. 2020). Mixing leaf litter tribution of these types of trees to carbon storage in the subsoil
materials of contrasting quality will likely favor the are needed to corroborate this recommendation.
cooccurrence of an increased number of ecological niches
associated with a diversity of decomposer macro- and micro-
fauna, which will stimulate nutrient cycling (Handa et al. Supplementary Information The online version contains supplementary
2014; Delgado-Baquerizo et al. 2016) thereby improving co- material available at https://doi.org/10.1007/s10113-021-01776-0.
coa mineral nutrition and probably pod production. In this
Acknowledgements This research work was funded by the UN
aspect, Gyau et al. (2014) reported the native Terminalia Environment Programme-Global Environment Facility (UNEP-GEF)
ivoriensis and T. superba (Combretaceae)—30–40 m tall— through the project (No GF/2715-2) "Conservation and Sustainable
44 Page 12 of 14 Reg Environ Change (2021) 21: 44

Management of Belowground Biodiversity - CSM-BGBD". Further, our Bisseleua DHB, Missoup AD, Vidal S (2009) Biodiversity conservation,
appreciation goes to UNESCO’s Keizo Obuchi Fellowship Progamme ecosystem functioning, and economic incentives under cocoa agro-
2004 granted to Yao K. Michel, the LAMTO ecological research station, forestry intensification. Conserv Biol 23:1176–1184. https://doi.org/
which assisted in the sample pretreatment and conservation, the 10.1111/j.1523-1739.2009.01220.x
Ecological Laboratory of University of Pierre and Marie Currie (ENS, Blaser WJ, Oppong J, Yeboah E, Six J (2017) Shade trees have limited
UPMC, Paris VI, France), which hosted the granted fellowship for benefits for soil fertility in cocoa agroforests. Agric Ecosyst Environ
performing soil C, N, and microbial analyses, and the Laboratoire Sol- 243:83–91. https://doi.org/10.1016/j.agee.2017.04.007
Eau-Plante/ESA/INP-HB Yamoussoukro (Côte d’Ivoire), which per- Bonanomi G, Cesarano G, Gaglione SA, Ippolito F, Sarker T et al (2016)
formed leaf litter and the other soil chemical analyses. Many thanks to Soil fertility promotes decomposition rate of nutrient poor, but not
the farmers and the support staff of Goulikao village who hosted the nutrient rich litter through nitrogen transfer. Plant Soil 412:397–411.
CSM-BGBD project. https://doi.org/10.1007/s11104-016-3072-1
Buyer JS, Baligar VC, He Z, Arévalo-Gardini E (2017) Soil microbial
communities under cacao agroforestry and cover crop systems in
Peru. Appl Soil Ecol 120:273–280. https://doi.org/10.1016/j.
References apsoil.2017.09.009
Carrera AL, Bertiller MB, Larreguy C (2008) Leaf litter fall, fine-root
Adeniyi SA, de Clercq WP, van Niekerk A (2017) Development of a production, and decomposition in shrub lands with different canopy
composite soil degradation assessment index for cocoa structure induced by grazing in the Patagonian Monte, Argentina.
agroecosystems in southwestern Nigeria. Solid Earth 8:827–843. Plant Soil 311:39–50. https://doi.org/10.1007/s11104-008-9655-8
https://doi.org/10.5194/se-8-827-2017 Cotrufo MF, Soong JL, Horton AJ, Campbell EE, Haddix ML et al
Ahenkorah Y, Halma BJ, Appiah MR, Akrofia GS, Yirenkyi JEK (1987) (2015) Formation of soil organic matter via biochemical and phys-
Twenty year results from a shade and fertilizer trial on amazon cocoa ical pathways of litter mass loss. Nat Geosci 8:776–779. http://www.
(Theobroma cacao) in Ghana. Exp Agric 23:31–39. https://doi.org/ nature.com/naturegeoscience
10.1017/S0014479700001101 Cuchietti A, Marcotti E, Gurvich DE, Cingolani AM, Harguindeguy NP
Alfaro-Flores A, Isabel Morales-Belpaire I, Schneider M (2015) (2014) Leaf litter mixtures and neighbour effects: low-nitrogen and
Microbial biomass and cellulase activity in soils under five different high-lignin species increase decomposition rate of high-nitrogen and
cocoa production systems in Alto Beni, Bolivia. Agrofor Syst 89: low-lignin neighbours. Appl Soil Ecol 82:44–51. https://doi.org/10.
789–798. https://doi.org/10.1007/s10457-015-9812-z 1016/j.apsoil.2014.05.004
Anderson JM, Ingram JSI (1993) Tropical soil biology and fertility, a Dawoe EK, Isaac ME, Quashie-Sam J (2010) Litterfall and litter nutrient
handbook of methods, 2nd edn. CAB International, New York. dynamics under cocoa ecosystems in lowland humid Ghana. Plant
https://doi.org/10.2307/2261129 Soil 330:55–64. https://doi.org/10.1007/s11104-009-0173-0
Andres C, Comoé H, Beerli A, Schneider M, Rist S et al (2016) Cocoa in de Gannes V, Bekele I, Dipchansingh D, Wuddivira MN, De Cairies S et
monoculture and dynamic agroforestry. In: Lichtfouse E (ed) al (2016) Microbial community structure and function of soil fol-
Sustainable agriculture reviews. Springer International Publishing, lowing ecosystem conversion from native forests to teak plantation
Switzerland, pp 121–153. https://doi.org/10.1007/978-3-319- forests. Front Microbiol 7:1976. https://doi.org/10.3389/fmicb.
26777-7_3 2016.01976
Asare R, Anders R (2016) Tree diversity and canopy cover in cocoa Delgado-Baquerizo M, Maestre FT, Reich PB, Jeffries TC, Gaitan JJ et al
systems in Ghana. New For 47:287–302. https://doi.org/10.1007/ (2016) Microbial diversity drives multifunctionality in terrestrial
s11056-015-9515-3 ecosystems. Nat Commun 7:10541. https://doi.org/10.1038/
Asare R, Markussen B, Asare RA, Anim-Kwapong G, Ræbild A (2018) ncomms10541
On-farm cocoa yields increase with canopy cover of shade trees in Dickson RW, Fisher PR, Argo WR (2017) Quantifying the acidic and
two agro-ecological zones in Ghana. Clim Dev 11:435–445. https:// basic effects of fifteen floriculture speciesn grown in peat-based
doi.org/10.1080/17565529.2018.1442805 substrate. HortScience 52:1065–1072. https://doi.org/10.21273/
Asigbaase M, Dawoe E, Lomax BH, Sjogersten S (2021a) Biomass and HORTSCI11926-17
carbon stocks of organic and conventional cocoa agroforests, Faboya IO, Adebola SI, Awotoye OO (2015) Assessment of decomposi-
Ghana. Agric Ecosyst Environ 306:107192. https://doi.org/10. tion rate and soil nutrient status under different woody species com-
1016/j.agee.2020.107192 bination in a tree plantation. Agric For Fish 4:46–54. https://doi.org/
Asigbaase M, Dawoe E, Sjogersten S, Lomax BH (2021b) 10.11648/j.aff.20150402.14
Decomposition and nutrient mineralisation of leaf litter in small- FAOSTAT (2015) The agricultural production, 2015. Available at: http://
holder cocoa agroforests: a comparison of organic and conventional faostat3.fao.org/download/Q/*/E, April 20, 2015
farms in Ghana. J Soils Sediments 21:1010–1023. https://doi.org/10. Fierer N, Bradford MA, Jackson RB (2007) Toward an ecological clas-
1007/s11368-020-02844-4 sification of soil bacteria. Ecology 88:1354–1364. https://doi.org/
Attignon SE, Weibel D, Lachat T, Sinsin B, Nagel P et al (2004) Leaf 10.1890/05-1839
litter breakdown in natural and plantation forests of the Lama forest Fontes AG, Gama-Rodrigues AC, Gama-Rodrigues EF, Sales MV, Costa
reserve in Benin. Appl Soil Ecol 27:109–124. https://doi.org/10. MG et al (2014) Nutrient stocks in litter fall and litter in cocoa
1016/j.apsoil.2004.05.003 agroforests in Brazil. Plant Soil 383:313–335. https://doi.org/10.
Barrios E, Valencia V, Jonsson M, Brauman A, Hairiah K et al (2018) 1007/s11104-014-2175-9
Contribution of trees to the conservation of biodiversity and ecosys- Franzluebbers AJ, Haney RL, Hons FM, Zuberer DA (1996)
tem services in agricultural landscapes. Int J Biodivers Sci Ecosyst Determination of microbial biomass and mineralizable nitrogen fol-
Serv Manag 14:1–16. https://doi.org/10.1080/21513732.2017. lowing rewetting of dried soil. Soil Sci Soc Am J 60:1133–1139.
1399167 https://doi.org/10.2136/sssaj1996.03615995006000040025x
Bayala J, Sanou J, Bazié HR, Coe R, Kalinganire A et al (2020) Freud E, Richard J (2000) Les champs du cacao- un défi de compétitivité
Regenerated trees in farmers’ fields increase soil carbon across the Afrique-Asie, Karthala, CIRAD
Sahel. Agrofor Syst 94:401–415. https://doi.org/10.1007/s10457- Frey SD, Six J, Elliott ET (2003) Reciprocal transfer of carbon and nitro-
019-00403-6 gen by decomposer fungi at the soil–litter interface. Soil Biol
Reg Environ Change (2021) 21: 44 Page 13 of 14 44

Biochem 35:1001–1004. https://doi.org/10.1016/S0038-0717(03) Loranger G, Ponge JF, Imbert D, Lavelle P (2002) Leaf decomposition in
00155-X two semi evergreen tropical forests: influence of litter quality. Biol
Gnahoua GM, Oliver R, N’Guessan KA, Ballé P (2013) Production et Fertil Soils 35:247–252. https://doi.org/10.1007/s00374-002-0467-
retombées minérales des litières chez quatre espèces de 3
légumineuses arborées, utilisées en amélioration de jachères, en Malik AA, Chowdhury S, Schlager V, Oliver A, Puissant J et al (2016)
zone forestière de Côte d’Ivoire. J Appl Biosci 72:5800–5809. Soil fungal:bacterial ratios are linked to altered carbon cycling. Front
https://doi.org/10.4314/jab.v72i1.99665 Microbiol 7:1247. https://doi.org/10.3389/fmicb.2016.01247
Gyau A, Smoot K, Kouame C, Diby L, Kahia J et al (2014) Farmer Mbow C, Smith P, Skole D, Duguma L, Bustamante M (2014) Achieving
attitudes and intentions towards trees in cocoa (Theobroma cacao mitigation and adaptation to climate change through sustainable
L.) farms in Côte d’Ivoire. Agrofor Syst 88:1035–1045. https://doi. agroforestry practices in Africa. Curr Opin Environ Sustain 6:8–
org/10.1007/s10457-014-9750-1 14. https://doi.org/10.1016/j.cosust.2013.09.002
Han M, Zhu B (2020) Changes in soil greenhouse gas fluxes by land use Milz J, Kouamé AG (2013) Études sur les espèces d’arbres d’ombrage et
change from primary forest. Glob Chang Biol 26:2656–2667. dispositifs agro-forestiers adaptés à la cacaoculture dans la zone
https://doi.org/10.1111/gcb.14993 périphérique du Parc National de Taï en Côte d’Ivoire. PROFIAB,
Handa I, Aerts R, Berendse F, Berg MP, Bruder A et al (2014) German development agency – GIZ GmbH
Consequences of biodiversity loss for litter decomposition across Murphy J, Riley JP (1962) A modified single solution method for the
biomes. Nature 509:218–221. https://doi.org/10.1038/nature13247 determination of phosphate in natural waters. Anal Chim Acta 27:
Hartemink AE (2005) Nutrient stocks, nutrient cycling, and soil changes 31–36. https://doi.org/10.1016/S0003-2670(00)88444-5
in cocoa ecosystems: a review. Adv Agron 86:227–253. https://doi. N’Drin DC (2018) Analyse de la corrélation entre les paramètres
org/10.1016/S0065-2113(05)86005-5 dendrométriques et la diversité spécifique de la forêt classée de la
ICCO (2018) Annual report 2017, International Cocoa Organization, Téné. Master dissertation. ESA/INP-HB, Yamoussoukro, Côte
Abidjan, Côte d’Ivoire d'Ivoire
Jagoret P, Michel-Dounias I, Malézieux E (2011) Long-term dynamics of N’Gbala FN, Guéi AM, Tondoh JE (2017) Carbon stocks in selected tree
cocoa agroforests: a case study in central Cameroon. Agrofor Syst plantations, as compared with semi-deciduous forests in centre-west
81:267–278. https://doi.org/10.1007/s10457-010-9368-x Côte d’Ivoire. Agric Ecosyst Environ 239:30–37. https://doi.org/10.
Jagoret P, Snoeck D, Bouambi E, Ngnogué HT, Nyassé S et al (2017) 1016/j.agee.2017.01.015
Rehabilitation practices that shape cocoa agroforestry systems in Nacro HB, Benest D, Abbadie L (1996) Distribution of microbial activ-
Central Cameroon: key management strategies for long-term exploi- ities and organic matter according to particle size in a humid savanna
tation. Agrofor Syst 92:1185–1199. https://doi.org/10.1007/s10457- soil (Lamto, Côte d’Ivoire). Soil Biol Biochem 28:1687–1697.
016-0055-4 https://doi.org/10.1016/S0038-0717(96)00246-5
Kähkölä A-K, Pekka Nygren P, Leblanc HA, Pennanen T, Pietikäinen J Naik SK, Maurya S, Mukherjee D, Singh AK, Bhatt BP (2018) Rates of
(2011) Leaf and root litter of a legume tree as nitrogen sources for decomposition and nutrient mineralization of leaf litter from differ-
cacaos with different root colonisation by arbuscular mycorrhizae. ent orchards under hot and dry sub-humid climate. Arch Agron Soil
Nutr Cycl Agroecosyst 92:51–65. https://doi.org/10.1007/s10705- Sci 64:560–573. https://doi.org/10.1080/03650340.2017.1362104
011-9471-z Norgrove L, Hauser S (2000) Leaf properties, litter fall, and nutrient
Kallenbach CM, Frey SD, Grandy AS (2016) Direct evidence for inputs of Terminalia ivorensis at different tree stand densities in a
microbial-derived soil organic matter formation and its ecophysio- tropical timber-food crop multistrata system. Can J For Res 30:
logical controls. Nature Commun 7:13630. https://doi.org/10.1038/ 1400–1409. https://doi.org/10.1139/x00-063
ncomms13630 Ochoa-Hueso R, Plaza C, Moreno-Jiménez E, Delgado-Baquerizo M
Kassin KE, Yoro G (2009) Rapport d’activité du programme gestion (2021) Soil element coupling is driven by ecological context and
durable des sols et maîtrise de l’eau (GDSME). Station de atomic mass. Ecol Lett 24:319–326. https://doi.org/10.1111/ele.
Recherche du CNRA Gagnoa, 17 p 13648
Koko LK, Snoeck D, Lekadou TT, Assiri AA (2013) Cacao-fruit tree Ofori-Frimpong K, Asase A, Mason J, Danku L (2007) Shaded versus
intercropping effects on cocoa yield, plant vigour and light intercep- unshaded cocoa: implications on litter fall, decomposition, soil fer-
tion in Côte d’Ivoire. Agrofor Syst 87:1043–1052. https://doi.org/ tility and cocoa pod development. Symposium on multistrata agro-
10.1007/s10457-013-9619-8 forestry systems with perennial crops. CATIE Turrialba, Costa Rica,
Koné AW, N’Guessan AK (2015) Rapport de formation des formateurs pp 17–21
en agroforesterie et en compostage en cacaoculture. PROFIAB, Olson JS (1963) Energy storage and balance of producers and decom-
German development agency – GIZ GmbH posers in ecological systems. Ecology 44:322–331. https://doi.org/
Koné AW, Yao MK (2021) Soil microbial functioning and organic car- 10.2307/1932179
bon storage: can complex timber tree stands mimic natural forests? J Pérez-Flores J, Pérez AA, Suárez YP, Bolaina VC, Quiroga AL (2018)
Environ Manage 283:112002. https://doi.org/10.1016/j.jenvman. Leaf litter and its nutrient contribution in the cacao agroforestry
2021.112002 system. Agrofor Syst 92:365–374. https://doi.org/10.1007/s10457-
Koutika L-S, Epron D, Bouillet J-P, Mareschal L (2014) Changes in N 017-0096-3
and C concentrations, soil acidity and P availability in tropical mixed Rajab YA, Leuschner C, Barus H, Tjoa A, Hertel D (2016) Cacao culti-
acacia and eucalypt plantations on a nutrient-poor sandy soil. Plant vation under diverse shade tree cover allows high carbon storage and
Soil 379:205–216. https://doi.org/10.1007/s11104-014-2047-3 sequestration without yield losses. PLoS One 11(2):e0149949.
Limtong P, Ishizuka K, Takahashi M, Sunantapongsuk V, Tummakate P https://doi.org/10.1371/journal.pone.0149949
(1998) Dynamics of soil microbial activities in different vegetation Rojas MJ, Caicedo V, Jaimes Y (2017) Biomass decomposition dynamic
types of the seasonally dry tropics. In: Sassa K (ed) Environmental in agroforestry systems with Theobroma cacao L. in Rionegro,
forest science, Forestry Sciences, vol 54. Springer, Dordrecht, pp Santander (Colombia). Agron Colomb 35:182–189. https://doi.org/
253–260. https://doi.org/10.1007/978-94-011-5324-9_27 10.15446/agron.colomb.v35n2.60981
Liu CLC, Kuchma O, Krutovsky KV (2018) Mixed-species versus Rousk J, Brookes PC, Bååth E (2009) Contrasting soil pH effects on
monocultures in plantation forestry: development, benefits, ecosys- fungal and bacterial growth suggest functional redundancy in carbon
tem services and perspectives for the future. Glob Ecol Conserv 15: mineralization. Appl Environ Microbiol 75:1589–1596. https://doi.
e00419. https://doi.org/10.1016/j.gecco.2018.e00419 org/10.1128/AEM.02775-08
44 Page 14 of 14 Reg Environ Change (2021) 21: 44

Ruf OF (2011) The myth of complex cocoa agroforests: the case of Vance ED, Brookes PC, Jenkinson DS (1987) An extraction method for
Ghana. Hum Ecol Interdiscip J 39:373–388. https://doi.org/10. measuring soil microbial biomass C. Soil Biol Biochem 19:703–
1007/s10745-011-9392-0 707. https://doi.org/10.1016/0038-0717(87)90052-6
Saj S, Nijmeijer A, Nieboukaho JDE, Lauri P-R, Harmand J-M (2021) Vanhove W, Vanhoudt N, Van Damme P (2016) Effect of shade tree
Litterfall seasonal dynamics and leaf-litter turnover in cocoa planting and soil management on rehabilitation success of a 22-
agroforests established on past forest lands or savannah. Agrofor year-old degraded cocoa (Theobroma cacao L) plantation. Agric
Syst. https://doi.org/10.1007/s10457-021-00602-0 Ecosyst En viron 219:14–25. https://doi.org/10.1016/
Sauvadet M, Saj S, Freschet GT, Essobo J-D, Enock S et al (2020) Cocoa jagee201512005
agroforest multifunctionality and soil fertility explained by shade Veldkamp E, Marcus Schmidt M, Powers JS, Corre MD (2020)
tree litter traits. J Appl Ecol 57:476–487. https://doi.org/10.1111/ Deforestation and reforestation impacts on soils in the tropics. Nat
1365-2664.13560 Rev Earth Environ 1:590–605. https://doi.org/10.1038/s43017-020-
Schroth G, Läderach P, Martinez-Valle AI, Bunn C, Jassogne L (2016) 0091-5
Vulnerability to climate change of cocoa in West Africa: patterns, Wardle DA, Ghani A (1995) Critique of the microbial metabolic quotient
opportunities and limits to adaptation. Sci Total Environ 556:231– (qCO2) as a bioindicator of disturbance and ecosystem development.
241. https://doi.org/10.1016/j.scitotenv.2016.03.024 Soil Biol Biochem 27:1601–1610. https://doi.org/10.1016/0038-
Snoeck D, Koko L, Joffre J, Bastide P, Jagoret P (2016) Cacao nutrition 0717(95)00093-t
and fertilization: Relevant agronomic basics and fertilizer issues. In: Wessel M, Quist-Wessel PMF (2015) Cocoa production in West Africa, a
Lichtfouse, E (ed) Sustainable Agriculture Reviews, vol. 19. review and analysis of recent developments. NJAS Wageningen J
Springer International Publishing, pp 155–202. https://doi.org/10. Life Sci 74-75:1–7. https://doi.org/10.1016/jnjas201509001
1007/978-3-319-26777-7_44
Yao MK, Angui PKT, Konaté S, Tondoh JE, Tano Y et al (2010) Effects
Soil Survey Staff (1999) Soil taxonomy A basic system of soil classifi-
of land use types on soil organic carbon and nitrogen dynamics in
cation for making and interpreting soil surveys 2nd edition
mid-west côte d’Ivoire. Eur J Sci Res 40:11–22. http://www.
Agicultural handbook n°436, 886 p
eurojournals.com/ejsr.htm
Swarnalatha B, Reddy MV (2011) Leaf litter breakdown and nutrient
release in three tree plantations compared with a natural degraded Zhang H, Yuan W, Dong W, Liu S (2014) Seasonal patterns of litterfall in
forest on the Coromandel Coast (Puducherry, India). Ecotropica 17: forest ecosystem worldwide. Ecol Complex 20:240–247. https://doi.
39–51. https://www.semanticscholar.org/paper/ org/10.1016/jecocom201401003
Tate R (2000) Soil microbiology, Second edn. Wiley, New York
Van Straaten O, Corre MD, Wolf K, Tchienkoua M, Cuellar E et al Publisher’s note Springer Nature remains neutral with regard to jurisdic-
(2015) Conversion of lowland tropical forests to tree cash crop plan- tional claims in published maps and institutional affiliations.
tations loses up to one-half of stored soil organic carbon. PNAS 112:
956–960. https://doi.org/10.1073/pnas.1504628112