2017 International Symposium on Cocoa Research (ISCR), Lima, Peru, 13-17 November 2017

Title: How habitat heterogeneity affects pollinator’s communities in cocoa-based agroforestry systems?

Authors: O. Deheuvels1, L. De Waal2 and L. Bagny Beilhe3

Institutions:
1
CIRAD, Lima, Peru.
2
CIRAD, Montpellier, France.
3
CIRAD, Turrialba, Costa Rica.

Abstract: In the humid tropics, a significant amount of the agricultural landscape where cocoa (Theobroma
cacao) is grown is managed as agroforestry systems. As pressure to intensify cocoa production is
increasing, the current worldwide trend for the intensification of cocoa production aims at significant
reductions of the shade canopy. However, this trend implies losing the potential to produce valuable
ecosystem services. Among those services, pollination is a limiting factor of cocoa production that has been
investigated in the 70’s and 80’s essentially by trapping methods. The genus Forcypomia sp. is
acknowledged to be the main insect responsible for cocoa pollination, when other insects such as ants and
Trips sp. would play a secondary role. Regulation and habitat provision for cocoa pollinating insect
communities are poorly documented in the literature. Most pollinating species are known from trapping
methods in the immediate surroundings of the tiny cocoa flowers and no study has succeeded so far in
observing and describing the insects actually visiting the inside of the cocoa flowers. In this investigation,
we characterized habitats for pollinating insects at plot scale, in a 1.5 ha cocoa-based agroforestry system
located in the Peruvian Amazon. Local climate, topography, leaf litter’s biomass, composition and water
content, cocoa trees and associated plant diversity and structure, were described, as well as farmer’s
practices. Based on habitats characterization, we compared the pollinating insects’ communities of two
contrasted habitats within the same plot. We used a digital video recording system that allowed us to
monitor and record all insects visiting cocoa from 6:30 am to 05:30 pm during the main flowering season.
Each habitat was sampled in 2 or 3 different locations when possible and 20 to 30 flowers were monitored
in each repetition, resulting in a total amount of 180 monitored flowers. Our results showed that the diversity
and the frequency of insects visiting cocoa flowers are influenced by habitat quality. The heterogeneity of
habitat often found in cocoa-based agroforestry system is mostly due to farmer’s practices relying on
opportunistic shade management. Pollinating insects’ communities do not always rely specifically on the
Forcypomia genus but rather depend on habitat quality. These results open good perspectives for the
ecological intensification of cocoa production in Agroforestry Systems.

Introduction
During the past two centuries, the impact of human activities on landscapes and biodiversity at global scale
has considerably increased (DeFries et al., 2004). More food production is needed to feed the growing
human population, and terrestrial natural habitats are being massively converted into agro-ecosystems
(August et al., 2002; Forman, 1995; Sala et al., 2000; Laurance, 1999). In tropical areas, perennial crops
such as oil palm, tea, rubber, coffee and cocoa represent a significant amount of the cultivated land
(Neufeldt et al., 2012), especially at the forest margin of the Amazon (Laurance et al., 2001). Some of these
crops may have strong negative impacts on ecosystem services, affecting local biodiversity, soils and
waters. Others, such as cocoa-based agroforestry systems in the humid tropics, have more limited impacts
and provide interesting synergies between socio-economic and environmental challenges. (Schroth et al.,
2004 ; Vandermeer et al., 1998). These cropping systems, where the main crop is associated with a number
of other cultivated plants on the same plot, often offer higher plant biodiversity levels and improved
ecosystem services than in mono-cropping systems (Deheuvels et al., 2014, 2012; Malézieux et al., 2009;
Chen et al., 1999). In particular, the presence of different strata, among the cultivated plant species, creates
micro-climate conditions that can be favorable for wild plant and animal species (Martin-Chave et al., 2016;
Burgess, 1999). Because they provide a wide variety of habitats and food, agroforestry systems are known
to be of importance for the populations of pollinating animals (Varah et al., 2013; Jha and Vandermeer,
2010; Klein et al., 2007). Most of the area (95%) where cocoa is grown is cultivated by small farmers (Rice
and Greenberg, 2000) on farms smaller than 10 hectares (Nolte, 2014; MINAGRI, 2003) with low
investment capacity, family workforce and low risk strategies often including agroforestry. There, the
design of the cocoa plantation is rarely conventional and both the cocoa trees and the associated plants often
present a heterogeneous distribution on the cocoa plot (Gidoin et al., 2014; Matey et al., 2013; Ngo Bieng
et al., 2013; Deheuvels et al., 2012). Since before the early 70’s, the cocoa tree was known to be a species
pollinated by insects (Glendinning, 1972), with only 5% of its flowers receiving enough pollen to get
fecundated. A number of insects families are hold responsible for the pollination of the cocoa tree, such as
Cecidomyiidae (Garibaldi et al., 2011), Ceratopogonidae (Young, 1982; Winder, 1978), but also ants,
mealybugs, Thrips and Cicadella (Orwa et al., 2009). Following Frimpong et al. (2011) and Young (1982),
we make the hypothesis that the heterogeneity of plant distribution in cocoa-based agroforestry systems
reflects farmer’s practices
2017 International Symposium on Cocoa Research (ISCR), Lima, Peru, 13-17 November 2017
and has an influence on the composition and abundance of cocoa-pollinating insects at the micro-habitat
level. In the Peruvian Amazon, we studied micro-habitats at plot level in one single cocoa-based
agroforestry system, including topography, plant composition and structure, and local climate. We then
monitored and compared pollinating insect populations in two major micro-habitats during the dry season.
In this communication we present very preliminary results of our study to be published in 2018.

Materials and Methods

Study site: the study was conducted between April and July 2017, during one of the two annual flowering
periods of the cocoa trees, on a 5 years old cocoa plantation. This plantation was cultivated as an
agroforestry system by a local farmer and located in the Peruvian Amazon, Ucayali region, Irazola District.
There, the cocoa trees were 95% CCN-51 cocoa clone and 5% Peruvian undescribed native varieties, in
association with fruit and timber trees, as well as other service trees. There, the humid warm tropical climate
is appropriate for cocoa cultivation, with an 80% average annual relative humidity, a 2500 mm average
annual rainfall concentrated from November to March, and a 26°C average annual temperature. The surface
of the cocoa plot was 1.5 ha and was measured with a GPS. The last treatments on this plot were applied in
2015: 1 herbicide application and 1 insecticide application.
Sampling units and habitat biological and topographic descriptors: a grid has been built that covered
the whole plot to serve as a reference for coordinates of each plant and sampling sites. Starting from the
origin point (South: 8°50.958'; West: 75°06.898'), the grid was based on 10 m x 10 m geo-referenced square
sampling units which elevation was registered at the center. On each 100 m² sampling unit, we identified
each living plant (cocoa and non-cocoa) and reported their coordinates on the grid. Total canopy cover was
assessed with a densitometer at 1m above the ground at the center of each sampling unit. Total height,
canopy shape and height, as well as dbh at 1.3 m above the ground were reported for each living plant. In
the case of cocoa trees and plants lower than 1.3 m, dbh was measured at 0.3 m above the ground. Each
100 m² sampling unit was divided into four 5 m x 5 m sub-sampling units at the center of which leaf litter
was sampled and weighted fresh and after drying 1 hour at 105°C.
Local climate monitoring: daily rainfall, temperatures (min/max), and relative humidity (min/max) were
registered at the plot level. Maximum wind speed was registered every 30 minutes from 6:30 am to 5:30
pm during the pollinating insect monitoring periods.
Monitoring of the pollinating insects: three sampling units were selected per micro-habitat. They were
chosen for presenting the highest distance from the borders of the habitat. On each sampling unit, 10 cocoa
flowers were monitored each one from 6:30 am to 5:30 pm, using a video camera system described by
Steen et al. (2011). Each flower was selected for monitoring 24h before opening and isolated. Each insect
visiting a given flower was registered and identified at Family level. Insects were then classified according
to their activity in relation with the reproductive parts of the flower, i.e. touching or not-touching male or
female reproductive parts.
Statistical analysis: a multivariate analysis followed by a cluster analysis, using the R- Vegan package
(Oksanen et al., 2017). allowed identifying micro-habitats. Kuskall-Wallis tests were performed for
quantitative variables and Chi² tests for qualitative ones, in order to identify the habitat discriminating
variables. Non parametrical tests (Kruskall-Wallis) were conducted to compare the number of potentially
effective pollination activity between the two habitats. The effect of the wind on the composition of
pollinating insects has been tested by Spearman non-parametrical correlation test.

Preliminary results
The Figure 1 presents the results of the cluster analysis and the repartition of micro-habitats 1 and 2 within
the cocoa plot. A transition area between the two habitats is clearly visible.

Figure 1. Cluster analysis and its translation on the geo-referenced grid and on each 100 m² sampling unit
of the cocoa-based agroforestry plantation.
2017 International Symposium on Cocoa Research (ISCR), Lima, Peru, 13-17 November 2017
Among the 55 variables tested, these two micro-habitat significantly differed according to 12 variables
described in Table 1.

Statistic
Habitat descriptor Variable df P
value
Slope inclination and orientation (*) 106,6 2 < 0,001
Topography inside the
Relative elevation coefficient (*) 84,7 8 < 0,001
plot
Micro topography (*) 56,9 1 < 0,001
Soil cover including % of opened pods on the ground 7,6 1 0,006
dead biomass and % of leaf litter on the ground 12,0 1 0,001
vegetation under the
% of medium sized herbaceous plants 16,0 1 < 0,001
cocoa trees
Associated plants % of service plants 8,2 1 0,004
% of shade trees with spherical shaped canopy 8,7 1 0,003
% of shade trees with inverted pyramid shaped
9,5 1 0,002
canopy
Shade plant structure
% of short plants [0 - 0,65 m] in total height 9,1 1 0,003
% of plants with short trunk [0 - 0,50 m] 13,2 1 < 0,001
% of plants with high dbh [0.67 – 4.71 m] 8,6 1 0,003
Table 1. Variable showing significant differences between Habitat 1 and Habitat 2, based on the X²-test
for qualitative variables (*) and the Kruskal-Wallis chi-squared test for other variables.

Figure 2 presents the insect orders observed in the two micro-habitats and show that in both cases
hymenoptera (mostly ants) represent more than 92% of the visiting insects.

Micro habitat 1 Micro habitat 2

Figure 2. Composition of insects visiting cocoa flowers in micro habitats 1 and 2 at the Order level.
NI = unidentified insects.
The only significant correlation between wind speed and the visiting frequency of pollinating insects has
been observed for the Hymenoptera order (Table 2).
Order Rho S Statistic P
Aranae < 0.001 1519000 0.9812
Blattoptera -0.11 1687600 0.1156
Diptera -0.09 1657600 0.1978
Hymenoptera 0.21 1198500 0.00203**
Heteroptera 0.002 1524600 0.9765
Lepidoptera -0.11 1687600 0.1156
Orthoptera -0.04 1575700 0.6087
Table 2. Spearman correlation test between wind and cocoa flower visiting frequency for each taxa
registered by video camera.
2017 International Symposium on Cocoa Research (ISCR), Lima, Peru, 13-17 November 2017
Discussion
Cocoa based agroforestry systems are often cultivated without any identifiable design for shade plants
distribution. In addition, the annual mortality of the cocoa trees and the associated plants results in a high
heterogeneity of plant population structure in these systems. Finally, all dead plants are not always replaced
by the same species if replaced at all. This process is well described in Jagoret (2011). It results in a complex
spatial heterogeneity of plant distribution at plot level, which offers a variety of micro habitats for small
animals, including pests and pollinators (Ngo Bieng et al., 2013; Gidoin et al., 2014).
This plant distribution heterogeneity combines with the topographic condition of the plot. Small intermittent
rivers often cross cocoa plantations and it is not rare to find steep slopes of different inclinations and
orientations, some parts of the plantation receiving more sunlight than others.
We evidenced this spatial heterogeneity due to both topographic and plant distribution on one single cocoa
plot in the Peruvian Amazon.
We also evidenced that the famous Forcypomia cocoa midge (Diptera, Ceratopogonidae) is not always the
main pollinating agent for cocoa. A huge variety of insects have been described in the literature as cocoa
pollinating agents in Africa, Latin America and Asia, including Ceratopogonidae, but also Cecidomyiidae,
Thysanoptera, Aphidae, Psyllidae, Formidae, Hemiptera and Apidae (De Reffye et al., 1980; Lucas, 1981;
Boussard, 1981; Paulin et al., 1983; Young, 1985; Mavisoy, 2009). Plot management, especially
topography interacting with plant distribution, is certainly playing an important role to regulate the
populations of pollinating insects at plot level.
The data collected in this study are still under analysis and our approach will also include the possible
effects of land uses surrounding the cocoa plantation.

References
August P., Iverson L., Nugranad J., 2002. Human Conversion of Terrestrial Habitats. In: Applying Landscape Ecology
in Biological Conservation. New York, Springer, pp. 198‑224.
Boussard B., 1981. Pollinisation. Arbres fruitiers et cacaoyers. Café, Cacao, Thé, 25 (4), pp. 297‑304.
Burgess P.J., 1999. Effects of agroforestry on farm biodiversity in the UK. Scottish Forestry, 53 (1), pp. 24‑27.
Chen J., Saunders S.C., Crow T.R., Naiman R.J., Brosofske K.D., Mroz G.D., Brookshire B.L., Franklin J.F., 1999.
Microclimate in Forest Ecosystem and Landscape Ecology: Variations in Local Climate can be used to Monitor and
Compare the Effects of Different Management Regimes. BioScience, 49 (4), pp. 288‑297.
Deheuvels O., Rousseau G.X., Soto Quiroga G., Decker Franco M., Cerda R., Vilchez Mendoza S.J., Somarriba E.
2014. Biodiversity is affected by changes in management intensity of cocoa-based agroforests. Agroforestry Systems,
88: 1081–1099. DOI 10.1007/s10457-014-9710-9.
Deheuvels O., Avelino J., Somarriba E., Malézieux E. 2012. Vegetation structure and productivity in cocoa-based
agroforestry systems in Talamanca, Costa Rica. Agriculture, ecosystems and environment, 149: 181-188.
DOI:10.1016/j.agee.2011.03.003.
De Reffye P., Parvais J.-P., Coulibaly N., Gervais A., 1980. Etude de la Pollinisation du Cacaoyer à Partir du Trafic
des Insectes. Modèle Mathématique et Simulation. Café, Cacao, Thé, 24 (2) : 83‑100.
DeFries R.S., Foley J.A., Asner G.P., 2004. Land-use choices: Balancing human needs and ecosystem function.
Frontiers in Ecology and the Environment, 2 (5), pp. 249–257.
Forman R.T., 1995. Some general principles of landscape and regional ecology. Landscape ecology, 10 (3): 133–142.
Frimpong E.A., Gemmill-Herren B., Gordon I., Kwapong P.K., 2011. Dynamics of insect pollinators as influenced by
cocoa production systems in Ghana. Journal of Pollination Ecology, 5 (10), pp. 74–80.
Garibaldi L.A., Muchhala N., Motzke I., Bravo-Monroy L., Olschewski R., Klein A.-M., 2011. Services from Plant-
Pollinator Interactions in the Neotropics. In: Ecosystem services from agriculture and agroforestry: measurement and
payment. London, Taylor & Francis, Earthscan.
Gidoin C., Avelino J., Deheuvels O., Cilas C., Ngo Bieng M. A. 2014. Shade tree spatial structure and pod production
explain frosty pod rot intensity in cacao agroforests, Costa Rica. Phytopathology, 104 (3): 275-281.
http://dx.doi.org/10.1094/PHYTO-07-13-0216-R.
Glendinning D.R., 1972. Natural Pollination of Cocoa. New Phytologist, 71 (4), pp. 719‑729.
Jagoret, P., Michel-Dounias, I. and Malézieux, E. 2011. Long term dynamics of cocoa agroforests : a case study in
central Cameroun. Agroforestry Systems, 81: 267–278
Jha S., Vandermeer J.H., 2010. Impacts of coffee agroforestry management on tropical bee communities. Biological
Conservation, 143 (6), pp. 1423‑1431.
Klein A.-M., Vaissiere B.E., Cane J.H., Steffan-Dewenter I., Cunningham S.A., Kremen C., Tscharntke T., 2007.
Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society B: Biological
Sciences, 274 (1608), pp. 303‑313.
2017 International Symposium on Cocoa Research (ISCR), Lima, Peru, 13-17 November 2017
Laurance W.F., 1999. Reflections on the tropical deforestation crisis. Biological conservation, 91 (2), pp. 109–117.
Laurance W.F., Cochrane M.A., Bergen S., Fearnside P.M., Delamônica P., Barber C., D’Angelo S., Fernandes T.,
2001. The Future of the Brazilian Amazon. Science, 291 (5503), pp. 438‑439.
Lucas P., 1981. Etude des Conditions de Pollinisation du Cacaoyer au Togo. Café, Cacao, Thé, 25 (2), p. 8.
Malézieux E., Crozat Y., Dupraz C., Laurans M., Makowski D., Ozier-Lafontaine H., Rapidel B., Tourdonnet S.,
Valantin-Morison M., 2009. Mixing plant species in cropping systems: concepts, tools and models. A review.
Agronomy for Sustainable Development, 29 (1): 43‑62.
Martin-Chave A., Mazzia C., Beral C., Capowiez Y., 2016. Effects of Agroforestry Microclimate Over Spiders and
Ground Beetles Daily-Activity.
Matey, A., Zeledón, L., Orozco, L., Chavarría, F., López, A. and Deheuvels, O. 2013. Floristic composition and
structure of cacao orchards and forest patches in Waslala, Nicaragua. Agroforestería en las Américas 49 : 61-67.
Mavisoy M., 2009. Evaluación de la abundancia de ceratopogonidos (Díptera) polinizadores de cacao (Theobroma
cacao L.) en la hojarasca de 7 árboles de sombra, Talamanca-Costa Rica. Ingeniero Agroforestal, Universidad de
Narino-Facultad de Ciencias Agricolas, Talamanca, Costa Rica, 20 p.
http://repositorio.bibliotecaorton.catie.ac.cr/handle/11554/8063.
MINAGRI, 2003. Caracterización de las Zonas Productoras de Cacao en el Perú y su Competitividad. Lima, p. 207.
http://www.infocafes.com/descargas/biblioteca/19.pdf.
Neufeldt H, Dawson I A K, Luedeling E, Ajayi O C, Beedy T L, Gebrekirstos A, Jamnadass R H, Kenig K, Sileshi G
W, Simelton E S, Sotelo Montes C, Weber J C, 2012. Climate change vulnerability of agroforestry. p. 31.
Ngo Bieng, M.-A., Gidoin, C., Avelino, J., Cilas, C., Deheuvels, O. and Wery, J. 2013. Diversity and spatial clustering
of shade trees affect cacao yield and pathogen pressure in Costa Rican agroforests. Basic and applied ecology, 14(4):
329-336. http://dx.doi.org/10.1016/j.baae.2013.03.003.
Nolte G.E., 2014. Peru: Cocoa Update and Outlook | USDA Foreign Agricultural Service. Peru, United States
Department of Agriculture, p. 5. https://www.fas.usda.gov/data/peru-cocoa-update-and-outlook.
Oksanen J., Blanchet G.F., Friendly M., Kindt R., Legendre P., McGlinn D., Minchin P.R., O’Hara R.B., Simpson
G.L., Solymos P., Stevens M.H.H., Szoecs E., Wagner H., 2017. Vegan : Community Ecology Package. R package
version 2.4-3, p. 12.
Orwa C., Mutua A., Kindt R., Jamnadass R., Anthony S., 2009. Theobroma cacao. World Agroforestry Center,
Agroforestree Database : a tree reference and selection guide version 4.0, p. 5.
http://www.worldagroforestry.org/treedb/AFTPDFS/Theobroma_cacao.PDF.
Paulin D., Decazy B., Coulibaly N., 1983. Etude des Variations Saisonnières des Conditions de Pollinisation et de
Fructification Dans Une Cacaoyère. Café, Cacao, Thé, 27 (3), p. 12.
Rice R.A., Greenberg R., 2000. Cacao cultivation and the conservation of biological diversity. AMBIO: A Journal of
the Human Environment, 29 (3): 167–173.
Sala O.E., Chapin F.S., Iii, Armesto J.J., Berlow E., Bloomfield J., Dirzo R., Huber-Sanwald E., Huenneke L.F.,
Jackson R.B., Kinzig A., Leemans R., Lodge D.M., Mooney H.A., Oesterheld M., Poff N.L., Sykes M.T., Walker B.H.,
Walker M., Wall D.H. 2000. Global Biodiversity Scenario for 2100. Science, 287 (5459): 1770‑1774.
Schroth, G., Fonseca da, G.A.B., Harvey, C.A., Gascon, C., Vasconcelos, H.L. and Izac, A.-M. N., 2004. Agroforestry
and Biodiversity Conservation in Tropical Landscapes. Ed. Island Press, Washington. 525 p.
Steen R., Aase A., Thorsdatter A., 2011. Portable digital video surveillance system for monitoring flower-visiting
bumblebees. Journal of Pollination Ecology, 5 (13): 90–94.
Vandermeer J., van Noordwijk M., Anderson J., Ong C., Perfecto I., 1998. Global change and multi-species
agroecosystems: concepts and issues. Agriculture, Ecosystems & Environment, 67 (1): 1–22.
Varah A., Jones H., Smith J., Potts S.G., 2013. Enhanced biodiversity and pollination in UK agroforestry systems: UK
agroforestry systems. Journal of the Science of Food and Agriculture, 93 (9): 2073‑2075.
Winder J.A., 1978. Cocoa Flower Diptera; Their Identity, Pollinating Activity and Breeding Sites. Tropical Pest
Management, 24 (1): 5‑18.
Young A.M., 1982. Effects of Shade Cover and Availability of Midge Breeding Sites on Pollinating Midge Populations
and Fruit Set in Two Cocoa Farms. The Journal of Applied Ecology, 19 (1): 47–63.
Young A.M., 1985. Pollen-collecting by stingless bees on cacao flowers. Cellular and Molecular Life Sciences, 41 (6):
760–762.