Floristics of cacao agroforestry in the Mata Atlântica of southern Bahia, Brazil

Floristics of cacao agroforestry in the Mata Atlântica of southern Bahia, Brazil

Katherine J. Young, MFS1


Sustainable land use practices are urgently needed to aid the conservation of biological diversity and natural resources as well as for social and economic development to sustain local communities living within natural or semi-natural ecosystems. Cabrucas, a traditional system of agroforestry management for cacao, entails clearing the native forest understory and planting cacao trees under the shade of the forest canopy. In this study, I documented the species richness and diversity of four cabruca stands in southern Bahia, Brazil. Of the 4,435 trees measured in the cabrucas, I found a total of 103 different tree species $$5 cm DBH, in 33 families and 61 genera. Almost a fifth (17%) were neither Theobroma nor Musa species. Most of these individuals were in the families Leguminosae, Rutaceae, Moraceae, and Meliaceae. Future analyses will assess how similar these cabruca stands are to adjacent secondary forest stands of the same age class to elucidate opportunities for applying forest restoration principles and silvicultural techniques to the management of landscapes dominated by agricultural production.

Práticas sustentáveis de uso da terra são necessárias e urgentes para apoiar a conservação da diversidade biológica e recursos naturais, bem como para o desenvolvimento social e econômico de comunidades locais que vivem em ecossistemas naturais ou semi-naturais. As cabrucas são sistemas tradicionais de manejo agroflorestal para o cacaueiro que envolvem a limpeza do sub-bosque da floresta nativa e o plantio de cacaueiros sob a sombra do dossel da floresta. Neste estudo, foram mensuradas a riqueza e diversidade de espécies de quatro cabrucas no sul da Bahia, Brasil. Das 4.435 árvores medidas nas cabrucas, o estudo revelou um total de 103 diferentes espécies de árvores $$5 cm DAP, em 33 famílias e 61 gêneros. Quase um quinto (17%) de todas as árvores não eram nem Theobroma nem Musa spp. A maioria desses indivíduos pertencia às famílias Leguminosae, Rutaceae, Moraceae e Meliaceae. Análises futuras avaliarão a similaridade entre áreas de cabruca a fragmentos de florestais secundários adjacentes e de mesma classe etária, a fim de elucidar oportunidades de aplicação de princípios de restauração florestal e técnicas de manejo silviculturais no manejo de paisagens dominados pela produção agrícola.


The Atlantic Forest in Brazil is one of the most biodiverse forests in the world with over 8,000 vascular plant species, approximately 40% of which are endemic to the region (da Silva and Casteleti 2003, Galindo-Leal and Câmara 2003). However, very little of this rich forest remains: just 9% of the original extent of the moist tropical forest in the cacao-growing region of southern Bahia is left. Despite this situation, the area holds more old-growth and secondary forest patches with more endemic species and greater species richness than any other part of the Atlantic Forest, largely as a result of traditional agroforestry management known as cabruca (Alger and Caldas 1994, Thomas et al. 1998, Rambaldi and Oliveira 2003, Rolim and Chiarello 2004). In cabrucas, native overstory tree species are retained for their shade cover while the understory is cleared and replaced with high quality cacao (Theobroma cacao L., Malvaceae s.l.). Yet, the future of conserving native species within this unique agroforestry system is uncertain.

As human populations continue to increase locally in the cacao-growing region of Bahia, and world chocolate consumption is expected to grow by 2–3% per year (Lass 2004), the pressure on farmers to intensify cacao production is also likely to increase (Schroth and Harvey 2007). Over the last two decades, farmers have frequently opted to simplify the native shade canopies of cabruca agroforestry systems and replace them with early-successional and/or exotic tree species (Rolim and Chiarello 2004), or opt out of cacao cultivation entirely and convert the cabrucas to other agricultural land uses such as pasture or annual crop production, that are generally less compatible with biodiversity conservation (Schroth and Harvey 2007, Rayner et al. 2011). Sustainable land use practices are urgently needed to facilitate the conservation of biological diversity and natural resources, regenerate degraded agricultural fallows adjacent to native forest stands, while addressing economic pressures faced by local communities.

A driver of the high biodiversity of cabruca systems is the selective management and assisted regeneration of native overstory shade species, and selective plantings of food-bearing species within available niches amongst cacao stems, increasing complexity in vertical structure and compositional diversity (Young 2017). As part of a larger study comparing the structure and composition of cabruca forest to natural forest, I conducted vegetation surveys of four cabruca forests. Here, I document the diversity and floristics of the trees of these cabruca agroforestry systems.


Study site

I surveyed cabrucas in the tropical moist broadleaf forest zone of southern Bahia, Brazil (approximately 14–15°S and 39°W), at an elevation of 15–150 m asl. Mean annual temperature is 24°C (75.2°F) (Instituto Nacional de Meterologia, Brasil), and average annual precipitation 86.6–209.48 mm, with a distinct 4–5 month rainy season (World Bank 2016). Soils are classified as haplorthox oxisols typically high in iron and with low fertility (Piotto et al. 2009). Serra do Conduru State Park has an overall topography of undulating to rolling (10–30% slope; Piotto et al. 2009).

The cabruca sites are located within or adjacent to Serra do Conduru State Park: three to the east of the park (near Serra Grande) and one to the north of the park (15km outside of Itacaré). Cabruca sites had a minimum of 20 ha of managed cacao cultivation, $$40 yrs of cacao cultivation following clearing native forest for cacao cultivation, and similar site topography and soil to minimize site heterogeneity

Sampling design

At each of the four cabruca sites I established three 0.5 ha fixed area plots (total sampling area 6 ha). Cabruca plots were located in the center of at least two hectares of contiguous cacao cultivation to minimize edge effects, with 20m between each plot. Within each plot we established ten parallel transects (10m x 50m) along either N-S or E-W coordinates to minimize heterogeneity of spacing density that occurs on steep slope inclinations. The first transect of each plot was placed 10m from the corner of the plot. Within each transect I identified and measured all woody trees $$5cm diameter at breast height (DBH, measured using a calibrated diameter tape (cm) at approximately 1.37m above ground level). To account for the consistent presence of bananas and plantains (Musa spp.) in these agroforests, I also included monocots, Musa spp., and members of the Palmaceae family $$5cm DBH. Unknown species were flagged, and leaf samples were collected, dried, and mounted as herbarium vouchers at Instituto Floresta Viva for later identification.



I recorded a total of 4,435 trees in the four cabruca sites in a total of 6 ha. Mean density (±SD) per 0.5 ha plot was 740±85 trees >5 cm DBH ha-1 (range = 624–902). Over all trees, mean DBH was 15.5±15.8 cm (1–308 cm).

Within each site, however, most trees were Theobroma cacao (3,483) or Musa (198). Excluding these stems, the total number of trees was 754, mean density per plot was 126±35 (80–174), and mean DBH was 36.1±30 (4.5–308.0).


Within all the cabruca plots, I documented a total of 33 families, 61 genera, and 103 species. Nineteen individual trees remained unidentified to family, 18 to genus, and 346 to species. The mean number of families per plot was 14 (range = 9–19); the mean number of genera per plot was 18 (10–25); and the mean number of species was 20 (10–29).

Figure 1. The number of trees in the top 18 families in four cabruca stands in southern Bahia, Brazil. Most Malvacaeae are Theobroma cacao; most Musaceae are Musa spp.

Figure 1. The number of trees in the top 18 families in four cabruca stands in southern Bahia, Brazil. Most Malvacaeae are Theobroma cacao; most Musaceae are Musa spp.

Over all plots, the most common families, excluding Theobroma (Malvaceae) and Musaceae, were Leguminosae (104 trees), Rutaceae (94), Moraceae (82) and Meliaceae (61) (Fig. 1). A total of five families had only one individual and 17 families had 10 or more stems.

Over all plots, the most common genera were Citrus (94), Erythrina (65), Cedrela (61), and Artocarpus (54). A total of 14 genera were represented by a single individual, and 22 genera had 10 or more stems.

Over all plots, the most common identified species were Cedrela atlantica (61), Artocarpus heterophyllus (54), Cordia trichotoma (48), and Citrus tangerina (48). A total of 54 species were present as a single stem; 22 species had 10 or more stems (Fig. 2).

Figure 2. Dominance-diversity curves for four cabruca stands in southern Bahia, Brazil. Each point is a species. Most species had very few individuals; Theobroma cacao and Musa were the two most abundant. Note the log scale on the x-axis.
Figure 2. Dominance-diversity curves for four cabruca stands in southern Bahia, Brazil. Each point is a species. Most species had very few individuals; Theobroma cacao and Musa were the two most abundant. Note the log scale on the x-axis.

A constrained correspondence analysis of the families in each plot showed that sites grouped together (Fig. 3). The first axis explained 29% of the variation, and the second axis 22%. It is clear that the families with single species are driving the separation of sites in this analysis.

Figure 3. Biplot of a community correspondence analysis of four cabruca stands in southern Bahia, Brazil. Three plots (P) were located in each of four stands (C).

Figure 3. Biplot of a community correspondence analysis of four cabruca stands in southern Bahia, Brazil. Three plots (P) were located in each of four stands (C).


The goal of this study was to provide a baseline ecological analysis to test the underlining “mimicry” principle behind agroforestry design and management, and quantify the similarities in structural and compositional traits between a typical multi-strata cabruca cacao agroforest and an adjacent native secondary forest. In 6 ha of cabruca agroforestry plots, I found a high diversity of non-Theobroma species, with 33 different families and 103 different species. This richness is comparable to other cabruca systems. For example, Sambuichi and Haridasan (2007) found 180 tree species and Lobão and Valeri (2009) 101 tree species in other studies from Bahia.

Despite the high abundance of Theobroma cacao and Musa, the floristics of these cabruca systems reflected the general patterns found in the Atlantic Forest, with high abundance of Leguminosae, Moraceae, Meliaceae, among others (Oliveira-Filho and Fontes 2000). However, whether these agroforestry systems “mimic” natural forests remains to be seen. Cabrucas, rather, may be simplified in structure and floristic composition as compared to secondary forests. However, there is ample growing space available in the understory, subcanopy, and canopy to increase agrobiodiversity by selecting and planting native ethnobotanically valuable species in available vertical gaps in the strata (Young 2017). Thus, these agroforests could be managed like secondary forests following assisted natural regeneration and successional forest stand dynamics to improve vertical assemblages of multi-functional species at each stratum. These forest would then better incorporate native ecological principles into agroforestry design and management.


This research was conducted in collaboration with Universidade Federal do Sul da Bahia, Instituto Floresta Viva (Serra Grande, Bahia, Brazil), the New York Botanical Garden Institute of Economic Botany, and Yale School of Forestry & Environmental Studies with generous financial contributions from Tropical Resources Institute (TRI), the Yale School of Forestry & Environmental Studies, the MacMillan Center for Latin America and Iberian Studies, and the Carpenter-Sperry travel grant. I am grateful for the invaluable support from my academic and research advisors, Mark Ashton and Simon Queenborough (Yale University), Charles Peters (New York Botanical Garden, Institute of Economic Botany), and Prof. Daniel Piotto (Universidad Federal do Sul da Bahia), and for the hard work and dedication of my field research assistants, Jucelino Oliveira Santos, Ronildo, Edilson Damaceno, and William W. Young. Special thanks to my research site hosts at Fazendas Pedro do Sabiá, Lagoana, and São Francisco, to Wayt Thomas and Lawrence Kelly (New York Botanical Garden) for their counsel on botanical identification and research methods, to Timothy Gregoire, Craig Brodersen, Marlyse Duguid, Meghna Krishnadas, and colleagues in the Ashton Lab Group for their statistical and research advice, constructive criticism, and feedback (Yale University); to Ilana Stein (UC-Berkeley) for her technical advice and support in the field; and to Ajit Rajiva (Yale University) for his GIS and R troubleshooting skills.


Alger, K. & Caldas, M. 1994. The declining cocoa economy and the Atlantic Forest of Southern Bahia, Brazil: conservation attitudes of cocoa planters. Environmentalist 14, 107–119.

Galindo-Leal, C. & Câmara, I.D.G. 2003. Atlantic Forest hotspot status: an overview. The Atlantic Forest of South America: Biodiversity status, threats, and outlook 1, 3–11.

Lobão, D.E. & Valeri, S.V. 2009. Sistema cacau-cabruca: conservação de espécies arbóreas da floresta Atlântica. Agrotropica 21, 43–54.

Oliveira-Filho, A.T. & Fontes, M.A.L. 2000. Patterns of floristic differentiation among Atlantic forests in southeastern Brazil and the influence of climate. Biotropica 32, 973–810.

Piotto, D., Montagnini, F., Thomas, W., Ashton, M., & Oliver, C. 2009. Forest recovery after swidden cultivation across a 40-year chronosequence in the Atlantic forest of southern Bahia, Brazil. Plant Ecology 205, 261–272.

Rambaldi, D.M. & Oliveira, D.A.S. (orgs.) 2003. Fragmentação de Ecossistemas. Causas, Efeitos Sobre a Diversidade e Recomendações de Políticas Públicas. MMA/SBF, Brasília.

Rayner, J., Buck, A., & Katila, P. 2011. Embracing Complexity: Meeting the challenges of international forest governance. A Global Assessment. Report by the Global Forest Expert Panel on the International Forest Regime. International Union of Forest Research Organizations.

Rolim, S.G. & Chiarello, A.G. 2004. Slow death of Atlantic forest trees in cocoa agroforestry in southeastern Brazil. Biodiversity & Conservation 13, 2679–2694.

Sambuichi, R.H.R. & Haridasan, M. 2007). Recovery of species richness and conservation of native Atlantic forest trees in the cacao plantations of southern Bahia in Brazil. Biodiversity & Conservation 16, 3681–3701.

Schroth, G. & Harvey, C.A. 2007. Biodiversity conservation in cocoa production landscapes: an overview. Biodiversity & Conservation 16, 2237–2244.

da Silva J.M.C. & Tabarelli, M. 2000. Tree species impoverishment and the future flora of the Atlantic forest of northeast Brazil. Nature 404, 72–74.

Thomas, W.W., Carvalho, A.M.V., Amorim, A.M., Hanks, J.G. & Santos, T.S. 2008. In: Thomas, W.W. (Ed.) Diversity of woody plants in the Atlantic coastal forest of southern Bahia. The Atlantic Coastal Forests of Northeastern Brazil. Memoirs of the New York Botanical Garden 100, 21–66.

World Bank Group, 2016. Average Monthly Temperature and Rainfall for Brazil at location (-14.18,-38.96) from 1991–2015. Climate Change Knowledge Portal. Climate Change Group, World Bank. index.cfm?page=country_historical_climate& ThisCCode=BRA

Young, K.J. 2017. Mimicking nature: A review of successional agroforestry systems as an analogue to natural regeneration of secondary forest stands. In: Montagnini F. (ed) Integrating Landscapes: Agroforestry for biodiversity conservation and food sovereignty. Advances in Agroforestry, 12. Springer.

  1. Katherine (`Kata’) works at the intersection of sustainable agricultural production and ecological restoration of tropical environments, as an international advisor on regenerative agro-ecological land use management and climate change adaptation. Kata has 15 years of professional experience working with multi-stakeholders in vulnerable communities in Latin America, South Asia, and Africa. Her work has been conducted in collaboration with smallholder farmers, farmer field schools, farmer associations and cooperatives, civic and non-profit organizations, private sector actors and corporations, and public sector representatives (local, regional, and national governments, and international-level). In 2017, Kata received a joint Masters in Forest Science from Yale School of Forestry & Environmental Studies and the New York Botanical Garden, Institute of Economic Botany. She holds a B.S. in International Agriculture & Rural Development from Cornell University (’15), with a strong interdisciplinary foundation in international development, soil and agro-ecological management, gender and development sociology, and participatory research and farmer-to-farmer learning modalities. Kata currently resides in Washington, D.C., where she works as a consultant and volunteers as the coordinator for her neighborhood’s community garden.

TRI Profiles