In the Anthropocene, meeting production demands through agriculture systems has been a primary challenge for governmental sectors and farmers, especially given that this comprises the most adversely affected sector by environmental changes including global climate change (Foley et al., 2005; Tilman et al., 2011; IPCC, 2014). Therefore, discussions on the concept of sustainable intensification (SI) in agriculture have significantly expanded in recent decades, especially focusing on prioritizing the development of practices that boost productivity at the same time provide social and environmental benefits (Tilman et al., 2002, 2011; Rockström et al., 2017). In particular, cocoa (Theobroma cacao) is a key crop able to offer SI benefits when cultivated under climate-friendly agroforestry systems, given that emergent trees are maintained in the property to provide shade for cocoa tree development (Street and Legon, 2014; Niether et al., 2020). As a result, cocoa agroforests can store significant amounts of carbon, depending on the management practices adopted by the farmer. In addition, it comprises the third most worldwide traded agricultural product, with millions of smallholder farmers involved in its production (Ariza-Salamanca et al., 2023). Cocoa naturally thrives in shaded environments, given that it comprises a shade-tolerant Amazonian tree, yet shade requirements for optimal growth are debated among scientists and producers to date (Cabala Rosand, de Santana and de Miranda, 1976; Asare et al., 2017, 2019). While reduction of shade level in established cocoa plantations has initially shown significant yield increases, long-term trials highlight various detrimental effects such as decreased lifespan of the cocoa trees, increased pest and disease damage, and greater demand for agrochemical inputs (Ahenkorah, Akrofi and Adri, 1974; Ahenkorah et al., 1987). However, growing global demand for cocoa beans has driven the expansion of cocoa-producing regions through deforestation, and a shift from agroforestry to full-sun monocultures in many production areas, involving significant use of agrochemicals and, at times, of irrigation to increase productivity (Andres et al., 2016; Schneider et al., 2017).
Traditional cocoa agroforest transitions to low-shade plantations became prominent in the mid-1980s, with the widespread assumption among farmers that local shading is negatively associated with production (Asare et al., 2010). The removal of shade trees is particularly alarming in biodiversity-rich regions, where the structurally complex agroforests are known to host many native species and deliver various ecosystem services (Harvey et al., 2006; Schroth and Harvey, 2007). This is the case of southern Bahia, in Brazil, where cocoa agroforests (locally known as "cabrucas") provide habitat for native and threatened biota of the Atlantic Forest biome, one of the global biodiversity hotspots (Faria et al., 2006, 2023; Cassano et al., 2009; Schroth et al., 2011). Cabrucas covers approximately 6,000 km² in southern Bahia, surpassing the native forests in extent (Landau and Hirsch, 2008), and contributing to curb deforestation in several municipalities. As a result, approximately two-thirds of the regional above-ground carbon stock in southern Bahia is estimated to be stored by cabrucas (Schroth et al., 2015).
In the 1960s and 1970s, the Executive Committee of the Cocoa Farming Plan (CEPLAC), the government agency responsible for promoting cocoa production in Brazil, initiated an extensive program of tree removal in southern Bahia, aiming to maximize cocoa yields with minimal shading and encouraging fertilizer application (Cabala Rosand, de Santana and de Miranda, 1976; Johns, 1998). Despite short-term yield boosts, the economic reason for reducing shade in cocoa systems was questionable, given the wide range of ecological and economic benefits provided by shade that are lost or minimized, particularly in sustainable agriculture contexts (Beer, 1987). Currently, this local intensification is supported by a state law stating that each cabruca must maintain a minimum of 20 native trees per hectare (State Decree No. 15180/2014, Bahia), a significant reduction in tree abundance compared to the traditional cabrucas that typically harbor ~ 200 trees/ha (Sambuichi et al., 2012; Schroth et al., 2016). Nevertheless, the Bahia region has maintained a mean cocoa productivity ranging between 230–300 kg/ha− 1 of dried cocoa beans since the 2000s (IBGE, 2019), which is significantly lower than the global average of approximately 450–500 kg/ha− 1 (FAOSTAT, 2022).
High levels of shade are often identified as significant contributors to low regional cocoa productivity in southern Bahia. Nonetheless, research indicates that it is feasible to surpass the below-average yield levels (585 kg/ha− 1) while maintaining relatively high above-ground carbon stocks (e.g., 65 Mg/ha) and shade levels (e.g., 55% shade) through effective management practices, including the application of fertilizers (Schroth et al., 2016). Meanwhile, the absence of technical assistance (for instance, 75% of producers in southern Bahia received no technical support between 2011 and 2017) leads to limited input usage and inadequate management practices (Chiapetti et al., 2020). Regarding management in the Bahia region, the most common practices include weed control and both heavy and light pruning of the cocoa trees, whereas 53% of farmers never used any type of fertilizer on their property (Chiapetti et al., 2020). Therefore, understanding the synergy between production, carbon storage, and shading becomes crucial to ensure the adoption of appropriate management practices that optimize a wide range of ecosystem services, including carbon sequestration, and simultaneously maintaining shade trees and stocking high levels of carbon.
Here, we intend to contribute to maximizing biodiversity conservation, while benefiting the livelihoods of cocoa producers in agroforests of southern Bahia. For this, we (i) investigated the main determinants of cocoa productivity, including, key landscape and local factors (i.e., landscape forest cover, local vegetation structure, aboveground carbon stock, shade levels, and management practices) aiming to identify how producers can optimize yields; (ii) estimated aboveground carbon stocks in surveyed cabrucas and subsequently assessed the influence of landscape and local factors on such reservoirs. We hypothesized that sustainable intensification would be possible, as management practices would emerge as the primary factor affecting cocoa productivity. We finally hypothesized that cabrucas have the potential to achieve higher productivity of cocoa than the world average productivity of 585 kg/ha− 1, while also maintaining shade trees and carbon stocks above the regional average in southern Bahia.