Agricultural diversification can enhance climate resilience, biodiversity conservation, and livelihood in global farming systems. Diverse agroforestry systems with cocoa have been shown to provide all these benefits, but the often-lower yields compared to monocultures limit agroforestry adoption by smallholder farmers. Cocoa yield is pollination-limited, and here, we quantified the effect of hand pollination in cocoa on fruit set, fruit abortion or cherelle wilt, pest and diseases, and number of mature fruits. Experiments were conducted in Bahia, Brazil, along a shade gradient from low [10–30%] to high [70–100%] canopy cover and with cocoa trees grafted with high-yielding varieties. We found on average 331% fruit set, and 300% mature fruit increase (i.e. ripe pods) per tree by enhancing pollination by hand as little as 10% of the flowers/tree, compared to the control (i.e. with only natural pollination). Fruit set, fruit losses, and mature fruit development linked to hand pollination was higher in low compared to highly shaded cocoa areas. We found 31% higher fruit set and 37% higher number of mature fruits in grafted than un-grafted trees. Further, when comparing efforts invested in hand pollination, minimum labor (∼5 min/tree to pollinate 10–30% of the flowers/tree) led to a 871% fruit set and a 750% mature fruit increase under low canopy cover, and intermediate labor (∼15 min/tree to pollinate 40–60% flowers/tree) to a 629% fruit set under high canopy cover. As ∼5–15 min/tree hand pollination can substantially enhance fruit set and number of mature fruits in low as well as high shade management, we recommend performing hand pollination particularly in agroforests under 40–50% canopy cover to create win-win opportunities for high productivity and climate resilience. Future research should focus on the wide range of agroforestry, tree grafting and innovation strategies in all major production regions to capture the long-term variability of hand pollination as a basis to scale-up hand pollination for sustainable cocoa production globally.

农业多样化可以增强全球农业系统的气候适应能力、生物多样性保护和生计。多种可可农林复合系统已被证明可提供所有这些好处，但与单一种植相比通常较低的产量限制了小农对农林复合的采用。可可产量受授粉限制，在这里，我们量化了可可手工授粉对坐果、落果或黑枯病、病虫害以及成熟果实数量的影响。实验是在巴西巴伊亚州进行的，沿着从低 [10–30%] 到高 [70–100%] 树冠覆盖的遮荫梯度，并在可可树上嫁接了高产品种。我们发现平均 331% 的坐果率和每棵树 300% 的成熟果实增加（即成熟的豆荚）通过手工授粉仅增加 10% 的花/树，与对照相比（即只有自然授粉）。与高度遮荫的可可区域相比，与人工授粉相关的坐果率、果实损失和成熟果实发育在低地更高。我们发现嫁接树的座果率和成熟果实数量比未嫁接树高 31% 和 37%。此外，当比较手工授粉的投入时，最低劳动力（约 5 分钟/树授粉 10-30% 的花/树）导致 871% 的坐果率和 750% 的成熟果实在低冠层覆盖下增加，并且中间劳动（约 15 分钟/树，为 40-60% 的花/树授粉）在高冠层覆盖下结出 629% 的果实。由于每棵树人工授粉约 5-15 分钟可以在低荫和高荫管理下显着提高坐果率和成熟果实的数量，我们建议进行人工授粉，尤其是在树冠覆盖率低于 40-50% 的混农林中，以创造高生产力和气候适应力的双赢机会。未来的研究应侧重于所有主要产区的各种农林业、树木嫁接和创新战略，以捕捉人工授粉的长期可变性，以此作为在全球范围内扩大人工授粉以实现可持续可可生产的基础。