The influence of floral resources and microclimate on pollinator visitation in an agro-ecosystem

As agriculture expands to meet the needs of a growing global population, natural ecosystems are threatened by deforestation and habitat fragmentation. Tropical agroforestry systems offer a sustainable alternative to traditional agriculture by providing food for production while also supporting biodiversity and ecosystem services. Previous studies have shown that these systems may even improve crop pollination, but the mechanisms of how these improvements occur are still poorly understood. Using coffee as a focal crop, we explored how microclimatic conditions affected nectar traits (sugar and caffeine concentration) important for pollinator visitation. We also studied how microclimate, floral traits, floral availability at the coffee plant level, availability of floral resources provided by other plant species in the agroecosystem (neighborhood floral availability), and the presence of other bees affected the amount of time bees spent foraging on coffee flowers and the proportion of coffee pollen carried on their bodies. We explored these factors using the two dominant coffee species farmed on Puerto Rico, Coffea canephora and C. arabica, under sun and shade management. We found that high nectar sugar concentration and temperature were important predictors of short floral visits (<15 seconds), while increased number of bees and open coffee flowers were important predictors of longer floral visits (16-180 seconds). High nectar caffeine concentration was an important predictor of longer visits on C. arabica flowers while the opposite was observed for C. canephora flowers. For both species, high coffee floral availability was the main predicting factor for the proportion of coffee pollen on the bees bodies. Surprisingly, neither neighborhood floral availability nor the type of coffee plantation (agroforest/shade or sun) were important predictors of bee visitation. These results suggest non-coffee flowering plants in coffee plantations were neither competitors nor facilitators of coffee plantes for pollinators. Additionally, most of the bees surveyed were carrying 80% pollen from one species (C. arabica or C. canephora), likely resulting in little heterospecific pollen deposition between Coffea and non-Coffea flowers. Shade trees in coffee plantations do not detract from pollinator visitation to coffee flowers, suggesting that the provision of multiple ecological and wildlife conservation benefits by shade trees is not in conflict with a growers ability to maximize the benefits of insect pollination on fruit production.