Quantifying crop pollinator dependence and its heterogeneity using multi-level meta-analysis.
Biotic pollination can benefit crop production, but its effects are highly variable. To maximise benefits from this ecosystem service, we need a greater understanding of the factors that cause variation so that ecological intensification can be more effectively applied. We focus on understanding the benefits of pollination to faba bean Vicia faba. We use a literature review followed by multi-level meta-analysis to estimate overall benefits of pollination to faba bean yield and to quantify variation (heterogeneity) in these benefits associated with different contextual factors (e.g. plant genotype, growing environment). Our overall estimate of pollination benefit to faba bean yield, expressed as the percentage yield reduction without pollination, is 32.9% (confidence interval: 21%-43%). Based on the prediction intervals, which include the heterogeneity in pollination benefit, there is an 80% chance that pollination will increase yield of a faba bean crop. Half of all heterogeneity in pollination dependence was due to differences between plant genotypes. The number of beans per plant showed similar pollination dependence to yield mass per plant while pod number and number of beans per pod underestimated yield benefits. There was weak evidence to suggest pollination benefits vary between pollinator species, with honeybees showing a smaller yield increase. Differences in the experimental method used to assess pollination benefit did not significantly affect the estimate, including the growing environment, measurement scale, or whether the effects of experimental pollinator enclosures were controlled. This suggests that simplified experimental studies comparing yield of open-pollinated and enclosed plants can provide reliable insights into pollination benefits across a large range of plant genotypes and landscapes. Synthesis and application. We found high variability in pollination benefits both between and within publications in our meta-analysis. Plant genotype, how yield was measured, and pollinator species affected the level of pollination benefit. Despite variability in pollination benefits due to various contextual factors (both inside and outside of grower control), there is a high likelihood that biotic pollination will increase faba bean yield. Our findings support ecological intensification and specifically the management of pollinators to maximise pollination benefits to faba bean production.