Consistency in the relationship between landscape diversity and crop yields

Keywords: landscape diversity, crop yields, agricultural systems, ecosystem services
Abstract Type: Paper Abstract

Abstract

A prominent review of biodiversity effects on ecosystem function published in Nature in 2012 posited several consensus statements that suggest that increasing diversity produces consistent positive effects on ecosystem function, including provisioning ecosystem services. Moreover, they suggest that this relationship is nonlinear and saturating and that diversity in functional traits are a key factor in increasing ecosystem function. However, they also state that exceptions to these relationships exist and point to a need to explore the boundaries that constrain diversity effects. Today many of these consensus statements remain untested and the boundaries of diversity effects are still not well defined. In the agricultural domain, landscape-scale observational work, has shown mostly positive effects of increasing diversity on crop yields although these effects differ across the crops, landscape metrics, and geographic contexts evaluated. In this study we aim to quantify the bounds of the relationship between landscape diversity and crop provisioning ecosystem services. Using a Bayesian stacking approach we produce mixture models that represent the consistency, or lack thereof, of estimated functional relationships between county landscape diversity and yields for five major crops grown in the U.S. across a range of landscape metrics, land use classification schemes, and approaches to computing metrics. Overall, we observe a weak positive and nonlinear relationship between increasing landscape diversity and crop yields across all conditions, however significant variation across crops, metrics, and classifications are observed that highlight both the bounds of diversity effects and the challenges in measuring and modeling these effects.