Predicting alterations to diversity-productivity relationships across changing climates: complementarity increases with precipitation

Background/Question/Methods

Changing environmental conditions, such as drought, can have variable effects on biodiversity-productivity relationships, making it difficult to predict how these patterns will scale across landscapes. Predictions should be based on understanding of mechanisms driving the relationship (e.g. pathogens, resource partitioning), though manipulative tests are limited. We addressed this in a grassland biodiversity manipulation experiment by testing the interactive impacts of plant richness, precipitation, and community composition on productivity. We manipulated plant diversity, selecting from 17 plant species in three families (Poaceae, Asteracae, and Fabaceae) commonly represented in native tallgrass prairies. We manipulated species richness by randomly selecting 1, 2, 3, 5, or 6 species. To increase functional and phylogenetic dissimilarity within communities we assembled mixtures from one family or multiple families. Precipitation was altered to 50% and 150% of ambient conditions by rainfall manipulation shelters. To quantify productivity, we observed plot cover and harvest biomass strips at peak productivity 2018-2021. We partitioned the relative yields into the complementary (many species are productivity) and the selection (strong competitors are most dominant). We assessed whether pathogens, as estimated by potential pathogen divergence (PPDq), and resource partitioning, as estimated by functional trait diversity (FDq), correlated with increased complementarity (the main driver diversity-productivity relationships).

Results/Conclusions

Plant diversity generated higher total plant yields (F[3,229.47]=57.55, p< 0.001), a relationship that strengthened with experimental duration (F[3,702]=33.33, p< 0.001). Complementarity increased with richness (F[1,170]=10.57, p< 0.001) and was significantly higher with greater precipitation (F[1,152.10]=8.87, p=0.003). However higher selection effects in droughted treatments reduced differences in relative yields between precipitation treatments and offset them in models predicting total yields. Potential pathogen divergence (PPDq) positively predicted complementarity (F[1,412.07]=18.34, p< 0.001), consistently across the 2nd, 3rd, and 4th years. Functional trait diversity (FDq) increased complementarity. While this relationship strengthened with time in single-family mixtures (F[3,242.78]=6.98, p< 0.001), it weakened with time in multiple-family mixtures (F[3,243.48]=3.85, p=0.01), even though complementarity increased overall with time in these plots (F[3,243.81]=12.03, p< 0.001). Together these results suggest that pathogen dilution is an important mechanism generating greater plant yields with diversity, and that resource partitioning also contributes for some plant species combinations. Our findings help us mechanistically understand where biodiversity might generate the greatest productivity benefits. The pathogen dilution mechanism, for example, predicts that environments conducive to severe plant pathogen impacts, such as high precipitation, will be when and where ecosystems are most sensitive to biodiversity loss.