ABSTRACT Understanding the context‐dependent nature of diversity‐productivity relationships is essential for predicting community responses to environmental change and informing restoration strategies. We conducted a reciprocal common garden experiment across a rainfall gradient (mean annual precipitation: 500–1200 mm year −1 ) in the US Great Plains to examine how local ecotypes of dominant grasses e.g., Andropogon gerardi Vitman and Sorghastrum nutans (L.) Nash and experimental drought (rainout manipulation) shape diversity‐productivity relationships. We applied generalized linear mixed models to assess the effects of aboveground live biomass, seed source, and rainfall manipulation on grassland species richness, phylogenetic diversity, and functional diversity. Functional diversity was positively correlated with aboveground live biomass at the dry site. Dominant species ecotypes influenced diversity‐productivity relationships in a site‐specific manner. The DRY ecotype reduced species richness as aboveground live biomass increased at the driest and mesic sites while the MESIC‐ecotype increased phylogenetic diversity with rising aboveground live biomass at the mesic site, with no effects observed on functional diversity. We further applied generalized boosted regression trees to assess the relative importance of aboveground live biomass, seed source, and rainfall manipulation in shaping grassland diversity. Across all sites, aboveground live biomass was the strongest predictor of diversity, whereas the effects of seed source and rainfall were weaker and more variable, likely reflecting the dominant role of locally adapted grasses in structuring communities and the limited translation of rainfall manipulation into consistent compositional change. These findings highlight the central role of dominant species attributes in shaping diversity‐productivity relationships and stress the importance of considering local adaptation and site context in restoration and biodiversity‐function research.
Ren et al. (Wed,) studied this question.