Abstract
Functional traits are often used as species-specific mean trait values in comparative plant ecology or trait-based predictions of ecosystem processes, assuming that interspecific differences are greater than intraspecific trait variation and that trait-based ranking of species is consistent across environments. Although this assumption is increasingly challenged, there is a lack of knowledge regarding to what degree the extent of intraspecific trait variation in response to varying environmental conditions depends on the considered traits and the characteristics of the studied species to evaluate the consequences for trait-based species ranking. We studied functional traits of eight perennial grassland species classified into different functional groups (forbs vs. grasses) and varying in their inherent growth stature (tall vs. small) in a common garden experiment with different environments crossing three levels of nutrient availability and three levels of light availability over 4 months of treatment applications. Grasses and forbs differed in almost all above- and belowground traits, while trait differences related to growth stature were generally small. The traits showing the strongest responses to resource availability were similarly for grasses and forbs those associated with allocation and resource uptake. The strength of trait variation in response to varying resource availability differed among functional groups (grasses > forbs) and species of varying growth stature (small-statured > tall-statured species) in many aboveground traits, but only to a lower extent in belowground traits. These differential responses altered trait-based species ranking in many aboveground traits, such as specific leaf area, tissue nitrogen and carbon concentrations and above-belowground allocation (leaf area ratio and root : shoot ratio) at varying resource supply, while trait-based species ranking was more consistent in belowground traits. Our study shows that species grouping according to functional traits is valid, but trait-based species ranking depends on environmental conditions, thus limiting the applicability of species-specific mean trait values in ecological studies.