Plant traits are differentially linked to performance in a semiarid ecosystem.

A central principle in trait-based ecology is that trait variation has an adaptive value. However, uncertainty over which plant traits influence individual performance across environmental gradients may limit our ability to use traits to infer ecological processes at larger scales. To better understand which traits are linked to performance under different precipitation regimes, we measured above- and belowground traits, growth, and reproductive allocation for four annual and four perennial species from a coastal sage scrub community in California under conditions of 50%, 100%, and 150% ambient precipitation. Across water treatments, annual species displayed morphological trait values consistent with high rates of resource acquisition (e.g., low leaf mass per area, low root tissue density, high specific root length), and aboveground measures of resource acquisition (including photosynthetic rate and leaf N concentration) were positively associated with plant performance (reproductive allocation). Results from a structural equation model demonstrated that leaf traits explained 38% of the variation in reproductive allocation across the water gradient in annual species, while root traits accounted for only 6%. Although roots play a critical role in water uptake, more work is needed to understand the mechanisms by which root trait variation can influence performance in water-limited environments. Perennial species showed lower trait plasticity than annuals across the water gradient and were more variable as a group in terms of trait-performance relationships, indicating that species rely on different functional strategies to respond to drought. Our finding that species identity drives much of the variation in trait values and trait-performance relationships across a water gradient may simplify efforts to model ecological processes, such as productivity, that are potentially influenced by environmentally induced shifts in trait values.