When the growing gets tough, the tough stop growing: Intra-specific trait variation reveals divergent physiological constraints of drought on tree range boundaries
Species geographic range shifts are projected to be one of the most prevalent ecological responses to anthropogenic climate change, and have already been documented worldwide. Yet our knowledge of the ecological forces that actually constrain range boundaries is insufficient to mechanistically project when, where, and why range boundaries will move. We explored clinal variation in drought tolerance and drought avoidance-related plant functional traits across the elevation range of a gymnosperm (ponderosa pine - Pinus ponderosa) and an angiosperm (trembling aspen - Populus tremuloides) in Southwest Colorado, USA to determine the physiological mechanisms through which drought stress controls the low elevation range boundary of each species. Specifically, we quantified tree-to-tree and range-wide variation in xylem tension, hydraulic traits (branch conductivity, resistance to embolism, hydraulic safety margin) and morphological traits (SLA, leaf/needle size, tree height, wood density, leaf area:sapwood area ratio).
We found that elevational patterns in morphological and hydraulic trait variation differed between aspen and ponderosa pine, even as both species showed similarly large decreases in radial growth at their dry range edges. Most traits varied significantly in one species or the other, but rarely in both species. Trembling aspen grows more carbon intensive branch and leaf tissues in order to increase drought tolerance at its dry range edge, implying a range constraint enforced by the increasing cost of growth. By contrast, ponderosa pine appeared to avoid drought stress at low elevations by limiting transpiration via stomatal closure. This indicates a dry range boundary constrained by increasingly limited carbon assimilation. Our results suggest that the same climatic factor can drive range boundaries through different physiological mechanisms, which has implications for process-based modeling approaches to tree biogeography. Our results also show that comparing intra-specific patterns of trait variation across ranges, something that is rarely done in a range limit context, can provide a useful approach to identifying these mechanistic differences.