Decoupled Climatic Drivers of Tree and Ground-Layer Carbon Uptake in Mountain Ecosystems Around the World.

One of the key ecological processes affected by climate change is plant carbon uptake. However, there is substantial uncertainty about how plant carbon uptake will respond to warming in mountain ecosystems, which are known for sharp temperature gradients and abrupt shifts in vegetation structure. Specifically, we lack an understanding of whether these response trajectories over time will be linear or non-linear, and how they might vary among mountain ecosystems globally. Here, we measured ecosystem Gross Primary Productivity (GPP) along forest-tundra elevational gradients in the mountain regions of five countries (Argentina, Australia, France, Sweden, USA) to infer future trajectories of carbon uptake, and whether any non-linear changes might occur. We also examined the role of microclimate in driving GPP responses. We found that whole-ecosystem GPP increased with increasing macroclimatic temperature (decreasing elevation), but this response was dominated by a sharp non-linear increase at the transition from tundra to forest (i.e., the treeline). In contrast, ground-layer GPP was largely independent of macroclimate, but often responded strongly, and linearly, to microclimate (growing degree days > 5°C, mean growing season temperature). This pattern reflected a frequent decoupling of microclimate from the expected temperature-elevation relationship, likely driven by such processes as cold-air drainage, limited near-surface air mixing, and shading by trees. The contrasting responses of GPP among global temperature gradients indicates strong context dependence at both local and continental scales, although in a few cases, biomass and leaf nitrogen were important moderators. These findings suggest that future shifts in carbon uptake in mountains will be mainly controlled by tree range expansion. Our results highlight the need to consider species responses on different spatial scales, and to increase representation of undersampled regions to capture the full breadth of ecological responses to climate change.