Does canopy structure affect carbon cycling resistance to disturbance?: Insights from an ecosystem experiment

Sustaining the temperate forest carbon (C) sink amidst rising disturbance prevalence requires understanding of the canopy structural features that stabilize C cycling processes. Lidar-based descriptions of vegetation quantity, distribution, and heterogeneity within forest canopies are strongly correlated with C cycling processes such as net primary production (NPP); unknown, however, is which canopy structural features forestall or shift C cycling tipping points triggered by increases in disturbance severity and different sources of disturbance.

In May 2019, we implemented the Forest Resilience Threshold Experiment (FoRTE) at the University of Michigan Biological Station, girdling >3600 trees in 16 hectares to achieve 45%, 65%, and 85% gross defoliation in four forest ecosystems varying in pre-disturbance canopy structure. At each defoliation level, we implemented separate "top-down" and "bottom-up" disturbances, targeting canopy and subcanopy trees, respectively, to simulate different disturbance types. Before and after disturbance, we measured wood NPP (NPPw), soil respiration (Rs), and leaf net CO₂ assimilation (A) under saturating light, and we used terrestrial lidar to derive measures of canopy structure. The principal goal of FoRTE is to identify the mechanisms that underlie C cycling stability - or decline - and, through coupled modeling experiments, determine what information is integral to simulating forest responses to disturbance.

Our initial results suggest that pre-disturbance canopy structure affects the resistance of some but not all C cycling processes. Forests with less complex canopy structures generally exhibited a larger magnitude of change in NPPw. In contrast, Rs displayed low resistance to rising disturbance severity, decreasing non-linearly, but the response was not dependent upon pre-disturbance canopy structure. Canopy composition and complexity influenced the degree of change in above-ground C cycling processes, with low complexity, oak-dominated stands displaying more rapid crown and leaf physiological degradation than biologically diverse, structurally complex stands. We conclude that pre-disturbance canopy structural characterization could enhance the prediction and management of C cycling stability, particularly for above-ground processes.