COS 7-1
Quantifying strength and scale of feedback mechanisms to prevent terrestrial regime shifts
Ecologists are increasingly adept at identifying ecological thresholds once they have been crossed, but the ability to identify thresholds a priori remains elusive. If we cannot identify thresholds before they are crossed, many ecosystems could undergo abrupt and difficult to reverser regime shifts, sometimes to undesirable socialecological states. Theoretical models suggest that as ecosystems approach a threshold negative feedbacks defining the current state become weaker, evinced as spatial auto-correlation and stronger antagonistic interactions between competing functional groups. However, these theoretical concepts have not been tested in terrestrial ecosystems. To fill this gap, we quantified how the strength and scale of reinforcing feedback mechanisms changes over the course of a 28 year experiment manipulating grassland fire frequency. Our analyses use an extensive multi-scalar network of observational plots with annual resolution (n =320), which allows us to observe non-linear regime shifts, create null models from plots that are far from critical thresholds, and quantify if and when feedback strength change at different scales.
Results/Conclusions
Treatments burned every 4 to 20 years undergo an abrupt transition to shrub-dominance after 20 years, linked to grass-cover and fire frequency thresholds. Leading up to the abrupt transition, shrub and grass cover a small scales become negatively correlated, indicative of declining resilience and a weakening of the fire intensity feedbacks that maintain grass dominance. Five to six years later, grass*shrub anti-correlation spreads to larger scales, suggesting that the spatial of footprint of shrub-favoring feedbacks is growing. In a sub-set of the 20-year burn treatments, annual fires were restored before this spread of feedback mechanisms to larger scales. This timely intervention halted the trend of increasing shrub expansion, whereas plots retaining a 4- or 20-year fire frequency saw a 4-fold increase in shrub cover over the same decade. These results open new doors to managing resilience in terrestrial systems: we may be able to predict and prevent unwanted critical transitions by measuring feedback dynamics and altering exogenous pressures before declines in resilience spreads to larger scales. Prominent parallels include the tropics and tundra, where small-scale feedback mechanisms are breaking down, but large-scale buffering processes have yet to fail.