Disturbances regimes are defining characteristics of ecosystems and regions. Understanding and describing disturbance regimes is a central theme in seminal branches of ecology. Disturbances control ecosystem processes, biogeochemical fluxes, and landscape heterogeneity. Many climate-sensitive disturbances (Fire, land-use change, pest and pathogens) spread contagiously. However, the earth system models we use to predict future climate-ecosystem interactions simplify or ignore the effects of sub-pixel size contagious disturbances. We derived a novel spatially implicit method of scaling contagious disturbances. A method for characterizing disturbance regimes emerged from this derivation: We describe disturbance regimes as a function of size distributions and the ratio of interior-edges to total-edges. We first validated our derivation against a spatially explicit simulation. We then explored the theoretical implications of these emergent metrics by looking at how they captured the disturbance regimes across US state, ecoregions, and mechanism. We used the remotely-sensed disturbance product LANDFIRE to look at the size, interior-total ratio, and mechanism of disturbance across 4 ecoregions within 2 US states (Oregon and Florida) for 2014. We compared size distributions with Kolmogorov–Smirnov tests. We also examined how the relationship between size distribution and interior-total ratio changed between disturbance mechanisms and regions by comparing fitted curves using a maximum likelihood test.
Results/Conclusions
Our spatially implicit scheme was abe to dynamically update contagios disturbance within acceptable levels of error when compared to a spatially explicit simulation after 1000 years. We made 66 pairwise comparisons of size distributions among 12 disturbance types, and 3 comparisons among state and two ecoregions. The size distributions of Florida and Oregon were found to be different (p < 0.001), and two ecoregions within Oregon were found to be different. The two ecoregions within Florida were not found to be significantly different (p < 0.01). Comparisons of fit curves found a significant effect of US state (p < 0.001) and ecoregion nested within state (p < 0.01). We also found a significant effect of disturbance type (p < 0.0001), but not US state nested within disturbance (p > 0.1). Comparisons of size distributions and curves showed that anthropogenic disturbances have different spreading patterns than natural disturbances. Our metrics were able to capture important differences in disturbance mechanism and overall disturbance regime. We recommend further evaluation of these metrics in future work.