The phenomenon of woody cover bifurcation, that is, the co-occurrence of high and low woody cover under similar precipitation conditions, has been greatly discussed, and sometimes observed, in the world’s tropical savannas. Theory suggests that competition-facilitation relationships, fire and grazing, may all lead to alternate stable states of savanna vegetation structure. However, bifurcations are likely to emerge at different spatial scales in response to specific feedback mechanisms (e.g. fire-induced bifurcation may be observed at landscape scales, depending on fire propagation characteristics, while competition-facilitation mechanisms will likely induce bifurcation at much finer spatial scales consistent with interactions between individual plants). Here we carry out a multi-scale analysis using very high resolution satellite data across African savannas to examine scale-dependencies in spatial patterning that may reveal potential mechanisms.
Results/Conclusions
Examination of the spatial patterns of woody cover across 3 orders of magnitude, from 100-103 m, revealed distinct patterns and potential drivers of bifurcation across spatial scales. At fine spatial scales (100–101 m) canopy aggregation is associated with individual canopies, and local scale aggregation suggesting competition-facilitation relationships. At coarser spatial scales (102–103 m) the occurrence of larger patches of high tree cover and open grasslands may reflect fire-induced bifurcation. However, the density of woody plants can also be affected by local topo-edaphic conditions which may confound detection of alternative states (since theory assumes that alternative states emerge under constant climatic and edaphic conditions). To conclude this analysis, we explore correlations between the presence and length scales of apparent bifurcations and the underlying variability in the soil and geomorphic template to infer whether and when apparent bifurcations are truly the emergent and self-organized consequences of positive feedbacks, or may more simply be explained by variations in the underlying soil and geomorphology. Our analysis contributes to improved understanding of savanna ecology and alternative state theory, while providing new insight into the dynamics of African savannas as impacted by varying climate and disturbance processes.