Testing a model of biological soil crust succession

Question Biocrust composition is important for the diversity, structure and function of dryland ecosystems globally, but there is no universally applicable model of biocrust succession, and the models that do exist may not apply everywhere. We ask: (1) how do biocrust species vary in their response to time since livestock exclusion (Time) across environmental gradients; (2) is there evidence for ruderal , mid ‐ and late successional responses to Time in biocrust species; (3) can exploration of species attributes within response groups reveal potential mechanisms for biocrust succession; and (4) do our data support an existing standard conceptual model of biocrust succession described in the literature. Location Semi‐arid woodlands of northwest Victoria, Australia. Methods We use data from a space‐for‐time study of biocrust composition following livestock exclusion and develop a hierarchical regression model of succession amongst a moss‐dominated biocrust community. Results Time since livestock exclusion (Time) was the most important driver of biocrust community dynamics in our study region; species varied widely in their response to Time. Succession within this moss‐dominated biocrust community did not follow the general sequence predicted by the standard model of biocrust succession presented in the literature. We classified species by their successional response from model coefficients. All late successional species were short‐statured and closely attached to the soil, whereas nearly all tall‐statured and loosely attached species were classified as mid‐successional. Conclusions This study demonstrates the utility of hierarchical models for investigating successional processes at the species and community level, for testing a standard conceptual model of succession and for suggesting hypotheses about underlying trait‐based mechanisms. We propose an alternative model of biocrust succession that distinguishes between succession in favourable and unfavourable environments. We propose species response to Time is mediated by species structural density, which in turn relates to a suite of correlated traits including growth rate and longevity. This alternative conceptual model is useful for its generality.