Miguel Berdugo, Blai Vidiella Rocamora, Ricard Solé, Fernando T Maestre

With ongoing climate change, the probability of crossing environmental thresholds promoting abrupt changes in ecosystem structure and functioning is higher than ever. In drylands (sites where it rains less than 60% of what is evaporated), recent research has shown how the crossing of three particular aridity thresholds (defining three consecutive phases, namely vegetation decline, soil disruption and systemic breakdown) leads to abrupt changes on ecosystem structural and functional attributes. Despite the importance of these findings and their implications to develop effective monitoring and adaptation actions to combat climate change, we lack a proper understanding of the mechanisms unleashing these abrupt shifts.Here we revise and discuss multiple mechanisms that may explain the existence of aridity thresholds observed across global drylands, and discuss the potential amplification mechanisms that may underpin hypothetical abrupt temporal shifts with climate change. We found that each aridity threshold is likely involving specific processes. In the vegetation decline phase we review mainly physiological mechanisms of plant adaptation to water shortages as main cause of this threshold. In the second threshold we identified three pathways involving mechanisms that propagates changes from plants to soil leading to a soil disruption: erosive mechanisms, mechanisms linked to an aridity-induced shrub encroachment and mechanisms linked to nutrient cycling and circulation. Finally, in the systemic breakdown phase we reviewed plant-plant amplification mechanisms triggered by survival limits of plants that may cause sudden diversity losses and plant-atmospheric feedbacks that may link vegetation collapse with further and critical aridification. By identifying, revising and linking relevant mechanisms to each aridity threshold, we catalogued a set of specific hypotheses and recommendations based on identified knowledge gaps concerning the study of mechanisms of threshold emergence in drylands. Moreover, we were able to establish plausible factors that are context dependent and may influence the occurrence of abrupt changes in time and we created a mechanistic-based conceptual model on how abrupt changes may emerge as aridity increases. This has importance for focusing future research efforts on aridity thresholds and for developing strategies to track, adapt to or even revert these abrupt ecosystem changes in the future.