Plant competition and facilitation after extreme drought episodes in Mediterranean shrubland: does damage to vegetation cover trigger replacement by juniper woodland?

QUESTION: We analyse the contribution of plant–plant interactions, particularly the outcome of plant competition and plant facilitation, on vegetation dynamics as a result of extreme drought episodes. These events will likely become more frequent under climate change, can induce vegetation die‐off and alter community dynamics. We study succession in a shrubland that tends to be replaced by juniper (Juniperus phoenicea) woodland. Due to drought, shrubland die‐off may result in competition release favouring J. phoenicea juveniles, and accelerating shrubland replacement. Alternatively, deleterious abiotic stress may increase after loss of vegetation cover protection. LOCATION: Mediterranean coastal shrublands, South Spain (Doñana National Park). METHODS: Field estimates of plant growth, production of needle‐like leaves, water‐use efficiency (WUE; leaf δ¹³C) and N leaf content of J. phoenicea juveniles in relation to plant size, drought‐induced damage, cover and habit characteristics of surrounding vegetation, and drought‐induced defoliation of the surrounding vegetation. RESULTS: Juniperus phoenicea juveniles growing beneath a dense vegetation canopy, particularly trees and large shrubs, were less damaged during the extreme drought episode. Plant size correlated negatively with damage. Post‐drought growth was higher in juveniles partially released from the vegetation canopy, supporting the existence of a balance between competition and facilitation. Cover of pines, large shrubs and spiny shrubs favoured growth of juveniles. Needle‐like juvenile leaves were more abundant in plants covered by the surrounding vegetation or in moderately damaged plants, but less abundant in plants without damage. Higher leaf δ¹³C values – indicating water stress – were measured in plants more damaged by drought and in those without canopy protection, or under vegetation strongly affected by drought. Leaf N content was lower in undamaged plants and individuals covered by surrounding vegetation. CONCLUSION: We did not find evidence that gaps opened by drought promoted growth of the potential replacing J. phoenicea. Thus, drought‐induced enhancement of successional replacement of shrublands with woodlands was not supported; instead, our findings foresee shrubland prevalence under future climate change conditions. Plant facilitation will play a relevant role in this process. Thus, we herein extend the relevance of plant–plant interactions to extreme drought episodes related to climat