Assessing the long-term contribution of nurse plants to restoration of Mediterranean forests through Markovian models


*Correspondence author. E-mail:


  • 1Based on short-term experimental data, facilitative interactions between woody plants (nurse-recruit interactions) have been described as essential for the restoration of Mediterranean forests. However, the long-term effects of nurse plants on vegetation dynamics are unknown. This study aims to project post-fire vegetation dynamics from easily retrieved data, and to asses the long-term contribution of nurse plants to forest restoration.
  • 2In an area burned 20 years ago, we compared post-fire dynamics in three states of regeneration: pioneer scrubland; spontaneous pine regeneration stands; and late successional scrubland. For each regeneration state, we obtained an interaction matrix with the frequency of recruitment of a given species under the canopy of every other species in the community. These matrices provided the raw data to develop Markov chain models of community dynamics. We used sensitivity analyses to explore how small shifts in replacement probabilities between species (or between functional groups) may affect the similarity of the projected community to an undisturbed reference community.
  • 3Plots in the pioneer state had the lowest frequency of facilitative interactions. Matrix projection showed that, under the current frequency of facilitation, these pioneer scrublands would remain so in the long term. By contrast, the building state had the highest frequency of facilitation, and its projection suggested that it should reach a steady state very similar to the reference community. These results confirm that nurse facilitative effects are fundamental for a secondary successional trajectory of the post-fire dynamics.
  • 4Sensitivities showed that secondary succession may be launched in the pioneer state by increasing the frequency of seedlings of tall shrubs, evergreen and deciduous trees under small shrubs.
  • 5Synthesis and applications. Our Markov chain successional model is an analytical tool which can be applied rapidly and easily to determine successional trajectories for forest restoration. It allows: (i) evaluation of post-fire dynamics and identification of areas in need of intensive intervention (i.e. where secondary succession remains arrested leading to stasis in the pioneer state); (ii) assessment of the role of long-term facilitative nurse effects on restoration; and (iii) identification of species-pair combinations and functional nurse groups of value for further planting efforts.