Evaluating potential restoration strategies for vascular epiphytes with niche‐based community mechanistic models

Ecological restoration is a global priority, with intergovernmental commitments such as the Kunming‐Montreal Global Biodiversity Framework and the Bonn Challenge seeking to restore the world's forests and their biodiversity. Central to these aims is the restoration of tropical forests, which harbour over half of all terrestrial species. However, one highly diverse and threatened group yet rarely considered in restoration is vascular holoepiphytes: vascular plants germinating and growing on other plants. Despite making up 9% of vascular plant species globally and in places over a third of species locally, they are often overlooked in restoration, omitting a substantial amount of diversity.

We address hypotheses regarding timing, population size and age structure of restored epiphyte populations using mechanistic models simulating virtual epiphyte communities within three‐dimensional virtual forests.

In epiphyte communities 19 years after introduction, we demonstrate that more species are retained when introducing (a) adult plants into (b) older forest stands. However, different functional groups benefit from nuanced restoration strategies, with higher proportions of faster growing, light‐when‐mature epiphytes in communities introduced into 5‐ to 10‐year‐old forests and at high initial population sizes of 100 individuals per quarter hectare, while short‐dispersing, slow‐growing, heavy‐when‐mature epiphytes are proportionally more abundant with restoration into 20‐ to 40‐year‐old forests.

Synthesis and applications . Generally, delayed restoration of adult‐only epiphyte populations into older restored forests improves conservation outcomes. However, restoration can be optimised, initially introducing large populations of fast‐growing, light‐when‐mature species in younger forests followed by the later introduction of slow‐growing, heavy‐when‐mature species in forests older than 20 years. As epiphytes have a slow recolonisation rate, our results identify optimal strategies for restoring a variety of functional groups of a diverse yet difficult‐to‐reestablish ecological guild into regenerating forests.