Species diversity reduces risk in tropical forest restoration: a portfolio effect across heterogeneous sites

Rachele Quaglino, Rebecca J Cole, Gerald Quiros Cedeño, David A. Rodríguez, Jeff Tingle, Daniel Maynard, Thomas W. Crowther, Leland K Werden , Anton Weissenhofer

Assisted restoration is essential for recovering degraded tropical forests. Biodiversity-stability theory predicts that functionally diverse communities should produce more reliable, predictable outcomes than monocultures, but broad-scale field experiments in tropical restoration are scarce. We established 120 plots spanning five planting treatments (two monocultures – of a fast-growing nitrogen-fixer or a fleshy-fruit species – and 2-, 6-, and 12-species mixes), with species selected at each richness level to maximize functional trait diversity, across four sites in southwestern Costa Rica. After three years we measured height, canopy cover, carbon stocks, survival, and structural heterogeneity. Inga edulis monocultures (the same nitrogen-fixer at all four sites) achieved the highest mean carbon stocks (5.4 Mg C ha-1) and fastest height growth (3.7 m vs 1.7-2.2 m in mixtures) but survived less and far less predictably across sites (78%, 56-100%, versus 90%, 83-95% in mixtures). Which species survived consistently therefore mattered more than which grew fastest. Each site's mixtures comprised different locally adapted species yet converged on consistent outcomes, so predictability appeared to be a property of diverse mixtures generally, not of any single shared species. Diverse mixtures were three- to four-fold less variable in survival across sites and developed up to 140% more within-plot structural heterogeneity than monocultures, suggesting a dual benefit of more structurally complex habitat within plots and more predictable establishment across them. Because functional diversity saturated at roughly six species, six-species mixes captured most of this benefit, with only marginal gains at twelve. Our results provide evidence for a biodiversity portfolio effect (i.e., biodiversity insurance) in assisted tropical restoration, whereby functionally diverse plantings trade peak productivity for predictability, buffering against the unpredictable, low-survival outcomes that drive replanting costs, at little added expense.