Experimental biodiversity enrichment in an oil-palm plantation

Abstract

Large-scale conversion of tropical forests has led to dramatic losses of biodiversity and associated ecosystem services and functions. One particularly severe example of forest and biodiversity loss is Indonesia. The country’s exceptionally high biodiversity is seriously threatened by land conversion, for which oil-palm cultivation is a major driver. To increase biodiversity and ecosystem functioning in oil-palm-dominated landscapes, reduced-impact farming systems such as agroforestry systems have been proposed. But in regions where oil-palm plantations already dominate the landscape, this increase can only be achieved through systematic ecological restoration. However, knowledge about the underlying ecological and socio-economic processes, constraints, and trade-offs of ecological restoration in oil-palm landscapes is very limited. To bridge this gap, I established a long-term biodiversity enrichment experiment, together with colleagues from Germany and Indonesia. We planted experimental tree islands in a conventional oil-palm plantation and systematically varied plot size, tree diversity, and tree species composition. We used six multi-purpose tree species that are native to the region. To enhance the light availability for the planted trees, we reduced the oil-palm density within the tree islands by thinning. In this thesis, I present the design of the experiment and give a broad picture of initial effects induced by the experiment, taking into account initial ecological effects and changes in oil-palm yields. The initial phase after planting is most critical to achieve restoration goals in the future, as it constitutes a bottleneck that trees have to surpass for long-term establishment. In addition, from the farmer’s perspective, the initial phase can be a decisive hindrance, given the time delay from the establishment of the tree islands to the emergence of benefits. First, I describe environmental variables and biotic characteristics of the associated vegetation, invertebrates, and birds of the experimental sites prior to the establishment of the experiment, as well as initial experimental effects on the fauna. One year after establishment of the experiment, tree plantings had an overall positive effect on the bird and invertebrate communities at the plantation scale. Diversity and abundance of invertebrates also responded at smaller scales, and were positively affected by the size of the tree islands. Based on these results, I expect a further increase of biodiversity and associated ecological functions in the future. Subsequently, I report on the establishment success of the trees, and identify the most important determinants for tree growth and survival. Most trees established well; however, I found strong differences between the tree species considering relative height increment, relative basal area increment and survival. Most species with high growth rates also showed high survival rates. At the plot level, site conditions, plot size, and diversity level of the planted trees only weakly affected mean tree growth and survival. At the level of individual trees, I found significant neighborhood effects between the species. Tall neighboring trees and distance to living oil palms significantly benefitted tree performance. It is yet too early to decide on the overall suitability of tree species regarding their effectiveness to achieve restoration goals, but differences in initial tree performance are an important information to evaluate the species in the future. Lastly, I present effects of the experiment on oil-palm yields. After two years, I found enhanced yields on the individual level of oil palms both inside and directly adjacent to the experimental plots. Estimating net yield changes for different sizes of tree islands including foregone yield of removed oil palms and spillover effects, I found evidence that – in particular for larger tree islands – yield gains at least compensate for the reduced number of oil palms. These results, obtained during the early phase of tree island establishment, are promising in terms of identifying sustainable management options for oil-palm plantations that reconcile ecological and economic functions. Initial effects were stronger and, particularly from the economic point of view, more beneficial than I had expected. However, so far, I observed mostly weak neighborhood effects and a weak influence of experimentally altered variables, and I expect these effects to become more pronounced over time. Insights gained from future long-term monitoring of the experiment I present in this study may bridge knowledge gaps towards the elaboration of management guidelines for oil-palm landscapes that are both ecologically improved and economically viable. As one part of the puzzle, this thesis represents an important contribution to an overall evaluation of the experiment, and can thus, moreover, add new insights beyond our project that help to advance restoration science.