The rapid expansion of high-yielding cash crops like oil palm has led to significant deforestation and land-use change in megadiverse tropical lowland rainforests. While oil palm plantations have boosted economic development and improved livelihoods, they've also caused dramatic ecological damage, reducing species diversity and altering ecosystem functioning. The UN Decade on Ecosystem Restoration highlights the urgent need for ecological restoration in agricultural landscapes to balance biodiversity, ecosystem functioning, and local livelihoods. However, trade-offs between these factors can hinder restoration efforts. Agroforestry systems and tree islands offer potential mitigation strategies, but empirical evidence on their impact on biodiversity, ecosystem functioning, and agricultural productivity is crucial for their viability. This study investigates the ecological restoration outcomes of establishing tree islands in an oil palm landscape, aiming to quantify their effects on biodiversity, ecosystem functioning, and agricultural productivity.

Existing literature emphasizes the negative ecological impacts of oil palm expansion, including reduced biodiversity and impaired ecosystem services. Studies highlight the need for ecological restoration in agricultural landscapes, focusing on balancing biodiversity conservation with agricultural productivity. The concept of tree islands or agroforestry systems as restoration strategies has been explored, with some evidence suggesting their potential benefits. However, a lack of large-scale, empirical data on their effectiveness in oil palm landscapes necessitates further research. This study addresses this gap by conducting a large-scale experiment to assess the impacts of tree islands on biodiversity, ecosystem functioning, and oil palm yields.

A large-scale, interdisciplinary ecosystem restoration experiment was conducted in a 140-hectare industrial oil palm plantation in Sumatra, Indonesia. Fifty-two tree islands were established, varying in size (25, 100, 400, and 1600 m²) and planted tree diversity (0-6 species). Control plots represented conventionally managed oil palm monocultures. Over three to five years, ten indicators of above- and below-ground biodiversity (bacteria, fungi, plants, animals) and nineteen indicators of ecosystem functioning (oil palm and tree productivity, invasion resistance, pollination, soil quality, predation/herbivory, carbon/nutrient cycling, water/climate regulation) were assessed. Multidiversity and multifunctionality were calculated to provide an holistic overview. Vegetation structure indicators were also measured to understand the mechanisms driving the observed effects. Structural equation models (SEMs) were used to analyze the effects of tree island area and planted tree diversity on restoration outcomes, mediating through vegetation structural complexity and tree dominance.

Tree islands exhibited significantly higher biodiversity and ecosystem functioning compared to control plots. However, the effects varied across indicators. Natural regeneration increased tree and bird species richness, while decreasing abundant seed species diversity. Ecosystem functioning showed strong increases in water infiltration, litter input, insectivorous bat/bird activity, and soil fertility. Multidiversity and multifunctionality were consistently higher in tree islands. Larger tree islands showed greater restoration benefits for both biodiversity and ecosystem functioning, mediated by increased tree dominance, leading to improved habitat quality and quantity. The effect of planted tree diversity on biodiversity depended on the indicator, with contrasting responses mediated by vegetation structure. Structurally complex habitats benefited some species, while more open habitats favoured others. Oil palm yield showed a 24% decrease at the local scale within tree islands due to reduced palm density, but no significant difference was observed at the landscape scale due to compensatory yield gains from surrounding palms.

The findings directly address the research question of the effectiveness of tree islands in restoring biodiversity and ecosystem functioning in oil palm landscapes. The significant improvements in biodiversity and ecosystem functioning, without compromising landscape-scale oil palm yield, demonstrate the viability of this restoration strategy. The mechanism-based insights from the SEMs highlight the importance of both island size and planted diversity in creating structurally complex habitats that support diverse communities. The study contributes to a growing body of evidence supporting the ecological and economic benefits of agroforestry systems in oil palm landscapes. The scale of the experiment provides robust evidence of the potential for wide-scale implementation.

This large-scale study provides strong evidence that integrating tree islands into oil palm landscapes significantly enhances biodiversity and ecosystem functioning without negatively affecting overall agricultural productivity. Larger islands and diverse plantings are particularly beneficial. Future research should focus on understanding the long-term impacts of tree islands, optimizing island design for different landscapes, and exploring the socioeconomic factors influencing adoption by stakeholders, particularly smallholders.

The study was conducted in a single landscape in Sumatra, Indonesia. While the large scale of the experiment increases robustness, the findings might not be generalizable to all oil palm landscapes. The relatively short duration of the experiment (3-5 years) may not fully capture long-term ecological changes. Further research across diverse oil palm landscapes is needed to confirm the generalizability of the findings.