Dirk Hölscher

Large-scale conversion of tropical rainforests to oil palm plantations is a major driver of biodiversity loss in Southeast Asia, particularly threatening native forest-associated plant species and their associated biota. Integrating tree islands within monocultural oil palm landscapes has emerged as a promising approach to mitigate biodiversity loss. However, their effectiveness in sustaining natural regeneration and long-term recovery of native forest-associated plant species remains unclear. We analyzed temporal trajectories of diversity, abundance, and composition of naturally recruiting woody species in tree islands embedded in a monoculture-dominated landscape. We used a large-scale biodiversity enrichment experiment with 52 tree islands, varying in area (25 - 1600 m²) and planted tree diversity (0, 1, 2, 3, or 6 species), established in a conventional oil palm plantation in Sumatra, Indonesia, in 2013. Across four censuses conducted between 2 and 8 years after the onset of natural regeneration, we found 82 species of recruiting woody plants (trees and shrubs) belonging to 64 genera and 34 families. At the landscape (gamma) scale, species richness and abundance increased by ∼42% and ∼69%, respectively, from 2018 to 2024, with shifts in community evenness and continued species arrival. After 8 years, neither richness nor abundance saturated, indicating continued accumulation of recruiting plant species. Tree islands promoted natural regeneration of native diversity, with 90% of species and individuals being native. Native forest-associated species (N = 35) increased steadily at the landscape scale, while localized dominance of alien species did not indicate arrested succession. At the local (alpha) scale, larger island area exerted a sustained positive effect on recruiting diversity. Higher initial planted tree diversity promoted native recruitment, particularly of forest-associated species, at earlier stages, an effect that diminished over time. Synthesis and applications. Tree islands promote natural regeneration and accelerate recovery of native forest-associated diversity over extended periods, providing an effective strategy to enhance forest-associated biodiversity within oil palm landscapes. Preserving the last remaining lowland rainforests, irreplaceable reservoirs of native forest-associated species, remains essential to support continued recruitment into tree islands that can create stepping-stone habitats and a more wildlife-friendly agricultural matrix.

The genus Ravenala (Strelitziaceae) comprises six species of arborescent herbs endemic to Madagascar and iconic to this global biodiversity hotspot. We examined how Ravenala density responds to environmental gradients, land‐use type, accessibility and canopy disturbance. The study was conducted in north‐eastern Madagascar, a region heavily affected by deforestation in recent decades. We carried out species‐specific terrestrial sampling in different land‐use types (fallow land, vanilla agroforests and forest fragments) in human‐modified landscapes and in old‐growth forest. Visual detection in high‐resolution satellite imagery extended observations of Ravenala at genus level across human‐modified landscapes. Both, the terrestrial and the satellite‐based approach indicated high but variable Ravenala densities. Ravenala density in human‐modified landscapes increased with distance from settlements. The four encountered Ravenala species differed in response to disturbance and other factors. R. grandis and R. menahirana declined under higher disturbance levels, whereas R. madagascariensis increased in density with greater disturbance and reached highest densities in fallow land. In the old‐growth forest, both R. blancii , which was recorded exclusively in the old‐growth forest, and R. menahirana were less frequent in plots containing thicket‐forming vegetation (bamboo and wild cardamom). Elevation responses diverged, with R. blancii increasing and R. menahirana decreasing in density at higher elevations. Divergent responses of different Ravenala species to habitat and disturbance suggest that reduction of human‐induced canopy disturbance in human‐modified landscapes and protection of old‐growth forests are needed to safeguard all four encountered species.

Biodiversity generally enhances ecosystem productivity, but whether such effects persist or intensify during climate extremes is unclear. Here we synthesized data from 75 biodiversity experiments across grasslands and forests to assess how aridity and soil nutrients modulate plant diversity effects under drought and heat extremes. Biodiversity most strongly enhanced productivity under extreme drought in more arid grasslands but had limited effects in forests under such conditions. More arid grasslands showed enhanced complementarity effects under extreme drought, whereas less arid grasslands favoured selection effects driven by productive species. Heat extremes did not produce comparable context-dependent changes in plant diversity effects across ecosystem types or aridity gradients. Soil nutrients did not have any detectable influence under either drought or heat extremes, suggesting that as climatic stress intensifies, hydric limitations override edaphic constraints. Our synthesis identifies when and where plant diversity most strongly enhances productivity, showing that its effects are the greatest under extreme drought in more arid grasslands.

Tropical rainforests harbor exceptional biodiversity and function, but are increasingly threatened by agricultural expansion. Whether landscape heterogeneity mitigates these impacts, as in temperate systems, remains unclear. Here, we quantified how local land use and landscape heterogeneity shape multidiversity and ecosystem multifunctionality, using 34 biodiversity metrics and 21 functions across 128 plots in Sumatra. Relative to rainforests, plantations reduced multidiversity and multifunctionality by ~25%, with stronger aboveground declines, lower plant and animal but higher microbial diversity. Contrary to temperate systems, landscape heterogeneity did not buffer local land-use effects but exacerbated declines in multidiversity and multifunctionality, and benefits of surrounding-rainforest cover were confined to rainforest fragments rather than plantations. Our results highlight the irreplaceability of continuous tropical rainforests, and the limited transferability of temperate-based landscape conservation strategies.

Pyrogenic carbon capture and storage is increasingly recognized as a scalable carbon dioxide removal strategy. Flame curtain pyrolysis (Kon‐Tiki) enables decentralized biochar production, but its net climate benefit depends on the magnitude of methane emissions caused during production. Here, we compared emissions from flame curtain pyrolysis and open burning of crop residues, which is considered the baseline scenario in many cases. We tested five feedstocks across a range of moisture contents and continuously measured temperature and concentrations of CO ₂ , CO, CH ₄ , and C ₃ H ₈ in the flue gas. Emission factors were calculated via carbon mass balance, and regression analysis was used to identify drivers of the variability of methane concentrations. Kon‐Tiki pyrolysis resulted in lower CO and CH ₄ emissions than open burning, particularly for wheat straw. Dry feedstocks (≤ 15% moisture) yielded low methane emissions (< 5 g kg ⁻¹ biochar), whereas moist feedstocks (≥ 25% moisture) produced up to tenfold higher emissions. Kon‐Tiki methane emissions were best explained by flue gas temperatures, showing an exponential decline of methane emissions with increasing temperatures. Overall, replacing open burning with Kon‐Tiki pyrolysis is a promising strategy to foster biochar production and reduce greenhouse gas emissions. Nevertheless, managing feedstock moisture and maintaining high temperatures are critical to minimize methane emissions and to enable Kon‐Tiki kilns to serve as an effective transitional technology for decentralized biochar production and carbon dioxide removal.

Agroforestry systems help address global challenges by supporting biodiversity, rural livelihoods, and climate change adaptation. While Arabica coffee ( Coffea arabica L.) is traditionally grown in agroforestry systems, full-sun plantations have been promoted too, fueling the debate over optimal shade tree integration for coffee production. The present study assessed coffee yields in relation to canopy cover and carbon stocks at the plot level in the mid-hills of Nepal. We established monitoring plots and quantified coffee yields through cherry counting. The canopy cover was estimated from canopy photos. Shade trees were identified and their above-ground carbon stocks were measured using tree height and diameter at breast height. We found coffee yields ranging from 36 to 1081 kg ha ⁻¹ , with an overall mean of 364 kg ha ⁻¹ . Canopy cover varied from 0 to 90% across all plots. Shade trees stored an average of 43.5 Mg ha ⁻¹ of carbon in their above-ground biomass. The total mean carbon stock—including above-ground, below-ground, and soil organic carbon to 30 cm depth—was 138.8 Mg ha ⁻¹ . Medium canopy cover levels supported the highest yields. In contrast, both low (< 25%) and high (> 70%) canopy covers resulted in significant yield reductions. The agroforestry systems with the highest above-ground carbon stocks were associated with dense canopy covers (> 70%) and reduced coffee yields, highlighting a trade-off between maximizing carbon sequestration and maximizing coffee yields.

Tree diversity often increases stand-level growth, but whether neighbourhood diversity effects on individual tree growth change with climatic conditions remains unclear. Here, using 852,170 records of 113,701 individuals from 129 species in 15 tree diversity experiments across four biomes, we address this knowledge gap with a synthesis of tree growth data spanning a broad climate gradient. We examine how neighbourhood-scale (defined as a focal tree and the adjacent trees) taxonomic and functional diversity effects on tree growth vary with climate, both spatially (across sites) and temporally (within sites). Increasing species richness and trait dissimilarity from monospecific to high-diversity neighbourhoods enhanced individual tree growth by 7–13% on average. The positive diversity effect increased from dry to wet climates, contrasting with most prior studies, but was unaffected by interannual climatic variation within sites. Given that tree–tree interactions are ubiquitous and likely to interact with climate in both young and old forests, our findings suggest incorporating neighbourhood diversity as a management tool to enhance forest productivity, while considering underlying mechanisms and interactions with climate, thereby facilitating targeted and site-specific climate and biodiversity benefits.