Across 40,049 species in 24 groups of organisms, smallholder rubber and oil palm farms in Sumatra scored about half as high as nearby forest on a plot-level diversity index, and the species-poor farm plots often held only leftovers of richer ones.
Peer reviewed · Observational field study (32 plots, one Indonesian province) · Smallholder farms only
Turning lowland forest in Sumatra into rubber and oil palm monocultures strips out species above and below ground, and it also changes how the surviving communities differ from place to place, according to a study published in Nature on 18 November 2026. Nathaly Guerrero-Ramírez, Jan Salecker, Holger Kreft and 44 colleagues, led from the University of Göttingen, compared 32 plots in Jambi Province. Monoculture plots did not become uniformly alike. Above ground, poorer plots tended to hold subsets of richer plots’ species. The data come from two areas sampled mostly in 2013, and industrial plantations were not included.
What you need to know
- Finding: Compared with forest, an index of diversity averaged across 24 groups was roughly half as high in a typical monoculture plot and about 78% lower when all monoculture plots were pooled.
- Method: Field teams counted trees, plants, birds, bats, butterflies, canopy and soil animals, and used DNA to identify soil fungi, protists, bacteria and archaea on 32 plots of 50 m × 50 m in forest, jungle rubber, rubber and oil palm.
- Source: Guerrero-Ramírez, Salecker et al., “Tropical land-use intensification restructures biodiversity”, Nature, 18 November 2026 (open access, CC BY 4.0).
- Why it matters: Counting species misses part of the damage: farm plots can look varied while they are really emptied-out versions of one another.
- Limitation: It is a snapshot comparison of different sites in two areas of one province, not a before-and-after record, and it excludes large industrial plantations, where the authors expect stronger effects.
Evidence at a glance
| Item | Assessment |
|---|---|
| Evidence status | Peer reviewed. Observational field study comparing land uses side by side (space-for-time design); no experiment. |
| Primary source | Nathaly R. Guerrero-Ramírez, Jan Salecker and 45 co-authors, “Tropical land-use intensification restructures biodiversity”, Nature, published online 18 November 2026, doi:10.1038/s41586-026-11298-6. Received 27 May 2022, accepted 12 October 2026. |
| Evidence strength | Moderate, because 24 groups were sampled with one standard design and the richness declines are large and consistent, but there are only 8 plots per land use in two areas of one province, and most sampling happened in a single year. |
| Principal limitation | The study covers smallholder farms at a deforestation frontier, with some forest still nearby. Results may not hold for large industrial estates or for older farming regions. |
| Coverage at a glance | 0 independent credible newsroom reports found as of 21 November 2026, 08:15 UTC (SciQuest English-language web searches). The University of Göttingen press release (18 November) is carried by EurekAlert!, Phys.org and Mirage News; copies of the same release were not counted, and an automated rewrite site was excluded. Coverage is not evidence that a finding is correct. |
Sources: the Nature paper (main text, methods, figure legends, funding and competing-interest statements); University of Göttingen press release no. 171; SciQuest coverage search at the time stated.
Is this a big deal?
It is a solid, careful step rather than a surprise. That forests hold more species than plantations was already well documented. The new part is scale and depth: few tropical studies have counted this many kinds of life on the same plots, and fewer have asked how the mix of species shifts from plot to plot. The answer that monocultures do not simply converge on one generic community is a useful correction to a common assumption, but it rests on one region.
What scientists found
The forest plots, which the authors describe as degraded primary forest, were the richest by far. Of all recorded species, 5,861 (15%) turned up only in forest. Above ground the share was higher: 1,510 species, or 30% of everything found above ground, appeared in forest and nowhere else.
Jungle rubber, a traditional system in which rubber trees are planted into secondary forest, kept much of that richness. Pooled across its plots, the diversity index fell by about 20% compared with forest. About 60% of above-ground species (3,057) were found only in forest, jungle rubber or both. In rubber and oil palm monocultures the pooled index dropped from 0.91 to 0.20, about 78% lower.
Plot by plot, the index fell from 0.59 in forest to 0.47 in jungle rubber, 0.26 in rubber and 0.28 in oil palm. Every above-ground group studied had significantly fewer species in oil palm than in forest. Two soil groups went the other way: bacteria gained species in all three farm types and archaea in oil palm, which the authors suggest may come from liming and fertiliser raising soil pH and available nitrogen.
Why “different” can still mean “emptier”
The authors had expected the farms to show biotic homogenization, where communities in different places become more alike as specialists vanish and the same hardy generalists spread everywhere. That pattern was not consistent. Significant homogenization showed up in 9 of the 24 groups in oil palm, but 12 groups became more different between plots instead. Only 10 groups responded the same way in all three farm types: understorey plants, fungi and bacteria homogenized, while canopy bugs, flies, parasitoid wasps, spiders, butterflies, leaf-litter beetles and oribatid mites became more varied.
The authors then split those differences in two. Think of two music playlists. In “turnover”, each playlist has different songs. In “nestedness”, the short playlist is just the long one with tracks deleted. Both pairs look different, but only the first adds variety. Above ground, rubber and oil palm plots shifted from the first kind toward the second. The playlist comparison has a limit: real plots also differ in how many individuals of each species they hold, which the main analysis ignored.
“We found that the sharp decline in the number of species does not automatically lead to increased similarity in community composition in monocultures,” Guerrero-Ramírez said in the university’s press release. The paper warns that this kind of apparent variety may be an early stage of simplification that ends in homogenization later.
How they found it
The work comes from EFForTS, a German–Indonesian research centre funded by the German Research Foundation (DFG) that has studied rainforest conversion in Jambi since the plots were chosen in 2012. The 32 plots lie in two areas, Harapan and Bukit Duabelas, and each area has four plots apiece of forest, jungle rubber, smallholder rubber and smallholder oil palm. At selection the rubber farms were 7 to 16 years old and the oil palm farms 8 to 15.
Different specialists sampled each group. Canopy insects and spiders were knocked down with insecticide fog in the 2013 dry season, yielding 372,146 individual arthropods. Bats were caught in mist nets and harp traps, birds were counted and recorded with automated microphones, and trees and understorey plants were censused. Soil cores supplied mites, springtails and centipedes, and DNA sequencing gave units for fungi, protists, bacteria and archaea. Those microbial units account for about 34,000 of the 40,049 “species”, so the headline count is mostly microbes.
To compare 12 kinds of centipede with 16,413 bacterial units, the team rescaled each group from 0 (lowest value seen) to 1 (highest) and averaged across groups. That is the “multidiversity” index in the graphic. So a fall from 0.59 to 0.26 means the average rescaled score roughly halved, not that half of all species vanished.
Why it matters to you
This won’t change anything in your daily life this year. It matters for how conservation success is measured. A survey that finds farm plots looking quite different from one another could wrongly conclude that the farmland still holds plenty of variety. This study shows that some of that variety is made of leftovers.
The authors suggest two practical directions: protecting the remaining traditional agroforests such as jungle rubber, and adding native trees or gentler management inside existing monocultures. Both are suggestions drawn from this and earlier EFForTS work, not results tested here.
What the evidence does not show
It does not track change over time at any one place. Forest and farm plots are different sites, so the comparison assumes they started out alike. Most groups were sampled in 2013 and 2014; bats were caught in 2015, soil fungi sampled in 2016 and butterflies in 2017.
It also cannot speak for industrial plantations. The study’s oil palm plots had about 13.6% forest within 1 km and received about 6.6 kg of nitrogen per hectare per year, compared with about 0.45% forest and 141 kg of nitrogen reported for industrial oil palm in an earlier study the paper cites. The authors expect stronger homogenization on large estates and as smallholders intensify.
Some of the explanations are interpretations. Statistical models linked differences between plots more strongly to vegetation, leaf litter and microclimate than to distance, which the authors read as habitat filtering (the environment deciding which species can live there). For leaf-litter spiders, mesostigmatid mites and springtails, differences between plots were only weakly tied to the measured conditions, and the causes there remain open.
How this fits what we already knew
Earlier EFForTS papers had found steep losses in individual groups after forest conversion, and multi-group comparisons of 3, 9 and 14 groups on the same plots. A 2016 Nature study of grasslands found that land-use intensification homogenized communities across many groups. The Sumatra results suggest the tropical story is messier, with homogenization in some groups and differentiation in others.
SciQuest has covered other signals of land use leaving a mark on wildlife, from microplastics in tapirs and anteaters living near development to a century of range loss in European bees.
Independent perspective
We did not obtain comment from an independent researcher before publication. Nature published a Research Briefing on the paper with an expert opinion section, but it is behind a paywall and we have not read it. We will add outside comment if we receive it.
How other headlines framed it
We found no independent newsroom coverage at the time of writing. The University of Göttingen release, republished by EurekAlert! and Mirage News, leads with “unexpected effects” and presents the 50% and 78% figures as a drastic decline in species richness. Phys.org ran the same release under a headline stressing that species counts alone miss the change. Nature‘s Research Briefing frames it as reshaping in “expected and unexpected ways”. None of these explain that the percentages describe a rescaled index rather than a share of species lost.
What happens next
The authors call for long-term monitoring, because a shift from differentiation to homogenization would only show up over years. They also point to plantations now entering their second and third planting cycles as moments to add native trees. The analysis data are public on Figshare, so other groups can test the patterns against industrial estates or other regions.
In one sentence
In Sumatra, replacing forest with rubber and oil palm cut diversity across 24 groups of life, and much of the variety left between farm plots turned out to be emptier copies of richer ones.
Quick check: did it click?
- Does “about 50% lower” mean half of all species disappeared from monoculture plots? No. It is a drop in a 0-to-1 index averaged across 24 groups.
- What is nestedness? When a species-poor site holds a subset of a richer site’s species, so the sites differ only because one has lost species.
- Which farm type kept the most forest species? Jungle rubber, the traditional agroforest, whose pooled diversity index was about 20% below forest.
Read next
- Islands hold about a fifth of the world’s bee species: why the places with the most species are not always the places with unique ones.
- How long do spiders live? Spiders were one of the groups that became more varied between farm plots.
Sources and disclosure
- Guerrero-Ramírez, N. R., Salecker, J. et al. Tropical land-use intensification restructures biodiversity. Nature (2026), published 18 November 2026. Analysis data: Figshare.
- University of Göttingen. Tropical plantations: unexpected effects on biodiversity. Press release no. 171, 18 November 2026.
- Kreft, H. and Guerrero-Ramírez, N. Tropical land-use intensification reshapes biodiversity in expected and unexpected ways. Nature Research Briefing (2026); not read in full (paywalled).
- Gossner, M. M. et al. Land-use intensification causes multitrophic homogenization of grassland communities. Nature 540, 266–269 (2016).
Disclosure: This article was drafted with AI assistance and checked against the sources above. The graphic redraws values from the paper’s main text and Fig. 2; its right-hand panel is a simplified illustration, not study data. The microbial share (about 34,000 of 40,049) is a SciQuest sum of the group totals in the paper’s methods. The study was funded by the German Research Foundation (DFG) and others; the authors declare no competing interests. Spotted an error? See our corrections policy.
Discover more from SciQuest
Subscribe to get the latest posts sent to your email.

Leave a comment