Read the full paper (open access) here: doi.org/10.1111/geb.70141
As for my fellow statistics/R/reproducible code lovers, you'll find the detailed commented R code (and data) for all analyses and figures here 🙃: zenodo.org/records/1704...
🧵9/9

We find that 🌳 demography is not uniform, even within a species’ range.
Accounting for this mosaic of mortality can improve management & conservation of forests facing intensifying stressors, and stresses the crucial role of permanent plots across environmental gradients. @umramap.bsky.social
🧵8/9

🌎 Take-home message:
To predict forest futures, we need to go beyond climate envelopes — including disturbances, pests, and ontogeny — if we want realistic models of mortality and carbon cycling.
@ird-fr.bsky.social
🧵7/9

In contrast, in low-mortality years, competition was the main mortality agent, suggesting a prevalence of gap dynamics under low mortality years and regions.
High-mortality patches, though, were dominated by disturbance agents like storms and insects.
🧵6/9

🔥 Pinus elliottii (slash pine) and P. palustris (longleaf pine) showed more localised spatial structures.
Extreme weather events (storms, mostly) and fire dominated their high-mortality years (especially 2013–2023).
🧵5/9

🌳 Pinus taeda (loblolly pine) stood out:
Its survival has clearly declined over the past two decades across its natural range.
This pattern was spatially consistent and mostly linked to weather, insects, and competition.
🧵4/9

We found that mortality patterns form a mosaic across space and time — and change in nonlinear ways with tree size (ontogeny).
👉 Most interspecific differences appear in saplings (2.5 to 10 cm DBH) and near maximum tree size.
🧵3/9

Tree mortality isn’t random — it depends on who you are, where you grow, and what happens to you.
We analysed demographic data from ~130,000 trees across 14,500 plots (years 2003–2023), looking at mortality agents like competition, storms, fire, insects, and disease.
🧵2/9

This follows our earlier Ecology Letters study showing that dead trees can influence neighbours even more than living ones.

Forests have long memories — the "ghosts" of trees past still shape who grows, survives, and coexists. 🌳🌱

🔗 doi.org/10.1111/ele....
[3/3]

We’ve long known that young trees struggle near adults of the same species — this mechanism helps maintain diversity.
But what if that effect doesn’t end when the adult tree dies?

Turns out this "legacy effect" can last at least 5 years — even after the tree is dead 🌲
[2/3]

With #COP30 next month in Belém 🇧🇷, we hope this will be a wake-up call: tropical forests are under growing pressure from climate change.
Safeguarding them demands ambitious and fair climate action, including stronger support for countries that protect these vital ecosystems for us all. 🌳🌡️📈
[5/5]

Of course, this doesn’t mean these forests have lost all climate value ➡️ they remain immense carbon stores and irreplaceable biodiversity havens.
But it shows that keeping global warming well below 1.5 °C is now even harder—and more urgent. 🌏
[4/5]

Context: in 2022, our @nature.com study found a long-term increase in tree mortality in these forests, likely due to rising atmospheric water stress.
🔗 nature.com/articles/s41586-022-04737-7
This new paper follows up: aboveground biomass is now a net carbon emitter — driven by #ClimateChange.
[3/5]

Based on 50 yrs of data from ~11 000 trees, we found that the aboveground woody biomass—once absorbing C—now now releases nearly 1 tonne of C per hectare each year.
The cause? Human-driven climate change: hotter extremes, droughts & cyclones are killing more trees than the forest regrow. 🌡️🌪️
[2/5]

📄 Link to study: www.nature.com/articles/s41...
With @josbarlow.bsky.social , @fplmelo.bsky.social , @erikaberenguer.bsky.social , @ird-fr.bsky.social , Universidade Federal de Pernambuco, @ecioxford.bsky.social , Lancaster University, Universidade Estadual de Santa Cruz, and several others.
[5/5]

These findings show how #causality can be inferred from observational data in #Ecology. Habitat loss, fragmentation, edge effects, degradation, and the functional make-up of tree communities have interdependencies, and require a #CausalModel to disentangle #causation from spurious association!
[4/5]