Scientists studying rare rocks buried deep beneath central Australia have uncovered how one of the world's most promising new sources of niobium came to be. Niobium is a critical metal used to strengthen steel and support clean energy technologies, and its origins are tied to dramatic geological events that unfolded more than 800 million years ago. The research, led by Curtin University, shows that these newly identified niobium-rich rocks formed during the early stages of a massive continental breakup. These pathways opened during a period of tectonic stretching and rifting that eventually split Rodinia apart. Their results show the carbonatites were emplaced between 830 and 820 million years ago, during a key phase of continental rifting before Rodinia fully broke apart. "Using multiple isotope-dating techniques on drill core samples, we found that these carbonatites were emplaced between 830 and 820 million years ago, during a period of continental rifting that preceded the breakup of Rodinia. "This tectonic setting allowed carbonatite magma to rise through fault zones that had remained open and active for hundreds of millions of years, delivering metal-rich melts from deep in the mantle up into the crust." Curtin co-author Professor Chris Kirkland, also from the Timescales of Mineral Systems Group, said the study highlights how advanced analytical techniques can untangle extremely complex geological timelines. But determining when and how they formed has historically been difficult due to their complex geological histories," Professor Kirkland said. "By analyzing isotopes and using high-resolution imaging, we were able to reconstruct more than 500 million years of geological events that these rocks experienced. "This approach allowed us to pinpoint when the carbonatites formed and separate those original magmatic events from changes that happened later in the rocks." New Evidence Shows Humans, Not Glaciers, Moved the Stones An Ocean That No Longer Exists Could Explain the Origins of Central Asia's Mountains Stay informed with ScienceDaily's free email newsletter, updated daily and weekly. Keep up to date with the latest news from ScienceDaily via social networks: Tell us what you think of ScienceDaily -- we welcome both positive and negative comments.

This is the most complete skeleton yet of our ancestor Homo habilis A partial skeleton dating back more than two million years is the most complete yet of Homo habilis, one of the earliest known species in our genus Adapted from "New partial skeleton of Homo habilis from the upper Burgi Member, Koobi Fora Formation, Ileret, Kenya," by Grine, et al., in The Anatomical Record; January 13, 2026 Its large brain and flat face—attributes found in today's humans—have long set the species apart from earlier hominins such as Australopithecus africanus. A new study analyzing the uniquely complete skeleton, however, suggests H. habilis' body looked much less modern. Additionally, H. habilis was small—perhaps even smaller than Lucy, a 3.2-million-year-old hominin specimen known for her tiny size. “A finding like this does give hope,” says William Harcourt-Smith, a paleoanthropologist at the American Museum of Natural History, who was not involved in the study. “It's been tough with Homo habilis, as there are very limited, scrappy remains. It shows us that hard work in the field, and constantly looking for them, reaps important dividends.” If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today. Over the next couple of years, the study authors followed a trail of bone fragments downslope for several meters, discovering more teeth and a series of larger bones from the upper body. “Fortunately for us, the teeth are one of the most diagnostic parts of the skeleton for identifying hominin species,” says study co-author Carrie Mongle, an assistant professor of anthropology at Stony Brook University. Still elusive is the build of H. habilis' lower body. Because H. habilis is one of the earliest members of our genus, understanding more about this species helps to shed light on the evolution that led to our own, researchers say. “This study underlines how crucial individual fossil discoveries can be,” says Rebecca Wragg Sykes, an honorary archeology researcher at the University of Cambridge and the University of Liverpool in England, who was not involved in the study. K. R. Callaway is a freelance journalist specializing in science, health, history and policy. If you enjoyed this article, I'd like to ask for your support. Scientific American has served as an advocate for science and industry for 180 years, and right now may be the most critical moment in that two-century history. If you subscribe to Scientific American, you help ensure that our coverage is centered on meaningful research and discovery; that we have the resources to report on the decisions that threaten labs across the U.S.; and that we support both budding and working scientists at a time when the value of science itself too often goes unrecognized. In return, you get essential news, captivating podcasts, brilliant infographics, can't-miss newsletters, must-watch videos, challenging games, and the science world's best writing and reporting. There has never been a more important time for us to stand up and show why science matters.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript. (2026)Cite this article The Strychnos alkaloids have long been regarded as landmark targets for chemical synthesis due to their captivating architectures and notorious biological properties. However, the design of approaches that access multiple family members in an asymmetric, concise and atom-economical fashion remains an important challenge. Here we show that thiophene S,S-dioxides (TDOs) offer a modular, rapid entry to Strychnos natural products via inverse electron demand Diels–Alder cascades. We demonstrate that exceptional levels of stereocontrol can be achieved in asymmetric TDO cycloadditions, affording tricyclic indolines of utility in medicinal chemistry research and enabling the stereoselective synthesis of eight Strychnos alkaloids by the shortest routes described so far, including a synthesis of the iconic family member brucine. Using a machine-learning approach, computational studies provide insight into the source of stereoinduction and reveal an intriguing and unexpected spontaneous cheletropic extrusion of SO2. The Strychnos indole alkaloid natural products have long excited the scientific community, not least because of their potent biological properties and architectural complexity (Fig. Robinson infamously said of the flagship member of the family, strychnine, that “for its molecular size, it is the most complex substance known”1, and the landmark synthesis of this natural product by Woodward et al.2,3 inspired the design of many synthetic approaches to these alkaloids4,5,6. In addition to the structural challenges, such endeavours have also been stimulated by their potential biological applications, such as the identification of alstolucines B and F as candidates for the re-sensitization of taxanes in multidrug-resistant cancers7. However, synthetic approaches that are asymmetric, sufficiently concise to produce useful quantities of natural product and also sufficiently flexible to enable the ‘collective' synthesis of multiple family members are rare. Additional demands of synthesis ideality, such as the avoidance of protecting groups and the minimization of waste8,9, impose further challenges that continue to limit the wider exploitation of the family. In this context, the six-step racemic synthesis of strychnine by the Vanderwal group10 and the collective asymmetric synthesis of six indole alkaloids by MacMillan et al. (including a twelve-step asymmetric approach to strychnine)11 represent pioneering contributions, with recent elegant approaches having also been described by the Zhang and Snaddon groups12,13. a, Examples of Strychnos alkaloids include strychnine, akuammicine, lagumicine and alstolucine B. b, Previous inverse electron demand cycloaddition approaches to indole alkaloids include the use of Zincke aldehydes10,14 and oxadiazoles15. c, This work: the three modular strategies that were explored to access tetracycle 4, a key precursor for the collective synthesis of eight Strychnos alkaloids, via asymmetric cycloaddition cascades of TDOs. X, substituent; TBS, SiMe t2 Bu; R, substituent; Hal, halide; R*, chiral substituent; P, protecting group; DA, Diels–Alder. We questioned whether an entry to the Strychnos family might be designed to meet the stringent demands of contemporary synthetic chemistry, specifically a modular, asymmetric strategy that would enable the construction of the Strychnos core in a handful of steps, and thereby access multiple natural products in a scalable, straightforward manner. We recognized that the polycyclic indoline core of the alkaloids is an attractive motif for ring formation via cycloaddition, in which an indole precursor would serve as an electron-rich 2π component. This concept was exploited by Vanderwal and co-workers using tethered Zincke aldehydes (Fig. 1b)10,14, and also by the Boger and Padwa groups via [3 + 2] cycloadditions with tethered carbonyl ylides15,16. However, these elegant strategies suffer from drawbacks, including the need for harsh reaction conditions when employing aromatic cycloaddition partners such as oxadiazoles (for example, heating at 180 °C for 48 h (ref. 15)), the use of hazardous diazo compounds for carbonyl ylide synthesis16 or the formation of racemic products10 that necessitate late-stage chromatographic resolution15. We hypothesized that thiophene S,S-dioxides (TDOs, Fig. 1c)17 could address these limitations. These heterocycles are in many ways ‘ideal' diene substrates for inverse electron demand Diels–Alder (IEDDA) reactions as they are non-aromatic and should therefore exhibit enhanced reactivity compared with traditional substrates such as pyrones or tetrazines, while still having the benefits of electron deficiency, an intrinsically s-cis constrained diene and a powerful entropic driving force that renders the cycloaddition irreversible (that is, the loss of SO2 through cheletropic extrusion). However, the deployment of TDOs in target-oriented synthesis is surprisingly rare, with applications so far being restricted to aromatic carbocycles such as the relatively simple illudalane triterpenes, the antiplatelet agent beraprost and the indoline core of the dictyodendrins18,19,20. We questioned whether these underexploited dienes could also be used to assemble the saturated carbocycles found in the Strychnos core, and whether we might simultaneously develop asymmetric TDO cycloadditions, which would render the collective synthesis enantioselective. Three complementary TDO-based approaches were envisaged to access the Strychnos core, which vary strategically in the order of synthetic events (Fig. First (Path I), tryptamine could undergo alkylation with side chain 1 suitable for downstream completion of the targeted natural products, followed by addition–elimination (substitution) reaction with a halogenated, chiral TDO 2a to give intermediate 3. The TDO in 3 would react with the tethered indole via a stereoselective intramolecular IEDDA–SO2 cheletropic extrusion pathway to form tetracycle 4 after partial reduction of the intermediate dienamine cycloadduct 5. Further functionalization of 4 would afford the Strychnos natural products. As an alternative, we considered intermolecular approaches such as Path II, in which temporary protection of the tryptamine side-chain nitrogen atom precedes intermolecular cycloaddition with TDO 2a to give diene 6; cleavage of the nitrogen atom protecting group should then lead to intermediate 5 by intramolecular substitution of the TDO-derived halogen atom, thereby converging on Path I. Finally (Path III), we considered the direct cycloaddition of N-protected tryptamine with TDO 2b to form adduct 7. TDO 2b would necessarily feature a disposable substituent (X), which is essential to prevent heterodimerization of the thiophene ring system in the course of its oxidation17. Subsequent to the cycloaddition cascade, deprotection and cyclization of the tryptamine side chain in 7, cleavage of the superfluous substituent X and N-alkylation with side chain 1 would also afford advanced intermediate 4. Here we describe the realization of all three strategies, which culminated in a collective asymmetric synthesis of eight natural products of the Strychnos family. In doing so, we also established asymmetric intermolecular cycloadditions of TDOs, accessing tricyclic indoline scaffolds that are of particular interest for medicinal chemistry applications. Computational studies revealed the source of the stereoselectivity observed in both the intra- and intermolecular IEDDA reactions, and suggested that the intramolecular TDO cycloadditions may benefit from an entropic driving force, where cheletropic extrusion of SO2 occurs in a spontaneous manner along the cycloaddition pathway without the formation of a discrete intermediate [4 + 2] cycloadduct. With a view to developing an asymmetric route to the Strychnos family, we first addressed the development of hitherto unrealized asymmetric TDO cycloadditions. Investigations began with intermolecular [4 + 2] cycloadditions between indoles and enantioenriched thiophene S,S-dioxides (prepared by peroxyacid oxidation of the corresponding thiophene21) equipped with inexpensive, readily available chiral substituents. We found that the reaction of indole (8a, Table 1, R1, R2, R3 = H) with a TDO substituted with an acyl camphorsultam group22 (9a, R4 = 5-Cl) in CH2Cl2 at 0 °C afforded tricycle 10a as a single diastereomer in high yield. The stereochemistry of 10a was established by single-crystal X-ray diffraction analysis, which revealed the indoline ring junction to possess the absolute stereochemistry required for the natural stereoisomers of the Strychnos alkaloids. This method was tested on a range of substituted indoles and showed broad functional group tolerance of both electron-rich (for example, methoxy) and electron-poor substituents (for example, ester, nitro and nitrile) at various positions around the indole ring, as well as useful functionalities such as boronic esters and halides and substitution of the nitrogen atom. All tricyclic products were isolated as single diastereomers in good-to-excellent yields (10b–10m, 68–93%). Variation of the thiophene S,S-dioxide was also well tolerated, including a bicyclic TDO (10n–10p, 54–68%). With a view towards the proposed intermolecular cycloaddition approach to the Strychnos alkaloids, we also evaluated the use of 3-substituted indoles in this chemistry. These reactions were also successful, albeit requiring higher temperatures to effect the cycloaddition due to the steric demand imposed by the additional indole substituent, affording cycloadducts 10q and 10r in 90% and 91% yields, respectively, again as single diastereomers. Having established highly selective, asymmetric intermolecular TDO cycloaddition reactions, we turned to apply them to the synthesis of the Strychnos alkaloids, exploring both the intra- and intermolecular routes. We first tested an intramolecular (tethered) cascade towards akuammicine (Fig. A key issue was whether the excellent stereocontrol imparted by the camphorsultam group in the intermolecular cycloadditions would be maintained in an intramolecular setting. In the event, the reaction of TDO 9a with amine 11 (prepared by the N alkylation of tryptamine with known side chain 1223) at room temperature, followed by warming to 80 °C, afforded a single diastereomer of the IEDDA cycloadduct 13. This intermediate dienamine was reduced in situ using acetic acid and sodium cyanoborohydride24,25,26 to afford diastereomers 14 and 14′ in a 1:1.4 ratio and 68% overall yield from 9a (see Supplementary Table 1 for details). The adverse selectivity of this step likely derives from the preferred delivery of hydride to the less-hindered concave face of the intermediate iminium ion24,25,26. To complete the synthesis of akuammicine, the camphorsultam amide was subjected to methanolysis27 to give the corresponding ester 15 (67%), with the camphorsultam moiety recovered in high yield. Heck cyclization28,29 of 15 proceeded to afford (–)-akuammicine in 75% yield in four steps from tryptamine (six steps in the longest linear sequence (LLS) that includes assembly of the side chain 12). This represents the most concise and atom-economical assembly of akuammicine described so far. An intermolecular cycloaddition approach to 13 (corresponding to Path II in Fig. 1c) could also be achieved by tert-butoxycarbonyl (Boc) protection of the side-chain amine in 11 (95%). Reaction of the resulting carbamate-protected indole with TDO 9a gave an intermediate cycloadduct that upon treatment with trifluoroacetic acid (TFA), followed by basic work-up to trigger cyclization of the free amine onto the dienoyl chloride, also afforded dienamine 13 in 75% yield (see Supplementary Section 2.2 for details). a, Syntheses of akuammicine, norfluorocurarine, lagumicine, alstolucines B and F, and echitamidine. b, Synthesis of strychnine. c, Synthesis of brucine. Reagents and conditions: (i) MeCN, room temperature (r.t.); (ii) MeCN, r.t. to 80 °C (70 °C for 28); then AcOH, NaBH3CN, 65 °C (55 °C for 28); (iii) NaOMe, dimethyl carbonate, CH2Cl2, 0 °C; (iv) Pd(OAc)2, PPh3, Et3N, 70 °C; (v) Boc2O, Et3N, CH2Cl2; (vi) CHCl3, 80 °C; TFA, r.t.; (vii) trifluoroacetic anhydride, H2O2, 0 °C to r.t.; (viii) H2 (1 atm), Pd/C, MeOH–EtOAc, r.t.; (ix) Na2CO3, MeCN, r.t.; (x) K2OsO4·2H2O, NMO·H2O, tBuOH–H2O, r.t.; then Dess–Martin periodinane, tBuOH, CH2Cl2, r.t.; (xi) SmI2, THF–MeOH, 0 °C to r.t.; (xii) NaBH4, MeOH, 0 °C to r.t.; (xiii) DIBALH, CH2Cl2, −78 °C; (xiv) p-anisaldehyde, NaBH(OAc)3, AcOH, 1,2-dichloroethane, 0 °C to r.t.; (xv) DIBALH, CH2Cl2, −78 to 0 °C; (xvi) Pd(OAc)2, Bu4NCl, NaHCO3, EtOAc, r.t.; (xvii) PhSH, TFA, 45 °C (50 °C for brucine); then NaOAc, Ac2O, AcOH, malonic acid, 120 °C; (xviii) Pd(OAc)2, LiCl, K2CO3, N,N-dimethylformamide, 105 °C; then (±)-camphorsulfonic acid, iPrOH, 50 °C. The poor selectivity observed in the reduction of the dienamine cycloadduct 13 prompted us to test the alternative intermolecular asymmetric cycloaddition strategy (Path III in Fig. We identified 4-bromo-TDO 9d (as used in the synthesis of 10p, Table 1) as a potential candidate for this route as its bromine substituent would protect the corresponding thiophene from unwanted dimerization during oxidation17 and would also likely be removed under mild conditions later in the synthesis. We found that intermolecular cycloaddition of N-Boc-tryptamine 16 with 9d (used directly from the oxidation of thiophene 17) afforded a single diastereomer of dihydrocarbazole 18. In situ addition of TFA effected Boc deprotection, and subsequent addition of the tryptamine side-chain amine onto the dienoyl motif upon basic work-up gave a single diastereomer of the pyrrolidine product 19 in 57% yield (from 17). We found that the allylic bromide resident in 19 could then be selectively cleaved by hydrogenolysis without reduction of the adjacent alkene; attachment of the vinyl iodide side chain 12 to the debrominated product 20 smoothly afforded 14, thus converging on the intramolecular cycloaddition route, but now with complete control of the stereochemistry at the three contiguous stereocentres within the tetracycle. Including the previously implemented endgame, this alternative synthesis of (–)-akuammicine proceeded in seven steps in the LLS (20% overall yield). Four additional natural products (lagumicine, alstolucines B and F, and echitamidine) were synthesized from akuammicine using established chemistry30, while (–)-norfluorocurarine was prepared from 14 by initial reduction to aldehyde 21 with diisobutylaluminium hydride (DIBALH) (69%), followed by Heck cyclization (74%). Pyrrolidine 20, prepared via the intermolecular cycloaddition route, also provided access to strychnine (Fig. The alkylation of 20 with side chain 22 (69% yield from 19), followed by para-methoxybenzyl (PMB) protection11 of the indoline nitrogen atom gave tetracycle 23 (81%). From here, exhaustive reduction with DIBALH afforded diol 24 (76%), Heck cyclization–lactol formation of which proceeded smoothly to afford the PMB-protected Wieland–Gumlich aldehyde (57% yield). This was converted into (–)-strychnine via deprotection of the PMB group, followed by treatment with malonic acid, acetic anhydride and sodium acetate (50% over two steps)11. Overall, the synthesis of (–)-strychnine was completed in ten steps via this intermolecular cycloaddition route (LLS, including the synthesis of side chain 22), which represents the most concise asymmetric approach reported so far (see Supplementary Information for details of the intramolecular TDO cycloaddition approach, which proceeded in seven steps LLS from tryptamine, but with inferior stereocontrol). Despite the rich history of the Strychnos alkaloids, there is one family member that has thus far eluded chemical synthesis: brucine. This dimethylcatechol analogue of strychnine has been known for over 200 years, and has been employed by chemists as a chiral resolving agent since Fischer's seminal report in 189931; its catechol derivative has been used as a chiral ligand in asymmetric catalysis32. The challenge for any synthesis of brucine relates to the highly electron-rich nature of the indoline ring, which confers sensitivity towards oxidation and electrophilic degradation. To implement the TDO cascade approach to brucine, we first required N-PMB-5,6-dimethoxytryptamine 25 (Fig. Due to the aforementioned instability issues, precedents for the synthesis of highly oxidized tryptamines are not well documented, but after extensive investigation we found that 25 could be accessed by Larock indole synthesis33,34 from iodoaniline 26 and alkyne 27, which afforded 25 in good yield (61% over two steps). The successful progression of 25 towards brucine proved possible only via the intramolecular TDO cascade: alkylation of 25 with side chain 22 proceeded uneventfully (92%), but due to the reduced stability of the 5,6-dimethoxyindole, the cycloaddition with 9a benefited from a lower reaction temperature (70 °C) and a longer reaction time, affording 28 after reduction of the intermediate dienamine (28% yield from 25, 1.3:1 diastereomeric ratio (d.r.)). A similar endgame sequence to that employed in the synthesis of strychnine then accomplished the total synthesis of (–)-brucine in nine steps LLS from commercial 2-iodo-4,5-dimethoxyaniline. Attempts to implement the alternative intermolecular TDO cascade proved unsuccessful, with complex mixtures observed under the cycloaddition conditions, presumably reflecting the challenge of deploying a highly electron-rich indole. To better understand the basis of the remarkable levels of asymmetric induction imparted by the TDO camphorsultam side chain in the inter- and intramolecular cycloadditions, we performed quantum mechanics (QM) calculations and molecular dynamics (MD) simulations driven by MACE machine learning interatomic potentials (MLIPs) to model the cycloaddition step35. MLIPs ‘learn' the high-dimensional potential energy surfaces (PES) from QM calculations, effectively mapping atomic positions to energies and, often, forces with an accuracy comparable to QM and efficiency comparable to simple empirical force fields36,37. We first explored the facial selectivity of the intermolecular cycloaddition reaction of TDO 9a with indole at the CPCM(MeCN)-DLPNO-CCSD(T)/def2-TZVP//CPCM(MeCN)-B2PLYP-D3BJ/def2-SVP level of theory (353 K/1 M, Fig. Interestingly, we found that the preferred pathway involves a stepwise cycloaddition via TS1, proceeding through a series of shallow energy minima before yielding the experimentally observed product diastereomer 10a (Fig. In contrast, the alternative diastereomer was found to form via a concerted cycloaddition pathway, in which the transition state (TS2) is 3.3 kcal mol−1 higher in energy (Fig. The favoured pathway via TS1 initially leads to Inter1 (Gibbs free energy of reaction ΔG° = +10 kcal mol−1), which undergoes a low-barrier cyclization (TS3, activation energy ΔG‡ = 14.4 kcal mol−1) to form the [4 + 2] cycloadduct Inter3. Exergonic extrusion of SO2 via TS4 (ΔG‡ = 12.6 kcal mol−1) affords the observed product 10a. a,b, Intermolecular IEDDA cycloaddition of TDO 9a with indole (a) and intramolecular IEDDA cycloaddition of N-methyltryptamine-derived TDO 29. The free-energy profiles were computed at the CPCM(MeCN)-DLPNO-CCSD(T)/def2-TZVP//CPCM(MeCN)-B2PLYP-D3BJ/def2-SVP level of theory (353 K/1 M). c, MLIP-MD simulations revealed a concerted asynchronous [4 + 2] process involving a short-lived intermediate (Inter′) and a dynamic stepwise extrusion of SO2. \({\bar{{{r}}}}_{\mathrm{CS}}\), mean of the two C–S bonds. d, Time evolution of the C–C and C–S distances after Inter4. Median (Mdn) values of the C–C and C–S bond lengths (\({r}_{1}^{\mathrm{CC}}\), \({r}_{2}^{\mathrm{CC}}\), \({r}_{1}^{\mathrm{CS}}\) and \({r}_{2}^{\mathrm{CS}}\)) are shown as solid lines, with the shaded regions representing the interquartile range (IQR; that is, the first to the third quartile) across trajectories. Inset: distributions of time gaps between the formation of \({r}_{1}^{\mathrm{CS}}\) and \({r}_{2}^{\mathrm{CS}}\), and between \({r}_{2}^{\mathrm{CC}}\) and \({\bar{r}}_{\mathrm{CS}}\). The dashed lines indicate mean values of 32 and 69 fs, respectively. MLIP-MD was trained using the MACE architecture35, with the CPCM(MeCN)-B2PLYP-D3BJ/def2-SVP level of theory used as the ground-truth method (Supplementary Section 4.1.2). We were curious to understand the energy difference between TS1 and TS2 (2.7 kcal mol−1) as both exhibit stabilizing hydrogen bonding and no obvious steric hindrance. Distortion interaction analysis38 indicated that the higher energy of TS2 compared with TS1 arises from the greater distortion of the TDO 9a (3.2 kcal mol−1 higher in TS2, see Supplementary Section 4.6, and Supplementary Fig. Although this increased distortion is partially compensated by a stronger interaction energy in TS2, the net result is that the electronic activation energy of TS2 remains 2.7 kcal mol−1 higher than TS1. The intramolecular cycloaddition cascade of the tryptamine–TDO adduct 29, which features a simplified N-methyltryptamine side chain, was found to proceed via a shallower stepwise pathway (Fig. This held true regardless of which face of the TDO the tethered indole occupies, with an intramolecular hydrogen bond forming between the indole N–H and the sultam SO2 group seeming to favour TS5, which leads to the experimentally observed diastereomer, over TS7 (difference in activation energies ΔΔG‡ = 8.3 kcal mol−1). The PES of the resulting intermediate Inter4 is shallow, such that the second C–C bond formation occurs via a low-energy transition state (TS6, ΔG‡ = 1.7 kcal mol−1). Beyond TS6, an unexpected, spontaneous extrusion of SO2 leads directly to the dienamine product Inter5 (Fig. Attempts to locate the expected SO2-bridged [4 + 2] cycloadduct (analogous to Inter3) were unsuccessful. Intrigued by this finding, we explored the dynamics of the [4 + 2] mechanism through downhill MD simulations using an MLIP (Fig. 3c and Supplementary Section 4.5 for further details). For this study, 500 trajectories were initialized from TS5 at 353 K and propagated downhill for 5 ps in the direction of either the reactant or product states. Half of the trajectories reverted to the reactant state, while the remainder progressed towards the product state (Supplementary Fig. Among the latter, 159 trajectories successfully reached the product, while the remaining 91 were trapped at Inter4. These trajectories revealed a highly asynchronous dynamic stepwise process for the cycloaddition reaction, with one C–C bond substantially advanced in the TS5 region (\({r}_{1}^{\mathrm{CC}}\) ≈ 1.84 Å) compared with the other C–C bond (\({r}_{2}^{\mathrm{CC}}\) ≈ 3.0 Å). At TS6, the second C–C bond formation has progressed (\({r}_{2}^{\mathrm{CC}}\) ≈ 2.44 Å) and a transient sulfur-bridged [4 + 2] cycloadduct (Inter′) is observed in which the C–S bonds remain intact. However, as the system progresses towards the product state, the C–S bonds elongate and SO2 is extruded. The asynchronicity of this process is evidenced in the evolution of the relevant C–C and C–S distances with time (Fig. The average time gap between the formation of the two C–C bonds (defined as bond lengths <1.6 Å) is 748 ± 491 fs, ranging from 109–2,459 fs. At around 700 fs, as the second C–C bond starts to form (\({r}_{2}^{\mathrm{CC}}\) shortening to 1.6 Å), the C–S bonds begin to elongate, with \({r}_{1}^{\mathrm{CS}}\) lengthening more than \({r}_{2}^{\mathrm{CS}}\). The time gap between the cleavage of the C–S bonds averages 33 ± 13 fs (ranging from 4–80 fs); this elongation continues until SO2 is extruded. Cascade (or domino) reactions and telescoped synthetic procedures offer useful means to increase the efficiency of target-oriented synthesis. Thiophene S,S-dioxides offer an excellent opportunity to engineer such events as their ability to undergo cycloadditions that are rendered irreversible by the in situ extrusion of SO2 offers a direct means to construct cyclohexadienes. However, TDOs have rarely been exploited in synthetic contexts, and their use in asymmetric cycloadditions has not been reported. Here, TDOs equipped with a cheap, common chiral camphorsultam side chain have been shown to offer efficient, stereoselective and general access to polycyclic indolines, where the cycloaddition step proceeds under remarkably mild conditions (0 °C) for indoles that are unsubstituted at the 2- and 3-positions. The observation that these cycloadditions proceed with exceptional stercocontrol (>20:1 d.r.) suggests that TDOs offer a mild and general method for polycyclic indoline synthesis. Application of this chemistry to various tryptamine derivatives provides an entry to the Strychnos alkaloid framework, which enabled the asymmetric synthesis of eight Strychnos alkaloids via the most concise routes reported so far. This includes the historic natural product brucine which, despite being known since the early 1800s, had yet to succumb to total synthesis. Key to the wider application of TDOs is an understanding of the basis of their reactivity and selectivity in asymmetric cycloaddition reactions. Computational studies using QM calculations and MD simulations driven by MLIPs provided insight into the source of stereocontrol imparted by the camphorsultam side chain in both inter- and intramolecular cycloaddition pathways, and also revealed highly asynchronous cycloaddition pathways or even stepwise ‘Michael addition' mechanisms for the reaction of the TDO with the indole substrate, along with spontaneous extrusion of SO2. This results in sound principles for the design of further reactions exploiting TDOs as dienes. Overall, these studies demonstrate that TDOs offer broad potential for application in the synthesis of structurally complex architectures across synthetic and medicinal chemistry. To a stirred solution of trifluoroacetic anhydride (64.3 ml, 463 mmol, 10.2 equiv.) at 0 °C was added H2O2 (30 wt% in H2O, 16.3 ml, 160 mmol, 3.5 equiv.) The resulting mixture was warmed to room temperature and stirred for 15 min. To the resulting mixture at 0 °C was added (5-chlorothiophen-2-yl)[(3aS,6R,7aR)-8,8-dimethyl-2,2-dioxidotetrahydro-3H-3a,6-methanobenzo[c]isothiazol-1(4H)-yl]methanone (16.4 g, 45.6 mmol, 1.0 equiv.). The resulting mixture was warmed to room temperature and stirred for 14 h before it was concentrated under reduced pressure. The crude residue was recrystallized from CHCl3–MeCN (3:1) to afford compound 9a (15.7 g, 40.1 mmol, 88%) as a light-yellow solid. To a stirred solution of 11 (217 mg, 0.64 mmol, 1.0 equiv.) in MeCN (26 ml) at room temperature was added 9a (250 mg, 0.64 mmol, 1.0 equiv.). The resulting mixture was stirred for 14 h before additional 11 (261 mg, 0.77 mmol, 1.2 equiv.) The resulting mixture was warmed to 80 °C and stirred for 40 h before it was cooled to 65 °C and AcOH (0.55 ml, 9.61 mmol, 15.0 equiv.) The resulting mixture was stirred for 15 min at 65 °C, and then NaBH3CN (401 mg, 6.38 mmol, 10.0 equiv.) in MeOH (6.5 ml) was added. The resulting mixture was stirred at 65 °C for 30 min before it was cooled to room temperature, quenched by the slow addition of a saturated aqueous solution of NaHCO3 (20 ml) and then diluted with water (10 ml) and CH2Cl2 (20 ml). The layers were separated and the aqueous layer was extracted with CH2Cl2 (3 × 30 ml). The combined organic layers were washed with water (50 ml) and brine (50 ml), dried (Na2SO4) and concentrated under reduced pressure. Flash column chromatography (silica gel, pentane–Et2O 9:1→3:1) afforded compound 14 and its C3a epimer 14′ (276 mg combined mass, 0.44 mmol, 68%, 1:1.4 d.r.) as a brown foam, along with recovered tryptamine 11 (240 mg, 0.71 mmol). These diastereomers could be separated for the purposes of characterization and the subsequent reduction of 14. To a stirred solution of Boc-tryptamine 16 (1.00 g, 3.84 mmol, 2.0 equiv.) in CHCl3 (19.0 ml) at room temperature was added 9d. The resulting mixture was warmed to 80 °C and stirred for 23 h, before it was cooled to 0 °C and TFA (9.0 ml) was added. The resulting mixture was warmed to room temperature and stirred for 2.5 h before it was cooled to 0 °C, quenched by the slow addition of a saturated aqueous solution of Na2CO3 (200 ml) and then stirred for 20 h. The layers were separated and the aqueous layer was extracted with CH2Cl2 (3 × 50 ml), and then the combined organic layers were dried (Na2SO4) and concentrated under reduced pressure. Flash column chromatography (silica gel, pentane–EtOAc 9:1→4:6) afforded compound 19 (585 mg, 1.10 mmol, 57% over two steps) as a yellow foam. To a stirred solution of indole (0.2 mmol) in CH2Cl2 (0.1 M) at 0 °C was added thiophene S,S-dioxide (0.1 mmol). The resulting mixture was stirred for 38 h before it was concentrated under reduced pressure to afford the crude material, which was purified by flash column chromatography using pentane–EtOAc eluent. The data generated in this study, including details concerning experimental procedures and characterization data, are available in the Supplementary Information. Crystallographic data for structure 10a reported in this Article are available in the Supplementary Information, and have been deposited at the Cambridge Crystallographic Data Centre (CCDC 2408040). Copies of this data can be obtained free of charge at https://www.ccdc.cam.ac.uk/structures/. & House, M. C. “For its size, the most complex natural product known.” Who deserves credit for determining the structure of strychnine? Woodward, R. B. et al. The total synthesis of strychnine. Woodward, R. B. et al. The total synthesis of strychnine. He, W., Wang, P., Chen, J. & Xie, W. Recent progress in the total synthesis of Strychnos alkaloids. Cannon, J. S. & Overman, L. E. Is there no end to the total syntheses of strychnine? Lessons learned in strategy and tactics in total synthesis. & Solé, D. Synthesis of strychnine. Teijaro, C. N. et al. Synthesis and biological evaluation of pentacyclic Strychnos alkaloids as selective modulators of the ABCC10 (MRP7) efflux pump. Gaich, T. & Baran, P. S. Aiming for the ideal synthesis. Trost, B. M. The atom economy—a search for synthetic efficiency. Martin, D. B. C. & Vanderwal, C. D. A synthesis of strychnine by a longest linear sequence of six steps. Jones, S. B., Simmons, B., Mastracchio, A. & MacMillan, D. W. C. Collective synthesis of natural products by means of organocascade catalysis. Zhou, W. et al. A bridged backbone strategy enables collective synthesis of strychnan alkaloids. Hutchings-Goetz, L. S., Yang, C., Fyfe, J. W. B. & Snaddon, T. N. Enantioselective syntheses of Strychnos and Chelidonium alkaloids through regio- and stereocontrolled cooperative catalysis. Martin, D. B. C., Nguyen, L. Q. & Vanderwal, C. D. Syntheses of strychnine, norfluorocurarine, dehydrodesacetylretuline, and valparicine enabled by intramolecular cycloadditions of Zincke aldehydes. Campbell, E. L., Zuhl, A. M., Liu, C. M. & Boger, D. L. Total synthesis of (+)-fendleridine (aspidoalbidine) and (+)-1-acetylaspidoalbidine. Padwa, A. Domino reactions of rhodium(II) carbenoids for alkaloid synthesis. 1-Oxides and 1,1-dioxides of thiophenes and selenophenes and related compounds. Formal syntheses of dictyodendrins B, C, and E by a multi-substituted indole synthesis. Wang, Z.-S. et al. De novo synthesis of dihydrobenzofurans and indolines and its application to a modular, asymmetric synthesis of beraprost. Park, K. H. K., Frank, N., Duarte, F. & Anderson, E. A. Collective synthesis of illudalane sesquiterpenes via cascade inverse electron demand (4 + 2) cycloadditions of thiophene S,S-dioxides. Nenajdenko, V. G., Moiseev, A. M. & Balenkova, E. S. A novel method for the oxidation of thiophenes. Synthesis of thiophene 1,1-dioxides containing electron-withdrawing substituents. Oppolzer, W., Chapuis, C. & Bernardinelli, G. Camphor-derived N-acryloyl and N-crotonoyl sultams: practical activated dienophiles in asymmetric Diels–Alder reactions. Huh, C. W., Bechle, B. M. & Warmus, J. S. Development of a scalable synthetic route towards a 2,2,6-trisubstituted chiral morpholine via stereoselective hydroalkoxylation. Azzouzi, A., Perrin, B., Sinibaldi, M.-E., Gramain, J.-C. & Lavaud, C. Stereoselective preparation of tri and tetracyclic amines as potential intermediates in Aspidosperma alkaloid synthesis. A short formal total synthesis of strychnine with a samarium diiodide induced cascade reaction as the key step. Maertens, G., Deruer, E., Denis, M. & Canesi, S. Common strategy for the synthesis of some Strychnos indole alkaloids. Asmari Bardazard, K. et al. Regioselective synthesis of enantiopure 1,2- and 1,3-dispirooxindoles along with a DFT study. & Vanderwal, C. D. A synthesis of alsmaphorazine B demonstrates the chemical feasibility of a new biogenetic hypothesis. Rawal, V. H. & Michoud, C. A general solution to the synthesis of 2-azabicyclo[3.3.1]nonane unit of Strychnos alkaloids. & Vanderwal, C. D. A sequential cycloaddition strategy for the synthesis of Alsmaphorazine B traces a path through a family of Alstonia alkaloids. Fischer, E. Ueber die Spaltung einiger racemischer Amidosäuren in die optisch-activen Componenten. Kim, H. Y., Li, J.-Y., Kim, S. & Oh, K. Stereodivergency in catalytic asymmetric conjugate addition reactions of glycine (ket)imines. Larock, R. C. & Yum, E. K. Synthesis of indoles via palladium-catalyzed heteroannulation of internal alkynes. Kamakolanu, U. G. et al. Discovery and structure–activity relationships of nociceptin receptor partial agonists that afford symptom ablation in Parkinson's disease models. Batatia, I., Kovács, D. P., Simm, G. N. C., Ortner, C. & Csányi, G. MACE: higher order equivariant message passing Neural networks for fast and accurate force fields. Zhang, H., Juraskova, V. & Duarte, F. Modelling chemical processes in explicit solvents with machine learning potentials. Unke, O. T. et al. Machine learning force fields. Bickelhaupt, F. M. & Houk, K. N. Analyzing reaction rates with the distortion/interaction-activation strain model. thanks Studienstiftung des Deutschen Volkes for a scholarship. thanks the ASAN Foundation for a scholarship. J.P. thanks the Marie Skłodowska-Curie actions for an Individual Fellowship (GA no. thank the EPSRC for support (EP/X028674/1). thanks the EPSRC Centre for Doctoral Training in Theory and Modelling in Chemical Sciences (EP/L015722/1). The computational work was performed using the University of Oxford Advanced Research Computing (ARC) facility and the Cirrus UK National Tier-2 HPC Service at EPCC (http://www.cirrus.ac.uk), funded by the University of Edinburgh and EPSRC (EP/P020267/1). These authors contributed equally: Kun Ho ‘Kenny' Park, Jisook Park. Chemistry Research Laboratory, Department of Chemistry, University of Oxford, Oxford, UK Kun Ho ‘Kenny' Park, Jisook Park, Nils Frank, Hanwen Zhang, Peilin Tian, Yasmine Biddick, Fernanda Duarte & Edward A. Anderson Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar and J.P. conceived the project. The experimental work was carried out by K.P., J.P. and P.T. carried out the computational analysis. acquired the X-ray crystal structure of 10a. The project was supervised by E.A.A. wrote the initial draft of the paper, which was reviewed and edited by all authors. Correspondence to Fernanda Duarte or Edward A. Anderson. The authors declare no competing interests. Nature Chemistry thanks Sylvain Canesi and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1–11, Schemes 1–5, Tables 1–18, experimental procedures, computational details and copies of NMR spectra. The xyz coordinates for all computed structures. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Reprints and permissions Collective asymmetric synthesis of the Strychnos alkaloids via thiophene S,S-dioxide cycloadditions. Version of record: 23 January 2026 Anyone you share the following link with will be able to read this content: Sorry, a shareable link is not currently available for this article. Provided by the Springer Nature SharedIt content-sharing initiative © 2026 Springer Nature Limited Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript. Nature Energy (2026)Cite this article Charge transfer across interfaces constitutes the rate-determining step in electrochemical systems. Sluggish kinetics triggers side reactions and hazardous surface morphologies, as represented by dendritic/dead lithium (Li0) in Li-metal batteries (LMBs), especially under ultrafast charging (UFC). Here we report an approach to accelerate interfacial charge transfer by redesigning the solvent molecular structure into a distinctive planar coordination of lone-pair electrons (LPEs) with alkaline cations (Li+ or Na+). This planar-aligned electron channel (PAEC) greatly strengthens the coupling between LPEs and Li+, promoting Li+/Li0 redox reaction kinetics and reversibility. The designed electrolyte dramatically enables stable cycling of industrial 2 Ah Li||LiNi0.8Mn0.1Co0.1O2 pouch cells at an ultrahigh rate of 4 C, achieving 100% full charge within 15 min at a charging power density of 1,747.6 W kg−1. We establish a link between the solvation electronic structure and charge-transfer dynamics, highlighting a potential strategy for electrolyte design under extreme electrochemical conditions. This is a preview of subscription content, access via your institution Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription cancel any time Subscribe to this journal Receive 12 digital issues and online access to articles $119.00 per year only $9.92 per issue Buy this article Prices may be subject to local taxes which are calculated during checkout All data for this study are provided within the main Article and its Supplementary Information files. Source data are provided with this paper. Xu, K. Electrolytes, Interfaces and Interphases (Royal Society of Chemistry, 2023). Wan, G. et al. Solvent-mediated oxide hydrogenation in layered cathodes. Hou, S. et al. Solvation sheath reorganization enables divalent metal batteries with fast interfacial charge transfer kinetics. Zhou, Y. et al. Strongly correlated perovskite fuel cells. Ramaswamy, N. & Mukerjee, S. Alkaline anion-exchange membrane fuel cells: challenges in electrocatalysis and interfacial charge transfer. Xu, K. Electrolytes and interphases in Li-ion batteries and beyond. Xiao, P. et al. Insights into the solvation chemistry in liquid electrolytes for lithium-based rechargeable batteries. Xu, W. et al. Lithium metal anodes for rechargeable batteries. Energy Environ. Zhang, S. et al. Oscillatory solvation chemistry for a 500 Wh kg−1 Li-metal pouch cell. Zhang, Q.-K. et al. Homogeneous and mechanically stable solid–electrolyte interphase enabled by trioxane-modulated electrolytes for lithium metal batteries. Niu, C. et al. Balancing interfacial reactions to achieve long cycle life in high-energy lithium metal batteries. Jie, Y. et al. Towards long-life 500 Wh kg−1 lithium metal pouch cells via compact ion-pair aggregate electrolytes. Xu, Q. et al. Li2ZrF6-based electrolytes for durable lithium metal batteries. Liu, Y. et al. Self-assembled monolayers direct a LiF-rich interphase toward long-life lithium metal batteries. Jagger, B. & Pasta, M. Solid electrolyte interphases in lithium metal batteries. Meng, Y. S., Srinivasan, V. & Xu, K. Designing better electrolytes. Ruan, D. et al. Solvent versus anion chemistry: Unveiling the structure-dependent reactivity in tailoring electrochemical interphases for lithium-metal batteries. JACS Au 3, 953–963 (2023). Lu, D. et al. Ligand-channel-enabled ultrafast Li-ion conduction. Shen, X. et al. The failure of solid electrolyte interphase on Li metal anode: structural uniformity or mechanical strength? Energy Mater. Holoubek, J. et al. Toward a quantitative interfacial description of solvation for Li metal battery operation under extreme conditions. Natl Acad. Zhang, Z. et al. Capturing the swelling of solid-electrolyte interphase in lithium metal batteries. Chen, K.-H. et al. Dead lithium: mass transport effects on voltage, capacity, and failure of lithium metal anodes. Lu, D. et al. Failure mechanism for fast-charged lithium metal batteries with liquid electrolytes. Energy Mater. Jiao, S. et al. Behavior of lithium metal anodes under various capacity utilization and high current density in lithium metal batteries. Chen, K. et al. Correlating the solvating power of solvents with the strength of ion-dipole interaction in electrolytes of lithium-ion batteries. Li, Z. et al. Non-polar ether-based electrolyte solutions for stable high-voltage non-aqueous lithium metal batteries. Yu, Z. et al. Molecular design for electrolyte solvents enabling energy-dense and long-cycling lithium metal batteries. Choi, I. R. et al. Asymmetric ether solvents for high-rate lithium metal batteries. Jiao, S. et al. Stable cycling of high-voltage lithium metal batteries in ether electrolytes. Ren, X. et al. Enabling high-voltage lithium-metal batteries under practical conditions. Chen, S. et al. Unveiling the critical role of ion coordination configuration of ether electrolytes for high voltage lithium metal batteries. Chen, S. et al. Strongly solvating ether electrolytes for high-voltage lithium metal batteries. Li, R. et al. Upgrading electrolyte antioxidant chemistry by constructing potential scaling relationship. Wiberg, K. B. & Rablen, P. R. Comparison of atomic charges derived via different procedures. Lu, T. & Chen, F. C. Meaning and functional form of the electron localization function. Acta Phys. Glendening, E. D., Landis, C. R. & Weinhold, F. Natural bond orbital methods. WIREs Comput. Lu, T. & Chen, F. Multiwfn: a multifunctional wavefunction analyzer. Lu, T. & Chen, F. Calculation of molecular orbital composition. Acta Chim. Lax, M. The Franck-Condon principle and its application to crystals. Article MathSciNet Creutz, C. & Taube, H. Direct approach to measuring the Franck-Condon barrier to electron transfer between metal ions. Wagner, M. et al. An electride with a large six-electron ring. Dong, H., Feng, Y. & Bu, Y. Electron presolvation in tetrahydrofuran-incorporated supramolecular sodium entities. Chen, X. et al. Phase transfer-mediated degradation of ether-based localized high-concentration electrolytes in alkali metal batteries. Cao, X., Jia, H., Xu, W. & Zhang, J.-G. Review—Localized high-concentration electrolytes for lithium batteries. Boyle, D. T. et al. Transient voltammetry with ultramicroelectrodes reveals the electron transfer kinetics of lithium metal anodes. ACS Energy Lett. Yuan, X., Liu, B., Mecklenburg, M. & Li, Y. Ultrafast deposition of faceted lithium polyhedra by outpacing SEI formation. Chen, Y. et al. Failure process during fast charging of lithium metal batteries with weakly solvating fluoroether electrolytes. Zhang, Y. et al. Unveiling the impacts of charge/discharge rate on the cycling performance of Li-metal batteries. ACS Energy Lett. Dolomanov, O. V., Bourhis, L. J., Gildea, R. J., Howard, J. A. K. & Puschmann, H. OLEX2: a complete structure solution, refinement and analysis program. M. J. et al. Gaussian 16, Revision C.01 (Gaussian, 2016). Glendening, E. D., Reed, A. E., Carpenter, J. E. & Weinhold, F. NBO Version 3.1 (Gaussian, 2003). Lu, T. A comprehensive electron wavefunction analysis toolbox for chemists, Multiwfn. Berendsen, H. J. C., van der Spoel, D. & van Drunen, R. GROMACS: a message-passing parallel molecular dynamics implementation. Jorgensen, W. L., Maxwell, D. S. & Tirado-Rives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. Fischer, J., Paschek, D., Geiger, A. & Sadowski, G. Modeling of aqueous poly(oxyethylene) solutions: 1. Atomistic simulations. Shimizu, K., Almantariotis, D., Gomes, M. F. C., Pádua, A. A. H. & Canongia Lopes, J. N. Molecular force field for ionic liquids V: hydroxyethylimidazolium, dimethoxy-2-methylimidazolium, and fluoroalkylimidazolium cations and bis(fluorosulfonyl)amide, perfluoroalkanesulfonylamide, and fluoroalkylfluorophosphate anions. Dodda, L. S., Cabeza de Vaca, I., Tirado-Rives, J. & Jorgensen, W. L. LigParGen web server: an automatic OPLS-AA parameter generator for organic ligands. Nucleic Acids Res. Martinez, L., Andrade, R., Birgin, E. G. & Martinez, J. M. PACKMOL: a package for building initial configurations for molecular dynamics simulations. Hoover, W. G. Canonical dynamics: equilibrium phase-space distributions. A Gen. Phys. Parrinello, M. & Rahman, A. Polymorphic transitions in single crystals: a new molecular dynamics method. Yeh, I.-C. & Berkowitz, M. L. Ewald summation for systems with slab geometry. Download references This study was supported by the National Key R&D Program of China (no. 2021YFA1201800), the National Natural Science Foundation of China (nos. We are grateful for resources from the Instruments Center for Physical Science and the Center for Micro and Nanoscale Research and Fabrication at USTC. The solvent screening and simulations were performed on the robotic AI-scientist platform of the Chinese Academy of Sciences. XAS data were collected at beamlines BL12B-a and BL12B-b in the National Synchrotron Radiation Laboratory (NSRL) in Hefei, China. Single-crystal X-ray diffraction measurements were carried out on an XtaLAB PRO diffractometer in the Core Facility Center for Life Sciences, USTC. We acknowledge B. Jiang, J. Hong and Z. Cui for their helpful discussion about CDFT calculations. We also thank Hubei Lishizhen Pharmaceutical Research Co. for the scale-up synthesis of the MTP molecule. These authors contributed equally: Digen Ruan, Shunqiang Chen. Hefei National Research Center for Physical Sciences at the Microscale, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei, China Digen Ruan, Shunqiang Chen, Jiasen Guo, Dazhuang Wang, Yuan Li, Jun Ma, Zhihao Ma, Zihong Wang, Ruiguo Cao, Shuhong Jiao & Xiaodi Ren School of Physics, Hefei University of Technology, Hefei, China Weiduo Zhu College of Chemistry and Molecular Sciences, Wuhan University, Wuhan, China Bing Huang School of Life Sciences, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, China Zhongliang Zhu Department of Chemistry and Biochemistry, The Ohio State University, Columbus, OH, USA Yiying Wu SES AI Corp., Woburn, MA, USA Kang Xu Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar conceptualized this work. and S.C. collected the experimental data. conducted the theoretical calculations. synthesized the solvent molecule. grew the single crystals, and Z.Z. characterized and analysed the single-crystal structures. tested electrolyte physical properties. contributed to data collection and discussions. prepared the manuscript with input from all the co-authors. supervised the work. Correspondence to Kang Xu or Xiaodi Ren. The authors declare no competing interests. Nature Energy thanks Jiheong Kang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1–63, Supplementary Tables 1–8 and Supplementary References. MTP-LiOTF single crystal. G2–LiOTF single crystal. Source data for Supplementary Figs. Source data for Fig. Source data for Fig. Source data for Fig. Source data for Fig. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. Reprints and permissions Ruan, D., Chen, S., Guo, J. et al. Molecularly aligned electron channels for ultrafast-charging practical lithium-metal batteries. Nat Energy (2026). Download citation Received: 06 June 2025 Accepted: 16 December 2025 Published: 23 January 2026 Version of record: 23 January 2026 DOI: https://doi.org/10.1038/s41560-025-01961-z Anyone you share the following link with will be able to read this content: Sorry, a shareable link is not currently available for this article. Provided by the Springer Nature SharedIt content-sharing initiative Nature Energy (Nat Energy) ISSN 2058-7546 (online) © 2026 Springer Nature Limited Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript. Climate and atmospheric changes are impacting forest function and structure worldwide, but their effects on tropical forest diversity are unclear. Here, using 406 permanent plots spanning four decades of intact lowland and montane forest dynamics, we test for long-term change in species richness and assess the influence of climate and other variables. We show that, at a continental scale, species richness appears stable, but this masks substantial regional variation. Species richness increased in Northern Andean and Western Amazon plots, yet declined in the Central Andes, Guyana Shield and Central-Eastern Amazon. Overall, warmer, drier and more seasonal forests lost species, while those at higher elevations, in less fragmented areas and with faster rates of tree turnover experienced increases. Region-specific drivers, particularly precipitation seasonality and demographic factors, modulated these trends. The results highlight the diverse ways in which Amazon–Andes forests are changing and underscore the critical need to preserve large-scale ecosystem integrity to maintain local tree diversity. By doing so, Northern Andean forests in particular could serve as an important refuge for species increasingly displaced by climate change. However, climate change and land-use change are threatening the stability of these ecosystems and the services they provide5,6,7,8,9,10. Over recent decades, temperatures have increased in this region, precipitation patterns have become more extreme and variable, deforestation has expanded and forest fires have become more frequent7,11,12,13,14,15. Under these increasingly stressful conditions, plant species have two feasible short-term responses to survive: (1) migrate—shift their distribution range in response to changing environmental conditions, or (2) acclimate—utilize their physiological tolerance to maintain function under the new conditions. If species do not manage to migrate or acclimate, their populations will decrease and eventually they may go extinct16. The response of plant species to climate change could lead to changes in forest structure, composition, diversity and species richness at the local scale17,18,19. The Andes and other tropical mountains are undergoing a process of thermophilization, where higher-elevation forests are incorporating new lower-elevation species that expand their ranges upslope, and current low-elevation species are increasing in relative abundance20,21,22,23. However, lower-elevation forests face the possibility of biotic attrition (a net loss of species), as there is no species pool from even hotter areas able to migrate and fill the new thermal niches24,25,26. While the wet tropics have been suggested to have the highest rates of plant extinction, based on literature reviews27 and model predictions28, we do not know whether this translates into consistent losses of local richness within the different regions in the Andes–Amazon area. Despite widespread threats across the Andes–Amazon area, climate change and other large-scale disturbances are not distributed evenly across space7,9,29. Moreover, geographical features—such as increased topographical variation, which may provide a potential advantage for species persistence by offering more suitable environmental conditions—are also unevenly distributed30,31. At the local scale, stressors, such as increasing temperatures and declining rainfall, have been related to mortality-driven compositional shifts, particularly in steep elevational gradients21,30,32,33. Baseline temperature and precipitation regimes have also been shown to relate to the probability of plant species suffering thermal or drought damage34,35. Fragmented areas are also vulnerable to diversity losses, while increasing fire frequency reduces regeneration and species richness13,36,37. However, although several mechanisms have been shown to drive changes in (neo)tropical forest diversity, most studies so far have been limited to local or regional scales and/or lack long-term assessments of tree richness and diversity at consistently monitored sites. Indeed, long-term compositional changes have often been estimated using modelling approaches and have rarely been addressed using field data (but see refs. Here we use 406 long-term floristic plots, measured for different time periods since 1971 across 10 countries in South America to estimate the magnitude and direction of tree richness change through time and to identify their drivers. Across this vast space, ranging from −17 to 8.5 latitudinal degrees and −80 to −47 longitudinal degrees, we explore the change in richness through time for the combined area and independently for each of six predefined regions (based on their geomorphological and biogeographical history and contemporaneous geoecological features), as we hypothesize that different regions are responding in different ways, forced by different drivers. Using consistent methods to identify the spatial distribution of diversity change and the factors that contribute to it at a larger scale is crucial to understanding the current status of the Amazon and Andean forests, predicting future patterns and informing conservation efforts. With this comprehensive plot compilation and a set of climatic and structural variables, we intend to answer the following questions. First, using the complete dataset, we ask: (1) How is tree species richness changing across the Andes–Amazon area? and (2) How are changes in richness related to baseline climate, climate change, landscape context and forest structure? We predict an overall stability of richness, with local increases and declines balancing each other out. However, we expect the change in richness to be associated with several large-scale variables. In particular, we expect a more pronounced decrease in richness in warmer, drier forests at lower elevations given the thermophilization trend where species are ‘migrating' towards higher elevations that usually tend to be colder and wetter. Similarly, we predict a richness decline in forests that are becoming warmer or drier as dealing with this climate becomes physiologically more challenging. We also expect a decrease in species richness in forests with high fragmentation due to the reduced source of colonizers and habitat connectivity. A summary of all the predictors tested and our hypothesized relationships with species richness change is presented in Table 1. and (4) Which of the selected predictors explain the change in species richness for each region? We expect to find a longitudinal gradient in diversity change across the six Andes–Amazon regions driven by the most pressing stressors in each region. In particular, we hypothesize (a) an increase in richness in the Andes as a consequence of thermophilization and a decrease of richness in the Amazon, particularly in the drier and warmer Central-Eastern regions (Guyana Shield, Central-Eastern Amazon and Southern Amazon), due to biotic attrition; (b) temperature will thus be a crucial factor in the Andean trends, while precipitation could be more important in the Amazon; and (c) landscape integrity will have an important role in the more degraded Southern and Central-Eastern Amazon regions Half of our plots (203) declined in richness, and 146 increased. Richness change varied widely across plots (range −1.95% to +3.3% per year) but had no consistent direction at the Andes–Amazon scale (bootstrapped mean richness change 0.036, mean confidence interval (CI) −0.09 to 0.16, mean t statistic 0.579, mean P value 0.56, degrees of freedom 179) (Supplementary Fig. We found a negative relationship between richness change and longitude (slope −2.39, adjusted R2 = 0.047, P < 0.001). At −64.5°, which coincides broadly with the transition between the Eastern and Western Amazon, the change in richness shifts from positive (West) to negative (East) values (Fig. There was no significant relationship with latitude. a,b, Relationship between plot location and richness change per plot: longitude in decimal degrees (a) and absolute latitude in decimal degrees (b). Each point represents a plot, and its colour corresponds to the region. The solid line represents statistically significant (P < 0.001) linear regression. In the bivariate regressions with the complete dataset, we found that maximum temperature, precipitation seasonality and precipitation seasonality change had significant negative relationships with richness change (Fig. Temperature change exhibited a hump-shaped relationship with richness, decreasing slightly where temperatures cooled and more markedly where warming was faster. Annual precipitation, stem abundance change, landscape integrity, elevation and identification effort change had positive significant relationships (Fig. The bootstrapped regressions corrected for spatial bias in plot location supported the representativity of the overall trends found as slope direction and significance coincided for most of the variables (Supplementary Table 2 and Supplementary Fig. The regression with annual precipitation, although always positive, was on average not significant in the bias-corrected analysis, and the one with landscape integrity was typically positive but not significant, probably because of the confounding effect of decreasing tree cover with elevation in the Andes. Bivariate regression between richness change (% yr−1) and the different predictors. Solid lines represent statistically significant regressions (P ≤ 0.05). Shaded ribbons around lines represent 95% CI. For extended statistical results, see Supplementary Table 1. Species richness declined with increasing precipitation seasonality, but this decline was steeper for less seasonal forests. Species richness in less seasonal forests increased with annual precipitation. We found marginal support for an interaction between the temperature variables, suggesting that warmer forests experiencing further warming lost more species, whereas cooler forests even showed a slight increase in richness. Richness change was directional in five of the six regions (Fig. Species richness significantly increased in the Northern Andes and Western Amazon, while the Central Andes, Central-Eastern Amazon and Guyana Shield experienced significant declines. Although the Southern Amazon did not show a significant trend, the mean change was negative and included some of the most extreme negative values. The direction of these changes coincided across the other diversity indices tested (Supplementary Table 4), although the significance of the change was more variable because different indices reflect slightly different aspects of diversity change (Supplementary Note 1). a, Map showing the distribution of the 406 plots (arrowhead symbols located at plot coordinates) in the six regions (NA, Northern Andes; CA, Central Andes; WA, Western Amazon; GU, Guyana Shield; CEA, Central-Eastern Amazon; SA, Southern Amazon). Symbol colour and angle represent richness change direction, and symbol size is proportional to the magnitude of change for each plot. Black circles represent no net change. Background SRTM represents elevation in m a.s.l. b, Richness change (% yr−1) per region expressed as proportional change in relation to the initial census. For extended results, see Supplementary Table 4. We used a multigroup piecewise structural equation model (SEM) analysis to identify the relationship between the predictor variables and the richness change directly and indirectly. 4) showed a good fit to the data (Fisher's C = 4.232, P = 0.375). The individual R2 for the component models were 0.18 (mortality), 0.27 (stem abundance change) and 0.30 (species richness change). Complete model results are presented in Supplementary Table 5. Diagram illustrating the relationships between the independent variables in the model, with richness change as the final response variable, and stem abundance change and mortality rate as intermediate response variables that may also influence richness change. Both panels are part of the same SEM, but for easier interpretation, they show general and region-specific relationships separately. a, Significant relationships constrained across the study area, with arrowhead colour indicating negative or positive effects. The effect of annual precipitation on stem abundance change (marked by asterisk) is constrained to 0. The effect of mortality rate on stem abundance change (marked by hash sign) is positive and significant across regions but not constrained. Non-significant constrained relationships are shown in grey. b, Significant relationships in specific regions, with arrow colour indicating the region, width representing the standardized effect size (in mm × 2) and stroke style denoting the effect sign (solid, positive; dashed, negative). For standardized effect sizes of all variables, see Supplementary Table 5. Many of the relationships between climate and environmental variables with stem abundance change and mortality rate were constrained (indicating a similar effect) across regions (Fig. For stem abundance change, five out of eight variables were constrained, with two of these being significant; for mortality, four out of the eight variables were constrained, with two of them significant. For richness change, 5 out of 11 variables were constrained, with 4 being significant. Regarding the intermediate factors mediating indirect effects, mortality rate had a significant negative effect on richness in the Central Andes, and stem abundance change had a significant positive effect on richness change in all regions. We computed the indirect effects that the predictors had on richness change mediated by the structural variables when each path coefficient was significant (Fig. Only significant (P ≤ 0.05) direct effects are shown. Indirect effects were calculated by multiplying the significant (P ≤ 0.05) standardized coefficients within each of the possible three indirect pathways (via stem abundance change, via mortality rate, and via mortality rate × stem abundance change) and then adding them. Region colours are shown in the top line for coherence with Fig. For extended results, including specific P values, see Supplementary Tables 5 and 6. Maximum temperature had a total negative effect on richness across regions, while precipitation had a general positive effect. Precipitation seasonality had a strong negative effect in the Andes but positive in the Southern Amazon. Temperature change had a very small negative effect in the Central Andes, Western Amazon and Central-Eastern Amazon. Precipitation change had a large positive effect in the Guyana Shield. Precipitation seasonality change was variable, having a large negative effect in the Northern Andes and Southern Amazon but a positive effect in the Central Andes. Stem abundance change had a positive effect in all regions, while mortality had a negative effect. Landscape integrity had a strong positive effect in the Southern Amazon, weaker positive effects in other regions and a negative effect in the Central Andes. Change in identification effort had a positive direct effect in all regions except the Southern Amazon, while the time frame had very small positive effects in five regions and a negative effect in one region. We found no overall trend in species richness change across 406 forest-dynamics plots distributed across the tropical Andes and the Amazon. However, this large-scale result masks important regional variations, with richness increasing in the Northern Andes and Western Amazon, while decreasing in the Central Andes, Central-Eastern Amazon and Guyana Shield. This masking or obscuring issue has been raised for global estimations of diversity change based on local trends, and some even question the relevance of these large-scale averages39,40,41. In any case, the absence of a significant overall trend in richness change may also indicate a temporary disequilibrium between current environmental conditions and large-scale vegetation responses42, which should not be misinterpreted as resilience. Lag effects could occur on the leading edge, where trees slowly colonize newly suitable habitats, delaying potential richness gains. Alternatively, lags at the trailing edge could indicate a temporary persistence of species, artificially inflating current richness estimates42. Lowland areas of the Amazonia are expected to experience greater lags due to the long migration distances required to remain at equilibrium with their optimal conditions43. By contrast, mountain regions are thought to have an extinction debt, allowing temporary species accumulation44,45. Our findings on species richness change align with this longitudinal gradient, revealing negative trends in the Eastern regions and positive trends in the Western regions. We first discuss these large-scale patterns, followed by the regional findings that help explain these trends. Over the past 40 years, more than 90% of our plots (368/406) have experienced warming with a mean rate of 0.028 ± 0.018 °C per year (321/406 during the individual monitoring periods). Faster-warming forests in the Central-Eastern and Southern Amazon (0.05 ± 0.02 °C per year) are losing species at a higher rate than forests experiencing more moderate warming. In addition, forests in warmer areas within the Andes–Amazon area are also losing more species (Fig. This pattern reflects the contrasting conditions and biotic responses of the Andes and Amazon forests and, supported by the higher rate of species accumulation with increasing elevation (Fig. 2), provides further evidence for thermophilization in the region21,30,47. This phenomenon is also supported by the temperature interaction, where the impact of heating in driving species richness loss depends on the baseline temperature, with hotter forests being more sensitive to a given rate of heating. This trend indicates that precipitation change modulates richness responses to temperature. Forests that are more seasonal—and especially those becoming more seasonal—showed declines in species richness17, with the strongest negative effects in currently less seasonal or wetter forests (that is, higher annual precipitation; Extended Data Fig. These results agreed with findings from the Andean mountain tops, where seasonality across the latitudinal gradient is strongly linked to richness changes, with more aseasonal peaks near the Equator showing richness gains38. While we did not find a latitudinal trend across the study area (Fig. 1), we observed differences between the Northern and Central Andes, which we discuss in detail below. 2), as expected because more species from the regional pool have a chance to recruit. This pattern extends beyond individual plots, as forests in less fragmented landscapes (higher landscape integrity), surrounded by more contiguous forest, are more likely to show increases in species richness. By contrast, forests that become more isolated from surrounding fragments tend to experience a decline in species richness36,48,49. The Western Amazon and Northern Andes are gaining species, while the Central Andes, Guyana Shield and Central-Eastern Amazon are experiencing species loss (Fig. Generally, mortality rates rose in more seasonal and fragmented forests, while stem abundance declined in warming forests and in forests experiencing higher mortality rates. The relationship between richness change and environmental variables revealed many region-specific drivers, with some variables having opposite effects in different regions, highlighting the context-dependent processes in our vast study area. This means that a greater decline in the number of individuals in a plot (in proportion to the initial number) was associated with a more negative change in species richness, and vice versa. Changes in individual abundance are crucial for enabling compositional change, as more recruits increase the likelihood of detecting new species from the local pool50. However, the entry of new species does not necessarily imply a shift in composition outside the existing regional pool, and species loss could reflect local extinctions or shifts within the same pool. Further analysis is needed to determine whether these species are new or are part of the regional pool. All regions showed a negative trend in stem abundance, with the Eastern Amazon (Guyana Shield, Central-Eastern and Southern Amazon) experiencing sharper declines than the Western Amazon and the Andes, which showed higher variability. This is contrary to the results of previous research showing an increase in stem density across 50 Amazonian plots from 1979 to 200251. Although this discrepancy may simply reflect the differing sample sizes and geographical extents of the studies, it could also indicate a recent change in the stem density trend driven by rising temperatures. Mortality directly affected only the Central Andes, with its effects on other regions mediated through stem abundance change. Across regions, warmer and drier areas are linked to lower rates of richness change. Regional temperature gradients, particularly elevation gradients in the Andes, play a crucial role in richness change. We found an increase in richness in the Northern Andes, which agrees with the reported compositional change caused by the thermophilization process in the area21,30,52 and with research showing a warming-related increase in mountain-top diversity38,44. The encroachment of lower-elevation, warm-adapted species, which would initially be rare in the community, would lead to a potentially temporary increase in the number of species supported by the extinction lag of cold-adapted species that cannot tolerate the new conditions and will eventually become locally extinct53,54. We expected that both Andean regions would share the same pattern; however, the Central Andes showed a decline in richness. Our results suggest that the faster-warming Northern Andes region14 could be more suitable for range shifts than the more moderate—and even cooling—Central Andes (Extended Data Fig. Across the Andes, precipitation and its seasonality are highly variable, being affected by local orography, orientation and cloud cover14; however, on average, the Central Andes are drier and more seasonal than the Northern Andes, and they are also becoming drier and more seasonal at a faster rate (Extended Data Fig. We hypothesize that migrating lower-elevation species, particularly those distributed in the Western Amazon, are more likely to succeed expanding into higher elevations of the wetter and less seasonal Northern Andes than in the Central Andes. Furthermore, the negative relationship between richness and landscape integrity in the Central Andes probably results from the confounding effect of decreasing tree cover with elevation. The Western and Central-Eastern Amazon presented a very similar breakdown of driver effects. In both regions, changes in stem abundance were the primary ecological drivers, with minor indirect effects from climate variables, largely mediated by the change in stem abundance. In these regions, forests that are warmer, drier or becoming warmer or drier exhibited declining richness, as these conditions reduce the number of individuals. In the Southern Amazon, where there was no significant trend in richness change, and in the Guyana Shield, which showed a negative trend, precipitation and its seasonality played predominant roles. In the Southern Amazon, which is highly seasonal, there is evidence that forests that were more seasonal at baseline tended to gain species; however, increases in precipitation seasonality were associated with richness declines. Nevertheless, in the Southern Amazon (the area with some of the most fragmented forests), landscape integrity exerted the strongest direct effect on richness change: forests embedded within larger, contiguous forested areas tended to gain species, whereas more fragmented forests tended to lose them. Landscape integrity also had a negative relationship with mortality rate across all regions, indicating that higher landscape integrity supports tree survival, thereby increasing tree abundance, which, in turn, positively impacts richness. This agrees with previous findings on the damaging effects of deforestation and/or degradation in surrounding forests, underscoring the importance of preventing forest fragmentation to support biodiversity conservation55. It also highlights the conservation priority of the Western Amazon–Northern Andes corridor, which appears to be the most feasible pathway for range shifts that could support species persistence. This study provides a comprehensive assessment of tree richness change in the Andes–Amazon forests using long-term field data. However, we acknowledge that we are working in one of the most diverse and dynamic areas of the planet56, and, as such, there are limitations to our analyses. First, the dataset lacks a historical baseline, so initial conditions may be influenced by uncertain processes39. To minimize bias, we used strict plot selection criteria, excluding plots with any sign of fire or large disturbances and directly including identification effort change and time between censuses as predictors in our analyses. The change in identification effort positively influenced richness change across regions: as more individuals are identified, we encounter more species. Monitoring time had only a small effect on species richness change, where shorter intervals capture more noise relative to the signal than longer intervals. Second, climatic and environmental data extracted from global databases add uncertainty, especially in topographically complex areas like the Andes. At an even finer scale, it is impossible to know the real climate experienced on the forest floor by each individual tree; further investment in microclimate monitoring in these structurally complex forests is crucial to improve our understanding of climate change effects. Third, we are including only trees with a diameter at breast height (DBH) greater than 10 cm and are ignoring the potential contribution of smaller size classes to changes in diversity. Finally, there are multiple factors not accounted for in the study that can have important roles in diversity trends. For example, it was beyond the scope of this study to evaluate the roles of past forest history, including Indigenous management, in current richness trends, nor did we evaluate the potential role of biotic pressures (for example, herbivory and pathogens), nor that of conservation efforts and compensation mechanisms, including carbon and biodiversity benefits. Further research should address more complex compositional questions, such as evaluating the taxonomic and functional identities of species being lost or recruited, and whether this indicates that the Andes–Amazon is undergoing taxonomic homogenization, functional homogenization or both. In conclusion, across the study area, hot, dry and seasonal forests and those becoming warmer and more seasonal are losing species, while forests with higher tree density and higher landscape integrity are gaining them. Our large-scale findings emphasize the critical role of temperature and temperature change in shaping tree richness in the Andes–Amazon area. However, at the regional level, precipitation and its shifts in distribution and annual amounts play more important and region-specific roles, outweighing the influence of temperature57. This study highlights the uneven impact of changing environmental conditions on tree diversity across different tropical forests, as well as the varied importance of climate and environmental variables across the different regions and scales. Our results underscore the key role of the Northern Andes as a refuge for tree species facing increasingly unsuitable climatic conditions in the Amazon. Finally, our findings highlight the tight relationship between preserving tree abundance and preserving diversity, emphasizing the enormous threat posed by land-use change, which indiscriminately reduces both tree abundance and regional species diversity. Plot establishment and resurveys were performed by well-trained field teams that followed a detailed protocol that included geolocating plot boundaries, marking subcorners with permanent polyvinyl chloride tubes, taking tree subplot and coordinate data, tagging trees with numbered aluminium tags, and noting and painting the point of measurement. Post-field quality control was carried out by database managers and the field team leader. We selected all plots within the study area (Andean or Amazonian country in areas lower than 4,000 m above sea level (a.s.l.)) that had been censused at least twice. We did not include plots located in the Chocó and the Northern Venezuela regions because of insufficient sample sizes to represent these areas. To avoid the confounding effects of successional trends on diversity change, we included only plots in forests that were undisturbed or had experienced disturbance at least 50 years prior (identified as equivalent to long-term successional forest). For the same reason, we excluded plots that had been recorded on ForestPlots.net as swamp or seasonally flooded forests or as having a history of fire or of large disturbances. We also excluded plots that had been flagged for having taxonomic identification issues. We obtained curated datasets for each census and plot. Hereafter, we refer to these two censuses as ‘initial' and ‘final'. We ensured that plot area and location exactly matched on both censuses and that the plot sampling strategy was standardized across time. For instance, we excluded palms when they were not measured in every census. To standardize methodologies, we removed from the dataset subplots (delimited sections within a plot) in which the protocol required a minimum tree DBH greater than 10 cm for inclusion. We also removed all individuals smaller than 10 cm DBH and those belonging to the families Cyclanthaceae and Araceae. Species taxonomic identification was carried out in the field and in the herbaria where reference collections with vouchers are deposited. Any change in an individual's identification was applied across all censuses. To minimize the impact of the change in identification effort (the proportion of individuals identified to species level) between censuses, we restricted our analyses to plots that (1) had more than 50% of the tree individuals identified to species level in the initial census, (2) had a difference in the proportion of identified individuals between first and last census smaller than 10% and (3) had at least 50% of the recruits in the final census identified to species (when there were more than 20 recruits). We used the taxonomic name resolution service (TNRS) tool59 (https://tnrs.biendata.org) and R package60 to standardize species names. We looked for potential explanations such as spelling errors in the Tropicos (https://www.tropicos.org) and WFO (https://wfoplantlist.org/plant-list) lists, and we either manually modified the accepted name for these species or used only their genus ID if there was no clear option. As the treatment of morphospecies was not curated or standardized across the dataset, we converted any morphospecies codes into ‘Genus indet' format to group morphospecies into genera across the dataset. Given that the plot size varied widely, we grouped plots that were less than 0.5 ha in area if they had other plots within a 7-km radius with no indication of large differences (that is, similar elevation, forest type, soil classification and so on). We will refer to these plot groupings as ‘plots', given that they are treated as a single unit. We also reduced the size of our biggest plots (plot areas of 25 and 9 ha) by selecting two 1-ha subplots on opposite corners and treating them independently. We then eliminated plots that had intervals of less than 4 years between the two selected censuses, because we considered this time elapsed to be too short to provide mid-to-long-term diversity change information. Finally, after preliminary exploration of plot distributions, we removed plots with ten or fewer species in either the initial or final census, as adding or removing even a single species could produce extreme percentage changes (±10%). After the selection process, our dataset compiled information from 406 plots (or grouped plots) covering ~420 ha (range 0.25–3 ha, mean plot size 1.04 ± 0.26 ha) with a cumulative monitoring time of 4,847 years (range 4.01–44.2, mean 11.94 ± 8.01 years). We divided the study area into six regions roughly following previous studies61,62,63,64. 5 shows the relative floristic similarity of our plots and regions. We calculated species richness as the number of fully identified species in each plot and census (SP). We calculated the change in species richness (% yr−1) as richness change = (((SPinitial − SPfinal)/SPinitial) × 100)/time; where SPinitial and SPfinal are the richness in the initial and final censuses, respectively, and time is the time interval between the initial and final censuses (in years). Palms (family Arecaceae) were included in the analyses (when included in both the initial and final censuses) as their exclusion did not have a significant effect on the results (Supplementary Fig. To test whether there was a significant change in richness through time, we used two-sided t-test analyses on richness change both for each of the regions independently and for the combined database. Given that the number of plots was unevenly distributed among the regions, to avoid sampling bias in the combined dataset analysis, we randomly sampled 30 plots per region and carried out a two-sided t-test with this subset. We repeated this process 1,000 times and obtained the averages of the t-test means and P values. To assess potential linearity issues in the relationship between changes in the number of individuals and species richness, we calculated the change in species richness after rarefying both the initial and final censuses to the minimum number of individuals observed in either census (that is, whichever is lower) (package vegan). 7) supports the use of richness change and stem abundance change as independent variables (see ‘Predictor variables' section) in the subsequent analyses. Additional diversity indices and their change through time for each region were calculated using the vegan R package65 and tested in the same way as richness change (Supplementary Note 1). Despite the considerable identification efforts by all research groups involved in this project, many tree individuals remain unidentified (Indet indet), identified only to the genus level (for example, Ocotea indet) or classified as morphospecies (for example, Ocotea sp1, Ocotea sp2 and so on). These morphospecies codes were maintained through the multiple censuses and retroactively changed to a full species name in the database if one was given; however, the morphospecies criteria were not standardized across plots, nor were they curated. Consequently, some changes in species diversity are not captured due to these exclusions; however, we speculate that such unreported changes are probably caused by a small number of individuals that recruit or die without being identified across multiple censuses and are unlikely to be dominant members of the community. To support the use of species-level data despite potential issues such as mistakes, changes in botanists and changes in the species concept through time, we calculated the change in genus richness in the same way as the change in species richness (but using the individual's genus-level information, thus including morphospecies). 8) and for each region independently (Supplementary Table 9). Despite the challenges of working at this scale, we believe it was important to use this very valuable information and to try to address its shortcomings instead of reducing the available information by working at the genus level. To characterize the average climate and changing patterns for the Andes–Amazon area, we downloaded climatic data from TerraClimate66. We selected this product for its temporal resolution (monthly from 1958 to 2020), its spatial resolution (~4 km) and the availability of data for maximum temperature and annual precipitation. We used the ‘climateR' package67 to download the TerraClimate monthly data from 1979 to 2020 (inclusive) for each of our plot locations based on their coordinates. For each year, we used the monthly data to calculate the maximum ‘maximum temperature' (°C), the sum of annual precipitation (mm) and the seasonality of precipitation (using monthly cumulative precipitation, coefficient of variation (CV) = 100 × (standard deviation/mean))68. This way, we include any lagged effects of climate on forest dynamics, but we do not include post-census climate events that are not relevant in the database. We also calculated the change of each variable in the complete 1979–2020 time period for reference. To characterize the geography and structure of the area where each of the plots is located, we extracted elevation and landscape integrity from available datasets. Tree cover is expressed as the percentage of pixel area covered by trees in 2015 (0–100%). We calculated stem abundance change as the annual rate of proportional change in tree abundance per plot. To calculate this, we first computed the number of live individuals for each plot and census. Due to species accumulation curves, this variable is crucial in determining richness change, and, as such, it is treated as an endogenous variable in the SEM. To account for a potential change in the identification effort (for example, a large increase in individuals identified to genus level only), we calculated the change in the percentage of individuals per plot that were identified to species level in each plot (that is, the change in the percentage of identified individuals). We also included the time frame (years) between the initial and final censuses as a sampling variable. Further descriptors of each variable in each region can be found in Extended Data Fig. We explored second-order polynomial relationships for all variables and compared them with linear regressions using analysies of variance. Only temperature change (%) had a better fit using polynomial regression. Mortality rate was log transformed to better fulfil linear model assumptions. To assess the potential interference of spatial bias in the dataset, we bootstrapped each individual linear regression 100 times using random sets of 30 plots per region at each time, and we compared the direction of the slopes and their significance with those obtained from the complete dataset. Finally, we performed regression analysis with interacting climate variables and richness change. In all cases, model residuals were checked to verify the fulfilment of the linear model assumptions. This analysis evaluates the relationships for the combined dataset and each region separately to constrain coefficients with homogeneous effects across regions, leaving the remaining variables to vary freely. We excluded elevation because of the high correlation with maximum temperature (Supplementary Fig. 11) as the piecewise framework is unable to integrate correlated errors in its estimates. The SEM was estimated using three component linear models whose response variables were (1) mortality rate, (2) stem abundance change and (3) species richness change (Supplementary Fig. We tested for normality, heteroscedasticity and the variable importance factor of each of the three component models to verify that the model assumptions were met, and that the inclusion of other variables with moderate correlation (Supplementary Fig. 11) did not create multicollinearity issues (variance inflation factor <4). The structure accounted for both the direct effect of mortality rate on changes in species richness—reflecting its role as a disturbance force that opens space and provides light for recruitment—and the indirect effect of mortality through its influence on changes in stem abundance, acting as a demographic force. Mortality rate was untransformed in order to facilitate the interpretation of results. Changes in identification effort were included only as a predictor of richness change, but not of stem abundance change or mortality rate, as there is no causal connection between these variables—an observation supported by the directed separation tests (P > 0.05) automatically performed in the piecewise analyses (Supplementary Table 10). We maintained this relatively simple partitioning of indirect paths to balance intrinsic uncertainty, the number of predictor variables, and a reduced sample size per region when applying the multigroup approach. To estimate the indirect effects for each predictor and region, we multiplied the standardized path coefficients for each significant path (for example, maximum temperature → stem abundance change → richness change), considering paths via stem abundance change, mortality rate, and the longer combined path through stem abundance change and mortality rate. We computed these indirect effects only when each path coefficient was significant (P ≤ 0.05). Then, we added the indirect effects obtained by the three potential pathways and added them to estimate the total indirect effect of each predictor on richness change for each region. Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. The datasets generated and analysed within this study are owned and managed by many co-authors. Data are available from the corresponding author on reasonable request and with permission of relevant data owners. Source data are provided with this paper. Global hotspots of species richness are not congruent with endemism or threat. Feldpausch, T. R. et al. Tree height integrated into pantropical forest biomass estimates. Zemp, D. C. et al. Self-amplified Amazon forest loss due to vegetation–atmosphere feedbacks. Long-term decline of the Amazon carbon sink. Gatti, L. V. et al. Amazonia as a carbon source linked to deforestation and climate change. Boulton, C. A., Lenton, T. M. & Boers, N. Pronounced loss of Amazon rainforest resilience since the early 2000s. Tovar, C. et al. Understanding climate change impacts on biome and plant distributions in the Andes: challenges and opportunities. Flores, B. M. et al. Critical transitions in the Amazon forest system. Vulnerability of Amazonian forests to repeated droughts. Aide, T. M. et al. Woody vegetation dynamics in the tropical and subtropical Andes from 2001 to 2014: satellite image interpretation and expert validation. Feng, X. et al. How deregulation, drought and increasing fire impact Amazonian biodiversity. Observed and projected hydroclimate changes in the Andes. The responses of tropical forest species to global climate change: acclimate, adapt, migrate, or go extinct? Olivares, I., Svenning, J. C., van Bodegom, P. M. & Balslev, H. Effects of warming and drought on the vegetation and plant diversity in the Amazon Basin. Ter Steege, H. et al. Estimating the global conservation status of more than 15,000 Amazonian tree species. Gomes, V. H. F., Vieira, I. C. G., Salomão, R. P. & ter Steege, H. Amazonian tree species threatened by deforestation and climate change. Lenoir, J., Gégout, J. C., Marquet, P. A., de Ruffray, P. & Brisse, H. A significant upward shift in plant species optimum elevation during the 20th century. Widespread but heterogeneous responses of Andean forests to climate change. & Lin, C. T. Responses of alpine summit vegetation under climate change in the transition zone between subtropical and tropical humid environment. Tanner, E. V. J., Bellingham, P. J., Healey, J. R. & Feeley, K. J. Hurricane disturbance accelerated the thermophilization of a Jamaican montane forest. Colwell, R. K., Brehm, G., Cardelús, C. L., Gilman, A. C. & Longino, J. T. Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. & Silman, M. R. Land-use and climate change effects on population size and extinction risk of Andean plants. Still little evidence of poleward range shifts in the tropics, but lowland biotic attrition may be underway. Humphreys, A. M., Govaerts, R., Ficinski, S. Z., Lughadha, E. N. & Vorontsova, M. S. Global dataset shows geography and life form predict modern plant extinction and rediscovery. Malcolm, J. R., Liu, C., Neilson, R. P., Hansen, L. & Hannah, L. Global warming and extinctions of endemic species from biodiversity hotspots. Espinoza, J. C. et al. Hydroclimate of the Andes Part I: main climatic features. Duque, A., Stevenson, P. R. & Feeley, K. J. Thermophilization of adult and juvenile tree communities in the northern tropical Andes. Changes in tree functional composition across topographic gradients and through time in a tropical montane forest. Feeley, K. J., Bravo-Avila, C., Fadrique, B., Perez, T. M. & Zuleta, D. Climate-driven changes in the composition of New World plant communities. Esquivel-Muelbert, A. et al. Compositional response of Amazon forests to climate change. Rowland, L. et al. Death from drought in tropical forests is triggered by hydraulics not carbon starvation. 3 (United Nations Sustainable Development Solutions Network, 2021). Costa, J. G. et al. Forest degradation in the Southwest Brazilian Amazon: impact on tree species of economic interest and traditional use. Cuesta, F. et al. Compositional shifts of alpine plant communities across the high Andes. Gonzalez, A. et al. Estimating local biodiversity change: a critique of papers claiming no net loss of local diversity. Cardinale, B. J., Gonzalez, A., Allington, G. R. H. & Loreau, M. Is local biodiversity declining or not? The undetectability of global biodiversity trends using local species richness. Svenning, J. C. & Sandel, B. Disequilibrium vegetation dynamics under future climate change. Accelerated increase in plant species richness on mountain summits is linked to warming. Extinction debts and colonization credits of non-forest plants in the European Alps. ter Steege, H. et al. Mapping density, diversity and species-richness of the Amazon tree flora. Feeley, K. J., Davies, S. J., Perez, R., Hubbell, S. P. & Foster, R. B. Directional changes in the species composition of a tropical forest. Berenguer, E. et al. in Amazon Assessment Report 2021. 19 (United Nations Sustainable Development Solutions Network, 2021). Laurance, W. F. et al. An Amazonian rainforest and its fragments as a laboratory of global change. Pärtel, M., Szava-Kovats, R. & Zobel, M. Dark diversity: shedding light on absent species. Concerted changes in tropical forest structure and dynamics: evidence from 50 South American long-term plots. Feeley, K. J. et al. Upslope migration of Andean trees. Extinction debt of high-mountain plants under twenty-first-century climate change. Steinbauer, K., Lamprecht, A., Semenchuk, P., Winkler, M. & Pauli, H. Dieback and expansions: species-specific responses during 20 years of amplified warming in the high Alps. The drivers and impacts of Amazon forest degradation. Forest turnover rates follow global and regional patterns of productivity. The taxonomic name resolution service: an online tool for automated standardization of plant names. & Efren, P. TNRS: Taxonomic Name Resolution Service (2024). ter Steege, H. et al. Hyperdominance in the Amazonian tree flora. Tuomisto, H. et al. Discovering floristic and geoecological gradients across Amazonia. Draper, F. C. et al. Amazon tree dominance across forest strata. The Andes through time: evolution and distribution of Andean floras. & Hegewisch, K. C. TerraClimate, a high-resolution global dataset of monthly climate and climatic water balance from 1958–2015. Karger, D. N. et al. Climatologies at high resolution for the Earth's land surface areas. & Nelson, A. D. Hole-filled SRTM for the globe: version 4: data grid. Gorelick, N. et al. Google Earth Engine: planetary-scale geospatial analysis for everyone. Sheil, D. & May, R. M. Mortality and Recruitment rate evaluations in heterogeneous tropical forests. Tropical forest tree mortality, recruitment and turnover rates: calculation, interpretation and comparison when census intervals vary. Lefcheck, J. S. piecewiseSEM: Piecewise structural equation modelling in R for ecology, evolution, and systematics. Sensitivity of South American tropical forests to an extreme climate anomaly. Stropp, J., Ter Steege, H. & Malhi, Y. Disentangling regional and local tree diversity in the Amazon. Seasonal drought limits tree species across the Neotropics. Fortier, R. P. et al. Hotter temperatures reduce the diversity and alter the composition of woody plants in an Amazonian forest. A., Abatzoglou, J. T. & Mynsberge, A. R. Changes in climatic water balance drive downhill shifts in plant species' optimum elevations. ter Steege, H. et al. A spatial model of tree α-diversity and tree density for the Amazon. Tropical biodiversity loss from land-use change is severely underestimated by local-scale assessments. & Martínez-Ramos, M. Impact of forest fragmentation on understory plant species richness in Amazonia. Chronic disturbance in a tropical dry forest: disentangling direct and indirect pathways behind the loss of plant richness. Rosenblad, K. C., Baer, K. C. & Ackerly, D. D. Climate change, tree demography, and thermophilization in western US forests. This project is possible thanks to the work of the RAINFOR, Red de Bosques Andinos, Madidi Project and PPBio networks, as well as numerous individuals and institutions, including field assistants, students, botanists, tree climbers and grant holders devoted to the understanding of tropical forests. We acknowledge the vital contributions to generating the South American long-term forest record made by colleagues no longer with us, including A. Gentry, D. Neill, E. Armas, J. Singh, N. D. Cardozo, S. Patiño, T. Erwin and T. Lovejoy. We acknowledge the contributions of ForestPlots.net, a meta-network and cyber-initiative developed at the University of Leeds to develop collaborative forest science, and the ForestPlots.net Collaboration and Data Request Committee (B.S.M., E.N.H.C., O.L.P., T.R.B., B. Sonké, C. Ewango, J. Muledi, S.L.L. B.F. is currently supported by the Royal Society Dorothy Hodgkin Fellowship (DHF\R1\241021) and was previously supported by the EU Marie Curie-IF 892383 (RESCATA). is funded by a CNPq postdoctoral scholarship (153601/2024-8). is supported through the Universidad San Francisco de Quito. Fieldwork was funded by the GEF Project EcoAndes (ID4750), Fundación Futuro (UDLA001) and the Swiss Agency of Development and Cooperation, SDC (grant number 81028631). acknowledges the US National Science Foundation award number 2020424: ‘AccelNet: International Tropical Forest Science Alliance (ITFSA): a multi-network science and training initiative to accelerate understanding of the role of tropical forests in the Earth System'. J.C. acknowledges the ANR Investissement d'Avenir grants: CEBA (ANR-10-LABEX-0025) and TULIP (ANR-10-LABX-0041). thanks the Dirección General de Biodiversidad, SERNAP, Madidi National Park and local communities for their support with permits, access and collaboration in Bolivia, especially C. Maldonado, M. Cornejo, A. Araujo, J. Quisbert and N. Paniagua. Fieldwork was funded by the National Science Foundation (DEB 0101775, 0743457 and 1836353), with additional support from the Missouri Botanical Garden, National Geographic Society (NGS 7754-04, 8047-06), I-CARES at Washington University in St. Louis, Comunidad de Madrid, CSIC, Centro de Estudios de América Latina, and the Taylor and Davidson families. is supported through the University of Miami's Smathers Endowment for Tropical Trees. was funded by the European Union's Horizon 2020 Research and Innovations Programme through the CHARTER project (grant number 869471) and the Marie Sklodowska-Curie Postdoctoral Fellowship project BIPOLAR (grant agreement number 101152158), and by the NERC TundraTime project (NE/W006448/1). received financial support from a Bolsa de Produtividade em Pesquisa grant (307178/2021-8) from CNPq, and the plots were financed by PPBio (441260/2017-9) and (573721/2008-4) grants by CNPq as well as an INCT grant from FAPEAM (722069/2009). It is an output of the ForestPlots.net Research Project 102 ‘Species Responses to Climate Change in the Amazon to Andes region (RESCATA)'. The development of ForestPlots.net and data curation has been funded by several grants, including NE/B503384/1, NE/N012542/1 – ‘BIO-RED', ERC Advanced Grant 291585 – ‘T-FORCES', NE/F005806/1 – ‘AMAZONICA', NE/N004655/1 – ‘TREMOR', NERC New Investigators Awards, the Gordon and Betty Moore Foundation (‘RAINFOR', ‘MonANPeru'), ERC Starter Grant 758873 – ‘TreeMort', EU Framework 6, a Royal Society University Research Fellowship and a Leverhulme Trust Research Fellowship. We have incorporated additional acknowledgements in Supplementary Note 2. B. Fadrique, T. R. Baker, A. C. Bennett, R. Brienen, D. Galbraith, E. Gloor, A. Levesley, N. C. Pallqui-Camacho, J. Peacock, G. C. Pickavance & O. L. Phillips Instituto Nacional de Pesquisas da Amazônia, Manaus, Brazil Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI, USA Instituto de Investigación en Cambio Global (IICG-URJC), Universidad Rey Juan Carlos, Móstoles, Spain Departamento de Biología y Geología, Física y Química Inorgánica, Universidad Rey Juan Carlos, Móstoles, Spain Naturalis Biodiversity Center, Leiden, the Netherlands Grupo de Investigación en Servicios Ecosistémicos y Cambio Climático, Fundación ConVida, Medellín, Colombia UNELLEZ-Guanare, Programa de Ciencias del Agro y el Mar, Herbario Universitario (PORT), Mesa de Caracas, Venezuela Departamento de Biologia, Universidade Federal do Amazonas, Manaus, Brazil MODEMAT Foundation for Mathematical Modeling and Education, Quito, Ecuador Institute of Environment, Florida International University, Miami, FL, USA Applied Ecology and Conservation Lab, Universidade Estadual de Santa Cruz, Ilhéus, Brazil INRAE, AgroParisTech, Université de Lorraine, Nancy, France Woodwell Climate Research Center, Falmouth, MA, USA Federal University of Acre, Rio Branco, Brazil INRAE, UMR ECOFOG, Kourou, French Guiana Centre de Recherche sur la Biodiversité et l'Environnement UMR5300, Université de Toulouse, CNRS, IRD, Toulouse, France National Park Service, Fredericksburg, VA, USA Instituto de Geociências, Universidade Federal do Pará, Belém, Brazil Universidade Estadual de Roraima, Boa Vista, Brazil Facultad de Ciencias Forestales, Universidad Nacional Agraria La Molina, Lima, Peru Universidad Nacional de la Amazonia Peruana, Iquitos, Peru CIRAD, UMR EcoFoG (AgroParistech, CNRS, INRAE, Université des Antilles, Université de la Guyane), Kourou, French Guiana Department of Geography, University College London and NERC National Centre for Earth Observation, London, UK Center for Research on Biodiversity Dynamics and Climate Change (CBioClima), Institute of Biosciences, São Paulo State University (UNESP), Rio Claro, Brazil Andrew Sabin Center for Environment and Sustainability and Department of Biology, Wake Forest University, Winston-Salem, NC, USA Biology Department, University of Miami, Coral Gables, USA Museu Paraense Emílio Goeldi, Belém-Pará, Brazil CREAF, Bellaterra (Cerdanyola del Vallès), Spain Resource Management, HAWK University of Applied Sciences and Arts, Goettingen, Germany Royal Botanic Gardens, Kew, Richmond, London, UK Centro Ecológico INKAMAZONIA, Valle de Kosñipata, Cusco, Peru Asociación para la Investigación Tropical, Cusco, Peru Unique Land Use GmbH, Freiburg im Breisgau, Germany Embrapa Recursos Genéticos e Biotecnologia, Parque Estação Biológica, Prédio da Botânica e Ecologia, Brasilia, Brazil IBAM - Instituto Bem Ambiental, Belo Horizonte, Brazil Grupo MYR ESG solutions, Belo Horizonte, Brazil Instituto de Ciências Biológicas, Programa de Pós-Graduação em Ecologia, Universidade Federal do Pará, Belem, Brazil Universidad Estatal Amazónica, Facultad de Ingeniería Ambiental, Puyo, Ecuador Universidad Técnica del Norte, Ibarra, Ecuador Systematic and Evolutionary Botany Laboratory, Department of Biology, Ghent University, Gent, Belgium Instituto de Investigaciones Forestales de la Amazonía, Universidad Autónoma del Beni José Ballivián, Riberalta, Bolivia Department of Biology, Washington University, St. Louis, MO, USA Grupo de Ecosistemas Tropicales y Cambio Global, Universidad Regional Amazónica Ikiam, Tena, Ecuador Center for Biodiversity and Global Change, Yale University, New Haven, CT, USA Broward County Parks and Recreation, Oakland Park, FL, USA Biological Sciences, Florida Atlantic University, Boca Raton, FL, USA Collections, Conservation and Research, Field Museum of Natural History, Chicago, IL, USA Centro de Ciências Biológicas e da Natureza, Universidade Federal do Acre, Rio Branco, Brazil Servicios Ecosistémicos y Cambio Climático, Fundación Con Vida and Corporación COL-TREE, Medellín, Colombia Wildlife Ecology and Conservation, University of Florida, Gainesville, FL, USA Jardín Botánico de Missouri, Oxapampa, Pasco, Peru Department of Science, Chemistry Section, Institute for Nature Earth and Energy, Pontifical Catholic University of Peru, Lima, Peru Programa de Pós-graduação em Ciências Biológicas, Universidade Federal Rural da Amazônia UFRA/CAPES, Belem, Brazil Universidade Federal do Amazonas, Manaus, Brazil Instituto de Pesquisa Ambiental da Amazônia, Brasília, Brazil Universidade do Estado de Mato Grosso, Caceres, Brazil Instituto Federal de Educação, Ciência e Tecnologia do Acre, Campus Baixada do Sol, Rio Branco, Brazil Instituto de Ecología y Biodiversidad, Santiago, Chile Missouri Botanical Garden, St. Louis, MO, USA School of Science and Engineering, James Cook University, Cairns, Queensland, Australia Iwokrama International Centre for Rain Forest Conservation and Development, Georgetown, Guyana Institute of Research for Forestry Development, Universidad de los Andes, Merida, Venezuela Herbario Selva Central HOXA, Oxapampa, Peru Van Hall Larenstein University of Applied Sciences, Velp, the Netherlands Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar BF and OLP developed the concept. FCo, FCu, GA, LC, TRB, FD, AEM, HtS, MB, SB, EG, FD, PMF, JP, GD, WEM and KJF provided support on analyses and concept development in addition to contributing forest monitoring data. All other authors contributed forest monitoring data and editorial support. The authors declare no competing interests Nature Ecology & Evolution thanks Jon-Arvid Grytnes, Agustina Malizia and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Interaction plots showing the results for the regression of the interaction between a) Precipitation seasonality and Precipitation seasonality change, b) annual precipitation and precipitation seasonality, c) precipitation seasonality change and annual precipitation, and d) temperature change and maximum temperature (regression results in Table SI3) for the whole dataset (n = 406). Line and point colours indicate the second predictor categorised into three groups (mean, +1 SD, −1 SD). Error bars represent the most extreme data points, which are no more than 1.5 times quantiles 1 and 3 of the data (represented by the box limits). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Fadrique, B., Costa, F., Cuesta, F. et al. Tree diversity is changing across tropical Andean and Amazonian forests in response to global change. Anyone you share the following link with will be able to read this content: Sorry, a shareable link is not currently available for this article. Provided by the Springer Nature SharedIt content-sharing initiative Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript. Nature Human Behaviour (2026)Cite this article Our experiences contain countless details that may be important in the future, yet we rarely know which will matter and which will not. This uncertainty poses a difficult challenge for adaptive decision-making, as failing to preserve relevant information can prevent us from making good choices later on. One solution to this challenge is to store detailed memories of individual experiences that can be flexibly accessed whenever their details become relevant. By allowing us to store and recall specific events in vivid detail, the human episodic memory system provides exactly this capacity. Yet, whether and how this ability supports adaptive behaviour is poorly understood. Here we aimed to determine whether people use detailed episodic memories to make decisions when future task demands are uncertain. We hypothesized that the episodic memory system's ability to store events in great detail allows us to reference any of these details if they later become relevant. We tested this hypothesis using a novel decision-making task in which participants encoded individual events with multiple features and later made decisions based on these features to maximize their earnings. Across 5 experiments (total n = 535), we found that participants referenced episodic memories during decisions in feature-rich environments and that they did so specifically when it was unclear at encoding which features would be needed in the future. Overall, these findings reveal a fundamental adaptive function of episodic memory, showing how its rich representational capacity enables flexible decision-making under uncertainty. This is a preview of subscription content, access via your institution Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription cancel any time Subscribe to this journal Receive 12 digital issues and online access to articles only $9.92 per issue Buy this article Prices may be subject to local taxes which are calculated during checkout The data are available via GitHub at https://github.com/jonathanicholas/nm2025_emdm. The code is available via GitHub at https://github.com/jonathanicholas/nm2025_emdm. Tulving, E. in Organization of Memory (eds Tulving, E. & Donaldson, W.) 381–403 (Academic Press, 1972). Eichenbaum, H. The hippocampus and mechanisms of declarative memory. Google Scholar Bartlett, F. C. Remembering: A Study in Experimental and Social Psychology (Cambridge Univ. Anderson, J. R. & Milson, R. Human memory: an adaptive perspective. Google Scholar Schacter, D. L., Guerin, S. A. & Jacques, P. L. S. Memory distortion: an adaptive perspective. Trends Cogn. Google Scholar Shohamy, D. & Adcock, R. A. Dopamine and adaptive memory. Trends Cogn. Google Scholar Lengyel, M. & Dayan, P. Hippocampal contributions to control: the third way. Advances in Neural Information Processing Systems (eds Platt, J. C. et al.) 889–896 (Curran Associates, 2008). Gershman, S. J. & Daw, N. D. Reinforcement learning and episodic memory in humans and animals: an integrative framework. Google Scholar Nagy, D. G., Orbán, G. & Wu, C. M. Adaptive compression as a unifying framework for episodic and semantic memory. Google Scholar Rescorla, R. A. & Wagner, A. R. in Classical Conditioning II: Current Research and Theory (eds Black, A. H. & Prokasy, W. F.) 64–99 (Appleton-Century-Crofts, 1972). Shohamy, D. & Daw, N. D. in The Cognitive Neurosciences 5th edn (eds Gazzaniga, M. S. & Mangun, G. R.) 591–603 (Boston Review, 2014). Knowlton, B. J., Mangels, J. & Squire, L. R. A neostriatal habit learning system in humans. Google Scholar Schultz, W., Dayan, P. & Montague, P. R. A neural substrate of prediction and reward. Google Scholar Bayer, H. M. & Glimcher, P. W. Midbrain dopamine neurons encode a quantitative reward prediction error signal. Google Scholar Shohamy, D., Myers, C. E., Grossman, S., Sage, J. & Gluck, M. A. The role of dopamine in cognitive sequence learning: evidence from Parkinson's disease. Google Scholar Zaghloul, K. A. et al. Human substantia nigra neurons encode unexpected financial rewards. Google Scholar Sutton, R. S. & Barto, A. G. Reinforcement Learning: An Introduction 2nd edn (MIT Press, 2018). Wilson, R. C. & Niv, Y. Inferring relevance in a changing world. Google Scholar Leong, Y. C., Radulescu, A., Daniel, R., DeWoskin, V. & Niv, Y. Dynamic interaction between reinforcement learning and attention in multidimensional environments. Google Scholar Niv, Y. et al. Reinforcement learning in multidimensional environments relies on attention mechanisms. Google Scholar Miller, G. A. The magical number seven, plus or minus two: some limits on our capacity for processing information. Google Scholar Bornstein, A. M., Khaw, M. W., Shohamy, D. & Daw, N. D. Reminders of past choices bias decisions for reward in humans. Google Scholar Duncan, K., Semmler, A. & Shohamy, D. Modulating the use of multiple memory systems in value-based decisions with contextual novelty. Google Scholar Mason, A., Madan, C. R., Simonsen, N., Spetch, M. L. & Ludvig, E. A. Biased confabulation in risky choice. Nicholas, J., Daw, N. D. & Shohamy, D. Uncertainty alters the balance between incremental learning and episodic memory. Google Scholar Montaser-Kouhsari, L., Nicholas, J., Gerraty, R. T. & Shohamy, D. Differentiating reinforcement learning and episodic memory in value-based decisions in Parkinson's disease. Google Scholar Plonsky, O., Teodorescu, K. & Erev, I. Reliance on small samples, the wavy recency effect, and similarity-based learning. Google Scholar Daw, N. D. & Dayan, P. The algorithmic anatomy of model-based evaluation. Google Scholar Kool, W., Cushman, F. A. & Gershman, S. J. When does model-based control pay off? PLoS Comput. Google Scholar Foundations of Human Memory (Oxford Univ. Ratcliff, R. A theory of memory retrieval. Sederberg, P. B., Howard, M. W. & Kahana, M. J. A context-based theory of recency and contiguity in free recall. O'Reilly, R. C. & McClelland, J. L. Hippocampal conjunctive encoding, storage, and recall: avoiding a trade-off. O'Reilly, R. C. & Rudy, J. W. Conjunctive representations in learning and memory: principles of cortical and hippocampal function. Rudy, J. W. & O'Reilly, R. C. Conjunctive representations, the hippocampus, and contextual fear conditioning. Lee, S. W., O'Doherty, J. P. & Shimojo, S. Neural computations mediating one-shot learning in the human brain. PLoS Biol. Packard, M. G. & McGaugh, J. L. Inactivation of hippocampus or caudate nucleus with lidocaine differentially affects expression of place and response learning. & Packard, M. G. Competition among multiple memory systems: converging evidence from animal and human brain studies. Stewart, N., Chater, N. & Brown, G. D. Decision by sampling. Biele, G., Erev, I. & Ert, E. Learning, risk attitude and hot stoves in restless bandit problems. Shadlen, M. N. & Shohamy, D. Decision making and sequential sampling from memory. Wang, S., Feng, S. F. & Bornstein, A. M. Mixing memory and desire: how memory reactivation supports deliberative decision-making. WIREs Cogn. Hertwig, R. & Erev, I. The description–experience gap in risky choice. Trends Cogn. Schlichting, M. L. & Preston, A. R. in The Hippocampus from Cells to Systems: Structure, Connectivity, and Functional Contributions to Memory and Flexible Cognition (eds Hannula, D. E. & Duff, M. C.) 405–437 (Springer, 2017). Whittington, J. C. et al. The Tolman-Eichenbaum machine: unifying space and relational memory through generalization in the hippocampal formation. Biderman, N., Bakkour, A. & Shohamy, D. What are memories for? The hippocampus bridges past experience with future decisions. Trends Cogn. Squire, L. R. & Zola, S. M. Structure and function of declarative and nondeclarative memory systems. Poldrack, R. A. et al. Interactive memory systems in the human brain. Daw, N. D., Niv, Y. & Dayan, P. Uncertainty-based competition between prefrontal and dorsolateral striatal systems for behavioral control. Lee, S. W., Shimojo, S. & O'Doherty, J. P. Neural computations underlying arbitration between model-based and model-free learning. Greve, A., Cooper, E., Tibon, R. & Henson, R. N. Knowledge is power: prior knowledge aids memory for both congruent and incongruent events, but in different ways. van Kesteren, M. T. R., Ruiter, D. J., Fernández, G. & Henson, R. N. How schema and novelty augment memory formation. Trends Neurosci. Rouhani, N. & Niv, Y. Signed and unsigned reward prediction errors dynamically enhance learning and memory. Jang, A. I., Nassar, M. R., Dillon, D. G. & Frank, M. J. Positive reward prediction errors during decision-making strengthen memory encoding. Howard, M. W. & Kahana, M. J. A distributed representation of temporal context. Polyn, S. M., Norman, K. A. A context maintenance and retrieval model of organizational processes in free recall. Kahana, M. J. Computational models of memory search. Bornstein, A. M. & Norman, K. A. Reinstated episodic context guides sampling-based decisions for reward. Aka, A. & Bhatia, S. What I like is what I remember: memory modulation and preferential choice. Michelmann, S., Staresina, B. P., Bowman, H. & Hanslmayr, S. Speed of time-compressed forward replay flexibly changes in human episodic memory. Wimmer, G. E., Liu, Y., Vehar, N., Behrens, T. E. J. & Dolan, R. J. Episodic memory retrieval success is associated with rapid replay of episode content. Liu, Y., Mattar, M. G., Behrens, T. E. J., Daw, N. D. & Dolan, R. J. Experience replay is associated with efficient nonlocal learning. McFadyen, J., Liu, Y. Differential replay of reward and punishment paths predicts approach and avoidance. Wimmer, G. E., Liu, Y., McNamee, D. C. & Dolan, R. J. Distinct replay signatures for prospective decision-making and memory preservation. Carmichael, L., Hogan, H. P. & Walter, A. An experimental study of the effect of language on the reproduction of visually perceived form. Nagy, D. G., Török, B. & Orbán, G. Semantic compression of episodic memories. Preprint at https://arxiv.org/abs/1806.07990 (2018). Zhou, D. et al. A compressed code for memory discrimination. Preprint at bioRxiv https://doi.org/10.1101/2025.10.12.681901 (2025). Kerrén, C., Reznik, D., Doeller, C. F. & Griffiths, B. J. Exploring the role of dimensionality transformation in episodic memory. Trends Cogn. Nicholas, J., Daw, N. D. & Shohamy, D. Proactive and reactive construction of memory-based preferences. Hoffman, M. D. & Gelman, A. The No-U-Turn sampler: adaptively setting path lengths in Hamiltonian Monte Carlo. Bürkner, P.-C. Brms: an R package for Bayesian multilevel models using Stan. Vehtari, A., Gelman, A. Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Sivula, T., Magnusson, M., Matamoros, A. & Vehtari, A. Uncertainty in Bayesian leave-one-out cross-validation based model comparison. Preprint at https://arxiv.org/abs/2008.10296 (2023). We thank members of the Mattar Lab as well as K. Jensen, Q. Lu, N. Biderman and D. Shohamy for their helpful comments and feedback on the project. This work was supported by the National Science Foundation SBE Postdoctoral Research Fellowship under grant no. Department of Psychology, New York University, New York, NY, USA Jonathan Nicholas & Marcelo G. Mattar Search author on:PubMed Google Scholar Search author on:PubMed Google Scholar conceived the research and study design. conducted the data collection and analysis and drafted the paper. reviewed and edited the paper. Correspondence to Jonathan Nicholas. The authors declare no competing interests. Nature Human Behaviour thanks Silvy H. P. Collin, Richard Henson and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Supplementary Tables 1–9 and Figs. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. Reprints and permissions Nicholas, J., Mattar, M.G. Episodic memory facilitates flexible decision-making via access to detailed events. Nat Hum Behav (2026). Accepted: 20 November 2025 Version of record: 23 January 2026 DOI: https://doi.org/10.1038/s41562-025-02383-3 Anyone you share the following link with will be able to read this content: Sorry, a shareable link is not currently available for this article. Provided by the Springer Nature SharedIt content-sharing initiative Nature Human Behaviour (Nat Hum Behav) ISSN 2397-3374 (online) © 2026 Springer Nature Limited Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Guinea-Bissau will implement a universal birth-dose policy for the Hepatitis B vaccine in 2027.Credit: Enrique Lopez-Tapia/Nature Picture Library/Alamy Public health authorities in Guinea-Bissau say that they have suspended a controversial US-funded hepatitis B vaccine study that has raised questions about who has authority over clinical research conducted in Africa. At a press conference held today, officials from Guinea-Bissau's ministry of health said that the study was being suspended pending a technical and ethical review by its national public health institute. Hotly anticipated US vaccine meeting ends with confusion — and a few decisions The continuing row shines a spotlight on long-standing tensions for clinical research trials in Africa. African scientists say that the Guinea-Bissau study shows how political pressure, funding interests and fragmented oversight can push local health priorities aside. But several scientists have argued that by randomizing newborns to not receive the vaccine, the trial denies a safe and life-saving intervention to infants in Guinea-Bissau, which has a hepatitis B prevalence of about 19%. The immune systems of newborns are immature and about 90% of people infected at birth go on to develop chronic, lifelong infections which may lead to liver disease and early death. “They're trying to use African children to prove a case for reducing vaccines in the US,” says Seye Abimbola, a professor of health systems at the University of Sydney in Australia, who researches decolonizing global health. Bandim Health did not respond to Nature's request for comments on the study design. Hepatitis B vaccine guidance set to be rolled back for US babies: what the science says RFK Jr demanded a vaccine study be retracted — the journal said no Hotly anticipated US vaccine meeting ends with confusion — and a few decisions Volunteer scientists work ‘nights and weekends' to guide vaccine advice in US Exclusive: key NIH review panels due to lose all members by the end of 2026 Exclusive: key NIH review panels due to lose all members by the end of 2026 Hepatitis B vaccine guidance set to be rolled back for US babies: what the science says An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday. Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.