Introduction

Continental arcs, where oceanic lithosphere is subducted beneath continental lithosphere, are thought to release more CO2 than other types of volcanoes due to decarbonation of sedimentary carbonates stored in the upper continental plate1,2. As such, the rise and fall of continental arcs pace with Earth’s greenhouse and icehouse intervals since at least the Cryogenian2, strengthening their role in modulating atmospheric CO2 levels and climate change. Elevated arc-related CO2 degassing and associated uplift of continental arcs would then promote silicate weathering and the drawdown of atmospheric CO2, stabilizing the climate system on multimillion-year time scales3,4,5.

During the Ediacaran–Cambrian transition (E–C), subduction along the Proto-Pacific and Proto-Tethyan margin of Gondwana became active following the final breakup of the Rodinian supercontinent and the initial formation of Gondwana6,7. An increased continental arc length and fraction of young zircon grains indicate enhanced continental arc volcanism during the E–C transition2,8,9, consistent with an increase in atmospheric CO2 and climatic warming2,10. At this time, the rapid rise in animal diversity and abundance (the “Cambrian Explosion”, ca. 540–515 Ma) also occurred11,12. To sustain complex life, nutrient availability (e.g. P, Ca) and oxygen are considered as two important limiting factors13,14,15,16, both of which could be affected by enhanced weathering of mountains linked to collision-dominated tectonics during the assembly of Gondwana17,18. However, a large part of the Gondwana assembly-related collisions (mainly before 610 Ma)7 occurred too early to be responsible for the Cambrian Explosion. In contrast, increased subduction-dominated continental arc volcanism along the margin of Gondwana temporally overlapped with the Cambrian Explosion2,8,9. Fast-eroding features of continental arcs, characterized by crustal thickening and intense mantle involvement, may fuel a more efficient feedback loop via CO2 degassing and climatic warming, enhanced weathering of mantle-derived juvenile rocks, and hence the rapid delivery of nutrients to the oceans3,8,19. Along with this feedback process, removal of long-term subduction-related CO2 would result in increased oxygen levels through efficient organic carbon burial that impedes O2 back-reactions10,20. Juvenile rocks from large igneous provinces (LIPs) or arc volcanism typically contain 1.5–2 times higher phosphorus content than continental crust21,22, and are highly susceptible to weathering23. Their rapid weathering would release large amounts of P into ocean, favoring primary productivity, efficient organic burial and O2 production21,22. However, it is not clear whether subduction-driven climatic warming accompanied by enhanced weathering of arc-related juvenile rocks may have played a significant role in the Cambrian biogeochemical cycle and its animal radiation. Additional geochemical evidence is thus required to explore their potential relationship.

Given the fast-eroding features of continental arcs developed upon either a mantle-involved unradiogenic or an ancient continental radiogenic basement, they play an important role in controlling the strontium (Sr) isotopic composition (87Sr/86Sr) of seawater through chemical weathering over geological times4,19. Like Sr, the Os isotope compositions (187Os/188Os) of seawater reflect a balance between two highly distinctive end-member fluxes: unradiogenic sources of 187Os/188Os ~ 0.13 (e.g. extraterrestrial flux, hydrothermal flux and mantle-derived juvenile material weathering) and radiogenic sources of 187Os/188Os ~ 1.4 (e.g. weathering of continental crust)24,25. Thus, variations in seawater 187Os/188Os over geological times can be applied to trace changes in the weathering of juvenile rocks and continental crust24. Besides the Os-Sr isotopes used to trace weathering sources, seawater Li isotope records could provide unique information on silicate weathering rates and regimes in response to climatic changes26,27. δ7Li values of modern rivers vary significantly (3–42‰; average 23‰)27 and depend on the silicate weathering intensity denoted as W/D, which is the ratio of the chemical weathering (W) rate to the total denudation rate (D, consisting of the erosion (E) + chemical weathering (W) rates)27. Low W/D regimes are characterized by high primary rock dissolution, known as congruent weathering with a low dissolved δ7Li and high Li flux, while moderate W/D regimes are linked to secondary mineral (largely clay) formation driving high dissolved δ7Li and a lower Li flux, known as incongruent weathering27,28. In high W/D regimes, there is little rock dissolution and pre-formed secondary clays redissolve, resulting in low dissolved δ7Li and a very low Li flux27,28. The riverine Li makes up over 50% of the modern ocean Li input, with the remainder from mid-ocean ridge hydrothermal inputs and balanced by Li sinks into marine authigenic clays and low-temperature alteration of the oceanic crust28,29. Given the fairly long residence time of Li-Os-Sr isotopes in the modern ocean, they can provide insights into the relationship between weathering dynamics and major paleo-environmental fluctuations25,26,30.

Here, we measured Li isotopes in carbonates and Re-Os isotopes of black shales and cherts that potentially record seawater isotope compositions from two drillcores covering the main interval of the Cambrian Explosion (ca. 560–515 Ma) in the Yangtze Block, South China. With compiled Sr isotope ratios, our work provides critical evidence of subduction-driven weathering dynamics and discusses their role in seawater chemistry and the Cambrian Explosion.

Results

The E-C transition witnessed widespread continental arc volcanism along the Proto-Pacific and Proto-Tethyan margin of Gondwana7 (Fig. 1a), such as the Avalonia–Cadomian arcs (750–500 Ma)31, the Ross and Delamerian arcs (590–480 Ma)6,32, and Proto-Tethyan arc (550–490 Ma)33. Some newly generated Large Igneous Provinces (LIPs) were also emplaced around the Central Iapetus on the Laurentian and Baltican cratons (ca. 535–533 Ma)34. The South China, consisting of the Yangtze Block and the Cathaysia Block, was surrounded by an open ocean during the E–C transition period7.

Fig. 1: Tectonic, paleogeographic and stratigraphic correlation during the E-C transition.
figure 1

a Reconstructed global plate tectonics at ca. 580–530 Ma, modified after ref. 7. The distributions of continental arcs were compiled from ref. 9 and their corresponding durations were summarized in Supplementary Table S1. Records of Large Igneous Provinces (LIPs) were adopted from ref. 34. SC South China, NC North China, Ta Tarim, AC Avalonia–Cadomian, Dem Delamerian, GTM Gondwana Proto-Tethyan margin, Ros Ross, Sal Saldanian, Tim Timanide, Pam Pampeanas. b Locations of study sections, phosphorite distribution and paleogeographic map of the Yangtze Block during the E-C transition (adapted after ref. 42). c Stratigraphic and biostratigraphic correlation across the Yangtze Block from shallow shelf to slope-basin regions, including 1: Organic carbon isotope and bio-fossil records of Xiaotan section38,65; 2: ZK4803, this study; 3: Dingtai section (ref. 38; 4: Organic carbon isotope of Ganziping section37; 5: ZK4411, this study. Organic carbon isotopes are from ref. 66; 6: Organic carbon isotopes of Longbizui section67; 7: Bahuang section37. U–Pb ages for the Xiaotan section are from the nearby Meishucun section41; Re-Os ages for the Ni-Mo ore layer and the V-rich layer are from ref. 39 and ref. 40, respectively; U–Pb ages for the Ganziping and Bahuang sections are from ref. 37; U–Pb age for the top of Doushantuo Member IV is from ref. 35. Locations of these typical sections are shown in (b).

Full size image

Based on biostratigraphy, radiometric ages (U-Pb and Re-Os), carbon isotope stratigraphy, and marker layers, we have reconstructed a regional stratigraphic framework across the shallow carbonate platform, transitional belt and deepwater slope-basin regions of the Yangtze Block (Fig. 1b, c). The regional black shale (~10 m) of the Doushantuo Member IV is constrained by a U-Pb age of 551.1 ± 0.6 Ma from an ash bed at its top35. Above this layer, the E–C boundary is clearly defined by the globally recognizable basal Cambrian negative carbon isotopic excursions (BACE, N1), and the appearance of small shelly fossils (SSFs) in the shallow shelf Meishucun and Xiaotan sections36 (Fig. 1c). The Liuchapo Formation slope-basin region straddles the E–C boundary, supported by two U–Pb ages of 542.1 ± 5.0 Ma and 542.6 ± 3.7 Ma from ash beds in the Ganziping and Bahuang sections37 (Fig. 1c), corresponding to the boundary of the Dengying/Yanjiahe Formations in shallow sections. Widespread black shales over the whole Yangtze Block act as a regional marker layer38, constrained by Re-Os ages of 521 ± 5 Ma and 520.3 ± 9.1 Ma for Ni-Mo-V ore layers at the bottom39,40. Thus, the depositional age of the Yanjiahe Formation in the shallow shelf is constrained at ca. 540–521 Ma, time-equivalent to the middle-upper Liuchapo Formation in deep water. Phosphorites are widely distributed in the carbonate platform, with a depositional age of ca. 541–525 Ma based on the U-Pb age of 526.5 ± 1.1 Ma above the phosphorite deposits41,42. Trilobites and more complex, large-bodied animals (e.g., Chengjiang and Qingjiang Biota) also first appeared in the shelf sections at ca. 521–518 Ma38,43 (Fig. 1c).

Core samples covering late Ediacaran to early Cambrian (ca. 560–521 Ma) were collected from carbonates of the Dengying and Yanjiahe Formations in the shallow shelf of drillcore ZK4803 (31°28′22“N and 111°11′25“E) and black cherts and shales of the Doushantuo Member IV and Liuchapo Formations in the slope region of drillcore ZK4411 (28°02′32“N and 109°04′33“E) (Fig. 2a, b). We measured and calculated Re-Os isochron ages of 557.3 ± 9.1 Ma (2σ, n = 9, MSWD = 1.6) for the Doushantuo member IV black shales with a high initial 187Os/188Os ratio (Osi) of 0.825 ± 0.086, and 529.1 ± 6.0 Ma (2σ, n = 8, MSWD = 2.6) for the Liuchapo cherts with evidently low Osi value of 0.341 ± 0.014 (Fig. 2b, c). The continuous Osi records of 560–523 Ma is presented based on their Re-Os data and estimated depositional ages. In general, Osi values stay at a high level with an average of 0.82 at 557 Ma, and 0.62–0.68 at ca. 541–540 Ma, and then rapidly decrease to ~0.34 at ca. 540–525 Ma. After this, they return to high values of 0.76–0.94 at ca. 525–523 Ma (Fig. 2b and Supplementary Data 1). Interestingly, δ7Licarb values of leached carbonate samples show the same varying trends as Osi, i.e., a general decreasing trend of δ7Licarb from ~18‰ to ~14‰ at ca. 550–540 Ma, and further to the lowest value of 6.3‰ at ca. 540–525 Ma and then back to a high value of ~17‰ at ca. 525–515 Ma (Fig. 2a and Supplementary Data 2).

Fig. 2: Geochemical profiles of ZK4803 and ZK4411 drillcores.
figure 2

a Li isotope data for ZK4803 carbonate samples. b Organic carbon isotope, Eu/Eu*, and initial 187Os/188Os isotope data for ZK4411 core samples. Organic carbon isotope data are from ref. 66. c Re-Os isochron ages for the Liuchapo and Doushantuo Member IV Formations.

Full size image

Following the typical diagenetic characteristics of modern marine carbonates44,45, additional elemental ratios (e.g., Al/(Mg+Ca), Li/(Mg+Ca) and Sr/(Mg+Ca)) and their relationships with δ7Li are used to evaluate the possible overprint of δ7Li values of ancient carbonates by detrital contamination during carbonate leaching, carbonate mineralogy, and diagenetic alterations. These observations suggest that carbonates in our drillcore are most likely formed during marine fluid-buffered diagenesis (Supplementary Figs. S1 and S2), and their δ7Li values are considered to approach the ambient seawater Li isotopic signals with a minor Li isotope fractionation of –0.5‰ to –2‰44,45. More details are given in Supplementary Information.

Discussion

Congruent weathering of subduction-driven thickened crust at ca. 550–540 Ma

The relationship of tectonic, phosphorite, geochemical records, and biodiversity patterns at ca. 560–510 Ma is compiled in Fig. 3. In particular, to avoid ambiguous interpretations of a single isotope due to local geological processes, the Li-Os-Sr isotope system is adopted to provide more robust evidence on changes in weathering dynamics spanning ca. 560–510 Ma. The compiled Osi values for late Ediacaran sediments (ca. 560–540 Ma), similar to those of modern seawater (187Os/188Os ~ 1.06)24, suggest a radiogenic Os input from weathering of continental crust (Fig. 3a). A significant increase in seawater 87Sr/86Sr at ca. 550–540 Ma also indicates a transient radiogenic Sr flux of continental crust weathering (Fig. 3b and Supplementary Data 3). This sharp excursion in seawater 87Sr/86Sr cannot be solely explained by the secular trends related to enhanced erosion of earlier collisional orogens during the Gondwana assembly before 600 Ma18, indicating an occurrence of distinct geodynamic events. Based on compiled timing and average arc length data of continental arcs along the margin of Gondwana (e.g., ~6750 km of Avalonia-Cadomian arc, ~5, 775 km of Gondwana proto-Tethyan margin arc), the total arc length increased from ~11, 500 km in pre-550 Ma to 21, 100 km in post-550 Ma (Supplementary Table S1 and Fig. 3c). It is illustrated that both the expansion of active subduction zones along Gondwana’s margin and the increase in continental arc length after ~550 Ma (Fig. 3c) coincide with rising seawater 87Sr/86Sr. Oceanic lithosphere subduction beneath continental lithosphere and associated crustal thickening, topographic uplift, climatic warming and erosion3 thus most likely caused the transient increase in seawater 87Sr/86Sr. The negative shift of δ7Lisw from ~18‰ to ~15‰ at ca. 550–540 Ma (Fig. 3d) may be also a response to the erosion of subduction-driven thickened crust, because exposure of uplifted fresh rocks and rapid runoff promote congruent weathering with a low riverine δ7Li and high Li flux46. Based on a dynamic Li box model and mass-balance equations (see model description in Supplementary Information), a 3‰ shift requires both a decrease in riverine δ7Li to 7.7‰ and an associated 1.5×river flux, assuming an isotopic fractionation between seawater and Li sinks (Δsink = 8‰) based on the Δsink = 5–10‰ for Precambrian oceans29. Therefore, the radiogenic Os–Sr and negative δ7Lisw shifts at ca. 550–540 Ma is most likely caused by congruent weathering of subduction-driven uplifted crust.

Fig. 3: Relationship of tectonic, phosphorite, geochemical records (δ13Ccarb-Os-Sr-Li isotopes), and biodiversity patterns at ca.
figure 3

560–510 Ma. a Compiled Os isotope records from this study and previous work39,40,59,68. The orange dashed line indicates an interval without Os isotope data (ca. 550–540 Ma), and the solid line shows the average Os isotope trend. Endmember values (black dashed lines) are from ref. 24. b Compiled Sr isotope records. Data sources: South China69; Mongolia70; Morocco30; Kotuikan River71; Lena River72; Spain73. The dark pink line shows the LOWESS regression for seawater Sr isotope composition. c Continental arc length (Summarized in Supplementary Table S1), subduction length10, LIPs34 and arc magmatism records including Avalonia-Cadomian (AC) arcs (550–525 Ma)31,50, Gondwana proto-Tethyan margin (GTM; ca. 547–522 Ma)33, Ross-Delamerian arcs (Ros-Del; ca. 540–520 Ma)32,51, Pampeanas arc (Pam; ca. 545–520 Ma)52, and Saldanian arc (Sal; ca. 537–533 Ma)53. d Li isotope records of carbonate samples from drill core ZK4803. The purple line shows the best-estimated curve using Locally Weighted Scatterplot Smoothing (LOWESS), bounded by a 95% confidence interval (grey field). e Estimated tonnes of P2O5 from global-scale phosphorite deposits13 and evolution of seawater Ca2+ concentration14,58. The two seawater Ca2+ data points and the estimated trend (dashed line) are from refs. 14,58. f Number of phyla and classes globally12. g Carbonate δ13Ccarb records36. BACE, ZHUCE, and SHICE represent Basal Cambrian, ZHUjiaqing, and SHIyantou Carbon isotope Excursions, respectively.

Full size image

Enhanced erosion of uplifted arc-related juvenile rocks at ca. 540–525 Ma

The strikingly negative shifts of Li-Os-Sr isotopes during ca. 540–525 Ma (Fig. 3a, b, d) reflect a unique weathering regime at the onset of the Phanerozoic. The negative shifts of Os-Sr isotopes suggest a significant unradiogenic Os-Sr source from either weathering of juvenile rocks or hydrothermal input. Most of our Eu/Eu* values are around 1 (Fig. 2b) and the Post-Archaean Average Australian Shale (PAAS)-normalized rare-earth element (REE) patterns are distinct from typical hydrothermal fluids that show positive Eu anomaly (Supplementary Data 4 and Fig. S3), suggesting limited hydrothermal activity in our study area. Seawater Sr isotope is highly sensitive to hydrothermal activity and can be further used to evaluate hydrothermal activity26. Following a dynamic box model (Fig. 4 and Supplementary Tables S2–S4), a 240% increase in hydrothermal flux (Fhyd) would be required to produce the observed negative Sr isotope shift. However, such a hydrothermal input fails to produce the observed negative shift of Osi (Fig. 4a). In contrast, as shown in Fig. 4a, b, a 375% increase in weathering flux, in the case of juvenile rocks-dominated weathering (Fjuv), can well explain the negative shifts of both Os and Sr at ca. 540–525 Ma, considering different residence time relative to that of the modern ocean (e.g., 1×present and 0.5 × present). Therefore, the unradiogenic Os–Sr shifts at ca. 540–525 Ma is most likely caused by the weathering of mantle-derived juvenile rocks.

Fig. 4: Box model results for Li-Os-Sr isotopic shifts observed at ca. 540–515 Ma.
figure 4

a The response of Osi to juvenile rocks weathering and hydrothermal flux. b The response of Sr isotope to juvenile rocks weathering and hydrothermal flux. c The effect of hydrothermal and riverine Li flux as well as δ7Liriv on seawater Li isotope. d Estimated bioavailable P flux in a 375% increase in Fjuv. Fjuv: Li-Os-Sr flux of juvenile rocks weathering; Fhyd: hydrothermal Li-Os-Sr flux. Rriv: riverine Li isotopic composition; Friv: riverine Li flux; τres: seawater residence time; Δsink: isotopic fractionation factor between Li sinks and seawater. More details are given in Supplementary Information.

Full size image

The continuous decline of δ7Lisw from ~15‰ to ~10‰ at ca. 540–525 Ma (Fig. 3d) requires a prolonged low δ7Li flux from either riverine or hydrothermal inputs. A 240% increase in Fhyd, as indicated by Sr isotope ratios, only cause a minor negative δ7Lisw shift of 1.5‰, further eliminating the possibility of hydrothermal flux (Fig. 4c). With different scenarios of Δsink = 5‰ or 8‰ and residence time (1×present and 0.5 × present), a ~ 5‰ negative shift of δ7Lisw requires both 2× increase in riverine flux (Friv) and significantly low riverine δ7Li (δ7Liriv) of 3–5‰ as compared to the avg. 23‰ of modern rivers (Fig. 4c), and this is characteristic of congruent weathering with high rock dissolution and erosion rate (i.e., low W/D)27. Mineral dissolution rates decrease sequentially with olivine > Ca-plagioclase > pyroxene > Na-plagioclase > K-feldspar > muscovite > quartz47,48. Consequently, juvenile rocks—primarily composed of pyroxene, plagioclase, olivine, and volcanic glass—weather rapidly and highly congruently relative to old felsic crust23. This behavior reflects their geochemical freshness and aluminum-deficient composition, which suppress secondary mineral precipitation (e.g., clays)48. Based on the dominated weathering of mantle-derived juvenile rocks as suggested by Os–Sr isotopes, the decreasing δ7Lisw may further indicate an enhanced erosion of these juvenile rocks.

Enhanced erosion of LIPs flood basalt is expected to yield coupled negative shifts of Li–Os–Sr isotopes, which have been observed in the Permian-Triassic transition48 and Oceanic Anoxic Event 226. However, LIPs during this interval were restricted to the Central Iapetus, covering only ~0.18 × 10⁶ km² and lasting ~2 Myr (535–533 Ma)34 (Fig. 3f), making them unlikely to explain a ~ 15 Myr negative shift in seawater Li–Os–Sr isotopes. In contrast, subduction zones and continental arcs were globally distributed with an average arc length over 20,000 km (Fig. 3c and Supplementary Table S1) and lasted for over tens of millions of years during the E–C transition9. Continental arcs could become isotopically-juvenile by tectonic processes such as oceanic terrane accretion, upper plate extension, slab breakoff, and ridge subduction19, which is believed to have exerted a strong control on unradiogenic Sr weathering flux and low seawater 87Sr/86Sr of mid-Cretaceous (125–85 Ma)19. It is inferred that Ediacaran-Cambrian continental arcs perhaps shared similar characteristics. First, Gondwana-margin arcs constituted external orogenic systems (e.g., Circum-Pacific), where ongoing arc magmatism could develop isotopically juvenile signatures49. Second, as shown in Fig. 3c, magmatic records from disparate continental arcs broadly cluster around 540–525 Ma. For example, arc magmatism activities were recorded in Avalonia-Cadomian arcs (avg. 6750 km) at ca. 550–525 Ma31,50, Gondwana proto-Tethyan margin arc (avg. 5775 km) at ca. 547–522 Ma33, Ross-Delamerian arcs (avg. 7300 km) at ca. 540–520 Ma32,51, Pampeanas arc (avg. 700 km) at ca. 545–520 Ma52, and Saldanian arc (avg. 1300 km) at ca. 537–533Ma53. Their emplacements during upper-plate extension could likewise generate isotopic juvenility. Tectonically uplifted fresh rocks in a warmer climate promote rapid mineral dissolution, enhancing runoff with low riverine δ⁷Li and high Li fluxes46. Intense continental arc volcanism satisfies these conditions through: (i) more efficient CO2 degassing driving climatic warming2, and (ii) subduction-driven uplift of highly weatherable juvenile arc rocks, both of which facilitate high erosion rates of isotopically juvenile rocks to produce our observed negative Li-Os-Sr isotopic shifts at ca. 540–525 Ma.

Negative Li-Os-Sr isotopic shifts are also overlapped with the peak of global-scale phosphorite deposits13 (Fig. 3e), suggesting a link beween enhanced erosion of arc-related juvenile rocks and enhanced P flux. Assuming a continental area of 9 × 107 km2, an average chemical weathering rate of 2 t/km2/yr for granitic crust and an average granitic P concentration of 650 ppm, the total background reactive P flux would be 1.5 × 109 mol P/yr, if 40% weathered P becomes bioavailable21. The continental arc area is estimated to be 6 × 106 km2 (~6.7% of continental area) based on the arc length of over 20,000 km and a mean width of 300 km (50–600 km of arc system)54. Mantle-derived rocks generally have P contents 1.5–2 times higher than granitic crust21,22. Assuming a continental arc area of 6 × 106 km2, physical erosion rate of ~80 t/km2/yr (referring to weathering rates of 63–170 t/km2/yr in high-runoff Reunion Island)55, and P concentration of 975 ppm (1.5 × granitic P concentration of 650 ppm), the total reactive P flux from physical erosion of continental arc would be 6 × 109 mol P/yr, which is nearly four times of the background P flux by granitic crust and also consistent with a 3.75 × juvenile weathering flux suggested by our Os-Sr model results (Fig. 4d).

Switch to incongruent weathering of continental crust at ca. 525–515 Ma

Positive Os-Sr isotopic shifts at ca. 525–515 Ma (Fig. 3a, b) suggest a switch of dominant weathering from juvenile rocks to continental crust. The coupled positive shift of δ7Lisw further indicate a dramatic change from congruent weathering of juvenile rocks to incongruent weathering of continental crust (Fig. 3d). Dynamic Li box model results reveal that a high δ7Liriv (up to 22‰) is required to explain this positive shift. Considering the arc juvenile magmatism additions may become weak and the long-term erosion of uplifted juvenile rocks at ca. 540–525 Ma would have reduced the high-relief terrains, it is reasonable to infer that the development of floodplains became prevailing and favorable for uptake of 6Li by clays at the time, leading to a high δ7Liriv. Actually, the abnormally low δ7Li values of marine mudstones (down to –5‰) also support the large-scale clay production from incongruent weathering of continental crust (i.e., continental clay factory)16 and their delivery to oceans after 525 Ma.

Linkage between subduction-driven weathering dynamics and the Cambrian Explosion

The Cambrian Explosion is characterized by the appearance of small shelly fossils mainly at ca. ~540–525 Ma, and a rapid radiation of large-body, skeletonized animals at ca. ~521–518 Ma12,36 (Fig. 3f). The phosphorus cycle is thought to modulate biospheric productivity and ocean–atmospheric oxygen levels over geological times56. Enhanced erosion of arc-related juvenile rocks as evidenced by Li-Os-Sr isotopes may facilitate an unprecedented flux of bioavailable P to the oceans, which may have significantly boosted primary productivity and amplified both photosynthetic oxygen production and organic carbon export20. In particular, increased organic carbon burial, usually occurring in favorable facies such as deep-water shelf to form organic-rich shales and evidenced by widely positive δ13Ccarb values (Fig. 3g), would lead to a net accumulation of atmospheric oxygen57. Concurrent with these processes, the oxygenation of shallow-water habitats is also enhanced via air–sea gas exchange, fulfilling the oxygen requirement of early animals and accelerating animal diversification. Moreover, juvenile rock erosion yields higher Ca2+ fluxes per unit volume than continental crust weathering23. This enhanced Ca2+ supply may have potentially contributed to the advent of skeletonized animal biocalcification—a hypothesis supported by coincident rise in seawater Ca2+ concentrations14,58 (Fig. 3e). A switch to continental clay formation at 525–515 Ma and their delivery to the oceans could have increased efficient clay-bounded organic carbon burial16 and facilitated further oxygen accumulation required by large-body animals as evidenced by the wide occurrence of organic-rich of shales over the whole Yangtze Block38 (Fig. 1c).

In summary, a three-stage model is tentatively proposed that links subduction-driven weathering dynamics, seawater chemistry, and the Cambrian Explosion during 550–515 Ma based on sedimentary Li-Os-Sr isotopic records (Fig. 5). The enhanced erosion of juvenile arc-related rocks increased the input of key nutrients (e.g., P) into oceans, which stimulated marine primary productivity. This elevated productivity, in turn, drove higher export and burial of organic carbon. The sequestration of organic carbon facilitated net accumulation of atmospheric oxygen, thereby increasing oxygen levels that ultimately satisfied the demands of Cambrian animal diversification. Our work provides a distinct perspective on the controls of deep earth activity on surface biogeochemistry cycle at this critical period of animal diversity.

Fig. 5: A three-stage model showing tectonic-driven weathering dynamic and seawater chemistry.
figure 5

a Congruent weathering of subduction-driven thickened crust at ca. 550–540 Ma; b Enhanced erosion of arc-related juvenile rocks promotes rapid delivery of bio-essential nutrient (e.g., P, Ca) to global oceans at ca. 540–525 Ma; c A switch to incongruent weathering of continental crust with secondary clay formation and resultant efficient clay-organic carbon burial and O2 accumulation at ca. 525–515 Ma. The Li-Os-Sr isotopic endmembers are in black for ancient crust weathering, in orange for weathering of arc-related juvenile rocks, and in blue for seawater.

Full size image

Methods

Sample preparation

Twenty-one fresh core samples of drillcore ZK4411 and forty-two carbonate samples of drillcore ZK4803 covering 560–515 Ma were selected for Li-Os isotope analysis, and thirty-seven core samples are selected for rare-earth elemental analysis. Core samples were cut into fresh chips and then manually ground to 200 mesh size using an agate mortar and pestle to avoid any metal contaminant. The powered samples were dried and then preserved in a desiccator before instrumental analysis.

Re-Os isotope analysis

The Re-Os chemical separation and analytical methods followed the procedure described by Yin et al. (ref. 59). Approximately 0.2–0.6 g of powdered sample was digested together with 185Re and 190Os enriched spikes in sealed Carius tubes. Prior to sealing, the tubes were frozen in an ethanol–liquid nitrogen freezing bath (–50 to –80 °C) and then sealed using an oxygen–propane torch. Digestion was carried out at 220 °C for 48 h. Following digestion, osmium was isolated via CCl4 solvent extraction, back-extracted into concentrated HBr, and further purified by microdistillation60. The remaining solution, containing Re, was processed and purified through anion-exchange chromatography. Rhenium concentrations were determined by inductively coupled plasma mass spectrometry (ICP-MS; Thermo Elemental X2 Series). Osmium isotopic measurements were conducted by negative thermal ionization mass spectrometry (N-TIMS) as OsO3⁻ ions loaded onto Pt filaments, using a Thermo-Finnigan Triton instrument operated in electron multiplier mode at the State Key Laboratory of Deep Earth Processes and Resources, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Reproducibility and accuracy of the Os isotopic measurements were monitored by repeated analysis of the DROsS standard solution, which yielded a mean 187Os/188Os ratio of 0.16094 ± 0.00007 (2σ, n = 4), consistent with previously reported values25.

Total procedural blanks (TPBs) averaged 0.41 ± 0.22 pg (2σ; n = 4) for Os and 2.9 ± 0.4 pg (2σ; n = 4) for Re, with an average 187Os/188Os ratio of 0.286 ± 0.082 (2σ; n = 4). All sample data were corrected for the corresponding procedural blank of each analytical batch. The contribution from the blanks was generally negligible. Analysis of the black shale reference material SGR-1b yielded mean values of 187Os/188Os = 1.7857 ± 0.0098, Os = 0.444 ± 0.026 ppb, and Re = 34.67 ± 0.34 ppb (2σ; n = 2), which are consistent with certified and literature values59. Regression analyses were performed using IsoplotR61 (https://www.ucl.ac.uk/~ucfbpve/isoplotr/home/). Initial 187Os/188Os values (Osi) were calculated from the measured 187Os/188Os and 187Re/188Os values, and the estimated ages (Supplementary Data 1), with a 187Re decay constant of 1.666 × 10−11/yr62. Total uncertainties include measurement, blank, decay constant, and spike calibration uncertainties, which are reported at the 2σ level.

Li isotope analysis

We targeted the Li isotopic composition of the carbonate core samples. A chemical leaching method using 0.3 M dilute acetic acid was adopted63. After cleaning using deionized water, approximately 200 mg of bulk carbonate sample was leached with 0.3 M acetic acid for at least one hour. After centrifuging, the supernatants were dried and re-dissolved in 1 mL 8 M HCl and dried again. The samples were then dissolved in 2 mL 0.2 M HCl for column purification. Purification of Li for isotope analysis was achieved using an organic solvent-free single-step liquid chromatography procedure in a super-clean lab at Hefei University of Technology. Large columns (4 mL AG 50W-X12 resin, 100–200 mesh) were used to ensure that the column was not saturated for calcium and other cations, which is important for low-Li samples. Separations were monitored with ICP-MS analysis to guarantee both high Li yield [>99.8% recovery] and low Na/Li molar ratio (<0.5)64. The Li yield was calculated using the following equation: Yield (%) = collected Li/(collected Li + pre-yield Li + after-yield Li) × 100%. The lithium isotopes were measured on a Neptune plus MC-ICP-MS at Hefei University of Technology. The set-up was optimized to achieve both high sensitivity (7Li signal > 6 V for 50 ng/g Li solutions, Jet+X cones) and high stability. Solutions for MC-ICP-MS analysis were introduced through a PFA nebulizer (50 μL/min) coupled with a quartz Scott-type spray chamber and were matrix-matched to 10 or 50 ng/g according to the bracketing standards to ensure the best precision and accuracy. The total procedure blanks determined for the combined sample digestion and column procedure were about 0.03–0.14 ng Li. Compared with the ~10–100 ng Li used for our analysis, the blank correction is not significant. Each measurement was bracketed with a L-SVEC standard. The in-run precision of 7Li/6Li measurements is ≤0.2‰ for one block of 30 ratios. The external precision, based on long-term monitoring of two in-house standards, LiQC = +8.8 ± 0.3‰ (2 SD, n > 500) and LiUSTC-L = −19.3 ± 0.2‰ (2 SD, n > 200), is better than 0.5‰. Analysis of international rock reference materials yield δ7Li values of +4.3 ± 0.4‰ (2 SD, n = 8) for BHVO-2, −0.8 ± 0.2‰ (2 SD, n = 29) for GSP-2 and +5.9 ± 0.4‰ (2 SD, n = 9) for AGV-1, which are within the uncertainty of previously published values64.

Rare-earth elemental analysis

Concentrations of rare earth elements were measured by inductively coupled plasma mass spectrometry (ICP-MS, Thermo Scientific Element XR) at the State Key Laboratory of Isotope Geochemistry of Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. About 40 mg of powdered samples (200 mesh) were digested with a mixture of HNO3 + HCl + HF+HClO4 in high-pressure Teflon cups. The organic-rich samples were firstly combusted in a muffle oven at 650 °C for 4 h to completely remove organic matter before dissolution by acids. The analytical standards GSR1 (granite), GSR3 (basalt), SGR-1b (oil shale) plus replicate samples were jointly used to monitor analytical precision, which was better than 7% for rare earth elements.