Coastal El Niño triggers rapid marine silicate alteration on the seafloor

The study addresses how quickly marine silicate alteration on continental margins can respond to extreme weather events and enhanced terrigenous input. Marine silicate alteration encompasses marine silicate weathering (consuming CO2 and increasing alkalinity) and reverse weathering (forming authigenic clays that consume alkalinity and release CO2). Historically, reverse weathering was thought to be slow (thousands to hundreds of thousands of years), but recent work indicates authigenic clay formation can occur within months to years, especially where terrestrial mineral supply is high. Here, the authors test whether coupled dissolution–precipitation reactions can proceed even faster—on the order of weeks to months—by examining the Peruvian margin before and after the 2017 coastal El Niño. They use stable silicon isotopes (δ30Si) and Ge/Si ratios as tracers of silicate dissolution and authigenic clay precipitation to evaluate the coupling between terrestrial erosion and marine sedimentary processes and the implications for alkalinity and CO2 cycling.

Prior studies established that reverse weathering in marine sediments can be an important long-term regulator of alkalinity and CO2 but was traditionally considered slow. Recent observations show rapid formation of authigenic clays in tropical settings with high terrestrial inputs. Silicon isotopes are sensitive to silicate mineral reactions because light Si isotopes are preferentially incorporated into secondary minerals; germanium behaves similarly to silicon in the ocean, with both mainly sourced by rivers and hydrothermal inputs and removed by biogenic uptake, adsorption to Fe-Mn oxides, and authigenic clay formation. On the Peruvian margin, previous work suggested authigenic clay formation may be limited by terrigenous Al delivery. Baseline pore fluid δ30Si values in earlier years (2008, 2013) reflected diatom dissolution signals with limited evidence for extensive authigenic clay formation.

• Study area and sampling design: Surface sediments and pore fluids were collected across the Peruvian margin, including shelf sites, an oxygen minimum zone (OMZ), and sites below the OMZ. Shelf pore fluids from 2017 (immediately following a coastal El Niño event) were compared to samples from 2013 and literature data from 2008.
• Sediment and pore fluid sampling: Sediment cores (up to 40 cm) were retrieved using a multiple corer and subsampled under argon at 4°C. Bottom water was filtered through 0.2 µm filters. Pore fluids were separated by centrifugation and filtered. Fluids (bottom water and pore waters) were acidified with suprapur HNO3 to pH <1 and stored in acid-cleaned LDPE bottles for analyses of Si isotopes, Ge, and Al. Sediments were freeze-dried for XRD and prepared as thin sections for in situ Si isotope analyses of bulk sediment and authigenic clays.
• Benthic chamber incubations: Dual-chamber benthic landers (28.8 cm diameter chambers) were partially inserted into the sediment (20–30 cm overlying water; volume ~12–18 L; area 651.4 cm²). Prior to incubation, chambers were flushed with ambient water to remove trapped particles. Incubations lasted approximately 32 hours, with eight sequential 12 mL samples collected in acid-cleaned quartz tubes. Post-recovery, fluids were acidified to pH <2 and analyzed for dissolved Si, Ge, Al, and Si isotopes.
• Major element analyses: Freeze-dried sediment powders were digested using HF (40%), HNO3, and HClO4 (60%) and analyzed by ICP-OES (VARIAN 720-ES) for pore fluid Si and bulk solid Al and K. Accuracy was checked with blanks and standards (SDO-1, MESS-3). Reproducibility was ≤5% using IAPSO seawater standard.
• Germanium and aluminum in fluids: Al and Ge in bottom and pore fluids were determined by ICP-sector field MS (Thermo ELEMENT XR) at high mass resolution (~10,000), with 1:20 dilution in 2% HNO3 and yttrium as internal standard. Procedural blanks were below LOQ (Si 0.08 mg L−1, Al 0.9 µg L−1, Ge 0.006 µg L−1). CRM recoveries for Al matched certified values; relative uncertainty was 5–10% for all elements.
• Fluid silicon isotope analyses: Si was purified via single-column cation exchange (AG 50W-X8 resin) following established protocols. Approximately 4 µg Si aliquots were processed; low-Si samples were evaporated prior to column chemistry. Samples and standards were Mg-doped to stabilize mass bias and adjusted to ~1 ppm Si for analysis. Measurements were performed on Neptune Plus and NuPlasma MC-ICP-MS systems. On the NuPlasma with Aridus II, ~21 µM Si yielded 28Si signals of 3–4 V; blanks ≤3 mV. On the Neptune Plus (wet plasma), 28Si signals were ~−6.0 V; blanks <20 mV. Reference materials (Big Batch, IRMM-18, Diatomite, in-house standard) reproduced literature δ30Si values; several samples were cross-checked on both instruments with agreement within error. Results are reported in δ30Si relative to NBS 28.
• In situ silicon isotope analyses of solids: Femtosecond laser ablation (196 nm, ~150 fs) coupled to a Neptune Plus MC-ICP-MS at GFZ Potsdam was used to measure in situ δ30Si in authigenic clays and bulk sediment thin sections. Aerosols were transported in He; medium mass resolution (m/Δm > 5000) mitigated isobaric interferences (e.g., 14N16O on 30Si). Faraday detectors with 1012 Ω amplifiers were used, tuned to ~7 V on 28Si for samples and NBS 28. Reference materials (BHVO-2, GOR132-G, ML3B) matched published δ30Si values.
• X-ray diffraction (XRD): Measurements used a Phillips diffractometer (CoKα, Ni filter, 40 kV, 35 mA). Triplicate runs provided ±2% reproducibility. Phase quantification employed MacDiff 4.25, Panalytical HighScore, and QUAX; standard deviations were ±1–3% for well-crystallized and ±5% for other phases.
• Electron microprobe (EMP): Elemental distributions in 0.5 cm depth thin sections were examined using a Jeol JXA-8200 in EDS mode (15 kV, 10 nA). Due to amorphous biogenic silica and authigenic clays, measurements were semiquantitative.
• Reaction-transport modeling: A 1-D model simulated Si turnover for 2013 and 2017, incorporating molecular diffusion, bioturbation, burial/compaction, and reactions. Mass balance equations described solids and solutes, including species SiO2 in biogenic opal and authigenic phases, K in sediments, isotopologues 30SiO2 in solids, and dissolved H4SiO4, H4 30SiO4, K, and Ge. δ30Si and Ge/Si in pore fluids were derived from modeled mole fractions. Model equations and parameters are detailed in the Supplementary Information.

• Following the 2017 coastal El Niño, shelf pore fluids exhibited exceptionally high δ30Si values up to +3.72‰ (vs. +1.33‰ in 2013 and +1.21‰ in 2008) and elevated Ge/Si ratios (up to 2.87 µmol mol−1), indicating extensive and rapid authigenic clay formation.
• Uppermost sediments (<5 cm) in 2017 had significantly higher Ge/Si ratios and Al concentrations than during earlier campaigns, consistent with enhanced terrigenous input and availability of Al for clay formation.
• Field observations documented deposition of a new sediment layer up to 4 cm thick enriched in terrigenous quartz and feldspars, supplying reactive minerals (notably albite) capable of dissolving within weeks to months in seawater.
• EMP analyses provided direct evidence for authigenic clays in 2017, showing enrichments with average elemental ratios: K/Si = 0.22 ± 0.07 (1 SD, n=11), Fe/Si = 0.13 ± 0.08 (n=10), Mg/Si = 0.40 ± 0.15 (n=10), and Al/Si = 0.49 ± 0.12 (n=11). Sediment matrix Al/Si was lower: 0.09 ± 0.10 (n=5) in 2017 and 0.09 ± 0.06 (n=6) in 2013.
• δ30Si vs. Ge/Si relationships show most samples align with mixing between seawater and dissolution of diatoms and feldspar. Only 2017 shelf samples deviate toward much higher δ30Si and Ge/Si, reflecting strong authigenic clay precipitation.
• Modeled authigenic clay Si precipitation rates increased from 13 µmol cm−2 yr−1 in 2013 to 291 µmol cm−2 yr−1 in 2017 on the shelf, a more than 20-fold increase.
• Ge was strongly enriched in 2017 relative to other reducing margin settings, attributed to differential uptake rates where Si is preferentially removed into clays faster than Ge during concurrent dissolution–precipitation.
• Benthic chamber δ30Si on the shelf in 2017 deviated from seawater more than at other stations, consistent with intensified sediment–seawater interactions and clay formation.
• Model results indicate that under present conditions at the Peruvian margin, reverse weathering exceeds marine silicate weathering by a factor of two (Table S10).

The data demonstrate a rapid response of marine silicate alteration to extreme rainfall and runoff associated with a coastal El Niño. Enhanced delivery of reactive terrigenous minerals (e.g., albite) to the shelf accelerated dissolution in surface sediments and fueled authigenic clay precipitation, recorded by extraordinarily high pore fluid δ30Si and elevated Ge/Si ratios. This coupled dissolution–precipitation signifies a dynamic reverse weathering regime capable of shifting alkalinity and CO2 on short timescales (weeks to months). The concurrent elevation of Al and Ge, direct identification of clay phases by EMP, and reaction-transport modeling collectively confirm that the post-event geochemical anomalies are best explained by rapid authigenic clay formation. These findings resolve the research question by showing that the timescale of authigenic clay formation can be much shorter than previously assumed and tightly linked to episodic terrigenous inputs. The broader relevance lies in the coupling between terrestrial erosion and marine element cycles: increased frequency of extreme precipitation events under climate change will likely intensify such coupled weathering, influencing cation budgets, diatom preservation (through conversion to less soluble phases), and short-term CO2 cycling. However, the net outcome for alkalinity and CO2 depends on the balance between dissolution (alkalinity production, CO2 consumption) and reverse weathering (alkalinity consumption, CO2 release), which may vary with temperature, pH, and redox conditions.

This study shows that coastal El Niño-driven floods can trigger rapid marine silicate alteration on continental shelves by delivering reactive terrigenous minerals that dissolve and promote authigenic clay formation. Pore fluid δ30Si and Ge/Si anomalies, elemental enrichments, direct mineralogical evidence, and modeling collectively demonstrate that authigenic clay precipitation can intensify within weeks to months, accelerating reverse weathering and influencing alkalinity and CO2 cycling on human-relevant timescales. The results challenge long-held assumptions about the slow pace of authigenic clay formation and highlight close coupling between land erosion and marine sedimentary processes. Future work should: (i) constrain the geographic extent and recurrence of such rapid alterations across diverse margins; (ii) quantify the sensitivity of coupled weathering to changing pH, temperature, and oxygen under climate change; (iii) refine models to predict net alkalinity/CO2 impacts under varying mineral supply regimes; and (iv) expand time-resolved observations across extreme events to capture transient dynamics.

• Regional and event-specific focus: Findings are based on the Peruvian margin and a single extreme event (2017 coastal El Niño), which may limit generalizability to other settings or event types.
• Predictive uncertainty: The net effect of climate-driven changes (warming, acidification, deoxygenation) on the balance between dissolution and reverse weathering is difficult to predict, with potentially opposing influences on authigenic clay formation rates.
• Modeling assumptions: Reaction-transport modeling relies on assumptions about mineral endmembers (e.g., albite abundance and dissolution behavior) and parameterizations that introduce uncertainty.
• Temporal coverage: Comparisons are based on snapshots (2008, 2013, 2017) rather than continuous time series, potentially missing short-lived variability before and after the event.