Seyfferth et al. 2025 — Rice water management trade-offs: As and Hg vs. Cd across a soil-redox gradient
This 2-year field study (University of Delaware, 2016–2017) grew rice across six paddies managed under a continuous gradient of soil redox conditions, spanning fully flooded to non-flooded furrow-irrigated conditions, and measured grain concentrations of inorganic and organic arsenic species, total and methyl mercury, and cadmium alongside methane emissions. The central finding is a fundamental trade-off: drier water management significantly reduces grain As and Hg (both total and methylated species) and cuts CH4 emissions, but substantially raises grain Cd, with the driest paddy exceeding the CODEX limit of 0.4 mg/kg Cd in both years. Because row rice and alternate wetting-and-drying (AWD) irrigation are being rapidly adopted in the US and globally as climate-smart practices, the study warns that historical risk assessments anchored in flooded-rice Cd data may materially underestimate the Cd risk in the emerging production system.
Key numbers
Soil EH gradient: Paddy 1 (most flooded) averaged −177 mV (Year 1) and 100 mV (Year 2); Paddy 6 (least flooded) averaged 623 mV (Year 1) and 564 mV (Year 2).
Grain total As: highest in Paddy 1 (most flooded); decreased with increasing EH. Inorganic As remained below the CODEX limit of 0.2 mg/kg across all paddies in both years.
Grain iAs: stayed below CODEX 0.2 mg/kg in all paddies both years; also below the FDA action level of 100 µg/kg for infant rice cereal across all paddies.
Grain total Hg: highest in flooded paddies; decreased with drier conditions, with more reduction in Year 1 than Year 2. No CODEX limit for Hg in rice.
Grain MeHg: also decreased with drier conditions, particularly in Year 1. MeHg species accumulate preferentially in polished grain (vs. bran).
Grain Cd: Paddy 6 (driest) exceeded the CODEX limit of 0.4 mg/kg in both years despite very low soil Cd (background soil Cd = 0.093 ± 0.013 mg/kg). Grain Cd increased 50–97% in the driest paddy relative to the most flooded paddy.
Grain yield: driest conditions reduced yield by 25–40% relative to flooded controls.
Grain Zn: unaffected by water management, suggesting homeostatic allocation. Grain Mn and Cu decreased with drier conditions, potentially increasing Cd bioaccessibility.
As speciation: grain organic As (oAs, principally DMA) decreased with drier conditions in both years. Methylated As was proportionally highest in most flooded paddies.
Hg speciation: grain MeHg also decreased with drier conditions, especially Year 1. iHg accumulated preferentially in bran; MeHg in polished grain.
Root plaque: proportion of lepidocrocite (reactive Fe oxyhydroxide) decreased with drier conditions; lepidocrocite proportion was positively correlated with grain As and Hg, independent of soil EH.
LODs: total As and Cd in plant tissue = 0.004 mg/kg; total Hg = 1 µg/kg; MeHg = 0.12 µg/kg; iAs and DMA = 0.004 mg/kg (some samples below LOD in driest paddies, assigned half-LOD).
Methods
2-year outdoor field study, 6 unlined paddies at University of Delaware RICE facility, same soil throughout (Typic Hapludult, sandy clay loam). Rice (Oryza sativa “Jefferson”) transplanted late May, harvested early September. Grain dehusked and polished by benchtop methods; polished grain and bran analyzed separately. Total As and Cd by ICP-MS (microwave acid digestion, TMG HNO3). As speciation (iAs and oAs species) by HPLC-ICP-MS (Jackson 2015 separation method). Total Hg by triple-quadrupole ICP-MS (acid digestion, Dartmouth Trace Metal Analysis Core). MeHg speciation by species-specific isotope dilution GC-purge-and-trap ICP-MS (Brooks Rand MERX system). QA/QC: NIST 1568a rice flour for total As, Cd, micronutrients; NIST 1568b for As species and total Hg; NIST 2976 mussel tissue for Hg speciation. Recoveries: iAs 102%, DMA 95%, total Hg 83%, MeHg 90–91%. Porewater sampled weekly with rhizon samplers; analyzed for pH, EH, Fe(II), total As, iAs, oAs, Cd, S, Mn, DOC. Partial correlation analysis used to disentangle multicollinear drivers of grain metal levels while controlling for soil EH.
Limitation: single location, single cultivar, sandy clay loam soil with unusually high Cd mobilization response relative to flooded baseline. Not generalizable to all soil types; acidic soils most susceptible to the flooded-to-oxic Cd release mechanism (CdS oxidation).
Implications
Certification: the finding that non-flooded rice can exceed the CODEX Cd limit of 0.4 mg/kg despite very low background soil Cd is a supply-chain risk that HMT&C standards for rice-containing products should anticipate. Historical market-basket surveys anchored in conventionally flooded US rice may understate Cd risk as row rice adoption accelerates. The As-Cd trade-off is irreducible under current agronomy; any water-management intervention that reduces As will increase Cd risk in susceptible soils.
Courses: the redox-driven speciation story (flooded conditions mobilize iAs and promote methylation of As and Hg; oxic conditions release Cd from CdS) is the canonical mechanistic explanation of why rice grain metal profiles depend on irrigation practice. This is foundational content for the supply-chain module.
App: the ingredient page for rice should note that grain Cd risk is elevated in row-irrigated and AWD rice relative to conventionally flooded rice. The app’s contamination profile for rice (iAs, tAs, tHg, MeHg, Cd) should distinguish irrigation management as a known variance driver.
Microbiome: the flooded-paddy sulfate-reducing bacteria that drive MeHg production in soil also influence Hg speciation in grain; this connects the soil microbiome to grain MeHg exposure routes relevant to the microbiome pages.