Oxygen input controls the spatial and temporal dynamics of arsenic at the surface of a flooded paddy soil and in the rhizosphere of lowland rice (Oryza sativa L.): a microcosm study

Presented at 2. International Conference on Rhizosphere, Montpellier, FRA (2007-08-26 - 2007-08-31)International audienceThe impact of oxygen (O-2) input at the soil surface and in the rhizosphere of rice (Oryza sativa L.) on the spatial and temporal dynamics of arsenic (As) was investigated in a flooded paddy soil. A soil microcosm and root-mat technique were designed to mimic submerged conditions of paddy fields. Water-filled containers with (planted) or without (unplanted) 27-day-old rice seedlings were fitted for 20 days on top of microcosms containing an As-affected soil (Bangladesh). After the initial establishment of strongly reduced conditions (-230 mV) in both planted and unplanted soils, the redox potential gradually increased until the day 8 to reach+50 mV at 2 mm from the surface of unplanted soils only. This oxidation was associated with an accumulation of NH4-oxalate extractable As (25.7 mg kg(-1)) in the 0.5-mm top layer, i.e. at levels above the initial total content of As in the soil (14 mg kg(-1)) and a subsequent depletion of As in soil solution at 2 mm from soil surface. Root O-2-leakage induced the formation of an iron (Fe) plaque in root apoplast, with no evidence of outer rhizosphere oxidation. Arsenic content reached 173 mg kg(-1) in the Fe plaque. This accumulation induced a depletion of As in soil solution over several millimetres in the rhizosphere. Arsenic contents in root symplast and shoots (112 and 2.3 mg kg(-1), respectively) were significantly lower than in Fe plaque. Despite a large As concentration in soil solution, Fe plaque appeared highly efficient to sequester As and to restrict As acquisition by rice. The oxidation-mediated accumulation of As in the Fe plaque and in the oxidised layer at the top of the soil mobilised 21 and 3% of the initial amount of As in the planted and unplanted soils, respectively. Soil solution As concentration steadily decreased during the last 16 days of the soil stage, likely indicating a decrease in the ability of the soil to re-supply As from the solid-phase to the solution. The driving force of As dynamic in soil was therefore attributed to the As diffusion from reduced to oxidised soil layers. These results suggest a large mobility of As in the soil during the flooded period, controlled by the setting of oxic/anoxic interfaces at the surface of soil in contact with flooding water and in the rhizosphere of ric