samplingWe sampled soils from all of the 10 paddy fields in eachof the 7 regions. Topsoil samples (Ap + Arp horizons withthicknesses between ∼10 and ∼30 cm) were taken with aplastic corer of 7.5 cm diameter (n = 9 cores per field). Forcomparison we sampled topsoil (top 20 cm of the mineralsoil) from five non-rice sites that were situated close-by thepaddy fields using a metal corer (n = 5 or 6 cores per site).In PH 1, fields with different fruit trees and/or vegetableswere sampled. In PH 3 and VN 3, forest sites were sampled.In VN 2, fields with vegetables, fruit trees or forest weresampled. In the other regions we could not find sites of analternative land-use type. All soil core samples were air-driedand then shipped to the soil laboratories of the Martin-Luther-University in Halle (export and import permits were obtainedbefore transfer), where they were oven-dried (40◦C), pooledto obtain one sample per paddy field (same dry mass of soilwas taken from all of the cores per field) and sieved (< 2 mm)for laboratory analysis.