In this study, we collected topsoil samples (0‒10 cm) from each of 54 subsites, including sites in direct adjacency (< 10 cm) and in 1 m distance to plants, along an aridity gradient across the Coastal Cordillera in the Atacama Desert. The soluble salts anions (NO^3‒, Cl^‒, and SO₄^2‒) and cations (Ca²⁺, Na⁺, Mg²⁺ and K⁺) were tested. And we performed soil sequential P fractionation and the oxygen isotope values of HCl-extractable P_i (δ¹⁸O_HCl-_Pi). 

1) Soil salt ions analysis

The soluble salts were extracted by Milli-Q water at a soil/water ratio of 1:10. Then the anions (NO^3‒, Cl^‒, and SO₄^2‒) concentrations in the extracts were analyzed by ion chromatography (Metrohm IC 850), and those of cations (Ca²⁺, Na⁺, Mg²⁺ and K⁺) were determined by Inductively Coupled Plasma Optical spectroscopy (ICP-OES, iCAP 6500).

2) Soil sequential P fraction

Sequential P fractionation was performed according to the Hedley sequential fractionation scheme as modified by Tiessen and Moir. In brief, 0.5 g of each air-dried soil sample was sequentially extracted with four reagents including (i) 30 ml deionized water with two anion exchange resin stripes; (ii) 30 mL 0.5 M NaHCO₃ solution at pH 8.5; (iii) 30 mL 0.1 M NaOH; and (iv) 25 mL 1 M HCl. Finally, aqua regia was used at 130 ⁰C to extract residual P as chemically most stable P form. After fractionation, the concentration of inorganic P (P_i) in the extracted solutions was measured by the molybdenum blue colorimetric method and measured at 890 nm by the Photometer Analytik Jena SPECORD 205. Total P (P_t) was measured by ICP-OES (iCAP 6500), and the difference between P_t and P_i was used as organic P (P_o).

3) Oxygen isotope composition of HCl-extractable P_i

For analysis of the oxygen isotope values of HCl-extractable P_i (δ¹⁸O_HCl-_Pi), we followed the method presented by Tamburini et al. (2010), which produced organic matter-free silver phosphate (Ag₃PO₄) from acid extracts for oxygen analysis. Twenty g air-dried soils were successively extracted by 0.5 M NaHCO₃, 0.1 M NaOH and 1 M HCl at a 1/10 soil/solution ratio (Amelung et al., 2015). We discarded the alkaline extracts to remove most of the organic P compounds and polyphosphates, so that only the 1M HCl extract was used for further purification process. The phosphate was precipitated as ammonium phosphomolybdate and magnesium ammonium phosphate successively. Cation exchange resin (Dowex 50X8, 200‒400 mesh, Sigma-Aldrich, Darmstadt, Germany) was used to extract all cations from the solutions and finally precipitated as silver phosphate (Ag₃PO₄). The Ag₃PO₄ crystals were dried in an oven at 50 ℃ for a few days and yellow euhedral Ag₃PO₄ crystals formed after 1 or 2 days. The δ¹⁸O value of Ag₃PO₄ was measured at the Plant Nutrition group at ETH Zurich. Each sample was analyzed in duplicates (resulting in a total of n=108) using a Vario PYRO Cube (Elemental, Hanau, Germany) with a C-based reactor coupled in continuous flow to an Isoprime 100 isotope ratio mass spectrometer (Isoprime, Manchester, UK). The furnace was kept at 1450 ℃ and the produced reaction gases were concentrated by a purge and trap chromatography system. Calibration was performed against an internal Ag₃PO₄ standard (δ¹⁸O=14.1‰) and two international benzoic acid standards (International Atomic Energy Agency; IAEA 601, δ¹⁸O=23.1±0.5‰ and IAEA 602, δ¹⁸O=72.2±0.5‰). All values were reported in the conventional delta notation relative to Vienna Standard Mean Oceanic Water (VSMOW) after calibration.