Advanced Atomic- and Molecular-level Techniques Reveal Alterations to Soil and Stream Organic Matter from Agricultural Sulfur Applications

Sulfur (S) applications to crops are a critical part of soil and nutrient management, but the environmental fates and consequences of agricultural S use are largely unknown. Organic S is the most abundant form of S in both natural and agricultural soils, and the quantity, speciation, and molecular composition of the dissolved organic S (DOS) pool affect S transport and reactivity both within and beyond agricultural fields. In this study, we combine three advanced organic matter characterization techniques to identify the agricultural DOS atomic, molecular, and isotopic fingerprint compared to non-agricultural (i.e. forested and grassland) land. We collected soil porewater and stream water samples from vineyards that receive S applications of 22-40 kg S ha-1 yr-1 and adjacent forested and grassland areas within the Napa River Watershed, California, U.S. We isolated dissolved organic matter (DOM) using PPL solid phase extraction, which enabled parallel measurements of acidic S DOM species molecular composition by negative-ion electrospray ionization Fourier-transform ion cyclotron mass spectrometry (FT-ICR MS), atomic oxidation state by S K-edge X-ray absorption near-edge structure (XANES) spectroscopy, and S stable isotopic composition using isotope ratio mass spectrometry (34S-DOS). We found that non-vineyard soil porewater and streams had 34S-DOS values enriched by 3-8VCDT relative to inorganic sulfate (34S-SO42-) and 3-7% relative abundance of S-containing heteroatoms (CHOS + CHONS). In contrast, vineyard soil porewater and streams had 34S-DOS values depleted by 1-5VCDT relative to 34S-SO42- values, a higher relative abundance of S-containing heteroatoms (9-13% CHOS + CHONS), and a shift in the proportions of reduced to oxidized organic S functionalities. We hypothesize that microbial sulfate reduction in soil anoxic microsites and/or during periods of intermittent to sustained soil saturation may explain the observed differences between agricultural and non-agricultural DOS. Combined, our analyses suggest that agricultural S additions stimulate microbial processes that alter the composition of DOS compared to non-agricultural areas, influencing the ultimate fates of agricultural S additions at field-to watershed scales.