Mobilization of grassland soil arsenic stores due to agronomic management.

The fate of arsenic in mineral soil stores over time is poorly understood. Here we examined arsenic loss over five decades from a managed grassland soil profile through analysing archived material from a long-term slurry (LTS) experiment at Hillsborough, Northern Ireland. A randomized block experiment was established in 1970 where a perennial ryegrass sward was seeded onto the site and subjected to control (no fertilization) and fertilization treatments using conventional (i.e. at farmers recommended application rate) mineral NPK fertilizer, and pig and cow slurry treatments. Soil (0-5, 5-10 and 10-15 cm), slurry applied, and sward off-take was archived each year. A mass-balance calculation found that control soils lost no arsenic down the 15 cm depth soil profile, the NPK treatment had a 10 % loss, while cow slurry caused 25 % loss, and a 35 % loss was observed for pig slurry. For treatments with arsenic loss, removal was linear over the 50 years of study in 2 out of the 3 blocks, with the 3rd block showing little or no change. Principal Component Analysis (PCA) found that arsenic was most positively associated with soil magnesium, manganese and nickel, while negatively associated with pH, organic carbon, phosphorus and silicon. Laser ablation - inductively coupled plasma - mass spectrometry (LA-ICP-MS) of soil found that arsenic association with lead mineralogy could potentially explain why there was a gradient in arsenic loss across the experimental plots. Slurry and atmospheric inputs, and sward off-take had little impact on the soil arsenic mass-balance. The findings suggest that leaching loss down the soil profile was the mechanism of loss of arsenic. The applicability of the LTS experimental site arsenic findings to other soils is discussed, as is the implication for the global biogeochemical cycling of those soils.