Mismatches between ammonium and nitrate losses at the field and watershed scales suggest contrasting controls in two agricultural watersheds

Nitrogen (N) fertilizer enhances crop production, but field runoff impacts water quality in adjacent freshwaters. Planting winter cover crops reduces nitrate-N losses during the fallow period, but less is known about impacts on ammonium-N. From 2016–2023, we sampled biweekly from the Shatto Ditch and Kirkpatrick Ditch Watersheds in Indiana (USA) to compare the impact of cover crops on dissolved inorganic nitrogen at the field-, edge-of-field, and watershed-scales. We measured soil ammonium-N and nitrate-N, biomass, and organic matter in fall and spring. Cover crops reduced soil ammonium-N at Shatto and soil nitrate-N in both watersheds. Tile losses and watershed yields of ammonium-N occurred on scales orders of magnitude lower than nitrate-N. Tile ammonium-N losses from cover cropped fields ranged from 97 % lower to 31 % higher at Shatto, and 45 % lower to 75 % higher at Kirkpatrick compared to those without. Cover crops reduced field-scale nitrate-N losses at Shatto by 58–87 %, but losses at Kirkpatrick ranged 99 % lower to 15 % higher. Tile flow explained interannual variation in nitrate-N losses, while field-scale ammonium-N losses were driven by soil and microbial interactions and mobilization during storms. Watershed-scale ammonium-N and nitrate-N yields correlated with runoff (Kendall τ=0.45 and 0.39, respectively). While nitrate-N yields mirrored runoff, ammonium-N yields exhibited a step-functional increase, pointing to the importance of storms as a driver of loss. As Midwest crop production adapts to fluctuating environmental conditions, we demonstrate how applying cover crops over a multi-year period can mitigate ammonium-N losses. • Cover crops reduced soil ammonium and nitrate pools, but only nitrate was related to biomass. • Tile ammonium losses varied with cover crops, changing with flow, soils and environmental conditions. • Watershed scale ammonium yields exhibited stepwise increases, suggesting storms drive loss. • Ammonium loss at field and watershed scales was 2–3 orders of magnitude lower than nitrate.