An intensity gradient of extreme rainfall governs nitrogen and phosphorus losses from contour ridge systems in the rocky mountain region of Northern China

Extreme water erosion-induced nutrient loss poses a widespread environmental threat, contributing to land degradation and waterbody eutrophication. The increasing changes in extreme rainfall aggravate the complexity and uncertainties for nutrient loss, particularly when accompanied with agricultural tillage practice. However, field investigations of nitrogen (N) and phosphorus (P) losses along with the intensity gradient of extreme rainfall in agricultural system remain limited. To address this gap, this study examined the dynamics of N and P losses via runoff and sediment across an intensity gradient of extreme rainfall, categorized into four gradients: R50 (daily precipitation between 50 mm and the 90th percentile), P90 (90th–95th percentile), P95(95th–99th percentile) and P99 (>99th). Field observations were conducted over three years (2021–2023) using in-situ runoff plots of sweet potatoes (SP) and peanuts (PT) contour ridge systems. Results reveal that extreme rainfall contributed most significantly to nutrient loss, accounting for 56.71 %–64.77 % of total loss compared to ordinary rainfall. Nutrient loss associated with sediment exhibited greater sensitivity to extreme rainfall than those via runoff. Under extreme rainfall, nutrient loss was primarily driven by R50 events in SP and by P90 events in PT. Notably, R50 and P90 events contributed to over 60 % of nutrient loss, attributable to their greater precipitation and higher rainfall kinetic energy. Despite contributing only 6.71 % of total precipitation, the short-duration, high intensity P95 event accounted for 10.53 %–15.02 % of nutrient loss across all events. While, the long-duration, moderate-intensity P99 event exerted a comparatively limited effect on nutrient loss. Phosphorus loss exhibited 3.76–3.99 times higher sensitivity to extreme rainfall gradients than nitrogen loss, reaching a high-risk level. Structural equation modelling confirmed that sediment yield and extreme rainfall kinetic energy were the dominant factors driving nutrient loss, outweighing the effects of runoff and other extreme rainfall variables. These findings enhance the understanding of nutrient loss dynamics under extreme rainfall, and provide guidance for developing targeted conservation practices to mitigate agricultural non-point source pollution in watersheds facing increasing climate extremes. • Extreme rainfall dominated nutrient loss, far exceeding the impact of ordinary rainfall. • Nutrient loss peaked at R50/P90 events and was further amplified by P95 event. • Phosphorus loss exhibited higher sensitivity to extreme rainfall than nitrogen loss. • Sediment yield and rainfall kinetic energy were the primary drivers of nutrient loss.