Legacy-buffered groundwater nitrate limits predictability from short-term hydroclimatic windows across intensive...

Groundwater nitrate contamination in intensive croplands often persists despite strong short-term weather variability, suggesting that near-surface signals may not translate directly to aquifers. We tested this across a four-site south-to-north transect in Henan Province, China, under a wheat-maize rotation. For each sampling event, we quantified antecedent hydroclimatic windows from daily meteorological data, and measured depth-integrated (0-50 cm) soil inorganic N stocks, soil N-cycling enzyme activities, groundwater NO₃^--N, and nitrate isotopes (δ¹⁵N-NO₃^- and δ¹⁸O-NO₃^-). Principal component analysis summarized hydroclimatic variability into two dominant gradients: a dry-frequency axis (PC1) and an evaporative-demand axis (PC2). Drier antecedent conditions (higher dry-frequency scores) were consistently associated with greater 0-50 cm soil NO₃^- storage and coordinated changes in enzyme indices related to N turnover; the dry-frequency gradient accounted for 41% of the variance in soil NO₃^--N stocks. In contrast, groundwater nitrate showed weak short-term coupling across all window lengths, with only 19% of NO₃^- variability explained and no significant contemporaneous associations with short-term soil indicators or hydroclimate axes. Isotope patterns showed strong site- and stage-dependent scatter: groundwater δ¹⁵N-NO₃^- varied widely (3.5-17.5‰) while δ¹⁸O-NO₃^- remained narrow (7.0-8.9‰), consistent with mixing among multiple nitrate sources and transport lags that buffer legacy signals in the aquifer. These results help explain why groundwater nitrate remains poorly predictable from short-term surface indicators, and they support management strategies that prioritize sustained reductions in long-term N surplus and monitoring designs that account for subsurface storage and time lags.