Winter-only grazing governs plant leaf economic traits and productivity pathways, with fertilization effects contingent on grazing regime

Grazing and fertilization are major drivers of grassland ecosystem functioning, yet the mechanisms through which their interaction shapes plant functional traits and productivity remain only partially understood. We carried out a split-plot field experiment in alpine wetland meadows, combining seasonal versus year-round grazing with three nitrogen addition levels (0, 5, and 10 g N m –2 yr –1 ). We measured community-level leaf traits, soil physicochemical properties, biodiversity metrics, and aboveground biomass (AGB), and used ordination, machine learning, and structural equation modeling. Relative to year-round grazing, winter-only grazing markedly increased community-weighted leaf area (CWM_LA; about 20% increase, P < 0.05) and generally elevated specific leaf area (SLA), shifting trait expression toward more acquisitive strategies. Fertilization responses were weaker and strongly dependent on grazing: under winter-only grazing, nitrogen inputs promoted more conservative leaf traits in grasses (reduced SLA, higher LDMC), whereas under continuous grazing, trait shifts were minimal. Principal component analysis indicated that grazing was the dominant source of trait variation, with fertilization adjusting community placement along the leaf economic spectrum. Random forest models highlighted available phosphorus, total phosphorus, soil pH, and CWM_SLA as the main positive predictors of AGB, while excessive nitrogen input showed a negative relationship with biomass. Variation partitioning revealed that management factors alone explained about 45% of biomass variation (P < 0.05), surpassing soil, traits, or diversity. Structural equation modeling showed that grazing reduced AGB indirectly by decreasing soil nutrients and plant diversity, whereas fertilization directly lowered biomass (b = –0.326, P < 0.05), likely due to acidification and altered community structure. These findings offer insight into sustainable grassland management, suggesting that adjusting grazing regimes may be more effective than fertilization in improving ecosystem functioning under non–N-limited conditions.