Compositionally Distinct Fertilization Regimes Produce Divergent Responses Across Soil Properties, Fresh-Shoot Chemistry and the Shoot Metabolome in a Mature Tea Plantation

Fertilization regimes in tea plantations must be evaluated across soil conditions, harvested-shoot chemistry and practical input requirements. We compared an unfertilized control with five compositionally distinct regimes in a randomized complete block field experiment. The field plot was the experimental unit, with three independent plots per treatment, and the first observation occurred after the first fertilizer application. At the autumn-2024 sampling, 12 of 15 soil and conventional fresh-shoot traits showed a treatment effect after Benjamini–Hochberg correction, but the direction and statistical support of pairwise differences varied among endpoints. Of 4609 metabolic features, 4069 passed pooled-QC criteria and 3845 were fully detected across the 18 biological samples. A total of 495 unique quantitative-event features contributed 988 feature–contrast events and were summarized in seven empirical modules. Only M1 alkaloid overrepresentation remained significant after correction within the preselected responsive subset; M1 contained 32 selected features, including seven putatively annotated alkaloids, and its unadjusted plot-level association with conventionally measured caffeine was weak. This is limited exploratory compositional evidence, not a confirmed biological mechanism. A total of 9 of 105 correlations among conventional indicators retained correction, whereas none of 35 candidate-metabolite-to-shoot-trait correlations did. Fertilizer material costs ranged from 187.5 to 7500 CNY mu−1 among the five fertilized regimes under the stated price assumptions. These costs and an owner-reported farm-operation scenario provide practical context but are not treatment-specific profitability estimates. The results support regime-specific evaluation under the tested formulations, rates, site and observation period, without establishing fertilizer-source superiority, an organic–mineral interaction or a causal soil-to-metabolite mechanism.