Orchard context dominates first-year soil–plant variation during repeated liquid humus application in commercial citrus orchards

Humic-based products are increasingly applied in perennial cropping systems, yet their early soil–plant responses under commercial conditions may depend strongly on orchard context. We evaluated the first annual cycle of repeated liquid humus application in two contrasting commercial sweet orange orchards in southern Spain. Thirty trees were monitored within control and treated commercial field areas following five fertigation applications (300 L ha -1 ). Within each orchard, comparisons involved one control and one treated field area, without independent field-area replication or pre-treatment baseline measurements. Soil physicochemical properties, soil/rhizosphere-associated metabolomic profiles, foliar nutrition, fruit quality, postharvest weight loss and canopy-standardized production were assessed using complementary univariate and multivariate approaches. Orchard context dominated system structure, explaining 25.4% of the variation in the shared-metabolite matrix; larger Site components in the all-union matrices (50.7–61.4%) also incorporated differences in analytical feature representation. Edaphic, metabolomic and foliar profiles were strongly associated globally (Spearman’s ρ = 0.662–0.758), although most relationships were not detected within individual orchards. Treatment explained smaller but statistically significant fractions of global individual-metabolite variation (5.4% in the shared matrix and 2.45% in the all-union individual-metabolite matrix), but these signals were not reproduced within either orchard and therefore do not indicate a consistent metabolomic response across orchard contexts. Other control–treated field-area differences were orchard-specific: treated trees in Andarax had higher foliar Fe and B concentrations and greater accumulated postharvest weight loss, whereas treated trees in COARE had higher foliar Na and lower fruit titratable acidity. No statistically supported control–treated difference was detected for either predefined canopy-standardized production endpoint. These results show that first-year field-area-associated differences were selective and orchard-specific, emerging against strong orchard-level structure without a reproducible within-orchard metabolomic response or statistically supported production difference.