Ecological risk associated with potentially toxic elements in agricultural soils across coastal and highland valleys

Soil elemental composition in heterogeneous agroecosystems is shaped by interacting environmental and anthropogenic controls. This study evaluated the spatial variability of potentially toxic elements (PTEs: Cu, Cr, Fe, Mn, Mo, Ni, Pb, V, Zn, As, and Cd) across coastal (Chancay and Pativilca) and highland (Mantaro and Tarma) agricultural valleys of Peru. A stratified sampling design was combined with multivariate analyses (PCA, PERMANOVA, PERMDISP, and variance partitioning) and ecological risk assessment using integrated indices (PLI, mCd, SRI, and Nemerow index). The first two principal components explained 50.2% of total variance (PC1 = 36.8%; PC2 = 13.4%), reflecting distinct soil–geochemical and climatic–spatial gradients. The Valley × Zone interaction significantly structured elemental composition (R² = 0.049, p = 0.011), whereas crop type showed no significant effect (p = 0.838). Variance partitioning indicated that soil physicochemical, climatic, and spatial/topographic predictors jointly explained 60% of total variation (adjusted R² = 0.597), with the three-way shared fraction accounting for 28%, highlighting strong coupling among pedogenic, climatic, and topographic drivers. Ecological risk indices revealed clear spatial differentiation between systems. Highland valleys exhibited greater contamination intensity, spatial heterogeneity, and more frequent high-risk categories according to PLI, mCd, and SRI. In contrast Coastal valleys showed more homogeneous and diffuse accumulation patterns associated with long-term agricultural intensification. These findings underscore the need for regionally adapted soil monitoring frameworks that incorporate environmental gradients in the assessment and management of PTE-related ecological risk in agricultural landscapes.