Iron oxide-rich laterites, soils, and regolith formed from the weathering of ultramafic rocks represent untapped unconventional resources for the critical minerals platinum (Pt) and palladium (Pd), but the fundamental surficial geochemistry of these elements remains poorly understood. Depletion of Pd relative to Pt occurs in some weathering zones in semi-arid climates. The accepted model attributes this Pt-Pd chemical fractionation to preferential complexation of Pd by dissolved chloride (Cl). However, similar fractionation is not observed in laterites of humid equatorial regions despite substantial wet deposition of Cl. The established mechanistic model for Pt and Pd behavior during weathering thus inaccurately predicts the distribution of these critical minerals in many settings, hindering global resource assessment. We show through mineral-fluid partitioning experiments coupled to element-specific spectroscopy that this canonical explanation for Pt-Pd fractionation is invalid: Cl complexation does not differentially mobilize Pd versus Pt. Instead, mineral-specific interfacial reactions control Pd and Pt accumulation. Modeling of Pt-Pd fractionation in representative weathering zone profiles demonstrates subequal retention in goethite-rich settings and Pd depletion in hematite-rich zones, accurately predicting trends observed in soils and laterites. Iron oxide mineralogy, reflecting modern and past regional climate conditions, is likely the primary determinant of Pt and Pd endowment in weathering zone resources. This new model for Pt and Pd mobilization and accumulation behavior provides a mechanistic foundation for exploration and recovery of Pt group elements from ultramafic regolith deposits.