Abstract: Organic contaminants in the vadose zone pose long-term threats to groundwater, ecosystems, and human health. Accurate characterization of their distribution and movement is critical for environmental risk assessment and remediation strategies. Low-field nuclear magnetic resonance (NMR), a noninvasive and cost-effective geophysical method sensitive to hydrogen-bearing fluids, holds great potential for characterizing and tracking organic pollutants in the vadose zone. However, the interpretation of NMR signals is challenging because vadose zone organic contaminants exist as multiphase fluids. In this work, we aim to analyze dynamic pore-scale displacement of water by petroleum hydrocarbons (represented by n-decane) in porous media visualized by X-ray computed tomography images, and further analyze the NMR responses under varying hydrocarbon-water distribution states. The results indicate that, compared to 1D T₂ spectra, 2D D-T₂ spectra can quantitatively reveal that the non-wetting phase (hydrocarbons) preferentially occupies larger pore spaces and the central (bulk) regions of the pores during the displacement process. This work enhances the accuracy of interpreting organic contaminant transport and distribution using vadose zone NMR data.