SEM and Machine Learning Root Porosity Dataset from Mississippi River Delta Plain Sites

This dataset contains scanning electron microscopy (SEM) images and machine learning (ML)-based segmentation results for quantifying internal porosity of Spartina roots collected from four marsh sites in the Mississippi River Delta Plain (MRDP). The dataset supports analysis of biomechanical trade-offs between tensile strength and porosity in wetland plant roots. The workflow combined high-resolution SEM imaging with a Random Forest classifier trained on manually labeled images to segment root cross-sections into background, voids, and tissue. Post-processing included pore size filtering (removing small speckle noise and retaining the largest, structurally relevant voids above the 98th percentile).