SAR Backscatter and In Situ Soil Moisture Dataset Across Multiple Soil Textures in the Berambadi Watershed, India

This dataset contains 677 paired observations of Synthetic Aperture Radar (SAR) backscatter and in situ soil moisture collected in the Berambadi watershed, Karnataka, India. The data combine Sentinel-1 (C-band) and ALOS-2 PALSAR (L-band) acquisitions with ground-based soil moisture measurements (0–5 cm depth) obtained using calibrated Hydra Probe sensors.

Each sample represents a spatially matched pair between a SAR pixel and field measurements averaged from three probes within the pixel footprint. The dataset spans five USDA soil texture classes: clay, loamy sand, sandy clay, sandy clay loam, and sandy loam.

SAR variables include co-polarized and cross-polarized backscatter coefficients (σ⁰, dB), along with derived features such as polarization ratio, total power, and SAR contrast. Additional field observations include land cover classification (bare vs vegetated) and Leaf Area Index (LAI).

This dataset is associated with the following publication:
Rao, S., Gupta, V., Murugesan, S., Zribi, M., & Muddu, S. (2026). An Information Proxy Framework That Explains Why Soil-Specific Radar Calibration Succeeds in Sandy Textures but Fails in Clay. Remote Sensing Letters. https://doi.org/10.1080/2150704X.2026.2647288

The dataset supports the analyses presented in the manuscript; full methodological details, preprocessing steps, and modeling procedures are described in the associated publication and are not repeated within this dataset.