Synthetic aperture radar (SAR) is a powerful tool for understanding planetary surfaces as the intensity of backscatter depends on the chemical and mineralogical (e.g., dielectric) and physical (e.g., roughness) properties of the surface. For sedimentary deposits in arid to hyperarid deserts imaged at decimeterscale wavelengths, grain-scale properties directly influence the radar backscatter. Thus, fine-scale granulometric information can be assessed through such remote sensing. A Mie scattering model informs our expectation that SAR backscatter should increase with coarser grain size and greater roundness. We tested the response of radar backscatter to grain shape and size sorting based on field measurements near Death Valley, California, USA. Results show a strong positive correlation between C-band (lambda = 5.55 cm) and Lband (lambda = 23.5 cm) backscatter to grain size over the tested median b-axis size range of 0.5-18.6 cm. Very weak or negligible effects of roundness and sorting are indicated. Endmember measurements of classic roughness parameters (i.e., root mean square height and correlation length) at the study sites also predict the radar backscatter, though with less statistical power. The field results are consistent with a Mie scattering model if the higher backscatter magnitude in the model is explained by the shorter radar wavelength. Additionally, the limited success of the adthe importance of further studying radar backscatter for very rough surfaces with realistic properties. Nonetheless, the strong positive correlation between grain size and radar backscatter supports the remote charenvironments. With SAR-based estimates of grain size, shape, sorting, and the spatial variation of such properties as provided in this work, the style and intensity of geologic processes on remote terrestrial and planetary surfaces can be effectively studied.