<p>The physical properties of snow, such as its stiffness, reflectivity, and thermal conductivity, are critical components in feedback processes of the Earth’s system and useful proxies for various applications in environmental science, ranging from avalanche forecasting to meteorology. It is therefore important to efficiently and accurately determine snow properties not only in the laboratory, but also during field campaigns. One promising approach is to measure the snow’s optical properties and deduce material properties via theory; from which physical properties can in turn be derived. Most notably, this applies to the determination of the snow’s specific surface area from total diffuse reflectance measurements. The retrieval of another important snow parameter, its mass density, from diffuse reflectance measurements has remained elusive. Here, we outline a theoretical description within the diffusion approximation of the radiative transfer theory to retrieve the density of dry snow from optical measurements via spatial truncation of the diffuse reflectance. Using our model, we determine snow density profiles from partial diffuse reflectance images given prior knowledge of the snow’s specific surface area. Beyond field measurements, our results are mappable to other applications relying on sub-surface light scattering, including remote sensing and biomedical applications.</p>