The Pitfalls of Ignoring Topography in Snow Retrievals: A Case Study with EMIT

Snow surfaces play a key role in climate change, as increased melting over the past decades significantly contributes to global sea level rise. Melting dynamics are controlled by the amount of solar radiation absorbed by the surface. Radiative forcing increases when snow is darkened by the accumulation of small light-absorbing particles (LAPs) such as dust or algae. Consequently, quantifying these particles is essential for predicting melt rates on ice sheets and mountainous glaciers, and for assessing the associated impacts on climate change. A new generation of orbital VSWIR imaging spectrometers provides the measurements needed to achieve this objective by featuring narrow, continuous spectral channels that are able to resolve subtle LAP absorption features. In particular, NASA’s Earth Surface Mineral Dust Source Investigation (EMIT), launched in July 2022 and installed on the International Space Station (ISS), aims to improve our understanding of the Earth’s mineral dust sources and their climate impacts. These impacts include dust deposition on snow surfaces in mountainous regions, such as the Western US and the Andes in South America. Recent work has demonstrated that a simultaneous inversion of atmosphere and surface state using optimal estimation (OE) shows promising potential to quantify snow biogeophysical properties from space. However, accurate retrievals require precise accounting for surface reflectance anisotropy and observation/illumination geometry. In this contribution, we present a modification of the algorithm by using a discrete anisotropic surface-atmosphere radiative transfer model that couples the MODTRAN code with a combination of Mie scattering calculations and the multistream DISORT program. The model provides a physics-based parameterization of the surface, including illumination and observation angles, whose consideration is particularly important for strongly forward scattering snow surfaces. The new implementation leads to a well posed retrieval problem by significantly reducing the number of state vector elements. We apply the approach to selected EMIT images from Patagonia, South America, resulting in a detailed outline of retrieval sensitivity to topography. In particular, we highlight increased retrieval errors in snow grain size of up to 200 μm and in dust mass mixing ratio of up to 75 μg/g when topographic characteristics are not accounted for. Furthermore, we demonstrate differences in LAP radiative forcing of up to 400 W/m^2 in cases of inaccurately quantified LAP concentration. And finally, we evidence that erroneous assumptions about surface topography are one of the major causes for the formation of the ``blue hook" in remotely sensed retrievals of snow reflectance. These findings will be essential for updating melt runoff and climate model input, but also for the conception of retrieval algorithms for upcoming global spaceborne imaging spectroscopy missions, including NASA’s Surface Biology and Geology (SBG).