Global Optical Snow properties via High-speed Algorithm With K-means clustering (GOSHAWK)

Snow surface albedo is a crucial component to the energy balance of our seasonal snowpack on planet Earth, reflecting most of the incoming solar radiation, and maintaining cool snow surface temperatures. Snow surface albedo, as well as other optical properties (fractional snow cover and specific surface area (SSA)), can be modeled using imaging spectroscopy measurements. The added spectral information, as compared to multispectral remote sensing, enables us to reduce errors by providing information to solve spatially heterogeneous mixed pixels. However, there is a computational burden to solve due to the added complexity of hundreds of spectral bands at 30 meter resolution. To help address these challenges, we present a fast open-source algorithm for computing snow surface properties from imaging spectroscopy, which we refer to as Global Optical Snow properties via High-speed Algorithm With K-means clustering (GOSHAWK). In this brief paper, we present the current algorithm methodology as well as validation to net-radiometers across North America.