Snow Property Inversion From Remote Sensing (SPIReS): A Generalized Multispectral Unmixing Approach With Examples From MODIS and Landsat 8 OLI
Edward H. Bair, Timbo Stillinger, Jeff Dozier
Abstract
Spectral mixture analysis has a history in mapping snow, especially where mixed pixels prevail. Using multiple spectral bands rather than band ratios or band indices, retrievals of snow properties that affect its albedo lead to more accurate estimates than widely used age-based models of albedo evolution. Nevertheless, there is substantial room for improvement. We present the Snow Property Inversion from Remote Sensing (SPIReS) approach, offering the following improvements: 1) Solutions for grain size and concentrations of light absorbing particles are computed simultaneously; 2) Only snow and snow-free endmembers are employed; 3) Cloud-masking and smoothing are integrated; 4) Similar spectra are grouped together and interpolants are used to reduce computation time. The source codes are available in an open repository. Computation is fast enough that users can process imagery on demand. Validation of retrievals from Landsat 8 operational land imager (OLI) and moderate-resolution imaging spectroradiometer (MODIS) against WorldView-2/3 and the Airborne Snow Observatory shows accurate detection of snow and estimates of fractional snow cover. Validation of albedo shows low errors using terrain-corrected in situ measurements. We conclude by discussing the applicability of this approach to any airborne or spaceborne multispectral sensor and options to further improve retrievals.