Improved Snow Albedo Evolution in Noah-MP Land Surface Model Coupled with a Physical Snowpack Radiative Transfer Scheme

The widely used community Noah-MP land surface model currently adopts snow albedo parameterizations that are semiphysical in nature and have systematic biases which impact the accuracy of weather and climate modeling systems that use Noah-MP as the land component. We hypothesized that integrating the snowpack radiative transfer scheme from the latest version of the Snow, Ice, and Aerosol Radiative (SNICAR) model can improve the physical representation of snow albedo processes and reduce corresponding land model uncertainties. Therefore, we evaluate Noah-MP simulations employing the SNICAR scheme and compare model accuracy to a Noah-MP simulation using the default semiphysical Biosphere-Atmosphere Transfer Scheme (BATS) scheme using in situ spectral snow albedo observations at three Rocky Mountain field stations. The agreement between simulated and in situ observed ground snow albedo is significantly enhanced in NoahMP–SNICAR simulations relative to NoahMP–BATS simulations (root-mean-square error reductions from 0.116 to 0.103). Especially, NoahMP–SNICAR improves modeled snow albedo variability for fresh snow and aged snow-pack (correlation increase from 0.42 to 0.67). The underestimated variability of snow albedo in NoahMP–BATS is a result of inadequate representation of physical linkages between snow albedo evolution and environmental/snowpack conditions (temperature, snow density, snow water equivalent, and light-absorbing particles), which is substantially improved by the NoahMP–SNICAR scheme. This new development of NoahMP–SNICAR physics provides a means to improve snow al-bedo accuracy and reduce corresponding uncertainties while providing new modeling capabilities such as hyperspectral snow albedo and effects of snow grain size, snow grain shape, and light-absorbing particles in future studies.