Boreal forests cover about a fifth of seasonally snow-covered land over the Northern Hemisphere. Enhancement of longwave radiation beneath coniferous forests has been found to impact the surface energy balance and rates of snowmelt. Although the skill of model-simulated snowmelt has been shown to be lower for forests than for open areas, model intercomparisons and evaluations of model parameterizations have not yet focused on longwave enhancement. This study uses stand-scale forcing for the simulation of subcanopy longwave radiation by Community Land Model version 4.5 (CLM4.5) and to drive SNOWPACK, a snow model featuring more complex canopy structure, as a benchmark model for CLM4.5. Simulated subcanopy longwave radiation and longwave enhancement are assessed using measurements from forest stands located within perennially snow-covered regions. These forest stands, of varying canopy density, cover the range of boreal plant functional types in CLM4.5. CLM4.5 is found to overestimate the diurnal range of subcanopy longwave radiation and longwave enhancement, and simulation errors increase with decreasing cloudiness and increasing vegetation density. Implementation of a parameterization of heat storage by biomass reduces simulation errors but only marginally affects the amplitude of diurnal ranges. These results reaffirm previous findings that simulation of subcanopy longwave radiation can be improved by partitioning the vegetation canopy into two layers. Moreover, this study reveals the variations of simulation errors across meteorological conditions and vegetation density, the latter of which is the most important parameter for longwave enhancement independent of vegetation type.