Actinometric measurements of photolysis rate coefficients within artificial snow have been used to test calculations of these coefficients by two radiative transfer models. The models used were based upon the delta-Eddington method or the discrete ordinate method, as implemented in the tropospheric ultraviolet and visible snow model, and were constrained by irradiance measurements and light attenuation profiles within the artificial snow. Actinometric measurements of the photolysis rate coefficient were made by observing the unimolecular conversion of 2-nitrobenzaldehyde (NBA) to its photoproduct under ultraviolet irradiation. A control experiment using liquid solutions of NBA determined that the quantum yield for conversion was $phi$ = 0.41 ¿ 0.04 (¿2σ). Measured photolysis rate coefficients in the artificial snow are enhanced in the near-surface layer, as predicted in the model calculations. The two models yielded essentially identical results for the depth-integrated photolysis rate coefficient of NBA, and their results quantitatively agreed with the actinometric measurements within the experimental precision of the measurement (¿10%, ¿2σ). The study shows that these models accurately determine snowpack actinic fluxes. To calculate in-snow photolysis rates for a molecule of interest, one must also have knowledge of the absorption spectrum and quantum yield for the specific photoprocess in addition to the actinic flux. Having demonstrated that the actinic flux is well determined by these models, we find that the major remaining uncertainty in prediction of snowpack photochemical rates is the measurement of these molecular photophysical properties.