Experimental observations have shown photochemical production in the upper layers of firn. We use a modeling approach for Antarctica in summer, calculating the actinic flux in snowpack, and estimating NOX production from nitrate photolysis. Assuming nitrate photolysis is about as efficient for ice as for aqueous solution, and that all nitrate in the ice is available for photolysis, we find good agreement with measured fluxes of NOX from the snow surface at Neumayer Station, Antarctica. We estimate fluxes for other sites, and confirm that they could significantly affect boundary layer chemistry, especially where an atmospheric surface temperature inversion is present. We find that there is considerably more NOX production in the upper snowpack than in the entire tropospheric column above it. Photolysis explains a proportion of the nitrate loss observed from snow at some sites, but other processes (sorption/desorption) are necessary to explain the magnitude and depth profiles observed.