We present a study on diurnally and annually averaged values of the actinic fluxes used in one-dimensional (1-D) photochemical models, as well as a 3-D radiative transfer model, based on Monte Carlo calculations with application to the atmosphere of Titan. This study shows that the commonly used value ⟨&thgr;⟩=30¿ for the mean incident angle at the equator in photochemical models of Titan is not the best choice, though changing the value has no dramatic effects on photochemistry. The results of the 3-D code give direct access to the photolysis rates at any point in the atmosphere. The necessity of 3-D values in a deep atmosphere such as Titan's is demonstrated particularly for high-latitude winter conditions. These 3-D photolysis rates are used to model the latitudinal variations of the chemical composition of Titan's atmosphere in a 1-D photochemical model adapted to different latitudes. This study shows that these kinds of simple photochemical models cannot reproduce the observed latitudinal behavior and that we need to develop real 2-D photochemical models of Titan's atmosphere. ¿ 1999 American Geophysical Union