Predicting light availability in forest understory is a key step in the modeling of forest dynamics. Aspect is often the main source of environmental variation in mountain forests, and so aspect-induced differences in crown geometry and transparency can be expected to affect light transmission through the canopy and modify understory light predictions. To gain a better understanding of the effects of aspect on the light interception capacity of forest trees, we determined crown allometry and crown openness (CO) in 120 adult trees of three common conifer species in the Pyrenees (mountain pine, Scots pine and silver fir), sampled in eight montane-subalpine forests that each included two contrasting aspects. CO was calcu-lated from digital photographs and several crown outlines were automatically determined to prevent user bias. We also calculated crown irregularity as the difference between crown surface area for the tightest and loosest outlines. Predictions of understory light availability obtained from a forest dynamics model were compared with actual values obtained from 115 hemispherical photographs. Crown length and CO varied across species following previous rankings of shade tolerance. Both pines had longer and wider crowns in the north aspect, whereas fir crown geometry was not affected by aspect. CO depended largely on the method chosen for determining crown outline, ranging from 0.10 to 0.56, and the tightest outline provided the best predictions of understory light (slope = 0.89, R 2 = 0.46). CO was also significantly affected by tree size and plot aspect (p < 0.001): crowns in southern-oriented plots were more open for mountain pine and silver fir, whereas Scots pine showed the opposite trend. Predictions of understory light were significantly improved when crown geometry and openness were allowed to vary according to plot aspect (slope of the regression: 0.95, R 2 = 0.50). Our results suggest that aspect should be explicitly considered when modeling ecological processes and dynamics in mixed mountain forests given its influence on both abiotic conditions and crown responses to them. Ó 2012 Elsevier B.V. All rights reserved.