Canopy architecture, the geometric organization of aboveground vegetation, provides an essential link between patterns observable by remote sensing and fundamental ecological processes. As part of the international Boreal Ecosystem-Atmosphere Study (BOREAS); (1) we developed a hierarchical approach for sampling canopy architecture, and (2) we acquired a comprehensive data set for characterizing canopy architecture for the major BOREAS study sites. The approach involves a series of four sets of measurements at different spatial scales, ranging from the ecosystem to the leaf level: (1) regional characterization, involving measurements of the climate, ecosystem, and landscape features; (2) stand characterization, involving measurements of crown geometry (diameter at breast height (DBH), height, and crown extent), individual tree location, and understory cover; (3) tree vectorization, involving detailed sampling of the three-dimensional distribution of canopy elements and crown form; and (4) characterization of canopy geometry as seen from beneath, involving acquisition of a multitemporal catalog of hemispherical photographs. The last three sets of measurements were then used to reconstruct the three-dimensional geometry of the canopy. By comparing simulated hemispherical views upward from beneath this reconstructed canopy with in situ hemispherical photographs, the methodological approach was validated. Simulated photographs faithfully reproduced patterns observed for in situ hemispherical photographs, in particular, for gap fraction distributions of the middle ranges of zenith angles (20¿--70¿). Moreover, simulation of the gap fraction for this middle portion of the zenith angle was insensitive to exact mapping of the stand. The hierarchical data acquisition approach, involving mapping of tree locations and tree reconstruction, permits realistic representation of canopy material distribution. Our approach and our comprehensive data set provide a solid basis from which to integrate data gathered at the stand and tree scales, and a powerful tool for the simulation of the light regime anywhere in the canopy. ¿ 1997 American Geophysical Union