Estimating vertical canopy cover with terrestrial and airborne laser scanning LAURI KORHONEN*†, HARRI KAARTINEN‡, ANTERO KUKKO‡, SVEIN SOLBERG§ and RASMUS ASTRUP§ † University of Eastern Finland, School of Forest Sciences, P.O Box 111, FI-80101 Joensuu, Finland ‡Finnish Geodetic Institute, Geodeetinrinne 2, P.O. Box 15, FI-02431 Masala, Finland §Norwegian Forest and Landscape Institute, P.O. Box 115, N-1431 Ås, Norway Forest canopy cover (CC) is an important ecological variable and also a basis for the international definition of forest. CC is defined as the proportion of forest floor covered by the vertical projection of the tree crowns, i.e. unbiased CC measurements should be made using vertical observations. Only gaps between the crowns should be considered. If instruments having a non-zero angle of view are used to map the canopy, sides of the crowns will also be observed, which leads to overestimation of CC. Both airborne (ALS) and terrestrial (TLS) scanning lasers measure the canopy with non-vertical laser beams, i.e. CC estimates are likely to be biased. We measured CC at 16 plots located in Eastern Finland and Southern Norway with a sighting tube to obtain an unbiased field CC, and compared these results to ALS and TLS-based estimates. In case of ALS, the simple proportion of single and first canopy echoes estimated CC very well with a small overestimation (absolute RMSE 3.7–7.6%, absolute bias -3.4–-4.4%) due to relatively narrow nadir angle. The TLS scanners used phase comparison method and had a hemispherical field of view, so instead of trying to calculate the proportion of canopy echoes, we used the points above the height threshold (1.3 m) to create a raster map of the canopy. In the initial image, the brightness of the 4- cm pixel was related to the number of echoes at the pixel. The image was first median filtered, then processed with morphological operations to reduce noise and remove within-crown gaps, and finally binarized to separate covered and open pixels. CC was then estimated as the proportion of canopy pixels. Although TLS created a detailed canopy map close to scanning points, just a small number of echoes were received from more distant crowns, which led to underestimation of CC by 42%–0.1% (absolute RMSE 8.0–17.9%, absolute bias 6.8–13.1%). We conclude that ALS can be safely used in CC estimation despite a minor bias. TLS allows detailed canopy mapping, but a dense network of scan points is required to cover the entire plot. *Corresponding author. Email address: lauri.korhonen@uef.fi