Mapping vegetation height over large areas presents a prob-lem of scale: height varies with the individual tree or stand, but the resolution of available datasets is too low to char-acterize this variability sufficiently for many applications. We address this problem by fusing 1 km resolution canopy height data derived from satellite-based laser altimetry with higher-resolution land-cover data, resulting in 30 m resolution estimates of canopy height. These are downscaled further to 1 m resolution by simulating individual trees. A web service architecture is used, which allows processing to occur on demand without preprocessing large datasets. We compared the resulting canopy volumes to reference airborne lidar data from 262 randomly located 1 km 2 areas within nine study sites. Results at 30 m resolution show an RMSE of 33 percent of the mean reference volume and an R 2 of 0.77; at 1 m the RMSE is 66 percent and the R 2 is 0.38. Introduction Vegetation height is a key measurement used to estimate a variety of ecological and biophysical variables, including above-ground biomass, surface roughness, and stem volume. Global large-footprint lidar data from the Geoscience Laser Altimeter System (