A revised terrain correction method for forest canopy height estimation using ICESat/GLAS data

Abstract

Although spaceborne Geoscience Laser Altimeter System (GLAS) can measure forest canopy height directly, the measurement accuracy is often affected by footprint size, shape and orientation, and terrain slope. Previous terrain correction methods only took into account the effect of terrain slope and footprint size when estimating forest canopy height. In this study, an improved terrain correction method was proposed to remove the effect of all aforementioned factors when estimating canopy height over sloped terrains. The revised method was found significantly better than the traditional ones according to the canopy height tested using small footprint LiDAR data in China. It reduced the RMSE of the canopy height estimates by up to 1.2m. The effect of slope on canopy height estimation is almost eliminated by the proposed method since the slope had little correlation with the canopy heights estimated by revised method. When the footprint eccentricity is small, the canopy height error due to the footprint shape and orientation is small. However, when the footprint eccentricity is large enough, the height estimation error due to footprint shape and orientation is large. Therefore, it is necessary to take into account the influence of footprint shape and orientation on forest canopy estimation.