Vertical Height Errors in Digital Terrain Models Derived from Airborne Laser Scanner Data in a Boreal-Alpine Ecotone in Norway

摘要

It has been suggested that airborne laser scanning (ALS) could be used for operational monitoring of vegetation changes in the alpine tree line caused by climate change. Because the vegetation is low in such tree-less areas close to the alpine zone, the accuracy of the digital terrain model (DTM) becomes crucial for early detection of, e.g., pioneer trees representing an ongoing tree migration given that the height of the vegetation may be on the same order of magnitude as the DTM uncertainty. The goal of this study was to assess and exemplify the vertical height errors of DTMs derived from ALS data under varying flying altitudes and pulse repetition frequencies (PRF). Important effects in the analysis were local terrain form, terrain surface, ground vegetation height, and terrain slope, because they may be correlated with recruitment patterns of pioneer trees. Based on 426 ground control points collected in a boreal-alpine ecotone, a standard deviation of 0.07–0.08 m was found for the lowest flying altitudes and lowest PRFs. For the highest PRF the standard deviation was 0.13 m. There were statistically significant mean errors for the different terrain forms and ground vegetation heights (−0.11 to 0.13 m).