Monitoring tree occupancy and height in the Norwegian alpine treeline using a time series of airborne laser scanner data

The main objective of this study was to demonstrate a method for monitoring tree occupancy and height in the alpine treeline ecotone using a time series of ALS data. We applied data collected in a longitudinal survey, comprising three spatially consistent campaigns from the years 2008, 2012 and 2018, on 25 sites along the Scandinavian Mountain Range (60-69 degrees N). We compared ALS-based estimates of tree occupancy and height with corresponding field-based estimates and provided ALS-based estimates of uncertainty. Cross validation of a longitudinal model for predicting tree occurrence probability from ALS data revealed an overall accuracy of 83 %. ALS data were useful for predicting the height of pioneer trees, despite sparse laser points. Both models needed to account for the time of measurement. ALS-based estimates of tree occupancy were 4.6, 6.7 and 6.0 % for the three measurement occasions, respectively, and corresponding field-based estimates were 4.3, 4.6 and 5.0 %. ALS-based estimates of tree height were 2.2, 2.1 and 2.2 m, respectively, and corresponding field-based height estimates were 2.3, 2.2 and 2.2 m. Overlapping confidence intervals of ALS-based estimates for both variables and for all three measurement occasions indicated no statistically significant changes in either of the studied variables. The proposed method can be used to monitor alpine treeline ecotones and to provide accompanying uncertainty estimates to inform whether changes are significant.