Abstract
In this work, we present a study about the application of bi-temporal, large interval aerial laser scans for constructing of tree growth models and estimating site index quality based on the measured increments from the laser scans. We compared two LiDAR scans with 14 years of difference in the national park area, where most areas are unmanaged. We derived the increment curve based on the Chapman-Richard growth formula. We used site index estimates from forest management plans from the national scale as the ground truth (both absolute and relative). We constructed three predictive models for site index estimates from bi-temporal scans, in modalities with and without stand age. Including the stand age improved all models, but even without the age, the models performed relatively well for differentiation between better and worse sites. At this moment, it is not directly possible to estimate age from remotely sensed data, but consistent monitoring, with laser scanning or photogrammetry, undoubtedly detects the harvest or dieback, so in the future, age can be considered as a variable easily estimated from remotely sensed data and so remote sensed material are viable source for understanding of forest growth and production.