Puistu esimese rinde võrastiku alguse kõrguse hindamine lennukilidari mõõdistusandmete järgi / A simple model to estimate forest canopy base height from airborne lidar data

Abstract

Abstract Airborne laser scanner (ALS) measurements from two test sites in Estonia were used to estimate forest canopy-base height (HVL). The ALS data was collected by Estonian Land Board using Leica ALS50-II scanner. The HVL was estimated by using mode value and standard deviation of the ALS pulse reflection position height distribution. The pulse reflections which had height less than 0.5 m over the estimated digital terrain model were excluded from the analysis. In situ measurements of canopy base height (HVA) were carried out in 20 mesotrophic Norway spruce and silver birch forest stands in Järvselja and in 45, mostly Scots pine dominant, mesotrophic forest stands in Aegviidu. Determination coefficients of linear regression between HVL and HVA for both test sites were over 0.8 and the residual standard errors of the models were less than two meters. The influence of forest understory vegetation to the estimation of HVL was tested by excluding the near-to-ground vegetation reflections which had height less than 1.5 m. The test results revealed no significant impact of forest understory to the HVL models. The cross validation showed that the HVL models were independent of test sites and tree species composition. The Järvselja data based HVL model had 1.3 m negative bias if applied to Aegviidu forests and the Aegviidu data based HVL model had 1.4 m positive bias if applied to Järvselja forests. In the Aegviidu test site, difference of HVL models of coniferous and deciduous stands was tested and the difference was found not to be significant