Harmonizing multi-temporal airborne laser scanning point clouds to derive periodic annual height increments in temperate mixedwood forests

Abstract

When combining multi-temporal airborne laser scanning (ALS) data sets, forest height growth assessments can be compromised due to variations in ALS acquisitions. Herein, we demonstrate the importance of assessing and harmonizing the vertical alignment of multi-temporal ALS data sets used for height growth calculations. Using four ALS acquisitions (2005–2018) in a temperate mixedwood forest, we developed an ALS data harmonization approach and quantified the impact of the harmonization on derived height periodic annual increment (PAI), comparing the ALS-derived PAI to PAI derived from non-harmonized ALS data sets and field measurements. We found significant differences in PAI derived from harmonized and non-harmonized data, and these differences were greater for shorter growth intervals. Data harmonization resulted in a consistent PAI series that reduced uncertainties associated with the different ALS acquisitions. Although overall there was a strong relationship between field and ALS height measures ( R2 ≥ 0.88), we found a weak relationship between the field- and ALS-derived PAI ( R2 = 0.12). We identified systematic errors in field-based tree height measures in plots with complex crowns, tall trees, and restricted visibility. We demonstrate the need for harmonizing multi-temporal ALS data sets for the generation of PAI and, likewise, highlight the need of carefully scrutinize field-measured heights and associated increments.