Airborne laser scanning for stand-wise forest inventories: a case study in Lithuania

Abstract

This paper presents the approaches used to integrate airborne laser scanning (ALS) data into Lithuanian stand-wise forest inventory together with some results to estimate the volume of forest compartments in a test area located in central part of Lithuania. Digital color infrared aerial images, LiDAR point clouds, based on 1 point/m2 scanning density, field measurement data on 439 circular sample plots were made available for the study area. Two phase sampling with nonparametric most similar neighbor estimator was used to get continuous surfaces of volume per 1 ha estimates, which were later aggregated for forest compartments. Three aggregation techniques were studied: (i) direct aggregation for the compartment; (ii) segmentation of ALS data, volume estimation for segments and generating statistics for compartments based on the segments and (iii) segment based aggregation adjusted by the area of segments. ALS data based volume estimates were found to be around 10% larger than the ones available from updated stand registers, originating prom previous stand-wise forest inventories. Such differences are explained by the tendency of stand-wise inventories to decrease volume estimates. Recommendation made is to incorporate ALS into conventional stand-wise inventories without major changes of their design.