Nationwide airborne laser scanning based models for volume, biomass and dominant height in Finland

• DOI: 10.14214/sf.1567
• Publication Date: 2016-07-08T06:10:51Z
• Authors: Eetu Kotivuori; Lauri Korhonen; Petteri Packalen
• Keywords: Laser Scanning; Sample (material)

<ja:p>The aim of this study was to examine how well stem volume, above-ground biomass and dominant height can be predicted using nationwide airborne laser scanning (ALS) based regression models. The material consisted nine practical ALS inventory projects taken from different parts Finland. We used field sample plots data create regional models for each response variable. final had one or two predictors, which were chosen on the root mean square error (RMSE), cross-validated. Finally, we tested much predictions would improve if calibrated with a small number plots. Although forest structures differ among Finland, volume performed quite (leave-inventory-area-out RMSE 22.3% 33.8%, difference [MD] –13.8% 18.7%) compared (leave-plot-out 20.2% 26.8%). However, model (RMSE 5.4% 7.7%, MD –2.0% 2.8%, exception Tornio region – 11.4%, –9.1%) nearly as 5.2% 6.7%). results show that provided means than real at level, because structure device have considerable effect predictions. Large MDs appeared especially in northern Local calibration decreased did not benefit calibration.</ja:p>