Carbon estimation using sampling to correct LiDAR-assisted enhanced forest inventory estimates

Abstract

Light Detection and Ranging (LiDAR) scanning has been increasingly applied in forest ecosystem surveys. Data from LiDAR describe forest structure and provide attribute information for forest inventory. These attributes can potentially aid in the estimation of biomass and carbon by providing sampling covariates. Therefore, this study explored the accuracy of estimating carbon storage by calibrating LiDAR attributes using list sampling with a ratio estimator. Standing tree carbon and down woody debris carbon were estimated across 10 broad forest types. LiDAR-derived gross total volume was used as a listing factor and big BAF samples to collect field data. Gross total volumes were “corrected” using a ratio estimator. The results show that standing tree carbon was 58.5 Mg C × ha-1 (± 2.9% SE), and dead woody debris carbon 1.8 Mg C × ha-1 (± 7.2% SE). With the exception of one forest type, these estimates were comparable to those derived from the carbon budget model of the Canadian forest sector (CBM-CFS3).