Post-stratified change estimation for large-area forest biomass using repeated ALS strip sampling

Abstract

Post-stratified model-assisted (MA) and hybrid (HY) estimators are used with repeated airborne laser scanning (ALS) strip sampling and national forest inventory field data for stratum-wise and overall estimation of aboveground biomass (AGB) stock and change. The study area covered the southern portion of the Hedmark County in Norway. Both MA and HY estimation substantially reduced the uncertainty in AGB change when compared with estimation using the field survey only. Relative efficiencies (relative variance) of 4.15 (MA) and 3.36 (HY) for overall estimates were found. The results suggest the MA estimator for single-time estimation and the HY as more appropriate for change estimation by cover class. With the HY estimator, a nested post-stratification scheme is demonstrated, combining cover classes with change classes, which enables detailed reporting for change according to cause within each cover class, and has the potential to improve the estimation precision. Finally, parametric bootstrapping is demonstrated as an empirical alternative to estimate the model-error component in the HY estimator. The model error estimated with parametric bootstrapping converged to the analytically determined value of the HY estimator within 1000 bootstrap samples.

A Model-Dependent Method for Monitoring Subtle Changes in Vegetation Height in the Boreal–Alpine Ecotone Using Bi-Temporal, Three Dimensional Point Data from Airborne Laser Scanning

“…In the current study, we adopted parametric bootstrap, as in Ene et al [61]. The technique has also been demonstrated recently in remote sensing studies by Bollandsås et al [60] and Strîmbu et al [62]. In the following, it is assumed that the model parameter estimates of the predictive models (β in the models in Equations (1) and (2) are independent (see Section 2.8) and follow asymptotically multivariate normal distributions, i.e.,…”

Modelling above-ground biomass stock over Norway using national forest inventory data with ArcticDEM and Sentinel-2 data

“…The total estimate for the studied period was smaller than all ΔAGB from previous five-years NFI cycles from 2010 until 2018 (see Figure 3), mainly due to a larger number of plots with AGB loss during the studied period (2014-2019) compared to the earlier periods. The direct estimate of yearly average ΔAGB was (0.03 t ha -1 year -1 ) was smaller than the estimates by Naesset et al (2013) in productive forests between 2000 and 2010 (1.78 t ha -1 year -1 ), the estimates by Bollandsås et al (2018) in montane forests in the period between 2008 and 2012 (0.21 t ha -1 year -1 ) or the estimate by Strîmbu et al (2017) for an areal extent similar to this study for the period 2006 -2011 (1.27 t ha -1 year -1 ). Even though the abovementioned studies were conducted in different areas and for different periods, the reported estimates are in line with the decreasing ΔAGB trend visible in our study area in the period 2010 -2020 ( Figure 3).…”

Comparing 3D Point Cloud Data from Laser Scanning and Digital Aerial Photogrammetry for Height Estimation of Small Trees and Other Vegetation in a Boreal–Alpine Ecotone

“…Under the model-dependent inferential framework, the variance due to uncertainty of the model parameters can be approximated either by using a closed-form formula based on a first-order Taylor series approximation (see e.g., [57]), or by using Monte Carlo simulations in the form of, for example, parametric bootstrap. Both approaches to estimation have been adopted in forest-and vegetation-related studies in recent years (e.g., [50,59]). Ref.…”