Återgå till huvudnavigering Återgå till sök Gå direkt till huvudinnehållet

A poststratified ratio estimator for model-assisted biomass estimation in sample-based airborne laser scanning surveys

  • Anna Ringvall
  • , Göran Ståhl
  • , Liviu Theodor Ene
  • , Erik Næsset
  • , Terje Gobakken
  • , Timothy G. Gregoire

Publikation: Bidrag till tidskrift › Artikel i vetenskaplig tidskrift › Peer review

Sammanfattning

To estimate the aboveground biomass (AGB) for large areas, two-stage sampling designs using airborne laser scanning (ALS) as a strip sampling tool in combination with subsampling of field plots have been successfully applied in several studies. However, the studies have pointed to problems in the proposed estimator, partly related to the unequal length of flight lines in irregularly shaped areas. In this article, we present a model-assisted ratio estimator for such two-stage designs utilizing the area of the ALS strip as the auxiliary variable. The proposed estimator is further developed for estimation in subpopulations and for poststratified estimation. When deriving a variance estimator of the poststratified estimator, we considered the dependencies between estimates from different strata that arise since flight lines extend over several strata. An evaluation by simulated sampling in an artificial population based on data from a survey in Hedmark County, Norway, showed that the proposed estimators and their variance estimators performed well in the case of simple random sampling in both stages. In such cases, the ratio and poststratified estimators improved the precision of AGB estimates by 30% and 70%, respectively, in comparison with the earlier suggested estimator.
OriginalspråkEngelska
Sidor (från-till)1386-1395
Antal sidor10
TidskriftCanadian Journal of Forest Research
Volym46
Nummer11
DOI
StatusPublicerad - 2016

Nyckelord

  • two-stage sampling
  • large-scale surveys
  • LULUCF reporting
  • REDD

Citera det här