Estimating forest biomass components with hemispherical photography for Douglas-fir stands in northwest Oregon

Abstract

Accurately and quickly identifying inventories of forest biomass has become increasingly important for a variety of reasons. Current allometric equations require time-consuming tree-level measurements, but ground-based remote sensing could lead to faster estimates of forest biomass. Hemispherical photography (HP) is one potential technology that could estimate forest biomass quickly and efficiently. This analysis is based on a study in northwest Oregon where 15 Douglas-fir ( Pseudotsuga menziesii (Mirb.) Franco) plots were destructively sampled, and 60 HPs (four per plot) were taken. One photograph was taken after removing each quartile of a plot (by basal area). Two subsets of bone-dry biomass were measured and estimated: (i) crown and branch biomass (CBB) and (ii) total aboveground biomass (AGB). AGB ranged from 136 to 423 Mg/ha, and CBB ranged from 26 to 68 Mg/ha. A regression analysis between actual and HP estimated biomass showed that the average of the top two and top three 18° zenith angles resulted in the highest correlation and lowest RMSE for both CBB and AGB, while an estimate of plant area index over the top three zenith angles had the lowest correlation. HP estimates are compared with two allometric equations: one based on a regional study and one based on a national compilation of studies.