An approach to estimating forest biomass change over a coniferous forest landscape based on tree-level analysis from repeated lidar surveys

摘要

Forests represent a significant opportunity for carbon sequestration, but quantifying biomass change at the landscape scale and larger remains a challenge. Here we develop an approach based on repeated tree-level analysis using high-resolution airborne lidar (around 8 pulses/m(2)). The study area was 53 km(2) of actively managed coniferous forestland in the Coast Range Mountains in western Oregon. The study interval was 2006-2012. Tree heights and crown areas were determined from the lidar data using point cloud clustering. Biomass per tree was estimated with allometry. Tree-level data (N = 14,709) from local USDA Forest Service Forest Inventory and Analysis plots provided the basis for the allometry. Estimated biomass change over the 6-year interval averaged -1.3 kg m(-2) year(-1), with the average gain in undisturbed areas of 1.0 kg m(-2) year(-1). Full harvest occurred on 3% of the area per year. For surviving trees, the mean change in height was 0.5 m year(-1) (SD = 0.3) and the mean change in biomass was 45.3 kg year(-1) (SD = 6.7). The maximum bin-average increase in biomass per tree (57.3 kg year(-1)) was observed in trees of intermediate height (35-40 m). In addition to high spatially resolved tracking of forest biomass change, potential applications of repeated tree-level surveys include analysis of mortality. In this relatively productive forest landscape, an interval of 6 years between lidar acquisitions was adequate to resolve significant changes in tree height and area-wide biomass.