Dispersed and piled woody residues volumes in coastal Douglas-fir cutblocks determined using high-resolution imagery from a UAV and from ground-based surveys.

Abstract:

After forest harvest significant amounts of woody residues are left dispersed on site and some subsequently piled and burned. Quantification of residues is required for estimating C budgets, billable waste, harvest efficiency, bioenergy potential and smoke emissions. Trofymow (et al 2014 CJFR) compared remote sensing methods to ground-based waste and residue survey (WRS) methods for residue piles in 4 cutblocks in the Oyster River (OR) area in coastal BC. Compared to geospatial methods using 15cm orthophotos and LiDAR acquired in 2011 by helicopter, the WRS method underestimated pile wood by 30% to 50% while a USFS volume method overestimated pile wood by 50% if site specific packing ratios were not used. A geospatial method was developed in PCI Geomatica to analyze 2-bit images of logs >15cm diameters to determine dispersed wood residues in OR and compare to WRS methods. Across blocks, geospatial and WRS method wood volumes were correlated (R²=0.69), however volumes were 2.5 times larger for the geospatial vs WRS method. Methods for dispersed residues could not be properly compared as individual WRS plots were not georeferenced, only 12 plots were sampled in total, and low-resolution images poorly resolved logs. Thus, a new study in 2 cutblocks in the Northwest Bay (NWB) area acquired 2cm resolution RGB air-photography in 2014-15 using an Aeryon Sky Ranger UAV prior to and after burn pile construction. A total of 57 dispersed WRS plots and 24 WRS pile or accumulation plots were georeferenced and measured. Stero-pairs were used to generate point-clouds for pile bulk volumes. Images processed to 8-bit grey scale are being analyzed with a revised PCI method that better accounts for log overlaps. WRS methods depend on a good sample of plots and accurate determination of stratum (dispersed, roadside, piles, accumulations) areas. Analysis of NWB blocks shows WRS field methods for stratum area differ by 5-20% from that determined using orthophotos. Plot-level wood volumes in each plot and stratum determined by geospatial and WRS methods will be compared. While geospatial methods for residue determination is a 100% sample, compared to sample-based WRS method, difficulties in resolving logs in the images may mean the best method for determining residues requires a combination of geospatial and ground plot measurements.

.