Quantifying timber value and carbon in structurally complex woodlands using terrestrial laser scanning

Abstract

The use of domestically produced timber in construction is key to achieving net-zero targets. Recent advancements in lidar (light detection and ranging) remote sensing technologies offer the potential to transition from qualitative assessments of stem form in forest inventories to systems that provide high-resolution, quantitative measurements of tree architecture enabling the accurate prediction of timber product output. Such improvements in the quantitative surveying of trees in complex stands—those challenging to measure using conventional inventory techniques—provide an opportunity to increase the utilisation of home-grown timber for high-value harvested wood products (HWPs). This could add value to trees that might otherwise be used for chip, pulpwood, or firewood, increasing the storage time of the carbon sequestered during trees’ growth. Furthermore, improvements to the surveying of complex trees may enable improvements to the management of marginal areas of woodland such as those found on farmland. In this study, I present a novel application of Terrestrial Laser Scanning (TLS) to quantify variation in stem form metrics which are relevant to timber properties in marginal woodlands in Wales. TLS surveys were conducted in 11 plots at 8 different sites along a gradient of forest complexity. TLS-derived measurements of stem diameter, tree height, stem lean, stem straightness and crown base height were compared against manual tree measurements collected using conventional techniques. TLS accurately replicated manually measured metrics, including tree DBH (RMSE 2.2cm (7.52 %), R2: 0.97, p < 0.001), tree height (RMSE 2.45 m (12.75 %), R2: 0.87, p < 0.001), and stem lean (RMSE, R2: 0.276, p < 0.001). Stem form metrics obtained from processing of TLS data were subsequently used to predict potential timber product output, based on variations in stem straightness and position within the tree. Operational sawlog length and grade (green—higher-quality or red —lower-quality) were assigned based on straightness criteria and each log’s position within a tree. An estimated total of 127 sawlogs were harvestable from the 60 surveyed trees across all sites comprising 40 green and 87 red sawlogs with a total value of £6,002. The highest value plot, a stand of large, straight oak trees with a clear understorey, contained a potential total timber value of £3,349. The lowest-value plot, an under-managed stand of densely planted Sitka spruce, was worth only £77. TLS-determined timber valuations showed good agreement with those of an expert consultant (ordinal logistic regression t = -3.64, coefficient = -0.023 ± 0.006 SE, p <0.001). Finally, the assessment of tree carbon content (Mg C) showed strong agreement between the two methods, though TLS consistently underestimated tree carbon content when compared with allometrically-derived carbon values (RMSE 0.186 Mg C (65.5%), R2 = 0.94, p < 0.001). By demonstrating the potential use of lidar instruments for predictive assessment of forest product outputs and carbon, the results of this study present a case for the development and utilisation of lidar instruments in forest inventory applications, even in complex woodlands. By enabling more precise surveying of trees, lidar surveying presents an opportunity for finer-resolution timber product procurement, enhancing the value retention of otherwise underutilised woodlands.

Date of Award	19 Nov 2025
Original language	English
Awarding Institution	Bangor University
Sponsors	Woodknowledge Wales
Supervisor	John Healey (Supervisor), Marielle Smith (Supervisor) & Matthew Drummond (Supervisor)