Predicting maximum branch diameter from crown dimensions, stand characteristics and tree species

Forest resource inventories must include wood quality information to support the optimum use of wood fibre. The objective of this study was to develop models relating maximum live branch diameter (MBD), which affects lumber value, to tree and stand characteristics that can be measured through current and emerging remote sensing technologies. Using non-linear mixed effects models for six Canadian conifer species, as well as for three broad-leaved species, MBD was related to crown radius, tree height, crown length, stand basal area, and basal area of trees larger than the subject tree. Models that included only individual tree characteristics (crown radius, tree height, and crown length) did not perform as well as models that additionally included stand characteristics (stand basal area and basal area of larger trees). Models that took into account tree species performed better than models that did not; in particular, broadleaved species had much thicker branches than conifers. The best model did not show bias with respect to independent variables and had root mean square error of 0.32 cm. For the best model, prediction error was not related to silvicultural treatment. These model characteristics strongly support the potential to successfully predict MBD from remotely sensed data.