Effects of turnover and internal variability of tree root systems on modelling coarse root architecture: comparing simulations for young Populus deltoides with field data

The behaviour of the FracRoot model, which describes coarse-root architecture based on proximal root diameters using a recursive algorithm, was studied using field data on Populus deltoides Bartr. ex Marsh. in a short-rotation plantation. Roots were described as a branching, hierarchical network of connected links. Diameters of daughter links were estimated from the diameter of the mother link using scaling parameter p and allocation parameter q, which were based on sharing of the root cross-sectional area by daughter links. Parameters were estimated from complete root excavations. The length of each link and the vertical and horizontal branching angle distributions were recorded. Parameter p values were distributed log-normally, and q values followed a beta distribution both for the whole root system and within 5 mm link diameter classes. Including the variability of p and q in the model did not significantly improve root length estimates compared with the use of mean p and q values over all branching points. Including a coarse-root turnover factor based on field-observed evidence on root mortality improved the model fit to field data. Root length was more sensitive to parameter values and turnover factor than root mass. Field observations and the importance of the turnover factor to simulation accuracy suggest that coarse-root turnover should be considered in root research, at least under conditions of strong competition or other external stress.