A Dynamic 3D Coarse Root Architecture Model to Study Anchorage of Forest Trees

Storms cause half of primary damages to European forests, mainly by anchorage failure. Anchorage is largely determined by 3D coarse root architecture during the tree lifespan. A modeling approach is required in order to study the time trends in mechanical resistance of root system including change due to soil constraints and wind acclimation.

We modified and parametrized the "RootTyp" model (Pagès et al., 2004) to simulate the tree root system development until mature age by using the datasets of 3D tree root systems for Pinus pinaster. This database was structured in a chronosequence from 1 to 50 years old.

As suggested in its name, the model RootTyp is based on predefined root categories which can be examined for their potential contribution to anchorage. Six root types were identified using a k-means clustering algorithm and internal cluster validity indices from four growth variables observed in the database: branching angle, tropism, tropism of mother root and growth potential. To complete these first associated sets of parameters, some parameters difficult to measure (e.g. maximum root length for adult specimen) were estimated from literature or using a differential evolution optimization method.

As root anchorage can be altered by waterlogging or a hard pan, we took into account specific soil characteristics and the resulting alterations of root system development. Five standardized variables (e.g. branching rate, tropism) along a depth gradient were observed. The soil model coefficients influencing the growth processes were calibrated to simulate hard pan effects.

The model performance was assessed using two methods. The first was a qualitative evaluation, where the geometrical aspects and the biological realism of the model were visually compared to observed root systems. The second comprised statistical tests on global (e.g. total root volume) and specific (e.g. taproot diameter distribution) root variables.