An approach to modelling tree root architecture in virtual urban growing conditions

Abstract

Understanding the architecture of tree roots is an important component of urban forestry management practice. Tree roots are structurally and functionally important to the survival of trees, and this can be even more so in urban environments where underground space for roots is limited. Tree root architecture models can provide a complementary approach to traditional on-site field investigation methods. Root architecture models are unique in that they can simulate the spatial arrangements of root system structure explicitly, and allow investigators to create hypothetical simulations to test their assumptions about what may be driving root growth. The use of root architecture models in the literature is extensive and may be applied in diverse streams of investigation, but their application to tree root systems is less common. This research demonstrates a root architecture model, Rootbox, as a case study in the application of plant architecture models to simulate tree root growth in urban conditions. Model parameterization was based on conformity of root simulations to tree root architecture reported in the literature. The model is deployed in four hypothetical urban soil scenarios, which are representative of planting sites commonly observed in urban settings. The analysis demonstrates that plausible tree root system architectures – specifically, commonly observed growth attributes - can be produced by Rootbox, but only after several adaptive changes to both the source code/model design are made. Custom soil models can integrate with the simulation to represent urban conditions by modifying both the growth direction and elongation of portions of the root architecture, and thus offer greater control over the output architecture. Rootbox offers a flexible method of simulating the architecture of tree root systems, but further research should focus on optimizing the model’s parameters and functions to enable greater user control over model output.