Plant–Soil Modelling

Abstract

Plant–soil models have been rapidly developing to address several of the world's growing needs (e.g. food security, climate change, and sustainable infrastructure). With the vast pool of literature available, this article provides a broad overview of the different modelling methodologies and techniques tailored for specific needs and expertise. We partition three categories of contemporary modelling methodologies: distribution based, architecture based, and image based. We overview the different modelling techniques employed for root growth, nutrient acquisition, and water uptake and designate modelling schemes for potential beneficiaries. Subsequently, we explore modelling work pertaining to root–soil mechanical interactions. We describe the mechanics associated with roots growing through soil and benefits provided by roots with respect to slope reinforcement. Lastly, we describe how these different models can enhance our understanding of fundamental problems.

Overcoming the Challenges to Enhancing Experimental Plant Biology With Computational Modeling

“…For example, if water is stored throughout the depth of the soil, deeper rooting growth patterns are preferable; while if the soil has a low water-retention capacity, dense and shallow rooting is preferable (Leitner et al, 2014;Tron et al, 2015). Roots can be represented by the root length/surface density (unit length/surface of root per unit volume of soil) as a function of soil depth and time (Ruiz et al, 2020a;Fletcher et al, 2021) or image-resolved geometries (Ruiz et al, 2020b). The function of root, root hairs and soil aggregate geometries can be studied using image-based modeling (a mechanistic approach) using high-resolution 3D imaging of roots in soil, typically X-ray computed tomography.…”

“…Root systems which had root tips in close proximity obtained the most additional phosphorus uptake due to organic-acid exudation. In summary, mechanistic mathematical models are also powerful vehicles to incorporate multi-scale processes, heterogeneous data such as soil, and complex geometries and are a future direction of focus for the field (Roeder et al, 2011;Bucksch et al, 2017;Hong et al, 2018;Ruiz et al, 2020b). Additionally, mechanistic models can be coupled with imaging studies and growth models to link observed plant structure to underlying function.…”

A 3D image-based modelling approach for understanding spatiotemporal processes in phosphorus fertiliser dissolution, soil buffering and uptake by plant roots

“…This could provide a complementary approach to focus and enhance more costly and time consuming field trials. Image-based modelling (using experimental images to define the geometry of models) offers vital advantages over models where root structure are represented by parameters or functions for the problem of assessing fertiliser dissolution rates and P availability and can complement results from imaging studies 27 . When using the latter modelling approach, root geometry including proximity to high P sources is averaged and thus important processes for fertiliser efficiency are overlooked.…”

“…When using the latter modelling approach, root geometry including proximity to high P sources is averaged and thus important processes for fertiliser efficiency are overlooked. In particular, the details of root competition for the same soil/fertiliser P supply within a discretised area/length are averaged without consideration for localised root interactions 27 . With image-based modelling, not only is explicit root fertiliser proximity calculated from experimental data, but competition between roots (of the same plant) for local P supplies are simulated 22 , allowing accurate assessment of P uptake efficiency, in the sense of unit of P absorbed per unit of photosynthate spent of root growth, from fertilisers.…”

“…Including such geometric detail in image-based models comes at large computational cost. Thus, image-based models are less suitable for research questions where multiple-physics and/or larger spatial scales need to be considered, and models that represent roots by parameters or functions are the preferred modelling approach 27 .…”