Revisiting crop root systems ideotypes for water uptake : new tools and models

Root water uptake is critical for yield determination in agriculture. Water stress is indeed the most significant environmental stress in crops worldwide. It is even expected to become more and more severe in many regions of the globe while nowadays already millions of people remain chronically undernourished. In such context, understanding the plant mechanisms to efficiently take up water in a heterogeneous changing environment would be of great benefit in the quest for more resilient food production systems. Several authors proposed plant root system ideotypes for water uptake defined as ideal plant models, which are expected to yield a greater quantity or quality of grain when developed as cultivars. Such plant models are characterized by an ensemble of properties (alternatively called phene states) that are responsible for their improved performance in terms of transpiration, such as root structural (e.g. root length, elongation rates) and functional (e.g. radial or axial hydraulic conductivity) properties. However current definitions of crop root system ideotypes are of limited in terms of interest for several reasons. They usually do not contain quantitative characteristics and do not include the time dimension. They are not associated to specific pedo-climatic conditions while it has been demonstrated that no root system can perform ideally in any situation. Furthermore root water uptake strategy is not uniquely defined by an ensemble of phene states: other combinations of phene states could lead to similar water uptake behaviors. The main objective of this thesis is to advance the current understanding of plant optimal water uptake and growth strategies for maximising crop yield. Such plants would minimize the water stress over their crop cycle and particularly at critical development stages. New definitions of crop quantitative eco-ideotypes for root water uptake, i.e. ideal root systems that avoid water stress over their crop cycle or, at least, at flowering were proposed. Such plant models display optimal water uptake and growth strategies that are determined by macroscopic parameter trajectories in time. In turn, these trajectories and their changes are entirely explained by an ensemble of phene states. As no plant can perform ideally in any situation, the ideotypes are systematically associated to specific pedo-climatic situations. To reach these eco-ideotypes, focus was first put on water flow in root system hydraulic architecture.The current numerical resolution method of the water flow equation was improved by integrating the analytical solution valid at the segment scale in a root system hydraulic tree model. Plant macroscopic parameters were also shown to be analytically calculable based on functional and structural parameters at root and root system scales. An online flexible software, MARSHAL, was also elaborated designed to easily generate contrasted root system hydraulic architecture of maize crop. This unique software is compatible with current functional-structural models to bridge the gap between local root phenes, macroscopic parameters and plant performance in contrasted pedo-climatic situations The second research section of the thesis was dedicated to the hydraulic property quantification of maize, ryegrass and lupine roots using a combination of local or global measurements of proxies of water uptake and inverse modelling. In all cases, good correspondence between measurements and simulations were obtained, as well as uncertainty of the fitted parameters. Finally the novel analytical relationships between root local phene states and macroscopic parameters were used to explore the parametric space given in the literature and to generate corresponding macroscopic parameters. The sensitivity of hydraulic and architectural plant-scale parameters (root system volume, convex hull, root system conductance and depth of standard uptake) to local phene states were then analyzed. These macroscopic parameters, together with their changes with time, define the plant strategy for water uptake that can be adapted to certain pedo-climatic conditions as shown through simulations of the water flow in the soil-plant-atmosphere continuum using R-SWMS. The model predicted cumulative transpiration were successfully correlated with observed plant yields. The coupled soil-plant model was further used to assess the impact of individual local phene change, which leads to contrasted effects depending on the pedo-climatic situation. Finally macroscopic parameter trajectories were shown to explain plant performances and strategies.