Search

International audience

Canopy light interception determines the amount of energy captured by a crop, and is thus critical to modeling crop growth and yield, and may substantially contribute to the prediction uncertainty of crop growth models (CGMs). We thus analyzed the canopy light interception models of the 26 wheat (Triticum aestivum) CGMs used by the Agricultural Model Intercomparison and Improvement Project (AgMIP). Twenty-one CGMs assume that the light extinction coefficient (K) is constant, varying from 0.37 to 0.80 depending on the model. The other models take into account the illumination conditions and assume either that all green surfaces in the canopy have the same inclination angle ([Formula: see text]) or that [Formula: see text] distribution follows a spherical distribution. These assumptions have not yet been evaluated due to a lack of experimental data. Therefore, we conducted a field experiment with five cultivars with contrasting leaf stature sown at normal and double row spacing, and analyzed [Formula: see text] distribution in the canopies from three-dimensional canopy reconstructions. In all the canopies, [Formula: see text] distribution was well represented by an ellipsoidal distribution. We thus carried out an intercomparison between the light interception models of the AgMIP–Wheat CGMs ensemble and a physically based K model with ellipsoidal leaf angle distribution and canopy clumping ([Formula: see text]). Results showed that the [Formula: see text] model outperformed current approaches under most illumination conditions and that the uncertainty in simulated wheat growth and final grain yield due to light models could be as high as 45%. Therefore, our results call for an overhaul of light interception models in CGMs.

Simulating leaf development from initiation to maturity opens new possibilities to model plant–environment interactions and the plasticity of plant architecture. This study analyses the dynamics of leaf production and extension along a maize (Zea mays) shoot to assess important modelling choices. Maize plants from two cultivars originating from the same inbred line, yet differing in the length of mature leaves were used in this study. We characterized the dynamics of the blade and sheath lengths of all phytomers by dissecting plants every 2–3 days. We analysed how differences in leaf size were built up and we examined the coordination between the emergence of organs and phases of their extension. Leaf extension rates were higher in the cultivar with longer leaves than in the cultivar with shorter leaves; no differences were found in other aspects. We found that (i) first post-embryonic leaves were initiated at a markedly higher rate than upper leaves; (ii) below ear position, sheaths were initiated at a time intermediate between tip emergence and appearance, while above the ear position, sheaths were initiated at a high rate, such that the time interval between the blade and sheath initiations decreased for these leaves; and (iii) ear position also marked a change in the correlation in size between successive phytomers with little correlation of size between upper and lower leaves. Our results identified leaf extension rate as the reason for the difference in size between the two cultivars. The two cultivars shared the same pattern for the timing of initiation events, which was more complex than previously thought. The differences described here may explain some inaccuracies reported in functional–structural plant models. We speculate that genotypic variation in behaviour for leaf and sheath initiation exists, which has been little documented in former studies., We compared leaf development and extension in two maize cultivars with same genetic background but differing in their size. We showed coordinations linking the initiation, differentiation and extension of leaves with their emergence. Some coordinations markedly differ before and after tassel initiation, resulting in successive leaves increasing in length up to ear position and then decreasing in length. Both cultivars showed remarkably similar coordinations, yet differed by an higher leaf extension rates in the taller cultivar. The work contributes to clarify the features that organize the timing of events and the pattern of successive leaf sizes along a maize shoot.

International audience; Introduction Rhizodeposition, i.e. the release of organic materials by roots, represents a significant portion of plant’s carbon (C) budget, ranging from 5% to 15% of net photosynthesized C (Pausch and Kuzyakov, 2018). Various rhizodeposits can be released by roots, e.g. soluble exudates, secreted mucilage, sloughed cells, or volatile organic compounds. Despite their short lifetime, some of these products have been shown to favor plant growth, e.g. by increasing water and nutrient uptake. Among rhizodeposition processes, exudation has been suggested to depend on the concentration of carbohydrates inside the roots (Personeni et al., 2007). However, rhizodeposition not only depends on the availability of C in the roots, but also on the architecture of the root system, and many have shown that rhizodeposits are more concentrated in specific areas, such as root tips. Consequently, a Functional-Structural Plant Model (FSPM) would theoretically represent the best framework for simulating the spatial and temporal dynamics of rhizodeposition, as it can describe the evolution of both the metabolism and the architecture of the plant. The objective of this work is to create such a framework by coupling a whole-plant FSPM, a 3D root architectural model, and a new model simulating rhizodeposition.Modelling approachOur strategy has been to combine the FSPM CN-Wheat (Barillot et al., 2016), which describes the main processes of C and nitrogen (N) acquisition and transformation by an individual wheat plant and the 3D growth and development of its aerial organs, with the model ArchiSimple (Pagès et al., 2014) that simulates the development of the 3D root architecture for a range of plant species, and the new model RhizoDep, which calculates a full C balance in each part of a root system in order to simulate local rhizodeposition fluxes. The complementarity of the three models is illustrated in Figure 1: i) CN-Wheat is used to calculate the amount of C allocated from the shoots to the roots, ii) ArchiSimple provides the 3D structure of the root system, and iii) RhizoDep distributes the C provided by the shoots within the 3D root system and simulates the actual growth, respiration and rhizodeposition of each root element based on C availability. The major link between the three models lies in the exchange of C between aboveground and belowground tissues, which is driven by gradients of sucrose concentration in the different compartments of the plant.Preliminary results & short-term perspectivesThe coupling of the three models has been started using the OpenAlea platform and its Multiscale Tree Graph formalism (Pradal et al., 2008). First simulations were done using the allocation of C to the roots simulated by CN-Wheat as an input to the root model based on the effective coupling of ArchiSimple and RhizoDep. These simulations show how rhizodeposition is intrinsically dependent on the architecture of the root system and on the total amount of available C. For completing the coupling, several issues still need to be tackled, e.g. how N uptake and metabolism should be spatialized in a 3D root system, how it may be regulated by local C and N availability, and how rhizodeposition can modify soil N availability. However, this modelling approach has already led to a first prototype able to simulate rhizodeposition processes on a dynamic, 3D root system that is fully integrated within the functioning of the whole plant. Its refinement will offer unique opportunities to study the possible link between rhizodeposition and the environmental factors affecting plant growth, e.g. atmospheric CO2 concentration or soil N availability.

International audience

National audience; The potential of agricultural soils to store carbon (C) has been claimed to require a large amount of nitrogen (N) and therefore a high soil N availability. This assumption is based on the observation that stable soil organic compounds are characterized by a narrow range in C:N ratios. However, N-fertilizer additions have also been shown to increase respiratory C losses from soils. Furthermore, most of the organic C naturally stored in soils originate from plant C inputs, which occur both by aerial litter deposition and by root turnover and rhizodeposition. Root C inputs have been shown to contribute 2-3 times more to soil C sequestration than shoot-derived C inputs. However, root C inputs may be limited or even decrease when soil N availability increases beyond a certain level. A trade-off may therefore exist between the amount of soil available N and the potential of soil C sequestration associated to the inputs of root C. The existence of such trade-off and its possible consequences for the optimization of soil C sequestration will be discussed according to evidences from scientific literature and to a modelling approach describing how root C inputs to soil can evolve with soil N availability.

International audience; The extraction of desirable heritable traits for crop improvement from high-throughput phenotyping (HTP) observations remains challenging. We developed a modeling workflow named Digital Plant Phenotyping Platform, D3P, to access crop architectural traits from HTP observations. D3P couples the ADEL (architectural model of development based on L-systems) wheat (Triticum aestivum) model, that describes the time course of the three-dimensional architecture of wheat crops, with simulators of images acquired with HTP sensors. We demonstrated that a sequential assimilation of the green fraction derived from RGB (Red Green Blue) images of the crop into D3P provides accurate estimates of five key parameters (phyllochron, lamina length of the first leaf, rate of elongation of leaf lamina, number of green leaves at the start of leaf senescence and minimum number of green leaves) of the ADEL-Wheat model that drive the time course of green area index and the number of axes with more than three leaves at the end of the tillering period. However, leaf and tiller orientation and inclination characteristics were poorly estimated. D3P was also used to optimize the observational configuration. The results, obtained from in silico experiments conducted on wheat crops at several vegetative stages, showed that the accessible traits could be estimated accurately with observations made at 0{degree sign} and 60{degree sign} zenith view inclination with a temporal frequency of 100 {degree sign}Cd. This illustrates the potential of the proposed holistic approach that integrates all the available information into a consistent system for interpretation. The potential benefits and limitations of the approach are further discussed.

International audience; The building sector is clearly undergoing the digital transition with the willingness to move from paper to absolute digital. Our research aims to study the transformation of the activity of “lifting of reservations” of a construction site manager using a business application. To grasp the complexity of this activity, a systemic approach seemed relevant and we proposed a theoretical articulation between the situated and incorporated dimensions of the activity and its historical-cultural development. We carried out an ethnographic investigation, for six months, in immersion on a construction site, in the form of an observant participation along with a shadowing observation. The main results indicate that the activity of "lifting of reservations" is part of a culture marked by the importance of the exchanges and the body. The business application transforms this activity which is sometimes strengthened, and other times thwarted. The construction site managers must adapt the use of the application to the site situations.; El sector de la construcción está claramente haciendo la transición digital con la voluntad de pasar del papel al totalmente digital. Nuestra investigación tiene como objetivo estudiar la transformación de la actividad de “levantamiento de reservas” de una directora de obra utilizando una aplicación informática. Para entender la complejidad de esta actividad, un enfoque sistémico parecía evidente y propusimos una articulación teórica entre las dimensiones situadas y encarnadas de la actividad y su desarrollo histórico-cultural. Llevamos a cabo una investigación etnográfica, durante seis meses, en inmersión en una obra, con una participación observadora y una observación en “shadowing”. Los resultados indican que esta actividad forma parte de una cultura caracterizada por la importancia de los intercambios y del cuerpo. La aplicación experta transforma esa actividad que a veces se desarrolla y otras se impide. Los directores de obra tienen que adaptar el uso de la aplicación a las situaciones de la obra.; Le secteur du bâtiment entre davantage dans la transition numérique avec la volonté de passer du papier au tout numérique. Notre recherche a pour but d’étudier la transformation de l’activité de “levée de réserves” d’une conductrice de travaux utilisant une application métier. Pour saisir la complexité de cette activité, une approche systémique nous a semblé pertinente et nous avons proposé une articulation théorique entre les dimensions situées et incarnées de l’activité et son développement historico-culturel. Nous avons réalisé une investigation ethnographique, durant six mois, en immersion sur un chantier, sous la forme d’une participation observante en parallèle d’une observation en “shadowing”. Les principaux résultats indiquent que l’activité de "levée de réserves" s’inscrit dans une culture marquée par l’importance des échanges et du corps. L’application métier transforme cette activité qui se trouve parfois développée et parfois contrariée. C’est aux conducteurs de travaux d’adapter l’usage de l’application aux situations du chantier.; O setor da construção está claramente a entrar na transição digital com a vontade de passar do papel para o completamente digital. A nossa pesquisa visa estudar a transformação da atividade de "levantamento de inconformidades" de uma diretora de obra usando uma aplicação especializada. Para apreender a complexidade dessa atividade, uma abordagem sistémica pareceu relevante e propusemos uma articulação teórica entre as dimensões situadas e incorporadas da atividade e seu desenvolvimento histórico-cultural. Realizamos uma investigação etnográfica, durante seis meses, em imersão num estaleiro, sob a forma de uma observação participante em paralelo com uma observação de tipo “shadowing”. Os principais resultados indicam que a atividade de "levantamento de inconformidades" faz parte de uma cultura marcada pela importância dos intercâmbios e do corpo. O aplicação especializada transforma essa atividade que por vezes é desenvolvida e por outras vezes é contrariada. São os diretores de obra que têm de adaptar o uso da aplicação às situações do estaleiro.

Plants are the main source of organic carbon in soils. Besides litter incorporation, most of plant carbon fluxes to soil occur belowground, through the release of organic compounds by roots and the decay of root tissues. Despite the importance of these processes for soil carbon sequestration and for soil biological functions, our understanding of such fluxes has been hampered by the difficulties associated to their measurement in actual soil environments and their integration to plant growth models. Our aim is to develop new modelling approaches in order to accurately describe trophic fluxes from roots to soil and their spatial and temporal evolution. Functional-structural plant models (FSPM), which take into account both plant physiology and plant architecture, may be well adapted to such a modelling strategy, but also bring new challenges in terms of processes coupling. This work will present our current strategy to simulate and integrate root exudation, mucilage emissions, root cells desquamation and root senescence into a common model, and will highlight some of the main knowledge gaps associated to the simulation of these processes.

International audience; Phenotyping is rapidly evolving from a small set of destructive measurement and visual notations of few simple traits to massive characterizations derived from high-throughput non-invasive proximal or remote sensing techniques. Meanwhile functional structural plant modeling (FSPM) integrates the physiological and morphological information from organizational scales to the canopy level. The combination of new phenotyping techniques with FSPMs is expected therefore to estimate a set of FSPMs parameters corresponding to traits of interest. LiDAR (Light Detection And Ranging) is recently exploited for detailed 3D description of the canopy structure, especially over dense canopies with small elements such as wheat and barley. In this work, we propose to use a model-assisted phenotyping approach to improve our understanding of the interaction between laser beam and canopy. It leads us to develop inversion algorithm to retrieve canopy traits. A discrete LiDAR scanning simulator was first developed based on PlantGL, a 3D plant modeling python library. The footprint and the geometrical configuration of the LiDAR are explicitly accounted for. The LiDAR simulator was validated over an artificial crop made of 45 artificial plants. Actual LiDAR measurements were performed over the same scene. Results proved that the simulator generates a 3D point cloud having the same statistical properties as those derived from the actual LiDAR measurements. Then a synthetic experiment was completed to demonstrate the potentials of model assisted phenotyping. 3D wheat canopy scenes were generated with AdelWheat model for two contrasting development stages corresponding to thermal time 500 °C • day and 1500 °C • day across a wide range of combination of the model parameters (242 cases replicated 20 times). The scenes were transformed into 3D point clouds using the LiDAR simulator. A set of 50 independent cases were generated in addition to evaluate the performances of the method. Results demonstrate that emerging properties could be retrieved with a good accuracy from the 3D map including the leaf area index (LAI) (R2 = 0.85 and rRMSE = 6.11%). The retrieval of other parameters will be discussed with due attention to the complexity of the comparison of simulated 3D point clouds with the measured ones.

poster abstract; Plant density is governed by the sowing density and the emergence rate. For a given plant density, the uniformity of plant distribution may significantly impact the competition between plants as well as weeds. Plant density and uniformity is therefore a key factor explaining production, although a number of species are able to compensate low plant densities by a larger development of individual plants during the growth cycle. However, as the lack of dedicated device, manual field counting in wheat crops is still extensively employed as the reference method although tedious and time consuming. Further, very little work has been devoted to document the plant distribution pattern along the row. The main objective of this work is to propose a high-throughput method to estimate plant density and the distribution pattern from high resolution images taken at the emergence stage. Wheat was selected as the material because it is one of the major crops cultivated over the world with relatively high plant density (150-400 seeds·m-2). High-resolution RGB images were captured from ground-based or unmanned aerial vehicle (UAV) platforms over several wheat experiments in France. They covered a large range of densities from 100 to 600 plants·m-2 over various cultivars at Haun stage around 1.5. Computer vision techniques were exploited to develop a pipeline for assessing the plant density and the distribution patterns of plant positions along the row. The method provides a good estimation of the actual plant density with R2» 0.95 and rRMSE » 12% over ground-based imagery while performances degrade only slightly for the UAV imagery with R2» 0.86 and rRMSE »19%. The inter-plant spacing along the row direction was found following a gamma distribution and the positions of the successive plants are independent from each other. The distribution parameters may be retrieved from the images, subsequently characterizing the uniformity of plant distribution along the row direction.

poster abstract; International audience; The upper leaves of wheat are the main sources of assimilate for grain production. As such, they are a specific target for fungicide treatments. The efficacy of a treatment can be assessed by the quantity of fungicide deposited on target leaves relative to the spraying application rate per unit area. For a given spraying system and under favorable meteorological conditions, foliar deposit depends on the individual areas and spatial arrangement of leaves within the canopy, which determine the interception of the spray; and on the physio-chemical characteristics of the product, which determine its fate on leaf surfaces. The objectives of this study are to model the deposition of pesticide on the upper leaves of wheat and to identify the plant architectural traits that influence the foliar deposit as the plant develops. For this we couple a 3D architectural model of wheat (ADEL-Wheat), which captures the dynamics of the spatial arrangement of individual leaves at different stages of plant development, with the projection algorithm of the Caribu light interception model that allows estimating the deposit of fungicide within the canopy. A three-year experiment was performed (ECHAP project), with detailed assessments of plant architecture for three genotypes, and of tracer-deposit (tartrazine) on the four upper leaves, for two spraying dates. The architectural measurements were used to calibrate the wheat model that was tested against the foliar deposit data. We used the model to further analyse the origin of the differences of deposit between leaves and between treatment dates observed in the experiments. The analyses reveal that the dynamics of leaf bending, assumed to be a function of leaf age, has a strong influence in spraying efficacy, especially for the upper two leaves. We therefore propose that optimal application strategy could be reasoned not only by considering the degree of expansion of targeted leaves, but also their dynamic of bending.

oral communication abstract; Context and aims: The 3D structure of plants is an important aspect of plant-interface with the environment, involved in the capture of resources, the exposition to stresses and physical transfers both within the plant and at canopy level. For these reasons describing plants in 3D is an important aspect of functional structural plant modelling. Despite of this, the representation of plants in FSPM pays little attention to the dynamic aspects of geometry. This work analyses how the 3D geometry of wheat leaves and axes changes from plant emergence to maturity. Methods: Experiments were conducted on three commercial winter wheat cultivars of contrasted stature (erectophile and planophile). Plants were grown in compartments installed outdoor at population densities of 70p/m² and 200p/m². Sixteen plants per treatment were tagged and we recorded weekly the number of emerged leaves on the main stem, their status (green or senescent) and digitalized the main stem leaves and all axes with a Polhemus digitizer. Results: Both main stems and tillers axes started their extension nearly horizontal and straightening progressively to reach a vertical direction only close to the time of flag leaf ligulation. The dynamic of leaf curvature depended on leaf position: lower and higher leaves showed a progressive droop down with rates and range gradually changing with leaf position and varying between genotypes, intermediate leaves showed a complex pattern with a transient straightening due to the movement of the bearing axe during their elongation. Conclusion: Our work brings a comprehensive view of leaf and axes movements and highlights the gradual straightening of the stems that is of high importance to understand leaf position and orientation. We plan to derive parameterizations for the movements observed and to evaluate with 3D models their implication in terms of light capture. We expect from that work to improve the simulations of plant-environment interactions.