Every year, #DigitAg launches a PhD campaign and supports new theses, from blue skies to applied research. Discover our 2021 PhD opportunities!
PhD positions – Campaign 2021
- Life & Environmental Sciences
- Life & Environmental Sciences + Applied Mathematics
- Life & Environnemental Sciences + Engineering Science
- Humanities & Social Sciences
- Humanities & Social Sciences + Engineering Science
- Humanities & Social Sciences + Life & Environmental Sciences
|Study and implementation of an original optical approach combining diffuse polarization spectrometry and Mueller’s method for plant monitoring with the identification of specific phenotypic traits: application to vine water stress.
Keywords: phenotyping, spectrometry, polarisation, Mueller’s matrix, vine, hydric stress
In a fast-changing global environment, the ability to produce plant material adapted to diverse, changing and sometimes extreme agronomic conditions is an absolute priority. In particular, agriculture must consider possible ways of adapting and managing water resources to maintain sustainable agricultural production. To do this, one of the levers identified is the creation of new varieties that are more tolerant to abiotic stresses. While genotyping capacities have exploded in recent years, identifying the mechanisms involved through phenotyping remains very complex. Thus, the ability to produce relevant phenotyping information remains the limiting factor for the progress of varietal selection. However, instrumental optics, and in particular measurement systems based on visible and near-infrared spectrometry, have demonstrated in recent years that they offer undeniable potential for addressing these demands. But while some gene expression parameters are now accessible, there are other, crucial parameters that are not yet directly measurable. In particular, the measurement of some characteristics of the plant’s response to water stress. The aim of this thesis will be to study, implement and evaluate a new and original optical approach that combines polarized light spectroscopy and the Mueller matrix. This technique, which is precise and more targeted, will make it possible to better decrypt the spectra in order to extract and select the main traits and indicators that characterize the varieties and their resistance to hydric stress. This method will be tested on different varieties of grape varieties subjected to different types of hydric stress with measurements taken at the leaf level to avoid the first level of spectral disturbance linked to the plant cover.
|Development of a mathematical method based on proximity data sensors for the early detection of pathologies in farm animals.
Keywords: Stochastic differential equations – non-parametric statistics – individual-based model – animal welfare – sensors
Animal production is one of the agricultural sectors with a heterogeneous operational deployment of ICTs. Ruminant breeding is a good example of this contrast with sectors such as dairy cattle, which are well supplied, and sectors with a more modest level of commercial digital equipment, such as suckling sheep. While there is no lack of sensors and associated tools, it is more a question of the relevance and reliability of the information that is derived from them, thanks to sufficiently precise and predictive algorithms that still need to be developed. The use of these new sensors creates the need for new methodologies, such as the modeling of social interactions from spatio-temporal data including advanced statistical inferences, to produce more accurate predictive information in real time. We propose to work on an interdisciplinary approach, using mathematical modeling, inferential statistics and the study of social behaviors.
The aim of the thesis is to adapt these approaches to ruminant (cattle and sheep) movement data, in order to identify breaks in the social structure of a group of farm animals that could be early indicators of individual pathologies. For this purpose, the PhD student will set up new non-parametric statistical estimators for a new model of diffusion process interactions (based on existing models) that will allow the production of specific algorithms for clustering and/or detection of disruptions.
|Combining participatory approaches and constrained modelling to design agroecological systems
Keywords: agroecological transition, artificial intelligence, simulation, co-design, participatory evaluation, orchard, market gardening
Addressing the challenges of agriculture in a context of global changes, loss of biodiversity and depletion of fossil resources requires a profound renewal of agricultural systems. Agroecology proposes to base the design of agricultural systems on the enhancement of ecological functionalities. This requires reintroducing biodiversity into agrosystems, adapting practices to the local context and combining knowledge (scientific, expert and operational). Agroecology experiments around the world have shown that increasing the complexity of agrosystems increases their resilience, reduces dependence on synthetic inputs, provides ecosystem services and thus improves their performance. But these experiences have also shown the difficulty of designing these complex agrosystems: agro-ecological design remains a major obstacle for actors in the agro-ecological transition, due to the lack of appropriate tools and approaches. This challenge for agricultural researchers also calls for research in artificial intelligence to design such tools. The objective of this thesis is to articulate concepts and methods from these two disciplinary fields in order to equip this stage of agroecological design with (i) constrained reasoning models to explore the combinatorics of spatio-temporal arrangements of agrosystem elements, and (ii) a participatory design device using models as an intermediate object, making it possible to clarify constraints, combine knowledge and stimulate creativity to co-design agroecological farms. It will be implemented on five situations of market garden orchard design, in collaboration with research and higher education institutes (Institut Agro, INRAE, Ecole du paysage), farmers, associations (GRAB, Domaine du possible), private sector companies (Potagers & compagnie) and local authorities. This diversity of situations will make it possible to increase the genericity of the knowledge produced and will facilitate, with the participatory approach, the transfer of the co-built tools to the actors of the agro-ecological transition.
|Model inversion from crowdsourced agricultural data: application to the estimation of crop and soil parameters in viticulture
Keywords: Crowdsourcing, water status, model inversion, vine
This PhD topic is positioned in the continuation of the research work carried out within the framework of the ApeX-Vigne project. A previous PhD allowed the initiation of a crowdsourcing project in agriculture and more specifically in Viticulture. This project is based on the use of a free and open source mobile application allowing professionals of the wine industry to collect standardised, timed and geo-referenced observations of the vine water status at the plot level. Since 2018, Apex-Vigne provided each year a large database (approx. 7000 measurements/year) at a regional scale. The aim of the PhD thesis is to take advantage of the massive nature of the data collected in order to estimate cultivation parameters that are difficult to measure (soil water holding capacity) but essential for the implementation and calibration of decision support models based on the water balance for irrigation management. The PhD thesis will consider several scales of application: plot, domain, region. The originality of the proposed approach is to integrate crowdsourced observations in a water balance model inversion approach. The scientific novelty of the approach comes from the type of data considered: imprecise, asynchronous, heterotopic, massive. Our approach thus makes the hypothesis that the imperfection of crowdsourced data will be largely counter balanced by the volume of data available. The considered approach is based on 3 steps of which each one is original : i) the development of a transfer function allowing to relate crowdsourced observations to a classical output variable (OV) of a water balance model like the predawn water potential, ii) the spatial consolidation of the OV, this step makes the hypothesis that the OV is spatially auto-correlated, it is at this level that the imperfection of the data will be taken into account by geostatistical approaches, iii) the model inversion itself, based on OV estimates and climate data of the vintage, the thesis will aim at inverting a water balance model in order to produce estimates on crop parameters (It will focus as a first step on the estimation of a parameter that is difficult to access : the soil water holding capacity.
|How digitisation reshuffles the cards between uncertainty and dependencies ; A comparative analysis France/Italy of farmers’ strategies
Keywords: Resources dependency theory, digitalization, signalisation, phytosanitary quality, short food supply chain, France, Italy
The project aims to analyze the role of digital tools in strengthening producer-consumer relations and promoting more environmental friendly practices. It is in line with the European “farm to table” strategy of the Green Pact for Europe, which aims to offer healthy consumption to consumers by supporting alternative practices. Digitalization can allow farmers reducing the number of intermediaries, promoting the quality of their products, valuing the specificity of their production and thus obtaining a higher remuneration. However, digitalization does not guarantee the contractual stability of traditional channels. This is why we wonder if digitalization can constitute a way to reduce uncertainty and manage dependency to customers and suppliers. From a theoretical point of view, this doctoral work is based on the theory of dependence on resources (Pfeffer and Salancik, 1978). Our thesis is that digitalization enables the farmer to regain control over the resources that its survival and development require. From an empirical point of view, we mobilize longitudinal exhaustive data from all French and Italian farms (Census of Agriculture and structure survey). We measure the level of digitalisation of producers, the certifications adopted and their marketing methods. In methodological terms, this project is based on simultaneous triple equation models and matching type models. France and Italy are subject to the same requirements in terms of the reduction of phytosanitary products but have different commercial and production strategies. These data will be supplemented by qualitative surveys to appreciate how do farmers estimate their dependence to consumers, distributors, suppliers and complementers and to which extent they consider digitalization as a way to facilitate their access to environment resources.
|Supporting the transformation of an agro-ecosystem for a sustainable management of the commons. The case of the relationships between water management, sanitary risks and quinoa production practices on the Bolivian Altiplano.
Keywords: Agent-based Model, Participatory Modelling, Interactive simulation, Social Equity, Quinoa, Water, Commons
This thesis is part of the Wasaca project (Wastewater irrigation: a sustainable agriculture adaptation to climate changes over the Bolivian Altiplano?) funded by Agropolis Fundation. In the framework of WP3 (“Engaging stakeholders in the adoption of sustainable agricultural practices”), the topic aims to set up a participatory modeling approach called “ComMod” around the availability of water resources, linked to environmental and health issues, to collectively consider sustainable trajectories for 2050 and 2100.
|Cooperatives: a way towards the digital transition of agriculture?
Keywords: Agricultural cooperatives – Digital transition- Agroecology – Innovation – Governance
Digital tools have been developing in agriculture for several years. Some authors see it as a way to support its Agroecological Transition. However, the compatibility between some digital tools and agro ecological approaches is still debated. On the other hand, in France, cooperatives represent an important part of the agricultural and agri-food activity. At the national level,75% of farmers (90% in Occitanie) depend on a cooperative, making these structures key players in the digital transition. They are characterized by a great diversity of sizes and structures, despite a trend towards concentration.