Development of a decision support tool to optimize grazing of spontaneous vegetation based on a "functional" classification of pastoral resources

My name is Elisa Deschamps and I'm a PhD student at UMR Selmet in Montpellier. I'm working on pastoral vegetation, and more specifically on the development of a decision-making tool to optimize grazing of spontaneous vegetation.
I'm an agricultural engineer specializing in the study of livestock systems, and thanks to my various internships, I've been able to discover the world of research. In particular, I did an internship with a PhD student at the University of Gembloux in Belgium, working with UMR Pegase on the feeding practices of dairy goats. I'm particularly interested in research topics related to pastoral livestock farming, as they are interdisciplinary and bring together a wide range of stakeholders (breeders/shepherds, land managers, research, etc.) in a given area. What's more, pastoral areas are facing a number of challenges (fire, climate change, etc.) that are having an impact on the resource, and it seems essential to me to improve our understanding of the mechanisms that govern animal choices, so as to be able to determine the feed value of an area for a herd. I'd also like to take advantage of this thesis to discover the teaching-research profession by taking part in tutorials or presenting my work to students.

I'm working on characterizing pastoral vegetation with a view to making better use of it in a changing context. Indeed, in the context of the agro-ecological transition, better valorization of these resources could help improve the food autonomy of pastoral farms and the ecological footprint of livestock farms. To make the best use of heterogeneous and variable pastoral resources, farmers need to be able to identify, at any given time, what constitutes a "resource" and the functions of the resources present in the animals' diet, taking into account the seasonal dimension and the animal species. At present, there are many tools and data available for grassland, which is not the case for rangeland.
The aim of this thesis is to propose a combination of approaches to describe the diversity of pastoral resources and facilitate grazing management. It explores the potential of SPIR to classify grazed feeds, the coherence between herders' knowledge and numerical analyses, and the creation of a digital decision support tool. The research is based on herd observations, SPIR sampling and analysis of plant samples, and surveys of farmers.

I trained as an agricultural engineer in Montpellier, specializing in livestock systems, which made me aware of the issues facing pastoral systems. It was therefore natural that I chose to pursue my work on these themes.

• Starting date : 1st November 2023
• Research unit: Selmet
• University: Institut Agro Montpellier
• PhD school: GAIA
• Scientific field:  Agronomy, animal sciences
• Thesis management : Lucile Sautot, UMR Tetis, AgroParisTech
• Thesis supervisors :  Magali Jouven, UMR Selmet, Institut Agro et Denis Bastianelli, UMR Selmet, Cirad
• Funding: PEPR Agroécologie et Numérique
• #DigitAg : Thèse labellisée - Axe 3 : Capteurs, acquisition et gestion de données, Axe 6 : Modélisation et simulation (systèmes de production agricole), Challenge 4 : Des productions animales durables, Challenge 1 : Le challenge agroécologique, 

Keywords: Grazing management, decision support, pastoral resources, NIRS, classification

Abstract: Pastoral systems value semi-natural environments through grazing, characterized by heterogeneous and variable spontaneous vegetation, often including woody species. Difficult physical and climatic conditions, exacerbated by climate change, profoundly affect the availability and seasonality of pastoral resources and pose challenges for pastoralists in managing grazing. In the context of agro-ecological transition, the efficient use of these semi-natural environments could contribute to improving food autonomy and the ecological footprint of livestock production. In order to make the best use of heterogeneous and variable rangeland resources, pastoralists must be able to identify at any given time what constitutes a resource, and the functions of the resources present in the animals' diet, taking into account seasonality and animal species. While there are many tools available to manage grazing on pastures, the heterogeneity, diversity and variability of rangelands have so far prevented the development of equivalents for rangelands. Existing tools for the characterisation of rangeland resources are based on typologies mainly documented by experts and do not integrate local specificities in terms of temporal and spatial distribution of resources. More recently, digital technologies such as remote sensing with increasingly precise data and near-infrared spectroscopy (NIRS) have offered complementary approaches that allow the characterisation of pastoral resources at a given moment. The aim of this thesis is to propose a combination of approaches to describe the diversity of pastoral resources and facilitate grazing management. It explores the potential of SPIR to classify grazed forage, the coherence between herders' knowledge and numerical analysis, and the creation of a digital decision support tool. This research is based on herd observations, sampling and SPIR analysis of plant samples, and interviews with herders. Two study sites in Hérault and Ardèche will be studied, with seasonal monitoring of herds and regular interactions with herders.

Contact: elisa.deschamps [AT] inrae.fr
Social network: LinkedIn