OUACHENE Naomi

Ph.D. student at INRAE Email: naomi.ouachene@inrae.fr Address: INRAE, UMR SAS, 65 rue de St-Brieuc, CS 84215, 35042 Rennes Cedex, France Expertise: Applied Statistics, Agronomy Keywords: Farming system, Dairy farm, Environmental system, Productivity, Interaction, Statistical modelling

Research

Ph.D. subject: Study of environmental and economic performances of dairy farms. Assess dependencies among descriptive variables of these farms using copula

From 1970 to 2004, greenhouse gas (GHG) emissions due to human activity increased by 70%; this increase is driving climate change and disrupting the global ecosystem. Consequently, improving environmental performances is crucial. More specifically, as 14.5-18% of global GHG emissions come from livestock production, improved understanding of dependencies among descriptive variables of dairy farms may help decrease environmental impacts of these farms. The goal of the thesis is to formalize multivariate relations among these variables to search for mechanisms that may decrease environmental impacts while maintaining farm productivity. To this end, multivariate copulas have been chosen because they are well suited for multidimensional and non-linear assessments. They can also highlight specific dependencies that occur only in the tails of distributions and thus may occur only during extreme events. In contrast, agronomic research to date has tended to focus mainly on interactions among mean values of variables, not their extreme values. Simulation models have been explored in other studies, but as they need large amounts of data, they are difficult to implement in a systemic assessment. Finally, other multi-criteria analysis approaches could be considered. Survey data on French dairy farms, collected by the French Livestock Institute (IDELE), will enable statistical modelling of these multiple interactions.