Development and validation of a grassland production index based on the use of satellite data time series - Application to an insurance product in France.

Summary The index method is considered in the field of agricultural insurance to measure the impact of climate on agricultural production. It is likely to characterize the impact of the random event of drought or flooding. To a lesser extent, this method can be used for other climatic events such as frost. Until now, the deployments of these index methods have been based on data from the measuring stations of the meteorological network of a country. These were mainly rainfall, temperature and in some cases potential evapotranspiration. This type of measurement is technically well mastered and accepted by insurance professionals and their clients. However, these measurements are difficult to spatialize if the number of observation points is not sufficient or the environment is climatically complex. With the development of satellite remote sensing, we have a set of data with variable characteristics (spatial, spectral and temporal resolutions) and complementary allowing a continuous monitoring of the state of vegetation over large areas and at a more or less fine scale (greater than a kilometer to a few hundred meters). In the proposed thesis, the objective is to analyze how remote sensing can be integrated with an index-based insurance. The application framework of the thesis is the forage insurance in France.