Section: New Results
Estimation of hail damages on vineyards from SPOT images
In the framework of a research contract with the LYNX company, the Clime team analyses the detection and quantification of damages caused to vineyards by hail events. A sequence of SPOT images, acquired before and after the event, is used for detected and measuring changes of vegetation indices and to relate them to production losses. This is an application of great potential interest for insurance companies.
The researches address the following issues:
Proving the feasibility: changes of vegetation indices are compared at the parcel level to production losses, as estimated by insurance experts, or to objective loss measurements, obtained from harvesting data. The potential of satellite data has been proved, as well as the inaccuracy of insurance expertises.
Analysing the causes of variability: a sequence of 6 monthly images is used to assess the influence of the observation dates on the loss estimation; the nature of grapes (Chiraz, Sauvignon, etc.) is also addressed on a well documented site.
Definition of an operational prototype, able to forecast production losses with a limited number of satellite images and without requiring harvesting data nor insurance expertises. The key issue is the ability to forecast the production of a parcel from satellite data. Encouraging results have been obtained by using an adaptation of a yield production model specifically designed for wheat. This model is fed by vegetation information throughout the vegetative period, obtained by fitting a phenological model to the satellite measurements.