To define the spatial extension of vineyard mortality over a region and to precise the respective contribution of its different possible causes requires to visit a sufficient number of fields in order to include the variety of vegetal material, soil types and agricultural practices which can be encountered in the studied region. To reduce the cost of these investigations, we propose to evaluate the vineyard mortality by estimating, for each field, the rate of missing vinestocks from high resolution aerial photographs. In order to identify possible causes of vineyard mortality, this information is then combined with maps of candidate factors, e.g. soil, age of vineyard, vine-variety, weeding technique, training system, by using a Geographical Information System. The proposed approach is tested over four sampling areas of 356 ha and 196 vineyard fields which are representative of the lower Peyne valley (Herault -South of France). The main results are as follows :
- The rates of vineyard mortality are satisfactorily predicted from aerial photographs (RMSE = 6.5 p. cent),
- The average rate of vineyard mortality is high (e.g. 35 p. cent of the fields have more than 12 p. cent of death vinestocks),
- The rates of vineyard mortality are extremely variable both from a sampling area to another and from a field to another within a given sampling area.
- They seem to be more in relation with soil and agricultural practices than with age or vine-variety. Particularly, colluviosols with redoxic horizon developped on colluvion of miocene marine sediments reveal rates of vineyard mortality which are significantly higher than other soil types
AttachmentsNo supporting information for this article