Use of probabilistic expert elicitation for assessing risk of appearance of grape downy mildew

M. Chen,F. Brun,M. Raynal,C. Debord,D. Makowski,M. Chen,F. Brun,M. Raynal,C. Debord,D. Makowski
DOI: https://doi.org/10.1016/j.cropro.2019.104926
IF: 3.036
2019-12-01
Crop Protection
Abstract:<p>Grape downy mildew (GDM) is a major disease of grapevine and the date of appearance of its first symptoms is a determinant information for the protection of the vineyard. Probabilistic elicitation of experts has been used here to estimate this date. In 2017 and 2018, 29 experts were elicited to provide probability distributions of dates of GDM appearance between April and June, for different plots. The results of these elicitations show that the experts' forecasts and their uncertainty change over the season. The elicited dates tend to be earlier at the beginning of the season and later at the end of the season, by about 18 days in average. In April 2017 and 2018, most of the elicited dates are too early compared to observed dates of GDM symptom appearance. However, this bias becomes negligible in the month of May. Our results indicate that probabilistic elicitation is a useful tool for communicating expert judgments and uncertainties in a transparent manner in plant disease risk assessment and epidemiological alert bulletins.</p>
agronomy
What problem does this paper attempt to address?