Is the Developmental Stage of Grapevine Clusters a Good Predictor of Ontogenic Resistance to Plasmopara viticola ?

Reduction patterns in grapevine berry cluster susceptibility to downy mildew (DM) due to ontogenic resistance were investigated under different viticultural conditions in northern Italy to develop mathematical equations for predicting the infection risk of Plasmopara viticola . Clusters were inoculated with sporangial suspensions of P. viticola at different developmental stages (DS), and disease severity was assessed twice weekly. The best fit was obtained using time-dependent logistic equations in which time was expressed as cluster DS (R 2 = 0.876), days after DS65 (full flowering) (R 2 = 0.880), or accumulated degree-days (base 0°C) after DS65 (R 2 = 0.882), all with a mean absolute error below 9%. The fitted curves consistently showed high susceptibility before and during early flowering, followed by rapid increase in ontogenic resistance after full flowering, sharp decline in susceptibility, and asymptotic approach to complete resistance. The inflection point, representing the fastest decrease in susceptibility, occurred at approximately DS72, 10 days after DS65, or 245 degree-days. Bands were empirically generated so as the Mean Absolute Error for the excluded out-of-band observations remained below the 2% disease severity threshold. Bands were widest during the rapid decline in susceptibility, reflecting variability among viticultural conditions. The DS-based equation, together with the bootstrapping variability in its parameter estimation and experimental error, was used in a Monte Carlo framework to provide a stochastic distribution of the predicted DM severities. Unlike deterministic approaches, which provide a single estimate of susceptibility at a given developmental stage, stochastic approaches produce a range of possible susceptibility levels with associated probabilities.