Testing and modelling the effects of climate on the incidence of the emergent nut rot agent of chestnut Gnomoniopsis castanea

Gnomoniopsis castanea is an emerging fungal pathogen causing nut rot of sweet chestnut, Castanea sativa. This study was aimed at testing and modelling the effects of climate on disease incidence. Up to 120 ripe nuts were collected in 2011 from trees in each of 12 sites located in the northwest of Italy. The incidence of G. castanea in each site was expressed as the number of infected nuts out of the total number of nuts sampled (%), determined by combining the results of morphological identification of isolates obtained from nuts, and their typing through a newly developed taxon‐specific molecular assay. Disease incidence ranged from 20 to 93%, depending on site. Geostatistical analyses revealed that, despite the clustering of sites (P < 0·05), disease incidence was not spatially autocorrelated (P > 0·05). This finding suggests that the disease is influenced by site‐dependent factors whose scale (c. 7·5–15·6 km) is consistent with the climate variability throughout the sampling region. Multivariate analyses on maximum, mean and minimum temperatures and on rainfall showed that warmer temperatures were associated with higher levels of disease incidence. The temperatures of months before nut harvesting were selected as predictors for partial least squares regression (PLSR) models (GnoMods) of G. castanea incidence. External validation on data collected either on sites or in years not used for model fitting showed the good predictive abilities of the GnoMods (Spearman's ρ_obs/pred > 0·72, P < 0·05). The above findings support a relationship between climate and incidence of G. castanea, providing statistical tools to forecast disease incidence at site level.