Evaluation of Six Models to Estimate Ascospore Maturation in Venturia pyrina

Abstract

Estimates of ascospore maturity generated by models developed for Venturia pyrina in Victoria, Australia (NV and SV), Oregon, United States (OR), and Italy (IT) or for V. inaequalis in New Hampshire, United States (NH-1) or modified in Norway (NH-2) were compared with observed field ascospore release of V. pyrina from 21 site–year combinations. The models were also compared with ascospore release data from laboratory assays. In the laboratory assays, the forecasts of the NH-1 and NH-2 models provided the best fit to observed spore release. Under field conditions, the lag phases and slope coefficients of all models differed from those of observed release of ascospores. Identifying the precise time of bud break of pear to initiate degree-day accumulation was problematic at both Australian sites. This resulted in a higher deviance between bud break and first released ascospore compared with the sites in Norway and Belgium. Linear regressions of observed release against forecasted maturity generated similarly high concordance correlation coefficients. However, where differences were noted, they most often favored models that included adjustment for dry periods. The NH-2, IT, and NV models using pooled data also provided the most accurate estimates of 95% ascospore depletion, a key event in many disease management programs.