Abstract: Phenology is an important indicator of climate changes and informs us about the compliance of the variety with climatic conditions. Plant phenology is most often predicted by the sum of degree days, but the accuracy of forecasts is insufficient, especially in conditions of climate change. Regression analysis of 71 grape varieties from Y. I. Potapenko Don ampelographic collection (Russia) in 1981-2014 revealed relationship between grape phenodates (bud break, flowering) and dates of transition air temperatures above temperature thresholds of these phases. On this basis, we have created a simple dynamic model of grape phenology. The time from flowering to ripening was assumed constant for variety. In was realized as a computer program SEASONS. The purpose of this work was to verify and validate the model. Model was parametrized and verified for 12 grape varieties in 1981-2014. Model was validated on the test period of 2015-2017. The approximation error of the SEASONS was 3-5 days for the date of bud break and flowering and 6-9 days for ripening, or 5.1-8.0% for production period duration. The forecast error for production period for the model was 4%, on the base of DD was 11%. This result confirms the quality of the model and its suitability for forecasting.
Keywords: grape, phenology, temperature thresholds, dynamic model