2019 ASHS Annual Conference

Weather conditions have a significant influence on the development and interannual yield of lowbush blueberries (19 to 3826 kg/ha for the years 1988 to 2018). A bioclimatic model that would predict the phenological development and yield of lowbush blueberries would be very useful for planning field activities (e.g. pollination, pest control), harvesting operations and marketing of this crop. In addition, such bioclimatic models would be very useful in climate change studies to provide future projections for this crop in eastern Canada. Several weather factors can influence the productivity of lowbush blueberries. An increase in fall temperatures could delay hardening and provide less protection against winter frosts. A lower snow cover would increase the risk of plants being exposed to extreme cold. Higher temperatures in the spring could anticipate bud break and make plants more vulnerable to late spring frosts. The severity of such damage varies mainly according to frost intensity and the phenological stage of the crop. To build this bioclimatic model, lowbush blueberry yield data obtained for the Saguenay Lac-St-Jean (SLSJ) regions from 1988 to 2018 were used. The daily weather data used for these years was obtained from the Canadian climate database at a resolution of 10 km. To determine links with weather data, the following variables were selected: minimum, maximum and average temperatures, precipitation, potential evapotranspiration, estimated snow depth and estimated soil temperature at 5 cm. A bioclimatic model that predicts in-season phenological stages of lowbush blueberries was also developed, implemente and used for the yield model. For the bioclimatic model corresponding to the day after the "Mature Fruit - Harvest" stage, the variables with the highest correlation with yield are the cumulated weighted temperatures during the flowering period, the number of frost days during the flowering period, the maximum snow depth in March and the average rainfall during flowering. The proposed model explains 64% of the interannual variability in yield between 1988 and 2018. Excluding the year 1998, in which a major killing frost occurred during the flowering period and resulting in a yield of only 19 kg/ha, the model using the same weather variables explained 82% of the variability in yield. Such a model is therefore very promising as a decision-making tool for lowbush blueberry producers of SLSJ regions in eastern Canada.