Phenology Modeling and Frost Risk Assessment with Climate Change for Temperate Fruit Crops

A phenology model utilizing daily min/max temperatures was developed for fruit crops in New York State including apples and grapes. The model  includes sequential stages to first meet chilling requirements, followed by degree day (DD) accumulations to trigger spring growth. This basic model was substantially improved by using a running average of the previous 10 days’ temperatures to modify potential DD increments.  Warmth following prolonged cold weather counts less towards development.  This rule affects predictions of budbreak to a greater degree than flowering stages.  Compared to 10-45 years of observations for different cultivars, 6 phenological stages of apple and 2 for grape are predicted by the model with a root mean square error ranging from 1.5-3.5 days across all stages and datasets, and averaging between 2 and 2.5 days across stages for different cultivars. Several stages of apple development between budbreak and petal fall are addressed in the model, spanning a period of about 4-6 weeks, and with progressively higher temperature thresholds for frost damage.  The phenology model was used to compare historic and future predictions of frost damage at different stages of development based on climate change projections averaged from an ensemble of 16 GCMs.  Predictions encompassed both modest increases and decreases in overall frost damage frequency depending on cultivar-specific characteristics of chilling requirement and DD thresholds for development. A strong qualitative shift is predicted for all cultivars in the most likely phenological stages to be affected.  In historical analyses, the stages of pink, bloom, and petal fall are most likely to experience damaging frost. In future scenarios of high emissions, the stages of greatest vulnerability shift by the 2080s to earlier developmental stages of green tip, ½ inch green and tight cluster.  This is due to the interplay of several factors affecting the rate at which budbreak and subsequent development all shift to earlier mean calendar dates under a warming climate. These dynamics affect all stages with a net beneficial effect of frost likelihood decreasing during flowering.  The earlier budbreak stages, however, eventually shift from April into early March, and are then more subject to the greater inherent variance in mid-latitude temperatures which occurs in winter versus summer. Seasonal patterns of variance in daily temperature are related to seasonal patterns of insolation and related atmospheric circulation which are likely to be similar during climate change despite rising global and regional mean temperatures.