One of the richest and most underutilized phenological datasets in the world is the viticulture record from Bordeaux, France. Wine production began in this region as early as first century and harvest and vintage quality information are available for some premiere chateaux from the mid-nineteenth century. Phenological data for the Bordeaux region as a whole are available since 1952, a period coincident with high quality weather data.
Bordeaux produces some of the world's finest wines. Despite the importance of wine to the regional economy, little research has been done to examine the impact of daily weather variations on viticulture variables over the course of an annual growth cycle. After removing technological trends, it is believed that most of the remaining year-to-year variability can be ascribed to the climatic conditions over the course of the year. The purpose of this research is to compare Bordeaux grapevine phenology to three weather variable sets: 1) a mean climatology (temperature, precipitation, and insolation); 2) several growing degree-day measures; and 3) circulation and air mass type frequencies based on regional and local synoptic climatologies.
Phenological dates for three grapevine stages--floraison (flowering), veraison (ripening), and harvest—are recorded for the Bordeaux region from 1952–1996. A phenological event is deemed to occur when at least 50 percent of the plants exhibit that growth stage.
Two temporal synoptic climatologies are produced by applying cluster analysis to the principal component scores of the climate data (a procedure called the Temporal Synoptic Index)--the first using four daily observations of the following weather variables recorded at the Bordeaux weather station: temperature, dew point temperature, station pressure, cloud cover, and the u and v components of the surface wind vector; and the second using daily 5 by 5 degree grid cell values of sea level pressure over the North Atlantic, western Europe, and the Mediterranean. Before running the procedure, both data sets are adjusted according to each phenological season. In addition, phenological-season mean climate characteristics are recorded for each year as well as traditional growing degree-days using both triangular and sinusoidal estimates for diurnal temperature variations.
Using phenological date as the dependent variable, multiple linear regression is used to relate three sets of independent climate variables for each phenological season: 1) annual relative frequency departures of each air mass and circulation type; 2) mean climate parameters; and 3) growing degree-days.
Preliminary results indicate that there is a stronger relationship between air mass and circulation frequencies and phenological timing than with mean climate characteristics or growing degree-days. A small sample of synoptic events are identified that affect viticultural potential throughout the year by regulating the timing of the plant’s physiology. These findings suggest that a variety of weather variables impact grapevine phenology beyond the standard measures of accumulated heat and moisture. Furthermore, similar air masses can have significantly different impacts depending on the growth stage of the plant.