Analyzing Weather Patterns to Predict Wine Quality for Sonoma County Pinot Noir.

摘要

Every year the average quality of wine produced in any particular region rises or falls to some degree. Differences in climate conditions each year are generally considered to be the main cause of this annual variation. In an effort to correlate weather with wine quality, software was developed to analyze weather patterns and make predictions on wine scores. Various weather data including temperature highs and lows, precipitation, temperature range and degree day were correlated to regional vintage wine scores using a genetic algorithm, linear regression, and an artificial neural network. Different calculations of weather data within a wide range of time periods throughout the growing season were evaluated in order to find ones that had a mathematical effect on the wine scores.;Pinot Noir from Sonoma County and neighboring areas was focused on, due to its thin skin and climate-sensitive nature. Sonoma County was selected for its large Pinot Noir production, numerous small AVA's, and climate differences among these distinct regions. Climate data was acquired from 14 different weather stations around Sonoma County up through 2007, and wine quality was represented by wine scores from Wine Spectator magazine. Thirteen separate regions were differentiated, and weather data and vintage scores were created for each one. Each region was individually analyzed for correlations between its weather data and vintage wine scores. Along with this, all of the regions were combined together and evaluated as a single area. Each region had approximately 17 years of data, with the total combination consisting of a set of 164 points, one for each year of each region.;High correlations between specific weather periods and vintage scores were found for each region, and accurate wine score predictions were made based on these. The weather periods and calculations used to make these predictions were studied, and consistent trends were found that linked them to different stages of berry development.