Fall 2013

Document Type

Master's Thesis (Open Access)

Degree Name

Master of Science (M.S.)


Science & Environmental Policy


Wine grape quality is heavily influenced by a combination of soil properties and site topography. We used anthocyanin content and yield data from Vitis vinifera L. cv. Cabernet Sauvignon from a vineyard near Madera, California collected during the 2007 growing season. We compared sets of hypotheses regarding the anthocyanin content of winegrapes and vineyard yields as a function of vineyard soil and topographic properties. Each hypothesis was expressed as a regression model predicting a response variable (yield or anthocyanin content) from one or more predictor variables. We used a multiple working hypotheses approach to compare these models using information theoretic criteria (AIC). There was substantial evidence that soil properties affected both anthocyanin content and yield. The top four anthocyanin models received 94% support while the top yield model received 68% support of all models considered. The null models received no support (AICw = 0.00). The predictive 2 power of both the model-averaged anthocyanin content and yield was relatively small (R2 = 2 0.04, R2 = 0.07, respectively). It is likely that greater predictive power could be achieved through the use of more finely-detailed spatial maps and data from additional vineyards.