Abstract Joel Harms | Jan Adamowski | Viacheslav Adamchuk | Nathaniel Newlands | Simone Castellarin

Mapping Global Future Potential for Pinot noir Cultivation under Climate Uncertainty using Generative AI

Joel Harms,* Jan Adamowski, Viacheslav Adamchuk, Nathaniel Newlands, and Simone Castellarin
*McGill University, 21111 Lakeshore Road, Sainte-Anne-de-Bellevue, Quebec H9X 3V9, Canada (joel.harms@mail.mcgill.ca)

This study addresses the effect of climate change on the global wine industry, specifically focusing on the suitability of regions for cultivating major international grape varieties. With the increasing challenges posed by climate change, understanding how shifts in climate may affect the quality and production of single-varietal wines is crucial. We propose using a climate-based wine variety recommendation system, using deep-coupled autoencoder networks, to predict regions that will likely undergo improvements or declines for key grape varieties. We tested this approach by predicting possible future Pinot noir regions globally. The system was fine-tuned and evaluated using vintage scorings from representative regions over the past ~30 years, using scorings from multiple respected wine critics. Future predictions are mapped using existing climate models under +2C and +4C scenarios from the TerraClimate Dataset. We use derived climate indicators to identify regions with the greatest potential for Pinot noir. Our findings indicate significant shifts in the suitability of regions, particularly in areas previously considered too cold. This study demonstrates the practical application of wine-recommendation systems in adapting to changing climate conditions and provides valuable insights for the wine industry. By fine-tuning these systems for specific tasks, such as predicting suitable regions for a specific varietal, the wine industry can proactively address the challenges posed by climate change and make informed decisions for sustainable viticulture. This research highlights the wide-ranging possibilities of wine-recommendation systems, showcasing their potential to enhance decision-making processes within the wine industry amid evolving climate conditions.

Funding Support: This research was supported by funds from the Canada Graduate Scholarships-Master’s (CGS M) Program by the Natural Sciences and Engineering Research Council of Canada held by JH.