EMS Annual Meeting Abstracts
Vol. 21, EMS2024-70, 2024, updated on 05 Jul 2024
https://doi.org/10.5194/ems2024-70
EMS Annual Meeting 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.
Poster | Wednesday, 04 Sep, 18:00–19:30 (CEST), Display time Wednesday, 04 Sep, 08:00–Thursday, 05 Sep, 13:00|

Vintage Port quality under climate change: A machine learning approach using Twentieth Century Reanalysis and CMIP6 data.

Helder Fraga, Nathalie Guimarães, and João A. Santos
Helder Fraga et al.
  • Centre for the Research and Technology of Agro-Environmental and Biological Sciences (CITAB), Inov4Agro, University of Trás-os-Montes e Alto Douro (UTAD - PT501345361), Vila Real, Portugal (hfraga@utad.pt)

The Douro region is renowned for its exceptional wines, notably the renowned Port Wine. Vintage years, occurring approximately 2-3 times per decade, signify outstanding quality, driven by optimal climatic conditions that enhance grape attributes. However, climate change presents challenges as rising temperatures and extreme weather events impact viticulture. This study evaluates the climatic influence on vintage years and forecasts climate change effects for forthcoming decades, utilizing machine learning algorithms. Historical vintage data spanning from 1850 to 2014 was gathered alongside monthly climatic variables including temperature, precipitation, humidity, solar radiation, and wind components from the 20th Century Reanalysis dataset. Diverse machine learning algorithms were deployed for classification, augmented by statistical analysis to pinpoint pertinent climate variables. Model training and evaluation employed cross-validation, followed by hyperparameter tuning for the most effective models. Future climate projections from 2030 to 2099, under various socio-economic scenarios (IPCC SSP2, SSP3, and SSP5), were integrated, with quantile mapping bias adjustment applied to refine future climate data. Historical data revealed vintage years occurring 23.6% of the time, averaging two vintage years per decade, with a slight upward trend. Crucial climate variables influencing vintage year occurrence were identified, including precipitation in March, air temperatures in April and May, humidity in March and April, solar radiation in March, and meridional wind in June. The study found promising results from Logistic Regression, SVC, and XGBClassifier models. This research offers valuable insights into the nexus between climate variables and wine vintage years, empowering winemakers to make informed decisions regarding vineyard management and grape cultivation. The projections underscore the importance of adaptation strategies in confronting the challenges posed by climate change to the wine industry. This research was funded by National Funds by FCT – Portuguese Foundation for Science and Technology, under the projects UIDB/04033/2020 (https://doi.org/10.54499/UIDB/04033/2020) and LA/P/0126/2020 (https://doi.org/10.54499/LA/P/0126/2020).

How to cite: Fraga, H., Guimarães, N., and A. Santos, J.: Vintage Port quality under climate change: A machine learning approach using Twentieth Century Reanalysis and CMIP6 data., EMS Annual Meeting 2024, Barcelona, Spain, 1–6 Sep 2024, EMS2024-70, https://doi.org/10.5194/ems2024-70, 2024.