Towards the chemical verification of terroir using spectroscopy and machine learning for wine classification

Abstract

Terroir influences winegrape production and thereby wine quality, style, product image, and ultimately the reputation of a region. Typically defined in terms of climate, soil, topography, and the like, there could also be chemical verification of terroir based on the effect of biophysical factors on grape composition. This could be especially useful when focusing on finer scales of terroir, such as subregion or even single vineyard. To enable the implementation of data-driven terroir classification, this study employed an absorbance-transmittance and fluorescence excitation-emission matrix (A-TEEM) approach to determine the molecular fingerprints of bottle-aged Shiraz research wines produced from five subregions of the Barossa Valley in South Australia. Classification models were developed from A-TEEM data using extreme gradient boosting discriminant analysis (XGBDA) with cross-validation, yielding 100% accuracy for prediction of vintage year and 99.5% accuracy for subregion. Using an external validation approach based on splitting the data into training and testing sets, vintage year and subregion classification accuracies remained impressive, at 98.8% and 93.8%, respectively. Addressing model stability over time, classification of a subset of the bottle-aged wines using a previously developed XGBDA model yielded 100% correct class assignment according to vintage year and over 90% accuracy according to subregion for wines from 2018 and 2021. Importantly, 2021 wines were not included in the original model, which highlights the robustness of the approach when analysing new wines as well as those that have aged since the generation of the model. As such, the influence of terroir on wine molecular fingerprints was conserved over time and upon ageing of wine in bottle. Considering the close proximity of the subregional sites, this work emphasises the potential of A-TEEM and machine learning to objectively classify terroir at a fine scale based on potentially subtle differences in wine composition.

References

Bramley, R. G. V., Ouzman, J., Sturman, A. P., Grealish, G. J., Ratcliff, C. E. M., & Trought, M. C. T. (2023). Underpinning terroir with data: Integrating vineyard performance metrics with soil and climate data to better understand within-region variation in Marlborough, New Zealand. Australian Journal of Grape and Wine Research, 2023, 8811402. https://doi.org/10.1155/2023/8811402

Brillante, L., Bonfante, A., Bramley, R. G. V., Tardaguila, J., & Priori, S. (2020). Unbiased scientific approaches to the study of terroir are needed! Frontiers in Earth Science, 8, 539377. https://doi.org/10.3389/feart.2020.539377

Fan, S., Bindon, K. A., Gilmore, A. M., & Jeffery, D. W. (2025). Chapter Two – Fluorescence spectroscopy for grape and wine compositional analysis and quality control. In D. Cozzolino (Ed.), Advances in Food and Nutrition Research (Vol. 115, pp. 45-130). Academic Press. https://doi.org/https://doi.org/10.1016/bs.afnr.2025.01.004

Koljančić, N., Furdíková, K., de Araújo Gomes, A., & Špánik, I. (2024). Wine authentication: Current progress and state of the art. Trends in Food Science & Technology, 150, 104598. https://doi.org/https://doi.org/10.1016/j.tifs.2024.104598

Ranaweera, R. K. R., Bastian, S. E. P., Gilmore, A. M., Capone, D. L., & Jeffery, D. W. (2023). Absorbance-transmission and fluorescence excitation-emission matrix (A-TEEM) with multi-block data analysis and machine learning for accurate intraregional classification of Barossa Shiraz wine. Food Control, 144, 109335. https://doi.org/https://doi.org/10.1016/j.foodcont.2022.109335

Ranaweera, R. K. R., Capone, D. L., Bastian, S. E. P., Cozzolino, D., & Jeffery, D. W. (2021a). A Review of Wine Authentication Using Spectroscopic Approaches in Combination with Chemometrics. Molecules, 26(14), 4334. https://www.mdpi.com/1420-3049/26/14/4334

Ranaweera, R. K. R., Gilmore, A. M., Bastian, S. E. P., Capone, D. L., & Jeffery, D. W. (2022). Spectrofluorometric analysis to trace the molecular fingerprint of wine during the winemaking process and recognise the blending percentage of different varietal wines. OENO One, 56(1), 189-196. https://doi.org/10.20870/oeno-one.2022.56.1.4904

Ranaweera, R. K. R., Gilmore, A. M., Capone, D. L., Bastian, S. E. P., & Jeffery, D. W. (2021b). Authentication of the geographical origin of Australian Cabernet Sauvignon wines using spectrofluorometric and multi-element analyses with multivariate statistical modelling. Food Chemistry, 335, 127592. https://doi.org/https://doi.org/10.1016/j.foodchem.2020.127592

Ranaweera, R. K. R., Gilmore, A. M., Capone, D. L., Bastian, S. E. P., & Jeffery, D. W. (2021c). Spectrofluorometric analysis combined with machine learning for geographical and varietal authentication, and prediction of phenolic compound concentrations in red wine. Food Chemistry, 361, 130149. https://doi.org/https://doi.org/10.1016/j.foodchem.2021.130149

Ranaweera, R. K. R., Gilmore, A. M., & Jeffery, D. W. (2024). Fluorescence Spectroscopy for Red Wine Authentication. In M. Á. Pozo-Bayón & C. Muñoz González (Eds.), Wine Analysis and Testing Techniques (pp. 23-38). Springer US. https://doi.org/10.1007/978-1-0716-3650-3_3

Ríos-Reina, R., Camiña, J. M., Callejón, R. M., & Azcarate, S. M. (2021). Spectralprint techniques for wine and vinegar characterization, authentication and quality control: Advances and projections. TrAC Trends in Analytical Chemistry, 134, 116121. https://doi.org/https://doi.org/10.1016/j.trac.2020.116121

Schartner, M., Beck, J. M., Laboyrie, J., Riquier, L., Marchand, S., & Pouget, A. (2023). Predicting Bordeaux red wine origins and vintages from raw gas chromatograms. Communications Chemistry, 6(1), 247. https://doi.org/10.1038/s42004-023-01051-9

Schmidtke, L. M., Bastian, S. E. P., Bindon, K., Bonada, M., Boss, P. K., Bramley, R. G. V., Danner, L., Petrie, P. R., Gonzaga, L. S., & Collins, C. (2024). Exploring interactions between vineyard performance, grape and wine composition and subregional boundaries—The terroir of Barossa Shiraz. Australian Journal of Grape and Wine Research, 2024(1), 2622516. https://doi.org/https://doi.org/10.1155/ajgw/2622516

Souza Gonzaga, L., Capone, D. L., Bastian, S. E. P., & Jeffery, D. W. (2021). Defining wine typicity: sensory characterisation and consumer perspectives. Australian Journal of Grape and Wine Research, 27(2), 246-256. https://doi.org/https://doi.org/10.1111/ajgw.12474

Wang, H., & Jeffery, D. W. (2024). Machine learning model stability for sub-regional classification of Barossa Valley Shiraz wine using A-TEEM spectroscopy. Foods, 13(9), 1376. https://www.mdpi.com/2304-8158/13/9/1376

Acknowledgments

Access to the research wines was made possible due to a Wine Australia funded project (UA1602) and is gratefully acknowledged. John Gledhill from WIC Winemaking Services is thanked for his assistance with wine sample collection. Technical support with the Aqualog instrument provided by Ruchira Ranaweera, Adam Gilmore, and Andrew Jane is appreciated. H.W. received financial support from the Mortlock Honours Scholarship and project funding from the School of Agriculture, Food and Wine, The University of Adelaide.