Distinguishing Terroir Effects Using NMR and ECVA Analysis

An aside before this post begins….

This is the 100th article reviewed by The Academic Wino!!!  We’ve posted over 130 total posts on this blog, but this one is the 100th peer reviewed article presented!  I can’t believe I’ve read 100 papers so far!  Here’s to the next 100!


Throughout the food and beverage industry, particularly in the high-end wine business, fraud is an ever present and serious threat to the authenticity of the product and the industry.  Wine, in particular, is relatively easy to defile, since it is very chemically complex and changes may go unnoticed if not examined thoroughly.  Wine consists of hundreds of compounds that vary depending upon many factors, including but not limited to grape variety, environmental conditions, and winemaking techniques.  All of these combined result in a wine that is unique, and that can be analyzed through several methods to determine its authenticity, including infrared spectroscopy, X-ray absorption, and dielectric fingerprinting.

The primary compounds in wine are formed primary as a result of the alcohol and malolactic fermentations of grape must.  In addition to having good quality grapes, control of the fermentation processes (as well as other steps in the winemaking process) are critical in producing a high quality wine.  To monitor quality of wine during these processes, several methods may be employed, though the use of Nuclear Magnetic Resonance, or NMR. has more recently been investigated as a possible method for doing so. 

Previous studies have shown that by using NMR analysis, they were able to discriminate between grape samples from different environments from different locations in southwest France.  This suggests that NMR analysis may be a successful method for analyzing authenticity and/or quality of wine all the way down to the level of “terroir”; a term which encompasses the specific environmental characteristics of a particular site that all contribute to create a unique finished wine.

The goal of the article presented today was to analyze the winemaking process from wines made in Rioja, Spain by NMR metabolomic fingerprinting and advanced chemometrics to evaluate the chemical differences between specific events during the winemaking process, the vintage, the geographical origin, as well as specific wineries.


9 winemaking cooperatives were selected for this study, including three from Rioja Baja, five from Rioja Alta, and one from Rioja Alavesa.  Specifically, the cooperatives selected were from Arnedo, Alcanadre, Arenzana de Abajo, Navarette, Haro, San Asensio, Uruñuela, and Labastida.  Vintages studied were from 2006, 2007, and 2008.  Five samples from each cooperative at different time points during the fermentation process were collected.  Time points were 1) before alcoholic fermentation; 2) at the end of alcoholic fermentation; 3) the beginning of malolactic  fermentation; 4) middle of malolactic fermentation; and 5) after malolactic fermentation. 

In total 111 samples were obtained from three vintages, nine winemaking cooperatives, and five fermentation time points.

Nuclear Magnetic Resonance was performed on all wine samples.

Principle Components Analysis (PCA) was performed on the three main wine NMR spectra analyses, separating the samples into three different regions (aromatic region, carbohydrate region, and organic acid region).  Using PCA analysis on the NMR spectra accounts for different types of functional groups (similar molecular compound structures) that help improve chemometric performance (i.e. the ability to extract information from chemical analysis data).

Extended Conical Variable Analysis (ECVA), which functions to determine which region on the NMR spectra is responsible for separation among different groups, was also performed.


  •  The transition from grape must to wine was evident on the NMR spectra as the disappearance of carbohydrate signals in the must and appearance of alcohol and organic acid signals in the wine.

o   The aromatic content during this stage remained constant.

  • In wine, carbohydrates did not completely disappear, but left a complex fingerprint.
Figure 2 from  Lopez-Rituerto et al, 2012.

  • During alcoholic fermentation, the major variation in the NMR spectra between samples was in the ethanol content.
  • During malolactic fermentation, the major variation in the NMR spectra between samples was in malic acid and lactic acid (not surprisingly).
  • When differentiating between subareas inside Rioja, the NMR spectra showed two distinct separations between Rioja Alta plus Rioja Alavesa and Rioja Baja.

o   Rioja Alavesa could not be distinguished between Rioja Alta, possibly due to its very close proximity to the area.

o   The best region in the spectra to distinguish these subareas was the aromatic region, though the PCA analysis was only able to explain 40.2% of the variation.

  • The three vintages studied were distinguished on the NMR spectra using the aromatic region, though not clearly.
  • Using ECVA allowed the researchers to reduce the error rate and misclassifications to 0 or near 0 in all cases.
  • Using ECVA in conjunction with NMR, all three vintages were clearly distinguished.
  • Using ECVA in conjunction with NMR successfully distinguished not only between subareas of Rioja, but also between individual wineries.

o   The two compound signals that showed the clear distinction between individual wineries in Rioja were isobutanol and isopentanol.


The results of this study suggest that isobutanol and isopentanol may be important biomarkers for differentiating wine from individual wineries in a wine region.  Also, different stages of the winemaking process may be effectively analyzed and distinguished using NMR analysis.

Results show that wines can be differentiated using NMR analysis to different time points during the fermentation process, as well as in different subareas of a wine region, and also different vintages.  Combining NMR analysis with ECVA analysis, wines can be distinguished as specifically as the individual winery level.  These results may be very important in distinguishing between wines if needed for authenticity confirmation or fraud investigations. 

This study only investigated one wine region, thereby further studies would be needed in other wine regions to determine if this type of analysis is applicable on a global scale, or if the results are just a regional phenomenon.

I’d love to hear what you all think about this topic!  Please feel free to leave your comments below!

Source: Lopez-Rituerto, E., Savorani, F., Avenoza, A., Busto, J.H., Peregrina J.M., and Engelsen, S.B. 2012. Investigations of Rioja Terroir for Wine Production Using 1H NMR Metabolomics. Journal of Agricultural and Food Chemistry 60: 3452-3461.

DOI: dx.doi.org/10.1021/jf204361d

I am not a health professional, nor do I pretend to be. Please consult your doctor before altering your alcohol consumption habits. Do not consume alcohol if you are under the age of 21. Do not drink and drive. Enjoy responsibly!

6 comments for “Distinguishing Terroir Effects Using NMR and ECVA Analysis

  1. July 12, 2012 at 4:16 pm

    Hello Becca, and thank you for another great article breakdown. I think that this article does a great job of identifying a biomarker signature by which one could possibly distinguish between different wine regions or even wineries. What the researchers now need to do however, is to now test this signature (eg. isobutanol + isopentanol) on a new dataset, perhaps for 5-10 more years. This will allow them to determine whether their biomarker signature can actually distinguish between wineries, as is suggested, or whether they are just distinguishing between two lots of wine. They would need separate, but consistent biomarker signatures for each winery, that are maintained across vintages in order to make the claim that they can distinguish between wineries. Right now they have preliminary evidence that has provided them with a biomarker signature discovery, and now they need to apply that to a new set of data (new vintages of wine). I'm interested to see their followup articles!
    BTW, I've been a reader of yours for some time now and I really appreciate your work. My day job is biomarker discovery and development for oncology drugs and my (sometimes day) and night job is wine, for Vinopanion – WineLog. Your blog satisfies both of my cravings!

    Vinopanion: Ward Kadel – @drXeNo

  2. Wineknurd
    July 14, 2012 at 2:38 pm

    Ward, I am an analytical chemist by day and connoisseur of all things fermented by night and feel exactly as you do about TAW! These types of articles especially, since I routinely work with NMR and data interpretation.

  3. Wineknurd
    July 14, 2012 at 2:51 pm

    Congratulations on The Academic Wino Centennial Article! Looks like a good start for these researchers, but I have seen other analyses such as these that start to fail once the sample sizes increase and encompass wines outside a specified region. I have spoken to researchers at UC Davis as to the difficulty of being able to define a distinct chemical fingerprint for wine types and styles. The best they can do so far is to determine the likelihood of fraud, at best when testing a suspect wine against a verified "authentic" standard of the same wine and at worst by comparing it to similar wines of that style and vintage. Wine analyses have been ongoing internationally for decades and think, we can smash protons into each other to produce the fundamental particles of physics but the best method we have to determine a style of wine is still by tasting it. Wine is a complex animal indeed!

  4. July 17, 2012 at 3:03 am

    Cheers to the intersection between wine and science, Wineknurd & Becca!

  5. July 17, 2012 at 3:13 pm

    I'm happy to please the chemistry nerds out there 🙂 Thank you both for all your kind words about my blog! I'm so happy to have such awesome readers as you both!

  6. July 17, 2012 at 3:16 pm

    Thank you! Wine is complex, indeed! There was that one study that looked at a noninvasive approach to testing the authenticity of wine, which didn't require opening it up and tasting it: http://www.academicwino.com//2011/09/combating-win… I'm certain research will find other noninvasive methods…it's just a matter of time!

Comments are closed.