Science and the Future of the Wine Industry

In this exclusive interview with Antonio Graça head of Research and Development at Sogrape, Meininger’s editor at large, Robert Joseph picked his brain about topics ranging from Portuguese and international grape varieties, to gene-editing, the role of wild yeasts in terroir character, robot tasters – and his latest climate research project that brings together wine, wheat and olives.

Reading time: 17m 55s


Antonio Graça at Mundus Vini (Photo: Ad Lumina / Ralf Ziegler)
Antonio Graça at Mundus Vini (Photo: Ad Lumina / Ralf Ziegler)
The Knowledge Broker

Antonio Graça has been director of Research and Development for the Portuguese wine company Sogrape since 2018. For the last seven years he has also been secretary of the OIV Sustainable Development and Climate Change Viticulture Commission. His responsibilities at Sogrape have included developing a global approach to sustainability, launching and managing a broad range of research projects with Portuguese universities and enterprises, and establishing a management model for business research and development for the wine industry. A widely-respected taster, he is one of the Mundus Vini competition’s top panel chairs.

Meininger’s: You are an academic who works for a large wine company. Can you add a bit more flesh to the bones of the statement? What do you actually do?

Graça: I would not dare call myself an academic out of respect to the several academics that work every day for the advancement of wine science. I made my university studies to obtain an Enology degree and, after working as an oenologist and winemaker in the Douro, was offered the opportunity to build and head Sogrape’s Research and Development department.

My role is bidirectional. On the one side, I make sure my colleagues have access and capacity to use the most advanced scientific knowledge in their daily activity to please and surprise wine consumers in a sustainable way. Conversely, I provide academics with a playground to test ideas and hypotheses by granting access to a wide variety of vineyards, grapes, wine types and process data while steering their theoretical approaches into practical tools and methods. You could say I’m a knowledge broker. In addressing challenges, I know who has or where the science is for each specific issue in the grape and wine business, and I engage to materialize it as value for Sogrape and, in many cases, also for the wider wine sector.


Meininger’s: On March 29, a Climate Change project called MED-GOLD in which you are involved is going to release some of its findings. What is it, and what is it seeking to achieve?

Graça: Project MED-GOLD is one of the most exciting projects I have been involved with. It is an international collaborative project of scientists, software programmers and agri-food professionals uniting a total of 16 organizations from seven countries. It spans three food sectors of the Mediterranean basin: wine, olive oil and pasta and represents an investment of nearly €5m by the European Commission’s Horizon 2020 funding program to create climate services for those three sectors.

Uniting scientists, software programmers and agri-food professionals of 16 organizations from seven countries and three food sectors to create climate services

A climate service is a customized product based on climate data that provides a useful function for someone or something. The beauty of MED-GOLD is it is creating, testing and validating services that were co-designed with, and for, users from those sectors, from farmers to technicians to consumers, allowing them to plan and engage climate change adaptation with some of the best climate science of the world on their side.

MED-GOLD places the huge amounts of climate data accumulated by Europe’s Copernicus satellite earth observation program - a loose equivalent to NASA and NOAA in the US - at their fingertips in a clear, easy-to-understand and quick-to grasp-way. Using these innovative services, their decisions become climate-smart, with no more relying on memory, inadequate data from unrepresentative far-away weather stations, or local unverified amateur devices. MED-GOLD provides not only some of the most reliable data in the world, but also clear and understandable measures of their quality, informing on how reliable they are.

MED-GOLD: ICT Platform
MED-GOLD: ICT Platform

Meininger’s: How does MED-GOLD relate to the International Wineries for Climate Action?

Graça: They are very different but complement each other. IWCA is an organization working to mitigate the contribution of the global wine sector to climate change by certifying self-implemented measures to reduce greenhouse gases emissions from wine companies. MED-GOLD is a research project producing tools enabling the wine sector’s adaptation to climate change and therefore assisting its survival in a warmer world - one that, at least, ours and the next generation must, beyond any doubt, endure even if IWCA efforts are successful.


Meininger’s: MED-GOLD covers grapes, olives and durum wheat. What are the advantages and disadvantages of a project that is not specifically focused on wine?

Graça: Several synergies were raised from engaging the three sectors together. Many models developed by scientists for one crop can be easily adapted to others, several common issues were important to address jointly because, more than anything else, there is a lot of psychology involved.

One of the first obstacles to overcome was communication. Climatologists use technical jargon that can easily be misunderstood by people outside their science. For example, ‘skill’ is a measure of quality of a prediction, ‘anomaly’ is a deviation from the normal, not necessarily something bad. For a climatologist, a model that consistently produces forecasts that get it right 40% of the time, is a reliable model because it is consistent but, for a grower it is unreliable because it is wrong 60% of the time, and therefore they cannot use it to make correct decisions.

Grape growers, olive growers or wheat growers were consensual about this and, it was really important that scientists faced that consensus. Having the three sectors together in the project allowed for the consortium to firstly address these fundamental common issues, finding ways for scientists and users to understand each other and consensually define goals. That was the first essential step that took us to engage the second obstacle: trust or lack thereof.

Climate is not weather. Weather is what you experience, while climate is a statistic analysis of weather over a long period, typically 30 years.

Farmers (and many a citizen) are naturally suspicious of climate predictions because they think about them as they do about the weatherman. But climate is not weather. Weather is what you experience, while climate is a statistic analysis of weather over a long period, typically 30 years.

When non-scientists think about climate, their minds play tricks on them. You say the weatherman is no good because you only remember those forecasts that are important for you (like if it will rain on the day you will harvest or spray). And you will far more readily remember those occasions when it was forecast that it would not rain but it rained, than those others when the forecast is right. And you don’t even take note of the days that were forecast to be rainy but were not, because our brains are hardwired to better remember our causes for loss than for satisfaction. It’s a primal survival mechanism.

Relying on facts and hard data.

The workaround is not to rely on our minds and memories, but on facts and hard data describing what really happened, what was forecast and was the forecast right or not. Over a long term, you start to get a picture of the trend and you understand if you can or cannot rely in the forecasts you have for the decisions you make.

As I said, a lot of psychology is involved. And the project’s success owed much to having in the consortium a social science specialist from the University of Leeds, a climatologist who is also a family farmer growing grapes in Italy, an agronomist who has been modelling the dynamics of pests in agriculture, supercomputing specialists, companies providing decision-support systems for agriculture and an experienced winemaker with a long-standing interest in weather and climate forecasting.

These cross-fertilizing approaches were key elements in sorting out the trust issue as they fueled mutual understanding. This has been important enough to prompt us to publish a scientific paper on our findings about trust in climate services. At the end of the day, MED-GOLD has the potential to make farmers shift their traditional deterministic approach in making decisions to a probabilistic one and, in that way, to be much better equipped to meet the challenges of climate change.

MED-GOLD (Photo: Polina Rytova/unsplash)
MED-GOLD (Photo: Polina Rytova/unsplash)

Meininger’s: The name MED-GOLD refers to the Mediterranean and the project is funded by the EU. How relevant is it going to be to the wine industry in northern Europe, in countries like Germany that are a long way from the Mediterranean – or to the wine industry in the New World?

Graça: When MED-GOLD’s results are showcased at the end of March, people will see that one of the best results that has been created, demonstrated, tested and validated by professionals from the three sectors, is a tool we named ‘MED-GOLD Dashboard’. Very simply, it is a set of software modules that perform magic.

We have the Copernicus Climate Data Store (CDS), a huge data lake with hundreds of petabytes (one PB equals a million Gigabytes) of climate data and incorporating eight more petabytes every year from sensors onboard satellites, airplanes, ships and from thousands of weather stations and sea buoys across the world. These data describe what is going on in the entire atmosphere of the planet every day. The dashboard was built in a way that it can go to the CDS, fish out the data it needs, say from a particular area of the world during a specific period, perform calculations to obtain indicators relevant for the crop sector (for example the sum of rainfall during spring) and transform the resulting information as an interactive map that can be easily visualized and manipulated by anyone with almost no training, just following onscreen guidance.

The dashboard and the back-office software platform are the project’s prototype service, a webservice providing that workflow over Iberia with wine and olive oil indicators, and over Italy with one for durum wheat. In this prototype, we can navigate, month by month, all the way back to 1951 and see how the temperature, rainfall, growing season average temperature, spring rain, harvest rain, annual count of days with more than 35 °C, the annual count of heatwave days, the hydrological balance of soils, etc.

MED-GOLD: Dashboard
MED-GOLD: Dashboard

It also provides forecasts for the next 7 months and long-term projections of the same indicators until the end of the century. We even created two risk indices for viticulture, one for sanitary risk and another for heat-related risk, both based on climatic drivers. This is a prototype, but the system can be scaled up and provide the same type of information in the same easy way for any place in the world, since the source, the CDS, includes data from the whole globe.

And there’s more. Already, during the project, our partner, the Universidad Militar de Nueva Granada in Colombia, explored the possibilities of deploying the same approach for coffee production. Additionally, the OIV has shown a great interest in having such a system available for all wine regions of the world. This would be a major legacy of the project and a great contribution from Europe to global agriculture. News about this will be revealed on March 29th as well.


Meininger’s: Talking to many wine producers and viticulturists today, much of the interest seems to be in recent Vitis vinifera crossings such as Marselan and PIWIs that marry Vitis vinfera with other species of Vitis. How does the MED-GOLD work on traditional indigenous Vitis vinifera relate to those trends?

Graça: There are several ways. MED-GOLD tools allow to visualize how the ideal climate conditions for each variety will be geographically distributed in the future. This is precious information when deciding to plant a new vineyard, choosing a rootstock, a variety or a clone, if you know their climatic requirements. We can learn if the place you own will still have an adequate climate for your traditional varieties or if you should change for others that are more resilient and tolerant to climate stresses. Or if, instead, you will better off selling your land and buying somewhere else where the projected climate will be more favorable for the type of wine you became known for.

Meininger’s: From your research, which grape varieties - Portuguese or international; old or new - would you be backing in the context of climate change?

Personally, I believe the greatest value lies in varieties with higher adaptation capacity, which translates as the ones with the highest genetic diversity. This diversity allows for the selection of their genotypes that are best adapted to different situations which, in turn, will help the variety to be resilient to differing environmental pressures, such as those arising from climate change.

This doesn’t have much to do with being a local or international variety; it relates more to the length of time that has passed since they first appeared in the world and, especially, for how long have they been cultivated. Diversity is created from multiple sequential multiplications from an original scion. The longer a variety has been around and the longer it has been used in cultivation (like quite old and diverse ones such as Pinot Noir or Arinto), the higher the number of vegetative (cane-based) replications it has undergone.

Each replication inserting small replication mistakes in the DNA molecule, creates similar individuals with slightly differing characteristics, which may amount to slightly higher grape production, slightly sweeter berries, slightly more heat-tolerant leaves. After centuries or millennia, the combined set of individuals of those ancient varieties has a large enough variability to allow them to adapt to different contexts of environmental or market requirements (for example, the production of wines with lower alcohol, or of sparkling wine from a variety traditionally used for fortified wine).

If enough of their diversity has been conserved (70 genotypes being the absolute minimum), we can study them, identify which genotypes (aka clones, though this is not the correct scientific term) behave in which way, and select the one best suited to the region and desired wine type.

This is what we have done in Portugal with almost 70 varieties so far - and the reason we were able to develop and propose polyclonal selections as a response to climate change. Conducting this kind of work over the last 40 years offers us an edge when facing future prospects. Sogrape has conservation vineyards for 197 genotypes of Touriga Nacional and 200 of Moscatel Galego Branco, for example. This allows us to view the future of Sandeman, Ferreira and Offley, our Port brands as well as our Casa Ferreirinha Douro wines confidence.

It has also led us to rethink and to experiment with field blends from old vineyards that have given us two new wines. Casa Ferreirinha’s Castas Escondidas - literally ‘hidden varieties’ and Sogrape’s fabulously elegant Legado both come from very old Douro vineyards with rich field blends made up of over 40 varieties. As far as our data show, this diversity has rendered them almost impervious to climate change.


Meininger’s: Sogrape is the only wine company involved with MED-GOLD. Why is that?

Graça: Sogrape identified climate as a top threat for its business a long time ago. I personally have been collecting maturation data from the regions in Portugal where we grow grapes, for more than 30 years. We have our own weather station network that is kept in good work conditions reporting high-quality, validated weather data every 15 minutes which is shared with partner organizations like the Royal Netherlands Meteorological Institute. It’s been running for more than 10 years, and we can clearly see, from its data, the climate drift happening.

We became very conscious that climate adaptation is the big game ahead and set out to find the best science on it. We first joined an earlier European climate project, EUPORIAS, where we learned the difficulties and intricacies of climate services. We also understood that climate scientists needed and welcomed professionals from business sectors that were willing to up their own climate knowledge and understanding and bring their research results into practical usage.

We also made some good friends, which is never too difficult when you work in the wine business. After all, scientists also need to eat, drink and relax, now and then. It was from the interactions during that project that the idea that eventually became MED-GOLD was born. Some of the people in EUPORIAS are also in MED-GOLD, so it was just a natural continuation.

There were other projects in which we also participated. One that has just ended was funded by the World Climate Research Program, an initiative of the World Meteorological Organization run by the United Nations. It addressed the difficult issue of forecasting climate in the interval between seasonal terms (up to 1 year) and the long-term (over 30 years) for several economic sectors. Agriculture was one of these, and Sogrape was the company delivering agri-food requirements for those types of forecasts. On a different, but also very rewarding, case we teamed up with UK’s MetOffice (also a MED-GOLD consortium member) to make music from water stress measurements in the Douro… Really!

Grape varieties
Grape varieties

Meininger’s: The wine world is generally split up according to regional and varietal styles based on a specific set of grapes. How are producers, marketers and, most importantly, consumers going to react to totally unfamiliar grape varieties when it is already hard to get them to accept grapes like Viognier that are thought to be harder to pronounce? Does this mean that the future of wine will increasingly take the form of blends?

Graça: Blends are certainly a valid proposition under the increasing variable and extreme climate we are witnessing. Having worked the first half of my professional life in Douro, I certainly am aware of the possibilities and security provided by blends. But you do not have necessarily to let go of your traditional varieties.

For me, the greatest challenge is for regions that made a name for a specific sensory style based on just one or two varieties. Those will probably suffer for some time until a new balance is attained by either enriching (or re-enriching) their varietal mix or by starting to propose something radically different but also attractive for consumers.

Consequences of climate change: Enriching the varietal mix or starting something radically different?

The optimist in me says that there are more opportunities to develop new businesses under climate change than there are threats to be tackled. If we accept that hard-to-pronounce varieties are an obstacle to business development, we would all be selling Cot or Zinfandel, and Portugal, Georgia and many other countries should forget about mentioning how their wines are made.

This is not what we are seeing. We see an increasingly serious interest from average consumers in discovering new tastes and varieties and I believe wine tourism is going a long way in making them acquainted with exotic names, and even proud they can pronounce them correctly. I do not think articulation will stand in the way of wine adapting to climate change, but it is totally possible that we will see a rising interest for blends among growers and winemakers.


Meininger’s: What are the implications for typicity?

Graça: I believe we are getting past the moment when consumers associated a grape variety with a specific smell and flavour. Today you have so many styles of Sauvignon Blanc and so many other grapes that can be made to resemble it, that taste standardization does not seem a valid proposition anymore. Everywhere we are seeing consumers either not caring for the variety on the label (which is a great opportunity for branding) or eager to try any fuzzy-named variety just for the sake of discovery.

Regional typicity will have to be redefined…

Inversely, regional typicity will have to be redefined if climate variability is so strong that you can no longer really talk about a particular place’s typical sensory character because there will be so many years when you can’t get it. This is when blends may really come in handy. And maybe not just blends of different varieties and places, but also blends of different years, as happens in Port and Champagne. After all, these two regions have always been climatically challenged, and blending was a sensible way of getting style consistency.

…Blends may really come in handy.

Appellations will have to evolve as well to accommodate these changes and provide a future for those businesses who base their value, fully or partially on the appellation. This is something that is becoming painfully evident these days in Bordeaux and several other well-known wine regions. It is surprising that so many geographical indications (GI) recognize the direct influence of climate in the taste of their products yet have no regulatory provisions to adapt to a changing climate.

I am not just talking about wine GIs. The same goes for cheese, meat products, olive oil and many more. The insidious nature of climate change in our lives is simply not triggering defensive behavior until it is too late. This is something that needs to change, and, at Sogrape, we are doing our bit to bring the GI and climate issue onto the table. Stay tuned…


Meininger’s: Apart from the MED-GOLD project, you have done a lot of research into the native yeasts to be found in specific vineyards. Nowadays, there is a growing belief that wines have to be made with native yeasts if they are to reflect their terroir. How true do you believe this to be?

Graça: In my opinion, from my experience and the science I know, not a lot. For decades all wine regions were making wines using a few selected strains of commercial yeasts and no one questioned if there was terroir on wines…at least not because of the yeast. On the other hand, in the incredible amount of micro-biodiversity studies that erupted in the last 5 years, we are seeing a lot of information on the yeasts present, but very little knowledge about what they do.

There are patterns of ecological distribution that seem to be associated with the type of vineyard management, not necessarily good or bad, just different. Other studies have shown clearly that species and strains present in the vineyard can vary widely from one year to the next, seeming to reflect again a climatic effect, which makes sense. All this seems to suggest that there is no causality between a specific set of species and strains and the taste of wine from a certain place in every year.

I think that there may be some effect in the wine if micro-biodiversity is purposefully managed to change taste, but I do not see a lot of future in it, as we need balanced ecosystems for sustainable vineyards and wine regions, not tailor-made ones (which would be unbalanced) to suit taste preference or innovation.

In any case, I cannot see how the random and constantly varying spatial and temporal distribution of yeast species and strains in soils, grapevines and grapes can consistently become part of a fixed sensory profile in the final wine. This seems contradictory. But I am aware that we still know very little about the effects of biodiversity on the sensory aspects of wine. Additionally, we never had a moment in history when we had the technology we have today, applied to understanding how vineyards inserted in balanced ecosystems perform in terms of wine taste. There may be surprises ahead.

Gene editing (Photo: ipopba/Adobe Stock)
Gene editing (Photo: ipopba/Adobe Stock)

Meininger’s: Another area of huge controversy is GM both in vines and yeasts, and the younger technology of CRISPR gene editing. Can you describe the difference between them and the pros and cons you think they have for the wine industry?

Graça: CRISPR is a GM technique that consists in editing the genome of an organism as if one was using a text editor. You can delete, cut and paste single genes or sections of the DNA strand. In this way you can silence or boost different functions of the organism.

By changing the place of some sequences in the DNA strand, you can create characteristics and functions the natural organism did not have. It is different from transgenics or cisgenics in that you do not insert DNA from another species or variety. But it is potentially possible to copy, cut and paste sequences within one organism to mimic genes typical of another variety or species.

At the end of the day, according to current EU legislation, the result is always a GMO, a genetically modified organism. And GMOs are not forbidden, they are just required to be labelled as such, and so are all food products coming from GMOs.

Often, this discussion is, for me, a bit bizarre because we see people protesting a supposedly existing ban of GMOs in Europe, which simply does not exist. However, while defending the use of GMOs they reject the disclosure of its use in labels and, by doing so, they fuel the very same fear that makes the general public reject GMO-labelled products. At Sogrape we believe and practice consumer transparency, we do not agree with misleading or misinforming consumers. A CRISPR-modified organism is a GMO. If you use it, you should come clean about it.

It is my opinion that we do not have yet enough information to be able to make an educated judgement about the possibilities and threats that CRISPR and similar techniques, designated as new genomic techniques, could have in the wine sector.

A lot more research is needed, and it is unfortunate that Europe has been quiet in this area for so long while others are taking the lead. For me, the main concerns are the fact that once someone changes an organism with CRISPR, if they choose not to register it, no one can discover if the organism was changed artificially, or if the change resulted from a natural mutation.

This is a critical issue because it hampers accountability on the part of the CRISPR user. And accountability is important here because new genomic techniques have the power to produce new organisms at a speed far greater than it is possible to study and assess their benefits and consequences. This is not compatible with the precaution principle. That’s why more research is needed so that we understand exactly what it can and cannot do to all aspects of grape growing and winemaking and the surrounding ecological and human systems while also shedding some light on how to move forward. Before this is done, I believe all discussion of its use in grapes or wine is just a risky speculation.

Developing sensors and algorithms that can come very close to the way humans feel aromas, tastes and flavors.

Meininger’s: Finally, looking forward, as a long-standing panel president at the Mundus Vini competition, how do you see the future of wine tasting and assessment? Do you believe that AI and quantum computing will make it possible to replace the human palate when it comes to judging typicity and quality and identifying faults such as oxidation, Brettanomyces, and TCA?

I have no doubt that we already have electronic tools capable of detecting faults in wine, that is, when they are at such a high level that there is no more discussion it is a fault and not an added layer of complexity!

I have been studying recent developments in artificial intelligence and I am quite convinced that indeed it will be possible to develop sensors and algorithms that can come very close to the way humans feel aromas, tastes and flavors.

However, the experience of tasting is a subjective one, even when we try hard to objectively score wines in the best and fairest way possible. Emotion is a subjective experience and requires consciousness. This is where it gets tricky for, as far as I know, no one has been able to explain, in a widely consensual way, what consciousness is, or where it is in our brain.

The best attempt I saw was a proposal that consciousness could be an emerging property. An emerging property is something that exists only when a large enough number of similar elements are bound together and working in sync. A good example is wetness. We all know water in liquid state is wet. But a single molecule of water is not, even at 20 °C, and neither are two or three molecules. It takes millions and millions of water molecules bound together at the atomic level until wetness appears.

In the same way, consciousness could be an emerging property of neurons, appearing when a sufficiently large number of neurons are bound together and working rhythmically. No one has been able to demonstrate it yet, so until that happens, it will not be possible to build a conscious machine, that would thus be able to have subjective experiences such as being able to judge a wine like a human does. Yet, we already have been able to make artificial neurons and get them to work together and even to communicate electrical impulses, so discovering what consciousness is lies at the border of us becoming able to make a tasting machine. And at the border of vastly more complicated ethical issues…

The results of MED-GOLD will be officially launched on March 29/30 at an online Showcase Event. For free tickets go to MED-GOLD.EU



Latest Articles