Doing it by the Numbers
Paul Taggart takes a look at a fancy, new piece of kit that could make wine judges obsolete.
The tasting team at WineNZ is a great bunch of guys – but when we take a lunch break at the EIT café they hoover up the food as if they haven’t eaten for a week.
So when I caught up with a boffin in Christchurch who has access to a remarkable new machine at AgResearch that performs a type of metabolomics where thousands of molecules are measured after undergoing vaporisation in an electric current, giving information about a wide range of characteristics based on molecular composition, I thought I could be on to something. Imagine if I could replace the guys with a computer.
The machine in question is a Waters REIMS (rapid evaporative ionisation mass spectrometer), that has been designed around real time analysis of metabolites in tissue samples.
It can gather information about a sample’s origin (provenance), genotype, ripeness, quality and vigour/health. And it can do it in seconds, with no messy spitting.
The remarkable piece of kit has all sorts of potential uses. It was developed to detect the difference between cancerous cells and healthy tissue during surgery, and in the food industry it has been successfully put through its paces determining the origin of meat and reliably establishing species of fish used in restaurants (detecting red cod being sold as blue cod, for example). But as far as I’m aware, this is the first time the British machine has been used in a bid to reduce the costs of a magazine’s wine tasting bill.
So, based on the entrants for several recent WineNZ tastings, we sent 20 Sauvignon Blanc, 30 Pinot Gris and 27 Pinot Noir samples to AgResearch in Lincoln, where they started pouring them in to the REIMS to get a detailed molecular fingerprint for each wine.
This is serious stuff and the work was carried out under the supervision of Dr Neil Pattinson, who has had chief executive and board positions with a number of medical and scientific startups in Christchurch and Wellington, as well as having a small Canterbury winery. Dr Alastair Ross and Ines Weissenbacher from AgResearch ran the instrument and did the expert analysis. Alongside WineNZ’s desire to ‘value engineer’ its modus operandi, there were more serious experiments, involving a group of non-professional wine tasters, to determine which specific elements in wine pull the pleasure triggers with consumers.
My personal interest was around pinot noir regionality – which cannot reliably be determined by wine tasters, however skilled or experienced they are. If you ever read a review twittering on about a Central Otago wine’s distinctive characteristics, the reviewer more than likely knew what the wine was before tasting it, so they had a prompt to talk about schist and merino wool influence, or a hint of soot from the Earnslaw’s funnel. It is utter BS. I’m not saying Central Otago pinot isn’t great – much of it is spectacular, but so is the best of Marlborough, Nelson, North Canterbury and Wairarapa, and they aren’t really different in style.
However, this is where the REIMS machine hits its straps. The machine can tell Central Otago from North Canterbury, with Marlborough identified as sitting somewhere between those two extremes. However, it is not a strong distinction, as even between Central Otago and North Canterbury only minor differences were found (only approx. 20 compounds out of 3000 were different between them).
With pinot gris, regional differences were stronger. For example, Central Otago and North Canterbury were clearly differentiated with over 60 molecules found to be different. Central Otago was also separated from Nelson and Marlborough, with the latter two regions virtually indistinguishable. Interestingly, with sauvignon blanc, REIMS couldn’t pick up differences between Nelson, North Canterbury or Marlborough.
But does it have a commercial use? Potentially it could pick up mis-labelling or fraud, but that isn’t a serious issue in this country – and with drinkable pinot noir coming out of most of our regions, the risk of a Hawke’s Bay pinot noir being labelled as a product of Bannockburn is pretty unlikely. WineNZ tasting team member Barry Riwai has taken part in a couple of informal experiments during our tastings over the years, attempting to pick regionality in the wines he has been judging. While he had no great success with pinot noir, he was a lot sharper with sauvignon blanc – certainly differentiating Marlborough wines from other New Zealand regions and even sucessfully picking out sub-regions within Marlborough.
It has to be said these REIMS experiments were on a small scale and a lot more work would be required to see where this might go, but the pinot noir results were interesting. However, I reckon the human judges still have their noses ahead as far as sauvignon blanc goes.
And I’m glad about that – wheeling a REIMS through the EIT campus for a coffee and sandwich wouldn’t be much fun, and I’m sure the conversation over lunch wouldn’t be as good as it currently is.
As far as a genuine commercial purpose for the machine – Dr Pattinson thinks that finger-printing flavour preference profiles of consumers could be the direction most worthy of exploration.
After initial scepticism that it might be a wee bit Big Brother being labelled, let us say, an F19 because I like Greywacke pinot noir, Clearview chardonnay and Black Barn sangiovese montepulciano, I came round to seeing the sense behind the idea.
For example, one of the pleasures of my job is to sample quite a few of the top wines after each of the WineNZ tastings. Most of the time these wines are spectacular. Occasionally though I’ll have a glass of a five-star winner – often pinot gris or riesling – and it doesn’t ring my bell at all. The wines are technically perfect and varietal, but don’t suit me. This is probably one of the main reasons riesling doesn’t have the following it deserves, because it has a range of styles, and sweetness, so it can be a lottery picking a bottle. But if I had been run through the REIMS programme and had my taste profile established – let us stick with me being an F19 for a few minutes – then I could pick up a bottle of Clearview Three Rows and confidently expect to see F19 as the “Taste Spectrum figure” on the back label.
More importantly if there were two rieslings on a restaurant menu and one was a C12 and the other was an F17, I’d know the second one was pretty close to my marker.
It might seem a little far fetched, but the work currently being done in Canterbury could see it become a reality.
Some might say it would take the mystery and romance out of wine. However, it could also prevent people spending $40 on a bottle they don’t like.
Only the most boring of souls would fastidiously stick to their number. When they felt like living on the edge, an F19 could ramp it up and try an F21, or even kick over the tracers and drop down in to the Es.
At this stage the boffins are still doing the groundwork, and who knows where this might go.
Anyone who wants their own REIMS machine – they’re north of $800,000, plus you need someone brainy to run it. I’m thinking I’ll stick with Simon, Barry, Matt,Ant and Dave.