By: Thomas J. Payette, Winemaking Consultant
Representative Sampling
While in the lab, and tasting wine at blending session etc, the winemaking team spends considerable time on getting procedures and processes correct to run each particular test on a sample of wine or juice. Yet, equally as important, if not more, are we obtaining data and flavors that represents that particular batch or tank of wine? Has the sample, in any way, been compromised so that it does not reflect the tank or lot of wine intended? One would not want to take winemaking action on a wine only later to find out the action was not necessary.
How often does a winemaker go into the cellar, select a sample of wine and then run a particular test, tasting analysis or blending trial on the sample drawn? Are these test(s) run with confidence? Do these numbers reflect the true tank’s contents? Do we need to sample each individual 60 gallon barrel and spend countless hours in the laboratory? Not necessarily unless the winemaker suspects a certain flaw to be identified with a particular vessel. If that is the case, it is best to “quarantine” that particular lot of wine until the proper blending time making sure not to infect other batches.
The author wants to emphasize for the purpose of this article the lab test results are correct and the lab technicians are not at fault. The numbers reflect the sample given to the lab but the sample is not representative of the complete batch of wine in that tank. Take the opportunity to resample a wine that shows suspicious numbers in the lab. Remain open-minded and always quiz yourself to the possibilities of improper sampling.
Below are some pointers to help with the scope of sampling properly
Tank Sampling
Sample Valve: Perhaps one of the easiest situations to monitor a large quantity of wine, yet, this process must be taken seriously with particular attention to the contents.
If taking a sample from a stainless tank understand where the sample is coming from in reference to the contents of the tank. If a cloudy sample is taken from a sample valve near the bottom of the tank, understand it may not be cloudy throughout the whole tank and most likely very clear at the top of the tank. Was the sample valve cleaned after it’s previous use? Was the valve flushed of its spoiled contents to bleed off any high VA or bacterially loaded wine, prior to sampling, that may have formed in the unsealed body of that valve? This flushing action of the sample valve, due to the positive pressure in the vessel, has little risk for cross-contamination and it is recommended in order to obtain a representative sample of the tank’s contents.
From the top lid or manhole: Perhaps one of the best ways to sample in well managed cellars if all the tanks are kept topped up and a catwalk exist to each man way. This gives the winemaker a chance to visually inspect the wine tank contents while taking the sample. Caution should be expressed not to sample the surface of the wine but to get the collection flask well under the surface of the wine to collect a representative sample. The surface of the wine may have a lower SO2 reading and false numbers other than these that do not reflect the majority of the wine. Please keep in mind this sampling choice could be a large source of cross-contamination if not done properly because certain items may contact each wine as samples are collected.
Ball Valves: Known to be the largest offender of panic and false a test results from the lab. Ball valves often have high spoilage counts, if not cleaned properly, lending toward off values most particularly with Volatile Acidity and Sulfur Dioxide just to name two. Other tests may give false reports in terms of the tank’s actual contents. One should flush ball valves diligently when taking samples and one should be able to clean these valves well while not in use.
Butterfly Valves: These offer an excellent source to sample a tank’s contents if a sample valve is not present. Care should be taken to flush these valves, too. This flushing is more to remove solids and less to remove any potentially spoiled wine. The butterfly valve often will collect solids in them and deliver an unrepresentative sample unless flushed prior to sampling.
Barrel Sampling
This is often the easiest yet most winemakers try and dodge this exercise making large decisions after tasting one barrel in a particular batch. One needs to isolate the vessels of a certain lot of juice or wine. With proper record keeping and a logical marking of each container this process is not too bad. One may sample an equal portion from each vessel of a particular lot (recommended for blending session exercises) or fifty percent of the vessels for that lot. Many times the winemaker can hedge this knowing a particular test or cellar action will be run in the future. One could use the fifty percent rule, or even less, if only addressing minor actions. Knowing the wine will be racked from barrel in 6 months may lend toward running further tests, double checking current data and making larger corrections at that time if needed.
Stratification: If taking a sample from the very top of a tank – does the sample represent the complete wine tanks contents? If taking a sample from near the bottom – does this represent the wine tanks content? Should mixing be used to make the tank contents uniform?
Mixing: It may not always be prudent to mix a tank of wine for sampling purposes. Much of the lees and solids may have settled to the bottom of the tank and mixing the tank would only re-suspend those solids. Certain times the winemaker may want to mix the tank prior to sampling may include: Prior to bottling, just after a racking or blending, just after and during additions and anytime a true representative sample is known to be needed for a particular winemaking decision.
Blending Sessions: Getting prepared for a blending session is a time to make sure your samples are very representative and broken down into areas of distinction – perhaps even inside various “same” lots. For example: Mountain Fruit Cabernet New French Oak, Mountain fruit Cabernet Old French Oak, Mountain Fruit Cabernet New American Oak, and so on.
All of these samples in the example may be from the same raw material but the cooperage influence incorporated into them has made them very different. These differences make blending sessions a joyful challenge and yet offers the best chance to have the flexibility and control needed during a blending session. This sampling will give the blending session the greatest flexibility and control to the outcome of the final blend. After wine samples are taken from each vessel – make sure to mix the sample so the actual sample taken will be representative of the complete number of vessels sampled and that incomplete mixing will not adversely affect the blending session’s outcome in the cellar.
Fining trials: This is an important time to have a representative sample. Mixing a reasonably clean wine, free of sediment, is desired to make sure this important refinement tool is employed properly prior to fining the tank’s complete contents.
Sampling collection beakers, vessels and containers: Make sure to take samples in clean containers free from any debris or residues. An example may be the adverse reaction to a sample taken in a beaker that was recently used during a Sulfur Dioxide addition or used to dissolve meta-bisulfite. If residual sulfur dioxide remains in that container, it may adversely affect the lab test results and needed additions may be overlooked. The lab test result will show ample quantities, when in fact, the actual tank contents sample may have indicated otherwise.
Temperature Measurement and Stratification:
Outside of the sampling topic, keep in mind when looking at a thermostat on a tank, where the thermocouple is inside that tank. If the reading is from a lower area in the tank, it may not be given a representative reading of that tank’s true temperature. This is especially important when cold stabilizing wines. Mix the tank prior to seeding (if seeding is the practice used for cold stabilizing) the wine. With large capacity tanks, one may notice during mixing the temperature may rise. Another temperature stratification check, prior to mixing, is taking a temperature reading from the top of the tank’s contents. Notice any difference? One will often see a difference in warm cellars with tall and large capacity tanks.
Summary / Miscellaneous
Representative sampling applies well beyond the wine cellar. This principal has huge applications in the vineyard when sampling the raw fruit to determine when to pick a certain variety or block. This concept is often reflected in grape berries sampling – a potential article in itself not to be dealt with in this article.
With above knowledge, keep in mind how the wines were sampled and how important that sampling technique may be when a particular decision is being made. When in doubt – resample and re-run the test in question. The winemaker is encouraged to make sure to think about the sample he or she has and to think what is actually inside the tank. Keep a keen sense of when tasting or when chemical data from a sample does not “measure up” to what is expected. Be sure to investigate all angles before proceeding with drastic processes toward a tank of wine. It may just be the sample!
The data collected, whether blending, tasting or chemical analysis in the lab, can only be as effective as the sampling. The samples content should be directly related to the tank and it should represent as closely as possible the contents. Always keep in the winemakers’ mind how a sample represents the tank contents while tasting, testing and blending.
Know your wine or juice sample and what it represents!
