Practical Winery
58-D Paul Drive, San Rafael, CA 94903-2054
phone:415/479-5819 · fax:415/492-9325

January/February 2002

By Christian Butzke
Associate Specialist in Cooperative Education, UC Davis

Imagine a world where all winemakers can sleep soundly at night and the winery lab staff feels appreciated. Also where consumers can rest assured that all wines are stable and healthy and show the finest qualities of fruit and style. Then wake up, smell the barrels, and realize that only a small number of winery labs in and outside California have lab quality systems in place that would allow for such peace of mind.

That is why the Winery Laboratory Proficiency Testing Guidance Subcommittee of the American Society of Enology & Viticulture (ASEV) established a continuous interlaboratory proficiency testing program for the American wine industry in late 1999. The subcommittee’s work, under the guidance of ASEV’s Technical Projects Committee, represents the first practical cornerstone of a lab quality system for American wineries.

The concept of proficiency testing was outlined by Sue Weeks in her Winery Laboratory Quality column, “Proficiency testing — Support groups for laboratories,” (PWV May/June 2001). The ASEV program is guided by the subcommittee and managed by Chris Czyryca of Collaborative Testing Services, who prepares statistical summary reports for the participants.

It is important to point out that participation in a proficiency testing program alone does not automatically lead to better lab performance. Below, we review the reasoning behind such a program, discuss the current state of proficiency in winery labs, and contemplate where we might go from here.

From a vintner’s viewpoint, interpretation of the test results is up to each of us, as we aim to establish performance criteria for basic wine analyses and, consequently, our winery labs.

Why should your winery participate?
Whether your winery bottles millions of cases and has a team of enologists and lab technicians available, or you grow the grapes, make the wine, and measure volatile acidity personally to make 100 cases, the criteria for sound winemaking practices are essentially the same.

In addition, regulatory requirements and liabilities for selling your wine will become more and more uniform for all marketplaces, whether you export your wine to the United Kingdom or sell it out of your own tasting room.

Adopting ISO standards
Euro-conspiracy, government plot, or a quality issue?
Now that the U.S. delegation has pulled out of the Old World-biased OIV (Office International de la Vigne et du Vin), it is even more important that our wine industry closely observes international competition and stays ahead of the curve when it comes to ensuring the quality and integrity of American wines.

In a time of increasing wine imports, stagnant wine exports, flat overall consumption, and a still looming worldwide economic recession, it can be a costly mistake to ignore trends that bring global competitiveness to a local level.

Quality systems, ISO accreditation, performance criteria — all this lingo sounds like some sort of paper-pushing regulator’s idea of fun, generating more paperwork and less time available to get the actual job done.

But in truth, these terms stand for processes that should lead you and your entire winery staff to take a step back and remember the most important task of your winery: making wine of the highest quality. That is where the international standards for quality management of analytical laboratories (ISO 17025) come in handy, as they outline a ready-made lab quality system that all wineries can use to their advantage.

To make things even easier for you, ASEV, through its Technical Projects Committee, has produced a Winery Quality Systems CD-ROM that is available to all members at cost. The multimedia CD highlights the principles of lab quality management, including practical winery examples.

The CD can help staff at wineries of all sizes understand and implement an internationally accepted quality system appropriate for their own environment. The goal may range from implementing the first basic steps to full ISO accreditation. The latter is already required in some export markets and is being quickly adopted and implemented by all our relevant government agencies, including BATF, FDA, USDA, U.S. Customs Service, and others.

Five basic steps to a Quality System
Following are the essential steps required to initiate a lab quality system at your winery:

  1. Preparation of a Quality Manual that details the agreement between senior winery management, winemakers, and the laboratory regarding the quality and performance goals of the laboratory and the resources (equipment, staff, etc.) required to meet those goals.
  2. Method validation constitutes another important first step. Often, a winery laboratory inherits methods or adopts them because they exist in a publication. ISO emphasizes that a method is valid only after its performance has been demonstrated to be adequate for your specific winery laboratory, utilizing your staff and equipment. Furthermore, fully validated methods make it possible to educate winemakers about the performance limitations of individual methods of analysis. See Sue Weeks’ column “Method validation — Making sure it works in your laboratory,” (PWV September/October 2000:76-77) for details.
  3. Documentation of lab procedures must be done in a consisten format. This is necessary for uniform implementation of individual methods of analysis, and is essential to the generation of quality data.
  4. Training for new and existing staff is available through courses at independent labs, community colleges, or UC Davis extension classes. You can even do it within your winery using your documented set of lab procedures.
  5. Proficiency testing is the last piece of the quality puzzle. Internal proficiency testing against standards and control samples verifies ongoing caliber of performance, while external proficiency testing, such as ASEV’s program, measures the performance of the winery laboratory against industry peers.
No worries, mates!
Advances in the Australian wine industry over the past 20 years were not due to luck, but to enormous investments in winegrowing and winemaking technology, wine research, and extension programs. One of Australia’s successful techniques for optimizing wine quality while improving cost and production efficiency was to establish the Interwinery Analysis Group and lab proficiency testing scheme in 1983.

Almost 20 years behind its closest global competitor, the American wine industry is only now on the right track to make up for this competitive disadvantage. However, if the speed of California’s advances in wine quality are any indication, it shouldn’t take long for our wineries to catch up.

Preharvest Wine and Grape Quality Air and Road Show
One of the Proficiency Testing Guidance Committee’s methods to promote the concept of winery lab quality has been an annual statewide road show, based on the Australian model and introduced by the Australian Wine Research Institute and other extension groups.

In the first Preharvest Wine and Grape Quality Air and Road Show, Gordon Burns of ETS Laboratories, Patricia Howe (teaching lab manager at UCD Dept of Viticulture & Enology), Sue Weeks of Bibber International, and I traveled to the North Coast, the Sierra Foothills, and the Central Coast in August 2001.

The program provided an effective platform for conveying the general concepts of a lab quality management system and included small group discussions of particular analysis problems. In addition, the committee organized Winery Interlab Group meetings to present and discuss test results during the Unified Symposium and the 2001 ASEV Annual Meeting.

The Proficiency Testing Guidance Committee plans to continue the road shows just before the 2002 crush, and PWV readers are encouraged to suggest class locations that are convenient to reach for most winery lab staff in their counties.

Current state of affairs

So far, there have been six rounds of testing in the ASEV Laboratory Proficiency Testing Program for nine standard wine analyses. Proficiency was tested for analysis of ethanol, total and free sulfur dioxide, titratable acidity, volatile acidity, specific gravity, pH, residual sugar, and malic acid.

Wineries were free to choose their methods of analysis, and the participants remain anonymous. Between 32 and 55 wineries have participated in each round, representing an estimated 90+% of U.S. wine production capacity.

Test rounds were conducted in February, May, August, and November of each year to minimize interference with lab obligations during crush. For each round, one 750 ml bottle each of two different commercial wines of similar composition were distributed to participating wineries. Each wine came from one bottling run at a large winery to ensure the most consistent sample composition.

Since real wines were tested, the accurate, true values for each measurement are known only to Mother Nature. Any comparison of measurements is based on the averages of the winery participants as a group of peers.

With respect to individual analyses (Figure I), alcohol, specific gravity, and titratable acidity measurements appeared to be precise and were well-correlated among participating winery laboratories. The increased availability of autotitrators may be one reason for improved proficiency in TA.

The coefficients of variation for both sulfite measurements, especially free SO2, pH, volatile acidity, residual sugar, and malic acid were undesirably high. However, for some of the analyses, results were biased toward the most popular methods of analysis — such as Ripper for total SO2, Cash still for VA, and enzymatic assays for RS and malic acid — all of which contributed proportionally to the grand mean.

While some results may be suitable for internal production control if reproducible, the range of results even without considering outliers was extensive. This is a concern, as the program participants represent the most advanced labs in the industry, and those that are most interested in improving their performance.

The reasons for the unsatisfactory results need to be investigated further, and all possible sources of error for each method of analysis must be determined, documented, and subsequently minimized.

Below are some statistics and comments on individual measurements; methods are listed in order of reported user frequency. The coefficient of variation (CV) is the error for each analysis relative to the mean of the whole group of winery labs.
Ethanol (EtOH). Methods: 51% Ebulliometer, 31% gas chromatography, 8% near-infrared spectroscopy, 10% distillation/density. Proficiency: Labs were most proficient in alcohol analysis compared to the other analyses. They achieved CVs between 1.2% and 1.8%, despite use of a wide range of equipment and analysis methods.

The application of ebulliometers, which were used by three out of four wineries in 1998, declined in favor of automated assays. Results confirm that non-standardized methodology, if validated, does not compromise the ability to produce acceptable results.

However, the industry trend to make delicate alcohol adjustments to wines for sensory or tax reasons would require an even greater proficiency in order to make such decisions more precise and reproducible.

Total SO2 (TSO2). Methods: 68% Ripper, 24% Aeration-Oxidation, 3% enzymatic assay. Proficiency: CVs ranged from 5% to 16% between test rounds. The ratios between concentrations of bound (total minus free) and free SO2 show a reasonable to generous use of sulfites (52 to 106 mg/L) throughout the winemaking process, with insufficient amounts of free and consequently molecular SO2 after bottling in three out of four commercial samples tested.

Free SO2 (FSO2) Methods: 48% Ripper, 48% Aeration-Oxidation, 4% colorimetric segmented flow analysis. Proficiency: This possibly most important wine analysis regarding microbial stability and general ageability was performed with CVs between 12% and 29%.

The amounts of free SO2 in the tested wines were between 10 mg/L and 26 mg/L independent of pH. The consequences of additional errors in pH analysis are discussed below. Molecular SO2 based on wine pH ranged from 0.14mg/L to 0.92 mg/L, suggesting a need for improvements in proper SO2 adjustments.

Titratable acidity (TA) Methods: 54% manual titration, 46% automatic titration. Proficiency: The essential analysis for a wine’s perceived tartness improved significantly over the preliminary survey (8%) with CVs between 2% and 4%. This coincided with increased use of autotitrators over manual titration (13%:87% in 1998 to 46%:54% in 2000).

Volatile acidity (VA) Methods: 81% Cash still, 10% enzymatic assay, 3% capillary electrophoresis, 3% gas chromatography, 3% colorimetric segmented flow analysis. Proficiency: The concentration of acetic acid and other volatile acids in wine is a federally regulated wine quality criterion (limits: 0.14 g/100 ml for red, 0.12 g/100 ml for white table wines).

Concentrations of VA in the wines tested were between 0.02 and 0.08 and thus far below those legal limits, implying good winemaking practices. However, the CVs ranged from 7% to 25%, which would make it difficult to confirm compliance in some instances.

Specific gravity (SG) Methods: 59% hydrometer, 35% density meter, 6% pycnometer. Proficiency: CVs were between 1.3% and 4% even if one compares the standard deviation between labs to the maximum range in specific gravity that can be expected in table wine. This assumes a range between 0.970 (14% v/v ethanol, no RS) to 1.015 (9% v/v ethanol, 50 g/L RS) kg/L (20C / 20C).

pH Methods: All labs used electronic pH meters. Proficiency: Labs achieved CVs between 6% and 9%, given the logarithmic nature of the pH measurement. This consideration increases the relative errors but recognizes the critical consequences of small inaccuracies in pH on the biological and chemical stability of the wine, especially the amounts of molecular SO2 available.

The molecular SO2 concentration for a given free SO2 content at a particular pH [formula: molecular SO2 = free SO2 / (1+10(pH-1.83))] determines the microbial stability of a wine. The suggested values have been between 0.6 to 0.8 mg/L (at 14%v/v and 12 %v/v ethanol, respectively) for protection against lactic acid bacteria.

Concentrations below 0.6 mg/L may lead to a potential instability in case of a failure or absence of a sterile filtration step at bottling, especially in wines that have not undergone a complete malolactic fermentation. An 0.1 underestimate of the pH represents a 26% overestimate of actual hydrogen ion concentration, which results in a proportionally higher requirement for free SO2.

Under commercial winemaking conditions, errors in the analysis of pH and free SO2 add up because of the impact of the hydrogen ion concentration on the equilibrium of the different ionic forms and the molecular form of sulfur dioxide in an aqueous solution such as wine.

The combined CVs for both analyses were between 19% and 39% for all labs and test rounds. If a wine is considered microbially stable because it is dry and has completed ML, one can be less concerned about the levels of molecular SO2. However, given the poor performance for both RS and malic acid analyses (see below), no winemaker should rely on those results alone.

Residual sugar (RS)
Methods: 69% enzymatic/spectrophotometric assay, 26% copper reduction assays (Gold Coast, Lane-Eynon, etc.), 5% HPLC. Proficiency: The analysis of residual fermentable sugar is crucial to determine microbial stability of a table wine and to adjust its sensory properties for the consumer.

The CVs improved significantly over the preliminary survey (140%), ranging from 5% to 39%. Given the dual importance of the analysis, this still appears to be unacceptable, even considering that some labs from wineries that produce only dry table wines were not accustomed to analyzing wines with relatively high amounts of RS.

The slight improvements may be attributed to an increased use of the enzymatic assay over copper reduction assays (51%:46% to 69%:26%) over the past three years. More important, wineries overcame problems reporting the actual sugars measured and the magnitude of their RS results.

L-Malic Acid (MA) Methods: 97% enzymatic assay, 3% capillary electrophoresis. Proficiency: Similar to the enzymatic determinations for RS, the CVs of the assay for malic acid varied in a wide range of 4% to 39%. Given the corresponding consequences of uncertainty over incomplete or deliberately prevented malolactic fermentations, in particular, CO2 evolution in the bottle, and changes in perceived acidity, this performance requires major improvements.

The whole picture

Lab proficiency tests in 2000 and 2001 showed similar levels of proficiency throughout the six test rounds. This was not a surprise, since a proficiency testing program alone is not sufficient to improve a laboratory’s analytical proficiency.

As mentioned above, a complete lab quality management system is required to achieve this goal. However, it is positive to note that labs have become more proficient in reporting the units of measurement correctly. Statistical results of the proficiency testing program so far suggest a continuing need for improved laboratory quality management systems.

With the proficiency testing program in place, the ASEV Technical Projects Committee will be able to address remaining parts of the winery lab quality management system, and proficiency surveys will confirm anticipated improvements.

As a next step, wineries need to establish performance criteria for the different analyses of wine that reflect the significance of their accuracy and precision for the quality and stability of any wine.

Performance criteria?

So far, the proficiency testing program has judged lab performance solely by statistical comparisons. Why? Because uniform performance criteria for wine analyses have not yet been established. To make proficiency testing most meaningful, it should be based on wine quality and stability requirements that reflect the sensory, biological, and chemical properties of the wine components analyzed.

As stated in the beginning, when it comes to performance criteria, (winery) size doesn’t matter. Every winery lab should be confident in the data it creates, and every winemaker should be able to rely on those data to make production decisions without nagging doubts. We all need to agree on acceptable errors for each of our measurements to assure safe, high quality winemaking practices.

Your input is invited!