An agricultural research center in southwestern Germany has assembled scientific affirmation of a nebulous tenet long championed by enophiles: Terroir exists, and it matters.

Preliminary research results were presented to the British wine trade and press during a seminar sponsored by the German Wine Institute in London on Nov. 7 last year. Speaker Andrea Bauer of the Service Center of Rural Areas (DLR)-Rheinpfalz in the city of Neustadt said, "Everybody talks about terroir, but we want to provide a scientific basis for these discussions. We're trying to determine which components of, for example, soil and climate influence the sensory characters of wine, and how they do so."

The results showed that wines made from sites with similar soils but located many kilometers apart had more in common than wines made from different soil types located closer together.

As a medium for this work, Bauer and research director Dr. Ulrich Fischer focused exclusively on Riesling. She explained that wines from this variety were appropriate for their studies because they are recognized as being particularly expressive of their origins--not least because their production traditionally avoids character-altering techniques such as oak-influenced fermentation or malolactic acid conversion.

The research began with the 2004 harvest in Germany, and it enlisted the assistance of a dozen growers across the Pfalz region--including leading estates such as Bassermann-Jordan, Rebholz and Christmann. From the 2005 vintage, the work expanded to include 13 more estates across the Mosel, Ahr, Nahe and Rheinhessen (shown above) regions.

Vineyard sites were selected for the program according to criteria including proximity to weather stations, a minimum plant age of 10 years, and also specific soil profiles--such as sandstone, basalt, limestone and slate. Regarding soil types, one goal of the work was to determine whether there were discernable similarities across wines from well-separated sites having comparable soils, as well as any consistent differences in wines from nearly adjacent sites with different soil compositions.

Controlling the winemakers

Another aspect of the research went even further to permit site-driven aspects of the wines to manifest themselves--by controlling for the influence of the winemakers. "We harvested about 100kg of grapes from each test site, alongside the participating estate owners," Bauer said. "But we vinified these batches ourselves under standardized conditions at the DLR's experimental cellar. So each wine was made in the same type of vessel and with the same yeast strain, before racking and bottling in an identical manner."

Not that the commercial winemakers were excluded from participation. Bauer said she also examined samples of wine produced from the test sites by the participating wineries themselves--who agreed to use the same yeast strain as the laboratory. "Analyzing this set of samples enables us to see what impact the style of vinification has on the character of the wines," she said. "Does the winemaker actually overpower the impact of terroir? That was a question in which we were interested."

At this point, though, Bauer said she can't make any general statements or highlight any consistent patterns in how the wines from the estates differed from those made by the DLR. But she did share some interesting observations on this front. "As one example," she said, "Two wines produced by different estates from vineyards whose primary soil component was sandstone had very similar structure in terms of acidity, and both seemed to show very fresh, lean aromas and very little in the way of ripe tropical fruit characters. So there appears to be something from the sites that is not overwritten by cellar technique."

Such a sensorial comparison points to the centerpiece of Bauer's research. "After bottling these Rieslings each March," she said, "we conducted qualitative analysis each June with expert panels trained to use sensory descriptors that we developed to highlight the differences between the various samples. Each of the wines was tasted in duplicate or triplicate."

Charting the results

This very clinical approach offered distinct results. Bauer said, "Diagrams that aggregate the results clearly showed that wines produced on slate were citrussy with markedly sharp acidity. Basalt showed richer fruit flavors and smoother acidity. Limestone showed more intense color and more tropical fruit and even honey aromas. Sandstone showed mineral and citrus characters on the nose, and pronounced acidity on the palate.

"Even when comparing wines from, for example, the very similar soils and steep slopes of the Birkweiler Kastanienbusch vineyard in the southern Pfalz and the Urziger Wurzgarten site in the Mosel--which are more than 200km apart--we saw only slight differences. The wine from the warmer south was bigger or more textured on the palate. But more interesting were the similarities. Sensory analysis showed a grapefruit and smoke profile in both, which seems to be evidence of the overt role of soil in wine character."

Bauer added that the influence of climate is another important part of this research; but she said that the data had not been adequately analyzed in time for the London seminar. In a follow-up interview conducted for this feature in Wines & Vines, Bauer said she is now examining various climatic factors to determine how they influence sensory characteristics in the wines.

"For example, I've looked at test site temperatures gathered on a daily basis across the entirety of each year," she said, "including a lot of different maximums and minimums. And that's because other research has been done that suggests aroma composition is influenced by the degree of daily temperature variation."

She also considered levels and frequencies of precipitation, as well as factors such as the phenological development of the vines at each test site. "I tried to keep tabs on the dates of bud burst and flowering and veraison," she said. "And in analyzing the weather data, we looked at these growth intervals. We considered, for example, the period from the beginning of floraison to the start of veraison against our climatic data, rather than purely in relation to the calendar. After all, the grapevine doesn't really care about the date."

Bauer also said that the research she presented was limited to the 2004 and 2005 vintages, since she had to draw a line under data collection in order to complete her Ph.D. in a reasonable timeframe. "I know it's a very narrow window of time," she said. "But we have compared our results with relevant weather station data since 1991. We want to be sure that our markers for 2004, for example, are typical for these test sites, and not influenced by some form of climatic extremes."

She further added that the project has continued with the 2006 and 2007 harvests. "And it's scheduled for future years, too," she said. "It doesn't make much sense to investigate site-specific sensory markers for just a few vintages. Terroir is something that shows itself through continuity. More vintages are needed to back up what we have learned.

"Though I do believe we've already made a significant contribution," she said. "Sensory analysis showed consistent differences across the wines produced from different vineyard site types, and also very clear similarities among those produced from similar site types. There were some variations, but they never overrode the essential profiles or characters or, in other words, the terroir of these places."

Gary Werner is a freelance wine industry journalist based in London, and a regular contributor to publications including Decanter, Harpers, The Drinks Business and Wine & Spirit. Contact him through edit@winesandvines.com.