In the water’s path
Erosion comes in two forms: the seen and the unseen.
Rill erosion is the most obvious form of erosion and typically occurs in heavier downpours of more than 30 mm (1.2 inches) over a 24 hour period. They begin in spots where soil aggregates are weakened, and will collapse with weight and friction of the water above it, forming the aqueduct-like channels into which the runoff will funnel. Rills often generate in flow zones, gathers in the depressions between rows. Here water can consolidate, growing in volume and velocity as moves with increasing rapidity down the hillside. With the water growing in mass and speed, larger and larger soil particles are pulled with it, releasing from both the bottom and sides of the rill, developing their typically U-shaped trough. As rills go unrepaired, they can grow substantially, that can be difficult to control, if measures are not already in place to prevent them.
Sheet erosion (aka surface erosion) is a precursor to, and happens simultaneously with, rill erosion. In this case, rainwater runoff moves in sheets across the surface of the vineyard, but between and through the vines in places where rills won’t, or have not yet, formed. Surface runoff has a less concentrated volume of water than the runoff that travels through rills, so it yields a lower speeds and less velocity. Because of this limited velocity, the water of surface runoff is capable of carrying particles with a lower suspension velocity than rills are capable. These may include sands, but unless the downpour was heavy, would primarily include clays and silts. In less intense storms (< 20mm) surface runoff can cause sheet erosion, but these actions are considered slightly erosive, typically transporting finer materials in weak aggregates. From year to year, soil loss to sheet erosion goes largely unnoticed as the topsoil loss directly beneath the vines disappears down the hillside forever.
2006 Study of erosion in Vosne-Romanée, Aloxe-Corton, and Monthélie
Two sibling studies, preformed by the same research team, illustrates very well the processes of erosion (detailed in part 4.2), and how it affects the wine we drink. These are multi-discipline studies by conducted by the team of Amélie Quiquerez, Jean-Pierre Garcia, and Christophe Petit from the Université de Bourgogne, and Jérôme Brenot from Géosciences Université de Rennes.
The first of the two studies was published in the Bolletino della Società Geologica Italiania 2006, with contributions by Philippe Davy, Université de Rennes. Entitled “Soil erosion rates in Burgundian vineyards (link).” It examined the erosion rates in the villages of Vosne-Romanée, Aloxe-Corton, and Monthélie. I highly encourage you to look at these important studies to get their analysis, which in some ways is limited by the rigors of science which require the researcher to prove what they already know to be true. My overview of the information revealed by their study applies my own perspective and insights.
The researchers selected three steep, upper-hillside vineyards from which to gather data, all which carried essentially the same average grade, with a mean of 10.5% for Vosne and Aloxe-Corton, with Monthelie the steepest, with a mean slope of 10.7%. Additional selection criteria were all three were they must meet these three (very traditional Burgundian) vineyard practices.
- The rows ran vertically down the hillside.
- None of the plots were allowed to have grass grow between the vines.
- Frequent plowing or tractor crossings (up to 15 times per year)
However, I note two marked differences between the vineyards.
- How much the slope changed within the plot boundaries.
- The length of the slope.
The study’s most uniform slope was a vineyard in Vosne, with a fairly consistent 10% to 12% grade. It also had, by far, the longest slope studied, at 130 meters.(1) This longer slope length, one might expect, would allow water to gain volume, speed, and velocity. These three factors all increase the runoff’s ability to carry larger and heavier particles with higher suspension velocities. Conversely, it was the only slope studied which had a murger (stone wall) at its base, slowing the runoff enough to allow sedimentation to occur, and it would appear to be the only plot with a level spot for sedimentation to rest.
Although unnamed by the study’s author, I have concluded this vineyard is Les Damaudes on the Nuit-St-Georges border. Clues to its identity include a maximum elevation of 345 meters – the highest in Vosne, and uniform slope of 10-12%. Identifying the parcel location is possible as well, as only one location in Les Damaudes is long enough to fit the 130-meter plot length of this study. When subtracting in the dirt roads at the top and bottom of the vineyard, which are natural erosional breaks, the total length is 126 meters. This vineyard was studied in-depth, over a multi-year period, and spawned the two studies I will detail in this article.
The vineyard in Aloxe-Corton may contain a significantly steeper section than the vineyard in Vosne, with a 17% grade, but overall the Aloxe-Corton vineyard had the same average gradient as the plot in Vosne, at 10.5%. This indicates that part of that vineyard had to contain no more than a 5% grade. Additionally, this vineyard was the shortest plot at 53 meters, meaning as long as fast-moving runoff could not enter the plot freely from above, runoff should not be able to attain the same velocity as it might in Vosne. Because of this, we might anticipate erosion lower erosional levels. There is no specific information that might allow us to identify this vineyard. And while the author Jérôme Brenot included a photo and a brief reference to the grand cru vineyard of en Charlemagne (regarding rill erosion down to the limestone bedrock), the lieu-dits of en Charlemagne is in neighboring Pernand–Vergelesses, not Aloxe-Corton.
The slope in the study with the steepest section, by far, was in Monthélie. The plot there reaches a maximum pitch of 24.5%, but the average gradient is only slightly greater at 10.7%, which again indicates much of the vineyard is gentile in its declivity. This vineyard, which would become a 1er cru shortly after the study was published, is the vineyard of Le Clou des Chênes,(2) and this parcel appears to share a border with Volnay’s ez Blanches vineyard. The study measured nearly twice the plot-wide erosion at 1.7 mm (± 0.5 mm year) as they did in either Vosne or Aloxe-Corton. However, in some locations within Le Clou des Chênes had far greater erosional levels: measuring as deep as 8.2 mm (± 0.5 mm) per year.
Notable is that the time under vine is much shorter, having been planted 32 years before the study. This makes the losses all the more alarming for these steeper slopes because the knowledge of how to resist erosion has improved so dramatically in the past twenty years.
Data collection and methodology
Determining these numbers involved a massive data collection effort, imputing vine measurements on a meter by meter scale. With 10,000 plants planted per hectare, this translates into thousands of data points are required to arrive at the final calculations.
Soil loss was determined by measuring the exposed main framework roots from the current soil level to the point of the graft cut. The graft is typically made 1 cm above the soil level at the time of planting, and with this measurement original soil level at the time of planting can be established (NEBOIT, 1983; GALET, 1993). By dividing this measurement by the number of years since planting, a relatively accurate average rate of erosion can be established. This method of using plants to give a historical record is called dendrogeomorphology, which is a geologic adaptation of dendrochronology, the study of trees and plants to determine the historical climatic record.
An unequal field of study
In the end, there was a single factor that differentiated these study vineyards: the road and the stone wall below the Les Damaudes vineyard in Vosne. Because of this road and wall, it also was the only vineyard that had an area at the base of the slope that was able to retain alluvial sediment. This proved to be an important last gasp defense regarding soil loss and allowed that sediment to be returned to the slope. With this material, workers could fill the rills in Vosne, that would grow into gullies down to the base rock in Aloxe-Corton and Monthélie.
The return of the sediment to fill the rills was preformed bi-annually in the Les Damaudes parcel. However, the owners of this vineyard were lucky rather than preventative. The wall was built as the headwall of the small clos that surrounds the vineyard below, and the access road that runs between the vineyards proved to provide the necessary flat collection area for the alluvium.
Inexplicably, the author chose to simply say that in Monthélie, the practice of returning soil to the hillside had never been done, whereas in Vosne it had been practiced every two years. Strictly speaking this was true. However when looking at satellite images of the vineyard, this statement appears somewhat disingenuous. In reality, the decision to plant the entire area Le Clou des Chênes in long rows without any roadways or other vineyard breaks, when coupled with the parcel’s physical position on the hill, created a highly erodible vineyard in which no level “toe of the slope upon which might sediment gather. Returning sediment, that doesn’t exist, to the hillside is simply not possible. That does not excuse the vineyard owner from not removing vines to build a walls or taking other erosion prevention measures, but it also gives and indirectly assigns blame for this lack vineyard maintenance. The Aloxe-Corton parcel (where ever it was) is not mentioned as the owners never having returned alluvial sediment to the hillside, although this was apparently the case.
In 2006, the researchers took the adjacent photograph of Le Clou des Chênes, showing that rills had developed into gullies due to the lack of effective intervention by the grower. They also included photo looking up towards the Bois de Corton (which I have not included), with a rill/gully that extends down to the raw limestone base rock below. In each photo, the vines roots can be clearly seen, having been exposed by the continuing erosion of these gullies.
Study design: did the study reveal unexpected results?
In some ways, the wall below the parcel in Vosne was problematic to the study. The stone wall, and ability the return of the sediment by the grower directly impacted the amount of erosion recorded. The study’s author reports this in the write-up as: “by a factor of two”. It not clear that the researchers anticipated this would be such a weighty factor when they formulated the study, since the focus of the study did not seem to take into account the effectiveness of wall in diminishing erosional forces. However the effect of the wall and the “anthropogenic factors” (meaning in these studies: the actions by man of returning the sediment to the hillside) certainly did have a dramatic effect on reducing the total soil lost, and the authors rightly took the opportunity to underscore the roll and value of murgers and clos as a primitive, but effective form of erosion control. (4)
But because of the wall (and the author’s eventual focus on it), other opportunities were lost. Since Les Damaudes in Vosne possessed the longest slope which also had the most consistent gradient, knowing how those factors affected erosion would have been instructional. Had the erosion measurements been made before the anthropogenic resupply of the sediment to the slope, this information might have been gained. But since the measurements were taken after the rills were filled, ascertaining the impact of degree of slope and the length of the run can not be readily determined if the Vosne parcel is included in the analysis.
Further analysis of meter by meter grid data, might answer some of these questions surrounding how much erosion is affected by increasing slope gradient and increasing slope length. Here the shorter Aloxe vineyard could have been compared to the top 53 meters of the steeper Monthélie vineyard. What were the erosional differences within these sections? What was the difference between erosion between the upper slopes and the lower slopes of the vineyards. Could these differences have been attributed to gradient or soil type? What were the soils left behind in the inter rows? Were they significantly different to the soils directly under the vines where the soil is more protected from rain strike and rill erosion? Then, if the full length of the Aloxe vineyard could be included, would there be greater erosion on the steeper sections where gravity has more effect? What about on the lower sections of the plot where increase water volume, speed and velocity might be expected to increase? It does not appear that these questions were asked by the study’s researchers in 2006.
It would be interesting if the data still exists and can be analyzed to examine those questions as well. It certainly would shed a more quantitative light on erosional forces on Burgundian hillside vineyards.
In the opinion of the study, while in the short-term, erosion didn’t affect the vines production as long as the root system was not exposed, over time, the overall surface soil level declined despite the best efforts in Vosne to return the alluvial sediment to the hillside. At the time of the study, the most alarmed of growers had begun been attempting to restrict erosion by allowing grasses to grow between rows, shortening the length of rows and rebuilding walls. The authors suggest these processes be applied to all hillside vineyards.
The study of a single rain event in Vosne-Romanée
The second study released by this team in 2007 is far more detailed, focusing solely on the Damaudes vineyard. Entitled, “Soil degradation caused by a high-intensity rainfall event“ (3) the paper details soil loss related to a single storm on June 11, 2004. This study is much more focused and is far more precise and instructional in its findings.
The study’s centers on the erosional path, volume, and sediment type, as well as the net erosion levels measured in the vineyard after workers had returned sediment to the hillside, post-storm.
Soil analysis of the plot
The soils native to the vineyard are within this description from the text of the study. The prose is tight and dense so I will quote the author, Emmanuel Chevigny, here.
“The texture is rather homogeneous over the whole plot and is composed of 40% of clays and silts, 50% of gravels (2 mm to 10 mm) and a low sand and boulder content. The topsoils are ploughed (Mériaux et al., 1981). The argillaceous aggregates with polyhedral blunted to grained form are slightly structured. No pedogenetic segregation has been observed.”
The soil, as described, is a marl, with what I would think has a surprisingly high clay content for being this high on the slope. A better breakdown of clay and silt would be informative, because (as detailed in Part 2.1 and 2.2 regarding soil formation), clay is metamorphosed from limestone and other materials, and very fine in size, while silt is larger (between 0.0039 to 0.0625 mm), and not metamorphosed. Silts are often parented from quartz, which unlike limestone is not prone to chemical alteration, and thus will not produce clay minerals. The origin of this silt must have been transported from farther up-slope, having arrived in Les Damaudes through erosion.
The vineyard’s soil has a low sand content.
The author then writes about argillaceous aggregates, which are clay aggregates. In this sentence, they are writing about the type of soil structure found in the vineyard. Clays tend to form into blocky structures, where each clay units sides is the same shape or a cast of the aggregate next to it. In other words, when the blocky structures form, they are literally cast so that they fit together like a puzzle. Here he is saying that the edges of these casts of the aggregates have been blunted making them more grain like. There is a soil type, classified as granular (grain-like), that is common to soils in grasslands with a high organic content, and Chevigny is clearly saying these are not granular soils.
Lastly, Chevigny notes that the researchers observed no pedogenetic segregation, meaning they could observe no identifiable soil creation nor the beginnings of soil horizons (sedimentary layering). This would lack of soil generation could be caused, in part, by plowing which disrupts soil horizons and encourages the erosion of weak young soils that have not developed into stronger aggregates. More on the concept of what soil is and pedogenesis later.
The gravel, or scree, which constitutes 45 percent of the vineyard’s soil makeup, (by definition) has slid into the vineyard by gravitational erosion, from higher on the hill. With the clearing of land and subsequent planting of the vines, this gravel has long ago been plowed into the clay-silt mixture. It is never mentioned by the study author, whether the scree is primarily limestone or not. Limestone is not a factor for these researchers, the particle size is squarely considered to be the issue.
By the numbers
While study revolves around the analysis of a tremendous amount of numerical data, to examine each piece of analysis is beyond the scope of this article, but their findings are none-the-less important and tells the story of erosion within a Burgundian vineyard very well. Below I’ve listed what I see as the most important changes to the hillside following this particularly heavy storm system:
- Both rill and sheet erosion occurred, but rill erosion accounted for approximately 70% of all soil lost from the hillside.
- A total of 13 rill erosion were noted, some forming a mere 30 meters from the upper plot boundary, that ran in straight lines down the slope, each time in the inter-rows.
- Rills occurred across 59% of the inter-row area
- The rills were U-shaped with strong vertical walls.
- Estimated soil loss from the rills alone was 4.77 meters
- A rill erosion for this rain event is estimated at 7.8 cubic meters (.275.5 cubic feet) and weighing roughly 6 metric tons (13,227 lbs)
- An estimated 1.6 meters erosional material was deposited into 7 alluvial fans at the base of the plot.
- The sedimentary fans consisted primarily of very fine sand to coarse sand that was between 63 μm (roughly the thickness of paper) to >2 mm. Only 10% of the fan sediment was silt clay fractions of less than 63 μm
- Fan #4 had a total sediment area one-half of a meter cubed (.5m3).
- The two rills that fed fan #4 had a total eroded area of .93m3 *
- If 10% of the rill volume is sand, then 70 percent of the fan debris came from the rills while a remaining 30% must have come from surface erosion which fed into the rills and were deposited into the fans.
- Topographic soil loss in inter-rows with rills was 3.9 mm, or 48 metric tons per hectare (105,800 lbs) even after anthropogenic resupply of fan sediment to the hillside.
- Mean (average) soil in non-rill effected vineyard area, was 1.4 mm, or 24 metric tons per hectare (52,900 lbs)
*1 cubic meter is equal to 1000 liters, or 6.29 oil barrels or 264 U.S. fluid gallons.
Storm size and frequency
Annual rainfall in the Côte de Nuits is between 700 and 900 mm (27 inches to 34.4 inches) per year writes Chevigny, citing the Météo France weather service’s Atlas climatique de la Côte d’Or 1994.* The study also cites that storms with rainfall of more than 30 mm per day, occurred 10 times between 1991 and 2002. Nine of these rain event dropped between 30 and 50 mm, (1.1 inches to 2 inches) and a single storm dropped 63 mm (2.5 inches) of rain water per event/day. Based on this, we might expect that there have been 50 such events between planting and the 2006 study.
The storm event of June 11, 2004, was uniquely powerful because 40 mm fell in a two-hour period, which caused causing 3 times the annual erosion rate established by the 2006 study of 1 mm per year. Perhaps most importantly, the erosion of this single event is averaged into that 54 year period. This indicates that some years little erosion occurred. Because the study only includes storm records from 1991-2002, we can’t estimate the distribution of erosion over the span of these 54 years.
With global warming, storm intensity seems to be on the upswing in Burgundy, just as scientists have noted in other parts of the world. The severe hail events of 2012, 2013 and 2014, which centered over the hapless villages Pommard and Volnay, resulted in total crop loss for some growers. In the Côte de Beaune, where precipitation and hail has recently been at its most extreme, has also been remarkably varied in its distribution. According to Jancis Robinson, in July of 2013, Volnay saw 57mm of rain (2.25 inches), while neighboring Monthelie only got 9.4mm. Needless to say, with this high degree of weather localization, these data figures are representative of the rainfall collection points only. There were likely areas of Monthelie that got much more rain, and areas Volnay that got much less rain than the data collection sites. The massive storms of late November 2014 that saw 200-300 mm of rainfall along the Mediterranean coastline and into Austria, the Dijon saw 95 mm of rainfall over a 24 hour period. So, in terms of storms, it would appear that while the Côte d’Or gets regular, low volume rain events, it is by and large, relatively protected from major storm fronts.
*Current monthly statistics are can be found here, and the average rainfall in Dijon as of 2015 is 775 mm (30 inches) per year.
The sediment at the “toe of the slope”
When examining sediment in the alluvial fans, researchers discovered that it was made up of 90% sand and 10% fine sediment. Fan number four, on which researchers focused their examination, contained nearly one meter of alluvial material. The fact that it contained little silt or clay, indicates that when the water became backed up at the stone wall, its movement did not slow enough for particles smaller than 63 μm (which includes all clays and silts) to fall out of suspension. This suggests there was a significant depth of water Then as the runoff began to gather enough volume to circumvent the murger, and continue downslope, it gained sufficient speed and velocity to quickly form rills in its path around the wall. The runoff carried virtually all particles smaller than fine sand out of the vineyard.
Study inconsistencies, and outdated or generic source material
Between the two articles, the explanation of soil and bedrock type differs. It is not clear why the authors of both studies would quote articles that are 35 to 45 years old, and that generic to the region rather than performing a shallow excavation themselves, in order to obtain information specific to that vineyard.
“The slopes are composed of Middle to Upper Jurassic limestones and marls (Mériaux et al, 1981) …“For example, the sandy-clayey screes (grèze litée) reach 3 meters on Comblanchien limestones in Vosne-Romanée.”
In the second study they write:
“The hillslopes develop on Middle to Upper Jurassic limestones and marls, and are covered by colluvium soils of argillaceous-gravelly nature and formed by Weichselian cryoclastic deposits (grèze litées) reaching up to 3 m thick (Journaux, 1976).”
Writing of Comblanchien as a class of limestones is a red flag, as it is distinctly a singular type of limestone. Adding to the confusion is the soil percentages that at first appear to be attributed to the vineyard, are actually from the 1981 Mériaux et al study and generic to the Côte d’Or. Later in the study, the percentage of sand is increased to >20% (from 10% sand and larger stones). 50% gravel content in the vineyard, which is cited in early in the text, is reduced to 45% later in the study write-up.
Computer modeling projects grain-size transition
Because the researchers must begin their work with the soil percentages they observe, this 45% gravel, 40% clay/silt and 15% sand, was their starting point. It was quickly recognized that outgo of clay minerals, coupled with the simultaneous retention of sand would eventually change the vineyard make-up, so they developed a computer program to predict future changes in grain size distribution of the soil composition. Computer models showed after only 4-5 rain events of similar magnitude as the one in 2004, there would be significant changes to the soil makeup. The results of those projections are to the right.
Chevigny encapsulates their findings with this statement.
“…the results of our simulation clearly show that repeated rainfall events modify significantly and very rapidly surface soil grain-size distribution: after only a few events, the top soil has lost more than 30% of its fine material.”
The ultimate effect of this would be the loss of organic materials, nutrients and ultimately soil sustainability.
Study conclusion: vineyard practices enhance rill development and erosion
While the wall slows the net output of soil volume from exiting the plot, the most soils most viable for farming are being lost, while simultaneously, the soil texture and particle size are being irrevocably changed as the sand sediment is returned to the hillside, and disked back into the soil.
It is forwarded by the author, that this action, is part of the problem since rills continue to re-emerge in the same locations, year after year. They submit that ill propagation in the inter-rows is heightened by tilling and repeated passes tractors and foot traffic, and the regularity of rill spacing are evidence of this. These practices, he writes causes decreased soil porosity (compaction) and restricts rainwater infiltration. Such wheeled ‘passage’ creates flow zones which increase the volume and velocity of runoff in a concentrated area, multiplying the quantity and size of material the runoff can carry. The evidence of these anthropologically created flow zones is the re-emergence of rills that return, repeatedly, in the same inter-rows, despite workers attempts to eliminate them by filling the rills and disking those areas.
It is clear the effort must be made to properly identify the flow zones and attempt to eliminate them but to do so is to understand their formation to begin with, and limit or eliminate that activity altogether.
For me, the results of the computer modeling and projections are not surprising. While this research team and Burgundian winemakers can only look forward to what is next, we have the opportunity to use this information to hypothesize what came before. This will allow us to see the true arc of geomorphological progression in the vineyards, and thus how winemaking styles have and will continue to change in Burgundy.
Next UP: Turning our understanding of the limestone Côte on its head
(1) Vineyards typically are areas with no breaks or obstacles to slow or impede storm runoff, so longer vineyards tend to suffer more greatly from erosion. However, this was not identified as an erosional factor in the study write-up. The length of this Vosne vineyard was listed in the first study at 130 meters, while in the second study it was written as 126 meters.
(2) Le Clou des Chênes’ increased prestige and vineyard value can be a tremendous incentive to better maintain a vineyard. The vines and vineyard appeared to be in good health in 2012, the last time googlemaps car drove up this stretch of road. Still, no murgers had been built as of that time.
(3) published by Emmanuel Chevigny of the Université de Bourgogne in 2007