Managing Cold Injured VinesWhat We Learned in 2003

Bob Pool and Steve Lerch
Department of Horticultural Sciences
NY State Agricultural Experiment Station
Cornell University
Geneva, NY 14456

Global warming or not, cold in the Finger Lakes can still be a problem. The late winter of 2003 produced considerable winter bud kil in the Finger Lakes of New Yorkl. Despite the reality of damage, there is some good news. One bit is that the worst injury seems to be at the Geneva Experiment Station Vineyards. That’s only good because we don’t have to sell grapes to survive, and we can take advantage of the opportunity to evaluate relative hardiness of new clones, the impact of rootstock and training system on survival, and so forth. Even at Geneva, the cold damage wasn’t universal; it was almost entirely limited to vinifera varieties and was largely confined to our highest farm.

When did it happen?

Average "official" daily minimum temperatures at Geneva, NY and those for the winter, 2002/2003.

I’ve only lived in the Finger Lakes for 33 years, but I’m still waiting for an average year. The 2002 growing season was noted for drought. Dr. Lakso reported the impact of water stress on photosynthesis in the Prejean Vineyard last September. Non-irrigated vines had less than half the photosynthetic rate of irrigated vines. I bring this up, because we first cut buds in the last week of December. At that time we found that almost all varieties were showing around 20% bud death. The injury was predominately at the basal nodes, and the buds looked very dry as if they had been dead for some time. I assumed that these buds probably had never become hardy and had been killed when temperatures first hit the teens or low 20’s. I also wondered if there was more than simple coincidence to the fact that those nodes where leaf pulling had been done were the ones most likely to be dead.
We cut buds several times during the winter and we tried to determine bud-killing temperature using differential thermal analysis (DTA). We found the DTA profiles unusually hard to interpret. This happens when there is a mixture of live and dead buds, and when cold hardiness of the live buds is less uniform than usual. Regardless, our bud cutting showed no additional bud damage up to mid-February.
Steve Lerch made the last freezing run on February 12 and got the results shown in table 1. In spite of the fact that the winter cold hardening conditions had been excellent, the measured killing temperatures were a little higher than typical for the date. We got the coldest temperature of the season a few days’ later (-4°), and subsequently cut buds. We were still only seeing values in the 20% dead range.
On February 26, the official low temperature was –3.7° degrees. However, thermometers in one of our vineyards had readings in the range of –11° to -13°. A few days later we cut buds and found 90% or more dead buds. Fortunately, our other vineyards had neither the colder temperatures nor the bud injury.

There are two kinds of low temperature events, air mass and radiation. This was a radiation event. This is significant because within site variation is much more important for a radiation freeze. In a radiation freeze, the night sky is very clear and there is almost no wind. Thermal radiation from the warm air at or near the ground radiates to the dark sky and is lost from the vineyard. Lowest temperatures are found at ground level, and temperature increases with elevation above ground level. Often there will be an inversion layer where a warm air layer is suspended over the colder, lower air.
Because cold air is denser than warm air, cold air flows downhill. So long as nothing interrupts this flow, the cold air will move out of the vineyard. Unfrozen lake water will be warmer than the air temperature. It will warm the cold air coming from the land and generate a convection current that reinforce the flow of cold air out of the vineyard. Keuka Lake was frozen for much of the winter, preventing this pumping action. As a result, some low elevation Keuka vineyards appeared to have more rather than less cold injury than higher elevation vineyards this year. This concept is covered in more detail at on another grape page.
What does this mean? It offers you an opportunity to evaluate the air drainage of your vineyard. Possibly you can divert cold air from entering your vineyard and enhance the flow of cold air out of your vineyard. One thing is to make sure that grass is mowed low. This not only enhances airflow, it actually lowers the effective floor height. You can see areas where you should avoid planting more cold tender varieties, and you can consider other steps to take.

In our vineyards, the relationship between proximity to the vineyard floor and low temperature is obvious. Most live buds on low head VSP vines are high in the vine near the upper catch wire. This is about the same height as our mid-wire cordon trained vines that had much greater bud survival. In the picture above, one can see that the live buds on the VSP trained vine row which is lower than the row of Cordon trained vines in the background. There is much greater bud survival on these higher located and higher trained vines.

Another fact about radiation events is that the grower does have some tools to fight the cold. Wind machines and helicopters are often used to combat spring freezes by moving the warmer, upper air closer to the ground. The strong relationship between height above the ground and bud survival suggests that there was warm air above the vineyard that might have been moved to the lower sections.

How we treated the vines

Our goals were to maximize crop for the 2003 growing season and minimize retraining and pruning problems in 2004. Most vines were VSP trained. We use catch wires during the summer to keep the shoots vertical, and last spring we kept those wires in place. For pruning, we cleaned up the cordons and/or removed one cane wherever there were double wrapped canes. We trimmed the tops of vertically oriented canes just above the top pair of catch wires. We also spent a lot of time tying up suckers to use as future replacement trunks.
In one block containing, Chardonnay grafted to different rootstocks, we accounted for every node retained on the vine. For each cane we recorded the number of nodes, and for each node whether it had a shoot or a fruitful shoot. For the purposes of this analysis, we assumed that fruitful shoots came from primary buds and other shoots from secondary or tertiary buds. For other vines, we just noted the number of nodes and fruitful or non-fruitful shoots on each cane.

Rootstock Block

Almost 3,000 canes examined on 280 vines. On average vines had 15 canes that were 8 nodes long. This means there were about 120 nodes per vine. Average bud kill in this block was 78%, meaning the average vine should have had about 26 shoots. Our normal goal is 24 shoots/vine, so in many cases there were more than enough shoots to get trellis fill, and the primary job was to keep the best 24 shoots.
Figure 3 illustrates bud survival in the block on a cane basis. About 1/3 of the canes had no shoots and a little more than 1/3 had 1 or 2 shoots per cane. Twenty percent of the canes had 3 to 5 shoots and only about 10% of the canes had more than 5 shoots.

Figure 3. Relative frequency of canes with different numbers of shoots.

Factoid 2. Canes with persistent laterals are a problem. They indicate high vigor (and hence less hardiness) and their fruitfulness is unpredictable. Buds on the cane node may often be unfruitful, but in those cases the first node on the lateral will tend to be very fruitful. Persistent lateral canes should be spurred, but the combination of cane node bud and the lateral node bud should be counted as a single unit when balance pruning.

This factoid may or may not be well known, but at the time that balance pruning concepts were strongly stressed, it was considered an important “fact”. When vinifera varieties began to be grown commercially, there was a lot of concern about persistent laterals. Our early observations were that, unlike labrusca varieties, the buds at both the cane node and the lateral node tend to grow shoots and be fruitful on vinifera varieties. Thus, a spurred lateral should be treated as having two buds, not one. There still was concern about cold survival. In high vigor situations, there were plenty of occasions when the cane buds failed while lateral nodes survived winter. Fortunately, as growers have gained experience with vinifera, these instances of very excessive vigor have become more rare.

Figure 5. Bud survival at nodes with or without persistent laterals.

In assessing the event we are discussing here, we kept track of bud survival on persistent laterals as well as of cane node buds. The relatively low vigor status of these vines is indicated by the fact that only 1.4% of the nodes had persistent lateral canes. These were trimmed to 1 node. The survival data is shown in Figure 5. The data support the concepts given above. First, bud survival on nodes with persistent laterals was slightly lower than on nodes without persistent nodes (18% vs. 27%). Secondly, survival of buds on persistent laterals was much higher than buds on canes. These data suggest that spurring lateral canes might be good practice, especially when increased shoot counts are desired.

Factoid 3. Pruning cold damaged vines to spurs is not a good idea. Because spur pruning retains only basal nodes that develop in a more shaded region of the vine, they will be neither as cold hardy nor as fruitful as nodes developing further out on a cane.

Figure 6 shows both the total number of nodes in the vineyard at different cane positions (node 1 is at the base of cane), and the number of nodes with shoots. Again, the data both supports and does not support the conventional wisdom. Note that the proportion of live nodes increases as node position increases. Thus average percent bud survival will be higher on canes than on spurs. However, there are many more canes with 3 to 5 nodes than there are with more nodes (more short than long canes). Thus, spurring all canes to 4 or 5 nodes can result in a similar number of shoots as retaining a lower number of 10 – 12 node canes. Of course, retaining all canes as was done in this case will result in a still higher number of shoots.
We should also remember that if we had actually cut all the canes back to 3 to 5 nodes, then the proportion of buds that developed into shoots might have been increased. This is because bud break at nodes 3 to 5 would not be inhibited by shoots developing at higher node positions which may have happened in this vineyard.

Figure 4. Number of canes with nodes (open bars) and live nodes (solid bars) at different node positions. Node 1 is at the base of the cane.

Factoid 4. Maintaining cordons on winter-injured vines can be a problem. There will be high survival at the ends of the cordons and lower survival in the vine center, or alternatively, Maintaining cordons on winter injured-vines can be a problem because canes developing at the ends of the cordons do not mature as well as those developing in the head of the vine.

Figure 7. Percent bud survival at different cane positions on the vine. Canes are numbered consecutively from south (node 1) to the north end of the vine (node 24).

Figure 7 reveals no obvious pattern of bud survival by cane position. It looks as if survival of the first two canes on the south end of the vine might be lower than average and the opposite true at the north end. Statistical analysis failed to reveal any significant difference among node positions or a correlation between node position and survival.
This doesn’t mean that we are off the hook in relation to cordons. Although there is no obvious pattern, there may well be many spur positions without shoots on these vines. These will have to be filled in with base shoots or the cordons may need to be replaced. Relative survival of cordons and canes will be discussed further below.

Factoid 5. Low vigor vines or vines growing on low vigor rootstocks will resist winter cold better than more vigorous vines.

This is a question we have been chasing for many years. All the data discussed up to now came from a rootstock experiment that had been in the ground for more than 10 years. Although the vines had not been managed as strictly in the last two years as previously, we were anxious to see if insights into vine hardiness would be revealed.
Table 2 summarizes the results with 22 different rootstocks. The stocks are divided into three vigor categories based on the average vine size during the previous 12 years. The data do not show any real relationship between average vine vigor and survival. If anything, low vigor vines perform less well than medium or high vigor vines, which had higher bud survival, shoot number and tended to have more fruitful shoots.

This is consistent with our previous experience in this vineyard, and by performance of individual rootstocks within a given vigor category. Vines of superior and inferior survival were found in each vigor category. This is also shown in figure 8 where box plots are made for the mean number of fruitful shoots/vine for rootstock vines in each vigor category. There were 7 different rootstocks each in the medium and high categories. The low vigor category contained 8 different rootstocks. Note that there vines with 40 or more fruitful shoots/vine in each category as well as vines with 30 or fewer.

Figure 5. Box plots of fruitful shoots per vine in 2003 for rootstocks in different vigor categories. Line for 50% means half the stocks produced fewer fruitful shoots/vine and half the stocks produced more fruitful shoots/vine.

Training Systems and Survival

We did not map each node on each cane when assessing survival in our training system experiments, but we did evaluate each cane, and obtained yield records. Table 3 shows the results for three varieties trained to five different training/pruning combinations.

Table 3. Effect of variety and training system on node production and bud survival

Several patterns emerge. Lyre trained vines tend to have more canes, more nodes/vine, and more fruitful shoots than other systems. Mid-wire cordon trained vines had high values for canes/vine and the lowest percentage of dead nodes. When cordon pruning is compared to cane pruning both VSP and Lyre trained vines have better survival values for cordon pruning. Scott-Henry trained vines were not superior. In all cases the lower (down pointing) parts of the Scott-Henry vines had little growth or survival

Effect of training systems on yield components.

Average yield ranged from 1.9 tons/acre (Cabernet Sauvignon – VSP cordon) to 7.9 tons/acre (Cabernet Franc – mid-wire cordon) (table 3). These yields justify the extra expense involved with modifying the pruning practices for 2003 and the extra cost pruning/training that will be incurred in 2004.

Lyre training was generally superior to VSP or Scott-Henry (tables 3, 4). Cordon pruning was superior to cane pruning. Scott-Henry produced the lowest yields for Chardonnay and Cabernet Franc vines, but was better adapted to Cabernet Sauvignon. The strategy was too successful with Cabernet Franc. An almost 8 ton/acre yield was too high to sustain wine quality. Even the three treatments with 5 or more tons/acre may have been marginally over cropped.

Training System

Performance of the various training systems is highlighted in Table 4 The last data column in the table is labeled % dead arms. Growth from each arm was evaluated at the end of the season for signs of trunk injury. Signs included collapse and withering of vegetative growth, reddening of foliage or premature foliage drop. The data indicate that Scott-Henry and Lyre cane pruned vines were most injured. Mid-wire and VSP cordon pruned vines had the lowest damage levels.

Table 4 . Effect of variety and training system on yield components in 2003.

Performance of the various training systems is highlighted in Table 5. The last data column in the table is labeled % dead arms. Growth from each arm was evaluated at the end of the season for signs of trunk injury. Signs included collapse and withering of vegetative growth, reddening of foliage or premature foliage drop. The data indicate that Scott-Henry and Lyre cane pruned vines were most injured. Mid-wire and VSP cordon pruned vines had the lowest damage levels.

Table 5. Effect of training system on survival and yield components. Data is average of three varieties

Performance of varieties and various clones of varieties

White wine varieties

Nine clones had less than 75% injury including (in ascending order) Auxerrois, Chardonnay clones 7, 95, 4, 76, Tocai Friuliano, Chardonnay cl 96, Fresia and Chardonnay cl 78. Arneis, Trousseau, Viognier, Sauvignon blanc, and Semillon had more than 90% shootless nodes (table 1).
Only Auxerrois had more than the desired 24 fruitful shoots/vine. Chardonnay clones made up 7 of the 11 clones with more than 10 fruitful shoots per vine. Rkatsitelli, Fresia and Tocai Friuliano were the others. Sauvignon blanc and Semillon had less than 1 fruitful shoot per vine. Auxerrois, Rkatsitelli and Clone 7 and 76 of Chardonnay had less than 10% trunk injury. Trousseau, Fresia, Viognier, Semillon and Sauvignon blanc had more than 80% trunk injury.

Table 6. Cane and node production and survival of white varieties and clones at Geneva, 2003.