At the start of production (3rd year), the mean mortality rate was 16.1 % (±2.4 %) of the replacement vines planted over 4 years (data not shown). This is considered a fairly good rate in the context of previously published work1. However, this mean average hides a wide disparity in mortality between plots. Cumulative mortality of replacement vines at the theoretical start of production ranged from 0 to 90 %, with one-quarter of batches showing a mortality rate below 5 % and an equal number at above 20 % mortality. Multi-year monitoring of plant survival revealed that 60 % of mortality occurred in the year following dead vine replacement, followed by 23 % and 10 % in the second and third years (Figure 1A). At the end of the first year, 14 % of batches had no dead vines and 39 % had a mortality rate below 5 %. The mean mortality rate across the batches was 11.3 % (±1.6 %). The mean mortality rate then fell to 3,92 % (± 1,3) in the second year and 1,85 % (± 1,4) in the third year (Figure 1A). This result highlights the importance of taking care of replacement vines, particularly in the year they are planted, both in terms of treatments against disease and in terms of irrigation to reduce water and heat stress.

Figure 1. Mortality rate of batches from the year of planting of the replacement vines (1F) to the year following the first theoretical year of production (4F) (df = 3, χ² = 22.7, p <0.001) (A) and as a function of the time of year (excluding first leaf, 1F), distinguishing between growing-season (GS) mortality and mortality following winter rest (WR) (df = 1, χ² = 15.9, p <0.001) (B). The letters above the bars indicate significant differences between categories.

Intra-annual variability in the mortality rate was also observed. After the first year, there was a significant difference in mortality depending on the time of year (Figure 1B). Mortality at the end of the growing season (GS) is much lower than mortality corresponding to the absence of budburst at the end of the winter rest (WR). This result could be attributed to a lower availability of reserves for certain plants. This unexpected result is interesting, as growers mainly attribute plant mortality to a significant water deficit leading to a plant desiccation in summer or to tillage when the soil is worked mechanically.

There is a significant vintage effect between batches planted in 2019 with the lowest mortality (4.8 % in average) and batches planted in 2021 with the highest mortality (17.3 % in average) (Figure 2A). Weather conditions during the respective seasons could explain these differences. Spring 2021 saw particularly high rainfall. In contrast, 2019 saw the lowest rainfall of the four years with just 454 mm between January and October, compared with 641 mm over the same period in 2021. In 2021, cumulative rainfall was 50 % higher than the 30-year average between April and June, which could have led to root asphyxia resulting in increased mortality.

Figure 2. Mortality rate of replacement vines in the first year by vintage (df = 3, χ² = 10.9, p = 0.01) (A) and by plot age (df = 1, χ² = 0.03, p = 0.86) and grape variety (df = 3, χ² = 1.4, p = 0.23) (B). The letters above the bars indicate significant differences between categories.

On the other hand, the age of the plot in which the replacement vines were planted did not seem to have any impact on the mortality rate in the first year, contrary to the generally accepted idea that competition from older vines compromises the success of dead vine replacement. However, only a few batches were planted in young vineyards less than 15 years old in this study. No grape variety effect was demonstrated (Figure 2B).

Figure 3. Mortality rate of replacement vines in the first year by soil type (df = 4, χ² = 9.1, p = 0.06). The letters above the bars indicate significant differences between categories.

A significant soil type effect was found among the 5 most common soil types in the study. In the first year after planting, mortality was 5 times higher on CALCOSOLS than on COLLUVIOSOLS (Figure 3). Mortality was also higher on LUVISOL, but the difference is not significant. For the CALCOSOLS, it could be hypothesized that the amount of clay in these soils could lead to temporary waterlogging, particularly in wet years. The waterlogging induced by clay in the CALCOSOLS could have caused root asphyxia and reduced nitrogen mineralization. Mortality due to tillage may be higher on heavier soils, which could explain the lower mortality observed on BRUNISOLS and PEYROSOLS. While it was noted that plants were uprooted by passing machinery, it was not possible to draw any conclusions on this point. However, our two-yearly scoring system was not ideally suited to assessment of this variable. Losses due to this variable could only be estimated by taking readings after each soil maintenance operation.