Memoire Online - The use of short-term solutions against grape sunburn within a context of climate change in the Médoc vineyard

As mentioned earlier in this report, climate change is causing the recent multiplication of both drought and canicular events during summer in the Médoc region. Therefore, more and more grape sunburn has been observed, resulting in yield and economic losses. Global warming has caused a modification of the typical Médoc climate, forcing the producers to change their production habits. Producers indeed have the choice between choosing to change the way they produce their wines (change the grape variety for example) at the risk of producing wines with different organoleptic profiles, or to adapt to climate change by finding short term solutions to reduce sunburn.

Château Margaux being an iconic figure of the Médoc vineyard, and being known worldwide for its exceptional wines, it seems impossible for them to entirely change their wines profiles. Their only option is to find solutions to adapt at a production level, without affecting their grapes and wines quality.

In the long-term, they plan on reorienting their plots of land to reduce the grapes exposure to the sun on the afternoon. Before being able to do so, they need to find short-term solutions.

The objective of this study is to find short-term solutions to adapt and reduce the effects and consequences of grape sunburn on Château Margaux's wine production.

What short-term solutions can be applied to a Médoc vineyard to significantly reduce its losses linked with grape sunburn?

- Will the spraying of a kaolin-based solution on the grapevine allow to protect the vineyard against grape sunburn?

- Is there a significant difference between the tested methods? Which solution seems to be the most efficient against grape sunburn?

- Will berry quality be impacted by kaolin treatments and early canopy defoliation?

- What effects the implementation of these potential solutions will have on the current vineyard management strategy?

The mission of this report will be to prove the efficacity of two different methods to potentially reduce grape sunburn. Both methods will be tested in the vineyard to evaluate their effect on grape sunburn, for the 2022 vintage.

It can be proposed to test both the use of a kaolin-based solution, as well as the use of early moderate plant defoliation before berry set on the vineyard to evaluate their impact on sunburn.

Based on the bibliographic review conducted, and on the studied problems, different hypotheses were formulated:

A study was therefore conducted at the scale of Château Margaux in order to verify those hypotheses.

For the experimentation to start, different choices had to be made. Among the decisions, the parcel choice and studied modalities were the most important, as they can affect the results of the study. The parameters that affected those decisions will be described in the experimental set-up.

The choice of parcel for this study was made based on different factors. The row orientation, the grape variety, and the frost damage were the main variables that were considered for this choice.

First, to expect some significant results, parcels with a great intensity of sunburn damage from the previous years were chosen, to make sure the results wouldn't be altered by the lack of sunburn.

In 2020, studies were conducted at Château Margaux to evaluate if row orientation had an impact on grape sunburn. As a result, they found that the best orientation to limit grape sunburn was North-East / South -West (Porte, 2020). In order to make sure that the plot chosen for the study would be the worst scenario case, plots with different exposures were chosen, allowing more grape sunburn.

The first parcel chosen is called «Jean Brun Ouest» (JBO) and is oriented East / West. It is part of the «Devant le Château» production block. The second parcel chosen is called «Les 4 Vents Sables» (L4VS) and is oriented North-West / South-East. It is part of the «Plateau» production block. A map with every parcel's orientation is available in Annex 6, and a map with the blocks is available in Annex 4.

Both grapevine parcels produce red grapes from the Cabernet Sauvignon grape variety. Both parcels are planted with a density of 10 000 plants/ha and are older than 10 years old. Those parcels were chosen for their soil homogeneity, to avoid biases.

This specific variety was chosen to study because it represents 52% of the total vineyard and is therefore the most important part of the grape production at Château Margaux. Two parcels were chosen rather than one in case differences in grape sunburn symptoms were observed between orientations.

In order to make sure that the grape production would be homogeneous in the plots chosen, a frost damage counting was performed beforehand. The frosts occurred at the very beginning of April, and the counting was done during this same month, to evaluate if many fruit-bearing bud might be

affected. To do so, a scale from 1 (<30% of buds dead) to 3 (>75% of buds dead) was established. A few rows in the parcels were then chosen where a note to each vine plant was given. Finally, the results were calculated and mapped, so that the company has an idea of which plots were affected by frost.

Both chosen parcels weren't affected by frost, which is an additional reason as for why they were chosen.

For each parcel of land, three modalities were studied. The first is the kaolin sprayed modality (K), the second is the early defoliation modality (ED), and the last one is the control modality (C). To compare modalities, it was decided to combine them all on one same parcel of land, instead of allocating one parcel per study. By doing so, it reduced the uncertainty of the results linked with the homogeneity and type of soil.

To reduce the incertitude linked with the heterogeneity of the plot of land, each modality was repeated 3 times per parcel. By doing so, there were a total of 9 blocks in one parcel, and 18 blocks in total (Figure 6 and Figure 7).

Figure 6: Scheme of the experimental plan of Les 4 Vents Sable, including the repartition of the modalities in the parcel, the
plots chosen, and the captors location

Figure 7: Scheme of the experimental plan of Jean Brun Ouest, including the repartition of the modalities in the parcel, the
plots chosen, and the captors location

Inside each block, a plot of 10 selected plants was followed during the rest of the study. Plots were chosen so that there wasn't any missing plant around and inside the plot, to avoid false results. For example, one missing plant just in front of the studied plot would increase its sun exposure period, and therefore induce error into the results.

The kaolin modality received different sprayings during the season. The spraying solution was prepared with a 20kg/ha dose of kaolin power (Sokalciarbo by AgriSynergie), water, and adjuvant (Vizir by AgriSynergie) at a dose of 20mL/100L of water as advised on the notice. The solution was mixed using a Mixbox mixing tank (Annex 7). The dates of treatment were defined based on the conditions observed. According to the kaolin powder's technical data sheet (Agrisynergie, 2022), the sprayings had to be done between bunch closure and veraison, before any heat wave, and after a leaching above 15mm. The maximum number of sprayings mustn't exceed 4 per year, and the last spraying had to be done at least 15 days before harvest to avoid residues (E-Phy, 2022).

The kaolin sprayings were done on the 14^th of June, the 7th of July, the 22^nd of July, the 29^th of July and the 9^th of August. A calendar is available in Annex 8, and dose calculations were reported in Annex 9.

Figure 8: Photograph of Cabernet Sauvignon leaves before (on le left) and after (on the right) the first kaolin spraying in

For the early grapevine defoliation, a few leaves were suppressed on only one side of the canopy, the North-Eastern side, exposed to sun on mornings. The process was moderate and not drastic, in order to gradually expose the grapes to the sun (Gaviglio 2022). Mostly the side shoots of the plant were suppressed, to allow a better sun exposure, without exposing too much the bunches (Figure 9).

According to bibliography, the defoliation had to be done right after the flowering period, during the fruit setting stage to obtain berries with a higher sun resistance (Serrano, 2018). The defoliation was done in June the 1^st for the Jean Brun Ouest parcel, and June the 2^nd for Les 4 Vents Sable parcel. Early defoliation was performed for both experimental parcels when more than 50% of the parcel was at the fruit setting stage.

For each studied variable, its definition will be reminded, it will be linked with one of the hypotheses/problems, the protocol and the material needed will be presented, and finally the date of data acquisition will be given.

Vintage characterization is done by analyzing the rainfall as well as the different types of temperature for a defined period. The meteorological data will help characterize the 2022 vintage to make sure the conditions were conducive to sunburn.

Meteorological data were provided by Sencrop weather stations located in the «Enclos» and «Plateau» blocks (their location can be found in Annex 4), as well as a global Margaux station. Values of under-shelter temperatures, humidity, rainfall and wind speed were collected for the 2022 growing season.

The following meteorological data: minimum temperature, maximum temperature, average temperature, and rainfall were recorded for each day to characterize the specific conditions of each parcel. Meteorological data collected of each parcel defined the climatic conditions along the growing season.

Based on the collected data, the number of days from March to the end of August with a maximum temperature higher than 30°C was calculated, in order to compare it with the previous years. Sunburn usually happens when temperature is higher than 30°C, so this computation will give information about previous sunburn events. The objective of this measure is to verify the hypothesis that the 2022 weather conditions were conductive to sunburn.

The vegetation zero is the minimum temperature from which a plant can develop. For grapevine, the vegetation zero is 10°C (Ephytia, 2022), meaning that grapevine usually develops faster when temperatures are higher than 10°C. The 10°C base temperature sum is therefore an indicator for grapevine growth over a defined period. Red grape berries are more sensitive to sunburn when mature, due to their darker color, and lower radiation reflectance. The objective of this measure is to verify the hypothesis that the 2022 weather conditions impacted the growth, and therefore the phenological stages precocity of grapevine, resulting in higher sunburn sensitivity.

Since grapevine growing season starts in March, along with the budburst, and ends with harvest, the weather data was only analyzed between March and August 2022.

A plant's phenological stage is characterized by the plant's development during its life cycle. The phenological study of a plant consists of observing the date at which those stages appear (Roussey et al., 2021). The objective of this variable is to define the precocity of both studied parcels, to evaluate their potential impact on the results.

Phenological stages of reference parcels were evaluated using a counting method. The most important stage that needed to be followed for our study is the fruit setting stage when the flowers start to form berries after a successful fertilization.

For the fruit setting stage, the percentage of bunches for one reference plant that were set was counted and calculated. Then, at the parcel's scale, an average was calculated. By repeating this evaluation at different dates, the exact date of the mid-fruit setting stage was defined, to start defoliating the grapevines. The exact same calculation method was done for the other stages.

The results were then compared to previous years, to define the precocity of the 2022 vintage.

Plant vigor is observed by an increase in plant height and density through time (Short and Woolfolk, 1956).

This variable reflects the canopy density and is therefore supposedly negatively correlated with sunburn. This variable was measured to make sure that they weren't any difference between the 3 studied modalities in terms of vigor, that could potentially affect the results.

To evaluate the vigor of the grapevine plants, aerial images were used. The vigor was estimated with an Enhanced Vegetation Index (EVI) based on the exact measurement of reflected light off the plants at different wavelengths, using Vineview (Vineview, 2022). The EVI values range from 0 to 1. Values close to 0 correspond to a bare ground, whereas values close to 1 correspond to a complete vegetation cover (Fraga et al., 2014).

The EVI maps extracted from Vineview dated from May 2022, so that the data was before any kaolin treatment or defoliation.

Vegetation porosity can be defined as the measure of blank spaces in the canopy's vegetation, and is the fraction of the blank portion per the non-blank portion of the canopy (Bélanger, 2017). The higher the porosity, the higher the sun exposure, and the higher sunburn sensitivity (Southey and Jooste, 1991). Vegetation porosity is evaluated by measuring the Leaf Area Index (LAI) for different plants.

The objective of this measure is to control the homogeneity of the followed plots, to make sure that significant differences aren't observed between modalities that could affect the results.

To conduct this measure, the height, length and percentage of blanks in one plant's foliage were evaluated. For each block, this measure was conducted on 1 plant every 3 plants (3 plants out of the 10 reference plants). The measures were done in August the 3^rd, during the grapevine maturation process, when it was evaluated that canopy had stopped its growth.

Grapevine hydric condition can be estimated my measuring the water potential of the vine. There are different types of water potentials. Both potentials are measured using a Scholander pressure chamber (Annex 10). The grapevine water status of the plant is defined by estimating the cell capacity to retain water, using the pressure of a neutral gas applied on the leaf, at different points of the soil-plant-atmosphere system (Ojeda and Saurin, 2014). The more the plant will be stressed, the more pressure will need to be applied to extract sap. The pressure needed represents the current hydric state of the leaf (Dufourcq, 2022; Deloire et al., 2005).

Both measures of water potential will be conducted in order to answer to the hypothesis that spraying kaolin can positively impact the grapevine hydric state in stressing conditions.

Predawn Leaf Water Potential (PLWP) is an estimator to assess soil water availability for species like grapevine at the scale of a parcel (Suter et al., 2019).

This measure gives an estimation of the basic hydric condition of the parcel, to make sure that there are no signs of water stress that could influence results. It is done predawn (from 2am to 6am) and represents a state of balance between the vine plant hydric state and the soil hydric condition. It gives a threshold reference of the current soil water availability for the plants at the scale of the parcel (Deloire et al., 2005).

Only the PLWP for the reference plant of the control modalities of the parcel (3 reference plants) was measured, so that it gives an idea of the evolution of the global hydric condition of the parcel.

For this measure, one leaf of each control modality is cut off the plant, and the petiole is inserted through the lid of the pressure chamber, its cut end remaining exposed. The chamber is then activated, and the petiole is examined until liquid is observed on its surface. When there is liquid, the chamber is turned off and the pressure on the gauge at which water was observed corresponds to the PLWP.

The PLWP measure was taken at different periods during the growing season to obtain an evolution of the water availability through the summer. The PLWP was measured in June the 16^th, July the 11^th and 13^th, and July the 26^th.

The Midday Stem Water Potential (MSWP) is a measure that represents the state of water tension during the plant transpiration, under stressing conditions. This method is usually more precise than the PLWP and is used to compare the hydric constraints of different modalities (Dufourcq, 2022).

The objective of this measure is to compare it between modalities, to define whether or not using kaolin helped reducing the hydric stress of the plant.

This measure is done at the solar noon (around 2pm) after at least 4 days without rain, so it remains precise. The objective of the stem water potential is to give an idea of the grapevine hydric condition during the experimentation period (Dufourcq, 2022).

The chosen leaf is placed in an opaque bag for at least two hours. Then, the leaf is cut off the plant, and the petiole is inserted through the lid of the chamber, its cut end remaining exposed. As for the PLWP, the chamber is then activated, and the petiole is examined until liquid is observed on its surface. When there is liquid, the chamber is turned off and the pressure on the gauge at which water was observed corresponds to the stem water potential.

For this experiment, one reference plant per block for each modality was defined (9 reference plants per parcel). The MSWP of each reference plant was measured, in the Scholander pressure chamber.

Like for the PLWP, the MSWP measure was taken at different periods during the growing season, on June the 13^th, July the 12^th, and August the 11^th.

Leaf surface temperature can be defined as the measure of grapevine's canopy temperature. This measure was proven to be a good indicator of water stress.

Stomatal closure is one of the earliest responses to hydric stress for grapevines. In constant environmental conditions, the leaves' temperatures should decrease because of stomatal conductance increase. The higher the leaves' temperatures, the less the vine plant has access to water (Grant et al., 2016). In normal conditions, grapevine should be able to reduce its leaves temperature with transpiration through stomatal conductance. When the hydric stress becomes important, plant transpiration ceases, and the leaves temperatures rise. Leaves temperature is therefore a good indicator of grapevine water stress (Jackson et al., 1981).

The objective of this measure is to verify the hypothesis that kaolin allows the leaves to reduce their temperatures by reducing the stress of the plant. By comparing the leaf surface temperature between modalities, we can potentially put into light differences of hydric stress.

Additionally, the temperature of sun-exposed leaves was taken during different afternoons in the season, at the bunch level. 30 leaves per modality were randomly chosen (10 leaves per block x 3 blocks) and their temperatures were measured using a manual infrared thermometer (IM8823, iMesure).

For the choice of date to measure the temperatures, the external air temperature should be around 30°C, which is the temperature around which leaves stop their photosynthesis and rise in temperature as they cannot continue to perform evapotranspiration. As kaolin reflects light, studies on grapevine reported that it can reduce the leaves temperature up to 3-8°C (Agrisynergie, 2022), enough for the plant to continue its photosynthesis.

Temperature was measured when kaolin was still present on the leaves, and not after a period of rain where it could have been washed-out. After the first heatwave in June, it rained, washing kaolin off the leaves, not allowing the measure leaf temperature with kaolin. Consequently, this measure was conducted after the other kaolin treatments on July the 11^th to the 13^th, July the 26^th and 28^th, and August the 9^th.

Bunches microclimate was evaluated throughout the season. The main variables evaluated were the bunch temperature and the received luminosity.

The bunch temperature can be defined as the temperature measure of the bunch of grapes at a specific moment of time during the growing season.

Bunch temperature was used in order to verify the hypothesis that applying a kaolin-based particle film on grapes and leaves will reflect a small part of solar radiation, allowing the bunches to get cooler. It was also used to verify the impact of early defoliation on bunch temperature, as no bibliography clearly stated a significant difference.

The microclimate temperature of the fruit zone was measured on the 9 plots of each parcel (3 modalities x 3 repetitions). To measure this temperature, temperature and luminosity captors (HOBO Pendant MX2202, ONSET) were placed on the afternoon sun-exposed side of the row (Figure 10).

The captors were positioned so that they were in the same exposure and temperature conditions as the
studied bunch of grapes. The chosen position of the captors was near bunches of grapes with high

sunburn risk. Those bunches are indeed shady in the morning but exposed on the afternoon. One HOBO captor per block was positioned, representing 9 captors per parcel.

Additionally, 4 under-shelter temperature and relative humidity Tinytag captors were positioned on each parcel. They were placed at different places of the parcel so that they give the mesoclimate of the parcel. Both captor types were programmed so that the data is captured continuously during the season, every 15 minutes. Tinytag captors' results were used to obtain the under-shelter temperature and humidity at a smaller scale than the weather stations to verify if there were differences at the scale of the parcels.

Before activating both the HOBO and Tinytag captors on the parcels, they were calibrated to make sure their data was similar and remove the ones with odd results.

To verify the bunch temperature difference between modalities, the temperature of 30 bunches of grapes per modality was taken with the infrared thermometer (IM-8823, iMesure) at given times. The bunches were chosen randomly and had to be sun-exposed during the afternoon. The measures took place during high-temperature days before and during fruit ripening on the 13^th to 17^th of June, and on the 3^rd of August during maturation, when kaolin was visible on the canopy. Different measures were done in June to calibrate the bunch temperature model at a more precise degree, and later measures were done in August to evaluate the bunch temperature difference after Véraison.

The measure of luminosity can be defined as an estimation of the received luminosity by the bunches of grapes, a factor that can greatly influence the apparition of sunburn symptoms.

The objective of this measure is to give an idea of the received solar radiation for each modality, at the bunch level. This measure should verify the hypothesis that applying a kaolin-based particle film on grapes and leaves will reflect a small part of solar radiation, and that moderate early defoliation will allow sooner and higher sun exposure.

The luminosity data was used to verify those hypotheses, by comparing acquired luminosity between modalities.

Using the HOBO captors, luminosity received by the canopy was measured every 15 minutes of every day of the growing season. This data was used to enrich the bunch temperature calibration model to compare modalities, in order to find for each modality, the correlation degree between luminosity and bunch temperature. The objective of this measure is to give a more precise bunch temperature for each modality, based on the fact that luminosity influences bunch temperature and sunburn symptoms.

Quantifying grape sunburn will help to evaluate the influence of both kaolin and early defoliation on the symptoms. The objective of this measure is to verify if both methods will positively impact sunburn symptoms apparition.

This measure will have the role to verify the hypothesis that the difference in terms of grape sunburn symptoms will be significant between the control modality, the kaolin treated modality, and the defoliation modality.

In order to quantify sunburn symptoms, counts in the vineyard were done throughout the season on both bunches and leaves. For each parcel, 10 grapevine reference plants per block (30 per modality) were observed and their symptoms were evaluated.

Bunch counting is a measure that estimates the number of bunches one grapevine plant will produce. The objective of this measure is to evaluate the average number of bunches per grapevine plant, in order to weight the results, based on the average number of bunches per reference plot.

To do so, for each parcel 5 plots of 10 consecutive plants were defined in different parts of the parcel, and for each plant the number of clusters were counted. By doing so, it gives an idea of the average number of grape bunches per parcel, and therefore an idea of the future yield.

A first count occurred in June the 14^th. Thinning operations took place on both parcels between July the 5th and July the 8^th, leading to another count on July the 13^th.

As mentioned earlier, grape sunburn can lead to two forms of symptoms: the first one being the browning of the berries, and the second one being the withering of the bunch. Bunch sunburn can greatly affect the quality of the berry, as well as the yield, as it can dry up entire bunches.

The objective of this measure is to verify the influence of the modalities on the bunch sunburn symptoms, and to define whether or not the applied methods are working.

To quantify the sunburn symptoms on the bunches of berries, each affected bunch out of the reference plots were counted and the intensity of sunburn symptoms in percentage was visually estimated. The frequency was then calculated by dividing the number of affected bunches by the total number of bunches. Finally, damages caused by sunburn (Frequency*Intensity) were estimated, to give an idea of the losses linked with sunburn.

To obtain the total number of bunches per block, the previously calculated average number of bunches per grapevine plant per parcel (3.4.1) was used. It was then multiplied by 10 (as there were 10 plants per block) to obtain the total number of bunches.

The first bunch sunburn symptoms evaluation counts took place on: June the 22^nd and 27^th. Other visual counts were performed later in July the 20^th and August the 17^th after two main sunburn episodes. In between, another evaluation was performed in July the 8^th and 11^th because thinning was done by hand on the 5^th and 8^th of July for JBO and L4VS respectively.

Although leaf sunburn doesn't have a negative effect on the berry yield and quality, it is a great indicator of heat stress undergone by the plant.

Quantifying leaf sunburn will allow us to verify the effects of both kaolin and early defoliation on the canopy. Under hot and dry conditions, stomata must close to prevent dehydration (Brodribb and Holbrook, 2003). However, as kaolin should reduce canopy temperature by causing stomatal opening, it could potentially cause superior leaf dehydration.

In order to quantify the sunburn symptoms on the leaves out of the 10 reference plants per plot, the number of plants with leaf sunburn were counted and sunburn intensity was visually estimated. By doing so, it gives an intensity and frequency evolution during the followed period of time.

The frequency was calculated by dividing the number of affected plants by the total number of plants per block (10). Just as for the bunch sunburn evaluation, total damages were calculated by multiplying the frequency of leaf sunburn by the average intensity of affected plants. The leaves sunburn symptoms evaluation counts took place on: June the 8^th, July the 26^th, and August the 17^th.

Berry quality evaluation is a process that will help verify the hypothesis that berry quality won't be negatively impacted by kaolin treatments and early canopy defoliation. Quality evaluation will be based on different criteria such as: the berries mass and volume, and the measure of primary and secondary metabolites per modality.

Berries mass represents the weight of the berries while their volume can be defined as the space they occupy.

Because kaolin and early defoliation can significantly reduce sunburn symptoms, they can consequently reduce the number of withered berries, and therefore increase the average berries volume compared to the control modality. However, the volume difference linked with sunburn often isn't significative on grapevine, as berries are small (Brillante et al., 2016).

The objective of this measure is to verify the hypothesis that berry physiology won't be negatively impacted by kaolin treatments and early canopy defoliation.

To evaluate the mass and volume of the berries, the berries that were picked for primary and secondary analysis were used. They were weighted to obtain an approximation of the mass of 100 berries per modality.

The average berry volume was precisely calculated using a Dyostem (by Oenosens). Berries were inserted in the machine for each modality, and their volume was precisely measured.

Grape quality is determined by the contents of the primary and secondary metabolites (Pavlouek and Kumta, 2011).

The objective of evaluating the berries metabolites between modalities is therefore to verify the hypothesis that berry quality won't be negatively impacted by kaolin treatments and early canopy defoliation.

To do so, primary and secondary metabolites have been measured and studied during the maturation phase of the berries, to make sure no significant differences could be observed between modalities.

Primary metabolites come from the primary metabolism that regroups synthesis paths necessary to growth and plant development. Grape primary metabolites involve sugars and organic acids (Chaabani, 2019; Pavlouek and Kumta, 2011).

The objective of comparing primary metabolites levels between modalities is to verify that both treatments won't affect wine quality.

The total acidity of a berry should reduce as it is maturing, while the sugars and pH should increase, as a response to the ripening process. Consequently, the higher the sugar and pH levels, and the lower the acidity levels, the more mature the berry.

To evaluate the primary metabolites per modality throughout the season, tests were run on berries samples. Samples of 200 to 300 berries per modality were collected in plastic freezer bags. Berries were randomly picked on both sides of the canopy, at 4 levels of the bunches.

The samples were then taken to Château Margaux's laboratory where analyses were conducted. The berries were pressed to extract their juice. The pH and density were measured using an automatic titrator. Total and malic acidity were measured using a sequential titrator. Malic acidity measure has a major consequence on the final produced wine acidity. If the measures of malic acidities are significantly different modalities, it will affect the final produced wine.

The sugar concentration in the berries was followed by modality, using the juice density measure.

From July the 21^st to August the 22^nd, analyses were conducted on the berries once a week to follow their evolution until they reach full maturity.

Secondary metabolites come from the secondary metabolism that regroups synthesis paths that aren't necessarily linked with plant growth. Grape secondary metabolites involve phenolic compounds and aromatic substances (Chaabani, 2019; Pavlouek and Kumta, 2011).

As seen earlier in this report (4.2.4), grapevine defoliation is supposed to increase the berry's production of polyphenols and anthocyanins due to a higher sun exposure and protect the berry from sunburn symptoms. To make sure that the berries of the early defoliation modality have a stronger protective skin than the other berries from different modalities, they were analyzed. Measuring the accumulation of polyphenols and anthocyanins in the berries will also give information on potential maturity differences between modalities during the season.

At the same maturity level, the higher the polyphenols and anthocyanins level, the thicker the berry skin.

To evaluate the primary metabolites per modality throughout the season, tests were run on berries samples. Samples of 200 berries per modality were collected in plastic freezer bags. Berries were randomly picked on both sides of the canopy, at 4 levels of the bunches.

The samples were then taken to an external laboratory (Excell) where analyses were conducted. The berries were crushed and centrifuged to analyze their juice using spectrophotometry. For both the anthocyanins and the phenolic compounds, the concentration in the berry was measured, and indexes were calculated. As a result, total anthocyanins, Folin-Ciocalteu index⁶, and total polyphenols / total polyphenols index levels were given.

Analyses were conducted on August the 3^rd and the 10^th, to make sure berries were ripe enough.

When trying to implement new solutions into the current organization, it is important to take into account the managerial and organizational implications linked with these new practices.

The objective of measuring managerial and organizational implications is to answer the following problem: what effects the implementation of these potential solutions will have on the current vineyard management strategy?

To evaluate the managerial implications linked with the implementation of such preventive solutions against grape sunburn, an interview of the vineyard manager of the company was conducted to understand how it will affect the vineyard's activity.

⁶ Measure of optic density based on phenolic compounds oxidation. Reflects the total level of phenolic compounds in berries.

To evaluate the impact of climate change on the production management and on the wine typicity, an interview was conducted with the vineyard manager of Château Margaux.

The interview was completed on July the 19^th, and its content transcribed in Annex 2.

For the characterization of the 2022 vintage, different graphs and tables were produced. To analyze the weather stations data, a graph was modelized representing for each weather station the maximum, minimum, and average temperatures, and rainfall evolution throughout the growing season. To compare the last growing seasons to our results, a table was done to compare the number of days with a maximum temperature higher than 30°C for the last 5 growing seasons. A 10°C base temperature sum was also calculated for the last 5 growing seasons to see when the plant's development was at is highest and lowest, and reported in a graph for comparison. To calculate the 10°C base temperature sum, the average temperatures of each day of the studied period above 10°C were added.

The phenological stages dates results were reported in a table for each parcel, and their results were compared to one another.

To evaluate the plant vigor of out three studied modalities, the EVI values were extracted from the studied plots in Vineview and an average EVI value was calculated for each modality out of their three repetitions. The values were then reported in a table. Then, a one-way ANOVA analysis was conducted using XLSTAT to determine if there are any significant differences between the modalities (XLSTAT, 2022a). If plots were noticed as significantly different from the others in terms of vigor, they were removed from the other results.

For the porosity measure, once the measurements were taken, the Leaf Area Index (LAI) was calculated

The LAI results per modality were reported in a table and a one-way ANOVA analysis was conducted using XLSTAT, to highlight any possible significant differences between modalities (XLSTAT, 2022a). The means resulting from the ANOVA were compared using the LSD function of the Fisher Test on XLSTAT, at a trust level of 0,95.

Both types of water potential measures (PLWP and MSWP) were reported in histogram graphs. Bounds between different hydric intensity values were also implemented into the graph, based on data provided by a study on vine water status (Leeuwen et al., 2009). Differences between MSWP values for each modality were evaluated with a LSD Fisher test at a trust level of 0,95 by conducting an ANOVA, in order to define if values are significantly different between modalities, at different periods of the season.

The leaf temperature data was measured then entered into databases. A one-way ANOVA was then conducted on these measures using a LSD Fisher test at a trust level of 0,95 between the three modalities for each parcel, using XLSTAT, to highlight any significant differences in leaf temperature (XLSTAT, 2022a).

Before implementing HOBO and Tinytag captors into the vineyard, they were calibrated. The calibration consisted in leaving the captors in the same conditions for a few hours, and comparing their results, in

order to make sure that they were homogeneous. The results were modelized in a graph using Excel where we plot the temperature, light and humidity data depending on the time.

The Tinytag captors' data was exported, and the results were plotted using an Excel graph for each parcel. Captors were compared to one another to verify potential intra-parcel temperature differences. The comparison dates chosen (between June the 17^th and the 19^th) correspond to the first heatwave event in June, to make sure that captors data do not differ with high temperatures.

The IR manually taken temperatures between modalities were not only used for the bunch temperature model, but also for punctual comparison between modalities, during hot sunny days. The results were reported in a graph where measured bunch temperatures were compared between modalities for each studied parcel. A one-way ANOVA was also conducted on those results, using XLSTAT, to verify if the bunch temperature was significantly different between modalities (XLSTAT, 2022a).

To acquire the bunch temperature for the studied period of time, the IR manually taken temperatures needed to be compared to the HOBO light and temperature data at given days and hours to produce a database. In the database, data was analyzed by modality and parcel at different stages of the growing season.

Based on the HOBO and IR temperature and light data, multiple linear regression analysis for each modality (6 in total: 3 modalities * 2 parcels) were done using XLSTAT to give the degree of influence of each variable on the bunch temperature (XLSTAT, 2022b). By doing so, XLSTAT gave us correlations between data in order to produce an equation for the bunch temperature. Before using the formula produced, the p-values were checked to make sure they were significant. Once the IR bunch temperature equations were obtained, they were used to produce a bunch temperature calibration model to apply to the HOBO data. The modelized bunch temperatures between modalities were then compared to define if there were significant differences justifying the use of kaolin or canopy defoliation. Those bunch temperatures were compared using a graph in Excel, and the control modality was used as a reference.

The bunch luminosity data was used for the bunch temperature calibration model where it was implemented into the multiple linear regression analysis as a variable.

Bunch counting data was used to be implemented into the calculation of sunburn symptoms frequency and damages on both bunches and leaves. The frequency was calculated by dividing the number of bunches with sunburn in the reference plot by the total number of bunches in the reference plot (calculated by multiplying the average number of bunches per plant by the number of plants by reference plot). The number of bunches was also reported in a table.

The quantification of bunch and leaf sunburn symptoms were processed the same way. Graphs were produced in order to compare the evolution of sunburn frequency, intensity and damages between modalities throughout the season. One-way ANOVAs were also done using XLSTAT and LSD Fisher tests at a trust level of 0,95 were done to verify the damage significance between modalities, to be able to conclude if both methods were actually efficient (XLSTAT, 2022a). The results were reported in tables.

Primary and secondary metabolites, as well as mass and volume data were used to produce graphs using pivot tables in Excel, where the evolution of all measures were plotted throughout maturation. To evaluate the sugar concentration in the berries, a table was used to convert the measured juice density value in sugar level (g/L) (Schaeffer 2018).

The interview conducted for the managerial and organizational implication part was used to report information about the consequences of implementing new solutions in the vineyard practices.

For each LSD Fisher test conducted using an ANOVA in XLSTAT, two groups with no letter in common are considered as statistically different (p-value < 0,001). The values reported in tables in the results regroup the average, the error margin, and the group letter.

After the study was set up, it started along with essential measures for each parameter. The objective of this part is to remind the studied data for every parameter, present the measured indicators, compare the data to references, and finally interpret the data to answer our problems and validate our hypotheses.

Vintage characterization is an essential factor that will help define the conditions of the study and their influence on its results. The 2022 vintage will be characterized based on its weather conditions as well as on the dates of its phenological stages.

To characterize the 2022 vintage weather conditions, two graphs were produced with data from the two weather stations near the studied parcels. Because the data from both weather databases are very similar due to the geographical proximity of both parcels, only the data from the Enclos weather station will be presented on Figure 11, while the data from the Plateau station will be available in Annex 11.

Figure 11: Evolution of the maximum, minimum and average temperatures as well as the rainfall for the 2022 growing
season, from March the 1^st until August the 22^nd, based on the Enclos weather station data.

Based on Figure 11, the 2022 wine growing season in Margaux can be characterized by important rainfall episodes between March and April with low temperatures. The season also faced a frost episode in April between the 3^rd and the 5^th. Those episodes didn't prosper during the season, and temperatures rose until they reached heat waves mid-June, mid-July, and mid-August. Overall, the 2022 growing season can be characterized as rainy until May, then hot and dry from June to September. The maximum reached temperature was 41,4°C on July the 18^th.

From 30°C (red line on Figure 11), the sunburn risk increases (3.1). According to the Margaux weather station database, there were 34 days during the 2022 wine growing season where the temperatures were higher than 30°C and the sunburn risk was growing. Consequently, the 2022 growing season can be characterized as favorable to sunburn development, making this study more significant.

In order to compare the risk of sunburn linked with temperature peaks, the last 5 growing seasons were compared, and the number of days that reached a maximum temperature higher than 30°C were reported in Table 2, according to the Margaux local Sencrop weather station.

Table 2: Comparison of the number of days where the maximum temperature (Tmax) was higher than 30°C, for the last 5
growing season, from March the 1^st to August the 22^th, according to the Margaux Sencrop weather station data

More days above 30°C were observed in 2022 than in the last 5 years. In 2021, the number of days with a maximum temperature above 30°C were lower than the other years, resulting in less sunburn. This could lead to think that the 2022 growing season was more affected by grape sunburn.

To compare and evaluate grapevine's plant growth over the 2022 growing season, the 10°C base temperature sum was calculated for the last 5 growing seasons in Figure 12.

Figure 12: Comparison of the 10°C base temperature sum for the last 5 growing seasons, from March the 1st to August the
22th, based on the Margaux Sencrop weather station data

Based on Figure 12, the 10°C base temperature sum seems to be higher during the 2022 growing season compared to the other seasons. It is higher than the average temperature sum between 1996 and 2015. The 2021 and 2019 seasons seem to be lower in terms of temperature sum than the other years.

Thus, the phenological stages of the 2022 growing season could have been shifted resulting in a higher berry sunburn sensitivity.

Overall, the 2022 growing season in Margaux can be characterized by important heatwaves causing intense plant growth. This season was hotter than the last 5 seasons, and the 10°C base temperature was higher than the other seasons. All the results show that the 2022 vintage meteorological data were conducive to grape sunburn.

The dates of the key phenological stages for the two studied parcels were reported in Table 3. The dates of those stages for the last 4 growing seasons for both parcels were reported in Annex 12.

Based on the Annex 12, the mid-bud burst dates were tardive as temperatures remained low during April and March, but the mid-flowering date was early due to high temperatures in May. A frost-freeze event occurred in April, but no late frost event was observed. The mid-ripening dates were also early, due to the high 10°C temperature sum, compared to the other years. As the mid-ripening stage was shifted, the berries started to change color earlier, and were therefore more sensitive to solar radiation, affecting their sunburn sensitivity.

Globally, both parcels' phenological stages were close to one another. The parcels can therefore be compared without their phenological stages affecting the results.

The EVI maps from Vineview (Annex 13 and Annex 14) show that the Jean Brun Ouest parcel is more vigorous than Les 4 Vents Sable. They also show that there are zones inside the parcels where the vigor isn't homogeneous with the rest. For example, in Jean Brun Ouest, the East part of the parcel is vigorous whereas the West part isn't.

The left of Les 4 Vents Sable parcel is an important non-vigorous zone. To make sure that our observations remain homogeneous, plots with higher EVI values were chosen, apart from this zone, so that it doesn't affect the study. To verify that the EVI differences between modalities were minor, a Fisher Test at a trust level of 0.95 was conducted using the LSD function. The results of this test were reported in Table 4 and Table 5.

Table 4: Results of the LSD Fisher Test on Enhanced Vegetation Indexes (EVI) of the different modalities in the JBO parcel

Table 5: Results of the LSD Fisher Test on Enhanced Vegetation Indexes (EVI) of the different modalities in the L4VS parcel

Based on the averaged EVI data per modality and on the Fisher Test results, the differences of EVI between modalities aren't significant (all modalities were classified in the same group).

Consequently, there are no significant differences in plant vigor between modalities that could potentially affect the results. Plant vigor is therefore homogeneous in the studied plots and won't impact the results of sunburn symptoms.

To make sure that the different modalities were studied in similar conditions, the canopy porosity values calculated from the photographs taken were reported in Table 6 and Table 7.

Table 6: Results of the LSD Fisher Test on Leaf Area Indexes (LAI) of the different modalities in the JBO parcel

Table 7: Results of the LSD Fisher Test on Leaf Area Indexes (LAI) of the different modalities in the L4VS parcel

Based on the results from the Fisher test, there aren't any significant differences between the average vegetation porosities on both parcels (all modalities were classified in the same group).

Consequently, it can be concluded that all the modalities inside both parcels are on average equally exposed to sun. Porosity won't affect the results on sunburn symptoms between modalities.

To determine the hydric state of the grapevine plants, two types of indicators were used to evaluate the potential impact of kaolin spraying and vine defoliation: water potentials, and leaf temperature.

To start with, the Predawn Leaf Water Potential measure was evaluated throughout the season, and the values were reported in Figure 13.

Because of the weather conditions of the 2022 vintage, the hydric constraint grew at the scale of the parcel from June to August. The JBO parcel has on average a higher water deficit than the L4VS parcel.

Both parcels were in stressing conditions by the end of the growing season. Consequently, it can be concluded that there were signs of water stress during the season, that could potentially influence the results.

Then, the Stem Water Potential was evaluated throughout the season as well, and the values were reported in Figure 14.

For every modality on both parcels, the water constraint was initially low and became moderate to high later in the season.

By comparing this variable between modalities, it shows that for both parcels the water constraint is on average lower for the kaolin modality, in stressing conditions. To verify this hypothesis, an LSD Fisher test was conducted, and its results were reported in Table 8 and Table 9.

Table 8: Results of the LSD Fisher Test on August the 11th Stem Water Potentials (SWP) of the different modalities in the

Table 9: Results of the LSD Fisher Test on August the 11^th Stem Water Potentials (SWP) of the different modalities in the

Based on the Fisher test results, the kaolin modality is significantly less stressed than the early defoliation modality for the L4VS parcel. However, based on the PLWP results, the JBO parcel was more stressed than the L4VS parcel, making it harder for kaolin to improve water use.

Consequently, it can be concluded that both water potential measures validated the hypothesis that spraying kaolin can slightly improve the grapevine hydric state up to a certain degree of stressing conditions.

The leaf temperature data was taken with an IR thermometer for each modality and the results were reported in Figure 15.

Figure 15: Leaf temperature per modality for both parcels, taken with an infrared thermometer, between July the 11th and

Figure 15 shows slight differences in leaf temperature between modalities. It can be observed that the kaolin modality is on average cooler by 2 to 3°C than the other modalities. However, based on the uncertainty bars on Figure 15, the temperature differences do not seem to be significant.

In order to verify this hypothesis, an ANOVA was conducted using an LSD Fisher test, and its results were reported in Table 10 and Table 11.

Table 10: Results of the LSD Fisher test on leaf temperature of the different modalities in the JBO parcel

Date	Kaolin	Early Defoliation	Control
July the 11th	43.62 #177; 0.54 B	45.35 #177; 0.54 A	45.91 #177; 0.54 A
July the 13th	36.50 #177; 0.75 B	38.53 #177; 0.75 AB	39.80 #177; 0.75 A
July the 26th	29.47 #177; 0.83 B	30.12 #177; 0.83 B	33.31 #177; 0.83 A
July the 28th	35.57 #177; 0.57 B	37.38 #177; 0.57 A	37.42 #177; 0.57 A
August the 9th	31.33 #177; 0.76 B	34.27 #177; 0.76 A	36.09 #177; 0.76 A

Table 11: Results of the LSD Fisher test on leaf temperature of the different modalities in the L4VS parcel

Date	Kaolin	Early Defoliation	Control
July the 11th	32.12 #177; 0.71 B	36.03 #177; 0.71 A	33.97 #177; 0.71 B
July the 13th	27.90 #177; 0.55 B	29.51 #177; 0.55 A	28.66 #177; 0.55 AB
July the 26th	23.54 #177; 0.49 A	24.87 #177; 0.49 A	24.18 #177; 0.49 A
July the 28th	27.33 #177; 0.75 B	30.25 #177; 0.75 A	32.14 #177; 0.75 A
August the 9th	39.03 #177; 0.51 B	41.61 #177; 0.51 A	41.93 #177; 0.51 A

Based on the ANOVA results, leaf temperature measurements showed a significative difference between modalities based on their group classification. Overall, it can be observed that the kaolin modality is significantly lower in leaf temperatures than the two other modalities for most measurements. Significative differences in terms of leaf temperature between the early defoliation and the control modalities were rarer, and both modalities were considered close.

Consequently, this does prove that kaolin reduces significantly leaf temperature, and validates the hypothesis that kaolin allows the leaves to slightly reduce their temperatures by reducing the stress of the plant, even by just a few degrees.

As seen earlier in this report (4.2.3), kaolin-spraying can influence bunches temperature, due to a higher reflection of the radiation. To verify if kaolin significantly reduces the bunch temperature, bunch temperature was measured during the growing season to conduct statistical comparisons between modalities. Moreover, those measures helped calibrate a bunch temperature model for the growing season.

In order to diminish the risks of error, the captors chosen to be placed in the plots were calibrated to make sure they wouldn't produce wrong values. The IR thermometer, the HOBO captors, and the Tinytag captors were all calibrated to reduce risks of errors.

To reduce the risks of error of berry and leaf temperature between modalities, all the measures were conducted with the same thermometer. For one given parcel, only one person was in charge to take all the measurements, to avoid the uncertainty linked to the observer.

Before implementing HOBO captors in the reference plots to acquire data, they were calibrated. Two additional captors were ordered in case some weren't working. The results of the temperature and light data were reported in Figure 16 and Figure 17.

Figure 16: Comparison of temperature data between 20 potential usable HOBO captors, on the 2^nd and 3^rd of June

Figure 17: Comparison of light data between 20 potential usable HOBO captors, on the 2nd and 3rd of June

Based on those graphs, results are overlaid, and the maximum difference in temperature between captors was 0.002°C, and 0.01 for the luminosity data. No captor showed values significantly different than the others, so the captors needed for the reference plots were randomly chosen.

Consequently, all differences between values under 0.002°C and 0.01 luminosity will not be considered as significant for interpretation.

Before starting the study, there were 10 TinyTags but the study only needed 8. This calibration helped to determine which ones were going to be used and which ones weren't.

The temperature and humidity data were plotted on Figure 18 and Figure 19 using an Excel pivot chart.

Figure 18: Comparison of temperature data between 10 potential usable TinyTag captors, on the 23^rd and 24^th of May

Figure 19: Comparison of relative humidity data between 10 potential usable TinyTag captors, on the 23^rd and 24^th of May

Based on the results, the values of the TinyTag number 7 were too different from the rest of the captors. Captor number 7 was therefore not included in the study.

When comparing all the other values, excluding captor number 7, a maximum difference of 0,458°C and 4,553%RH was observed between captors. Captor number 10 was also excluded, whose temperature values were slightly under the rest of the lot.

Therefore, all differences between values under 0,458°C and 4,553%RH will not be considered as significant for interpretation.

The parcels under-shelter temperature and humidity were measured using Tinytag captors at different positions. As both parcels results were similar, only the results from JBO were reported in Figure 20, and the rest of the results are available in Annex 15.

Figure 20: Tinytag captors temperature and humidity results on the JBO parcel between July the 17th and the 19th

Based on Figure 20, no significant differences were observed between captors in terms of temperature and humidity at the scale of the parcels.

Consequently, this verifies the hypothesis that the under-shelter temperature and humidity values are homogeneous and couldn't have affected the measured temperature differences between modalities.

The bunch temperature was punctually manually measured for each modality at different moments of the growing season. The infrared measured temperatures were compared at one precise moment for the three modalities during warm and sun-exposed days, and the results were reported in Figure 21.

Figure 21: Comparison of average bunch temperature per modality at different times of the day, on the sun-exposed side of
the canopy, taken by an infrared manual thermometer, between June the 13^th and August the 3rd

At the beginning of the measures, a slight difference in temperatures was observed between modalities. During the berries ripening, the same tendency was observed, but the differences stayed minimal.

The defoliation of the grapevine reduces its porosity. Consequently, the early defoliation modality should have higher bunch temperatures than the other modalities, because of the higher sun exposure. Based on the graph, the bunch temperatures differences do not seem significantly different between modalities.

In order to verify this hypothesis, the significance of the values was evaluated by conducting an ANOVA using the LSD Fisher test method. The results of the Fisher test can be found in Table 12 and Table 13.

Table 12: Results of the LSD Fisher test on bunch temperature of the different modalities in the JBO parcel

Date	Kaolin		Early Defoliation		Control
June the 13th	27.75	#177; 0.31 A	26.83	#177; 0.31 B	27.71	#177; 0.31 A
June the 15th	36.70	#177; 0.34 A	36.07	#177; 0.34 A	36.70	#177; 0.34 A
June the 16th	22.12	#177; 0.24 A	22.01	#177; 0.24 A	22.43	#177; 0.24 A
June the 17th	30.06	#177; 0.35 A	30.30	#177; 0.35 A	30.64	#177; 0.35 A
August the 8th	39.36	#177; 0.34 B	41.35	#177; 0.34 A	40.68	#177; 0.34 A

Table 13: Results of the LSD Fisher test on bunch temperature of the different modalities in the L4VS parcel

Date	Kaolin	Early Defoliation	Control
June the 13th	29.78 #177; 0.31 A	28.71 #177; 0.31 B	28.92 #177; 0.31 AB
June the 15th	32.4 #177; 0.39 B	34.05 #177; 0.39 A	34.24 #177; 0.39 A
June the 16th	20.63 #177; 0.11 A	20.85 #177; 0.11 A	20.83 #177; 0.11 A
June the 17th	25.78 #177; 0.26 A	25.80 #177; 0.26 A	25.66 #177; 0.26 A
August the 8th	34.63 #177; 0.44 B	35.99 #177; 0.44 A	35.67 #177; 0.44 A

The results show that the punctual bunch temperature measures weren't always significantly different between modalities as they were often classified in the same group. There aren't enough measures where the kaolin modality was significantly lower in bunch temperature than the other modalities to conclude that kaolin reduces significantly bunch temperature. When the measures are significantly different, the differences are very low (1 to 2°C).

Consequently, it doesn't verify the hypothesis that applying a kaolin-based particle film on grapes will reduce bunch temperature, due to the lack of measures. The result might have been significant if measures were done continuously for one day. To verify this hypothesis, the bunch temperature model measured temperatures all day long.

To obtain the precise bunch temperature model for each modality, the model was first calibrated, then applied to data during heatwave events in the season.

To produce a bunch temperature simulation, the bunch temperatures weren't measured during nighttime, and only the daytime temperatures differences will be taken into account. The model is based on the pre-véraison bunch temperatures.

To model the average bunch temperature for each modality at different moments of the day, the followed equation was synthetized by XLSTAT:

CTemp_HOBO: Coefficient Temp_HOBO given in the model's parameters CLight_HOBO: Coefficient Light_HOBO given in the model's parameters Temp_HOBO: Temperature of the bunch microclimate recorded by the HOBO captor Light_HOBO: Light of the bunch microclimate recorded by the HOBO captor

For every of the 6 produced models, the HOBO temperature and light variables of the bunch microclimate were considered in the model as being significative.

Figure 22 represents the result of the bunch temperature model for the control modality of the L4VS parcel. The other results of the significant linear regressions can be found in Annex 16 to Annex 20.

Figure 22: Multiple linear regression model from XLSTAT between the IR thermometer bunch temperature and the HOBO
captor recorded light and temperature data for the control modality in the Les 4 Vents Sable parcel

Figure 23: Evolution of bunch temperature on the JBO parcel between the 17th and the 19th of June 2022

Figure 24: Evolution of bunch temperature on the JBO parcel between the 12th and the 15th of July 2022

The bunch temperature models for each modality were applied to different periods during the 2022 summer where temperatures were high, and sunburn happened. The results for three heatwaves were graphically represented on Figure 23, Figure 24 and Figure 25. The results on both parcels were similar, so only the JBO parcel results will be analyzed in this part, and the L4VS results can be found in Annex 21.

Figure 25: Evolution of bunch temperature on the JBO parcel between the 10th and the 12th of August 2022

For the JBO parcel, the followed bunches of grapes were exposed on the South part of the canopy and reached a temperature peak between 2 and 4 PM. During the first heat wave (Figure 23), the bunches reached a maximum of 38,8°C for the kaolin modality, 43°C for the early defoliation modality, and 40°C for the control modality.

During the second heatwave (Figure 24), the bunches reached a maximum of 39,2°C for the kaolin modality, 45,8°C for the early defoliation modality, and 43,9°C for the control modality.

Lastly, during the August heatwave (Figure 24), the bunches reached a maximum of 40,08°C for the kaolin modality, 46,47°C for the early defoliation modality, and 42,89°C for the control modality.

Nevertheless, when comparing where the temperatures reached a plateau, it can be observed that the early defoliation modality and the control modality values are very similar, and only the kaolin modality is significantly cooler.

Overall, the differences between modalities were similar before and during véraison. The bunch temperature difference between the highest temperature of the control and the kaolin modalities is 1,2°C for the June heatwave, 4,7°C for the July heatwave, and 2,81°C for the August heatwave.

Those results verify the hypothesis that applying a kaolin-based particle film on grapes and leaves will reflect a small part of solar radiation, allowing the grapes to get cooler.

Sunburn symptoms were evaluated on both bunches of grapes and leaves. The measure of bunch sunburn symptoms will verify the influence of the modalities on berry sunburn, and therefore will help to define whether those methods can save grape yield and quality or not. Additionally, the leaf sunburn measure is an indicator of the sunburn pressure as well as the hydric state of the plant.

Due to grape thinning operations in July, two counts occurred. The results were reported in Table 14.

Table 14: Evolution of the bunch number per parcel before and after thinning operations

Based on the results, the final number of bunches per plant was reduced to obtain similar values. Consequently, the number of bunches per plant isn't significantly different between modalities.

The first evaluation of bunch sunburn symptoms took place in June, after a high-temperature week where the ambient temperature went up until 42°C. Other visual counts were performed after each high-temperature episode. The frequency as well as the intensity of bunch sunburn were represented for both parcels for each modality in Figure 26.

Figure 26: Evolution of bunch sunburn frequency and intensity on JBO and L4VS, from June the 22nd to August the 17^th

On Figure 26, the bunch sunburn symptoms seem to have decreased mid-July. This is due to thinning operations that took place on July the 5^th on JBO and the 8^th on L4VS. A lot of bunches affected with sunburn were removed, explaining the differences between the first and the second measure.

The frequency as well as the intensity of bunch sunburn rose for all the modalities of JBO and L4VS from the 8^th and 11^th of July until the 17^th of August. The control, early defoliation and kaolin modalities frequencies respectively rose by 12%, 1,9%, and 1,9% for JBO and 9,2%, 4,3%, and 3,1% for L4VS during this period.

From these measures, it could be concluded that overall, the kaolin and the early defoliation modalities had a noticeable effect on reducing bunch sunburn frequency.

In order to evaluate the losses linked with sunburn for each modality, the bunch damages (frequency*intensity) were calculated and represented on Figure 27.

Figure 27: Evolution of damages linked with sunburn on bunches per modality, from June the 22^nd to August the 17^th

According to Figure 27, yield loss (damage) estimation due to grape sunburn seemed to be significantly reduced by kaolin spraying and by early plant defoliation by the end of the season. By August the 17^th, losses due to sunburn were estimated to 0,69% on JBO and 2,25% on L4VS for the control modality, 0,52% on JBO and 0,51% on L4VS for the kaolin-sprayed modality, and 0,31% on JBO and 0,51% on L4VS for the defoliated modality.

As seen on Figure 27, the results are more significant on the L4VS parcel due to higher intensity and frequency. Only the L4VS parcel will therefore be analyzed.

On L4VS on the 27^th of June, early defoliation seemed to have reduced sunburn symptoms by 85%, while kaolin reduced symptoms by 67%. If the 11^th of July results are taken as a base zero due to thinning operations, new damages on the 20^th of July are evaluated to: 0,22%, 0,12% and 1,02% for respectively early defoliation, kaolin and control modalities. On the 20^th of July, early defoliation therefore reduced damages by 78%, while kaolin reduced them by 88%.

With the 11^th of July results as a base zero, damages on the 17^th of August are evaluated to: 0,3%, 0,22% and 2,14% for respectively early defoliation, kaolin and control modalities. On the 17^th of August, early defoliation reduced damages by 86%, while kaolin reduced them by 90%.

An ANOVA was conducted between modalities, to make sure that the damage differences were significant. The results were reported in Table 15 and Table 16.

Table 15: Results of the LSD Fisher test on berry sunburn damages of the different modalities in the JBO parcel

Table 16: Results of the LSD Fisher test on berry sunburn damages of the different modalities in the L4VS parcel

Date	Kaolin		Early Defoliation		Control
June the 27th	0.002	#177; 0.002 A	0.001	#177; 0.002 A	0.005	#177; 0.002 A
July the 11th	0.003	#177; 0.001 A	0.002	#177; 0.001 A	0.001	#177; 0.001 A
July the 20th	0.004	#177; 0.002 B	0.004	#177; 0.002 B	0.011	#177; 0.002 A
August the 17th	0.005	#177; 0.004 B	0.005	#177; 0.004 B	0.023	#177; 0.004 A

While a significant difference can be observed between the control modality and the others (they were classified in different groups), both the Kaolin and early defoliation modalities weren't classified as significantly different.

Consequently, this verifies the hypothesis that kaolin as well as early defoliation both have a positive impact on grape sunburn.

After each heatwave of the growing season, evaluations of leaf sunburn damages was performed additionally to the bunch sunburn evaluation. The evaluations took place between July the 8^th and August the 17^th. The frequency as well as the intensity of leaf sunburn were represented for both parcels for each modality in Figure 28.

Figure 28: Evolution of leaf sunburn frequency and intensity on JBO and L4VS, from June the 22nd to August the Xth

On Figure 28, the leaf sunburn symptoms seem to have globally increased between July and August. The frequency as well as the intensity of leaf sunburn rose up for all the modalities of JBO and L4VS from the 8^th and 11^th of July until the 17^th of August.

Overall, all the modalities seemed to all have been subject to intense leaf sunburn due to the high-temperatures and low rainfall episodes.

Figure 29: Evolution of damages linked with sunburn on leaves per modality, from June the 22^nd to August the 17^th

According to Figure 29, leaf damage estimation due to grape sunburn doesn't seem to be significantly different between modalities, even if small differences can be observed between modalities. By the end of the season, leaf sunburn damages were estimated to 26% on JBO and 16% on L4VS for the control modality, 22% on JBO and 10,5% on L4VS for the kaolin-sprayed modality, and 19% on JBO and 16% on L4VS for the defoliated modality.

An ANOVA was conducted between modalities, to verify if the damage differences on leaves were significant. The results were reported in Table 17 and Table 18.

Table 17: Results of the LSD Fisher test on leaf sunburn damages of the different modalities in the JBO parcel

Date		Kaolin	Early Defoliation		Control
July the 8th	0.004	#177; 0.004 A	0.008	#177; 0.004 A	0.014	#177; 0.004 A
July the 26th	0.180	#177; 0.042 A	0.137	#177; 0.042 A	0.201	#177; 0.042 A
August the 17th	0.217	#177; 0.032 A	0.190	#177; 0.032 A	0.257	#177; 0.032 A

Table 18: Results of the LSD Fisher test on leaf sunburn damages of the different modalities in the L4VS parcel

Date		Kaolin	Early Defoliation		Control
July the 8th	0.035	#177; 0.008 AB	0.053	#177; 0.008 A	0.023	#177; 0.008 B
July the 26th	0.044	#177; 0.016 A	0.074	#177; 0.016 A	0.083	#177; 0.016 A
August the 17th	0.105	#177; 0.019 A	0.160	#177; 0.019 A	0.157	#177; 0.019 A

According to the ANOVA, no significant difference (they were all classified in the same group) can be observed between the three modalities. Therefore, leaf sunburn symptoms were overall the same for each modality.

Consequently, kaolin and early defoliation do not have a significant impact on leaf sunburn symptoms. And while kaolin reduces leaf temperature and berry sunburn, it does not significantly cause higher leaf dehydration.

As a result to both measures, even if all the modalities faced different bunch sunburn damages, their leaf sunburn symptoms were similar. This can be interpretated the fact that in the same stressful environment where sunburn should thrive, the kaolin and early defoliation modalities are efficient to protect the berries.

For each modality, the mass and volume of 100 berries were measured from Véraison, to evaluate any potential difference. The results from both parcels being similar, only L4VS results were reported in Figure 30. The JBO results are available in Annex 22.

Figure 30: Evolution of mass and volume of 100 berries between July the 21st and August the 22nd, for the L4VS parcel

Based on Figure 30, the berries' size and volume seem to be different between modalities by the end of the season. The peak of volume hasn't been reached yet, as berry maturity usually stops later in September.

Indeed, the kaolin modality seems to present bigger berry sizes and volumes. Theoretically, the use of kaolin should lower the berries temperature during hot days, reducing the berries transpiration resulting in a berry volume rise of 7%, as it has been studied on Sauvignon Blanc and Merlot (Coniberti et al., 2013; Shellie and Glenn, 2008).

Consequently, the mass and volume measures verify the hypothesis that berry physiology wasn't negatively impacted by kaolin treatments and early canopy defoliation. On the contrary, kaolin treatments significantly increased the berries' size and volume.

Primary and secondary metabolites were measured during the berry maturation on both studied parcels for each modality. The results of those measures and analyses were reported in graphs. The analysis results from the external laboratory (Excell) are available in Annex 23Annex 23.

As both parcels had similar results in terms primary metabolites, only L4VS results were analyzed in this part, and results from the JBO parcel were reported in Annex 24.

Figure 31: Analysis results of berry maturity per modality, between July the 21st and August the 22^nd, for the L4VS parcel

As it can be observed on Figure 31, by the 22^nd of August, the modalities levels of pH, sugar, malic acid, and total acidity are similar.

On the 29^th of July, it can be observed that the early defoliation modality seemed more advanced by the others, due to its higher values in pH, and lower acidity. However, no significant difference can be observed by the end of the maturation process.

Consequently, this verifies the hypothesis that kaolin sprayings and early defoliation do not impact berry maturity, and therefore wine quality.

The total phenolic compounds were evaluated using the Folin Index. The polyphenols and anthocyanins levels were measured and the TPI was calculated. The results were reported in Figure 32.

Figure 32: External laboratory analysis results of anthocyanins and phenolic compounds per modality, between August the

According to the results, neither the total phenolic compounds nor the total anthocyanin concentration significantly differ between kaolin and control modalities, meaning that kaolin spraying did not have a significant effect on both compounds. However, early defoliation levels seem slightly higher than other modalities for both metabolites, but the difference isn't significant.

Globally, it can be concluded that the measured physico-chemical characteristics on the different modalities weren't significantly influenced by neither the kaolin spraying, nor the early defoliation of the canopy. Consequently, both potential solutions against grape sunburn won't affect the quality of the harvested berries, and therefore the quality of the produced wines.

However, it has been observed that early defoliation increases polyphenol and anthocyanins production in berry skin, validating the hypothesis that exposing plants to solar radiation sooner helps forming a stronger protective skin that will be able to resist sunburn, without affecting grape quality.

Assumed from the results of this study, grape sunburn adaptation comes with important management changes and implications. Including sunburn preventive solutions into the vineyard work calendar represent additional workload that needs to be managed. For example, implementing kaolin sprayings and early defoliation imply to free some time, and reduce time allocated with other types of tasks.

Thanks to the interview conducted with the vineyard manager (Annex 2), early defoliation could be included in the existing work calendar, if it is done early and at the good phenological stage, to make sure that the bunches have enough time to adapt to sun exposure. Kaolin sprayings could also be managed if it is proven that it doesn't affect the organoleptic parameters of the wines. However, both solutions will necessitate additional work as they cannot be merged with other existing tasks. For example, kaolin powder cannot be implemented inside their regular vineyard treatments they already conduct, as it might affect the application process.

In terms of expertise, processes, and workforce, the limits that Château Margaux will face to implement those solutions are: workforce availability, machinery availability, weather conditions, and workload conducted at the same time.

In terms of managerial adaptation, this means that the company will be most likely to hire supplementary working force to complete this job.

Overall, both solutions will induce management changes, will come with challenges, and will be time and cost-bearing.

The study took place during the 2022 growing season. The conditions, results, as well as the limits of the study were reminded in order to conclude on the efficiency of the studied potential solutions against grape sunburn.

Regarding the conditions of the study, weather conditions as well as the parcels homogeneity were two main factors that have impacted the results, and that differ between vintages.

Weather conditions can greatly influence this study. For example, this study was already conducted in 2021 with three modalities: control, kaolin spraying on bunches and leaves, and kaolin spraying on bunches only. However, the 2021 weather conditions weren't keen to sunburn development, as it rained a lot, and there wasn't any heatwave.

Consequently, it was harder to decipher a significant difference between modalities, as they weren't much affected by sunburn, and it was harder to define a good date to apply kaolin on the canopy. It is however important to conduct this study on different types of weather conditions, so that it can be adapted and improved.

The results found in this study on weather conditions confirm the bibliography results linking higher sunburn intensity with higher light and temperature conditions (Schrader et al., 2009; Gambetta et al., 2021; Araújo et al., 2018). It also confirms the link between temperatures higher than 30°C and sunburn (Pastore et al., 2013), based on the results from the last 5 growing seasons.

Rain can also greatly affect this study on the kaolin modality, as the kaolin clay can be washed away by rain.

High temperatures such as the ones faced during the 2022 growing season were perfect for this study, as they surely caused sunburn on both parcels. However, the season was dry, and caused water deficit, intensifying the sunburn symptoms, even on the kaolin and early defoliation modalities, as it was found in bibliography (Cook et al., 1964; Gambetta et al., 2021). Additionally, a climate like the 2022 vintage reflects the potential climatic conditions the Médoc vineyard will face in the future due to climate change.

For the 2022 study, as the rain episodes were rare between June and August, the kaolin sprayings lasted for long periods of days. Visually, both the leaves and bunches were covered by kaolin on both sides of the canopy, and the defoliation allowed the bunches to be significantly more sun exposed than the control modality.

The results of porosity and EVI clearly show that the modalities were homogeneous in terms of plant physiology, reducing uncertainties limiting biases.

The EVI values between modalities weren't significantly different, which reflects a global homogeneity in terms of plant vigor for both studied parcels.

However, those results do not consider the zones of the parcels where the vigor was low, and only considered the studied plants. It can therefore be imagined that significantly different EVI values could have been observed if the studied plots were chosen differently.

No link between vigor and sunburn were found, as opposed to what was found in other studies (Smart, 1985), due to the homogeneity of the studied plots.

Regarding the vegetation porosity values, the Leaf Area Indexes (LAI) between modalities weren't significantly different either. The studied plots were consequently equally exposed to sun, limiting biases linked with sun exposure, as it was found in bibliography results (Southey and Jooste, 1991).

It can be said that homogeneity between modalities was hence controlled based on the porosity and EVI values, limiting biases linked with plant growth.

Overall, the results of the study were conclusive and showed that both kaolin sprayings and early defoliations had an impact on grape sunburn. Both solutions being however different, they affected grapevine physiology as well as berries quality differently.

In this part, the effects of both studied solutions against grape sunburn on grapevine physiology will be discussed. More precisely, their consequences on the plants' hydric state and sunburn symptoms will be resumed.

Based on the measures of both water potentials and leaves temperature, the kaolin sprayings have significantly improved grapevine hydric state, by reducing its temperature. The results found confirm the bibliography results, showing that kaolin can diminish hydric stress (Glenn and Puterka, 2010; Glenn et al., 2010). Kaolin sprayings can therefore be considered as a good solution against grapevine hydric stress. However, early defoliation only seemed to have increased water stress, by increasing canopy porosity due to removed leaves.

In this study, kaolin reduced leaf temperature between 2 and 3°C on average, which is low, as the bibliography results predicted a 3 to 8°C reduction (Agrisynergie, 2022).

Consequently, it can be concluded that kaolin was the best modality out of the two studied solutions to reduce water stress and improve grapevine physiology.

The bunch temperature results showed significantly differences between modalities, contrarily to leaf temperature. However, when the bunch temperature model was applied to high temperatures episodes, differences in temperature peaks were observed between modalities. The kaolin modality was the coolest, while the early defoliation modality was the hottest.

As grape sunburn is mostly a result of high solar radiation and temperatures, it could have been expected that the early defoliation modality had more sunburn symptoms. However, based on the sunburn evaluation results, the control modality was significantly more affected than the two other modalities. The results helped validate the hypothesis that both kaolin and early defoliation can reduce grape sunburn symptoms on bunches by acting differently on grapevine physiology.

According to studies, kaolin sprayings should increase light reflectance and reduce the plant's sensitivity to sunburn (Yazici and Kaynak, 2009; Lobos et al., 2015). On the other hand, early defoliation should increase sunlight exposure, resulting in higher skin thickness and lower sunburn sensitivity (Pastore et al., 2013; Solovchenko, 2010).

The results found in this study confirm both hypotheses, as damages linked with sunburn were significantly reduced by both kaolin sprayings and early defoliation.

In order to verify the potential impact of the modalities on berry quality, some analyses were conducted, and results were analyzed.

Based on the results, kaolin sprayings had a positive effect on grapevine physiology. Indeed, kaolin increased berry size and volume compared to other modalities.

The early defoliation modality, due to a higher and sooner berry sun exposure, presented a slightly higher anthocyanin and polyphenols concentration than the other modalities. However, the difference between modalities being unsignificant, other factors might have helped reduce sunburn sensitivity for this

modality. For example, the aeration of the canopy linked with defoliation could have potentially helped reduce bunch temperature and therefore sunburn.

When looking at other primary and secondary metabolites, no significant difference can be observed between the three modalities. It can therefore be imagined that based on those results, neither kaolin nor early defoliation have negatively impacted berry quality. However, no conclusion can be taken on the final wine quality, as other parameters are to be taken into account, such as unmeasurable compounds and berry taste, that can only be evaluated by conducting tastings.

As a conclusion on the study results, it can be said that both methods worked differently on grapevine physiology but were both efficient on grape sunburn. While kaolin proved to improve the plant's hydric condition and reduce bunch temperature, early defoliation built up the berries' tolerance to sun exposure.

Both studied solutions at the scale of the vineyard can present some technical challenges linked with their application.

Kaolin sprayings are made using a kaolin clay powder mixed with water and adjuvant. When preparing the mix, the solution had lumps that were hard to incorporate. Homogeneity was reached eventually after adding kaolin powder gradually, to avoid lump formation. To incorporate the powder into the solution, it had to be mixed longer. Consequently, the adjuvant caused the appearance of bubbles. For a prepared volume of 200 liters, an estimated additional volume of 150 liters of bubbles was formed, reducing the part of the product that could have been used. The bubbles formation also made it harder to see the level when filling the sprayer tank.

The sprayer nozzles weren't clogged by the solution, as kaolin was rarely used. However, when applied at the vineyard scale, there is a risk that the nozzles get clogged due to the thick and abrasive property of the mix. This can cause a technical problem as the solution could be to change the nozzle size between regular and kaolin treatments, being time consuming.

Regarding early defoliation, no specific technical problem was observed. However, it is important to define the degree of defoliation to be reached, or it can result in drastic bunch exposure leading to more sunburn symptoms. The defoliation needs to be moderate to only increase the exposure enough to strengthen the berries. Defoliation also needs to be done at a specific phenological stage to ensure higher sun exposure for newly formed berries. If defoliation is completed too late in the season, it can result in sudden berry sun exposure, leading to higher sunburn symptoms.

Another technical limit linked with early defoliation is the row orientation of the parcel. As the vineyard is composed by differently oriented parcels, it makes it harder to define the defoliation intensity for each parcel.

There is also an uncertainty linked with the worker, leading to more or less leaves left on plants.

Picking berries for quality evaluation was done by different samplers, resulting in a possible bias in results. A solution to reduce this bias could be to only assign one person to sample the integrity of the modalities. However, this solution is time consuming, especially at the scale of two parcels, making it almost impossible to implement.

To limit this bias as much as possible, the company is implementing a specific picking protocol, so that gaps between samplers are minimized.




^{Costs linked with kaolin sprayings}
^{Dose kaolin (kg/ha)}	⁸⁰
^{Price of kaolin per kilo (€/kg)}	^1,6
^{Price of kaolin per hectare (€/ha)}	¹²⁸
^{Spraying time for 0,3 ha (h/0,3)}
^{Cost of sprayer driver per hour (€/h)}
^{Cost of kaolin per hectare (€/ha)}

^{Profits linked with kaolin sprayings}
^{Number of plants per hectare (ha)}	¹⁰⁰⁰⁰
^{Number of bunches per grapevine plant}	^5,35
^{Average weight of a Cabernet Sauvignon bunch (kg)}	^0,15
^{Estimated yield (kg/ha)}	⁸⁰²⁵
^{Damage difference between control and kaolin} (%)	^0,20%
^{Yield gain (kg/ha)}	^16,05
^{Berry mass per bottle of wine (kg/0,75L wine)}	¹
^{Berry mass per litter of wine (kg/L)}	^1,33







^{Profits linked with early defoliation}
^{Number of plants per hectare (ha)}		¹⁰⁰⁰⁰
^{Number of bunches per grapevine plant}		^5,35
^{Average weight of a Cabernet Sauvignon bunch (kg)}		^0,15
^{Estimated yield (kg/ha)}		⁸⁰²⁵
^{Damage difference between control and defoliation} (%)		^0,40%
^{Yield gain (kg/ha)}
^{Berry mass per bottle of wine (kg/0,75L wine)}
^{Berry mass per litter of wine (kg/L)}

In terms of management, the feasibility of both solutions was evaluated to make sure that it could be implemented at the scale of the vineyard.

Overall, kaolin sprayings management at the scale of the study wasn't time consuming. The mix preparation took about 30 minutes for a 200 liters volume. However, if kaolin sprayings are applied at the scale of the vineyard, they will represent additional work and will be harder to manage. The first management limit will concern the use of the sprayers, as they are also used at the scale of the vineyard to perform other tasks such as tillage or canopy trimming during the growing period. Managing how and when the sprayers are going to be used might be time consuming but is necessary to ensure the efficiency of the vineyard work management.

The dose and dates of treatments will also have to be reasoned based on weather conditions, canopy coverage, and workload. It might also represent additional work and time to make sure the conditions are optimal to spray.

Finally, as kaolin must be evenly sprayed on both sides of the canopy, it requires two sprayer passages, increasing the treatment time by 2.

Regarding early defoliation at the scale of the vineyard, the work itself do not represent a management problem. However, it will require hiring seasonal workers to defoliate but will also require the presence of employees to oversee their work.

Overall, for both kaolin sprayings and early defoliation methods, time and workload management represent the main limits in terms of project feasibility.

In order to make sure that using kaolin and early defoliation methods on the vineyard are worth it in terms of loss reduction, an economic study was conducted. The objective of this study was to define if the cost of kaolin and early defoliation would overcome the economic gains linked to sunburn reduction.

The gains were estimated based on the calculated yield difference between the control modality and the other two modalities. The yield gain was then converted in number of produced bottles. As both studied parcels produce Pavillon Rouge wine, the average price of a bottle was calculated based on their selling price.

The results of this study can be found in Table 19 and Table 20, where costs and gains of both methods were reported, based on the latest damage results on both studied parcels.

Table 19: Estimated economic margin made from both studied modalities based on the results on the JBO parcel

Table 20: Estimated economic margin made from both studied modalities based on the results on the L4VS parcel

^{Costs linked with kaolin sprayings} ^{Spraying time for 0,3 ha (h/0,3)} ^{Profits linked with kaolin sprayings} ^{Damage difference between control and kaolin} (%) ^{Yield gain (kg/ha)} ^{Margin (€/ha)}
^{Costs linked with early defoliation} ^{Profits linked with early defoliation} ^{Damage difference between control and defoliation} (%) ^{Yield gain (kg/ha)} ^{Margin (€/ha)}

¹⁰⁰⁰⁰ ^5,35 ^0,15 ⁸⁰²⁵ ^1,80% ^144,45 ¹ ^1,33 ^108,61 ¹²⁵ ^13576,13

According to both tables, the gains linked with sunburn reduction overcome the costs of both studied

methods. Early defoliation seems to be more expansive than kaolin, therefore less profitable for the

Overall, the results show that both methods are profitable when applied at a larger scale. Even if the

damages results on the JBO parcel weren't significant, kaolin sprayings still manage to help gain

However, those results can highly vary based on the vintage weather conditions. Indeed, if the damage

difference between modalities is lower, it could potentially induce economic losses. Moreover, due to

some parcels' orientation, both methods might not work at the scale of the entire vineyard.

Consequently, those results need to be pondered as they highly depend on row orientation and could

vary based on which parcel is studied, the size of the sprayer used, and the planting density.

This study highlighted the effects on sunburn symptoms of two short-term solutions. However,

due to the limited duration of the study, it hasn't been possible to define the impacts of both kaolin and

In order to have been more precise on the organoleptic consequences of kaolin and early defoliation on

Château Margaux's wines, it would have necessitated to continue to conduct this study longer.

If the study had lasted longer, we would have been able to define whether those solutions can be applied

at a large scale in the vineyard, without affecting the typical profile of Château Margaux's wines or not.

conduct Berry tastings could also have been implemented to define the impact of kaolin and defoliation

However, in continuity to this study, the company will vinify the modalities separately, so that they will

be able to conduct tastings to compare the effects of the studied modality on wine organoleptic profile.

Additionally, the study size was limited. The studied plots represented 30 plants per modality per parcel,

which hardly represents the integrity of the parcels. There is consequently an additional uncertainty

Based on the study's results and limits, propositions have been formulated. For each proposition, the objectives, means, costs and risks have been discussed.

To overcome the study duration limit, it can be advised to continue the study on the 2022 vintage. The primary and secondary metabolites analysis will continue until harvest to make sure that no significant differences are observed.

It can also be advised to harvest and vinify separately the modalities of each parcel to produce different wines. When they will be ready, the wines could later be blind tasted and compared to one another to define if there is a significant organoleptic difference between the three studied modalities. If there aren't, then it means that both solutions can successfully be applied to the integrity of the vineyard, without affecting the typicity of Château Margaux's wines.

The 2022 study wasn't the first conducted on grape sunburn by Château Margaux but is part of a longterm project to reduce grape sunburn at the scale of the vineyard. It can therefore be advise to prolongate the study to the 2023 and 2024 vintages in order to make sure that kaolin and early defoliation work on different types of climates.

To continue and perfect the study on grape sunburn, it can be advised to reconduct the study for the next growing seasons, by improving the experimental set-up. The objective of this proposition is to increase the study precision and improve the knowledge and management capacities linked with sunburn solutions.

To do so, the study could be applied at a larger scale of the vineyard, on different grape varieties and parcels orientations. By also choosing more reference plants per modality to study, it will increase the representativity of the results at the scale of the vineyard.

Moreover, the intensity of both methods can be modulated, and trials can be conducted. For example, different defoliation intensities could be tested as different modalities, based on the parcels' orientations. For kaolin sprayings, it could be advised to test different concentrations based on weather conditions. By doing so for a few years, the applied dose can be modulated so that time management becomes optimal.

The cost of this proposition will be close to this year's study cost, but the investment return will be larger as it should help at a long-term scale to significantly reduce/avoid berry sunburn.

The only risk linked with this proposition is the absence of sunburn linked with bad weather conditions, and therefore the inability to evaluate the efficiency of the studied methods.

Based on the positive results of the studied, it can be proposed to extend the solutions at the scale of the vineyard. While early defoliation results mainly depend on row orientation, kaolin sprayings can be easily applied at the scale of the vineyard independently from row orientation.

The objective of this proposition is to reduce sunburn at the scale of the vineyard, and reduce economic losses linked with this issue.

To implement this proposition, visual observations will have to be done to define which parcels need to be treated with kaolin. Based on the results, a kaolin mix will be prepared and sprayed depending on weather conditions.

The cost of this proposition will be estimated at 13 614,02€/ha, what kaolin sprayings cost for this study multiplied by the number of hectares of the vineyard (94ha). Regarding the investment return of this proposition, it can be estimated as the return obtained with the kaolin sunburn study multiplied by the number of hectares treated (= at least 13 614,02€/ha).

However, the risk of this proposition is that kaolin sprayings might not work as well on the integrity of the vineyard as they worked on the two studied Cabernet Sauvignon parcels. It can be imagined that other grape varieties are less sensitive to sunburn, and that the investment return will be less important,

if not negative. In order to overcome this risk, it can be advised to first continue and improve the experimental set-up at a larger scale before extending it at the scale of the vineyard.

To overcome technical limits linked with kaolin sprayings and early defoliation, technical resources propositions have been formulated.

For kaolin sprayings technical limits, filters can be included in the sprayers to limit the risk of nozzle clogging. Rather than having to change nozzles before each treatment, implementing filters is less time consuming, as they can be left for regular treatments too.

Regarding the adjuvant bubbling property, it is linked with air incorporation during the mixing process. To overcome this issue, it can be advised to reduce the mixing speed and to incorporate kaolin into the mix slower and in lower quantities.

To overcome technical problems linked with early defoliation at the scale of the vineyard, two propositions can be made. The first one being to conduct a trial on defoliation based on row orientation, in order to link ideal defoliation intensity with sun exposure. The second proposition is to provide the correct formation for seasonal workers so that they understand the degree of defoliation that is supposed to be reached, reducing problems linked with intense defoliation.

Managing time and workload is essential to continue this study or apply it at a larger scale. As seen previously (7), kaolin sprayings as well as early defoliation solutions need to be correctly managed.

Based on the vineyard workload, sprayers and workforce might not be available to perform kaolin sprayings during the growing season. Consequently, two solutions exist. Either the vineyard manager can decide to reduce time allocated with other tasks such as tillage or canopy trimming, and prioritize kaolin sprayings in the work calendar, either kaolin sprayings can be performed by external workers. While the first solution seems more complicated in terms of time management, it is less costly as it doesn't necessitate to spend additional money by hiring workforce and lending material. In comparison, there are no direct economical losses linked with tillage and canopy trimming, even if they are essential to perform, while economical losses linked with sunburn are significative.

To manage the dose and dates of treatments, different solutions can be implemented. First, the weather conditions of the season will need to be followed. If long high-temperature episodes are predicted, a kaolin treatment must be done in prevention. Then, observations can be performed at the scale of the vineyard based on weather conditions, to verify if kaolin has been washed out of the canopy depending on rainfall events. The dose of treatment will also depend on the weather conditions. If high temperature periods are predicted before a rain episode, then the dose could be divided by 2, as it is expected to be washed out and renewed after rain. Kaolin treatment management should therefore be as precise as possible, to ensure higher protection of the vineyard.

Finally, managing early defoliation will necessitate hiring and overseeing seasonal workers, representing additional money and work. However, vineyard employees are usually already busy completing other tasks, and do not have time to oversee early defoliation. Like for kaolin sprayings, solutions include reducing time allocated with other tasks to free some time for defoliating or hire external workers to complete their tasks while they oversee early defoliation.

To continue the study on grape sunburn, other types of experiments could be conducted such as evaluating the consequence of different inter-row cover crops on grape sunburn.

Cover crops can absorb an important part of solar radiation (Varlet-Grancher et al., 1989). Consequently, when implemented in the vineyard's inter-rows, it can reduce part of the reflected solar radiation, diminishing direct reflection on bunches of grapes.

To do so, different types of cover crops should be implementing at the scale of the vineyard, and measures should be conducted in order to evaluate their reflection degree, and therefore their impact on sunburn.

In terms of budget, this trial will have costs linked with the purchase of specific seed mixes, soil preparation and cultivation, as well as tools used to evaluate reflected radiation. However, if the study significantly diminishes canopy temperature by finding the ideal inter-row cover crops, the investment return could be important.

Nevertheless, this solution only works on specific types of soils. If the soil type is lighter than the cover crop, then it is useful to implement it, as the nude soil's solar radiation reflection will increase radiation received by the canopy.

Moreover, this solution has limits. Implementing cover crops could potentially result in higher competition between the grapevine and the crops in terms of soil nutrients.

Climate change will bring new challenges to wine production. The 2022 vintage is a good example of this hypothesis, as long heatwave episodes caused severe drought and grape sunburn, affecting both the yield and quality of the production.

In order to overcome the future effects of climate change, institutional changes seem necessary. For instance, the evolution of the appellations' system to include technical and agronomical innovations is inevitable to limit losses and quality changes linked with climate change. It can therefore be imagined that the PDO specifications as well as the national legislation will evolve to allow for instance shade netting and vineyard irrigation up to a certain level at the scale of the Médoc vineyard. Insurance systems for environmental induced losses (such as frost and hail) could also be implemented to strengthen the vineyards capacity to face climate change.

However, if the PDO specifications do not evolve in the next few years, it can be advised that Château Margaux should stop being part of the PDO and start using alternative methods to limit the consequences of climate change without affecting their wine typicity. Thanks to their renown, Château Margaux won't necessarily be affected in terms of sales and profits by their declassification, due to their already existing notoriety.

The use of short-term solutions against grape sunburn within a context of climate change in the Médoc vineyard

5. Problem and hypotheses