21st GiESCO International Meeting: ‘A Multidisciplinary Vision towards Sustainable Viticulture’
EXPLORING THE FACTORS AFFECTING SPATIO‐TEMPORAL VARIATION IN GRAPEVINE
POWDERY MILDEW
Romane MELYON‐DELAGE1, Benjamin BOIS2, Sébastien ZITO2, Mario REGA2, Guillaume GARIN1, Amelia CAFFARRA1 1 2 itk, Cap Alpha, Avenue de l’Europe, 34830 Clapiers, France CRC, Biogéoscience, Université de Bourgogne Franche Comté, 2 Bld Gabriel, 21000 Dijon, France *Corresponding author:amelia.caffarra@itk.fr
Abstract: Context and purpose of the study ‐ The spatial distribution of powdery mildew is often heterogeneous between neighboring plots, with higher disease pressure in certain places that can be considered as disease “hotspots”. The position of hotspots can vary over the years, even if some plots consistently present a higher vulnerability over time. This spatio‐temporal heterogeneity makes it difficult to obtain accurate prediction by epidemiological models that are fed by meteorological variables coming from weather stations that are not in close proximity to the vineyards or are spatialized over large cell grids. The aim of the project was to explore the role of environmental/agronomic factors involved in powdery mildew pressure variation in time and at the plot and regional scale. Material and methods ‐ To do so, a series of variables were monitored in 10 different vineyards, in the Beaune region (Bourgogne) of France, over two years. These factors included agronomic descriptors and weather variables. Weather data were acquired at the plot scale. Disease symptoms were observed weekly on leaves and grapes, highlighting inter‐plot variation in disease indicators. Results ‐ The factors that most impacted this variability were relative humidity, rain, leaf wetness, vigor and phenology. A more in‐depth study of the interactions between these factors will help to disentangle the complex effects of the environment on powdery mildew inter‐plot heterogeneity. Relative humidity and leaf wetness appeared as the most closely correlated variables to powdery mildew onset and pressure. Undoubtedly these results need to be further confirmed and quantified through more extended surveys, but they indicate interesting directions for the improvement of predictive models of powdery mildew
Keywords: Correlation, Heterogeneity, Humidity, Leaf Wetness, local effects
1.Introduction
Powdery mildew (caused by Erysiphe necator) is one of the most common and harmful diseases of
grapevine. If its optimal conditions are met, a new generation of this disease can be completed in less
than a week (Wilcox, 2003). These explosive germinations combined with the fact that symptoms are
visible only once the disease is well established, make its control difficult (Pearson &Gadoury, 1987). In
addition, the spatial distribution of the disease is often heterogeneous between neighboring plots, with
higher disease pressure in certain places that can be considered as disease “hotspots” (Calonnec et al,
2009; Mammeri et al, 2014). This spatio‐temporal heterogeneity makes it difficult to obtain accurate
prediction by epidemiological models that are fed by meteorological variables coming from weather
stations that are not in close proximity to the vineyards or are spatialized over large cell grids. Indeed,
local disease pressure might be related to micro‐climatic factors but could also be the result of local
environmental/agronomic variables that are rarely considered by models (Delp, 1954; Sall, 1980; Caffi et
al, 2011), such as the type of soil, agricultural practices and the interaction with the plant: e.g. the
growth, the physiology and genetic specificities of the variety.
To date, the effects of local factors on disease pressure have not been analyzed in depth, and the
triggers of powdery mildew hotspots have not been clearly identified. However, an exploration of these
effects is of the foremost importance to improve the understanding of the disease and to support the
development of predictive models of powdery mildew. Therefore, the general aim of this study was to
explore the role of the factors involved in powdery mildew pressure variation in time and space at a
June 23 - 28, 2019 | Thessaloniki | Greece
GiESCO Thessaloniki | 513
21st GiESCO International Meeting: ‘A Multidisciplinary Vision towards Sustainable Viticulture’
local scale. To do so, we monitored a series of environmental and agronomic variables, including
powdery mildew observations, in 10 different vineyard plots near Beaune (Burgundy), over the growing
seasons of 2017 and 2018.
2.Methods
Statistical correlation and multivariate analyses were conducted to explore the relationships between
environmental variables and disease pressure. This paper presents the description of the data collected
in 2018 (for more details on 2017, see Weens, 2017) and the multivariate and correlation analyses
conducted on the joined 2017‐2018 datasets.Observations were conducted on 10 plots in the Côte d'Or,
all located in the same area near Beaune so that local effects could become evident. The plots were not
treated to avoid the confounding effects of pesticides on disease pressure. Each plot was characterized
with its soil type, its variety, its training method, its cover crop, whether it was in the plain or on the hills
and its orientation. Disease pressure monitoring was carried out in the field through visual notations
from April to July. At the same time, meteorological factors and plant‐related factors such as vigor and
phenological stages were measured. These variables were suspected to contribute to disease
heterogeneity, as suggested by previous literature. They can be divided into:
1.
2.
3.
Meteorological variables, either measured at reference stations at 2 m above soil on each plot,
and also in the canopy (micro‐climate) on three specific plots (CRC_20, CRC_21, CRC_28). They
included temperature, relative humidity, rain, leaf wetness (only for the three micro‐climate
stations in the canopy).
Agronomic (plant and “situation”) variables: phenology, plant vigor (measured as weekly
growth in twig length and width), variety, training method, cover crop, plain or hills, orientation
year (2017 or 2018).
Disease variables: percentage of leaves and grapes affected by powdery mildew.
The data was used to calculate indicators (see below) and organized to compute correlation matrices.
The correlation coefficients (Pearson) obtained in this way were used to select a set of “meaningful
indicators” highly correlated to disease occurrence that were then used in a principal component
analysis (PCA). In order to perform these analyses, static descriptive indicators of the plots (such as
grape variety) or synthesis indicators for dynamic variables such as temperature were used.The
indicators in the statistical analysis were the following:
1.
2.
-
Meteorological indicators: Average daily temperature, minimum and maximum temperature,
relative humidity and leaf wetness. These variables were averaged either over the month of
May (typical time of the year for primary contaminations) or weekly with a specific focus on the
th
st
nd
th
three weeks before the week of the first symptoms: 15 – 21 of May (week 1), 22 – 28 of
th
th
May (week 2) and 29 of May – 4 of June (week 3).
Agronomic indicators:
Plant related: Plant vigor (average weekly growth), date of phenological phases
Disease related: First date of symptom onset, final percentage of affected leaves or grapes (on
the 4th of July)
Plot related: Variety, training method, cover crop, plain or hills, orientation
Year
3.Results and discussion
Unlike 2017 that had low powdery mildew pressure, year 2018 was characterized by a moderate
powdery mildew pressure. The disease developed relatively late (first symptoms between the end of
May and the beginning of June over the region). For the plots, disease pressure ranged between weak
and moderate/strong with a maximum of 30% symptoms in July before the beginning of phytosanitary
treatments (Figure 1). The date of first observable symptoms was clearly heterogeneous, varying
between the 4 and 23 of June (CRC_22 vs CRC_28 on grapes). Dynamics of symptom progression also
varied, with some plots showing a faster rate of disease increase than others. Interestingly, whereas
final disease pressure was higher in 2018 than 2017, the date of first symptoms (first week of June) was
similar for the two years under study.
The correlation analysis highlights a strong year effect, and stronger correlations of powdery mildew
with meteorological than with agronomic indicators. Indeed, disease biology in general is known to be
strongly affected by weather variables like temperature, humidity and leaf wetness, which also have a
June 23 - 28, 2019 | Thessaloniki | Greece
GiESCO Thessaloniki | 514
21st GiESCO International Meeting: ‘A Multidisciplinary Vision towards Sustainable Viticulture’
large inter‐annual variability (Rapilly, 1991). In the case of powdery mildew, temperature hastens
germination and promotes spore production, and relative humidity and leaf wetness promotes
infections (Delp, 1954; Chellemi&Marois, 1991a; Chellemi&Marois, 1991b).
These effects were confirmed by the correlation coefficients and by the PCAs (Fig 2 and 3) showing
positive correlations between temperature and humidity and disease indicators. Almost all indicators of
temperature were positively correlated to disease onset on grapes, but not on leaves, implying that
warmer temperatures might correspond to later grape infection. This might seem counter‐intuitive,
given the promoting effect of temperature on spore release (Chellemi&Marois, 1991b). However, this
effect might not be directly due to temperature (which over the study period always stayed in a
favorable range for powdery mildew) but to relative humidity, which is highly negatively correlated to
temperature. The promoting effect of humidity and free water on leaves on disease could be confirmed
by the highly significant correlation between leaf wetness and disease onset on leaves. Indeed, whereas
some caution must be exerted due to the small size of the leaf wetness dataset (3 data points), this
factor seems to explain not just final disease pressure but also the date of onset of leaf symptoms.
The only agronomic factors that appeared to be related to powdery mildew are the year of observation
and phenological timing and vigor. These two plant variables are both related to growth. Phenological
timing has already been indicated as a factor affecting powdery mildew (Gadoury et al, 2003; Caffarra et
al, 2012). On the other hand, vigor is known for favoring diseases in general, as it produces denser
canopies, more susceptible tissue, and more humid conditions for the disease to develop (Calonnec et
al, 2013). In the study, vigorous plots had later dates of grape infection, contrary to expectations.
However, as vigor is positively related to temperature and temperature is negatively related to
humidity, it is impossible to say whether the relationship between vigor and disease onset is causal, or
whether it is due to the covariance of indicators.
4. Conclusions
In conclusion, the present study highlighted, at least for 2017‐2018, a stronger correlation of powdery
mildew with climate and plant growth than with static plot descriptors. The high degree of temporal
(year effect) and spatial (plot effect) heterogeneity of powdery mildew corresponded to an equally high
degree of variability in local weather conditions during the period of primary infections. Such variability
and the high correlations with local powdery mildew pressure encourages a more precise weather
monitoring on plots and inside the canopy. Relative humidity and leaf wetness appeared as the most
closely correlated variables to powdery mildew onset and pressure. Undoubtedly these results need to
be further confirmed and quantified through more extended surveys, but they indicate interesting
directions for the improvement of predictive models of powdery mildew, that are actually unable to
predict disease hotspots. Ultimately the ability to build precise tools to optimize phytosanitary
management depends on our understanding of the environmental effects underlying the physiology of
this disease.
5.Litterature cited
CAFFARRA, A., RINALDI, M., ECCEL, E., ROSSI, V. AND PERTOT, I., 2012. Modelling the impact of climate change
on the interaction between grapevine and its pests and pathogens: European grapevine moth and
powdery mildew. Agriculture, Ecosystems & Environment, 148, 89‐101.
CAFFI, T., ROSSI, V., LEGLER, S. E. AND BUGIANI, R., 2011. A mechanistic model simulating ascosporic
infections by erysiphe necator, the powdery mildew fungus of grapevine: dynamic model for
ascosporic infections. plant pathology. 60, 3, 522-531.
CALONNEC, A., BURIE, J‐B., LANGLAIS, M., GUYADER, S., SAINT‐JEAN, S., SACHE, I. AND TIVOLI, B., 2013. Impacts of
plant growth and architecture on pathogen processes and their consequences for epidemic
behaviour. European Journal of Plant Pathology. 135, 3, 479-497
CALONNEC, A., CARTOLARO, P., CHADOEUF, J., 2009. Highlighting Features of Spatiotemporal Spread of
Powdery Mildew Epidemics in the Vineyard Using Statistical Modeling on Field Experimental Data.
Phytopathology. 99, 4, 411-422.
CALONNEC, A., CARTOLARO, P., NAULIN, J.‐M., BAILEY, D. AND LANGLAIS, M., 2008. A host‐pathogen simulation
model: powdery mildew of grapevine. Plant Pathology. 57, 3, 493-508.
June 23 - 28, 2019 | Thessaloniki | Greece
GiESCO Thessaloniki | 515
21st GiESCO International Meeting: ‘A Multidisciplinary Vision towards Sustainable Viticulture’
CHELLEMI, D. O., MAROIS, J. J. 1991A. Development of a demographic growth model for Uncinulanecator
by using a microcomputer spreadsheet program. Phytopathology. 81, 3, 250‐254
CHELLEMI, D. O., MAROIS, J. J. 1991B Sporulation of Uncinulanecator on grape leaves as influenced by
temperature and cultivar. Phytopathology. 81, 2, 197‐201
DELP, C. J. 1954. Effect of temperature and humidity on the grape powdery mildew fungus.
Phytopathology. 44, 615‐626
GADOURY, M. O., ROBERT, C., FICKE, S. A., WILCOX, W. F. 2003. Ontogenic Resistance to Powdery Mildew in
Grape Berries. Ecology and Population Biology. 93, 5, 547‐555
MAMMERI, Y., BURIE, J.B., LANGLAIS, M. AND CALONNEC, A., 2014. How changes in the dynamic of crop
susceptibility and cultural practices can be used to better control the spread of a fungal pathogen at
the plot scale? Ecological Modelling. 290, 178-191.
PEARSON, R.C AND GADOURY, D.M., 1987. Cleistothecia, the Source of Primary Inoculum for Grape Powdery
Mildew in New York. Phytopathology. 77, 11, 1509.
RAPILLY, F. 1991. L'épidémiologie en pathologie végétale: mycoses aériennes (Epidemiology in plant
pathology: aerial mycosis). Paris: Editions Quae (INRA), 1991. 318 p. 2‐7380‐0297‐8
WEENS V., 2017. L’hétérogénéité de l’oïdium de la vigne – Etude des facteurs influençant son
développement. (Hetereogeity of grapevine powdery mildew. Study of the factors affecting its
development). Mémoire de fin d’étude, supervised by BOIS, B., CAFFARRA, A. and GARIN, G., Nancy,
France : Université de Lorraine.
SALL, M.A., 1980. Epidemiology of Grape Powdery Mildew: A Model. Phytopathology. 70, 4, 338
WILCOX, W. F. 2003. Effects of humidity on the development of grapevine powdery mildew.
Phytopathology 93, 1137‐1144
2.a
2.b
Figure 1. Dynamics of powdery mildew on grapes (2.a) and leaves (2.b) in 2018.
June 23 - 28, 2019 | Thessaloniki | Greece
GiESCO Thessaloniki | 516
21st GiESCO International Meeting: ‘A Multidisciplinary Vision towards Sustainable Viticulture’
Figure 2. PCA for both years with the variables that were found to be highly correlated to disease
pressure.
Figure 3. PCA for 2018 with the variables that were found to be highly correlated to disease pressure.
June 23 - 28, 2019 | Thessaloniki | Greece
GiESCO Thessaloniki | 517