Search PMN  


PDF version
for printing

Symposium Homepage

Peer Reviewed

© 2010 Plant Management Network.
Accepted for publication 28 January 2010. Published 26 May 2010.

Development of an Advisory System for Grapevine Powdery Mildew in Eastern North America: A Reassessment of Epidemic Progress

Michelle M. Moyer and David M. Gadoury, Department of Plant Pathology and Plant-Microbe Biology, New York State Agricultural Experiment Station, Cornell University, Geneva, NY 14456, USA; Peter A. Magarey, Magarey Plant Pathology, PA & CC Magarey Pty Ltd, Loxton, South Australia, AUS 5333; and Wayne F. Wilcox and Robert C. Seem, Department of Plant Pathology and Plant-Microbe Biology, New York State Agricultural Experiment Station, Cornell University, Geneva, NY 14456, USA

Corresponding author: Michelle M. Moyer.

Moyer, M. M., Gadoury, D. M., Magarey, P. A., Wilcox, W. F., and Seem, R. C. 2010. Development of an advisory system for grapevine powdery mildew in eastern North America: A reassessment of epidemic progress. Online. Plant Health Progress doi:10.1094/PHP-2010-0526-02-SY.


Historical data indicated that year-to-year variation in severity of powdery mildew fruit infection in New York was not directly related to annual variations in temperature or relative humidity through their known effects upon conidial germination, colony expansion, or latent period. Although severity of fruit infection varied from 0.02 to 61% among years, these environmental factors generally remained in the optimal range for the pathogen, with only slight year-to-year variations during the brief period that fruit were susceptible to infection. Vineyard experiments in New York and Australia indicated that colony growth was slower and sparser than would be expected from previous lab studies. Additional experiments indicated that transient exposure to temperatures near 4°C for a few hours had a debilitating effect on mildew colonies and produced growth similar to that observed on ontogenically-resistant tissues: slow and sparse with epidermal necrosis. Additional studies on seasonal release of ascospores from overwintering cleistothecia indicated that a large proportion of inoculum was often released prior to local grapevine budbreak at multiple sites in the eastern USA. These findings may represent key elements in forecasting potential severity of fruit infection in cool climates such as New York.


For over 150 years grapevine powdery mildew has been a resilient and global adversary of grape production. The disease, caused by the fungus Erysiphe necator (syn. Uncinula necator), has been studied intensively since the 1850s when it first swept across Europe. After discovering how to control powdery mildew using sulfur, many felt that the disease was no longer a threat. As noted by P. Viala in 1893 in Les Maladies de la Vigne (16): "Concerning powdery mildew; we know what is needed to control the disease. There are a few things that are unknown, but they are of secondary importance." What Viala and others did not foresee was a shift in standards towards stringently low levels of powdery mildew on fruit, spurred by research confirming a wine maker’s worst fear: taint and spoilage of wine as a consequence of even trace levels of powdery mildew.

Current market demands for European-style wines and centuries of experience growing the mildew-susceptible Vitis vinifera varieties from which the wines are produced have insured that powdery mildew will continue to threaten the industry. Growers today face extraordinary pressures to keep susceptible, high-value grapes free from disease. Control provided by several fungicides has been compromised over the last 30 years due to resistance development leaving a greatly reduced arsenal for disease management. Proper timing of fungicide applications has become increasingly crucial to effective pathogen suppression. But what exactly does a grower need to know in order to more effectively time control measures? While various models have been developed for powdery mildew as it is experienced in generally warm and dry Mediterranean climates, do they answer the questions that are being asked by growers in the cooler and higher-rainfall climates? What information do such growers want the model to provide?

To identify these potential questions, we surveyed researchers, advisory personnel and growers across the Northeast region of the USA. Their responses framed the objectives for future model development. Our objectives are best defined by the following questions: (i) What are key warnings of a "bad year" for powdery mildew, and (ii) when is the first fungicide spray needed and when has an infection period occurred?

What are the key warnings of a "bad year" for powdery mildew?

Powdery mildew is ubiquitous but severe fruit infection occurs only during certain years. In some years, severity on clusters may be as low as 1% and in others over 60% (Fig. 1). What are the host, pathogen, and/or environmental factors that set the stage for either mild or severe fruit infection? Previous work described an exponential relationship between temperature and various growth responses of powdery mildew, making it the first major environmental parameter to be exploited in forecast models. To explore the usefulness of temperature as a parameter to forecast severity of fruit infection for New York, we derived equations from data reported by Delp (3) on the effects of temperature on germination. Our hypothesis was that the relationship of temperature to growth might be similar to that of germination, regardless of how "growth" is measured (e.g., germination, latency, colony expansion, etc). A model derived from such "growth" data might explain some of the observed year-to-year variation in severity of fruit infection and thus help potentially explain, and ultimately predict, when conditions are conducive for severe powdery mildew outbreaks. This is the approach followed by most of the existing forecast and advisory models for grapevine powdery mildew.


Fig. 1. Grapevine powdery mildew outbreaks in New York vary substantially from year to year. Fruit infection can range from complete loss (A) to very little disease in unsprayed vineyards (B). However, supposedly clean fruit may also be plagued with diffuse infection (inset), which can alter wine quality.


A series of three equations were developed to rate the favorability of hourly temperatures, assigning a severity score between 0 and 1 for every hour of the day. Ten growing seasons of historical weather data were selected based on how severe (> 40%) or how mild (< 4%) fruit infection was (Table 1) during each season, and these were analyzed from the period between budbreak (earliest possible colony establishment) to veraison, with emphasis on time periods between critical phenological stages (i.e., berry set, when ontogenic resistance influences fruit susceptibility).

Table 1. Powdery mildew cluster severity ratings for selected
years on susceptible, unsprayed Vitis hybrid 'Rosette.'

Year Surface area infected (%)
1986 48.2                          
1988 0.02                          
1989 47.7                          
1991 2.9                          
1992 50.10                          
1993 3.4                          
1994 61.04                          
1999 4.00                          
2001 2.50                          
2003 46.70                          

The analysis demonstrated that over the duration of the above period, or during selected two-week segments of the above period, temperatures were equally favorable year-to-year for germination of conidia, and thus derived severity ratings accumulated at similar rates from year to year (Fig. 2). Similar results were obtained when we applied this approach using a model based on the relationship between temperature and the duration of the latent period (3). This reinforces field studies in which it was found that while absolute levels of foliar disease differ from year to year, severity and incidence increased at similar rates. Ultimately, this means that using previously derived relationship to temperature as the sole predictor or indicator for disease progression in New York is unlikely to provide additional useful information. Temperatures here remain within the favorable range, and variation in this parameter is not associated with either mild or severe years. Furthermore, New York does not experience the high temperatures that are important in mortality of the pathogen, and which help moderate epidemics in Mediterranean climates.


Fig. 2. Temperature-based severity accumulations do not help explain fruit disease severity. Major differences in disease levels on fruit occur despite year-to-year similarities in temperature, expressed here in the form of severity accumulations.


Previous research (7) had indicated that the number and intensity of rain events during the critical period of fruit susceptibility could partially explain yearly variations in severity of fruit infection. However, there are many other environmental factors that also correlate with such events: reduced UV radiation, increased humidity, and moderated temperatures. Severe disease years are often typified by prolonged rain events with cooler daytime temperatures and more stable night temperatures, all of which tend to favor the growth of powdery mildew colonies.

When is the first fungicide needed and when has an infection period occurred?

Primary inoculum. In the Northeastern USA, E. necator over-winters as cleistothecia in grapevine bark crevices. Dormant infected buds that give rise to flag shoots and are key inoculum sources in Mediterranean climates do not survive the low winter temperatures in New York. Ascospores have previously been trapped in vineyards following rain events of greater than 2.5 mm coincident with temperatures greater than 4°C (6). While ascospores have only been trapped in the Northeast in vineyard air from budbreak to shortly after bloom, autumn and mid-winter releases have been reported in warmer Mediterranean climates. Are there years in New York during which rain before budbreak releases ascospores, thereby depleting the overwintering inoculum?

To address the foregoing question, populations of cleistothecia from New York were overwintered at vineyards located in New York, North Carolina, New Jersey, and Washington during 2005-2006. The experiment was repeated and expanded in 2006-2007, when populations of cleistothecia from New York were overwintered in vineyards in New York, Washington, New Jersey, Georgia, North Carolina, and Virginia; and populations of cleistothecia from New Jersey, North Carolina, Washington, and Pennsylvania were overwintered in New York. In both overwintering seasons, samples were shipped overnight to New York to be assessed for the potential to discharge ascospores in a laboratory setting. Ascospore release potential was assessed every 1 to 2 weeks from January until mid-June in both years.

In general, dates of budbreak in 2006 were progressively earlier as one moved from north to south among the overwintering sites. Cleistothecia originally collected in New York but overwintered at these locations showed distributions of ascospore release that were correspondingly shifted in time. At all sites (including New York) ascospores from New York cleistothecia were released and germinated in lab tests well before local budbreak (Fig. 3), reaching between 60 to 100% cumulative release prior to those local budbreak dates. Interestingly, severity of cluster infection was relatively high in 2006 in New York, which might indicate that severe fruit infection is related to early release and a consequent early abundance of foliar mildew colonies. However, given how rapidly the populations could progress from 0 to 100% dehiscence, the pathogen walks a fine line between early establishment and pre-budbreak depletion of primary inoculum. This offers an intriguing hypothesis to explore as an explanation for the observed year-to-year variances in severity of fruit infection: is the severity of fruit infection profoundly affected by asynchrony of host and pathogen phenology (Fig. 4)?


Fig. 3. Cumulative ascospore release from New York populations overwintered in New York in 2005-2006, 2006-2007, and 2007-2008. Vertical lines indicate yearly budbreak for Vitis vinifera ‘Chardonnay’ in Geneva, NY. In all three years, a substantial proportion of cleistothecia were able to release ascospores well before budbreak, and many of the cleistothecia dehisced in the field before the availability of host tissue.



Fig. 4. Ontogenic resistance could potentially alter epidemic fate. Schematic of potential powdery mildew epidemic outcomes in relation to the timing of arrival of primary inoculum and the development of ontogenic resistance in grape clusters. In grapes, a substantial delay in inoculum arrival could allow the fruit to escape infection.


North Carolina and New York represent the two extremes in terms of timing of release: distributions from all other populations were intermediate to New York and North Carolina (data not shown). As in 2005-2006, once a population had produced mature ascospores, release would quickly progress towards 100%. Irrespective of whether a cool-climate population (New York) was overwintered in a warmer site (North Carolina) or a warm-climate population (North Carolina) overwintered in a cooler site (New York), substantial ascospore release occurred before local budbreak, although local populations were usually more synchronized with the local host phenology compared to imported populations.

In summary, cleistothecia from several local and imported populations released germinable ascospores in laboratory assays prior to emergence of host tissue in fields across a broad viticultural region. However, this merely indicated the potential release under vineyard conditions. Do rain events suitable for ascocarp dehiscence (6) occur before budbreak under vineyard conditions? A review of historical weather data for New York alone indicated that such events are indeed common even in mid- to late winter and would be expected to occur even more frequently as one moved southward within a region, to areas with warmer winter temperatures. We are presently investigating the frequency and intensity of such pre-budbreak rain events for their potential to prematurely deplete the supply of ascospores, in particular for those years in which disease severity of fruit is low.

Secondary inoculum. Despite the existence of a diverse body of literature related to environmental factors and the increase of powdery mildew, there was a paucity of information on precisely how colonies develop during the critical early weeks of the growing season. Not only does temperature at this time fluctuate greatly within and below the optimal range for E. necator, but the first 6 to 8 weeks after budbreak contain the period during which the fruit are at risk (e.g., from immediate prebloom to 2 weeks postbloom). We therefore inoculated emergent, expanding and fully-expanded leaves beginning at 15 cm of shoot growth. Interestingly, colonies grew far more slowly than expected from previous controlled-environment studies (1,2,3), seen as an increased difference between expected and observed latent periods represented in Fig. 5. Many inoculations of highly susceptible tissues failed to produce colonies. Those that successfully became established were associated with substantial epidermal necrosis on inoculated leaves, in particular at the sites of appressoria (Fig. 6). Colonies often took up to 3 weeks to sporulate, if they reached that stage at all (Table 2). This response was observed repeatedly over two growing seasons, using different methods of inoculation, with multiple isolates, on two cultivars, and has now also been observed in a climate similar to that of the Northeastern USA: Tasmania, Australia (Dr. Kathy Evans, University of Tasmania, personal communication).


Fig. 5. Differences between actual and predicted latent periods of field-developed colonies, and the relationship between overnight minimum temperatures. As minimum temperatures no longer fall below 10°C, actual latent periods begin to approach predicted latent periods.



Fig. 6. Erysiphe necator colonies developing on field tissue in the early part of the growing season are typified by diffuse growth, low fecundity, and host-cell necrosis at the site of appressorium formation.


Table 2. Predicted and actual 2007 sporulation dates of field inoculations of Erysiphe necator in Geneva, NY, USA, and Loxton, SA, Australia.

Site Inoculation
date adjusted

Geneva, NY 29 May 8 Jun 3 Jun 9 Jun
Geneva, NY 11 Jun 21 Jun 17 Jun 21 Jun
Geneva, NY 3 Jul 16 Jul 9 Jul 13 Jul
Geneva, NY 11 Jul 25 Jul 18 Jul 23 Jul
Geneva, NY 10 Aug 20 Aug 16 Aug 20 Aug
Loxton, SA 25 Sep N/A  11 Oct*  7 Nov*
Loxton, SA 4 Oct N/A 13 Oct  7 Nov*
Loxton, SA 9 Oct N/A  3 Nov*  7 Nov*
Loxton, SA 19 Oct 7 Nov  3 Nov*  7 Nov*

 * Temperatures reaching lethal highs were dealt with by resetting biological start points.

Recently, our lab reported that the latent period of E. necator is density-dependent (8), i.e., colonies arising from single conidia exhibit latent periods approximately 50% longer than those that develop from 10 or more conidia. The former (a single spore) represents the more natural deposition of inoculum in an epidemic, while the latter (a dense spore deposit) is an artifact of purposeful inoculations in research. Thus, previously reported latent periods of as short as 5 days for this pathosystem (1,2,3), possibly represent gross underestimates of the true latent period under vineyard conditions. Even with the use of dense spore deposits in the inoculations, observed field latent periods were generally 30 to 50% longer than what would be predicted based upon average hourly temperatures (1,2,3).

What then are the true effects of temperature on development of powdery mildew? It would appear that previous studies have accurately described the relationship between temperature and development within and perhaps slightly below the optimal range. Some of these studies have also examined the lethal effects of high temperature on the pathogen. However, no one has considered either the transient or long-term impacts of commonly occurring acute cold events upon pathogen survival and development. Our preliminary studies suggest that brief exposures to cold temperatures (as little as 3 h at 4°C) can actually debilitate and damage preexisting colonies, causing hyphal segment death as seen through vital staining with fluorescein diacetate (Fig.7). Previous studies have often assumed that suboptimal temperatures have a neutral effect on development, and models based upon these studies usually involve extrapolation of a minimal growth response. However, our results suggest that acute low temperatures not only have a negative effect, but that the magnitude of the effect may approach that reported for exposure to high temperature (1,2,3).


Fig. 7. Fluorescein diacetate is a stain that helps visualize esterase activity. Erysiphe necator colonies that were subjected to an acute cold exposure have gaps of fluorescence in the hyphal strands, indicating damage.
(A) Control colonies do not have noticeable damage.
(B) Cold treated colonies (8 h at 2°C) have a greatly reduced fluorescent area and noticeable gaps in fluorescence in hyphal strands.


In many respects, the host and pathogen responses that typify development of mildew colonies during the earliest phases of an epidemic mimic those our lab has reported for ontogenically resistant tissues (9). This raises another interesting question: Do normally susceptible tissues express a form of resistance if they are subjected briefly to sub-optimal conditions? Reports from other pathosystems suggest that plants pre-conditioned in cold temperatures express genes that enhance resistance to powdery mildews. Does low temperature directly affect the pathogen, does it somehow alter tissues that are ordinarily (and ontogenically) highly-susceptible, or both? Due to the biotrophic nature of powdery mildews, it will be a challenge to tease out the confounding factors. We are continuing our investigations of how acute cold events occurring both before and after infection alter the establishment, survival, and spread of E. necator.


The design of this project was driven by grower-identified priorities. A critical analysis of weather parameters successfully exploited to forecast disease in Mediterranean climates revealed several unexpected aspects of epidemic development in the comparatively cooler climate of the Northeastern USA; including asynchrony of host and ascocarp development, lack of correspondence of the previously applied temperature relationships to early epidemic development in the Northeastern USA, a protracted latent period and attrition during unexpectedly slow and sparse colony development. Additionally, we have explored a potentially effective means to counter a potentially catastrophic infection of fruit at a critical stage of susceptibility. Our long-term goal is the assembly of these various parts into a unified, coherent and comprehensive forecasting and advisory system for improved management of powdery mildew in the Northeastern US. To quote van der Plank (18) "Chemical industry and plant breeders forge fine tactical weapons; but only epidemiology sets the strategy."

Literature Cited

1. Chellemi, D. O., and Marois, J. J. 1991. Sporulation of Uncinula necator on grape leaves as influenced by temperature and cultivar. Phytopathology 81:197-201.

2. Chellemi, D. O., and Marois, J. J. 1991. Development of a demographic growth model for Uncinula necator by using a microcomputer spreadsheet program. Phytopathology 81:250-254.

3. Delp, C. J. 1954. Effect of temperature and humidity on the grape powdery mildew fungus. Phytopathology 44:615-626.

4. Erickson, E. O., and Wilcox, W. F. 1997. Distributions of sensitivities to three sterol demethylation inhibitor fungicides among populations of Uncinula necator sensitive and resistant to triadimefon. Phytopathology 87:784-791.

5. Frank, J. A., Cole, H. J., and Hatley, O. E. 1988. The effect of planting date on fall infections and epidemics of powdery mildew on winter wheat. Plant Dis. 72:661-664.

6. Gadoury, D. M., and Pearson, R. C. 1990. Ascocarp dehiscence and ascospore discharge in Uncinula necator. Phytopathology 80:393-401.

7. Gadoury, D. M., Seem, R. C., Magarey, P. A., Emmett, R., and Magarey, R. 1997. Effects of environment and fungicides on epidemics of grape powdery mildew: Considerations for practical model development and disease management. Vitic. Enol. Sci. 52:225-229.

8. Gadoury, D. M., Wakefield, L. M., Seem, R. C., Cadle-Davidson, L., and Dry, I. B. 2004. Preliminary studies of signaling and sporulation in Uncinula necator. Phytopathology 94:S33.

9. Gadoury, D. M., Seem, R. C., Wilcox, W. F., Henick-Kling, T., Conterno, L., Day, A., and Ficke, A. 2007. Effects of diffuse colonization of grape berries by Uncinula necator on bunch rots, berry microflora, and juice and wine quality. Phytopathology 97:1356-1365.

10. Gee, L., Stummer, B., Gadoury, D. M., Biggins, L. T., and Scott, E. 2000. Maturation of cleistothecia of Uncinula necator (powdery mildew) and release of ascospores in southern Australia. Aust. J. Grape Wine Res. 6:13-20.

11. Gubler, W. D., and Thomas, C. 1999. Control of powdery mildew using the UC-Davis powdery mildew risk index. Online. January 1999, APSnet Feature, American Phytopathological Society, St. Paul, MN.

12. O'Hara, P., Ayres, P. G., Hughes, M. A., Dunn, M. A., and Smith, R. J. 1998. The influence of powdery mildew (Erysiphe graminis f.sp hordei) on the accumulation of transcripts from low-temperature-responsive genes in barley. Physiol. Molec. Plant Pathol. 52:353-369.

13. Ough, C. S., and Berg, H. W. 1979. Powdery mildew sensory effect on wine. Am. J. Enol. Vitic. 30:321.

14. Pearson, R. C., and Gadoury, D. M. 1987. Cleistothecia, the source of primary inoculum for grape powdery mildew in New York. Phytopathology 77:1509-1514.

15. Sall, M. A. 1980. Epidemiology of grape powdery mildew: A model. Phytopathology 70:338-342.

16. Viala, P. 1893. Les maladies de la vigne. Georges Masson, Paris, France.

17. Wong, F. P., and Wilcox, W. F. 2002. Sensitivity to azoxystrobin among isolates of Uncinula necator: Baseline distribution and relationship to myclobutanil sensitivity. Plant Dis. 86:394-404.

18. van der Plank, J. E. 1963. Plant Diseases: Epidemics and Control. Academic Press, London, UK.