2005 Downy Mildew Epidemic Summary

2005 Cucurbit Downy Mildew Epidemic Summary

G. Holmes, C. E. Main, T. Keever, M. Miller, P. Pyle

North American Plant Disease Forecast Center (NAPDFC)

North Carolina State University

February 2005

The North American Plant Disease Forecast Center (NAPDFC) at North Carolina State University has completed the 2005 downy mildew forecasts. The forecasts were issued on the internet every Tuesday and Thursday, with additional forecasts for significant weather events. The HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model available through NOAA’s Air Resources Laboratory (ARL) is used to simulate air parcel trajectories that represent pathogen spore transport. Timely information on pathogen spore transport and atmospheric conditions along the trajectory pathways are valuable to cucurbit growers across North America in managing the downy mildew disease epidemic. The cucurbit downy mildew web page offers maps of past and present disease locations, up to date forecasts, disease control recommendations, archive forecasts, and the current season epidemic history and control measures. The disease was highly active and aggressive in all locations in 2005.

On average during the 2005 growing season, the eastern US experienced slightly below normal to normal precipitation amounts and slightly above normal to normal temperatures. The Ohio Valley was the region that was hardest hit with warmer and drier conditions.

Downy mildew forecasts began on April 5. A total of 350 forecasts were issued. 63 counties and 18 states mainly in the eastern United States were affected by the disease. The first occurrence was in late February from southern Florida on cucumber, squash, and later on cantaloupe. The disease was found to be highly aggressive, attacking past season resistant cucumbers. Growers in the area struggled to control the epidemic, with most losing nearly 100% of their crops.

Downy mildew began to make its way northward into southern GA, affecting Pierce County near June 9 and Brooks County near June 10. The report from Pierce County suggests the disease was two weeks old. For this time period, this would suggest the time window of infection was approximately May 19-23. The primary lesions would have become visible near May 26-30. The NAPDFC performed backward trajectory analysis from Pierce County (Mershon), GA for the time period of May 17 – 31. No significant events came from the southern FL area where disease was known to exist. An early transport event may have occurred before the forecasts began to be generated. One possible event could have occurred early in May. The April 28 forecast out of southern FL suggests a Moderate Risk to southeastern GA for May 1. Pathogen spores may have been transported to southern GA at this time. This time frame is not within the window of infection and is therefore not considered a highly possible transport event because weather conditions were mostly favorable during the month of May in the region with a stationary frontal boundary in Florida, followed by a dry period, and a cold frontal passage. At this time, the cause of the southern GA outbreak cannot be determined with much confidence.

Tropical Storm Arlene began to develop in the western Caribbean in early June and made landfall on the coast of Alabama and Florida on June 11 (Figure 1). The NAPDFC performed trajectory analysis after the storm moved northward through the Ohio Valley and posted the information on the website the week of landfall. The storm analysis consisted of trajectory and spore deposition simulations from the southern Florida and southern Georgia known source locations. Figure 2 shows a sample 48 hour spore deposition map from the Mershon, GA source location valid June 11. Deposition analysis suggested a short distance of spore deposition only reaching central and northern Alabama, northern Mississippi and western Tennessee. The maximum spore deposition locations were near the source locations, due to the heavy precipitation associated with the tropical system.

The most interesting transport event of the 2005 downy mildew forecast season was the movement of the disease from southern Georgia to New Jersey. The New Jersey disease report to the NAPDFC suggests that the first sign of infection was June 8. This would create a window of infection from approximately May 28 – June 3. After further reanalysis of weather and trajectory data, the May 31 forecast out of southern Florida is the only worthy candidate of a possible transport event. However, as the transport event traversed northward, it encountered rain all along the East Coast. Risk assessments were confined to South Carolina and further southward due to the likelihood of spore washout in these areas. Reports also mention that cucurbit transplants from Florida were planted near May 3. Due to the time period from the transplanting to the disease reports the NAPDFC does not have a satisfactory explanation regarding the infection time for this outbreak. Downy mildew became a major problem to growers in the mid Atlantic after this disease report. A total of seven counties from four states near the first disease report in Cumberland County, New Jersey were reported with six of them being received in less than two months from the first report on June 8.

A report of downy mildew attacking cucumber in southeastern North Carolina, Duplin County, was received on July 4. We performed reanalysis of the spore deposition out of the known southern Georgia and Florida sources to further investigate the transport event. The Mershon, GA trajectories did not support the transport event to Duplin County. The forecast generated on June 23 from southern FL utilized figures 3 (a) Palm Beach County, FL and (b) Collier County, FL to issue risk assessments along the trajectory pathway. The image shows trajectory pathways from southern FL locations initiated from three different heights; red – 10 m, blue – 100 m, and green 250 m. A Low Risk was forecasted for cucurbits in southeast North Carolina for these events. However, the NAPDFC believes this transport event was the most worthy candidate for infection in southeastern NC. Four additional eastern NC counties also reported downy mildew during the forecast season.

Downy mildew then began to spread into western NC. The report was received from Haywood County on July 14. Analysis of the situation, using historical data (or reanalysis data) suggested that events from the Mershon, GA source were responsible. Two transport events from June 26 and 27 have minimal deposition over western NC. The NAPDFC issued a Low Risk of infection because of the likelihood of spore washout earlier in the event. A Moderate Risk of infection to cucurbits in western North Carolina from Mershon, GA was issued for July 6 in the July 5 forecast. This event occurred too late, however, due to the fact the primary lesions were found near July 4. Two additional counties, one in northern GA and western NC also reported downy mildew during the forecast season.

During the month of August, 13 counties in Michigan were attacked by downy mildew. Downy mildew in MI was first reported to the NAPDFC on August 9. The electronic report estimated that symptoms first appeared on July 5. Assuming that symptoms began on the 5^th, the likely date of infection would have been around June 26. Thorough investigation of transport events traveling over MI in the month of July support an earlier infection date. According to results, 4 of the 31 days in July yielded trajectories that originated over known areas of downy mildew infection. Of the identified events, only the period of July 16 and 17 stands out as a good candidate for a possible infection scenario.

The transport event ending the 16 and 17 of July originated over the NJ/DE hotspot region on the July 14. The 70+ hour transport event was very long, but archived weather data suggest that sufficient cloud cover may have been present to prevent significant irradiation of spores mid-transport. In addition, precipitation data in MI support the likelihood of airborne spore washout. Figure 4 shows a 72 hour reanalysis trajectory, initialized from the NJ/DE region starting at 0800 UTC July 14. The colored lines represent various starting heights above the source location, red – 100 m, blue – 500 m, and green – 1000 m. All trajectories move northwestward, with the lowest height (100 m) moving through the lower MI area.

The long-range transport of spores from the Mid-Atlantic coast to the Michigan area was a rare event. The event described above by the NAPDFC occurred in the tropical air mass left behind by Hurricane Dennis. The post-Dennis Low most likely set up enough east-to-west transport to allow for such a long distance transport event. Typically, mean flow is west-to-east across this region preventing such transport events from occurring.

The exact date of infection for the Michigan and surrounding areas is not certain. Observations and reports occurred after many disease cycles. When this happens, it becomes very difficult to establish the time of the formation of primary lesions. The window for time of initial infection remains very wide. Thus, the NAPDFC cannot determine the transport event responsible with a high degree of confidence. Additional reports from a total of six counties in northwestern Ohio and northern Indiana were received in August and early September respectively. Due to the close proximity of these sources to the large area of infection in Michigan, pathogen spore transport most likely originated from Michigan.

On August 16, the NAPDFC received three additional reports from southern Indiana, eastern Kentucky, and Connecticut. The spread of downy mildew to the lower Ohio Valley is difficult to quantify. During the infection time periods for both the Hardinsburg, IN and Paintsville, KY reports, forecasted transport events from Mershon, GA and Canton, NC affect this region. However, spore death was forecasted for the transport events, due to solar exposure, therefore Low Risk assessments were issued for these areas. Three additional counties near the Hardinsburg source location and one near the Paintsville source were also reported. The North Haven, CT downy mildew report suggested a first occurrence on August 1. Low risk assessments were issued for this region during this time period. The NAPDFC is not certain of a likely transport event to this region.

A week after the CT report, Long Island, New York reported downy mildew on August 23. The symptoms of downy mildew were first observed between the 15^th and 18^th on a variety of cucurbits including cucumber and winter squash. Based on the reports from observers, the NAPDFC estimated a window of possible infection between the 3^rd and 9^th of August.

The NAPDFC performed a careful reanalysis of air parcel trajectories from the suspected source area of southern New Jersey. During the suspected window of infection, 2 periods, spanning a total of 4 days, were identified as having the potential to impact the Long Island, NY area.

Air parcels originating from the New Jersey downy mildew hotspot area passed over Long Island on August 4, 6, 7, and 8. The air parcel trajectories suggest that there was an almost constant risk of spore deposition for the period in question. However, deposition of viable spores requires rainfall and protection from heavy solar exposure. Surface and radar data suggests that a typical summer precipitation pattern was in place. Mornings were typically clear of clouds. Cloud formation would begin around midday and there would be a chance of rain or thundershowers towards the afternoon. Surface and satellite data suggests that cloud cover most days would have been enough to sustain any live spores long enough to travel from New Jersey to Long Island with only moderate spore death in route. Manually digitized radar data from Plymouth State University shows that there was plenty of precipitation over Long Island to induce deposition of live spores. Surface data from the Long Island area didn’t correlate well with the radar data. However, the data did indicate very light and isolated rain during the suspected window of infection.

The NAPDFC believes that even though there was not a clear precipitation event to attribute live spore deposition in the Long Island area, there was enough rain and live spores over the area to signify the period in question as the most likely time of infection. Rather than being the result of a single transport event, infection in this case resulted from several transport events with light deposition of spores over four separate and mostly contiguous days. Figure 5 shows multiple transport events on August (a) 4 (b) 6 (c) 7 and (d) 8 originating from the NJ region traversing across the Long Island area. Air parcel initialization heights over the source are distinguished by the pathway color. Red represents 100 m, blue 250 m, and green 500 m. The trajectories shown are a product of archived meteorological data. Other transport scenarios are not likely as they would not fit in with the timing of when initial symptoms occurred.

The NAPDFC forecasted a Low Risk for this particular event. At the time we believed that strong solar exposure combined with a lack of forecast precipitation would mean that virtually all spores would die before any potential deposition. However, early formation of clouds and precipitation, on a scope not well predicted by operational numerical weather forecast models, allowed for live spore deposition over Long Island. Given the weather information available at the time of forecast, the NAPDFC issued risk level was reasonable. Infection under the scenario described in this analysis is uncommon. Similar transport situations occurred frequently throughout the 2005 forecasting season with no infection having been reported.

The epidemic continued to spread, affecting Barnwell County, South Carolina and Ontario County in western New York State. The two reports were received on August 25 with first signs of infection in mid August. The Mershon, GA source location is most likely the origination of the infection in Barnwell, SC. The August 9forecast from Mershon, GA issues a Moderate Risk to cucurbits in SC for the August 10 trajectory from Mershon. Four more counties in southeastern South Carolina were also affected later on in the season with significant impact. Low Risk assessments to the western NY area were issued due to solar exposure or earlier spore washout. At this time, the NAPDFC does not have a worthy transport event candidate for this region.

Downy mildew was reported on September 1 from Holcomb, Missouri. The NAPDFC believes the likely source of the transport event was Hardinsburg, IN. The August 16 forecast from Hardinsburg issued a High Risk of infection for southern IN and western KY. Due to the likelihood of spore washout, a Low Risk assessment was issued for the southeastern MO area.

On October 18, a rare report from Utah was received. Downy mildew had attacked watermelon and cantaloupe throughout the growing season. The report suggested that the first occurrence was most likely in August and growers have not seen the disease this active in nearly 40 years. At this time the NAPDFC does not have a candidate for this disease location. However, it most likely did not originate from the known source locations in the eastern US due to the normal west to east atmospheric flow patterns in North America.

In early November, reports of downy mildew affecting the second growing season in southern FL were received. The latest report was received on December 7 and suggested the disease was still active. Due to the winter season in most growing locations, this source area is not a threat to other cucurbit production areas in the eastern US.

The 2005 cucurbit downy mildew growing and forecasting season was active throughout most of the eastern US. The disease was found to be highly aggressive in some areas, especially southern FL. Timely forecasts from known source locations were issued throughout the heart of the growing season. The overall effectiveness of forecasting epidemic spread not only relies on the accurate prediction from atmospheric models, but also accurate knowledge of the field conditions. Scouting and reporting must occur in the early stages of disease development to be of use in determining the exact circumstances of infection. Reports later in the disease cycle are helpful for establishing the strength of a source location and aid the NAPDFC in determining which areas require special attention and which areas are no longer a threat due to spray treatment and harvest. To best fulfill our mission, the NAPDFC needs reports early and updates often.

The 2005 season was active in the fields and our website. The website received approximately 50,000 visitors throughout the season. The NAPDFC thanks all the reporting Coordinators and growers for their participation during the 2005 forecast season. If you wish to become part of the Forecast System, or need further information, contact Dr. Gerald Holmes or Thomas Keever at the NAPDFC directly. You may also refer to the cucurbit downy mildew website at: http://www.ces.ncsu.edu/depts/pp/cucurbit/.

List of Figures

Figure 1. Tropical Storm Arlene reanalysis track from the Caribbean to the Ohio Valley. Image courtesy of Unisys weather.

Figure 2. 40 hour accumulated spore deposition map from Mershon, Georgia starting at 1400 UTC June 11.

Figure 3. Model forecasted trajectory events from (a) Palm Beach County, FL and (b) Collier County, FL. Trajectories are valid 1400 UTC June 25 through 1400 UTC June 27. Red – 10 m starting height, blue – 100 m, and green 250 m. The bottom of the image shows a time series of model derived precipitation amount.

Figure 4. NAM reanalysis trajectory valid 0800 UTC July 14 through 0800 UTC July 17, initiated from DE at three different heights; red – 100 m, blue – 500 m, green – 1000 m.

Figure 5. Archived meteorological data derived trajectories from the southern NJ source location to Long Island, NY. Trajectories are valid (a) 1400 UTC August 4, (b) 1400 UTC August 6, (c) 1400 UTC August 7, and (d) 1400 UTC August 8.

Figure 6. 2005 Cucurbit Downy Mildew epidemic status map at the end of the 2005 disease season.

The image “http://weather.unisys.com/hurricane/atlantic/2005/ARLENE/track.gif” cannot be displayed, because it contains errors.

Figure 1. Tropical Storm Arlene reanalysis track from the Caribbean to the Ohio Valley. Image courtesy of Unisys weather.

Figure 2. 48 hour accumulated spore deposition map from Mershon, Georgia starting at 1400 UTC June 11.

Figure 4. NAM reanalysis trajectory valid 0800 UTC July 14 through 0800 UTC July 17, initiated from DE at three different heights; red – 100 m, blue – 500 m, green – 1000 m.

Figure 6. 2005 Cucurbit Downy Mildew epidemic status map at the end of the 2005 disease season.