You are here

2009 Annual CDM Summary

-A A +A

Summary of the 2009 Cucurbit Downy Mildew Epidemic in North America

Thomas Keever
Forecaster / Meteorologist, CDM ipmPIPE
North Carolina State University


2009 was the second year of Cucurbit Downy Mildew (CDM) epidemic forecasting within the new CDM ipmPIPE initiative. The CDM ipmPIPE is a joint effort of academia, the USDA, and industry headquartered in the Plant Pathology Department at North Carolina State University. The CDM ipmPIPE provides opportunities not previously available for cucurbit downy mildew forecasting. A network of carefully monitored sentinel plots located in southern Ontario and the central and eastern United States (plus 2 in CA) and southern Ontario was available to aid the tracking of epidemic spread. Collaboration with the NC State Climate Office yielded improvements in the website, increased efficiencies in the reporting and presentation of outbreaks, and provided additional tools for analysis and evaluation of the forecasts. Please consult our website at for additional and more specific information.


Sentinel Plots

The network of sentinel plots, available initially in 2008, was again part of the CDM forecasting framework in 2009. Eighty-five sentinel plots were planted in 24 states and 1 Canadian province. California was added in 2009. Each plot contained cucumber (2 varieties), squash (2 varieties), cantaloupe, watermelon, and pumpkin. Plots were scouted weekly for signs of disease. Fifty-nine of the 85 sentinel plots reported CDM; disease was not observed in 26 sentinel plots. Of the sentinel plots that did not report disease, most were in areas that were a large distance from other outbreaks, in regions where the weather during the growing season was more unfavorable for disease development, or in regions where they were part of a mix of weak and/or widely scattered outbreaks. However, some plots reporting no CDM were relatively close to counties reporting disease activity.

As was the case in 2008, tracking of the 2009 CDM epidemic was greatly enhanced by the presence of the sentinel plot network. Nearly half of the initial reports from new states or new regions came from the sentinel plots. Tracking the evolution of a plant disease epidemic is a cooperative effort that requires timely information from the field. Diligent scouting at regular intervals is essential. The sentinel plot system provided both of these related elements.


Epidemic Overview

The CDM ipmPIPE Forecast Center received 205 unique reports from 157 counties in the United States and southern Canada in 2009 (“unique reports” excludes sentinel plot updates). This included 24 states east of the Rocky Mountains, southern Ontario, and California. There were 59 reports from the sentinel plots, 90 from commercial fields, 22 from other research plots, and 28 from home gardens. The numbers for unique reports and counties represented an increase from the 2008 values of 174 and 130, respectively.

Similar to previous years, the large majority of outbreaks occurred from the Great Lakes into the Northeast and along the growing regions of the Eastern Seaboard. Scattered to loosely-clustered reports were found from the central Gulf Coast through the Ohio Valley. Isolated outbreaks were reported in the Midwest, TX, and CA. Only 1 county in TX reported CDM from the southern Plains growing region.

North Carolina and Michigan led the 2009 CDM reports with 21 counties each. Ten counties had confirmed reports in Virginia, Ohio, and Alabama. New York and FL had 8; South Carolina had 7. Six counties in PA reported CDM. The states of GA, IL, KY, MD, and WI each had 5 counties reporting CDM. Four counties in TN and NJ reported outbreaks. CA, DE, IN, and southern Ontario reported CDM in 3 counties. The other states that reported CDM in 2009 had 1 or 2 counties.

As in 2008, infections on cucumbers dominated the 2009 CDM epidemic. Nearly every region that reported disease had confirmed CDM on cucumber. The only exception was TX. Of 90+ commercial field reports, only 20 did not involve a variety of cucumber, and it is likely no cucumbers were planted in most of these. Only 1 sentinel plot that reported CDM did not have a report on cucumber. Also, a very large majority of ‘first reports’ from states or growing regions occurred on cucumber. This included all but 3 of the initial reports that occurred in the sentinel plots (the rest were on squash).

In the commercial fields, reports on squash arrived from the growing areas along the East Coast. CDM was observed on cantaloupe in NC and LA. Infections in fields of giant pumpkin were reported in the Mid-Atlantic. Watermelon reports were restricted to TX and VA. All known outbreaks on these other hosts were dominated by the impact on cucumber.

In the sentinel plots, most of the reports in 2009 occurred initially on cucumber, or concurrently with another cucurbit host (typically squash). Disease then spread to other hosts in the plot during periods of favorable conditions. In nearly all cases, multiple hosts in a plot later became infected. Many plots had at least 5 of the 7 cucurbit hosts infected by CDM. As in 2008, there was variation among the plots, as well. This was likely due to other factors in addition to the CDM (weather favorability, host susceptibility, condition/growth stage of the plants, presence of other diseases, etc).

Confirmed outbreaks in home gardens occurred in a dozen states scattered over most major growing regions. The affected host was usually squash, cucumber, or cantaloupe. Reports of CDM in research plots arrived mostly from sites in the South or Great Lakes regions. These mostly frequently involved squash or cucumber.


Epidemic Chronology

Forecasts in 2009 began on March 2 from sources in southern FL. CDM had been present here on squash and cucumber for some time. The epidemic progressed very slowly for the next several months, limited by lack of host availability and dry weather conditions. Outbreaks were reported at a research plot in Marion County, FL on April 22 and a week later at the nearby sentinel plot. Both disease reports were on cucumber. The sentinel plot in Hillsborough County, FL reported CDM on May 27 on squash, cucumber, and cantaloupe. Symptoms here may have appeared as early as May 13, though it appears that live spores were also introduced by a subsequent event during mid-May. One of these events (Low risk) is shown below.

 sFL trajectory

Disease activity began in to increase in late May and June. The first report from GA arrived on May 29 from a commercial field of cucumber in Lowndes County near the GA / FL border. Symptoms were estimated to have appeared on May 27. Though there may have been an earlier introduction, airborne spores were most likely deposited during the High risk May 23 event from north-central FL, shown below.

nFL traj


CDM was discovered on watermelon in a large commercial field in Hidalgo County, TX on June 1, in the growing region of the lower Rio Grande Valley. Information from the report indicated that disease was probably present in other fields in the area. Similar to previous years, examination of transport events does not reveal any plausible explanations due to movement from the known sources. Sources in Mexico are suspected, but cannot be confirmed. This was the only outbreak of CDM reported to the Forecast Center from the southern Plains in 2009. 2009 was a dry year in this region.

 California’s first report of CDM occurred on June 5, in a commercial field of cucumber. Symptoms were estimated to have appeared on June 1. Origins are unknown. However, this outbreak was not due to transport from sources in the eastern United States. There is no evidence to support transport from regions to the east, and the central and eastern U.S. / southern Canada CDM pathosystem is separated from the one along the West Coast by the Rocky Mountains. It is possible that this outbreak was due to a transport event from Mexico but this cannot be confirmed.

A period of active weather in the Southeast occurred from approximately May 20 to May 29. A low pressure area in the Gulf of Mexico caused 4 to 5 days of variable to favorable conditions for epidemic spread. Unsettled weather continued for another 4 to 5 days, with afternoon / evening showers in a moisture-laden air mass ahead of a cold front. New infections in several different regions / states resulted, including GA (described above), AL, LA, and NC.

The first outbreak of CDM in North Carolina was reported on June 5, earlier than in previous years. Disease was confirmed on cucumber in a commercial field in Franklin County in the northeast Piedmont area. First symptoms were estimated at June 4. This outbreak is closely related to the second one confirmed in NC from Columbus County (southeast NC), reported from a commercial field of cucumber on June 12 with first symptoms estimated at June 2.

The scenario for epidemic spread to NC that is best supported by the available evidence is one in which viable spores were deposited at both NC locations during the transport events of May 27. One event came from north-central FL, tagged with Weakly Moderate risk for the near-coastal counties of SC and southeast NC. The other event came from the new infection in southern GA, which would be discovered and reported 2 days later. Both of these events moved northeast. Deposition would have occurred during the second day of travel; this is less common, but weather conditions supported this possibility. The GA trajectory is shown below.

sGA traj

Other scenarios cannot be completely ruled out, but appear far less likely. These might have involved spore contribution from (then) unknown sources, infection first at one site and then later at another, or earlier infection at both locations. Beyond the southern GA source, there is no evidence to support the first option. Other nearby outbreaks at the sentinel plots in southern GA (Tift County) and northern FL (Gadsden County) were initiated at the same time, or later, and symptoms would not have been missed during the regular scouting at these locations. The other two options would have involved earlier infections due to prior transport events to the northeast from either known or unknown sources southwest of NC. The idea of earlier infection has some modest support from the field observations which indicate that may have been possible. However, neither the forecasts nor later analyses show any plausible transport scenarios that fit this description.

Louisiana’s first report came from a home garden in Vermillion Parish in the south-central part of the state. The date of first symptom appearance was estimated to be June 2 in the report of June 10, but may have been earlier. This fits in well with an introduction of spores during the May 20 – May 24 period, when the low pressure in the Gulf drew airborne spores westward from the sources in FL.

Periods of favorable weather during June gave rise to increased disease activity in the Southeast, localized and short-range spread near the known sources, and extension of the epidemic northeastward along the East Coast.

The first reports from SC came from Charleston County on June 15 (butternut squash in the sentinel plot) and Clarendon County June 18 (cucumber in a commercial field). First symptoms were estimated to have occurred on June 13 and June 12, respectively. The forecasts of June 5 and 6, with Moderate risk to SC cucurbits, yield the best explanation for the two outbreaks in SC. Either or both events may have caused the outbreaks. The June 5 event from southern GA is shown below.

sGAnFL traj

Virginia’s first report came from Suffolk County in the southeastern portion of the state. CDM was observed on cucumber in the local sentinel plot and reported on June 19th. First symptoms were estimated to have appeared approximately 3 days earlier. The High risk events of June 11 were the most likely culprits for this outbreak, as well as for those from Dinwiddie and Northampton Counties reported several days later. The events of June 9 (Moderate risk for southeast VA) may have contributed, as well. Airborne spores from Columbus County, NC were likely involved, though that source would not be discovered / reported until June 12. The June 11 event from the known NC source is shown below.

neNC traj


CDM was confirmed in MD in a commercial cucumber field and reported on July 6. Dorchester County reported on July 9. First symptoms were estimated at June 30 in both locations. The most likely scenario for epidemic spread to this area is by way of the forecast transport events of June 17 and 18 from eastern NC (Weakly Moderate risk one day, Strongly Moderate risk the other). The June 17 event is illustrated below.

seNC traj


Attention now turned to the west. CDM was discovered in Chatham-Kent County in southern Ontario in a commercial field of cucumber. The outbreak was reported on June 29. First symptoms were estimated to have occurred on June 25.  It is possible that the disease was present prior to this date in this region. As in previous years, thorough examination of the evidence has not yielded a plausible method of epidemic spread to this area via transport from known sources. It is strongly suspected that some other method of introduction was involved, whether by transport from unknown or unreported sources or by some other means.

The dominant theme during the heart of the summer and for the rest of the 2009 epidemic was a familiar one. The CDM epidemic spread in the Great Lakes region via a number of short-range events, initially to OH and MI and then to surrounding regions. New outbreaks occurred in the Mid-Atlantic and Northeast during periods of favorable weather. Some of these outbreaks were due to transport from the south, others were due to eastward transport from the expanding Great Lakes infections. Isolated outbreaks arose in the Ohio Valley and Midwest.

The first report from OH was in a commercial field of cucumber in Sandusky County. The report was received on June 30 with first symptoms estimated to have appeared on June 29. This outbreak, and likely the second OH find in Henry County (July 6 report, July 4 first symptoms), were most likely the result of short-range transport from southern Ontario in late June.  

Similarly, the initial reports from MI appear to have been mostly due to short-range transport from nearby infections in southern Ontario and/or northern OH. Three counties reported CDM on July 14, all from the eastern side of lower MI and all on cucumber: Monroe (home garden), Saginaw, and Tuscola (large commercial fields for the latter 2). First symptoms were very recent in all cases. Viable spores may have been deposited on one or more occasions. Weather conditions varied between unfavorable and mixed during early July. Evidence is insufficient to definitively select one event over another.

The first report from IL in 2009 was from a home garden in Hancock County (west-central IL). CDM was confirmed on cucumber on July 23. Origins of this outbreak are unknown.

The first outbreak of CDM in PA was reported on July 13 in Westmoreland County (southwest PA). Symptoms were estimated to have first appeared in this commercial cucumber field on July 7. By far, the most plausible candidates for this outbreak were the forecast transport events of July 1 (Moderate risk western PA) and July 2 (Strongly Moderate risk). Spores likely originated from the infections in southern Ontario and/or northern OH. Conditions were mixed to favorable. The July 1 event from that region is shown below.

sONnOH traj


The first outbreak from WI was reported on July 31. Disease symptoms were just beginning to appear on cucumber in a research plot near the Columbia / Dane County border. This outbreak has several potential origins. The most probable is transport from sources in MI or northern OH on July 20/21. The northern OH event is shown below (Weakly Moderate risk). The active but not yet discovered source(s) in southwest lower MI likely played a large role, as well.

nOH traj


To the East, the first outbreaks in DE and NJ were reported on July 13 and 14, respectively. The Sussex County, DE outbreak (sentinel plot, cucumber, first symptoms July 13) arose from transport from the nearby MD sources. The Cumberland County, NJ outbreak (fields, cucumber, first symptoms July 9) was due to transport and deposition from the MD sources on July 5. The MD sources were not discovered or reported until July 6, so they were not yet part of the forecast rotation. Had they been, elevated risk to NJ cucurbits would have been indicated. This July 5 event is shown below.

eMD traj


The July 9 transport events from eastern NC were the most probable cause of the first outbreaks in west-central NC (Rowan County) and western NC (Henderson County). The Rowan County find was on cucumber in the sentinel plot, reported on July 14. At the time of the report, symptoms were just beginning to appear. Moderate risk was forecast for cucurbits in this region. The Henderson County outbreak, also on cucumber in a sentinel plot, was reported on July 20. First symptoms were estimated at July 18. This area was at Low risk. There is far less support for other transport episodes in explanation of these infections. The July 9 events are shown below.

eNC trajs

The first report from NY arrived from Suffolk County (Long Island) on July 28. Cucumbers in the sentinel plot were estimated to have shown first symptoms one week prior to the report. The events of July 16 hold the most weight regarding the origin of this outbreak. Moderate risk was forecast for cucurbits up and down the eastern Seaboard; nearly all the transport events from the mid-Atlantic indicated transport toward that region. One of these events is shown below.

DE traj


The months of August, September, and October saw the 2009 epidemic reach maturity and then wane as autumn gathered strength and the growing season wound down. New outbreaks occurred in the Northeast, including the first reports from MA (Hampshire County, commercial cucumbers) and CT (Hartford County, commercial cucumbers). Scattered reports were received from the OH Valley and TN Valley regions, with some epidemic spread occurring before the growing season ended.

The first outbreak in TN was in the Jackson County sentinel plot on cucumber. It was reported on July 31, with first symptoms estimated the same day. Spores most likely arrived from sources to the south. However, evidence for this is scant and no definitive candidate transport events have emerged to support its origination.

The first outbreak in KY was in McLean County (western KY), on cucumber in the sentinel plot. It was reported on August 19. First symptoms were estimated at August 14. The first outbreak in IN was in Knox County, in the southwest portion of the state. CDM was reported on cucumber in a research plot on August 10, with first symptoms estimated at August 7. The transport event of August 3 out of western TN may have been the origin of both of these infections. Subsequent outbreaks in central and eastern sections of TN and KY were likely due to these initial sources farther west. Conditions were mixed for this August 3 event, shown below.

wTN traj


By mid-October, frosts, freezes, and lack of host had ended the disease season over the northern areas. Some CDM activity persisted in the southern regions. Official forecasts ended on October 30.

The 2009 CDM epidemic was large in scope, affecting nearly all of the major growing regions. The epidemic gathered momentum in late May in the Southeast, when a period of favorable weather conditions coincided with increased host availability. The CDM ipmPIPE center received over 200 unique reports from over 150 counties in the United States and southern Canada during the growing season. Counties with confirmed infections were clustered more heavily in the Mid-Atlantic and Great Lakes, but were more scattered in the Northeast and the Ohio Valley southward to the Gulf Coast. Only one county reported CDM from the southern Plains region. Other recent epidemics have featured fewer reports and some regional differences, though the general prominence of reports from the Great Lakes and Mid-Atlantic has been consistent.