Summary of the 2010 Tobacco Blue Mold Epidemic in North America
Thomas Keever, Lead Forecaster
North American Blue Mold Forecasting System
NCSU, Department of Plant Pathology
The Tobacco Blue Mold Forecasting System has concluded its 15th year of forecasting epidemics in North America. Forecasts were produced at regular intervals from March through mid-September. The forecasting website provided visitors with forecasts of potential epidemic spread and up-to-date information about the status of the epidemic. We hope the forecasts have proved useful to the growing community via improved crop management and assessment of the possibility of developing tobacco blue mold. Two hundred thirty-five forecasts were produced on 73 days during the growing season.
In terms of the number of counties with confirmed disease, and likely also in terms of disease incidence and impacts, 2010 was one of the lightest continental epidemics since full-season forecasts began in 1996. Nineteen counties in six states reported blue mold, including North Carolina (12 counties), Virginia (1 county), Kentucky (1), Pennsylvania (2), Connecticut (2), and Massachusetts (1). These outbreaks were found in 5 growing regions: eastern NC, the southern Appalachian Mountains, central KY, southeast PA, and southern New England. The reports from eastern NC dominated the map.
Meteorologically, the 2010 season was dominated by dry, hot conditions in the tobacco-growing regions east of the Rocky Mountains, especially during the summer months. Some areas did receive normal or above-normal rainfall for short periods, only to lapse into drier conditions. Many locations set or neared records for sustained heat. The reason is the pattern that became established in the middle and upper levels of the atmosphere in June and persisted through the summer. A trough of low pressure existed in the western U.S. (tendency toward wetter, cooler weather) while a ridge of high pressure situated itself over the southern / eastern U.S. (tendency toward drier, hotter weather). This pattern varied and/or relaxed for short intervals, but the overall theme was unrelenting and lasted until the last week of September. The drier, warmer conditions plus good grower practices likely contributed heavily to the low severity of the 2010 epidemic.
There was no report from the northern FL / southern GA region, which has become typical (only 1 year in the past seven has included a report from this area). There were several instances early in the growing season in which spores may have been brought to this region from Cuba. Some transport episodes were similar to those of past years, in which conditions for survivable transport were unfavorable but the arrival of any airborne spores (alive or dead) would have coincided with rainfall. These events presented Low Risk to U.S. tobacco. For two other events, risk to tobacco in northern FL was Weakly Moderate or Moderate. Disease severity in Cuba was unknown / uncertain in 2010. As became apparent, none of these episodes enabled blue mold to get a foothold in the southernmost growing areas.
The first report of blue mold in the U.S. in 2010 came from Wilson County, NC on May 28. Estimated date of first symptoms was May 25. During the next week, numerous additional reports arrived in the Forecast Center spread across five counties in eastern NC. Each indicated symptoms beginning during the last week of May. Three other reports from three new counties in NC arrived between June 8 and June 16. First symptoms from these outbreaks occurred between May 30 and June 8. The timing of all these outbreaks coincided with the one wetter-than-normal period in eastern NC that lasted from mid-May through the first week of June.
The origin of the NC outbreaks remains undetermined. Presuming an airborne introduction of spores and based on the information from the field, extensive study was undertaken to determine from where any spores may have come. There were no other known sources in the U.S. at that time. Spores were not introduced from Cuba. Trajectory analysis did not reveal the potential for transport from other growing areas, where disease may have existed but been undiscovered or appeared earlier than estimated, as well.
Blue mold was reported on June 22 from both southern New England (Hartford and Tolland Counties, CT) and southeastern PA (Lancaster County). Estimated date of first symptoms was June 6 for the CT outbreaks, June 14 for the PA outbreak. Grower response to the presence of blue mold was strong in both cases once disease was confirmed.
The disease report for the CT outbreaks indicated that disease was observed in the field and in a greenhouse, with inoculum estimated to have been introduced first in the greenhouse in late May. Examination of the outbreak lends strong support to this assessment. Each day from May 30 through June 1, transport events from eastern NC moved to the northeast, with airborne spores brushing by central CT late in each event. Conditions for survivable transport varied for the 3 events. Showers occurred over the outbreak region during two of the events. No other events approached this region during an appropriate time frame. The NC sources, having just been reported, were not yet in the forecast rotation, but later analysis indicates these events would have produced Low or Weakly Moderate Risk to the CT growing region.
The keys in this instance are disease becoming established first in the greenhouse and the fact that there were several transport events in a row that moved near/over this region. It is easier for disease to become established in a greenhouse; all it requires is for some lives spores to become trapped within the structure long enough to come to rest on the host. In addition, experience has shown that the risk of epidemic spread to a given location can increase substantially when events clustered closely in time traverse similar paths to each other. This is true even when conditions for each event are variable, or even unfavorable. Both of these keys were present for the CT outbreak situation. A simulation of one of these events is shown below, from May 30.
The origin of the PA outbreak is fairly straightforward in terms of event identification. Given the field observations, timing of the first symptoms, and the lack of any other events in the appropriate period, airborne inoculum was introduced via the events of June 8 and/or June 9 from the sources in eastern NC. The June 8 events encountered less favorable conditions (Low Risk in the forecasts) but had greater spore deposition; the June 9 events had more favorable conditions but lesser spore coverage. Deposition from the June 8 event from NC is shown below.
Reports of outbreaks in Lee and Moore Counties, NC arrived at the Forecast Center June 25. First symptoms were estimated around June 14. Multiple farms were affected in each county, and both old and new lesions were present. Live spores were almost certainly deposited from events originating in the many NC sources just to the east. This may have occurred on one or more occasions in early June. Possible events include those of June 5 (Weakly Moderate Risk), June 6 (Moderate Risk), or June 8/9 (Low/Moderate, illustrated above). The events of June 15 (Weakly Moderate) may have contributed as well.
A new growing region was involved in the next outbreak reported to the Forecast Center, Alleghany County, NC (very close to the Ashe County line) in the southern Appalachian Mountains. The report arrived on July 7; estimated date of first symptoms was June 22. Unfortunately, as was the case with the initial eastern NC outbreak, the origin of this infection is unknown. No transport event from any known source, or from a source that became known later in the season, drew near enough to northwest NC to warrant serious consideration. Backwards-in-time trajectories, generated for many days surrounding the suspected infection date, consistently indicated transport into this area from the northwest, west, or southwest. No tobacco is grown in many of these potential origin locations. In those areas that grow tobacco (TN, KY, southern OH, etc), only one county (in KY) reported blue mold in 2010, and initial infection there occurred many weeks later than the one in Alleghany County. In short, no plausible scenario exists that would explain the spread of blue mold into northwest NC in 2010, based on the information available.
The first reports from Massachusetts (Hampshire County) and Virginia (Washington County) came into the Forecast Center on July 26. Estimated dates of first symptoms were July 19 and July 20, respectively. Both of these outbreaks occurred near existing sources in CT and northwest NC, respectively. Short-range transport on one or more days from July 9 through July 13 was almost certainly the origin of these infections. Forecasts on July 8, 10, and 13 described the risks to tobacco in these growing regions, which was Moderate or higher at least once during that period.
The last major growing region affected in 2010 was the Ohio Valley. Blue mold was confirmed in Shelby County, KY and reported on July 31. Estimated first date of symptoms was July 17. Less than 1% of a 17 acre field was affected. This was the only county west of the Appalachians to report blue mold this year, which was likely another first during the Internet forecasting era. Despite intensive analysis, the origin of this outbreak is as elusive as that of the eastern NC or southern Appalachian Mountain infections. The same themes apply: no evidence of transport from known sources, transport that would have had to occur from regions with no reported disease or no tobacco, and rainfall at the outbreak site within a suitable time frame to produce the observed symptoms but no corresponding arrival of live spores.
The last report of 2010 arrived on August 5 from Chester County, PA. Tobacco in this county adjacent to Lancaster County first showed symptoms around July 26. Infection occurred most likely during the favorable weather of July 19, 20, and 21st. Tobacco here was at Moderate to High Risk in that period.
The 2010 disease season was the lightest on record since full-time Internet forecasting began in 1996. Only 19 counties in the U.S. reported blue mold; most outbreaks were non-severe. Growing regions in the GA/FL region, the southern Appalachian Mountains, and the Ohio Valley were either unaffected or minimally affected. Hot, dry weather dominated the eastern U.S. growing regions in 2010 with few periods of disease-favorable weather. Disease reports and other information from the field indicated good grower practices in general and decisive response to the presence of blue mold once discovered. The overall unfavorable weather and the behavior of the growing community were likely the most significant contributors to the gentleness of the 2010 disease season. After the 11 counties in eastern NC were reported by mid-June, only 8 more counties in the entire eastern U.S. and southern Canada reported blue mold for the rest of the growing season. Tracking this light epidemic proved difficult. The first U.S. report (eastern NC) was late once again (late May) with no discernable origin. Epidemic spread to other regions was at times reasonably explainable (southern New England, southeast PA) and in other cases mysterious (southern Appalachian Mountains, Ohio Valley; no airborne introduction plausible).
We wish to thank all of our cooperators in the U.S. and abroad for their participation in 2010. If anyone would like more information about TBM forecasting or other aspects of the program, please contact Dr. Mila or the Forecast Center.
Dr. Mina Mila
Asst. Professor, Tobacco Specialist
mina_mila@ncsu.edu
Thomas Keever
Lead Forecaster / Meteorologist
thomas_keever@ncsu.edu