Thomas Keever: Forecaster, North American Blue Mold Forecasting System
During March and April of 1996, forecast trajectories and outlooks were produced for a number of source areas in Mexico, southern Texas, Cuba, and Honduras. Some were selected for their general interest to the tobacco-growing community, but the Texas and Cuba locations were of special importance because blue mold was active in these areas and both might pose a threat to the major U.S. production regions.
The threat posed by these sources during this early period tended to be highly localized with the long-range threat subdued compared to a 'typical' year. Of the 101 risk-rated outlooks, 91 were Low risk and the other ten were rated Moderate. Further, there was only a few times when conditions were favorable for disease development in Florida via long-range transport and deposition. Instead of the disease becoming firmly established in the eastern U.S. in March or April, the first official report of blue mold was May 9th!
The most likely reason for this delay in the introduction of the disease to the eastern U. S. is the jet stream pattern that became established in the fall of 1995 and persisted through the spring of 1996. During this time, these upper-level, high-velocity wind currents were flowing across the northwestern states, dipping southeastward through the central plains and into the southeast U.S., then turning to the northeast over the coastal states. Since the surface weather features tend to follow the upper-level currents, this allowed the frequent introduction of cool, dry air into the Southeast behind the surface frontal systems that passed through the region. The general weather scenario, especially during April, was one in which a high pressure area centered over Georgia or Florida would remain in place for days at a time. Then, as a frontal system guided by the upper-level steering currents approached the region, there would be some rain or showers for a short while followed by more dry weather. Not only were the dry conditions generally unsuitable for disease development, but the clockwise circulation around Highs centered near the region meant that surface winds were seldom directed toward the production areas. Winds around the Cuban sources blew frequently from the east, pushing trajectories into the Gulf of Mexico; winds in the south Texas region were often from the southeast or south, so airborne spores typically travelled into the central plains. The quite fortunate result was that conditions were generally unfavorable for disease development in south Georgia or Florida during March and April and the opportunities for survivable spore transport into the region were pretty limited.
It's difficult to determine when blue mold first gained a foothold in Florida. There was evidence at the time of the first official report that the disease had been present previously in north Florida, but the activity was extremely light and had gone undetected. In addition, there is strong evidence to support the theory that blue mold was introduced to Kentucky this season by seedlings brought in from Florida transplant operations in late April. Given these facts, it seems likely that there was some successful spore transport, deposition, and infection of Florida tobacco around the time of one (or more) of the rain events in March or April. However, the generally unfavorable conditions kept the disease from developing beyond a few scattered lesions, and its presence went unnoticed until May.
A review of the disease outlooks from this early-season period reveals several instances when the disease may have had an opportunity to establish itself in Florida via long-range transport. The first is from 14 March 96. Spores released from infected wild tobacco in south Texas may have survived transport eastward into south Georgia or north Florida. Blue mold there was at its strongest at this time and this was the one occasion this season when spores from south Texas may have been carried that far east. (Incidently, the pervasive drought conditions this spring in the southern plains and the Southwest worked to destroy the blue mold on the wild tobacco there; the disease presence declined steadily through the remainder of the season.) Other opportunities for transport and effective deposition in Florida may have occurred the last weekend in March, this time from Cuba. A front moved through the Southeast and stalled around north Florida, dumping copious amount of rain in the northern half of the state. No trajectories were run that weekend, but the possibility definitely exists and any spores rained out would have encountered favorable conditions for infection. Lastly, the most salient example in April is from the forecasts of the 29th. Trajectories from Cuba led directly to Florida's production areas, where conditions were again favorable for disease development.
Looking at the biology of the situation, it seems that one of these last two occasions was the most likely culprit. Conditions were not very favorable for long-range transport and later deposition during the mid-March episode from southern Texas. Perhaps more important, though, is the fact that the blue mold found in southern Texas is quite probably Ridomil-sensitive, since there is no selection pressure in that region to develop any resistance to the product. Florida growers typically use Ridomil during the months of March and April. Thus, any spores arriving from southern Texas would have been stopped. Spores coming from Cuba in late March or late April, on the other hand, may very well have been resistant to Ridomil, so the use of the product couldn't stop the disease.
May was a month of transition for the forecasts and for disease development. Blue mold was reported in Florida on May 9th. With this confirmation of the disease in a major production area, the Central America and Mexico source locations were dropped from the regular outlooks. However, since the activity in Florida was light, one source site was retained in each of Cuba and south Texas. Even with the confirmation of blue mold in Florida, the disease situation during the first several weeks of May was apparently benign. The weather was mostly dry in the Southeast, nearly all risks were low, and there were no High Risk outlooks. This all changed, though, around the 21st. A tropical disturbance moved across Florida and brought favorable development conditions to that state and to south Georgia. The next day, blue mold was reported in Kentucky, the first news of the disease in the mid-Atlantic/southern Ohio Valley growing region. Near the 24th a stationary front became established in that same region, beginning a month-long period of very favorable weather for disease development. Finally, on the 26th, Dr. William Nesmith confirmed the presence of blue mold in six Kentucky counties while raising the possibility that blue mold might be widely distributed there. He stated that the disease had been present in greenhouse or outdoor-float systems for three to five weeks and sporulation had been occurring for at least two weeks.
The events of late May signaled the beginning of the next disease development phase, which has continued through mid-August. With the exception of later June and early July, it has been characterized by mostly favorable weather conditions for blue mold development and the appearance of the disease in various locales in the eastern U.S. growing region, especially the burley production areas. Trajectories were run from reported source areas in west-central, east-central, central, and southern Kentucky during late May and June. By June 19th, blue mold was present in 31 counties in central Kentucky so the centralized location of Lexington was chosen to represent them all. A source site was added for southeast Georgia (Hortense) on June 5th. Eastern North Carolina became involved on June 10th after blue mold was reported in Robeson County; the disease was confirmed in several other nearby counties later in the month, and trajectories were also run for Edgecombe County in northeast North Carolina beginning the 26th. In early July, source sites were added for the burley-growing areas of the southern Appalachian Mountains. Northeast Tennessee was added on the 3rd, Yancey County, NC, on the 8th, and northeast Georgia at Young Harris, on the 10th. The disease continued to develop in the mountains, involving many counties over time. Blue mold was also discovered in southwest Ontario in early July, with a few forecasts prepared for this site in August.
Risk assessments for late May and beyond reflected the favorable environment for disease development. Of the 166 source outlooks produced in June, July, and August, 30% were High Risk and about 40% were rated Moderate, with around 30% Low. This is drastically different from those produced in March through May (also 166, ignoring 12 others for which the risk was uncertain), when Low risk sources accounted for over 85% of the total. The Kentucky and north Florida/south Georgia source regions dominated the High Risk scenarios during late May and June. In July and August, most High Risk situations occurred for the eastern North Carolina sources or for the three disease sites in the southern Appalachian Mountains. The outlook from central Kentucky was High Risk several times, though, and was rated Moderate quite frequently.
The timing of blue mold's introduction to southeast Georgia and southeast North Carolina is not readily discernable from consideration of the outlooks. Georgia's occurrence (reported June 1) almost assuredly originated with spores from north Florida and effective deposition and infection could have taken place at any time during middle or late May. A similar situation exists for southeast North Carolina's outbreak, which was reported about a week later. Conditions were only occasionally favorable for development along the coastal areas leading up to the time of discovery. Different scenarios are possible; infection with spores from Florida may have occurred earlier in May and gone undetected because the activity was so light, or transport and infection may have occurred later in the month and been discovered soon after lesions became active. It's highly unlikely, though, that spores from Kentucky or some other location were carried to this area and deposited while still alive. One additional note: other than the handful of counties in southeast North Carolina and Edgecombe County in northeast North Carolina, a large majority of the flue-cured growers in North Carolina and Virginia went unscathed by blue mold this year. This is somewhat surprising to your Forecaster. Although the weather in June and July was not as favorable for development here as it was in Kentucky, the conditions were pretty similar. Given that the southeast NC counties were well-positioned to send spores to other growing areas via southerly or southwesterly winds, it seems odd that there wasn't more disease development, especially in nearby counties of North Carolina. Perhaps the initially light activity remained light through the period, and that, combined with preventative efforts by the growers, kept blue mold from advancing.
The appearance of blue mold in the burley production areas of Kentucky coupled with the beginning of favorable weather conditions for disease development produced a very unfortunate situation in late May. Considering the discovery dates of blue mold in various locales, and from that, the possible arrival times of inoculum that may have been responsible for the ensuing disease development, it appears that airborne spores from central Kentucky later infected a number of the burley growing areas in the Ohio Valley, southern Appalachian Mountains, and Canada. In Kentucky itself, localized and short-range development occurred steadily in June and most of July, then escalated into a major epidemic during later July into early August. Dr. William Nesmith, Plant Pathologist at UK, initially reported blue mold in six counties on May 22nd. By June 19, that number had grown to 31, and by mid-July the disease was widespread and quite strong throughout central Kentucky and present elsewhere. At the height of the epidemic, more than 100 counties were reporting active blue mold!
Blue mold development in the burley-growing regions of Indiana and Ohio could have begun at any time from late May onward. Forecast trajectories often led over these areas and the weather was favorable for disease development. The most likely period was early to mid-June, when blue mold was becoming more widely established in Kentucky and the number of airborne spores was increasing rapidly.
Unlike the disease development along the east coast, examination of the forecasts/outlooks reveals several significant instances in which inoculum was likely introduced to Canada and the southern Appalachian Mountains. Blue mold was reported in southwestern Ontario, Canada, at the beginning of July, with the oldest lesions appearing to be several weeks old. This correlates well with the forecasts of June 5th and 6th from Kentucky. A slow-moving cold front was approaching from the west. Conditions were very favorable for survivable transport and effective deposition ahead of the front, and the spore cloud centers from Cynthiana and Elizabethtown were forecast to go directly over the Canadian growing areas. It had also been two weeks since the initial report of blue mold in Kentucky, so it's likely that infected fields in a number of counties were sending spores in this direction.
Another likely scenario for blue mold development occurred the next week, on June 13th. At this time, trajectories for four source areas in Kentucky were generated, and all of them indicated pathways leading to the east and then southeast, which would take any airborne spores across the burley regions of the southern Appalachian Mountains. As stated above, 31 counties in Kentucky were reporting blue mold on the 19th, so one week before that, the disease was likely present in many areas. In addition, there was only one other occasion when trajectories from Kentucky were moving in that direction before the disease was reported in northeast Tennessee and the central NC mountains (July 3rd and July 6-7th, respectively); and, in that case, conditions were mostly unfavorable for development. So, although the risk assessments at the time were three Lows and one Moderate, it's very likely that this was indeed when the first inoculum was introduced to the mountain growing areas.
Once blue mold was present in the mountains, there were plenty of opportunities for further disease development due to the generally favorable conditions beyond the first week of July. At least once a week, the three mountain sources were rated High Risk. However, nearly all of these cases involved localized or short-range threats. If any long-range transport and deposition did occur from the southern Appalachians, there was no apparent disease development resulting from it. Other outbreaks that occurred in southwest Virginia, Pennsylvania, and even Missouri were most likely due to spores released from Kentucky.
The regular forecasting schedule was discontinued after August 8th. Many growers had harvested or were near-harvest, and blue mold was present throughout much of the remaining growing areas. There were a total of 344 outlooks produced, newly available on 63 days, over a span of almost five months.
Return directly to