Forecasting Long-distance Movement of Cucurbit
Downy Mildew

G.J. Holmes and C.E. Main

Department of Plant Pathology, North Carolina State University, Raleigh, NC 27695

T.Z. Keever

Marine, Earth and Atmospheric Sciences, North Carolina State University, Raleigh, NC 27695

Additional index words. Pseudoperonospora cubensis, spore movement, epidemiology, NOAA

Abstract. Downy mildew is an important disease of cucurbits worldwide. The pathogen (Pseudoperonospora cubensis) is believed not to overwinter in the north, but rather survives winters in the south and is reintroduced from there to the north each year. A system has been developed to forecast the long-distance movement of P. cubensis spores by way of large-scale weather systems affecting the source(s) of inoculum. The system depends on actual observations of qualified plant pathologists to confirm disease outbreaks and utilizes weather forecasting models and pathogen epidemiology to determine the likelihood of disease outbreaks along weather trajectories. This is the first year the system is being evaluated. As of 25 Sept. there have been 6 disease reports and 29 forecasts. Forecasts predicted three high-risk scenarios, 33 moderate-risk scenarios, and 21 low-risk scenarios.


Downy mildew, caused by Pseudo- peronospora cubensis, is an importantdisease of cucurbits that occurs worldwide. The pathogen is not believed to overwinter north of the 30th latitude, but rather is introduced annually from southern latitudes where it can survive winters. (The 30th latitude passes near Gainesville, Fla.; San Antonio, Tex.; and El Rosario, Baja, Calif.) While local spread of the disease is dependent on local weather conditions, long-distance movement of the pathogen occurs associated with large-scale weather systems. This paper describes a disease forecasting system that tracks long-distance spore movement of P. cubensis from known sources of inoculum to other potential sites of infection and discusses its advantages and disadvantages. This is the first year the system is being evaluated and a full report was not available at the time this was written since several weeks of the fall season remained.

Materials and methods

Disease reports. The forecasting system is dependent on timely and well-documented reports of new downy mildew occurrences. A network of 40 state representatives (mostly plant

pathologists and horticulturists) from the U.S. and Mexico was established. Each representative was asked to report any occurrence of the disease in their area via an electronic report form on the World Wide Web ( The web-based disease report asks for important information including the geographic location of the source, and source characteristics. Multiple source sites grouped closely together (e.g., within the same county) are reported as a single, central location. If a source ceases to exist, this information should also be reported. The forecaster uses the most recently reported and important continuing source sites to initiate each new set of forecast trajectories. Forecasts are typically made in the afternoon on Tuesdays and Thursdays of each week. If an urgent situation presents itself, forecasts can be made available more frequently.

Forecast model. Spore transport in the atmosphere is calculated using the HY-SPLIT trajectory model from NOAA's Air Resources Laboratory (ARL) in Silver Spring, Md. For the necessary meteorological input, HY-SPLIT commonly uses outputs of either the ETA model or the AVN (Aviation) model, members of the family of Nu

Cucurbitaceae '98

merical Weather Prediction (NWP) models used to forecast short-term weather conditions in and around the United States. These meteorological data are generated initially by the National Center for Environment Prediction (NCEP), a branch of NOAA, then modified by ARL to conform to HY-SPLIT's input requirements. The data of primary interest are the forecast wind fields in the atmospheric boundary layer. In nearly all cases, HY-SPLIT trajectories are provided by ARL via an automated Electronic Mail system. The trajectory is a plot of the future atmospheric pathway of a parcel of air likely to contain spores; in other words, the prediction of the spatial and temporal positions of a spore cloud CENTER for the future 2 days, following release from a source site.

Outlook. Each day's forecasts are archived on the Website by year, date and source location. A current forecast includes maps of the latest trajectories available. A trajectory weather section describes the recent past, present, and near future weather conditions at the source and along the anticipated spore pathway. The following are factors important to sporulation at the source: temperature, rainfall, cloud cover, survivability during transport (cloud cover related to UV radiation and desiccation effects), and deposition from rainout and washout (potential rainfall). General weather conditions in the United States that may influence the movement of spores or provide opportunities for subsequent infection are discussed in a regional weather section. An Outlook section combines these elements describing the likelihood of inoculum spread and disease risk, 48 hours into the future.

Map description. The source site is represented on the map by a geographic point (a red dot). The time-labeled dashes on the pathway represent the spore cloud position at 6-h intervals. Chronological time is given in universal time (UTC; formerly known as Greenwich mean time; subtract 4 h to get eastern daylight time). Header labels detail the time and date of the trajectory start (14Z 02 JUN UTC), and the time and date of the meteorological forecast file used to compute the trajectory (00Z 02 JUN model name FORECAST INITIALIZATION). The latitude and longitude of the disease site used for the source is given along

the left edge of the map.

At the terminal end of the trajectory on the regional map is a pressure value (in parentheses) which indicates the altitude of the source from which the trajectory began. The small rectangular graph underneath the map shows the vertical motion of the spore cloud center. This vertical motion is determined by using the vertical velocity output of the ETA or AVN model (from NCEP). The solid line represents the pathway, with the dashes on the solid line corresponding to the time dashes on the horizontal trajectory shown on the map. To the left of the vertical pathway are several numbers and a label reading HPA. The numbers are values in hectaPascals, which are numerically equivalent to millibars. Both are measures of atmospheric pressure, a common method to indicate altitude. The values decrease as height increases, since air pressure decreases with altitude. Average sea-level pressure is 1013 millibars (or hectaPascals).

Results and discussion

Many factors are involved in predicting the future movement of fungal spores and the associated weather conditions. The HY-SPLIT model represents state-of-the-art atmospheric trajectory analysis and reflects the accuracy of the meteorological data (i.e., winds). Good HY-SPLIT trajectories may deviate from the true path by about 15% of the transport distance; many trajectories are off from 30% to 35%. Complex weather situations can produce errors of 50% or more. Research with HY-SPLIT and similar models using historical data of actual epidemic spread provides good qualitative guidelines for evaluating the accuracy and usefulness of the forecasts. Each downy mildew forecast includes a measure of the anticipated confidence in its trajectory pathway. It is important that users of this information also continue to pay close attention to local weather conditions associated with the forecasts.

This year, downy mildew was first reported on squash near Immokalee, Fla., on 30 Mar. Since then, 6 reports of the disease have been received and 29 forecasts issued as of 25 Sept. From these forecasts, 3 high-risk situations were predicted (all in Fla.), 33 moderate-risk situations, and 21 low-

Cucurbitaceae '98

risk situations. Probably, long-range transport scenarios are few. Most forecasts involve localized and/or short-range disease development potential. This year has been exceptionally dry and may not be representative of a typical year.

The limiting factors in the forecasting system are field scouting and reporting. The benefits are 48-h advance warning of potential disease outbreaks, and sensitivity of the fungal population to fungicides.

The system requires minimal input once the models are running and the Website has been established. The Website contains a great deal of information related to downy mildew (e.g., epidemiology, management guidelines, diagnostic guide, list of state representatives and other im

portant links). We expect this effort to continue, expand and improve each year as users recognize its benefits and become familiar with the system.

Literature cited

Davis, J.M. and C.E. Main. 1984. A regional analysis of the meteorological aspects of the spread and development of blue mold on tobacco. Boundary-Layer Meteorol. 28:271­304.

Davis, J.M. and C.E. Main. 1986. Applying atmospheric trajectory analysis to problems in epidemiology. Plant Dis. 70:490­497.

Davis, J.M. and C.E. Main. 1989. The aerobiology of the sporangiospore of Peronospora tabacina, p. 264­267. In: Proc. 18th Conf. Agr. and For. Meteorol., Amer. Meteorol. Soc., 7­10 Mar., Charleston, S.C.

Cucurbitaceae '98