Predictive Models for Turf Grass Diseases

Paul Vincelli, Extension Plant Pathologist(pvincell@uky.edu)

University of Kentucky

These models attempt to provide weather-based forecasts of several important turfgrass diseases in Kentucky. All are based on research published in scientific journals and are thus considered to be based on sound science. However, use the predictions of these models with a certain degree of caution, for the following reasons.

1. Valid models that are effective on some sites may not perform well in others.

2. All of these models are weather-based; none can account for how one=s management practices influence disease, which can have a profound impact on disease development.

3. None of these models has yet received adequate validation under Kentucky conditions. In fact, this is one of our objectives in making these available to the public.

Users of these models are invited to provide comments on predictive successes or failures of these models to forecast disease activity to

Brown Patch Model #1

Forecasts brown patch on perennial ryegrass. Limited evaluations suggest that this model (as modified slightly from the original publication) is the accurate for predicting brown patch activity on perennial ryegrass for conditions in Central Kentucky. However, these evaluations also suggest that the model is not accurate for predicting brown patch activity on creeping bentgrass or tall fescue.

Algorithm:

E = -21.5 + 0.15 RH + 1.4T - 0.033T², in which RH is the mean relative humidity and T is the minimum daily air temperature. As originally published, a threshold value of E > 6 constituted a brown patch warning. My evaluations in 1999 suggest that we will have greater forecasting accuracy with a threshold value of 5.

Reference: Fidanza, M.. A., Dernoeden, P. H., and Grybauskas, A. P. 1996. Development and field validation of a brown patch warning model for perennial ryegrass turf. Phytopathology 86:385-390.

Brown patch Model #2

Forecasts brown patch disease on cool-season grasses. Limited evaluations in Kentucky suggest that this model issues too many Afalse negatives@. A false negative is a forecast of Ano disease@ when, in fact, disease occurs.

Algorithm:

Relative Humidity: >95% (=leaf wetness) for 10 hr or more

Soil Temperatures: mean >70^o F (21^o C); minimum > 64^o F (18^o C)

Air Temperatures: mean > 68^o F (20^o C); minimum > 59^o F (15^o C)

Rainfall/Irrigation: 0.1 inches (3 mm) received in the 36 hr preceding the tenth

hour of high relative humidity.

Temperatures apply to the 24 hr preceding the 10^th hour of high relative humidity. Disease can also occur at somewhat lower air temperatures following heavy rainfall and extended hours of leaf wetness. If temperature falls below 59^o F (15^o C) in the 48 hr following the environmental conditions necessary for disease development, no disease will occur.

Reference: Schumann, G. L., Vittum, P. J., Elliott, M. L., and Cobb, P. P. 1997. IPM Handbook for Golf Courses. Ann Arbor Press, Inc, Chelsea, MI.

Pythium Model #1

Forecasts Pythium cottony blight on cool-season grasses. In Kentucky, typically this disease is of greatest concern on high-maintenance creeping bentgrass and perennial ryegrass. This model was developed and evaluated in Ohio. Research there and in Kentucky indicates that it is rather conservative, in that it has a high chance of forecasting subsequent Pythium outbreaks. However, it also has a somewhat high rate of Afalse positives@, where the disease is forecasted but no disease occurs.

Algorithm: Predicts cottony blight activity on a day (noon of the previous day to noon of the current day) when the maximum air temperature > 82^o F, the minimum air temperature > 68^o F, and hours of relative humidity > 90% or hours leaves wet > 9.

Reference: Shane, W. W. 1994. Use of disease models for turfgrass management decisions. Pages 397-404 in Handbook of Integrated Pest Management for Turf and Ornamentals. Anne R. Leslie, ed. CRC Press. 660 pages.

Pythium Model #2

Forecasts Pythium cottony blight on cool-season grasses. In Kentucky, typically this disease is of greatest concern on high-maintenance creeping bentgrass and perennial ryegrass. This model was developed and evaluated in Pennsylvania. Research in Ohio suggests that this model is more likely than Pythium Model #1 to issue a Afalse negative@ forecast. A false negative is a forecast of Ano disease@ when, in fact, disease occurs. However, it is more likely than Pythium Model #1 to identify days when Pythium disease activity is unlikely.

Algorithm:

The risk is high when a maximum temperature > 86 F is followed by at least 14 h of relative humidity exceeding 90% with a minimum temperature > 68 F. Days run from noon to noon.

Reference: Nutter, F. W., Cole, H., Jr., and Schein, R. D. 1983. Disease forecasting system for warm weather Pythium blight of turfgrass. Plant Disease 67:1126-1138.

Anthracnose Model

Forecasts foliar anthracnose on Poa annua; it has not been evaluated for forecasting accuracy on creeping bentgrass, the other host commonly affected in Kentucky. No information is available on the model=s forecasting accuracy for the crown-rot phase of the disease, which is more destructive than the foliar-blight phase. A weakness of this model is that is incorporates data on duration of leaf wetness. Such data are unavailable and are often not possible to obtain accurately in turf ecosystems. We therefore are estimating leaf wetness duration from data on relative humidity, which introduces a potential source of forecasting error. This model can be used to accumulate points up to a predetermined threshold based on local conditions.

Algorithm:

Prediction factors are based on hours of leaf wetness (LW) and average daily temperature (T).

$ Hours of leaf wetness (LW)

$ Average daily temperature (T)

Points are accumulated daily using this equation:

ASI = 4.0233 - 0.2283LW - 0.5308T - 0.0013LW² + 0.0197T² + 0.0155LWXT

In sites with a history of the disease, infection may be possible whenever ASI >2.

Reference: Danneberger, T.K., Vargas, J. M., and Jones, A. L. 1984. A model for weather-based forecasting anthracnose on annual bluegrass. Phytopathology 74:448-451