Research and IPM

Models: Diseases

Crop: Processing Tomato

Disease: Late Blight
Pathogen: Phytophthora infestans

Note: Before using a model that was not field tested or validated for a specific location, the model should be tested for one or more seasons under local conditions to verify that it will work in the desired location. See "Validation Work" below.

Model developer and citation

Bugiani, R., Cavanni, P. and Ponti, I. 1993. An advisory service for the occurrence of Phytophthora infestans on tomato in Emilia-Romagna region. Bulletin OEPP/EPPO 23: 607-613.

Cavanni, P., Ponti, I. and Marinelli, M. 1990. A forecasting model for late blight of tomato. Informatore Fitopatologico 40:17-25 (in Italian)

Ponti, I., Cavanni, P, Mazzini, F. and Libe, A. 1985. Verification of forecasting criteria for late blight of tomato. Informatore Fitopatologico 35:13-21 (In Italian).

Sensor location

Off-site, at regional weather stations.

Input variables

Environmental: Daily temperature, rainfall, and relative humidity.

Calculated: Average daily temperature in C (T_av), daily minimum temperature, average daily relative humidity (RH_av) and 24 and 48 hour rainfall accumulation (R).

Pathogen: Number of sporangia trapped and disease observed in untreated plots placed in strategic locations throughout the tomato growing areas.

Model description

This model generates infection potential indexes (IPI) that predict the most probable inoculum increase of Phytophthora infestans in the environment. In Italy, IP indexes are used along with indicator plants plus spore traps to warn farmers about when to start spraying. The model does not give recommendations about subsequent fungicide applications.

To calculate daily IPI, relative indexes for T_av, RH_av and R are calculated independently and combined by multiplying the T index by either the R index or the RH index. IPI only needs to be calculated when days have T_min > 7 C, T_av 9-25 C, total rainfall > 0.2 mm or RH_av > 80%. Favorable weather conditions for Phytophthora infestans produce a positive IPI. Functions for index calculations are:

T_Index = (-2.19247+0.259906 T - 0.000139 T³ - 6.095832 10^-6 x T⁴) Fc.

Fc = correction factor = 0.35 + 0.05 x T_min

T = daily average temperature (C)

RH_Index = -34.9972725 + 0.751 RH - 0.003909 RH² (see note below)

RH = daily average relative humidity.

R_index = 0.006667 + 0.194405 R + 0.0002239 R².

R = rainfall (mm) accumulated during the last 48 hours.

Note: The RH_Index as shown in Bugiani, et al. (1993) erroneously shows the constant (-34.9972725) as a positive number. After reviewing the other papers, it has been changed to the negative constant shown here.

The cumulative daily IPI over a defined period is used by the model to evaluate late blight risk.

A network of unsprayed plots of 100 m² are planted in strategic locations with a susceptible tomato cultivar (Heinz) and used to record and follow late blight epidemics. These plots are scouted weekly and used to report late blight attacks in advance.

Spore traps installed at the ground level are used to monitor sporangia of Phytophthora infestans in the air every two hours throughout the tomato growing season. Rapid increase of the amount of sporangia collected has been positively correlated with the occurrence of the first symptoms during the next few weeks.

Cumulative IPI plus disease data from unsprayed plots and populations of sporangia are considered together to recommend control strategies for specific areas.

Action threshold

IPI values below 15 indicate no risk of disease according to historical data from Emilia-Romagna, Italy. An IPI above 15 indicates high risk of disease and is used to determine when fungicides should start being sprayed, but the model does not give any recommendations about further treatments. This threshold is still being evaluated where it was developed. Model developers recommend adjustment of threshold for specific locations. Data from spore traps and disease in unsprayed plots should also be considered.

Model validation

Cavanni, P., Ponti, I. and Marinelli, M. 1990. A forecasting model for late blight of tomato. Informatore Fitopatologico 40:17-25. (In Italian).

Ponti, I., Cavanni, P, Mazzini, F. and Libe, A. 1985. Verification of forecasting criteria for late blight of tomato. Informatore Fitopatologico 35:13-21. (In Italian).

Model implementation

This model has been implemented since 1990 by a Regional Plant Protection Service of Emilia-Romagna. Tomato growers have achieved a 50% reduction in fungicide sprays by following this model's recommendations.

Bugiani, R., Cavanni, P. and Ponti, I. 1993. An advisory service for the occurrence of Phytophthora infestans on tomato in Emilia-Romagna region. Bulletin OEPP/EPPO 23: 607-613.

Current limitations of model

The model provides a risk index of initial disease but gives no recommendations about subsequent spray timing.

To provide correct recommendations, the model, which is driven by weather data, should be combined with data from unsprayed tomato plots and sporangia traps.

Future directions of model

The action threshold is considered provisional. Specific action thresholds could be developed for specific locations after a period of model evaluation.

Representativeness of the ambient weather stations for specific microclimates could be also considered.

Some models that were developed to forecast late blight on potato are being evaluated for tomato late blight in California.

Top of page