Research and IPM

Interactive Tools and Models: About the Citrus Thrips Scarring Prediction Model

This research is the work of Heinrich Schweizer and Joseph Morse, Department of Entomology, University of California, Riverside, CA 92521. UC IPM Information Systems staff prepared the model for use on this Web site.

Reference articles:

• Schweizer, H. and J. G. Morse. 1999. Predicting citrus thrips damage on navel oranges in the San Joaquin Valley. Citrograph 84(3): 3, 6-7.
• Heinrich Schweizer and J. G. Morse. 1997. Estimating the level of fruit scarring by citrus thrips from temperature conditions prior to the end of bloom. Crop Protection 16(8): 743-752.

The following is a summary of the research from which citrus thrips damage estimates were calculated on the UC IPM Web site. Complete methods, analysis, and results are reported in the journal article in Crop Protection (above).

Citrus thrips, Scirtothirps citri, cause scarring on fruit rinds as they feed on small citrus fruitlets soon after petal fall. Because scarred fruit may not meet cosmetic standards for fresh sales, the damage can cause significant economic losses to California orange growers.

Schweizer and Morse developed and validated multiple regression models to estimate the degree of fruit scarring by citrus thrips. The models used heat and chill degree-days, and regression coefficients and indicated that cool weather during early March (2 March-16 March) and warm weather during bloom were associated with high levels of thrips scarring.

The study used data from:

• Annual pesticide evaluation trials in a Navel orange (Citrus sinensis) block at the Lindcove Citrus Research and Extension Center near Exeter, California. The trees were Early Atwood Navels on Rubidoux Trifoliate rootstock planted in 1970.
• Fruit scarring levels from untreated trees were used to estimate citrus thrips damage. Data for the years 1982-1996 were used to develop the models, and data from 1972-1981, and 1994-1996 were used to validate them.
• From mid-1981 through 1996, daily minimum and maximum air temperatures from a weather station located at Lindcove were used. Values from nearby Lindsay were used to estimate temperatures for 1972 through mid 1981.

Different models using heat and chill degree-days during several periods were developed and evaluated, based on the fact that:

• Relative rates of development for citrus thrips and the tree depend on temperature.
• Overwintering eggs may require specific temperature patterns to enter, remain in, and terminate diapause.
• Egg mortality may depend on temperature.

Calculations:

• Heat and chill degree-days were calculated using a double sine method.
• 14.6°C was used as a lower threshold for citrus thrips development.
• The fraction (p) of severely scarred fruit was transformed to z= ln(-ln(1-p)).
• The variable z was regressed against heat and chill degree-day sums for selected periods.
• For model validation, observed percentages of severely scarred outside fruit were compared with predictions made by the best three models.

The authors also studied the economic effect of using the scarring predictions to make citrus thrips treatment decisions.

Scarring predictions

The most promising model suggests that:

• Cool weather in early March and warm weather during bloom are associated with high levels of thrips scarring
• Relatively warm weather during early March and relatively cool weather during bloom are associated with less scarring.

Models performed well in predicting fruit scarring for the specific field, somewhat overestimating the percentage of scarred fruit, but followed the observed trend.

Potential explanations for link between weather pattern and thrips scarring

• Eggs may require chilling in early March to ensure timely hatching
• Warm weather during bloom may be required for timely adult emergence and high oviposition activity on suitable fruitlets, which leads to high fruit scarring.
• Cool weather in late September may be crucial for eggs to enter an appropriate overwintering state.
• Cold weather in late October may increase egg mortality.

Economic evaluation

• Using results from any of the tested models resulted in fewer treatments and higher grower returns than was found with regular application of prophylactic pesticide treatments.
• Fruit sampling for larvae, to determine the need for protective pesticide treatments, was more expensive and left less time between data collection and treatment decisions than using the model.

Limitations of the model

• The model is based on data from one field, but different levels of thrips scarring have been observed in different fields in a single year.
• The model is intended to provide an estimate of citrus thrips damage on Navel oranges in the San Joaquin Valley only.
• During the last 4 seasons (1998-2001), the model has considerably overestimated the level of thrips scarring. Using its predictions without additional information from field monitoring would have led to unnecessary thrips treatments.

For more information, see the Crop Protection (1997, Vol. 16, No. 8, pp. 743-752) article, or contact Joe Morse.

Current citrus thrips treatment decision-making strategy

• UC recommends that citrus thrips treatment decisions be based upon field sampling of thrips larvae on small fruit.
• A major drawback with this method is that by the time economic thrips levels are detected, treatments need to be applied almost immediately to be effective. Also, accurate field sampling is laborious. Thus, sampling is expensive and provides little advance warning of the need for treatment.
• This model attempts to estimate citrus thrips damage from weather data that accumulates prior to petal fall.
• The model is a first step in developing a reliable citrus thrips damage prediction with advance warning (e.g., at least one week in advance of when thrips treatments would be applied, if needed).

What to do with the predictions

The primary purpose of posting the model on this Web site is to collect feedback from citrus growers and PCAs. The model has not been fully tested and will probably require substantial improvement before the authors can recommend using its predictions without reservations.

Check historical predictions

• In case you have records of citrus thrips scarring or fruit infestation levels, you can use historical model predictions from a nearby weather station and compare those to your records.
• Develop a feeling of the accuracy or inaccuracy of the model predictions. Contact Joe Morse to let us know your opinion.

Help improve the model

Because of the limited validation of the model, we are interested in acquiring further data for future improvement of the prediction model. Two types of data may be useful for this purpose:

• Fruit infestation by immature citrus thrips before any insecticide treatments have been applied.
• Levels of fruit scarring by citrus thrips on trees that have not been treated.

It is very important that the data have been collected before treatment and from untreated trees, in the case of fruit infestation and fruit scarring, respectively. There is no easy way to accurately estimate the potential for citrus thrips damage if trees are treated.

If you are interested in evaluating the model, we suggest leaving a corner of your orchard (e.g., 3 rows by 4 tree wide) untreated. Comparing thrips damage from the untreated trees to the treated trees will enable you to judge whether treatments have resulted in appreciable damage reduction. Comparing fruit scarring on untreated trees to model predictions will allow evaluation of the model performance.

In case you have such data or you are willing to cooperate in collecting such data, please contact Joe Morse.

Make decisions—only if you accept the risk; see disclaimer

Disclaimer: The University of California and the Authors of this model do not assume any responsibility and cannot be held liable for any loss that may result from using predictions from this model. Validation of the model has been limited, and you are encouraged to test the model for one or more seasons to see if it is helpful to you.

• Use the model as a supplement to the thrips scarring history in a particular field.
• In case your historical records of thrips damage are consistently higher or lower than model predictions. adjust the predictions accordingly.
• Use model predictions to identify seasons and orchards with high fruit-scarring risk.
• Then optimize labor allocation for accurate field sampling.
• Optimize equipment allocation to high risk areas for eventual treatments.
• Use field sampling data to back up your decision on whether or not to apply treatments against citrus thrips.

Top of page