Olive disease warnings adjust to the variety—but missed cases matter

One storm, different vulnerabilities

Peacock eye disease gets its name from the dark, yellow-ringed spots it produces on olive leaves. Caused by the fungus Venturia oleaginea, it can make leaves fall prematurely. Repeated defoliation reduces the canopy available for photosynthesis and can impair tree productivity.

Growers commonly manage the disease with fungicides, but a calendar cannot tell them whether a particular spell of weather actually favors infection. Nor does it distinguish between olive varieties that readily develop disease and those with greater resistance.

In a bioRxiv preprint that has not been peer reviewed, Yotam Gilat and colleagues describe an expanded version of PedMan, the Peacock eye disease Manager. The system combines weather information with cultivar susceptibility to guide decisions about whether and when to spray. The central question is whether more resistant varieties can receive fewer warnings without missing occasions when treatment helps.

Narrowing the warning window

PedMan was originally developed for highly susceptible olive cultivars. Its weather rule identifies potentially favorable infection conditions when rainfall reaches ≥15 mm in one day or over consecutive rainy days, and the average daily minimum temperature during those rainy days falls between 12 and 18 °C. This is not the daily average temperature.

To adapt that rule, the researchers simulated alerts using four growing seasons of weather records from 11 Israeli stations. They tested higher rainfall thresholds, narrower temperature windows, and combinations of both. Their targets were approximately 25% fewer alerts for moderately susceptible cultivars and 50% fewer for moderately resistant ones.

Keeping the rainfall threshold unchanged while narrowing the temperature window met those targets: 13–17 °C produced 25.1% fewer simulated alerts, while 14–16 °C produced 51.3% fewer. These percentages describe warnings in historical weather simulations—not measured reductions in fungicide use or disease.

The expanded system also considers season and the presence of symptomatic leaves for moderately resistant cultivars. For cultivars classified as highly resistant, it recommends no spraying. That last rule makes the accuracy of resistance classifications especially consequential: resistance is not immunity, and the paper notes that cultivar responses can differ among regions.

Comparing sprayed and unsprayed trees

The new field work comprised two small randomized experiments during the 2024/25 growing season and 22 commercial-grove observations during 2025/26. Researchers compared sprayed and unsprayed trees both when PedMan recommended treatment and when it advised withholding it.

The main outcome was visually estimated leaf loss, assessed around the tree canopy. In the commercial observations, 20 trees per treatment were evaluated within each subplot. The researchers also used mixed-effects analyses to address spatial grouping among trees.

All 14 spray recommendations in the new observations were associated with statistically lower leaf loss in treated trees. For Coratina, reported mean leaf loss was 18.0% without spraying and 9.4% with spraying. For Picual, the corresponding figures were 25.6% and 11.5%. These results support a treatment benefit under the conditions studied; they do not directly establish higher olive yields.

Combining the 24 new evaluations with nine from a companion study yielded the headline result: 30 of 33 recommendations matched the authors’ statistical criteria, or 90.9%. The reported 95% confidence interval was 76.8–96.9%, reflecting the limited dataset.

All 20 spray recommendations were supported by significant treatment benefits. Of 13 no-spray recommendations, however, three were contradicted by significantly lower leaf loss in sprayed trees. In practical terms, the errors were missed opportunities for protection, rather than recommendations to spray without a detected benefit.

What the headline leaves unresolved

The definition of a correct decision matters. The authors counted a no-spray recommendation as correct when the treatment comparison was not statistically significant. But failure to detect a difference does not establish that treatments are equivalent: a small or variable study can miss a real benefit.

There is also an unresolved contrast within the preprint. Individual comparisons found significant spraying benefits in two observations labeled as the highly resistant cultivar Askal, contributing to the reported missed cases. Yet cultivar-level mixed-model comparisons did not find a significant benefit for Askal. These analyses address different levels of the data; together, they caution against treating the no-spray rule as settled.

Leaf loss is also an indirect disease measure and can have other causes. Most new evaluations were commercial observations rather than randomized experiments, and the study did not establish a quantified reduction in spraying against a calendar-based program.

The useful advance is therefore a testable way to incorporate host susceptibility into weather warnings, not a universal spray prescription. Independent evaluation across more seasons, cultivars and locations—particularly outside Israel—is needed to determine whether fewer alerts consistently preserve disease control.