Weather-Driven Forecasting Brings Precision to Grape White Rot Control
White rot in grape berries has long been managed with repeated fungicide sprays, a strategy that often follows habit more than infection biology. A new weather-driven mechanistic model for grape white rot, built around the fungus Coniella diplodiella, moves the field toward timing sprays around risk rather than routine calendars.
The practical value is immediate. When growers can identify infection windows instead of treating every period of uncertainty as dangerous, they gain a more rational way to protect yield and reduce unnecessary applications.
How the Model Reconstructs the Disease Cycle
The model was designed around three linked epidemiological steps: production and dispersal of primary inoculum, infections caused by primary and secondary conidia, and symptom onset followed by production of secondary inoculum. That structure matters because white rot does not unfold as a single event; it is a sequence in which weather influences each stage differently, from inoculum availability to infection success to epidemic amplification.
Rather than relying on empirical spray timing alone, the model uses available literature and disease knowledge to translate meteorological conditions into biological risk. In plant disease management, that shift is more than technical. It changes the logic of intervention from reactive coverage to predictive intervention, especially in seasons where warm, wet conditions can accelerate fungal spread through the vineyard canopy and onto susceptible berries.
Performance Across 16 Independent Epidemics
Investigators validated the model against 16 independent epidemics drawn from different grape-producing regions in China between 1982 and 2021. Those vineyard-by-year combinations gave the model a broad test across weather patterns and disease pressure, which is exactly where simple rules often fail.
The results were strong. For infection events, the model reached an accuracy of 0.971, a sensitivity of 0.929, and a specificity of 0.991. In plain terms, it rarely missed true infection periods and rarely flagged false alarms. For seasonal disease progress, the agreement between observed and predicted severities remained high, with a concordance correlation coefficient of 0.946, a root mean square error of 0.112, and a coefficient of residual mass of -0.047.
Those numbers point to a model that does more than classify isolated events. It also follows the trajectory of disease development across the season, which is essential for practical field use because growers need to know not only when infection begins, but how quickly disease intensity may build after weather turns favorable.
Why Weather-Based Risk Tools Matter in the Vineyard
Repeated fungicide use has a cost when it is untethered from disease risk. Unnecessary sprays raise production expenses, can increase selection pressure for resistance, and may still fail if the application misses the true infection window. A mechanistic forecasting tool gives growers and advisers a way to align treatment with the pathogen’s actual biology.
The model’s transparency is an important feature. Because its structure is built around defined epidemiological processes, it can be expanded as new data become available. That makes it more adaptable than a black-box forecasting system that produces risk scores without showing how those scores arise. In practical terms, a transparent framework can be examined, recalibrated, and extended to fit different regions or disease scenarios.
One especially important implication is seasonal flexibility. The model performed robustly under diverse weather and disease pressure conditions, which is where many forecast systems lose reliability. White rot epidemics do not look identical from one year to the next, and the same vineyard can face very different risk profiles depending on rainfall, temperature, and infection opportunity. A mechanistic system that remains stable across those shifts is more likely to earn field trust.
A Platform for Broader White Rot Surveillance
The conclusion points beyond Coniella diplodiella alone. The model’s architecture is open enough to incorporate other white rot species, including Coniella vitis, if future epidemiological data support that extension. For growers and disease advisors, that could mean a more unified framework for white rot management rather than separate tools for closely related pathogens.
The current study also carries value for regions where historical epidemics are already documented but decision support has lagged behind the biology. By pulling together prior literature and observed disease patterns, the model shows how existing knowledge can be turned into a working forecasting system rather than left scattered across papers and extension notes. In agriculture, that kind of synthesis can be as useful as a new active ingredient, because the main obstacle is often not knowing what to spray, but knowing when spraying is justified.
What This Means for Fungicide Timing
For growers, the central consequence is risk-based scheduling. A high-performing forecast system allows protective measures to be concentrated around periods when primary inoculum, infection conditions, and subsequent sporulation are aligned. That is where epidemic control is won or lost.
The model’s strengths are clear, but its value will depend on how it performs outside the validation set and how easily it can be integrated into real vineyard operations. Its transparent design gives it a strong starting point, while the remaining work lies in regional refinement, operational deployment, and testing against other white rot pathogens and field decision thresholds.
PMID 43035345 and DOI 10.1002/ps.71480
The study appears in Pest Management Science with PMID 43035345 and DOI 10.1002/ps.71480. Its main contribution is a disease forecast model that translates weather into biologically grounded white rot risk, replacing empiric spray timing with a structure that can be tested, adjusted, and expanded.
- • PubMed PMID 43035345: https://pubmed.ncbi.nlm.nih.gov/43035345/
- • DOI 10.1002/ps.71480: https://doi.org/10.1002/ps.71480