🍎 Bitter Rot Model Predictions

Page Description

THIS IS EXPERIMENTAL, DO NOT USE THIS DATA FOR ANYTHING OTHER THAN RESEARCH PURPOSES.

This dashboard displays daily weather data and EXPERIMENTAL predictions for bitter rot infection risk and spray history across three trial locations: Highland, NY, Pittstown, NJ, and Winchester, VA. The weather data including temperature, precipitation and leaf wetness durations and subsequent infection risks are all based on forecasted data and may not reflect what actually was recorded by ground-based weather stations.

Last Updated:

More Information

Bitter rot of apple, caused by Colletotrichum spp., is a major disease that thrives in warm, wet conditions. To predict infection risk, multiple weather-based models have been developed, each using temperature and leaf wetness duration thresholds to estimate when conditions are favorable for infection. This experiment compares four different bitter rot infection models to evaluate their predictive accuracy and applicability in a commercial orchard setting.

1. Wilson (1990) Model
   This model was created for predicting Colletotrichum acutatum species complex infections on strawberries and is currently deployed in disease warning systems for strawberry. A model calculated risk value (exact values not shown) above 0.5 (risk threshold = 0.5) is considered high and "INFECTION!" is displayed in the table under the Wilson column.
2. Cosseboom and Hu (2024) Model
   This model was created for predicting Colletotrichum fioriniae infections on wine grapes. The risk threshold is set at 0.45.
3. Crusius et al. (2002) Model
   This model was created for predicting Glomerella leaf spot caused by species within the Colletotrichum gloeosporioides species complex on apple. The risk threshold is set at 0.15.
4. Katsurayama et al. (2010) Model
   This model was created for predicting Glomerella leaf spot caused by species within the Colletotrichum gloeosporioides species complex on apple. The risk threshold is set at 0.15.

Risk thresholds: Risk thresholds were set either according to previous model validation trials or according to model behavior with historical weather data. In short, thresholds were chosen to draw a line between predicting bitter rot infections from the most conducive weather conditions, while not predicting infection too frequently. If certain models do not perform well or predict infection too regularly, the thresholds of the models can be adjusted.

Weather data source and model implementation: Each model uses hourly weather data—specifically temperature and leaf wetness—collected from the National Weather Service. Leaf wetness is calculated according to forecasted precipitation probability and relative humidity.

Spray columns: These columns record preventative fungicide applications that were made in a replicated trial at research stations.