Infection forecasts are at the heart of VineForecast. In this article you will learn more about the modelling, interpretation and use of disease forecasts in VineForecast.
The calculation 🧮
The weather 🌦
The development of diseases is largely determined by the weather and so all forecasting models for Powdery or Downy Mildew are based on weather data. In order for risk indices for infections to be modelled correctly, the weather must first be considered.
VineForecast obtains raw data from the German Weather Service (e.g. temperature, precipitation, humidity, etc.). Ideally, however, this data is only resolved to 1 - 2 km. If this raw data were fed into a disease model, the accuracy would be very low. To increase the accuracy, VineForecast uses a method from climate physics.
The Downscalingor regionalisation, makes it possible to include regional influencing factors in the weather forecast. Among other things, VineForecast includes the local topography with a resolution of 25 metres when adjusting weather data. This allows the effects of different altitudes or slope inclinations to be included. The effect is illustrated by the following example of temperature adjustment.
Infection modelling 🍄
In addition to the weather, the Phenology another crucial basis for calculating infection. VineForecast calculates the phenology on the basis of temperature sum models. For example, the BBCH stage is calculated based on research by Molitor are calculated for different grape varieties. For the number of leaves, we orientate ourselves on the Geisenheim model from Schultz. The infection prognosis is adjusted depending on the stage of the season. Around flowering, for example, the susceptibility to infection with Powdery Mildew is considered to be significantly increased.
The modelled phenology is not set in stone, but you have the option of changing the Correct model.
After VineForecast has calculated the two influencing factors (regionalised weather, phenology) separately, these can now be converted into infection risks using disease models from viticulture research.
For Powdery Mildew We orientate ourselves on the OiDiag Index according to W. K. Kast and make specific adjustments based on further research and experience.
With the Downy Mildew forecast VineForecast uses different models depending on the phase. Oospore germination and primary infection are analysed according to the methodology of Caffi & Rossi calculated. Secondary infections are modelled using the leaf wetness hours (BNG) indicator. The risk index is calculated from the sum of the temperatures during hours with leaf wetness. If the value is above 50, a low risk of infection is assumed. If the BNG is higher than 150, a high risk of infection can be assumed.
The effect that the adaptation of weather data to local conditions has on the disease forecast is reflected in the following image. The figure shows the calculation of BNG on the Moselle.
The figure on the left shows that a weather forecast with raw data leads to a very inaccurate resolution of the disease forecast. The figure on the right shows the prediction of BNG by the VineForecast model. The effects of topography on the risk of infection are visible here. For example, VineForecast reveals higher infection risks in the side valleys of the Moselle due to the regionalisation described above, while the model with the raw data does not recognise these effects.
Weather stations generally use models with a resolution of at least 1 x 1 km for forecasting and would refer to the calculations in the figures on the left. This means that although the station can guarantee an accurate recording of the actual state, it is very inaccurate when forecasting the future. See here for more information:
Why no weather station helps with preventive plant protection
The representation 📊
Of course, VineForecast does not generate complex maps for the users of the software, but communicates the calculation described above as simply and purposefully as possible. The risk display is shown in the form of a traffic light scheme:
No risk of infection
Low risk of infection
Medium risk of infection
High risk of infection
Potential infections can occur from a low risk of infection. However, it only becomes really dangerous from a medium risk of infection and it is definitely advisable to act here.
In the dashboard, you can choose between the overview and the detail page. On Prognoses -> Diseases gives you a quick impression of the current overall situation. Here you can see all the infection risks for your vineyards summarised. If you move the mouse pointer over the bars in the sidebar on the right, you will also see an indication of how many vineyards fall into which category.
On the details page, you will always see six small coloured rectangles for each plot, which are based on the colour traffic light described above. The first triangle represents the calculated risk of infection from yesterday. The second triangle is the current risk of infection. Rectangles 3 - 6 are the forecasts for the next four days.
Quick tip: On the map you can quickly see the risk of infection for your vineyards. At the top left you can quickly change the date displayed.
You can switch between the disease forecasts for Powdery and Downy Mildew in the top right-hand corner of the map. The display on the map and in the sidebar should then adapt.
