From pagoscope to new tools: towards even more accurate frost forecasting
"We can put all the means to produce, but if we don't anticipate the frost enough, we have lost everything. "
"Without it [Sencrop Station], I would not have had any production this year and I would have been devastated, because I was one of the few with a frozen orchard."
These testimonies show the value of good frost forecasting in agriculture. But while good frost forecasting is crucial, it is difficult to find a reliable tool to predict it correctly. Historically, the Pagoscope was a tool used to anticipate night frost. But how does it work? Is there no more reliable tool available today?
The Pagoscope, a tool for forecasting night frost
Every farmer who has been affected by frost from near or far knows that spring frost is difficult to predict accurately, especially at night. It can devastate many farms, especially vineyards or fruit trees. In order to anticipate it, a tool was created many years ago to know whether or not there is a risk of frost for the following night: the Pagoscope.
The Pagoscope, also known as a psychrometer, is a device for predicting morning frost, which is used 30 minutes before sunset. It is simple to use: it consists of a dry thermometer, a wet thermometer and a reading table in the form of an abacus. Once the two temperatures have been entered into the abacus, the combined reading of these two temperatures gives an indicator of the risk of frost for the following night. Depending on the results, the risk of frost is then certain, probable or improbable.
Pagoscope, a tool with limitations
The Pagoscope is a good way to help farmers better manage the risk of future frost, but it has some limitations. Indeed, the frost risk forecast is limited to the coming night, requires a daily manual check, is only used when there is no wind (since the Pagoscope does not predict advective frost, caused by the passage of a cold air mass) and can lack accuracy.
Sencrop conducted a study to compare the reliability of the Pagoscope measurements and weather forecast models with actual frost observations. The results are shown below: the cumulative number of stations that recorded the correct event (frost or no frost) predicted by each forecast model.
It can be seen that the cumulative number of stations that recorded the same event as the one forecasted by Pagoscope is significantly lower than the others. As the Pagoscope made the least accurate frost forecasts, it can be said that the reference weather forecasting models are a more reliable way of forecasting the risk of frost. On the other hand, Pagoscope's ""probable frost"" forecast for the risk of frost is still rather vague. It has been chosen here to model this category as 50% frost risk. The weather models only predict a certain or improbable risk of frost, making it easier for the farmer to read the forecast.
From this graph, it can be said that the best performing weather models in terms of frost prediction are : Meteoblue, Icon7 and Arome2 (although the performance of all models remains very close).
This is why new digital tools have appeared in recent years to better support farmers in their decision making, particularly in their management of frost. Among these solutions is Sencrop, which provides a more reliable forecast of the imminent risk of frost, but also a forecast of the risk over several days, using a comparison of data from different reference forecasting models. Moreover, with this solution, you no longer need to go and check your temperature readings: simply connect to your application to access your data!
This makes it a more powerful, accurate and reliable tool.
A new tool for reliable frost forecasting by Sencrop
The Sencrop solution brings together in a single application dry and wet temperature forecasts provided by the various reference forecasting models, and records live weather data from your plots to enable you to create alerts. These two tools combined are now popular with wine and fruit growers.
Focus on these two frost forecasting tools by Sencrop:
Frost alerts
Sencrop offers a frost alert system that allows you to be warned straight away when a drop in temperature in your plots is detected. It allows you to activate its protection systems at the right time.
Sencrop alerts are pre-set to activate as soon as the dry temperature of your weather stations is below 2°C and the wet temperature below 1°C, for 15 minutes. You can then create your alert from these settings or change them to customise the alert.
On the application, choose your mode of notification (SMS, phone call or email) to avoid missing the shooting window for the protection of your crops.
The frost alert system works thanks to the weather reports from your Sencrop stations. This way, you can benefit from live local weather reports and be prepared for an optimal forecast of the risk of frost!
Comparative tool for minimum temperature forecasts
To go even further, Sencrop offers you a frost forecasting solution. To do this, we have brought together 6 of the most reliable weather forecasting models on the same application (depending on the country, the number of forecasting models may vary). You can then :
- follow your temperature data history
- see the different forecasts in terms of dry and wet temperatures for the next 4 days
When one of the forecasting models indicates a dry temperature below 0, a risk of frost is indicated by a flake icon (❄) for the day in question, as is the case for Friday in the illustration below.
This allows you to analyse trends, assess the forecasted risks and make the right decisions to organise your frost control more serenely.
Conclusion:
The solution offered today by Sencrop is more reliable, is based on real weather forecasters, provides a longer forecast and uses station data so ultra-local data.
By detecting the exact moment your crops face frost, this is what will save you money, as you only activate your antifreeze systems at the right time!
14 days to try our frost features
Create your profile, connect to a station near your plots and get a free, no obligation trial.