The evening before a long-awaited trip or a meticulously planned garden party, you checked the weather app one last time - just quickly make sure that nothing can go wrong.
But what is that?
A gray rain cloud has now moved in front of the sun symbol that gave us confidence yesterday.
The app shows a 70 percent chance of rain, so tomorrow seems to be gray and wet, and the plans are in the dumps.
But is that actually true?
Weather forecasts are nothing more than calculations of probabilities.
Computers use the current weather conditions and extensive historical data to determine scenarios for the coming hours and days.
Because weather models have become more sophisticated and the data available thanks to satellites has become more extensive, the quality of forecasts has also improved.
Rain probabilities are often used to indicate the risk of precipitation for an entire day.
However, if you open a weather app and look at the forecasts for the respective day, you quickly see that even with high values, it doesn't necessarily mean that the whole day will be rainy.
A more detailed and often optimistic preview can be found in the hourly view.
However, hour-by-hour forecasts are difficult and less reliable.
Basically, the larger the rain areas and the shorter the forecast horizon, the more reliable the forecast.
According to information from the German Weather Service, major rainfall events can usually be easily predicted for the next 48 hours.
Small-scale showers in particular are more difficult to calculate - especially with long forecast periods or low amounts of precipitation.
The atmospheric processes at such locations are complex, which makes precise predictions difficult.
Even small changes in local weather conditions can have a major impact on precipitation patterns.
Neither when it will rain nor how large the affected area will be can be predicted with certainty.
It can happen that a shower is predicted with a high degree of probability, but in the end it doesn't rain at all.
The amount of precipitation is therefore an important indicator for better assessing the upcoming weather.
For example, if the probability of rain is 70 percent but only minimal amounts of precipitation are specified, there is only a risk of a few drops for a short time, if at all.
The situation is different with a probability of rain of 70 percent and a forecast of significant amounts of precipitation.
In the future, artificial intelligence may further improve weather forecasts.
At the end of 2023, Google's subsidiary Google DeepMind presented current plans for an AI application.
Other companies such as the Chinese company Huawei and the European Center for Medium-Range Weather Forecasts (ECMWF) have also entered the race for the best weather forecasts.