There is hardly anything that is not digital these days. At least it makes many things faster and simpler. While in the past it was difficult to gather weather data over e.g. the radio or teletext in order to have an approximate plan, whether it's going to be windy or not, things have changed today. You can load countless of forecast apps on your mobile phone today and of course every one of them promises to be the best. But which one is the best? Unfortunately this question is not so easy to answer!
Forecast models are nothing more than mathematical calculation models. A huge amount of measuring data flows into these computing models such as data from weather sations, satellites, measuring buoys, ships or weather balloons. From the state of the atmosphere at a given starting point, the state at later times is calculated with an equation. Physical variables such as temperature, air pressure and density play an important role here. Here we are already at the first problem, in the atmosphere there is chaos and if one of these variables changes, it changes the whole prediction. Therefore, the functions are also calculated at regular intervals with current data to make the prediction as accurate as possible!
Basically, the different forecast models differ in their way of calculating the forecast, in the size of the areas for which they make forecasts and the times they are updated each day. Each of these models lays a grid around the globe. These grids are always square and vary in size depending on the model. Depending on which model is used, different predictions result with the same input data.
The models that are used by the well-known prediction providers are GFS, WRF, ECMWF, NEMS and ICON. It must be said that the providers use different grid sizes of the respective models e.g. Windfinder uses the GFS model with a grid of 13km² and Windy the GFS model with a grid of 22km². A detailed description of the models would go beyond the scope of the contribution but we still want to try to give you a brief overview of the differences.
|1||GFS - global hydrostatic model, uses pressure as vertical coordinates in the forecast
|2||WRF - mesoscale numerical weather prediction system designed for both research and operational forecasting applications
|3||ECMWF - global nonhdydrostatic model, uses altitude and more accurately accounts for topographic effects
|4||NEMS - nonhdydrostatic multiscale model
|4||ICON - uses icosahedral grid model, provides a nearly homogeneous coverage of the globe
If you ask metrologists for predictions you always get the same answer, everything from 4 days' prediction is more of a trend. The accuracy of a prediction decreases with each day that goes further into the future. So you should not be completely crazy about a 7 day forecast but just check the predictions regularly.
Every model has its advantages and disadvantages. For example, a global model such as the GFS can work very well on coast regions, where there are hardly any major obstacles such as mountains. At the same time, the model can work very poorly in areas such as Lake Garda, as there are high topographical obstacles. In addition, the grid of this model is simply too big to take in the fine fall winds in mountains for example. For such areas, a model with smaller grid system is more suitable. It is therefore always worthwile to compare several predictions and to pay attention to local forecasts!