One of the most treasured values of science is its ability to predict events. Celestial mechanics is particularly remarkable, since it allows us to predict, for instance, that on 14 May 2887 an annular solar eclipse will be visible from my city immediately after sunrise.
In meteorology, where there is a lot of practical interest in forecasts, the situation is quite different. Despite the remarkable advances that have been made in this field, predictions still cannot be made far enough in advance, even about very intense and widespread phenomena. Thus, hurricane Matthew pummelled across the Caribbean Sea between 28 September and 10 October 2016, but the phenomenon was not predicted until four days before, and even then it was only assigned a 70% probability.
Further, it does not look like data resolution or calculation-power are to blame for this. These parameters are constantly improved, but this does not significantly affect the timeliness of meteorological forecasts. Therefore, it is natural to wonder if there might be an intrinsic limit on how long in advance predictions can be made in this field.
It is not only about the famous «butterfly effect», which results from the limited precision of the data. What we are putting forward is the possibility that the future might be unpredictable even if the data were infinitely precise!
It can be argued that meteorological processes are very complex. In view of this, it is advisable to bypass non-essential complications and consider a simpler system – the simpler the better – where the question we are posing continues to make sense: whether or not there is an intrinsic limit to the time extent of predictions.
AW: Hurrikan "Ida"
Mit Zitat antworten
