Why is the weather forecast sometimes “wrong”? Whose fault is it? We asked Riccardo Ravagnan, Meteo Forecast & Services Manager at Meteomed.
A “MATRYOSHKA” SYSTEM
“The relationship between different natural variables (humidity, wind, temperature, etc.) is not linear,” Ravagnan began, “the parameters on which we base mathematical calculations make our task of predicting the future difficult. There are an infinity of motions in the atmosphere that affect different dimensions of reality. For example, we can observe movements on a planetary scale (think of the fast upward flow of the jet stream) and movements occurring on smaller scales (such as breezes), all the way to those affecting infinitely smaller volumes (molecular movements). We can think of it as a kind of ‘Matryoshka doll’.
Each doll represents a volume in which a different movement takes place. Unfortunately, the movements within these ‘boxes’ are not observable and describable completely, and our analytical schematization carries with it gross errors (simplification of phenomena). Nature is a very complex system, and its description, improved over the years, to date does not allow us to predict with exact accuracy the time course of certain phenomena.”
Ravagnan continues: “Then there are physical processes that the model cannot represent in detail because they take place within scales smaller than those solvable by the model (radiative processes, convection, fluxes from the ground, cloud microphysics). These cannot be neglected as they affect the flow that develops at larger scales (parameterizations). Thus, if between two ‘dolls’ in which motion occurs there were motions within smaller ‘boxes,’ the model would deviate from a fair representation of the fluid atmosphere.”
OVERNIGHT? 92% CHANCE OF “GETTING IT RIGHT”
Because the models are based on numerical calculations, rounding the values of meteorological quantities could cause the results to diverge over short distances:
“Nowadays, the degree of reliability of a weather forecast is 92 percent for the next day and tends to decrease over time. The atmosphere is a chaotic fluid. Physically it means that there is a limit of predictability beyond which the system under consideration forgets the starting information.
Occasionally, however, it is possible to paradoxically guess one-week predictions and ‘mug’ the next day’s. It is a phenomenon known in this context as ‘intermittency’ and results from the fact that the atmosphere system has instabilities. For example, initial states containing regions of strong instability (e.g., convective, barocline or barotropic) are less predictable than those containing little instability. Where the system is more unstable it is less predictable.”
METEOROLOGY? IT’S STILL AN EXACT SCIENCE
Forecasts developed through models, therefore, express the probability of the occurrence of weather events (not certainty) and are the only means we have to get a more accurate idea of the evolution of weather conditions. “By this can we say that meteorology is an inexact science? No, meteorology is a branch of physics and as such is definable as an exact science. It studies dynamical systems from imposed initial conditions to solve the mathematical equations that describe phenomena and are included in predictive models. However, this discipline is related to Lorenz’s deterministic chaos theory.
Knowing the initial conditions of a physical system and the mathematical equations governing its reactions is not the same as being able to predict its future evolution as accurately as the initial elements are known. Just the atmosphere is the ideal prototype of everything in physics that is governed by deterministic laws but exhibits chaotic behavior.”