Air traffic can be affected by adverse weather conditions, such as storms. The formation of convective clouds, in which warm air rises rapidly, is one of the most unexpected weather situations. This type of event can be associated with strong winds and storms and represent a risk for aviation, causing flight diversions, increases in waiting time in the air and, ultimately, problems in airport operations.
In this context, Spanish researchers have developed an algorithm capable of predicting the formation of this type of clouds and storms in the most frequented airport in Spain and one of the European aerodromes that registers the highest number of passengers per year, that of Adolfo Suárez Madrid- Playing cards.
The algorithm incorporates artificial intelligence techniques to make predictions of convective phenomena in Barajas 12 hours in advance, which could help air traffic controllers face these adverse situations
The study, coordinated by the University of Córdoba (UCO), and in which the Polytechnic University of Madrid, the University of Valladolid and the University of Alcalá de Henares have participated, incorporates techniques of artificial intelligence to formulate accurate predictions within a 12-hour time horizon, which could help air traffic controllers in making decisions in such situations.
According to co-author David Guijo from the UCO, the new system is based on traditional weather forecasting models, to which historical data from the Madrid airport itself are added, collected during the period from 2011 to 2015. Techniques are subsequently applied machine learning on this data to get the weather forecast.
The information has been provided by the Madrid-Barajas radiosonde station, managed by the State Meteorological Agency (AEMET), and by other meteorological stations in the capital, from which different variables at different heights such as temperature have been assessed. , wind or water vapor content.
The algorithm is capable of inferring meteorological predictions at the airport itself based on four different categories: sunny days, cloudy days, days with the presence of convective clouds and days with storms, these last two situations being the ones that present the most problems in take-off and landing. of airplanes.
“The combination of various different sources of information is one of the novelties of the study, which allows us to better characterize the problem and make more precise predictions”, highlights Pedro Antonio Gutiérrez, another of the researchers from the AYRNA group, led by Professor César Hervás.
The research, led by the AYRNA group of the UCO specialized in artificial intelligence, is part of the Hamlet project, an initiative in which the University of Alcalá also participates and which aims to develop predictive algorithms to address problems related to health and environment.
D. Guijo-Rubio, C. Casanova-Mateo, J. Sanz-Justo, PA Gutiérrez, S. Cornejo-Bueno, C. Hervás-Martínez and S. Salcedo-Sanz. “Ordinal regression algorithms for the analysis of convective situations over Madrid-Barajas airport”, Atmospheric Research 2020
Rights: Creative Commons.