| |
 |
UNIVERSITY OF BUCHAREST
FACULTY OF PHYSICS Guest
2024-11-24 21:15 |
 |
|
|
|
Conference: Bucharest University Faculty of Physics 2024 Meeting
Section: Atmosphere and Earth Science; Environment Protection
Title:
Using neural networks in temperature meteorological forecast
Authors:
Adrian BERLIC (1,2), Mihai DIMA (1), Mihaela CAIAN (1,2)
Affiliation:
1) University of Bucharest, Faculty of Physics, 405 Atomiștilor Street, 077125, Măgurele, România
2) National Meteorological Administration, 97 Șoseaua București - Ploiești, Bucharest, România
E-mail
adrian.berlic@gmail.com
Keywords:
weather forecast, temperature, data modelling, neural networks, artificial intelligence
Abstract:
We employ Artificial Intelligence (AI) technologies in weather forecasting, with focus on temperature predictions. The main objective is to assess the applicability of Neural Networks for medium to long-term meteorological and climate forecasting. Neural Networks were chosen due to their ability to learn high-level representations directly from data, capturing complex patterns governing the climate system without explicitly defined causal relationships. The study spans three distinct geographical areas in Romania—Sulina coastal, Fundulea plains, and Făgăraș mountains—to ensure a comprehensive evaluation across diverse topographies and climatic conditions. Preliminary results indicate significant potential for neural networks in temperature prediction, showing satisfactory accuracy and underscore their utility in modeling and predicting temperature fluctuations based on historical data sets. This highlights AI's potential to revolutionize weather forecasting, paving the way for developing more precise and reliable prediction systems capable of managing the inherent complexity and variability of climate systems.
|
|
|
|