Title | ||
---|---|---|
XNow: A deep learning technique for nowcasting based on radar products’ values prediction |
Abstract | ||
---|---|---|
The problem of forecasting severe weather events is one of the most challenging topics in meteorology, as severe phenomena are becoming more and more frequent in many regions of the world. The short term weather analysis and forecast is called nowcasting and has a major role in prevention of risks and management in crisis situations. Radar data is one of the essential materials used by meteorologists for making short-term weather forecasting and for issuing early warnings for severe weather. We are proposing a convolutional neural network model XNow for short-term prediction of radar products’ values that would be useful for issuing nowcasting warnings and for providing climatologists the possibility to obtain relevant information about the short term changes in radar products values. Experiments are performed on real radar data provided by the Romanian National Meteorological Administration. An average normalized root mean squared error less than 3% was obtained, highlighting the effectiveness of XNow. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1109/SACI49304.2020.9118849 | 2020 IEEE 14th International Symposium on Applied Computational Intelligence and Informatics (SACI) |
Keywords | DocType | ISBN |
Weather nowcasting,data mining,deep learning,Xception | Conference | 978-1-7281-7378-8 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ioana Angela Socaci | 1 | 0 | 0.34 |
Gabriela Czibula | 2 | 80 | 19.53 |
Vlad-Sebastian Ionescu | 3 | 0 | 0.34 |
Andrei Mihai | 4 | 0 | 0.34 |