<i>NowDeepN</i>: An Ensemble of Deep Learning Models for Weather Nowcasting Based on Radar Products’ Values Prediction

• Main Authors: Gabriela Czibula, Andrei Mihai, Eugen Mihuleţ
• Format: Article
• Language: English
• Published: MDPI AG 2021-12-01
• Series: Applied Sciences
• Subjects: weather nowcasting; machine learning; deep neural networks; autoencoders; Principal Component Analysis
• Online Access: https://www.mdpi.com/2076-3417/11/1/125

One of the hottest topics in today’s meteorological research is <i>weather nowcasting</i>, which is the weather forecast for a short time period such as one to six hours. Radar is an important data source used by operational meteorologists for issuing nowcasting warnings. With the main goal of helping meteorologists in analysing radar data for issuing nowcasting warnings, we propose <inline-formula><math display="inline"><semantics><mrow><mi>N</mi><mi>o</mi><mi>w</mi><mi>D</mi><mi>e</mi><mi>e</mi><mi>p</mi><mi>N</mi></mrow></semantics></math></inline-formula>, a supervised learning based regression model which uses an ensemble of <i>deep artificial neural networks</i> for predicting the values for radar products at a certain time moment. The values predicted by <inline-formula><math display="inline"><semantics><mrow><mi>N</mi><mi>o</mi><mi>w</mi><mi>D</mi><mi>e</mi><mi>e</mi><mi>p</mi><mi>N</mi></mrow></semantics></math></inline-formula> may be used by meteorologists in estimating the future development of potential severe phenomena and would replace the time consuming process of extrapolating the radar echoes. <inline-formula><math display="inline"><semantics><mrow><mi>N</mi><mi>o</mi><mi>w</mi><mi>D</mi><mi>e</mi><mi>e</mi><mi>p</mi><mi>N</mi></mrow></semantics></math></inline-formula> is intended to be a proof of concept for the effectiveness of learning from radar data relevant patterns that would be useful for predicting future values for radar products based on their historical values. For assessing the performance of <inline-formula><math display="inline"><semantics><mrow><mi>N</mi><mi>o</mi><mi>w</mi><mi>D</mi><mi>e</mi><mi>e</mi><mi>p</mi><mi>N</mi></mrow></semantics></math></inline-formula>, a set of experiments on real radar data provided by the Romanian National Meteorological Administration is conducted. The impact of a <i>data cleaning</i> step introduced for correcting the erroneous radar products’ values is investigated both from the computational and meteorological perspectives. The experimental results also indicate the relevance of the features considered in the supervised learning task, highlighting that the radar products’ values at a certain geographical location at a time moment may be predicted from the products’ values from a neighboring area of that location at previous time moments. An overall <i>Normalized Root Mean Squared Error</i> less than <inline-formula><math display="inline"><semantics><mrow><mn>4</mn><mo>%</mo></mrow></semantics></math></inline-formula> was obtained for <inline-formula><math display="inline"><semantics><mrow><mi>N</mi><mi>o</mi><mi>w</mi><mi>D</mi><mi>e</mi><mi>e</mi><mi>p</mi><mi>N</mi></mrow></semantics></math></inline-formula> on the cleaned radar data. Compared to similar related work from the nowcasting literature, <inline-formula><math display="inline"><semantics><mrow><mi>N</mi><mi>o</mi><mi>w</mi><mi>D</mi><mi>e</mi><mi>e</mi><mi>p</mi><mi>N</mi></mrow></semantics></math></inline-formula> outperforms several approaches and this emphasizes the performance of our proposal.