Author:
Dr. Sheshang Degadwala ,Vedant Sukhadia
Abstract
Weather nowcasting, the short-term forecasting of weather conditions, is a critical area of study due to its impact on safety and planning. Recent advances in machine learning (ML) and deep learning (DL) have significantly improved nowcasting accuracy, particularly in precipitation prediction. This review examines modern techniques and models used for weather nowcasting, focusing on their advancements and challenges. Motivation for this study arises from the need for more accurate and timely weather predictions to mitigate adverse effects on daily life and infrastructure. Despite advancements, current methods face limitations such as high computational costs, data quality issues, and the challenge of generalizing models across different geographic regions. The aim of this review is to provide a comprehensive overview of recent developments in nowcasting techniques, highlight their strengths and weaknesses, and discuss future directions for research. The objective is to synthesize the latest findings, assess the performance of various models, and propose potential improvements. This review underscores the importance of continued innovation and integration of new technologies, including transfer learning and hybrid models, to enhance the effectiveness of weather nowcasting systems.
Reference65 articles.
1. A. Asperti, F. Merizzi, A. Paparella, G. Pedrazzi, M. Angelinelli, and S. Colamonaco, “Precipitation nowcasting with generative diffusion models,” Springer Nature, 2023, [Online]. Available: http://arxiv.org/abs/2308.06733
2. S. Imran, T. Anuradha, and R. Bharat, “Radar Based Precipitation Nowcasting Prediction by Using Deep Learning Techniques,” E3S Web of Conferences, vol. 405, pp. 1–9, 2023, doi: 10.1051/e3sconf/202340504003.
3. W. Liu, Y. Wang, D. Zhong, S. Xie, and J. Xu, “ConvLSTM Network-Based Rainfall Nowcasting Method with Combined Reflectance and Radar-Retrieved Wind Field as Inputs,” Atmosphere, vol. 13, no. 3, 2022, doi: 10.3390/atmos13030411.
4. G. Czibula, A. Mihai, and E. Mihuleţ, “Nowdeepn: An ensemble of deep learning models for weather nowcasting based on radar products’ values prediction,” Applied Sciences (Switzerland), vol. 11, no. 1, pp. 1–27, 2021, doi: 10.3390/app11010125.
5. A. Bihlo, “Precipitation nowcasting using a stochastic variational frame predictor with learned prior distribution,” arxiv, 2019, [Online]. Available: http://arxiv.org/abs/1905.05037