Temporal and Spatial Satellite Data Augmentation  for Deep Learning-Based Rainfall Nowcasting

This is a Preprint and has not been peer reviewed. The published version of this Preprint is available: https://doi.org/10.2166/hydro.2024.235. This is version 1 of this Preprint.

Add a Comment

You must log in to post a comment.


Comments

There are no comments or no comments have been made public for this article.

Downloads

Download Preprint

Authors

Özlem Baydaroğlu , Ibrahim Demir

Abstract

Climate change has been associated with alterations in precipitation patterns and increased vulnerability to floods and droughts. The need for improvements in forecasting and monitoring approaches has become imperative due to flash floods and severe flooding. Rainfall prediction is a challenging but critical issue owing to the complexity of atmospheric processes, the spatial and temporal variability of rainfall, and the dependency of this variability on several nonlinear factors. Because excessive rainfall is the cause of natural disasters such as floods and landslides, accurate real-time rainfall nowcast is critical for the necessary precautions, control, and planning. In this study, rainfall nowcasting has been studied utilizing NASA Giovanni satellite-derived precipitation products and the convolutional long short-term memory (ConvLSTM) approach, which is a variation of LSTM. Due to data requirements of deep learning-based prediction methods, data augmentation is performed using interpolation techniques. The study utilized three types of satellite-derived rainfall data, including spatial, temporal, and spatio-temporal interpolated rainfall data, to conduct a comparative analysis of the results obtained through nowcasting rainfall. This research examines two catastrophic floods that transpired in Türkiye Marmara Region in 2009 and Central Black Sea Region in 2021, which are selected as the focal case studies. It also explores the suitability of a nowcast model for various flood events, while also examining the impact of data augmentation on the nowcast.

DOI

https://doi.org/10.31223/X5QQ39

Subjects

Atmospheric Sciences, Climate, Meteorology, Physical Sciences and Mathematics

Keywords

Rainfall flood, nowcasting, Deep Learning, data augmentation, interpolation, rainfall, flood, nowcasting, Deep learning, data augmentation, Interpolation

Dates

Published: 2023-10-03 00:41

Last Updated: 2023-10-03 07:41

License

CC BY Attribution 4.0 International

Additional Metadata

Data Availability (Reason not available):
https://giovanni.gsfc.nasa.gov/giovanni/