Preprints
https://doi.org/10.5194/gmd-2022-19
https://doi.org/10.5194/gmd-2022-19
Submitted as: development and technical paper
11 Feb 2022
Submitted as: development and technical paper | 11 Feb 2022
Status: a revised version of this preprint was accepted for the journal GMD and is expected to appear here in due course.

RAP-Net: Region Attention Predictive Network for Precipitation Nowcasting

Zheng Zhang1,, Chuyao Luo1,, Shanshan Feng1, Rui Ye1, Yunming Ye1, and Xutao Li1 Zheng Zhang et al.
  • 1Dept. of Computer Science, Harbin Institute Technology, Shenzhen, China
  • These authors have contributed equally to this work.

Abstract. Natural disasters caused by heavy rainfall often cause huge loss of life and property. Hence, the task of precipitation nowcasting is of great importance. To solve this problem, several deep learning methods have been proposed to forecast future radar echo images and then the predicted maps are converted to the distribution of rainfall. The prevailing spatiotemporal sequence prediction methods apply ConvRNN structure which combines the Convolution and Recurrent neural network. Although ConvRNN methods achieve remarkable success, they ignore capturing both local and global spatial features simultaneously, which degrades the nowcasting in regions of heavy rainfall. To address this issue, we propose a Region Attention Block (RAB) and embed it into ConvRNN to enhance forecasting in the area with strong rainfall. Besides, the ConvRNN models are hard to memorize longer historical representations with limited parameters. To this end, we propose Recall Attention Mechanism (RAM) to improve the prediction. By preserving longer temporal information, RAM contributes to the forecasting, especially in the middle rainfall intensity. The experiments show that the proposed model Region Attention Predictive Network (RAP-Net) significantly outperforms state-of-the-art methods.

Zheng Zhang et al.

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on gmd-2022-19', Anonymous Referee #1, 02 Mar 2022
    • RC2: 'Reply on RC1', Anonymous Referee #2, 17 Mar 2022
      • AC2: 'Reply on RC2', Chuyao Luo, 05 Apr 2022
    • AC1: 'Reply on RC1', Chuyao Luo, 05 Apr 2022
  • RC3: 'Comment on gmd-2022-19', Anonymous Referee #3, 19 Mar 2022
    • AC3: 'Reply on RC3', Chuyao Luo, 05 Apr 2022

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on gmd-2022-19', Anonymous Referee #1, 02 Mar 2022
    • RC2: 'Reply on RC1', Anonymous Referee #2, 17 Mar 2022
      • AC2: 'Reply on RC2', Chuyao Luo, 05 Apr 2022
    • AC1: 'Reply on RC1', Chuyao Luo, 05 Apr 2022
  • RC3: 'Comment on gmd-2022-19', Anonymous Referee #3, 19 Mar 2022
    • AC3: 'Reply on RC3', Chuyao Luo, 05 Apr 2022

Zheng Zhang et al.

Zheng Zhang et al.

Viewed

Total article views: 532 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
398 115 19 532 5 5
  • HTML: 398
  • PDF: 115
  • XML: 19
  • Total: 532
  • BibTeX: 5
  • EndNote: 5
Views and downloads (calculated since 11 Feb 2022)
Cumulative views and downloads (calculated since 11 Feb 2022)

Viewed (geographical distribution)

Total article views: 500 (including HTML, PDF, and XML) Thereof 500 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 20 May 2022
Download
Short summary
In this paper, we develop a model to predict radar echo sequences and apply it in the precipitation nowcasting field. Different from existed models, we propose two new attention modules. By introducing them, the performance of RAP-Net outperforms other models especially in those regions with middle and high-intensity rainfall. Considering these regions would cause more threats to human activity, the research in our manuscript is significant to prevent natural disasters caused by heavy rainfall.