IWRA Proceedings

< Return to abstract list

Spatiotemporal Bias Correction of Satellite Precipitation Using Convolutional Neural Networks for Rainfall-Runoff-Inundation Modeling in the Tonle Sap Lake Basin, Cambodia

IWRA World Water Congress 2025 Marrakech Morocco
Towards Innovation and a Smart Water Future
Author(s): Giha LEE (Dr., Professor)
Giha LEE (Dr., Professor)
Dept. of Construction and Disaster Prevention Engineering, Kyungpook National University, S. Korea
Email Address: leegiha@knu.ac.kr

Poster: PDF

Abstract

This research examines how the CA-UNET technique can improve hydrological simulations.

The proposed framework enhanced the quality and reliability of CHIRPS rainfall data with a CNN-based approach, yielding Corrected-CHIRPS with superior temporal and spatial correlation relative to the observed dataset.

Applying the CNN-based bias-corrected precipitation dataset within hydrological modeling could aid watershed water resource management and mitigation of floods in data-scarce regions.