Multidisciplinary and Interdisciplinary (M)
Session Sub-category General Geosciences, Information Geosciences & Simulations(GI)
Session ID M-GI27
Title Data-driven approaches for weather and hydrological predictions
Short Title Data driven study in weather prediction
Main Convener Name Shunji Kotsuki
Affiliation Center for Environmental Remote Sensing, Chiba University
Co-Convener 1 Name Daisuke Hotta
Affiliation Meteorological Research Institute
Co-Convener 2 Name Yuki Yasuda
Affiliation Institute of Science Tokyo
Co-Convener 3 Name Thomas Sekiyama
Affiliation Meteorological Research Institute
Session Language 
E
Scope 
In the digital era, data-driven techniques are transforming our understanding and prediction capabilities of complex earth systems. This session aims to explore the cutting-edge methodological and applicational studies for weather, climate and hydrological predictions.
Key themes includes: (1) methodological studies to deepen data-driven approaches for geoscience problems, (2) machine/deep learning studies applied for extreme weather-related disasters, (3) climate predictive analysis to discern climate variability, trends, and anomalies, (4) integrating remote sensing and ground data to refine prediction models.
This session aims to foster a rich dialogue among experts, highlighting both the advancements and challenges in data-driven environmental modeling. Participants will gain insights into current best practices and envision the future trajectory of this rapidly evolving domain.
Presentation Format Oral and Poster presentation
Invited Authors Shigenori Otsuka (RIKEN Center for Computational Science)
Masanobu Inubushi (Tokyo University of Science)
Time Presentation No Title Presenter
Oral Presentation May 29 AM1
9:00 - 9:15 MGI27-01 Global precipitation nowcasting with ConvLSTM and adversarial training Shigenori Otsuka
9:15 - 9:30 MGI27-02 Conditional Deep Diffusion Modeling for GSMaP Inpainting Daiko Kishikawa
9:30 - 9:45 MGI27-03 Deep learning approach to subseasonal prediction of the western North Pacific subtropical high: transfer and multitask learning Yuki Maeda
9:45 - 10:00 MGI27-04 Sequential analysis of tipping in high-dimensional complex systems with partially known dynamics Tomomasa Hirose
10:00 - 10:15 MGI27-05 Automatic Front Detection Using Deep Learning: Leveraging Temporal Data and Local Explanations with Attention Mechanisms Takumi Matsuda
10:15 - 10:30 MGI27-06 ClimaX-LETKF: A pure data-driven artificial intelligence-based ensemble weather forecasting system Akira Takeshima
Oral Presentation May 29 AM2
10:45 - 11:00 MGI27-07 Synchronization in Turbulence and Its Significance for Data-Driven Approaches Masanobu Inubushi
11:00 - 11:15 MGI27-08 Multi-Model Ensemble and Reservoir Computing for Efficient River Discharge Prediction in Ungauged Basins Mizuki Funato
11:15 - 11:30 MGI27-09 Leveraging Japan's National Streamflow Records for End-to-End Data-Driven Hydrological Modeling at National Scale Tristan Hascoet
11:30 - 11:45 MGI27-10 Precipitation super-resolution using diffusion model with d4PDF Ryo Kaneko
11:45 - 12:00 MGI27-11 Toward enhancing the ensemble Kalman filter with a diffusion model Takumi Honda
12:00 - 12:15 MGI27-12 Real-Time 3D Super-Resolution for Urban Micrometeorology Using Diffusion Models and Schrödinger Bridge Yuki Yasuda
Presentation No Title Presenter
Poster Presentation May 29 PM3
MGI27-P01 Manifold learning-aided offline parameter estimation of an Earth system model using observation of changing climate Amane Kubo
MGI27-P02 Weather field super-resolution using Restricted Boltzmann Machines Ryo Kaneko
MGI27-P03 Probabilistic Ensemble Generation Using a Diffusion Model Trained on JMA MSM Data Natsumi Saito