Benchmarking of Data-Driven Causality Discovery Approaches in the Interactions of Arctic Sea Ice and Atmosphere CyberTraining: Big Data + High-Performance Computing + Atmospheric Sciences ...

• Main Authors: Huang, Yiyi, Kleindessner, Matth¨Aus, Munishkin, Alexey, Varshney, Debvrat, Guo, Pei, Wang, Jianwu
• Format: Article in Journal/Newspaper
• Language: English
• Published: UMBC HPCF 2020
• Subjects: UMBC High Performance Computer Facility HPCF; Arctic; Sea ice
• Online Access: https://dx.doi.org/10.13016/m28nd1-di6h; https://mdsoar.org/handle/11603/21276

The Arctic sea ice has retreated rapidly in the past few decades, which is believed to be driven by various dynamic and thermodynamic processes in the atmosphere. The newly open water resulted from sea ice decline in turn exerts large influence on the atmosphere. Therefore, this study aims to investigate the causality between multiple atmospheric processes and sea ice variations using three distinct data-driven causality approaches: TCDF, NOTEARS and DAGGNN. We find that the static graphs produced by NOTEARS and DAG-GNN are relatively reasonable. The results from NOTEARS indicate that relative humidity and precipitation dominate sea ice changes among all variables, while the results from DAG-GNN suggest that the horizontal wind fields are more important for driving sea ice variations. However, both of them produce some unrealistic edges. In comparison, the temporal graphs generated by the three methods are not physically meaningful enough. It also turns out that the results are rather sensitive to the choice ...