Graph Neural Network as Computationally Efficient Emulator of Ice-sheet and Sea-level System Model (ISSM)

TL;DR

This paper introduces a GPU-accelerated graph neural network that emulates the Ice-sheet and Sea-level System Model with high accuracy and significantly improved computational speed, enabling rapid future predictions of ice sheet changes.

Contribution

It presents a novel GCN-based emulator for ISSM that runs efficiently on GPUs, outperforming traditional CNNs and enabling faster ice sheet modeling.

Findings

GCN achieves correlation > 0.998 in reproducing ice thickness and velocity.

GCN runs 34 times faster than CPU-based ISSM.

Enables rapid future scenario predictions for ice sheet changes.

Abstract

The Ice-sheet and Sea-level System Model (ISSM) provides solutions for Stokes equations relevant to ice sheet dynamics by employing finite element and fine mesh adaption. However, since its finite element method is compatible only with Central Processing Units (CPU), the ISSM has limits on further economizing computational time. Thus, by taking advantage of Graphics Processing Units (GPUs), we design a graph convolutional network (GCN) as a fast emulator for ISSM. The GCN is trained and tested using the 20-year transient ISSM simulations in the Pine Island Glacier (PIG). The GCN reproduces ice thickness and velocity with a correlation coefficient greater than 0.998, outperforming the traditional convolutional neural network (CNN). Additionally, GCN shows 34 times faster computational speed than the CPU-based ISSM modeling. The GPU-based GCN emulator allows us to predict how the PIG will change in the future under different melting rate scenarios with high fidelity and much faster computational time.