Integrating Fourier Transform and Residual Learning for Arctic Sea Ice Forecasting

Louis Lapp,Sahara Ali,Jianwu Wang

DOI: https://doi.org/10.48550/arXiv.2312.13445

2023-12-21

Abstract:Arctic sea ice plays integral roles in both polar and global environmental systems, notably ecosystems, communities, and economies. As sea ice continues to decline due to climate change, it has become imperative to accurately predict the future of sea ice extent (SIE). Using datasets of Arctic meteorological and SIE variables spanning 1979 to 2021, we propose architectures capable of processing multivariate time series and spatiotemporal data. Our proposed framework consists of ensembled stacked Fourier Transform signals (FFTstack) and Gradient Boosting models. In FFTstack, grid search iteratively detects the optimal combination of representative FFT signals, a process that improves upon current FFT implementations and deseasonalizers. An optimized Gradient Boosting Regressor is then trained on the residual of the FFTstack output. Through experiment, we found that the models trained on both multivariate and spatiotemporal time series data performed either similar to or better than models in existing research. In addition, we found that integration of FFTstack improves the performance of current multivariate time series models. We conclude that the high flexibility and performance of this methodology have promising applications in guiding future adaptation, resilience, and mitigation efforts in response to Arctic sea ice retreat.

Atmospheric and Oceanic Physics

What problem does this paper attempt to address?

The paper attempts to address the problem of accurate prediction of Arctic Sea Ice Extent (SIE). Specifically, the paper aims to propose a new framework that combines Fourier Transform and Residual Learning techniques, along with multivariate time series and spatiotemporal data, to improve the prediction accuracy of Arctic Sea Ice Extent.

### Background
Arctic sea ice plays a crucial role in the polar and global environmental systems, including ecosystems, communities, and economies. Due to climate change, Arctic sea ice is continuously decreasing, making accurate predictions of future sea ice extent critically important. However, existing prediction models often underestimate the rate of sea ice decline, necessitating improvements.

### Research Objectives
1. **Propose a new prediction framework**: Combine Fourier Transform and Residual Learning to propose a new prediction framework capable of handling multivariate time series and spatiotemporal data.
2. **Improve prediction accuracy**: Validate through experiments that the new framework can achieve or exceed the performance of existing studies in predicting Arctic sea ice extent.
3. **Application prospects**: Explore the potential applications of this method in guiding future adaptation, resilience, and mitigation measures.

### Methods
1. **Data preprocessing**: Preprocess 2D and 3D data, including linear detrending and Fourier Transform.
2. **FFTstack**: Propose an iterative system based on Fourier Transform (FFTstack) for modeling periodic time trends.
3. **Residual Learning**: Train a Gradient Boosting Regressor on the residuals generated by FFTstack to further improve prediction accuracy.
4. **Architecture design**: Develop two architectures to handle 2D multivariate time series data and 3D spatiotemporal data, respectively.

### Results
1. **Performance metrics**: Evaluate model performance using metrics such as RMSE, NRMSE, and R².
2. **Comparative analysis**: Compare with models from existing studies, showing that the 2D GBR (de-cycled) model outperforms other models on all performance metrics.
3. **Long-term correlation**: The 2D GBR (de-cycled) model demonstrates good stability in long-term predictions.

### Conclusion
The FFTstack framework proposed in this study performs excellently in predicting Arctic sea ice extent, not only improving prediction accuracy but also showing long-term prediction correlation. This method can be applied to related research, such as predicting sea level rise, species decline, or abnormal weather patterns, helping to address the challenges posed by the reduction of Arctic sea ice.