As Arctic Ice Vanishes, Scientists Turn to AI to Improve Melting Forecasts

Arctic sea ice is melting rapidly, making traditional forecasting methods outdated. Scientists are turning to AI and machine learning to improve predictions.

AI models like PolArctic and IceNet are showing early promise in sea ice forecasting, sometimes outperforming physics-based models.

AI models can provide higher resolution, localized forecasts to aid Arctic shipping and resource development. But the "black box" nature of AI poses explainability challenges.

Long-term climate modeling still relies heavily on physics and statistics. Combining these with machine learning may yield the most accurate forecasts.

Incorporating Indigenous knowledge into training data could further boost AI model accuracy, says Alaska Native AI expert Leslie Canavera.