New algorithm could predict Arctic sea ice up to 9 months ahead, offering new climate insights(opens in a new tab)
New algorithm could predict Arctic sea ice up to 9 months ahead, offering new climate insights
Event
NYU Abu Dhabi releases Random Analogue Predictor for Arctic sea‑ice forecasting up to nine months ahead
Researchers at Mubadala ACCESS, NYU Abu Dhabi released the Random Analogue Predictor (RAP), a data‑driven algorithm that forecasts Arctic sea‑ice extent up to nine months in advance using historical records. RAP’s ensemble forecasts match the skill of the 34 models in the Sea Ice Prediction Network and provide uncertainty estimates. The team views RAP as a benchmark for more complex models and a low‑cost tool that could aid UAE polar research.
Coverage
Phys Eurekalert New NYU Abu Dhabi algorithm could predict Arctic sea ice up to nine months ahead ...(opens in a new tab)
Researchers at NYU Abu Dhabi’s Mubadala ACCESS developed the Random Analogue Predictor, a data‑driven algorithm that can forecast Arctic sea ice extent up to nine months ahead, providing forecast skill comparable to complex seasonal models and an estimate of uncertainty.
Nyuad New NYU Abu Dhabi Algorithm Could Predict Arctic Sea Ice Up to Nine Months Ahead(opens in a new tab)
A new algorithm from NYU Abu Dhabi can forecast Arctic sea‑ice extent up to nine months in advance, matching the skill of current complex models while providing uncertainty estimates.
Nature Random analog prediction provides a benchmark for seasonal Arctic sea ice extent forecasting(opens in a new tab)
A new random‑analog predictor algorithm has been developed to forecast seasonal Arctic sea ice extent. Using only the historical sea ice record, the method produces ensemble forecasts with negligible bias and an RMS error under 0.6 × 10^6 km², achieving skill comparable to Sea Ice Prediction Network models. Its simplicity, interpretability and self‑derived uncertainty estimates make it a useful baseline for physics‑ and AI‑based sea‑ice models.
