Collections »

About this collection

Extent

1 digital object.

Cite This Work

Hoffman, Lauren; Mazloff, Matthew R.; Gille, Sarah T.; Giglio, Donata; Bitz, Cecilia M.; Heimbach, Patrick; Matsuyoshi, Kayli. (2023). Data From: Machine learning for daily forecasts of Arctic sea-ice motion: an attribution assessment of model predictive skill. UC San Diego Library Digital Collections. https://doi.org/10.6075/J0X06774

Description

With the aim of using machine learning as a tool to predict and understand sea-ice motion in the Arctic on one-day timescales, these data include processed satellite and reanalysis measurements of sea-ice velocity, sea-ice concentration, and wind velocity. Also included are outputs from statistical model predictions. Finally, we include all files required to download and process raw data, run statistical models, and plot analyses of outputs.

Date Collected

• 1989 to 2021

Date Issued

• 2023

Author

• Hoffman, Lauren A.

Advisors

• Giglio, Donata
• Gille, Sarah T.
• Mazloff, Matthew R.

Contributors

• Bitz, Cecilia M.
• Heimbach, Patrick
• Matsuyoshi, Kayli

Funding

Outputs from machine learning models were funded by Office of Naval Research grant N00014-20-1-2772.

Geographic

• Arctic Ocean

Topics

• Machine learning
• Operational forecasting
• Prediction
• Sea-ice velocity
• Wind velocity

Format

View formats within this collection

• data

Language

• English

Identifier

Identifier: Cecilia M. Bitz: https://orcid.org/0000-0002-9477-7499

Identifier: Donata Giglio: https://orcid.org/0000-0002-3738-4293

Identifier: Lauren A. Hoffman: https://orcid.org/0000-0002-9563-8925

Identifier: Matthew R. Mazloff: https://orcid.org/0000-0002-1650-5850

Identifier: Patrick Heimbach: https://orcid.org/0000-0003-3925-6161

Identifier: Sarah T. Gille: https://orcid.org/0000-0001-9144-4368

Related Resources

Primary associated publication

• Hoffman, L., M. Mazloff, S.T. Gille, D. Giglio, P. Heimbach, C. Bitz, K. Matsuyoshi (2023). Machine learning for daily forecasts of Arctic sea-ice motion: an attribution assessment of model predictive skill. Submitted.

Source data

• JRA-55 based surface dataset for driving ocean-sea ice models (JRA55-do), wind velocity: https://climate.mri-jma.go.jp/pub/ocean/JRA55-do/
• Polar Pathfinder Daily 25 km EASE-Grid Sea Ice Motion Vectors, Version 4: https://nsidc.org/data/nsidc-0116/versions/4
• Sea Ice Concentrations from Nimbus-7 Passive Microwave Data, Version 1: https://doi.org/10.5067/8GQ8LZQVL0VL

Collection image

• Image source: Lauren Hoffman, "Performance of machine learning model built to predict sea-ice motion in the Arctic." Maps showing the performance of a machine learning model built to predict sea-ice motion in the Arctic. (Left) A convolutional neural network (CNN) makes skillful predictions of sea-ice motion in the Arctic, where lighter colors are indicative of a higher predictive skill in the map of the first panel. (Right) A CNN typically outperforms a traditional linear regression (LR) model in predicting Arctic sea-ice motion, where red is indicative of areas where the CNN outperforms LR in the map of the second panel.