Artificial intelligence to help predict Arctic sea ice loss

IceNet determine. Supply: British Antarctic Survey

A brand new AI (synthetic intelligence) instrument is about to allow scientists to extra precisely predict Arctic sea ice situations months sooner or later. The improved predictions may assist new early warning methods that shield arctic wildlife and coastal communities from the results of sea ice loss.

A global workforce of researchers led by the British Antarctic Survey (BAS) and The Alan Turing Institute, printed this week in Nature Communications, describes how the IceNet AI system addresses the problem of offering correct Arctic sea ice forecasts for the approaching season create. one thing that scientists have missed for many years.

Sea ice, an enormous layer of frozen seawater that seems on the North and South Poles, is notoriously tough to foretell due to its advanced relationship with the environment above and the ocean beneath. The sensitivity of sea ice to rising temperatures has resulted within the Arctic summer season sea ice space halving over the previous 4 a long time, a lack of space roughly 25 occasions the scale of Nice Britain. These accelerating adjustments have dramatic penalties for our local weather, arctic ecosystems, and indigenous and native communities whose livelihoods are tied to the seasonal sea ice cycle.

IceNet, the AI ​​prediction instrument, is almost 95% correct at predicting whether or not sea ice can be current in two months – higher than the main physics-based mannequin.

Lead writer Tom Andersson, an information scientist on the BAS AI Lab and funded by the Alan Turing Institute, states, “The Arctic is a area on the entrance traces of local weather change and has warmed considerably over the previous 40 years. IceNet has the potential to shut an pressing one Gaps in sea ice prediction for Arctic sustainability efforts and runs a thousand occasions quicker than conventional strategies. “

Dr. Scott Hosking, Principal Investigator, Co-Head of the BAS AI Lab and Senior Analysis Fellow on the Alan Turing Institute, says, “I am excited to see how AI is driving us to rethink our environmental analysis. Our new sea ice prediction framework combines knowledge from satellite tv for pc sensors with outcomes from local weather fashions in ways in which conventional methods merely could not. “

Not like conventional prediction methods that try to mannequin the legal guidelines of physics instantly, the authors developed IceNet based mostly on an idea referred to as deep studying. By way of this method, the mannequin “learns” from 1000’s of years of local weather simulation knowledge, together with a long time of observational knowledge, how the ocean ice is altering in an effort to predict the extent of the Arctic sea ice months sooner or later.

Tom Andersson concludes, “Now we have proven that AI can precisely predict sea ice. Our subsequent purpose is to develop a each day model of the mannequin and run it publicly in actual time, in addition to climate forecasting early warning system for dangers related to fast sea ice loss. ”

Constructing a greater mannequin of arctic ecosystems

Extra info:
Seasonal arctic sea ice forecast with probabilistic deep studying, Nature Communications (2021). Offered by the British Antarctic Survey

Quote: Synthetic Intelligence to Predict Arctic Sea Ice Loss (2021, August 26), accessed on August 26, 2021 from

This doc is topic to copyright. Apart from honest commerce for personal research or analysis functions, no half could also be reproduced with out written permission. The content material is offered for informational functions solely.

Source link

Leave a Comment