Dataset Information

Seasonal Arctic sea ice forecasting with probabilistic deep learning.

ABSTRACT: Anthropogenic warming has led to an unprecedented year-round reduction in Arctic sea ice extent. This has far-reaching consequences for indigenous and local communities, polar ecosystems, and global climate, motivating the need for accurate seasonal sea ice forecasts. While physics-based dynamical models can successfully forecast sea ice concentration several weeks ahead, they struggle to outperform simple statistical benchmarks at longer lead times. We present a probabilistic, deep learning sea ice forecasting system, IceNet. The system has been trained on climate simulations and observational data to forecast the next 6 months of monthly-averaged sea ice concentration maps. We show that IceNet advances the range of accurate sea ice forecasts, outperforming a state-of-the-art dynamical model in seasonal forecasts of summer sea ice, particularly for extreme sea ice events. This step-change in sea ice forecasting ability brings us closer to conservation tools that mitigate risks associated with rapid sea ice loss.

SUBMITTER: Andersson TR

PROVIDER: S-EPMC8390499 | biostudies-literature | 2021 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Seasonal Arctic sea ice forecasting with probabilistic deep learning.

Andersson Tom R TR Hosking J Scott JS Pérez-Ortiz María M Paige Brooks B Elliott Andrew A Russell Chris C Law Stephen S Jones Daniel C DC Wilkinson Jeremy J Phillips Tony T Byrne James J Tietsche Steffen S Sarojini Beena Balan BB Blanchard-Wrigglesworth Eduardo E Aksenov Yevgeny Y Downie Rod R Shuckburgh Emily E

Nature communications 20210826 1

Anthropogenic warming has led to an unprecedented year-round reduction in Arctic sea ice extent. This has far-reaching consequences for indigenous and local communities, polar ecosystems, and global climate, motivating the need for accurate seasonal sea ice forecasts. While physics-based dynamical models can successfully forecast sea ice concentration several weeks ahead, they struggle to outperform simple statistical benchmarks at longer lead times. We present a probabilistic, deep learning sea ...[more]

PMID: 34446701

Similar Datasets

Project description:Quantitative camera surveys of benthic megafauna were carried out during the expedition ARK-XXVII/3 to the Eastern Central Arctic Basins with the research icebreaker Polarstern in summer 2012 (2 August-29 September). Nine transects were performed for the first time in deep-sea areas previously fully covered by ice, four of them in the Nansen Basin (3571-4066m) and five in the Amundsen Basin (4041-4384m). At seven of these stations benthic Agassiz trawls were taken near the camera tracks for species identification. Observed Arctic deep-sea megafauna was largely endemic. Several taxa showed a substantially greater depth or geographical range than previously assumed. Variations in the composition and structure of megabenthic communities were analysed and linked to several environmental variables, including state of the sea ice and phytodetritus supply to the seafloor. Three different types of communities were identified based on species dominating the biomass. Among these species were the actiniarian Bathyphellia margaritacea and the holothurians Elpidia heckeri and Kolga hyalina. Variations in megafaunal abundance were first of all related to the proximity to the marginal ice zone. Stations located closer to this zone were characterized by relatively high densities and biomass of B. margaritacea. Food supply was higher at these stations, as suggested by enhanced concentrations of pigments, organic carbon, bacterial cell abundances and nutrients in the sediments. Fully ice-covered stations closer to the North Pole and partially under multi-year ice were characterized by lower concentrations of the same biogeochemical indicators for food supply. These stations nevertheless hosted relatively high density and biomass of the holothurians E. heckeri or K. hyalina, which were observed to feed on large food falls of the sea-ice colonial diatom Melosira arctica. The link between the community structure of megafauna and the extent and condition of the Central Arctic sea-ice cover suggests that future climate changes may substantially affect deep ocean biodiversity.

Dataset Information

Seasonal Arctic sea ice forecasting with probabilistic deep learning.

Publications

Seasonal Arctic sea ice forecasting with probabilistic deep learning.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets