A 20-Year Climatology of Sea Ice Leads Detected in Infrared Satellite Imagery Using a Convolutional Neural Network

Abstract

Sea ice leads, or fractures account for a small proportion of the Arctic Ocean surface area, but play a critical role in the energy and moisture exchanges between the ocean and atmosphere. As the sea ice area and volume in the Arctic has declined over the past few decades, changes in sea ice leads have not been studied as extensively. A recently developed approach uses artificial intelligence (AI) and satellite thermal infrared window data to build a twenty-year archive of sea ice lead detects with Moderate Resolution Imaging Spectroradiometer (MODIS) and later, an archive from Visible Infrared Imaging Radiometer Suite (VIIRS). The results are now available and show significant improvement over previously published methods. The AI method results have higher detection rates and a high level detection agreement between MODIS and VIIRS. Analysis over the winter season from 2002–2003 through to the 2021–2022 archive reveals lead detections have a small decreasing trend in lead area that can be attributed to increasing cloud cover in the Arctic. This work reveals that leads are becoming increasingly difficult to detect rather than less likely to occur. Although the trend is small and on the same order of magnitude as the uncertainty, leads are likely increasing at a rate of 3700 km2 per year with a range of uncertainty of 3500 km2 after the impact of cloud cover changes are removed.

Breaking the Ice: Exploring the Changing Dynamics of Winter Breakup Events in the Beaufort Sea

“…This is similar to Hoffman et al. (2022) who observed a small, but significant increase in pan‐Arctic leads from satellite data over the same period, despite large uncertainty due to the increasing cloud cover in the Arctic.…”

“…Meanwhile, recent observational data based on MODIS imagery (Willmes et al, 2023) show a significant positive trend of 2% per decade in lead frequencies in the Beaufort Sea over the period from 2002 to 2021 (during April only). This is similar to Hoffman et al (2022) who observed a small, but significant increase in pan-Arctic leads from satellite data over the same period, despite large uncertainty due to the increasing cloud cover in the Arctic.…”

The Application of a Convolutional Neural Network for the Detection of Contrails in Satellite Imagery

“…Finally, as validation, a photograph is shown looking northward from southern Wisconsin, taken at 4:05 P.M. local time (2105 UTC). The primary advantage of the detection method is that it uses techniques previously demonstrated to be effective in satellite remote sensing applications [12,13], and that it can be applied to this new application without changing the model architecture. The technique also is also relatively simple to run in an operational sense because the detections are based on a single brightness temperature difference image rather than more complex multi-spectral time series of images [18].…”

“…The first step in the formation of a contrail detection method is to utilize a dataset of contrail images that can be used to train, test, and validate a detection method [18]. The detection architecture is based on the same AI architecture that was used to detect quasi-linear sea ice lead features as described by Hoffman et al in 2021 [12] and updated in 2022 [13]. Both the contrail and sea ice lead detection methods uses a particular kind of convolutional neural network, U-Net, to perform image segmentation, which was first described by Ronneberger et al in 2015 [11].…”

Temperature and Relative Humidity Profile Retrieval from Fengyun-3D/VASS in the Arctic Region Using Neural Networks

“…Climate change in the Arctic has been the focus of scientific research in recent years [1][2][3][4], but the polar climate in this area of the planet is extremely harsh, with low temperatures and ice for many years [5]. The Arctic region has long winters, extremely low temperatures, strong air-sea exchanges, and high humidity, which is manifested as fog.…”