Sea-Ice seminar
June 26 – h 12:00 CEST

Title:
Historical humpback whales catch records and the assessment of early 20th-century sea ice edge in climate models

Speaker:
Marcello Vichi
Professor at the University of Cape Town

Abstract:

The assessment of historical environmental conditions in the Southern Ocean is limited by the poor availability of oceanographic records during the 20th century, up until the advent of remote-sensing observations. Whale catch data can help to fill this gap, because they represent a wealth of information on the habitat distribution of formerly commercial whales and the underlying environmental conditions at the time of capturing. These data have been used (and severely criticized) in the past as proxies for the summer sea ice extent. Using historical whale records in the oceanographic context is hampered by the different whaling practices and scattered or incomplete written documents. However, the selection of certain species that have known relationships with their habitats can give us indications of the larger-scale Southern Ocean features. This is the case of humpback whales, which have been severely decimated during the first decades of the 20th century. This species is usually found in open waters adjacent to the sea ice edge, making it a suitable candidate for assessing the northward extent of sea ice in a period of complete data absence at the circumpolar scale. We report on the interdisciplinary approach used to combine historical whale catches with sea-ice model simulations from the Climate Model Intercomparison Project phase 6 (CMIP6) during the historical period 1920-1970. Models were ranked according to their skills to reproduce current climate seasonality of sea ice extent. We found that the simulated sea-ice edge from the best-performing models is located northward of the regions in which humpback whales where historically caught in November and December. The high-ranking models, which also show reduced decreasing trends in line with the satellite products, have a higher extent at the beginning of the last century, possibly to compensate for the anthropogenic-driven decline. On the other hand, there is good confidence that the minimum extent in February is well captured by the models, and this dataset can also be used in the next CMIPs to validate the pre-industrial simulations. This Southern Ocean story informs us that tuning and assessing climate models only using satellite-era observations may lead to an overestimation of the Antarctic sea-ice extent earlier in the century, which may impact on our future predictability.