Reconstructing historic Southern Ocean sea ice from Norwegian ship logs


Whaling around Antarctica in the first half of the 20th century had a devastating effect on the global population of blue and fin whales. These depredations left behind impoverished seas, but also intriguing relics in the form of handwritten records of conditions in the Southern Ocean.

Intense whaling in the Southern Ocean throughout the early- to mid-20th century is notorious for its effect on whale populations. Although it is an iconic example of unsustainable economic activity, the whaling left behind an extensive stratum of documents and accounts on a wide range of subjects, from economics to safety management. This rich historic source material has been, and still should be studied. One largely underexplored area where one can successfully mine for knowledge is sea ice and climate of the Antarctic region.

Photo: Foto: Hvalfangstmuseets fotoarkiv
By: Dmitry V Divine, Svetlana Divina, Elisabeth Isaksson, Ivar Stokkeland and Harald Dag Jølle // Norwegian Polar Institute. Clive Wilkinson // CSW Associates Data Services, UK

Our understanding of the Antarctic climate suffers from a lack of high-quality instrumental observations from the region throughout a substantial part of the 20th century. The first observational network in the region was established during the International Geophysical Year of 1957, whereas regular sea ice data are not available until the start of satellite-based monitoring in 1979. This limits the capabilities of climate models and climate reanalysis products for the region in the pre-1980 period, with implications for understanding the signature and effects of ongoing anthropogenic warming in the Antarctic region.

Collection of journals with meteorological observations conducted onboard Norwegian factory ships in the Southern Ocean during 1932-1939. Photo: Norwegian Polar Institute library

Complementary data sources such as ships’ logbooks have already proven to be a successful tool in reconstructing past marine climate. Although recent years have seen significant efforts to recover information from ships’ logbooks, data from the Antarctic region are largely yet to be digitised and analysed. In contrast to logbooks from infrequent Antarctic expeditions, accounts from commercial vessels are much more abundant and represent a promising source of valuable climate information.

This article is based on a project that aimed to reconstruct past sea ice extent and climate variability in the Southern Ocean using data sources previously unavailable to modern tools and methods of data analysis.

Historical data example: fragment of a page in the logbook of Antarctic for the afternoon of 15 March 1931, showing information on weather (green boxes) and sea ice (red boxes). The vessel is reported to be drifting at the ice edge (“beliggende ved iskanten”) or in ice pack (“beliggende i isen”) at S69.5°, E26.55°. Weather after conversion to modern units: air temperature about -5°C; easterly winds 9-13 m/s, overcast, snow flurries.

Historical sources reveal climate

Once economic activity in the region started in the late 19th century, the number of vessels from various national whaling fleets operating in the area during austral late spring to early fall grew until the introduction of restrictions on the Southern Ocean whaling in the late 1960s. In the early 1930s, dozens of vessels could annually be active in the whaling grounds of the Southern Ocean. When these ships were sailing at high latitudes, relevant information on marine climate, including sea ice, would be registered on a regular basis in a captain’s/first mate’s logbook (“Dekksdagbok”), a catchbook (“Fangstdagbok”) and inspector’s diaries and reports. The frequency of these notes and the quality of the observations vary between the documents and types of vessels, but they were generally more consistent for factory ships.

Analysed vessel positions at which ship logs provide information on weather and the state of the ice cover in the 1930s.

For analysis, we primarily selected accounts providing relevant information on at least a daily basis. Naturally, not all documents have survived, but several relevant logbooks from Norwegian vessels were archived in the Vestfold library and the Whaling Museum in Sandefjord. During a recent recovery effort many logs were inventoried, imaged, and made available for processing and analysis.

Rediscovering forgotten treasures

The idea of opportunistically using commercial whaling to establish a meteorological observational network in Antarctica emerged already during preparations for the second International Polar Year (1932-1933). The national committee led by H U Sverdrup (at the time a professor at the Geophysical Institute of the University of Bergen), with support from the Norwegian Meteorological Institute and through a collaboration with Hvalfangerforeningen (Whalers Union of Norway), agreed to establish up-to-date meteorological stations onboard ten Norwegian factory ships.

A centralised purchase of standardised meteorological instruments followed, accompanied by dedicated training of a few crew members onboard each vessel to ensure that the observations would be conducted in accordance with international standards. Though the number of vessels participating in this programme decreased over time, the legacy of this initiative now comprises a collection of more than 50 meteorological journals covering seven whaling seasons in the period of 1932-1939.

Positions of sea ice observations/sea ice edge for November (month close to the maximum sea ice extent in the Southern Ocean) with inferred sea ice concentrations during the 1930s shown together with modern median satellite-based sea ice presence, including the 15% ice edge over 1981-2010. Note that most of the blue circles indicating historic presence of ice are located within the area where sea ice is present in modern times. In the Indian sector of the Southern Ocean (60 to 90°E) some of the positions recovered from daily ships logs and reported as being next to the ice edge may lie outside the contemporary seasonal ice zone.

Even though these observations were used in the early research on the climate of the region, they seem to have been forgotten, and the compete collection was never transferred into any modern digital climate data archive. Given a high demand for good-quality climate data for this period and region, an effort for their digitalisation, reanalysis and publishing is needed.

Antarctic sea ice cover before 1979

Sea ice is a convenient climate indicator and straightforward to observe. Though discrimination between ice types requires specific knowledge, the mere fact that sea ice presence is registered in a particular location tells a lot about the state of the climate system on spatial scales from local to regional.

Previous attempts to infer 20th century sea ice changes in the Southern Ocean were based on the analysis of whale catch positions. These studies suggested that Antarctic sea ice has receded substantially (by 3-8 degrees latitude) since the 1930s. However, analysis of logbooks shows that whaling could occur both inside and well outside the ice pack. This speaks for using the more reliable actual reports of sea ice or open water conditions instead. Strong intraseasonal changes in the Antarctic sea ice extent should also be taken into consideration, to avoid the effect of seasonal biases in the inferred status of the sea ice pack.

As of December 2021, more than 30 documents in Norwegian from the 1930s had been transcribed and analysed in the first phase of the project. Accounts from the factory ships Svend Foyn, Antarctic, New Sevilla, Hektoria, Thorshammer, Terje Viken and the research vessel Norvegia provided about 4000 actual observations of air temperature, wind force and direction, state of the sea surface, and sometimes sea surface temperature. About 1300 observations contained information on sea ice or open water conditions, providing a valuable insight on sea ice conditions during this period.

Why focus specifically on the 1930s?

While it would be easiest to point to the generous number of archival sources available for this period, this is far from being the principal answer. By coincidence, these sources bear witness to conditions in the Southern Hemisphere during the 1920s-1930s when the Northern Hemisphere experienced what is now called “the early 20th century warming”. In the Arctic this prominent climate swing led to a substantial reduction of sea ice extent and positive air temperature anomalies that were not exceeded again until the past two decades.

This motivates assessing a possible Antarctic manifestation of this phenomenon, and comparing the relative phasing of climate evolution between the two hemispheres. This may have implications for understanding and predicting the future effects of ongoing anthropogenic warming on the Antarctic region.


Analysis of the historic sea ice observations recovered so far does not provide evidence for a substantial recession of sea ice in the Antarctic region since that time. We note that past sea ice observations in late winter in this region extended 2-3 degrees latitude farther north than the modern average sea ice edge position as shown by satellite observations of 1981-2010. However, this comparison also shows that past ice edge position/ice observations mainly lie within the natural variability over the satellite period. These results motivate further exploration and exploitation of available Norwegian historical maritime sources. This would help fill the gaps in our knowledge about pre-satellite era sea ice and climate variability in the Southern Ocean. In addition to encouraging the analysis and publication of results, we recommend that these invaluable data be integrated into existing climate databases, thus making them available for future research.

This would help fill the gaps in our knowledge about pre-satellite era sea ice and climate variability in the Southern Ocean. In addition to encouraging the analysis and publication of results, we recommend that these invaluable data be integrated into existing climate databases, thus making them available for future research.