The availability of large water-bodies so high in the Northern Hemisphere are the ultimate ‘blueprint’ for the Northern Hemisphere climate, which needs to be recognised in the first place. However such plain statement cannot explain the ‘climate revolution’ in the region in the early 20 th Century. Calling this initial event ‘Arctic Warming’ might even be misleading, if the warming commenced primarily over the northern part of the North Atlantic, and even more precisely, in the Spitsbergen region. On the other hand all ocean space and seas within the Arctic Circle are relevant for the arctic weather and arctic climate, although on a very different scale due to size, depths, sea ice, freshwater, sun-less periods, and sea currents, to mention only few decisive factors. Detecting the time and source of the sudden start of the Arctic Warming requires to have a look at each of the specific sea areas in question. This needs to keep in mind that the ‘explosion’ of air temperatures at the Spitsbergen archipelago at 80º North during winter 1918/19, must have been generated by ‘something’. Although no option, including an internal atmospheric process, should be excluded, the most likely source is the sea area between the direction 135º (SE) and 270º (West) of Spitsbergen, which is usually sea ice-free throughout the year and belongs to the Barents Sea, the Norwegian Sea and the Greenland Sea. The source of the warming was presumably due to either internal processes within the water bodies, or influenced by more warm water coming from the Atlantic Gulf current, or both. The latter came with the Norwegian Current and West Spitsbergen Current, formed by water flowing from the Gulf Current after it had passed the Iceland – Faroe – Scotland line, enhanced by North Sea water, and continental run off rain and melt water. However, once the warm Atlantic water has passed the north of the Scottish Hebrides and Faroe, and travelling northwards, things tend to complicate.

Due to the size of the Arctic Ocean, which is much larger than the Northern North Atlantic[1], and due to its considerable depths of more than 3000 metres, one would assume, that its immediate impact on the weather should succeed by far the influence of the Northern North Atlantic. That is not the case. At least during the winter periods in the early 20 th Century the Arctic Ocean was permanently covered with sea ice, which diminish any heat release from the sea to the atmosphere to a small factor, thus making this huge water body ‘continental’, characterized by cold and stable weather conditions, and clear skies.[2]. Although the relatively warm ocean surface water (slightly below zero degrees C) has an influence on the regional air temperatures. Its impact, compared to an open sea area, is remote.

Despite the fact that the water body of Arctic Oceans is based on many complex features, and highly influenced by internal transformation, e.g. sea ice freezing and melting, and external influences, e.g. fresh water from rivers, there is not one aspect from which a sudden warming could have been generated. None of the seas in question can be excluded with such unequivocal certainty from the list of potential contributor for the early Arctic Warming as the Arctic Ocean.

Scenario 2 – Barents Sea

It is not easy to assess how much the up to ca. 500 m deep Barents Sea, with an average depth of 200 metres, might have contributed to the ‘Severe Warming’ . Generally speaking, the Atlantic water ‘disappears’ in the East of the North Cape. In the northern part polar water flows from Northeast to Southwest and partly joins the Spitsbergen Current in the south of Spitsbergen and Bear Island. The North Cape Current, and its subsequent currents, which supply the eastern part of the Barents Sea with Atlantic water, may have contributed to the warming in the long run. In so far it needs only to be noted that the air temperatures in North Norway increased only modestly since 1920[3]. Neither Vardö[4] (close to the North Cape), nor any Russian Station reported any exceptional winter temperature rise. O n a 10-year mean basis (1911/20 and 1921/30), a significant increase of 6°C was observed at Franz-Joseph Land[5], whereby the highest water temperatures in the top 200 m of the Barents Sea at 70-72°N; 33°E, north of the Kola Peninsula, shall have been reached during the period 1935-39[6]. That all confirms a slow process but hardly any facts that the Barents Sea had significantly generated a temperature jump in its southern part (North Cape), or in the eastern part (Kola Peninsula ).

The observed warming at Franz-Joseph-Land can also be hardly connected to the sea, which is partly supplied via the North Cape current and subsequent currents along the Kola Peninsula. In addition the Barents Sea between Spitsbergen (East) and Franz-Joseph-Land (West) is not very deep and is governed by very cold currents flowing southwest towards South Spitsbergen and the Bear Island. (See also next section).

According to Wagner[7], the mean water temperatures in the Barents Sea increased with +1.8°C from 1912/18 to 1919/28. From 1916 to 1925, the annual mean water temperature was as it follows[8]:

1914 = -0.3°C	1915 = + 0.7°C	1916 = -1.1°C	1917 = -1.5°C	1918 = -1.6°C
1919 = + 0.6°C	1920 = +1.0°C (?)	1921 = +1.0°C	1922 = +1.9°C	1923= +1.0°C

Wagner’s additional observations ascertain a ‘rise’ of 2-3°C, at water depths of 100 and 200 m, during the last 30 years (1895 and 1927). However, a general observation of ‘over 30 years’ is of little help in this case, presumably also his general assessment that the Barents Sea ice border retreated significantly since 1919[9], even though it is undisputed that the retreat occurred gradually over a two-decade period.

There are no significant indications that the Barents Sea contributed significantly in the early stage of the Arctic Warming. The presumably most relevant aspect is that no particular strong winter warming had been observed at the southern parts of the Barents Sea, albeit a trend change was observed since 1919. The question would remain, whether the trend change had been caused by higher temperatures in the North, or alone by the Barents Sea, or was due to changes over the European Continent, or had been a combination of all three factors. But this does not need to be answered here.

In the west of Spitsbergen, the seawater has a temperature of 5°C and a salinity of 34.90 –35.00 mg. A significant part of the warm Atlantic Gulf water that has reached Spitsbergen ‘turns left’ in the south-western direction, at the position of 75-77° North, and flows either as Greenland current down to Newfoundland and back in the Atlantic, or goes down into the huge Greenland Sea Basin with depths of 2,000 metres (max. ca. 3,500 m), or circles for some time at the sea surface water layer or in sub surface layers.

Although the Greenland Sea represents a huge water body, the option for being a serious contributor to the extreme warming event in winter 1918/19 is remote. The West Greenland Current is colder and less salty than the Atlantic water of the Spitsbergen Current, and the current is to a considerable extent covered with sea ice during the winter season (see:Annex C). This sea area may have contributed to the warming many months later, but not longer than until ca. 1930[10]. This clearly indicates that the Greenland Sea, at least near Greenland, had received extra warm water only temporarily.

The southern part of Norwegian Sea from 65ºN (ca. Iceland/ middle Norway) to 40ºN (Lisbon – New York) a zone where the North Atlantic Oscillation (NAO) represents the average strength of the eastern Atlantic westerly flow in the region did not deviated from long term average[11]. Neither was any other observation in the sub-polar Atlantic been made which would require a more detailed investigation. Our interest will therefore focus more on the northern part of the Norwegian Sea, together with the currents passing through the Norwegian Sea from South to North.

The Norwegian Sea Basin is up to ca. 3,000 metres deep, and separated from the Arctic Ocean by a ridge of about 600 metres under the sea surface. This allows huge water exchange from north to south and vice versa. A key role for the moderate temperature conditions in the region derives from huge warm water supply by a branch of the Atlantic Gulf currents which is ‘pushed down’ to lower depths after passing by the Shetland Islands, Faroe Island and Iceland ridge (approx. 500 m).

A considerable part of the Atlantic water moves via the currents towards the basin of the Arctic Ocean. Actually, due to the high salinity of the warm Atlantic water and the cooling process, the water becomes very dense and ‘falls’ over a ridge (with a depth of 600 m below sea level) in the Arctic Basin. Before the Spitsbergen current reaches the ridge, at about 80° North, the water, has a depth of 20 metres, a salinity of about > 35 per mile, and a temperature of up to 7°C.

But also the basin of the Norwegian Sea – is a reservoir for warm Gulf water, reaching depths of 800 metres. Status and dynamics of the Norwegian Sea is also strongly influenced by other factors, particularly wind, rain, melt water, and the low saline water form the North Sea. Any increase in temperature, or enlargement of the ’warm water part’, or ‘change of dynamics’, would quickly be reflected in temperatures in Europe or elsewhere in the Northern Hemisphere.

 

Summary

The brief overview on the possible potential of the most relevant sea areas for the early Arctic Warming could show that two, out of the discussed four areas, can be excluded with high certainty as serious contributor, namely the Arctic Ocean and the Greenland Sea, and with less certainty the Barents Sea. The by far highest potential has the northern part of the Norwegian Sea and the Spitsbergen Current. This preliminarily assessment shall be verified with further information in the next section.

Footnotes:

[1] Here meant within the lines: Greenland, Iceland, Scotland, Norway, Spitsbergen, and Greenland.

[2] See: http://en.wikipedia.org/wiki/Arctic_Ocean#Climate

[3] Manley, Gordon; ‚Some recent contributions to the study of climatic change’, in: Quarterly Journal of R. Met. Soc., Vol. 73, 1944, Fig.4.

[4] John Daly, http://www.john-daly.com displays temperature data from stations in North Norway and refers to Vardö data as follows: Willis Eschenbach, willis@taunovobay.com , did a closer examination of Vardø and also found the same discontinuity around 1920, amounting to 0.73°C.  When that artificial discontinuity is discounted, the temperature rise is only +0.12°C per century, a tiny result for a region that according to the models should have undergone rampant warming in the last century .

[5] Kirch, Regina; ‚Temperatureverhältnisse in der Arktus währen der letzten 50 Jahre), Meteorologische Abhandlungen, Bd. LXIX, Heft 3, p.22, Fig 27.

[6] Lamb, H.H.; ‚Climate – Present, Past and Future’, Vol. 2, London, 1st ed. 1977, 2nd ed. ca. 1980s, p.528.

[7] Wagner, Arthur; ‘Klimaaenderungen und Klimaschwankungen’, Braunschweig, 1940, p. 50.

[8] Ditto, Table 10

[9] Ditto, p. 47

[10] See next section: Bjerkness, J; ‘The Recent Warming of the North Atlantic’; in: Bolin, Bert, ‚The Atmosphere and Sea in Motion’, Oxford 1959, p. 65ff.

[11] Lennart Bengtsson, Vladimir A. Semenov, Ola M. Johannessen, The Early Twentieth-Century Warming in the Arctic—A Possible Mechanism, Journal of Climate, October 2004, page 4045-4057.