THE TEMPERATURE OF THE ATLANTIC OCEAN 



1. Introductory Remarks on the Charts and the 

 Formation of the Two-Degree Field Averages 



The ideal objective in a description of the surface 

 conditions of an entire ocean would be to treat the 

 various phenomena such as temperature, salinitj^, 

 density, and drift collectively in order to reveal their 

 relationships. 



Such a direct procedure, however, is impossible in 

 the science of oceanography because of the great 

 differences in the accuracy and numbers of observations 

 of the individual components of the data. For example, 

 the ratio of available observations of temperature to 

 those of salinity in the Atlantic Ocean is about 19:1. 

 (See plates I, II, III, IV.) 



For this discussion of temperature there were drawn, 

 besides a chart of the annual average (plate V), 12 

 charts for each monthly average (plates VI-XVII), 

 similar ones for the annual anomaly (plates XVIII, 

 XIX), and in addition the monthly anomalies for 

 several months (plates XX-XXIII). The monthly 

 charts, which are to form the foundation of this report, 

 were drawn on the basis of the monthly averages for 

 the one-degree fields. In drawing the isotherms, all 

 observations, insofar as the surface unit of the one- 

 degree field and the distribution of the material 

 allowed, were taken into consideration as strictly as 

 possible. For this reason, the appearance of the charts, 

 differs from other presentations of the surface tempera- 

 ture, such as the charts of the North Atlantic Ocean, 

 published by the Conseil Permanent,' or the Atlas for 

 the North Sea," which Zorell ^ has treated critically. 

 This procedure in drawing the isotherms was based on 

 the fact that all descriptions which use synoptic material 

 from partial areas, such as the temperature charts of 

 Church * for the Gulf Stream area, or the surface charts 

 in the Bulletins Hydrographiques and the Bulletins of 

 the Ice Patrol Service, show a much more active and 

 complicated configuration than the average charts. 

 Even though the arrangement of bodies of water of 

 varying temperature, as it is shown on these monthly 

 charts, does not everywhere completely correspond to 

 reality, nevertheless, the character of the temperature 



I O. und V. O. Pettersson, Cartes synoptiques de I'eau de surface de rOcean Atlanliqite 

 du Nord. Cons. Perm. 1931. 



^ Atlas de ta temperature el salinite moj/enne de t'eau surface de la Mer du Nord et 

 de la Manche. Cons. Perm. Intern. Kopenhagen, 1933. 



3 F. Zorell. Ein neuer Alias fur Temperatur und Salzgehalt im Oberfldchenwasser 

 der Nordsee. Ann. d. Hydrogr. Bd. LXII. Berlin, 1934. 



* Phil. E. Church, Surface Temperatures of the Gulf Stream and its Bordering 

 Waters. The GooRr. Review. Vol. XXII, No. 2, New York, 1932. S. a. Inter- 

 nationa! Ice Observation and Ice Patrol Service in the North Atlantic Ocean. 

 U. S. Treasury Department. Coast Guard. Bulletin No. 26. Washington, 1938. 



distribution is brought out better by them than by 

 strongly generalized illustrations. 



The charts for the annual average and for the 

 anomalies, however, were treated differently. Since 

 these values are to be regarded as "abstract" in the 

 sense in which Zorell uses the word, the drawing of the 

 isotherms for them could take a more equalized (ad- 

 justed) form than for those of the monthly charts. In 

 addition, both the aimual average and the anomalies 

 were calculated not from the one-degree field averages, 

 but from averages for fields of two degrees of longitude 

 and latitude. The reasons for this procedure were as 

 follows: Even if observations had been on hand for 

 every one-degree field through all 12 months and the 

 annual averages from the various areas consecjucntly of 

 equal value, it would still have been impossible, con- 

 sidering the available data, to carry out calculations for 

 annual averages and the anomalies for each one-degree 

 field. Therefore, the temperature values for the inter- 

 secting points of every second degree of longitude and 

 latitude were interpolated with the assistance of the 

 data in the Dutch atlases. These temperature values are 

 valid for a field of an area of 4 one-degree squares. By 

 analogy to the ten-degree field, which consists of 10 

 times 10 one-degree squares, this unit was called a two- 

 degree field. These two-degree field values, which 

 appear in Appendi.>c I, form the foundation for further 

 treatment of the temperature. They are more equalized 

 than the one-degree field values and can be regarded as 

 independent of the number of observations. From them 

 the annual average was obtained by forming arith- 

 metical means. Attention was given to the number 

 of fields provided with values. 



These values were likewise used for the purpose of 

 forming the averages for the zones of 2 degrees of 

 latitude for the individual months as well as for the 

 year. (See Appendix II.) These zonal averages * may 

 be subject to slight inaccuracies, which are fully dis- 

 cussed in "Das Beobachtungsmaterial und seine 

 Aufbereitung," part I (untranslated) of Temperatur, 

 Salzgehalt und Dichte an der Oherflache des Atlantischen 

 Ozeans. In the high north and south latitudes, where 

 observations are lacking, the averages are more or less 

 uncertain, since in places where no material whatever 

 was available, extrapolation had to be relinquished. 

 The monthly and annual surface temperature anoma- 



s The expression "zonal average" is, strictly speaking, not adequate, since on the 

 coasts, for e.iample, where the two-degree field is partly obscured by land, the tem- 

 perature value was regarded as valid for the whole field. Considering the small size 

 of the unit ol area in proportion to the whole zone, this neglect seems justified. 



