The gradual increase in the greenhouse effect which may be observed over the last century, and after the Second World War in particular, has led to a growth in interest in year-to-year climatic variability. As the calculations show, the climatic conditions of the 1980s were globally the warmest over the period of instrumental observations (Jones 1994; Parker et al. 1994), and this prompted many scientists to investigate the problem of global warming. Hundreds of articles addressing this issue have been published in numerous academic journals in recent years.
It has long been known that polar areas are of key importance in the shaping of the global climatic system. There was once even a popular saying that the Arctic is the "kitchen" of global weather conditions. Contemporary climatic models confirm the above assertions, showing that as the concentration of C02 doubles, the greatest climatic warming should occur in polar areas
(IPCC 1990, 1992). This warming can be predicted to be particularly high in the winter, exceeding 12°C in some regions of the Arctic and Antarctic. This means that the increase in T in these areas will be several times greater than the average for the whole globe. The rate of the formation of cold deep-sea water in the high latitude oceans also has a very important climatic role, a fact which has been demonstrated in numerous articles (Gordon 1986, Manabe & Stouffer 1988; Stocker et al. 1990; Stocker & Mysak 1992). Walsh and Chapman (1990) maintain that in all likelihood it is also dependant on the variability of the climates of the Arctic and Antarctic. However, as has already been pointed out, the number of publications which study the short-term variability of the climate in the Arctic is surprisingly modest. Far more often one finds information about the climate of the Arctic in analyses discussing the climate over larger areas, for instance, hemispheres or the whole globe. They are, however, strongly generalised and hardly ever make possible a detailed examination of the spatial variability of the climatic changes in this region (cf. for instance, Yamamoto 1980; Jones & Kelly 1983; Jones et al. 1986, 1988; Parker & Folland 1988; Karoly 1989; Alekseev & Svyashchennikov 1991; Kukla et al. 1992; Parker et al. 1994). A certain, slight increase in research efforts has been noted as late as in the 1990s (Barry et al. 1993; Chapman & Walsh 1993; Kahl et al. 1993a, b; Przybylak & Usowicz 1993, 1994; Zhadin & Sutyrina 1993; Dmitriev 1994). Thus, the main objective of this subsection is to attempt to provide detailed information about the contemporary variability of T in the Arctic.
188.8.131.52 The variability of T in the Arctic since the beginning of the instrumental observation period to the middle of the 20th century
Up the 1920s the only permanently operating meteorological stations were those in Greenland (which started observations in the second half of the 19th century) and the station Green Harbour, located on the western shore of Spitsbergen (which started observations in 1911). A few more stations were set up in the 1920s (Jan Mayen, Bjornoya, Malye Karmakuly, Ostrov Dikson, and Barrow). Most Russian stations were set up in the 1930s while the Canadian ones were established as late as in the late 1940s. Consequently, the discussion of the variability of T in the Arctic up to the middle of the 20th century is possible only for the areas where there were stations in operation. Certain incomplete data pertaining to the climate in the remaining area of the Arctic can be obtained only from the observations carried out by expeditions.
In the 1980s some works were published which presented thermal conditions in the "Arctic" since the middle of the 19th century in the form of monthly, seasonal and annual mean values of anomalies relative to various long-term periods (Kelly & Jones 1981a-d, Kelly & Jones 1982; Kelly et al. 1982; Jones 1985a) or in the form of monthly means (Alekseev & Svyashchennikov 1991). The word "Arctic" is used here in inverted commas because the thermal conditions attributed to it pertain, in fact, to the area 65-85°N, which - as Figure 1.1 shows - is significantly different from the real area of the Arctic. Moreover, as has already been mentioned, in the 19,h century the only operating stations were those in Greenland. As a result, mean T values of the "Arctic" were calculated predominantly on the basis of the data from the stations located in the Subarctic. According to Jones (1985a), the mean for 1851 was calculated using data pertaining to as little as 6% of the area between 65-85°N. In 1874, 10% coverage was exceeded and by the start of the 20th century this had gone up to 20%. It was not before 1951 that the number of stations increased sufficiently for the spatial data coverage (in the form of grid points encompassing 5° latitude and 10° longitude) to exceed 50%; (the fact conducive to this increase was the setting up of meteorological stations in the Canadian Arctic between 1945 and 1950). The maximum data coverage in the area discussed occurred at the turn of the 1950s. Then a gradual decrease in the number of stations occurred, with the consequence that since 1981 spatial data coverage again fell below 50%. Thus, the data from the 19,h century and the beginning of the century presented in the publications quoted above pertain, in fact, to the area limited to Greenland and the Subarctic part of northern Eurasia. Because of the above considerations, the employment of the term "Arctic" to refer to the thermal conditions in this area is, in the author's opinion, inappropriate and can be misleading, especially for readers less familiar with the topic, who may not be able to gain access to data connected with the production of the above thermal series. This is a real possibility as the differences pointed out are not explained in every article, although there are references given to the literature where the source data base is described.
Summing up, it must be said that the temperature series of the "Arctic" calculated by the authors mentioned above are, due to the changing data coverage of the area, not fully comparable. Comparability is reduced with the increase in the difference in the data coverage of a given region. It should be added that the changes in the spatial data coverage are of less importance for the approximate assessment of a long-term trend, where the time-scale is from several decades to hundreds of years (Jones - personal communication). Bearing in mind all these reservations, the changes of T in the latitude band 65-85°N in the 1851-1950 period are shown below.
The coldest decade in the period under discussion was 1881-1890. Subsequently, according to the data published by Kelly et al. (1982), a brief warming occurred, which lasted till the beginning of the century. The increase in T at that time was 0.65°C. In the second half of the first decade
Was this article helpful?