Comparing the values of trends over the periods from 1961 to 1990 and from 1971 to 1990 (Table 6.6), we may observe a pronounced decrease in the values over the latter period. As a result of this, a significant decrease in the amount of statistically significant trends may also be noted. In the case of 40-and 30-year periods, statistically significant trends for annual P occurred at about 10 stations, while in the last 20 years of the period analysed only one station revealed such a tendency. A slight change in the spatial distribution of regression coefficients was also observed in the Arctic. A decrease in P occurred at all ATLSRs stations, despite the increase in P that dominated here over the earlier period. As regards western and central parts of SIBR, the tendency changed from decreasing to increasing. The area encompassed by positive trends in P increased slightly between 1971 and 1990. Nevertheless, the extent of increasing and decreasing tendencies in P covers a similar area of the Arctic. Fluctuations of P

In order to examine the fluctuations of P in the Arctic, two methods recommended by Gregory (1976) have been used, i.e. that of running means (Figures 6.9 and 6.10) and that of cumulated deviations from the mean. However, the latter was presented in a modified manner, thus creating the so-called difference-integral curves (Figure 6.14) whose ordinates were calculated according to the formula cited by Kozuchowski (1985):

Pt - annual total of precipitation in a given year, P - long-term mean of annual total precipitation.

The application of running means leads to a significant smoothing of the course of "raw" values, as a result of which it is easier to analyse the variability of P. Running means were calculated for the anomalies in P determined in relation to the means for the period 1951-1990. Due to the long distances between the Arctic stations and to different origins of P, there is a significant diversity in the fluctuations of P. These fluctuations are characterised by irregular swings of different amplitudes and an asynchronic course. Analysing five-year anomalies of P at the stations of the longest observation series, one can notice a significant similarity in the fluctuation of P at stations in Greenland, despite their location on opposite coasts. Such a situation is conditioned by the advectional origin of most of the P occurring here. The area of southern Greenland is influenced by the activity of the lows forming over and around Iceland. At the remaining stations, where P originates locally or regionally to a greater degree, its fluctuations run in a different manner. It very often happens that periods of a surplus in P at one station contrast with a scarcity of P at the other. Such a situation occurs at Barrow and Coppermine which, in Arctic terms, are located relatively close to each other. On the other hand, it is possible to find places more remote from one another than those mentioned above that are characterised by a good synchronisation of P fluctuations (cf. Barrow and GMO E.K. Fedorova stations). Taking into consideration the nature of P, an element which is changeable in time and space, and the significant distances between the stations, we may thus understand the predominant lack of any synchronicity of short-term fluctuations in five-year anomalies. As may be observed from Figure 3 in Przybylak and Usowicz (1994), such a situation may even arise in the area of ATLR. A greater compatibility may be observed for long-term (lasting for several decades) fluctuations. Examining the main maximums and minimums observable in the course of running means (Figure 6.9), it is possible to determine approximate precipitation cycles. They are most pronounced in the areas with the greatest advectional precipitation, i.e. in ATLR and BAFR. It is highly probable that the cycle at the Greenland stations lasts for 60-70 years, while at Jan Mayen it lasts for 15-20 years. Long-term (about 40-year) cycles may also be observed at Barrow and Coppermine stations. The problem with the periodicity of P will be discussed in greater detail in the next sub-chapter.

Looking at Figures 6.10a-i, we will describe the fluctuations of seasonal and annual anomalies of P in particular climatic regions and sub-regions; however, our analysis will be confined to the period 1951-1990, for which the data from all the stations are available. Ten-year running means in the study area were marked with thick unbroken lines. Even though their courses are even smoother than those of the five-year running means discussed above, a significant diversity of the fluctuations of P in various areas of the Arctic is still observable.

In ATLSRw (Danmarkshavn), the maximum P begins to be clearly pronounced in the 1950s, while the minimum P lasts for the greater part of the 1960s and for the beginning of the 1970s. Fluctuations in annual P are shaped predominantly by the changes in autumn and especially in winter P totals. The highest summer and spring P was observed at the end of the 1970s and the beginning of the 1980s, while the lowest P values occurred at the same time as annual P. In the last 20 years of the study period, fluctuations of summer P increased considerably, while those of autumn and winter P decreased (Figure 6.10a).

In ATLSRs (Jan Mayen), two maximums of P occurred. The former could be noted in the 1950s, and the latter in the 1970s. The course of annual P fluctuation depends predominantly on the winter and spring values of P (Figure 6.10b). The former maximum occurs in all seasons of the year except summer, whereas the latter is most pronounced during winter and summer, and is weakest in autumn. This is the reason why the 40-year trend is negative only in this season. A decrease in the fluctuation of P, especially in winter, may be observed between 1971 and 1990.

In ATLSRn (Ostrov Vize) where the influence of cyclonic circulation is sporadic, courses of the fluctuation of annual totals are most similar to the summer ones, and least similar to autumn P. It is worth noticing that in this sub-region, the maximum of P in all seasons occurred in the 1960s, i.e. during the period when the minimum P was observed in the region analysed above. Since the 1970s, we may notice in this region a considerable decrease in annual P (by over 100 mm) (Figure 6.10c). This led to a significant increase in the variability of P in that period (Figure 6.7). It must also be noticed that 10-year anomalies of P in this sub-region are the highest in summer, while in the previous two sub-regions they are the highest in winter and in autumn (Figures 6.10a and b).

Similar to ATLSRn, the highest fluctuation of P in ATLSRe (Mys Kamenny) occurred in the 1960s. Fluctuation of the annual P totals depends to a large extent on summer P (Figure 6.10d). Despite the occurrence of irregular and asynchronic fluctuations, all seasons of the year are characterised by a clear decreasing tendency, especially between 1961 and 1990. It is also in this period, and especially in the last two decades studied, that there occurred a pronounced increase in the stability of winter P and a slight decrease in summer P (Figure 6.7).

In SIBR (Ostrov Kotelny) 10-year running means of P display a strong decreasing tendency in all seasons of the year, but most significantly in winter and autumn (Fig 6.10e). Thus, fluctuations of annual P are most closely related to its courses in these two seasons of the year. Summer P is decidedly characterised by the greatest amplitude, frequency, and regularity of oscillations. However, because of the aforementioned significant regularity of oscillations in relation to the norm, the 10-year mean anomalies of summer P are smaller than those of winter. A pronounced change in the value of fluctuations in the four decades analysed does not occur in this region.

Fluctuations of annual and summer P are most similar to each other in PACR (Barrow). It should be noticed, however, that the course of the 10-year winter deviation of P differs slightly from the curves (Figure 6.10f). Greater differences occur in transitional seasons; in spite of this, long-term changes proceed in a similar way in all seasons. A pronounced decrease in the variability of P can be observed in this region practically for all seasons between 1971 and 1990.

In CANSRs (Coral Harbour A), the fluctuations of annual and autumn P are most similar - a similarity which rarely occurs in the Arctic (Figure 6.10g). The minimum P occurred here in the 1950s, which is mostly due to its low values in autumn. The maximum that occurred at the end of the 1960s and at the beginning of the 1970s is not significantly developed and is mainly due to high values of P in autumn rather than in summer. The increase in their stability in the last two decades studied is most pronounced for winter and annual P.

The most compatible courses of the anomalies of annual and summer P occur in CANSRn (Resolute A). The highest annual P occurred from 1981 to 1990, mostly due to its high values in autumn (similar to CANSRs) (Figure 6.10h), As regards the minimum observed in the mid-1970s, it was caused by low summer P. It is in this sub-region that there occurs a significant incompatibility of fluctuations in particular seasons. In the 1980s one may observe a decrease in summer and annual variability as well as an increase in winter P. The increase occurs only when compared with the decade 19711980 (Figure 6.7).

In BAFR (Clyde A) the course of 10-year anomalies of annual P generally resembles the courses of summer and spring anomalies; however, they also remain similar in the remaining seasons of the year (Figure 6.10i). The highest annual and seasonal P occurred between 1981 and 1990, and the lowest P may be noted at the end of the 1950s and the beginning of the 1970s. Its significant decrease and sudden increase were observed before the first and after the second minimums, respectively.

The profile of the fluctuation of P can be extended and complemented using the aforementioned difference-integral curves. Drawing them, one may determine exactly the periods of deviations from the norm of P which were positive (increments of the variable) and negative (its falls) and the moments when they changed. We are not able to determine this solely by using running means. Difference-integral curves of winter, summer, and annual P demonstrate their cumulated relative deviations from their respective means for the period 1951-1990 (Figure 6.14). An analysis of these curves allows us to assume that the changes in seasonal and annual P proceed in a relatively compatible way in certain regions of the Arctic. This is clearly observable at Ostrov Vize, Barrow, and Clyde A stations. In those regions of the Arctic where advectional P prevails, a greater compatibility of the courses of cumulated relative anomalies of winter and annual P is confirmed (e.g. at Jan Mayen and Danmarkshavn). This compatibility was described in the analysis of running means. The curves of summer and annual anomalies of P (e.g. at Mys Kamenny, Coral Harbour A, and Resolute A) are very similar in the areas of continental climate characterised by a dominance of anticyclonic circulation. Comparing difference-integral curves drawn for stations in the Arctic (Figure 6.14) and in

Poland (Figure 45 in Kozuchowski 1985), it may be observed that, in the period 1951-1990, the changes in annual P in the Arctic rarely demonstrate the oscillations lasting several years which are so characteristic of P in Poland. In contrast, long-term fluctuations (usually longer than 40 years) are observable in the Arctic. However, it is difficult to evaluate the periods of these fluctuations on the basis of an analysis of the curves (e.g. because of series which are too short).

Resolute A

1960 1970 19B0 1990

Figure 6.14. Difference-integral curves for the seasonal and annual totals of P for selected Actic stations over the period 1951-1990.

1960 1970 19B0 1990

Figure 6.14. Difference-integral curves for the seasonal and annual totals of P for selected Actic stations over the period 1951-1990.

Changes in P proceed in similar ways in ATLSRw and ATLSRs. Winter and annual P exceeded the norm in the 1950s, and dropped below it in the subsequent decade. Since about 1970, they have been oscillating at the level of the long-term mean, except for winter P at Danmarkshavn which revealed a deficit at the end of the 1970s and the beginning of the 1980s. Summer P

proceeds slightly differently here. In the first 20 years negative anomalies occurred here, while later they were followed by positive ones, to remain around the zero level in recent years.

In ATLSRn, the period of surpluses in P in all seasons analysed lasted until about the mid-1970s. Later, there occurred a significant deficit which reached its maximum in the first half of the 1980s.

Summer and annual P in ATLSRe change in a similar way to P in ATLSRn except that here the period of positive deviations from the norm ends earlier, especially when annual totals are taken into consideration. Three negative and two positive periods of their cumulated anomalies may be distinguished from the curve presenting winter P (Figure 6.14).

SIBR (Ostrov Kotelny) and PACR (Barrow) are characterised by a similar course of the difference-integral curves of P. Periods when P exceeded the norm lasted until the second half of the 1960s, and were followed, with rare exceptions, by negative anomalies. More complicated changes were characteristic of summer P in SIBR, which displayed irregular oscillations lasting several years.

In the Canadian Arctic, P oscillated around the long-term mean over the period 1951-1990. It is only on the curve ofthe cumulated anomalies ofwin-ter P that periods when they were positive (mainly the first 20 years) and negative (the last two decades) may be distinguished.

Changes in P proceed differently in BAFR (Clyde A) than in ATLSRn (Ostrov Vize) (Figure 6.14). The period of deficit, observed in all seasons of the year, lasted here until the mid-1970s. In the last 15 years of the study period (and especially up to about 1982) significant positive anomalies of P occurred.

Was this article helpful?

0 0

Post a comment