I. N. Zavyalova

A Climatology of Arctic Clouds

Cloudiness is one of the most important factors affecting the climate, determining to a significant extent the details of the moisture, sensible heat, shortwave and longwave radiation regimes at high latitudes. For example, model calculations have shown that an increase in cloudiness in the lower and mid levels produces a cooling effect, and the reverse gives warming. [8]. The influence of low-level cloudiness on the formation of the near surface thermal regime is sufficiently great that an increase of 10 percent (1/10th) may be sufficient to compensate for the effect of warming caused by a doubling in the atmospheric concentration of CO2 [24].

Climatic research on cloudiness of the north polar regions has been carried out by the Arctic and Antarctic Research Institute (AARI) in St. Petersburg. Recent advances in the amount of information that can be obtained from satellites have resulted in an underestimate of the importance of surface-based observations. However in addition to the exciting advantage objective observations, it is clear that the analysis of clouds involves considerable complexity. Currently, there are three basic sources of information about arctic cloudiness: standard observations from surface stations (including balloon and other aerological observations), satellite data, and observations by a Flying Meteorological Observatory (FMO).

Surface based observations are characterized by a number of limitations, producing overestimates of the cloud amount. First, during observations from the earth's surface the field of view of the observer contains not only low clouds (whose area and amount is of particular importance) but also the lateral surfaces of distant clouds especially those which are located low in the sky near the horizon. Second, the cloudiness in the polar regions may be overestimated at the expense of ground fog or cloud-like structures produced by atmospheric temperature inversions [4]. It is particularly difficult to determine cloudiness during polar night. It is also essential to take into account the subjectivity of the observers. This is especially apparent in the Arctic, due to the frequent changes in observational personnel, which is typical in the Arctic.

Most of the factors indicated above are not present in information obtained from satellites. However, accumulated tests using satellite data allowed a side-by side evaluation of the advantages associated with the ability to make objective evaluations along with certain disadvantages associated with this observational system.

Principal among these is that it is essentially impossible to obtain information about the multilayer structure of the clouds and the height of the lower boundaries of the layers. Besides this, it is difficult to obtain a time series of the characteristics of the cloudiness obtained using various instrument arrays whose sensitivities can change with time and for which the algorithms used to determine the characteristics of the clouds varies depending on the instruments used [7]. Satellites do not reliably measure thin cirrus clouds, clouds over snow and ice, and translucent fogs [7, 17, 27]. They are also unreliable at night [7, 27].

All these considerations result in a decrease in the amount of cloudiness in the polar regions reported using satellite data compared to the values from surface-based observations. The greatest divergence between them is observed northward of 55 degrees north latitude. The difference in the results grows steadily with latitude and at 85 degrees north reaches 2.5 to 3.5 tenths [21, 22]. The deviation is particularly large for partial cloud cover. The significant differences in the data sets obtained using different methods is also noted by Mokhov and Schlesinger [26]. Thus, the information currently available from satellites does not have the necessary precision. Consequently, the use of satellite data in climatological research on cloudiness at high latitudes is not yet advisable, and as before the results of surface-based observations are better suited for this sort of application.

An important addition to this is information from FMOs. One of these observatories operated every year in the Russian Arctic from 1948 to 1964, during the period when the conditions for flying are most favorable, from June through September. In some cases flights were also carried out in other months including the winter. Usually these flights were carried out from several starting points in both the western and eastern parts of the Russian arctic to the north and then back again. Included in the observations were valuable results on the characteristics of clouds and fog in these regions.

Many cloud parameters are determined more precisely from aircraft observations than from surface-based observations. This relates primarily to the form of the clouds, their vertical extent (especially for mid-level and high clouds), their boundaries, horizontal extent etc. FMO data presently in hand give a rather general but one-time description of the spatial and microphysical structure of the clouds and fog.

In the present paper, results from surface-based observations are used for interpretation of their spatial and temporal variations to provide climatological characterization of arctic cloud conditions, and information about the cloud forms and their parameters are presented from FMO observations. In this context, it is necessary to remember that the data sets obtained from FMOs, despite their precision, should at present be considered as approximate because of the comparatively small number of observations.

For the characterization of the spatial and temporal distributions of cloudiness in the Arctic, we have used observational data from the polar stations of Russia and other countries covering a 40 year period, NP drifting station data, and also observations from ship transects The drifting station observations were reduced using the methods of Prik [19], which combines data from groups of stations. The selection of groups and their average location varies for different months. Ship-based observations are interpolated onto a rectangular measure 5 degrees latitude and 10 degrees longitude in the Greenland, Norwegian, and Barents Seas, and 1 degree x 5 degrees in the remaining seas of the Russian Arctic.

The following indicators are used as the basic climatological indicators of the amount of clouds: the mean value, the frequency of various degrees of cloudiness, and the mean number of clear and overcast days. Data have been obtained from the sets of observations and combined over the entire time interval. To specify the frequency of various degrees of cloudiness, the amounts are combined into three groups: clear skies (0 to 2 tenths), partly cloudy (3 to 7 tenths) and overcast (8 to 10 tenths). The frequency of clear and overcast conditions is determined as a percentage of the multiyear total number of observations for each month. Partly cloudy conditions make up the remainder.

The generally accepted division of the cloud conditions for clear (0 to 2 tenths) and overcast (8 to 10 tenths) does not reflect the structure of arctic clouds with sufficient precision. An analysis of the frequency of each tenth of total cloudiness, carried out using the observations from the net of Russian polar and drifting stations shows that within the gradations described above, extreme conditions tend to prevail (completely clear skies (0/10) and complete overcast (10/10)). Further, in the summer months and in the fall the cloud coverage is nearly always 10/10ths. For example, in July complete overcast occurred 71 to 92 percent of the time corresponding to a frequency of 72 to 95 percent for the entire 8/10 to 10/10ths category. In the winter and spring months there was a characteristically sharp prevalence of either 0/10 or 10/10ths cloudiness which occurred 80 to 90 percent of the time.

As a result, the mean value of the cloudiness is actually quite different from its modal value and does not always reflect the level of the most frequently observed cover on the horizon. This is especially apparent in the winter months, when the frequency of cloudiness closest to the mean value is in fact less than 5 to 10 percent [9]. Therefore, to provide an accurate objective characterization of cloudiness in the Arctic, a single parameter is not sufficient and a more complex characterization is needed.

A day is considered to be clear when the total of the scores for 4 (8) intervals throughout that day does not exceed 7 (14) tenths, and a day is overcast when the total of the scores is greater than 33 (66) tenths. Since 1969, additional conditions have been added, in which a day is considered clear unless the cloudiness in one of the intervals exceeds 5 tenths [18]. This leads to a slight reduction (1 to 2 days) in the mean number of clear days during the winter months. All the characteristics indicated above are considered both for total cloudiness (including clouds at all levels) and for low level clouds (only clouds in the lower layers of the cloud deck).

The morphology and solidity of the cloud cover over the Arctic is determined primarily by the atmospheric circulation and the influence of the underlying surface. The homogeneity of the latter contributes to a comparatively small spatial variability in the cloud cover, but strong seasonal changes in the regime of the centers of activity in the atmosphere result in a considerable change in the cloud-cover patterns from winter to summer.

The annual changes in cloudiness are characterized for the most part by the variations from May through October, while conditions are least variable from November through April. A consequence of this is that during the changeover from the winter regime to the summer, particularly from April to May, a particularly strong increase in cloudiness is observed. In the majority of the arctic regions the most clear-cut maximum is expressed in August through October, and the minimum is in January through March (Figure 1). The amplitude of the yearly variation in total cloudiness over the southern part of the Barents Sea is about 1 tenth, over the peripheral seas it is 2 to 3 tenths, and over the maritime central part of the Arctic Ocean it reaches 5 tenths. However, in connection with the diversity of the circulation processes in various parts of the Arctic, the annual variation in mean cloudiness is somewhat more complicated. For example, over the Barents, Kara, and Laptev Seas, a secondary increase in cloudiness is observed in July that is connected with a weakening in the cyclonic activity in that month. Over the Chukchi Sea this secondary decrease in the annual cycle of cloudiness is displaced to June, and it is also connected with the decrease at that time in the number of cyclones passing through the area (3 to 4 per month), whereas in May and July up to 5 to 6 cyclones are observed per month.

A characteristic feature of the spatial and temporal structure of the fields of mean cloudiness are such that from May through October the largest values of total mean cloudiness (in excess of 8 tenths) are recorded over the marine regions of the Arctic Ocean increasing to 9 to 9.5 tenths over the region around the north pole. Only southward over the continent does the amount of cloudiness decrease, to about 7 tenths (Figure 2a). The arctic value and the stability of the cloudiness is a result of the intensity of cyclonic atmospheric activity as well as the influence of the cold ocean, covered with melting ice, above which the air masses are rapidly transformed, resulting in the formation of local inversions, fog, and low level clouds.

In winter, the amount of cloudiness over most of the Arctic Ocean decreases sharply. The value in January is 4 to 5 tenths (Figure 2 b). The smallest mean cloudiness is measured at that time in the northern part of the East Siberian Sea, which contributes to a tendency to anticyclonic circulation in that region, and also its separation from sources of water vapor. The region of maximum cloudiness (approximately 9 tenths) is displaced from the north polar area to the southern part of the Barents Sea [1, 9].

The spatial and temporal distribution of mean and total cloudiness closely correspond to the distribution of the frequency of occurrence for overcast and clear sky conditions. In summer, the maximum occurrence of overcast skies (8 to 10 tenths) with the total cloudiness exceeding 90 percent is characteristic for the north polar region and the northern part of the Barents Sea. To the south the probability of overcast conditions decreases to 75 percent along the Asian coastal region, to 65 to 70 percent over the southern part of the Barents Sea, and to 60 to 65 percent along the coasts of Canada and northern Greenland (Figure 3a).

From May to September, the frequency of overcast skies changes less and reaches a maximum in September or October. In November, cloud frequency begins to increase rapidly, and in January through March reaches its maximum value. During these months, the minimum frequency of overcast skies of 35 to 40 percent is recorded over the Laptev Sea, over areas of the northern East-Siberian Sea, and over the islands in the Canadian Archipelago (Figure 3b). The most frequently overcast weather in winter (more than 70 percent) is observed over the southern part of the Barents Sea. It is caused by the active cyclonic activity along the troughs of the Icelandic minimum. In connection with this, the intensity of Atlantic cyclones over the eastern archipelago of Novaya Zemlya quickly weakens, and the frequency of overcast skies over the Kara Sea decreases to 65 percent in the vicinity of the strait of Karskiye Vorota, and to 40 to 45 percent along the coast of Severnaya Zemlya. From March to May, a significant increase occurs everywhere over the entire region both in total cloudiness and in the frequency of overcast skies.

Clear skies (0 to 2 tenths of total cloud cover) over the central Arctic Basin in the summer is recorded infrequently, in only 10 percent of all cases, but in the area around the north pole it is even less frequent (less than 5 percent). In the coastal zones of the peripheral seas, the frequency of clear weather usually does not exceed 15 percent, and only in the south-eastern part of the Kara Sea does it exceed 20 percent. From summer to winter the frequency of clear skies over the central part of the Arctic Ocean grows by a factor of 5 to 10 times and increases significantly in excess of 50 percent. For the winter months, sharp contrasts are characteristic in the frequency of clear skies over the Barents and Greenland Seas, where it quickly decreases from 35 to 40 percent at the northern periphery to 8 to 10 percent in the central parts of these seas. In the Chukchi Sea the probability of clear weather in winter is approximately 35 percent [1,9].

The relative annual variation of different types of cloud conditions normalized to 100 percent, described above, is illustrated in Figure 4. This Figure presents the data from three polar stations located in different parts of the Russian Arctic.

The percentage frequency of clear and overcast skies describes the total duration of these conditions. The number of clear and cloudy days shows the distribution of clear and overcast weather in the course of a month. Overall the number of overcast days in the Arctic is significantly greater than the number of clear days. Over the Laptev and East Siberian Seas and the central parts of the Arctic Ocean, for example, overcast days occur 3 to 4 times more often, over the Chukchi Sea 4 to 6 times more often, and over several parts of the Barents Sea cloudy conditions are from 10 to 15 times more frequent. However, because of the strong annual variation, the ratio of the number of clear to overcast days varies seasonally: in summer, the difference increases, in fact in winter the number of clear days exceeds the number of cloudy days over the regions with prevailing anticyclonic weather regimes.

In the course of the year, the greatest number of cloudy days is observed in summer. In the region near the pole, this number reaches 26 to 28 days per month on average. In individual years, cloudy weather has been recorded during every month. Toward the continent, the number of overcast days quickly decreases and in the coastal zone it reaches 18 to 20 days per month.

In general, everywhere over the marine regions of the arctic seas, typical conditions are such that more than half of all days are overcast from May to October. The greatest number of overcast days during this period exceeds 25 per month, and sometimes reaches 30 per month. Beginning in November or December, the number of overcast days quickly decreases and in January does not exceed on average 6 to 8 over the greater part of the Arctic Basin, and 8 to 10 over the Chukchi Sea. Only over the central part of the Barents Sea is the number of overcast days nearly indistinguishable from the number in summer, consisting of 18 to 20 days per month. The strong annual minimum is absent. From January to March the number of overcast days is approximately constant, but in April a strong increase begins [1, 9].

The annual variation in the number of clear days has opposing characteristics for which the relative contrasts between winter and summer are substantially larger than for the number of overcast days (although the absolute values are less). In winter over most of the central part of the Arctic Ocean, the mean number of clear days reaches 8 to 10, but in certain years 15 to 20 per month have been recorded. Over the Chukchi Sea there are 5 to 7 clear days, and over the Barents Sea as a rule no more than 2 clear days are recorded per month. During the summer months, over all maritime parts of the Arctic Ocean the number of clear days does not exceed 2 per month, but over the central part 2 clear days per month have been measured only 2 to 5 times per decade.

Low level clouds show the same basic characteristics as the total cloudiness. Since corresponding results are usually not given in the foreign literature, less information is available and data have been obtained only from the Russian polar stations. The regular behavior patterns observed for the annual variations of total cloudiness are essentially preserved in the data for the low level clouds (see Figures 1 and 4). However is should be noted that while the average values are less than for total cloudiness, the amplitude of the annual variation for low level clouds is larger, consisting of 3 to 7 tenths compared with 2 to 5 tenths for total cloudiness (See Figure 1).

The frequency of overcast skies with low clouds is significantly less than for total cloudiness, but the frequency of clear skies is greater. In the southern regions of the Russian Arctic from June through October this difference is approximately 20 percent, but in winter when low clouds are present less often, it increases to between 30 and 35 percent. Over the maritime part of the Arctic Basin in summer and autumn the frequency of low level overcast skies in 15 percent less than for total cloudiness, but the frequency of clear skies is approximately 15 percent greater. In winter, these differences increase by more than a factor of two. Low level cloudiness is usually observed in the form of extensive cloud banks, covering more than half of the sky. Under these conditions there is a high frequency of low cloudiness, in an amount exceeding 5 tenths, which for different parts of the Arctic, ranges from 20 to 30 percent in the wintertime and 70 to 90 percent from June through September.

The question of stable clear and overcast weather both for total and for low level clouds is of great interest. For such an evaluation, coefficients of stability, K, are usually used, which is the ratio of the percentage of the number of clear or overcast days relative to the total number of days per month corresponding to the frequency of clear (0 to 2 tenths) and overcast (8 to 10 tenths) sky conditions. The greatest stability for clear weather occurs typically during the winter and spring months, when it is often recorded several days in a row. Note that clear weather with respect to low level clouds (for which K approaches unity) is more stable than for total cloudiness (where K is less than 0.5 for the most part). In summer, extended periods of clear weather are exceptionally rare. Overcast weather, on the other hand, is most stable from June through October, when the coefficients of stability approach unity.

The variation in cloudiness in a 24-hour period is complex. It is different for different types of clouds. As noted by Vize [23], in autumn and the beginning of winter thin clouds are characteristic for the Arctic. The stars often shine through them, and they can go unnoticed at night. Thus the amount of cloudiness recorded for the nighttime decreases. Only at high latitudes during polar night do the illumination conditions remain the same over 24-hour intervals, so the daily variation is not distorted. According to the results of Vize, the analysis of 24-hour variations in cloudiness during the drift of the G. Sedov, and also according to data from the northernmost polar station on Rudolf Island and from NP drifting stations located to the north of 85 degrees north latitude, the cloudiness during polar night for the nighttime and early morning hours is a little larger than for the daytime hours. The change in amplitude is very small.

Analogous diurnal variation occurs during the period of polar day, when the atmospheric illumination conditions over the course of a day are constant and the observed diurnal variation is not distorted. On this basis, Vize concluded that since the diurnal variation of cloudiness in winter and summer is the same, then it must be the same at other times of the year. The analysis of long-term observations from a series of polar stations showed that the most typical amplitude of the diurnal variation, was less than the degree of precision on the observations (1 to 1.5 tenths).

On the whole, cloudiness is a constant climatic characteristic whose mean annual oscillations are relatively small. According to data from the polar stations, the standard deviation (s) for mean monthly values of cloudiness in the winter is, as a rule, at the level of 1.0 to 1.5 tenths, and the maximum can occur in different months. The smallest variability in cloudiness is recorded at the end of summer or the beginning of autumn (s values are about 0.5 to 0.6 tenths). The coefficients of variation for the amount of cloudiness are small: 0.2 to 0.3 in the winter months and no more than 0.1 in the summer and fall. The standard deviation for the mean daily amount of total cloudiness exceeds the indicated mean monthly values. In winter s values range from 3.0 to 3.5 tenths and in summer from 1.6 to 2.7 tenths.

An analysis of the parameters of variation for total cloudiness is of particular interest from the point of view of investigation of short period climate fluctuations, specifically because the cloudiness is closely related to changes in the radiation flux and the hydrothermal regime in winter for both the surface and the atmosphere.

According to research by Birman and Pozdnyakova [3], based on diurnal observations from 1966 to the present time, a general tendency is observed towards an increase in the mean annual values of cloud fraction in both hemispheres. However, in the northern hemisphere, this growth is slightly less than in the south (the corresponding slopes from linear regression are 0.26 and 0.60). Against the general background of increased cloudiness three specific periods in which cloudiness increased were recorded: 1996, 1975, and from 1988 continuing into 1989.

The characteristics of the zonal distribution of cloudiness from 1984 to 1989 are such that in the polar latitudes of the northern hemisphere, positive and negative anomalies occurring from 1984 to 1986 occur alternately and produce no significant trend, but from 1987 the negative anomalies are prevalent. Such conclusions may arise as a result of the uncertainties in satellite observations [26].

There is a great deal of interest the future tendencies of cloud-cover variations in the Arctic regions. Expert evaluation about the expected climatic conditions in the Arctic in the period up to 2005 (taking into account anthropogenic influences) has shown that in the Russian Arctic the annual amount of total cloudiness will oscillate within existing limits with peaks and troughs in various years [16].

In the present work, the quantitative characteristics of the form of clouds are described by the following parameters: horizontal extent, elevation of the upper and lower boundaries, and thickness.

The horizontal frequency of clouds is obtained from aircraft observations. The heights of the upper and lower boundaries are determined from relatively sharp deterioration and improvement in visibility during flight. Cloud thickness is given by the difference between the heights of the upper and lower boundaries. All mean characteristics for these parameters are obtained by summing the values for each cloud type and dividing each total by the number of observations.

According to the data from surface-based observations, the clouds with the greatest extent in the Arctic are stratus clouds (St) characteristic of stable air masses. They exist throughout the entire year. Note that homogeneous air masses with constant thermal stratification can cover vast areas of the Arctic, and the horizontal extent of stratus clouds can be hundreds and even thousands of kilometers. Most often, however, their extent does not exceed 600 kilometers . In the western part of the Russian Arctic these conditions are observed 80 to 90 percent of the time, 40 percent of which are up to 200 kilometers in extent [13]

Since thermal conditions of this sort in the Arctic are mostly characteristic of the summer period, the frequency of Stratus clouds grows considerably in the transition from the cold to the warm part of the year. Independent of the time of year, the frequency of this cloud form is higher in regions far away from the basic centers of atmospheric activity. In the series of points, this includes up to 40 percent of all observations. In 70 percent of the cases, stratus cloud coverage reaches a value of 8 to 10 tenths.

The lower boundary of stratus clouds is located on average at a height of 170 meters in summer and 350 meters in winter, but the thickness of summer clouds (400 meters) is larger than for winter clouds (150 meters) [8, 20]. Variations in the cloud boundaries are quite substantial. However, analysis of the frequency of the heights of these boundaries at various levels in the atmosphere shows that in 88 percent of the cases the lower boundary of stratus clouds is located in the 600-meter layer adjacent to the surface. In 74 percent of the cases, the upper boundary of this cloud form does not exceed 1000 meters [12]. Stratus clouds in the warm half of the year are shrouded by the fog, that frequently occurs in the Arctic, but sometimes they descend right to the surface.

A characteristic feature of the location of the top and bottom boundaries of stratus clouds is that they become lower in the summer going from the coastal regions to high latitudes. For example, the mean height of the upper boundaries of stratus clouds in the zone 70 to 75 degrees North is 770 meters, but in the region around the North Pole the mean height is 544 meters. The lowering of the bottom boundary is connected with the lowering of the level where water vapor condenses in the regions under study. The upper boundary of the clouds is determined by the inversion height, which is a retarding layer that blocks the vertical transport of water vapor [6].

The second most extensive cloud type in the Arctic is stratocumulus (Sc), which as a rule also forms in stable air masses but as a result of wave motions in the inversion layers located in the lower troposphere. The frequency of this type, although also close to that of stratus clouds, is more often greater, especially in summer. The minimum frequency of Sc clouds is recorded in the cold part of the year (12 to 18 percent), but by May their frequency is already 28 to 34 percent, and at the beginning of autumn it increases to 47 percent. In 60 percent of the cases the coverage of this cloud type is measured at 8 to 10 tenths. The lower boundary of Sc is located on average at an altitude of 450 meters in summer and 650 meters in winter. The vertical extent of this cloud type is from 450 meters in winter to 600 meters in summer [8, 20].

The extreme values of the boundaries of stratocumulus clouds varies over a wide range, as is the case for stratus clouds. Most often the lower boundary of Sc clouds is located between 200 and 600 meters (35 percent), and the upper boundary lies between 600 and 1400 meters (50 percent)[13].

The horizontal extent of Sc clouds is on the whole the same as for stratus clouds. However, the extent of Sc up to 200 kilometers is more frequently observed. For example, in the western part of the Russian Arctic such an extent is observed in about 30 percent of the cases [12].

Other lower layer cloud types, as well as clouds of vertical development are unusual in the Arctic. The frequency of these cloud types does not exceed 10 percent. Note that for surface-based observations, the high frequency of St and Sc clouds masks the clouds with vertical development. Nevertheless, during FMO observations, cases of the appearance of cumulonimbus clouds (Cb) have been repeatedly recorded above the level of the stratocumulus. The summer and autumn are the most favorable for their formation. During this time there is sufficient contrast of the air temperature above the sea and the land (or the ice) to cause convection as colder air is transported from the continent or the ice pack to the relatively warm sea surface. According to observational data, this contrast is 4 to 5 degrees Celsius or larger [10]. The presence in the cloud formation layer of retarding (stable) layers, alternating with unstable stratification restricts the vertical development of such clouds. Cb clouds form over water surfaces, usually do not have anvils, and their vertical thickness is relatively small (500 to 2000 meters) [10]. Frontal cumulonimbus clouds are distinguished by their large thickness. There are recorded cases during aircraft missions when soundings up to 6000 meters did not reach the tops of these clouds.

The mean height of the lower boundary of nimbostratus clouds (Ns) is about 500 meters in the summer and autumn [20]. At that time of year, nimbostratus clouds have the greatest thickness compared with other cloud types, however their thickness is not particularly large in general terms. On average it is 1500 meters [20], and in individual cases it is no greater than 4000 meters [11]. The horizontal extent of Ns is approximately 250 kilometers [12].

In the Arctic, there is a particular variety of cloud type, relative to stratus clouds, that has been recorded, the so-called "shroud clouds." They appear during the cold part of the year as thin whitish layers which are formed beneath inversion layers as a result of condensation. The height of the lower boundary of shrouds varies from 30 to 200 meters. Shroud clouds very frequently appear in the Arctic during polar night, especially over drifting ice [4].

The correct calculation of mid-layer cloud amounts from surface observations is possible only in the absence of low level clouds and other weather phenomena that cause degradation in visibility in the lower atmospheric layers. In the summer, investigations of mid-level clouds are made difficult as a result of the exceptionally frequent occurrence of fog and low-level clouds, and in winter by the presence of polar night and snow storms [5, 7, 8]. However, results of observations from FMO's show evidence of the substantial occurrence of mid level clouds. For example, from 183 flights in the summer and autumn in the upper levels of the upper boundaries of low level clouds, 96 percent of the clouds measured were altocumulus (Ac) and altostratus (As). In the spring, the frequency of these cloud forms is about 35 percent [5].

For surface-based observations, which record only the visible parts of Ac and As clouds , altocumulus clouds are noted in 10 to 30 percent of the cases. Their frequency in the summer is less than in the cold part of the year [7]. This last circumstance is caused not only by the high degree of cover by lower level clouds but also by the close connection between the frequency of inversions in the middle troposphere and the presence of Ac clouds [8]. The maximum frequency of this sort of inversion (80 percent) occurs from May through October, that is, during the period of maximum frequency of low level clouds. For this reason, the effect of masking by low level clouds, even though it decreases the observed frequency of altocumulus, particularly in the warm part of the year, does not distort the picture of the actual annual distribution. Altocumulus clouds most often are observed at concentrations of 1 to 2 or 8 to 10 tenths. This cloud type is usually located in the layer from 2600 to 3050 meters with a mean thickness of less than 500 meters [5].

In the Arctic, altostratus clouds are observed more often than altocumulus clouds. Their intra-annual frequency is closely connected with cyclonic activity and thus in the warm part of the year their frequency is 2 to 3 times less than in the cold part. Altostratus clouds are most often recorded in combination with other cloud types, but in the remaining cases they are measured basically at concentrations of 8 to 10 tenths [7]. This cloud type is located in a layer from 2350 to 3150 meters with a mean thickness of 800 meters.

The extreme location of the boundaries of mid-level clouds varies over a wide range, in excess of 4000 meters. For example, the lower boundary of these clouds has been measured at heights of 1000 and 5500 meters. In 95 percent of the cases, the lower boundary of As and Ac clouds has been measured in the layer from 1.4 to 3.4 kilometers . The greatest frequency (69 percent) of the upper boundary of As is typically in the layer from 2.2 to 3.4 kilometers. For Ac clouds, in 72 percent of the cases the frequency of the upper boundary has been measured in the layer from 2.6 to 3.4 kilometers. Note that there are a relatively large number of cases where the upper boundaries of both cloud types is higher than 4.6 kilometers (for As 16 percent and for Ac 10 percent). Most often frontal systems are present when these very high upper boundaries occur. In the course of FMO flights there were several cases, where the upper boundary of the frontal clouds of mid level were not reached by soundings up to 6000 meters above the aircraft. According to available data, clouds of frontal systems in the coastal regions can reach altitudes of 7000 to 9000 meters, and their thickness exceeds 1000 meters. The thickness of midlevel internal clouds is small and seldom reaches 1000 meters, with an average thickness that equals 420 meters [5].

The horizontal extent of mid-level clouds is on average about 600 kilometers. The greatest extent has been recorded for opaque As opacus (opaque) at 1015 kilometers, and the minimum for Ac translucidus (translucent) at 210 kilometers [5]. For altostratus as well as for stratus, the largest variations in horizontal extent are typically seasonal. The extent grows on average 150 to 200 kilometers from the winter to the summer [13].

The shortcomings with surface-based observations of mid-level clouds outlined above are true to an even greater extent for observations of high-level clouds. These clouds are primarily connected with atmospheric fronts [2]. The most prevalent upper level cloud type is cirrus (Ci), making up 80 percent of the upper level cloud cover. Cirrostratus (Cs) is recorded only 16 percent of the time, and cirrocumulus (Cc) in 4 percent of the cases [7]. This frequency distribution for the upper level cloud types testifies to the fact that in the Arctic during the formation of clouds in the upper troposphere, a complex character for the vertical motions is most often observed. Ordered vertical motions in the upper troposphere are not dominant, as would be expected, if one takes into account the prevalence of decaying fronts over the Arctic.

Complete or nearly complete cloudiness of the upper layers is characteristic only for Cs. For example, from FMO observational data in August and September 1959 over the Laptev Sea and part of the central part of the Arctic Basin, a cloud coverage of 8 to 10 tenths consisted of 92 percent for cirrostratus, 50 percent for cirrocumulus, and 29 percent for cirrus. At the same time, the frequency of clouds up to 5 tenths was 2 percent for Cs, 38 percent for Cc, and 55 percent for Ci [2].

The lower boundaries of high level clouds in the warm half of the year have been observed at heights ranging from 5 to 9 kilometers with a highest frequency at 5 and 6 kilometers (67 percent). During the cold half of the year these boundaries lie in a layer from 4 to 7 kilometers for which the maximum frequency of 86 percent has been measured between 5 and 6 kilometers. The upper boundaries of high level clouds during the warm half of the year were located in a layer from 5 to 11 kilometers with a maximum between 7 and 9 kilometers (61 percent), and during the cold half from 6 to 10 kilometers with the maximum between 6 and 7 kilometers (43 percent) [2].

According to Zavarina and Dyuzheva [12] the mean heights for the lower boundaries of high level clouds appear to be the same over the arctic seas and in the central part of the Arctic Basin: for cirrus clouds the value is 5.6 kilometers and for cirrostratus it is 5.8 kilometers. The maximum heights were 6.5 and 8.5 kilometers respectively, and the minimum heights for the both types of cloud systems were 4.0 kilometers.

The vertical extent of high level clouds varies from a few hundreds of meters to 4 kilometers. The greatest number have a thickness of clouds from 1.1 to 3.0 kilometers, occurring 70 percent in the warm period and 50 percent in the cold period. Warm fronts are characterized by a more pronounced vertical extent for the clouds, while the values are lower for cold and occluded fronts. On average, the high-level internal clouds had the smallest thicknesses. With increasing latitude, the heights and thicknesses of high-level clouds decrease as they do for clouds at other levels [2].

The horizontal extent of high-level clouds varies over a wide range, and is defined primarily by the synoptic conditions during their formation. The cloud zones with the most significant horizontal extents are associated with warm fronts. In a number of cases, favorable conditions for the formation of high level cloudiness are also created during anticyclonic atmospheric activity. The mean width of zones of high clouds is 250 kilometers. The following zone widths include most of the cases and have the frequencies indicated: up to 100 kilometers about 32 percent, 100 to 200 kilometers about 24 percent, 200 to 300 kilometers about 24 percent, and greater than 800 kilometers about 3 percent.

The prevalence of one or the other cloud form over the Arctic changes depending on season as well as region. This dependence is related to the details of the atmospheric circulation and differences in the characteristics of the underlying surface [15].

Thus, the characteristic features of the spatial and temporal structure of arctic cloud cover are such that the greatest cloudiness in the Arctic is recorded during the period from May or June through October. During this time the characteristic features are that the values of mean cloud cover are greatest (total cloudiness over the region around the north pole is 9 to 9.5 tenths, decreasing to 7 tenths over the continent); the frequency of sustained cloudy overcast is highest (for total cloudiness over the region around the pole and the southern part of the Barents sea - greater than 90 percent); and the number of overcast days is at a maximum (more than 26 in the vicinity of the pole, 16 to 20 in the coastal regions). Consistent with these conditions, the frequency of clear skies and the number of clear days have their lowest values (over all maritime areas of the Arctic Basin they do not exceed 2 days per month).

Overcast weather in the summer and autumn is exceptionally stable (the coefficient of stability approaches unity). The greatest stability for clear weather is characteristic of the winter and spring months, when clear conditions are often recorded for several days (24-hour intervals) in a row.

There is a clear and well defined annual cycle of cloudiness that has a maximum in August through October and a minimum in January through March. During this cycle, the changes in cloudiness over the course of a given 24-hour period are not significant and in the majority of cases do not exceed 1 tenth. The greatest amplitude of the annual variation characteristically occurs in the central part of the Arctic Ocean (up to 5 tenths). In the Arctic, stratiform clouds occur with the greatest frequency, especially low level stratus. In the summer, these clouds are recorded in almost half of the cases, and, together with fog, they occur more than 70 percent of the time.

Compared to clouds of analogous type at temperate latitudes, arctic clouds characteristically have a relatively small lateral extent and occur at lower levels. Distinguishing features of arctic clouds also include the seasonal and latitudinal changes of the cloud parameters (lateral extent and thickness). Maximum values of these parameters occur in the summer. As a rule there is a decrease in the values with increasing latitude. The causes of this sort of phenomenon include changes in the vertical stratification and circulation conditions of the atmosphere.

The highest priority tasks for future investigations should be studies of the dynamics of cloud types on the basis of modern observations together with the parameters obtained from time series analysis and parameterization of cloudiness for the purpose of modeling of the radiation regime. Since the variability of clouds over the arctic region can have a significant influence on the gradient of planetary albedo and the exchange of energy at the earth's surface [25], monitoring of the polar cloud cover is important for climate-change studies. As a result, checking the stability of cloudy skies must be included in regional monitoring of arctic climate.

References

1. Atlas of the Oceans. The Arctic Ocean, -Moscow: Publisher MO USSR, VMF, 1980, 184 pp.

2. Baranov, A. M., On upper level clouds over the Arctic, Trudy AARI, Vol. 239, 111-120, 1962.

3. Birman, B. A. and Pozdnyakova, T. G., Climate monitoring of global cloudiness and radiation, Meteorology and Hydrology, No. 4, 104-110, 1991.

4. Bryazgin, N. N., Cloud shrouds in the Arctic, Problems of the Arctic and Antarctic, 9, 75-77, 1961.

5. Burova, L. P. and Voskresenskiy, A. I., Meteorological conditions of ice formation in altostratus and altocumulus clouds, Trudy AARI, Vol. 239, 95-103, 1962.

6. Voskresenskiy, A. I. and Chukanin, K. I., Meteorological conditions of ice formation in stratus and stratocumulus clouds, Trudy AARI, Vol. 228, 124-134, 1959.

7. Voskresenskiy, A. I. and Bryazgin, N. N., On monitoring of Arctic cloudiness, -V, in the book: Monitoring the Climate of the Arctic, Leningrad, Hydrometeorological Publishers, 96-105, 1988.

8. Voskresenskiy, A. I. and Karimova, G. U., The frequency and degree of fractional coverage of mid-level and high level clouds in the Arctic during the periods MGG and MGSS, Trudy AARI, Vol. 266, 66-89, 1959.

9. Gredenova, N. K. and Petrov, L. S., Cloudiness over the Arctic Ocean, Trudy AARI, Vol. 381, 68-77, 1983.

10. Dergach, A. L., Investigations of Cumulonimbus clouds without anvils, Trudy AARI, Vol. 228, 124-134, 1959.

11. Dergach, A. L., Results of observations of frontal clouds, Trudy AARI, Vol. 228, 100-112, 1959.

12. Zavarina, M. V. and Dyuzheva, O. G., Horizontal extent of clouds in the Arctic, Problems of the Arctic, Issue 6, 71-80, 1959.

13. Zavarina, M. V. and Romasheva, M. K., The thickness of clouds over the arctic seas and the central Arctic, Problems of the Arctic, Issue 2, 127-132, 1957.

14. Zyabrikov, V. A. and Shapovalov, V. F., On an example for evaluating the amount of clouds using satellite- and surface-based observations, Trudy GGO, Issue 440, 6-9, 1980.

15. Climate of the Polar Regions, S. Orvig editor, translated from the English, Leningrad, Hydrometeorological Publishers, 443 pp., 1973.

16. The Climate Regime of the Arctic at the Transition From the 20th to the 21st Century, Leningrad, Hydrometeorological Publishers, 200 pp, 1991.

17. Kondratyev, K. Ya., The Radiation Balance of the Earth, Aerology and Clouds, Results of Science and Technology, Meteorology and Climatology, Vol. 10, 316 pp., 1983.

18. Instructions for Hydrometeorological Stations and Posts, Issue 3, Part II, Leningrad, Hydrometeorological Publishers, 115 pp., 1969.

19. Prik, Z. M., Climatological processing of the results of meteorological observations carried out at the drifting stations, Trudy AARI, Vol. 328, 4-21, 1976.

20. Radiation and Cloudiness in the Atmosphere, E. M. Feygelson ed., Leningrad, Hydrometeorological Publishers, 280 pp., 1981.

21. Sonechkin, D. M., On the comparison of data on the degree of cloudiness from satellite- and surface-based observations, Trudy GMTs USSR, Issue 11, 47-55, 1967.

22. Titov, V. I., Cloud climatology from satellite data, Trudy VNIIGMI MTsD., Issue 108, 3-26, 1983.

23. Scientific studies of the drift expeditions on the Main North Sea Route from the MV "G. Sedov" (1937-1940), Vol II, Meteorology, Leningrad, Glavsevmorputi Publishers, 416 pp., 1951.

24. Hunt, B. C., An examination of some feedback mechanisms in the carbon dioxide climate problem, Tellus, 33, No. 1, 78-88, 1981.

25. Key, J., Cloud cover analysis with Arctic Advanced Very High Resolution Radiometer data: 2. Classification with spectral and textural measures, J. Geophys., Res., 95, D6, 7661-7675, 1990.

26. Mokhov, I., Schlesinger, M., Analysis of global cloudiness. Comparison of ground-based and satellite-based cloud climatologies, 1st Scientific Conference on International Global Atmospheric Chemistry [IGAC], Eilat, Israel, Apr 18-22, 1993. Also published in J. Geophys. Res., 99, D8, 17,045-17,065, 1994.

27. Welch, R. M., Kuo, K. S., and Sengupta, S. K., Textural characteristics of cloud- and ice-covered surfaces in the polar regions, IGARSS '89, Remote Sensing: An Economic Tool for the Nineties. Also presented at the 12th Canadian Symposium on Remote Sensing, Vancouver, July 10-14, Vol. 5, 2773-2776, 1989.

†Translators Note: Reference 15 is translated from an English edition. References 24, 25, 26 and 27 are in English. All other references are in Russian.

 

Tables

Table 1: Annual variation of mean total and low level cloudiness, tenths

Table 2: Annual variation of the frequency of clear, partly cloudy, and overcast cloud conditions for total and low level cloudiness

 


 

Image 1

Figure 1. Annual variation of mean total (A) and low (B) cloudiness, tenths at (1) Amderma; (2) Golomyanny and Domashny Islands; (3) Dixon Island; (4) Kotelny Island; and (5) Uelen.

 


 

Image 2a

Image 2b

Figure 2. Spatial contours of mean total cloudiness in (a) July and (b) January in tenths (see reference [1])

 


 

Image 3a

Image 3b

Figure 3. Spatial contours of the frequency of stable overcast skies for total cloud cover in (a) July and (b) January (from the data presented in reference [1]. Values are in percent.

 

 

 

Image 4

Figure 4. Annual variation in the frequency of clear, partly cloudy, and overcast cloud conditions for total and low level cloudiness at (A) Amderma, (B) Golomyanny and Domashny Islands, and (C) Kotelny Island. 1 - white, frequency of clear skies (0-2 tenths); 2 - gray, frequency of partly cloudy skies (3-7 tenths); 3 - black, frequency of overcast skies (8-10 tenths)