Arctic meteorology

Modern meteorological conditions and climate in the Arctic is to a high degree regulated by the advection of warm North Atlantic waters into the Nordic Seas, the Norwegian- and the Greenland Sea . Maritime climate conditions therefore prevail over much of the Arctic Ocean, coastal Alaska, Iceland, northern Norway and adjoining parts of Russia.  In these areas, winters are cold and windy. Summers are cloudy and cool with mean temperatures ranging from 4 to 8^oC over land areas. Annual precipitation is generally between 600 mm and 1300 mm (w.e.), with a cool season maximum (largely snowfall) and about five to seven months of continuous snow cover (e.g., Berry et al. 1993; Barry and Chorley 1998). Shallow permafrost (0-250 m) characterise these regions. Forests are usually absent or found only close to sea level in sheltered positions due to low summer temperatures and/or windy conditions (Humlum and Christiansen 1998).

Arctic interior continental climates have more severe winters and precipitation is usually small. The coldest part of the Northern Hemisphere is located in northeast Siberia near the city Verkhoyansk (Lydolph 1977), where present mean winter (DJF) air temperature is –43^oC. Although frost may occur in any month, long summer days usually provide three months with mean temperatures above 10^oC, and at some sites in the continental interiors summer temperatures may exceed 30^oC. In such regions, forests extend 200-1000 km north of the southern limit of permafrost and, consequently, permafrost extends far beyond the traditional warm limit of periglacial environments (the tree line). Permafrost is widespread and typically reaches 300-600 m thickness.

In winter, arctic weather is dominated by the frequent occurrence of inversions where warm air overlies colder air near the terrain surface, decoupling surface winds from stronger upper layer winds. For this reason, surface wind speeds tend to be lower in winter than one might expect and cold (and dense) air tend to accumulate in topographic lows. In summer, inversions are less frequent and weaker, and the movement of low-pressure systems (cyclones) periodically dominate Arctic weather, even in central Siberia and in the Arctic Basin .

The Arctic is characterized by "semipermanent" patterns of high and low pressure (Serreze et al. 1993; Serreze et al. 1995; Serreze and Barry 1998). These patterns are semipermanent because they appear in charts of long-term average surface pressure. They can be considered to largely represent the statistical signature of where transitory high and low pressure systems that appear on synoptic charts tend to be most common. This pattern is relatively weakly developed in summer, but stronger in winter.

The Icelandic Low is such a semipermanent low-pressure centre located between Iceland and southern Greenland . It is most intense during winter, while in summer it weakens and frequently splits into two centres, one near Davis Strait and the other between Iceland and SE Greenland. Travelling cyclones formed in the subpolar latitudes in the North Atlantic usually slow down and reach maximum intensity when they pass the area of the Icelandic Low. The Aleutian Low is another semipermanent low-pressure centre, located near the Aleutian Islands in the Northern Pacific Ocean . Also most intense in winter, the Aleutian Low is characterized by many strong cyclones. Like the Icelandic Low, travelling cyclones tend to slow down and intensify while passing the Aleutian Low. Areas of significant winter cyclonic activity (storm tracks) are found in the North Pacific and North Atlantic. These channel heat, momentum and moisture into the Arctic, and significantly influences upon the high latitude climate.

Winter cyclones in the Eurasian Arctic occur most frequently in the Barents and Kara Seas region, bringing in pulses of warm air, causing rapid warming and snow melt even in the middle of winter. Over the North Atlantic Arctic, the highest frequency of cyclones occurs east of Greenland after having passed through the Icelandic Low. Cyclones are also common in Baffin Bay between Greenland and Canadian Arctic.

The summer distribution of air pressure and frequency of cyclones is different from that of winter. With more uniform temperatures over the northern parts of the Atlantic and Pacific oceans, summer cyclones tend to be weaker than their winter counterparts and the semipermanent Icelandic and Aleutian Lows weaken. In July and August, few strong cyclones move to the Arctic Ocean from the northern Atlantic , while several weak cyclones move towards the pole from the midlatitudes of Siberia and Canada .

Left diagram shows winter (DJF) sea-level pressure (SLP) averaged over the period 1900-2001. Isobars are spaced every 3 hPa with red colors used for SLP values greater or equal than 1013 hPa and blue colors used for lower values. Numbers at circumference indicate SLP values in hPa. Right diagram shows the modern distribution of permafrost in the Northern Hemisphere. Continuous permafrost is shown by dark blue color. Discontinuous and sporadic permafrost is shown by light blue color. Red and black arrows show main surface air flow (warm and cold, respectively) as generated by the 20^th century pattern of SLP. The overall windsystems set up by the average winter sea-level pressure appears to represent one of several controls on the present distribution of permafrost in the northern hemisphere.

The Siberian High is an intense, cold anticyclone that forms over eastern Siberia in winter (see figure above). Prevailing from late November to early March, it is associated with frequent cold air outbreaks over East Asia. Strong cooling in this region results in the lowest air temperatures in the Northern Hemisphere. A persistent anticyclone or high-pressure ridge called the Arctic High, also known as the Beaufort High, is located over the Beaufort Sea and the Canadian Archipelago in winter and spring. The North American High is a relatively weak area of high pressure that covers most of North America during winter. This pressure system tends to be centred over the Yukon, but is not as well defined as its continental counterpart, the Siberian High. In the winter and spring, anticyclones in the Russian Arctic move mainly from the circumpolar regions through the eastern parts of the Barents and Kara seas. Some also move into the Barents Sea from the northern coast of Greenland . The sea level pressure distribution in summer is dominated by subtropical highs in the eastern Pacific and Atlantic oceans, with relatively weak pressure gradients in polar and subpolar regions. Arctic anticyclones are less common and generally weaker in summer.

Unsolved climatological and meteorological issues in the Polar Regions

The meteorology of the Polar Regions is still poorly understood compared to other regions, and both better observational data and a more thorough analysis of existing data sets are needed to remedy this situation. In particular, the potential control exercised by local temperature inversions on registered Arctic surface air temperatures during the winter season should be investigated for individual meteorological stations. Many of these are located in or near settlements, which tend to be localised in topographic lows, in valleys, along rivers, etc. For such stations the frequent occurrence of winter temperature inversions during periods of calm causes them to be located in a shallow layer of very cold air, thereby recording extraordinary low temperatures not necessarily representative for the region as such. If the average number of calm conditions with temperature inversions during winter is reduced in periods with increased cyclonic activity, this will be recorded as a temperature increase, and vice versa. In summer, inversions are less frequent and weaker, and this potential source of error therefore smaller. New meteorological stations located at high altitudes will help solve this methodological problem (Christiansen and Mortensen, 2002).

The urban heat-island effect in the Arctic deserves separate scrutiny to improve the quality of existing meteorological records. At the village of Barrow, Alaska, Hinkel et al. (2003) recently demonstrated the existence of a strong urban heat island during winter. During winter the urban area averaged 2.2 °C warmer than the hinterland. The strength of the local heat effect increased as the wind velocity decreased, reaching an average value of 3.2°C under calm (<2 m/s) conditions and maximum single-day magnitude of no less than 6°C. Barrow has grown from a size of about 300 residents in 1900 to more than 4600 in 2000.

A central issue in Polar Region climate dynamics is to understand how climates in the Northern and Southern hemispheres are coupled. The strongest of the rapid temperature changes observed in Greenland (so-called Dansgaard-Oeschger events) during the last glaciation have an analogue in the temperature record from Antarctica (Blunier et al. 1998). A comparison of the global atmospheric concentration of methane as recorded in ice cores from Antarctica and Greenland permits a determination of the phase relationship (in leads or lags) of these temperature variations. Greenland warming events around 36 and 45 ka BP before present are lagging behind their Antarctic counterpart by more than 1 ka. On average, Antarctic climate change leads that of Greenland by 1±2.5 ka over the period 47±23 ka BP (Blunier et al. 1998). Also on shorter time scales, there appears to be an out-of-phase between climatic development in the Arctic and the Antarctic, such as demonstrated by the late 20^th century cooling in the Antarctic and the contemporary warming in the Arctic (Ingólfsson et al. 2003).

Blizzard in Longyearbyen, Svalbard, 8 April 2003.

Another pressing meteorological issue is the distribution of precipitation in the Arctic, itself representing a complex problem, subject of long-standing debate and compounded by the paucity of meteorological stations. While air temperatures today are registered at Arctic meteorological stations with relative small technical difficulties, except for sites with icing conditions, precipitation is considerably more complicated to measure correctly, especially when in solid form. Many Arctic meteorological stations have simply avoided measuring precipitation due to severe problems by doing so. In addition, little is known about the local and regional effect of altitude and topography on precipitation. Also the local and regional importance of redistribution of snow by wind is usually virtually unknown (Humlum 1987; Humlum 2002; Humlum et al. 2003; Nordli and Kohler 2003). Finally, much of the information that does exist on precipitation within the Arctic tends to be widely scattered in the scientific literature and is often viewed only in the context of a particular local problem, with little emphasis on the regional amount of precipitation (Humlum 2002). In addition, high-latitude trends in measured precipitation are influenced by gauge undercatch. At a meteorological station exposed to warming, the fraction of annual precipitation falling as snow diminishes, and vice versa. As the gauge undercatch is substantially larger for solid than for liquid precipitation, this implies that a fraction of any observed positive precipitation trend is fictitious, caused by reduced undercatch in the precipitation gauges (Førland and Hanssen-Bauer 2000). For these reasons, we have shunned from discussing precipitation in this contribution.

The general problem of reliable records on Arctic precipitation, however, remains, and also has implications for knowledge on duration and thickness of the snow cover, significant for the ground thermal regime (Ballantyne 1978; Humlum et al. 2003). Snow plays a key role in protecting plants and animals from cold dry winter conditions. It is also important for the seasonal water cycle. Variations in the snow cover may therefore have profound impact on biological activity and geomorphic activity in the Arctic. In addition, the snow cover also has a direct effect on the distribution of permafrost on both local and regional scale (Humlum et al. 2003). In arid parts of the Arctic land regions the average winter snow cover is thin and the ground surface cools rapidly during the winter. Conversely, in more maritime areas the snow cover usually is thicker and reduces heat loss from the ground surface during winter. Interannual variations in the establishment of the snow cover are also important. A dry and cold autumn enables enhanced cooling of the active layer and topmost permafrost, while high snowfall during late winter and late onset of snow melt protect the ground against thawing in early summer. The combination of these two meteorological phenomena is likely to be beneficial for conservation and growth of permafrost. Variations in the timing and duration of seasonal snow cover presumably also have an influence on active layer thickness, but the effect is still not known in detail (Humlum et al. 2003).

A specific problem adheres to the lack of knowledge on mountain climate in general. Despite the fact that high-relief areas (mountains) account for about 20 per cent of the earth’s land surface, the meteorology of most mountains is still little known. Meteorological stations are few and tend to be located at conveniently accessible sites, often in valleys or along coasts, rather than at points selected to obtain representative data. Precipitation distribution in mountain areas has been a subject of debate and controversy since the publication on orographic rainfall by Bonacina (1945). The problem is compounded by the above-mentioned paucity of high-altitude meteorological stations and the additional technical difficulties of determining snowfall contributions to total precipitation, especially at windy sites. As recognized early by Salter (1918) from analysis of British data, the effect of altitude on the vertical distribution of precipitation in mountain areas is highly variable between even nearby geographical locations.

This poor understanding of the dynamics and characteristics of mountain climate is particularly pronounced for the Arctic (Humlum 2002). This is unfortunate; partly because of existing predictions of an amplified response of northern high-latitude regions to various climatic forcing mechanisms (see above), partly because most geomorphic activity is not controlled by temperature only, but very much also by precipitation and wind, e.g. frost weathering, gelifluction, active layer processes (e.g., Etzelmüller and Sollid 1991) and the dynamics of glaciers (e.g., Dowdeswell et al. 1997). Also any kind of biological activity is likely to be influenced by the amount of precipitation.