Skip to main content
 Previous Next
  • Zoom In (+)
  • Zoom Out (-)
  • Rotate CW (r)
  • Rotate CCW (R)
  • Overview (h)
The Climate of the American Northlands: Encyclopedia Arctica 7: Meteorology and Oceanography
Stefansson, Vilhjalmur, 1879-1962

The Climate of the American Northlands

EA: Meteorology
[F. Kenneth Hare]


Introduction 1
Dynamic Climatology of the American Arctic 8
The Temperature Regime 35
Precipitation, Humidity and Cloudiness 48
Conclusion 56
References 57

EA: Meteorology
[F. Kenneth Hare]

The Climat ology and Meteorology e of the American Arctic Northlands
No special quality separates the Arctic climates from those of
temperate latitudes. As one passes from the mild climates of the southern
United States to the wholly Arctic climate of the Canadian north, one en–
counters no fundamental climatic boundary, no line beyond which the laws
governing atmospheric motion change abruptly. Not even the fact of cold–
ness is wholly Arctic; the Arctic summer is far from cold until we pene–
trate almost to the Pole itself. In brief, the term “Arctic” is arbitra–
ry in meaning insofar as climate is concerned. Whatever limits we choose
to apply to the Arctic will necessarily be in some degree conventional.
The commonly accepted southern limit of the true Arctic is the
“tree line”; where such a line can be clearly distinguished, much can be
said for it as the least conventional boundary. It will be seen a little
later that windchill rather than low temperatures is the real measure of
the Arctic cold from a physiological standpoint. Within the relatively
sheltered region of the forest - even in the thin taiga of the sub-Arctic
lands - the force of the surface wind is much reduced, and the magnitude
of windchill similarly cut down. Out on the tundra beyond the tree-line
there are few obstructions in the path of the wind, so that windchill is
correspondingly much greater than in the nearby forest. The effect of this
reduction in windstrength is also seen in the check the forest offers to
blowing snow, a characteristic weather type of winter in the tundra.
In yet another way the tree-line is a significant divide. The
animal life of the forest differs greatly from that of the tundra. There are, of course, certain migratory species like the caribou that systemat–
ically use both forest and tundra, but in general the two great vegetat–
ional formations have distinct suites of animal life. This has economic
repercussions for trappers and hunters, quite apart from its scientific
But the tree-line has certain disadvantages which partially off–
set the arguments set out above. What does one mean by the term tree-line ?
Is it the northern limit of tree-growth? If so the line is exceedingly
sinuous, winding far north along the main rivers and passing far south on
the interfluves. Or is it the northern limit of continuous forest forma–
tion? This at once raises the question “what is forest?” and the argument
swings full circle. Between the dense stands of the Boreal or Canadian
Forest on the southern part of the Canadian Shield and the open tundra
there lies in fact a very broad belt of open woodland or sub-arctic [: northland.] [: ]
[: ] The characteristic association
in th e is [: ] formation is of widely spaced white and black spruce and tamarack with a rich floor
of reindeer moss (Cladonia sp.) : in places there are wide stretches of open mossy tundra
within the [: taiga] wooded zone itself, the countryside having a parkland aspect - the
“Parktic” as it has been dubbed.
This passes northwards into the scatter–
ed trees and wide treeless patches near the treeline with no clear-cut
boundary between them. In places, moreover, the trees near the tree-line
are dwarfed or even recumbent forms which make a mockery of the term tree.
None-the-less, when all these arguments have been reviewed, we
still accept the existence of the tree-line, however complex it may be on
the ground itself, and the arguments in favour of accepting it as the sou–
thern Arctic limit are overwhelmingly strong. It is clear, then, that the
best climatological climatological limits for the Arctic are those which pertain along the tree-line, and it is this convention which is commonly accepted.
The time-honoured equivalent for the tree-line is the mean iso–
therm of 10° C (50° F.) for the warmest month: this limit dates back to
Supan or beyond, though it is normally associated with the name of W.
Köppen (1) whose classification of climates remains the most widely known
of quantitative schemes. On a world vegetation map the rough and ready
equivalence of the two lines may be confirmed by anyone.
A more refined limit has been suggested by O. Nordenskjöld (2),
who was dissatisfied with the Köppen-Supan line. He suggested that the
line satisfying the identity
W = 9 − 0.1K,.......(1)
where W = mean temp. (°C) of warmest month
where K = mean temp. (°C) of coldest month
offers a more effective limit, and a closer fit to the tree-line. In Fah–
renheit degrees, this identity becomes
W = 51.4 − 0.1K......(2)
Nordenskjöld’s objective was to allow for the great contrasts which exist
between continental and maritime areas. He had been impressed by the lux–
uriant forests of Tierra del Fuego, where no month has a temperature ex–
ceding 50° F., but where the winters are exceedingly mild. The effect of
his formula is to make necessary a greater degree of summer heat in the
continental interiors along the Arctic limit. On fig. 1 are indicated for
North America the tree-line (following Halliday, 3), the 50° F. warmest
month line, and Nordenskjöld’s line. The latter fits the tree-line more
closely than Köppen’s limit.
The sub-Arctic, if it is defineable at all, is still more vague
and less concrete than the Arctic. Plainly it is in general tree-covered Fig. 1 Some important divisions of the northlands
This diagram shows some of the devices used in delimiting the Arctic
and sub-Arctic. The sub-Arctic limit shown is that line north of
which less than four calendar months have a mean temperature of over
50°F (10°C.). Tree-line follows various sources. land if not forest. We might define it as the southern limit of perma–
frost, the northern limit of true Boreal forest, or in a host of equally
uncappable ways. The real significance of the term is probably economic;
the writer likes to think of the sub-Arctic as that area of forest or
woodland north of the northernmost lands which may be cleared and farmed.
At present true farming settlement in North America has not been pushed far north of the
line of approximately 75* days between first and last killing frosts: the
marginal settlements in the clay belt of northern Quebec and Ontario, the
pioneer lands of Northern Saskatchewan and the Tanana-Yukon valleys lie
close to this line. It is true that vegetable gardens have been maintain–
ed north of this limit for some time in the Mackenzie Valley, but there is
all the difference in the world between vegetable gardens and the settled
rural community drawing its subsistence wholly from the soil.
A more commonly accepted division is that suggested G. T.
Trewartha (4): he regards the line where four summer months have mean tem–
peratures above 10° C (50° F) as the sub-Arctic limit (see fig. 1) : in many areas this
is close to the 75 frost-free days line . (see fig. 1). This is the line
separating Köppen’s cold microthermal from warm microthermal climates (viz.
Dcf, Dew, etc. are sub-Arctic) so that there already is some sanction for
it. The line is not far distant from the boundary between the true Boreal
forest and the sparser woodland or [: ] [: northland] taiga belt . to the north.
The climate of the American Arctic is affected by certain broad
physiographic controls. These are exerted by (i) the great mountain bar–
riers of the Alaskan and Canadian Rockies and associated ranges; (ii) the
ice-cap plateau of Greenland, associated in some small degree with the
ranges of Ellesmere and Baffin Islands; (iii) the seas and channels that
criss-cross the Canadian Arctic. We must examine each of these in turn.
(i) the Western Cordillera comprise the complex series of ran–
ges, valleys and plateaus between the Mackenzie Lowlands and the Pacific,
including the whole of Alaska except the narrow Arctic coastal plain. The
role of the Western Cordillera is essentially that of a barrier which ei–
ther inhibits the flow of the lower atmosphere across the ranges entirely,
or else [: ] radically alter ed s the physical properties in of the
airmass concerned. In effect the mountains separate a thoroughly maritime
coastal climate in Alaska and British Columbia from an extreme type of
continental climate in the Mackenzie Valley. The main climatic divide is
the Alaskan Range and the British Columbian Coast Range, for the Yukon
and Tanana valleys have almost as extreme a climate as the Mackenzie country it–
self. The precise way in which the barrier action works will be discussed
in several places in the text.
(ii) the great ice-cap plateau of Greenland also acts as a major
climatic barrier, though it stands second to the Western Cordillera in this
respect. Much controversy has surrounded the role of the Greenland ice-cap
in the control of world climate, and special treatment of this topic will
be necessary later in the essay.
(iii) the Arctic seas and channels profoundly influence the cli–
mate of a wide area of North America. In winter their surface is largely
ice-covered, and they function in large part as an extension of the contin–
-ental surface. Some small amount of heat escapes through the ice itself,
as well as through the leads of open water commonly found even in mid–
winter: nevertheless observations at coastal points make it clear that the
ameliorating influence of the sea is small in winter. Even Hudson’s Bay,
formerly believed to contain extensive open water throughout winter, is
now known to be frozen over almost completely from January [: ] onwards. * The only extensive open
water channels from January to May appear to be (a) parts of Hudson Strait
and (b) the eastern half or Davis Strait and Baffin Bay. Both these areas
are remarkably warm for their latitude in consequence.
The second “season”, as it were, is the season of ice melt - in–
cluding in most areas June, July and much of August. Even though the pack
ice may break up, melt or founder in June and early July, the very cold
ice-melt water keeps sea temperatures in the 30’s and low 40’s throughout
the summer. The cold sea surfaces act as a conspicuous climatic influence
all through the warmer season: they are responsible for the extensive sea–
fog and stratus clouds so typical of the eastern Arctic’s brief, damp and
cloudy summer, and for the similar weather experienced in coastal Alaska.
The role of the sea begins to change late in August, and through–
out the Fall and early winter differs radically from the part it plays in
summer. At this season the Arctic is traversed by increasingly frequent
and vigorous outbreaks of fresh polar air; as these come into contact with
the open water - now relatively warm - of the seas and larger lakes, they
are heated and moistened from below. Strong convection is set up, and cum–
ulus cloud, snow flurries and poor visibility result. As the cold air tra–
vels southwards across the Canadian Shied, each lake sets up its own little
pile of cumulus as long as the water surface remains unfrozen.
Significant though these physiographic controls may be, the
fundamental control of the Arctic climates remains dynamic, and it is to
the dynamic climatology that we must turn first. This expression was
coined by Tor Bergeron (5) to denote the explanatory description of cli–
matic phenomena in the language of the newer methods of meteorological an–
Dynamic Climatology of the American Arctic
The Arctic in the General Atmospheric Circulation
Though the daily weather map indicates wide variations in the
pattern of the general atmospheric circulation from day to day, certain
broader patterns appear to persist or recur at frequent intervals. The
complex circulation made up by these persistent or recurrent features is
what we call the general circulation of the atmosphere ; this latter is
perhaps best regarded as a statistical generalisation of the more complex
reality, or as a logical device by which we reduce that reality to meas–
ureable terms. If we accept this generalization for working purposes, it
is plain that we can recognize the following classification of atmospher–
ic circulations:-
(i) primary circulation patterns, viz. those of the general cir–
culation itself;
(ii) secondary patterns, or those large scale phenomena which
last only for a brief period; and
(iii) tertiary, or local features of the circulation such as
land and sea-breezes, katabatic winds and the like.
We are here concerned with the primary features, the rest being deferred
to a later section.
Figs. 2-3 show average sea-level pressure maps of the entire
Arctic for January and July. Certain features are exhibited in common by
both maps:
(i) The Polar High Pressure Belt
Fairly central to the picture is a belt of relatively high
pressure. In July it lies centred on the American side of, but quite
close to, the Pole itself, whereas in January it forms only a ridge link–
Fig. 2. Mean Sea-Level Pressure in January. (Courtesy H. G. Dorsey, Jr.)
The prevailing low-level flow of air is parallel to the isobars, its
vectorial mean velocity being greatest where the isobars are closest to–
gether. Hence mean circulation over the Eastern Canadian Arctic consists
of a great, cyclonically-curving flow of Arctic air passing southward
at high speeds. Over the western Arctic, however, quieter conditions
prevail, and the surface air is locally derived. In coastal Alaska
stormy conditions prevail, with air derived from the Pacific.
The positions of the European and American Arctic frontal zones are
indicated by the customary symbol. Fig. 3 Mean Sea-Level Pressure in January
This diagram is uniform with fig. 2, except that dashed isobars are
shown for the values midway (i.e. 2.5 mo.) between the standard isobars.
in order to indicate the form of the gentler pressure gradients typical
of summer.
Note the extended form and northern position of the Siberian-American
Arctic frontal zone. -ing pronounced high pressure areas over Siberia and the lower Mackenzie
Valley. The high pressure belt appears on all other monthly mean maps,
and must hence be regarded as a primary feature of the general circula–
tion. We refer to it as the polar or Arctic high pressure belt .
It is necessary to be very clear about the precise status of
the polar high pressure belt. Early meteorological theory presupposed
that the Poles would be covered by permanent low pressure, but the first
expeditions to penetrate with meteorological equipment well into high
latitudes made it clear that pressure was low only in sub-polar latitudes,
tending to rise as the poles were approached. It is equally fallacious,
however, to talk of “the permanent polar anticyclone”; intense anticy–
clones are quite rare over the true polar belt, and are seen much more
often over high latitude continental regions. The facts are quite simple:
the polar belt is normally covered by a belt of relatively high pressure,
much higher than the very low pressure characteristic of sub-polar lati–
tudes; nevertheless, the polar high pressure belt is highly variable in
position, intensity and in significance. Rarely does it achieve the sta–
tus of a major anticyclone.
The January map calls for some further comment. In that month,
as in all the winter months, mean pressure is far higher in the Mackenzie
Valley and over Siberia than in the polar belt itself. This is a direct
effect of the intense radiative cooling affecting these vast regions in
the long, dark winter. Whenever dry, clear air lies over the cold, con–
tinental interiors in winter, conditions are favourable for the formation
of cold anticyclones, that is of anticyclones in which the high pressure
is due to the high density of the cold surface layers. Over Siberia, in–
-tensely cold throughout winter, high pressure is virtually continuous,
and is associated with the vast outflow of polar air known as the Asiatic
north-east Monsoon. Over the Mackenzie Valley, however, cyclonic activ–
ity of Pacific origin is by no means uncommon, so that mean pressure is
much lower than in Siberia.
It is clear, none the less, that these continental high pressure
areas effectively replace the polar high pressure belt during the winter.
(ii) the Sub-Polar Low Pressure Belt . Over the Arctic as a whole,
both January and July mean pressure maps (Figs. 2-3) indicate low pressure
in latitudes 60° - 65°; this also is a feature of the entire year, though
the intensity of low pressure is much greater in the winter half-year. We
apply the term Sub-Polar Low Pressure Belt to this primary circulation pat–
tern. In detail we observe that there are two main cells of low pressure.
One of these lies on the January map between Greenland and Iceland, where
it is also found throughout the winter half-year. The term “Icelandic Low”
is applied to this cell, though in July lowest pressure lies further west
over Southern Baffin Land. The second main cell lies over the Aleutians
in winter: hence the term “Aleutian Low”. In summer the centre of lowest
pressure is much further west, lying over the hilly country of northern
Siberia. Both cells, then, appear to retreat westward towards the nearby
continent in the warmer months, when they are also least intense.
The twin sub-polar low pressure cells are obviously primary fea–
tures of the general circulation. The reason for their existence, however,
involves the secondary circulations. They represent “graveyard” regions
into which the main streams of travelling cyclones of the northern hemis–
phere discharge. A constant succession of mature, occluded cyclones of
considerable intensity approaches the two regions concerned, there to de–
-cline and ultimately to vanish, each cyclone being replaced by another al–
most at once. The low pressure of the average map is due to the almost
constant presence of these intense but declining systems. The paradox in–
volved in our classification of circulations into primary and secondary
orders is now apparent, for the Icelandic and Aleutian lows, themselves
primary systems, are plainly the integrations of numerous secondary cir–
culations. None the less the recognition of the two cells as primary fea–
tures is a necessity.
The surface mean pressure maps, then, show a structure resolva–
ble into two primary features, a polar high pressure belt and a twin-celled
sub-polar low pressure belt. This structure , [: ] is essentially sup–
erficial, and it will now be appropriate to consider the upper air. Modern
meteorological theory attaches increasing significance to the higher-level
circulations, which are being investigated and explored by means of a vast–
ly increased network of radiosonde stations. In the Arctic we have hardly
entered the exploratory phase, for systematic observation of the upper air
in high latitudes is of very recent date, especially in America. Today
there are many stations carrying out upper air observations in the north–
land, but in most cases their records are woefully brief.
A glance at a map of mean pressure at any level above about 10,000
feet and in any month shows us that the surface configurations are extremely
shallow. Instead of relatively high pressure over the poles, we find a
gigantic low pressure system, usually twin-centred, with centres over the
Alaskan-Siberian and Greenland sides of the Arctic. Around this polar low
a vast system of westerly winds circulates throughout the year apart from
brief interruptions lasting only a day or two. These “circumpolar wester–
lies” are the dominating feature of the atmospheric circulation of middle and
high latitudes. As yet, however, we known little of their direct influence on
the surface weather of the Arctic.
Though our views on the upper level circulations are hazy, we
know enough about their significance in lower latitudes to say with con–
fidence that the study of the upper air is the most urgent and crucial
task facing the student of the Arctic climates.
Airmass Classification
Most of our ideas about the circulation of the extratropical at–
mosphere are intimately bound up with the frontal and airmass concepts in–
troduced by the group of Norwegian meteorologists at the Meteorological In–
stitute at Bergen - notably by V. Bjerknes, his son J. Bjerknes and the
Swede, T. Bergeron (6). For an introduction to these concepts the reader
is referred to the cited literature. The following account will assume a
general familiarity with the field.
The internationally accepted classification of airmasses was or–
iginally proposed by Bergeron (7) and stands today little changed from the
original form. The classification comprises three terms: (i) a latitudinal
term, expressing the zonal origin of the airmass; (ii) a term expressing
the degree of “continentality” of the airmass; and (iii) a term expressing
the change in thermodynamic stability, i.e. resist e a nce to convective turbu–
lence, of the airmass. Table I lists the airmass types with special refer–
ence to their occurrence in the American Arctic.
Table I overleaf.
Table I The Airmasses of the American Arctic
Symbol Airmass Type Area of Origin Properties(winter) Properties(summer)
cA Continental Arctic Permanent ice-pack,
Arctic archipelago
(winter only)
Very stable, in–
tensely cold and
very dry.
Does not occur
mA Maritime Arctic Arctic Ocean,
Arctic archipelago
Does not occur in
American Arctic
Stable, cool and
moist, Becomes
unstable inland.
cP Continental Polar N.Canada, Yukon,
central Alaska,
Very stable, in–
tensely cold and
Fairly stable,
cool and rather
mP Maritime Polar,
Pacific south and
west of Aleutians
Relatively warm,
Moist on coast,
dry east of Rockies
Cool and moist
(dry east of
Atlantic east of
Relatively warm
and moist
Cool and moist
mT Maritime Tropical Sub-tropical waters
of Atlantic.
Warm and very
moist. Occurs only
Warm and very
moist. Rare ex–
cept aloft.
As this table suggests, the air over the American Arctic comes
from certain definite source regions, each of which imparts definite char–
(1) Arctic airmasses are those which originate over the Arctic
Ocean and the Arctic Archipelago of Canada. They are sometimes lumped to–
gether with the Polar airmasses, but it seems desirable to separate them
from airmasses formed over sub-polar latitudes (to which the label “polar”
has been applied: historical accident explains the queer inversion of names).
In winter this entire vast area is snow- and ice-covered; “leads” in the
sea-ice are few and shrotlived in most cases, though there are a few pat–
ches of fairly permanent open water. The ice of the permanent Arctic pack
and in the channels between the islands is thick enough to reduce the up–
-ward flow of heat from sea to air to very small proportions. The surface
is thus effectively continental, and continental Arctic is an important
winter airmass. In summer, however, the general thaw creates a wet surface
over much of the source region: the sea-ice in part breaks up, in part be–
comes wet as it thaws in situ (as in the permanent pack). On the islands
the soil is kept wet by capillary rise from the permafrost layer a few in–
ches below the surface, or by sluggish drainage. Under summer conditions,
then, a vast area including the Arctic Ocean and much of the Arctic Archi–
pelago presents a damp, cool surface; the term “maritime Arctic” is some–
times applied to the airmasses formed over it, but in practice there is very
little contrast between such air and the maritime polar formed over high
latitude oceans. Recognition of a separate class of “Arctic” air loses
much of its value in summer.
The properties of Arctic air resemble those of Continental Polar
air in winter and maritime Polar in summer. In winter the outstanding char–
acteristic is the remarkable stability: an intense inversion separates the
very cold surface layer from the actually and potentially much
warmer air of the middle troposphere. Fig. 4 shows a typical temperature–
height curve in winter on air.
There is a shallow unstable surface layer
from 500 to 2000 feet thick associated with minor turbulence and sometimes
with very thin stratus cloud. Above this layer there is an abrupt inversion,
and temperature rises with increasing height to as much as 5000 or 6000 feet,
at which levels the highest temperatures are normally encountered. There–
after temperature decreases in normal fash io n to the tropopause.
This characteristic winter stratification re du ces turbulence to
a minimum above the shallow surface layer. The extreme dryness typical of
the airmass also inhibits cloud formation, and the normal weather associated
with such air is cloudless and extremely smooth in flight. At all levels, however, Arctic air in winter is very cold, calling for special precautions
both for movement by land or by air.
In summer, the characteristic stability is much reduced. The sur–
face layers are moist and rather unstable, and layers of stratus or strato–
cumulus are by no means uncommon. There is normally a weak stable layer
only two or three thousand feet above the surface, and above this layer the
airmass is fairly free of cloud and turbulence. Spectacularly good visib–
ility is often reported in mA air in summer both above and below the low
cloud layers. Inland over the larger islands and on the mainland, however,
the shallow stable layer is rapidly dispersed be solar heating: cumulus
cloud and even showers are often reported in southern districts in Arctic
(ii) Continental Polar Air is formed in areas of anticyclonic cir–
culation over the mainland, especially in Alaska, the Yukon and the Macken–
zie Basin. In winter the air closely resembles cA air, but the shallow un–
stable surface layer is absent, temperature rising from the surface right
up to the top of the inversion at levels varying from 3,000 to 8,000 feet.
cP air is apt to be even colder than cA air at low levels; temperatures of
below −70° F. have at times been recorded in this airmass. Weather is clear,
intensely cold and dry; unless blowing snow is present, visibility is un–
limited and turbulence negligible at all levels away from fronts.
In summer the opportunity for the creation of this airmass is re–
duced; prolonged spells of anticyclonic circulation are uncommon over the
source regions, which are in any case quite warm at this season. cP air
tends to be dry, fairly stable (because of its dryness) and, by Arctic stan–
dards, quite warm. Typically it gives clear, bright weather with scattered
cumulus cloud. If it moves east of its source region it picks up moisture
over the lakes of the Shield or Hudson’s Bay and rapidly becomes indisting–
uishable from the mA air dealt with above.
(iii) Pacific-type maritime Polar air is extremely common in
all western parts of the Arctic at all times of year. Its properties,
however, vary widely from place to place. The ultimate source for such
air is northern Siberia, but it derives its immediate properties from long
travel across the high latitudes of the North Pacific. mP air arrives on
the Alaskan coast from south-west; in winter it is relatively warm at low
levels, and quite moist, but it is very unstable, being cold aloft. The
heavy rains and snows of winter on the Alaskan Range and coastlands fall from
the warmer varieties of this airmass. In summer however, it is much more
stable, and habitually brings extensive sheets of stratus or even sea-fog
to the coastal belt.
In winter mP air does not normally penetrate inland across the
Alaskan Range into the Yukon and Tanana valleys, being deflected parallel to
the coast; the interior valleys are in such cases normally covered by cold
cP air, and remarkable contrasts in temperature are observed between the
two sides of the mountains. Occasionally, however, a current of mP air
bearing considerable moisture crosses the mountains at high levels, and ap–
pears aloft over the Yukon, the Tanana and the Mackenzie. It reveals its
presence by layers of altostratus or altocumulus cloud; light snow may fall
over a wide area. Occasionally such air may be drawn to the ground level
in the interior, with spectacular rises of temperature; such episodes are
analogous to the Chinooks of Alberta (see below).
In summer Pacific mP air enters Alaska and the Mackenzie Valley
with greater ease; the currents are deeper and more persistent. Though
on the coast the air is cool and moist, in the interior temperatures be–
come quite high, the moist, cloudy low-level air is dispersed, and weather
becomes very pleasant: indeed on occasion mP air may give temperatures in excess of 80° F. as far north as Whitehorse and Fort Good Hope. There is
usually considerable turbulence over the mountains, and everywhere at low
In spite of the crossing of the mountains, Pacific mP air is the
main moisture bringer in the inland western Arctic, a land of markedly de–
ficient rainfall. When involved in frontal activity with other airmasses,
it may give wide areas of light but persistent precipitation.
(iv) Atlantic maritime Polar air is common only in the Greenland–
southern Baffin Land-Ungava region, though on occasion it may penetrate
west to Keewatin and the Canadian prairies. Its deep penetration into the
continent requires an abnormal pressure distribution. Such deep penetra–
tions may have amazing effects on temperature distribution.
In its source region, the high latitude belt of the Atlantic,
mP air is moist, unstable, mild in winter, very cool in summer. If it
moves west onto the Labrador, Baffin or Greenland coasts it is always
chilled, and becomes superficially stable. Thick layers of stratus dev–
elop at low levels, with sea-fog patches in summer. Drizzle or freezing
drizzle is also common.
In winter such air carries its relative warmth and thick cloud
cover far inland on occasion. In February 1947 an enormous and very per–
sistent flow of mP air covered much of Canada as far west as the Mackenzie
and as far north as Lancaster Sound. Everywhere the sky was clouded, with
drizzle or light showers. In some regions average temperature was more
than 20° F. above normal for the month. Though such invasions are infer–
quent, they are very striking in their effects.
An unusually high frequency
during recent winters has produced a
[: striking] amelioration of inter temperatures
in the Eastern Arctic.
In summer Atlantic mP air is very cool and moist. As it passes
westwards across the cold surface waters of the Labrador current, Hudson’s Strait and Bay and the Davis Strait it becomes saturated by chilling, and
sea-fog is widespread below the stratus sheets typical of the airmass. Ex–
posed coasts are normally blanketed thickly, and inshore navigation pres–
ents real hazards. Inland, however, the heat of the sun rapidly disperses
both fog and low cloud layers by day, and the inherent instability of the
air reasserts itself: cumulus cloud and showers are very general. The
writer recalls seeing large cumulo-nimbus passing over Goose Bay, Labrador
in such air while the marine stratus still hung thickly over nearby Lake
(v) maritime Tropical air (mT ) is a real rarity at the surface
in the Arctic; on very infrequent occasions it may enter Hudson’s Bay or
the Ungava region in the warm sector of travelling cyclones, but such oc–
casions are rare enough to make them acts of God rather than features of
the climate. Aloft, however, it is all otherwise: currents of air of ap–
parently tropical origin may enter the eastern Arctic in the middle and
upper troposphere at times of weak circumpolar westerlies, revealing their
presence by extensive sheets of altostratus or altocumulus. Recent high–
level soundings by radio-sonde also suggest that stratospheric currents of
tropical origin may also enter the Arctic. Too little is known, however,
of these tropical invasions to permit us to speak of them with any cert–
(vi) Returning continental Polar air (cPW ) is not included in
the Table I, as it does not represent a generically distinct airmass. When
cP or cA air streams southwards across North America, they are almost al–
ways colder than the surface over which they pass; the symbol “K” (German
‘Kalte’) is added to the airmass label, viz. cPK, cAK. The circulation
pattern which causes such outbreaks is the development of a cold anticy–
-clone over the western Arctic; and when [: ][: a] s such an anticy lc cl one moves away
southeastwards (the customary route) conditions are favorable for a broad
return current of cP or cA air northwards: such currents are composed of
air which has swept far south across the continent before returning north
across the western Arctic. They are in general substantially warmer than
the surface over which they pass, and are labeled “cPW” (from the German
‘warme’). Further east the rapid decline in intensity of the anticyclone
usually restricts the magnitude of the return current.
In winter the warming effect on such far travelled air is small,
for the continental interior is then much chilled. Temperatures in the
Yukon and Mackenzie Valleys rise substantially, but by no means so rapid–
ly as they do when mP air arrives. In summer, however, the continental
interior is hot and relatively moist; the cP air is warmed and moistened
and returns to the north-west as the warmest airmass experienced by those
regions. Temperatures may rise to 90° F. in many districts once or twice
in a year in cPW air. The high humidity permits cumulus formation, and a
high degree of low level turbulence is characteristic, especially over the
mountains. Showers and thunderstorms, however, are exceptional unless the
air is undercut by a fresh cold outbreak.
Fronts and Frontogenesis
The term “frontal surface” is applied by meteorologists to the
sloping surfaces of discontinuity in the density and velocity fields which
separate dissimilar airmasses. The line along which such a surface in–
tersects ground-level or any other standard surface horizontal plane is called a “front”.
These surfaces play a main part in the formation of cyclonic depressions
and of most forms of bad weather. For a more detailed discussion of their
physical characteristics the reader is referred to the standard treatises (8).
In discussing the fronts affecting the American Arctic, we must
distinguish between: (i) the semi-permanent zones of frontogenesis (that
is, of front-formation) within the Arctic itself, and (ii) fronts further
afield whose associated disturbances may enter Arctic regions. In fact
much the most severe disturbances affecting the Arctic are those which
develop externally, so that we cannot restrict the discussion to the home–
grown variety. The fronts concerned are listed below:-
A. Fronts formed internally within Arctic Territory -
(i) the American Arctic Frontal Group;
(ii) the European Arctic Frontal Group;
(iii) the lateral frontogenetic zones of the Greenland region.
B. Other Fronts affecting the American Arctic -
(iv) the Atlantic Polar Frontal Group;
(v) the Pacific Polar Frontal Group.
Frontogenesis, the dynamical process by which fronts are created
or intensified, is greatly favoured by the existence of strong horizontal
temperature gradients; hence within the uniformly cold Arctic regions it
is neither common nor intense. The American Arctic, however, is peripheral,
and comes into sharp contact with non-Arctic lands and seas in certain
areas. Internal frontogenesis is hence in North America, at least, a sig–
nificant feature.
The American Arctic frontal group * is a good case in point. The
term “frontal group” is here preferred to the commonly used term “front”
to indicate that we refer to the numerous fronts forming in this zone ra–
ther than to any individual system. The fronts of this group form between
3 cP, cA and mA air over the Arctic Basin or the Mackenzie Valley, and mP
air from the Pacific; in winter the temperature contrast between these
airmasses is very great, and frontogenesis correspondingly intense. In
summer the contrast between cP and mP air is inconsiderable, but mA re–
mains very much colder than mP, so that the fronts are most notable over
Alaska and north-east Siberia.
The “mean” or standing winter position of the American Arctic fronts
lies from southern Alberta through the Stikine Range, across the head–
waters of the Liard River, then to the St. Klias and Alaskan Ranges, and
finally to the coast near the mouth of the Yukon. Thus the Yukon and Tan–
ana Valleys are habitually covered by cP and cA air, and have frigidly
cold weather; by contrast the Pacific slope of the St. Elias and Alaskan
Ranges have the much milder weather characteristic of mP air. Further
south the same contrast is visible between coastal British Columbia and
the Liard and Peace River Basins. To what extent, however, the surface of
separation is a true front is questionable; for much the greater part of
its length, the boundary consists of solid rock, and not of an intangible
frontal surface. If, on the other hand, the normal condition is disturbed
in any way, and the boundary shifts, a clear-cut surface of separation at
once becomes visible; on these grounds the term “American Arctic Front” is
easily defensible.
Displacements of the American Arctic Front may lead to striking
anomalies in winter weather. If, for example, mP air penetrates into the
Dawson region, the normal sub-zero weather is replaced by weather as warm
as that normally experienced on the coast. In February 1938 , for instance
(9) temperatures between 40° and 50° F. were general in the Yukon through–
out the latter part of the mouth. Similar invasions of mP air eastwards across the Rockies and Stikine Mountains produce the familiar “Chinooks”
of the Peace River country and southern Alberta. Such invasions are much
rarer in the Mackenzie Valley below For Simpson because of the added bar–
rier of the Mackenzie Mountains.
The American Arctic fronts in winter are often associated with
cyclogenesis, though the precise pattern of circulation in such cases is
obscure. The arrival of a Pacific cyclone, usually occluded, over the
coasts of Alaska or British Columbia at this season is accompanied by
gales, thick cloud and heavy precipitation - snow on all but the lowest
ground; from such storms the great glaciers are fed. The forward movement ,
however, of the cyclones is arrested by the coastal ranges, and their cen–
tres normally (though by no means always) become stationary near the Aleu–
tians or over the Gulf of Alaska. The Yukon-Tanana trench gets a thin
overcast and often light snow, but the weather is much less severe than be–
yond the mountains. In some cases, however, the cyclonic centre actually
does appear to penetrate inland. In either case the frontal system tends
to penetrate eastwards across the British Columbian plateaus. The “Amer–
ican Arctic” front now comes into play: the mP airmasses south of the storm
centre move east across the Cordillera, either displacing or overrunning
the cP or cA air ahead. The movement inland reveals itself by sheets of
cloud and sometimes by light snow. The mP air only comes down to ground
level on most occasions south of the Liard River - giving the familiar
Chinook. Over the Meckenzie below Fort Simpson overcast skies, light snow
and easterly cP winds are all that are observed.
Though the precise details of such penetrations are as yet income–
pletely understood, the eventual result is indisputable. In the great maj–
ority of cases the passage of a Pacific disturbance - whether upper cold
front, occlusion, or non-frontal trough - inland leads to fresh cyclogen–
-esis east of the Rockies, generally in the Peace River basin or further
south. Such storms appear to derive their energy from the Arctic front,
on which they are initially centred; they travel eastwards across the
prairie provinces into Ontario and Quebec with marked warm sectors of mP
air and with easterly winds of cP or cA type to the north. Their chief
contribution to Arctic weather is in bringing spells of fresh easterly
winds with overcast skies and light snow to the southernmost districts, es–
pecially Keewatin district, Hudson’s Bay and Ungava. Some of these “Alberta”
lows, as they are called, travel eastwards along a more northerly path,
bringing light snow and strong winds to wide areas.
In summer cyclogenesis along the Arctic front is less frequent,
and exists in its own right, being as a rule independent of Pacific storms,
which are infrequent in mid-summer. Their eastward movement of Arctic front cyclones across the
American Arctic, including Alaska, is accompanied by considerable low
cloud and some rain, though the available moisture is too small for real
downpours. Such storms are the chief source of disturbed weather in the
north and west Arctic in summer.
Before we leave the American Arctic frontal group and its assoc–
iated disturbances, we must stress, however, the great complexity of atmos–
pheric circulation which is characteristic of the whole western Cordilleran
belt in North America. The foregoing account is an attempt to generalise
processes which almost defy generalisation. Furthermore we are far from
an adequate comprehension of the barrier role of these great mountains and
their intervening valleys. Scepticism is the only correct frame of mind
in which to read this account. It should always be remembered, however,
that one of the two major groups of cyclonic storms affecting the American
Arctic is associated with this little-studied airmass boundary.
The Atlantic Polar Frontal Group provides the other main storm
group, which affects chiefly the eastern Arctic. Fronts of this group are
formed between cP or cA air and Atlantic cPW or mT airmasses. The zone
of frontogenesis tends to lie along the Atlantic Coast in winter, shifting
inland towards the Great Lakes-central Labrador region during the hot seas–
on. Fronts are rarely created in this zone; what happens is that cold
fronts leading outbreaks of cP or cA air become subject to intensification
as they enter the zone. At all seasons fronts of this group are intense,
and because of the high humidity of the Atlantic air involved, possess great
reserves of energy available for cyclone formation.
The Atlantic Polar Frontal Group is one of the two most fertile
sources of cyclonic disturbances in the northern hemisphere. A procession
of such storms passes north-eastwards across the eastern U.S., Canada, or
the western Atlantic. Some of them penetrate east to Europe, and are lost
to the Arctic. The majority, however, eventually reach the ‘graveyard’
region of the Icelandic Low over Southern Greenland or Iceland, and so af–
fect the eastern Arctic. A few (especially in the warmer season) actually
move northwards or even northwestwards into the Canadian Arctic.
The Atlantic Polar Front cyclones are of considerable intensity
for seven or eight months in the year. The summer storms, which originate
as a rule over central North America, are rainy and sometimes extensive, but
they do not as a rule not usually cause extensive widespread gales. In winter , however most cy–
clones of this group give gales over wide areas, and not a few produce vio–
lent storms in the turbulent seas around the southern tip of Greenland.
Widespread rain or drizzle accompanies the summer storms, usually with ex–
tensive stratus sheets over the seas and onshore coasts. Thunderstorms
may occur in Ungava and Labrador. Winter cyclones of the group bring con–
-siderable snow to Ungava, Labrador and southern Greenland, but blowing
snow may be widespread over the entire eastern Arctic (see below).
The remaining frontal groups affect the Arctic chiefly indirect–
ly. Most of the cyclones which reach Alaska are formed on the Pacific Po–
lar Front, and are the Pacific analogues of the systems dealt with in the
preceding paragraph. The storms of the European Arctic frontal group
rarely or never affect the American Arctic, though they give a violently
stormy climate to the Greenland s , Barents and Laptev Seas. Minor fronto–
genesis may occur between ice-cap air and mP air over the warm seas around
Greenland: such fronts are sometimes associated, for example, with severe
deepening of old Atlantic cyclones as they move up the Davis Strait or
Denmark Strait. But these influences are small by comparison with the
greater systems previously discussed.
A Regional Division of the Arctic Northland on Dynamical Grounds
The foregoing account will have made clear that real differences
in dynamic climate exist between various parts of the American Arctic northland . It
remains to emphasise these differences, for they are fundamentally signi–
ficant in the physical climatology. We can distinguish these four regions:-
(i) the Pacific northwest , including those parts of British Col–
umbia and Alaska which lie west or south of the Coast, St. Elias and Alaskan
Ranges, together with the Aleutians and southwesternmost mainland districts
of Alaska. These areas are united in being exposed throughout the year to
Pacific maritime airmasses and to the full fury of the Pacific cyclones
and their associated fronts. Because of the mildness of their winters, it
may be doubted whether we can even call them sub-Arctic. Their nearest
equivalent climatically is the Atlantic coast of Norway.
(ii) the Western Arctic , including the rest of Alaska, the Yukon,
the Mackenzie Valley and a part of the mainland between the Mackenzie and
Hudson’s Bay. Dynamically, we can also include in this region much of the
Alaskan-Siberian sectors of the Arctic Ocean, including the western third
of the Canadian Archipelago. This huge territory is characterized by a re–
latively quiet climate. Of the two main streams of cyclones, only one,
the American Arctic Frontal group, affects the region and these storms are
normally in a relatively immature condition. Moreover the region is very
apt to be covered by extensive high pressure systems, especially in winter.
(iii) the Eastern Arctic , comprising the eastern two-thirds of
the Archipelago, the Hudson’s Bay coastlands and Ungava-Labrador. The di–
viding line between the Eastern and Western Arctic is hard to define on
dynamical grounds; it lies nor far from the 100th or 105th degree west.
This region is characterised dynamically by the disturbed character of
the weather; the two main streams of cyclones converge on the eastern Arc–
tic, which they reach at the peak of their intensity, especially in the
cooler season. Gales , blowing snow and high windchill are characteristic
of winter, and low temperatures of summer. Once again we can trace this
American region north and north-east into the Arctic Ocean; in fact the
100° W meridian, if traced beyond the pole as the 80° E. meridian, effect–
ively divides the whole Arctic into two halves: a stormy “east”, including
the European Arctic, and a relatively settled “west”, including the regions
between Siberia and Alaska. This great contrast is ultimately a reflection
of accessibility: warm, moist maritime air, the trigger which sets off the
cyclonic storms, can enter the Arctic along a broad front on the Atlantic
flank, but the Pacific is cut off by the mountains of Alaska and north–
east Siberia.
(iv) Greenland : this remarkable ice plateau absolutely demands
individual treatment. So little is known of its climate, and so much con–
troversy has surrounded it, that a detailed treatment here is unavoidable.
Greenland: the Enigma of Arctic Climatology
The central ice-cap of Greenland, occupying as it does the great
bulk of th at land-mass, has remained throughout the past hundred years a
virtual blank on the weather-map. Our knowledge of its climate rests upon
the reports of the handful of explorers who have crossed it, and on the
two or three brief periods of observation by icecap meteorological stations.
In place of real knowledge we have had a generation or more of heated con–
troversy, surrounding the theory of the so-called “glacial anticyclone”.
This controversy has recently (10) been likened to a debate between expl–
orers (who have seen for themselves) and absentee meteorologists (who
persist in ignoring the evidence provided by the explorers). This is a
gross oversimplification, but it does contain the seeds of the truth; the
controversy surrounds different interpretations of the evidence, that of
the explorers - which is essentially inexpert - and that of the meteorolo–
gists, not by any means all of whom are lacking in experience in Greenland.
Nearly all the explorers who have crossed or penetrated the ice–
cap refer (11) to a remarkable wind and weather regime. They speak of
strong and monotonously regular downslope winds blowing out from the cen–
tral core, and pouring out through the coastal fjords to the surrounding
seas. Peary (11) described the flow as being analogous to the flow of water
down a slope. In a sense these storms represent very large scale paral–
lels of the “Gletscherwinde” of the Alps.
In 1910 (12) the well-known geologist, W. H. Hobbs, presented
in its earliest form the theory of the glacial anticyclone, by which he explained not only the Greenland climate but also the general character–
istics of Pleistocene ice-cap climates. With added experience he pub–
lished a greatly expanded version of this theory in his book “The Glac–
ial Anticyclones” (13, 1926) and his views have remained substantially
unchanged ever since (14). Though he has been hotly attacked, Hobbs has
maintained his position with great tenacity; the ‘glacial anticyclone’ has
become an established part of the legend of Pleistocene geology.
We may summarise the main features of Hobbs’ theory as follows:
(i) An ice surface is a good reflector and a good radiator;
hence the air resting on the ice-cap should be subject
to intense radiative cooling which cannot be counteract–
ed by absorption of insolation.
(ii) The cooling must lead first to an accumulation of very
cold air in the central regions of the cap, and hence to
rising pressure. Finally the accumulated mass of cold
air must break out to the coast in a spectacular down–
slope surge or “stroph” of the glacial anticyclone. To
replace the outflowing air, general subsidence begins,
and at very high levels there are radially inflowing winds.
The subsiding air, however, is warmed thermodynamically,
and eventually destroys the cold of the central regions.
In a few days, the stroph is finished, and the ice-cap
settles down to a fresh period of refrigeration, after
which the cycle is repeated.
(iii) In effect, this view postulates that a fixed, permanent
anticyclone lies over the ice-cap; in it the normal clock–
wise circulation of the anticyclone is replaced by a gig–
(iii) contd.
-antic downhill, gravity-impelled divergence of the chilled
surface air. The presence of such an anticyclone over the
cap must plainly inhibit the passage of cyclones, and it is
a fundamental tenet of Hobbs’ theory that cyclones do not
penetrate more than a few miles inland. He has published
several papers (15) designed to prove this.
(iv) The alimentation of the ice-cap, however, presents a prob–
lem, for if cyclones do not cross it, from where does the
snow come to make good the large annual loss by ablation?
In Hobbs’ view, hoar-frost, now snow, feeds the Greenland
ice, as it did the Pleistocene sheets. The radiative cool–
ing of the subsiding air at the ice-surface causes the con–
densation of huge quantities of “hoar-frost”, enough to make
good all losses by ablation.
(v) Hobbs’ final point - and in some ways the most pregnant –
is that the “strophs” of the glacial anticyclone set off
the main frontal cyclones of the Atlantic; the ice-cap acts,
in fact, as the “North Pole of the winds”, to use his own
picturesque phrase. At all seasons the air over the ice–
cap is the coldest in the Arctic, and hence must replace
the polar cap itself as the source of true Arctic outbreaks.
All aspects of this remarkable theory have been hotly debated
for more than a generation. It is not too much to claim that Hobbs’ work
was directly responsible for the bulk of the exploration of the Greenland
interior carried out after the first world war. The University of Michi–
gan sent several expeditions to the west coast between 1926 and 1933 under the leadership of W. H. Hobbs himself and L. M. Gould. These expeditions set
up stations both on the coast and the ice-cap, carrying out an elaborate
series of surface and upper air observations. In 1930-31 Alfred Wegener
organised a remarkable excursion which maintained stations on both the flank s
as well as in mid-ice (Eismitte) for a calendar year. The Wegener results
provide by far the most important evidence yet to hand concerning the ice–
cap climate. In the same year an official British party established an
ice-cap station on the east slope, rather to the south of the Wegener sta–
tion. Thereafter there was a period of inactivity until the United States
established stations during 1944.
As a result of these studies the con c s ensus of meteorological op–
inion is that the glacial anticyclone is in large measure based on a con–
fusion of fact with deduction. To summarise the present position:-
(i) the radial wind system , the basis for Hobbs’ views, is more
complex than early accounts allowed. Stations established
on the slopes (16) reveal that the surface wind most common–
ly flows downslope, but by no means invariably so. At the
Eismitte station, where, according to Hobbs, the wind should
have been light, variable or easterly, winds were actually
strong. All stations showed periods of upslope flow, and
considerable variability was general. The University of
Michigan station on the west slope indicated that the out–
blowing winds might extend to 4 km. (13,000 ft.) or above.
The general view today is that the radial wind system is
a katabatic circulation; the cooled air flows down the slope
under gravity, eventually reaching sea-level along the num–
erous fjords with which the coast abounds. General subsid–
(i) contd.
-ence in the central regions of Greenland must occur to feed
the katabatic flow. In this sense the term “anticyclone”
is justified, since anticyclones are regions of subsiding
air. In every other sense, however, the term is a misnomer:
anticyclones are vast eddies expressed as disturbances of the
horizontal pressure field; winds blow clockwise along the is–
obars, not radially outwards from the centre.
Several writers have expressed the opinion that the kata–
batic winds blow only during periods of feeble general cir–
culation over the ice-cap. H. G. Dorsey (17), who operated
the U.S. ice-cap station during 1944, indicates that the
katabatic winds are absent if there is an appreciable upslope
component in the regional circulation. Much the same con–
clusion is drawn by Mirrlees (18). Both these writers pro–
duce evidence that exceptionally strong flow down the east
slope occurs in the rear quadrants of cyclones in the vici–
nity of Denmark Strait or Iceland; in other words, excess–
ively strong “strophs” of ice-cap air require the joint act–
ions of the katabatic flow and a parallel regional flow.
(ii) The alimentation of the cap by hoar-frost has been similarly
disputed. Matthes (19), in a closely argued critique of
Hobbs’ views, points out that the air subsiding into the
central regions of the Greenland “anticyclone” can contain
only negligible amounts of moisture; furthermore, it must
have been rendered relatively drier by the adiabatic warming
postulated by Hobbs. The present writer has examined the
climatological records of stations in the Yukon and Macken–
(ii) contd.
-zie Valleys during prolonged winter spells of quiet, clear
anticyclonic weather, with marked radiative cooling. Though
many instances can be found in which such conditions persis–
ted for three or more consecutive weeks, there is no record
of measurable precipitation in the form of hoar-frost from
any station.
The condensation of rime (liquid cloud or cloud droplets
frozen onto solid surfaces) is a well known phenomenon in
moist airmasses. Mount Washington Observatory affords num–
erous striking and photogenic examples every year, and the
process has actually been observed in Spitzbergen over ice
surfaces. It is very possible that rime accretion is appre–
ciable on the lower slopes of the Greenland cap. Hobbs,
however, has specifically ruled that moist maritime air does
not penetrate more than a few miles inland, and rime cannot
possibly form from the subsiding air of the ice-cap.
On the other hand, there is abundant evidence that there
is frequent heavy snow on the ice-cap. Georgi (20), obser–
ver at Eismitte, was almost overwhelmed by the heavy snow
of August 1930, when he established the station; Courtauld,
hero of the celebrated six-month vigil on the ice-cap of
the British Arctic Air Route Expedition, was actually buried,
and had to be rescued through the ventilating shaft of his
quarters. Moreover, the periods of snow fell during spells
of upslope winds, and were preceded by cloud sequences sim–
ilar to those observed at lower levels during frontal pas–
sages. It is difficult to believe that experienced obser–
(ii) contd.
-vers like Georgi and Courtauld could have confused snow and
Most of these observers concluded that cyclones could and
did cross the ice-cap, and that the snow was normal cyclonic
snow. Dorsey, however, (17) felt that it was unrealistic to
talk of sea-level pressure systems crossing a 10,000 ft. bar–
rier; he suggested, with convincing evidence, that it is the
higher level perturbation that overlies a surface cyclone
that crosses the ice-cap. This is the mechanism, for exam–
ple, by which Pacific cyclones appear to enter central North
America across the Western Cordillera. Dorsey argued that
it was the height of the barrier that blocked the free move–
ment of cyclones, and not the presence of a fixed anticycl–
one. A similar blocking action is exerted by many mountain
systems such as the Alps, Pyrenees and Caucasus.
As inspection of the remarkable observational diagram
drawn up by Georgi (20) makes it impossible to doubt the
force of these objections. The record shows the rapid fluc–
tuations of pressure, temperature the wind velocity typical
of a disturbed cyclonic climate. To quote Georgi (21)*: …
…. “the wind regime over the ice-cap is much more complic–
ated than the simple model of the Glacial Anticyclone allows”..
…. To a reader of Georgi’s diary, this seems a conservative
(iii) Dorsey (17) has also pointed out the basic fallacy in Hobbs’
argument that the Greenland cap is the source of the cold
waves which touch off the Northern hemisphere’s chief cycl–
onic storms. It is true, he admits, that the ice-cap air
is the coldest in the hemisphere on a year-round basis. Be–
cause of its high level, however, it possesses great reser–
ves of the potential energy of gravity, which is realised
when the air comes down to low levels. Actually ice-cap air
enters the circulation of the Atlantic very much warmer than
the typical continental polar airmasses of Canadian proven–
In sum, then, the idealised model of a glacial anticyclone ap–
pears quite inadequate to account for the observed facts; the theory has
elements of the truth, but in general goes too far in deducing mechanisms
that do not exists. Its greatest value has been in the interest the theory
has aroused in Greenland meteorology, which is even now being explored more
Physical Climatology of the Arctic and Sub-Arctic
We have now examined the dynamic side of Arctic climatology –
the side concerned with processes, explanations, inter-relations and eq–
uilibria. The main bulk of the subject remains untouched - that which
the layman understands as climatology, the plain facts of temperature,
humidity, precipitation and cloudiness. Physical climatology is essential–
ly a secondary science; it deals with results rather than causes events
rather than eventualities. Nevertheless it is the aspect of climatology
which brings us closest to grips with climate, for it deals with quanti–
ties directly sensible to the human body.
The Temperature Regime
The Mean Temperature Distributions
Since the Arctic is defined in terms of cold, it is appropriate
that we should begin with a review of the temperature regime. Figs. 5-6 4-5
give mean air temperature for January and July - midwinter and midsummer
- and will serve as a basis for our earlier discussion. It should be
stressed that the fragmentary nature of Arctic climatological information
makes these maps in some degree conjectural. This is especially true of
the interior of Greenland, and also of the hilly areas of the Western
Cordillera. The remarkable inversion of temperature normally found over
the western Arctic in winter renders the run of isotherms in hilly coun–
try difficult to interpolate. It is highly probable, for example, that
an isolated hill 4,000 or 5,000 feet high would be warmer at the top than
at the foot.*
5 Fig. 4 Mean Air Temperature at ground-level, January (°F.)
Note the great cold over the central Arctic and the warm gulf over the
open water belts in Baffin Bay, Hudson Strait and Davis Strait. Note
also the remarkable contrast in temperature between coastal Alaska
and the Yukon-Tanana trenches. Fig. 5 Mean Air Temperature at ground level in July
The outstanding characteristic is the contrast between the cold eastern
Arctic and the warm interior of the sub-Artic Mackenzie, Yukon and
Tanana Valleys.
January is the depth of winter, and the temperature pattern for
the month may be taken as typical of the whole season, though naturally
the intensity of the cold varies from month to month. The main features
of the map are these:-
(i) Over the Arctic pack and the northernmost islands mean tem–
peratures are generally apparently a little below –30°F., though naturally much of
this expanse is little known climatologically. From the nucleus of cold
two broad cold wedges extend southward. That over Greenland is essent–
ially an effect of altitude and of the excessive radiation from the ice–
cap. That over North America, however, represents a winter extension of
the Arctic deep into the continental interior over ground of little alti–
tude. It reflects both the normal tendency for great winter cold in con–
tinental areas and also the easy access for Arctic airmasses offered by
the flatness of the terrain.
Great cold extends over the Yukon and Tanana Valleys, the Arctic
Coast Plain of Alaska and the entire Canadian Arctic; conditions are a
little warmer in Labrador and southern Baffin Island, but in most areas
temperatures lie below –20° F. Though the evidence is very flimsy, we
may hazard a guess that the coldest area on the mainland lies in north–
eastern Keewatin around and to the north of Baker Lake; average tempera–
tures of below –30° F. * may eventually be confirmed here. The recent cy–
cle of mild winters has made the short-period averages of the newer Arc–
tic weather stations rather misleading, and the lines shown on the map
have been drawn with careful attention to the possibility that recent records may prove untypical.
attempt a longer view than is possible on the existing statistics.
The warmer conditions of Baffin’s southeast district and of
coastal Labrador reflect the presence nearby of open water, and the oc–
casional advection of Atlantic mP air. The remarkable gulf of warmth off
6 the west coast of Greenland has a similar origin; temperatures in the
Davis Strait – Baffin Bay region are some 25° 35° warmer on the east than
the west (Baffin) shore. A much smaller “gulf” extends westwards along
Hudson Strait, but does not extend west of Southampton Island.
Hudson’s Bay exports little influence in mid- and late winter.
It was formerly supposed that the Bay remained largely open throughout
the winter (22), but recent evidence suggests that the area of open water
becomes negligible by early January, and that the Bay pack does not break
up until late in May. The fall maps of temperature show that the Bay
markedly amolierates temperatures, but the influence ceases by the end
of the year. Despite onshore winds, Port Harrison, Que. has the same
January mean temperature as Churchill, Man. in the same latitude. This
could not be the case if there were open water in between; it is incon–
ceivable that temperatures of –19° F (the actual means) could exist in an
airmass that has crossed 600 miles of water at a temperature of 31°F.
Flights across the Bay in the spring of 1948 and the winter of 1948-49 revealed a largely unbroken
pack . , criss-crossed by large pressure ridges, signifying a considerable thickness.
The Pacific coast of Alaska has a radically different winter re–
gime. Though mean temperatures are everywhere below 32° F (except loc–
ally in the Aleutians), no part of the coast has really severe weather,
and the relative warmth extends to the divide between the Coast and the
Tanana-Yukon trench. As we have already said, this warmth is due to the
almost continuous advection of mP air, and not to the presence of unus–
ually warm water offshore.
The role of the western mountains is hard to decipher, and the
isotherms are dashed where they run along the main hill axes.
The July map (fig. 6 5 ) is typical of the summer distributions.
It at once reveals the fundamental division of the Arctic into east and
west that is so often repeated in many fields. The west is conspicuous–
ly warm; in fact true Arctic conditions are confined to a narrow strip
along the Arctic coast. The Archipelago, however, and the whole eastern
Arctic are truly Arctic at this season: cool, damp and hostile to most
forms of exploitation. The 50° F. isotherm runs north from the Aleutians,
then follows a course along the Brooks Range to the Mackenzie Delta, and
thence to Coppermine. There it begins its southward movement, passing
just north of Baker Lake and Eskimo Point before descending to latitude
56° N over the Bay. Over Ungava it swings north again, leaving, however,
the region north of the Payne River to the north. It passes out to sea
somewhere south of Hebron, Labrador, but then runs parallel to the coast
to pass south of Belle Isle.
The coldness of the eastern Arctic and the Archipelago is direct–
ly caused by the presence of a huge area of cold water surfaces, which
chill all airmasses crossing them. The break-up of the ice in Hudson’s
Bay, the channels in the Archipelago and in Baffin Bay in May, June and
July, keeps a steady stream of loose, disintegrating floes moving south
and east out into the Atlantic and down the Labrador Coast. Even in the
late summer, when many of the seas and channels clear fairly completely,
recent ice-melt waters keep sea-surface temperatures very low. Over much
of the Eastern Arctic the sea does not warm up beyond 40° F., while in
the channels of the far north it can rarely exceed 35° F. East coastal
Greenland is also washed by extremely cold water streaming southwestwards.
The result is apparent on fig. 6 5 . Air temperatures over the
archipelago average close to 40° F. and are uniform over wide areas. The most outstandingly cool region is Hudson and Davis Straits, where Reso–
lution Island has the astonishingly low mean for July of 38° F., though
well south of the Arctic Circle. The east coast of Hudson’s Bay and the
Atlantic coast of Labrador are among the coldest places on earth for
their latitude at this time of year. How quickly the sun can dissipate
this cold in the interior is very visible in Labrador, where the coastal
belt of chill is only a few tens of miles wide.
Since almost all Eastern Arctic stations are coastal, it is hard
to commit oneself as to conditions in the uninhabited interiors. Ungava
offers us proof that conditions are far warmer inland, but observations
from more truly Arctic lands are lacking. Nordenskjold (2) has commented
upon the Holsteinborg region of western Greenland, where a wide area of
land separates the ice-cap from the sea. He comments remarks that the tempera–
ture seems to increase in summer as one goes inland towards the ice. It
is highly probable that a similar increase occurs in such large land–
masses as Baffin and Victoria Islands, though the effect may be lessened
by the permafrost and the waterlogged soil.
The march of the seasons can perhaps best be described by refer–
ence to the onset of freezing conditions in the fall and of persistent
thaw in spring. The best indicator of such conditions is the 32° F. iso–
therm for mean daily temperature; though thaws may occur a month or more
in spring before the long period mean temperature rises above 32° F., deep
thaw is deferred until this temperature is passed. Similarly in the fall
deep penetration of frost may be expected to begin only after mean temp–
erature falls below freezing.
Fig. 7 6 shows the onset of spring thaw. By mid-March thaw has
affected only the southernmost part of the Alaskan coasts. A month la–
ter, by April 15th, persistent thaw has normally extended to the heads
of the southern Alaskan valleys, and is entering the Yukon. The upper
Liard and Peace Valleys experience the thaw a little later. It penet–
rates into the Tanana and Yukon Valleys by April 30th, and by May 5th
has extended to all but the northernmost parts of the Mackenzie Basin.
By May 20th, all the western Arctic south of the Brooks Range and the
Arctic Coast has normally cleared.
The Eastern Canadian Arctic and the Archipelago experience the
thaw much more tardily. By May 1st. it has entered Labrador and south–
central Quebec and has reached the south shore of James Bay. By mid- May
it has begun to move into Hudson’s Bay, though Churchill and Port Harri–
son are still frost-bound. In Ungava thaw spreads to Ungava Bay by the
15th. Thereafter the northward spread is more rapid and highly erratic.
Over northern Hudson’s Bay and most of the Archipelago persistent thaw
waits until the end of May or early June, and in the northernmost islands
may be deferred into mid-June or later.
In Greenland thaw comes to the south much earlier. By mid-May
it has almost reached Disko Island and is well north of Angmagssalik.
The northern coasts, however, are much slower, and in Peary Land thaw
does not occur until late in June.
The southward march of persistent autumn freeze-up is also best
indicated by the 32° F. mean isotherm (see fig. 8 7 ). By September 1st.,
frost is renewed in the northernmost regions, and by the 15th, northern
Baffin Land and almost all the Archipelago are affected: the Greenland
coast, however, is some two or three weeks later. A month later (Oct. Fig. 6 The Northward March of Persistent Spring Thaw
The isolines show the dates on which mean air temperature rises above
32°F. Note the retarded spring around ice-covered Hudson Bay. Fig. 7 The Southward March of persistent frost in fall
Note how far south the freeze-up goes by mid-October in Ungava-Labrador. 15th) the freeze-up has extended across all interior Alaska and the Mack–
enzie Basin north of Lake Athabaska, has penetrated south of Churchill,
and is south of 55° N. in the plateau of Labrador. By the end of the
month all parts of the sub-Arctic have begun the winter frost.
In most parts of the north there is no true summer frost-free
season: frosts can and usually do occur in each month of the year. For
possible agricultural development this frost-free season is of import–
ance, as it defines the effective season of growth. In certain favoured
parts of the northwest, the period free from soil frost is long enough
to permit outdoor growth of certain economic plants and the growth of
good grass pastures in favourable localities:-
(i) in the Mackenzie and Liard Valleys there is an area around
Fort Simpson and the lower Liard valley with an average frost-free seas–
on (viz. days between the last spring frost and the first fall frost) of
80 - 90 days (9), comparable with that of the Peace River Territory. The
frost-free season diminishes rapidly eastwards as well as northwards; cer–
tain parts of the pioneer areas of Northern Saskatchewan have a shorter
frost-free season, as does all the rest of the Mackenzie Valley, even up–
stream from Great Slave Lake. It should be noted that this relatively
favoured area has, however, an erratic frost regime: at Fort Simpson the
frost-free season has been as long as 120 days and as short as 42. Frosts
may occur as late as July.
(ii) A similar favoured area occurs in the upper Yukon Valley
(Dawson 74 days) and in the Tanana Valley, where a frost-free season of
90 days is available in places. Here also the risk of late frost extends
into July.
Diurnal and Inter-Diurnal Variations
We may quickly disabuse ourselves of the idea that temperatures
are in any sense invariable under Arctic conditions. Both cyclical and
non-periodic variations are significant.
The true diurnal variation in temperature is the cyclical varia–
tion depending on the radiative inequality between day and night. We
might expect that such variations would be important in the Arctic at the
equinoxes, and much reduced at the solstices, when day and night are of
very unequal extent. Table II gives mean deviations from the daily mean
for three Arctic Stations during the second Polar Year (1932-33).
Table II. The Amplitude of the Diurnal Variation of Temperature at Cop
permine, Chesterfield Inlet and Cape Hope’s Advance, 1932-3
(Degrees Fahrenheit)
N.B. The amplitude is the difference between the highest and
lowest mean hourly values of temperature, and indicate s the
magnitude of the cyclical diurnal changes; non-periodic chan–
ges are eliminated.
Jan. Feb. Mar. Apr. May June July Aug. Sept Oct. Nov. Dec.
Coppermine 1.4 3.2 8.1 14.9 6.5 6.2 8.0 9.4 6.8 4.2 1.2 2.2
Chesterfield Inlet 1.3 4.5 8.1 8.2 9.2 5.6 8.4 6.4 4.2 2.2 1.4 1.5
Cape Hope’s Advance 2.5 2.3 3.3 3.7 4.6 4.5 5.3 3.2 2.5 1.7 0.9 1.3
Authority: Canadian Polar Year Expeditions, 1932-3,
Report Vol. 1, 1940.
These figures by no means confirm out speculation regarding
annual variation. The Coppermine figures represent the true high latitude conditions [: ] , as
this station alone lies north of the Circle. There is a strong suggest–
ion here of solar control; there is a decided maximum (approx. 15° F.) in
mid-April (about three weeks after the equinox) and a secondary maximum
in mix-August (five to six weeks before the equinox). Chesterfield and
Cape Hope’s Advance, however, appear to show merely a pronounced winter minimum and early summer maximum. Little or no trace of equinoctial
maxima can be detected.
It is noteworthy that all three stations show little true diur–
nal variation in winter. Oscillations of winter temperature in the north
are due to advection rather than to any clear-cut diurnal radiative
rhythm; temperature maxima and minima can hence come at any time, showing
no systematic preference for a particular hour.
Non-periodic changes in temperature of dynamic origin are hard
to disentangle from the periodic, but we can get some idea of their mag–
nitude by considering the difference between the mean amplitude of the
cyclical variation and the mean daily range of temperature (which is the
difference between the mean daily maximum and minimum temperatures).
Table III gives representative figures for a number of northern stations.
Table III. Mean Daily Range of Temperature at certain northern stations
(Degrees Fahrenheit)
J F M A M J J A S O N D Yr
oppermine 14 14 15 17 15 14 16 13 10 11 14 14 14
t.Simpson 17 21 24 24 23 23 24 22 20 15 14 16 20
hester.I. 13 12 14 16 12 12 16 13 9 9 14 13 13
ond Inlet 13 13 16 18 16 11 12 10 7 10 11 13 13
p.Hp’s.A. 9 11 10 11 9 9 12 11 8 6 10 10 10
Authority: “Meteorology of the Canadian Arctic”
Canadian Meteorological Service, 1944.
“Climatic Summaries”,
Canadian Meteorological Service.
The outstanding interest of Table III is that it demonstrates
how widely variable winter temperatures are; in spite of the negligible
diurnal variation at this season, there is an average non-cyclical var–
iation of sufficient magnitude to give daily ranges of over 10° F. In
summer the daily range is also much greater than the diurnal amplitude, though at this season the inequality is less marked. It is clear, however,
that advection and other dynamic processes are active throughout the year
in creating variability of temperature in the north.
The interdiurnal variations are also considerable throughout the
year. Though the actual figures have not been computed for all stations,
the mean difference between temperatures at the same hour on successive
days appears to average from 6° to 10° F. in all parts of the Arctic
throughout the year: the only area and time where small variations occur
seem to be the Hudson’s Strait area in summer. These differences must re–
flect the effect of importation of airmasses of differing temperatures;
they cannot have a local radiative origin.
Year-to-year Variability
In yet another respect the Arctic climates exhibit erratic ten–
dencies. One year may differ strikingly from the next, the present decade
from the immediately past, and so on. These non-periodic tendencies show
themselves especially well in such areas as the Yukon, where it is not un–
common for a winter month to differ from the monthly mean temperature by
10° F. The Yukon, Tanana and to some extent the Peace River districts lie
close to the mean position of the American Arctic frontal zone, and a per–
sistent displacement of this front may play havoc with mean temperatures.
The effect, however, is by no means restricted to the western Arctic; Table
IV will prove this.
Table IV. Successive yearly values of mean daily maximum temperatures at
Pond s Inlet, illustrating year to year variability (°Fahrenheit)
January April July October
1931 –20.7 –3.4 53.2 23.3
1932 –26.8 3.0 44.8 21.5
1933 –24.5 1.7 45.6 22.5
1934 –18.2 2.3 47.5 x
1935 [: x] x [: x] x
1936 –15.3 x 49.8 13.9
1937 –21.3 –4.0 49.4 17.0
1938 –19.5 –4.5 48.9 22.2
1939 –9.6 8.5 47.5 17.7
1940 –3.1 7.4 46.8 22.7
1941 –6.3 [: x] 45.0 [: x]
The station, Pond ’s Inlet, is in northern Baffin Land, and the figures
make it clear that even in these latitudes the year-to-year differences
are striking. This is true more of the ice season than the thaw: the ub–
iquitous ice melt water stabilises summer temperatures to a large extent.
These figures emphasise the extreme danger of taking a single
year’s observations at an Arctic Station too seriously. One may well be
10° F. in error . if one uses them to estimate a mean.
Temperature Extremes
Popular beliefs about the northern climates are largely fed on
stories of extreme cold; naturally enough such stories find common accep–
tance, for they are often quoted by otherwise reliable authorities. There
is another school of Arctic enthusiasts, however, who do their best to
hide the grim realities of the climate by claiming, for example, that the
Dakotas are colder than the Canadian Arctic Archipelago - on the dubious
grounds that temperatures of –60° F. have been recorded in those states,
but not in the Archipelago! [: ] Neither approach helps the layman to get an
adequate picture of reality.
The lowest temperatures recorded on earth at ground-level have
come from the mountain valleys of north-east Siberia. Verkhoyansk, with
–93° F., and Oimekon *, with –90° F., have afforded the lowest available
readings to date. Readings almost comparable with these have been record–
ed in the Mackenzie Valley and Yukon region, but not further north or east.
The lowest authenticated record + is of –81.5° F. in the winter of 1946-47
at Snag in the Yukon; readings of –70° F. have occurred in many western
Arctic areas. It is interesting to compare the records of the Yukon sta–
7 -tions with those of the Siberian “cold pole”. Table V compares Verk–
hoyansk, Oimekon, Dawson City and Chesterfield Inlet. It is seen at once
that at Siberian Arctic is much colder, even though the observed ex–
tremes are not too different.
Table V. A comparison between mean monthly temperatures in the Siberian
and Canadian sub-Arctic (°F)
Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Yr
Verkhoyansk –58 –48 –22 + 8 + 35 + 55 + 59 + 51 + 36 + 6 –34 –52 + 3
Oimekon –61 –53 –24 + 2 + 34 + 55 + 60 + 53 + 37 + 3 –32 –53 + 2
Dawson City –21 –12 –4 + 29 + 46 + 57 + 60 + 55 + 42 + 26 + 1 –14 + 23
Chesterfield –27 –25 –16 + 2 + 21 + 37 + 49 + 47 + 38 + 22 – 2 –17 + 11
Authority: W. B. Schostakovitch: The climate of Verkhoyansk (trans.
from Russian manuscript)
G. Salischew: The cold pole of the earth, Geog. Rev.,
1935, pp. 684-685
H. Flohn: Zum Klima der Kältepole der Erde. Met. Rund–
schau, Jahr. 1, Heft 1/2, 1947, pp.25-26.
It is noteworthy that most of the very low temperatures report–
ed from North America come from the Yukon and the lower Mackenzie Valley,
where there are many deeply enclosed valleys. In quiet, anticyclonic
weather cold air drains down by gravity to fill these valleys with a very
dense, calm layer of frigid air. Exactly similar conditions occur in the
vicinity of Verkhoyansk and Oimekon, though there the continuity and dur–
ation are much greater. The high overcast which so often cuts off the sky
and stops radiative cooling in the Yukon is not common in the Siberian
Arctic. The cold in both regions is accompanied by calm or light winds,
which reduces its effect on exposed beings.
The eastern Arctic is very different. Here extreme cold is less
common. Temperatures of below –60° F. are unusual, and tend to occur, paradoxically enough, on the mainland relatively far south, especially in
northern Ontario and the plateau of central Quebec. The less severe cold
of the east may be attributed in part to the escape of heat through sea
or lake ice and through the occasional leads and open patches of water.
More probably, however, the greater vigour of the circulation is primar–
ily responsible: winds are usually strong, and the opportunity for very
prolonged radiative cooling does not present itself.
The strong winds, though they reduce the absolute value of the
cold, greatly increase its penetrating power; the windchill, that is, is
far greater than in the west on the same mean isotherm. It is not uncom–
mon to experience temperatures below –40° F. with windspeed in excess of
30 m.p.h. for periods of 48 hours or over. Such cold is far more impres–
sive as an environmental factor than are the very low temperatures of the
Yukon. The stormy eastern American Arctic climate may be claimed as the
severest in the northern hemisphere; elsewhere it is rivalled only by the
marginal Antarctic climates, where winds are even stronger and temperat–
ures lower.
Precipitation, Humidity and Cloudiness
After low temperatures, snow is the element popularly associat–
ed with the Arctic. Yet paradoxically enough really heavy snow is exp–
erienced only outside the true Arctic, especially in hill countries like
the Cascades and Pacific ranges generally and the mountains of Japan. Even
in lowland areas the heaviest snowfall in North America lies in the St.
Lawrence region rather than in the Arctic. Actually the northland is
markedly deficient in all forms of precipitation: drought is very common
and prolonged in all except a few small areas.
The reason for the paucity of Arctic precipitation is the low
moisture capacity of cold airmasses. Though relative humidities are us–
ually quite high in Arctic regions, the moisture content of cold air even
at saturation is very small. Only in summer is there an appreciable quan–
tity of water vapour in most regions. Actually humidity statistics have
little real value under such conditions, and they will only be discussed
here incidentally to precipitation.
We may discuss the precipitation regime of the northlands under
the region division described on p.
(i) The Pacific North-West is anomalous in precipitation charac–
teristics as in temperature. Its exposure to moist mP air from the Paci–
fic gives it a very heavy precipitation indeed. The heaviest falls occur
in the eastern Panhandle of Alaska (Ketchikan, 161.5 ins. per year), but the entire
coast from the Kenai Peninsula eastward is wet. Westwards from Cook Inlet
precipitation amounts decrease steadily to about 40 ins. in the drier
Special attention should be called, however, to the broad low–
land at the head of Cook Inlet; sheltered by the Kenai range from southerly (i) contd.
rain-bearing airstreams, the region has a remarkably low total fall (Mat–
anuska, 13.9 ins.) largely concentrated in summer season.
Almost the entire coast has an autumnal rainfall maximum, fol–
lowed by a rainy or snowy winter. Spring and summer are relatively dry,
though at no time is rainfall inconsiderable. Almost everywhere Septem–
ber and October are the wettest months; at this time of the year the air
is still relatively warm and moist, and when involved in the renewed vig–
our of fall cyclones, yields heavy rain, or snow at higher levels.
Winter precipitation occurs chiefly as wet snow except early and
late in the season in the south when rain is usual. Total snowfall ex–
ceeds 100 inches per annum in all exposed localities from Kenai eastwards
to Cross Sound, and at 3,000-4,000 feet on exposed slopes probably ex–
ceeds 250 ins., especially in the St. Elias Ranges. At altitudes above
about 6,000 feet, considerable snow also falls in summer. The great ex–
tent of Alaskan glaciers affords testimony as to the heaviness of the
(ii) the western Arctic . This region, we may remind ourselves,
extends from the Bering Sea, eastwards across Alaska north of the Gulf of
Alaska drainage slope, across the Yukon and Mackenzie Basins to an indet–
erminate eastern boundary not far from the 100th or 105th meridian.
Characteristic of this entire region are (i) a very scanty and
unreliable winter snowfall and (ii) a light summer rainfall punctuated by
considerable periods of drought.
Over most of the region winter snow accompanies the advection
inland of mP (Pacific) air at high levels. Total amounts are rarely heavy,
as such air loses a large part of its water vapour over the Pacific Ranges.
Total falls are largest in western Alaska, and least in the Arctic Coast Plain of Alaska. Inland maximum snowfall occurs in mid-winter, but in
coastal situations a fall maximum is all but universal. Total annual
falls range from 70-80 ins. in parts of western Alaska down to 35 ins.
on the Arctic Coast Plain and along the Canadian Arctic coast , ; 50 ins.
is a very typical fall over much of this region (cf. Montreal, 105 ins.).
Summer rains, in the western Arctic occur infrequently, and are
of only moderate duration and intensity. Along the Bering Sea coast July
and August rains exceed 3″ per month in places. Further inland the Tan–
ana valley also gets an appreciable fall (Fairbanks, July, 1.9 ins.,) Tan–
ana , July, 2.4 ins.) as does the Canadian Yukon (Dawson, August, 1.6 ins.)
the source of this rain being the cyclones which frequently travel in
from the Bering Sea at this time. The Arctic Coast plain and the Yukon
Flats, however, are dry even in summer, and are normally subject to pro–
longed drought (Barrow, July, 1.1 ins., Fort Yukon, July 1.1 ins.).
East of the Yukon, the source of summer rain is apt to be cPW
air from the southeast; here again we see relatively considerable falls
in the south (Fort Simpson, 2.0 ins. July) and lighter falls in the north
(Coppermine, 1.8 ins., August, Aklavik, 1.4 ins., July).
(iii) the Eastern Arctic , as defined previously, includes the
mainland and Archipelago east of about 100° - 105° W., but excluding Green–
land. For the purposes of the present discussion, however, we may include
coastal Greenland, though the ice-cap deserves special treatment. This
entire territory is fed primarily by moisture derived ultimately from At–
lantic sources.
The southern half of this huge region has an appreciable prec–
ipitation at all seasons, though a summer maximum is universal. From (iii) contd.
Churchill northeast towards Disko Island, r R egions to the southeast of a line form Churchill to Disko Island have
over 15 ins. per annum, while in pars of Labrador annual falls exceed
30 ins. The wettest region is southwest Greenland (Ivigtut, 44 ins.).
Northwest of the line mentioned, however, total falls decrease rapidly,
and over much of the Archipelago the fall does not exceed 10 ins., and may
be as low as 5 ins. per annum. Thus severe drought maybe reckoned as
normal in the true Arctic here as in the west.
Winter snowfall is considerable in the southeast. In certain
parts of Labrador, southeastern Baffin and southwest Greenland, it may am–
ount to over 100 ins. per year, falling between September and May. A
pronounced early fall maximum is usual, due to convection in cP and cA
air over still-open water surfaces; heavy snow-squalls are characteris–
tic of the early fall. Further north , however, snowfall decreases rapid–
ly, and is below 50 ins. per annum over most of the Archipelago. An Oct–
ober maximum is again very general, followed by a minor maximum in April
and May. Mid-winter is normally largely snow-free.
(iv) The Greenland Ice-Cap requires special treatment; unfort–
unately the almost complete lack of information makes it impossible to
deal with it adequately. M. Demorest (24) has assembled evidence to show
that the snow falls most heavily along the flanks of the cap rather than
in the centre, and that in general falls most heavily on the western and
southern flanks. R. F. Flint (25) has pointed out that the northern part
of the Cap must be less heavily snowed-upon than the south, and visual–
ises the maintenance of the northern ice by movement from the snowier
south. E. Sorge (26) has recorded that the annual increment to the firn
at Eismitte was about 39 ins. [: ] , and that the maximum increment (about 90 (iv) contd.
miles from the west coast) was about 65 inches [: ] . Since these figures rep–
resent the snowfall after evaporation, packing and wind-drifting, it is
apparent that the local snowfall even in the centre of the cap is very
considerable. Beyond this, however, we cannot go.
Blowing Snow
Though not strictly a form of precipitation, blowing snow is so
much a feature of the Arctic winter that we must give it some attention.
Much of the snow of the Arctic falls in the form of loose, uni-crystalline
or granular flakes which are very readily whipped up by a light wind. Ob–
servation shows that with winds in excess of 15 m.p.h. blowing snow is
nearly always present in some degree in the cold Arctic. The degree of
stability in the air (and hence of turbulence) and the characteristics of
the grain structure have also some influence on the incidence of blowing
snow, but as yet its mechanics remain little studied.
The hindrance caused by blowing snow to air and land communica–
tions can readily be imagined. The Musk-Ox force, for example, encoun–
tered visibilities of 1,000 yds. or below on 19 out of the 45 days in
which they were moving across the windy, snow-covered barrens between
Churchill and the Arctic coast. Visibilities were on one occasion reduc–
ed to below 25 yards for a period of 24 hours. Similar troubles were ex–
perienced by the Wegener observers on the Greenland cap, and by the Brit–
ish Arctic Air Route party in the same year. Table IV shows how extensive
was the effect of blowing snow at the latter party’s station.
Table VI over
Table IV. Significance of Blowing Snow as a factor affecting visibility
at the British Ice-Cap station, Greenland (67° N., 42° W.,
8,200 ft.) Observations at 1300 hrs., 45th meridian time .
1930 1931
Sept. Oct. Nov. Dec. Jan. Feb. Mar.
% observations with
visibility less than
1,100 yards -
13 17 46 45* 54 54 14
% due to blowing snow 0 10 40 28 45 46 7
% due to snowfall 4 6 3 7 6 0 7
% due to fog 9 3 3 7 3 8 0
*: 3% not accounted for
Authority: S.T.A. Mirrlees: “Meteorological results of the B.A.A.R.E.,
1930-31”, Met. Office, London, Geophysical
Memoirs, Vol. VII, No. 61, 1934.
Thus in January and February blowing snow caused visibilities
within the fog range on nearly half the observations taken.
The efficiency of snow removal by the wind over the treeless
barrens is considerable, and much of the surface has a very thin and pat–
chy cover. Near obstacles, in hollows or around buildings, however, the
obstruction to the wind causes huge drifts to accumulate. Buildings are
usually all but buried for much of the winter. It will be realised that
under such conditions the measurement of fresh snowfall is a highly ap–
proximate technique.
Cloudiness and Fog
The Northland is normally a cloudy region, especially in sea
areas. Because of the high frequency of frontal and cyclonic activity,
medium- and high-level cloud systems are abundant. Furthermore, the high
stability and superficial dampness of many of the airmasses affecting the
region leads to the formation of thin but extensive layer clouds at low
In general winter is the least cloudy season; clear skies are
frequent in the Western Arctic and the Archipelago. In places mean
cloudiness falls to as low as 2/10 at this time of year. Greenland and
the southeast Canadian Arctic, however, remain cloudy because of the high
storm frequency in this region. So also does the south coast of Alaska,
for a similar reason.
In spring there is a general rise in cloudiness, especially in
coastal and sea areas. Spring, summer and early fall are cloudy seasons
everywhere. Inland districts of the Western Arctic normally enjoy rela–
tively cloud-free conditions until September or October, by which date
average cloudiness has normally exceeded 6/10. In the east, however,
and in coastal Alaska, the maximum cloudiness tends to come earlier.
Western Alaska is very cloudy indeed, as is the Bering Sea (St. Paul’s
Island has a mean cloudiness of 9.4/10 in July and August). In the Eastern
Arctic a dual maximum of cloudiness tends to occur, the first in May or
June, and the second in September or October. The cloudiness of early
summer is due to extensive stratus sheets created by cooling of moist
airmasses by melting ice. The fall maximum arises from the reverse pro–
cess when cold outbreaks travel southwards across the relatively warm sea
channels: ; extensive sheets of turbulent strato-cumulus are created.
The high cloudiness of summer is related to a similar prevalence
of sea-fog, one of the most characteristic weather phenomena of maritime
Arctic localities. Arctic sea fog is primarily advective, that is, it
arises from the chilling of warm, moist air being brought northwards ac–
ross seas whose surface water is recent ice-melt, or is possibly still
charged with disintegrating ice. Such fog may occur anywhere, though it
is commonest in those areas which are most accessible to mP air, viz. the Alaska Coast, the Greenland Coast and the Hudson Strait, Davis Strait, Ba–
ffin Bay area. Because of its mode of formation, it is normally a wide–
spread phenomenon of spring and summer, though in southern Greenland and
Alaska it may also occur in winter.
Sea-fog normally occurs with relatively light winds. Stronger
winds tend to lift the fog into stratus sheets, from which drizzle falls
at intervals. Visibility below such sheets remains very poor, and is al–
most as severe a hazard to flying as is the true fog.
Table VII gives frequencies of fog at selected localities. It
emphasises the concentration on summer, as well as the fog-free character
of most inland sites. Normal nocturnal radiation fogs are frequent only
in isolated areas; they occur, for example, in the Matanuska Valley of Al–
aska (see Anchorage in Table VII) where mP air is subject in winter to ex–
tensive radiative cooling. Ice-crystal fogs may also occur in cold anti–
cyclonic weather, especially near the larger settlements.
Table VII. Days with fog (visibility less than 1,100 yds.) at selected
stations in the American Arctic .
Jan. Feb. Mar Apr May Jun Jul Aug Sep Oct Nov Dec Year
Nome 5 5 5 4 5 11 8 6 4 4 3 7 67
Anchorage 8 4 3 2 <1 1 1 2 2 2 4 5 35
Fort Yukon <1 <1 0 0 0 0 0 0 <1 <1 <1 <1 3
Churchill 1 1 <1 1 1 4 2 2 1 <1 <1 0 13
Resolution Is. <1 1 <1 1 5 10 13 12 7 2 1 <1 55
Craig Harbour 0 <1 0 <1 <1 <1 2 2 <1 <1 <1 <1 7
Godthaab 1 1 1 3 7 10 13 13 7 2 1 1 61
Julianehaab 1 1 2 4 8 11 15 14 9 5 3 1 71
The review of the climates of the American Arctic and sub-Arctic
just presented is in fact far from completed: one cannot properly assess a
climatic whole by describing its parts. The task of describing the cold
climates is made doubly difficult by the lack of parameters specifically
created for the purpose; almost all our ordinary climatological methods
break down under Arctic conditions. Thus the account given conspicuously
fails in many respects to paint an adequate picture.
It is highly probable that our view of the climates of the North–
lands will be increasingly directed along two channels in the immediate
future: first, the physiological channel, in which the Arctic climate is
discussed primarily as an environment for living things; and secondly
the dynamic channel, where we shall be interested in the weather processes
of the north, and their relation to the regions further south.
Until these avenues have been more quantitatively explored, lit–
tle can be added to the account given hero. The great interest now being
directed towards the Arctic, however, makes it highly probable that much
more will be learned in the next few years.
The author’s thanks are due to the many friends who
have assisted him in the study; especially to Pat Baird, J. Bird,
P. Dansereau and N. Polunin for reading all or part of the
manuscript. He also has to acknowledge the debt he owes to
Capt. H. G. Dorsey, Jr., of the U.S. Army, much of whose material
and experience have been used in the preparation of the article.

(1) W. Köppen: “Grundriss der Klimak u ü nde”, Berlin, 1931.

(2) O. Nordenskjold: “Geography of the Polar Regions” (with L. Mecking),
Amer. Geog. Soc. Spec. Pub. no. 8, 1928, pp.72-86.
For Holsteinborg, see p.83.

(3) W.E.D. Halliday: “A Forest Classification for Canada”. Dept. of
Mines and Resources, Ottawa. Forest Service Bul–
letin 89, 1937.

(4) G.T. Trewartha: “Introduction to Weather and Climate”, New York,
1943 (2nd ed.) p. 472.

(5) T. Bergeron: “Richtlinien einer dynamischen Klimatologie”, Met.
Zeit., Vol. 47, 1930, pp. 247-262.

(6) V. Bjerknes and
collaborators: “Physikalische Hydrodynamik”, Berlin 1933. Reprin–
ted Edwards Bros., Ann Arbor, Michigan.

(7) T. Bergeron: “ U Ü ber die dreidimensional verknüpfende Wetteranalyse”
Geof. Publ. Vol. V, no. 6, Oslo, 1928.

(8) S. Petterssen:
H. R. Byers: See for example:
“Weather Analysis and Forecasting”, New York, 1940.
“General Meteorology”, New York, 1944.

(9) A. J. Connor: Meteorology, in “Canada’s New Northwest”, North Pa–
cific Planning Project, C. Camsell, Director, Ott–
awa, Ont., 1947, pp. 145-155.

(10) W. H. Hobbs: “The Climate of the Arctic as Viewed by the Ex–
plorer and the Meteorologist”, Science, Vol. 108,
1948, pp. 193-201.

(11) R. E. Peary: “Journeys in North Greenland”, Geog. Jour. Vol. 11,
1898, pp. 213-240.

(12) W. H. Hobbs: “Characteristics of the Inland Ice of the Arctic
Regions”, Proc. Amer. Phil. Soc., vol. 49, pp. 57-
129, 1910.

(13) W. H. Hobbs: “The Glacial Anticyclones”, New York 1926.

(14) W. H. Hobbs: “The Greenland Glacial Anticyclone”, Jour. of Met.
vol. 2, pp. 143-153, 1945.

(15) W. H. Hobbs: “Nourishment of the Greenland Continental Glacier”
Jour. Geol., Vol. 52, no. 2, pp. 73-96, 1944.
and W. H. Hobbs: (and R. L. Belknap) “Errors in the Synoptic Weather
Charts which cover the Greenland Region”, Trans.
Amer. Geoph. Union, Part 3, pp. 482-490, 1944.

References (contd.)

(16) K. Wegener: Wissenschaftliche Ergebnisse der deutschen Grön–
land Expedition Alfred Wegener 1929 und 1930-31,
Vol. 4, Part 1. See also ref. 18

(17) H. G. Dorsey: “Some Meteorological Aspects of the Greenland Ice–
Cap”, Jour. Met. Vol. 2, no. 3, Sept. 1945, pp.

(18) S.T.A. Miv r lees: “Meteorological Results of the B.A.A.R.E., 1930-31”
Air Ministry, London, Geoph. Mem. Vol. VII no. 61,

(19) F. Matthes: “The Glacial Anticyclone Theory examined in the
light of recent meteorological data from Greenland,
Part 1”, Trans. Amer. Geoph. Union, Vol. 27, 1946,
pp. 324-341.

(20) J. Georgi: “Im Eis Vergraben”, Munich, 1933, pp. 52-87.
The celebrated Georgi diagram is most accessible
in W. H. Hobbs: “Rhythm in the Greenland Glacial
Anticyclone”, Trans. Amer. Geoph. Union, 1944,
pp. 491-494.

(21) J. Georgi: Op. cit. sup. (20) p. 19.

(22) U. S. Navy: “Ice Atlas of the Northern Hemisphere”, Washington,
Hydrographic Office, Dept. of the Navy, 1946.

(23) K. A. Salischev: “The Cold Pole of the Earth”, Geog. Rev. Vol. 25,
1935, pp. 684-685.
see also: (23) H. Flohn: “Zum Klima der Kaltepole der Erde”, Met. Rundschau,
1947, pp. 25-26. Jahr. 1, Heft. 1/2.

(24) M. Demorest: “Ice Sheets”, Geol. Soc. Am. Bull. Vol. 54, pp. 363-

(25) R. F. Flint: “Glacial Geology and the Pleistocene Epoch”, New
York, 1947, p. 45.

(26) E. Sorge: “The Scientific Results of the Wegener Expeditions
to Greenland”, Geog. Jour. vol. 81, 1933, pp. 333-

Reference has also been made to the following stan–
dard reference works, whose value cannot be exaggerated:-

R. de C. Ward, C. F.
Brooks and A. J. Connor: “The Climates of North America”, Band II, Teil J.,
in “Handbuch der Klimatologie”, W. Köppen and R.
Geiger, ed. 1936.

References (contd)

H. [: ] U . Sverdrup: “Klima des Kanadischen Archipels und Grönland”,
Köppen and Geiger (ed.) op. cit. sup., Band II,
Teil K. 1935.

Dept. of Transport
(Canada): Meteorology of the Canadian Arctic”, 1944.
“Climatic Summaries”.
“Aerological Data from Northern Canada”. 1949

F. Kenneth Hare
HomeThe Climate of the American Northlands : Encyclopedia Arctica 7: Meteorology and Oceanography
 Text Only
 Text & Inline Image
 Text & Image Viewer
 Image Viewer Only