CLIMATOLOGICAL PROBLEMS IN THE ARCTIC AND SUB-ARCTIC 
lies well north of the tree line in North America, as 
well as in most parts of Siberia. The northernmost trees 
in Keewatin do not get far beyond the line # = 33.0 
em. In Labrador-Ungava they locally attam H = 29.5 
em, but again tend to halt at the 33-cm value. 
Attempts to find climatic equivalents for the tree 
line assume that the latter is in fact “‘climax,”’ that is, 
has reached the northernmost limit of feasible tree 
growth. In many areas this is not the case. The trees 
have not yet reoccupied all the climatically suitable 
lands from which they were driven by the Wisconsin 
glaciation and its attendant cold. Thus Marr has shown 
that active northward migration of the tree line is in 
progress on the east coast of Hudson Bay [88]. Griggs 
has similarly described the spread of the Pacific and 
Boreal forests into the Alaskan maritime tundra [17], 
notably on Kodiak Island and the Alaskan peninsula. 
It seems quite evident that the tree line has not yet 
attained its equilibrium position in many areas; further- 
more the rapid fluctuation characteristic of recent sub- 
arctic climates makes it doubtful whether any such 
equilibrium can be struck at all. In the circumstances, 
it is unlikely that any rational climatic equivalent can 
be found for the tree line. Nordenskjéld’s line, arbitrary 
though it is, comes closer to the present position than 
any “rational” attempt. A glance at Fig. 1 will con- 
vinee the reader of this close fit. 
The “maritime tundra” already briefly discussed pre- 
sents many interesting problems. Along many coastal 
belts. of the sub-Arctic, treeless tundras, usually with 
a high proportion of typically arctic species, descend 
far south of the normal poleward limit of trees. Excel- 
lent examples are the Aleutian-Alaskan peninsula re- 
gion, the Labrador coast, parts of Iceland, coastal 
Fennoscandia, and certain localities on the coast of 
the Maritime Provinces of Siberia, facing the Okhotsk 
and Bering Seas. Similar coastal treeless formations 
occur still farther south, as in the moss-barrens of 
south-coastal Newfoundland, the coasts of Ireland and 
Scotland, and the Kuril Islands. These regions are 
different, however, in that floristically the vegetation 
has many nonarctic affinities; furthermore, human agen- 
cies may have destroyed pre-existing forests, except in 
Newfoundland. 
The best-studied case is that of coastal Labrador 
[28, 66], where maritime tundra forms a narrow belt 
extending as far south as the Strait of Belle Isle. Though 
rarely more than a few miles wide, the strip grades in- 
land into the Boreal forest through a well-marked tree 
line and a zone of stunted brushwood like the krummholz 
of the Vorarlberg and the thickets of Mount Washing- 
ton. It was long assumed that this tundra was the 
result of the chilling of summer temperatures by drift- 
ing ice in the Labrador current. It has been shown, 
however, that the “thermal efficiency” of the coastal 
climate is adequate to support a poor forest or wood- 
land cover [19, pp. 237, 238; 19a, p. 631]. The absence 
of trees is ascribed by Wenner to excessive transpiration 
(and hence physiological drought) induced by the strong 
coastal winds. Whether the explanation holds or not, 
957 
it is certainly true that the forest finds it more difficult 
to invade the coastal sub-Arctic than the interior. 
There are very few studies of the montane tree line 
in subarctic regions comparable with those on high, 
isolated summits in temperate and tropical latitudes. 
The Forest-Tundra Ecotone. This zone is a belt of 
variable width in which treeless and woodland associa- 
tions intermingle. In North America it has been de- 
limited in Labrador by Hustich [31], its southern bound- 
ary being thermally adjusted to the isopleths of 
E = 35-37 cm 
potential evapotranspiration [19a, p. 630]. Elsewhere its 
boundary has yet to be adequately mapped. In Siberia 
its extent is well known west of the Yenisei, where again 
its southern edge is at about H = 35 cm. It is clear that 
in these regions the ecotone is a belt whose position 
and width are thermally determined, thus conforming 
to the expectations raised by the work of Hustich, 
Erlandsson, and others discussed above. 
The Boreal Forest Itself. This zone can be divided 
into two broad subdivisions that have been recognized 
both in North America and in the Soviet Union. On 
the northern flank there is a belt of open woodland, 
characterised by widely spaced spruce (Pzcea spp.), 
pine (Pznus spp.), or larch (Larix spp.) set in a rich 
lichen floor, the dominant genus of lichen being Clado- 
nia (‘“reindeer” or ‘“‘caribou moss”). On the southern 
flank lies a broad belt of true forest, with close stands 
of spruce, fir, pine, larch, and some hardwoods. The 
ground is in deep shadow and is usually carpeted with 
mosses and shade-loving herbs. The two belts are fairly 
distinct, meeting along a mappable boundary; they 
go by various names, but the term ‘‘Boreal woodland”’ 
and ‘main Boreal forest” will be used for them here. 
The boundary between them is indicated for Canada 
in the official publication ‘Native Trees of Canada” 
[5], the Boreal woodland being indicated as “‘transi- 
tional” to the Arctic to the north. Hustich, however, 
distinguishes (1) a forest-tundra ecotone, already dis- 
cussed, (2) a “taiga”’ belt, or Boreal woodland of the 
lichen-floored type, and (8) a southern Boreal forest, 
equivalent to the main Boreal forest defined above 
[81]. Hustich’s divisions correspond structurally with 
those widely recognised for the Boreal forest of Finland 
and Soviet Russia, and are hence of great interest [53]. 
In Labrador-Ungava, it is at once clear that the wood- 
land/forest boundary coincides very closely with the 
isopleth of potential evapotranspiration H = 43 cm 
[19a, p. 630], which is that selected by Thornthwaite to 
divide the microthermal climates into warmer and cooler 
halves. Though a similar investigation of the Russian 
climate has not been carried out, Arkhangelsk, about 
100 mi south of the boundary, has # = 44.5 em, and 
Ust-Sylma, on the Pechora River some 50 mi within the 
Boreal woodland has H = 40.3 cm. Furthermore, it is 
clear that all the belts of the Boreal forest west of the 
Yenisei extend zonally along the isopleths of potential 
evapotranspiration, as they do in Labrador-Ungava 
[31]. Obviously the woodland/close-forest boundary is 
