dividual case, the local temperature variations 
might be much higher. As soon as the ground 
becomes bare of snow in early spring, the sur- 
face temperature increases rapidly at daytime 
because of the increase of the absorbed solar 
radiation. At night the temperature differences 
between various surface types are generally 
small. With high surface temperature and 
increased turbulent fluxes of sensible and latent 
heat from the bare ground, one may generally 
expect increased rates of melting as warm and 
moist air is advected over the remaining snow. 
On the microscale, stable internal boundary 
layers will prevail over the bare ground. 
We also observed rapid increases of the water 
temperature of the puddles due to high values of 
the solar radiation, allowing high rates of evapo- 
ration. Several times during the breakup period 
we observed fog-smoke over ice-free ponds and 
over the waterlogged tundra in relatively cold 
weather, indicating high rates of evaporation. 
During a 4-day period after snowmelt in June 
1971, Weller et al. (1972) found evaporation 
rates of 4.5 mm per day at Barrow. 
Melting of a snowpack with high albedo 
generally starts when the air temperature in- 
creases toward O°C. The importance of the 
vertical turbulent fluxes of sensible and latent 
heat in the melting process and, indirectly, the 
air temperature, is especially apparent during 
brief spells of warm weather in early spring. For 
instance, the rise of air temperatures above 0° C 
in connection with warm air advection on 5-6 
May 1972 at Prudhoe Bay induced melting and 
the formation of ice lenses within the snowpack 
(see Physical Characteristics of the Snow, above). 
After the air temperature decreased, the percolat- 
ing meltwater in the snowpack refroze again. 
During the main breakup period on the 
tundra, the influence of the air temperatures is 
not as evident as during the early spring period. 
On the other hand, there are reasons to always 
consider the melting above an extensive snow- 
field as the combined effect of the net radiation 
at the surface and the warm air advection since 
the total radiation budget of the surface- 
atmosphere system is always negative over a 
snow surface. In other words, the energy input 
from the snow surface into the atmosphere is 
too small to compensate for the radiation energy 
losses of the atmosphere. This is why melting at 
43 
the surface is closely related to advectional 
effects above the surface. 
Since the air temperatures at the end of May 
and at the beginning of June appear to vary 
around O°C in the coastal zone of the Arctic 
Ocean, with relatively small variations in differ- 
ent weather situations, the advection effects on 
the melting process may become less apparent 
than during the brief early melting periods. Also, 
as bare patches of the ground appear, a rapid 
warming of the surface takes place, changing the 
microclimate around and over the remaining 
snow patches. At this stage, the large-scale 
advection effects should become less important 
in relation to the advection on the microscale. 
Exactly when the snow starts to melt and 
exactly when it is gone is a matter of subjective 
judgment. Roughly, the melting takes place in 
2-4 weeks. During part of that period the snow 
surface is at the melting point. However, espe- 
cially when light melting occurs intermittently, 
it is difficult to judge whether or not the surface 
is melting. Any single meteorological parameter 
is then a rather poor indicator of the snow 
surface temperature. Part of the problem is 
related to the fact that snow is semitransparent 
to short-wave radiation. The short-wave incom- 
ing radiation that penetrates the surface is 
absorbed at depth, most of it within the first 
few centimeters. At the surface the emitted 
long-wave radiation is generally greater than the 
incoming long-wave radiation, and the net radia- 
tion budget of the uppermost surface layer may 
be negative. In situations with air temperatures 
slightly above or below O°C, the surface is often 
frozen, while internal melting occurs in the 
snowpack. The penetration of solar radiation 
into the snowpack and the physics of the melt- 
ing have been discussed in detail by Liljequist 
(1956). 
Toward the end of the breakup the flat 
tundra may be regarded as an extensive shallow 
lake, as far as the surface conditions go. The 
highest amounts of snow accumulation are 
found in shallow depressions and in the sastrugi. 
Ground observations and also observations dur- 
ing helicopter flights in the Prudhoe Bay area on 
4 and 5 June 1973 indicated that a considerable 
damming of the meltwater may be caused by the 
snow and possibly by the superimposed ice in 
the natural shallow drainage channels. As the 
