FOG 
The smaller number of particles is characteristic of the wet 
and less dense type of fog which consists of relatively large 
particles. The larger number is the characteristic of the very 
dense and dry type of low level industrial fog which con- 
sists of a great number of very small particles. The wet fog 
with 10” particles per cc may contain more condensed water 
per ec than the dry fog with 10* particles, but for the same 
amount of liquid water per cc the dry fog of numerous par- 
ticles is much more effective in reducing the visibility. 
Fog over Snow Cover. The formation of fog over snow 
introduces considerations which are not present in the 
absence of snow cover. The principal effect is that the 
snow cover acts as a fog-inhibitor. Petterssen [19] cites 
climatological evidence to the effect that fogs are rare 
Over snow-covered areas in winter, just when radia- 
tional conditions are nearly perfect. His evidence, how- 
ever, indicates that for temperatures below —40C fogs 
imcrease again and are frequent during such low tem- 
perature conditions when they are ordinarily composed 
entirely of ice crystals. Byers [2] comments on the 
frequency of fog particularly at very low temperatures 
and quotes from Frost’s Cliumatological Review of the 
the Alaska-Yukon Plateau, “At Fairbanks dense fog 
always accompanies temperatures of —45F or 
lower ....” 
However, important recent evidence is presented by 
V.J. and M. B. Oliver [18] to the effect that these low- 
temperature fogs form only around settlements and 
are almost nonexistent in regions only a few miles 
removed from them. Pilots accustomed to flying in the 
arctic winter confirm this, and 1t seems that many of 
the accepted statistics concerning arctic fogs at low 
temperatures are impeached by this evidence. The 
Olivers offer the hypothesis that smoke particles from the 
towns act as sublimation nuclei and are the necessary 
element for the formation of fog under these conditions. 
They suggest that the smoke contains sublimation 
nuclei not present in sufficient numbers otherwise. They 
also point out, however, that the sources of smoke are 
also sources of moisture which may be of primary im- 
portance. 
The reasons for the low incidence of fog over snow 
cover are completely treated by Petterssen [19]. Briefly, 
for temperatures above freezing he considers that the 
temperature of the snow surface remains near 0C and 
that the usual condition of an increase of specific 
humidity with elevation causes an eddy flux of moisture 
toward the surface, with the result that condensation 
takes place on the snow. This removes moisture from 
the air and acts as an inhibitor of fog but does not 
entirely prevent it. The higher the temperature above 
freezing, the greater the dissipating effect. (Although 
it is not mentioned by Petterssen, it is apparent that 
this process would have no such dissipating effect upon 
low stratus clouds.) At a temperature of exactly freezing 
for both air and snow surface no dissipating influences 
exist, since the saturation vapor pressure is the same 
over water and over ice at this temperature. For tem- 
peratures below freezing, Petterssen ascribes the dis- 
sipating process to the difference between the saturation 
vapor pressure over water and over ice. 
1181 
The Olivers [18] mention another dissipating influence 
on supercooled water fogs. 
A mixture of water fog and ice fog occasionally forms at 
temperatures between zero and freezing, when clearing skies 
permit rapid cooling by radiation. This condition lasts for 
only a few hours, during which period the water fog rapidly 
changes to ice fog or rime and the ice fog thins out and dis- 
appears even though temperatures continue to fall. This gen- 
eral weather situation also produces rapid accumulation of 
hoar frost on trees, wires, and other exposed surfaces. Ap- 
parently, the snow surface, trees, wires, etc., are much better 
radiators than air, and so become much colder than the air, 
thereby permitting rapid frost deposits, rapid drying of the 
air, and lowering of the dew point. 
It is well known that fogs contain supercooled water 
down to temperatures of —20C to —40C. The recent 
work of Schaefer [22] indicates that at —39C fog 
particles of 10- to 15-» diameter will freeze without 
outside stimulus. KGhler [14], some time ago, found that 
large drops became ice at higher temperatures than 
small ones, and Heverly [11] has recently produced 
quantitative results which seem to show that fog drop- 
lets with diameters of 50 » would begin to freeze 
spontaneously at about —28C or —30C. Even if it is 
argued that fog droplets are much smaller than this at 
these temperatures and consequently would have a 
spontaneous freezing point nearer to Schaefer’s —39C, 
there is still another effect which would lead to a mix- 
ture of ice and water particles below —28C. For ex- 
ample, if saturated air at a temperature of —28C is 
cooled further, as by radiation or upslope movement, 
the further condensation resulting will [11] “produce 
relatively fewer and smaller water droplets and more 
ice crystals.”’ Accordingly we may conclude that nearly 
all fogs in air at —28C will contain some ice crystals 
and that the proportion of ice crystals to supercooled 
droplets will increase rapidly for lower temperatures. 
Bergeron [1] has shown that a mixture of supercooled 
water and ice crystals is “colloidally unstable.”’ Due to 
the difference in saturation vapor pressure over water 
and over ice, the vapor pressure over the water droplets 
would be appreciably greater than that over the ice 
particles, and a flux would be set up causing evaporation 
of water particles and an increase in size of the ice 
particles. There is a tendency for mixed water and ice 
fogs to change over entirely to ice fogs by this process. 
Figure 3 is particularly interesting im the light of the 
above discussion. It is from the Olivers’ paper [18] on 
Alaskan fogs. The almost complete absence of fog at 
temperatures above —33C and the very rapid increase 
at lower temperatures seems to add credence to the 
hypotheses discussed above. All authorities agree that 
for temperatures below —39C fogs are composed en- 
tirely of ice crystals which are in equilibrium with a 
snow surface and therefore have no tendency to dis- 
sipate due to the surface condition. 
Physical Processes. The classical concept of the radia- 
tive and heat balance of fog situations is about like 
this: After nightfall, on calm clear nights, the radia- 
tional cooling of surface objects is much more rapid 
than that of the air. Consequently the surface layers of 
