FOG 
nights. This definition of radiation fog carries with it the 
implication that such fog does not form unless one of 
the two conditions mentioned above is present. 
Some misunderstanding of these principles has ap- 
peared; for instance, O’Connor [17] specifies a non- 
convective cloud cover, and states that fogs formed as 
a result of this type of restricted heating are usually 
post-coldfrontal. In reality, a large proportion of cloud 
covers which result in radiation fogs are strictly con- 
vective in origin, and although practically all fogs in 
clear air behind cold fronts are formed from this cause, 
they are very few indeed when compared with the total 
number of radiation fogs caused in this way. 
For forecasting purposes there is little point in dis- 
tinguishing between ground fog and stratus as two 
different types of fog, for in nearly all cases the fore- 
casting problem is identical. Whether ground fog or 
stratus results is entirely dependent upon (1) the ter- 
rain, (2) the characteristics of the air mass, and (3) 
whether the wind velocity exceeds a certain value. 
Accordingly, in the forecasting classification ‘ground 
fog” and “high inversion fog” are replaced by “‘re- 
stricted heating” and “air drainage fog.” 
Forecasting Methods 
The forecasting of fog is, above all other considera- 
tions, a local problem. There are many instances in 
which airport sites which have been moved only a few 
‘miles required definite revision of methods for fore- 
casting fog. This is especially true of locations at which 
an advective element is present. Each season usually 
requires a new technique, and the duration of the 
season must be determined from local data. 
Advection Fog (Sea Fog). As the name implies, fog 
of this type forms over the open water as a result of the 
advection of warm moist air over colder waters. This 
process obviously stabilizes the air and if continued 
long enough cools the air to its dew point and fog is 
formed. The forecasting of these occurrences for sea 
areas can best be done by combining a computation of 
future trajectories [2, 7] at various points with a care- 
ful plot of mean sea-surface temperatures. 
Radiation (Restricted Heating) Fog. There is only one 
type of fog which can be forecast by using a single tool 
for most localities, and even this type must be hedged 
with restrictions. The tool is George’s restricted heating 
radiation fog graph. This method has been tested at 
various locations in the eastern United States and 
found to be valid at all locations tested which (1) 
were not within about 50 miles of a seacoast or other 
large body of water (including some large rivers) or (2) 
presented no severe air drainage problems. Figure 4 
shows the general form of the radiation graph from 
Roche’s study [21] of Carolina stations. It 1s not en- 
tirely clear why a locality near a seacoast does not 
experience fog caused by restricted heating in the same 
degree as inland stations, but it seems likely that land- 
breeze effects at night added to the normal divergence 
caused by decrease in friction as the air moves from 
land to sea may very well thin the moist layer enough 
to overcome all but the most ideal situations. 
1185 
The presence of precipitation as a primary element 
in the formation of fogs has been recognized for some 
time. Counts [4] cites it as a basic precept in forecasting 
2 4 6 8 iKe) l2 14 
HOURS AFTER SUNSET 
Fig. 4.—Radiation fog graph devised by George and Brad- 
ley. Data are by Roche, for Carolina stations. S + D (ordi- 
nates) are the sum of the number of sunshine hours during 
the day plus the number of degrees between temperature and 
dew point at sunset. The units are mixed but retained in this 
inaccurate form for practical convenience. Abscissas give num- 
ber of hours after sunset that fog should form. 
fog in California’s San Joaquin Valley, and earlier in 
this article it was pointed out that Clapp found it an 
almost essential factor for the formation of fog at 
Cheyenne, Wyoming. Precipitation occurring during 
the day aids in the formation of fog in two ways: (1) 
it adds moisture to the lower air layers, and (2) the 
attending clouds act to prevent the normal diurnal rise 
in temperature. 
At one location (Atlanta, Ga.) a study [8] has been 
made of the difference in fogs and stratus clouds caused 
by daytime precipitation conditions and those formed 
when only cloudiness (without precipitation) was pres- 
ent. Figure 5 illustrates this difference. For geostrophic 
winds of 15 mph, the resulting stratus clouds in rain 
situations have an average minimum ceiling of about 
200 ft (Fig. 5a) and not infrequently they may lower 
to the surface as dense fog. For the same wind velocity, 
Fig. 5b shows the minimum ceiling in “no rain” situa- 
tions to be about 700 ft. For geostrophic winds below 
12 mph the “no rain” graph indicates that the fog 
forms and remains a true fog, that is, one at ground 
level. According to these data, the effect of rain is to 
add enough moisture to the lower layers to offset to 
some extent the dissipating effect of higher wind veloci- 
ties. 
Radiation (Air-Drainage) Fog. At various locations, 
unusual geographic conditions exist which cause the 
