FOG 
/ 
during fog conditions at New Orleans Airport [12]. 
Other localities require more complicated treatment 
than this. For instance, Figs. 8a, b, and c show the 
ife) 20 30 
"inp 180 
Fig. 7.—A typical method for forecasting advection-radia- 
tion fog for New Orleans Airport during winter (after Hil- 
worth [12]). The data are gradient winds taken from 
the afternoon upper-wind sounding. 
steps required with present methods for forecasting 
this type of fog at Charleston, 8. C., during the fall 
months. Other and still different methods are used for 
the remaining seasons. 
Pre-Warmfrontal Fog. The forecasting methods for 
this class of fog usually given in textbooks are rather 
general in nature and sometimes they are omitted 
altogether. It is difficult to find much literature on the 
subject which would be of real assistance to an in- 
vestigator wishing to learn how to forecast these oc- 
currences. Perhaps the biggest difficulty is the vaguely 
recognized fact that local geography again plays a 
dominant role. For example, pre-warmfrontal fogs occur 
very rarely at Cleveland, probably due either to down- 
slope compressive heating in the cold air or to the 
frictional divergent effects caused by the proximity of 
Lake Erie, or to both. At this locality, the clouds con- 
sistently remain well above the ground in pre-warm- 
frontal situations, while at Atlanta, Washington, D. C., 
and Chicago, fogs formed in this way are common. 
Some rather ingenious methods have been utilized 
for forecasting the formation of low stratus in warm- 
front situations. An example is Roche’s graph [21] for 
Charlotte, N. C., shown in Fig. 9. In nearly all such 
methods, however, the authors have used synoptic ma- 
terial which was readily available in a practical sense. 
Not much effort has been made to apply quantitatively 
any relationship between temperature of falling pre- 
cipitation and the surface air layers, which Petterssen 
has pointed out to be very important to the problem. 
Also, consideration of the stability of the cold air has 
remained largely a qualitative problem. 
Mizxing-Radiation Fog. This is not at all a usual 
type, but it is important to recognize the significant 
role played by dissolving fronts at many locations. 
Particularly in late spring and summer there are many 
areas which have little or no fog not associated with a 
weak frontal condition. Whether the formation of this 
kind of fog involves actual mixing of two dissimilar air 
masses is open to considerable question, but nocturnal 
radiation is certainly a requirement. Furthermore, there 
are many occasions when fog does form in a narrow 
1187 
band along a dissolving front even when no daytime 
clouds attend it. Forecasting methods are very general 
for this unusual class and depend largely upon clima- 
tology and the individual experience of the forecaster. 
b 
z40 (0) 
S 
S68 FOG 
wn 
uJ 
a 
tw 
a6 NO_FOG 
is 
: Pane 
a 
=10 }—t 
Wd 
3 CUAL] 
=3 oy 3 6 9 12 15 
DEWPOINT INCREASE(*F) 
(c) 
z 
72 
7. 8 to] Ni2 
& 3 NE 
= 
Zz 
[e) 
Qa 
= 
a 
54 
45 
5 10 15 20 
DEWPOINT DEPRESSION °F) 
Fras. 8a, 6, c—These graphs indicate a representative so- 
lution for advection-radiation fog a little more complicated 
than most. They are for winter cases at Charleston, 8. C. 
(after George). Figure 8a takes care of the gradient wind re- 
quirement as well as giving a preliminary estimate of time of 
formation from the numbered contour lines (in hours after 
sunset). Figure 8b eliminates certain instances which would 
otherwise produce an erroneous forecast for fog. Ordinates are 
dew-point depressions at maximum temperature and abscissas 
are obtained by subtracting the temperature of the dew point 
at maximum temperature from that at sunset. Figure 8c pro- 
vides the humidity element and another estimate of timing. 
Timing the Clearing of Fog or Stratus. Several methods 
have been devised to aid in estimation of the time 
required to dissipate fog or stratus. Probably the sim- 
plest and most direct method is the relationship between 
thickness of the stratus and time required to clear, 
given by Wood [26] and reproduced here as Fig. 10. 
