TROPOSPHERIC PROPAGATION AND RADIO METEOROLOGY 157 
be seen that at as small a height as 0.5 m above the 
sea M has a value much lower than at the surface 
itself. As the surface value of M is obtained by the 
assumption that the air in immediate contact ‘with 
the water is saturated with moisture, this indicates 
that 0.5 m above the water the moisture content of 
the air is still appreciably below saturation. The 
moisture in the lowest levels is subject to consider- 
able variations caused partly by turbulence, partly 
by the waviness of the sea surface. M curves, such 
as Figure 31, are obtained by averaging over several 
measurements. 
These ducts are much lower than the advective 
ducts discussed in the previous section; their height 
is about 12 to 15 m (around 40 ft). The effective 
decrease of M in the duct (apart from the sharp 
decrease in the lowest half méter) is of the order 
of 4 to 8 MU. 
The latter figure depends somewhat on the wind 
speed. There is a maximum decrease of 8 MU at Bo 
wind speed of about 8 m per sec (13 miles per hour) »”! 
‘and lower values for both lower and higher wind 
speed. The duct height in turn shows a very slight 
dependence on wind speed, increasing somewhat with 
increasing speed. 
These ducts are so low that they are not very 
effective for trapping of waves even as short as S 
band, presumably on account of strong leakage (see 
pages 149-150), and signal strength is not increased 
when an S-band transmitter or receiver is placed 
inside the duct. For K band, on the other hand, the 
trapping effect is marked; on raising the transmitter 
or receiver from the ground a maximum of signal 
strength is observed at about 9 m, but from there on 
the signal begins to decrease up to about 20 m (overall 
decrease 5 db) ; at greater heights the signal gradually 
rises again. 
These ducts appear to be a permanent feature at 
Antigua, at least during the season these observa- 
tions were carried on. This is probably true also for 
many locations in the trade wind belt. The daily 
variation of weather phenomena and of duct charac- 
teristics at such purely maritime locations seems to 
be insignificant. 
Nocturnal Cooling—Daily Variations 
A daily variation of surface temperature occurs 
only over land. During the day the heating is caused 
by the sun’s rays, and the cooling of the ground 
surface during the night is produced by radiation 
from the ground. The diurnal temperature variation 
of the sea is extremely small. However, shallow 
bodies of water sometimes have an appreciable 
diurnal variation. 
The radiation which-causes nocturnal cooling of 
the ground is temperature or heat radiation which is 
composed of waves in the infrared portion of the 
spectrum. It is the same kind of radiation that is 
given off by a hot stove or electric heater, but since 
the temperature of the earth is less than that of a 
stove the earth emits comparatively less heat radia- 
tion. Nevertheless, radiation is a very powerful agent 
in cooling the ground. From about sunrise until the 
late afternoon, the surface of the earth gains more 
heat from the sun and atmosphere than it loses by 
radiation to space; in the late afternoon and during 
the night, the surface loses more heat than it gains. 
The amount of heat radiated is very nearly inde- 
pendent ct the physical constitution of the ground 
but is dependent upon its temperature and increases 
very rapidly with a rise in ground temperature. 
The atmosphere has a “blanketing” effect upon 
the infrared radiation emitted by the ground. The 
atmosphere itself absorbs and emits infrared radia- 
tion, and the cooling of the ground may be greatly 
reduced by the action of the atmosphere. The 
blanketing effect is least with a clear sky and dry, 
cool air; it is somewhat stronger when, with a clear 
sky, the atmosphere is-very warm and humid, as in 
the tropics. A cloud will produce a distinct blanket- 
ing effect, and with a complete overcast of low cloud 
the blanketing is so pronounced that the nocturnal 
cooling of the ground is reduced to only a small 
fraction of its value with clear skies. 
The loss of heat from the ground is distributed by 
turbulence over the lowest layers of the atmosphere, 
thus giving rise to a temperature inversion. Inver- 
sions of this type are strongest in temperate and 
cold climates with a clear sky and cold, dry air 
overhead; they are less pronounced in the tropics 
with humid air and a clear sky and are practically 
absent with an overcast sky. A meteorologist, after 
some experience, can estimate the magnitude of an 
inversion to be expected with given local weather 
conditions. 
Temperature inversions, by themselves, can at 
best produce only weak ducts, but strong ducts may 
result when the inversion is accompanied by a suffi- 
cient moisture lapse. This requires that the air be 
dry enough to allow evaporation into it from the 
ground. In warmer climates where the transition 
between night and day is rapid, evaporation may 
set in in the early hours of the morning before the 
nocturnal inversion has been completely destroyed 
by the action of the sun. A strong duct will then be 
formed for a short period. This condition seems to 
be frequent during certain seasons in Florida. 
It is obvious that the shape of the M curve, when 
it deviates from the normal, may undergo rapid 
variations with the period ef a day. One example 
has just been quoted; another is illustrated by the 
advective ducts over the North Sea produced by the 
mechanism described on pages 155-156. These ducts 
usually form in the hours before midnight and last 
until the early hours of the morning. 
Fog 
Contrary to what might perhaps be expected, the 
