686 
network showed that in 182 of the traces there was a 
pressure rise associated with the arrival of the cold 
outflowing air. Since the rate and total amount of 
pressure rise depend on the slope of the cold-air mass, 
the temperature difference between the cold air and 
the displaced warm air, the depth of the cold air itself, 
and the speed with which the system travels, the most 
marked pressure changes are found near the cell-core 
and decrease with distance from it. The areal extent of 
the pressure dome is similar to that covered by the 
cold air. Therefore the pressure remains high for a 
period of from one-half hour to several hours, depending 
on the amount of cold air involved. 
The distinction between the pressure nose and the 
pressure dome can be made only with difficulty on the 
conventional week-long barograph traces, but on the 
twelve-hour recording drums used on the Thunderstorm 
Project the distinction is clear. 
At any given time the pressure nose was usually 
detected by but one station per cell m Ohio, indicating 
that the maximum diameter of the area in which it 
occurred was less than five miles It is thus apparent 
that it is caused by a transitory process which exists 
for only a brief period, namely the time of commence- 
ment of the mature stage when rain and downdraft 
reach the ground. Two effects which are thought to be 
important in creating the pressure nose are the weight 
of the suspended and falling water, and a lack of balance 
between convergence and divergence. Both of these 
effects are at a maximum at this time. An accumulation 
of an average of one gram of liquid water per cubic 
meter from the cloud base to 35,000 ft, a conservative 
value, would increase the surface pressure by about 
1 mb, other factors being equal. When the downdraft 
becomes established, it reverses at least part of the 
vertical circulation pattern and reduces the divergence 
above 20,000 ft, or perhaps even reverses it to con- 
vergence. Should this happen before the divergence 
(outflow) in the surface layers becomes well established, 
a net increase in mass convergence could result in a 
substantial surface pressure rise. For example, an un- 
compensated convergence of 1 hr (a conservative 
value for thunderstorm convergence) within a layer 
only 1000 ft thick at 20,000 ft would increase the 
surface pressure at a rate exceeding 1 mb in 3 min. 
Relationships of pressure changes to vertical motions 
in thunderstorms, which have been a favorite subject 
for theoretical treatment by meteorologists, could not 
be confirmed by data of the Thunderstorm Project. 
One of the difficulties is that a downdraft at the lower 
levels is under an updraft at a higher level, so that the 
vertical accelerations are opposed and tend toward 
compensating each other. 
After the brief pressure nose, the pressure remains 
at the value of the pressure dome which prevails for the 
particular thunderstorm. The dome persists through 
the dissipating stage of the cell, after which the pressure 
returns to the trend prevailing before the passage of the 
storm. In the case of a thunderstorm associated with a 
cold front or a fast-moving squall line, the pressure 
remains high or even continues to rise as a result of 
LOCAL CIRCULATIONS 
cold-air advection or the passage of a wave in the 
pressure and wind fields. 
Humidity. An extraordinary phenomenon is noted in 
the hygrograph traces from the surface micronetworks 
in the form of a sharp decrease in the relative humidity 
in the midst of the heaviest downpour of rain. With the 
onset of rain, the relative humidity rapidly approaches 
but usually does not quite reach 100 per cent, then, in 
many cases, drops suddenly to values as low as 60 or 
70 per cent, returning to near saturation again m a 
few minutes. Such a fluctuation, termed the humidity 
dip, and illustrated in Fig. 7, is associated with the rain 
area. In practically all cases the humidity dip occurred 
in a region of divergence in the surface wind, therefore 
in the outflowing downdraft. The temperature was 
usually decreasing, but if it had already reached its 
lowest point, as was sometimes the case, a rise in 
temperature of 2 or 3F would sometimes accompany 
the humidity dip. 
spa et aoa wae er eee eh eel a oo 
Fic. 7—Example of humidity ‘‘dip” as shown on hygro- 
gram during heavy rain under thunderstorm of August 14, 
1947 near Wilmington, Ohio. Time divisions are for 5-min 
intervals. 
The failure of surface air to become saturated during 
the heavy rain, together with the frequently observed 
relative-humidity dip, suggests that the downdraft air 
fails to maintain saturation as it descends, even in the 
presence of large concentrations of liquid water. Two 
processes are suggested: first, desiccation of the air by 
condensation on cold precipitation particles; second, a 
time lag in the evaporation of the waterdrops so that 
there is insufficient accommodation to the increase in 
the saturation mixing ratio as the air descends to lower 
levels. 
Measurements of raindrop temperatures at three feet 
above the ground show that in afternoon thunderstorms 
the rain temperature in the first few minutes of the rain 
period is several degrees lower than that of the ambient 
air. This would be sufficient to keep the relative humid- 
ity from reaching 100 per cent. If the second process 
mentioned above is taking place, one would expect that 
the air of the thunderstorm downdraft would be heated 
at a rate between the moist- and dry-adiabatic and 
would reach the ground in an unsaturated state. 
Development of New Cells. From a study of numerous 
cases of new cell development, the action of the cold 
outflowing air appears unmistakably as causing or con- 
tributing to the new growth. When two cells in the 
mature stage are within a few miles of each other, the 
cold outflows collide. The greatest frequency of new 
