1098 
hope for are of the order of 1000 nautical miles per day. 
Not until many more situations have been analyzed will 
an evaluation of the usefulness of the method be pos- 
sible. The results so far obtained demonstrate the 
feasibility of locating and tracking storms at distances 
exceeding 3000 miles and of making rough estimates 
regarding the size and character of these storms. 
The Storm Surge Method 
The preceding discussion has indicated that the swell 
spectrum, visible or invisible, is limited by the maxi- 
mum velocities of the winds in a storm area. This con- 
clusion is in accord with the hypothesis that swell was 
originally caused by the normal and tangential stresses 
applied by the wind to the sea surface, but it does not 
preclude the existence of other types of waves from 
storm areas of much longer period than those of the 
swell. To explore this possibility an instrument [11] 
has been built which is sensitive to waves whose periods 
lie between those of the swell and tides. It is essentially 
a modified tide gauge, and the filtering of the swell and 
tides is accomplished by means of capillaries and air 
MARINE METEOROLOGY 
Should it be possible to use storm surges in locating 
storms at sea, they would have the advantage of rela- 
tively high speed, \/gh in water of depth h, or 10,000 
nautical miles per day in mid-ocean. It is hoped, there- 
fore, that they may eventually provide a tool to the 
synoptic meteorologist. 
Discussion and Conclusions 
The relative merits of the three methods involving 
the propagation of ocean surface waves are summarized 
below. We shall also point out some of the disadvan- 
tages and advantages of these methods as compared to 
methods involving the propagation of sound waves 
through the sea bottom (microseisms) and electro- 
magnetic waves through the atmosphere (radar and 
spherics). 
Instrumentation. The height-period (HT) method re- 
quires no special instrumentation. The frequency-spec- 
trum analysis (FA) method requires an underwater 
recording unit connected by a cable to a shore recorder. 
Along rocky coasts and coral reefs the chafing of the 
cable may create quite a serious problem and the use of 
Taste I]. Dara ror FIvE SEVERE STORMS OFF THE CALIFORNIA COAST 
Earliest storm surges Maximum storm surges Maximum swell 
Date (1948) Warni 
Date and time a (min) Date and time & (min) Date and time & GS period. 
Feb. 9-10 9:0500 .04 14 9:2230 12 15.5 10:1500 6 8 34 
Feb. 21-23 21:1800 06 15 22:1100 15 15.5 23:0400 8 9 34 
Feb. 27-29 27:1400 06 12 28 :0600 12 14 29:1200 4.5 il 46 
Mar. 30-31 30: 1000 .10 20 30: 1600 14 12 31:1200 7.5 8 26 
Apr. 22-24 22:0800 -08 18 22:2200 -09 13 24:0700 9.5 8.5 47 
Mean: 37 
volumes. The instrument does not require the expensive 
and troublesome underwater cable and is therefore 
simpler and much less expensive than the underwater 
pressure recorder described in the preceding section. One 
such unit has been installed at the end of the Scripps 
Institution pier at La Jolla, California; a second unit 
has recently been installed in Hawaii. 
During the spring of 1948 five severe storms passed 
off the coast of California. In each of the five cases, 
waves of periods of approximately 15 min (storm surges) 
were received from 24 to 36 hours prior to the arrival 
of the heaviest surf and the strongest winds. 
Some of the data are summarized in Table II, giving 
the height H and period T of the storm surges, and also 
of the swell as it was recorded on the underwater 
recorder at a depth of 40 ft. The mean height of the 
storm surges was only 1 in. and it is quite clear that 
these waves could be detected only by special instru- 
ments. According to Table II the earliest arrival of the 
storm surges preceded the heavy swell by one and one- 
half days on the average. 
It must be emphasized that the study of these storm 
surges is in a preliminary stage, and that the manner in 
which they are generated is not at all understood. 
expensive and unwieldy armored cable may be required. 
A frequency analyzer must be available at each loca- 
tion if the method is to be used synoptically. The storm 
surge (SS) method involves a much simpler installa- 
tion, provided a pier or breakwater is available for 
securing the instrument. 
Storm Location. The bearing and distance of a storm 
can be located by the HT and FA methods from a 
single station. In the case of the HT method a visual 
estimate of wave direction should suffice in view of the 
uncertainties regarding the determination of distance. 
The theory underlying the method is insecure, and it is 
also necessary to make assumptions regarding the dura- 
tion of the storm. Furthermore, an estimate of deep- 
water wave height from shore observations may be 
subject to considerable error. The HT method has the 
advantage that it permits some estimate of the storm 
location from a single set of simple observations. 
The FA method, on the other hand, depends upon 
the determination of a shift im the frequency spectrum 
and requires therefore (intermittent) observations ex- 
tending over a minimum period of perhaps six hours. 
The determination of distance is based on sound physi- 
cal theory, and furthermore does not involve the deter- 
