MARCH 1935 MONTHLY WEATHER REVIEW 85
4. A relatively low quartile deviation is usually con- sidered favorable. A small departure below median probably would not often be disastrous. In areas where the quartile deviation is low we say the rainfall is of a
dependable type.
5. A knowledge of departure below the median es- pressed in perc,entage of the time should be helpful especially in subhumid and semiarid areas, or in places
where there is a tendency for a wide departure from the
median.
6. A knowledge of departure above median expressed in percentage of the time should be helpful especially
where superabundance of rainfall may be harmful to
crops or produce floods.
7. Intelligent long-time planning always takes into
account as many factors as possible. Rainfall is a vari- able factor in Nebraska that always must be consid-
ered. The more we know about it the better planning
we can do. An area with a wide rainfall variability may
present fewer hazards if the percentage of variability is
known and considered when plans for the future are
being made. This variability series of rainfall maps of
Nebraska may offer some possibilities in this connection
that previously could not so well be taken into account.
THE HURRICANE WARNING SERVICE AND ITS REORGANIZATION
By EDGAR B. CALVERT
[Weather Bureau, Washington, April 19351
Tropical cyclones are the meteorological monsters of
the sea. No other type of ocean storm approaches them in violence and dest>ruc tiveness, nor in t,lie persistency with
which they maintain their form and force. I t is a ra.re
thing for one of then1 that has fully developed to dissolve
and disappear over water surfaces in the Tropics. TVit’li
few esceptions t,liey conbinue as violent st’ornis until
they strike land or pass out of the tropics. They are
cudled hurricanes when they occur in the Ablantic. Ocean,
Gulf of Mexico, Caribbean Sea., and the Pacific Ocean ofl
the coast of Central America and Mexico; they are known
as “ty hoons” in the China Sea; “baguios” in t.he water
gal and other portions of the Indian Ocean. By whntever
name they are known and wherever tiiey occur they strike
terror into the hearts of seamen and of people who reside
along the low-lying shores subject to their visitation. No one person or organization can surely be c.reclite,d
with being the first to engage in a systematic forecasting
program for the purpose of giving warning of the approach
of hurricanes. William Reid began his studies of them when on duty on the island of Barbados in 1831. His
work published in 1838 is stsill held in esteem as a book of
reference. He studied hurricanes, plotted their courses,
o.nd formulated his “law of storms”, including rules for
niariners to maneuver so as to avoid their centers; and in 1847 he established the earliest system of displaying
warnings when tlie approach of a storm was indicated by tlie barometer.
The credit for the next warnings is probably due to Fat,her Benito Vifies, who for ninny yeaas was associ- ated with Belen College a.t Habma, Cuba. It is known
t’hat in 1870, when Father Vitie.s became director of the
college, he began to grapple wibh the problem of fore-
warning tlie people of tlie advent of hurricanes whic.11
threatened them. Before that time the inhabit,ants of
Cuba were accustomed to hear of these phenomena only
upon their new approach. This WRS t’lie same year that
the United States Congress made appropiiations for orgnnizing a nationa.1 meteorological service, control of whch wa,s vc.sted in the Signal Corps of the Army (lat,er
transferre,cl to the Agricultural Depart,ment as the Weather Bureau). The Signa.1 Service was scarcely in
posibion to issue hurricane warnings until A u p s t 6, 1873.
On that date arrangements for securing dally weather
reports by cable from Cuba and other islands of the West’ Indies went into effect with rec.eipt of observations from
Habana. Daily reports from Kingston began Sept,ember
18, and from Smtiago de Cub?, on September 29 of the
same year. Plans for obtnirung reports from Puerto
mea o P the Philippines and ‘‘ cyclones ” in the Bay of Ben-
Rico, Guadeloupe, and Barbados did not materialize that year as was espectecl. The met,eorological service of the Signal Corps did not start functioning until 1870, but almost from the begin-
ning the need for issuing warnings of hurricanes to the
people along the southern coasts of the United States WB,S
recognized; also that t,liis could be accomplished only by obt,aining current observat,ions from islands in the West
Indies. Plans for obhining such observations were dis- cussed in the reporbs of the Chief Signal Officer for t,he
fiscal years 1578, 1873, and 1874.
Father Viiies had nearly 3 years’ start on the Signal Service in organizing l i s hurricane warning service.
According to one of his conimentators (Rev. Walter
hcf. Drum, S. J.), the earliest authenticated date on which he issued a warning was September 11, 1875, but
it seems likely that he did so prior to that time. There, too, is uncertainty as to the date of the first hurricane
wwning issued by the Signal Corps. Father Viiies is
generally conceded to be the first meteorologist to fore-
cast hurricanes from observations of the upper as well
as the lower clouds; also, he was the first to announce that both the place of formation and the direckion of
movement of hurricanes change as the season advances.
The August 1873 issue of the MONTHLY WEATHER REVIEW refers to a forecast issued on August 23, 1873,
of stormy weather for tlie New England and Middle L4tlanticz coasts “with cautionary signals at Cape May, New Tork, and New London.” This article contains a
statement to tlie effect that the storni was of tropical origin, that it was by far the worst one since the estab-
lishment of the Signal Corps and that it “did not occur within the limits of our stations.” It is a safe deduc-
t,ion that August 23, 1873, was the date of the first
warning issued in connection with a storm of tropical
ori,gin but it would not be proper to claim that it was
a hurricane warning because the storm was extratropid
when the warnings were put out. Fa,ther Vifies is justly pla.ced in the front, rmlr of those
wlio have contributed to. our knowledge of tropical storms and have been untiring in their efforts to develop
means for forecasting them. Father Vifies’ hurncane forecasting studies were in four directions: First, to find some sign or group of
signs tha.t would invariably prove the existence of a
cyc.lone while it was yet a t a great distance from the observer; second, to get bearings in regard to the whirl-
wind, i. e., to determine what part of the horizon it
was coming from; third, to locate the trajectory or curve along which the cyclone would move, and to do t’his in
MONTHLY m A T H E R REVIEW MARCH 1935
time for the observer to betake himself to a safe distance
from that fatal curve; fourth, to determine the distance of tlie hurricane from tlie observer, its intensity, area,
nncl velocity. A serious setback in the developnient of the hurricane
wnrniiig servke occurred in 1S76 be.cause of discon-
tinuance of observa.tions from the Rest Indies, due to
lack of funds for cabling, but, this difficulty wns pa.rt8inlly
overc,oine by tlie generosit,y of the cable c,ompanie,s in
granting recluc,ed m 6e,s of one-third to one-half of com-
nie,rcial ra t8e,s. In lssl tlie reports from the We.st Inclies again were
suspe.nded partly because of lack of funds but principally
becmise of "doubt being thrown on the legality of spend-
ing any part of the sppropriations in mzi.intnining a stsation outside of tlie 1iniit.s of the United States." This diffi-
culty appears to have been reniored before the nest
se:i,son began. At that time reports were being received from six West 1ndia.n localities, namely, Bn.rbados, Guadeloupe, H:tbann, Kingston, Snnt8in.go de Cuba and St. Thonias. In 1SSS or 1889 other reports, from Puerto
Rico, S:mt,a Cruz, and Antigua., we,re se.c.urec1 t,hrough bhe
newly organized Cuban hiet,eorologic:il Service wliic,li
wis estsblishecl Janusry 15, 1889, under hhe directmion of
t81ie N a i d Observatory of Cuba. Froni thst t,irne on-
ward unt8il t8he ou hbre,n.k of the Sp".nish-Anieiic,nii JV:i,r
tlie records do not clisclose m y material chmge in tlie general operating plans of the hurricane warning service,.
In 189s there W'RS 1vha.t niny he termed the first reorgan-
ization of the hurricane warning service. It was clue to
the Spmish-Aniericnn War. Prior to thn.t time warnings had been confined to our own coa,stal areas, a.lt81iough
there wa.s court,esy exchange of hurricme infornia t,ion bet8ween the Weather Bureau and the Cuban Rleteoro-
logjcal Service and Belen College. bince 1Si3 observers a t West Indian points had been
local residents. They were pnid for taking the observa-
tions but were not subject to any officinl control by our Government. When war was declared in lS9S, it was
recognized a t once that tlie hurricme warning servic,e on the old basis would be wholly inadequate to proleck t.he
large fleet of nnral craft, transports and Unitfed States rnerc,linnt,nien opera t,ing in t8he waters subject to hurricme
visitations. Accordingly, a bill was drafted and sub-
mitted to Congress on June 16, 1S9S, authorizing the Weat,her Bureau to establish and opern.te observing sba-
tions t,liroughout the West Inclies and along the shore,s of
the Caribbean Sea. The n.ct niaking nppropriations for
t'he service was not approved unt8il July 7, 1895. Per- mission had to be 0btaine.d from vniious goveimment's for
placing trained obse~rvers and citizens of the United States
a t places from which observations were desired. It was
obvious that only those whose loyalty and efficiency were
beyond question could be given responsibility in such n.n
important undertaking. Much of this preliniinary work
wa.s clone before the appropriation act was passed on
July 7, lS98. By vigorous action the first party wis
st,arted on July 22, 189S, and before the iniddle of August
trained observers of the Weather Bureau had opened stntions a t Kingston, Jamaica; Port of Spain, Triiiitliid;
Willenistad, Curaca.o ; Santo Doniingo, Santo Doniingo ;
and a t Snntia.go de Cuba. Before the middle of Scp-
tember additional stat'ions were operating a t Bnsse-Terre,
St. Kitts; and Bridgetown, Barbados. The Weather Bureau forecasting center in the West Indies was located
a t Kingston, Jamaica.
The war soon was over; in fnc.t, before there was 011-
port,unity to issue warnings to our conibat8ant8 fleet.
Nevertheless, there neve.r was quest,ion as to the wisdom
of the nc.tion to provide means for guarding our ships against, hurricnnes, one of which unannounc,ed might
have ca,usecl eiiornious damage. Notwithsta.nding the ce,ssa.tion of liosbilities it was clec.idec1 60 complete the
original orgaaization plan. Stations were established a t
Roseau, Doiriiiiica; a t San Juan (soon nfter Puerto Rico
came under Unit'ecl St,ilt,es cont,rol) and at Colon. Original
intentions to place st'nt'ions a t St. Thomas and Barran- clnills were abandoned and the station a t Colon was
closed early in lS99. Hen.dqnarters of the forec.asting service was transferred t,o Hnbana on February 1, 1899. At bhe same t,iine plans were perfecte,cl for giving all parts
of the West 1ndie.s mid ships of all nationalit,ies the bene-
fit, of the hurricnne warning service of the Weather
Burenu. This responsibilit,y lins internationnl recogni-
tion t'o this c1a.y. Wit>liin a year aft.er t.he declnra.tion of pence on De-
ceinber 12, lSOS, observers began t,o be withdrawn from
t81ie West, Indian sta.t,ions n.s fast as capable local persons
co!ilcl be t,ra.ineel and qua.lifie,cl to tdie t,he observations.
ln 1902, at, t8he t,erniination of t,he American occupation,
tlie hurricane foremstsing work for the West Indies was
t misferred froin Habana to Washington. All Weather Bureau employees were wiblidrttwn escep t one who re- mained for t,he purpose, of comple,ting t,he substitution of
local observers for t>liose sent, froin tlie 'CTnit>ed States.
A forecast center for tlie issuing of hurricane warnings
and advices for Puerto Rico and contiguous areas was e,st&lishecl a t Snn Juan on June 1, 1919.
St,eady improvements were made in the efficiency of
the se,rvic.e nfter t,he Spanish War. The largest factor in this iniprovenient has been the development of the
radio. It hn.s added treimndously to t8he field of observa-
t'ion. Far greater dependence now is placed on ship re-
ports than on observations froin island stations. Ships make regular contribut,ions of two observations a day
and additional ones when needed. Last year more than
2 1,000 observat,ioiis were, received from ships in tlie soubh-
ern portions of t8he Nort81i Atlant,ic, tlie Gulf of Mexico aad the Caribbenn Sea during t,lie G months which con-
st,itute what, is known as tlie hurricane sea.son.
A third epochal change in the hurricane warning service will occur on July 1, 1935. Change,s a,s contemplated
on that dat,e, a.re ninde. possible by an item of $80,000
contained in t8he Weather Bureau portion of the agri-
cultural approprintmion bill for the nest fiscal year. Since the beginning of t,he hurricane warning servic.e in 1873, warnings and advices concerning tropical storms
hnve been issued from Washington, e,scept for the short
period during the Spanish War when warnings for the West Indies area w-ere issued from Kingston and Habana,
m d the service re,nde,red a t San Juan since 1919. Under
t.lie new reorganization e.ffective on July 1, 1935, no warn- ings of hropical storms will be issued froni Washington
except for such ns have moved north of labitude 35"N.
It is believed that more espe,clitious, sat,isfactory, and e,ffic,ient service cmi be rendered from forecast centers
located nearer to the scene of action than is Washington. Strictly speaking, nfter that dat,e there will be three
liurricane warning centers as follows:
Sun Jua,n.-Cnrihbe,an Sen nncl islands east of longi- hide i5OW. and south of 1at)itude 20°N.
New Orleans, La..-That portion of the Gulf of Mexic.0 and it,s c.oast,s we.st of longitude 85"W.
Ja.cksonville, Fla.-Re.maining port,ions of the Atlantic, Cnribbean Sea and Gulf of Mesico areas, islands and
coasts south of labitude 35ON.
MARCH 1935 MONTHLY WEATHER REVIEW 87
Jacksonville also will be a district forecast center to
include the States of North Carolina, South Carolina,
Georgia, and the Florida Peninsula. From that center will emanate all daily weather forecasts, warnings of cold waves, frosts, weekly outlooks, etc., for the States
named in addition to storm and hurricane warnings. At present all of this work is being done a t thewashington
forecast center. Jacksonville also will take over from the Washington center twice daily wind and weather fore-
casts for the coastal and contiguous ocean zones, Hatteras to Jacksonville; Jacksonville to the Florida Straits;
the eastern Gulf of Mexico (east of longitude 85OW.) and the Western Caribbean Sea (west of longitude 75OW.)
New Orleans is now a forecast center for extreme north-
west Florida, Alabama, Mississippi, Louisiana, Arkansas, Oklahoma and Texas and storm warnings for the Gulf
Coast from Apalachicola westward to Brownsville. Its principal additional work to be transferred from the Washington center will be the hurricane service for the
area as heretofore described; also daily wind and weather forecasts for the middle Gulf of Mexico (between longi-
tude 85"W. and 9OoW., north of the Yucatan Channel)
and for the Western Gulf of Mexico (west of longitude 9O"W.) will emanate twice daily from New Orleans in-
stead of from Washington. The reorganization program involves material changes
in the collection of observations from ships. Under existing arrangements observations are secured regularly
twice a day throughout the year by radio from a limited
number of selected ships in the Atlantic, Gulf of Mexico,
and Caribbean Sea. These reports are now transmitted
only to Washington. On and after July 1, 1935, selected
ships routed south of latitude 35ON. will transmit their observations addressed both to Washington and to
Jacksonville. The North Atlantic ships will continue
to report only to Washington. Another class of ships, known as cooperative, supplies observations only during the
hurricane season, June to November, inclusive. Hereto- fore service from this class of vessel consisted of two obser-
vations a day, taken at 0000 and 1200 G.C.T., and con-
taining only the universal group of data (day of week, octant of globe, latitude, longitude, direction of wind,
state of weather, wind force, barometric pressure, visi-
bility and temperature). After July 1, and during the
period June 16 to November 15, inclusive, they will take
2 additional observations, a t 0600 and 1800, G.C.T.,
making 4 in all. Observations from cooperative ships
will contain the universal data group and also a supple- mental group (swell, clouds, temperature difference
between air and water, ship's course, barometer change,
and past weather). Cooperative ship reports will be
addressed only to Jacksonville. A system of direct calls for special observations from
ships when a hurricane is in progress was inaugurated in 1934. It proved to be successful and it will be continued
in the reorganization program. This plan, briefly stated,
is as follows: The southern portion of the Atlantic, the
Gulf of Mesico, and the Caribbean Sea is divided into 5O
squares which are numbered. When a hurricane is known
to be in progress, commercial radio stations at Galveston, Port Arthur, Tampa, West Palm Beach, and Ensenada
(Puerto Rico) are informed of the areas (squares) from which special ship reports are desired and the time the
observations should be taken. Operators a t these sta-
tions contact all ships in the speciiied squares and ask that the observations be radioed to a specsed Weather
Bureau forecast center. In this way many valuable
134307-35---2
reports are secured which otherwise would not be avail-
able.
Another feature, not strictly new but on a more system- atized basis, will be cooperation with the Coast Guard, whereby weather observations, including data on swells
and tides, will be obtained from Coast Guard life saving
stations. As a rule observations from Coast Guard stations are not needed until the storm has approached
within a few hundred miles of the coast. They will be collected by telephone, telegraph, and radio and speedily
will be made available to the forecast centers a t New
Orleans and Jacksonville. I n a similar way there will be cooperation with Coast Guard ships both in forwarding
observations taken on board and also in collecting ob-
servations from other ships. By a like plan reports from Coast Guard stations, from Cape Fear River northward to the Delaware Breakwater, will be secured for the fore-
casters a t Washington.
An incidental but important part of the program will be the exchange of warnings and advices issued a t Washing-
ton, Jacksonville, and New Orleans so that each may have knowledge of what the others have done and thereby
avoid conflict and confusion. The Weather Bureau maintains storm warning display
stations a t numerous ports along the coasts affected by
hurricanes. A systematic plan for obtaining special weather, tide and swell observations from those places ie a part of the program. At the present time all hurricane warnings and adviso-
ries are included in a bulletin specially designed for the
use of masters of ships at sea and are broadcast twice daily through the naval radio station (NAA) a t Arling-
ton, Va. This bulletin is copied by other naval stations
and rebroadcast so that all ships in waters subject to tropical storms are warned. This service will be con- tinued. Warnings issued from San Juan, Jacksonville, and New Orleans will be telegraphed to Washington for
inclusion in the bulletin. In addition, a special bulletin
containing weather information, forecasts, hurricane warnings, etc., will be broadcast twice daily during the
hurricane season and on fixed schedules from the naval
radio station a t Key West, Fla. It is expected that com- mercial radio stations will cooperate similarly in the dis- tribution work. A unique and important feature of the reorganization
will be a teletypewriter set-up during the five active
hurricane months. This system is designed for speedy
collection of observations f r o p life saving and storm warning stations, the distribution of land and ship reports
from one forecast center to the other, including inter- vening first order stations, and immediate dissemination
of warnings and advices to the entire coastal area. This teletype circuit will connect bhe Weather Bureau forecast
centers at Jacksonville and New Orleans and its offices at Tampa, Miami, Key West, Pensacola, Mobile, Port
Arthur, Houston, Galveston, Corpus Christi, and Browns- rille. The teletype circuit will operate 24 hours a day,
every day in the week, including Sundays and holidays,
and will be used exclusively by the Weather Bureau. A forecaster will be on duty night and day a t Jackson-
ville and New Orleans and trained observers will be on watch constantly a t all the other offices having teletype
installations. In this way advantage will be taken of
every weather report that has signiiicance, the forecaster will be in continuous touch with the local Weather Bureau officials and through the latter the newspapers and the
public will be kept fully and promptly informed.
88 MONTHLY WEATHER REVIEW MARCH 1935
CHRONOLOGY
1831. Capt. William!Reid began hurricane studies on the island of
1838. Reid published his laws of storms.
1847. Reid established display of signals at approach of storms.
1870. Father Benito Vifies became director of Belen College and inauguratu a hurricane-forecasting service, for Cuba. 1870. February 9: United States Congress made appropriations for a national meteorological service,. 1873. August 6: Daily reports from Cuba and other West Indies islands first received. 1875. September 11: Father Vifies issued first hurricane warning.
1876. Set-back in development of hurricane-warning service due to discontinuance of reports from West Indies. 1881. West Indian reports again suspended; legality of expendi- tures outside of United States questioned. 1889. January 1: Meteorological service for Cuba organized under direction of Naval Observatory of Cuba.
Barbados. 1898. First reorganization of hurricane-warning service to protect American fleet during Spanish-American War; bill for providing funds submitted to Congress June 16; approved July 7. 1898. July 25: First observing station opened a t Kingston, Jamaica, which was made headquarters of hurricane- warning service. 1899. February 1: Headquarters of forecasting service transferred to Habana.
1902. Forecasting service for hurricanes transferred from Habana. 1902. National Meteorological Service of Cuba established.
1919. June 1: Hurricane-warning center for Puerto Rico established a t San Juan.
1935. July 1: Second reorganization of hurricane-warning service; service transferred from Washington to centers at Jack- sonville and New Orleans.
EFFECT OF THE ATLANTIC OCEAN ON TEMPERATURES IN EASTERN UNITED STATES
AS SHOWN BY TEMPERATURE-WIND ROSES'
By KATHARINE B. CLARKE
[5814 Thirty-sewnd Street NW., washington, D. C., March 19351
In the preparation of a thesis concerned with some effects of the Atlantic Ocean upon the climate of eastern
United States, a study was made of the effectiveness of
the ocean in moderating temperatures at various stations. The decrease in temperature ranges along the Atlantic
coast is obvious from maps of the average daily range of
temperature (figs. 82,83, and 84, p. 25, Atlas of American Agriculture, Pt. 11, Climate, Sec. B, Temperature, Sun- shine, and Winds, United States Department of Agri-
culture, Washington, D. C., 1925); there is a very much
smaller daily range along the immediate seashore than
inland. The Brownsville region, southern tip of Florida, Cape Hatteras and Cape Cod have in spring, summer, and
fall the small daily range of 9' to 13' F. A comparison of
the highest and the lowest recorded temperatures (fig. 3,
p. 7, and fig. 6, p. 8, Atlas of American Agriculture, loc. cit.) shows a pronounced moderating effect of the Atlantic Ocean along the coast, but indicates that this does not
extend westward beyond the Appalachian Mountains. To show, by some quantitative and graphical means,
the influence of winds from the direction of the Atlantic
Ocean upon the temperatures of coastal and inland sta- tions, temperature-wind roses were constructed: Data
used for these roses were the 8 a. m. readings of tempera-
ture and wind-direction published on the Washington daily weather maps. Seventeen stations in eastern
United States were chosen, and data for 20 years (1906-
25) for the months of January and July were compiled
and averaged. For each station the following data were
obtained: (1) Average 8 a. m. temperature; (2) fre- quency of winds from the cardinal and semicardinal
points; (3) average temperatures with winds from each direction; (4) the departure, from the 8 a. m. average
temperature , of these average temperatures for each
wind-direc tion. From this information the roses were constructed, as
illustrated by figure 1 , the January and July roses for
Boston: The center part is a simple frequency wind rose. On the 620 January days, Boston had a northwest wind
149 times and a southeast wind 23 times. At a con- venient distance from the center of the rose (the same distance for all directions and for all roses), a point was
chosen as a zero from which to plot temperature depar-
tures; minus departures were represented inside the zero
1 A part of a thesis submitted to the Faculty of Clark University, Worcester, Mass., June 1930, in partid fulNment of the requirements for the degree of master of arts in the department of geography.
point, and plus departures outside. A polygon, which is a perfect octagon, connects the zero points and repre-
sents the simple average 8 a. m. temperatures for the
620 days. A heavy line connects the points that repre- sent the departures; i t forms the temperature-wind rose, a distorted polygon. The amount and kind of distortion
represents the effect upon temperature of winds from the
different directions; on the Boston January rose, for example, the mean departure with a northwesterly wind
is minus 7.7', and with a southeasterly wind, plus 7.5'.
Unfortunately i t is almost impossible to make a cor-
rection for the difTere,nt latitudes from which the winds come; southerly winds usually bring warmer air, and
northerly winds colder air. The greatest interest is in
winds from easterly directions, as the purpose of the
JULY I
FIGURB I
roses is to measure, if possible, marine influence. Local conditions of topography will affect almost every station;
Albany, for example, has a very decided minus departure in winter with southeasterly and easterly winds as well
as with northeasterly winds, since easterly winds bring colder air from the nearby highlands. These temperature-wind roses were placed on a map of
the eastern United States in their respective positions.
Figure 2 shows the roses for January so placed; for pur- poses of better reproduction, the zero polygons have been changed to broken lines. The distortion of the solid-line polygons indicates that departures from the average are
greatest along the New England coast. The roses very
clearly show the plus departures with easterly, south- easterly, and even northeasterly winds. This plus depar-
ture with an easterly wind in January does not appear a t