h Y , 1931 MONTHLY WEATHER REVIEW 188
inform masters a t Sault Ste. Marie of the weather con- ditions on the east end of Lake Superior and the nort,h
end of Lake Huron. As previously stated, the lake ships have been equipped far beyond the requirements of law, but the major portion of the commercial fleets of both Unit8ed St.at,es
and Canada is still dependent upon t.he flag and lantern displays. In addition to these unequipped vessels of
the larger classes engaged in int.erlake trade there are numerous small craft, such as fishing vessels and yachts
to which the flag and lante,rn displays are the only available warnings of threatened storm.
I have no hesitancy, therefore, in stating t,hat no
thought sho~dd be given a t the present time to the with-
drawal or material contraction of the. primitive syst8eni
of flag and lantern displays. On the contrary, if this
service can be expanded to greater usefulness hi tQe saving of life and property the study of the Weat8her
Bureau should be directed to that e,nd. .For instanc.e, these signals show merely t,hat a storm may be espected from a certain point of the compass. If 8 simplerevision
of the code could be arranged to show the anticipated force of the expected storm, the additional informat.ion
would be valuable.
’
I make tlie further suggestion that the Weather Bureau might render ttn added se,rvice by the forecasting ob fluc,tuations in the, lake levels, particularly in the vicinity
of t,he shoal places governing the loading depths of the vessels. Some years ago Mr. Frank Jennin, the meteorol- ogist of t,he Weather Bureau at Alpena, Mich., made an
eshensive study of the effects of baromet’ric pressure on
the lake levels and the currents created by the transit of tfhe high and low pressure areas from one part of a lake to
mother. The nie,rits of Mr. Jermin’s deductions I am not coiiipet8e,iit to discuss int~elligently with you gentle-
Iiieii, but I haw s distinct recollection that Mr. Jermin
said that t,liese water-level fluctuations could be forecast with reasonable accuracy about six hours in advance of t,lieir c)c,currence,. If this he t.rue, may I not recommend
t,o my Weather Bureau frie,nds that consideration be
given to t,he issuance of advame information with refer-
ence t,o these fluct,uations? While the poe,t sings of the “bounding billows” and
the “we,t sheet and flowing sea,” the mariner reads with much gre.at,er concern the indications of his barometer and the reports and warning signals of the Weather Bureau.
The stories the.y tell may not be ever new, but the interest holds longer t h m i t does in other Twice-Told Tales.
SIGNIFICANCE OF AIR AND SEA TEMPERATURES OBTAINED ON CRUISE VI1 OF THE
“CARNEGIE”
Qn the tenth of hf‘ay, 1928, the nonmagnetic ship,
Carnegie, sponsored by the Department of Terrestrial Magnetism of the Carnegie Institution of Washington,
took departure from Newport News on its seventh cruise.
It was possible on this cruise to inaugurate a complete
meteorological program. From the middle of May, 1928,
to the middle of November, 1929, except for days in port,
air pressure, temperature and humidity, and sea tempera-
ture were recorded continuously and a definite efr‘ort was made to obtain as accurate records as possible. These
Camegie observations are particularly valuable because in some regions of the Pacific where the Carnegie cruised
meteorological data of known accuracy are scanty if not altogether lacking. Of the meteorological results which are now being
compiled a t the Department of Terrestrial Magnetism, those of sea and air temperature are the most complete
and accurate.
A continuous record of sea-water tein eratures a t a
mercury-in-steel bulb-and-capillary type of sea-water thermograph with daily movement. Sheets were changed
daily a t Greenwich mean noon. Immediately before
each change of sheet the temperature of the surface sea water was measured by the bucket method. This con-
sisted in lowering a canvas bucket into the sea about two feet below the surface, quickly hauling this to the deck
and measuring the water temperature by immersing a
stmdardized thermometer in the bucket. The tempern- ture so obtained was entered on the thermogram. In
amas where the sea-surface temperature was changing rapid€ y, aa in entering port or in calm weather, a mean of
several bucket readings was taken. The thermograms were scaled a t every full hour local
mean time. The dift’erences between thermograph and bucket readings have been recorded, and these values,
depth of approximately two meters was o E tained with n
1 Bered 011 a paper pmented before the American hleteorologkal society, Washingto-, Ma 4,1931. .41so cf. Brooks, Charles F. Meteorological Program of the Se!enth cruise of& Concgte, l928-lS’29, MONTELY \VPATBIB REVIEW, May, 1929, vol. 67, pp. l~Gl96.
used as a correction factor, applied to the hourly thermo- graph readings to obtain true sea-surface temperatures.
, Bucket temperatures were higher than thermograph
bemperatures by 0?8 C. to 0?9 C. a t lower sea tempera- tures and by 0 3 C. to 001 C. a t higher sea temperatures.
Comparing temperatures so obtained with the sea-sur- face temperatures measured a,t the oceanographic sta- tions wihh standardized reversing thermometers a differ-
ence grecter than 0:s C. never was found and a t over
half the st>ations the difference was less than O ?l C. Values of corrected sea-surface temperatures range from
G?4 C, (4305 F.), recorde,d just south of the Aleutian
Islands a t 12h July 8, 1929, to 3 0 2 C. (8604 F.) ap- pronching Pago Pago a t 14h, November 14, 1929.
For obtaining air temperatures several types of ap-
paratus were used. The Hartinnnn and Braun electric-
resistance multitherniograph was installed for the purpose
of obtaining lapse rates from deck to masthead. Three pairs of wet and dry bulb thermonieters were installed
a t various heights above sea level-one. pair in the Steven-
son shelter on de& 12 feet (3.6 meters) above sea level, another pair in a ventilated scre,en just above the cross- trees on the mainmast 72 feel, (21.9 mebers) above sea
level, sncl n third pair a t the ma.sthen.d on the mainmast
113 feet (34.6 meters) above sea level. These Thermom- eters were cttlibrntecl from time to time with an Assmann
aspiration psychrometer.
The usefulness of these Hftrtmann and Braun records has been lessene,cl be.ca.use corrections for dl the single thermometers c.tm not be obtained. It is evident that
the recorded values depend upon the effic,ienc.y of ven- tilat,ion of the screens, whic.h in turn is m0difie.d by direc- tion and velocity of the wincl. Unfortunately these wind-
records were lost in the destruction of the vessel.
An e,sasiination of these Hart,mann and Braun records has re,ve.aled a diurnal variation in the apparent lapse rates bet,we.en deck and crosst,rees (masthead-records
were too incomplete for use). This must be due to heat-
ing of the de.c,k-t,hermomete,r during the daylight hours.
at *h&slb&eh' pomiple to o8e these Bartmann and Braun
record< in ,coTec,tmg deck-tehperatures for overheating,
as d l be explaheg later. 'The Hegretti-Zafhbra ventilating recording psychrom- eter was 'located in the Stevenson screen on the quarter-
deck: . The recording wet and dry bulb thermometers of
&,his $s t y m p t were calibrated daily at Greenwich mean noon, Ij;s means of an Assmann sling psychrmneter, As soon,as the wet and dry bulb temperatures were read off oxi;the Assmann they were ntered directly on the Negretti-
ZambrgL thermogram. &he Negretti-Zambra traces
have been' spaled at local meap time and hourly values corrected, from Assmapn, reading have been obtained.
$'ram' these values tables of hourly air temperature, relafive hupidity, aQd vapor pressure have been com-
piled. €'artl$ as a contribution to climatology and partly to
qtudy the,, @urn81 variatiop of these elements, mean 4obrly.values for area5 have beep coniputed. The areas, f$enty-two in all, have been selected to represent regions
within ,yGcp small variations of temperature were found,
or regions, with like variations, such 'as the Gulf stream
crossing, "Tb periads,of observation for the areas vary
As mentioned previously, an examination revenled a
didMa1 '"Phrfition in the apparent lapse-rate betmeen deck and crosstree temperatures recorded by the Hart-
niann and Braun instrument due undoubtedly to over- heating of the deck thermometer during the day. Like-
wise a diurnal variation in differences of temperature
recorded by the Hartmann and Braun dry bulb at the chmtrees and the Negretti-Zambra dry bislb in the deck
screen has been discovered. The amplitude of this varia-
tion, however, is not as great as that of the differences in
the two Hattniann .and Braun thermometers presumably
bebause the Negretti-Zhmbra instrument was better venti-
lttted!
It has seemed justifiable to use these curves of differ- enbe$ for computing a correction to be applied to the day-
time hourly mean temperatures by areas recorded by the Ne&et&i-Zambra dry bulb. The curve of differences
during ' daylight hours between Negretti-Zambra dry
bulb on 'deck dnd Hartmann and Braun dry bulb at the croestrees (means for areas) has been applied as a correc-
tion to' the mean values of air temperatures. The result of qpplying these corrections is shown in Figure 6 .
Deshed h e ' bepresents mean air temperature as read
I roffi ' Negretti-Zrtlhbra dry bulb and corrected from
A6srhann I'eadhngs. Broken line represents what the dean' air temperature would be with a correction of the
mean differencks between Hartmann and Braun deck and crosstree temperatdes applied to the Negretti-
Zambra dry-bulb means. Full line represeats air temper- atutef! corrected for the mean differences of Hartmann
and Brauh crosstrees' and Negretti-Zambra deck dry-
bplb temperatures, which is accepted as the most accb- rate ai* temp'erature which can be obtained from the
data available.) There are theTefore corrected hourly
dace temperature and of air temperature
y of the ditferences betwee? air and sea
temperatures was undertaken. Of the dally means for the entire cruise it was found,that in 61.5 er cent of the
tqriperature. ' On the other 38.5 per cent of the days
mean> air temprature exFeeded mean sea temperature. €€ owever, the? daily means of air temperature were not
corrected for overbeatihg and are undoubtedly too high
for actual air temperatures over the sea. Moreover the investigations were all carried out during a summer
from 3 to 35 ii ays.
days, the mean sea temperature excee 0 ed mean air
season or in the tropics and therefore do not represe$t
true annual averages. In comparing the mean differences for areas betwke'n
air and sea temperatures, t'he mean air Cemperbture? corrected for radiation were wed. The difference of sea temperature minus sir temperature w49 never gs great' ad
2?0 C. In only two areas, crossing the, Gulf +ream a ~Q
in the Gulf of Panama, were pienn sea temperatures r$b<e than 1cIO C. higher tlian mean air temperatures.' his
large dieerence of 106 C. over the Gulf stream may be
attributed to the high water temperatures, A dieerFnce of 1'15 C. between mean air and sea temperatures ip the
Gdf of Panama play be explained by the fsct tbat during the entire 12 days of this series the wind was consigtentlq
from the southwest-from a region in which sea tempera-
tures only tt few hundred miles away were 4s much as Eo lower than in the gulf. Thus air considerab1y.cooler than gulf water temperature was imported.
It is, also, interesting to note that of the means for the
areas which include that part of the cruise from Japan
to Ssn Francisco, air temperatures appear to be slightly
higher on the pverage than sea temperatures. Diffprenceg are small, from 0 9 C. to 007 C. The vindg durip ,this part of the cruise usually had a southerly componenf'
In one other area, that centered off the coast-of Chile
approftimately on the western edge of the Perhbibn kt16 rent, the mean air temperature was O f 1 1 C. higher than the rather low mean sea temperature here.
Froin the corrected hourly means for areas tt study of
diurnal variation has been hade. From the litdrature concerning previous investigations of this subject it was espected that the diurnal variation in diderences b'etmeh
air and sea temperatures would be small-sbcht ld C.-.' and that in general air tempehtures would be lower thah' sea tem eratures during the'niglit and wduld abprotich
and pro pb ably exceed sea temperatures dinin$ the dky.' Mean hourly air temperatures for' areas corrected 'fm radiation and for nonperiodic change, and mean h0pfl.p. sen temperatures for areas kofiected for 'nonpkriodic'
change were used. For the areas yhich include that part of the cruise from Iceland to the Mid;North Atlhntic
(1 7 O north, 38O west) and from Barbados tb Caaad none of the 24 mean hourly sir temperatures exekeded the'medh
hourly sea temperatures. FOP all otlier aread the' niean
houply air tehperature at some time ddring' the day Pose higher than the mean sea temperature. Howevek, w h e ~ air temperatures not corrected for radiation were used 1k1
this comparison, for every area except that of the Gulf
stream and of the Gulf of Panama air temp8rathreA exceeded sea temperatures sometime during the da . This seems to indicate that if the e&ct of radiation cot&
be entirely eliminated the mean daily air temperathe would seldom exceed mean daily sea temp&atuie. ' It was noted that in sbme of the arbaa the air tempem- ture exceeded the sea temperature only d u ~g the ho- between 8 and 10 a. rn. In others the air temperdtare rose above sea temperature about 8 or 9 a. m. and 're-
mained above until late afternooh (see fig. 7). A atudy for the cause of this reveale'd that in the areas which hadi an air temperature greater than sea tempwaturb in the
morning and then fell below the rest of. the day, thdt the. air temperature was at a maximu about 10 'a: m. The most plausible explhn'ation for this seemed to 'be found in the cloudiness records from.the log extdct and from some atmospheric-electric observbtiens; which-lin-' dicated that days included in means which had an early
maximum were also days when thO sky clouded over in the late morning and remained cloudy the rest of'the
day, thus producing an effect comparable to that of a
niountain climate in summer.
..
'
(To face p. 184) M. W. R.. May. 1931
FIGUEE 1.-CrulseVII of the Carnegie, May ,1928 ,to November ,1828 (brokenlineshows portion not completed
I
FIGURE 2.-Sea-water thermograph used on Carnegie showing mercury bulb, metal shield, and recording mechahism
(To face p. 185) M. W. R.. May. 1931
FIQURE &-Two thermograms from cruise VI1 of the Carncpie (upper one wa8 obtained on the western edge of the Humboldt Cur- rent; lower one is typical of those recorded on calm days in the Tropics)
FIGURE 4.-The Carnegie showing positions of Hartmann & Braun wet and dry bulb thermometers
M. W. R.. May, 1931 (To face p. 184)
II -
I
FIGURE 5.--Stevenson screen open (on left a pair of wet and dry bulb Hartmann & Braun ther- mometers; in center, Negreth-Zambra ventilating recording psychrometer motor-box outside shelter; Assmann aspiration psychrometer at evtreme right)
FIGURE &-Showing result of correcting excessive daytime tem- peratures recorded on deck, by means of the diurnal variation in differences between temperatures at deck and crosstree
M. W. R., May. 1931 (To face p. 185)
FIGURE 7.-Typical diurnal curves of air and sea temperatures
FIGURE &-Wet and dry bulb lapse rates
9n oHer t b d e t e d h e to what degree' lapse raCes ob-
tdineti. .by means of thermometers a t different heights Above debklare useful; the wet and dry bulb temperature at"d8erent heights on days wheri an Assmann calibra-
tion with' the Hartiniadn and Braun thenhometers was
'mrtde;' 'werd plofted. These d a y t h e ldpse Pates are
shown in Figure-8. The most striking fact'is that these
h t e s hre: decidedly superadiebatio. lodal mean t h e , o b coast of
Iceland: Thd dry bulb a t the masthead Was 1?5 C. lower
'than the declk.de bulb, a lapse &qual to four times the diy adiabatic. ' The wet bulb lapse 'was 1:l C. between
deck and masthead or 'six times the safuraded adiabatic.
The teather was cloudy with a moderate NW. breeze,
sea moderate' wlth surface terriperature of 11
2) January 14, 1929, a t 1Ok local mean time, ehtering the por6 of Clallao: There 'wa$ a dry bulb temperature
laphe of 2?1 C. €Porn deck to crosstrees and of 0:5 C. frorn crosstrees' to tnhsthead, a total lapse of 2?6 C. in
36 rheters or seven times the dry adiabatic. The wet
bdlb lapee mte was l ?O C. between deck and crosstrees or nine times the saburated adiabatic. Whd was SSE.,
force 3, weatheI' cloddy, sea tmpePatiure*l8?8 C. (3) March 13, 1929, at l l h local mean time, approach-
ing the island of Tahitti: The dry bulb lapse rate was
Z:O C. from deck to crosstrees and 008 C. from crosstrees to masthead, a totd of 2?8 C. or seven times the dry adiabatic. We$ bplb lapse was J?l C in 35 meters or
six times the saturated adiabatic. Weather was squally with gentle NW, breeze. Sea-surface temperafure was
28?3 C. If the dock readings are ignored the lapse rates be- tween crosstees and masthead are respectively two, four,
(1) July 29, 1928, a t
c ).
and siq times the dw adiabatic. These are exceedingly
steep, suggesting that even the crosstree temperatures may have been affected by radiation from deck, sails, and shelter. It is entirely possible that such lapse rates could exist-rates as high as ten or twenty times the dry
adiabatic have been observed. Certainly these exces- sive lapse rates do not represent actual a b conditions
ovet the entire ocean fur any length of time. It would seem impossible for such unstable conditions to exist
throughout a layer of air 35 meters thick for mjr length
of time over any great area.
The results of the study of air and sea temperatures obtained on the Carnegie indicate that it is possible to
obtain entirely satisfactory sea-gurface temperatures with the sea thermograph comected by careful bucket readings.
It seems very probable, however, that air temperatures obtained on a ship at sea, particularly in the summer or
in the tropics al'e too high and do not represent actual conditions over the sea. Since differences between sea and air temperatures are usually less than l o C., for
purposes of studying the physical processes of the atmos-
phere it becomes necessary to have air temperatures
accurate to a tenth bf a degree. In otder to obtain temperatures of such degree of accuracy methods must
be devised for obtaining these con6iiuuous temperatures
free from the effect off local heating. Carnegie data in- dicate that if air temperatures a few meters above the sea, free from the effects of insolation on the shelter,
radiat.ion and heated air could be obtained, it would be found that even the mean hourly air temperatures
seldom exceed the sea temperatures. Certainly the sea experts a powerful temperature influence upon the atmosphere.
I $HE '&LECTEDSHIP PROGRAM FOR OCEAN-WEATHER REPORTING BY RADIO
By EDGAR B. CALVERT, Chief, Forecast Division.
[Weather Bureau, Wruhlngton]
~~During. ,the dastl two Tears the Weather Bureau has ' . reports from a s h i ~ available to mare than one meteoro- been actrvely eigaged in lfurthering its share of a project, international in scope,.intended to coordinate and improve
the work of. reporting meteorological conditions at sea by
radio., It is not intended that the new scheme, so far as the bureau's own service' is concerned, shdl supersede
the existing arrangement through which it secures by #radio, ohidly for its own purpose, a considerable daily
cdleckioq of reports fcom ships in the Pacific Ocean and during the 'hurricane season horn ships in ,the South
rAtdantic, Gulf of Mexico; and Caribbean Sea. Though
the new project is distinctively international in character
and is ,the Weather Bureau's contribution to a world-
wide program, it will serve to strengthen materially its
own radio weather service! from ships at sea.
Weather reports from ship have long been used in adyanckg knowledge of ocean meteorology and in supply- ing infomation concerning storms and other atmospheric
cQnditions over the oceans for the beneGt ,of navigation.
In the last quarter of a centqry ships' weather reports have been collected by radio in incceasing numbers,
t4eFeby msbling meteorplogical eexvices tp extend daily
synQptiF ,charts OVW .the oceans and, provide daily fore-
Gaets and, warnings and synoptic weather information by
r(ldiobroadc,asast fQr , use of ships &t sea.
GISTORICAL
I ,
A majority of ocean-going vessels traverse waters
from which weather reports are needed by the meteoro-
To make weather ' logical services of two or more nations.
lo$cal service, a system of interna'tiona€ exchanges must
be set up or officers of ships are charged with much addi-
tional work in taking observations and forwarding reports to each service separately. The need of coordination has long been recogpized.
hlore than 50 years before the invention of, wireless
communication, Lieutenant Mau sought more effec-
tive cooperation in acean meteoroogical Y work. He at- tended the First Meteorological Congress in Brussels k 1853, and advocated the establishment of a uniform mode
of making nautical and meteorological observations on board vessels of war. The result was that this confer-
ence undertook to use a uniform system of meteorological
observations both on land and sea all over the world.
High honors qrere bestowed on Maury both in thia and other countries because of his work in the fields of
oceanography and meteorology, but many may pot
remember that in 1565, he was appointed professor of ineteorolo y in the Virginia Military Institute a t Lex-
ington, wkch possibly was the first recognition in tbis way of 6he science of meteorology by any institution of le arnina .
Devaopnient of wireless communication early the present century brougb t many seriom camplicatiops
that did not enter into the program conceived by Iyfaury. The first wireless message received by the Wea$her Bureau containing a weather observation from a ship a t
sea was in December, 1905. Radio weather service from ships at sea was thereafter extended by the Weather Bureau m d the weather services of other countries, keep-