AUGUST, 1924 MONTHLY WEATHER REVIEM7 387
There a pears to be little doubt that t-he use of the
tirely practicable. When t.his has been thorou lily dem-
tuted in the experimental work, in order to make t.he work less tedious. It is not to be ex ected t.hat all orange growers will
the time of lighting their orchard heat,ers; indeefi such a sudden, radical change is not to be recommended. In all cases tlie fruit grower should continue to use accumt.e, sheltered thermometers to obtain the temperature of the sir in t,he orchard, and when the use of t,lie fruit t,lier- monieter is fiist begun it should be only to supplement the information obtained from the slielt,ered thermometer. The average fruit grower is likely to meet with minor diffi- culties in obtaining fruit temperat,ures a t first, and lie
mercurial P ruit thermometers by the fruit growers is en-
onstrated, other types of thermometers may Q e subsbi-
immediately adopt t, f ie fruit therniometer for re ulat.ing
should not depend on such readings until he is sure he thorough1 understands how to use them. Probab r y some growers will prefer to cont,inue to use the old methods of obtaining the tem eratum, if they feel
not t.lioroughly t.rustworthy. On tlie other hand, the use of t.he fruit thermometers will not be difficult in any wa after t,he orchard lieatin crew has become familiar wit
err, will consider them almost, indispensable in handling orchard heating. During t.lie winter of 1923-24 a large number of records was secured showing t,he temperature inside lemons on the trees during frost,y night,s, in the same manner as that in which the orange t,emperat,ures were secured. The results of these observations will be published Mer.
that tahe men charged with reading t P ie thermometers are
1 them, and it is believed t. H iat eventually most orange grow-
OSCILLATIONS OF THE ATMOSPHERIC CIRCULATION OVER THE NORTH ATLANTIC OCEAN IN THE 25-YEAR PERIOD, 1881-1905
By A. DEFANT
[Innsbruckl
[Translated from the German by W. 11'. Reed; nhstracted and ronhmd by B. M. Varneyl
#s-c/. 513 (2 6 I .d
In the pa er: " Die Verteilung des Luftdruckrs ubrr
month of the ye:rr new chi1rt.s of meiln air-pressure dis-
t.ribution for the wliolc rcgion covered by the dnily synoptic weather charts issued by the Deutsche Seewnrte and the Danish meteorolo icd service. Each chart is
1581-1905. By the basis of them we may obtain the air-pressure snomalies for that period in the region under consideration. This cxt.cnsive matcrinl furnishes an
excellent bn.sis for investigations of nonperiodic changes in the distribution of aiir pressure as well tis of oscillations of the atmos heric circulat,ion over the North AtJantic Ocean and aijncent lands. In the following pages will be set forth some of the most important results of inresti- gations carried out on the basis of these charts. They constitute only a partial elaboration of the mttterial from certain points of view, in order that the investiga- tion should not be too voluminous.
1. Air- rewii.re a.ieonmalics m7er the North Atlantic
the intersections of parallels (5' apart) wit meridians (10' apart), it appears that in the majority of cases t.hc distribution of the anomal belongs to a definite system which is a unit in itself. bence 'it a peared desirable to
the ocean, leavin out af consideration the mtensivc
60' and 10' west longitude ant1 from 75' to 10' north latitude. In all there are included 84 intersect,ion points,
so that the distribution of anomalies was given by 84
values. k
The fact that on the whole the anomaly in the direction of each arnllel had the same sign and the same mtlgni- tude ma es it np ear perniissible to form mean values in
we obtain for each month a mean distribution of i r - pressure aiiomdy in ii north-sou th direction betweeii
75' and 10' norbh latitude. Each value is the mean of the six values between GOu and 10' west longitude. These monthly values ive a t once a satisfactory view of
dem Nordnt. P mtischen Ozeu.n, etc.,"5 were given for en+
the mean of the 25 charts o B a given month for the period
R
Ocean.- If we compare anomalies with mean ressures at
confine the invest.igations to anoma f ies of pressurc? over
ad'oining region o B the European Continent. Tho Nwth At 1 antic was considered to include the region bet,ween
the direction ' R of t. e pwallels for the entire region. Thus
t,he kind and mngilitu f e of the departure of air pressure
4 Geografiska hnnaler, 1924, 11.1, pp. 13-41.
3 Dentschr. dfr Wiener Aknd.. Band 93, 1116.
i 11 the month consitleretl, while in their succession they nfford :I 1iist.ory of air-prcssure sliiftiiigs over the dtlnntic Ocean. This is t.hc first time tlint t i sunimwized repre- sentat.ion of irir-prrssure de artures from normal for a
period of 300 months has Kcen compiled for such an extended region of the cnrth. In this meridional clistrihutioii of air-pressure mom- dies over t,he Atlantic one mny very clearly discern the
n peartinct! nntl recurrence of certciiii charact,eristic types ossible, without making too liberal int.cvpretat,ion o the miiterial, to arraii e the 300
successive cases under four t.ypes, to which can e assigned indices of intensity of the departures occurring in them. In type A there lies over the N0rt.h Atlantic a remion of positiw pressure anomnly. It extends from t%e far north to about latitude 50' north, and on an avera e its
Sout,li of t.his estends a region of neaative anomal centered near l~bitiide 41)' north, which, gradual! diminishing, extends t.o the thermal e uator (10' nort
anonialies, :Ltiiios heric pressure in the north teing
Of the 300 cases? 113, or nearly 35 per cent., fall under this ty e.
hype U shows the opposite distribution; the negative anomaly reaches from the far north to latitude 50°,
with it.s center a t 65'. The positive anomaly extends to latitude 10' north, with its center near 40'. Type B is thus exactly opposed to type A. One hundred and thirty- seven months show the t pe B distribution, or 46 per cent of all cases. Under 1 and B occur in the aggregate
83 per cent of all cases. 17 per cent
determination, belong to two other t-ypes, which again are opposed t.0 each other. Ty e C is related to type r l : the positive anomaly usual f y extends from the far north to 35' north latitude and is centered between 55' ancl50' north. The negative anomaly includes the whole sou thern portion. Twenty-
five cases fall under type C. The esact,ly opposed type D, which is relat,ed to type B, appears in 21 cases.
If we combine types A and C, which show a pressure relatively too high in the north and too low in the south, they together include 138 months, or 46 per cent. Under
%
It is P
o r anomaly.
center lies near 65' north in the vicinity of Ice f and.
latitude). North ancl south have t % erefore op mite
relatively too hig f and in the south relatively too low.
The remain+ excepting four cases in which it was di P cult to make
388 MONTHLY WEATHER REVIEW Auaum, 1924
B and D, in which the anomalies are reversed, occur 158
cases, or 53 per cent. The division into these two o posite types is thus nearly equal. A yearly march in t i! e predominance of one or the other could not be determined; their distribution over the individual months is likewise practically uniforni. The nature and interrelations of the four types are shown in Figure 1.
From this figure, and from maps not here reproduced, it is seen that these types represent primarily oscillations in the intensity, and to a lesser degree in the position, of the two centers of action in the North Atlantic Ocean, namely, the Iceland LOW and the Azores HIGH.
A the Iceland LOW is considerably weakened and t e Azores HIGH is subnormal: the south-north gradient, which is a measure of the rate of atmospheric circulation in the re ion, is thereby diminished. Type B, with opposite fistribution of anomalies, shows an intensifi- cation of t,he pressure gradient between south and north and thus an increase in the intensity of the atmospheric circulation.
In tT
N. Int.
75' 70' 65' 60' 55' 50' 45' 40' 35' 30' 25' 20' 15' IO0
ma. l.-Typea of pressure anomaly over the North Atlanth Ocean
In connection with the more important types A and B, the question was studied as to what extent positive and negative anomalies compensate each other, or, in other words, whether the deficienc of air mass in the region
the r 'on of ositive anomaly. If this were the case, it wo d not e too hazardous to assume that these types are produced mainly by the shifting of air masses in a meridional direction. On the basis of a mathematical discussion of the data, it can be said that there is almost complete com ensation between north and south. Not only in the year P y means is this the case, but
also in the several seasons, though in the latter case not with the same completeness. For the formation of type A, airmasses must be shifted from south to north, and for type B from north to south. It is clear that from these strongly characteristic changes of pressure we may draw conclusions as to the intensity of the atmospheric circulation. In the first case the meridional pressure gradients become less steep and the atmospheric machine works with less force, while in the second case the gradients are stee ened and the machine works with more force.
2. de suecession. ofthe awerent types: Oscillations in atmospheric circulation in the years 1881-1805.-1n Table 1
(Table 4 in the original) the monthly values of air pres-
of negative anomaly probaby f corresponds to excess in
sure anomalies are set forth according to the four types for the 25 years. The attached indices give the intensity of the formation of the type on a scale 0-3, in which there was taken into consideration plainly the amount of the departures;O only in a few cases of unusually great anomaly was the index 4 used. The table shows that in certain years the types A and C predominate, indi- cating intensification of the atmospheric circulation; in others the tvpes B and D, which connote a weakening of the circulation. Because of their close relationship and for the sake of simplicity the types A and C on the one hand and B and D on the other will be combined in the following discussion. If from Table 1 we combine the frequency of the' ty es
uency mrtxima of the type A+ C correspond to the ?requency minima of the type B + D, since in each year the two values must amount to 12. Moreover, the march of frequency of both tvpes makes i t clear that the ressure anomlies do not follow each other arbitrarily, {ut that through a rather long series of years first the one and then the other predominates, each oscillation having
a period of several ye:m.
TABLE l.--.4nomalies of atmospheric pressure over the North Atlanlic Ocean, 1 SS1-1905
into A + C and B + D, i t becomes evident that the P re-
Tear
--
DI
8)
a¶
CI
bi
bl
aa eo
&I
di
bo
aalda bi ai
br CI
n di hi bi
br
do bo idh
eo
CL
-
J F
--
a¶ a1
bi bo dr bi bo bi da ar
('1 Cl
b: hi
HI ci
ndbo edbo br bo
cr bo cr aa
32 br bi br
82 a¶
c) bs
~1 bo ba ~g
br br bo n:
nr ar m a ¶ c l i ba bi hi ho BO
~-
-.-
M
-
01
B:
BO
81
81
a0
br
br
bl
a1
ai
81
no bi bi
bi br
ni aa
a1
br bi br
el
Cl
-
-
1\1
br
bo a1
a1
a1
bi
br
-
aa
80
81
aa
81
a0
ai bi
br br bo at
br
bi hi
hr br
R I
-
._
J
-
bu eo
bi hi
bi aa dl br bo
a1
ni
81
a1
br
a1
ai
ho hi
bi
dl
bl
al
81
ill
-
-
J
di
br
-
n1
n1
bi bi
Bo
a1 bl
bo b
bi br
bi br bi br
br
81
Ca
aa
a1
R1
du
do
-
-.
-4
bo bo bi bi
bi
bi bi
bi
ai
bi ai
-
a8
aa
Bo
aa
CI
dr bi
di
110
ni
a0
bo
31
ai
--
- __
s
ai
bi
bo dr bi
-
81
el ni
81
eo
br br BO
8)
di
____ CJ
dl bi bi
bz ai
hi bo
:I1
.-
0
br bi br br ca
br
81
8 )
bo ai
br BI
81
81
a1
a¶
dr hi
hi
bi IN
81
br a1
m
-
N
br br bi
ar
-
er
80
bi br
ai
ar b: bi
81
80
CI
hi
C l ?
hi
ar dl
bo
BI
ai
-
'The year runs from December.to November, incluslve, thus keeping the wlntcr months. sprlng months, etc., together. Eence December. 1885, for example, falls In the year 1888.
In order better to espress the intensity and ext.enb of tht' anomalies for the urposes of Table 3 and Figure 2, each case with index 0 &able 1) has been given the wei h t of 1 ,
year the sum of all cases was obbained bv adding the index number to the frequency number. S h e the type B + D showed opposite distribution to the type A+ C, if a negative si be applied to the B +D ty e, the sum
atmosp eric circulation in the re ion. These are given in the "difference" column of Ta le 2, (Table 5 in t,he original), the smoothed values being given i n the last columii, where a positive number indicates a predonii- nLtnc.e of the A + C type and thus a weakening of the circu- lation, and a negative number a pred0minnnc.e of the type B + D and an intensification of the circulation.
each case with index 1 the weight 2, etc., and B or each
(A+ C) - (B+%) becomes an ex ression o B the mean anomal for a given year and o P the intensity of the
$ E
6 No ernct statrinent or the criteria on which the indlces were bDscd is given by tbc RUthiJr. --Ed.
AUGUST, 1924 MONTHLY WEATHER REVIEW 389
TABLE 2.-Frequency of the types A+C and B+D, .icrith reslwct io their strength and dtwlopmsnt
. . . . ...
Difference
n
+in
-1 4
-i
-1
+S
+I +10 -9
-13
t-2 1-1.i
+I C
-15
+r
0 --r,
- 17 --8
--I
-1
-1.7
--s. 8
+IJ. 2 - ., I .
- -
+r,. IC
+:. 5
+li 5
-.I . 2
-8.2
+l. 5
+l'.O
+5. 0 -2. 5
-1.8
-11 5
-i. 2 -12.0 4. 5
-4. 2
- 1). ,-8
+.i. 8
+0. 8
-10.8 -13.3
+;. s
We find as periods of weakened atmospheric circu-
1855-1S85, especially 1SS5 and 1885;
1891-1893, especially 1593 and 1593; 1902-1903, especially 1902;
and as periods of intensified circulation t,he yenrs
18824884, especially 1882;
1889-1890, especially 1S90;
1894; 1897-1901, especially 1594 and 1598:
1904-1905, especially 1904,.
lation the years
Hence, if one considers only the more significant es- tremes, the smoothed values show a very uniform march and a mean interval from mnsimum to masimuni or from minimum to minimum of about eisht yea.rs. 3. Osdlatwne of the meridion.4~l pressitre grad.ie.nt. between 1atitud.es SOo and 6,'i' ?aorf?1~-~4no ther expression for the intensity of the atmospheric circulntion over the North Atlantic Ocean and a corn arison wit.h t,he results already given has been obtainea from a st,udy of t.he oscillations of the meridional pressure gradient. By deriving the monthly anomalies of meridional pressure gradient between m y two ptwallels, and letting a posi- tive si n indicate increase in gradient and a negative sign a iminution, we ex ress the de artures of tlic nienn meridional pressure gragent, froni t e normal.
From a table of these departures [not here reproducea] it is evident that extremely marked deviat.ions occur in the several months. The extreme values are -1S.9
mm. in January, 1S81, and 13.3 ium. in February, 1903.
The positive and negative values do not follow each other in an irregular manner, but maintain the same sign for several months and then give place to departures of
op osite sign. &here appears to be no simple relation in the sequence of these periods.
In the case of the annual means of the anomalies, the alternation of ty e agrees e-sactl with the alternation of the ty es A + 8 and 13 + D. d e s e annual means are expressecffor the purposes of figure 3 in percent.ages of the normal gradient, and are shown in the second curve.
4. West-ea.& gr~die~l.t i n the fc,r north and: the .ineridiunlrl
pressrue gradient over th.e North dtla,~.fic.-Inspection of
monthly anomaly charts prepared in connection with
\ 8
this study and of tables showin the monthl and annual anomalies of pressure radient fetween the K orth Atlan-
the meridional pressure gradients and the radient be.tween the North Atlantic and Europe. fuomaly values were found for the east-west gradient in the re 'on
tuclc, and the gra hic representation of them is presented
ns number 4 in Jgoro ?, a positive sign there indicating an increase nnd a iiepative sign a decrease in the pressure grdien t. of the type
gradient with the curve of east-west gradient, shows a
tic and Europe revea f ed rt striking parallelism between
from 0' to 40' west longitude and 60' to 75' north ff ati-
(A+ C) - (B + U), or even of the meri 3 ional pressure Comparison of the curve of frequenc
1881 5 1890 5 1900 5
+!6
12
8
I 4 Frequencyof the types
0 (a d - (btd)
-4
-8
I -I 2
-10 I1
-30 -16
-20
Msrfdlonal Pressure
0 Gradient.
+IO
+20
+30 -111
VOlUnlC Ash
Eruptions.
mm. - 20
-10 IV ,W-E Pressure Gradkbt
0 In Northern Europe.
100 +IO
80 +20
60 V Chuacter of the Ice
Year near Iceland. 40
20
0
Vii.. 2.-V.uintious irr ntmosplieric circulation over tllc! North Allnutic (Jcetlu, und relntril phenomenn
complete pttrallelism. Not only in the yearly means is this connection so well defined, but also when we com- >are the several seasons we find the same agreement. bhis fact can best be demonstritted by giving the corre- lation factors between meridional pressure gradient and the west-east, gradient.:
Winter ._____ +O. 76 f0.06 Spring-.. .. . . - ___. +O. 85 fO. 04 Summer ... . . ..__ +O. 59 fO. 09 Autumn .__ - +O. SO +O. 05
Pear ___________________ $0. 7 U fO. 003
l'lie connection seems to be least lainly shown in
hivh value of 0.8 with a probable error of 1/16 of T . The vJue for the yearly mean is calculated on the basis of
300 successive values, the probable error is 1/250 part of
I', .and t.he relat.ion is thus atremely close.
summer: in the other seasons the come f ation reaches the
390 MONTHLY WEATHER REVIEW AUGUST, 1924
Drcemher __._______ ._ +O. 778
January .___. ___ __ __. - + . 818 February ________.____ + . SS5 Pvlarrli ___________
~
___ + . S51 A!ril---- - - - - - - - - - - - - - - + . 755
Since in the formation of both the meridional and the west-east gradient there is ?ne oint in common, the
duced primarily by oscillations in pressure a t the common point. This oint coincides in general with the central
said that at times of weakening of this center of action there is conveyed thither for the production of the osi- tive ressure anomaly air both from the Azores ti h
pression, air masses are shifted chiefly to the south and west. A similar shifting of air masses on the occasion of
chan es in the pressure gradient between central Europe and a t e North Atlantic Ocean was to he expected, but investi ation showed that such is not the case. It tip-
gea are sionificant in diminishing the influence of the pressure chibution to the northwest-that is, over central Europe. 5. Relation of pressure gradient to oceanic circda f ion .in the northern part o the At7ant.i.c Oeean.4scillations of
produce and- be accom anied by corresponding oscilla-
circulation the Gulf Stream flow must be strengt ened and thereby its influence extended farther to the nort,li; especially in this case does the Irminger Current considerably in extent and volume toward Icelapd. polar front between the warm Gulf Stream Drift and the cold Polar Current thus lies farther north than normally. In contrast to this, in gears with weak atmospheric cir- culation the cold, ics-bearing Polar Current gains in force and estent and drives toward the south the weak- ened branches of the Gulf Stream, especially the western and eastern divisions of the Irminger Current. The polar front then lies south of its normal position and Iceland is encircled by the branches of the Polar Current. It is therefore to be espected that a year with weak ittmos- pheric circulation will be a year with abundant and long continued ice near Iceland, while with stronger circulation the quantity of ice will be less and the ice season shorter. On the basis of studies by W. Meinardus on the oscil- lations of the ice drift near Iceland, this line of reasoning was carried out in detail. The results are shown in the curve at the bottom of Figure 2. This curve is based on
fi ures showing the character of the ice year as given by deinardus, each month in which the ice was especially thick being assi ned a double weight. I t is clear that
weak atmospheric circulation, and that years having little ice are ears wit-h strong circulation. The correla- tion factor ietween the meridional pressure gradient over the North Atlantic and the character of the ice year near Iceland is: r equals -0.59 f 0.09
and between the west-east gradient over northern Europe and the character of the ice year near Iceland:
r equals+0.71 f0.06,
indicating that the two phenomena are very closely re- lated.
6. Relation between osdl.at.ions in the strength of the North Atlantic northeast trade and the atmospheric cir-
&tion in the temperate latitudes.-The pressure from latitude 30° southward gives a measure of t c force
close connection indicates that t R e anomalies are pro-
position of t E e Icelandic depression. Thus it can be
and P rom northern Europe. With deepening of the fe-
ears t a at pressure oscillations over the Mediterranean
atmospheric circu f ation over the North Atlantic must
tions in the water circu r ation. E i t h strong atmos heric
K
years with abun I f ant ice always coincide with years having
rdient
June ___..__._________ +O. 506
July- ____ __. ____ __ ---+ . 582 August, + .662
September ____________ + . 622 October ___--_____-___ + . 696
The Irminger Current constitutes the recurve on the northwest side of the Gulf Stream Drift south and southwest of Iceland.-Rd.
I Annalen der Hydrographic und Marit. Meteotologie, Jnhm. 1906.
AUGUST, 1924 MONTHLY WEATHER REVIEW 391
In order to see further how the relation, pointed out by Shaw, between the circulation in lower and hi her lati-
calculatel the correlation factor for the precipitation in England and the annual mean pressure gradient over the southern part of the North Atlantic Ocean (latitudes 30’ north to 10’ north). For the amount of rain over Eng- lmd there was taken the mean of the yearly totals in percenta es of 50-year means at the five stations of Greenwic 5 , Stonyhurst, Sertthwdte, Edinburgh, and Rothesay.” The correlation factor resulting was r equals 0.42. It is thus relathely small, but on closer ins ection of the months [table, not here reproduced] an B of the individual stations it is very noticeable t,hat in both series the estreme values-fall in the same ear, a
The combination of all the values reduced the correlation factor to the small one iven, but nevert.heless t,here
7. On the cau.sm cf osei7lntion.s .in the atmospheric circu.Znfion..-$tud of the nionthly pressure anomalies
these parts of the earth’s surface rather signific.ant oscillations in the atmospheric cirmlation take place, which, if a11 extremes are considered, appear to succeed each other irregularly in a period of some three to five years. A definite system in the succession does not appear to exist, while the shortness and variability of the period did not warrant the espectation that there is a
relation to the rather regular changes in sunspot numbers. 1x1 the individual months, of come, there appeared to be some indication of such a relation, but on the whole a
definite connection ,can not be h e d . The eriod
in the matter, as it, includes qnly two sunspot periods. In this epoch the sunspot mamma occurred in tho years 1S84, 1894, and 1905 and the minima in 1889 and 1901. W. J. Humphreya has dealt with the question of an influence on temperature, and thereby on the general conditions of the circulation in our atmosphere, exerted by gigantic volcanic em tions. He
matic changes on the earth have been caused by the eruption of ashes from volcanoes: Even if he seems to
hare gone somewhat too far in his line of reasoning-as
W. Koppen has demonskated in his paper, “Lufttem- peraturen, Sonnenflecke und Vulkansausbruche ” still
it appears that he was entirely correct in his belief that major eruptions of loose materials from qolcanoes are a
factor which is able to exert a profound influence upon the climabic conditions of our earth. In an earlier paper Is the writer sought to show how three factors are concerned in the determination of .cli- matic changes, name1 , solar radiation, earth radiation,
primarily takes the r81e of a regulator of meridional temperature distribution, and, due to that fact, it enters extremely intimately into the mechanics of climatic changes. Variations in the first two factors, however, are to be considered above everything else as prunary causes of climatic change. If one wishes to regard oscillations of the solar constant as too small to be of climatic significance (which may not be .the case, our knowledge on this point being still too lunited), we have
tudes, ap ears in the now more extended data, 5 t ere was
fact which increases the impression of a direct re T ation.
appears t,o be a re1at)ion in t 7 le sense meant by Shaw.
over the North 1 tlantic Ocean have shown t,hat over
investigated is too short to admit of drawing a conc P usion
In an eshaustive work
came to the conc.lusion that many, if not a1 f , of the cli-
and the circulat-ion o ? the atmosphere. The last factor
~~~
11 Seo Hellnm, G.. Uotersachungun fiber dle Schwankuneen der Niedem-.
18 Huaphreys, W. J., Volcanic dust ani other factors in the production Of CllmntiC Bulletin of Mount Weather Observa-
1) Ddant, A. pie Zfrkulation der Atinosphem in den prnlsslgten Breiten der E&.
Ahhandl. der pmusr. meteor
changes and their pwible celation to ice ages. tory, 6. part 1 .1 ~. &?e also Physics of the air. Part IV, chapters 3 and 4:
Geomaftshr Annder, !W1, H. 3.
Knstituts. Bd. 111, no. 1. le.
remaining as phenomena capable of sffecting the tem- erature equilibrium of the earth’s surface and of the
Ewer layers of the atmosphere, only disturbances in the optical qualities of the earth’s atmosphere. Enormous eruptions of loose material from volcanoes appear, as the most recent occurrences have forceful1 shown US, to
bumphreys has found, the enve ope of volcanic turbidity must be 30 times as effective in obstruc.ting solar r a d b tion as in the repression of earth radiation. The fine, long-continuing dust veil of H volcanic ash therefore have a reversed hothouse effect, continuance the the earth would effect might be aompmed to that of a small diminution of the solar constant. These considerations make it appear as not improbable that in the osc.illations of atmospheric circulation over the North Atlantic Ocean which we have recognized for the period 1581-1905,
there are indications which su port the above line of
influence due to the great eruptions of volcanic ash uring
be sufficient1 powerful to produce pertur 5 ations in the
eneral circu T ation of the atmos here. In such a case, as
P
reasoning, reflecting in particu P nr, that is to sy, .an
-
this period. In the st,udy of t,his question, only the greater ash eruDtions were considered, since only these, and not the eruptions with lava flows, ase determinin fwton.
which the masses or ashes and fine dust are too small and nre carried to too slight a height (lower part of the troposhere), can exert no lasting. influence. In the com ilation of these greater erupt,lons, use was made of g. Sapper’s *‘ Beitrage zur Geographie der tatigen Vulkane,’ in which he classifies t.he great ash eruptions according to their intensity and to the mass of material ejected.
Tabulating the eruptions duri the period 1881-
weivht of 4, to those of the second order a weight of 2, anzso on, we find the years 1553, 1886, 1888, and 1902 especially noteworthy. In Figure 2, below t,he curve for the meridional pressure gradient over the North Atlantic, these years with the greater eruptions of ashes are marked by vertical rectangles according. to their weights. We observe at once that in the vicinity of those years, but especial1 soon after them, marked
This holds especially for the eruption years 1886 and 1888, lS92 and 1902. The fact that there is missing the year 1883, in which by the Krakatoa eruption enormous amounts of volcanic dust got into the atmos- phere, is not of great significance, since in this giant eruption the dust masses were hurled far into the stmtos- phere, with the result that, in line with Humphrep’s view, the chief effect on the atmosphere must have been shifted one or probably even two years. Indeed, we see that from 1883 to 1888 there is acontinual decrease in the intensity of the atmos heric c!rculation,
between those Tears. Also the years 1893 and 1902 show a diniinution in atmospheric circulation. Henco
it certainly does pot a pear to be too hazardous to assert that the major ad! eruptions are accompanied by
a decrense in atmospheric. circulation.
Comparison of the two graphs shows us still another strilring phenomenon. In the first or second yean after those in which eruptions occurred and diminution in
Furthennore, the oreat number of small out f reaks, in
1905 and assigning to eruptions Y o the first order the
disturbances occur in t x e atniospheric circulation.
which agrees directly with the marke a volcanic activity
1s Zeitschr. fur Vulkanolopie. 1917, BmJ 111.
MONTHLY WEATHER REVIEW
atmospheric circulation appeared, there takes place an abru t increase in the strength of the circulation. We see tks at the end of each eruption period-in the year3 1889-1890 following 1888, in 1894 following 1892, and in 1903-1904 following 1902. Inspection of the indi- vidual casea showed that the conditions ap eared to be not inde endent of the magnitude of t % e eruption.
circulation are in general small, pressure gradients departing less than 2 per cent from the normal on the average. Which is to sa that in general the dis- turbances compensate ea& other, since there often occur lesser decreases and increases, which are probably the remains of earlier disturbances. But in the years of eruption the circulation is greatly disturbed, and there is found first a weakening of it which even in the avem e
normd pressure gradient mounts to 10 per cmt, whereas in the more marked cases it reaches, even on the average, almost 20 per cent. But the circulation soon returns to the normal intensity, and indeed we find in the year after the eruption even an intensification of it. This increase in intensity continues and in the second year reaches 7 per cent and in the more marked cases 17
per cent. Thereafter the circulation again approaches normal conditions. Disturbances in the equilibrium of the atmosphere due to great volcanic eruptions are thus extremely charactenstic. First the atmospheric turbidity caused
by the eruption induces a weakening of the general cmulation. This soon ceases, however, and gives place to an increase in i t intensit , which, after approximately two years, varying accorc&g to the strength of the volcanic outbreak, reaches a maximum, only to decline It appears as if the atmospheric circulation, w Y?. en thrown out of equilibrium by the disturbance result’ from volcanic em tions, proceeds to oscillate about%e position of equili%rium. The amplitudes and the periods of these oscillations depend both on the intensity of the volcanic upheaval and on the duration of the optical disturbances. After a marked weakening in circulation and after the cessation of turbidity there follows an increase in intensity, and so on. We may develop these conditions somewhat in detail, using the consideration which are found in the earlier investigation by the author, in order to reveal the causes of the pendulum swings of the atmosphere about its osition of e uilibrium, swings which may a pro riately
nom$ temperature radient between low and high
interchange of air between h q h and low latitudes should remain in a certain conditiou of equilibrium. If any one of these factors is altered, the condition of equilibrium is disturbed and the circulation now executes oscillations about its position of e uilibrium, as does every other
pulsations is to be sought in the inertia of current condi- tions, since both the more constant tropical circulation and the exchange of air between lower and hi her lati-
t,hemselves to the modified conditions of an altered tern- perature gradient or to chan ed conditions of insolation
Hence the circulation once thrown out of equilihriuni varies from it now in one direction, now in another. It will be now stronger, now weaker, and these “beak” will gradually dimmish in force. The damped oscilla-
Before t B e eruption the disturbances in the general
of all cases is very great, since the reduction of t a e
{e des’ nate 3 as “pulsations of the atrnospEere! The &nos P eric circulation is a current system which, with
latitudes, with nom 3 outward radiation and normal
system when so distur B ed. The chief cause of these
tudes of the Temperate Zone will not suddeny P adjust
and earth radiation, but wi 6 always lag behind these.
AUOUBT, 1924
tions about the position of rest must follow a period which
is determined b the structure of the atmospheric
words, we have here to deal with a phenomenon which is similar to the free oscillation of a system, and these pulsations of definite period can be designated ns such.
The atmosphere and its circulation are a t first in normal condition. By a great volcanic eru tion the tem erature in the lower latitudes is considerab T y lowered. thereby the meridional temperature gradient is automatically decreased, with the result that the closed circulation in the Tr0pic.s and subtropics and the exchan e of air in higher latitudes iinde oes a weakening. Tais weaken-
once initiated. Later, due to lessened eschange between lower arid higher latitudes, it will cause an mcrease in
ineridional te~npern.ture gmdient, extendin farther and
and causing an intensification of the circulation. After this again a weakening in the temperature gradient .ce and correspoudingly a weakening in the
subse uent to :i.n original weakenin of tce circulation,
and so on.
The eriod of t h e damped and subsiding pulsations
equations from which one must proceed being given in the author’s a er cited above. The mathematical
of course be considerable. It app$ars, however, that this fundaiiien tal period iiiay be obtained from observations. The years 1883-lSSS were characterized by great erup- tions, but in this interval the period can not well reach complete develo ment, since one turbidit in part over-
the resulting phenomena. But after this time until 1902
ver marked eruptions took place only in 1892 and 1898,
a n 8 the phenomena resulting from the great eruption eriod 1SS3-188s could develop almost undisturbed. !“he second curve of Figure 3 shows that in this period maxiplum follows minmuni dinost regularly, so that the mean period calculated from both m&uima and minima amounts to 3.5 years, which is, therefore, probably the natural period of the pulsations of the atmospheric circulation. The circulation when once thrown out of equilibrium swings back and forth and thereby causes lon period c.limatic changes. The ulsations become
Lows, since the ainplitude of the oscillations becomes smaller from swing to swing. Only when .a new disturbance, a new and mighty eruption, takes place, do the pulsations attain a greater amplitude. It is then a matter as to the point of time at which, in the fadin pulsations of an earlier disturbance, the new one takes p f ace whether the vibrations are now to proceed reinforced in the same phase, or, on the con- trary, with a shift in phase. In the ear 1892 the eru -
new disturbance is in the same phase as the earlier pulsations. In 1898, on the other hand, the eruption, only a weak one, to be sure, appeals not to have been in the same phase, hence the penod of marked disturbance, in 1903 is the fiivt to bring a revival of the pulsations with suddenly increasing amplitudes. It ia estraordinarily striking that the same period of
3.5 years, which we have here designated as the natural period of the atmospheric circulation when its equi-
circulation and t i e dimensions of the earth. In other
Let us consider one such case more closely.
ins will proceed of itsel 7 at first as the result of movement
fnrt,lier because of the gradual fading of t 5 e turbidity,
strengti P13 of the circulation. So, after a oiven interval
there 9 ollows an intensification and nffer this a weakening,
should 1 e obtainable in u purely theoretical manner, the
difficulties whic T R i evet the solution of, this problem may
l a p another an s there is disturbance in t Tl e evolution of
wra %- ually extinguished, however, aa t E e caae after 1888
tion takes place in a year of weakene d circulation and t % e
IIUQUBT, 1924 MONTHLY WEATHER BEVIEW 3g3
librium is disturbed, has, it would seem, a very general character. This has been demonstrated unequivocally in the work of the Solar Physics Committee, “Mean monthly values of barometric ressure,” by N. Lockyer.15
C. Braak has investi ated
temperature and the pressure conditions as wen as in precipitation, and that the periodic changes in the nor&: south gradients between Australia and the East Indies are t,he “faithful companions” of these pulsationa.
we are dealing with a phenomenon of extreme im ortance in the weather sequencm of the longer periods 0 time.
t 1 is in detail for the Dutch
East Indies and has s % own that it is reflected in both the I t is not improbable that in this approximate 3.5 year period in the pulsations of the atmoapheric circulation
P
~---___
U Lack er N.. Solar Physlcs Laborato South Kensington. 1908. u B~&. 6., ~~~o d i ~e h e Klh.d%d&-n Meteor, zeitsct~., 1~10. pp. 121-124. Die 8.tLjlihrlge Barometerperiode. Meteor. Zeltschr., 1912. p p 1-7.
TORNADO NEAR FITCHBURG, MASS., JULY 17, 1924
By CHARLES F. BROOKS
(Clark University, Worcester, Mass.. July 19, 1924)
About noon, July 17, 1924, with the arrival of the thundersquall on a marked wind-shift line, a tornado hit here and there don a path rtbout, 18 nliles long from near Templeton, throu 5 Gardner, West-minster, southern Fitchburg, and d n l o m , Mass. This course, averaging from west by south, when rojected toward the east-
Newbu ort, where torrential rain occurred shortly after. E r e and there dong the path there were of trees destro ed, factories nearly demolished, roo s and upper stones &own off, nnd chimnevs generally blown down. So localized was the damage, however, that little was to be observed, for exam le, from the main thorough-
Two were killed, and damage estimated at $500,000 to
$1,000,000 was done, accordin to the Worcester Telegram.
narrow and direct path in which destruction occurre{ would lead one to suspect the action of a tornado, eye observation of the funnel cloud b Leroy Moreland and
forest trees there leave no doubt as to the whirling nnture of the wind. With a geological compass I climbed over the fallen trees a t Whalom, and obtained their direckion of fall. Most were blown down from IL west soutli- wester1 direction, but near the northern ecl e of the
seen the trees were blown down from south and north as
well as from west. A barn that was hit, blew down nortli- wards, the north wall being blown out lower cnd first, and boards being carried a few tens of feet clenr of the general wreckage. A few details will be ap r0print.e. After blowingadown
southern part of Fitchburg, the storm for the next 2
miles before reacliin Wlialoni broke off or uprooted many lar e trees, bEw down chimneys and damagzd
a tent carnival blown rtway, aiid n sniall grove of ines
up. A proachin Whalom, a large elm, well rooted, but
The tree was about 3% feet in diameter, and was said to have been growing there for at least 168 years. The upturned roots reached to a height of nearly 20 feet. A
number of other trees were uprooted or blown‘ down in t h i s vicinit . A local resident said that another member of the ram& had seen II funnel cloud. At Whalom Park about half the trees (mostly pines 1 to 1% feet in dia- meter) were blown down by u rooting or breaking. They
northeast, passes near Lowel, P Lawrence? Haverhill, and
Y-
fare from Leominster to Fitc f iburg ac.ross the storm path.
While the local severity of t E e damage and the general1
others a t and near Whaloni, an K the criss-cross fnll of
zone o 9 destruction, where the funnel cloud a ad becii
or partly wrecking some F actories aiid tenements in thc
some roo H s. A ball park fence was partly blown clown,
demolished, the trees generally being broken off ha1 P way
expose B on a hi1 P , was blown down from the southwest.
lay mostly from southwest % y west (compass bearing
17 This &but a brier amount Cli plugs from the local press of im ortaut laces affected or nearby cities, and also a iew p\otographs are on file at the 8. 8. datther Rurea; office. Boston, MRSS. This published report.is based on D brief tour of observation through the region of patest damgge near Fitchburg, and on some of the newspaper acmuntn for other portions or the path or the heavy storm.
W.30130° S.), though a few la on top from about west northwest (compass W. 30’ d). Th~s was one-half to three-quarters mile south of the path of the funnel center. Across the road from Whnlom Park, in a grove one-quarter to one-half mile from the ath, about two- thirds of the trees were blown down. %ese were larger than those in Whnlom Park, ranging up to 2 feet in diameter. About as many were broken off as uprooted. The were blown down from the same directions as those in h a l o r n Park, though there were more from the west northwest, perhaps a quarter of the total, as compared with only a few in Whalom Park. A little farther north, at Sunn side Farm, I came upon
scribing the storm, d. Leroy Moreland, who with his father manages the farm, said he was between the two barns that were not blown down (the more southerly of three) *when he saw what he thou ht at h t was the smoke froni a bad fire in the woo % s westnorthwest of
him. coniinm straight towards him. It appeared white. 8, seeme8 to be bouncing up and down somewhat as it approached. Suddenly it turned at about a right angle just in time to avoid all but one of the barns. He had never before seen a cloud anythin like that. The noise
the storm, Mr. Moreland had to take shelter. He said the tornado struck at 12:20 or 13325 p. m. (At Fitch- burg, 2% miles away the time of the storm was re orted as 12:15.) b o t h e r ninn a t the same farm said i! e saw a whirling cloud a proaching, and that it had become estraordinarily dart. Others, a t Whalom Park, had not seen the tornado cloud. Anyway, trees would have re- vented their viewing it. One man with whom I t&ed said he had seen a small funnel cloud at Manchester, N. H., at about the time of the storm here. The mised forest. through which Mr. Moreland had seen the funnel cloud come was about half destroyed, an open gap being cut about 50 feet wide from the west- northwest where the center passed. South of this gap some individual trees and clumps were blown down from the southsoutlieast (compass bearing S. 5’ E.), and a few from the southwest, but the great majority lay from be- tween west by south and westnorthwest (compass W. to
W. 30’ N.). North of the gap about half the lnrge and small oaks, maples, birches, etc., were down, mostly from the westnorthwest (compass W. 30” N.). There were several, however, from the northnorthwest (compass bearing N. 10’ W.) under those from more westerly directions. F’rom 250 to 300 yards beyond this was the barn that was blown down at about the place where the funnel cloud turned. About 50 yards south of this barn a silo had gone down and the corner of 0 barn roof had been blown off, while a few yards farther all seven chimneys of the well-built farmhouse had been blown
the region within the ath of t E e funnel cloud. In de-
There was a ragged cloud mass, whirling violent1
was terrific. Unfortunately for !? urther observation of