-, This paper not to be cited without prior reference·totheauthor·

Bun(~sforschvn

the Belts~':.This·.latter .effect surely does:not. depend so much. on meteoro-•.h·•. ·__.• ,v _"'_'~'.'_'~"'"''~''''' "' ~.•_ _._ 0.' __ ,.. '" •.••. " __ " _ .• _. _•• ' ••:.. "

logical happenings as on a more or less strong influx of oceanic water

into"'the:northern North Sea and Skagerrak." Indeed" it can be shown that.,: \':' ! ',". •• •

in thc p6st~war period a major Baltic inflow has accompanied each inter-

annual salinity maximum in the deep water of the Skagerrak. The

hydrographie record from the Baltic deeps is less detailed in earlier

years but is sufficiently good in the inter-war period to show a simi1ar

concurrence of events.

The periodic recurrence of high-sa1inity conditions in the bottom

.layers of,the Skagerrak can in turn be 1inked ,to broader hydrographie and.... . .meteorological events•. It has recently been ~emonstrated (Di~kson 1971)

~ . . " .., . . , .,.that,. during this century, the European shelf seas (North Sea, Irish 'Sea,'. " . .- .. ' .EnglishChannel) have themselves experienced a striking inter-annual' .,

v~ia~ion betw~en ~eriods"of l~w and of high salinity; the timin~:~f'". '. ,... • I. \ " •

these high,and low sa1inity,periods is shown to be essentially similar in~ • I. .' • .• • .", • • .

a+1 ~e.as of·,the she1f seas. Each occurrence of high-sa1inity conditions

appe~r~,~o be associated with the periodic re-establishment .of an ano-malo.'lls.. ro:d persistent pressure pattern over the North Atlantic, ~ho,s,e

per.~istence has been suggested by Namias (1964, 1965) to be due to

ocean/atmosphere feedback. vlhen weIl developed, this pattern has two• • t. '. ,-

,_main elements - a low pressure-anoma1y cellover the west At1antic and a,". • • ,~," , • ., ~~'. • r : •

pressure-anomaly ridge over north-west Europe - andalthough in certain..... .' ", ..... , ..,' .

• :; • h 1

c~scs only one of these.cells may be strongly present, a common feature'. . .. '. "~ " . : '. '. ,'. .

is the addition to the 'normal' circulation of a southerly anomaly.wind.'. .'. ::.. . . . . . ~;. , ; : .. '. .?ver:.thc, .. (3astern Atlantic. Advective ,changes i~duced in .the ~astern

At1antic by ..this anomaly circulation, resulting in an incrE?ased sal"j; .

"transp~~t, to th~"~h~lf,sea~, a:re hel(i.,t~. be pr~arilY responsibl.e for.each sustained sa1inification of,these.sea~. Of importance to the pre-

sent discussion is the fact that the intcr-annua1 salinity variations

,', ...obscrved.at depth ,in ,the. Skagerrak are, cssentia11y identica1 to those of.' ."" •• ' ' ... ' •• ', 1 " ~ '.•'~ • • .: ' • • '. '. • :.:" .'.:' • •

.' the, ohelf seas .proper. ,It therefore appeal.'S .likely that the abnormal.'.. ...•. ~ . ." ..•• .". ',. : . ~ • . • • •• : , ... ~. r. .' .~.': • ..' . ." r "

moteorological conditions ..which giv~ .rise to an increasod saline trans-." ., .': .' ' •. :'.. :' ,~.: !.: _, .. 10 •••• ' I.'. '. '. ,. ";, ...,

port to..tJ:lo. shelf soas, are also responsiblo for producing an augmented•• ." •••• • "'. • •• • .' ' • \ • : ~, • • ', ••• ' '. " • I

.deep saline ,influx. to the .Skagerrak, ,thereby setting up a, conditio~ .• ~ , • ...., • • .' • j :.~. ; • " •• : .. .: .' .' , • ' 1 '.' '. • j" , '. • , '. :. . .',_ : •..

favourable, ir.not essential,.to ,the__ generation 01' major Baltic inflows.. , ," , . -:. . ... ' ."" ... I ""~ ..' ,. . . . ' . . ,J " '; : " '..' ," .' •

In,view of,their.radica1 effects..on.hydrographyand marine 1ife, it!" ..', . '.. • ~ ...' " .....,... , : .: - .' \ ••,. . ,.,'. ...: . : ' ,,' ...

bocomcs important. to develop.our ideas regarding the cause of Baltic.' . . . ~.. .. .' .....'. .:' .' . . , ';.. . "inf10ws to thc pointof predicting their occurrence. Apart from any

• ., , ..... '••••J. ,. J;..' .:' .. .' " '

; 2

•

•

•

•

"fish~ry or other biölogical advantage tobe,gained from such,a forecast,

theanticipation, of, inflows' would assist 'in planning,the"ß:tudy, of ,these

relatively'short-lived events. At,present, two methods,of inflow pre-

di6tion may be attempted. Firstly, we m~ try a simple forecast based on

the assumption that the fairly regular periodicity of inflows in the,

" 'recent 'past' will' provide:"a guide to their periodicity in the near future.

""Secondly; assulning that the chain of cause and effect relationships out-

"lined'aboveis valid, we'may'take advantageof the expected time-::lag

, 'between cause and ultimate effect so that, by a close monitoring of

"atD1ospheric trends~ wo may interposo a forecast when the supposed atmos-

. pheric 'cause' becomes established.

, Figuren 1 and 2(iv) give the basis for the first methode

Figure 2(iv) shows the post-war salinity variation, at near-bottom,depths

" for three stations in the Skagerrak and Kattegat (SI\AGENS REV, 38 mt

,

depth; LAESO NORD, 38 m; ANHOLT, 30 m). The datahav:e .been,smoothe,d

and are plotted as running 3-quarterly means of, the deviation,of salinity

from the 1941-69 normal. For convenience in descriptiontheconspicuous, .

salinity maxima have been labelIod A, B, e, D, E and F in Figure 2(iv).At thotime this method of prodiction,was attempted" salinity datawere

availablo only up to the end of 1910, so that the object was to predict

the 'occurrenco of peak F. ' '

From the available data it was cloar that, ovor the:period 1941-10,

salinity maxima in tho deeper layers of, the Skagerrak andKattegat had

recurred with considerablo rogularity. Figure 1a shows'theresults of

a spectral analysis of tho ANHOLT data over this period, indicat~ng a

fairly'conspicuous 5-yoar poriodicity. It would clearly'be'naive merely

to project tbis 5-year periodicity into the future without,establisbing

whether, over,the longer-term, this periodicity had been liable.to

change; we may investigate this possibility, howeve:t', by using the,

method of, complex demodulation (see Bingham, Godfrey and Tukey 1961,

Perkins 1910), which isolates the variation with time of any chosen,

'pe'riodicity within a time-series, and expressen this variation' in 'terms":';of'amplitude and,phase. Figures 1b and c show the complex demodulates of

: thc 5-year poriod variation of salinity anomaly at ANHOLT (28-30, m depth)

'over tpo period 1900-70. Each demodulate estimate applies, to a portion

", "of,: da;j;~J!10 years' in 'length' (Le. twice the peribd: of inte:r;est) ,. and, the

13:--estimates':werc achieved' by overlapping adjacen't portions "of.' data:, byone:";'hal:f, their length~··.. :,lIence each,demodulate estimate i8 'independent

except for the estimates .which immediately precede and~ followit. ' ',:'

3

·.

: " :'~\.r:l; is:'elear frornFiguxes 1b and' e that,'both the amplitude and phase

'of~the'dominantperiodieity'in'the deep salinity reeord at ANHOLT have

been subjeet to eonsiderable variation over the present centur,y. ,Tbe,

progressive,drift of phase shown between the first and fifthdecades,of

this century'woUld normallysuggest -that, the demodulation frequency ~:,

''';(O~2 'cycles' per year) was higher than the dominant 'frequency, o~ :the, data

---::in thc~e'years, with the reverse situation in subsequent decades. ,HCl.w-

ever;'since'the amplitude of,tbe salinity variation was low inthe,years

1900-20 i t fs probably more meaningful to regard the years, ,1920-10 (when

thc amplitude remainedconsistently high) as years when the phase,flue-

tuated about a relatively constant mean value. By either interpretat~on

it was thought that the phase stability was not suffieiently high to

'permit 'us to' assume that the recerit 5-year periodieity shown in the, ,

ANHOLT data would be maintained in the near future. This view was',sub-

! sequently justified by the lpremature l oeeurrenee of peak F in 1911", •

, :, 'since, if the reeent 5-year periodicity, had been strictly maintained,

this salinity maximum should not have developed unti11913.:

The seeond method of inflow predietion, outlined above, relies as

before on the'assumption that high-salinity eonditions in the deep part

of the Skagerrak are an essential preeondition of inflowto the Baltie;

the method is also based, however, on the theory that these salinity·

'maxima'are themselves attributable to adveetive ehanges indueed in the

eastern Atlantic by a speeific and reeurrcnt large-scale anomaly in the

atmospherie eirculation (Dickson 1911). If the recurrence'of. this.·,

eireulation anomalycould be detected by 0. elose monitoring of suitable

atmospheric'indices, thena short-range forecast·of Baltic infloweould

"be interposed. The meteorologieal event in' question has already been •

deseribed,'and fromthis it is evident thatits recurrenee should be

refleeted in an, 'increasing strength and: persistenee of anticyelonieity

over north~west Europe, coupled'with'an increase in the southerly:wind

eomponent.over·the eastern Atlantic. A valuable set of indices, based

"'on Lambts catalogueof'dailyweather types over the,British Isles frC?m

'1861~to the present ,(Lamb :1950, and in 'press), has· recently been eompiled

:. ;~;.,-;. by, Murray and Lew'iS' of...theHeteorological Office; Bracknell (l1urray' and

, Lewis 1966, Hurray and Benwel11970); " 01" these thc Cyclonicity ,Index

'·(reflecting'the.changing balance of cyclonicity and anticyclonicity,over

:·thc' British' Islcs), and 'thc Southerly' Mcridionality Index, (reflccting

thc relative strengths of southerly and northerly airflow over Britain)

are most sUitable for 'our use here~'.· Their:method of compilation i8 des-

eribed elsewhere (Murray and Lewis 1966).

4

....

•

•

..

" Figura 2 presents thc' synthesis of data whlch is the:basis for this

..' theo'ry, olBalti'c -inflow predietion. All tho dataare shown in thc formcf rUnning 3-quarterly means of anomaly (relative to some long-term mean) ,

'and 'eäeh'graph iileludes data up to the end of 1971. Individual points are

'.',: 'Plotted 'a'gainst the middle quarter of tho 3-quarterly poriod to, whiehthey

"refer (e.g. the last point on eaeh graph is a maan anomaly for,the 2nd,

" , 3:rd"and 4th quarters of 1971, and is plottod in the 3rd quarter).

Fi~e 2(i)'shbWS the ehange of Cyclonieity Indox anomaly sinee 1947. (nor-

';m~i pbriod 1947':"70); inereased antieyelonieity is shown by negative.. ~.- " ..,.values of the index, but the graph hasbeen invorted to show the,relation

with other parameters more elearly. Figuro 2(ii) shows the ehange of

Southerly 11eridionality Index anomaly over the same period (normal

period 1947-70), with inereased southerliness shown by positive values.

'Fi~o'2(iii)'~howsrunning 3-quarterly means of surfaee salinityano-

~ly"(nörmal period 1905-54) for ELBE I lightvessel (German Bight) and

VYL lightvessel (off Jutland); two oeales of salinity anomaly aro." :;. ~,.. [: ,'.

neeessary, sinee the amplitude of the salinit,y variation differed markedly

'~t" tho two locations, although the trends of variation are eosentially

identi~al. Figure 2(iv) (alreadYdeseribed) shows running 3-quarterly

means'of'salinityanomaly (normal period 1947-69) for the lightvessels

SKAGENS REV (at 38 m depth), LAESO NORD (38:m),and ANHOLT (30 m) •

. Finally, 'Fi~e 2(v) shows the similar,:pd.§l~.:".!l.ar'salinity variation at

lightvessels KATTEGAT SW (37 m), HALSSKOV REV' (20m) and GEDSER REV

(20 mh the normal periodis a.gain'1947-69 and different, seales of

salinity anomaly are again required"to'aeeommodate'the different ampli-

tudes cf salinity variation at the"threelightvesselßö ,The loeation of

eaeh of the named'lightvessels is given-in Figura 3. ,':; .', .,: :,:,. !~' .. ';.;

In Figure 2 conspieuous peaks are shown whieh,'are ,eommonto .eaeh

graph and whieh are thought to be eausally.eonneeted. ,Eaeh,peak in'a

'conneeted s~ries has been similarly labelIed •. ,Inthe ease, of peElks.A,

B, C, D and E the series is eomplete; eaeh ease,exhibits,the expeeted

sharp inerease in antieyelonieity ovar the BritishIslesaeeompaniedby

an increased southerly airflow off north-west Europe. ' These peaks are

'followed - generally after inereasing time-lags - by salinity maxima

.... in the North Sea, in the deepe~:~ayers of .the Skagerrak and Kattegat·

:;:, and in the Belts: The overall,time-lag, whieh is attributed largely to. ":... ;' .

adveetive proeesses, may be .ver,y small, as in the ease of seriesE, butj·:.',l '.;'.. .

- may be as great as 1~ to;,2 i years, as in the ease ofseries A, B and C.

, Clearly the range of any inflow,predietion will depend on the length· of

this time-lag.

5

, ......

In the case of serics F, thc chain;·of. cvents 1s apparently not yet. " .. I .

complete. : A sharp' rise in anticyclonici ty. over north-west Europc. occttr-, ; red in the latter part of 1971, and the airflow over Brit~in has changed

from a situation of strong northerlies to one of m?derate souther,lies •

. As a result·the decline in salinities following peak E has ,been unusually" .

-; .

short~lived, and salinities have once again risen to peak values in the• .' I ,'-

North Sea, and at depth in the Skagerrak and Kattegat. No salinity maxi-. ., ~l •

mum'is yet apparent, however, at the threc southcrn stations situated at

the mouth of thc Baltic. In this situation, past expcrience would suggest

that the 'preconditions' for Baltic inflow have been fulfilled and that

inflow itself is imminent, requiring only favourable loca~ meteorological

conditions to initiate surge. We have some indications, however, ~~at

the intensity of this inflow may be relatively weak.

In' i'ts strongly dcvelopedform - with a deep meridional tr0tl€€h .over

the west.'Atlantic and an intensc ridge over north-west Europ~, -thc, causa-

tive pressure-anomaly pattern shows a strong tendency to persist,

.,., apparentJ.Y through ocean/atmosphere feedback effects. (Tbe mechanism is

described in Namias (1964, 1965).) In one case the feedback ,loop resulted

in a'tenacious persistence,ofthe pressure-anomaly pattern for 1i years,

from thc autumn of 1958-to thc spring of 1960, ~d the sustained rise of

salinity anomalies that'resulted is clear in,Figure 2. Since the behaviour

of the atmospherc never exactly repeats itself, the'spc?ific pressure-

anomalY,pattern' referred to carlier is in fact better described.as a

•

collection of similar pressurepatterns which'share common elements to a

greater' or lesser degree, ,.and which are thereby capable of inducing the

same trends of change in European waters, acting. through a common

mechanism. In some circumstanccs only one of thc two main pressure.cells •

needs tO'be strongly prcscnt; for example, an. intense pressure-anomaly

ridge centred over north-west'Europe will effect much the Dame change in. . .. ~.

European: watero as would occur under.,the 'fully-developcd' pressw;epat-

'tern; when the western Atlantic trough is also present. Tbc 'one-cell'

.' case is liable to be less persistent than in ,thc.case of the ~ully,":,

developed pattern, however, -, Dince the full mcchanism promoting, occaIl/atmosphcre feedback in not'present. In all cases a great deal depends

upon' thc location and' configuration. of thc cells themselvcs. ',.

. :,: Tbe mcan' disposition cf 500 -mb ;height, anomaly for. July-December 1971

(Figure' 4) illustrates theparticular·,pressure pattern responsibl~for

peak F. Clearly this 'is not a,. well-developcd case of· thc prcssuro. pat-

tern described above, hut it.'bcai's enough similarity to that patte~ to

6

,~.

•

•

..

be described as ·a··weaker, one-celled veroion 'of, i t. "The western Atlantic

trough is almost complotely absent, being repreoentad merely by theweak

601 which separates·tha high prossure-anomaly colls ovar oas~ornNorth

.America and north-west Europe. ' The latter ridgo is intensely de'{eloped

however:and has'gonerated a strong southerly anomaly wind'ovar th~:eastern

Atlantic. Thissoutherly airflow, coupled with the favourable.w~s~erly

anticycloniccurvatuxe ,over the northern North Sea, is held respC?nsible

:'for the renowed salinification of the European ohelf shown.in Figure 2.

The study of past pressure-anomly patterns of this general, type,.' 'I'

(Dickson 1971): suggests that, in the presont caso, the looat~o.n

of the inflowing compcn8ation current as a rcsult (Dickson 1911). In

thc present case thismechanism'would be relatively unimportant.

Thus, in summary, the North'Atlantic pressure-anomaly pattern for

July-December 1911 was of the type known to be rcsponsible for inducing

6. forced salinification of theEuropean shclf seas. It was however.a,,'rather poorly-developed cxamplc of this type of pressure 'pattern, and:'

because'of this is likely to exhibit less persistence than forme~,

examples. "The establishment, of this pressure pattern has been followed

by'the:eXpected rise in salinity levels in the North Sea,and in the

deeper layers of the S~agerrak and Kattegat. In this situaUon past

experience would suggest that inflow to the Baltic itself is imminent,

but the location of the pressure pattern, couplcd with its likely lack

of persistence, would appear to suggest that this inflow will be,rela-

tively wenk 'in intensity.

TheauthoJ: is indebted to the Nautisk Afdeling, Danske Heteorologiske

Institut, Copenhagen for the prompt provision of salinity obs~rvations

from Danish lightvessols',and to the Climatology Division of tho

Meteorological Office, Bracknellf'or providing monthly'values of the,

atmospheric circulation indices.

REFEBENCES

BINGHAM, C., GODFREY, H. D. and 'TUKEY, J. i'T., 19~I~.::, .!1~~ern techniques

of power spectrum estimation. IEEE Trans. Audio Eleetro

, .

•

, '

Aeoust.,~ (2), 56-66~

DICKSON,R. R., 1911. 1J.. reeurrent and persistent pressure-anomaly:pat-

, tern as the 'prineipal eause of intermediate-scale hydrographie

variation in the European shelf seas. 'Dt. hydrogr. Z., ~, " •

" 91-119.

FONSELIUS,: S. H~, 1962. Hydrography of the BaItie deep basinS.' Rep.

Fishery Bd Swed., Hydrogr., No ~ :13, 40 pp.

~"FONSELIuS, S. H.~ '1961. Hydrography of the Baltic deep basins, Ir.; . ~. ", .. ~ ~

Rep." Fishery Bd Swed., Hydrogr. ~ No. 20, 31 pp. '

FONSELIUS; S. H~~ 1969.Hy~ographyof the Baltic'deepbasins" III.

""-Rep" FisherY Bd Swed., Hydrogr., No. 23, ,91 pp.

LABLA:ntA, Io' A~, 1961 ~'Distributionund 'aga composition; oi' cod inthe" . GöthInndD

•

•

..

LAMm, H. H. (in press). Geophys. Mem., Lond.MURRAY, R. and EEmVELL, P. R., 1970. PSCM indices in synoptic climatology

and long-range forecasting. Met. Mag. ~ Lond., .22" 232-245.MURRAY, R. and LEYTIS, R. P. W., 1966. Some aspects of the synoptic clima-

tology of the Eritish Isles as measured by simple indices.

Met. Mag., Lond., 95, 193-203.NAMIAS, J., 1964. Seasonal persistence and recurrence of European block-

ing during 1958-1960. Tellus,~, 394-407.NAMIAS, J., 1965. Short-per~od climatic fluctuations. Science, N.Y.,

,ill, 696-706.PERKINS, H. T., 1970. Inertial oscillations in the Mediterranean.

Ph.D. thesis, M.I.T. - Woods Hole Oceanogr. Inst., 155 pp.(unpublished manuscript).

SHURIN, A. T., 1961. Characteristic features of the bottom fauna inthe eastern Ealtic in 1959. lurnls bioI., Copenh., ~ (1959), 86-88.

SHURIN, A. T., 1962. The distribution of bottom fauna in the easternEaltic in 1960. ~trrnls bioI., Copenh., 11 (1960), 93-94.

WYRTKI, K., 1954. Der grosse Salzeinbruch in die Ostsee im November und

Dezember 1951. Kieler Meeresforsch., jQ, 19-25•

9

... ~ (a)~ .> 9c::~

CU:>-0-CU....c::0-::;)

..c....-V)"C

~

C"I....CU 1c::cu

-0....-ucuc..

tr.l

"0 0·5Period lyearsJ

• 00 0·3~cu

"C::;)- 0·2 (blQ.

E0

c.n 0·1:Ec::

01900 1920 1930 1940 1950 1960 1970

(cl300

.

• VIcucu....g'200"0

cuVI

~ 100C1.

01900 1910 1920 1930 1940 1950 1960 1970

Figure 1 (a) Spectral energy distribution on running 3-quarterly means of salinity. anomaly at 30 m depth, ANHOLT lightvessel, over the period

1947-70.(b) and (c) Complex demodulates of 5-year period variation of salinity

anomaly at 28-30 m depthj ANHOLT, over the period 1900-70.

....

•

•

. ......,

• KATIEGAT SW• HAlSSKOV REV

• GEOSER REV

• SUGENS REV• LAESO NORD

• ANHOLT

• VYl• ELeE 1

o

o

c .

A

, ! I 1950' I I , '19~51 ! I I '1960' ! ! I ''f36S' I I "970' I ,

Running 3-quarterlymeans (postwar period) of:(i) I Cyclonicity Index anomaly, British Isles (anticyclonic tendency

negative, hut graph inverted);(ii) Meridional Index anomaly (southerlies positive), British Islesi

(iii) Surface salinity.anomaly at lightvessels VYL and ELBE I,German Bight (for explanation of scales see text);

(iv) Near-bottom salinity anomaly at lightvessels SKAGENS REV,

LAESO NORD and ANHOLT, Skagerrak and Kattegat;(v) Near-bottom salinity anomaly at lightvessels KATTEGAT SW,

HALSSKOV REV and GEDSER REV, Danish Belts.

-05

t:'"

•

•

100600 r--:,'--;'-r--r-:~~--r---"""'T"'---'--"""'T"'-~--"r--~\I\II,\

\ c~,, ~ \\ ~/ \, / \, / \

" ,'\,--... / -- ,, ,," 'J" ".",,'" Skagens

' .... , ,," + Rev'100 fms" +

LaesoNord

Anholt+Nord

Vyl+

o55

Eibe I.+

Figure 3 Positions of lightvessels from which salinity observations were used.

I.

I.:I

Figure 4 500 mb height :momaly cbart (metres) for the period July-December 1971.