90 MONTHLY WEATHER REVIEW. FEBRUARY, 1919 CON(IBESS OF SCANDINAVIAN QEOPHYSICISTS IN GOTHENBURG AUGUST 28-91, 1918. By HANS PETTEHBBON, General Secretary. [nnterl: Gothenhug, Nov. 27, 181b.l At the invitation of Dr. G. Ekman, Prof. 0. Nordensk- Prof. 0. Petkersson and other scient,ists of Gotlien- , Sweden, a highly re Iresentatit-e congress of nhout anish, Norwegian, ai;( i Swedish geopli3-sicist.s nitr.t in that city duriii the last davs of August t,hiu gear. Representatives 7 rom Fiiilnntl hac1 also been invited hut, were unable to be present. At the opening meeting Prof. IIililebrniii!~si,u, or Upsala, was unanimously elected president,, su )piwtd 11:)- three vice resitleiits, viz! Director Ryder (, \ )e1iai:ii*k 1 , Prof. BjerLes (~onvny), and Prof. X(>rdrlis!;jijld (Swedeiij ; General Secretary, DP. €h i s Pet,t.erss:on. Prof. Bjerknes, of Bergen, openccl the lirst gtwvii.1 meeting with a paper on weather forecmting. rlcscrihin:; a new and most successful niethotl of shnrt,-ixnge pmg- nostics for yyiculturnl purposes estnblislied in west.ern Norway during the summer 191s. [I'ublisliecl pi). ---, of this REVIEW.] During t8he followiiig thy9 meetings i\-ere Iie!cl, pa.rt,iy general and partly by sections. I n nll, :XI lxiprs W:'O!T read, niauy of these of very great interese. One nft,ernoon was devoteil to the m ~l J ~:~ howali<.; phenomenon, n forenoon to. prognc~tks. and t h e last. nf~cr- noon wm occupied by very :iiiimat.c:.l diacussioii~ of futur:. Scandinavian cooperation within ilific'r:ivnt, I J ~ :ilii;hc+ ul the geo hysicnl science. ado tecl by t,lie cong:.ess in pleno nnl-l :i cc-mmit,trc f9r:iii?:I w i t l the object of calling togetlit:. a sei.viiJ coiigress in due time. Abstracts of t.he )apers mill before t h coiig:'e:;s will shortly be publisheJ. Final P y n number of resnlut,iorlr v.-cre, imw:! an11 RESOLUTIONS ADOPTED BY C9SGIt.?c ol)sarr:i tbns of t,lie int,ern:il iiiovenieiits in the sea rou!id t.he coasts of Scmc1in:ivi:i. compared with ~imu!t,~i:ieejus met.eoroIogica1 phenomena and tishery st.ati::t.ics. (Uovecl I J J ~ Dr. IIans Pettersson (Sweden) mid otlicrs.) [See Dr. Pettwssoii's paper on this subject o:i pp. l(10-105 of t.his Eavzew.] IS. In f:iwr of thc est,&lishing of aerological ohserva- t.io!is in tliXoreiit p:wt,s of Yc.nnrliiin.c.ia. (Moved by Prof. IIcssc!l!wrg, ( 'histinilin. V.. Exi;)h:isixiiig liie need for coopcratioii between t.hc gco; liy a icist,.; of tho SCiI,!id i; iarimi cou i i tries, both Kith reg:cnl i.o 1:il)orntor.v niid ticlrl resenrch, which pro- i.! t,ho C+cJvcr;i:ncuts of t,lic:sc countries itre retiuest.ed t o t';wilit.:it.e n.3 ~iiurli :~.s possible. I-!.. 1ii fnww oi' t4hc project tlraftd hv 0. Pettmsson th:; t, :I. !irst-cliiss sc:iciitific institmute shall be sstJ>lished iil ihthi~iil ~r g for oceniiogra,phy, ninriiie meteorology, :i i l d :iero!ogg wit.li the OhJeCk of studying the dynamics 01 t,he morement>s occurring in the ntmosphere mid the sm , which det,erinina t,he cliinitt,e a.nd the weather of the Yc..;tritliiinvi:tii cou:itries, awl of st.uclying t,he influence of t,!ieac? f:irt,ors on agriculture, fisheries, navigation, and aeronautics. (Y.ovetl 1)y 0. Pet8terssoii nncl seconded by 1 I ) rt?j)rcsc.iit:i t8ivcs for Deiiintwk, Sweden, and Norway. 111. In favor OF est,:i.?)liahiug It is possible t1in.t more t h p one so!ut.ion m:ip he found to t,he problein of stit,isf:ictory :ind prac h *:i I weat,her forecasting. It is ' dso possi1)l.e tha.t, nriioiig these we nitiy he nble to fitid met,liocls which o!)viztt.o t,lie nec.essit:y of a con1plet.e uiiderst.a!idiug of the I'hctioliie1ioil whose developmelit we are to forecast,. Persoildly, I have no interest in such methods. I ~i i i iiit.c.re,he following: All zit,nios- rocesses obey the l n w of physira, first of :dl pherlc those o P mechaiiics and tliermodviiamics. We me, there- fore, really in possession of dl tho t.Iieoret.ic:.i.l knoideclge necessary to determine future weather. It resides in the equations of niechniiics nnd t,lic?riiiodvn3mii.s, or more generally espre-swd, iu the eqwi.t,inns of physics. These contain t,he answer to all qurst.ions nhout, the future weather, if only the 011servat~)ia caii give us the concrete dntil wit,h which we are to deiil.:" The problem of dct~criiiil1ing t.hc fut.urc vreiither may therefore be re- clucctl t o t.he sohitmion of two special problciiis: 1 . Tlle prticticn.1, the oht>aining of the uec.essnrg con- cret e ila t.:3 t.hrougii observation. 3. Thr tlieorc:.ic;la,.t,lie evolving of t,he methods whereby t.ho li!io\vletlge contained in the equnt.ioiis can he applied to t.he o!)seruii t.ions. If we state t,he xohlein in it.s most precise form, as.a mat,liemnt.icnl pro L lem, we cwi not hope to solre it in t,he i i c :~ fut.ure. But if we put it into a more practknl form, W E will t,hen see t1ia.t t,he line ofp-ogress lies in the cont iaua ticxi of t.he niethocls which the meteorologists folkJWei1 when t.hey began the study of atmospheric procssses by driiwing synaptical ma )s, from which t,he founcl the ~)nric wiiicl law [BUYS nit ll ot's law], etc. d thcrct'ore need not ilouh t t,hiit, in this way we will arrive i1.t results which will give us Q better underst~unding, niid miiich will tdius react upon niet.eoro1ogica.l practice. cert,ain estciit this has already been confirmed. Cut blie ic iea of ong:iging myself in meteorological tice was untliouglit of, even a few months ago. $.E:; however, the critical tiiiie came, when duty demanded r 1 l o a FEBRU~RY, 1919. MONTHLY WEATHER REVIEW. 91 our best service in the aid of agriculture, I cloubt,ed my right to hold back anylonger. From the ioint of view of niy theoretical work it was not yet feasible to take such a step; and the conditions of the start mere as uiifavor- able as possible. It concerned the weather forecasts in the west of Norway, wliere even in time of peace we had not yet tried the problem, and where now in t,ime of war there is lacking even that which, judging from all our meteorologicd esperience? ought to be the niost inipor- tant, namely weather telegrams from tho wcst---from the the British Isles, the Fwoe Isles, ani1 Icelancl. That w1iic.h gave me hope, however, was some of our newest results. When .\-e wish to proceed rigorously in the tlieoret.ica1 way an uiiavoiclnhlt? deiiinnd is tlictt we take into accouiit all the variables upon which tho solu- tion of the probleni is based. A va.rinble which c:i.ii no longer be kept in the b ~k g r ~~i i d as before is the winil. representation. These maps present a series of in t,eresting peculiarit,ies, which had to be discassecl. Very conspicuous were what, we now call the lines of convergence and of diver- gence, i. e., lilies tn w7~ich. line .t !h d blf.ws .in, from both sides, or lines-frow. ,whir7$ t7t.t ,v+nd 7dotm ott5 on both siiIts. They had been found alrertdy on t,lie first innps of this kind, which were c~rawn a t Sttwkholm a t tlie hegiiining of this century by J. TV. Sanclst.rhi, x-ho w.\‘L~ iiiy assis- ant a t that time. Since then they have been increa.singl:l. the object of our n.ttention, pnrticulnrly of late, after 0~18 of niy ~ollahorat~ors hat! succeeded in cletliwiug t,he lttw of their ropagat,ioii n.s a. consequence of the gcnernl ln,w the law reads simply: A line of convergence x-ill nlwaq’s move to the right, a line of divergence alraj-s tjrJ tdic left., of an observer who loola along the line in the clirect,ion of the wind. Further, t h d o c i t y of propagation may, for the main term, be det,ermmetl Ly measuring the L ( divergence ’? and the li curl ” in the nearest surrountl- ings of the lines. To test these results, ohse.reat.ions froni a T-ery close net of stations hail to be preparetl. Ohservntions were therefore gatliered from the second and third cla.ss dinin- tplogicnl stations in t>he Scn.iiclinarinn countries, from the light-keepers’ journals, et,c. The n i a p confiriiietl t.he law with as grea.t an. accurac.y as it was posailde to oht.:>in, thou& not sufficient for a sharp cpnitntire test. l h t , on tEe other .linncI, tliey s1iowec1 mitli full slicwpiiess mi important enipiric law, viz, there belong to every cyclone which is not stntiomry hut in niot,ion two chara.ctcristic l i e s of conver ence. (See figure l*). They both come one runs almost normal. Rut the other one nlwiys dings nearer to the path of tlie c ~y h n a the iien.rrr it, conies to the cyclonic cent,er, and in t.he cenber itwlf it seems to have the smie line as the path. If we, t,lic.re- fore, have a suffic.ient number of observations t.0 he c t l h to draw this line right into the center of t.he cyclone, t,hen it will show us the iiiomrntary direction of propa.gation of the cyclone. For this reason we niap call it, the cyclone’s steering Zinc. At the same time it ought) t.o he possible on a suKxiently acc,urate map. by simple iiiew- urements near the one or t,lie other of the two lines, tu Among meteorologists it has not yet reached a maps of the lilies of flow a.nd of the strength of This represcn tation is iiow of the P orniation of rortices.2 In it qualitatiw form in froni the rig a t side of t,he pn.t,h of t.he c;vrlone. The ___.____ ___ .. -- I Cf. V. Bjerknes: Dynamic Meteorology and IIydrogrnph>-. YO~. 11. Iiincnmlic, * I. Bjerlnk.: TJe5r die Fortheweyng des Ponvcrgeenz- und Dirergenzlinien. Ncte- *GI. Shaw, Manual, pt. I x p . lOS, etc.-W. J . H. Washington 1913. orologische Zeitschritt, 1917. 3 5 . decide at what speed the cyclone is proceeding in the direction which is thus found. The importance of the two lines of conver ence will be properties. I t has been fouiid that the warm air, which IS carried to the cyclone and imparts the energy* flows in the sector between the two lines. Under the nopagation convergence tlie front of a cold wave. This involves also very important meteorological rop- sharper t-lian the wrtrni, is followed by violent rain. This falls from the warm a.ir wliicli is raised by the advancing cold wedge of a.ir. It often appears as thundershowers with violciit sts of wind, and in reality this line of con- long known as a frequent follower of cyclones. Also the sbeering line is associated with rain, n t any rate, suffi- ciently near t.lie. cyclone ce.nter. But this rain is not in the form of violent showers. It comes not onlv after the line but also up to 300 km. or niore before the h e . This still niore conspicuous when we look upon t % eir thermal of the cyclonic s stem, the steering line w Er; therefore mnrlr the front o ? a warm wave, and the other line of erties. The cold wave, which is, as a rule, niarkec 7 out vergenco is i rY onticd with that s c p d line which we have /-- FIG. 1. Cyclone. rniu hefore the line conies from the warni air which lides up along t,liat inclined plane which borders the under 5 ying cdd wcclge of air, which is swept away during tlie for- mtrd movement of the cyclonic system. Evident+ me are here at the beginning of n cyclone theory 1%-liicli may be based on real knowledge of the phenomena. But it is it practical view of the nintter that interests us to-tlay. I t is perfectly evident that these results will prove useful for weather forecasting within an extended :ireti, where we get the entire cyclone with both its an- tcnnns 011 the map. But, we are not able to try this free appl.ic.at~ion on maps which have normal extension until n.ftrr the war. It now concerns the application on a sniall scale, t.0 west.ern Norway, whose weather chiefly is governed by cyclones which pass outside the territory covered by our observa.tions. The question is, have we not,, in the ident,ification of the passing cyclones’ con- - *Not all oIit.--Ed. 92 MONTHLY WXATHER REVIEW. FEBBUABY, 1919 horizon ? If the sky is covered at the stmation, it is evident that if a clear stripe is seen far away on the sea margin the report of this clew stripe will he most important.. Rut we are more puzzled as to wlint slioulrl he re- ported on clear clays. In order t.0 get hints, I have visitcc-I most of our coast stations ~i i d talked with t,he people. It is well knowi that in ninny cases fishermen and seanicn can predict coming weather with great ~C.CUPBC . I have asked them concerning t io signs which they use. I t is not easy t,o conic to the bottom of the niatter. Evidently a comples of siudl s@s is intermixed wit#li much superstition. But. now nnd t,licii sonie safe hints hnve been rrceivqcl. \Vwn a storni is browing, a Clark stripe is penerJI~ seen on the I m i - zoii. A Tilot, at Lindcsniis gave me T vergence lines, a new important means of supporbing the forecasts under the present difficult conditions ? To serve the preliminary investigations hereof, we got weather reports from tlie signal s t d o n s of the navy district of Bergen. The resu1t.s proved enc.oumging, and w0 went on. The Government granted the necessary means, and by and by the net of st,ations as shown on I t is seen to be a consiclernbly c oser net than my hitherto used for ~~-~atlier-forcc~.stiii~. This is essenbid if we are to he able to disc.orer and identify the lines of convergence with full ac.curnc.y. And it is a step towtwd the fulfilment of blic theoretical demand, that a continuous image. of the state of the atmosphere should form the base. An important feature is the double series of barometer shtions, one on the islands farthest out to sea, and one at the heads of the fjords. Between them we have a great number of other 2 was erec,ted. -, -"- -- -+ -'- -6 - -I- -.- -s- -- --;- -1- -;-, -I.- -,,- =,,I -I)- ,*- -L ,*- -1.- -I- - ,.- 'I - -. 00 Kristiania U O T wu O+ o r 'n FIG. ?.-Meteorological stations. stations observing wind and weather without the use of instruments. On the outer islands the direction of the wind is taken very exactly by azimuth clink.. As the stations can never be too nuinerous, our ob- servers are instructed to give as complete informat.ion as possible, not only of the state of the atmos here at the instruments, barometer, thermometer, etc., can only give the measured uantities in one single point.. I3ut ob- servations of %?bat we call weather, permit inore coin- Our observeis have accorclingly %een instructed to report not only rain a t the place, but even rain f f i g within sight, and the region of the sky where it is seen. Further, they specify the cover of clouds, not o d in accorclaiice with the coninion cliiua- tological rules, g u t ailcl also the region of tlie sky where the clouds or the different forms of cloud appear. If the place itself, but even in its surroundings. E; UT coninion lete communications. FEBRUARY, 1919. MONTHLY WEATHER REVIEW. 93 the construction of the distance measurer shown in figure 3. It is to be kept 50 centimeters from the eye, so that the lowest divisional !ne $3 seen against the horizon. The figure at the divisional line coinciding with the edge of the stripe will then give the distance of tlie storin. If froni time to time we take the azimuth of the bank of clouds, and determine the distance, it must be possible to determine approsiinately the course of the storiii, and also if it will reach a given place, and wlwn. The instrument is now being distributed aniong the coast stations, and reports on tho sltu- ation of banks of storm clouds have already often proved useful. On theotlier hand, more coiiipleto reports, wliicli nuglit give the course and speed of the storm, have not yet been received. But no matter how far we can look in this way into macces- sible regions, fore- casts for ail outpost district such as the west coast of NOS- way will remain pre- carious. The equa- tioli will detorniine the future state at any timo, in every point of the atnios- phere, only 011 con- dition that in addi- tion to the exterior conditions, we know thcinitidstete of tlie entire atmosphere. And tho iiioraliniited is the region for wliicli the initial state is known, the morelinlited will also be not only the dis- trict for which ra- tional forecasts can be made, but even the period in which they ivill remain vahd. Acyclonesit- uated so far off that its bank of clouds can not be observed can very well reach the coast. in 6 to 8 hours. This entails A very instructive case occurred on August 14 and 15, in the morning. On the morning of the 14th rain w u reported froni the west coast north of Bergen. At the saiiie time W R ~ reported a “brewing up” 180 km. west of a station on the Roinsdal coast. This observation proved a consicleralde estcnsion of the district of rain to tho north on the sea. The 1,aromotcr was fallin rapidlv on the west coast, more slowly at tho coast of f hnsdal. This seemed to indicate a cyclone taking the course southward along the cowt. Tho circumstance that uuy 15 8a FIG. L-Lines of wind flow and rain arew, bug. 1445,1918. an im ortant restriction: The forecasts can be issued for a period which can still be of use: Based upon the ohsetva- tions made at S o’clock in the morning, forecasts are issued for the rest of the day, and even this will in somo cases be beyond ernlissible limits. which, in coiuiection with the corresponding clinrts of pressure, temperature, tendency of barometer, etc., have been used for the forecasts. very r imited period only. We have chosen the shortest I shall give a f ew esamples of charts of lines of flow, neither of the two lines of convergence could be identified seemed to point in the same direction, as lines of this kind are not to be found on the left side of the path of a cy- clonic track. The midday ch arts for tho same day, fi . 4, show a slow f91 of the barometer at the coast of Roinsdal, but a very mpicl fall on the west coast. This seems still to point toward a movement of the low pressure southward. But now tho chart of wind shows a mark- ed line of coiiver- ence that reaches k and on the coast of Romsclal. Tem- perature is two or t h e e degrees higher behind this line than in front of it. Con- s e u e n t l y it is steering line. And to judge from the situation of this line the cyclone should go toward the east, n o t toward t h e south. Thus the signs seem to be con- tradictory, and the uestion is which of to believe. The evening chart shows that the steer- iilg line has moved wit Y iout doubt the t R e two signs are we considerably toward the east. At the same time its character ns ti steering line is further marked by rain froni the inclined plane in froiit of it. Meanwhile the fall of the barometer lins ngnin been more rapid on the west coast than on the Roinsclal coast. The barograph at Bergeii shows a very sharp fall. This seems to indi- cate that the cyclone will trike the course toward the south. Are we now to place our confidence in the fall of the ba- rometer or me we to think that the cyclone will follow its steering line 1 94 MONTHLY WEATHER R.EVIEW. FEBRUARY, 1919 The mornin chart for the next day gives the answer: turbed toward the east, and now it has been followed by a squall line also, which has reached the west coast. The strange fall of the barometer on the stations of this coast has been the consequence of the propagation of this line toward the east, iiot of any propagation of the cyclone toward the south. As soon as the line had passed, the barometer also ceased to fall on the west coilst but the fall continued on the Ronisdal coast. Tlie two linea of convergeme now allow us to take a heariiig of the c_\-c.lonc on the sea. The lines are situated typically as on t,he schematic figure, the steering line with the regiou of rain in front of it, the squttll line with tlic region of rain behind it. The air is cold in front of the steering h i e niid behind the squall line, but warm in the sector l>etweeu the two lines. The circumstmce that nll rain has ccased in this sector is evidently due to a foehn effect, the air current here c.oming down t,he mount~iiiiu. On the next day the cyclone litis cont,iiiued its n i o ~e - ment tow-twds the ectst to the coast north of Tronclhjeni. Here it has subsided and nieiLnu-hile n new cyc?lone has been formed on the Swedish side of the momitairis. This seems to be the ordinary-way in which the cyclones pass across a chain of mount~uns. During the time the cj-clone has been pressing against the west and the Romsdal const, tho weather has been d r j in eatstern Norway, with some rain only in the uppcr vdleys underneat.11 t,he rmge of inountnins. 'l'he weather has probably also been genemlly dry in midtllo and southern Sweden and iu Denmark. I tlo not h o w the forecasts of the Swedish and Danish meteorologi&s for these days. But i t would not be surprising if led l ~y t,hc tendeiicy of bar?meters oil the Norwe&n coast, they hiid predicted ram in all districts. The meteorologists in Christiania a t all events fell into the trap. The tmgic reason was that many of the midday telegrams from t.hc coast of Romsdal were delayed so much t h t the trnc chart of wind, coiitainiiig the. obvious line of convergciice, did not yet exist a t the t h e when the Central in I3ergen sent its report to Christiania concerning the situn.t.ion. The forecast for easterri Norway liad therefore to be founded upon the fdl of the barometers, am1 concludecl with rain over the whole district, escept in the upper valleys, esactly the opposite of what happened. Not least through this erroneous forecast, we are Icd to strongly a,ccentuate: I. Tlie iniportaiice of a sharp watch on the Norwrgiiln coast, not only for the benefit of the forecasts there, but as much for the forecasts in eastern Norway and the whole of northern Europe. 11. The importance of a close net of stations, which alone makes it possible to identify the two characteristic? lines of convergence of the cyclones. 111. The importance of these two lines, the stoeriiig line and squall line, by which the cyclone itsevi,ndico,te.s its The experience of this summer has more than suffi- ciently shown us to what high degree our weather is con- trolled by the lines of convergence, which sweep over thc eountry during the pnssiiig of the cyclones. I will sup- plement the precedin esample with a simpler one. the southwest coast and a falling barometer on both sidcs of the southern point of the country. The c.liart of wind, @re 5, shows that a line of convergence has landed, and the fall of temperature characterized this as a squall line. The steering line could then point The steering f: me has continued its propagation uiidis- CoZCT8e. On the morning o 4 July 18 rain was reported from only to t,lie north, parallel to the coast, and the unknown oyckme then had to move trhis way. The forecast mas given wi tali the empiric lmowledge of the veloci t of propghon of tlie squall line: More rain for the sout, x ern districts, ribin soon for the nest northern districts, rain after midclny in the districts north of Sognefjord. The chart for 1 p. m. shows that tlie rain had reached the mouth of this fjord, and it continued northward durina the nft>ernoon. J d e r on I happened to visit B place w%ere either this or mother forecast of the snnie character had been useful. Alarmed by the forecast, t,he peasants began, in spite of tlie bright wenther, to fetch in the hay even in the forenoon, and just, in time, as they ilf t.er\\mds snw. Such forecnvts having the time of the occurrence of rain cle:~ly iii!lioateil, bin-e, of course, been issued only on liire nccosions, but when given t,liey litwe been successful. C'uiisirlcring the system of wen.t.her forecash , which lias in this way hem iinprovisec-! for western %orway, its wwult,s as A w-holl: hnrc, in spite o€ all difficulties, been surprising1-y sa t.isf:ictory. A verificat,ioii, worked out by / July 18 8a July 18 la FIG. 5.-!.iiies 01 wind lloiv and rrrin areas on July lS, 1918. compiwing the forecasts with the midclay and evening rc xxt.s, gives for the first iiioiith of the forecasts,'namel , J!I.y, the following percelltvilge of hits for the eig z t ciifterent tlidricts from the sout.11 point of the country to the dist,rict.s north of Trondhjem: 81.3, 79.2, 77.1, 93.6, SG.4, 79.5, SS.7, S4.0. An essentid ~'ctwxi why the results for tslie three southern tlistikts ha\-e turnetl out less favorable than for t,lie ot,hers, is that in most cases they had to be issued before tlie iirrirnl of the Danish t,ele ams. The averago per cent of hit,s is 83.7. Although t T- ie results thus have proved sttt,isfactory, owing to tlie war we must work for t.lie present,, as it were, with blinded eyes; nevertheless, we h a n every reason to expect that. in the futu1;e the area from wli~ch t,he o!mrvotions are ohtained will be restored to its former extent. The system, liomever, has certain failings which can not be remedied c.ven by this estrnsion. The smaller the cyclones thc more difkult it is to make forecasts of t,hem. Z)iniinut,ive cycIones are very fre uent. Form- erly they slipped undiscovered t.hroug8 the wide FEBRUARY, 1919. MONTHLY WEATHER REVIEW. - 95 meshes of the old net of observations. With our close net of stations we can, to a certain estent, follow ancl forecast them, ancl the closer beconies tlie network the farther wo can advance. But at last t,he liniit of what, is obtainable is reached. We can not mnke it close enough to predict every locsl rain shower. There is, however, one wny open by which the systeiii c:in be completed, and this way I hope will be tried by nest summer.. A similar view of the sky, only still niure est,cmive t,han from the co:ist, can be obtained froni f:iromhly situated mountains. Everyone who has t,ravelerl in t,he niountaiiis knows that we can see wmrlcring rain sho-t\-ers at a great distance. The question is then t,o find prnc- t.ical nirt,hods by which t,lieir tracks can be ileterniinetl, just 8s the course of ilist.ant b d i s of s:.orni clourls aw found from the const) sttit,ions. No\{-, if me st,uily blie view froin the t.op of n mounta.iii bp the use of gooil clitirts, we c m identify a grept number of point.s on the around, such a.s summits, ridges of mount,nins, riwrs, rakes, towns, .vil~tiges! a.nt1 cliurclies, etc.. lie c~inrt. cnii be espeoinlly prepnrecl for the purpose I)p inarking all the chanrcberistic lines and points we ti.re able ta see m i l hatching the regions, which can not he seen. Then when the showers of rain advance the observer will be able to draw on the chart tlie part of the front line of the shower which is turned to him. By drawing this line from hour to hour lie will see how the shower develops and propa- gates, which districts are hein threatened, ancl when the rain will reach them. The ofserver will then not only be able to give valunble information to the central offices but he ca.n also senil forecasts direct, to the threatened districts. From a coniparatioely small number of moun- tain stations, which cooperate and supplement each other, it must, in t.his w.ay, be possible to organize a system of forecasting local showers for the gretiter part of tlie country. ,i(7~:7eii.17.1i.)li., Octobi~r, 191S.-For the same eight dis- tricts as above. the percentage of verifications for the mont,lis of ~L I ~L I S ~ nni1'September have been the following: Ahgust, 85.1, '77.7, S3.3, 70.5, 55.2, 05.4, 90.7. 96.3; September 94.0, !j-l.0, 92.0, 96.0, 90.0, 94.0, SS.0, SS.0. l'he avcrage percentage of rcrification has thus been for tlie three slimmer months: J ~l y S3.7, August 56.3, Septemlwr 92. The decided progress at the end is clue a l )o \~ :ill to the eqwrience from h u ust 14 and 15, as based with increasctl confidence upon the steering line. from this date trhe forecasts have, in c 7 ubious CRSFS, been ON THE STRUCTUBE OF MOVING CYCLONES. I:y J. BJERKNES. [Ltatrgl: Drrgen, Octolwr, lals.] When the equations of hylrodynamics a.re to be applied directly to concrete atnios dieric motions, two conilit,ions should be fulfiuecl: ~1 i e dist,aiices betwceii t,lie st:i,t.ions Wiving tlie ohservations should be snitdl cilough to he con- iidered as differentials of spixcc, and tlie t,ime in t,crvitls between tlie successive periods of obserrntion slioiilil he sniall enough to be c.onsitlerec1 as cliff ereii t,inla of binio. Neither of these condi tiona is fultilleil by the obserrntioirs available. from clnily weatlies n i q s ancl priuted in year- boob. Distances amounting to hundreds of liilometcrs ancl time intervals of ais hours are too petit. In order to get a t least t ~i e first of tliese conclitioiis' ful- filled, to sonie degree n t lezlst, I have collected detailed data from tlie archives of the meteorological institutes in Norway, Sweden, ancl Denmarli, including observations from third-class stations. In addition I have esnniiriecl the comparatively very detailed dailg maps used for weather forecasting in western Norway during thc suni- mer of 1915, and conibinecl them with the siniultnneous study of the sky. In this way I have been led to some general results coii- cerning the structure of cyclones, which I shall outline in this paper, and shall consider in detail later. THE STEERING LINE AND THE SQUALL LINE. "lie lines of flow in a cyclone have approsiniately the character of logarithmic spirals. By increasing the num- ber of observations, however, several deviations from the regular spiral shape appear. A4niong a multitude of details, certain characteristic traits seem to recur more os less markedly in all cyclones yet esamined. Every cyclone that is not stationary has two lines of convergence, which are greater and more conspicuous than any otheiu, and are distinguished by characteristic thermal properties, as shown in figure 1. The firat of these lines comes in to t.lie center froni the front of the cyclone, lying in its entire estent on the right side of ita path. The tangent to the line at its terminus in the cyclonic center seems to be identicd with the path of the 1at.ter. -4s the line is thus giving tlie mon!entarg tlirection of pro moation of the cyclone, it may, with propriety, ]>e c d e ~ tlie strwirig Jim?. The otlier line of convergence comes in from the right rear of the cyclone, and is iden tical witrli tlie well-linown sqti.rrlb k i r i c, wllicli nccompanies cyclones. 'l'he steering line ancl tlie squnll line are intiniately related to the distribution of temperature, as they border FIG. 1.-Lines of flow in n moving cyclone. the warm ares, of the cyclone, or, as we may call it, its "wnrni sector." uall line move ac- cording to the law gence,' viz, in t.he toward the right, relatively to Both the steering line and the - - 1 J. B'erknes "' h e r die Fortbewegung der Iionvergenz- und Divergmzlinien/' Meteorohgisclie Zeitsrhrik 191i,10/11.