MONTHLY WEATHER REVIEW Editor, ALFRED J. HENRY CLOBED DECEMBER 3, 1929 OCTOBER, 1929 ISSUED DECEMBER 30,1929 VOL. 57, No. 10 W. I3. No. 99s THE PRACTICAL IMPORTANCE OF CLIMATIC CYCLES IN ENGINEERING By A. STREIFF, Consulting Engineer, M e m . Am. So,. C. E. $X./# 5-83 [J:ccksan, Mich.] Science has furnished a vast amount of informs tion regarding the ancient and recent changes of climate. Little, if any, of this knowledge has found its way into the practical arts, such as hyclraulic engineering. Works for the control of water are based on records of precipi- tation, run-off, and water levels and tlie fluctuations thereof; but thus far no practical use has been made of climatic science for that purpose. The reasons arp twofold : First, the theories o f cliiiiatic changes from the Paleozoic to the present are still largely controversial ; second, the amplitudes of recurring cycles are generally considered too snia11 to be of any iniportance. 1. SCIENCE OF CLIMATIC CHANGES STILL IN FORhIATI\E STATE. Ostt~crltl Hcer (1) found that :t uniform wnrm cliiiitzte esisted from the cnrly Curboniferous to tlie late Cre- taceous, I~ased on fossil flora found in Grcenlnncl and Spit,zberpen. St ignmis, Cnlnmi t cs, and ot licr C’arhon- iferous florn, RS iwll as Oleander, hIripnolia, Ficus from the Cretnveous, once grew in rthund:tncc in those pollir regions. -4 dirisioii of climate :i(mnditig to Intitucle is visible, according to Heer, only in the late Cretaccnuc: Heer states: “Nirgends ist die geringste S;pur eiricr Gletscherzeit wthr zii nehiiieri.” This view is opposed by nh?/chior L\reumayer. (2 ). who finds many IJ’OOfS i:f zonal distribution in the Jurnssic and the C‘retxeous; :,lw the Carboniferous flora is said to show zoiial c1iciracteristic.s. Doctor l l ’ i e l a d (m e ) slates that growth rings have been found in the upper Mid-Devonian Crillisylon, as well as in a Carboniferous Cordaites as f a r south as T e ~u s , which would iiidicate the esistence of setsons long before the Miocene (3). The strongest argunient agxinst Heer’s uniform climate is the occurrence of glaciation in the Carboniferous. On the other hand, Professor Hobbs holds that Heer’s niajor thesis is still correct. Similar opposite views prevail concerning recent cli- matic fluctuations. Brzickner was severely criticized bg the meteorologist Schreiber. (4) while, on the other hand, G. E. P. Brooks states that Briickner’s vieus were imme- diately accepted (5 ). Henry finds (6), using BriickIier’s methods on data of the United States since 1890, that the former’s results are only partly verified. The same differences are also found in recent opinion on the relation between the solar cycle and climate. Huntington (7) and Douglass (8) find relations well estab- lished, while Charles G. Abbott (3) and N. Norris Russell (10) do not. Regarding methods of analysis, divergent and the Miocene flora, :dso found in the polar re,‘ "loris. __ opinions also esis t . Bigclow’.r (9) and C‘lloiigh’s (1 1) views are not shared by G‘. T. Tl‘c~lker., .who apparently does not accept (12) variable cycles, and considers hitrnionic analysis app1ic:tble on weather data, a1t)hough the necessary mathematical conditions are in that ap- plication not satisfied, diiferent results from diflerent parts of tlie smie record are obtainecl therewith, and extensions do not agree with subsequent observations (13). Instead of n small number of reriable c j ~l e s , far enough apart to be distinctly separahle, C‘. E. P. Brooks prefers to discuss 18 hnnnonic elements in the Nile floods (14) and Brunf 44 hnrmonic eleinents in rainfall (15). Hence, finding that the science of climatic changes is still in formntive condition, practical engineering has thus far abst ninecl from applying controversial theories ant1 methotls. 2 . AMPLITUDES OF CLIRIATIC CYCLES ARE SMALL AS COM- PARED WITH LOCAL AND ?JOMENTART CHANGES. If cycle stiidies h,:d pro(luced results in meteorology, their use w ~u l d probahly liavc perietrsterl into the engineering field, hut thus fr,r they have been of no aid. It has tacitly been itssuiiictl that the same results would tilw be obtaiuecl in hydraulic engineering. Sccorciing to C’. F. A l u i t t i j j : Tlw primnry ohligation of the Weather Biireau is to forecu5t tlie weather, and tlie principle of cycles might be iisecl in this ~vork if the cyclical recurrences are real and the effects inipurtant. Real cycles of very siiinll ctmplitudes n-oulcl hare very little forecnsting value (3). The Briickner cycle of teiiiperature and pressure, as well as the 11-ye:ir cycle, hare amplitudes altogether too siiinll to be of any significance for forecasting weather. The variation in teniperctture is on the order of 1’ C. or less. On the other hand, a 15’ C. difference in the temperature of the same day is a coninion occurrence (16). The secular variation in pressure is not Inore than a few millimeters, and the claily changes during storms may be fifteen times as much. For elements which are measured by accumulation, such RS rainfall, the secular amplitudes are larger, al- though here d s o the nii~iy fortuitous d d y changes sur- pass in amplitude the secular variations of yearly rain- fall. For a g r i d t u r e a severe passing drought, hail- storm, or cloud-burst is of more inimcdi:~ te local concern than a slight change in total rainfdl over niany years. While the amplitudes of long rainfall cycles are greater than those of either temperature or presswe, it has been impossible to Ilialie practical use thereof in weather fore- 405 406 MONTHLY WEATHER REVIEW OCTOBEB, 1929 , because the erratic and fortuitous changes $ gations have established many possible sources of errors III rain-gage readmgs. The influence of gage size,. expo- sure, elevation above ground, and wind velocity is well known. The erratic distribution of rainstorms, such as occurred a t Cambridge, Ohio, on July 16, 1914, is amply recarded in the annals of the United States Weather Bureau and was recently investigated by Professor Knssairine, of Moscow (17). Study of long cycles in rainfall is hampered by the scarcity of reliable long records, as illustrated by Dcsmond Fitrgerald in the Cochituate (18) recozds of the Boston water supply, and by R. Siedek in discussing the old “Brander” rain gage of the Austrian hydrographic service (19). Summarizing, it may be said that because cycles in temperature, pressure, and rainfall are largely obscured and surpassed in amplitude by the local and moinent’ary changes, they are of no practical use in weather service. 3. BASIC DIFFERENCE BETWEEN HYDROGRAPHIC AND This conclusion of practical meteorology has been transferred sine qua non to engineering, although the conditions are modified to such extent m hydrographic data that the situation is completely reversed. Here secular amplitudes occasionally surpass the amplitudes of local and momentary changes, and, far from being ob- rainfa casti”fl are yet many times greater. Repeated investi- METEOROLOGICAL DATA. FIGUBE i .-Ampli(ying effect of N U -O ~~ scured by the latter, are often visible by mere inspection of the record. On account of this more favorable pro- portion between the secular and the fortuitous momen- tary amplitudes the chances for a successful elimination of controversy are more favorable in the application of climatic sequences to hydraulic engineering than to weather service. Although the recent fluctuations in climate are small, i t is not necessary that these changes should be large in order t P become of great practical im- portance in hyd~aulic engmeering. Weather service deals largely with momentary values and changes. The task of hydraulic engineering is the control and conservation of water; hence-it deals largely with cumulative values. Precipitation is collected on large drainage basins and partially runs off. The run- of% is again accumulated in lakes and storage basins. The storage quantities and water elevations represent a repeated accumulation, and by t h s cumulative process small secular differences, unimportant for weather service but sustained over long periods, become great and important effects. The difference in evaporation from and flow into a lake basin may be delicately small, but the unlimited time factor may nevertheless cause cumulative effects of enormous magnitude, The Great Salt Lake fell 14 feet between 1875 and 1905, or in 30 years, while the climate was merely somewhat drier than usual, and the daily weather changes with the usual hail, rain, and thupder storms probably were not noticeably different than at present with a rising lake level. At the same rate a fall of several hundred feet might, in a sufficiently deep lake, well take place in 500 years; post-glacial lacustrine changes might well occur within the limits of small secular changes of no account in weather service, but sustained through thousands of years. The post-glacial changes nevertheless were very great. The Great Salt Lake shrunk from 19,750 to 1,750 square miles since the retreat of the ice. The diluvial Lakes Agassiz and Iroquois disintegrated into the present much smaller Lakes Winnipeg and Erie (20). Nachtigal investigated the enormous extent of the diluvial Lake Tchad in Africa (a), while von Richthofen has carefully collected evidence of the much greater diluvial areas of Lakes Pangong, Chamoriri, Lob-nor, and many others in Turkestan and Tibet (22). According to Mucshketov, the Caspian Sea had more than twice its present size, and still has ancient beaches 180 feet above its resent level (23); a connection existed through the d n y t c h Valley with the Black Sea and also with Sea of Aral, then more than three timw its present size. The streams which feed such lake basins are also the cumulative result of rainfall on’ a large drainage area. The erratic nature of rain-gage data is th6roughly modi- fied in stream flow. A drainage area of 10,000 square ndes has a collecting surface 805,000,000,000 times greater than an 8-inch rain-gage. Ground storage aids in equal- izing unequal distribution of rain as to place and time. An example of the equalizing effect of drainage areas on a huge scale is furnished by the Amazon River. Flowing along the Equator, its northern and southern tributaries are alternately in flood during the first and second half of the year, so that the flow of the main stream is equalized throughout the year (24). In addition to the equalizing effect of large drainage areas, ground and lake storage, the river functions as an amplifyer of the variations of annual rainfall. This is illustrated in the examples of Figure 1, giving the relation between annual rainfall and run-off. As examples are chosen the Austtble River, in Michigan, compiled by the writer, and the Devil Canyon watershed in California, given by SOndeTegpT (25). The run-off increases faster than the rainfall; a variation of one-third in rainfall may double the run-off, the latter increasing three times faster than the rainfall. The river functions exactly in the nianner of an amplifying radio tube, the characteristic curves having the same shape. On account of these fundamental differences between hydrographic and weather data, the importance of ~l i - matic cycles, such as the Bruckner and the ll-year cycle; h is not to be found in meteorology, but in hydraulic en- ‘ gineering. I t is measurable with current meter and sur- veying instruments rather than with thermometer, -- barometer, or rain gage, These fundamental differences are still overlooked in hydraulic engineering. Supposedly in accordance with the great measure of fortuity prevailing in the ceaseless weather changes, a fortuitous sequence of stream flow has been adopted in hydraulic engineering. The actually existing amplification of the small secular changes due to the cumulative nature of hydrographic data is not clearly recognized. Because the climatic cycles thus far found were, rightfully, disregarded in practical met,eorology, they were also disregarded by the engineers, although the conditions are fundamentally and totally different. OCTOBER, 1929 MONTHLY WEATHER REVIEW 407 The difference was carefully noted, however, by Pro- fessor Bruckner. In Chapter IS of his work, Iilima- schwanlrungen seit, 1700, he discusses the “theoretical and practical import’ance of the climatic cycle.’’ In this chapter only hydraulic phenomena and activities covered by hydraulic engineering are quoted as examples illus- trating the iniportance of the climatic cycle. Bruckner describes tQe secular changes in glaciers, lakes, and swamps; the number of floods; channel de.pth for inte,rior navigation; ground wnter and sanitotion; agriculture, especially in the semiarid regions. . These subject,s are covered by drainnge, sanitary, irrigation, rivers and harbors engineering. In Alpine countlies glaciers are now carefully studie,d by hydroelectric engineers. The remainder of Chaper IS c.ontains a scientific dis- cussion of the change.s in level of oceaI?s,e the choice of climatologicnl mean values, and of existing literature. It is significant that the e.rudit,e, and ingenious author of Climatic Changes thus de,rnonstrate,d that the c,ycle which he found by studying the records of SO4 stsations and 36,900 observation years is principlly of impor- tance in hydraulic engineering. Although it has no value for the mete.orologist in forecnshig the weather, the old work of Bruckner, newly i,llterpreted and properly applie,cl becomes a veritnble handbook for the hydraulic engineer, 4. IDENTITY OF THE BRUCKNER CYCLE. The interpretation referred to involves the relat,ion. between the climatic and s o l u cycle. Finding succes- sively a cycle in h k e levels, rainfall, pressure, and teni- perature, Briichier naturally suspected the cause of these to be solar variation. Investigation of such reln- tion began shortly after the discovery of the solar cycle by Schwabe. Bruckner quotes 36 authors on t,he rela- tion between solar and terrestrial cycles. Since then the number has greatly increased, but, no unaniinous verdict has been reached. Briickner recognized the e,sistence of several long cycles. “Three systems of oscillations exist which nre superimposed upon and interfere wit,h each other * * *” (p. 323). He also state.s that the longest one is a “diluvial” cycle; only t’he two others control 1iist.oric times. However, using only a stat,istical compilation of 6-year avera.ges, he does not find a relat,ion with t,he Wolf numbers. He state,s, nevert,heless: “I do not deny blie influence of sun spots on weather. On the contrary, t,he above table shows distinc,tly that influence in certain details although the longer variations seem iiidepe,ndent * * *. Much speaks for the t8heory discussed a b n e , that this force resides in the sun, t,hat therefore t’hc solar radiation shows a 36-year period independent of sun spots. It. may cause surprise that suc.h oscillations of solar radiation hitherto have escaped attenbion : at least I do not know of any phenomena on the solar disli hav- ing a period of 36 years. But the measurement of solnr radiation is a t present (1890) still very imperfect. We even can not demonstrate difference,s which are known to exist and can be computed. The quantity of heat received on earth is one-fift,eenth gre.ate.r in perihelian than in aphelian, yet this difference has not yet been measured. How much easier, then, could a secular dif- ference in intensity have escaped att,ention, the amplitude of which is probably less than that and the duration many decades? ’) The difference which Bruckner finds between his own figures and the Wolf numbers is only apparent. It is mainly due to his exclusive use of 5-year averages, whic,h is insufficient to segregate the cycles. In a previous study (26) agreement between 6he two was illustrated with several examples. It is of interest to further examine Rruckner’s own figures of wine harvest and temperature. The curves are platted in accompanying Figure 2. By using 6he previously given method (26) they can be segregated into the “Bruckner cycle” and the “secular cycle,” as shown. It niay be seen that with a slight dis- placement of the wine-harvest curve which Bruckner already note,d, bot,h are in satisfactory agreeme,nt. In Figure 3, which is an ext’ension back to 1750 of Figure 3, page 293, M.W.R. 1926, these curves appear to run parnllel witah the Wolf numbers. Bruckne,r does not segregate the Bruckner from the secular cycle, and hence his length of period is different. He designates t,he interval 175G-lS05 as one period, whereas in reality i t consists of t,wo periods of the Bruckner cycle, but only one of the secular cycle. He arrives in this manner at a greatly vnrying length of the Bruckner cycle. Prior to the plat,ted records Briickner derives his length of period from mere tabulations of “cold” and “warm” periods, which should be re.garded uncertain. Adopting 1733 23 33 53 ’73 ’83 33 I603 73 23 33 ‘43 33 ‘63 ’73 193 FIGURE 2.-Bruckner’s figures of temperature and wine h:irrest the above separntion of the Briickner and secular cycle, the average length since 1770 is 24.5 years; since 1860, 32.6 years. The 33.6-year period seems to become the accepted solar cycle instead of the ll-year cycle; striking is the great regularity of the cycle since 1S60, as pointed out by Abbott (3). Also, we have since 1860 three times a high 1 1-year maximum following a low 1 l-year maximum. This has been referred to by Clements (3). Prior to IS60 the alternately high and low masimum of the ll-year cycle is lost. Whether this is due to greater uncertainty of the Wolf numbers prior to 1S60 or to an “evolution of the law,” n s discussed by the celebrated mathema- tician, Henry Poiiicarb (31), will be decided in the future. Should the permanency prove true, then the Bruckner cycle is the polarity cycle of the sun spots, superimposed on the secular cycle. The secular cycle is also variable. From all data available the three last periods are estimated 70, 60, and 90 years in length, or an average of about 73 years. The first period is derived from Douglass’s sequoia curve (1911, 11 trees). But is this reversal of polarity permanent? 408 MONTHLY WEATHER REVIEW OCTOBER, 1928 While Bruckner, with much ingenuity, has sought to systematize the climatic oscillations of long duration, it appears that the conditions are not as simple as lie describes them. The alleged reversal of the cycle over the oceans can not be generalized. Undoubtedly the Gulf St,ream has much to do with the reversal of the cycle over the oceans as lie found, for, while existing on the west and east shore of the Atlantic, i t is absent along the California coast. Apparently the Briickner cycle changes phase with geographical location in greater nieasiire than he found. available, but still more are required to properly estab- lish the relations here discussed. Tile abow should therefore be considered tentative and suhject tu future confirina tion. As Briickner already emphasized, the phase of the cli- matic cycle varies with geographical location. This Since Briickner’s time 40 more years of observ :I t’ 1011 art1 beach will c9we numerous seconclsry waves and ripples The division into continents ancl oceans, warm and cold ocean ciirrents, plateaus nnd mountnin ranges, influence the course of atmospheric circulation. A standard regi- nieii of circulntion can he visualized only as existing within wide limits. The records indicate greater regularity in those from plains than from mountainous regions. Sucli :t standnrd system of niiiltiannusl oscillations is sulmierged in the daily we:ither clianges, hut beconies ap- parent in the cumulative cflects of run-of€ and storage. While meteorologists, therefore, rightfully claim that a relation between sun spots nncl wenrher has not been suc- cassfiilly deiiionstrntcd, tli2t c.l:iini can not be transferred to liytlrographic- records. Tlitw, often in unison with the Wolf nu nibers, tiibplity oscil1:it ions ()f f nr-rea ching econom- ical iiiiport:inc.e, which follow n COUI’RC apparently (but only npparentlv) indqmideiit from the also economically important u enther chnnges with tlicir ceneeless succession seemingly is in wcorclance with the results of Kiil1merj who show the path of cyclones to vary with the sohr ~?7cle (32). In a previous study i t was shown that nnt only the Brucliner cycle can be identified in the Wolf nuinhers, hut the smaller cycles as well (26). Besides the two cyrles named above, the 11-ymr, the “double sirii-spot c!ycle ” (after Douglass) of half that, period, and the “T7oiiqil” cycle are the funduniental units t80 bc found in nenrly every record. The latter can oftmen he separated into tivo cycles of nearly one-quarter and one-eighth of the 1 1 -y ~:~ cycle; but, this is of no practical import. d .l C C . 7 I t should not he inferred that a universal ant1 rigid all- plicability should be at,tributecl to this conception. In some regions the relahioris are clearer than in others. A stone thrown into n quiet pond will cause a concentric and symmetric wave, but the irregularities of an adjacent of frnsis, hot nntl cold wuves, r:iiii, hail, mow, ancl thun- tlcr s t ~r :~i s , tlro:lgllts, :~n c l dot1ti-l)tlr&. The lnt,ters’ hcrtic si:cccs;si!~ii is ~iic~rg~tl intn resiJusl system of slow hut cc.rt:iin scqurnces, whirh at lenst in some regions can l w rorrcll:itec! c.learlp wit11 the Wo!f numbers. 5 . EFFECTS OF THE C‘LIhIATIC CICLES The lake and ground-water levels rise and fall with these cycles. The Great Salt Lalie (fig. 4 ) is again near- ing a period of expansion similnr to that prevailing be- tween 18(;0-1S’iT,. An increase in area of some 400 square miles around 1945 mny well be expected. The damage thereby caused to riparian owners, to roads, bridges, qunys, wh~rves, ancl docks is obvious. Riparian struc- tures are iisaally erected oblivious of the slow but sure rise and fall of the water level, which are not apparent to the uninformed. Only in locations where the liniits occur OCTOBER, 1929 MONTHLY WEATHER REVIEW 409 frequently, as happens along the rivers, are they heeded The trestle of the Southern Pacific Railroad across the Great Salt Lake will probably hnve to be raised, as well as tlie structures of hathin: resorts nenr Salt) Lalie C‘ity. The Great Salt Lake begnil to rise simu1tar:eously u itli the extension of the Rlorrnon settlements in Vtnh. Cyrus Thomas and EIoi~yh (27, pp. 71 m r i 03) escribed the im- provernent o f climnte to ~~iiciiltiire. * The rivers nntl brooks feeding the lake carried iixm water than i:su:d: :t greater amount could hc tlivcrted for irriptiim piirpows. This was supposedly (1iie t o increased area under calti- va ti on, c a usin g incre a sed e\* n pom t i on a n tl conseq 11 en t 1 ~- rainfrlll. The theory pro1 ed fiilqe when after 1S75 the lalie fell continiioiisly until 1005, and the irrigsted area had to he reduced. Since then the lake lias risen, ancl the nest period iintil around 1945 I d 1 probnhl:,. be chw-ac- terizetl by improved gr:>zirig :iud farming conditions all through the arid West. Typical for the last, clown\\ arc1 swing of the scculnr cycle, beginning around 1s75, niininiiim around 1905, :md nest masiniuni probahly iiroiind 19-15-1950, is the fol- lowing opinion based on extensive ohserration during the last 25 years (28): Prior to the advent o f tlie white ninii rrohion <\:IS n matter o f minor iinportaiice in the aricl Southire~t. At that time there \i n s generally an excellent grass carpet \\ hich retarded run-ufl at id prei eiited ernsiuii. Overgmzing by the cnttlr and htieep iiirlri.tr> has depleted. in fact, nliiiost e\termiiintcil, the gin\\ c:trpet 01 cr large areas, resrnting in erosion mid rleiiuclation. Rspirl re.tura- tion of overgrazed areas 11:~s tnketi Illace during n Yeriez IJf v e t yearb followiiig A serieb of dry ye2rs. * * * Kirk Bryan (29), finding geolopical evidence of erosion long before overgrazing by the white man’s cattle, states: “It, appears that these cyclic changes have a cornnion and doubtless climatic cause. The introduction of live- stock and the ensuing overgrazing should be regarded as a mere trigger pull, which tiriiecl a change ahout to take place.” At the desert bortler a persistent, cuniii- latire though sinall deficiency of rainfall just below the necessary niinimuni gradually overcomes the tenacity of life of the vegetation. Plant life perkhes, followed by wind erosion, denudation, and desolation. This uii- doubteclly is hastened but not caused by overgrazing. Briicliner stittes: “During rnininia the dry and ~varni years will be somewhatj Inore frequent8 than the moist and cool ones. A forecast of this cycle is of no signiii- cnnce for Europe, as here the variations of rainfall, which are of principal importance, are rather moderate. Of value such forecast wonk1 be for t,he contihental regions, where t,he variations of rainfall are much more accentu- ated: for Siberia, Aiistrdia, and, above all, for the inte- rior of North America.” Briickner also gives statistics to show that, while in the continental regions t,he wet period of tlie Briickner cycle accelerates agriculture, tlie reverse effect, occurs in the normally humid niaritinie climate of Europe. The Great, Lakes esliibit the Briickner cycle in close agreement with the rainfall ; but 3s the lake lei-el is pro- portional to the total inflow, tlie cycle of lalie level lags one-quarter period aqainst the cycle of rainfall. The 11-year cycle, of sinnll amplituclc in the rainfall of tliese regions, again is enlarge(1 in the 1:ike levels. At present n ninx~niuni prevails ; the next iiiasiniuni, espected to occ~ir iiround 1940, v-ill, on account of the rise in the Briicliner cycle, probably he almut a foot’ higher than the present niasiniuni. In this region the prnctic:il consequences of the faint cycle in climate are nieasurahle in inillions of dollars The d u e of 1 foot additional draft to navigation is estimated a t $5,000,000 (30). FIGURE 4.-Grent Salt Lnlie, Lake George, Anstralin. and Wolf nuinhers In another city a large public-utility corporation installed tit grtwt cost a booster pumping plant, for the circiilating n-atc.r of its siiperpower station, hut due to subsequent rise in level the large investment was never used. The vnrying yirltl o€ streams also affected water power. Another piihlic utility, fearing. continuation of the dry ye:irs prior to 1926, hurriedly installed adclitional boiler capacity to take care of the reduced hydroelectric output, but the large investment, proved superfluous for the oricin nl piirpose. Such are the far-reaching effects of faint cycles in cblimate, hnviiy an aiiiplitude of no importance in d d y v, cntlicr qervicc. 6. .APPLICATION OF CORRELATION WITH WOLF NUMBERS. ,S t f l ) B. Aicholso~i (3‘) n:lvocntetl the use of the C’lisr- acter ?';pres of So1:ir Phenumena, publi4iecl in Zwich, instead o f the Kolf numbers, for con?p:irison with tcrres- t r i d cyclrq. According to A h . Nicholson, the Kolf nuniher is equal lo tlie number of spots plus ten times the number o f groups; a c d it should therefore not surprise t h c invc\atig;ator of c.3-cles that such an arbitrary constant 410 MONTHLY WEATHER REVIEW OCTOBER, 1929 should fail to give consistent results throughout. Never- theless, a surprising agreement with terrestrial cycles may be found in some records. The unbroken sweep of open count,ry in the Great Lakes region seems especially favored with a regular systeni of atmospheric circulation. Broni the rnin-gape readings of t’he United States Weather Bureau the cycle was derived already 30 years ago by the prominent ligdro- logist, Robert E. Horton. In Water-Supply Pnper No. 30, 1899, A h . Ilorton discusses run-oil’ and water power FIGURE 5.-Ralnfall, Ausable region, Michigan of the Kalaniazoo River, and therein clearly recognizes a cycle which, after 30 more observation years, proves to be the double sun-spot cycle. Mr. Horton predicted the next minimum to take place in 1901. The forecast was niade in 1898, or three years in advance. Figure 5 shows that this forecast was esactly verified. Figure 5 is a graph of the rainfall in the Ausable region of the Southern Peninsula of Michigan, computed from data furnished by Mr. D. A. Seeley, United States meteor- ologist in Lansing. The average of 10 stations is resolved into the “Clough cycle” and a residual, which is coni- posed of the double sun-spot cycle superimposed on the Briicliner and the secular cycle. The 11-year cycle is very small in this record. It should be emphasized that an average of ninny sta- tions is needed to eliminfitc fortuitous errors, distributed, according to H a m , over an area not more thnn 200 miles in diameter. The Ausnble watershed is only smie 1,600 square miles. It may be seen that the cycle discovered by Hortoii has faithfully recurred since 1899. Indicating the ninsi- mum of rainfall on the secular cycle of the Wolf nunibers with a black dot, it may be seen that this occurs alnrtys before each maximum or niininiuni of the 1 l-year cycle. I t has faithfully recurred ten times during the last 54 years. We have here in reality a double sun-spot cycle. Transferring the cycle to Figure 3, the inverse relation of the Briicliner cycle in Wolf nunibers p.nd rainfall is distinctly seen. The secular cycle is here, as in ewry other record always independent of geographical loci;- tion. A continuation of these relations in the future, v;hi!e not known with ceriaint,y, is probable. The conclusion may be drawn that, around 1940-1950 the rainfall mny again have increased 30 per cent up to the values prevail- ing around 1580. The average run-off will then be doubled. Six water-power plants were built on the Ausable River since 1912, a t a cost of some $12,000,000. The output may he expected to increase 80 per cent in the above-named period. The effect on the returns on the invested capital is obvious. The same increase will be felt over a considerable part of the country, and will materially contribute to a renewed popularity of hydroelectric power, which is now laggin behind steam power on account of t’he greatly increase economy of the latter. The double sun-spot cycle in rainfall is, as stated above, amplified in stream flow. Figure 6 shows the flow of the RIuslwgon River, an adjacent, watershed. It may be seen that the double sun-spot cycle is plainly visible. By means of this and other graphs the deficiency of hydro- electric power in 193 1 is for one power company computed to be 1G0,000,000 kilowatt-hours, resulting in an increased expenditure for steam coal of some $500,000 for that year. A deficiency in peak capacity of 20,000 kilowatts is also estimated for that year. This means either a new investment of $3,500,000 or else purchase of power elsewhere. Nevertheless, the weather changes in 1931 will probably alternate as usual 1% ithout any apparent connection with the Wolf nunibers. The latter’s unniistaliable connection with stream flow, how-ever, will a t the same time influence ublic-utility engineering to the extent of millions of Xollars. w The continuity instead of the fortuity of stream flow enables a close estimatie of nest year’s run-off. AE described in MONTHLY WEATHER REVIEW, March, 1928, it is possible to make such estimates by extrapolation. The esaiiiple there given actually checked within 5 per cent Another estimate niade that year placed the mean flow of the Winnipeg River a t Great Falls in 1928 at 55 per cent of that during 1927. It actually was deter- 8 !9i3 19!4 19i5 :9!b 1317 1918 1913 1920 1921 1922 1923 19?4 1925 1926 1927 1928 FIGURE B.-Run-off, Muskegon River, hlich. mined by the Dominion Bureau of Reclamation and Power to be 60 per cent of thaB in 1937. The illustration Figure 7 is a graph of the run-oft’ of the Winnipeg Riyer. Figure 4, the elexation of the Great Salt Lake, the Aust,ra.linn Lake George, and the Wolf numbers, shows 1il.rewise close reititmion wit,h emh other. Of int,erest is the inverse relation of the 11-year cycle in the Wolf nuinbe.rs and the Great Salt Lake. The graph wm made in 1924, and it could then safely be announced that a drop in lake level would take place on account of that OCTOBER, 1929 MONTHLY WE relation. .This actually occurred, and a rise culminating in 1936 may well be expected, increasing the lake level at least 4 feet over the present. On account of the simul- taneous rise in the Briickner cycle, that figure will prob- ably be exceeded.’ I