MONTHLY WEATHER REVIEW 279 THE RECORD-BREAKING COLD WAVE OF MID-NOVEMBER 1955 IN THE NORTHWEST J. F. O'CONNOR AND C. R. FEAN National Weather Analysis Center, U. S. Weather Bureau, Washington, D. C. 1. INTRODUCTION A record-breaking cold spell hit Montana on the 11th of November and spread westward and southwestward through the Plateau, persisting in severity for a week, November 11-18, 1955, and causing heavy losses in unharvested crops and produce in unheated storage. A t Helena, Mont. the temperature remained below 0' F. for 138 consecutive hours. After this period the cold air pushed into eastern sections causing near-blizzard conditions in the northeastern Plains and Lakes region and producing cold waves as far east as the Appalachians. This very cold air was eventually responsible for the first general snow in the Northeast and the great contrasts in temperature it provided as it lay over New England on the 19th and 20th helped to spawn a most severe storm off New England. It is frequently the case that when unusually severe cold persists near one coast, the opposite extreme may occur near the other. This is due to the fact that below normai conditions are usually associated with abnormally deep persistent troughs aloft, while above normal condi- tions are associated with persistent ridges aloft. The distance normally found between long-wave troughs and ridges is such that a trough over the Plateau will accom- pany a ridge over t'he East and vice versa. During the period of the situation to be studied the 500-mb. height's in the Northwest were below normal by record amounts, while in the East they were 200 to 400 ft. above normal. Thus while much-below-normal conditions developed in the Northwest in mid-November, temperatures rose into the 80's in the East producing such late-season record high t'emperatures as 81' F. a t Huntington, W. Va., 79' a t Cincinnati, Ohio, and 78O a t Baltimore, Md. This marked warming in the East, in contrast to the extreme cold in the West, set up ideal conditions for cold waves in the East also, when conditions finally became favorable for the cold air in the Northwest to move eastward en masse. The cold Arctic air poured initially into Montana on the 11th causing one of the most severe cold spells in history for so early in the season. By the 13th it had produced subzero minima from North Dakota westward through the northern border States to northeastern Washington. Subfreezing minima occurred along the northwest coast and southward into California's Central Valley killing tender vegetables and unharvested grapes. On the 15th many stations reported record low tempera- tures for so early in the season, e. g., "29' F. at Helena, Mont., "11' a t Spokane, Wash., -3' at Boise, Idaho, and -9' a t Ely, Nev. On the same day Portland, Oreg., FIGURE l."Chart A shows departures of average surface temperatures from normal, November 12-17, 1955 inclusive (solid lines) and mean 500-mb. height contours (dashed lines) labeled in hundreds of fwt. Chart B gives lowest temperature minima reached during same period. 280 MONTHLY TT'EL\THER REVIEW NOVEMBER 1955 FIGURE 2.-(A) Surface chart and (B) 500-mb. chart on Kovembcr 9, 1955. On surface chart 1,000-500-n1h. thicknesses (dashed lines) are labeled in hundreds of fect. Previous 6-hourly positions of centers are indicated by "X". On 500-mb. chart isotherms (dashed) are labeled in "C. and height contours (solid) in hundreds o f fcet. Flags on wind shafts represent 50 knots, full barbs 10 knots, and half-barbs 5 knots. reported a record early-season low of 13' as subfreezing minima ext'ended southward along the coast to Eureka, Calif. At Salt Lake City, Utah, a "14' F. low on the 16th was 14' lower than ever before recorded during November. The average temperature on this day was lo, or 38' below normal, the greatest daily departure ever recorded at Salt Lake City during any month. In Washington it was the longest severe cold spell on record for November, wit'h Ast'oria having 135 consecut'ive hours below freezing. Wyoming experienced average tempera- tures nearly 25' below normal wit'h several stations in the Yellowstone Basin reporting average tcmperat'ures near or below zero. Farther east Kebraska had the coldest mid-November weather since 1940. St,. Joseph, Mo. had the coldest November on record wit'h 10' F. on the 16t'h a new daily record. Figure 1A gives the depart'urcs of average surface temperatures from normal during the period November 12-17 with the mean 500-mb. flow superimposcd. The 40' below normal in Montana for the period strongly contrasts with the 17' above normal in the Southeast. Figure IB gives the lowest minima report'ed by each station during this period. The protract'ed nature of t'his cold spell makes it con- venient to divide the period into three parts. 'The initial phase embraces the first surge of cold air into the Korth- west on the llt8h and l2t'h, the middle phase covers the period during which tlle cold Low aloft performed a loop and came back to its st'art'ing point over northeastern Washington, and the final phase starts with the 15th as tlle upper Low and t,he associated cold dome moved out of the Nortllwest~. I t will be seen that t'llis cold spell was not cl~aracterizcd bv abnormal surface anticvclonic conditior~s hut rather was due to the persistence of an abnormally cold Low at upper levels. The thickness of the air column directly under t>he upper low center bet)ween 1,000- and 500-mb. pressure levels was generally 1,400 to 1,600 ft. thinner (21' to 24' C. colder) than normal for almost a week. Such a deep layer of cold air had an important bearing on t'he extreme surfacc temperature anomalies observed, since the thickness of an isobaric layer is directly propor- tional t'o the mean temperature in the vert'ical. I t is our purpose to study and record some of t'he synop- tic aspects of the situation which produced such a record- breaking cold spell both as t'o intensity and duration at such an early datc. It is also planned to examine some details of the vertical structure of t'he cold dome. 2. ANTECEDENT CONDITIONS AND PROGNOSTIC INDICATIONS Figure 2A shows the conditions a t the surface a t 1230 GMT, and figure 2B shows the 500-mb. picture a t 1500 CMT on Wednesday, Xovember 9, about 36 hours prior to the first penetration of cold air a t t h e surface int'o Montana. At this time t'hcrc was a cold core of somewhat less than 15,800 ft. (fig. 2A) in the 1,000-500 mb. thickness field over Alaska. This is equivalent to a mean temperature of "36' C. in the column of air bclow the 500-mb. Low. A weak wave cyclone was just' entering the coast at the surfacc about 600 miles south of the upper Low, in such a wag.' relative to the upper contours that it was approach- ing the delta or exit region of an upper confluence pattern. Accordingly, most forecasters would look for a t least some deepening of the surface storm, not only because of the contours aloft, but because the wave cyclone would be approaching an area in the lee of the Continental Divide which is normally divergent. However, in order for the wave to deepen enough to produce appreciable cold ad- vection to its rear, while moving parallel to the upper flow, the upper heights would have to fall simultaneously. A southeastward movement of the upper Low and its asso- ciated cold core from Alaska would thus t'end to result. Therefore, it seems clear that upper heights would have t,o fall over southwestern Canada if these effects were to he realized. No height falls aloft had been in evidence prior to this time which could be extrapolated toward southwestern Canada. Thus, the signal t h a t t h e eold Low aloft was about to plunge southeastward into Wash- ington State, Idaho, and Aiontana, was certainly not clearly in evidence at this time. It should also be pointed out that in t,erms of large-scale phenomena the clue to the turn of events appears likewise elusive. There had been no sudden or abnormal deepening evident upstream or in fact anywhere over Asia prior to November 9 which could have dispersed eastward at the group velocity. Computations of moveme,nt of the 500-mb. Low using grid methods of averaging the geostrophic flow around the vortex such as the Wilson [I] adaptation of the Riehl- Haggard method, failed to indicate the nature of the en- suing development also. As can be readily guessed by merely looking at, the Sovember 9, 500-mb. contour pattern over Alaska and environs in figure 2B, tho existing momentum w~ould carry the Low east and northeast. In fact 24 hours later a filling Low cent,er northeast of Aklavik, N. W . T., did verify tlw computat'ion well. However the main 24-hour 500-mb. height fall plunged southeastward to British Columbia and increased in mag- 37030&-56-3 nitude from 600 f t ,. ovcr ~Uaska, to 1,000 ft. over British Columbia at' 1500 GAIT 011 the 10th. This 24-hour 500-mb. height fall area, in t'hc sul)scqumt~ 24 hours, again plunged sout'h-sollth~\.cstu-arci TOSS the initial contour pattern into Oregon t'o produw the 500-mb. pattern in figure 3B. Although the grid mclthotl did vcrify fairly well in 24 hours, it' sccmed to be a fortuitous rwult of residual low pressure ill the verifyillg :xrca, u~hilc the tleljelopmental kinetic ener{gg wa.s actually plwngilzg southeastward and southward, propagating the hcigllt falls nut1 the bulk of t8he cold dome with it. Of course it, might bo said that the grid method isn't t>xpwt,ed to predict this aspect of what occurred since it should only work for barotropic or srmil-)arot8ropic cwxlitions. Actually most, upper Lows are at least similarly baroclinic and even so respond motleratcly well to the mean flow a,round t,he periphery, so long as t#Iwre are no environmental factors operating to change tlw pcripheral mean flow. Of course, with suc-h rnvironmental fact'ors operating, the longer the c>ffects of t,lw initmid conditions are projected forward, the greater will bc the error in the final result. Zn retrospect however, the following clues, which were available on Kovcmbcr 9, seem to t.ake on added impor- t'ance in explaining somc of t,he changes in the peripheral circulat,ion which produced the t'remendous height falls over the northwestern St'at'es within 48 hours after the initial c~ndit~ions pictured in figure 2 : (I ) The major ridge a t 500 mb. had retrograded in the previous two days from a position inland over eastern Washington. Furthermore, t8here had been persistent, height, rises in the Gulf of Alaska for a t least 36 hours indicahg rat,her strong ant'icyclogenesis was taking place. 282 MONTHLY WE,\TI-IER REVIEW NOVEMBER 1955 (2) Cold advection at 500 mb. was indicated to\vartl southeast Alaska, t,ending t o lower heights over south- western Canada. (3) Warm advection over the Bering Sca pointed to further height rises over western Alaska, which would serve to keep the gradient strong on the west side of the upper vortex. This would favor propagating the momcrl- turn of the northerly flow southeastward, as t'he gradient t'o the southeast weakened due tlo falling heights in that area. (4) The fourth factor was the strong confluence zone at 500 mh. straddling the coast near 55' N. The entrance zone off the coast may be interpreted as an area of accu- mulation of air and the exit zone over southwestern Canada as one of depletion. This would tend to produce height rises and anticyclogenesis a t t h e entrance and lwigllt, falls near the exit of t'hc confluence zonn causing further retrogression of the ridge off t.he west coast, allowing t'he Alaskan Lorn to plunge southeast'ward. In tLtldit'ior1 to the failure of the grid mcthod of compu- tation, it should be point'etf out t'llat other auxiliary tools normally a~ailahlc for prog~losis also did not indicate what was to t'ranspire. One such tool, t'he advection of vorticity by the 500-mb. FjGrtoft [a] space mean flow (not shown) not only did not indicate what was to trans- pire, but in fact indicated that the vorticit'y would move to t3he northeast from Alaska. This is a common failing of the FjGrtoft method. It cannot forecast t'his impor- tant mechanism of plunging since this very process radically alters the space mean flow. The hemispheric numerical barot'ropic prognostic simi- larly did not predict the penetration of the cold upper Low into the Stat'e of Washington. However no develop- mental virtues have been claimed for the barotropic prognosis. Neither did the baroclinic numerical prognos- t'ic predict, this event' but this is of course not surprising since the vort'icit'y responsible for the devclopmcnt, originated beyond the northwest boundary of the data network. The program employed assumes t ' h a t t h e vorticity advection is zero at the boundaries. I t may well be that the baroclinic prognostic would have predicted t'his event if thc data nctwork included the Alaskan area, but this has yet to be shown. Extended forecasting methods of predicting readjustments in large scalc features also fell short, of predicting the onset of this cold spell in thc extreme Xorthwest, indicating indcfinite or conflicting evidence in the large-scale fcatures. It \vas not until 1500 GMT on tho 10th that indications were sufhcientl~ defirritc for this severe condition to be included in tlrc sw.\c ;io()- mb. prognosis for 0300 GMT, November 12. Sinw thc Gulf of Alaska region and t h Yulwrl are favored sites for the gcnerat'ion of dynamic instabilit,y [3] with cnormous effects on the weat,her ovcr the Unitcd States within 48 hours, much more study mustj hc: matie on this problem, and particularly in this area. 3. INITIAL PHASE OF THE COLD WAVE By 1830 GMT~ on Novembcr 10 (not shown), an Arctic cold front at the surface was crossing the Canadian border and about to invade Alontana, having moved south~vard in the rear o f a deep (982 mb.i surface Low in hIanito1)a. This deepening had resulted from thc w l -c d c surface n-uvtl cyclone wlrich had crossed tllc coast of southcast Alaska on the 9th dircctly bencatll tlrc strong upper l~cigl~t' f:rlls plunging southcast'w-artl. By 12XI GMT on thc 11 tlr (fig. 3A) the surface Arctic air was we!l er~t~~wchetl ilr the northwestern Plains while the 1500 G w r 500-mb. LOT\- was centered ovcr Brit.ish Columbia (fig. 3 3 ) w-ittr Princc George, B. C., reporting -45' C., indicating an al~norrnt~lly cold upper vort.ex indeed. BJ- this time 500-mb. hcights had fallen in 48 hours almost 2,000 ft. at Scatt'lr, Wash., while 1,000-500-mb. thicknesses were more t>han 1,400 f t ,. (21' C.) colder than norma,l in thc vicinitS of the upper vortex. This resulted in surface maximum t'emperatures in Montana on the 11th being 40'-50' F. lower than the previous day, with maxima in the north central areas not going above 10' F. At 500 mb. on the l l t ' h (fig. 3R) the gradient over t'he northwestern States was very wcak with average winds of about 30 knots, while t'he northerly current west of t'he Low was relatively strong, averaging 75 knots or possibly more. By 1500 GMT, November 12 (fig. 4B) t'lrc upper Low had been propagat>ed southward t'o the weak gradient area just north of Spokane, Wash., by the strong northerly jet along the coast of southeast Alaska and British Columbia. The 500-mb. temperature lowered to -43' C. a t Spokane and 1,000-500-mb. thickness departures from normal increased 400-800 ft. from the previous day half of the atmosphere near northeastern Washington and northern Idaho. This caused a further lowering of surface maximum temperatures over the northwestern States on the 12th. From t8he previous day's values of 10' to 20' F., maxima dropped t'o 0' t,o "10' F. in eastern Montana and into the 20's along t'he coast of Washington. Figure 4C shows the departures from normal of mini- mum tempera,turcs and figure 4E shows the departures of maximum temperatures in comparison with the 1,000- 500-mb. thickness dcparturc.s on the 12th (fig. 4D). The minima were 30' F. or more below normal while the maxima were 50' F. or more below normal at some areas in tlrc Sorthwest. By the morning of November 13, minima ranged from :!do below zero at Cutbank, Mont., to near zcro from the Dakotas westward through Wyoming :mtl nortlrcrn Kevada, except for the coastal regions. Scattle had a minimum of 1 3 ' F., the second morning the minimum ctroppctl bclom 15'. 4. MIDDLE PHASE OF THE COLD SPELL 'l'lr(> contours arourltl the Low ncar Spokane, Wash., a t 500 mb. on Sovembrr 12 (fig. 4B) exhibit, considerable ccccntricitJ- (80-knot winds west of the cent'er and 30 knots cast of the ccntcr) suggesting that the maximum q-clonic vorticit>y would propagate toward the weaker gradient area t o i t s cast'. This act'ually occurred but only in the form of a short-wave trough moving east- nortllcastwartl toward James Bay (southern Hudson B ~J -). At tllc same time a wcak-appearing short-wave FIGURE 5,"Track of 500-mb. Low during period ,RuTovemher 9-17. Date and time groups are separated by slants wit,h departures from norrnnl of central height values (feet) enclosed in paren- theses. Successive positions of center at 12-hour intervals are producing a departure of 1,600 ft. (24' c.1 in the lower indicated by "X". FIGURE B.--ru’orember 14, 1955. (A) Surface chart,, (14) 500-mb. chart, (C ) departure from normal of surface minil:Inm le~npcr:tt,llrrs, (D) 1,000-500-mb. t,hickness drpnrt,ure from normal, (E) depxrtllre from normal of surface mauilnum t,empc~rxtnrc~P. undulation of the contours was rounding the long-wave ridge in the strong westerly current in t’he Gulf of Alaska, having resulted from a 24-hour height fall of 600 ft’. in the ridge over this area. This perturbation was associated wit,h a weak “A” (CIT) type Low [4] at the surfacc in t h t Gulf of Alaska (fig. 4A). In the absence of such an approaching short II-ILVC~, height rises probably would have occurred in the stronger gradient area to t,he west, of the 500-mh. Low near Spokane, and the whole vortex would have moved east’wa,rd ending the cold spell in the h-orthwest. However t h e rapid southward t,ranslntion of t,his minor feat’ure down the east side of the long-wave ridge in the Gulf of Alaska. introduced height falls into the west and southwest quadrants of the main upper vortex. The net efTect on the main cold core circulation was to produce II small c.yclonic trajectory to its cer~tcr over Brit’ish Columbia (fig. 5 ) together with it,s associatccl 1.000-500-mb. thick- ness anomaly, so as to ret)nrn tllc (*old corc to north- ctlstemWashingt,on on the 14t,h. 1-31- this timc t ,l ~e surfacc “A” t,l-pe Low was in the central Plattau (fig. 6A) and the 1,000-500-mb. t,hickncss departures ww again more than 1,600 ft. (24’ C.) colder t,llan normal ovcr Washing- ton. The associated tlcpnrturcs from normal of t h e surfacc maximum and minimum tcmptraturcs (fig. 6c ant1 E) wCre thus ver;y similar to thosr reported t,wo clays earlier. For cxamplc, below zero minimum temper- atures were registered on the 15th in east,crrl Washington, Oregon, Idaho, Montnrla, Wyoming, northern Utah, and Ncvada. Seattle had it’s lowest minimum of 6’ F. during this phase of the cold spell. Ukiah, Orcg., a little south of Pmdleton, reported a “32’ E’. minimum. FIGURE 7.-November 15, 1955. (A) Surface chart, (B) 500-mb. chart, (C) departure from normal of minimum temperatures, (D) 1,000- 500-mb. thickness departure from normal, (E) departure from normal of maximum temperatures. The surface pressure pattern for 1230 GMT, November 14 in figure GA has long been recognized as the type whic~ll precedcs cold waves in Utah. In 1931, Hales [5 ] wrote that for a cold wave to sweep the Utah section of the Bockies, the center of the High must bc placed well to the west of the Divide with a vigorous Low emerging from the coast of Oregon or Wastlingt,on and passing to t.hc southeast through southern Utah. Such Lows are now generally referred to as “A” t ~p c Lows aft,er I. P. Krick and R. D. Elliott, who originall)- developed the CIT weat’llcr t’ypes [6]. A necessary con- comitant of these storms aloft’ is a tlechp cold trough ovcr the Plateau near 115’ W. and a stat,ionary long-wavc ridge near 145’ W. The first phase (day) of this type is when the Idow breaks off (“skagerraks”) in the Gulf of Alaska. In the current sit’uat’iorl this was on thc 12th. %70308--5G”-4 Average timing for this type requires that it be a deep storm near Moosonee, Ont., about 6 days later. The current analogue reached Moosonee on the 17th after passing across the Lakes as a severe storm. Perhaps the most famous “A” type analogues of recent memory were those of January 1949. The maps of January 7-9, 1949, both surface and 500 mb., were very similar to the current situation from the 9th to the 14th. However there the similarity ends. For subsequent to ,January 9, 1949, the cold cyclonic core plunged south- southwestward to southern California, stagnating there for a few days and causing snow a t sea level in the Los Angeles Basin. In the current situation the cold vortex behaved in a more conventional fashion. To return to some of the more remarkable details of the Novcmlwr 1955, situation, t>he surface map in figure 6A 286 MONTHLY NOVEMBER 1955 FIGURX 8.-.Departures from normal of surface minimum temperatures (" F.) shown in chart (A), departures of 1,000-500-mb. thicknesses (hundreds of feet) in chart (B), and departures of surface maximum temperatures (" 17.) in chart (C) for November 16, 1955. over the Plateau shows the extremely cold core of 16,200 ft. in the thickness lines in eastern Washington, with t>he strong pressure gradient in the rear of the surface TJow over Utah. This gradient intersected almost a t r i g h t angles the strong gradient of mean temperatsure lines packed between the Arctic front extending west from the surface Low and the core of the cold air farther north. Thus cold advection of a strength comparable to the most severe encountered even in mid-winter was indicated threatening Nevada and Utah antl points east and south- east. As mentioned before, in the ensuing two days Salt Lake City experienced temperatures lower than ever before so early in the season and the greatest departures of average temperatures from normal ever experienced in its history. The easternmost wave in the Gulf of Alaska on the surface chart of the 14th was not associated wit,h any upper height falls and therefore proved t'o be a stable wave and not a new "A" type wave. It, t,herefore had no bearing on ensuing development,^. 5. FINAL PHASE OF THE COLD WAVE During rc'ovember 14 the cold vortex drifted slowly east- southeastward into southwestern hiont'ana. Figure 7 shows conditions on the morning of November 15. AtJ 500 mb. (fig. 7R) cold advection toward the relatjively weak contour gradient was evident in the northern Plains suggesting this was an area of imminent Ileight falls. The speed maximum in the peripheral jet was located in the southwestern quadrant of the upper Low. At the same time a strong westerly stream had developctl from 42' to 50' N. along the meridian of ship "P", due t'o a strong 24-hour height fall of over 1,200 ft. just southeast of Kodiak. This picture clearly suggested large rises propagating into western quadrants of the upper Low i n the Northwest in the ensuing 36 hours. Together with indicated height falls east of the Low, the upper Low moved rapidly eastward in response to this isallobaric field. At t,he surface (fig. 7A) the "A" type Low was well organized as a wave cyclone over Iowa, and the Arctic a i r a t t h c surface was poised to push rapidly into the Plains. Tlrcsc circumst,alwcs contrast,ed sharply with the warm temperatures exist,ing or expected over the eastern part of the country. The departure from normal chart,s for the 15tfh show that maximum tempcrat'ures (fig. 7E) were in excess of 50' F. lower than normal in Wyoming contrast,ed with over 20' F. above normal from Ohio to Texas. Further- more minima (fig. 7C) wcrc 30' P. abovo normal in Arkan- sas that morning. Thus on the 16t)h as tho surface cold front moved east,- ward, sevcrc cold waves occurred as maximnm tempera- tures plunged 40' F. lower in tho west central Plains than t'llc prcvious day while t h y rose about 30' F. over lllinois antl Alissouri. Figure 8 shows tcmpcrat>ure and t3lickness departures from normal on thc 16t'h. Thicknesses (fig. 8s) were morc than 800 f t . above normal in Ohio producing maxima (fig. SC) morc t'han 20' F. above normal in the East, e. g.: St. Louis, Llo. reported a maximum of 81' F. Below normal thicknrsses of almost 1,500 ft. in Wyoming were associated with 30' to 40' below normal minima (fig. SA). The surface dorm deepened rapidly as it moved from Iowa to Sault Saintc Llarie, Mich., on the 15th and 16th due to strong cyclonic vorticity advection producing large height falls aloft moving rapidly eastward. This rc?sultcd in near-blizzard conditions with 2-9 inches of snow from the Dakotas to New England. On the 161% the cold wave edged toward the east coast and maximum temperatures from Missouri to northern Texas fell 40' to 50' below the previous day's values. By this time slow mnsming was in progress in the Korthwest. 6. THE 500-MB. TEMPERATURES Thc caoltlness of the vortex a t 500 mb. during this period is very impressive, even t'llough extreme temperatures for NOVEMBER 1955 MONTHLY WEATHER 287 constant pressure surfaces are not yet available in pub- lished form. An older publicat'ion, ''Extmreme Tempcra- tures in the Upper Air" [7] containing only very short records, shows that at 5 lrm. (about 16,400 ft.) thc lowest temperature measured up t80 that t,irnc over thc whole of North America was -51" C. in January over Alaska. For thc continental United St'at'es t'he lowest value was "48' C., again in January, at Bismarck, N. Dak. and Sault Sainte Marie, Mich. The minimum a t 5 km. over continental United St'at'es for Sovemhcr was "40' C. In tho November 1955 situat,ion, "48' C. was tllc lowest value report'cd when the Low was ill northwcstcrn Canada, while the continental United Stat'cs st'at'iolls reported 500- mb. temperatures as givcn in tablo 1 (for comparison the 500-mb. values for Nov. 1955 havc bccn corlvcrt'cd to 5 km. using a lapse rate of3' C./l,oOO f t . wllich is t'hc valuc that prevailed a t these levels during thc currcllt situation). TABLE 1,"Selected 500-mb. temperatures during cold wave in North- in ["] west, h'ovember 195.5, compared with previous records as tabulated Station 1 500 mh. __ Temp. "~ c. -42 -43 -41 -42 -40 -38 -42 -38 -38 -39 -41 - 39 -38 - 39 - 38 "37 - 41 -37 -40 ~ ~~ I'ime G M T )at( ~ 12 12 14 15 12 1 3 15 15 12 1 5 15 15 13 16 16 16 15 16 16 c. - 43 - 44 -42 - 43 -42 - 42 -40 "43 -40 -40 -41 -42 - 4d - 40 -41 - 39 -42 -3X - 40 I'ruvions Record I I I "32 -3 ; 5 -35 -:rear as a wllolc for some stations, partjicularly Salt Lalrc City, Utah, ant1 Boise, Idaho. 7. THE STRUCTURE OF THE COLD DOME In addition to its apparent record coldness throughout the troposphere t>hc upper vortcx possesscd some other characteristics worthy of note. Whilr perhaps not o f record depth, it was nevcrtheless all unusually deep Low. The 500-mb. height was 1,400 feet lower t'han normal near the center during much of the period from November 12 to 17. TROPOPAUSE ASSOCIATED WITH COLD DOME It also possessed an unusually low ta.opopause, near or below 24,000 feet throughout most of its history, from Alaska on November 8 tllrougll its sojourn irl t>hc North- FIGURE 9.-Contours of tropopause at 1500 GMT, November 12, 1955, labeled in thousands of feet. west from thc 12th to the 16th, retaining this same characteristic eastward to the North Atlantic. Figure 9 shows the tropopause topography a t 1500 GMT, November 12. The nadir was somcwhat south of the 500-mb. circulation center in figure 4B and seems to have been more closely associated with the vorticity maximum in extreme northeastern Oregon. A similar structure ex- isted two days later at 1500 GMT on the 14th, at which time, as can be seen from figure 5, the Low was completing a loop in its trajectory prior to moving eastward into the Plains. Although drawn as a continuous surface in figure 9, there seems to have been a definite break in the tropopause between Bismarck and St. Cloud as indicated on the cross section in figure 10. The only significant tropopause a t Bismarck and stations to the west over the cold dome had a potential temperature averaging about 310° A. This corresponds to the low Arctic tropopauses usually found over cold core Lows. At St. Cloud there were two definite tropopauses in the sounding with thc higher one at about 340" A. being the more significant. The less significant tropopause point was at about 310' A. and was therefore a diffuse continua- tion of the low Arctic tropopause. The higher point ap- parently was the same tropopause observed at 150 mb. at Green Bay, Wis., with a potential temperature of 365' A. It might therefore be inferred that the Arctic jet (to be discussed further below) from Salt Lake City to Rapid City at 300 mb. in figure 11 was located between Bismarck and St. Cloud in the region of the tropopause break, even though no wind reports arc available between these two stations. THERMAL PATTERNS AT 300 MB. As can be seen from the cross-section in figure 10 the tropope.use associated with the cold dome in the North- west was well bclow the 300-mb. level a t some stations. The thermal pattern at 300 mb. in the vicinit,y of the vortex associated with the tropospheric cold dome is therefore of some int'crest. This 300-nib. thcrmal pattern 288 100 v) a Q 200 - J -1 H - 30 0 Z - W a 3 v) cn w 600 a a 700 800 900 100 0 400 500 Tatoosh I. S e a t t l e S p o k a n e G r e a t Falls G l a k g o w B i s m a r c k St. C l o u d G r e e n B a y FIGURE 10.-East-west vertical cross-section through cold dome a t 1500 GMT, November 12, 1955. Plotting model gives temperature (" C.) to left. and potential temperature (" A,) to right. Wind force same as on 500-mb. charts. Double heavy solid lincs are frontal zones, single heavy solid line? arc tropopauscs, and thinner solid lines are isotherms labeled in " C. (fig. 11) resembled in marly respects the isothc>rms fourrtl by Kochanski [8] associated with stratospheric sinks a t 300 mb. in midwinter of 1949. As figure 11 slrows, a t~losetl -45' C. isotherm over Idaho tlelineated the warm ('or(' of stratospheric air. This isotllcrm was practically con- gruent with the 24,000-foot contour in the tropopausc topography. It was almost~cwmple~tely south of the -4FiOC. cold core a t 500 mb. and the tropospheric cold corc as rcpresent,rd by t'he 1,000-500-mb. tllickncsses i l l figurc 411. Actually the warm core a t 300 mb. seems t'o have bccn most, closely associated with the cyclonic vorticity maxi- mum rather than with the center of t,hc vortex. This warm core was surrounded by a colder ring whosc axis was along tlrc tropopause intersection with the 300-mh. sur- face. The colder ring was dclirreatctl in the west', ~rorth, and east quadrants by the -50' C. isotherm wllictl kinked iu the vicinity of t'lle t'ropopause intersection. This type of arralysis of the 300-mb. isot811ermalfield in t8he vicinity of dynamic vortices at 300-mb. is standard procedure at the Nat'ional Analysis Center l~ccausc i t seems to possess a high tlegrcc of rcdity as rcvcalctl by cross-scct'ions a r d is a rcquisit,c for tlrrcc~-tlirnc~~siorlnl consistency. lrlcrcasetl irlt'ensit,y of tlre st'ratosplleric sink was cvi- dent on later maps when tlrc upper Low was moving eastward through t'hc Great 1,akes. On the 17th the 300-mh. charts showcd a warm corc wit8h an astounding "3 5 " C. isotherm ahout, 300 miles in diurnctcr. Slani- waki, Quc., rcportctl "34" C., Buffalo, X. Y., "3 5 " C., while wintls of 200 linot,s in the associatctl jct axis were rcportcd in close proximit'y to ttrc sout'tl on ttrc equator- wartl sitlc of tlre cold ring JET AXES ASSOCIATED WITH THE COLD DOME Ttrc st,tcprst slope in ttrc tropopause topography from the Plains westward t'o the coast (fig. 9) was immediately to the rrort,lr of a jet axis a t 300 mb. (fig. 11). This jet axis appears to havc bccn close to t'he 29,200-ft'. contour of tlrc 300-mh. surfacc. This is about 1,000 ft,. lower than the jet, associated with t,hc polar front is normally found. It, is now prctt8y well accepted that t'he location of thc polar front' jet axis is nearly vertically above the 18,400-ft'. contour and t'hc -18" to "20" C . isotherm a t 500 m b . Assuming a tcrrlpcrature anywhere betwceri -40" C. a ~r t l -45" C . a t 300 mb., any combination of t l r t w 500- t111tI 30O-ml). t~cmperaturw gives 300-mb. FIGURE ll.--300-mb. analysis above cold dome at 1500 GMT, November 12, 1955. Height contours (thin solid lines) are labeled in hundreds of feet. Isotherms (dashed) are labeled in O C. Jet axes are heavy solid lines and tropopause intersection is dotted. contour value of ovcr 30,000 ft. directly above t,hc 18,400- ft,. contour a t 500 mh. Expcricncc with actual 300-mb. dat,a suggcst,s that, the normal height for t#hc polar front' jet axis at 300 mb. may be ncar 30,300 ft'. Ovtr thc southern Ylateau in the current inst8ancc, rlc~ithcr thc' w i n d data nor thc cont,our gradient,s suggested a wc.11 tlefinctl jet axis along the 30,300 ft'. cont80ur. Tllis wo~~ltl tcntl to imply t,tlat' t,he horizontal trmpwatmure field associat'ctl with the polar front was rather weak. A glance at, t,he t,hicltness lines in Figure 4A shows comparatively lit,t,lc thermal shear in the vicinit,y of the 18,400-18,600-ft. thickness zone, t,his being one criterion of frontal intensity at the Analysis Center. By t'he same token tmhc much strollger horizorlLa1 concentration of thicknesses between Salt Ilake Cit'y, Utah, and Boiw, Idaho associatcd with t ' l x hrct8ic front at the surfacc seems to have resulted in the secondary jet over the Plat,eau being the dominant one. This jet continued east'wartl along the 29,200-ft. contour (fig. 11) and apparently was associated with the main break in the tropopause bct,ween Bismarck, N. Dak., and St. Cloud, hlinn., as discussed above Alt'llough the polar front jet was weak and diffuse over the Plat,cau it apparently existed near the 30,300 ft. contour in great,er intensity over the eastern sections of the I!nit,etl St,ates due t80 the stronger baroclinic con- dition evidence in t,hr 1,000-500-mb. thickness pattern. It also appears to have been well south of the tropopause break. As can be seen from the cross-section, figure 10, however, the polar front a t 500 mb. was vertically below the tropopause break. The large horizontal separation bttwcen the 500-mh. polar front and its jet axis so much 290 MONTHLY NOVEMBER 1955 farther south over the Plains States is somewhat puzzling, but nevertheless seems to be a fact. It may have been due to an abnormal flatness of t8be frontal slope a t upper levels. FRONTS ASSOCIATED WITH THE COLD DOME As discussed above in relation to the jet axes at' 300 mb., the dominant baroclinic zone over the Plateau was north of Salt Lake City. The southern edge of t'his zone was near the 17,600-17,800-ft. thickness lines (fig. 4A). This southern edge is generally referred to as the Arctic front', i. e., the separation between CP and mP air. Figure 10 shows its vert,ical structure in cast-west cross-section. It was characterized by a pot'cntial temperature of 282-284" A. This is in line wit>h the value of about 286' A. suggested previously [9]. The CP dome reached a height of about 20,000 feet dircct>ly below the nadir of the warm Arct'ic tropopause, and was relatively shallow elsewhere. Figure 12 shows t'llat this front was the most prominent feat'ure of the air column at Great Falls, %font., at 680 mb. and was also a significant feature a t S t . Cloud a t 810 mb. Another break showed up in t'he soundings in t'he vicinity of an even lower potential temperature of about 270" A., for example at 710 mb. at Great Falls and 910 mb. a t S t . Cloud. It also appeared in all soundings in the vicinity of the cold dome and is thought t'o represcnt thc top of a much shallower strat'um of real Arctic air, urd was associat'cd with representative surface temper- atures below freezing. This discont,inuit,y is rarely car- r i d as a separate front in surface analyses since it repre- sent's merely a further intensification of the baroclinic temperature field associated with thc secondary or Arctic front,, and because i t is of such limited vertical extent. Above the cP dome as delineatcd by the Arctic front was a deep layer of mP air. Ordinarily, away from the proximity to a cold upper vortex the polar front is the vertical tcrmination of t,he mP air. In the current instance however, at' Great Falls, Mont., Edmonton, Alta., Spokane, Wash., and Boise, Idaho, t'he polar front was indistinguishable from the low Arctic tropopause; for example, in the Great Falls sounding a t 350 mb. The polar front is easily located on most soundings as a stable strat'um whose top is a t a potential temperature of about 298-302' A. or more, which is about the minimum potential temperature of tropical air in its source region. The east-west' cross-section in figure 10 shows t'he polar front very prominently at Bismarck, St. Cloud, and Green Bay, Wis., at about 300O A. wit'h the strongest vertical wind shea>r through this frontal zone. Above this frontal surface was a deep stratum of tropical air. From Glasgow wcst of Spokane there was no tropical air below the tropopause. As is evident in figure 12, the top of the mP FIGURE 12."Upper-air soundings for Great Falls, hlont., and St. Cloud, Minn., for 1500 GMT, h-ovember 12, 1955. Dashed lines indicate tops of frontal zones; slopes have no significance. FIGURE 13.-Upper-air soundings on a north-south scction through the center of the cold dome, at 1500 GMT, Sovember 12, 1955. Sloping hatched zones conncct frontal inversio~ls in adjacent soundil~gs. Slopcs have 110 significance. NOVEMBER 1955 MONTHLY WEATHER 29 1 air was near 580 mb. at St. Cloud with tropical air ahove, while at Great Falls t,ropical air was clearly absent. Figure 13 shows a series of soundings through the cold dome from Edmonton, Alta, southward to Las Vegas, Nev. The horizontal separation of the soundings on t’he pseudo- adiabatic chart has no rclat,ion to the geographical separation of the stations. The complex structure of tho air masses associated wit,h t’lle (:old dome is c~learly evident. There were four distinct air masscs in t’lle vertical. Real Arctic air was at, t>he surface at Edmont,on ant1 Spo- kane topped by an inversion surface at 270’ A. This is not carried as a separate front in t,he surface analysis. Above the real Arct>ic air was a sloping stratum of CP air topped by an inversion surface evident a t Boise, Spokane, and Edmonton with a potent,ial temperature of about, 285O A. This is the Arct’ic front in the surface analysis. Above t,his mas a sloping st,rat,urn of mP air topped 1 ))~ an inversion surface a t 300’ A. Above this stratum was tropical air with pot,ential tcmperaturc in (:xcess of 300” A. This stratum was missing a,tj Boise, Spoktmc, ant1 bG1- monton because of the low t,ropopnusc. 8. SUMMARY 1. I t appears t,l~at dynamic instabilit!y associated with an unusual confluence in the weskrlies normal to the coast, of west’ern Canada produced the rclrogression which permitt’ed a deep cold upper vortex to plunge southcast- ward from Alaska t,o Washington a(mss more than 2,000 feet of height in 48 hours. 2 . No current met’llods of prognosis succcssfrdly forecast) this event nor even hinted a t t,lre readjustment in large- scale features of the circulat’ion which occurrcd in t h o 48 hours subsequent to November 9. 3 . Record-breaking surface rlegat’ivc temperature anom- alies in t,hc northwestern Stjates were produced t)y tho persist’encc of a dynamic cyclonic vortex over t’his area. 4. Record-breaking upper-tropospheric low t’emper- aturcs maintained the deep cyclonic cold dome over this area by virtue of a continued flus of kinetic energy from the developing ridge upstream. 5. The strikingly warm core of stratospheric air a t 300 mb. seems to have been most closely associated with the cyclonic vorticity maximum rather than with the center of the vortex, and was well south of the tropo- spheric cold dome. 6. In thc south and east quadrants of the cold dome, tropical air was completely absent from the circulation in the upper troposphere within about 600 miles of the center due to the lowness of the tropopause. 7. A secondary jet associated with the baroclinicity of t’llc Arctic front appeared to be the most prominent aspect of its upper peripheral circulation. 8. Despite the apparent homogenity of its tropospheric wind and thermal circulations, soundings through the cold dome clearly indicated sloping stratification of Arctic, continental polar, and maritime polar air masses within the inner corc, with tropical air aloft no closer than about 600 miles. REFERENCES 1 . H. P. Wilson, ((A Test of a Grid Method of Forecasting t,he Motions of Lows at 500 mb. in Arctic Regions,” 6%. 2539 TEC 194, Meteorological Division, Dc- partme,rlt of Transport, Canada, Oct. 4, 1954. 2. R. Fj@rtoft, “On a Numerical Method of Integrating the Barotropic Vorticity Equation,” Tellus, vol. 4, 3 . J . Bjerknes, LLExtratropical Cyclones,” Compendium of Meteorology, American Meteorological Society, Bos- ton, 1951, pp. 577-598. 4. R. D. Elliott, “The Weather Types of North America,” Weatherwise, vol. 2 , Feb.-Dcc. 1949. 5 . W. B. Hales, “Characteristics of Cold Waves in Utah,” Utah Academy of Sciences, vol. 8 , July 1931, 6. California Institute of Technology, Meteorology Dept., Synoptic Weather Types of North America, Pasadena, Dec. 1943. 7. U. S. Weather Bureau, lLExtreme Temperatures in the Upper Air,” Technical Paper No. 3, July 1947, 100 pp. 8. Adam Kochanski, The Construction sf 300, 200, and 100-mb. Maps of the Northern Hemisphere, Dept. of Meteorology, University of California, 1950, 39 pp. 9. ,T. F. O’Connor, ‘LPractical Methods of Weather ilrlalysis and Prognosis,” NAVAER 50-1P-502, TJ. S. Office of Naval Operations, 1952, pp. 18, 40. SO. 3, A u ~. 1952, pp. 179-194. pp. 115-124. Blank page r e t a i n e d for p a g i n a t i o n