/ I NOAA Technical Report ERL 388-APCL 41 o "2te&#. J Meteorological Aspects of the Big Thompson Flash Flood of 31 July 1976 Robert A. Maddox, Fernando Caracena, Lee R. Hoxit, and Charles F. Chappell July 1977 U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration Environmental Research Laboratories THE PtIVWSVlVAMi/A STATE "'-^ LIBRA R(FS Digitized by the Internet Archive in 2013 http://archive.org/details/meteorologicalasOOmadd ^O WMOSP^, ^^^^^Tco' a o NOAA Technical Report ERL 388-APCL 41 Meteorological Aspects of the Big Thompson Flash Flood of 31 July 1976 Robert A. Maddox, Ferr^ando Caracena, Lee R. Hoxit, and Charles F. Chappell July 1977 Boulder, Colorado U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary o a. ^ Richard A. Frank, Administrator National Oceanic and Atmosphenc Administration Environmental Research Laboratories Wilmot Hess, Director For sale by the Superintendent of Documents, U.S. Government Printing OfiBce. Wasbineton, D.C. 20402 m. '■ill 4 ^^ ^\ ^^ '■^0k, '% •^•'Ipfc^ Contents Page Abstract ^ 1. Introduction i 2. Meteorological Conditions Prior to Storm Development ^ 2.1 Synoptic Scale Analyses for 1200 G\IT, 31 July 1976 4 2.2 Regional Analyses from 1800 to 2200 GMT, 31 July 1976 15 2.3 Analyses for 0000 GMT, 1 August 1976 22 3. Conditions During the Storm Period 33 3. 1 Local Area Analyses 3 J 3.2 Radar Coverage 37 4. Post-Storm Conditions ^i 4.1 Regional Scale Analyses from 0200 to 0600 GMT, 1 August 1976 ^7 4.2 Synoptic Scale Analyses for 1200 GMT, 1 August 1976 54 5. Physical Model of the Storm 60 6. Comparison of Conditions Associated With the Big Thompson and Rapid City Floods 63 7. Summary and Conclusions ^^ 8. Acknowledgments "^^ 9. References o3 111 115^" 110° 105° 100° 95° - — - -; 1 Huron \ \ • — -— ^" ! i Rapid City \ ; Lander • Casper • • — — - - — --v,- ~^ — - -- ~- — ---_ Rock 'a Springs • "Laramie. Cheyenne Sidney North Piatte • "\ Salt Lake Big Thompson-^^ F~orrCotrinr"»13'rover \ V""^ 40° - City : Canyo "EstesP;;k>^^Loveland-S«^''?9 v_ - {Grand ^ , ^A'Boulder •AkroiL_ Eagle Jp^ .Denver Junction /7v\ — '--- • •Goodland \ • I I i j N '^A'~Palmer^idge /"^ /\/\^ ""'-'" Arkansas ^— -, ATrinidad River Dodge City ~ ~ - -- — — 1 1 Front ; „ . , Range ) - - Continental j ^ " " ) n ; Divide 1 1 1 35° - Winslow • ( j Amariilo j -" i Mbuquerque i i 1 - 40° 110" 105° 100° Figure 1. Locations and topographical features in Colorado and surrounding states. IV METEOROLOGICAL ASPECTS OF THE BIG THOMPSON FLASH FLOOD OF 31 JULY 1976 Robert A. Maddox, Fernando Caracena, Lee R. Hoxit, and Charles F. Chappell Abstract. Analyses mid descriptions of meteorological conditions that pro- duced the devastating flash flood in the Big Thompson Canyon on 31 JuK 1976 are presented. The storm developed when strong low-level easterly winds to the rear of a polar front pushed a moist, conditionally unstable air mass upslope into the Front Range of the Rocky Mountains. Orographic uplift released the convective instability, and light south-southeasterly winds at steering levels allowed the storm complex to remain nearly stationary over the foothills. Minimal entrainment of relatively moist air at middle and upper levels, very low cloud bases, and a slightly tilted updnift structure contributed to a high precipitation efficiency. Meteorological conditions that produced the Big Thompson and Rapid City flash floods are compared and shown to have been very similar. A set of meteorological conditions is defined for the purpose of identifying the potential for flash floods in Front Range canyons. 1. Introduction During the evening hours of 31 July 1976, a destructive flash flood rushed through the Big Thompson Canyon west of Loveland, Colorado. U.S. Highway 34 — the primary route into Estes Park and Rocky Mountain National Park — parallels the river; the canyon had been extensively developed with businesses, motels, and campgrounds located along the scenic river bank. Larimer Count) officials estimated that between 2,500 and 3,500 people were in the canyon when the flood occurred (NOAA, 1976), and the toll was heavy: at least 139 people were killed, and property damage of about $35.5 million occurred. The storm produced very heavy rains in a narrow band along the Front Range from the Big Thompson drainage north- ward into Wyoming. Maximum amounts exceeded 12 in (305 mm) with much of the precipitation in the Big Thompson drain- age falling during the 4 h from 0030 to 0430 Greenwich Mean Time (GMT). (GMT is converted to Mountain Daylight Time by subtracting 6 h.) Significant flooding and damage occurred in Colorado on the Big Thompson and the North Fork of the Big Thompson River, within Rist Cain on west of Ft. Collins, on the lower portion of the Cache la Poudre River (Poudre Park to Ft. Collins), and on the North Fork of the Cache la Poudre River. Flash flooding was also reported over areas south and west of Wheatland, Wyoming. Fig. 1 shows Colorado and surround- ing states and identifies landmarks and towns that are referred to in the report. The Big Thompson drainage, areas iifFected 1 Golden >- i- Denver "r^ Figure 2. Big Thompson and North Fork of the Big Thompson drainages. Towns within and near the flash flood area are identified. Cumulative rainfall isohyets (black lines) for the period 31 July-2 August 1976 are shown. Terrain contours (orange lines) are in feet above mean sea level. The precipitation summary and isohyetal map were prepared by the National Weather Service Central Region Headquarters in cooperation with other Federal agencies. Figure 3. Mean July rainfall (inches) for Colorado (data from NOAA-EDS Climatological Summary for Colorado). The area shown in Fig. 2 is shaded. h\' the flash floodin of the Rapid City and Big Thompson flash flood events. This stnd\ describes the meteorological conditions that prodnced this localized, e\- tienieK hea\"\ rainfall along the Colorado Front Range dnring the evening and night of 31 Jnl\ 1976. These conditions are com- pared with those associated with the Rapid City, Sonth Dakota, flash flood of 9 June 1972, and a summary of common features is presented. Table 1. Comr>arison of maximum measured peak flows during flood events with the drainage areas involved Maximum Peak Flow (ft2 sec M 2,630 7,210 45,000 76,000 31,200 50,600 25,000 Drainage Area (mi2) Peak Flow (ft^sec-M Drainage Area (mi^) Location 0.3 1.0 6.9 22.9 60 91 692 8,767 7,210 6,522 3,319 520 556 36 Little Pinto Creek tributary, Newcastle, Utah Dark Gulch, Glen Comfort, Colo., during Big Thompson Flood Trujillo Arroyo near Hillsboro, N.Mex. Eldorado Canyon, Nev. Big Thompson River at mouth of canyon Rapid Creek at Rapid City, S.Dak. Animas River near Durango, Colo. "Many of the data are from Williams (1976). 2. Meteorological Conditions Prior to Storm Development A series of uppei-air analyses was used, in conjunction with hourly surface analyses, upper-air soundings, radar scope and satellite photographs, and stal)ilit\- charts, to study the evolution of meteorological conditions that culminated in the development of the Big Thompson s tor in. 2.1 Synoptic Scale Analyses for 1200 GMT, 31 July 1976 Detailed surface and upper-air analyses for 700, 500, and 300 mb are pre- sented in Figs. 4 through 7. Important sur- face features included a strong polar high pressure area centered over southern Canada, and a weak low pressure area lo- cated near (irand Jimction in western Col- orado. A double frontal structure at the periphery of the polar air mass (Fig. 4) stretched from the Great Lakes through Kansas and Colorado northward across central Montana. The leading front was characterized 1)\ wind shifts and a pressure trough, whereas the trailing front was marked, 1)\ a less pronounced pressure trough, lelativeK strong temperature gra- dients, and an increase in wind speed. Thermal packing was most pronounced along and to the rear of the trailing front across Kansas and Nebraska. Dewpoint temperatures weie high with values of 60°F (15.5°C) extending northwestward from Kansas into Colorado and Nebraska. A band of ver\- moist air lay just to the rear of the trailing front where dewpoints of >65°F ( 18°C) had moved into southwestern Nebraska. Earl\ morning shower and thundershower activit\ was oc- curring from Missouri to western South Dakota and also over much of the inter- mountain West. Upper-air features were dominated b\ a large, negativeK' tilted or "bent back ridge (the ridgeline sloped from NNW to SSE), which extended fiom southern Texas to western Canada. A closed high was present over the Central Plains. Moist conditions were present over the inter- mountain West and eastern slopes of the Rockies from 700 through 300 mb. A weak short-wave trough at 700 mb (Fig. 5) arced from near Salt Lake City, Utah, to El Paso, Texas. Light southeast- erly winds were present over jjouthwest- ern Nebraska, northwestern Kansas, and much of eastern Colorado. Hot, dry condi- tions at 700 mb over the southern plains indicated that deep convection would probably be suppressed south of the polar air mass. Two weak 500 mb (Fig. 6) short-wave troughs — one over Mexico and another over Arizona and New Mexico — were im- bedded in the southerly flow west of the ridgeline. A broad area of falling heights with a weak flill center over the Four Cor- ners area was associated with these troughs. A small, closed anticyclonic circu- lation over the Colorado mountains pro- duced westerly winds at 500 mb over Den- ver while winds were southerly at Grand Junction. At 300 mb (Fig. 7) winds were light southerly over Colorado while stronger (20 to 45 kt— 10 to 23 m s'^) south to south-southeasterly flow extended from Baja California northward to Utah. The LFM 500 mb vorticity analysis (Fig. 8a) defined a weak vorticity maximum over New Mexico. The 12 h forecast (Fig. 8b) indicated that the "bent back" ridge- line would move slightly eastward during the day, allowing weak south-southeasterly flow to become established over the Front Range. The two weak short waves merged into a single trough that was forecast to extend from southeastern Idaho to west Texas. Weak positive vorticity advection was forecast to occur over most of the Rocky Mountain region during the day, which would contribute to further de- stabilization of the air mass. Analyses of the Totals and Lifted Indi- ces (Fig. 9) depicted a potential for moder- ate to heavy thunderstorm activity over northern Arizona, much of Utah and Nevada, and from western Kansas into northeastern Colorado. The Totals Index is defined as 2(T85o - T5,,,,) - Ds.^,,, where T h5o is the 850 mb temperature, T500 is the 500 mb temperature, and D^so is the 850 mb dewpoint depression, all expressed in °C. Values of this index > 46 reflect favorable conditions for thunderstorm development, and values ^50 indicate a potential for moderate to heav\ storm activit\ . (See Mill- er, 1972, for a more complete discussion of this index.) The Lifted Index (L.I.) was computed b\' lifting a parcel possessing the mean thermodvnamic characteristics of the lowest 100 mb la>er adiabaticalh to 500 mb, and then subtracting its temperature from the environmental temperature at that level. The 1200 GMT Denver sounding (Fig. 10) was ver\ moist (average vapor mixing ratio for the lowest 100 mb layer was 12.0 g kg~^) below a temperature inversion at 670 mb. Winds above the inversion were light and variable, and winds in the cool air mass 5 Figure 4. Synoptic scale surface analysis for 1200 GMT, 31 July 1976. Frontal positions, pressure centers, and isobars for 2 mb intervals (1012 = 12) are shown in black. Isotherms for 5° F intervals are in orange. Dewpoint temperatures > 60°F are analyzed at 5°F Intervals with the high dewpoint region shaded orange. 6 13a 125" 120' 115" 110" 105" 100" 95" 90" 85" 80" 50'-' 45' 40° 35° 30° 25° ^306" 306' 308- 310' Oil. ^;;ri4T \ V / /J 087 \ ^>302 m\^ v\fl8 50° 45° 40" 35° 30° 25° ^^ <^^^^ 1 ^^ ^v~^ \ \ /\ /C-\. / 08i) M8 (' ^^^zy^^^''^^^^^^^ 1>-21 ^^^--^Zi^^^^^^^^rT^^^^^^ 312^ 1 31 { 07 "^■'7^10 />c^^O^N^^^ r4— t-ft p-'J^ =^^;:r Vsrt""^ ftl3f 1 ; i0oi66\| / 319=-^^r'-^^ 091,130 M y 316 ^. 08„ 149 \ . \ V 09V 149 1^ / 31^ \ I / \ / 7olo\ \\^ vV H^ \ ^,^^^^=^6 / ' \ \ ^v. / W \ ,r^-^ \ .-^orQie? / ■' \ \ ^y' N^'85 >iLi8^A V-, 5^' 9 \oi / ^jL^"^ / VC ^^6?-oi^ // / \xs 09^171 t\ / r 117-191 J >r /^-/^. .kJV>" U / N. /*" IXil68 / 09„ 189/^ f " / loA ^^ ^[V'/l '^Sr \7 184 / 9/*'^ X / \ / 318V \ ^^' ( i \ 1 , / \ \ 115° 110° 105° 100" 95° 90' Figure 5. 700 mb analysis for 1 200 GMT, 31 July 1 976. Height contours (drawn for every 20 m, 31 = 3100 m), short-wave troughs, and circulation centers are shown in black. Isotherms for 2°C intervals are in orange. Regions where T-T„< 6°C are shaded orange to indicate moist condi- tions. Figure 6. 500 mb analysis for 1 200 GMT, 31 July 1 976. Height contours (drawn for every 30 m, 570 = 5700 m), short-wave troughs, and circulation centers are shown in black. Isotherms for TC intervals are in orange. Regions where T-Tj< 6°Care shaded orange to indicate moist condi- tions. 8 130° 125° 120" 115 110" 105" 100" 95 90" 85 100° 95° Figure 7. 300 mb analysis for 1 200 GMT, 31 July 1 976. Height contours (drawn for every 1 00 m, 960 = 9600 m) and circulation centers are shown in black. Isotherms for 2°C intervals are in orange. Regions where T T^ ^ 10°C are shaded orange to indicate moist conditions. 125" 120" 115" 110" 105" 100" 95" 90' Figure 8a. National Meteorological Center (NMC) Limited Fine Mesh (LFM) vorticity analysis for 1200 GMT, 31 July 1 976. Vorticity field is shown in orange; LFM 500 mb height analysis, surface frontal positions, and 500 mb short-wave troughs are shown in black. Surface analysis and trough positions are from Figs. 4 and 6. l^elow were easterly with speeds of less than 10 kt (5 m s"^). The L.I. was -1, but the heiy;ht of the Level of Free Convection (LFC) found at 530 mb indicated consider- able lilting and/or heating would be needed to initiate deep convection. The high mois- ture content was the most umisual feature of the soundintz;. Precipitable water con- tents of().67 in (L71 cm) in the lowest 150 mb layer and 1.00 in (2.. 54 cm) in the layer from the surface to 500 mb, were approxi- mately 50^^ above Denver Jidv means (fromLott, 1976) of 0.40 in (1.02 cm) and 0.69 in (1.74 cm), respectivelv. A low over- cast at 1200 ft ACL (366 m) was reported at Denver at sounding time. The 1340 GMT Sterling, Colorado, sounding (Fig. 1 1) was taken during opera- tions of the National Hail Research Exper- iment (NHRE). A pronounced radiational 10 130" 125" 120" 115" 110" 105" 100" 95^ OCT 85_ 115' Figure 8b. NMC-LFM 12 h forecast of 500 mb heights (black) and vorticity (orange), valid 0000 GMT, 1 August 1976. inversion near the suiface was topped hv a weaker inversion at 725 nih. Winds in the cool air mass were hi^ht easterlv. The L.I. was +1 and the LFC was at 480 nih. Pre- cipitahle water contents of 0. 59 in (1.49 cm) in the lowest 150 ml) la\er and 1.04 in (2.64 cm) in the suiface-t()-5()() mh la\er were similar to the Denver values. Middle le\ el cloudiness was present in the Sterlimj; aiea at the time of the soundinti. The 1200 GMT synoptic analyses su,tz;gested that the sta,i:;e was set toi- the development of si 50 are shaded gray. Lifted Index values (orange) are shown for Colorado and portions of surrounding states. Ifnels in Idaho and Utah, and heavy rains with flash flooding were repoited in por- tions of Idaho, Oregon, Nevada, Utah, Wyoming, and Colorado. Synoptic weather conditions were quite similar to those associated with the Las Vegas flood of 3 July 1975 (Randerson, 1976). East of the Continental Divide the atmosphere was conditionally unstable, but high LFC's indicated that considerable lifting and/or heating would be needed to trigger deep convection. It was apparent that, should storms develop over the east- ern slopes or high plains, they would be very slow moving because of the light winds aloft. Unusually large amounts of precipitable water, combined with slow storm movement, suggested that thim- derstorms would have the potential to pro- duce heavy precipitation over localized areas. 12 Figure 10. Skew T/Log P plot of 1200 GMT, 31 July 1976, Denver upper-air sounding. Totals Index was estimated since Denver was above the 850 mb level. Wind speeds are in knots with a full barb = 10 kt and a half barb = 5 kt. Figure 1 1 . Skew T/Log P plot of Sterling, Colorado, upper-air sounding taken at 1 340 GMT, 31 July 1976. 13 Figure 12a. Regional surface analysis for 1800 GMT, 31 July 1976. Frontal positions, squall lines, and pressure analysis (altimeter setting shown at 0.05 in intervals) are in black. The 3 h pressure change field (at 0.5 mb intervals) is shown in gray. Dewpoints > 60°F are analyzed at 5°F intervals and shaded orange. Surface observations and pertinent remarks are plotted. 35°- NO ECHOES NO ECHOES 24 1 A470 K5 Oo \ Vo ^/ 1 SOLD "RW ■'O ° OTRW- NO ECHOES 14 no lOb 100' Figure 12b. Radar summary chart for 1735 GMT, 31 July 1976. Regional frontal positions are indicated. Figure 12c. GOES-1 photograph (visible channel, 1 km resolution at satellite sub-point) for 1800 GMT, 31 July 1976. 2.2 Regional Analyses from 1800 to 2200 GMT, 31 July 1976 Hourly surface charts were plotted and analyzed for a region that included Colorado and portions of neaihy states. All available surface reports were used, along with radar and satellite data, so that the movement and development of important weather features could he followed. In Figs. 12-17 these regional analyses are pre- sented, at 2 h inter\'als, along with coire- sponding radar summai\ charts, GOES satellite photographs, and additional Ster- ling upper-air data. At 1800 GMT the secondary' frontal surge (Fig. 12a) had moved into the north- eastern corner of Colorado, and surface pressures were rising over Nebraska and falling in western Colorado. Siuface winds were easterly and gusting from 20 to 25 kt (10-13 m s"^) to the rear of this trail- ing front. Dewpoint temperatures were >60°F (15.5°C) in the cooler air masses but were considerabK lower south of the lead- ing front. The pressure anaKsis indicates counter-gradient flow in parts of eastein Colorado and the southern Plains. This was thought to l)e a result of reduction techniques rather than an actual feature. Notice that the trailing front had just moved past Sterling. Radar and satellite- data (Figs. 12b and c) indicated that cumulus development was occurring over the mountains, especially in southern Colorado and New Mexico, and a sc|uall line had already developed over southwestern Utah. A large area of low and middle cloudiness, with some embedded 15 shower acti\ it\ , la\ ^eneralK to the rear of the trailing front oxer the plains. This cloudiness persisted through the after- noon, and the reduction in insolation helped to maintain the tliernial contrast acri,, ^ the trailing tiont. The 1920 GMT Sterling sounding iFig. 13) was taken about 40 kin to the lear of the tiailing front and exhibited impor- tant differences from the Sterling and Den- ver morning soundings. The L.I. had de- creased from + 1 to —4, while the LFC had lowered 160 mh to 640 mh. Precipitable water contents of 0.78 in (1.97 cm) in the lowest 150 mb la\'er and of 1.31 in (3.34 cm) in the surface-to-500 mb Ia\er were almost double the JuK means for Denver. Winds abo\ e a temperatiue inversion (considered to be of frontal origin) at 720 mb were west- erly at 10 to 20 kt (5-10 m s"i), indicating that Sterling was very near the upper ridge- line. Easteily low level flow had increased to 10 to 15 kt (5-8 m s^i). The Sterling sonde had sampled the air mass just behind the trailing front and found it to be condi- tionally very unstable with an unusually high moisture content. This air mass re- (juired lifting of approximateh' 140 mb to lelease its instability, and was moving westward and southwestward toward the Colorado Front Range at 15 to 20 kt (8-10 m The 2000 GMT regional analysis (Fig. 14a) showed that the leading front had re- mained nearly stationary except over southwestern Kansas and southeastern Colorado. Veering winds and decreasing dewpoints at Garden Cit\ and Dodge City, Kansas, indicated that the leading front had retreated northward in this region. The trailing front had moved southward and westward and was about to overtake the leading front across southern Kansas. l^adar and satellite data (Figs. 14b and c) showed that convective cloud and thun- derstorm activity had continued to in- ciease. The Utah stjuall line had intensified and was moving north-northeastward. Large thunderstorms had developed over northwestern New Mexico and southwest- ern Colorado. An intense thunderstorm had formed along the leading frontal bonndarx in southeastern Colorado. Small cumulus clouds dotted most of the Plains, and a line of small cumulonimbus and tow- ering cumulus clouds had begun to de- velop along the secondary frontal surge as it moxed onto the Palmer Ridge. By 2200 GMT (see Fig. 15a) the trail- nig front had overtaken and reinforced the leading front across Kansas. It had become better defined with higher dewpoints and stronger easterly winds behind it. High temperatures in eastern Colorado and Kansas had helped trigger a line of thun- derstorms along and to the rear of the front. The large storm in eastern Colorado had a top indicated b\ radar (Fig. 15b) of 56,000 ft MSL (17 km). The storms in this line appeared nearly circular on the satellite photograph (Fig. 15c), which indicated that the\ had developed in a low wind shear environment. Surface pressure had continued to fall west of the Continental Divide, and the surface low pressure area was centered north of Grand Junction. The pressure at Rawlins, Wyoming, onl\ 150 n mi (278 km) to the noitheast, was holding 10 mb higher than that at Giand Junction. Thunderstorm activity was widespread over the West with large, apparently intense storms indicated over northwestern New Mexico, south- western Colorado, and central LUah. Some thunderstorms had developed over the noith-central mountains in Colorado; how- ever, the\' remained well west of the Big Thompson drainage. A final sounding was taken at Sterling at 2202 GMT (Fig. 16) when the trailing front was located approximately 120 km to the west. During the 2 h and 40 min that had elapsed since the 1920 GMT sounding several important changes in the air mass had occiured. The mean vapor mixing ratio in the lowest 100 mb laver had decreased 1.3 g kg-' to 12.5 g kg-'. The L.I. had increased to a value of —2 (indicating more 16 stable conditions) and the LFC was now at 600 ml) (compared to the earlier 640 mh). Precipitable water amounts of 0.70 in (1.79 cm) in the lowest 150 mh layer and of 1. 14 in (2.90 cm) in the surface-to-500 mh layer had also decreased slightly from the 1920 GMT values. A plot of equivalent potential temper- ature (6'e) vs. height for the three Sterling soundings (Fig. 17) demonstrates the dif- ferences in moisture and stability charac- teristics of the air masses ahead of, im- mediately behind, and well behind the trailing front. Of most interest are the changes that occurred within the lowest 2 km. The 1340 GMT sounding showed a layer of high ^e values very near the surface with a rapid decrease to a minimum just above 2 km. Although Sterling was well within the cool air mass behind the leading front, the moist layer was actually very shallow. The 1920 GMT sounding indi- cated a dramatic increase in d^ throughout a layer extending from the surface to almost 4 km. Values of 0^, at the surface had in- creased from 343 K to 353 K, and ^e 'it 1 km AGL had increased from 334 K to 345 K. During the same period the siuface tem- peratme had increased only 5.9°C, while the temperature at 1 km AGL had actually de- creased 0.9°C. The large changes in 6^. were therefore primarily due to the arrival of the more moist air behind the trailing front. The zone characterized by high ^e and a deep moist laver was approximately 100 km in width. The 2202 GMT sounding, taken when the trailing front was about 120 km west of Sterling, showed a decrease of 4.5 K in mean 0^ foi- the lowest kilometer. Figure 13. Skew T/Log P plot of Sterling, Colorado, upper-air sounding taken at 1920 GMT, 31 July 1 976. 17 no" 30.15 Figure 14a. Regional surface analysis for 2000 GMT, 31 July 1976. Refer to legend of Fig. 12a for details. Figure 14b. Radar summary chart for 1935 GMT, 31 July 1976. 18 Figure 14c. GOES-1 photograph for 2000 GMT, 31 July 1976. 19 Figure 15a. Regional surface analysis for 2200 GMT, 31 July 1976. Refer to legend of Fig. 12a for details. 115" 110' 105 100" 95" , ' " '0 W!> o . 1 I \ o- \ o -•- ■ ] a 300 NO ECHOES \ a ^ o ^0 o RW- TRW 25 — N, TRW Qg ISOLD r)(\ h FEW ^ TRW i f- 180'^ ® ^ yi7o TRW++ ^y (K W^'K H RW- '\ L 0^ c Iv N >0 RW oy" \s 10 'X"'^ . V 40' "^"""-vy X \ N ! _.._— --K., L \ Y V 370 StVfRW 500 "" /•^ ) \ vC 560 TRW (D ( n/^^ /> ^-430 TRW+/+ y A 7 r\ / It % \ C^ y P^ \ 430 V^ / g ° o H 1 \ 4> TRWx/+ £r^ 0^ o > 400 • \ 1 35° ~ e( L / \ 1 NO ECHOES ol TRW+/+ . / / o o o\ 1 _ 40° 110 105" 100" Figure 15b. Radar summary chart for 2135 GMT, 31 July 1976. 20 ^^'L'^t Figure 15c. GOES-1 photograph for 2200 GMT, 31 July 1976. Figure 16. Skew T/Log P plot of Sterling, Colorado, upper-air sounding taken at 2202 GMT, 31 July 1976. 21 14 12 10 1920 GMT 8 - 6 - 4 - 2 - 2202 GMT 320 330 340 &e (°K) 350 360 Figure 17. Plot of equivalent potential temperature («,.) versus height for Sterling, Colorado, soundings. 2.3 Analyses for 0000 GMT, 1 August 1976 Synoptic and regional scale conditions at ()()()() GMT, 1 Aiiii;ust, were studied using a variety of charts and data. Surface analyses along with 700, 500, and 300 nib upper-air analyses are presented in Figs. 18 through 23.' By 0000 GMT the trailing front had overtaken and reinforced the leading front (see Figs. 18 and 19), except in south cen- tral Colorado where the leading front had become diffuse and was analyzed as under- going frontolysis. The low pressure center in western Colorado had deepened to 1005 nib and was located just north of Eagle. Siulace pressures had remained nearly constant along the Front Range, and had risen 1 to 3 nib over southwestern Ne- braska, northwestern Kansas, and north- eastern Colorado. A 1200 GxMT pressure difference of 2.9 mb between Grand Junc- tion and Sidney, Nebraska, had increased to 10.5 mb by evening. Twelve-hour changes in the 850 to 500 mb thickness (Fig. 20), which included 70 m increases over northwestern Colorado and 20 m de- creases over southern Nebraska, indicated a strengthening of the easterly pressure gradient through a large depth of the 22 Figure 18. Synoptic scale surface analysis for 0000 GMT, 1 August 1976. Refer to legend of Fig. 4 for details. troposphere. In response to this increasing pressure gradient low-level easterly flow had maximized in a broad hand horn cen- tral Kansas westward to northeastern Col- orado and eastern Wyoming. Surface re- ports included steady winds of 25 kt (13 m s~') at Akron, gusts to 21 kt(ll m s ') at Ft. Collins, and gusts to 24 kt (12 m s"^) at Denver. This strong, moist flow was oriented nearly normal to the Front Range. Surface obsei'vations and the radar and satellite data shown in Figs. 19h and c indicated widespread mountain thun- derstorm activity. A large squall line stretched from northern Nevada through southern Idaho to southeastern Utah. Salt Lake City reported gusts to 52 kt (27 m s~^) and % in (9.5 mm) hailstones as this squall line passed. A line of strong and visually impressive thunderstorms stretched from Colorado to southern Missouri along and behind the polar front. Even though tops on several of these storms grew to over 50,000 ft MSL (15.3 km), their intensities remained below severe limits although lo- cally heavy rains were likely produced. A large area of thunderstorms had de- veloped over east central Wyoming to the 23 Figure 19a. Regional surface analysis for 0000 GMT, 1 August 1976. Refer to legend of Fig. 12a for details. Q> o O "O O NO ECHOES 1 1 105 Figure 19b. Radar summary chart for 2335 GMT, 31 July 1976. 24 '■^i^^^. Figure 19c. GOES-1 photograph for 0000 GMT, 1 August 1976. Figure 20. Changes in 850-500 mb thickness (in meters) from 1 200 GMT on July 31 to 0000 GMT on August 1. Regions where thicitness decreased are shaded orange. 25 Figure 21. 700 mb analysis for 0000 GMT, 1 August 1976. Refer to legend of Fig. 5 for details. rear of the front, and the first cells in the area of Big Thompson were just l)eginning to develop over the mountains southwest of Ft. Collins. The smaller cumulus clouds over northern Kansas, southwestern Ne- braska, and northeastern Colorado had dis- sipated as boundary layer cooling began. Upper-air analyses indicated that the large, negatively tilted ridge had intensi- fied and developed northward duiing the day. The cutoff high within the ridge had drifted slightly southward over the Central Plains, and the ridgeline extended fiom western Kansas to cential Montana. Warm air aloft had suppressed development of deep convection over the plains south of the surface front. Winds aloft over eastern Colorado were east to south-southeasterlv at onl\ 10 to 25 kt (5 to 13 m s"'). Wanu 700 mb (Fig. 21) temperatures (^16°C) over western Colorado reflected strong summer afternoon heating. The 500 mb (Fig. 22) short wave extended in an arc from central Nevada to New Mexico as it continued to move north-northeastward. The position of this short wave suggests that falling pressures in western Colorado were probably caused by a combination of dynamical effects and the iifternoon heat- ing. High moisture content was evident over the western United States ahead of the trough. An unusually strong (for the 26 Figure 22. 500 mb analysis for 0000 GMT, 1 August 1976. Refer to legend of Fig. 6 for details. season) southerly wind band with speeds of 40 to 60 kt (21 to 31 m s'l) at 300 mh (Fig. 23) extended from Baja Cahfornia north- ward across Utah. The 0030 GMT infrared satellite photograph (Fig. 24) showed the large areal extent of clouds and convection over the western United States. An influx of mois- ture from the Inter-Tropical Convergence Zone was indicated by the band of clouds streaming northward out of the tropics and into the western United States. Vorticity and stability analyses along with sounding data are presented in Figs. 25 through 27. The 500 mb short-wave trough was clearly depicted on the 0000 GMT vorticity analysis (Fig. 25) with posi- tive vorticity advection indicated over a broad area from northwest Texas to Idaho and northern Nevada. Comparison with Fig. 8b shows that the position oi the short wave was accurately forecast although it was more intense than predicted. The Big Thompson storm was developing in a legion of minimum vorticity and weak positive vorticity advection. The active squall line in Nevada and Utah was just ahead of the northward moving short wave. The 0000 GMT stability analyses (Fig. 26) showed that the centers of instability, which were over Arizona and western Kan- sas in the morning, had merged to form one 130" 125"' 120" 115" 110'^' 105" 100" 95" 90" 85" 80° Figure 23. 300 mb analysis for 0000 GMT, 1 August 1976. Refer to legend of Fig. 7 for details. large area of unstable conditions that stretched from New Mexico to Montana. Thunderstorms were occurrinii; thioughout this region and Totals of 50 to 52, along with LT. values of -2 to -4, indicated the po- tential for moderate to heavy storm activity The 0000 GMT Denver sounding (Fig. 27) showed that the air mass structure had modified significantly during the day. The tempeiatuie inveision had lifted 80 mh to the 590 mh level. Stiong diui-nal heating was piohahly lesponsihle loi" this lifting since a dry adiahatic lapse late existed below the inveision. The low cloud laser had dissipated during the morning, and the mean vapor mixing ratio in the lowest 100 mb la\ei" had decreased from 12.0 to 9.5 g kg~^. Easterly flow had increased slightK and the LFC had lowered to 620 mb. The L.I. was —2 and thunderstorm actix it\ was developing in the vicinit\ . Precipitable water for the surface -to- 500 mb la\ er was 2.47 cm, little diffeient from the 2.. 54 cm value at 1200 CiMT. Strong iifternoon heat- ing and mixing had redistributed the mois- tuie thiough a deeper layer. The Denver 28 Figure 24. GOES-1 infrared (4 km resolution) satellite photograph for 0030 GMT, 1 August 1976. 29 130 125" 120" 115" 110" 105" 100" 95° 90" 85 Figure 25. NMC-LFM vorticity analysis (orange) for 0000 GMT, 1 August 1976. Synoptic surface analysis and 500 mb height analysis are in black. Position of the important 500 mb short wave is also shown in black. sonde was released at appioximately 2315 CMT, prior to passage of the trailinii; front. Tahle 2 compares sonndintz;s taken durinii; the day of 31 Jnl\ 1976. Note the considei"al)le difference between the 1920 C;MT Sterhnu soundint^ and the ()()()() C;MT Denver s()nn(hn^ollins ^, •Windsor J'/ y ♦•^fyi ' €' s/ /' E.res'^^ \ koveland "S^'^'^v / Grand / La^e I'MeekiyjN-j eerthoub ' 1 '' Granby ^i.I^^ able / Nedeflatiit (^jKPs '^ f^^^*""^ •Brighton C, /y^ !¥- 5000' / •Enip/j d 10000' 10000' 10 20 30 \Englewood^ 'O ^^ larkspur Kiowa • Elbert 7000' 750^ ^y Glendevev \ \ ^^ ,_^ 10000 fLoveland '^''"'"' 5000' righton Stapleton Airport _ ,, _ n *--^ Watkins .Bennel • Denver " ^^7.^^^^^ Castle Rock 10 20 30 Elizabeth Kiowa --,__ "Elbert 7000' Figure 31. Continued, (c) 0025 GMT, 1 August 1976. (d) 0100 GMT, 1 August 1976. in conjunction with surface ohsei'vations and the Table Mountain wind data, a soundini:; was constnicted for Loveland, Colorado, valid at OCXX) GMT. This sounding, shown in Fig. 32, is an estimate of the Big Thompson storm environment. The L.T was —6, and the mean vapor mixing ratio below the temperature inversion was 14.8 g kg~^. The LCL was at 730 mb (== L 1 km AGL), which agrees with observed low cloud heights at Ft. Collins. An additional 80 mb of lift was necessary to bring this air to its LFC. 3.2 Radar Coverage The NHRE 10 cm radar at G rover scanned the storm complex along the northern Colorado foothills until a few mi- nutes iifter 0100 GMT. During this period the storm's intensity peaked about 0045 GMT and then weakened temporarily. Re- flectivitv data from Limon and G rover radars were compared during the peak period. Limon (0° elevation angle) and Grover (1.9° elevation angle) radar echoes at 0045 GMT are superimposed on a map of the Big Thompson area in Fig. 33. Both radars scanned a section through the storms at elevations between 15,000 ft MSL (4.6 km) and 20,000 ft MSL (6.1 km). Limon radar showed only a VIP level 3 contour, which corresponds to reflectivities between 41 and 46 dBZ. However, Grover radar ob- served a level 5 (55-65 dBZ) with a meas- ured peak reflectivity of 64.6 dBZ. The Grover data contained more detail than the Limon data. This is partially explained by the narrower beam width of the Grover radar (1° conical beam compared with Li- mon s 2° conical beam) and Grover s loca- tion closer to the storm area. Fig. 33 suggests that Limon radar underestimated the true intensity in the core of the storms b\ about 15 dBZ. 37 Figure 32. Skew T/Log P plot of upper-air sounding constructed for Loveland, Colorado, at 0000 GMT, 1 August 1976. LCL and LFC levels and moist adiabat are shown for a lifted parcel with mean thermodynamic characteristics of lowest 100 mb layer. Alth()uii;h it underestimated rainfall intensities, Linion ladar data still provide an excellent description of the temporal and spatial vaiiation of precipitation over the Biti Thompson drainage. Fig. 34 shows the radar echo locations and intensities for the period 2300-0400 GMT (20 min inter- vals prior to 0100 CiMT and 10 min inter- vals thereafter). The first cells that developed around 0000-0030 GMT were several miles east of the maximum rainfall zone. These storms reached their peak intensities near 0045 GMT. Durinir the period 0100-0130 GMT the echoes in the foothills tended to merge, weaken slightly, and shift westward. The most intense rainfall was indicated south- west of the town of Drake by 0130 GMT. From about 0130 to shortly iifter 0300 GMT the "cloudburst phase of the storm occur- I'ed in the Big Thompson drainage around C;len Gomfort. After 0300 (;MT the area of intense rain moved slightly northwest over the tributaries of the North Fork of the Big Thompson. The rainfall intensity de- creased after 0400 GMT, and most of the precipitation shifted northward into the foothills west of Ft. Gollins. The rainfall amounts shown in Fig. 2 at the beginning of this report were for a 36 to 48 h period. Radar data and eyewitness accounts suggest that the intense rain in the Big Thompson drainage was over by 0400-0430 C;MT; however, showers per- sisted throughout the night. Rain showers were present again the following iifternoon and evening and were locally heavy, espe- cially west of Ft. Gollins. Therefore, the 36 to 48 h totals are an overestimation of the precipitation that contributed to the flash flooding. It was shown above that Limon radar echoes of the Big Thompson storm under- stated the true intensity of the storm; how- ever, if the reflectivity were off by a con- stant factor, it should be possible to correct 38 ■^^ J^ f V \ '^^ ^ -f Red Feather Lakes •Rustic -o"^ ! * V - ' • Golden " ■< Denver ^r„ Figure 33. Relief map showing 0045 GMT radar echoes from NWS radar at Limon, Colorado (orange), and from NHRE radar at Grover, Colorado (black). Limon contours are for VIP levels 1 , 2, and 3. Grover contours are at 10 dBZ intervals beginning at 15 dBZ with regions of reflectivity ^ 55 dBZ shaded orange. 39 2301 GMT Glendevey 1 0000'. 5000' 7 Glen ,' Ft. Collins / Haven r' Drake •C^/ '\ Estes Comfort \ \ / Grand V park '' / Berthoud / Lake ^, / /' ' \ Granby I Meeker > \ / ^'^''^ \ \ Table ,'' '\Wa-d /' I * *"'■ ' I * ,' Dr... Boulder Nederland righton t^ \ \ Jetfco /'~\ , . • ( \ ^.' Central / "c'j'l'/ Slapleton V_y /' 'Cltv / , Airport /O ' 'Empire i •tGolden * WatMs _^ 'v \ 'Denver/ ^1 / \ , \ ^ ^-* \ "■ EnglewooQ \ Evergreen, \ f^,,^^^y ,---'. Parker r^ 1 D ,— V Sedalia ^.^Bail.ey y ^_^ ^^v..'- (;ast^ 10000' lOOOO' ~\^ '"''/' '' \^~ ^ Kiowa ~~") /Deckers \ ^^^^ 6000' • Elbert .Rock Qj-"^'^ ^^ Larkspur 7500' 7000' 1 ^ ' 7500' ^,'^,.6000' 2320 GMT Glendevey \ / A V ! 10000' \ l^ ! J 1^ ] 5000' \ , f^'' ' '\ ,- "''' \.S' V, _. , \ 3 F't. Collins §/ 1 Glen 1 / 1 ^.f' /»=rv' Drake 1 •"""''" / Grand »Gt^ ; tLoveland •'^"^"=* ; Estes (Com^ifrt ) Park ^"^ / Berthoud / Lake "; / ,'' 1 Meeker i \ / . / ^v y^'^ \ .Table ,/ / 1 Granby \Ward ; ^'- L,-. /-X ^ \ ] ' { Boulder ^X,/ ^- .^ V^ Nederland \ ', •Brighton ~-~^ ^ ( ' \ \ Jefko O 500QP ; J Central / "pfa's' ^tapleton ^ (_5 •City / ' AirpM-~-^y X Q. / 'Empire ■' •iGolden */ Wattos .fi^i ^ ' n ''-'' 'Vi' ) Englewood V > — — -\ / \ Evergreen,^ N^ /^^^^^ ^ ^l J \ \ ,---';p;,ker ^N^^^^*^ / „. / ^ Sedalia ^^^ — \ / ', -^. Bailey P™' 0-._'/ r.,,\/ ) 10000' \^ --' /' ; t^ ^ '"^ Kiowa \ 10000' ~^ /Oeckers \ ^V^ 6000' , /V i / /'' \ Urkspu^---^;"- 10 20 30 ^vj ( / ■> ^ kn 7500' 7000' 2350 GMT Glendevey 10000' . Rustic ''Z <§"/ ) Glen 1 \^^ 1 '^ ' 'taven/ ri •blen 4^ /"T"' Drake i/ $/ ( Es'es Comfort 1 N / Grand \ Park '' /' Berthoud / I akp ~~^. / / • > Winrlsor koveland •'^™"^* Lake I Meeker i \ / ^^''^ \ \ .Table / Granby '^^^^^^ ] i ' Mt. ,^^ /' * ( Boulder Nederland \ \ I ' \ \ Jeffco / •Empire o X \ -Wyl iXLy ^ ~- Castle CI h ,r, ^ \ V' />< ,-;, .Rock E';"""" lOObo' 10000' \^ >-' /' ; i '- ^ Kiowa \ I I I I 10 20 30 km ^i / Deckers \ ^^ / ' • , \ ^^^ "■-- / / / ^, Larkspur • Elbert 7000' 0000 GMT Glendevey 10000-, \.S' "-- -^ r 'v ' F*t. Collins s/ ,' Glen 1 / 1 sf/ ,,''"a/Drake 1 •"'"^"' J"/ f' . .'Glen* / koveland •°"^"'^^ -^/ \ Es*es Comfort \ \ '' Grand 'v pgrk '' /' Berthoud / Lake "^^ / ,- • \ Granby I Meeker i \ Park \ .,Ward '■ 1^-^'- V -yO ''"' I ' ( Boulder ^ ^^. /' Nederland \ \ 10000' 10000' \ "i / Oeckers \ ^v^ . ""n, / / / ^, Larkspur ""-- Kiowa Elbert ^ Larkspu 7500' 7000 Figure 34. Time series plots of Limon radar echoes covering period 2301 GMT, 31 July 1976 through 0400 GMT, 1 August 1976. Time interval is approximately 20 min prior to 0100 GMT and approximately 10 min after 01 00 GMT. Level 1 and 2 returns are shaded gray, and level 3 returns are shaded orange. 40 0020 GMT 0110 GMT Ft. Collins \ •Windsor ♦ Loveland '^'"""'^ ; Benhoud GfanlJv fjghton Jeffco apleton ( y . 5000- Central '' '^p'lM^'' Staplel •C.tv ,_'^"" Anpo Golden * Watkins •Bennel •Denver Evc.gN Englewood Q. 0\ \ /' o „ , — J ^, Sedalia 10000' 10000 Parker \ \ / W , — y ^ Sedalia Bailey P:ne O^,^^., ^^^^,^ / / \ '''^\^^ • Rd*^*^ ^-^ / / \ ^^^ — ^ ) /Deckers \ ^>,^ ^•^ / / / ^ Larkspur Elizabeth \ Kiowa \ 6000' _, 'Elbert 10 20 30 km ''■ 7500' 7000' Figure 34. Continued. 41 Glendevey 10000' . 0120 GMT Stapieton Airport •\{]olc)en • Walkins •Bennet •Denver * 1 ^ ' 'T X>'''' 0132 GMT Gleiidevev \ / /'' V ' 10000' Rustic ^/ \ \ Lj J ; 1 5000' \ V-^f' V ,^, \ ( 3 Ft. Collins S/ /^ > Glen 1 \ ; i / Haven; n\i,„ \ 'Windsor .''' @en V koveland •'^''='=v / Grand / Lake ^•^ PaNi 'f\ A Befthoud /^»"<»^ Table / Granby ^ W/' ;*^'" ^~''-X /'^ /] r^ { Boulder \_y -^ ^^Ncderland \ ', •Brighton "^^^^ ' \ \ Jelko 5000' ( _y Ce/tr/T/ 'IX* Stapieton \./" i/Citv\ / / Airport \»Einpifsr \i "Golden * Watkms .Bennet y \ ^ r> \ •Denver C— Y \/ ^i-' > \ / ^ ^-*"'' I ^ Englewood \J \ E«e'9feen,^ \ A|,p„„ ^ ^ y \ \ ,---'iPacker ^x^ / y ',\, Badev Pine' .\^ ., \ ^V/ /,.- ,-.\ .S El-'feth \ lOOQO' 10000' \^ ^-' /' / 1^ ~ ^ Kio'wa \ / ~") /Deckers \ ^^^^ 6000' ^X ,''' / /' ''\ L*arkspur~~~--*J^^''" 10 20 30 ~-vJ 1 / \ ' km 7500' 7000' Ft. Collins , \ •Windsor ke I (•Loveland •'^™"='' etthoucl 0141 GMT Brighton Jnt/al / "^',' Stapieton •C.tl / / '^"<>°" •iGolden ' Watkms .Bennet •Denver / r "-.^ Barley P™' vO 10000' 10000' "\^ ' ' ^v "■") ^/'Deckers \ ,' / ) ''^'\ 'Rock J- "x ! ; /' 10 20 30 ^O r; / km '' ^^ Larkspur 7500' Elizabeth \ Kiouva \ 6000' • Elbert 1 >- / "?"■ /,.''«°°°' 0150 GMT Glendevey \ z'' ''' k \ ' 10000' /, 1 \ Rustic ^' 1 \ ' ) • ; 1 5000' \ , i£/ ~^ \ \ , -'-"' ' A. ,, \ / ) Ft. Collins \ ^ — ; Glen 1 I ; i .♦7 ''' 1 .*^ () Cleveland •°'''''"' ; Ei'es tfwory / Grand / Lajl«ll6*«I\d '/~+v •Brighton "^-^^ ( ( * ) A \\ Jeffco foOO' l.-' /CenL / T[\^ Stapieton •CitA 1 /, Arrport /^Empir Vi/ •iGoldcn • Watkms .Bennet ^■^ \ \ •Denver ' — - - . . ,^ Vyi ^ ^V 1 \ Englewood ^rvO, ^""'S'**"; \ Airport i Z^ \ \ ,— -IParker N / -. ' n , ' \ Sedaha \ ', -^. Bailey P'ne' x\._«, „ \ ^^x //-> ,'V^ •Rock Elizabeth \ 10000' 10000' \^ --' / 1 r~~~^^ Kiowa \ /' \ "~") /Deckers \ ^^v^ 6000' 1 1 1 1 V '' / /'' '~\ L'arkspur^~-~--*Jlbert 10 20 30 ^v.y f / \ km 7500' 7000' Figure 34. Continued. 42 0200 GMT Ft. Collins \ •Wiiidsof (•Loveland '^"''"'^ ' Berthoiid I7eiierlarfl| \ I\ 'Brighton "^-^^ 'J Jcffco 501 V' ' ' -* - ^ ™^ Airport iiGolden " Watkins .Bennet \ •Denver * ^ >- ( ' Enqlewood Evergreen, \ ^j^p^^, 0000' 10000' BailBv P™' V- *' c ^) / Deckers \ 10 20 30 ^^ Larkspur 7500' Elizabeth \ \ Kiowa y 6C -^__ 'Elbert 7000- 0210 GMT Glendevev \ •'^ ^^ I' 1 10000' „ y' \ r. Rustic /,' { \ > '-■ • ^^ \ I ^. — -' >> DGIeni \ ' Ft. Collins # rSi.Drak\ • Windsor O ^Loveland ■JV i / EstA/ComM' / Grand ' A ParL, / / / Berthoiid Lake h 'A f / * \ 'Greeley Gidiiby i^Meejii J V \ ¥'^ _l I Table , N^ard7\i'"'' ' „ ( BJiulder |lederlaf\ v, \ \ Jeftco • Brighton entral i "ll'^/ Stapleton City / , Auport Golden • Watkins .Bennet •Denver . , Englewood J Eyctgreem^ \ Airport ^ .^ ,---'iParker -_/' ^ Sedalia Bailey P™' ^-._*/ r , • I ^. uastie ci,„i • Rock 10000' 10000' I I I I 10 20 30 km "") /'Deckers \ ^^ Larkspur 7500' Kiowa • Elbert 7000' 1 ^ 1 7500' 1 .6000' 0220 GMT Glendevey \ . / .^ • V [ 10000'. Rustic /-/' • 1 \ \ 5000- \ i^\ " \\ -'-" '' 'v J '' i ) Ft. Collins oV ~/") Glen i / / Haven/ Drdke \ •Windsor ■'' ^ (Z|l n/i tLoveland • Greet y ■y/ ; \ &tesV9*7 tort i / Grand • T-^'^ih '' / ^' Berth oud / Lake / I'MsKerr S. 'i Table /' Granby > VWard ; XT V --^ /'^ c y^i ^-4- *tiulder ^^ / ■~ — -^^ ■^ li r"^^ Ulderl^td ' 1 • Brigh ton ^^ / ^^ 1 V \ Jettco \ \ • 5000' / y /^entral 1 Rocky / Flats Stapleton \/ ' /•Citv Airport Vi Farfiirp .' •\Gok 6" „ ' Walk ns •Ben net • Denver O/ ^"■^^v' /^ — > \ Englewood c ^v^ Evergreen 1 '^ Airport a»^ ) ^ ■ \ \ ^ , 'JParker X ^^ ~ — " \ \ / .' \ Sedaha \ \ "^^ Bailey Pine" \'^— *' Castle '-^\ • Rock Ell abeth ' 10000' 10000' \ " ^ Kiowa \ / \ ) /Deckers \ ^s. 6000' ' • , 1 1 1 ^X / 1 / \ • ^^ Larkspur \ ""^^ _ •Elbert 10 20 30 ^O km 7500' 7000- 0232 GMT 5000' /"'n'Pfrake cLoveland •'^™"=v V )/ Berthoud righton / •Empire , ■ Rocky •City Stapleton Airport Golden • Watkins .Bennet ', •Denver ' 1 ^ ^'-•'^ f ^ Englewood Evergreen,^ \ Airport / \ ""-^^ Bail^ey P;ne' 10000' 10000' \ '-'' I I 1 I ^^ I ; / 10 20 30 ^-.J i / km '^ Castle • Rock ^^ Larkspur 7500' Elizabeth ^ Kiowa \ 6000' -__ -Elbert Figure 34. Continued. 43 Glendevey • 10000- Rust.c ''/ '» L' * _ ii^^ \ /'.''■I."' \ ;?V '. . r\' ^/ J /" ^/ /,' -?/ r S' \ Esfes / Grand / Lake ^J Park* A^ 0240 GMT -' Ft. Collins f , \ "Windsor Take I tLoveland •'^™"'v /' Berthoud Granbv .~^v Ward ,' , ,^,- ^ ^ Boulder \y "^ ^^ Nederj/nd \ \ "Brighton ^^--^^ (/ ' V \ Jeffco 501 ' Central / '^fl^^ Stapleton •City / ' Airport / "Empire ' "iGolden * Watkins .Bennet > \ •Denver ■", / \ • 1 ^ ^-•' \ Enalewood \ Evergreen i^ \ Airport ,^ Bail^ey P™' 10000' 10000' "; /' \ Castle • Rock Deckers ■■^ Larkspur 7500- Elizabeth \ Kiowa \ 6000' "Elbert 1 ^ J 7500- ^.^-^^,6000' Q250 GMT Glendevey \ / / V 1 10000' \ 1^ ; • 1 V 1 5000' \ 'M A-^ I .'"" \ / .._•/ (, , \ ^n/ //-) ,-\ "Rock E':-^='^ ^, 10000' 10000' \^ -' /' ; \ ^--^ Kio'wa \ ~") /Deckers \ ^V^ 6000' 1 1 1 ^N { i / \ Urkspu^~---VE,'^»" 10 20 30 ^vj ( /■ \ ^ km 7500' 7000' 1 ^ 1 7500' ^^'"^-6000"^ 1 / ^-^ 0300 GMT Glendevey \ /^ '" • y ( 10000'. •- » 11 5000' \ '^^^ ^ /^ .--1---""' \.S't-- - '\ ■) "-' Ft- Collins / ~ ;Glen 1/ , 1 #f I /' ( (*A" < »Loveland '^""'^''< / Grand / LaJ>c ^v raHr / / Berthoud / (WeekerT \ / / V" ''> \ Table / Granby / '/Ward /' ; *"• 'v_,-. \ ' n A ( Boulder \ / -^ 1 \\ ij '• '• "^-v c jWljfleriand \ ', "Brighton ""---,^ —^ { ' \ \ Jetko 5000- y Central ' '^r,l^^ Stapleton "City / ' Airport / "Empire ■' "iGolden ' Watk ' \ \ •Denver ns •Bennet c^ 1 , ^-^^ ! y Englewood \ Evergreen,^ \ Airport ^ J \ \ ^ JParker X / . „. i \ Scda'lia \ ^^ T -^^ Bailey P™' ^>._'/ ^ , \ ^% / /-> ,'<\ "Rock Elizabeth \^ 10000' 10000' \^ >-' /' / I '- ^ Kiowa \ / ^v ""■) /Deckers \ ^^.^ 6000' , ;^^. / / / ^, Larkspur ~~-- ^•Elbeit 10 20 30 ^.J ! / \ '•^ 7500' kn 7000' 1 ^ 1 7500' y -6000' 0310 GMT 1 1 / /' Glendeuev '^ n! 10000'. Rustic //' ■■, 5000' <'p, Collins '"" c?/ / Gten 1 / , ,//^"i ■Drake \ "Windsor / L/,»GI« ( Estes Cm! )' \ Ifort 1 tLoveland •Greeley / Grand V Park V, ; Berth Dud / 1-aj.e ' Park \ \ Table Granby N ) ^< '".'"' /' 1 'Mt. \^^'-^ ''^ / / V Boulder \y ( J Nederland ^ \ "Brighton ^^^ \ \ Jetfco 5000' ^■^' Central 1 Rocky / Flats Stapleton /" •Citv Airport '' •Empire \ "iGold -" n * Watkms .Bennet 1 ^-- • Denver * ■ ' ^. /^ — 1 \ Englewood v^ Evergreeni \ Airport .— — ^^' ^-— -i'Parker \ / J \ \ ^-^^ Bail^ey Pine' "^~"--' Castle -^.jV "Rock Elizabeth ^ 10000' 10000' \ '"' /' / Kiowa \ / \ ) /Decker 6000' 1 1 1 IN.) 1 / ^^ Larkspur ~~,__ "Elbert 10 20 30 VJ km 7500' 7000' Figure 34. Continued. 44 0320 GMT Ft. Collins ', •Windsor koveland •'^™''* i'/ J Berthoud Granby \ / Ward /' ■ * "'' L,— ''\ r~> /' i' Boulder \ / "~ ^ \y / \ !• " ^ Nederland \ ', 'Brighton ^^.,^_^ ( * \ \ Jctfco 501 ) Central / "p'i'l','' Stapleton /' •City / ' Airport / •Empire ' •iGolden „ * Watkiris .Bennel ■■' \ \ •Denver \ ^ ''.*■'' I \ Enqlewood Eyergreen,^ \ f^„^^n ^ ^ ,y \ ^^ ^' iParker n^ Sedalia / \ "---^ Bail^ey Pi",e lOobo' 10000' \^ '"''/' ''' ^' ^~ - Castle • Rock Elizabeth \ Kiowa \ ~^ /Deckers \ N^ 6000' / / /' \ I Vi,„,,r"~~--^ 'Elbert I I I I ^"^s I / / 10 20 30 ^^; I / km ^^ Larkspur 7500' 7000' ; ^ 1 7500' y ,/^ y^^ -6000' 0330 GMT Glendevey \ / y * V 'y 10000'. Rustjc // s 5000' \ , '^^' ^^\ l\ , - "" V/ 1 \ J ^^ Collins ' ; Glen r / Haven; p, ake \ •Windsor ,^/ V' lo'Glen/ ), koveland ''^^'^^ / Grand / Lake y Park \^ rt 1 / Berth ' Table Granby \Ward /' 1 'Mt. V ^ /'N ' * ( B oulder \-/ -^.^ ^ Nederland \ \ • Brighton ^~ c \ Jeffco 5000' ^ ^^' Central Rocky Flats Stapleton • City / Airport J) '' •Empire ' •\Gol( en • <3 Walk ns .B ennet ^-' , \ • Denver \ /' --, \ Englewood 1 ^ Everg reeni \ Airport ^ "^v ( — \ J \ ^ , — 'JParker ( 1 \j / J \ Sedalia \ \ ^^-.^ Bail^ey Pi ne' ~v~"— *' Castle CI. \ '^^\ "Rock , 10000' 10000' \ > " ^ Kiowa ^^ \ ^) /Occke S \ ""v^ 6000' N / / / ^v • ^^^^ • Elbert 1 1 ■-. 1 / ^^ Larkspur 10 20 30 Vj I km 7500' 7000' 0340 GMT Glendevey \ ^ y • V sA 100OO'_.^ Rustic ^/ \ \ 1 ^A 1 ( >7 / -:'Glen ©y , Ft. Collins H Ei^*4l' Drake • Windsor nfort 1 tLoveland •Greeley / Grand/ V j^^ / Berthoud / LakeV. ^Jf^ 1 / \ Gtanby ,jf Meeker i ^ / V^"< \ \ Table / \Ward /' I * ""'■ V_,-.^ /'N,^ / '\ Boulder \ / "~ ^^ Nederland \ \ •Brighton ""-^_^ { ' \ \ Jeffco m y Central / 'lul'' Stapleton •City / , Airport / •Empire i •iGolden * Watkins .Bennet \ •Denver 1 ><^'-| \ V Evergreen 1 ^^ Englewood Airport ;— ,,:^.,,., W/ \ ^-^. Bailey Pine' O-.^-, j.^^,,^ 7 \ ^^^ ,' /'\ ,'^^ •Rock 10000' 10000' \ •■-' 1 / \ "— ^^ Sedalia 1 I I I 10 20 30 ~") /Deckers \ "-^^ / / / ^^ Larkspur Elizabeth \ Kiowa \ 601 -___ •Elbert 7000' 1 ^ J 7500' ^,^-^6000'' 0400 GMT Glendevey V-<#'V,-, \ J; F*t. Collins g/ } Glen 1 y, 1 0/ /Haven,' »/',,„ \ •Windsor jI''' X/'' ^ie^n'A tLoveland ''^'"''^ / GraiN ,'<\ .Rock Elizabeth \^ 10000' \^ '-' J 1 >^^- ^ Kio'wa \ / ~^ /Deckers \ ^V 6000' 1 1 1 ^N /' / /'' ^^\ L^rkspur---A^^- 10 kn 20 30 ^.j 1 / \ ^ 7500' 7000' Figure 34. Continued. 45 7 - js 6 .2 5 E 4 - 3 - 2 - 1 - 1 I I r~\ r~i I 1 r Glen Comfort I I I I I I I I I I I I I I I I 0100 0200 0300 0400 Time (GMT) 0500 0600 Figure 35. Accumulated rainfall (inches) curves for Glen Comfort and Glen Haven, Colorado, from 0040 GMT to 0630 GMT, 1 August 1976. Curves were developed using Limon reflectivity data. for this error and use the following radar ec}uation to calculate more realistic pre- cipitation rates: Z = 200Pi 6^ :i; wheie Z is the reflectivity factor (nini^ m"'*), and P is the precipitation rate (mm h^^). At Glen Comfort, a total of about 12 in (305 mm) of rain fell in the 48 h period. It is not known for certain how much of this fell from the Big Thompson storm complex, hut the amount was probably 10 in or more. For the purpose of the following calcula- tions it was assumed that the total rainfall accumulated during the period 0000 to 0600 GMT on 1 August at Glen Comfort was 10 in (2.54 mm). It was further assumed that reflectivities within the central half of a VIP level 3 area were 2 dBZ higher than the level 3 threshold value. Level 4 interior contovus fleetingly occurred in the storm area indicating that the assumed 43 dBZ for the core areas may have been conservative. An accumulated rainfall total for Glen Comfort was computed using the Limon reflectivities, equation 1, and the above assumptions. The resulting precipitation amount was onlv 0.98 in (24.9 mm) for the period 0000 to 0600 GMT. For this particu- lar event an intensity correction of +16.1 dBZ had to be applied to the Limon reflec- tivity data to yield the assumed rainfall of 10 in*. This adjustment agrees quite well with the approximate 15 dBZ difference between Limon and Grover reflectivities noted at 0045 GMT within the echo cores. With the 16. 1 dBZ correction applied to *The radar equation currently being used by NWS for convective precipitation is Z = 55P^^. Use of this relation yields an accumulated rainfall of 2.19 in (55.6 mm) at Glen Comfort during the period ()()()() to 0600 GMT. The required Limon reflectivity correction decreases to -I- 10.6 dBZ. 46 Linion data, VIP levels 2, 3, and 3+ (level 3 threshold +2 dBZ) represent rainfall rates of 1.21, 5.91, and 7.89 in h"! (31, 150, 200 mm h^), respectiveK'. These modified rainfall rates were used to construct the accumulated precipi- tation curves for Glen Comfort and Glen Haven shown in Fig. 35. The curve for Glen Comfort shows that about 7.5 in (191 mm) of rain prohahly fell in the period 0130 to 0245 GMT. The Glen Haven curve indi- cates that heavy rains began somewhat later in that area and continued for a longer period of time. These curves are consistent with the Grozier et al. (1976) report that the flood crest on the Big Thompson River just downstream from Drake occinred be- fore the ciest from the North Fork arrived at Drake. 4. Post-Storm Conditions 4.1 Regional Scale Analyses from 0200 to 0600 GMT, 1 August 1976 Regional surface analyses and corre- sponding radar summaries and infrared satellite photographs are shown for 2 h in- tervals for the period 0200 to 0600 GMT in Figs. 36 through 38. At 0200 GMT the front had pushed into the Front Range, the Big Thompson storm had developed (note that a squall line symbol was used on Fig. 36a to indicate the position of the Big Thompson echoes), and easterly winds of 20 to 26 kt (10-13 m s'^) were carrying cool, moist air into the storm area. The low pressure area over western Colorado had begun to fill as temperatures cooled and thunderstorm activity spread northward into that area. Thunderstorms over southwestern Colorado had organized into a squall line, which stretched from north of Grand Junction to east of Al- buquerque, New Mexico. This line of storms was moving northward ahead of and associated with the 500 mb short-wave trough. The northward moving squall line in Utah and the westward moving surface front were closing rapidly toward a juncture in southwestern Wyoming. The Big Thompson storm had a top of 62,000 ft MSL (18.9 km), measured by the Limon radar (see Fig. 36b), which is ex- tremely high for a thunderstorm over the Rocky Mountain area. The line of storms over the plains of Colorado and Kansas had weakened rapidly. These storms had de- veloped during maximum heating as the trailing front pushed into the region. Once developed, they had remained nearly stationary (see Fig. 36c). As the front continued to move southward, the bound- ary layer air mass changed from that sam- pled by the 0000 GMT Denver and Dodge City soundings to conditions similar to those depicted in the reconstructed Love- land sounding. Over the plains this effect, coupled with slight diurnal cooling, caused dissipation of the storms. Even though the air mass had become conditionallv much 47 105° 110' 105' 100' Figure 36a. Regional surface analysis for 0200 GMT, 1 August 1 976. Refer to legend of Fig. 12a for details. 1 1 5-' 110" 105" 100" 95" S'^y Ci, \ / o ^^ — / o %, NO ECHOES _, ) o' / RW- ; ==00 i "o FEW \ ^ J=^-\ ~4> <^°TRW-^ "^'■' - 7 1> "> ^ 25 / H^ Lr^ NE """"L 5 of ^^ TRW-H - __,^ , -■■ "— " A^ No ( \ V. -^RW-H-/- ~N RW- SCTD TRW-H- /\20 \ \ 520^ lis ®\ (D 2Vn^ 220 .r-^RW/- 520 ^ 550r;^ 15 Q ^^V. ^ 270 - ^^^ °0 ^^ >7 . ^V. - ' 4bUARW-l- ____g TRW-H-/- ; 200 "^ 10 \TRW/-i ^ 440 .._ / , Oo°^-^ >> TRW-H- TRW 35° OTRW (??o /(- 450 y NE 1 Z> 1 o OO qO TRW 00° TRW-H 1 1 - 35° 110' 10'y 100" Figure 36b. Radar summary chart for 0135 GMT, 1 August 1976. 48 r Figure 36c. GOES-1 infrared photograph for 0200 GMT, 1 August 1976. more unstable, there was no mechanism acting over the plains to lift this air to its LFC and release the instability. By 0400 GMT (refer to Figs. 37a, b, and c) the meso-high pressure system be- hind the Colorado s(juall line had inten- sified, and the low pressure area over northwestern Colorado had continued to fill. This squall line was moving northward toward the Big Thompson storm complex. The squall line moving northward out of Utah and the moist air moving westward across southern Wyoming met near Rock Springs, Wyoming. A large thunderstorm intensified rapidly in the area where the two features interacted. The intense nature of both the Big Thompson and Rock Springs storms was indicated by continu- ous lightning reported by observers in these areas and by the very bright and cold storm tops in the infrared satellite photo- graph. By 0600 GMT (Figs. 38a, b, and c) the Big Thompson storm had begun moving northward along the Front Range in re- sponse to the approaching short wave and associated squall line. The Rock Springs storm now appeared (in the satellite photo- graph) larger and more intense than the Big Thompson storm, as it moved north- ward across sparsely populated portions of western Wyoming. However, an impor- tant difference between this storm and the Big Thompson storm was that the stronger steering level winds over western Wyo- ming kept it moving steadily northward. 49 110° Figure 37a. Regional surface analysis for 0400 GMT, 1 August 1 976. Refer to legend of Fig. 12a for details. TRW )TRW TRW+ llO" 105" 100" Figure 37b. Radar summary chart for 0335 GMT, 1 August 1976. 50 Figure 37c. GOES-1 infrared photograph for 0400 GMT, 1 August 1976. 51 40< 35' 95° 60 ^^"xy 30.20 A V" Figure 38a. Regional surface analysis for 0600 GMT, 1 August 1 976. Refer to legend of Fig. 1 2a for details. TT . STC RW- \^ SCTD \ TRW+ 1 J . STC 0, 0.\ / r \ TRVy 1 I \ H 350 XfRW/+ \ X_''- NO ECHOES STC 0\ STC J ]\ -1 V " ■ t 440^ NMRS ® \ 1 1 I \ STC ! H ~ - .. .- -- L ! 1 © TRWx \ I 1o\ 280 NO ECHOES ®RW- TRW oRw- o jf^^ 1 o i <^ 1 o o 8 RW- TRW NA TRW 1 1 1 -- 110' 105" 100" Figure 38b. Radar summary chart for 0535 GMT, 1 August 1976. 52 Figure 38c. GOES-1 infrared photograph for 0600 GMT, 1 August 1976. 53 4.2 Synoptic Scale Analyses for 1200 GMT, 1 August 1976 Surface and upper-air analyses are sho\\ n, along with \ orticit\ and continuitv charts, for 1200 GMT in Figs. 39 through 44. Thunderstorm acti\'it\ had merged into one convective featiue during the early morning hours, and had continued to move north-northeastward ahead of the 500 mh short-wave trough. This arc of shower and thunderstorm activitx' had persisted through the earh' morning hours and stretched from central Idaho through western South Dakota to western Kansas (refer to Fig. 39). This line of convective activity was immediately followed by a meso-high pressure area, which in turn was trailed by a wake of low pressure (bubble high/wake depression). Positions of the squall lines are shown on the continuity map of Fig. 40. The or- ganization and movement of these lines of storms were strongly controlled b\' the movement of the 500 mb short-wave trough. Between 0600 and 1000 GMT thunderstorm activity merged into one 115° Figure 39. Synoptic scale surface analysis for 1200 GMT, 1 August 1976. Refer to legend of Fig. 4 for details. 54 large arc of storms along and ahead of the trough. One important effect of this short-wave tiough was its role in moving the storm complex northward out of the Big Thompson drainage. The upper level ridge (refer to Figs. 41, 42, and 43) had continued to build and increase in amplitude overnight ahead of the 500 mh short-wave trough, which stretched from northern California through western Wyoming to west Texas. Winds at 700 and 500 mh had veered and become south to southwesterly over Colorado. The short-wave trough was easily identifiable on the vorticity analysis shown in Fig. 44, and the sc^uall line lay in the legion of posi- tive vorticity advection just ahead of the trough. Although the synoptic scale pattern of a trough over the West and a ridge over the plains persisted for the next several days, the interaction of mesoscale weather fea- tures and orogiaphic features that had pro- duced the nearly stationary Big Thompson storm complex was not repeated along the Front Range of the Rock\' Mountains. Figure 40. Continuity positions of Utah squall line (black), Colorado squall line (orange), and Big Thompson storm complex (broken orange). Time in GMT is given for each line. 55 130° 125° 120° 115° 110° 105" 1 00° 95" 90" 85° 80° 50" -1^ "99^ ^ --^316 316 314 312 310\ 308 306 ^^ ) 1 50° ^^^in : MTii4/ / ^^>.^i::::>^ \/\ \\ \ \ \ \ \ /-^/o6ol<7/6 ^ 7^^^ ^^^^..^^.^ ^S. \ ^ \ /12Y+03/ .'^ / V ,X J49 \ \ \ 1 >^^i vWfclJl -T^X^ A H V\ \ \ y^3o/ /^ Af^ v\-iUui\A/^rNi ^ // 45° ^ / y) \ / osa 1 ' ^ ''^ \ ^ ) \ — Vnv Vr**V*y\ -^^"^ i^^n^ 45° ^ \ \ X Tn\ — \ \ /^^^f'W H-i^^\ \ „\,76 V^??OlSo ■ \ \ \\ \A^\ -2V\^ y^ I #l6^i?fi ^ — \~7\ \ X^a^V \ *^° \ \ \\ - - , ._ x. ; x '\ \ n\v \ \ \j 10-) "^tA V^^\\^H^^itf^^Y^ 1 V'^joa- — - ~-3rw^ \ X \ 1?^ \^x\\WT^ 'W^3p8 \iV\>^A>^^ Vt -hkt'Jjm'' 40 ^50097'' 40° 35° fc y. 07ol47 \ iLlM ^312, 35" \-i 317\^ioA,S;^ >. \ V \ =^320--\ 1^ \ ,-,„-y^'''^^ 1 ^o^^C" LA^;;i^ 318 \ ^316 18^/ 320 u'^-^A^--^" V V" 30° 314-^ \ o" 316-'"'^ * w Sic . \ ,,202 / >\ ./ ;/ .^ V"A' X12 XJ^j 30"' N . ^\ /V jf / \^i*A/ / /"''''Vi/ !>f ^\ "/"^°'oW / M / LV/^ / fi 25° -- \^^V 319^/ /^^l ^^ / \ 318^ / 1 V \ ) ( \ 316 1 1 \ 25° / 1 / \ \ 115 110" 105° 100° 95" 90° Figure 41. 700 mb analysis for 1200 GMT, 1 August 1976. Refer to legend of Fig. 5 for details. 56 130" 125" 120" 115" 110" 105" 100" 95" 90" 85" Figure 42. 500 mb analysis for 1200 GMT, 1 August 1976. Refer to legend of Fig. 6 for details. 57 1 1 5" Figure 43. 300 mb analysis for 1200 GMT, 1 August 1976. Refer to legend of Fig. 7 for details. 58 130" 125° 120° 115° 110" 105° 100° 95" 90" 85° Figure 44. NMC-LFM vorticity analysis for 1 200 GIMT, 1 August 1976. Refer to legend of Fig. 25 for details. 59 5. Physical Model of the Storm NHRE radar data were taken at 1.4° increments in elevation ani2;le, and these data were nsed in conjunction with the constructed Loxeland sounding (Fit:;. 32) to dexelop a phxsical model of the Big Thompson storm. A 0030 GMT GOES photograph was used to estimate the hori- zontal extent of the anvil. The model (Fig. 45) is valid for an approximate time of 0045 GMT (note the 5 to 1 compression of the horizontal scale). The cross section is along a NW-SE line (oriented from 314° to 134°) that passes through a point at 40°30'N, 105°21'W. The cross section lies nearly along the storm inflow. The stiong low-level inflow allowed a huge amount of mass to he processed by the storm. As the low-level flow ap- proached the Front Range, stratus and stratocumulus formed in the shallow laver between the LCL and the LFC. The 0000 GMT Ft. Collins surface observation indi- cated a thin, broken deck of cloud based at 4,000 ft AGL (1.2 km). When this low-level air was forced above the LFC, explosive convective growth occurred. Radai" data indicated that new cells formed in the in- flow and moved north-northwestward into the storm complex. Cloud base was effec- tivcK on the ground in the stoiin area. This, combined with the high incloud freezing level (5.8 km MSL) and height of the -25°C isotherm (9.6 km MSL), indicated an unusually deep (for Colorado storms) layei" for warm cloud coalescence processes to act. Since there was weak wind shear, rela- tively moist environmental air, and little flow relative to the storm al)ove the LFC, entrainment of drier air into the storm was suppressed. This reduced evaporative loss- es and increased the storm's precipitation efficiency. With cloud base at the surface, subcloud evaporation was also minimized, further enhancing the precipitation effi- ciency. Neither entrainment nor evapora- tive processes were able to produce strong downdrafts within the storm. It might be expected that light windshear and rapid cloud droplet growth would combine to quickly overload the up- draft with precipitation. However, notice that the echo contours slope toward the northwest with height. Vertical transport of strong, low-level easterly momentum into uppei" parts of the storm produced an updraft stiucture that sloped with height toward the northwest. The tilted updnift allowed large precipitation droplets to fall out of the rear of the updraft, enabling the system to exist in a nearly steady state. This model of the Big Thompson storm is considerably different from that developed by St. Amand et al. (1972) for the Rapid City, South Dakota, storm com- plex. Their model cloud sloped strongly to the east above 450 mb and invoked a recy- cling of water substance to help explain the high precipitation efficiency of the system. Deep convection that develops within a negatively sheared air mass (characterized 60 2 — Poudre Little South Buckhorn """" „ !°'l Creek Poudre River Big Thompson River Berthoud Wind^ (kt) - 60 - 55 - 50 - 45 - 40 - 35 ^ - 30 2 - 25 - 20 = 15 - 10 = 5 40 30 20 10 Distance (km) 20 30 40 50 Figure 45. Physical model of one of the initial ceils of the Big Thompson storm complex. LCL, LFC, winds, and levels of 0°C and -25°C isotherms are from interpolated Loveland, Colorado, sounding. Grover 0045 GMT radar reflectivities are shown with 10 dBZ contours beginning with 15 dBZ level. 61 Distance (km) Figure 46a. Grover radar reflectivity (dBZ) cross section through one of the initial cells of the Big Thompson storm complex at 0045 GMT. Winds (in knots) are from the Loveland, Colorado, interpolated sounding. Cross section was orientated from 31 4° to 1 34° through a point at 40°30' N and 105°21'W. Cross section lies nearly along storm inflow. Distance (km) Figure 46b. Grover radar reflectivity (dBZ) cross section through the severe Fleming, Colorado, hailstorm (after Browning and Foote, 1975). Winds (in knots) are from a nearby upper-air sounding. Cross section lies nearly along storm inflow. 62 by stronu easterly flow in low levels de- creasintz; with hei<2;ht) will normally tend to slope toward the east with height. West- waid nioxenient is greater near eloiid base than at middle and upper levels and pro- duces the slope. Howevei', if terrain lifting tri^^ers intense convection within a neii;a- tiveK sheared air mass, the structure and dynamics of the resultint:; storms ma\ he (jfuite different from those of a movinii; storm. The Olograph) effectively forces the lower portions of the cloud system to re- main quasi-stationary. The vertical trans- port of strong, low-level easterly momen- tum then produces an updnift structure that slopes along the direction of inflow. Grover radar data seem to verifv this hypothesis. When the thunderstorm cells initially formed they moved rapidly toward the higher terrain and sloped toward the southeast with height. However, once they moved over the higher terrain they became nearly stationary and developed a north- westward slope with height. An efficient unloading of the updnift in the lower half of the cloud would encourage large updraft velocities to develop within the glaciated upper cloud and allow cloud tops to grow to ver\' high levels. Indeed, radar observations indicated that tops of the Big Thompson storm complex were about 6,000 ft (1.8 km) higher than those of other storms on the eastern slopes and plains. Fig. 46 shows cross sections along the inflow of (a) the Big Thompson storm and (b) the Fleming storm (a severe hailstorm). The Big Thompson storm was tilted along the direction of inflow. There was little echo overhang and onl\ a hint of Weak Echo Region (WER). fiigh reflectivities >55 dBZ were low in the cloud, below 7 km. This suggests that warm cloud coales- cence processes were highU' efficient. The severe Fleming hailstorm de- veloped in a strongly sheared environment and bore little resemblance to the Big Thompson storm. (For details on the Flem- ing storm refer to Browning and Foote, 1975.) It displayed a large WER with a massive overhang of high reflectivity parti- cles. High reflectivities ^ 55 dBZ extended upward within the storm to over 10 km. Severely sheared storms are highly ineffi- cient precipitation producers (Browning and Foote, 1975; Marwitz, 1972). The negatively sheared Big Thompson storm was characterized b\ a high precipi- tation efficiency and b\ a lack of large hail or strong surface outflow. Similar featmes were characteristic of the Rapid City storm (Dennis et al., 1973). Schroeder's (1977) conceptual model of a flood producing thimderstorm over the Koolau Range in Hawaii is quite similai" to the Big Thompson model indicating that this t\pe of highly efficient precipitation system is probably not unicjue to the Front Range area. 6. Comparison of Conditions Associated With the Big Thompson and Rapid City Floods Meteorological aspects of the destruc- tive flood that occurred at Rapid City, South Dakota, on 9 June 1972 (Dennis et al., 1973) were very similar to those of the Big Thompson flood event. These features are compared since it is felt that the condi- tions that produce this type of intense east-slope rainstorm are recognizable 3 to 12 h in advance. Early recognition of this type of flash flood potential would allow more effective use of special statements and watches to alert the public. The surface anaKses at 1200 CMT (Fig. 47) prior to the floods show that sur- face conditions were dominated by large polar highs with lower pressures south and west of the threat areas. The dashed line on the Rapid City map (Fig. 47a) represents a weak wind shift line. Both cases were characterized by a zone of unusually moist air with high dewpoints noith of the polar 63 iroui cuul In hotter, drier air west and south of the trout. Fiiis. 48 aud 49 are the 700 aud 500 ml) auaKses at 1200 GMT. The couiuiou fea- tures of importance are tlie following: 1. The large, uegatixeK tilted ridge o\ er the plains with a trough over the west. 2. A short-wave trough south aud west of the threat areas. This shoit wave was moving northward up the back of the ridge. 3. Light south to southeasterly winds over the threat areas. 4. Abundant moisture over much of the western United States. Fig. 50 presents the 1200 GMT upper-air soundings taken at Rapid City and Huron, South Dakota, and also shows the 1200 GMT Denver sounding and the 1920 GMT Sterling sounding. In both cases, the soundings taken east-northeast of the threat areas were more representa- tive of the air mass in which intense thun- derstorms were triggered later in the day. Both the Huron and Sterling soundings were conditionally very unstable but with temperature inversions capping the unsta- ble, moist boundary layer. Both soundings were taken near the upper ridgeline and showed easterK' winds below the inver- sions. The surface analyses at 0000 GMT (Fig. 51), just prior to the floods, showed that strong easterly winds were pushing moist, conditionally unstable, boundar\ layer air westward onto higher terrain. A zone of moist air north of the polar front fed the storms, while drier air lay south and west of the front. Pressures were lower south and west of the front helping to maximize the easterly, upslope flow. Figs. 52 and 53 depict the 700 and 500 ml) analyses at 0000 GMT. Notice that at 700 ml) moist southeasterly flow exists in the threat areas ahead of the approaching short waves. The 500 mb short-wave troughs, which eventually moved the flood producing storms away from the moun- tains, were approaching from the south- west. Winds over the threat area were light southerly, and the movement and orienta- tion of the 500 mb short-wave troughs helped maintain lower pressures west of the threat area. The 0000 GMT Rapid City and inter- polated Loveland soundings are shown in Fig. 54. Both soundings were conditionally unstable with strong easterly winds and very high moisture contents within the lowest 1 to 1.5 km. Winds above 500 mb were light south to southeasterly. The Rapid City sonde sampled active convec- tion (Dennis et al., 1973) and thus did not exhibit the temperature inversion. Table 3. Comparison of upper-air parameters for Rapid City, Loveland, and typical severe thunderstorm soundings Location Time Mean vapor Planetary mixing ratio Lifted Index Boundary Layer Wind velocity (deg/kt) (GMT) (lovi/est 100 mb) (lowest 100 mb) (lowest km) 700 mb 500 mb 300 mg (g kg M (deg/kt) Rapid City 0000 GMT, 10 June 1972 14.0 -5 100/40 140/31 150/18 140/08 Loveland (interpolated) 0000 GMT, 1 Aug. 1976 14.8 -6 100/50 130/30 190/15 170/15 Typical plains severe storm ' 12 to 14 -4 to -6 200/35 225/40 235/50 245/70 •Data from Maddox (1976). 64 Table 3 compares the Rapid City and Loveland soundings with conditions as- sociated with severe Central Plains thun- derstorms (storms producing large hail and/or damaging winds and/or tornadoes). L.I. s and moisture contents are compara- ble; however, the wind fields are dramati- cally different. The wind veers with height in both types of soundings, but the heavy rain soundings are characterized by strong easteiK' winds near the sin-face and light southerly winds aloft. This "reverse shear contrasts markedly with the strongly sheared severe storm environment. The following list summarizes similar meteorological conditions associated with both the Rapid City and Big Thompson flash floods. A. Common large scale features 1. A middle and upper tropospheric long-wave trough lies over the western United States with a negatively tilted ridge just east of the threat area. 2. A weak 500 mb short-wave trough rotates northward in the long-wave trough and approaches the threat area. 3. Light southeast to south-southeast winds (5 to 8 m s ^) are present in the upper troposphere over the threat area. 4. A slow moving, or stationary, polar front lies just south of the threat area. 5. Unusually moist (vapor mixing ratio of 13 to 15 g kg"^) and strong (15 to 25 m s"^) easterly low-level flow moves into threat area. 6. High moisture content is present through a large depth of the tropo- sphere (siuTace through 300 mb). B. Common mesoscale features 1. Afternoon heating west of the threat area and cold advection east of storm area combine to intensify thickness and pressure gradients. Low-level wind flow maximizes about sunset. 2. A narrow band of conditionally un- stable air (L.I. = -4 to -7) moves southward and westward behind the polar front. This air mass is capped by a temperature inversion. 3. Orographic lift provides the mechanism needed to release the in- stability, and heavy rains fall over mid- dle elevations of the iiffected drainages. The moisture content of the low-level air is so great that the lifting needed to trigger storm development occurs east of the highest terrain. 4. Convective cells move slowly north-northwestward, and continued cell redevelopment on the southeast flank of the thunderstorm complex re- sults in a quasi-stationary precipitation system. C. Common microphysical characteristics 1. A high freezing level (5 to 6 km MSL) in the cloud and low cloud bases combine to produce a deep warm cloud layer in which precipitation particles may grow by coalescence processes. 2. Negligible vertical wind shear and light winds relative to the storm, at middle and upper levels, combined with moist environmental air and cloud bases at or near the surface, suppress evaporative losses and the formation of downdrafts. These conditions contri- bute to high precipitation efficiency. Most of these important large scale and mesoscale features can be detected and monitored if National Meteorological Center analyses and forecasts are used in conjunction with aviation teletype data, radar reports, and satellite photographs. As recently emphasized by Mogil and Groper (1976), when the detection level shifts from the large scale down to actual storm reports, the forecaster should rely heavily upon mesoscale analyses and in- terpretations to delineate the potential for excessive rainfall. 65 50" 45" 40" 35" 30" 25" 130" 125' 1 20" 1 1 5" 110" 105" 100" 95" 90" 85" 80" .' 1 6 8 50 ylo X ^<1_55 J 1 1 11^ \ \ 40 _^ 26 ^.24 ^^^ ,11 ^^^ .20 50° 45° 40° 35° 30° 25" =-^^45 ^ 1 50 V_ /U5 ^^ 6^ r=6is:^^ ^^^:^^^^^^i2 }/v \ =^^^^^ — C\^*Pf/ ,f V . 8'/ 1/ te/ 1^ ^Q 12 \\\ \ \ V 1 \ nlV^ \ =p "O / 1 1 TWrnL 6 1 la ?\\ 115" 110" 1 05" 100" 95" 90' Figure 47a. Rapid City synoptic scale surface analysis for 1 200 GMT, 9 June 1 972. Refer to legend of Fig. 4 for details. 66 50° 45° Figure 47b. Big Thompson synoptic scale surface analysis for 1200 GMT, 31 July 1 976. (Same as Fig. 4.) 67 45° Figure 48a. Rapid City 700 mb analysis for 1200 GMT, 9 June 1972. Refer to legend of Fig. 5 for details. 68 130° 125° 120° 115° il 0° 105° 100° 95° 90" 85" 80" '( \ ' v \ \ '\ \. ^07^. \ ^Vs^ ) 50° ^06" 306- 1 . -02 o oy ' \ -010139 ^ ^ ■J ^ ^>?02 45° orS>sa 1 ^^^--^^ 308- -^ ^C^=?^"X^ -«»-,<; /'I +01- \\w\5153 \^ \\^?"^^^l^n- 2p306 310- /l^ ^b^^ /^^ u /^"^Q 4 ^^\5^ -^ ^^T^^2\,^^>^ -^ .04^06^ / 312-=^=^^^^ C OTn 133 3l/ ^/ / JO o^ / osib 148 ; / <\^' 1 1 09^1 54N 8 P ont i\ Vwo"vv /^ — 1 r L 316 ) A 1' < iiiie-iy ^ y;j Xr-4^X^ ^ \^ ^l%Lf -^"^\ ^'^ vfe ~~~-:^3i4 ^T>316' OrPl67 9 Vol r'V \l 40° 08\3l5(l) 35° osfci'" r* / 2y-02 / / _ _^^0 " +00^ ' \\V 319; ~~"\ V r" r 175 03 04 / 30° - VlY 09„ni i_r^^V / 11^ 187 Q. 01 '>^^ / / ^ M O /\^ V s ^^318-^ ^ ^ J^m 316 '\\^ H Ht? 184 V^n .©168 y 09„ 189 y^ 1 25° - , V \,^ \ 318 \\ 12CX18S -021 I 318 *~''^rfe.Ai9i \ 13^-01 \ 115" no- 1 05° 1 00"' 95" 90" 45° 30° 25° Figure 48b. Big Thompson 700 mb analysis for 1200 GMT, 31 July 1976. (Same as Fig. 5.) 69 Figure 49a. Rapid City 500 mb analysis for 1200 GMT, 9 June 1972. Refer to legend of Fig. 6 for details. 70 Figure 49b. Big Thompson 500 mb analysis for 1200 GMT, 31 July 1976. (Same as Fig. 6.) 71 Figure 50a. Skew T/Log P plot of 1200 GMT, 9 June 1972, Rapid City upper-air sounding. Figure 50b. Sl ' w\ 50" 45" 40" 35" 30" 25 ^<^i^ V \ 7W\ ',85 ^ / T- / / ^^Sf V Wk^ J) \l- ' Ml 4 1 ^ V V\ r As 115" 110" 105" 100" 95 ■ 90' Figure 51a. Rapid City synoptic scale surface analysis for 0000 GMT, 10 June 1972. Refer to legend of Fig. 4 for details. 74 50° 45° Figure 51b. Big Thompson synoptic scale surface analysis for 0000 GMT, 1 August 1976. (Same as Fig. 18.) 75 Figure 52a. Rapid City 700 mb analysis for 0000 GMT, 10 June 1972. Refer to legend of Fig. 5 for details. 76 130° 125° 120° 115° 110° 105" 100" 95° 90° 85" Figure 52b. Big Thompson 700 mb analysis for 0000 GMT, 1 August 1976. (Same as Fig. 21.) 77 !iO 125" 120" 115" 110" 105" 100" 95" 90" 85 Figure 53a. Rapid City 500 mb analysis for 0000 GMT, 10 June 1972. Refer to legend of Fig. 6 for details. 78 Figure 53b. Big Thompson 500 mb analysis for 0000 GMT, 1 August 1976. (Same as Fig. 22.) 79 Figure 54a. Skew T/Log P plot of 0000 GMT, 10 June 1972, Rapid City upper-air sounding. q: 400 80 Figure 54b. Skew T/Log P plot of 0000 GMT, 1 August 1976, upper-air sounding constructed for Loveland, Colorado. (Same as Fig. 32.) 7. Summary and Conclusions A large thunderstorm complex drop- ped as much as 10 in of rain in the foothills west of Loveland, Colorado, during the evening hours of 31 July 1976. A resultant flash flood devastated the Big Thompson Canyon claiming at least 139 lives and pro- ducing property damage of about $35.5 million. The storm complex developed when a conditionally unstable and extremely moist air mass pushed upslope into the Front Range of the Rocky Mountains. This air mass lay to the rear of a polar front, and the associated temperature inversion pre- vented release of the potential instability until the air mass experienced strong oro- graphic lifting. Low pressiue centered west of the Front Range helped to increase the magnitude of the easterly upslope flow. The storm was triggered just west of a negatively tilted upper level ridge, where winds above the temperature inversion were light southeasterly. These weak steer- ing winds allowed the storm complex to remain nearly stationary over the Front Range for 2 to 3 h as new cells generated on the sourtheast flank and moved into the storm complex. A weak 500 mb short-wave trough, approaching the storm area from the south-southwest, moved the storm slowly northward into Wyoming and even- tually out onto the plains. Local area surface analyses indicated that prefrontal, high-based thunderstorms that moved over the Front Range west of Boulder may have limited the southward extension of the storm complex and helped to focus moist inflow into the Big Thompson drainage. Detailed radar analyses indicated that as much as 7.5 in of rainfall over the main fork of the Big Thompson may have fallen in a little more than 1 h. A physical model of the storm complex was developed using an interpolated Love- land, Colorado, upper-air sounding and 1.4° inciemental elevation angle data from the NHRE radar. A sloping updnift struc- ture allowed efficient precipitation unload- ing. A low shear environment above the temperature inversion, relatively moist air at high levels, and little flow relative to the stoiin helped minimize the eftects ol en- trainment on precipitation efficiency. Cloud base on or near the surface also minimized subcloud evaporation. As a result precipi- tation efficiency was increased, and strong downdrafts, which would have tended to force new cell development at other loca- tions, were subdued. Meteorological conditions associated with the Big Thompson and Rapid City flash floods have been compared. These events were very similar, and a set of im- portant meteorological features associated with this type of convective cloudburst has been presented. Recognition of these fea- tures should be of use in predicting the potential and issuing watches for this type of severe storm with geographical areas and lead times similar to those currently issued for expected tornado activity. 81 8. Acknowledgments The authors would Hke to thank the man\ persons and organizations who supphed data used in this study. The De- partment of Atmospheric Science, Col- orado State Univ., provided data for Ft. Collins. Dr. C. Glenn Cobb supplied Greele\ data measured at Ross Hall, Univ. of Northern Colorado. John M. West of Rockwell International Corp. supplied the Rock)' Flats data, and Frank Pratte of Wave Propagation Laboratory (NOAA-ERL) provided the data from Table Mountain. James F. W. Purdom of NOAA-NESS Ap- plications Group supplied copies of GOES imagery. The NHRE and the National Center for Atmospheric Research Field Observing Facility obtained and supplied the invaluable sounding, radar, and surface data from the NHRE site in northeastern Colorado. National Climatic Center (NOAA-EDS) personnel expeditiously filled several large data requests. John Asztalos of Waukesha, Wisconsin, provided copies of photographs that he had taken of the developing storm. 82 9. References Browning, K. A., and G. B. Foote, 1975. Air flow and hail growth in supercell storms and some implications for hail suppression. National Hail Research Experiment Tech. Rept. No. 75/1, 77 pp. Dennis, A. S., R. A. Schleusener, J. H. Hirsch, and A. Koscielski, 1973. Meteorology of the Black Hills flood of 1972. Institute of Atmos. Sci., Re- port 73-4, South Diikota School of Mines and Technology, Rapid City, 41 pp. Grozier, R. U., J. F. McCain, L. F. Lang, and D. C. Merriman, 1976. The Big Thompson River Flood of July 31-1 August, 1976. Larimer County, Col- orado, U.S.G.S. and Colorado Water Consei"va- tion Board Flood Information Report, Denver, 78 pp. Lott, G. A., 1976. Precipitable water over the Unit- ed States, Volume 1: Monthlv means. NOAA Tech. Rept. NWS 20, 173 pp.' Maddox, R. A., 1976. An evaluation of tornado proximit\- wind and stahilit\ data. Man. Wea. Ret., 104:133-142. Manvitz, J. D., 1972. The structure and motion of severe hailstones. Part III: Severely sheared storms. /. Appl. Meteor., ll:189-20l' Miller, R. C, 1972. Notes on analysis and severe storm forecasting procedures of the Air Foice Global Weather Central. Air Weather Sei^vice Tech. Rept. 200 (Rev.), 102 pp. Mogil, H. M., and H. S. Groper, 1976. On the short range prediction of localized excessive convective rainfall. Preprints, Conf. on Hydro-Meteorology, Amer. Meteor. Soc, Ft. Worth, 9-12. NOAA, 1976. Big Thompson Canyon flash flood of July 31-Augiist 1, 1976. Natiu-al Disaster Survey Report 76-1, U.S. Dept. of Commerce, Rock- ville, Md., 41 pp. Randerson, D., 1976. Meteorological analysis for the Las Vegas, Nevada, flood of 3 July 1975. Mon. Wea. Rev., 104:719-727. Schroeder, T. A., 1977. Meteorological analvsis of an Oiihu Flood. Mon. Wea. Rev., 105:458-468. St. Amand, P., R. J. Davis, and R. D. Elliott, 1972. Report on Rapid City flood of 9 June 1972. Report to South Dakota Weather Control Com- mission, Pierre, South Dakota, 37 pp. Williams, G., 1976. Application of the National Weather Service flash-flood program in the Western Region. NOAA Tech. Memo. NWS WR-103. 20 pp. 83 il U S- GOVERNMENT PRINTING OFFICE 1977 — 780-900 v/ ADDDDESSb^4Eb ■^;:| -iM >:;■;•!» im ;'■''":';';:'"'::::';:•;;.::? 1 ■ ■:Qlg ill |i