A MOUNTAIN WAVE EVENT WEST OF THE COLORADO PARK RANGE
Christopher N. Jones1*, Jeffery D. Colton1, Ray McAnelly2, and Michael P. Meyers1
1National Weather Service, Grand Junction, Colorado
2Colorado State University, Fort Collins, Colorado
During the overnight hours of 22-23 April 1999 wind gusts to 40 ms-1 occurred west of the Park Range (3000-3700 m MSL) in northcentral Colorado (Fig. 1) toppling numerous pine trees and damaging at least six homes in the vicinity of Steamboat Springs (2100 m MSL-SBS Fig. 1). The lower and middle tropospheric wind flow during the windstorm was atypically east-to-west. This event showed similarities to the 25 October 1997 Routt National Forest blowdown event where 13,000 acres of old growth forest were devastated by wind gusts in excess of 50 ms-1. Snook et al. (1999) and Meyers et al. (2000) detailed the nature and possible mechanisms responsible for this destructive storm. The research presented here will detail the synoptic and mesoscale environment of the 22-23 April 1999 windstorm through analyses of various National Centers for Environmental Prediction (NCEP) models, Regional Atmospheric Modeling System (RAMS) forecasts, and observed conditions. Comparisons will be made to the 25 October 1997 Routt National Forest devastation and post-storm research in an attempt to aid operational meteorologists in forecasting these rare events.
Mountain wave events occurring to the lee of a barrier have long been modeled and observed (Scorer 1949; Lilly and Zipser 1972). Lilly and Zipser (1972) conducted a thorough survey of a mountain wave event near Boulder, Colorado on 11 January 1972, which has been the focus of several subsequent studies and model comparisons (e.g. Doyle et al. 1998). One significant feature found in their research was the wave-like structure of the potential temperature () field associated with mountain waves and severe windstorms. Indications of wave-like motion were also found in an evolving vertical velocity field produced in the numerical modeling of Peltier and Clark (1979). Other observational research indicates the importance of crest-level flow across the barrier and the cross-barrier sea-level pressure gradient to the severity of mountain wave windstorms (Colle and Mass 1998). Through numerical modeling the presence of a critical level has been shown to be a possible contributor to severe windstorms by reflection of the mountain wave (Clark and Peltier 1984). Although common east of mountain barriers, such as the Colorado Front Range, mountain waves are observed far less often west of mountain barriers. In fact, the bulk of the theoretical and observational research pertaining to mountain wave formation has focused on westerly lower and middle tropospheric wind flow.
3. SYNOPTIC AND MESOSCALE ENVIRONMENT
Upper air analyses of 23 April 1999 0000 UTC (times hereafter UTC) depicted a nearly vertically stacked cut-off low near Las Vegas (LAS). Contours of the height field at 500 hPa showed the tightest gradient over northern and central Idaho. At the surface, high pressure was progressing south through the plains. The cut-off low exhibited little movement over the next 12 hours, remaining near LAS at 23 April 1200. However, surface high pressure had moved to the Texas panhandle and was associated with 12-hour 700 hPa height rises of 50-70m at Bismarck (BIS), Denver (DEN), and Riverton, Wyoming (RIW). Additionally, the tightest height gradient at 500 hPa shifted from Idaho to the northern Utah border. This southward shift caused the 500 hPa east winds at Boise (BOI) to increase from 12 ms-1 to 25 ms-1, and backed Salt Lake City (SLC) winds from 150 at 5 ms-1 to 090 at 12 ms-1. Winds at RIW and at the Medicine Bow (MBW) wind profiler in southcentral Wyoming were consistently from the east at 15 to 20 ms-1 during this 12 hour period. Evidence of the strong easterly winds across Wyoming was inferred from infrared satellite imagery in the form of mountain wave clouds west of the Wind River Range (not shown).
The cross-barrier sea-level pressure (SLP) gradient increased dramatically between 22 April 1800 and 23 April 1400. METAR observations from Cheyenne, Wyoming (CYS) and Craig, Colorado (CAG) were chosen to measure the cross-barrier SLP difference. CYS and CAG lie 169 km east and 72 km west of the Park Range crest, respectively. The SLP at CYS increased from 1009.0 to 1026.8 hPa, while measurements at CAG showed an increase from 1004.1 hPa to 1010.0 hPa. Thus, the cross-barrier SLP difference increased from 4.9 hPa to 16.8 hPa in 20 hours across a distance of 241 km. Colle and Mass (1998) found a common factor among the strongest west Cascade wind events to be a SLP difference of >8 hPa between Seattle and Yakima (169 km). Similarly, a cross-barrier SLP difference between RIW and SLC (320 km) of >8 hPa is typically enough to generate gap winds on the west slope of the Wasatch (Dunn 1999). A time-height section of the MBW profiler (Fig. 2) showed 20 ms-1 crest-level flow from the east at the onset of the damaging winds (23 April 0400). The presence of a critical level was seen approximately 3.5 km above ground level (AGL) where the wind became southerly, a direction parallel to the Park Range. Given that the MBW profiler resides over 2 km MSL, the presence of the critical level would be around 500 hPa. Mountaintop level in the Park Range is roughly 650 hPa.
Fig. 2. Medicine Bow, Wyoming wind profiler time-height section of 22-23 April 1999. Wind in knots. Y-axis in meters.
4. NUMERICAL MODELING
The various models of the NCEP suite available to the operational forecaster were all quite similar for this event. All of the NCEP models forecasted a cut-off low near LAS inducing strong easterly synoptic flow across southern Wyoming and northern Colorado between 0000 and 1200 23 April. The model that provided the best guidance to forecast the mountain wave event was the NGM, which consistently produced the strongest winds during the event. The NGM employs 80 km horizontal grid spacing and 18 vertical levels.
Fig. 3. Plan view of NGM 18-hour forecast of 700 hPa wind (knots) and height (m) valid 23 April 1999 0600 UTC.
The 22 April 1200 model run of the NGM gave strong indications to the possibility of damaging winds in northern Colorado and adjacent areas. A plan view of this region showed east winds at 700 hPa in excess of 25 ms-1 valid at 23 April 0600 (Fig. 3). Figure 4 is a cross-section from Fort Collins, Colorado (FNL) to CAG depicting wind and valid at the same time. The cross-section showed crest-level winds of 25 ms-1 and a mean-state critical level around 600 hPa in the Park Range. The wind profile matched closely that of the nearby MBW profiler seen in Fig. 2. It is important to note that wind speeds were nearly unchanged through the 24-hour forecast. Stability was difficult to infer from the cross-section, but slight isentrope packing could be seen just above crest-level.
Fig. 4. NGM 18-hour forecast cross-section of wind (knots) and potential temperature (K) valid 23 April 1999 0600 UTC.
Additionally, a less stable layer was observed in the
middle troposphere. Snook et al. (1999) illustrated how
stability conditions similar to these contributed to
mountain wave formation during the 25 October 1997
event. NGM forecasts of SLP valid at 18 and 24-hours
produced a maximum gradient between CYS and CAG of
11.0 hPa. These parameters forecast by the 22 April
1200 NGM model run gave several indications that
severe downslope winds were possible west of the Park
Range. However, the wave-like structure of the and w
fields observed in previous research was not seen in the
NGM solution. This was most likely due to the coarse
resolution and smoothed terrain surfaces of the model.
Using the RAMS forecast to complement the NGM
solution provided even more evidence of possible
mountain wave formation.
Fig. 5. RAMS forecast of near-surface wind (knots) vaild 23 April 1999 0800 UTC. Bold line denotes >50 knots. Shaded areas >60 knots.
RAMS is a local mesoscale model which was
run daily during the 1998-99 winter season at both
Colorado State University (CSU) and the NOAA Forecasts Systems Laboratory (FSL). RAMS is
supported through funding provided by the Cooperative
Program for Operational Meteorology, Education, and
Training (COMET), operated jointly in cooperation with
the National Weather Service (NWS), CSU, and FSL. A
nested-grid version of RAMS (Pielke et al. 1992) provided
by CSU was utilized for this research. The CSU model is
run daily at 0000 utilizing parallel processing for the
interactive nested grid version of the model. The model
includes two interactive, nested grids. The coarse grid
covers the western United States with 48 km horizontal
grid spacing; and the fine grid encompasses all of
Colorado and portions of Wyoming, Utah, South Dakota,
Nebraska, and Kansas at 12 km grid spacing. Both grids
utilize 26 vertical levels beginning at 250 m grid spacing
near the surface, gradually stretching to a maximum of
1000 m. One distinct advantage of RAMS is the more
advanced physics employed compared to that of the
NCEP model suite. RAMS includes a mixed-phased
microphysical scheme described by Walko et al. (1995).
Another distinct advantage of this high-resolution model
is the improved representation of complex terrain, which
is invaluable especially over mountainous regions.
Fig. 6. RAMS forecast of wind (knots) and potential temperature [(K) bold lines] valid 23 April 1999 0800 UTC. Shaded area >50 knots. Note wind maximum of 60 knots above SBS.
The 23 April 0000 RAMS model run was employed for this research. All forecast fields discussed here are valid at 23 April 0800, which approximates the mid-point of the severe wind event. Similar to that of the NGM, the RAMS forecast a tightening pressure gradient between CYS and CAG with a cross-barrier difference of 14.0 hPa (not shown). This value nearly equaled the observed cross-barrier pressure difference of 16.8 hPa. A plan view of near-surface winds (Fig. 5) painted 30 ms-1 from the Park Range west to around SBS. Several cross-sections were created using RAMS, running along an axis from FNL to just south of SLC (approximated by the 40.5N latitude). The wind and fields in Fig. 6 clearly showed the mountain wave pattern with descending easterly winds of 27-32 ms-1 west of the Park Range. The enhanced damage area was located almost directly below this wind maximum. The pattern also depicted a layer of instability in the middle troposphere from 5000-10,000 m MSL overlaying a stable layer above crest-level. The RAMS cross-section of the w component of the wind (Fig. 7) showed promise in replicating the wave-like couplet modeled by Peltier and Clark (1979). Three distinct downward maxima could be seen west of the Colorado Front Range. Of particular interest was the downward velocity maximum of 0.5 ms-1 displaced above SBS. Special note should be made of the intense downward velocity seen west of the Wasatch Front (far left, Fig. 7). Just prior to the valid time of the RAMS forecast a gust to 113 mph was recorded near Brigham City, Utah. This mark established a new record for lower elevations in the state of Utah (Dunn 1999).
Fig. 7. RAMS forecast of w (cm/s) valid 23 April 1999 0800 UTC. Downward w denoted by dark shading.
The model guidance provided by RAMS closely resembled the findings of previous observational and theoretical research. Several important mesoscale factors appeared to have congealed in the SBS vicinity. Crest-level flow was strong, and isentropes illustrated a mountain wave pattern and favorable stability profile. RAMS also displayed strong downward descent of the mountain wave in the w field, and the cross-barrier SLP gradient was significant. The use of RAMS, in concert with the solution produced by the NGM, provided excellent identification to the formation of a mountain wave west of the Colorado Park Range.
5. COMPARISON TO THE 25 OCTOBER 1997 ROUTT NATIONAL FOREST BLOWDOWN
The Routt National Forest blowdown of 25 October 1997 is the only other mountain wave event occurring west of the Continental Divide in the Colorado Rockies that has been examined in detail (Snook et al. 1999, Meyers et al. 2000). This event produced wind gusts in excess of 50 ms-1, which downed 13,000 acres of old growth forest in the Park Range. The toppled trees fell in a swath several miles wide and 20 miles long. The findings of various studies pertaining to this devastating blowdown will be discussed here and compared to those of the 22-23 April 1999 event.
Analyses of 25 October 1997 at 0000 indicated a deep, cut-off low in northeast New Mexico which produced lower and middle tropospheric easterly flow of 15-20 ms-1 across northern Colorado. Contours of the forecast height fields at 850 hPa and 700 hPa showed the tightest gradient to be across northern Colorado. Wind profiler data from Platteville, Colorado (PTL) and MBW confirmed this synoptic scale flow. Both the PTL and MBW profilers also showed the presence of a mean-state critical level around 2 km AGL (Wesley et al. 1999). Although the cut-off low was much farther east than that observed in the 1999 event, the resulting synoptic and mesoscale wind structure was quite similar. Surface observations during the 1997 event indicated a cross-barrier SLP difference of only 6.3 hPa between CYS and CAG. This value was far less than the 16.8 hPa difference observed in the 1999 event. Another observed parameter that varied greatly between the two mountain wave events was lower tropospheric temperature. Snook et al. (1999) showed how deep, very cold lower tropospheric air enhanced mountain wave formation during the 1997 blowdown. Mountain top temperatures of -20C were recorded in the 1997 event, whereas readings of around -6C were recorded during the 1999 event. Also, surface temperatures at the mid-point of each wind event were taken from METAR observations at nearby Hayden, Colorado (HDN, Fig. 1). Surface temperatures of the 1997 event were 8.0C colder than those observed in the 1999 event (-5.0C in 1997 vs. 3.0C in 1999). The extremely cold lower tropospheric air present in 1997 may have been a significant factor in the magnitude of that event.
Investigations of the 1997 event through numerical modeling were conducted using the RAMS model. Findings of these simulations indicated that strong synoptic easterly flow and deep, very cold lower tropospheric air were the primary meteorological factors enhancing the mountain wave (Snook et al 1998). Three grids were utilized in an attempt to isolate the relatively small areas where the greatest damage and highest wind gusts occurred: 1) a 15 km outer grid, 2) a nested 5 km grid, and 3) two 1.67 km inner grids within the boundaries of the 5 km grid. One of the 1.67 km grids focused specifically on the Park Range.
Fig. 8. RAMS 1.67 km grid lowest model level wind (knots) forecast over the Park Range valid 25 October 1997 1000 UTC. White lines are isotachs.
Fig. 9. RAMS 1.67 km grid cross-section of wind (knots) and potential temperature (K). Forecast valid 25 October 1997 1000 UTC. Black and white lines denote isentropes and isotachs, respectively.
Results of the 1.67 km RAMS simulations depicted 30 ms-1 east winds across the Park Range at the lowest model level (Fig. 8). A cross-section of wind and taken perpendicular to the Park Range showed a 40ms-1 maximum descending the west side of the barrier along with tight isentrope packing just above crest-level (Fig. 9). The mountain wave structure was evident in the isentrope pattern. Additionally, a wave induced critical level was observed in the middle troposphere, characterized by a wind minimum. Despite the varying horizontal grid spatial differences of the two RAMS forecasts (12 km in 1999 and 1.67 km in 1997), the overall structure of the wind and fields were quite similar. The difference was the magnitude of these features. The discrepancies existed primarily in the stronger jet maximum descending the lee of the barrier in the 1997 blowdown, and the tighter isentrope gradient near crest-level present during the same event.
This research was prompted by the occurrence of a severe mountain wave event west of the Colorado Park Range on 22-23 April 1999, the second such event within a span of two years in that region. Prior to the 25 October 1997 Routt National Forest blowdown little knowledge or literature existed regarding mountain wave events west of the Continental Divide in Colorado. A detailed examination of the mountain wave of 22-23 April 1999, and subsequent comparisons to the 1997 event, provide some insight to the synoptic and mesoscale characteristics common to this phenomena.
NCEP models served as a good diagnostic tool to infer the possible formation and timing of the 22-23 April 1999 mountain wave episode. The NGM depicted strong easterly synoptic flow near crest-level and the presence of a mean-state critical level evident in cross-sections. Confirmation of the strong crest-level flow and critical level was found in the MBW wind profiler. However, the coarse horizontal resolution of the NGM made it difficult to discern atmospheric stability and mesoscale forcing mechanisms. A local mesoscale model, RAMS, was used in concert with the NGM to forecast the magnitude and location of the mesoscale phenomena. The 12 km RAMS forecast added superior predictive value in forecasting the severe winds. RAMS forecast fields of wind and were able to show the mountain wave pattern and descending winds west of the barrier. Forecast profiles of middle tropospheric instability overlaying regions of crest-level stability, and forecasts of downward vertical velocity also gave operational forecasters strong indications of mountain wave formation. The results closely resembled the overall wind flow structure observed in the 1.67 km RAMS simulations of the 25 October 1997 event. Strong synoptic and crest-level easterly flow, and the presence of a critical level were common to both episodes. In addition, RAMS profiles of the 1997 event clearly showed a well-developed mountain wave within the stable layer.
Several factors appear to have caused the higher wind speeds observed in the 1997 event. First, Snook et al. (1999) showed how the presence of very cold lower tropospheric air could modify stability profiles, thereby enhancing mountain wave development by nonlinear effects. Mountain top temperatures in the 1997 event were 14C colder than those measured in the 1999 event. Secondly, forecasts of 700 hPa height contours point to the strongest winds occurring over northern Colorado in the 1997 event, and over northern Utah in the 1999 event. Also, RAMS forecast a descending jet maximum of 40 ms-1 during the 1997 event, whereas a 30 ms-1 maximum was forecast during the 1999 event. Both of these factors would favor the higher wind speeds observed in the Routt National Forest blowdown. Finally, more research needs to be conducted concerning mountain waves west of the Continental Divide and the relationship to cross-barrier pressure difference. Contrary to stability profiles and forecast winds which favor the 1997 event for higher wind speeds, the cross-barrier pressure gradient favored the 1999 event. The 1999 episode generated a 16.8 hPa SLP difference between CYS and CAG, a value more than 10 hPa greater than that generated during the 1997 event. Colle and Mass (1998) investigated 55 western Washington windstorms and found that 82% of variance of event severity was due to the magnitude of cross-barrier flow and cross-barrier pressure gradient. A comparison of the Colorado Park Range windstorms indicates that cross-barrier pressure gradient may not be a good predictor of event severity in that region.
Operational forecasters west of a mountain barrier should be attentive to cut-off lows south or southwest of the forecast area that could produce strong synoptic easterly winds. Wind profilers and the velocity azimuth display (VAD) wind profile of the WSR-88D can supplement upper-air reports to determine the presence of embedded wind maxima. This could prove especially helpful in data sparse regions of the west and between upper-air launch times.
The authors wish to thank Doug Wesley and John Snook for their insight and comments on this research. Additional thanks are given to Liz Page for providing archived data used to diagnose the event, and the staff at WFO Grand Junction for their support.
Clark, T.L., and W.R. Peltier, 1984: Critical level reflection and the resonant growth of nonlinear mountain waves. J. Atmos. Sci.,41, 3122-3134.
Colle, B.A., and C.F. Mass, 1998: Windstorms along the western side of the Washington Cascade Mountains. Part II: Characteristics of past events and three-dimensional idealized simulations. Mon. Wea. Rev., 126, 53-71.
Doyle, J.D., D.R. Durran, C. Chen, V. Grubisic, C.-Y. Huang, Y.-L. Lin, G.S. Poulos, D. Weber, M. Wurtele, M. Xue, 1998: An intercomparison of model predicted wave breaking for the 11 January 1972 Boulder windstorm. Preprints, Eighth Conference on Mountain Meteorology, Flagstaff, AZ, Amer. Meteor. Soc. 131-137.
Dunn, L., 1999: Downslope windstorm case study Wasatch Front - April 23, 1999. Western Region Technical Attachment NO. 99-09.
Lilly, D.K., and E.J. Zipser, 1972: The Front Range windstorm of 11 January 1972 - A meteorological narrative. Weatherwise, 25, 56-63.
Meyers, M.P., J.S. Snook, D.A. Wesley, and G.S. Poulos, 2000: A Rocky Mountain storm - Part II: The forest blowdown - observations, dynamics and modeling. Submitted to Wea. Forecasting.
Peltier, W.R., and T.L. Clark, 1979: The evolution and stability of finite-amplitude mountain waves. Part II: Surface wave drag and severe downslope windstorms. J. Atmos. Sci., 36, 1498-1529.
Pielke, R.A., W.R. Cotton, R.L. Walko, C.J. Tremback, W.A. Lyons, L.D. Grasso, M.E. Nicholls, M.D. Moran, D.A. Wesley, T.J. Lee, and J.H. Copeland, 1992: A comprehensive meteorological modeling system RAMS. Meteor. Atmos. Phys., 49, 69-91.
Scorer, R.S., 1949: Theory of waves in the lee of mountains. Quart. J. Roy. Meteor. Soc., 75, 41-56.
Snook, J.S., M.P. Meyers, G.S. Poulos, and D.A. Wesley, 1998: A local model comparison of the October 1997 Rocky Mountain blizzard and severe wind event. Preprints, Eighth Conference on Mountain Meteorology, Flagstaff, AZ, Amer. Meteor. Soc. 31-36.
Snook, J.S., M.P. Meyers, G.S. Poulos, and D.A. Wesley, 1999: Longer range numerical prediction of the Routt National Forest severe wind event. Preprints, 17th Conference on Weather Analysis and Forecasting, Denver, CO, Amer. Meteor. Soc. 137-140.
Walko, R.L., W.R. Cotton, M.P. Meyers, and J.L. Harrington, 1995: New RAMS cloud microphysics parameterization. Part I: The single-moment scheme. Atmos. Research, 38, 29-62.
Wesley, D.A., G.S. Poulos, J.S. Snook, and M.P. Meyers, 1999: Meso- scale snowfall and wind variability in a Front Range blizzard. Preprints, 17th Conference on Weather Analysis and Forecasting, Denver, CO, Amer. Meteor. Soc. 121-126.