Observations of a Gravity Wave Packet as it Moved Across Northwest Kansas


Llyle Barker
National Weather Service Office
Goodland, Kansas




On July 11, 1994, a series of wave clouds propagated across Northwest Kansas. The clouds appeared to be associated with a packet of small wavelength gravity waves. Fluctuations in several atmospheric variables were evident during the passage of the packet. The WSR-88D and the ASOS unit, both at Renner Field in Goodland, Kansas provided high resolution observations of the packet as it moved across the region.


The 0000 UTC 11 July 1994 numerical forecast model runs indicated a rather strong summer vorticity maximum and associated short wave trough to move eastward across Nebraska by 0000 UTC 12 July at 500 millibars. An associated cold front was expected to move into northeast Colorado overnight before stalling over extreme northern Kansas during the day. The bulk of the dynamics and cold air advection was predicted to remain across Nebraska. Between 0000 UTC and 0800 UTC 11 July, moist convection was active along and ahead of the boundary across eastern Wyoming, western Nebraska, and extreme northeastern Colorado. The thunderstorms were not particularly noteworthy, though several warnings were issued for the Nebraska Panhandle during the evening hours. By 0800 UTC, the activity had pretty much dissipated (Figure 1).

A strong ridge had built into the central high plains on July 10, allowing a subsidence inversion to develop. This feature was evident at the 750-700 millibar level on the 1200 UTC 11 July sounding from Dodge City Kansas (Figure 2). Winds overnight were from the south and southwest below this inversion as shown by the sounding and the Vertical Wind Profiler (VWP) product from the Goodland WSR-88D (Figure 3). Surface winds at Goodland varied little overnight with winds from 160-170 degrees at speeds of 13-17 knots. The Goodland VWP showed these winds increasing with height up to the inversion level. Above the inversion, the sounding and VWP indicated flow from the north at 10 to 20 knots. With the approach of the front, surface pressures reduced to sea level fell slowly overnight, from 1011.8 millibars at 0600 UTC to 1010.5 millibars at 1200 UTC.



    A. Visual

At approximately 1140 UTC, what at first seemed to be a roll cloud appeared on the northwest horizon as viewed from the National Weather Service Office at Goodland. At first glance, it appeared that the cold front might have accelerated southeastward in response to the earlier convection across western Nebraska. Such roll clouds have been experienced at the station during the passage of strong fronts during high moisture conditions. However, with most of the cold air moving across Nebraska, the discontinuity associated with the southern extent of the front was mainly in the form of wind direction rather than temperature or moisture. After reviewing surface data across western Nebraska and northeastern Colorado, it was discovered that winds continued to be out of the south and southwest at Akron, Colorado and Sydney, Nebraska, both stations well northwest of the "roll" cloud.

Figure 1. Goodland WSR-88D Base Reflectivity Product (248 NM Range, 0740 UTC).

Figure 2. 1200 UTC 11 July 1994 Dodge City Kansas Sounding.

Figure 3. Goodland WSR-88D Vertical Wind Profiler (VWP) Product (Precursor Environment: 1101-1200 UTC).

As the packet moved closer, apparently there was no circulation about a horizontal axis in the formation, and the cloud was reclassified as a rope cloud. As the formation continued southeast toward the station, the cloud elements were observed to be moving down the length of the cloud from southwest to northeast, similar to the environmental winds below the inversion. The length of the cloud was indeterminable as it stretched in a straight line from horizon to horizon.


As the approach continued, three more formations of the same type became visible behind the first rope cloud. Each cloud had a base of about 1500 ft with a vertical extent estimated visually at close to 4000 ft. The distance between each rope was between 5-10 miles.


As the fourth and final cloud moved southeast of the station, the early morning sun began to mix the layer, and by 1330 UTC only a few wisps of cloud material remained visible.


The passage of the series of rope clouds was easily detected by the Goodland WSR-88D. The radar observed the first of the series around 1130 UTC (Figure 4). By 1217 UTC (Figure 5), the entire set was visible on the Base Reflectivity product. By 1223 UTC (Figure 6), the first wave cloud was moving across the RDA. The extremely linear character of the clouds is particularly evident on the magnified image. By 1322 UTC (Figure 7), the final cloud had moved across the site and was dissipating rapidly.

Due to the proximity of the echoes to the ground the WSR-88D had a limited surveillance capability with these formations. Even with this in mind, the geographical extent of these rope clouds was rather impressive with lengths estimated to be over 50 miles. Due to the lack of significant width of the echoes and the increased power output, resolution, and lowest detectable signal of the doppler radar, the WSR-88D could easily pick out these cloud formations even while in Volume Coverage Pattern 21, while the co-located WSR-74c was unable to resolve them.

Figure 4. Goodland WSR-88D Base Reflectivity Product (124 NM Range, 1130 UTC).

Figure 5. Goodland WSR-88D Base Reflectivity Product (124 NM Range, 1217 UTC).

Figure 6. Goodland WSR-88D Base Reflectivity Product (124 NM Range - Magnification 8x, 1223 UTC).

Figure 7. Goodland WSR-88D Base Reflectivity Product (124 NM Range, 1322 UTC).

The WSR-88D velocity data (Figure 8) also displayed important features contained in the waves. The base velocity product clearly shows the southerly flow ahead of the packet and the series of wind shifts associated with the packet (see ASOS observation section below). The WSR-88D VWP product (Figure 9) from the time immediately following the wave passages the increase in RMS error in low gate wind direction, possibly caused by rapid indicated velocity fluctuations in the wind as the packet moved through.

Figure 8. Goodland WSR-88D Base Velocity Product (62 NM Range-Magnification 4x, 1235 UTC).

Figure 9. Goodland WSR-88D VWP Product (1200 UTC-1259 UTC).


As the cloud packet traversed the area, rapid changes in pressure and wind direction were observed in the time series of the Renner Field Goodland ASOS five minute observation data (Figure 10). With a time resolution of five minutes and a wave period of close to 10 minutes, some of the finer temporal details of the pressure and velocity pattern were lost. However the pressure tendency from this five minute sampling still shows the four waves as they propagated across the ASOS site. The increase in surface pressure reduced to sea level of 2.0 millibars in a 15 minute period with the first wave was quite impressive, as was the peak perturbation from initial conditions of 2.3 millibars. The data includes both a rise and a fall of 1.5 millibars between adjacent five minute observations. The actual extremes were probably even more substantial when consideration is given to the observation time resolution of five minutes and the ASOS pressure sensor data averaging of one minute.

Figure 10. Goodland ASOS five Minute Pressure and Wind Observations (1145 UTC-1325 UTC).

The wind data also shows the wave passages as directional shifts from the background southerly direction to the west northwest as the wave crest's pass. The five minute observation time resolution and two minute averaging of wind direction both played important roles in smoothing the data. Visual observations of the F-420 wind equipment indicated that the wind veered too as much as 330 degrees during the wave crests. Though not observed in the cloud field or distinctly in the pressure pattern, the wind direction changes hint at five waves rather than just the four visible. After the packet passed, winds returned to a southwest or west direction until frontal passage about 1630 UTC.

As the first wind shift occurred, the Weather Service Forecast Office (WSFO) with terminal forecast (FT) responsibility, WSFO Topeka Kansas, assumed that the wind shift was of a permanent nature and that the front had reached Renner Field. The FT was then amended. The forecaster on duty at Goodland than contacted the aviation forecaster at Topeka, to inform him of the events that were occurring. A second amendment was then issued placing the prevailing winds back to a southwest direction. This type of event shows the advantages of mesoscale analysis and the new generation of equipment in assisting with forecast update decisions.

What was also indicated by this event is a limitation of ASOS when reporting wind shifts. The first wind shift that occurs between hourly SAOs is retained to be remarked on the next hourly, even if additional shifts occur meanwhile (Department of the Navy et al. 1992). When multiple wind shifts occur within an hour, as in this event, or during a convective situation, the wind shift remark can be misleading. As ASOS is currently implemented, wind direction may be identical on adjacent hourly SAOs, however a wind shift remark may be retained on the second hourly. This could lead to confusion when examining observations.


The data from the Goodland ASOS, WSR-88D, and visual observations by National Weather Service employees lead to a conclusion that a packet of short wavelength gravity waves passed across Goodland, Kansas during the early morning hours of July 11, 1994. Though the wavelength was on the small size of the classic definition of mesoscale gravity waves (Uccellini and Koch 1987), they still exhibited many characteristics normally found with such phenomena. Koch et al. (1993) published a conceptual model of this type of event (Figure 11). The pressure changes, wind direction shifts, and even the rope cloud formation fit this model nicely.

Figure 11. Conceptual Model of Gravity Wave Effects. Sine Wave Represents Pressure Tendency. Thick Arrows Indicate Wind Direction. Thin Arrows Represent Vertical Velocity. Arrow C Shows Propagation Direction (Koch et al. 1993).

The most obvious source for these waves was the evening convection that occurred over western Nebraska. Much research has described thunderstorm outflow as a source for these type of packets (Bosart and Cussen 1973, Lin and Goff 1988, among others). Erickson and Whitney (1973) published a satellite picture that showed a similar wave packet moving through the Mississippi Valley region, created by thunderstorms over Kansas during the previous night.

Though the convection occurred some 200 miles northwest of Goodland, conditions were favorable for ducting the wave energy. The inversion evident on the 1200 UTC Dodge City sounding (Figure 2) was an important atmospheric condition conducive to wave refraction and energy trapping (Hooke 1986).

Another possible mechanism for enhancement of these waves was the potential for shear instability evident in the sounding and the Goodland WSR-88D VWP Product. Shear instability may arise when a discontinuity in current speed occurs at the interface of two atmospheric layers in which little mixing is occurring (Huschke et al. 1959). The wind fields may have large enough kinetic energy differentials across the layer boundary so that if an air parcel is displaced vertically, it may be able to extract some available kinetic energy to develop an oscillating motion about the equilibrium position (Hooke 1986). The ducting of this instability may be further enhanced by a neutral or slightly unstable layer overlying a stable layer such as an inversion (Bluestein 1993). Both factors are present in the 1200 UTC Dodge City sounding. The Goodland VWP product also indicates the large difference in wind direction and speed between the inversion layer and neutral layer above.

Though the shear instability may have been a factor in trapping the energy, wind speeds indicated by both the sounding and Goodland WSR-88D was likely too low to initiate wavelengths since those apparently associated with the rope clouds. The Glossary of Meteorology equates an estimated phase propagation speed of a shear instability induced wave with the average of the flow below and above the inversion (Huschke et al. 1959). It also derives a critical wavelength required for a shear instability initiated wave to develop. To satisfy the phase propagation equation, one must assume that the flow immediately above the inversion was originally stronger in the region of wave genesis, than what was indicated by the Dodge City sounding and Goodland WSR-88D VWP. One possibility for producing this type of environment would involve outflow from the Nebraska convection. This air would have warmed due to downslope flow as it moved southeast, then overrode a radiationally cooled layer over northeast Colorado. The outflow wind produced by the convection would have been significantly stronger nearer the source, possibly strong enough to produce the required values in the phase propagation equation, and then weakened as it moved further from the storms and into western Kansas. Even with this assumption, the derived critical wavelength for wave development is about 1-3 km rather than the observed 10 km. This inconsistency on the face of it seems to dismiss this mechanism as the initiator. However, Chimonas and Grant (1984) showed that it is possible for a series of waves induced by shear instability to interact to develop gravity waves of longer wavelengths than would normally be expected, so this mechanism can not be completely ruled out.

The depth of the region below the inversion at Dodge City and the height of the wind shift on the Goodland WSR-88D VWP correlate well with the height and depth of the rope clouds. ASOS measured the cloud bases during the event at 1600 ft above ground level. WSR-88D VWP data and base reflective elevation data showed cloud tops near 8000 feet above mean sea level, or approximately 4000 feet above ground level.

Satellite data would have been a valuable tool for confirming the source area for the packet. However, infrared images were the only images available and resolution was too low for detection of the waves. By the time visible images were downloaded, the clouds had dissipated. The new higher resolution GOES 8 sensors may be able to distinguish these type of features during future events.


A variety of data sources were used to analyze an unusual series of rope clouds that moved across northwest Kansas during the early morning hours of July 11, 1994. The clouds appeared to develop in response to a gravity wave packet. The waves were likely initiated by moist convection during the previous evening over the Nebraska Panhandle, with other factors possibly contributing. These waves moved into an environment conducive to long distance propagation.

The advent of new technologies such as ASOS, and the WSR-88D, and their increased spatial and temporal resolutions were invaluable in examining these formations. The event also served to point out the importance of mesoscale analysis in evaluating current forecast packages.

The impressive visual characteristics of the formation led to much speculation by the public and the local media. Several phone calls were received by the National Weather Service containing rather bizarre explanations of the phenomena. These ranged from cloud seeding experiments gone awry to allegations of government cover-ups. Most refused to believe that nature could produce such straight and perfect cylinders. Hopefully this paper has helped explain this event to be much less sinister in nature.


Bluestein H.B., 1993: Synoptic-Dynamic Meteorology in Midlatitudes, Volume II: Observations and Theory of Weather Systems. Oxford University Press, New York NY, 594pp.

Bosart L.F., and J.P. Cussen Jr., 1973: Gravity Wave Phenomena Accompanying East Coast Cyclogenesis, Mon. Wea. Rev., 101, 446-454.

Chimonas G., and J.R. Grant, 1984: Shear Excitation of Gravity Waves - Upscale Scattering from the Kelvin-Helmholtz Waves. J. Atmos. Sci., 41, 2278-2288.

Department of the Navy, FAA, and NOAA, June 1992: Automated Surface Observing System User's Guide, Washington D.C., 56pp.

Erickson C.O., and L.F. Whitney Jr., 1973: Gravity Waves Following Severe Thunderstorms. Mon. Wea. Rev., 101, 708-711.

Hooke M.H., 1986: Gravity Waves, Mesoscale Meteorology and Forecasting. Ed. P. Ray., AMS, Boston MA, pp 272-288,

Huschke, R.E., 1959: Glossary of Meteorology. AMS, Boston, 638pp

Koch S.E., F. Einaudi, P.B. Dorian, S. Long, and G.M. Heymsfield, 1993: A Mesoscale Gravity-Wave Event Observed During CCOPE. Part IV: Stability, Analysis, and Doppler-derived Wave Vertical Structure, Mon. Wea. Rev., 121, 2483-2510.

Lin Y.-L., and R.C. Goff, 1988: A Study of a Mesoscale Solitary Wave in the Atmosphere Originating Near a Region of Deep Convection. J. Atmos. Sci., 45, 194-205.

Uccellini, L.W., and S.E. Koch, 1987: The Synoptic Setting and Possible Energy Sources for Mesoscale Wave Disturbances. Mon. Wea. Rev., 115, 721-729


USA.gov is the U.S. government's official web portal to all federal, state and local government web resources and services.