The Evolution and Time Scale of Mesoscale Processes that Created an Intense Mesoscale Snowband on

15 March 2004 in Des Moines, Iowa

 

Emily Eisenacher and Dr. Charles Graves

Saint Louis University

St. Louis, MO

Submitted: April 2008

 

Corresponding author address: Emily Eisenacher, Department of Earth & Atmospheric Sciences, Saint Louis University, 205 O’Neil Hall, 3642 Lindell Blvd., St. Louis, MO 63108

Email: eisenaeb@eas.slu.edu

_____________________________________________________________________________________

 

ABSTRACT

 

Mesoscale snowbands cause locally heavy areas of snowfall; however, it is operationally difficult to predict the timing and location of snowband formation.  For this reason, forecasting snowbands becomes a matter of situational awareness and nowcasting, rather than predicting where the snowbands and locally heavy snowfall will set up before they form.  An investigation of the temporal and spatial radar characteristics of an intense snowband case along with an investigation of their environment provides more insight into the time tendency and the orientation of the snowbands compared to the mechanisms that produce heavy banded snowfall.  Mesoscale processes such as frontogenesis, conditional symmetric instability, and equivalent potential vorticity and their time tendency are compared to the spatial and temporal evolution of the radar reflectivity of the snowband.  This particular case study reveals that frontogenesis and conditional symmetric instability were present in the environment five and a half hours before the snowband formed, while equivalent potential vorticity was present in the environment approximately 30 minutes prior to the initial multiple snowband forming.  Further investigation of more cases is needed to identify a trend in the timing and onset of mesoscale processes in environments conducive to heavy banded snowfall.

______________________________________________________________________________

 

1.  INTRODUCTION

Despite all the technological advancements in meteorology over the past thirty years, forecasting the location and amount of localized heavy snow due to snowbands remains a challenge.  Banded snowfall is especially difficult to forecast, because the mechanisms responsible for producing it are not clearly understood or easy to diagnose operationally.  Additionally, their horizontal scale is too small to be accurately resolved by the coarse grid spacing of today's numerical models (Banacos 2003).  Therefore, forecasters are able to diagnose threat regions where heavy snowfall may potentially fall, but cannot accurately forecast the development of locally heavy snowbands.  Novak et al. (2006) outlined a forecast strategy to diagnose the potential for snowbands by diagnosing the synoptic scale days in advance, then the mesoscale processes hours in advance to determine a threat region.  As snowbands form, forecasters tend to react to the enhancement of physical processes responsible for snowbands, rather than anticipate the development of enhanced snowbands (Weismueller and Zubrick 1998). 

This article focuses on the time scale and evolution of mesoscale processes that form snowbands.  An investigation of the temporal and spatial evolution of the mesoscale processes will help forecasters better understand key parameters in the environment before snowbands form rather than reacting to them after they have already developed.

2.  DATA ANALYSIS

            The General Meteorological Package (GEMPAK) and NMAP2 were used to analyze surface, upper air, model data, satellite, and radar data. The surface, upper air, 80 km 3-hourly Rapid Update Cycle (RUC) II initialization model data were obtained via Unidata’s Internet Data Distribution.  Level II and III WSR-88D radar data were obtained from the National Climatic Data Center and satellite data were obtained from the SSEC data center at the University Wisconsin-Madison.

The mesoscale processes analyzed include frontogenesis, conditional symmetric instability (CSI), and equivalent potential vorticity (EPV).  The investigation on frontogenesis follows Keyser et. al (1988), and the investigation of EPV follows McCann (1995) and Moore and Lambert (1993). This case study relies on past research by Moore et al. (2005), Nicosia and Grumm (1999), Novak et. al (2004), and Shultz and Schumacher (1999), who investigated snowbands through case studies and contributed to the development of the key parameters used to forecast snowbands.  

3.  EVENT OVERVIEW

This case is characterized by an east-west oriented snowband over Des Moines, Iowa.  A radar animation, Fig. 1, depicts the main snowband’s formation, propagation, and dissipation.  The snowband formed about 0922 (all times in UTC) as multiple snowbands merged and lasted through 1822, when it began to rotate counterclockwise and break up.  A weak surface low pressure system moved south from Colorado into Kansas and snow fell over most of Iowa except for the extreme northeast portions.  The heaviest snow was over west-central Iowa with 46.7 cm (18.4 inches) in Sioux City and 39.6 cm (15.6 inches) in Des Moines.  Fig. 2 shows the event snowfall with amounts greater than 20 cm (8.0 inches) in orange. 

Although the surface system was not impressive, there was sufficient upper level forcing.  Fig. 3a and Fig. 3b show the initial low pressure system that formed over southeastern Colorado. Another low pressure center developed in southern Nebraska at 1200 and became the main system while the original low pressure center dissipated by 1500.  The inverted trough and the warm front were responsible for the snow over Des Moines.

This system experienced very little intensification and remained relatively quasi-stationary over the Midwest for its entire lifetime.  In addition to the surface features being weak, the upper-level features were benign.  By 1200, the system had a closed circulation at 850 hPa over Nebraska (Fig. 4), and a broad trough at higher levels.  Fig. 5 shows the broad trough at 700 hPa and Fig. 6 is the 500 hPa analysis with a small vorticity center approximately 14x10-5s-1 over the Upper Plains.  At 300 hPa, there was a jet streak entrance region in the vicinity of the system, providing upper level divergence Fig. 7.  This system could be classified as the frontal/weak cyclogenesis pattern from Banacos (2003) since it had a positively titled 700 hPa trough and downstream confluent flow.

4.  RADAR CHARACTERISTICS

The snowband in this event was embedded within a larger swath of snow.   Fig. 8 and Table 1 outline the structure and movement of the snowband with time.  The snowband started as multiple snowbands, merged, intensified, and then broke up into several snowbands which dissipated.  The snowband propagated northeast, and then became quasi-stationary over Des Moines, before dissipating.  Evidence of banding began when there were three snowbands present over Iowa at 0732.  By 0922 these three snowbands merged into one snowband that stretched northwest to southeast from Little Sioux, Iowa to Kirksville, Missouri.  The snowband continued to intensify and propagate northeast and stalled over west-central Iowa from 1021 to 1240.

The snowband remained over Des Moines through its life cycle and was most intense about 1600 with regions of 35 dBZ embedded.  After about 1822, the snowband began to weaken and break up into smaller snowbands and after 1957 it broke up further while continuing to rotate counterclockwise.  After 2356, the multiple snowbands dissipated to become part of the larger snow field.  Jones and Thorpe (1992) indicate that counterclockwise snowband rotation is indicative of weakening and clockwise rotation is indicative of intensification.        

5.  MESOSCALE ANALYSIS

a.  FRONTOGENESIS

Mid-level frontogenesis was analyzed in plan view and in a cross section.  A cross section was taken from St. Cloud, Minnesota to Oklahoma City, Oklahoma. The cross section showed mid-level frontogenesis between 700 hPa and 500 hPa on the order of 1.0 (K 100 km-1 3 hr-1) to as high as 2.0 (K 100 km-1 3 hr-1) Frontogenesis was actually in the environment over Omaha and Des Moines well before the snowband formed.  At 0000 the frontogenesis was 1.0 (K 100 km-1 3 hr-1) and it weakened slowly until 0800 when it regained strength to 1.0 (K 100 km-1 3 hr-1) again.  During this time period, the snow was approaching Des Moines and multiple snowbands formed at 0732.  The frontogenesis strengthened to 1.2 (K 100 km-1 3 hr-1) at 0900, which is almost coincident with the formation of the single snowband at 0922.

While the snowband was propagating northeast closer to Des Moines, frontogenesis strengthened to 1.4 (K 100 km-1 3 hr-1) by 1100 and the most impressive frontogenesis was 2.0 (K 100 km-1 3 hr-1) at 1200, found between 700 hPa and 500 hPa.  Fig. 9 shows the frontogenesis while the snowband was intense around 1600.  By 1700, the frontogenesis had weakened to 1.0 (K 100 km-1 3 hr-1) and continued to weaken through 2300.  There were two periods when the snowband rotated counterclockwise and weakened, at 1350 and after 1957.  During both of these periods frontogenesis was weakening and the snowband was weakening.

Analysis of 750 hPa and 650 hPa plan-view frontogenesis with a radar composite overlay led to an interesting correlation between the main storm and snowband compared to the frontogenesis structure.  Frontogenesis began to set up in the environment at 0400 approximately five and a half hours before the single snowband formed at 0922.   At 0700 the 750 hPa frontogenesis became more linear and the strongest frontogenesis of 0.8 (K 100 km-1 3 hr-1) was focused over the region where the snowband began to form. At 0800, the 750 hPa frontogenesis continued to match the orientation and location of the snowband and the 650 hPa frontogenesis was just to the north of the main storm.  The 750 hPa frontogenesis strengthened to 1.2 (K 100 km-1 3 hr-1) by 0900 and was oriented northwest to southeast in the same location and orientation of the snowband.  The intensification and organization of frontogenesis correlated with the formation of the single snowband.

At 0900, just before the snowband developed, the 750 hPa frontogenesis was stronger than the 650 hPa frontogenesis and the 750 hPa forcing was closer to the position of the snowband.  The 750 hPa frontogenesis was more compact and focused.   Fig. 10 shows the frontogenesis at 1600, while the snowband was intense.

The 750 hPa frontogenesis was stronger than the 650 hPa frontogenesis from 2000 through 2200.  The 650 hPa frontogenesis continued to be oriented more east-west, but the strong forcing associated with the 750 hPa frontogenesis was closer to the location and orientation of the snowband.  As shown in Fig. 11, at 2200, the 750 hPa frontogenesis was right over the snowband over Des Moines and the 650 hPa frontogenesis was more aligned with the main precipitation shield.  By 2300 the frontogenesis at 650 hPa and 750 hPa were both 0.8 (K 100 km-1 3 hr-1) and the 750 hPa frontogenesis was less focused and widespread.  At this point, the snowband had broken up and was almost absorbed into the main snow shield.

Novak et. al. (2006) and Banacos (2003), both use radar reflectivity with 700 hPa frontogenesis overlay.  It is know operationally that using one level for frontogenesis is not recommended, and analysis of multiple levels or even layers may be necessary.  In this particular case, a useful pattern was recognized.  It appeared that 750 hPa frontogenesis was a good indicator of the location and orientation of the snowband and the 650 hPa frontogenesis was an indicator of the general location and orientation of the main precipitation shield.  A through analysis of more cases is necessary to determine if this correlation between 750 hPa and 650 hPa frontogenesis is a trend in snowband events, or if it is erroneous.  The strongest 650 hPa frontogenesis was at 1400, after the main snowband was well developed and the strongest 750 hPa frontogenesis was at 2000 and 2100; coincident with the snowband breaking up and rotating at 1957.  The strongest plan-view frontogenesis was before and after the snowband reached peak intensity, not while the snowband was intense.  The time scale of frontogenesis in the environment was 19 hours and was well established approximately five and a half hours before the single snowband formed.

b.  CSI AND EPV

A cross section of RUC II θe (K), absolute geostrophic momentum (ms-1), and relative humidity greater than 80 % was taken normal to the thermal wind, from Duluth, Minnesota to Oklahoma City, Oklahoma.  Analyzes determines areas of conditional symmetric instability (CSI), weak symmetric stability (WSS), and convective instability (CI).  At 0300, WSS began to set up south of Des Moines, where there was a minimal amount of relative humidity greater than 80 %.  At 0400 the amount of moisture increased and there was CSI and WSS now in the vicinity of Des Moines. Therefore, CSI and WSS were in the environment five and a half hours before the single snowband developed.  The amount of relative humidity greater than 80 % slowly increased and CSI and WSS remained through 0900.

 By 1000, just after the single snowband formed, CSI and WSS were both present.   CI did develop, but was in the warmer air to the south closer to Oklahoma City.  Through 1600, saturation, CSI, WSS, and CI remained consistent and during this time the snowband reached its peak intensity (Fig. 12).  At 1700, there was CSI and CI, but little WSS; however, the WSS returned by 1800.  During this time, the CI moved farther north, but stayed just south of Des Moines.  Through 1900 to 2200, the area of CI continued to moved north closer to the CSI and WSS, and during this time, the snowband rotated counterclockwise, weakened, and broke into multiple snowbands that remained through 2300.  The release of CI most likely helped to sustain the multiple snowbands longer, rather than dissipating quickly.  By 2300, there was CI, less CSI, and no WSS present and the snowband had significantly weakened.

A plan view of 600 hPa saturated EPV with 750 hPa to 600 hPa mean layer relative humidity was examined to determine areas of reduced EPV that would indicate CSI (Fig. 13).  There is a close correlation between the location of negative EPV and dry air entrainment into the system as seen on water vapor imagery.  At 0700, a region of negative EPV and relative humidity greater than 80 % was located over Sioux Falls and Omaha; this is approximately 30 minutes prior to multiple snowbands forming at 0732.  Therefore, there was a very short time scale between the environment being conducive to reduced EPV and relative humidity and snowband formation. Of course this is only analysis of one level; analysis of a layer will provide better results.

The region of reduced EPV began to break up and at 0900; but there was an enclosed region of reduced EPV that remained over Omaha and Des Moines.   Recall there was one distinct snowband over Des Moines beginning about 922 and propagating northeast.  The enclosed region of negative EPV was present for approximately five hours and remained over Omaha and Des Moines through 1400.  At 1400, it was displaced to the north, closer to Sioux Falls and for several hours there was no reduced EPV over Des Moines.  Note that while the snowband was quasi-stationary and distinct, there was no reduced EPV and frontogenesis was weak.  Radar imagery after 1350 showed the snowband rotated counterclockwise and began to break up, until reduced EPV returned to the region in conjunction with a dry filament of air.  As the dry air approached the region the snowband reached its peak intensity.   The greatest intensity of the snowband from 1431 to 1822 corresponds to the nose of the dry filament impacting the region, see Fig. 14.

By 1900, a broad region of reduced EPV extended into western Iowa and covered southern Iowa by 2300, which is coincident with the dry filament covering the region.  Again the snowband began to break up at 1957, soon after the broad region of negative EPV returned to the environment and the dry filament completely covered Des Moines.  In this case, the orientation of the snowband and the main storm followed the orientation of the negative EPV, so it would be interesting to investigate if the orientation of the negative EPV guides the orientation of the snowband in future case studies.

To provide a more through investigation, EPV was investigated as a layer average from 600-400 hPa (Fig. 15).  The cross section in Fig. 9 shows that reduced EPV overlays the mid-level frontogenesis in approximately the 600-400 hPa layer.  Analysis of this layer is similar to just looking at the 600 hPa reduced EPV.  Reduced EPV, closely follows the orientation of the snowband.  During the time period from 1700 to 1900 the snowband lies within a tight region of saturated EPV.  This region represents the dry intrusion and the snowband follows the orientation of the saturated EPV contours.  Analysis of this layer shows how the snowband intensifies as the dry intrusion pushes into the Des Moines region.

6.  SUMMARY

The inclusion of radar data in the analysis helped determine the time scale and evolution of the mesoscale processes.  Fig. 16 is a graphical comparison of the evolution of the process at one point.  It shows that frontogenesis peaked before and after the snowband was intense and that EPV was close to zero from 1000 to 1400, while the snowband was quasi-stationary.   Frontogenesis, CSI, and WSS set up in the environment approximately five and a half hours before the single snowband formed, however; reduced EPV and relative humidity greater than 80 % set up in the environment approximately 30 minutes prior to snowband formation.  The intensification of the snowband was coincident with a dry filament of air intruding. Once the dry air was fully in place, over Des Moines, the snowband actually began to break up.  As seen in water vapor imagery, too much dry air cut off the moisture supply and the snowband weakened and dissipated. 

Fig. 17 is a composite diagram of the mesoscale processes at 1600 and shows the timing of the processes when the snowband intensified.  A vertical composite and plan view composite of mesoscale processes at the time of intensification were created to aid forecasters in identifying when a snowband will intensify, shown in Fig. 18 and Fig. 19.

7.  CONCLUSIONS

The evolution of mesoscale processes and the propagation characteristics of the snowband were correlated.  As frontogenesis increased the snowband responded by intensifying and rotating clockwise and as frontogenesis weakened, the snowband weakened and rotated counterclockwise.  Frontogenesis and the composite radar reflectivity overlay was useful in determining that the 750 hPa frontogenesis indicated the location and orientation of the snowband, while the 650 hPa frontogenesis indicated the location and orientation of the main precipitation shield.  Since only one case was analyzed it is speculative that the 750 hPa and 650 hPa pattern is observed in most cases, rather it may be a different level or even a layer that better defines the location and orientation of the snowband.  Intensification of the snowband coincided with the intrusion of dry air, but once the dry air was fully in place, the snowband weakened and dissipated.  This particular event had a distinct evolution and time scale of mesoscale processes.  Further understand the evolution and time scale of mesoscale processes in snowband events will only improve our understanding and situational awareness in the operational environment.

ACKNOWLEDGEMENTS:  Thank you Dr. Eichler for your valuable editing.  This article is dedicated to the late Dr. Moore, my thesis advisor, thanks for dedicating your life to seeing your students succeed.

 

REFERENCES

 

Banacos, P. C., 2003: Short-range prediction of banded precipitation associated with deformation and frontogenesis forcing. Preprints, Tenth Conf. on Mesoscale Processes, Amer. Meteor. Soc.,

Portland, OR, CDROM, P1.7.

 

Jones, Sarah C., and A. J. Thorpe, 1992: The three-dimensional nature of ‘symmetric’ instability.  Quarterly Journal of the Royal Met Soc, 118, 227-258.

 

Keyser, D., M. Reeder, and R. Reed, 1988: A generalization of Petterssen’s frontogenesis function and its relation to the forcing of vertical motion. Mon. Wea. Rev., 130, 477-506.

 

McCann, D., 1995: Three-dimensional computations of equivalent potential vorticity, Wea. Forecasting, 10, 798-802.

 

Moore, J., and T. Lambert, 1993: The use of equivalent potential vorticity to diagnose regions of conditional symmetric instability. Wea. Forecasting, 8, 301-308.

 

Moore, J. T., C. E. Graves, S. Ng, and J. L. Smith, 2005: A process-oriented methodology towards understanding the organization of an extensive mesoscale snowband: A diagnostic case study of 4-5

December 1999. Wea. Forecasting, 20, 35-50.

 

Nicosia, D. and R. Grumm, 1999: Mesoscale band formation in three major Northeastern United States snowstorms. Wea. Forecasting, 14, 346-368.

 

Novak, D. R., L. F. Bosart, D. Keyser, and J. S. Waldstreicher, 2004: An observational study of cold season-banded precipitation in Northeast U.S. cyclones. Wea. Forecasting, 19, 993-1010.

 

Novak, D. R., J. S. Waldstreicher, D. Keyser, and L. F. Bosart, 2006: A forecast strategy for anticipating cold season mesoscale band formation within Eastern U.S. cyclones.  ­Wea. Forecasting,

21, 3-23.

 

Shultz, D., and P. Schumacher, 1999: The use and misuse of CSI. Mon. Wea. Rev., 127, 2709-2932.

 

Weismueller, J. L., and S. M. Zubrick, 1998: Evaluation and application of conditional symmetric instability, equivalent potential vorticity, and frontogenetic forcing in an operational forecast environment.

Wea. Forecasting, 13, 84-101.