The COMET QPF Project

As part of the COMET cooperative we are examining cases involving rainfall amounts of over 2 inches within 24 hours over the Mid-Mississippi and Ohio River Valleys. Three National Weather Service Offices are involved: the St. Louis, MO; Louisville, KY; and Paducah, KY. Our specific goals over the next three years include the following:
 
• Complete a synoptic climatology of warm and cool season heavy precipitation events (including snowstorms) involving >= 2 inches of liquid precipitation in 24 hours over MO, IA, AR, IL, KY, IN, and OH. We will examine rain/snow events from 1990 up to the present. As part of this synoptic climatology we will organize the events according to the synoptic pattern (e.g., slow-moving N-S frontal system) and time of year. This work should parallel the work done by Maddox et al. (1979). However, we intend to be more definitive in terms of synoptic patterns at various pressure levels, sounding characteristics and the type of convection that is usually present. Other parameters will include CAPE (boundary layer and elevated CAPE), hodograph structure, and relationship to lower and upper level jet streaks.

 
• Evaluate current empirical QPF schemes that have been developed by NWS forecasters using either diagnostic or model data or both. Current schemes that are being investigated include those by: Al Johnson and Robert Ricks at Slidell, LA; Joe Rogash, currently at the SPC; Steve Amburn (SOO at Tulsa, OK); and John Henz (private meteorologist at Denver, CO). We are looking to verify the "science" in these schemes and see how they may help us in the quest to come up with a "number" in the QPF game.

 
• Diagnose key parameters and techniques which help to delineate the area of precipitation and the maximum value of the mean areal precipitation (MAP; not a point maximum) on a regional scale. We intend to construct, through many case study analyses, conceptual models for the short-term prediction of heavy rain/snow. This will involve both parameter evaluation and pattern recognition. We especially want to test numerical models (Eta, Meso-Eta, RUC, MASS), using PC-GRIDDS to show how often the forcing displayed in the models is more important to use than the precipitation fields. Since precipitation is still parameterized (stable vs. convective) it is subject to error. This is especially true with respect to the TIMING, AREAL coverage and the MAGNITUDE of the precipitation being forecast. Our goal, is to try to use the models to better forecast these aspects of the heavy rain threat.

 
• Develop a method for estimating the precipitation efficiency from sounding data, and possibly data from a numerical model. Precipitation efficiency is a very difficult parameter to get a handle on in the operational environment, yet it is critical for estimating the potential for heavy rain.

 

The following conceptual models have arisen from our past COMET research and are associated with heavy precipitation events. We encourage your comments and feedback (but not convective feedback!)
 
• Dynamics of upper-level jet (ULJ) entrance regions:

 
• Direct thermal circulation is typically found in the entrance region of the ULJ

 
• On the cold (cyclonic) side of the ULJ at low levels we expect northerly ageostrophic flow---this reinforces the low-level cold air and enhances frontogenesis

 
• On the warm (anticyclonic) side of the ULJ, well to the south, a southerly low-level jet (LLJ) often forms in response to the divergence in this quadrant of the ULJ

 
• Homan and Uccellini (1987) note that in ``a stratified baroclinic environment (which includes sloped isentropic surfaces), the response to divergence (PVA) in the middle and upper troposphere is not a localized ascent maximum concentrated directly beneath the upper-level divergence field. To a first approximation based on adiabatic constraints, the response must be generally confined to the flow along the sloped isentropic surfaces toward the divergent region.''

 
• They further note that ``the initial slope of the isentropic surfaces in the lower to middle troposphere and the evolution of the wind fields on these surfaces are important factors in specifying the vertical motion field that develops in response to divergence in the middle to upper troposphere, which, in turn, could significantly influence the distribution and intensity of precipitation.''

 
• In this sense the warm air advection pattern that forms in the lower to middle troposphere can be viewed as a manifestation of the response to divergence in the right entrance region of the ULJ.

 
• In this region of the ULJ the vertical wind shear is strong, the absolute vorticity is small, and there is strong positive theta-e advection. Elevated convective instability and/or CSI can easily form in this environment as momentum surfaces flatten and theta-e surfaces become more vertical or even fold.

 


 
 

• Conceptual model for a back-building upper-level jet streak


In many heavy precipitation events (snow and rain), we notice the development of a retrograding (or backbuilding) jet streak. Consider a deep trough that has developed with a jet streak downstream and another upstream. At 0 hr, the downstream jet streak appears to be of no concern as it is propagating away from the region of concern. The approaching upstream jet streak is of concern but is too far away to provide significant dynamics for the main surface system (not shown) located downstream. At t + 12 hours, the downstream jet streak has strengthened and retrogressed with respect to the mean flow, which is basically west-to-east. The main contributing factor to this phenomena is the strong cold air to the north which propagates southward in the developing trough and strengthens the horizontal temperature gradient. In the warm season, latent heat from a developing MCS in the right entrance region of the downstream jet streak will act to strengthen the temperature gradient which in turn will strengthen the jet streak.

A "backbuilding" jet has been documented on two different occasions in which heavy snow fell. On Jan 17, 1994, Louisville, KY received a record 20+ inches of snow as a surface low moved northeast out of the Southern Plains. On Jan 18-19, 1995, Columbia, MO and Springfield, MO received 20 and 18 inches of snow, respectively. In both cases, a jet streak downstream strengthened and retrogressed in the wake of a strong upper-level trough and resulting push of cold air.
 
 


 

More recent efforts focus on the influence of diabatic heating in generating a back-building or quasi-stationary jet streak. The 16-17 January 1994 snow event in Kentucky appears to be a case where feedback between latent heating and jet streak evolution provided a long-lived synoptic environment for the production of heavy snowfall. This concept will be tested using diagnostic analyses as well as simulations from the MASS model. In the following diagram, a conceptual model is provided of a back-building jet streak due to the influence of diabatic heating. In the left column, thin solid lines denote geopotential height, thick dashed lines are isotachs (both at, say, 300 mb), while the grey stippled area represents the region of latent heat release. The thick black arrows represent the low level jet stream, while the line A'-A'' denotes the cross section line used in the right column. On the right, the thin solid lines represent isentropes, while the dashed arrows denote the vertical and horizontal ageostrophic components of the direct thermal circulation.
 


 
 

• Conceptual model for the development of CSI/elevated convection


Many heavy precipitation events (rain and snow) are dominated by mesoscale instabilities embedded within larger synoptic scale systems. Often a series of mesoscale instabilities that consists of elevated convective instability, conditional symmetric instability, and frontogenesis in the presence of weak symmetric stability will develop from south to north respectively. These instabilities tend to form under conditions of relatively strong unidirectional wind shear (from the south or southwest), a cold stable boundary layer, and nearly saturated low-to-middle layers. The important aspects about these instabilities is that all or only one of the instabilities can contribute to narrow precipitation bands. Each instability has a length of approximately 200 km, a depth of 3-4 km, and a time scale of 3-6 hours.

Two examples of the existence of this instability spectrum were witnessed on February 15-16 1993 over Missouri and March 8-9 1994 over southern Missouri/northern Arkansas. On Feb 15-16 1993, over 18 inches of snow fell in southern Missouri, while on 8-9 March 1994, over 20 inches of snow fell along the spine of the Ozark Mountains.
 
 


 
 

• The Glass et al. (1995) conceptual model for heavy precipitation with an E-W boundary


``An idealized model is illustrated...to help identify/refine areas favorable for heavy convective rainfall. The shaded region in the figure represents the highest threat region. In this configuration, moisture convergence, and warm thermal advection is maximized along and north of a surface boundary where the low level jet (LLJ) and positive 850 mb theta-e advection are coupled. Cells repeatedly develop in this region, organize into meso-beta scale elements (MBEs), then move downwind with the 850-300 mb cloud-layer shear. Providing the low-level boundary remains quasi-stationary (a must for long-lived events) and the airmass is not significantly capped, this pattern of new cell growth will normally remain favored until the LLJ veers, altering the moisture convergence/advection pattern. Observations from individual cases indicate that when the LLJ/850 mb theta-e advection `couplet' remains stationary or veers little with time, backward propagation and regenerative MCSs are favored. Alternatively, when the LLJ/850 mb theta-e advection `couplet' veer clockwise with time, forward propagating and/or regenerative MCSs are more likely. Both patterns result in echo training with the axis of heavy rainfall typically extending downstream from the `couplet', parallel to the 850-300 mb thickness. Factors which determine the downstream extent of the heavy rain include the breadth and magnitude of the LLJ, magnitude of the 850-300 mb cloud-layer shear (strength of the 850-300 mb thickness gradient), and moisture/instability distribution of the environment. This conceptual model can be applied with either real-time data or using conventional and derived products from gridded model data. ''
 

-Glass, et al. (1995) see reference below

 


 
 

• Related work involves the choice of parcels used in stability and bulk condensate production computations. Dramatic differences in the perceived stability profile can be found depending upon the level from which the parcel is lifted.

 

 
 
 
 
 
 
 
 
 
 
 
 
 



• Synoptic Climatology of MO-IL for Significant Rainfall

 

 
 

Five Categories Based on 80 Events from 1990-1996, Inclusive
 

• Warm-Sector Convective Rainfall Events:

 

 

TYPE 1: NE-SW oriented slow moving cold front with heavy rain on the warm side (Maddox, synoptic type):
 



TYPE 3: Heavy rainfall within the warm sector of an extratropical cyclone:
 


 

• Cold-Sector Convective Rainfall Events (Elevated Thunderstorms):


TYPE 4: E-W oriented warm/stationary front with heavy rain on the cold side (Maddox, frontal type):
 



TYPE 5: Heavy rainfall on the cold side of a mesoscale outflow boundary (Maddox, mesohigh type):
 


 

• Cold-Sector Stable Precipitation Events:

 

 

TYPE 2: NE-SW slow-moving cold front with stable precipitation on the cold side:
 


 

HISTOGRAM OF EVENT TYPES FROM 1990-1996, INCLUSIVE:


 

• Characteristics of MCS Propagation Types

 
• Forward Progagating MCSs:

 

 

Maximum CAPE values downstream or coincident with MCS

850 mb theta-e ridge downstream or coincident with MCS

Moderate-strong 850-300 mb mean winds

Low-Level Jet (LLJ) coincident with MCS location

Progressive short-wave trough translating eastward

300 mb upper-level jet (ULJ) oriented E-W and positioned north of MCS

Strongest moisture transport and low-level moisture convergence located along or downstream from MCS
 

• Backward Propagating/Quasi-Stationary MCSs:

 

 

Maximum CAPE values along and upstream from MCS (typically to W-SW)

850 mb theta-e ridge along and upstream from MCS (typically to W-SW)

Quasi-stationary E-W surface boundary present (old front, outflow boundary, etc.)

Weak 850-300 mb mean winds

LLJ upstream from MCS location

Diffluent thickness (850-300 mb) pattern aloft

Typically upper-level ridge aloft

Veering winds with height dominate speed shear

Strongest moisture transport and low-level convergence located upstream from MCS
 
 

• Precipitation Efficiency Factors


Depth of the warm cloud with temperature > 0 C: enhances collision-coalescence process by increasing the residence time of droplets in cloud

Mid-upper level vertical wind shear: weak-moderate shear (< 3 X 10**-3 s-1) yields slower storm movement and decreased entrainment

Cloud scale vertical motion: is a function of CAPE; related to condensate production and residence time of droplets in cloud

Storm-relative mean inflow vector and mixing ratio in the surface-LCL layer

Vertical gradient of the saturation mixing ratio: related to condensate production in the warm cloud (highly correlated to the warm cloud depth

Mean environmental relative humidity: higher RH (>65%) results in less dry air entrainment into the cloud mass; enhanced by ITCZ "connection" or plume

Wide spectrum of cloud droplet sizes

Cloud base height: lower height means less sub-cloud evaporation
 
 

• Conclusions

 
• QPF Studies: What we have learned so far...

 

 

Low-level jet (LLJ) is critical for initiation and sustenance of training convection

Entrance regions of upper-level jet (ULJ) streaks often tied to the development of the LLJ and MCSs during the warm season

Exit regions of ULJs are coupled to LLJs and stable precipitation during the cold season

Surface boundaries (frontal or outflow) are very important to diagnose as they focus and force convection

Precipitation efficiency is enhanced by deep, warm cloud depths and high environmental RH and PWs

Max theta-e CAPE is a good parameter to evaluate for diagnosing elevated convection

Elevated convection occurs a great deal in the Midwest and is often confused with CSI for creating heavy bands of rain or snow

The vertical wind shear associated with heavy convective rainfall events typically displays strong veer and shear from the surface-850 mb and weak veer and shear above

Backward propagation is favored by low-level moisture convergence, instability and moisture located upstream from the convection; animation of satellite and WSR-88D imagery help to diagnose backbuilding storms

          Eta-32 Evaluation of Heavy Rain Events by Scott Watson

          Significant Rainfall Climatology for Kentucky/Southern Indiana by Stephen Klaus

          Strategies for Estimating the Areal Coverage and Magnitude of a Precipitation Event by James Moore
 

RECENT PAPERS & PUBLICATIONS
 

• Glass, F. H., 1998: A Climotology of Widespread Heavy Rainfall Events Across Missouri.  Preprints, 19th Conference on Severe Local Storms, Minneapolis, MN, American Meteorological Society, 540-543.

• Considine, S. R., F. H. Glass, and J. T. Moore, 1998: A Synoptic Climatology of Significant Rainfall Events Over the Mid-Mississippi River Valley. Preprints, 19th Conference on Severe Local Storms, Minneapolis, MN, American Meteorological Society, 544-547.

• Moore, J. T., A. C. Czarnetzki, and P. S. Market, 1998: Heavy Precipitation Associated with Elevated Thunderstorms Formed in a Convectively Unstable Layer Aloft.  Meteorological Applications, 5, 373-384.

• Market, P. S., and J. T. Moore, 1998:  Mesoscale Evolution of a Continental  Occluded Cyclone.  Mon. Wea. Rev., 126, 1793-1811.

• Moore, J.T., S.R. Considine, S.M. Rochette, F.H. Glass, and T.M. Tworek, 1998: A Comparison of Cool Season and Warm Season Heavy Rain Events in the mid-Mississippi River       Valley. Preprints, 16th Conf. on Weather Analysis and Forecasting, Phoenix, AZ, Amer. Meteor. Soc., 267-269.

 
• Market, P.S., J.T. Moore, S.R. Considine, T. Funk, and S.M. Rochette, 1998: A Synoptic Comparison of Two Heavy Precipitation Events in the Ohio River Valley. Preprints, 16th Conf. on Weather Analysis and Forecasting, Phoenix, AZ, Amer. Meteor. Soc., 358-361.

 
• Rochette, S.M., and J.T. Moore, 1996: Initiation of an Elevated Mesoscale Convective System Associated with Heavy Rainfall. Wea. Forecasting, 11, 443-457.

 
• Moore, J.T., S.M. Rochette, F.H. Glass, D.L. Ferry, and P.S. Market, 1996: Elevated Thunderstorms Associated with Heavy Rainfall in the Midwest. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 772-776.

 
• Funk, T. and J.T. Moore, 1995: Vertical Motion Forcing Mechanisms Responsible for the Production of a Mesoscale Very Heavy Snow Band across Northern Kentucky. Postprints, 4th National Winter Weather Workshop, Kansas City, MO, NOAA Tech. Memo. NWS CR-112, 19:1-11.

 
• Rochette, S.M., J.T. Moore, P.S. Market, F.H. Glass, and D.L. Ferry, 1995: Heavy Rain Events Associated with Elevated Thunderstorms in the Midwest. Postprints, 4th National Winter Weather Workshop, Kansas City, MO, NOAA Tech. Memo. NWS CR-112, 25:1-12.

 
• Glass, F.H., D.L. Ferry, J.T. Moore, and S.M. Nolan, 1995: Characteristics of Heavy Convective Rainfall Events across the Mid-Mississippi Valley during the Warm Season: Meteorological Conditions and a Conceptual Model. Preprints, 14th Conf. on Weather Analysis and Forecasting, Dallas, TX, Amer. Meteor. Soc., 34-41.

 
• Moore, J.T., S.M. Nolan, F.H. Glass, D.L. Ferry, and S.M. Rochette, 1995: Flash Flood Producing High-Precipitation Supercells in Missouri. Preprints, 14th Conf. on Weather Analysis and Forecasting, Dallas, TX, Amer. Meteor. Soc., (J4)7-(J4)12.

 
• Moore, J.T., C.H. Pappas, and F.H. Glass, 1993: Propagation Characteristics of Mesoscale Convective Systems. Preprints, 17th Conf. on Severe Local Storms, St. Louis, MO, Amer. Meteor. Soc., 538-542.

 
• Moore, J.T., and F.H. Glass, 1992: Mesoscale Convective Systems: Initiation and Propagation. Preprints, 4th Atmospheric Environment Service, Canadian Met. and Ocean Soc. Workshop on Operation Meteorology, Whistler, Canada, 215-224.

• At SLU
• Dr. James T. Moore, Principal Investigator
• Sean Nolan, PhD Student
• Scott Watson, MS Student
• Steve Klaus, MS Student
• John Gagan, MS Student
• Marc Singer, MS Student
• Heather Sumner, MS Student
• At NWS
• Ted Funk, Science and Operations Officer at NWSFO Louisville, KY
• Rob Cox, Forecaster, NWSFO Louisville, KY
• Larry Datillo, Hydrometeorological Technician, NWSFO Louisville, KY
• Pat Spoden, Science and Operations Officer at NWSO Paducah, KY
• Michael York, Forecaster, NWSO Paducah, KY
• Fred Glass, Lead Forecaster, NWSFO St. Louis, MO.